1-:^ nj §^=: □ a 2 r^ a m a □ PLANT SCIENCE MONOGRAPHS Edited by Professor Nicholas Polunin THE MICROBIOLOGY OF THE ATMOSPHERE PLANT SCIENCE MONOGRAPHS Uniform with this volume will be: * Biology of Mycorrhiza Encyclopaedia of Weeds and Their Control Grassland Improvement . Mangroves of the World Mutations and Crop Improvement *Plant Growth Substances Plant Life and Nitrogen . *Salt Marshes and Salt Deserts of the Wor *Seed Preservation and Longevity Sex in the Lower Ormnisms Id J. L. Harley L. J. King A. T. Semple V. J. Chapman A. Gustafsson J. L. Audus G. Bond V. J. Chapman L. V. Barton H. P. Papazian FURTHER TITLES ARE UNDER CONSIDERATION The feeding and clothing of the world's teeming millions can continue to keep abreast of population increases through the help of effective application of research in the plant sciences. The publication of this research, by which means a scientist or technologist makes his findings known to workers elsewhere, tends to be scattered in literally hundreds of botanical and agricultural journals emanating from most of the countries of the world. Often it appears in such polyglot arrays of fragments that it is extremely difficult to bring together even in some narrow 'line' of endeavour. Consequently advances are slowed and interests unnecessarily divided, scientific and human progress being thereby retarded. The present series of 'monographs' is designed to remedy these deficiencies in especially important or attractive specialities, by publishing individual book- length accounts of the entire background and current progress in their fields. Such detailed surveys, being fully documented and plentifully illustrated, should prove of real value to the world at large in constituting the bases for further advances on the ever-expanding horizons of scientific research, and so lead to improved productivity and, ultimately, standards of living. They are prepared by specialists usually of international reputation for their work in the field chosen, and often culminate a lifetime of active investigation. Being as up-to-date as possible, they will often embody significant advances not pre- viously published. *Already published and available FOR DETAILS OF A COMPANION SERIES SEE THE END OF THIS BOOK Plate i. — Representative fields from Hirst-trap slides illustrating diverse air sporas. (Magnification: all x looo. At this magnification i sq. cm. of field corresponds to 2-i cc. of air.) Frontispiece (a) Fine-weather air spora, Ascot, Berks., i2-oo hr., 27 June 1958. Showing grass and nettle pollen, Cladosporiutn, Epicoccutn, etc. (h) Damp-air spora. Ascot, Berks., 02-00 hr., 17 June 1958. Showing spores ofGanodernia, Sporobolomyces, Tilletiopsis, and other hyaline basidiospores. (c) Rain-type air-spora, Ascot, Berks., 04-30 hr., 13 June 1958. Showing various ascospore types, Sporobolomyces, soot, etc. {d) Air-spora in a wheat crop at night during a thunder-storm, Harpenden, Herts., 21-30 hr., 10 August 1959. Showing Puccinia graminis uredospores, Cladosporium, and unidentified ascospore. {e) Spores of MertiUus lacrymans from air of building with timber affected by dry-rot. From camera-lucida paintings by Maureen E. Bunce. ^k & ^ 7'i .0 7 i.^- v.- i a I # ')ii 1 # ^w • vlT* 7 (>X-^^^^ V- /^->. €» Plate i. — For details sec preceding page. PLANT SCIENCE MONOGRAPHS edited by Professor Nicholas Polunin THE MICROBIOLOGY OF THE ATMOSPHERE By P. H. GREGORY Ph.D., D.Sc. (London), D.I.C. Head of Plant Pathology Department, Rothamsted Experimental Station, Harpenden, England; formerly Professor of Botany, Imperial College of Science and Technology, University of London 1961 LONDON LEONARD HILL [BOOKS] LIMITED INTERSCIENCE PUBLISHERS, INC. NEW YORK LONDON Leonard Hill [Books] Ltd. 9 Eden Street, N.W.i NEW YORK Interscience Publishers, Inc. 250 Fifth Avenue, New York i FIRST PUBLISHED IN GREAT BRITAIN I961 © P. H. Gregory, 1961 DEDICATED to my wife Margaret Fearn Gregory PRINTED IN GREAT BRITAIN AT THE UNIVERSITY PRESS ABERDEEN PREFACE Aerobiology is usually understood to be the study of passively airborne micro-organisms — of their identity, behaviour, movements, and survival. One characteristic, which it shares with many other population studies in biology, is that the ultimate relevant unit consists of the individual cell or small group of cells. Analysis at the molecular or sub-atomic level is irrelevant to our present purpose. Like geography, aerobiology is an agglutinative study, drawing information from many kinds of scientific research. Although it already has its patron saint, Pierre Miquel, and its martyr, Fred C. Meier, aerobiology is best regarded as an activity whose material will in due course be incorporated into the main body of bio- logical science — without, I hope, any necessity for splinter societies, journals, and international conferences. This book amplifies and extends a course of Intercollegiate Lectures given to botanical students in the University of London in 1956. The theme, which has occupied me for over fifteen years, is as follows. Trans- port through the atmosphere is the main dispersal route for such organic particles as the spores of many micro-organisms. How do the properties of the atmosphere, and the properties of these particles themselves, affect their dispersal? How do the particles get into the air? How far, and in what numbers, are they dispersed? By what processes do they become grounded, so that they can continue growth ? What is in the air, and how can we measure it ? What are the practical consequences of this process for the micro-organisms themselves, and for man, other animals, vegetation, and crops ? Although there are one or two other books on airborne microbes, this is the first to treat the subject as a world-wide phenomenon. It is, perhaps, inevitable that it should be attempted by a mycologist. Few other biologists find their material so dominated by the atmosphere, and no other micro- organisms have so thoroughly exploited the possibilities of aerial dispersal as the fungi. One of the fascinations of the subject is the impact of facets of its knowledge on such apparently diverse topics as artificial rain-making, allergy, smoke screens, effluent of nuclear power-stations, crop protection, icing of aircraft, air hygiene, and many other topics. This book treats of the development and principles of aerobiology rather than applications; yet the stimulus to nearly all aerobiological work comes from applied science. In this book the term 'microbe' is used freely when a general word is wanted; but, like the word 'spore', it has admittedly been stretched beyond its normal meaning. Airborne pollen of flowering plants must be THE MICROBIOLOGY OF THE ATMOSPHERE Included and is safely covered by the term 'spore' (botanically : 'micro- spore') ; but are pollen grains and mushroom spores microbes ? There is no other commonly accepted word that covers quite what is meant by the word 'spore' as used here: 'propagule', 'disseminule', 'biota', 'diaspore'? We have isolated part of the continuum for study but find we are not well- equipped verbally for the task of dealing with it. The microbial population of the atmosphere is referred to here as the 'air-spora', using 'spora' as a word analagous to 'flora' and 'fauna'. Botanical nomenclature has presented some difficulties: authorities have not been given for specific names, and the names used by other authors have usually been quoted as given in the original papers — without necessarily attempting to guess what was meant, or following the nomen- clature fashionable in i960. I have converted other workers' numerical data to the metric system, and temperature to the Centigrade scale, to aid comparison, and have moreover assessed spore concentrations on the uniform basis of number per cubic metre. Frequently, in making general statements, I have omitted safeguarding, but tedious, escape clauses: this has been done to spare the reader who will understand that biological generalizations aboun-d in exceptions and complexities. Interpretations in this book are mostly my own responsibility, but I am grateful for help received from many people during its preparation. In particular I offer my thanks to the following : G. Samuel and W. Buddin for introducing me to dispersal problems in the field; F. C. Bawden for encouragement in the study of aerobiology and for reading this book in manuscript; E. C. Large for advice on planning the book; D. A. Boalch (and many other librarians) for continual help with the literature; members of the British Mycological Society for named specimens of fungi, and H. L. K. Whitehouse for mosses; A. Home, V. Stansfield, and F. D. Cowland for photography; R. Adams, G. C. Ainsworth, J. R. D. Francis, E. J. Guthrie, Elizabeth D. Hamilton, J. M. Hirst, C. T. Ingold, C. G. Johnson, F. T. Last, Kate Maunsell, T. Sreeramulu, and O. J. Stedman for discussion and help with aerobiological problems and applications; Audrey Baker, Beatrice E. Allard, and Marie T. Seabrook for clerical assistance; and Maureen E. Bunce for experimental help, revision of the manuscript, and preparation of many of the illustrations — especially the paintings for Plates i, 5, 6, and 7. I also wish to thank authors, editors, and publishers for permission to copy illustrations which are acknowledged in the text. Finally, fi)r the calculations involved in Figs. 24 to 27, and for those chapters needing the help of a mathema- tician, I have been fortunate in having the constant advice and willing help of my wife, Margaret F. Gregory, to whom I am most deeply grateful. Rothamsted Experimental Station Philip H. Gregory Harpenden^ Herts., England. September, i960. vi 5 W*^ ' VJO \.rf' .e. ^^^s^: CONTENTS Preface List of Figures in Text List of Plates List of Tables CHAPTER I Historical Introduction .... Speculations on the origin of disease . Early microscopists and the discovery of spores Controversy on spontaneous generation The germ theory of disease The hygienists and their investigation of the air The allergists ...... II Sedimentation in Still Air Factors determining velocit}' of fall Effects of sedimentation .... III The Atmosphere as an Environment . The atmosphere and its layers The troposphere ..... Laminar boundary layer Local eddy layer ..... Turbulent boundary layer Transitional or outer frictional turbulence layer Convective layer ..... Night radiation and temperature inversion Role of convection .... The stratosphere ..... Circulation of the atmosphere Air masses ...... IV Spore Liberation Problems of 'take-off' .... Take-off mechanisms in cryptogams etc. Viruses ...... Bacteria ...... Actinomycetes ..... Myxomycetes ..... Fungi ....... Lichens ...... Algae ....... vii PAGE V xi xiii xiv I I I 3 6 7 II 14 14 19 22 22 24 24 25 25 27 27 27 28 29 29 29 31 31 32 33 33 33 33 34 38 38 THE MICROBIOLOGY OF THE ATMOSPHERE PAGE Bryophytes . Pteridophytes Pollination of phanerogams Gymnosperms Angiosperms V Horizontal Diffusion Dispersion of the spore-cloud Diffusion as a result of atmospheric turbulence Field experiments on diffusion of spore-clouds Comparison of theories of W. Schmidt and Sutton VI Deposition Processes Mechanism of impaction ..... Wind-tunnel study of impaction Impaction on cylinders ..... Impaction on a rotating sticky cylinder Impaction on plane surfaces at various angles to wind Deposition on horizontal slides Deposition of Lycopodium spores on inclined plane surfaces Mean deposit on inclined slides Deposition of other spores on inclined plane surfaces Effect of thickness of slide .... Orientation of spores ..... Deposition on 9 cm. diameter Petri dish Retention and blow-off from clean surfaces . Deposition and retention on potato and bean leaves VII Natural Deposition .... Measurement of deposition Measurement of deposition coefficient, 'p Loss by deposition from spore-cloud . Deposition mechanisms outdoors Boundary-layer exchange Sedimentation Impaction Turbulent deposition Electrostatic deposition . Minor deposition mechanisms Rain-washing ('scrubbing', 'rain-out', 'wash-out' Relative importance of deposition mechanisms VIII Air Sampling Technique .... Gravity sedimentation methods . Sedimentation from still air . Sedimentation from wind Sedimentation from artificially moving air Inertial methods ..... Impaction using wind movement viii CONTENTS Forced air-flow impactors Adhesives Thermal precipitation Electrostatic precipitation . Comparison of methods IX The Air-Spora near the Earth's Surface Composition of the air-spora Taxonomic groups needing study in the air-spora Miquel's work on bacteria and moulds Effect of rain .... Diurnal periodicity Relative numbers of bacteria and moulds Recent study of fungi and pollen The air-spora at 2 metres above ground-level The air-spora at other heights near the ground Diurnal periodicit}' of fine-weather spora Seasonal changes Effect of localit}' Effect of weather Biotic factors Marine air The air of polar regions The origin of the air-spora X The Upper-air Spora Vertical diffusion Ground stations at different altitudes above sea-level The role of turbulence . Early studies of the upper air Sampling from balloons . The stratosphere . Sampling from aeroplanes Spores of green plants in the lower troposphere McGill Universit}" studies Flights over the Arctic Microbiology of air masses over northern Canada Air masses over Montreal Air masses over the North Atlantic Ocean \'ertical gradient over the ocean Summary ..... XI Deposition in Rain, Snow, and Hail Rain ..... Snow ..... Hail XII The Air-Spora of Enclosed Spaces Die-away of concentration . Spore movement in convection currents ix THE MICROBIOLOGY OF THE ATMOSPHERE XIII XIV XV page Intra-mural sources ....... i5« The air of different environments .... 159 Dwelling houses ....... 159 Hospitals ........ 159 Factories and workshops, schools, public buildings 160 Subways, mines, and caves ..... 160 Sewers ........ 160 Air-spora of farm buildings ..... 160 Air-spora of glasshouses ..... 161 Ships ......... 161 Deposition Gradients and Isolation .... 162 Factors complicating infection gradients 162 Empirical methods ....... 166 Diffusion and deposition theories .... 167 W. Schmidt's theory ...... 167 Development of Sutton's theory .... 168 Re-calculation of the deposition gradient 170 Calculation of Qx . 171 Application of gradients ...... 176 Characteristics of gradients ..... 178 Topographical modification of gradients 179 Gene dispersion ....... 179 LONG-DlSTANCE DISPERSAL 181 Controversy on the importance of the air-spora . 181 Theoretical discussion ...... 182 Observations ........ 184 Re-colonization of Krakatoa ..... 184 Quantitative studies ...... 185 Viability ......... 190 Physiological studies of viability .... 191 Aerobiology 194 The phenomena ....... 194 Implications of aerobiology ..... 198 Biological warfare ....... 198 Isolation, quarantine, and geographical distribution 198 Medical mycology and allergy .... 200 Palynology ........ 201 Evolution ........ 201 Beyond the atmosphere ...... 203 Future study of our atmosphere .... 204 Appendix I. Visual Identification (Plates 5, 6 and 7) . 207 Appendix II. Conversion factors 215 Bibliography 217 Subject Index 237 Author Index 246 FIGURES IN TEXT PAGE 1. Pasteur's gun-cotton filter for airborne microbes ... 5 2. Cunningham's aeroconiscope ....... 8 3. Diagrammatic representation of layers of the atmosphere with a logarithmic vertical scale ....... 23 4. Splash from impact of water drop (5 mm. diameter) falling with velocity of 440 cm. per sec. on a thin film of water ... 36 5. Anthesis of false oat-grass {Arrhenatherum elatius): (i) closed anther; (2) open anther; (3) spikelets on a calm day; and (4) spikelets in a wind ........ 41 6. Diffusion of spore-cloud during horizontal travel in wind . . 48 7. Test of agreement of W. Schmidt's and Sutton's diffusion theories with experiments using Lycopodium spores liberated over grass field at Imperial College Field Station, Ascot, Berks.: {a) graph of log CT against log x; (/;) graph of log o against log t . . -57 8. Diagram showing relation between concentration x = no. of spores per unit volume; wind-speed = u. ; area dose, A.D. = no. of spores passing through frame of unit area; and trap dose, T.D. = no. of spores deposited on unit area of surface . . 58 9. Streamlines of air and particle trajectories around a cylindrical obstruction ......... 59 10. Observed relation between E per cent and k = Vsu/idg. Solid lines from Gregory & Stedman (1953). Broken lines = values predicted by Langmuir & Blodgett (1949) for spheres, strips, and cylinders for ^ = 10^ . . . . . . .60 11. Diagram showing small wind-tunnel used in deposition study at Rothamsted Experimental Station, elevation view . . .61 12. Diagram illustrating gravity theory of particle deposition . . 65 13. Efficiency of deposition oi Lycopodium spores on zones across glass microscope slide at presentation angles from 0° to 90° as observed in wind-tunnel experiments ....... 66-67 14. Mean efficiency of deposition o^ Lycopodium spores on glass micro- scope slide (all zones) at presentation angles of 0° to 180° . . 70 15. Number of raindrops of various diameters passing per second through I sq. metre (horizontal), with rainfall of intensities varying from 05 to 25 mm. per hr. . . . . . . .86 16. Collection efficiencies of spores of diameters 4 to 40 /x by raindrops of diameters 003 to 50 mm. ...... 87 17. Diagram of the Hirst automatic volumetric suction-trap: (i) eleva- tion facing wind; (2) plan of section through orifice; (3) elevation in side view ......... 100 18. Diagram of six-stage Andersen sampler . . . . .102 xi THE MICROBIOLOGY OF THE ATMOSPHERE PAGE 19. Diurnal periodicity of total numbers of bacteria in air at the Observa- toire Montsouris, Paris, based on hourly readings between March 1882 and September 1884 . . . . . . .112 20. Mean diurnal periodicity curves of thirteen spore-groups expressed as percentage of the peak geometric mean concentration. From Hirst trap records at Rothamsted Experimental Station, summer 1952 .......... 118 21. Diurnal periodicity o( Cladosporium at London (South Kensington) and Harpenden (Rothamsted), based on Hirst trap catches from May to September, 1954 . . . . . . .121 22. Exponential form of the die-away of bacteria-carr}ang particles from the air of a room. Line A : In an observation military canteen after the occupants had left suddenly. Line B : Observations on the die- away following a group of sneezes in a small room . . -156 23. Diagram showing changes of circulation in a room according to relative temperature of walls and of inside air . . -157 24. Fraction of spore-cloud remaining airborne (allowing for loss of spores from spore-cloud bv deposition to ground), expressed as ax/Qo . . . " . . . . . . .171 25 to 27. Dilution of spore-cloud by eddy diffusion. For m = 175 (also m = 1-24 and 2-0), expressed as logarithms of d, dw, diw, D (and also for daw with width of 100 metres) 25. Deposition coefficient, p = 005 ..... 173 26. Deposition coefficient, p = 001 ..... 174 27. Deposition coefficient, p = o-ooi ..... 175 28. Infection gradients of potato late-blight {Phytophthora infestans) observed by Waggoner (1952) at Clear Lake, Iowa, compared with theoretical line for deposition downwind from a point source 177 Xll LIST OF PLATES 1. Representative fields from Hirst-trap slides illustrating diverse air sporas. (Magnification: all x looo. At this magnification i sq. cm. of field corresponds to 2-i cc. of air.) (a) Fine-weather air spora, Ascot, Berks., i2-oo hr., 27 June 1958. Showing grass and nettle pollen, Cladosporiinn, Epicoccum, etc. (/;) Damp-air spora, Ascot, Berks., 02-00 hr., 17 June 1958. Showing spores of Ganoderma, Sporoholomyces, Tilletiopsis, and other hyaline basidiospores. (c) Rain-t}'pe air-spora. Ascot, Berks., 04-30 hr., 13 June 1958. Showing various ascospore t}'pes, Sporobolomyces, soot, etc. (d) Air-spora in a wheat crop at night during a thunder-storm, Harpenden, Herts., 21-30 hr., 10 August 1959. Showing Puccinia graminis uredospores, Cladosporinni, and unidentified ascospore. (e) Spores of Meniliiis lacrjmans from air of building with timber affected by dry-rot. From camera lucida paintings by Maureen E. Bunce ......... Frontispiece facing page 2. Ehrenberg's illustration of sample of dust collected by Charles Darwin on the Beagle near the Cape Verde Islands, January 1833 . 12 3. Photographs by Worthington & Cole (1897) showing splash of a water drop weighing 0-2 gm. (coated with lamp-black) falling 40 cm. into a mixture of milk and water. ....... 36 4. Rosebury-Henderson Capillary Impinger in operation, with air entering first through a May & Druett Pre-impinger. (Inset shows the capillary in still conditions) ....... 96 Appendix I. Typical components of the air-spora at a uniform magnifica- tion. Mag. 1000 X . From camera lucida paintings by Maureen E. Bunce ........... 207 Plate 5. Phycomycetes, Ascomycetes, Fungi Imperfecti, Actinomy- cetes Plate 6. Basidiomycetes, Lichens, Bryophytes, Pteridophytes, M\tco- mycetes Plate 7. Pollens, and miscellaneous other objects Xlll LIST OF TABLES PAGE I Observed terminal velocities of pollens and spores . .16-17 II Pollen distribution at different altitudes .... 20 III Size of 5f/z<^7 pollen at different altitudes .... 21 IV Typical characteristics of anemophilous and entomophilous plants ......... 40 V Results of dispersal of spores of r/7/^?/fl frtr/V5 . . ■ 5^ VI Dispersal of mixed spores of Tilletia caries and Bovista plumhea 53 VII Calculation of parameters of Sutton's diffusion equation from Stepanov's data ........ 54 VIII Observed values of parameters in Sutton's diffusion equation from experiments on spore dispersal . . . -55 IX Efficiency of deposition on inclined slides in turbulent wind- tunnel . . . . . . . . -71 X Efficiency of deposition of Lycopodium spores on upper and lower surfaces of potato and broad-bean leaflets in turbulent wind-tunnel . . . . . . . -75 XI Deposition on grass of Lycopodium spores activated with iodine- 13 1 ......... 78 XII Deposition of spores on ground ..... 79 XIII Observed values of p and Vg for Lycopodium spores on hori- zontal microscope slide in wind-tunnel .... 80 XIV Percentage of total spores liberated near ground-level that were estimated to have been deposited on ground in open-air tests 80 XV Number of Lycopodium spores deposited on upper and lower surfaces of horizontal traps ...... 84 XVI Average density of spore deposit from three types of trap 2 m. above ground ........ loi XVII Estimated detection thresholds of concentration of Hirst trap and sticky microscope slides inclined at 45° . . .106 XVIII Means of monthly mean numbers of bacteria and moulds per cubic metre of outdoor air in Paris . . . .111 XIX Total number of pollen grains and spores per cubic metre in oak-birch wood . . . . . . . • 115 XX Diurnal periodicity in the air-spora on land . . 1 16-17 XXI Temperatures at which highest concentrations were recorded 120 XXII Numbers of microbes and distance from land . . . 126 XXIII Analysis of Pady and Kelly's (1954) data on two return flights over the North Atlantic . . . . -145 XXIV Geometric means of ratios of catches by rain-trap to dry-trap 151 xiv XXV XXVI XXVII XXVIII XXIX XXX THE MICROBIOLOGY OF THE ATMOSPHERE Spores brought down by thunder rain terminating 7-day dry spell, Rothamsted ....... Mukiple-infection transformation : percentages to infections . 'Probable flight range' based on Rombakis's modification of Schmidt's theory ........ Effect of gradient on distance of horizon of infection . Pollen trapped on lightships in Gulf of Bothnia . Stakman & Hamilton's (1939) data on long-distance dissemina- tion of Puccinia graminis ...... PAGE 164 168 184 186 187 XV I HISTORICAL INTRODUCTION The air we breathe, like our food and drink, varies in quality from time to time and from place to place. This fact was recognized many centuries before industrialized man assumed the right to pollute the atmosphere with poisonous chemicals and radioactive isotopes. In Britain we hold that, 'when the wind is in the East 'tis neither good for man nor beast'. Some places are noted for invigorating air, and some for relaxing air; but it is not yet clear whether these properties are associated merely with differences in temperature, humidity, and move- ment of a gaseous mixture consisting mainly of 78 per cent nitrogen, 21 per cent oxygen, and 0-03 per cent carbon dioxide with traces of the inert gases, or whether some other factor or factors are involved. Speculations on the Origin of Disease Classical \NTiters believed that the wind sometimes brought sickness to man, animals, and crops. Hippocrates, the father of medical science, held that men were attacked by epidemic fevers when they inhaled air infected 'with such pollutions as are hostile to the human race'. A rival, though perhaps not entirely incompatible, view held that epidemics were the result of supernatural agencies, and were to be m arded off or cured by taking appropriate action. Lucretius in about 55 B.C. held quite modern views. He observed the scintillation of motes on a sunbeam in a darkened room and concluded that their movement must result from bombardment by innumerable, invisible, moving atoms in the air. This brilliant intuition enabled him to account for many interesting phenomena, including the origin of pesti- lences. We now know that bodies which transmit human diseases through the air are larger than those which Lucretius thought of as atoms — the mosquitoes carrying malaria, for instance, or the droplets which spread the common cold and influenza viruses indoors. But in his concept of baleful particles carried in clouds by the wind, settling on the wheat or inhaled from the polluted atmosphere, Lucretius touched on some of the main problems existing in plant patholog}'' and allergy today. Early Microscopists and the Discovery of Spores After Lucretius, more than 1,500 years passed before men even began to be aware that the air teems with microscopic living organisms. A I THE MICROBIOLOGY OF THE ATMOSPHERE The discovery had to wait almost until the invention of the micro- scope. For a long time after Aristotle and Theophrastus, the lower plants lacking obvious seeds were believed to be generated spontaneously in decaying animal or vegetable matter. The same view was held of the origin of many of the lower animals. However, the minute 'seeds' or spores of several kinds of plants were observed in the mass long before the invention of the microscope allowed them to be identified and observed individually. What was more natural than to suppose that these minute particles were wafted about by the winds } The discovery of reproduction of ferns is attributed to Valerius Cordus {b. 1515, d. 1564), and spores of the fungi seem to have been observed soon after this by a Neapolitan botanist, J. B. Porta, although the rusty- coloured spore deposits under bracket-fungi on beech trees must always have been familiar to the countryman. It was P. A. Micheli {b. 1679, d. 1737), botanist to the public gardens at Florence, who first illustrated the 'seeds' of many fungi, including mushrooms, cup-fungi, truffles, moulds, and slime-moulds. Further, by sowing spores on fresh-cut pieces of melon, quince, and pear, and repro- ducing the parent mould for several generations, he showed that the spores of some common moulds were, indeed, 'seeds' of the fungi. He noted, however, that some of his control slices also became contaminated, and he concluded that the spores of moulds are distributed through the air {see Duller, 191 5). The hand-made lenses of Anton van Leeuwenhoek rendered visible the world of minute organisms whose existence had only been guessed at before, and whose significance in nature had scarcely even been imagined. He could just see bacteria, and in his letters to the Royal Society in 1680 he described some yeasts, infusoria, and a mould. From his experiments he came to doubt the current belief in spontaneous generation; it seemed more plausible to him to suppose that his 'animalcules can be carried over by the wind, along with the bits of dust floating in the air' (Dobell, 1932). The controversy over spontaneous generation was to last for a couple of centuries ; but, in the second half of the eighteenth century, ideas w^ere developed by Nehemiah Grew and E. F. Geoffrey on the function of the pollen of flowering plants. J. G. Koelreuter, in 1766, was perhaps the first to recognize the importance of wind-pollination for some plants and of insect-pollination for others. C. K. Sprengel in 1793 developed these views and concluded that flowers lacking a corolla are usually pollinated in a mechanical fashion by wind. Such flowers have to produce large quantities of light and easily-transported pollen, much of which misses its target or is washed out of the air by rain. Thomas A. Knight in 1799 reported that wind could transport pollen to great distances. By the beginning of the nineteenth century, therefore, it was recognized that pollen of many, but by no means all, species of flowering plants, and HISTORICAL INTRODUCTION the microscopic spores of ferns, mosses, and fungi — as well as protozoa — were commonly liberated into the air and transported by the wind. The potential sources of the air-spora had been discovered and identified in the main before the year 1 800, but their role remained obscure. Controversy on Spontaneous Generation * Leeuwenhoek had come to doubt the belief, dating from Aristotle, that flies, mites, and moulds were generated spontaneously by decaying animal and vegetable matter. To him it seemed likely that animalcules could be carried by the air, and this provided an alternative explanation to spontaneous generation. J. T. Needham {b. 1713, d. 1781) had claimed that minute organisms would appear in heated infusions; but L. Spallan- zani {b. 1729, d. 1799) showed, by a series of experiments, that when organic materials were subjected to sufficient heat-treatment (with various precautions against contamination) they would neither putrify nor breed animalcules unless exposed to air. From this Spallanzani concluded that the microbes were present in the air admitted experimentally to his sterilized vessels. A rearguard action was fought to explain away these results. J. Priestley (^. 1733, d. 1804) and L. J. Gay-Lussac {b. 1778, d. 1850) claimed that heating the vessels drove out the air and that it was shortage of oxygen., not lack of 'seeds', which prevented heat-sterilized materials from generating a microbial population. Meanwhile, Appert (1810) put heat sterilization on a commercial basis by applying it to food preservation ; but the controversy lingered on, even into the present century, although the experiments and polemics of Louis Pasteur were decisive. Pasteur showed that food could be conserved in the presence of oxygen and that preservation depends on the destruction by heat of something contained in the air. In 1859 F. A. Pouchet, of Rouen, had raised the objection that a very minute quantity of air sufficed to allow the development of numerous microbes in heated infusions, and that the air would have to be a thick soup of microbial germs. In reply, Pasteur (1861) sterilized a series of evacuated flasks con- taining nutrient medium. So long as the flasks remained unopened they all remained sterile; but, even when they were opened and air was ad- mitted, he found that one or two out of each batch would remain sterile on incubation. Pasteur replied to Pouchet, denying that only a minute quantity of air needs to gain access for a microbe population to develop and for putrefaction to take place. On the contrary, the cause of the phenomenon was discontinuous and a sample of 250 cc. of air might or might not contain germs. Pasteur then showed, by opening batches of about forty such flasks in various sites, that the quantity of airborne germs differed in different places. In the open air in Paris he obtained bacteria, yeasts, and moulds ; * See also Bulloch (1938) and Oparin (1957). 3 THE MICROBIOLOGY OF THE ATMOSPHERE but some flasks remained sterile. In cellars of the Observatoire, where the temperature was constant and the air still and dust-free, many more flasks remained sterile. On 5 November i860, Pasteur deposited at the office of the Academy- no fewer than seventy-three quarter-litre flasks, some of which he had opened to the air in batches of twenty at various heights ranging from the foothills of the Jura to high up on Mont Blanc, as follow s : Number of flasks Altitude Locality where air sampled Contaminated Sterile Country air, far from dwelling houses, on the first plateau of the Jura 8 12 850 metres Jura mountains 5 15 2,000 metres Montanvert, near Mer de Glace on Mt. Blanc i 19 The cause of this supposed 'spontaneous generation' was not only discontinuous but, moreover, its concentration decreased with height. F. A. Pouchet had admitted that among dust particles of vegetable origin there were some spores of cryptogams, but he held that these were too few to account for the phenomena of putrefaction. Pasteur decided that he would abandon Pouchet's method, which relied on examining spontaneous deposits of dust on the surface of objects, in favour of a new method of studying the particles by collecting from actual suspension in the air. Pouchet had drawn invalid conclusions from surface deposits because, according to Pasteur, the light air-movements which constantly play over surface deposits would pick up and remove the extremely minute and light spores of microbes more readily than they would any coarser particles. (It now appears, however, that the small numbers of the lighter bodies in surface deposits is due to the extreme slowness with which they are deposited, rather than to their preferential removal after deposition.) Pasteur's apparatus for extracting the suspended dust in the air, for microscopic examination, was quite simple (Fig. i). A tube of | cm. diameter was extruded into the open air through a hole drilled in a window frame several metres above the ground. The rear part of the tube was packed with a plug of gun-cotton to catch particles. Air was drawn through the apparatus by means of a filter pump, and the volume of air was measured by displacement of w^ater. Tests were made on air draw^n from beside the Rue d'Ulm, and from the garden of the Ecole Normale in Paris. During aspiration, solid particles were trapped on the fibres of the gun-cotton plug. After use, the gun-cotton was dissolved in an alcohol- ether mixture, the particles were allowed to settle, the liquid was decanted, and the deposit was mounted for microscopical examination. 4 HISTORICAL INTRODUCTION Fig. I. — Pasteur's gun-cotton filter for airborne microbes. a = gun-cotton plug, i cm. long, held in position by: b = spiral platinum wire. FF = window frame drilled to allow passage of: T = tube to exterior for sampling outdoor air. R (m.k.l.) = aspirator. Pasteur, as usual, had little interest in the specific identity' of his or- ganisms; he was no taxonomist. The particles exactly resembled the 'germs' of lower organisms. They differed in volume and structure so much among themselves that they clearly belonged to very many species or even groups, including bacteria, moulds and yeasts. Their numbers contradicted the general conclusion that the smallest bubble of air admitted to a heat-sterilized medium is sufficient to give rise to all the species of infusoria and cryptogams normal to an infusion. This view was sho\\Ti to be highly exaggerated, and Pasteur indicated clearly that it is sometimes possible to bring a considerable volume of ordinary air into contact with an infusion before living organisms develop in the latter. Pasteur had demonstrated visually the existence of an air-spora, he had pointed out that it should be measured while in suspension and not after deposition on surfaces, and he had made the first rough visual measurements of its concentration in the atmosphere of the City of Paris: a few metres above the ground in the Rue d'Ulm, after a succession of fine days in summer, several thousands of micro-organisms were THE MICROBIOLOGY OF THE ATMOSPHERE carried in suspension per cubic metre of air. He then abandoned the method — remarking, however, that it could doubtless be improved and used more extensively to study the effects of seasons and localities, and especially during outbreaks of infectious diseases. The Germ Theory of Disease We must now look back and trace the growth of the microbial theory of disease, that had been developing for more than a century. The minute growths of fungus noticed for centuries on mildewed or 'rusted' plants were believed to be a consequence of the diseases; the dusty powder on rusted wheat was regarded as a curiously congealed exudation of the diseased plant itself. But might this not be putting the cart before the horse ? Could the rust possibly be the cause of the disease instead of an effect? Perhaps the first to give reasonably affirmative evidence was Fontana (1767), who examined wheat rust with his micro- scope and described what he saw as a grove of parasitic plants nourishing themselves at the expense of the grain. As further crop diseases were studied it became clear that, in some, infection is acquired by planting in contaminated soil, while others are carried on seed and still others are spread in the wind by airborne fungus spores {see Large, 1940). The discovery that microbes can cause disease in man and animals came somewhat later, and the first animal pathogens to be recognized were again fungi — no doubt because they were easier to find than bacteria. In 1835, Agostini Bassi showed conclusively, by inoculation experiments, that a specific mould is the cause of the 'muscardine' disease of silkworms which was then threatening the silk industry of Piedmont. Next, histori- cally, came the recognition of the fungi causing favus, ringworm, and 'thrush' in man, as a result of the work of David Gruby and Charles Robin. Pasteur had demonstrated that microbes are normally abundant in the air. Many of them can cause fermentation or putrefaction when intro- duced into sterile organic substrates; and it was natural to speculate that others might be the causes of epidemics of some of the so-called 'zymotic' diseases whose etiology was then unknown. Medical workers soon began a systematic search among airborne microbes for the unknown causes of infectious diseases. The search was long, and on the whole unfruitful because most epi- demic diseases that attacked man were gradually traced to sources other than the outdoor air. However, in the course of the search, most of the im- portant characteristics of the air-spora were discovered — and then forgotten . The search occupied the last thirty years of the nineteenth century and coincided with the golden age of bacteriology. Listing the dates of contem- porary salient advances in bacteriology will help to give the background to this phase of aerobiology {see Bulloch, 1938). 6 HISTORICAL INTRODUCTION Pasteur, L. Microscopical and cultural demonstration of the existence of an air-spora, and the fermentation of urea by a Al/V/'oro^n/^ . . i Koch, L. Introduction of pure-culture methods, and demonstration of spore production in bacteria. Discovery of cause of anthrax Statement of Koch's postulates .... Introduction of gelatine to solidify media . Hansen, G. H. A. Discovery of cause of leprosy Neisser, a. Discovery of the Gonococcus . Koch, L. Discovery of the tubercule bacillus Discovery of the cholera Vibrio .... LOEFFLER, F. Discovery of bacillus of swine erysipelas NicoLAiER, A. Discovery of the tetanus bacillus . KiTSATO, S Yersin, a. ivanovski, d. Beijerinck, M. W Discovery of the bacillus of plague Discovery of filterable viruses in plants )I-62 1876 1878 1881 1874 1879 1882 1883 i88s 1894 1892 i8q8 The Hygienists and their In\t:stigation of the Air While the causes of infectious diseases of man and animals were being unravelled in laboratories and clinics, a series of field investigations into the air-spora was in progress to find whether fluctuations in number and types of microbes present in the atmosphere were connected with out- breaks of such diseases as cholera, t}^phoid, and malaria. Salisbury (1866) investigated the air-spora in connexion with malaria in the Ohio and Mississippi Valleys, by exposing sheets of glass above marshy places during the night and examining them microscopically. He observed small, oblong, Palmella-like cells singly or in groups on the upper side of the glass sheets, but never in the droplets which formed on the underside. He believed that these cells were produced from a grey mould growing on the surface of prairie soil, and were in fact its spores which were liberated at night and rose some 30 to 100 ft. in the air, none being present during the daytime. Their liberation could be prevented by covering the ground with a layer of quicklime or straw. Some form of the 'aeroconiscope', invented by Maddox (1870, 1 871), was in favour with many investigators in this period. The model used by Cunningham (1873) consists of a conical funnel, with the mouth directed into the Avind by a vane, ending in a nozzle behind which is placed a sticky microscope cover-glass on which were impacted dust particles driven into the cone by the wind (Fig. 2). Cunningham's studies were made in two Calcutta gaols where cholera and other fevers were rife, and where medical statistics were available. He sampled for 24-hour periods, and illustrations of representative catches of airborne organisms, mainly fungus spores and pollens, were published in a series of splendid colour plates. He found no correlation between these micro-organisms and the incidence of fevers in the gaols. Moist weather diminished inor- ganic dusts, but it appeared to increase the total number of fungus spores. The most intensive sustained analysis of bacteria and moulds in the atmosphere was made in Paris during the last quarter of the nineteenth 7 THE MICROBIOLOGY OF THE ATMOSPHERE century. Largely through the influence of the chemist, J. B. A. Dumas, the Observatoire Montsouris was launched as a State institution in 1871 to make records needed for meteorology and agriculture. The Observatoire was housed in a palace in the Pare Montsouris, about 5 km. south of the centre of Paris. One of its tasks was to be the microscopic and cul- tural study of the organic and inorganic dust in the air, including both Mucedineae (moulds) and bacteria. B A Fig. 2. — Cunningham's aeroconiscope. A = side view of apparatus (partly in section) ; B = face view of sticky surface behind apex of cone (on a larger scale). Observations were started in 1875 by M. Schoenauer. He was re- placed after a year or two by Pierre Miquel {b. 1850, d. 1922), the dis- tinguished bacteriologist, who continued in charge of the work for over a quarter of a century. During the course of the survey, various methods were tested and discarded or improved; but all aimed at estimating the HISTORICAL INTRODUCTION number of particles of various types contained in a measured volume of air. Moulds were at first estimated microscopically in a 24-28 hour deposit, obtained by impinging the air to be sampled on a glycerined glass slide which was placed horizontally 2 to 3 mm. above a downward-facing orifice. The diameter of the orifice was from 0-5 to 075 mm. Suction of 20 litres per hour was maintained by a water-operated pump (Miquel, 1879). Miquel found that this apparatus yielded about 100 times as many particles as the aeroconiscopes designed by Maddox and Cunning- ham, though for qualitative work away from the laboratory he still used a wind-operated trap of the Maddox type. Bacteria, especially bacterial spores, could not be satisfactorily counted microscopically and Miquel was forced to estimate them by cultural methods. At first he drew known volumes of air through liquid media (sterile beef extract, etc.), partitioning the liquid either before or after exposure into 50 or 100 vessels, and adjusting the volume of air sampled so as to leave from a quarter to a half of the vessels sterile — in order to get a reliable estimate of the number of bacterial particles in the volume of air sampled. The numbers of microbes in the air varied greatly in the same place at different times, and this variation was studied in relation to season, weather, district, and altitude. Miquel was the first to make a long-term survey of the microbial content of the atmosphere by volu- metric methods. In the Pare Montsouris, out-of-doors, Miquel estimated that the mould spores averaged about 30,000 per cubic metre in summer, some- times rising to 200,000 in rainy weather. In prolonged dry weather they decreased in number, and were only about 1,000 per cubic metre in winter, with very few indeed when snow was on the ground. While rain was falling the numbers of mould spores usually decreased considerably, but afterwards their numbers recovered quickly — in fact, much more quickly than did those of particles of inorganic dust. Resting stages (eggs) of infusoria were estimated at about i or 2 in 10 cubic metres of air. Pollen grains in June may make up 5 per cent of the airborne organic particles, while starch grains near habitations may account for i per cent. Bacterial numbers out-of-doors in the Pare Montsouris were at first estimated at about 100 per cubic metre; but improved culture media increased this figure by a factor of 7 to 10 times. The numbers of bacteria in the centre of Paris were, perhaps, 10 times as high again as in the Pare Montsouris, with larger numbers inside dwellings, and still more in crowded hospitals. The work showed signs of settling into a steady routine with the publication of Miquel's Les organismes vivants de Vatmosphere^ Paris, 1883. However, in 1883 and 1884 Miquel was stung into a burst of renewed activity by the intrusion of a rival centre for the study of hygiene which had been established in Berlin under W. Hesse, who used the new solid media which Miquel abhorred. With the collaboration of de Freudenrich THE MICROBIOLOGY OF THE ATMOSPHERE in field work, Miquel studied the microbial population of the air at high altitudes in the Alps by volumetric methods (1884, p. 524); with the help of a sea captain, M. Moreau, the air over the sea was studied on voyages to Rio de Janeiro, Odessa, Alexandria, and La Plata; the micro-organisms brought down in rain-water were caught, precipitated, and counted; hourly variations of fungus spores and bacteria in the air were studied on improved volumetric traps with sticky slides, or on paper impregnated with nutrient media and moved by clockwork. At Montsouris, fungus spores showed a diurnal periodicity with tw^o maxima at about 8 and 20 hours, regardless of wind velocity. When he pressed the study of changes in spore content of the air with passage of time still further, Miquel found that the hourly reading was merely a smoothing of still shorter-term variations. Trapping airborne bacteria at Montsouris on a moving paper disc imbibed with nutrient agar, Miquel (1885) observed a regular diurnal periodicity — with two maxima at approximately 7 and 19 hours averaging about 750 per cubic metre, and with two minima at approximately 2 and 14 hours averaging about 150 per cubic metre. This periodicity was not related to wind direction, and was not altered by moderate falls of rain. In the centre of Paris the bacterial content also showed two maxima and two minima, but there the minima were about equal to the maxima at Montsouris, and the times of the maxima were closely related to activities in the city such as sweeping the street, and to the passage of horse-drawn traffic . Miquel appears to have been overwhelmed by the richness of the information on the mould spore flora provided by his apparatus, for he promptly abandoned it, merely remarking 'the micrographer ^^ ho has the leisure could make some nice [curieuse] studies of this subject'. It was, however, not abandoned before the main elements in the mould-spora had been discovered by this excellent method. Interest in the mould-spora waned when it became clear that the devastating epidemic diseases prevalent from time to time in cities were not fungal in origin but were due to bacteria, and attention became ur- gently focused on drinking water as the source of many of the current epidemic fevers abounding in Paris. The laboratory at Montsouris then became the centre for the bacterial analysis of samples of drinking water sent from wells in Paris and other parts of France. Meanwhile, in Germany, the work of W. Hesse {b. 1846, d. 191 1) had proceeded along similar lines. Hesse's apparatus for air sampling con- sisted of a narrow horizontal tube, 70 cm. long and 3-5 cm. wide, con- taining a layer of Koch's nutrient gelatine. A known volume of air was aspirated slowly through the tube, and micro-organisms settled and grew on the medium. Most colonies developed near the entrance to the tube, and Hesse assumed that by the time the slow stream of air had reached the end of its 70 cm. course all micro-organisms had been precipitated 10 HISTORICAL INTRODUCTION by gravity. Hesse found that moulds penetrated much farther into his tubes than did the bacteria, and made the important deduction that mould-germs as found in the atmosphere are on the average lighter than the bacterial germs. This led him to conclude that, whereas fungus spores were usually present in the air as single particles, the aerial bacteria mostly occur in the atmosphere either as large aggregates, or attached to relatively large carrier particles of dust, soil, or debris (Hesse 1884, 1888). He also observed that most colonies consisted of a single species — bacteria usually in small colonies of pure culture, and fungi as isolated spores — and deduced that the airborne germs are not in the form of aggregates of different t\'pes. Hesse's method was also used in London by Frankland (1886, 1887) and Frankland & Hart (1887) on the roof of what is now known as the Old Huxley Building of the Imperial College of Science and Technology, and elsewhere. Simultaneous comparisons were made between the number of micro-organisms per 10 litres (as indicated by colonies growing on Hesse's tubes of peptone gelatine) and the number deposited on horizon- tal dishes of the same medium, expressed as the number deposited per unit area per minute. Tests were made both outdoors and inside crowded or empt}' buildings. Frankland noted that the number of colonies was greater when the mouth of the tube faced the wind rather than in other directions, so he standardized his method by always turning it at an angle of 135° to the wind. A control tube facing the wind but not aspirated was always used, and sometimes it had a substantial number of colonics. Frankland seems to have been the first to realize that aerodynamic effects are of major importance in techniques for trapping the air-spora. These methods for studying the air-spora were continued into the present century, notably by Saito (1904, 1908, 1922) in Japan, and by Buller & Lowe (191 1) in the Canadian Prairies. The Allergists The idea that men, other animals, and plants, could become infected by microbes which set up pathological changes, had been made acceptable by the analog}'' of sterile organic infusions that become infected with putrefying microbes. The idea became widely accepted during the latter half of the nineteenth century and, when once the cause of the common epidemic diseases had been established, advances in hygiene and therapy began to transform the social scene. Yet there remained some diseases for which no pathogenic or parasitic invader could be found. Some of these, such as pellagra and beri-beri, have now been traced to a variety of nutritional deficiencies. Another group, the so-called allergies, were at first difficult to grasp because a peculiar condition of the patient was a complicating factor. Allergic diseases, unlike those caused by invasion of the body by a pathogenic micro-organism, are due to a changed condition II THE MICROBIOLOGY OF THE ATMOSPHERE of an individual patient who has become sensitive and reacts adversely to substances, often in minute amounts, which normal individuals can tolerate. The substance or allergen can be taken into the body, for example in food, or by contact through the skin, or by inhalation from the air. Hay fever was one of these puzzles. Long before Pasteur's epoch, hay fever had been attributed to inhalation of pollen ; but it remained for Charles H. Blackley (1873), a Manchester physician, to prove by in- halation experiments on himself and others that this guess was correct, and to demonstrate by trapping methods that pollen was at times present in the air in large quantities. Blackley first tried Pasteur's gun-cotton filters and obtained some pollens, but too few to satisfy him. Finally he used four sticky horizontal microscope slides exposed under a roof supported by a square central post. The slides were placed at 'breathing level' (about 135 cm.), and he caught a maximum of 880 grains per sq. cm. per 24 hours on 28 June 1866. In 1867 his maximum was only 106, and in 1869 he placed his slides vertically in a vane shelter and gave no numerical data. He found that rain reduced the number of pollen grains caught to about 5 per cent of the number caught in dry weather. He ex- plored the air above the ground up to 1,500 ft. by means of kites, and found that vertical slides facing the wind caught nearly 20 times as much pollen at the higher altitude as at breathing level. Blackley showed by means of his sticky slides that the air contains enough pollen during the grass-flowering season for large quantities to be deposited on exposed surfaces. He also gave himself an attack of bronchial catarrh by inhaling Penicillmm and Chaetomiwn spores — an experiment which he said was too unpleasant to repeat. According to Durham (1942), after Blackley's pioneer work no progress was made with these studies until the period 1 910-16, when fresh in- terest was aroused by the discovery that injections of pollen extracts can be used to de-sensitize patients who are allergic to pollen. When the study of airborne allergens was again taken up in the present century, it was unfortunate that the technique chosen should have been the so-called 'gravity-slide' adopted by Blackley — a method which Pasteur had abandoned in 1861 and which Miquel had roundly con- demned as 'the simplest and most defective method' of collecting air- borne particles. By the early years of this century it became possible to assess the value of the ancient belief that the wind brings disease. Many diseases of crop, but very few diseases of man, have proved to be caused by minute particles carried on the wind. The particles are not some sort of invisible atoms as Lucretius thought; indeed, among the motes in the sunbeam, he may himself have been watching some of the baleful fungus spores and pollens which cause crop disease and respiratory allergy. 12 \ ^"' J'^/f'^^^^ } t\ w^ Ehrenberg's illustration of sample of dust collected by Charles Darwin on the Beagle near the Cape Verde Islands, January 1833. Plate 2 HISTORICAL INTRODUCTION Meanwhile evidence was accumulating that these particles might be carried by wind to distances vastly greater than had been imagined by the ancients. In dust deposited after transport for hundreds of kilometres by sirocco and trade winds, Ehrenberg (1849, 1872, 1872^?) found large quantities of protozoa and plant spores, and gradually he became con- vinced that viable micro-organisms could survive transport through the atmosphere. When the Beagle w^as near the Cape Verde Islands, Darwin (1846) found the atmosphere haz}' with dust from North Africa. In samples of this dust Ehrenberg found sixty-seven kinds of organisms — including freshwater infusoria and cryptogamic spores (Plate 2) — and Darwin at once grasped the importance of the phenomenon in the geographical distribution of organisms. 13 II SEDIMENTATION IN STILL AIR All the particles with which we are concerned are heavier than air. In still air they sink with characteristic and constant 'terminal velocity'. Stillness as a quality of air is only relative. In the laboratory we can make the air as still as possible by eliminating draughts and convection currents, only to find an intense underlying activity revealed by the scintillation of motes in a beam of light. The motes are small enough to be jerked irregularly by the impact of gas molecules; but they are too large to be transported bodily by molecular diffusion, and most of the phenomena of colloidal suspensions are irrelevant to the air-spora. We shall meet some analogies with the diffusion of a gas, however, in studying the diffusion of a cloud of spores in the atmosphere. In this study we usually ignore the underlying molecular activity of the medium, and consider a patch of air as 'still' if it is not being trans- ported bodily at more than a certain speed. Out-of-doors this speed might be 10 cm. per sec; in a room it might be i cm. per sec; and, under carefully controlled conditions in special apparatus, a higher standard might be expected. For the present we must leave the definition vague, and simply regard air as 'still' when, in a particular context, the effects of wind, turbulence, and molecular activity are negligible. Knowledge of the properties of small particles in still air throws light on the behaviour of spores in moving air out-of-doors. Factors Determining Velocity of Fall One effect of its molecular activity is that the air is viscous, i.e. it resists the movement of solid particles. A small particle liberated into the air from a resting position tends to fall with an acceleration due to gravity; however, the resistance of the air increases faster than the speed of fall, and a state of balance is soon reached in which the particle stops accelerating and continues to fall through the air at a constant terminal velocity. The terminal velocity of smooth spheres with diameters of between about I /x* and loo ju, is satisfactorily predicted by Stokes's law (for smaller particles Cunningham's correction becomes applicable, and larger par- ticles have to be treated experimentally). Stokes's law can conveniently be given in the form : * « == 1 oVo mm. 14 SEDIMENTATION IN STILL AIR 2 a — p Vs = -. .gr- 9 /^ where, in C.G.S. units at ordinary surface temperature and pressure: Vs = terminal velocity (velocity of sedimentation) in cm. per sec. ; a = density of sphere in gm. per cc. (water = i-oo); p = density of medium (air = 1-27 X lO"^ gm. per cc); g = acceleration of gravity (981 cm. per sec.^); ju, = viscosity of medium (air at i8°C. = i-S X iQ-^gm.percm.sec); r = radius of sphere in cm. {N.B. radius = \ diameter). For a water droplet falling in air, Vs = 1-2 X io~^ r^ cm. per sec, when the radius is expressed in microns (/it). A fog droplet of 10 p- radius (20 p. diameter) has a calculated terminal velocity of 1-2 cm. per sec. The pollens and spores with which we are concerned belong to the size-range where Stokes's law is valid, but they are seldom anything like smooth spheres. Stokes's law has given values of at least the right order, however, for spores whose terminal velocities have been measured experimentally. At first sight the pollen grains of some species of conifers appear to fall unexpectedly slowly, but these grains have conspicuous air sacs which greatly reduce the density of the individual particle. The diameters of particles constituting the air-spora vary from approxi- mately I ^ to lOO/Lt or more for the largest pollens and spores {see Appendix I, p. 207, Plates 5-7). Some spores are filamentous, perhaps one hundred times as long as wide. Although the densities of the spores of very few species have yet been measured, there are reasons for expecting them to be much less dense than mineral particles and indeed to resemble water droplets in density. The few determinations which have been made, relative to water = i, are as follows: Pohl (1937) (^angiosperm.\e; Almis glutinosa 0752 Bettila verrucosa o-8o8 Corylus avellana I -008 Dactylis glomerata 0-981 Fagus sylvatica 0713 Typlia augiistifolia 0747 Typha latifolia i-i6i (Gymnosperaiae) Junipenis communis 0-405 Picea excelsa 0-550 Pinus sylvestris 0-391 Pinus montana 0-496 Taxiis baccata 0-579 (PTERIDOPmiA) Lycopodium sp. I-I75 (Bryophyta) Polytricliiim sp. 1-53 Zeleny & McKeehan (19 10) 15 THE MICROBIOLOGY OF THE ATMOSPHERE (Fungi) Amanitopns vaginata 1-02 Buller (1909) Erysiphe polygoni (conidia) 1-094 Yarwood (1952) Lycoperdon sp. 1-44 Zeleny & McKeehan (1910) Peronospora destructor 1-34 Yarwood (1952) Piiccinia graminis tritici 0-807 to 862 Weinhold (1955) Uroiiiyces pltaseoli 1-36 Yarwood (1952) The properties of spores are not invariable, but may alter with external conditions — sometimes enough to have a marked effect on their terminal velocity. For instance, the spores of the toadstool Amanitopsis vaginata were recorded by Buller (1922) as falling at 0-5 cm. per sec. when ob- served immediately below the gill after liberation, but they became desic- cated on continuing to fall through dry air and soon slowed down to one third of their original speed. Durham (1943) gave laboratory determina- tions of densities of pollens, and for some the probable outdoor values which are shown in parentheses: Ambrosia elatior^ 0-63 (o'55); A. bidentata, 0-56 (0-50); Xanthiiim commune^ 0-52 (0-45); Iva xanthifolia, 079; Salsola pestifer, i-o (0-90); Acnida tamariscina^ i-o; Zea mays, i-io (i-oo); Phleum pratense, i-oo (0-90); Quercus imbricaria, 1-04; Juglans nigra, 0-93; Alnus glutinosa, 0-97; Fraxiniis americana, 0-90. Observed terminal velocities (vs) of spores and pollen grains are collected in Table I. TABLE I ELOCITIES OF POLLENS AND SPORES (cm. Vs per sec.) Author reference* 387 (11) 1-7 (10) 2-4 (10) 2-2-6-8 (9), (10), (II) 2-5 (10) 3" I (10) 5-5 (10) 9-9-22-0 (9), (10) 12-3 (II) 8-7 (10) 4-5 (II) 2-5 (10) 2-9 (10) 2-16 (II) 6-0-8-8 (9) 3-24 (II) 3-2 (10) 3-24 (II) Flowering plants Abies pectinata Alnus viridls Betula alba Carpiiuis betiilus Corylus avellana Dactylis glomerata Fagiis sylvatica Larix decidua Larix poJonica Picea excelsa Pinus cembra Piiins sylvestris Qiierciis robur Salix caprea Secak cereale Tilia cordata Tilia platypliylla Ulniiis glabra Pteridophytes Lycopodiiim sp. 1-76-2-14 (2)5(5) Bryophytes Polytrichum sp. 0-23 (2) 16 SEDIMENTATION IN STILL AIR TABLE I — contd. Fungi Alternaria sp. 0-3 (6) Amanita rube seem 0-15 (I) Aiiianitopsis vagiiiata o-29-o-6i (I) Boletus badius on (I) Boletus felleus 0-12 (I) Bovista plumbed 0-24 (5) Coprinus comatus 0-4 (I) Coprinus plicatilis 0-43 (I) Cronartium ribicola 2-03 (3) Erysiphe gramiiiis 1-2 (7) Galera tenera 0-21 (I) Helminthosporium sativum 2-0-2-78 (5), (6) Lycoperdon pyrifurme 0-05 (5) Ly CO per don sp. 0047 (2) Alarasmius oreades 013 (I) Monilia sitophila 016 (5) Paxillus involutus on (I) Pluteus cervinus 0-067 (I) Polyporus squamosus 010 (I) Psalliota campestris 013 (I) Puccinia coronata avenue II ID (4) Puccinia graminis secalis II i-o6 (4) Puccinia graminis secalis I 1-02 (4) Puccinia graminis tritici 1 1 0-94-1 -25 (4), (8) Puccinia graminis tritici I 1-06 (4) Puccinia triticina II 1-26 (4), (5) Russula emetica 016 (i) Tilletia tritici I -41 (5) Ustilago tritici 0-07 (5) Ustilago zeae 0-3 (6) *(i) Buller, 1909. (2) Zeleny & McKeehan, 1910. (3) McCubbin, 1918. (4) Ukkelberg, ^933- (5) Stepanov, 1935. (6) J. J. Christensen, 1942. (7) Yarwood & Hazen, 1942. (8) Weinhold, 1955. (9) Bodmer, 1922. (10) F. Knoll ex Rempe, 1937. (11) J. Dyakowska ex Erdtman, 1943. Two methods have been used for measuring terminal velocity. The simpler method is to time the fall over a short, measured distance in a small chamber of still air by direct observation with a horizontal micro- scope. It was used in the pioneer work of Buller (1909), and by Yarwood & Hazen (1942). So far this method has been used only for small, slowly- falling spores, because large ones travel too fast to be timed by direct observation. The method could no doubt be extended to fast-moving spores by photographing with a flash of known duration. The technique most generally used, however, has been to release spores or pollen at the top of a column of still air in a vertical cylinder and find the time they take to arrive at the bottom. This is the method used by Zeleny & McKee- han (1910), McCubbin (1918), Ukkelberg (1933), Stepanov (1935), and Weinhold (1955). McCubbin and Ukkelberg report results of similar type. The number of wheat-rust spores reaching the bottom of the tube in successive inter- vals of time showed a negative skew distribution. Ukkelberg was able to B 17 THE MICROBIOLOGY OF THE ATMOSPHERE show that part of this skewness was due to the presence of clumps of spores which fell faster than single units. It is also clear that, with both uredospores and aecidiospores of rust fungi, a large number of single spores fall very slowly. Measurements are needed to test whether, within one species, the single spores arriving first at the bottom are larger than those arriving at the end of the experiment. Another possibility is that small eddies may have hastened the fall of some spores and retarded that of others. A more serious defect of the method is that a vertical circulation of air by convection in the cylinder might bias the results by introducing a systematic acceleration or retardation of fall. This drawback could be overcome by establishing a small temperature difference between the top and bottom of the column, so that the stratified air would be stabilized as in a 'temperature inversion'. A thermostat may produce artefacts from convection currents set up by rhythmic temperature changes. BuUer (1909) emphasized the difficulty of reducing air to anything like stillness, even in closed beakers. In air, spores gain or lose water rapidly and the effect of spore hydra- tion on terminal velocity, noted earlier by Duller, is evidently complex. Weinhold (1955) showed that with uredospores of Puccinia graminis tritici, changes in volume and weight occurred within 3 minutes of transfer to air of different temperature and humidity. Weinhold reported that, contrary to expectation, spores stored at 5 per cent relative humidity fell at 1-25 cm. per sec, in spite of being smaller and less dense than spores stored at 80 per cent relative humidity, which fell at i-i cm. per sec. Increasing the humidity of air through which the spores fell increased the terminal velocity, which was: 1-03, 1-22, 1-23, and 1-54 cm. per sec. at relative humidities of 24, 45, 52, and 80 per cent, respectively. With increasing temperature, terminal velocity decreased from i-o6 cm. per sec. at 23'4°C. to 0-94 cm. per sec. at 39-9°C. We still lack observations on the rate of fall of highly elongated fungus spores found in such genera as Ophiobolus, Epkhloe, Geoglossum^ and Cordyceps, whose unusual shape makes Stokes's law inapplicable. Falck (1927) calculated terminal velocities for a number of species with approxi- mately elliptical spores on the assumption that the expected velocity Ve = vj i^(a/b), where v^ is the fall velocity of a spherical particle of the same volume, and a and b are axes of the ellipse. McCubbin (1944) stressed our lack of observations on asymmetrical spores, and provisionally suggested a method of calculating terminal velocity on the assumption that surface drag accounts for most of the retardation. He showed that observed terminal velocities of most spherical and oval spores fitted the r 1 length X width , . . . . approximate formula Vs = , where velocity is in mm. per 40 sec. and spore dimensions are in microns. Fusiform spores were treated 18 SEDIMENTATION IN STILL AIR as consisting of an intercalated cylinder (length = l/it) between two axial cones (each of axial length = x/x), Vs being again in mm. per sec. During fall in still air, an asymmetrical particle will assume a charac- teristic orientation. Hydrodynamical theory requires that the orientation assumed will be that in which the resistance of the air to the motion of the particle is greatest. This phenomenon can be observed with the naked eye if minute airborne particles of fibre are watched in a beam of light in a still, darkened room. We know very little as yet about spore orientation. Duller (1909) observed that some slightly elongated spores tend to fall with their long axis horizontal, as is to be expected for dynamical reasons. Sometimes factors other than shape seem to influence the orientation of an asymmetric spore. When Yarwood & Hazen (1942) watched the smooth conidia of Erysiphe graininis, measuring 32 X 20 /x, during fall in vertical glass tubes 7 mm. in diameter, they saw that half of the spores fell with the long axis horizontal and the other half with it vertical. This might indicate an uneven distribution of materials of different density in the cell contents; but, more likely, the vertical position was due to drag at the wall boundary, because if the tube is made even narrower, all the spores fall vertically. The present author has seen the filamentous ascospores of Cordyceps gracilis similarly oriented whilst being carried up by convection currents beside a vertical glass surface. Further, while watching the tailed spores of the puffball, Bovista phwihea, falling in a small chamber on the stage of a horizontal microscope, the tail was seen to trail behind the spherical spore. In Chapter VI it will be indicated that spores tend to be deposited with characteristic orientation on a surface. Stokes's law holds for smooth spheres. Few pollens or spores are spheres, but a large proportion of them are microscopically smooth. Others, when highly magnified, are seen to possess warts, spines or other projections, or even to be pitted. These roughnesses would be expected to increase friction during movement through air and to retard fall, but we have as yet no experimental evidence of this. Viewed over the whole range of spore and pollen size of, say, 4 to 100 /x diameter, and of terminal velocities of from 0-05 to 10 cm. per sec, it is clear that Stokes's law gives a good idea of terminal velocity in still air, but that asymmetry and surface roughness may play a part as yet unmeasured. Effects of Sedimentation The effects of spore fall in still air can be observed indoors, particu- larly if a room is left closed and unoccupied — a fact noted early in the study of air hygiene by workers using Hesse's horizontal tube method of air sampling, or some modification of it {see Chapter I, p. 10). Although all these investigations suffer from the defect of being based on highly 19 THE MICROBIOLOGY OF THE ATMOSPHERE selective culture media, all agree that wind or crowds stir up micro- organisms, and that these soon settle in buildings when the air is left undisturbed. At the Royal Institution in London, England, Tyndall (1881) made a close study of microbes in the air in relation both to the question of spontaneous generation and to the antiseptic surgery which was being developed by Lister at that period. Tyndall showed that the air of a darkened room scattered a powerful beam of light. Gas molecules did not appreciably scatter light. Scattered light always arose from suspended particles^ some of them too fine to be described as motes. By passing a beam of light through windows in the side walls of a glass-fronted box, he showed that, after a day or two, the air became optically empty, the par- ticles having settled on the floor and even on the roof of the box. At the same time Tyndall found that the air, previously full of microbes, had become sterile. The ability to generate life was associated with the presence of the light-scattering particles, and the air of small spaces could be sterilized by sedimentation. Tyndall had the curious idea that microbes remained associated in the air in clouds, much as fish are associated in shoals, and he explained that some of Pasteur's flasks (pp. 3-4) must have been opened within clouds, while others were opened between clouds of floating microbes. We now think of micro-organisms as distributed in the air at random (Home, 1935), but, under certain conditions, it may be that Tyndall was right. Outdoors the effects of terminal velocity are usually masked by the speed and turbulence of the wind. However, conditions are sometimes tranquil enough for its effects to be detected. One example was found by Rempe (1937), of Gottingen, who made a series of aeroplane flights both by day and by night to study the distribution of tree pollen over German forests. By trapping on sticky cylinders, he obtained evidence that pollen grains of different sizes and terminal velocities differed in their relative abundance with altitude, even by day (Table II). TABLE II POLLEN DISTRIBUTION AT DIFFERENT ALTITLIDES (After Rempe, ig^j ; day flight, A 6) Approximate Per cent of total p ollen at he Pollen diameter (m) 10- 40 metres 2,000 me Bettila (birch) 22 29-0 73-3 Carpinus (hornbeam) 37 55-0 lO-O Fagus (beech) 38 II-5 3-3 Others — 4-5 13-4 By night, it sometimes happened that pollen grains were partially sorted out according to size even within a single species, as shown by the 20 SEDIMENTATION IN STILL AIR mean diameters of birch pollen on the night flight, A lo (Table III). The size-range recorded by Rempe varied from 23 ^ to 27-5 |U., so it is evident that even at night the sorting effect was not great — a difference in altitude of 1,000 metres was associated with a drop of only 4-5 fi (or i/gth of the mean diameter), with an estimated terminal velocity differing between 1-6 and 2-3 cm. per sec. TABLE III SIZE OF Betula POLLEN at DIFFERENT ALTITUDES (After Rempe, igj/ ; night flight A 10) Altitude Mean diameter Estimated terminal (metres) (^) veloc ty (cm. per sec.) 1,000 23-0 1-6 800 24-5 1-8 500 267 21 200 27-5 2-3 10-40 27-2 2-2 Quite insignificant convection currents may be enough to counteract the terminal velocity of fall of small spores. Falck (1904) believed that the fruit-bodies of the larger fungi generate sufficient heat to induce convection currents which could carry their spores upwards. The temperature of an insulated mass of Polyporiis squamosus rose nearly io°C. in 10 hours, and he regarded parasitism by maggots as a heat-generating adaptation favouring dispersal. BuUer (1909) justly criticized this view, but field experiments are needed to determine whether the pilei of agarics modify air-flow by their own heat, by absorbing solar radiation, or by their shape generating stationary eddies in an air stream. The colours of agarics are usually considered to be functionless, but the presence of dark colours among species inhabiting burnt ground suggests that this character may have been selected during evolution. It is not impossible that both colours and shapes of agaric fruit-bodies are partly adaptive. 21 Ill THE ATMOSPHERE AS AN ENVIRONMENT Aerobiology is a synthesis: just as the geographer draws upon astronomy and geology, so the aerobiologist draws upon many sources. To under- stand the environment of the air-spora we must go to meteorology. A fuller account of the relevant features of the atmosphere than that given here can be found in works on dynamical meteorology, e.g. SutclifTe (1940), Geiger (1950), Sutton (1953), and U.S. Weather Bureau (1955), the last including many excellent diagrams. The Atmosphere and its Layers The atmosphere is usually recognized as layered; some of its main features are illustrated in Fig. 3, in which altitude is drawn on a logarith- mic, instead of a linear, scale in order to allow the various layers to be represented together on one page and to illustrate vividly how the proper- ties of the atmosphere change most sharply near the ground. Barometric pressure, density of the air, and (as a rule) temperature, decrease with increasing height above the Earth's surface. These changes are all quantitatively important in aviation, and calculations are based on a table of an agreed 'International Standard Atmosphere'. Changes with altitude in temperature, humidity, density, and viscosity will be complex in their effect on a suspended spore, but are not likely greatly to affect its terminal velocity. The three vertical panels of the diagram represent conditions in contrasting weather types. The central panel represents a dull, windy day, with a cloud layer shielding the ground from direct sunlight (conditions on a cloudy night are not very different). The right-hand panel represents a sunny day, and the left-hand panel a still, cloudless night. The thickness of each individual layer of air varies according to conditions ; the boun- daries between them vary in definiteness: sometimes transitions are imper- ceptible, but there is sometimes even a visible interface between layers. The layers are variously named in the literature and this may be confusing unless the following approximate equivalents are borne in mind {see p. 23). It is convenient to describe these layers in the reverse order, from ground-level upwards, beginning in the troposphere with the laminar boundary layer. 22 lOkm lOOm lOm ttt t THE ATMOSPHERE AS AN ENVIRONMENT STRATOSPHERE TROPOPAUSE CONVECTION LAYER \ OUTER FRICTIONAL TURBULENCE LAYE^ I m _ lOcm V t t t t i^ I TURBULENT BOUNDARY LAYER ii' LAMINAR BOUNDARY LAYER t t-J Still conditions > Wind velocity increasinq > CLEAR NIGHT CLOUDY DAY SUNNY DAY Fig. 3. — Diagrammaticrepresentationof layers of the atmosphere with a logarith- mic vertical scale. Nomenclature of atmospheric layers STRATOSPHERE TROPOPAUSE FREE 'Convective (convection) ATMOSPHERE layer Transitional or outer frictional turbulence layer TROPOSPHERE Turbulent boundary] PLANETARY layer BOUNDARY Surface LAYER Local eddy layer boundary laver Laminar boundary layer 23 the microbiology of the atmosphere The Troposphere The troposphere is the name given collectively to the lower layers of the atmosphere extending from the ground to a height of approximately 10 km., and is a region characterized by a decrease in temperature with increasing height — the temperature lapse. Air is relatively transparent to the short-wave radiation of sunlight which therefore heats the air very little as it penetrates the lower layers of the atmosphere. On a sunny day, solar radiation falling on the Earth's surface is in part temporarily absorbed, and in part reflected back as a radiation of longer wave-length that is more readily absorbed by air. This reflected radiation now heats the layer of air near the ground and the heat later becomes diffused through the lower layers of the atmosphere from below upwards. Air temperature is thus highest near the ground and decreases with increasing height, unless a 'temperature inversion' is formed under conditions described below. The normal temperature decrease (or 'lapse rate') is about o-6°C. per lOo metres. At the top of the troposphere is the tropopause — the boundary between troposphere and stratosphere. The troposphere comprises the five following layers. laminar boundary layer In contact with the surface of the earth and all projecting bodies is a microscopically thin layer of air held firmly by molecular forces. Except for molecular diffusion this layer is still and windless. Above this windless film the atmosphere is usually in motion, set going either by pressure differences of distant origin, or by convection currents produced by local heating. The lowest layer of moving air, next to the still layer, is known as the 'laminar boundary layer' (or sometimes the ' laminar' layer). This again is a thin layer, of the order of a millimetre thick, in which there is no turbulence and the air flows in streamlines parallel to the nearest surface; heat, gases, and water vapour can move across the streamlines by molecular diffusion. Wind speed is negligible at the still surface film, and in the laminar boundary layer wind speed increases linearly with height (momentum being transmitted through the layer by molecular diffusion only). Particles, droplets, or spores getting into the laminar layer will sink through it, following trajectories deter- mined by wind speed and gravity, and will come to rest at the Earth's surface. A laminar layer also exists at the interface around any solid body, and much of the foregoing description applies equally to the air layer at the surface of a leaf or stem. The thickness of the laminar boundary layer varies with the wind speed and with the roughness of the adjacent surface. In a high wind it may be thinned down to a fraction of a millimetre, and turbulent air from the 24 THE ATMOSPHERE AS AN ENVIRONMENT next higher layer may reach do\Mi nearly to the surface. In very calm weather the laminar layer may thicken considerably. In comparison with the relatively equable air at a metre or two above the surface, the eco-climate of the laminar boundary layer is violently changeable (Monteith, i960). Unless protected by a layer of vegetation, small organisms at ground-level may be subject to extreme heat from the sun's rays by day, followed by a rapid drop in temperature as heat is lost by radiation to a clear sky at night. The laminar boundary layer acts as a dust trap. Particles which have sunk through it and come to rest in the still or slowly moving air at the surface, are out of reach of eddies — until some unusual condition arises which thins the laminar layer enough for eddies to penetrate down and sweep away the dust particles. High winds may do this ; or local heating of the surface, perhaps on a micro-scale, may set up 'dust-devils' — smaller or larger whirlwinds raising dust into the air. LOCAL EDDY LAYER For biological purposes we need to add the 'local eddy layer'. Even in streamlined air, local stationary eddies may exist behind small rough- nesses; and, as will be shown on page 35, air-flow over a cup-shaped depression may set up a rotation pattern sufficient to throw dust up from the bowl. This layer is probably important in nature, where ideally smooth surfaces are rare. A special r\'pe of boundary at the top of a plant layer or crop has been called the 'outer active surface', or, in forests, the 'crown layer'. TURBULENT BOUNDARY LAYER In this layer, where flux of momentum decreases linearly with height, solid obstacles, arising at the surface in the laminar boundary layer, project into the wind and cause eddies which break away from the surface and travel downwind. A surface is aerodynamically smooth in conditions when the laminar layer is thick enough to submerge projections from the surface; but if the irregularities project through the laminar layer, the surface is considered rough. As the thickness of the laminar layer depends both on the wind speed and the stability of the atmosphere, it is clear that a particular surface such as a grass sward or a hairy leaf may be aerodynamically smooth under one set of conditions and rough under another. Each surface has a characteristic roughness parameter. Air-flow over calm water may be smooth ; but, except at extremely low wind-speeds, flow over land is normally rough and disturbed by surface irregularities which cause turbulence. Eddies of two types may occur : local or stationary eddies which may arise on both the windward and leeward sides of a bluff" obstacle, and eddies which break away and travel \^ith the wind in the obstacle's wake, 25 THE MICROBIOLOGY OF THE ATMOSPHERE The forward velocity of a turbulent wind is thus the net result of a com- plex movement; the wind has vertical and lateral components as well as the horizontal movement. Further, vertical and horizontal turbulence may differ in intensity (non-isotropic turbulence). Occurrence of mechanical or frictional turbulence depends on the wind speed being high enough, and the object large enough, to cause eddying. Whether or not flow will be turbulent can be calculated by the method of Osbert Reynolds, who found that flow is turbulent when the Reynolds number, , _ , length X wind velocity , , Re, denned as — 7-. -. — ; -. , exceeds about 2,000. kmematic viscosity Here 'length' is taken as a characteristic dimension of the object, and kinematic viscosity for air under average surface conditions is 0-14 cm^. sec.~^ Thus for a leafy bush 100 cm. high in a wind of 100 cm. per sec. we have „ 100 cm. X 100 cm. sec.-^ Re = 2 = 7,100 0-14 cm."^ sec.~^ so flow would be expected to be turbulent. In the turbulent boundary layer, properties such as temperature, amount of water vapour, and wind velocity, change much less rapidly with increasing height than in the laminar boundary layer beneath. Eddies mix the different parts of the layer much more rapidly than do the slow processes of molecular diffusion. Particles can also be carried by eddies upwards and laterally in a manner impossible in the laminar layer. In the turbulent boundary layer the wind velocity, temperature, and amount of water vapour show a change which is linear with the logarithm of the height. In this layer diurnal changes of temperature are less pronounced than in the laminar boundary layer underneath, and diurnal changes decrease still further with increasing height until, at the top of the next layer, they have almost disappeared. An increase in wind-speed increases the thickness of the turbulent boundary layer both downwards, by thinning the laminar boundary layer, and upwards, by pushing into the transitional layer as turbulence increases. The turbulent boundary layer is thinnest on clear calm nights and thickest on hot sunny days, when it may reach to a height of 150 metres. The turbulent boundary layer is the part of the atmosphere most familiar to us. While our feet are planted in the violently fluctuating climate at ground-level, our heads, and the weather-recording instruments of the conventional Stevenson's screen, inhabit the relatively equable turbulent layer. 26 THE ATMOSPHERE AS AN ENVIRONMENT TRANSITIONAL OR OUTER FRICTIONAL TURBULENCE LAYER Here frictional turbulence, generated in the layer below, still domin- ates vertical diffusion, but it dies out gradually until, at the top of the layer, both turbulence and diurnal temperature changes disappear. Both layers may be dusty, and the top of the transitional zone is sometimes visible as a distinct dust horizon at 500-1000 metres, marking the upper limit to which spores are raised by frictional turbulence (though much greater heights may be attained by convection). In dynamical meteorology this zone is defined as the region where the wind structure is determined partly by surface friction and partly by the Earth's rotation. CONVECTIVE LAYER This layer extends from about i km. above the ground to the top of the troposphere at about 10 km. As in all the layers constituting the tropos- phere, the temperature continues to decrease with height to the top of the convective layer, though diurnal temperature variation is almost absent. Frictional turbulence does not reach here, but, as already indicated, particles from the Earth's surface can be carried into this layer by large- scale convection currents when the ground is heated by sunshine. The height above the ground attained by a mass of heated air before it loses buoyancy, depends on the temperature gradient and water-vapour content of the air at the time, as explained in the standard works on dynamical meteorology. Ascent may be halted if there is a temperature inversion layer in the upper atmosphere. Under conditions of thermal instability, 'bubbles' of heated air may arise intermittently from areas where the ground or vegetation is being heated by the sun. These bubbles may rise to the convective layer, carrying spores and other particles as well as water vapour to the level at which cumulus clouds are formed, and at times reaching to the base of the stratosphere (Mason, 1957). NIGHT RADIATION AND TEMPERATURE INVERSION At night, wind speeds tend to diminish; the laminar boundary layer then becomes thicker than by day and the turbulent boundary layer may become thinner, being reduced to perhaps only 10 to 15 metres in thickness. These changes may be carried still farther if the sky is cloudless, thus allowing radiation from the ground to escape into space. Loss of heat by radiation cools the ground and this in turn cools the air lying nearest to the ground. Thus, instead of temperature decreasing with increasing height, a 'temperature inversion' is set up: over the cold air near the ground lies air at a higher temperature — up to a certain height, the top of the inversion, above which the usual lapse rate is again encountered. 27 THE MICROBIOLOGY OF THE ATMOSPHERE As the air is coldest and densest at the bottom of the inversion, gravity tends to prevent it from ascending and mixing with warmer air above. The air in the inversion becomes stratified according to temperature and remains very stable, in contrast to the instability that is apt to develop when the ground is heated. Such a layer of cold, heavy air may flow slowly downhill as a nearly laminar katabatic wind, filling hollows with cold air and aiding the formation of frost pockets (Geiger, 1950, p. 203). In a temperature inversion, spores and dust particles tend to settle out, leaving the air relatively clean although the air above the inversion may continue to carry a normal spore-load. ROLE OF CONVECTION When the surface of the ground is heated by sunshine the lowest layer of air may be heated in turn. If a large temperature lapse-rate is established, the atmosphere becomes unstable, because the less dense ground-layer of air tends to rise and carry its load of microbes upwards, being replaced by cooler air from above. The pattern of this overturn is not yet clear. A regular 'cellular' pattern of ascending and descending air has been suggested, but more recently the ascending air has been pictured as taking the form of 'chimneys' or of 'bubbles'. Glider pilots are familiar with the properties of warm ascending currents of air or 'thermals', as described by Yates (1953). In still air a glider sinks at about 90 cm. per sec. (about 20 times the terminal velocity of a pine pollen grain). On dull days thermals do not develop. They reach their maximum upward velocity of 3 metres per sec, or up to 25 metres per sec. in cloud, at midday in summer. Yates indicates that, depending on the type of soil, on Mind strength, and on sun height, a sizeable thermal is released from an area of 1-25 sq. km. every 5 to 15 minutes in summer. At a height of 300 metres, thermals may be 300 metres in diameter, though they are probably often smaller when lower, whereas at still greater heights a diameter nearer 2 km. was reported by Ludlam & Scorer (1953). Their vertical movement may cease at a temperature inversion, or may continue to from 10,000 to 50,000 ft. The temperature in a thermal appears to average i°-2°C. higher than the surrounding air through which it is ascending. The theory of thermals is still a matter of controversy, but there seems no doubt that air rises from some surfaces more often than from others. Green vegetation and wet soils may be relatively cool, while a ripe cereal crop, buildings, roads, or bare rock, may heat up rapidly in the sun and become the source of rising warm air. Thermals can also arise at a cold front, and glider pilots regard hilly country as the best source of thermals. The pattern by which cool air sinks to replace the ascending warm air is also a matter of debate. Downward draughts reported in the neigh- bourhood of thermal upcurrents appear to be comparatively feeble. Probably the downward movement occurs over a much wider area than 28 THE ATMOSPHERE AS AN ENVIRONMENT the thermal, as a slow sinking of the atmosphere. The sinking speed may be comparable with that of a fungus spore (Hirst, 1959), but the local rising velocity may commonly be 100 or more times this velocity. Some bird species soar in large thermals, as do dragonflies in smaller ones near the ground. Other birds haunt thermals to prey on the insects carried upwards (Scorer, 1954). The Stratosphere In this region, which extends upwards from the tropopause to the limit of the atmosphere, the temperature lapse-rate, characteristic of the troposphere, is zero or may even be reversed. The height of the tropo- pause varies with season, latitude, and other factors. The bottom of the stratosphere may be found at an altitude of about 10 km., though under special conditions it may reach temporarily to much nearer ground-level. The dust of the stratosphere is believed to be meteoric and to have entered the Earth's atmosphere from space. It is generally believed that terrestrial dust, including organic spores, is almost, if not entirely, con- fined to the troposphere — except for occasional incursions in air currents dragged up into the stratosphere by volcanic eruptions (or hydrogen bombs). However, recent studies of atmospheric circulation, discussed by Machta (1959), may point to exchange of air between troposphere and stratosphere — with rising air over the equator and descending air in middle latitudes. Circulation of the Atmosphere Under the influence of pressure difl^erences resulting from solar heating, and of friction between wind and the rotating earth, a general pattern of atmospheric circulation is set up. The surface winds sho\Mi in atlases are the ground-level part of a three-dimensional system that has not yet been well explored. The pattern still being worked out shows a complex circulation, with air over the Equator ascending and flowing discontinuously to the poles, which are themselves regions of generally subsiding air (Palmen, 195 1). Across the Equator there is relatively little interchange between northern and southern hemispheres. AIR MASSES The fact that air may have the same temperature and humidit}' over a wide geographical area has given rise to the concept of the discrete air mass, with properties different from adjacent air masses and separated from them by 'fronts'. When an air mass remains stationary for some time, it acquires a temperature and humidity dependent on the surface on which it rests. These characteristics will be retained for some time 29 THE MICROBIOLOGY OF THE ATMOSPHERE when the air mass moves into a new region. Air masses are classified, therefore, according to their origin, and we may have, for example: polar maritime, polar continental, tropical maritime, and tropical con- tinental types, as well as air masses of indeterminate origin (Belasco, 1952). The different air masses interest us, not only because they bring different kinds of weather but also because they might conceivably bring an air- spora characteristic of their place of origin. 30 IV SPORE LIBERATION Up to now we have considered only the physical properties of spores and of their environment. Spores, however, are parts of living organisms whose evolution has been extensively moulded by the environment. The air- spora comes mainly from species which are highly adapted towards using wind energy for their dispersal. The physical properties of the atmosphere make dispersal possible, but also set problems to organisms using it. Adaptations for wind transport have been evolved in many widely- separated taxonomic groups. The process of wind dispersal of spores has three principal stages, (i) Spore liberation. This chapter describes the processes by which pollen grains or spores 'take-off' into the air from the structure where they have been formed. (2) Dispersion. Chapter V describes the transport of spores by gentle air currents or strong winds, and the diffusion of an airborne spore cloud. (3) Deposition. Chapters VI and VII will deal with the pro- cesses by which spores leave the air and land on a surface — a necessary prelude to the germination of a pollen grain or a mould spore on its substratum. The spore or pollen output of many species is notoriously large. For instance, Pohl (1937) estimated, for the dominant species encountered by pollen analysts, the pollen production per stamen, flower, inflores- cence, and branch, revealing an annual production averaging many millions per square metre of ground covered {see also Erdtman, 1943). According to J. J. Christensen (1942), a field of wheat moderately affected by Puccinia graminis would produce at least 25 million uredospores per square metre. Duller (1909) estimated that one giant puff'ball {Calvatia gigantea) produced 7 million million (7 x 10^^) spores. The spore output of mosses and ferns is also potentially enormous. PROBLEMS OF 'tAKE-OFF' As described in Chapter III, the surface of the ground or plant is covered by a thin layer of still air and by the laminar boundary layer of slowly moving air ; a spore will fall through this composite zone under the influence of gravity. To tap the energy of moving air for dispersal a spore must overcome the adhesive forces which tend to keep it in contact with neighbouring spores or with the substratum. It must cross the still- and 31 THE MICROBIOLOGY OF THE ATMOSPHERE laminar-air layers, at the interface between the ground or other surface and the atmosphere, in order to enter the freely moving air of the tur- bulent boundary layer, where it stands a chance of being carried into higher layers of the troposphere. Many species that are distributed as spores have not solved this problem, but instead have become adapted for dispersal by some other agency such as water, insects, or other animals. There are more insect- pollinated (entomophilous) species of flowering plants than wind-pollin- ated (anemophilous) species, though in the temperate regions at least there are more wind-pollinated individuals because of the preponderance of grasses and anemophilous trees. We may wonder how important in practice is the occasional dispersal of a spore by some agency other than that to which it is adapted. However, it is a fundamental principle that the better a species is adapted to dispersal by one agency, the poorer are its chances of dispersal by another agency — unless, like many fungi, it produces spores of several distinct types that are specialized for different dispersal mechanisms. If we wish to control the dispersal process, a precise knowledge of the mechanisms involved is preferable to the vague idea that the spores will get there somehow anyway! Success in colonization or fertilization de- pends on logistics — on getting enough material to the right place at the right time. Energy is required to detach spores from their source. It may be an active process through which, by some explosive or hygroscopic mechan- ism, spores are discharged by energy operating through the parent struc- ture. Or it may be passive, by the energy of an external agent — usually wind or the kinetic energy of falling raindrops. Seasonal development of the parent structure and maturation of the spores commonly determine what organisms are in the air at a particular time, but other fiictors modify this pattern. The working of the various discharge mechanisms is more or less affected by external conditions, and the result is that the output of spores of a particular species varies greatly from time to time. Conversely, all the individuals of one species in an area may behave in unison, so that the composition of the air-spora differs vastly on different occasions. Take-off Mechanisms in Cryptogams etc. Spore- and pollen-liberation mechanisms have formed the subject of classical researches in biology for over a century. The wealth of informa- tion in the scattered literature on land plants is reviewed by Ingold (1939), and knowledge about bacteria by Wells (1955), but for protozoa and algae I know of no comprehensive treatment. In the present connec- tion we are concerned with those aspects of the mechanism which deter- mine when, and under what conditions, spores get into the air. 32 SPORE LIBERATION VIRUSES The viruses are little adapted to independent air dispersal. Some viruses infecting the animal respiratory tract are forced into the air on droplets during coughing and sneezing; but most bacterial and plant viruses, if they occur in the air at all, only get there on 'rafts' of debris or water droplets. Some of the so-called polyhedral viruses infecting insects are exceptional. Reports of outbreaks among pests of forest trees in eastern Europe speak of copious yellow deposits of the polyhedral bodies, shed by parasitized insects, which coat the surfaces of vehicles travelling through the forests. Study of air dispersal of viruses might explain some of the anomalies in the behaviour of insect viruses. To prevent contamination by an airborne infective particle of the dimensions of a virus may well require quite unusual experimental precautions. BACTERIA Moving air does not normally detach bacterial cells from the surface of a colony, at least when this is slimy, and in the absence of an active discharge mechanism natural processes capable of producing an aerosol of single bacterial cells are unknown. Mechanical disturbance of dust, clothing, surgical dressings, etc., however, carries into the air contamin- ated particles of substratum acting as 'rafts' and bearing clumps of bac- teria (Bourdillon & Colebrook, 1946). Rafts of soil or dust particles are raised by wind, by 'dust-devils' when the ground is heated by solar radiation, and by animal and human activity such as cultivation of bare ground. Rain splash, breakers, and sea spray continuously throw minute, potentially bacteria-laden, droplets into the atmosphere. Droplets expelled by coughing and sneezing are important indoors {and see p. 158), yet processes which put bacteria into the air are still not satisfactorily known. This is also true of the yeasts whose frequent abundance in the air remains unexplained, except for the Sporobolomycetaceae which show the ballistospore discharge mechanism (pp. 37-38). ACTINOMYCETES The mycelial organization of this group allows the Streptomycetaceae to develop aerial hyphae bearing &r^\ powder}'' spores — the first example of the sporophore elevation device, common in more elaborate organisms, for raising the spore-producing organ above the substratum and towards the moving layers of the atmosphere. Take-off conditions in the Actino- mycetes seem not to have been investigated. MYXOMYCETES (Mycetozoa, Myxogastrales) The slime-moulds are a group thoroughly adapted to wind dispersal. Some, such as Reticidaria^ merely expose a dr}'', powder)'' spore-mass on a cushion raised above the substratum. Others, such as Stemonitis and c 33 THE MICROBIOLOGY OF THE ATMOSPHERE Trichia^ expose small, dry spore-masses on stalks at most a few millimetres high. The spores are set free by twisting movements of hygroscopic elaters which take place with changes in air humidity (Ingold, 1939), or, in a few species, spores may be removed by eddies from shallow wind-cups. FUNGI Adaptations facilitating air dispersal show more diversity in the fungi than in any other group — except, perhaps, adaptations for seed dispersal among the flowering plants. They vary from the passive but quite effective processes in the Fungi Imperfecti, to the spectacular ballistic feats of the ascus gun. The various mechanisms have been summarized by Dobbs (1942, 1942^/) and Ingold (1953, i960), and they formed one of the main topics of the classical work of Buller (1909-50). In contrast, spores of many other species of fungi rarely get into the air but are carried by insects, on seeds, or in soil. Mere dispersal by in- sects may be relatively unimportant; but, where the insect actively inoculates the substratum or host, it is a mechanism comparable in efficiency with insect pollination of flowering plants. Passive liberation by the action of external energy depends on 'spore presentation' (Hirst, 1959)- (i) Shedding of spores under gravity. Stepanov (1935) concluded that spores of some CiinninghameUa species, and of some Fungi Imperfecti, including Botrytis cinerea, Monilia sitophila and Hehninthosporium sativum, as well as the macroconidia of Fusarium, could be shed under gravity. However, as he also showed that minor air currents released spores of some of these fungi, the effect remains uncertain. (ii) Shedding in convection currents. Stepanov (1935) placed open Petri-dish cultures at the bottom of glass cylinders 10 to 12 cm. high in which convection currents were induced by differential heating. Sticky slides or a surface of inverted sterile medium at the top of the cylinder trapped spores which might become detached and carried aloft by con- vection. With temperature differences of the order of io°C., conidia of Monilia sitophihi and Botrytis cinerea were freely transferred upwards, but CoUetotrichum Uni was not. Smaller temperature differences, such as resulted from the slight heat produced by a mould culture or an electric lamp shining on the floor, were ineffective. (iii) Blowing away ('deflation'). This occurs commonly with dry- spored fungi including moulds, smuts, and rust uredospores. The spores are often 'presented' on an elevated sporophore, any stem or leaf patho- gen usually being adequately raised on its host tissue. Qiiantitative studies so far are insufficient to lead to a theory of 'deflation'. Little is known about the quantitative effect of wind-speed on liberation, but there is good evidence that the higher the wind-speed the more spores are carried away. 34 SPORE LIBERATION Stepanov (1935) was apparently the first to use a small wind-tunnel to blow spores at controlled wind-speeds. Using either cultures or plants infected with pathogenic fungi, he found that the minimum wind-speed required to remove spores varied according to the organism being tested : for Botrytis cinerea it was o-36-o-5o metres per sec; for Alonilia sitophila^ Ustilago spp., uredospores of Piiccinia triticina, and Hehninthosporhim sativum it was 0-5 1-0-75 rnetres per sec; for aecidiospores of Puccmia coronifera and P. priugsheimia, o-'jb-i-o metres per sec; for Ciinning- hamella sp., 1-5-1 -75 metres per sec. On the other hand, Phytophthora hifestans and Fusarium cidnwrum spores were not removed at any speed tested up to 3-37 metres per sec More spores were removed in turbulent than in streamlined wind. A special structure facilitating blowing away is the 'wind-cup' described by Brodie & Gregory (1953). Flow of air over a cup-shaped structure produces a double eddy system which can effectively remove dry spores contained in the cup, as shown by wind-tunnel experiments with smoke and Lycopodium spores. Soredia were also removed from the podetia oiCladonia at 1-5 to 2-0 metres per sec, and spores were removed from the cupulate sporangia of certain Myxomycetes at 0-5 metres per sec. Certain Gasteromycetes, including the puffballs Lycoperdon perlatum and L. pyriforme^ and the earth-stars {Geaster spp.), have a 'bellows' mechanism consisting of a thin, flexible, waterproof wall covering the spore mass. Indenting this wall forces out a jet of air laden with spores. Contact with animals operates the bellows efficiently, but must be a relatively rare event in nature. Raindrops or run-off drops from trees also operate the bello^^ s mechanism, and as one fruit-body would be hit many thousands of times in a season, rain is probably the most effective mech- anism in the field (Gregory, 1949). In India W. H. Long & Ahmad (1947) find that the bellows mechanism of Tylostoma is operated by wind-blown sand grains in addition to raindrops. (iv) Mist pick-up. This is a mechanism that has only recently been recognized. Dry, or even humid, wind fails to detach spores of some fungi which are nevertheless readily removed from their conidiophores by collision with minute droplets carried by mist-laden air. This method is kno^\Tl to function with two important crop pathogens, Cercosporella herpotrichoides (Gl}Tme, 1953), and Verticillium albo-atrum (R. R. Davies, 1959), and it may play a part also in the dispersal of Cladosporium. (v) Splash dispersal. Spores of some species are 'presented' in sticky masses to which they adhere tenaciously in wind. However, spores may become incorporated in splash droplets (Plate 3, and Fig. 4) which are thrown up from the impact of a falling raindrop, or a drip from a leaf, hitting a liquid film containing spores (Gregory et al., 1959). Rain splash is thus another passive mechanism, quite different from mist pick-up by which slime-spored fungi may become airborne in the smaller droplets. Experiments suggest, however, that the larger splash-droplets, over 50 or 35 THE MICROBIOLOGY OF THE ATMOSPHERE 100 ju, in diameter, carry most of the spores which are dispersed in this fashion, and these droplets are massive enough to follow definite tra- jectories without being truly airborne. o TT-rrT-v-TT- -7-rrrrrrrTTTT- r = 0-002l r=0-007 2" t- = o / / / / ////// i / J I t=0-0035" r=o-ooq" '7~r77~7-7-T7~. r = 00007" rTy-m^TTTrr- \ =0-0055" t = 0-OI75" Fig. 4. — Splash from impact of water drop (5 mm. diameter) falling with velocity of 440 cm. per sec. on a thin film of water (drawn from stills from ultra- high-speed film made by Mr. E. D. Eyles at Kodak Research Laboratories, Harrow). (vi) The splash-cup mechanism. This is a device, studied particularly by Brodie (1951, 1957), which is widespread among lower as well as higher plants, by which the energy of falling raindrops throws relatively large bodies to distances of several feet. Examples are the peridioles of the birds-nest fungi (Nidulariaceae), the gemmae of Polyporus conchatus, and droplets bearing spermatozoids of the Bryophytes. However, the projectiles scarcely become airborne, for they follow a definite trajectory. (vii) Hygroscopic movements of conidiophores, which may result in detachment of spores during violent twisting, occur in a number of Fungi Imperfecti and Phycomycetes. The effect depends on rapid changes in atmospheric humidity and is often most marked in the morning hours. All active mechanisms for spore liberation depend on the fungus having sufficient water-supply. The more ephemeral fruit-bodies develop after rain and discharge spores for a short period only. More durable fungi can be dried but will discharge spores again when re- wetted; others again can draw on an extensive mycelium deep in the substratum and may be almost independent of the weather. Distances of ejection vary and have been compiled by Spector (1956, p. 153). (viii) Squirt-gun mechanism. This is found in many Ascomycetes in which the ascus, which contains the ascospores, typically swells at maturity 36 t = o sec t = O 0338 t ~ O- OOJ8 t ^ o-ofo44- t = O 0055 t = O- l52(o t = o ■ 0072 t = 0*970 t = O 0242 t = 02175 Photographs by Worthington & Cole (1897) showing splash of a water drop weighing 0-2 gm. (coated with lamp-black) falling 40 cm. into a mixture of milk and water. Mag. J X . Plate 3 SPORE LIBERATION and finally bursts at the tip, projecting the spores Into the air to a distance varying from a fraction of a millimetre to several centimetres. The larger the projectile, the further it tends to be shot (Ingold, 1956^; Ingold & Hadland, 1959). Four clearly-distinct t}-pes of liberation are recognized in the Asco- mycetes by Ingold (1953), as follows: 'i. In the Discomycete type the spore-producing surface, consisting of asci intermixed with parallel paraphyses, is more or less exposed, most often as a lining to a shallow cup-shaped apothecium. The extensive exposed hymenium allows opportunities for 'puffing' — the simultaneous bursting of numerous asci. '2. In the Pyrenornycete type the asci are contained in a small flask- shaped structure (perithecium) which opens to the outside by a minute ostiole. Before each ascus can discharge the spores, its tip must reach the ostiole, and the canal of the neck is usually so narrow that normally only one ascus can emerge at a time. '3. In the Erysiphales type the fruit-body is a cleistocarp. This is rather like a perithecium but is completely closed; there is no ostiole. In this t}'pe the swelling asci must first burst the wall of the cleistocarp before they can emerge and discharge their spores. '4. In the Myriangium t}'pe, though the hymenium is exposed in a structure like a small apothecium, the spherical asci are embedded in a plectenchymatous tissue and are free to discharge only when this gradually undergoes gelatinization.' Some Ascomycetes which lack explosive asci may liberate spores in slime to be dispersed by rain-splash. Other species, again, may be either explosively or slime-dispersed, according to the conditions obtaining. With still others, such as Chaetomium whose spores are regularly found in the air, the spore discharge mechanism is unknown. (ix) Squirting mechanisms^ which propel spores violently into the air, occur among Phycomycetes in Pilobolus^ Basidiobolus^ and Entomophthora muscae, as well as in the imperfect genus Nigrospora (Webster, 1952). (x) Roundifig-off of turgid cells acts as a discharge mechanism when the flattened double walls between two turgid cells suddenly separate. By this means spores of some Phycomycetes can be ejected up to a centimetre into the air. The same mechanism operates to eject aecidiospores when aecidia of rusts become moistened. Discharge of all these types is favoured by high humidities, and indeed aecidiospores of the rust fungi are dis- charged under conditions unlike those favouring dispersal of uredospores. (xi) Basidiospore discharge. This is a highly characteristic process which is found with the same essential features almost throughout the Basidiomycetes. The basidium is a cell producing one or more sterig- mata, at the end of each of which one basidiospore is formed asymmetri- cally. Typically, when the spore is mature, a drop of water is excreted at the hilum end of the spore and almost immediately the spore is shot off 37 THE MICROBIOLOGY OF THE ATMOSPHERE to a distance of a fraction of a millimetre or more. In species which form the basidia on exposed surfaces, as in many lower Basidiomycetes, the spore after discharge has a chance of being picked up by an air current. The higher Basidiomycetes often show great elaboration of a stalked fruit-body with the basidia lining the vertical surfaces of folds, gills, pores or spines. Here, in cavities protected from wind and adverse conditions, the basldiospores are discharged into still air and fall under the influence of gravity into the moving air-current below the cap-shaped or bracket-shaped fruit-body. Spore discharge in the higher Basidiomycetes often goes on continuously throughout almost the entire life of the fruit- body — to all appearances little affected by wind, temperature, or humidity, though it must be emphasized that accurate quantitative studies on the effects of these factors are lacking. Just how a basidiospore is shot off" the sterigma remains a major puzzle of mycology; several explanations have been advanced, but none seems entirely satisfactory. Nevertheless the process is highly efficient and basidiospores are a conspicuous com- ponent of the air-spora. The same mechanism occurs in the mirror-yeasts (Sporobolomy- cetaceae), which may possibly have evolved from lower Basidiomycetes (unlike the Saccharomycetaceae, which are clearly Ascomycetes). To avoid prejudging the issue by calling the spores of the mirror-yeasts 'basidiospores', the term 'ballistospores' has been coined to include all spores showing the drop-excretion discharge mechanism. A moist sub- stratum is necessary for spore discharge in the Sporobolomycetaceae. LICHENS The fungus component of lichens discharges ascospores from typical apothecia or pcrithecia, or basidiospores from basidia. Fragments of the thallus including both fungal and algal components are blown about freely. Rounded groups of algal cells surrounded by fungal hyphae, separating off" from the lichen thallus as soredia, are also blown away; but we know little as yet about the relative importance of these various modes of reproduction. ALGAE Adaptations facilitating take-off" into the air are unknown in the algae, though some of the simpler types of algal cells get into the air regularly. Pettersson (1940) suggested that Chlamydomonas nivalis is carried away from its habitat on snow-fields and glaciers in melt-water and becomes airborne by splash in mountain torrents. Lichen soredia may possibly aid the dispersal of algae when they become grounded in a habitat moist enough for the algal component to dominate the fungus, and where the resulting colony will be an alga instead of a lichen. Again, some terrestrial and epiphytic algae may crumble and blow away. 38 SPORE LIBERATION BRYOPHYTES Spores of mosses and liverworts are formed in sporangia which are typically raised on stalks above the substratum, but the structure of the sporangium is quite different in the two groups. The moss sporangium is a firm 'box' opening at the top, whereas the liverwort sporangium breaks open completely, exposing the spores in a mass of stiff threads (elaters). In the simpler liverworts the spores may be blown away by wind from the mass of elaters, or the elaters may twist hygroscopically, actively throwing spores into the air. In most leafy liverworts, however, a spring mechanism released by water-rupture in the dr\'ing elaters throws the spores into the air (Ingold, 1939, 1956), while in FruUania the sporangium explodes by an efficient spiral spring mechanism which also is released on drying. The mosses liberate spores from the stalked sporangium (capsule) by two principal methods. Sphagnum has an 'air-gun' mechanism (Ingold, 1939). An air space below the spore mass is compressed by transverse contraction of the dr}'ing sporangium wall, internal pressure increases and, finally, the top of the sporangium breaks, ejecting a spore cloud to a height of 15 or more centimetres. Most of the other mosses have flask-shaped sporangia, which open gently at the top when mature. In some genera the mouth of the spor- angium is surrounded by one or more rows of triangular teeth which move hygroscopically, closing the mouth at high humidities. To what extent spore liberation in nature depends on shaking of the sporangium in the wind, and what role is played by hygroscopic movements of the teeth in actively throwing out the spores, is still a matter of controversy; but evidently spore liberation is checked by high humidities and low wind-speeds. PTERIDOPHYTES Spores of Pteridophytes (ferns and their allies) are formed on the fronds m ithin a closed sporangium, from which they are dispersed into the air by a 'sling' mechanism depending on water-rupture under great tension as the maturing sporangial wall dries {see Ingold, 1939). Pettersson (1940), in Finland, found that effective scattering of fern spores takes place out- of-doors only when the relative humidity of the air falls to 76 per cent, or even to 60 per cent, according to the species. Pollination of Phanerogams Insects and wind are the chief agents of cross-pollination in flowering plants. Other pollinating agents that are effective in a far smaller number of species include water and humming-birds. There are probably ten times as many entomophilous (insect-pollinated) as anemophilous (wind-pollinated) species of flowering plants in the world as a whole. 39 THE MICROBIOLOGY OF THE ATMOSPHERE The characteristics of wind-borne pollen become clear when contrasted with insect-borne pollen (Table IV). There are many exceptions to the generalizations in this table and, in particular, some plants make the best of both methods. Both anemophilous and entomophilous plants often protect their pollen from the rain, and many store it within the flower for some time after shedding from the anthers. Anemophilous pollen is not generally shed into very calm or very damp air. TABLE IV TYPICAL CHARACTERISTICS OF ANEMOPHILOUS AND ENTOMOPHILOUS PLANTS Wind-pollinated Flowers Lack conspicuous and attractive petals, scent, and nectar. Flower Projecting into air: hanging from position bare branches before leaves open (catkins); on erect stalks (grasses, etc.); or at ends of branches (coni- fers). Prevention Male and female organs often in sep- of self- arate flowers or inflorescences, or on fertilization separate plants. If flowers herma- phrodite, one sex commonly matures before the other, or, if sexes are in separate inflorescences, the female is often above the male. Pollen Often shed into the air in vast quan- tities. Shape rounded, often nearly spheri- cal or ellipsoidal. Size-range narrower than entomo- philous pollen and seldom less than 15 M- Surface typically smooth as seen under the microscope, non-sticky, easily separating into single grains in Insect-pollinated Often with bright colours, scent; nectar attractive to in- sects. Tend to be exposed to view, but not exposing anthers to wind. Flowers usually maturing when plant in full growth and insects abundant. Flowers usually hermaphrodite, with structural or genetic bar- riers to selfing. Usually restricted pollen pro- duction with little shedding. Shape very variable. Size very variable, 3 to 250 /n, but often less than 15/11. Surface typically rough, spiny or war ted, often oily or sticky, tending to adhere in clumps. GYMNOSPERMS Conifer pollen, instead of being formed in stalked anthers as is that of Angiosperms, is produced in two or more pollen sacs on the lower side of the male cone-scales. The pollen grains are large and often bear two conspicuous air-filled bladders which decrease the density of the particle and so retard its fall under the influence of gravity. In PmuSy cone-scales of the erect male cone separate as they mature, and pollen shed from the paired sacs falls into small hollows on the upper surface of the cone-scale below. From these hollows the pollen is blown away when the wind reaches sufficient velocity. Some other conifers have hygroscopic mechanisms protecting their pollen from rain and allowing 40 SPORE LIBERATION its release only in dty weather. In Taxus, Thuja, Cupressus, and ^uniperus, the pollen is not winged. In ^uniperus the expanded ends of the cone- scales interlock closely in damp weather, separating again in dry air and allowing pollen to be blown out. ANGIOSPERMS Details of flowering-plant pollination mechanisms are given by Mari- laun (1895), Knuth {1906), Erdtman (1943, 1952, 1957), Wodehouse (1945) and others. Fig. 5. — Anthesis of false oat-grass {Arrhenatherum elatius): (i) closed anther; (2) open anther; (3) spikelets on a calm day; and (4) spikelets in a wind. (Repro- duced from Marilaun's: Natural History of Plants, by permission of Messrs. Blackie & Son, Limited.) (i) Grasses, rushes, sedges and their allies. The Gramineae, Cyperaceae, Typhaceae, and Juncaceae are typically wind-pollinated. From the raised inflorescences of grasses, the anthers are extruded on long filaments to which they are so lightly attached that they vibrate in the slightest wind. Often, as in Arrhenatherum, the end of each pollen-sac bends up (Fig. 5), forming a spoon into which pollen is shed from a slit, and where it ac- cumulates until blown away by the wind. Either damp or ver>- dry weather 41 THE MICROBIOLOGY OF THE ATMOSPHERE may delay both extrusion of the stamens and splitting of the anthers. Except for rye and maize, most cultivated cereals are self-fertilized and shed little pollen, but pasture grasses are free shedders. In central Europe, Marilaun (1895) found that different grasses flowered for brief periods of only 1 5 to 20 minutes daily, and at charac- teristic times of the day : hours 04-05 Poa, Koeleria, Avena elatior. 05-06 Briza, Deschainpsia caespitosa, Triticiim, Honleum. 06-07 Secale, Dactylis, Andropogon^ Brachypodiian, {Bromiis?), Festtica spp., Hole us (ist anthesis). 07-08 Trisetiiin, Alopecurus, Phkiini, Anthoxanthum. 08-09 Exotic types in Europe: Paniciiiii, Sorghum. 09-10 Set aria italica, Gynerium (Cortaderia) argenteum. 11-12 Agrostis spp. 12-13 Melica, Molinia, Nardiis, Elymus, Sclerochloa, some Calamagrostis spp. 14 A few B ramus spp. 15 A few Arena spp. 16 Agropyrum. 17-18 Deschampsia flexuosa . 19 Hokus (2nd anthesis). This timetable does not necessarily apply elsewhere, and, in Nebraska, Jones & Newell (1946) found a less precise timing and showed that anthesis is determined by temperature. They distinguish cool-season from warm-season grasses. Cool-season grasses include: Fcstuca elatior (anthesis at 13.30-15.00 hours); Agropyrum spp. (14.00-18.30 hours); Bromus inermis (14.30-19.00 hours); Poa pratensis (during the night); Secale cereale (02.30-11.30 hours, maximum 06.00-08.30 hours). The warm-season group includes: Boiitcloua gracilis (03.00-09.00 hours, maxi- mum 04.30-05.30 hours during darkness) ; ^m^A/oV dactyloides (06.30-13.00 hours, maximum 07.00-08.30 hours); Paniciini virgatum (10.00-12.00 hours, delayed in cool season); Zea mays (07.30-16.00 hours, maximum 08.30-11.00 hours). Hyde & Williams (1945, p. 89), from the cooler climate of Wales, report both discrepancies and agreements with Marilaun's timetable: Holcus lanatus (04.00-06.00 hours, but mainly at 14.00-19.00 hours); Cynosurus cristatus (05.00-06.00 hours); Arrhenatheriim (07.00-08.00 hours); Trisetum flavescens (before 08.00 hours); Festuca pratensis (08.00- 14.00 hours). (ii) Aquatic monocotykdonous herbs include a few other wind-pollinated plants, for example Triglochin and Sparganimn, while in the genus Potamogeton some species are pollinated by wind and others by water. (iii) Entomophilous herbs and low shrubs include some species in which the phase of insect visitation is followed by an opportunity for wind- pollination, the anthers first shedding pollen within the corolla; but as the flower matures, the elongating filaments protrude and scatter pollen in the wind. These types include the semi-parasites Bartsia and Lathraea 42 SPORE LIBERATION (Rhinanthaceae), and the heaths Calluna and Erica — but not Rhododendron^ which has very sticky pollen. (iv) Tropical and sub-tropical trees include few anemophilous species, but Casuarina and Myrothamnus are wind-pollinated, and some of the palms, although entomophilous, shed a good deal of pollen, which may be carried by the wind. (v) Nettles and their allies form an anemophilous group which do not store pollen after dehiscence of the anthers. The anthers dry as they mature, tissue tensions are set up, and suddenly, as the pollen sacs burst, the filaments uncoil, throwing pollen into the air. The process can be watched on a still, dry day when small puffs of pollen appear as the nettle flowers explode, but in damp air dehiscence of the anthers is inhibited. The mechanism occurs in Urtica, Parietaria, Moms, and Broussonetia. A sifter mechanism similar to that of the grasses occurs in Cannabis and Humiilus. (vi) Herbs with inflorescences elevated above the general level of the foliage include a number of anemophilous types such as Mercurialis. In Plantago and Globularia the anthers, w^hich are exposed in cups, close their slits in moist weather but shed their pollen in dry air. Upward- facing cups occur also in Poterium and Sanguisorba. Sifter mechanisms occur in some species oi Rimiex and Thalictrum. Other conspicuous pollen shedders occur in the Chenopodiaceae {Beta, Salsola, Chenopodium) and in the Amaranthaceae, and also in some groups within the Compositae — especially Ambrosia and Artemisia. (vii) Deciduous trees of temperate regions form a biological group. Typically the male flowers are aggregated into pendulous catkins, usually appearing shortly before the leaves expand. In Alnus, Betula, Castanea, Corylus, Fagus,Juglans, Populus {Salix, like Tilia, is both insect-pollinated and a wind shedder), and Qtiercus, pollen is protected from rain after shedding while temporarily stored on the upper scales of the flower standing underneath — until it is blown away by w ind in a manner reminis- cent of Pi?ius. Platanus closes its catkins by a hygroscopic mechanism, so that pollen is not merely protected from rain but can be blown away only in dry weather. Hippophae pollen is shed into the base of the flower while this is still in bud. At maturity the perianth lobes remain united at the top but separate at the base, leaving slits through which pollen can be removed by the wind. In another group, including Fraxinus, Buxus, Phillyrea and Ulmus, the anthers project as upward-facing cups from which pollen is removed by wind. The take-off mechanisms briefly sketched in this Chapter, with others (doubtless including some still undiscovered), are not mere curiosities of natural history. On the contrary, they are highly efficient processes that restrict spore liberation to limited meteorological conditions. Pollen and spores of mosses and ferns tend to be shed into dry winds. Ascomycetes 43 THE MICROBIOLOGY OF THE ATMOSPHERE and lower Basidiomycetes are more likely to discharge spores when the substratum is wet. Spore shedding in higher Basidiomycetes is less affected by air humidity and wind speed. Spores of some Fungi Imperfecti may depend on wind for removal, or on changes in humidity for hygro- scopic movements, or on rain for splash dispersal. Soil- and dust-borne bacteria and protozoa are probably borne aloft in high winds from heated or mechanically disturbed ground. The nature of the 'take-off' mech- anism profoundly affects the occurrence of different kinds of spores or pollens in the atmosphere — with consequent significance for hay- fever patients, seed-crops, plant diseases, evolution, and geographical distribution. 44 V HORIZONTAL DIFFUSION We have now described the particles composing the air-spora and the relevant properties of the atmosphere. What happens to the particles after they have been launched into the atmosphere? Common-sense tells us that they become dispersed — in the sense that their concentration per unit volume of air decreases with increasing distance from the point of liberation. Tyndall (1881) believed that airborne microbes float through the atmosphere in miniature clouds. He explained Pasteur's demonstration of non-continuity in the spontaneous generation controversy by postulat- ing that Pasteur sometimes opened his flask in the midst of a bacterial cloud and obtained life, and sometimes in the interspace between two clouds and obtained no life. In hospital practice, opening a wound during the passage of a bacterial cloud would have an effect very different from opening it in an interspace between clouds. It was not necessary to draw this conclusion, however, as Pasteur's results could be explained equally well if microbes were randomly dis- tributed in the air. Evidence for random distribution was obtained by Home (1935), who applied Fisher's x^ test to catches on 1,000 or more Petri dishes of sterile media which had been exposed in Kentish orchards by N. W. Nitimargi. The observed frequencies of total bacteria or total moulds, or of any genera or species tested separately, did not depart significantly from the Poisson distribution. Home concluded that micro- organisms are distributed at random in the air, and that, for making valid comparisons between populations of airbome microbes at different places and times, analysis of variance could legitimately be applied to plate counts. Dispersion of the Spore-cloud It is still convenient to speak of clouds of spores — not, indeed, keeping together in the manner of locust swarms, but tending to become dispersed while suspended passively in the atmosphere. Sampling a region small enough in relation to the size of the cloud may then reveal a random distribution of particles. Dispersion of the spore-cloud can be deduced from early observations on the distribution of rust on rye by Windt (1806), who observed that 45 THE MICROBIOLOGY OF THE ATMOSPHERE rust was severe near barberry bushes which are now known to be the akernate hosts for the fungus: 'the effects are striking and desolating in the distance of ten to twelve paces, I have also perceived them visibly at 50, 100, 150 paces and a final attack at above 1,000 paces.' Similarly, dispersion of the pollen cloud made it possible for Blackley (1873) to advise his hay-fever patients to keep away from grass fields during the flowering season of the grasses. Attempts to formulate the process of spore dispersion through the atmosphere have been based on geometrical, empirical, or meteorological considerations. The geometrical approach is suggested by analogy with the laws of radiation. Niigeli (1877) stated that the amount of dust which comes on an air current from one place falls off with the inverse square of the distance, whereas E. Fischer & Gaumann (1929) stated that, with linear increase of the distance, the chance of infection by rust spores decreases in cubic progression. Kursanov (1933) stated that, in the absence of wind, the number of fungus spores would fall off inversely as the cube of the dis- tance from the source. The ideas that underlie the geometrical approach are simple. Spores travel away from the point of liberation: at greater distances the volume of air which they can occupy increases as the cube of the distance or, alternatively, the surface of the ground on which they could fall increases as the square of the distance. A third possibility would be a simple inverse relationship with distance, as the areas of successive annuli around a point increase in arithmetical progression. The geometrical method is unsatisfactory because, although in a general way it illustrates the features of dispersion, it is not clear why spores should travel and spread out in the manner predicted. The particles interesting to us here are passively borne and do not behave like radiations, because the air which carries them is not in process of being continuously generated at some point in the atmosphere; consequently some totally different concept is needed. The approach by empirical curve-fitting has been based on field records of dispersal gradients, such as the scatter of seeds or seedlings on the ground, contamination of seed crops by foreign pollen, or the incidence of plant diseases. Using such data, a curve is fitted to the ob- served points, either graphically or by the statistical method of least squares, and an attempt is made to find an empirical formula to fit the curve. These methods will be referred to in detail in Chapter XIII, after the subject of spore deposition has been discussed. In general the empirical method has the advantage that an equation can usually be obtained, containing at most three parameters, which gives a good fit to any one set of field data. On the other hand, it is difficult to compare results obtained by different workers. The parameters are calculated from the data and correspond to no obvious natural phenomena ; consequently it is difficult to use empirical formulae to predict a dispersal pattern under conditions differing from the original one. 46 HORIZONTAL DIFFUSION In the long run a more ambitious approach seems essential, with the aim of developing formulae whose parameters correspond to factors of the environment, and which take into account the total number of microbes liberated (if known), allow for variations in weather, and use a standard unit of distance. Diffusion as a Result of Atmospheric Turblt^ence Watching the drift of smoke from a bonfire or factory will convince the observer that wind, instead of having a steady streaming motion, is characteristically turbulent as described in Chapter III. According to Brunt (1934), large numbers of small-scale eddies, whose periods are of the order of i second, are usually present in the turbulent boundary layer, and at least two-thirds of the eddying energy is associated with eddies of less than 5 seconds. The action of these very numerous eddies of varying size on the very numerous spores produced from plant sources, makes some regularit}' in the dispersal pattern possible. The study of eddy diffusion has proved difhcult, but it provides the most promising approach to the elucidation of dispersal. Before describing the methods in detail, a few general notions — familiar to physicists, but mostly unfamiliar to biologists — must be introduced. We are attempting to discover laws governing spore diffusion in the atmosphere. In nature this is often a complex process, as there are obstacles preventing the free flow of air. We therefore use a device familiar to physicists — making a simplified model in the hope that, if we can under- stand the process of diffusion under simple conditions, we shall be able to attack the more complex situations found in nature. The assumptions we have to make for a simplified model are as follows. (i) The field. Diffusion is assumed to be taking place in three dimen- sions in the atmosphere over a plane surface which is of indefinite extent, free from topographical irregularities — not necessarily 'smooth', but, if aerod}'namically rough, then uniformly so. (ii) Co-ordinates. To describe movement over the plane surface we need a system of co-ordinates. Their origin, 'O', is conveniently taken to be the point of liberation of the pollen or spores. The 'x'-axis is horizontal and positive in the down-wind direction, and the 'y'-axis is also hori- zontal but at right-angles to the direction of the wind. Lengths above and below the origin are measured on the vertical 'z'-axis. (iii) Sources. Particles are liberated from a source. The simplest form of source is a 'point source', and this may either liberate a number 'Q^ of spores at a single instant (an 'instantaneous point source'), or it may be a 'continuous point source' emitting Oospores per second. Instead of a point source we may have a 'line source'. For simplicity we assume that the line is horizontal, and is emitting Oospores per centi- metre of its (effectively) infinite length. The line source in turn may be 47 THE MICROBIOLOGY OF THE ATMOSPHERE instantaneous or continuous. Furthermore, we may have an 'area source' (emitting Q_ spores per square centimetre), or a 'volume source'. Real sources in the field that correspond approximately to these ideal sources would be a single plant (point), a hedge (line), a ground crop (area), and an orchard or forest stand (volume). The dimensions of the source must be treated as relative to their distance: thus a field would be regarded as effectively a point source when considered from distances many times its own width. Fig. 6. — Diffusion of spore-cloud during horizontal travel in wind. O = origin of co-ordinates at source of liberation; x, y, z= down-wind, cross-wind, and vertical axes, respectively. Growth of cloud is measured by increase in standard deviation after the centre of the cloud has travelled to three positions down-wind. All these sources may be instantaneous or continuous. The cloud from an instantaneous point source is a puff or spherical cloud, whereas the conical cloud arising from a continuous point source is familiar in the smoke plume from a chimney. A continuous point source can be viewed as made up of a succession of overlapping instantaneous emissions. (iv) Standard deviation. Suppose that a 'puff' of spores has been liber- ated at an instant from a point source into a wind and has become subject to the action of atmospheric eddies which move individual spores apart at random. After a short time the particles composing the cloud will show a scatter around their origin (Fig. 6). At any instant such a cloud has two characteristics which we could compute if we had all the data: (i) the mean position of the particles, i.e. the centre of the cloud, which can be expressed as a point on the system of x, y, and z ordinates ; and (2) the standard deviation, o-, of the particles from their mean position. When the 48 HORIZONTAL DIFFUSION cloud has travelled farther down-wind it will have a new mean position, and during the time the cloud has been travelling it will have been further diluted by eddies, its particles will have got farther apart, and consequently their standard deviation will have become larger. The next problem is to find a relation between the standard deviation and the distance travelled. How does a grow as x grows ? This is a prob- lem that has excited the interest of many workers since the First World War, who were attempting to predict the concentration of gas clouds, smoke screens, smoke trails, and crop pathogens. The pioneer in the subject was the Austrian meteorologist, Wilhelm Schmidt (191 8, 1925), who put forward a theory similar to those being developed almost simultaneously in Britain by G. I. Taylor and L. F. Richardson. Schmidt supposed that, with a given state of turbulence of the air, dijffusion of particles proceeds like the diffusion of heat in a solid, but with an atmospheric turbulence coefficient A/p replacing the coeffi- cient of thermal conductivity. He showed that for these conditions CT^ = 2(A//3)t, where t = time. His work is now mainly of historical interest, but we should note one interesting feature : according to Schmidt the standard deviation squared is proportional to the time during which diffiasion has been taking place, so that on his theory the standard deviation will not be constant at a given distance, but will depend on the time taken to reach that distance, i.e. upon the speed of the wind. Schmidt also assumed that the particles in the diffijsing cloud are brought to ground-level by their fall under gravity, and he used measured values of terminal velocity to fix dispersal limits for various organisms (cf. Table XXVII). Sutton (1932) recognized that diffiasion in the atmosphere diffisrs from molecular diffusion of heat in a solid in one important respect. Diffusion in a solid is constant (depending on the mean free path of the molecules) however long the diffusion has been going on. Diffusion in the atmosphere is much more complex, because atmospheric eddies are of a vast range of sizes, varying from a centimetre or so up to eddies that we recognize as fluctuations in wind direction, and even to cyclones and anti- cyclones. Sutton realized that the size of eddy effective at a given moment in diluting a cloud is of the same order as the size of the cloud itself at that moment. Thus a i-cm. eddy would not effectively dilute a cloud i-metre in diameter, and a 1,000-metre eddy would merely carry a i-metre cloud around bodily without diluting it. The eddy that dilutes a i-metre cloud is itself of the order of i metre. This led Sutton to an equation for the standard deviation which is fundamentally different from that of Schmidt : a2 = iC2(ut)"\ where t = time; u = wind-speed; 'C is a new coefficient of diffusion D 49 THE MICROBIOLOGY OF THE ATMOSPHERE with dimensions (L)^; and 'w' is a number varying between 1-24 in extremely stable, non-turbulent wind, and 2-0 under conditions of extreme turbulence. The value for normal overcast conditions with a steady wind is m = 1-75. Because wind-speed multiplied by time equals distance we can write Sutton's formula : a^ = IC^x"'. This suggestion that a^ is a function of the distance, x, is not unreasonable, because the surface roughnesses which generate eddies are spread out along the distance travelled by the cloud. It is moreover a tempting theory, because we need not know the wind- speed under which dispersal takes place. Values of C decrease with height because conditions at great heights are unfavourable for the formation of eddies. Values for m appear to increase with longer sampling periods, and Sutton suggests that m itself is a function of time. In making continuous observations over a long period on the density of a cloud, he suggests that the random element may become smoothed out, so that, over a sufficiently long period, m = 2. These possibilities should be borne in mind when the density formulae described below are applied to some biological data where the sampling period is very long. In practice it is found that near ground-level, diffusion takes place faster on the x- and y-axes than on the vertical z-axis. Turbulence is then said to be 'non-isotropic', and C has to be represented by its com- ponents : Cx, Cy, and Q. The number m is an indicator of the degree of turbulence of the air and is, as a first approximation, independent of the mean wind-velocity. It is primarily affected only by those factors which tend to damp out or increase turbulence, such as the vertical temperature gradient and the roughness of the ground. For conditions of spore dispersal tests it seems appropriate to assume values of Cy = 0-5-1 -o (metre)^, Q = o- 1-0-2 (metre)", and m = 1-75-2-0. Expressions for the concentration of particles in a cloud emitted from various types of source were deduced by Sutton (1932), and are analogous to heat-conduction equations, as follows : (i) An instantaneous point source^ such as a puff of Q^grams of smoke, or a number Q^of spores emitted at an instant of time. Here the concen- tration in the cloud is given by Q_ r r2 ^ 7rtC3xt"'^''P\ CH' where 'r' = distance from the centre of the puff or cloud. (ii) A continuous point source^ such as a factory chimney emitting Q_ particles per second. Here, to obtain an integral that can be handled conveniently, the assumption is made that the spread of the cloud laterally and vertically is small compared with its spread down-wind. When 50 HORIZONTAL DIFFUSION emission has continued long enough for the distribution to reach a steady state, the concentration is given approximately by The cross-wind concentration shows a 'normal' distribution of particles. On the axis of the cloud (y = z = O) the concentration is given by the simpler expression ^ ttOux'" and because, according to the theory, m cannot exceed 2-0, the fall-off in concentration on the axis of a point-source cloud cannot be more rapid than the inverse square, no matter how turbulent the wind may be. (iii) A continuous line source at right-angles to the mean direction of the wind, emitting (^particles per second per centimetre, and assuming the line to be of infinite lencfth a \ z ,2 ^ V(^)Cuxi™^''P\ C^x" Values obtained by Sutton suggest that, as a rough and ready rule, a finite line source behaves as a line of infinite length for distances of travel of the cloud up to 4 times the actual length of the line. For points on the xOy plane g ^ V(^)Cux^'« Sutton's statistical method does not exhaust the possible approaches to the problem of atmospheric diffusion, and attempts to find a still more useful model continue {see H. L. Green & Lane, 1957). From the theory of T. von Karman, Calder (1952) developed an equation which is said to give better predictions than Sutton's theory up to distances of 100 metres, but not for greater distances. It is also difficult to apply Calder's equations except to point sources. Another theory of dispersion, based on fluctuations in wind direction, is outlined by Sheldon & Hewson (1958), and a recent theory by Clarenburg (i960) will have to be taken into account. Field Experiments on Diffusion of Spore-clouds Several experiments have now been reported that give data from which it is possible to test the applicability of eddy diffusion theories to diffusion of the spore-cloud in a horizontal direction. Stepanov (1935) used artificial sources of spores that were liberated at a point in the open air. He trapped the spores on glass slides, coated with glycerine jelly, placed on the ground at various distances from the source and in various directions relative to the wind. At the end of the experiment cover-glasses were placed on the slides, and the number of single spores per unit area was counted (spore clusters were disregarded). 51 THE MICROBIOLOGY OF THE ATMOSPHERE In Experiment i (28 July 1933), on a lawn near the Middle Neva River, Elagin Island, Leningrad, approximately 1-2 X 10^ spores of Tilletia caries were disseminated into the air through gauze, at a height of about 80 to 120 cm. above the ground. According to anemometer readings the wind varied from 0-5 to 4-0 metres per sec, but sometimes fell to a complete calm; its direction was also variable. Two glass slides were placed at each trapping position, the numbers of spores trapped being shown in Table V. TABLE V RESULTS OF DISPERSAL OF SPORES OF Tilletia caries, Experiment i (Stepanov, 1935) Number of spores per cover-glass 18 X 18 mm. (average of 2) Angle of slide to wind -20° — 10° + 30° + 45° + 55° + 65° + 75° + 85° At 5 metres from place of dispersal of spores 204 435 964 1 198 659 341 365 20 At 10 metres from place of dispersal of spores 23 45 212 587 123 24 5 ID At 15 metres from place of dispersal of spores 4 19 207 87 77 26 26 9 At 20 metres from place of dispersal of spores 49 142 15 7 53 14 Experiment 2 (5 September 1933) was made at the same place as the previous one. This time a mixture of spores of Tilletia caries and Bovista plumbea was disseminated through a small sieve at a height of about 150 cm. Scattering of the spores occupied 15 minutes, after which 30 to 35 minutes were (perhaps unnecessarily) allowed to elapse for the deposi- tion of the spores. During this period the wind mostly varied from 2-3 to 3-0 metres per sec, but was sometimes calm. As shown in Table VI, three slides were placed at each trapping position. Approximately i-8 X 10" spores of Tilletia were used but those of Bovista were unfortunately not estimated. Stepanov's results led him to an empirical law of spore dispersal which was expressed as : y = C + a/sx, where 'y' = the distance at which the spores were trapped, x = the number of spores deposited per unit area of trap surface, s = area of trap surface, and 'C and 'a' are parameters dependent on the conditions of the experiment. The number of spores deposited is thus regarded as varying inversely as the first power of the distance from an origin of co-ordinates that is not coincident with the source. It will be shown later that Stepanov's formula, which is the first- fruits of the experimental approach to the problem, needs modification 52 HORIZONTAL DIFFUSION if it is to describe spore dispersal over a wide range of conditions (Gregory, 1945). First it will be necessary to re-examine Stepanov's results in the light of present knowledge of eddy diffusion. TABLE VI DISPERSAL OF MIXED SPORES OF Tilktia carks AND Bovistci pliimhea. Experiment 2 (Stepanov, 1935) Number of spores per cover-glass 18 x 18 mm. (average of 3) Tilktia Bovista Angle 5 m. 10 m. 20 m. 40 m. 5 m. 10 m. 20 m. 40 n -45° 3-0 0-3 0-7 00 0-3 0-0 0-0 0-0 -30° 128-0 2-3 0-3 0-0 7-0 0-3 0-0 0-0 -15° 43-3 547 47 0-3 7-0 4-0 0-0 0-0 0° 2o6-o 204-0 5-3 8-3 17-0 0-0 1-7 0-0 + 15° 6230 115-3 31-3 1-3 46-0 160 0-7 0-0 + 30° 8777 216-7 49-0 7-0 8i-3 20-3 6-3 0-0 + 45° 9II7 89-7 207-0 93 70-0 10-7 7-3 2-7 + 60° 2457 48-0 3-0 2-3 27-7 17-3 0-7 0-0 Stepanov's observational data enable us to test whether the standard deviation, o-, of the spores from their mean position agrees with Sutton's form: o^ = |Ox"', or with the older diffusion theories where a- = iKt. The data also allow us to estimate the parameters m and C, which can then be compared with values obtained by meteorologists for similar conditions. Examination of Tables V and VI shows that the spores at any one distance do not lie in a smooth normal frequency distribution, but are significantly clumped. This is probably because the duration of the dis- persal operation was insufficient to smooth out the action of a few large- scale eddies. The standard deviations of spores lying at each distance from the source have been calculated for Table VII, where for convenience the deviations from the mean position at each distance were measured along the arc with the point source as centre. The standard deviation at each distance was calculated from the usual formula, a = -\/[(x — x)"/(n — i)]. This is not strictly legitimate, because the trapped spores are a systematic instead of a random sample of the population and should be regarded as estimates of the ordinate of a normal frequency curve. However, the formula clearly gives a useful approximation — which would have been better if the traps had extended farther laterally and if data for some of the intermediate radii had not been missing. Both experiments were done in the same place, and with comparable wind velocities, and when values for log a are plotted against log x the points are found to lie reasonably close to a straight line. The slope of this line is not unity, as it would have been with the older diffusion theories, 53 THE MICROBIOLOGY OF THE ATMOSPHERE but corresponds with Sutton's formula for o-, where C = 0-64 (metre)*, and m = 1-76. Sutton's work was apparently unknown to Stepanov when these experiments were done, and so the data could not at the time be analysed in terms of eddy diffusion. However, the agreement between experiment and theory provides evidence that spore dispersal in air is mainly controlled by eddy diffusion of the t^'pe postulated by Sutton {see Table VII). The values for C and m obtained from Stepanov's experiments agree well with those found in spore dispersal tests by other workers (Table VIII), and with comparable data obtained by Richardson (1920) for dispersal of smoke from a point source over distances of tens of metres, where C = o-6 (metre)«, and m == 1-75. TABLE VII CALCULATION OF PARAMETERS FOR SUTTON's DIFFUSION EQUATION FROM stepanov's data (1935) Expt. No. Distance in metres 5 10 15 20 40 I Tilletia 2-25 2-97 5-03 6-64 — log 0-3522 0-4728 0-7016 0-8222 — 2 Tilletia log 1-81 0-2577 3-62 0-5587 — 4-87 0-6875 14-79 1-1699 Bovista log 1-790 0-2529 3-673 0-5651 — 5-343 0-7277 — log distance (metre) 0-6990 I -0000 1-1761 1-3010 I -6021 Equation for regression line : a = 0-5971 (S.D. 002), b ^- Whence C = 0-637 (nietre)^ o-88i2 (S.D and III = I 0-072), y 76 = 0-5971 + 0-88,2 (x -X). E. E. Wilson & Baker (1946) liberated Lycopodimn spores at 7-5 ft. above ground-level and caught them, not on glass slides on the ground as Stepanov had done but, because they were interested in diseases of fruit trees, on vertical sticky slides placed on three vertical posts at 1-5, 3-0, and 5-1 metres down-wind from the source, and at thirteen heights above ground at each distance; seven tests were done at wind speeds ranging from 17 to 7-2 metres per sec. Other tests measured the horizontal dispersion. Wilson & Baker calculated the standard deviation of the spores deposited at each distance in each test, and from their values for a we can now esti- mate the parameters for Sutton's equation. In a few individual experi- ments their values obtained for m lay outside the limits of 1-24 and 2-0 postulated by Sutton. But their mean value is m = 1-74 — thus agreeing well with Sutton's theory {m = 1-75) and with Stepanov's experimental value (;;/ = 1-76). Though their values for C^ agree well with Sutton's findings, those for Cy are much higher, but agree with those of Gregory, Longhurst & Sreeramulu (unpublished). 54 HORIZONTAL DIFFUSION Values for C and ;// calculated from various spore-dispersal tests are sho\Mi in Table VIII. Evidently for microbiological work we must use high values of C^., perhaps because we are concerned with longer sampling periods than Sutton. \Xt shall therefore choose Cy = o-8 (metre)*, and Q = 0-12 (metre)*, as standard in Chapter XIII where deposition gradients are considered in detail. There seems to be little in favour, however, of adopting m = 2-0 instead of m = 1-75. TABLE VIII OBSERVED VALUES OF PARAMETERS IN SUTTON's DIFFUSION EQUATION FROM EXPERIMENTS ON SPORE DISPERSAL Stepanov, 1935 Tilletia spores I Tilletia spores II Bovista spores Wilson & Baker, 1946 Lycopodium spores Wind speed 17 metres/sec. 27 3-2 4-6 (mean of 7 experiments) Gregory, Longhurst & Sreeramulu {unpublished) Lycopodium spores Wind speed 04 metres/sec. 1-05 1-28 1-63 264 Gregory {unpublished) Lycopodium spores Cz vertical (metre)7^ 0-20 o-ii 0-22 0-12 015 Cy horizontal (metre)V^ 1-67 0-35 0-89 0-66 4-26^ 0-30 0-40 0-97 0-28 0-58 089 1-40 0-68 026 1-49 1-90 165 170 1-51 1-97 0-88* 175 1-64 1-86 174 1-92 — 1-80 212 I 99 1-66 171 1-98 1-94 * abnormal values. Comparison of Theories of W. Schmidt ant) Sutton According to Schmidt's theory, a^ = sAt/p, so, because t = x/u, we have log a = Mog X + i log {ikjpu). If this relation holds true in field tests, plotting experimental data for log a against log x should give a line 55 THE MICROBIOLOGY OF THE ATMOSPHERE l-O 0-8 E en/ O ^ 02 o"-^ (a) Graph of log a against log x. 0-4 Ob 0-8 log T sees {b) Graph of log a against log t. Fig. 7. — Test of agreement of W. Schmidt's and Sutton's diffusion theories with experiments using Lycopodiiim spores liberated over grass field at Imperial College Field Station, Ascot, Berks. (Gregory, Longhurst & Sreeramulu, unpublished) 56 HORIZONTAL DIFFUSION of slope tan-1 1, that is 26°34'. However, according to Sutton's theory (T^ = |Cx™, therefore log cr = — log X + ^ log (iO) 2 If this holds true, plotting the observed values of log a against log x 111 should give a line of slope tan-^ — . For values of Sutton's m between 1-75 and 2-0, the line should slope at an angle between 40° 36' and 45°. Field tests with Lycopodium spores liberated over short grass by Gregory, Longhurst & Sreeramulu {impublished) allow a direct comparison to be made of the theories of Schmidt and Sutton. Spore-cloud concen- trations were measured near ground-level at distances up to 10 metres simultaneously at 24 points. Results plotted in Fig. 7^ show the lines sloping at angles varying between 40° and 46°. This is incompatible with Schmidt's theory which requires a slope of 26°34'. Furthermore, if log CT is plotted against log t (calculated from the distance and mean wind speed) according to Schmidt's theory a should be the same after a given time whatever the wind speed, but this is not so (Fig. ']b). On Sutton's theory at a given distance log a varies only over a comparatively narrow range of values depending on the parameter m. The results of these experiments are compatible with Sutton's theory which requires a slope of 40°36' for m = 1-75, and 45°o' for m = 2-0. In biological applications we are usually interested in the relation between diffusion and distance rather than between diffusion and time. As we often lack measurements of the variable wind velocities in which dispersion has occurred, Schmidt's theory would be inconvenient to handle. On Schmidt's theory a varies with time, on Sutton's o- varies with distance travelled. Sutton's theory not only fits experimental results well, but is also convenient because it does not require a knowledge of wind speed. 57 VI DEPOSITION PROCESSES We have now considered airborne micro-organisms as diffusing clouds. Before we can discuss processes by which they are deposited in the com- plex outdoor environment, we must deal with deposition processes under simplified, ideal conditions. The word 'deposition' is used in a general sense to include all processes by which airborne particles are transferred from aerial suspension to the surface of a liquid or solid. One form of deposition, the impaction of droplets or particles on surfaces, has been extensively studied both theoretically and in wind-tunnel experiments. It is highly relevant to the problems of spore deposition in nature and of sampling techniques which form the topics of Chapters VII and VIII. 1^ • 1^^ u > X %■ Fig. 8.— Diagram showing relation between concentration (x = number of spores per unit volume); wind-speed = u; area dose (A.D. = number of spores passing through frame of unit area); and trap dose (T.D. = number of spores deposited on unit area of surface). The relation between concentration of the spore-cloud, x, and deposi- tion on the surface (T.D. = trap dose) over which the spore-cloud travels, is illustrated in Fig. 8, together with the concept of 'area dose' (A.D. = the number of particles flowing through an imaginary frame of unit area cross-section at right-angles to the direction of the wind). Concen- tration of the cloud (x = number of spores per cubic metre) is the more fundamental measurement, and the one of greatest interest to the allergist, whose patients inhale volumes of air. The trap dose, which measures deposition on a surface, is of more interest to plant pathologists, plant breeders, and pollen analysts. The area dose is a useful concept in passing from the one measurement to the other. For a given concentration of particles per unit volume of air, the area dose must increase with wind- speed, but whether the trap dose will also be affected is a matter for 58 DEPOSITION PROCESSES experimental investigation. With a continuous source emitting during a limited time, the area dose will be the same as if the same total quantity of particles, Q_05 had been liberated in a number of successive instantaneous puffs arriving in a series of greatly fluctuating concentrations. We can conveniently express the percentage efficiency of a trapping surface as ^ Trap dose per sq. cm. E = T — ^-^1 ^ ^ X 100. Area dose per sq. cm. This convention expresses the efficiency with which a surface clears the spore-cloud to a height of one centimetre above the surface. Deposition on a surface takes place in several ways, including impaction and sedi- mentation, for sedimentation seldom acts alone. Mechanism of Impaction When a bluff object such as a cylinder is placed in wind, the oncoming air-stream has to flow around the obstruction, but airborne particles w^ill be carried some distance towards it by their own momentum before they are in turn deflected by the wind flowing around the obstacle (Fig. 9). Pig. g. — Streamlines of air and particle trajectories around a cylindrical obstruction (vertical cylinder seen in plan). E = streamlines carrying spores towards cylinder; d = diameter of cylinder; arrows on left show direction of wind of velocity = u. If all those particles were impacted whose trajectories in the free wind- stream would have passed through the obstruction, impaction efficiency would be 100 per cent, but apparently in practice it never is. The distance travelled by the particle towards the cylinder before being deflected by the air streamlines flowing around the cylinder is related to both the momen- tum of the particle and the size of the object disturbing the air-flow. Another effect, collection by direct interception, becomes important when the diameter of the particle is an appreciable fraction of the diameter of 59 THE MICROBIOLOGY OF THE ATMOSPHERE the cylinder. With low wind-velocities and with particles smaller than spores, Brownian diffusion may play a major role in deposition. Fig. 10. — Observed relation between E per cent and k = Vsu/^dg. Solid lines from Gregory & Stedman (1953). Broken lines = values for spheres, strips and cylinders as predicted by Langmuir & Blodgett (1949), for

F / F=~^^ v,^ ^^^^3 Fig. 12. — Diagram illustrating gravity theory of particle deposition. ABCD = trap surface; u = wind velocity; Vg = terminal velocity of particle; ABCDEFGH = rec- tangular skew prism containing particles whose trajectories would bring them to rest on trap surface; 6 = presentation angle; tan a = Vs/u. If 6 = 0°, then u sin = o, and deposition under the influence of gravity should depend on the terminal velocity, Vs, so that for a hori- zontal trap the volume of air sampled should be independent of wind- speed, and should depend only on the terminal velocity of the particles. From a cloud of uniform concentration, most trajectories should pass through a surface inclined at an angle d, when tan 9 = u/vs {6 = 45° when Vs = u). If the time-mean density of the spore-cloud = x spores per cu. metre, it will be apparent that the area dose A.D. = xVs — from which, if the deposition is by gravity as assumed, the expected area dose = (ioo/u)vs. Comparison of observed results with the expected value will give a convenient test of the validity of the theory. Our wind-tunnel experiments show that deposition on a horizontal flat surface is a fairly complex process depending on several factors besides the simple resultant of gravity and wind. The surface studied in greatest detail has been the 76 X 25 X 1-3 mm. glass microscope slide, as this has been extensively used in routine spore trapping. Experiments with other plane surface traps are reported in more detail than given here, by Gregory & Stedman (1953). The slide was placed with its long axis at right-angles to the wind and held by clips, placed at the two ends to avoid disturbing the air-flow. Its surface was orientated at various angles to the wind in diflferent experi- ments, the convention adopted being: presentation angle 0° = parallel with the wind; 45° when the leading edge was lower than the trailing E 65 £% 100 Presentation angle ,o -,50 30^ Fig. 13. — Efficiency of deposition of Lycopodium spores on zones across glass microscope slide at presentation angles from (left side of left-hand figure) to 90 (right side of right-hand figure) degrees as observed in wind-tunnel experiments. F% = efficiency as percentage area dose; A, B, C, D, E = successive half-centimetre £% 45° Presentation angle 60° 75° zones across slide from leading edge (A) to trailing edge (E); + = gravity positive; o = gravity neutral; — = gravity negative {from Gregory & Stedman, 1953). Repro- duced by permission from Annals of Applied Biology. THE MICROBIOLOGY OF THE ATMOSPHERE edge; and 90° at right-angles to the wind. (Presentation angle so defined differs from the aeronautical 'angle of incidence' in which at 45°, for example, the leading edge is higher than the trailing edge.) The effect of gravity was studied in two sets of experiments. In one set, with the long axis of the slide vertical (parallel with the z-axis) and the surface making various angles with the x, z-plane, the effect of gravity on deposition must be neutral. In the other set, with the long axis of the slide horizontal (parallel with the y-axis) and the surface making various angles with the x, y-plane, the effect of gravity must be positive at angles from 0° up to less than 90°, neutral at 90°, and negative at angles greater than 90° and up to 180° (gravity condition denoted by g+, go, and g_, respectively). Angles greater than 180° represent the back of the slide. Preliminary tests showed that trapping efficiency varied in different parts of the slide. Accordingly the slide was divided into five |-cm. zones, denoted : A, B, C, D, and E, respectively, from the leading to the trailing edge (Fig. 13). Results of the main series of tests are plotted in Fig. 13, where the efficiency of deposition expected on the gravity theory at 0° for each wind- speed, taking Vs for Lycopodium as 1-76 cm. per sec, is indicated by dotted lines. The observed values below E = o-i per cent are unreliable, but are given to show the trend. Values below E = o-oi per cent, including zero, are all plotted as o-oi per cent as they cannot reasonably be dis- tinguished with the data available. The curves obtained probably result from the interaction of several deposition mechanisms: sedimentation, impaction, turbulence, and edge effects. In certain sets of conditions, one or other of the mechanisms can be found acting singly; but for the most part deposition is interpreted as resulting from the simultaneous action of several mechanisms. DEPOSITION ON HORIZONTAL SLIDES (i) Deposition by sedimentation^ under the influence of gravity alone, is seen on the upper surface of a horizontal slide at the lowest wind-speed tested (conditions denoted by: 0°, 0-5 metres per sec, g+). Here deposition over the slide as a whole was very close to the expected value predicted by the gravity theory (E = Vs/u X 100 = 1*76/50 X 100 = 3-5 per cent), but even at this low wind-speed the bluff edge of the slide, 1-3 mm. thick, caused some edge shadow, shown as a reduced deposition just behind the leading (upwind) edge. That deposition was solely caused by gravity is shown by the absence of deposit on the underside of the hori- zontal slide, and on either side of a vertical slide held parallel with the wind (0°, 05 metres per sec, g_, and go). At the wind-speeds more usual outdoors of between i-o and 2-0 metres per sec, a surprising effect developed in these wind-tunnel experiments. With gravity positive, the bluff edge of the slide produced an edge shadow deflecting a large proportion of the approaching spores; 68 DEPOSITION PROCESSES at I • I metres per sec. the part of the surface behind the leading edge was almost free from deposit, and efficiency reached about 50 per cent of the expected value only on the rearmost zone. At 1-7 metres per sec. efficiency was almost zero over the whole slide. With gravity neutral or negative, efficiencies were also almost zero. (ii) Turbulent deposition. As the wind-speed was raised still further, the efficiency of the horizontal slide recovered ; but deposition cannot have been due to gravity sedimentation, because at 9-5 metres per sec. the amount deposited was almost the same on the under-side (g_) as on the upper-side (g^) of a horizontal slide, and also on the two sides of a vertical slide held parallel with the wind (go). At this speed turbulent deposition is seen in its almost pure condition. As indicated below, there is some evidence that this deposition may result from turbulence generated by the bluff edge of the slide itself. At 0° (horizontal slide) and higher wind-speeds, although deposition was turbulent, gTaviry appeared to interact with the process in some way that is at present obscure. With gravity neutral (go) and wind-speeds of 57 and 9-5 metres per sec, deposition was higher at the leading and trailing edges than in the middle of the slide. With gravity negative (g_) at 9-5 metre per sec, however, the leading edge showed an anomaly, having a deposit 8 to 10 times that found with gravity positive. With gravity negative at 57 and 3-2 metres per sec, deposition behind the leading edge was negligible. DEPOSITION OF Lycopodium spores on inclined plane surfaces Deposition on a horizontal microscope slide is best regarded as a special case of deposition on an inclined plane with the presentation angle 0°. A number of possible angles ranging from 0°, through 90° (vertical slide) to 180°, were tested in the turbulent wind-tunnel, and results are also sho\Mi on Fig. 13. (i) Impaction. Deposition by impaction should be zero at 0°, but it would be expected to occur to some extent at all other presentation angles, as at these angles the surface subtends the oncoming air-stream. With the slide vertical (90°), the gravity effect should be neutral, and at low wind- speeds of 0-5 to I- 1 metres per sec. the slide would not be expected to generate turbulence. Deposition under these conditions should be due to impaction only. In the tests, as wind-speed was increased (90°, 17 to 9-5 metres per sec), deposition increased over the whole surface and was more uniformly distributed; but even at the highest speed tested the deposit at the margin exceeded that at the middle. In this respect im- paction on a plane surface contrasts strikingly with that on cylinders, where the centre of the trace is always denser than the edges. (ii) Edge drift. The effect of the bluff edge of the slide in 'shading' the leading edge has been referred to above. Behind the edge shadow, a region of greater deposition caused by an edge drift might be expected. 69 THE MICROBIOLOGY OF THE ATMOSPHERE Fig. 13 shows that at 0° this edge drift fell behind the trailing edge, but that when the shde was inclined at 15° or 30° to the wind, the edge drift impacted on the slide. This is shown by the deposit on the leading edge which greatly exceeded the expected value over the range 15° to 60°, 1-7 to 9-5 metres per sec. (iii) Mixed ejfects. Over most of the range of zones, presentation angles and wind-speeds, the deposition was from a mixture of two or more mechanisms whose relative importance can be roughly assessed from the empirical results shown in Fig. 13. 90'-' Presentation angle 135° '\l-1m./sec. Fig. 14.— Mean efficiency of deposition of Lycopodium spores on glass microscope slide (all zones) at presentation angles of o° to i8o°. E = efficiency as percentage area dose (Gregory & Stedman, 1953). Reproduced by permission from Annals of Applied Biology. MEAN DEPOSIT ON INCLINED SLIDES In the foregoing paragraphs, efficiencies of various arrangements of a microscope slide, acting under different conditions as a spore trap, have been measured and interpreted in terms of different deposition mechan- isms acting on different surface zones of the slide. In practice, when scanning a slide (or Petri dish trap), we usually need a mean value of efficiency for the whole sampling area. Mean efficiencies for a microscope slide with its long axis parallel to the y-axis, orientated at different presen- tation angles to the x, y-plane (i.e. g+, or g_), are given in Fig. 14. They are taken from the same results as were used in the preceding figure. Each point of Fig. 14 represents values obtained in from one to eight experi- ments. 70 DEPOSITION PROCESSES The values used for Fig. 14 were obtained with highly turbulent wind. Partially streamlining the flow, by removing the turbulence-generat- ing obstruction from the tunnel, had little effect on the deposit — except with presentation angles between 0° and 10°, and with wind-speeds below 5 metres per sec, when efficiency of deposition was reduced, being least with the horizontal slide (0°). Efficiency of deposition on the back of the microscope slide was usually less than i per cent. Decreasing the width of the slide from the customary 2-5 to a mere 0-5 cm., increased efficiency most at the lowest wind-speeds. This narrow trap was most efficient at 5-5 metres per sec. and 90° presentation angle (vertical), and a fall in efficiency at 9-5 metres per sec. and 90° was com- parable with the anomalous reduction in efficiency with very narrow cylinders at higher wind-speeds (Gregory, 1951). TABLE IX EFFICIENCY (E %) OF DEPOSITION ON INCLINED SLIDES IN TURBULENT wiNTD-TUNNEL (Gregory & Stedman, unpublished) Presentation Wind-speed (metres/ sec.) angle 9-5 5-5 3-2 17 I-I 0-5 Lycoperdon {Calvatid) giganteum (c. 4 /x) 0° 0-07 i-i 0-6 1-2 56 3-3 90° 180" 0-04 i-i 17 I-I 6-1 2-2 Ustilago perennans (c. 6 to8/x) 0° 0-2 0-2 0-2 o-i 0-02 0-09 45° 0-3 0-4 0-3 0-04 O-I O-I 90° 0-3 0-2 0-6 0-2 o-oi 0-03 135° 0-4 0-2 0-4 0-09 0-04 0-04 180^ 0-2 0-2 0-2 0-02 0-02 Erysiphe graminis (conidia c. 25 X 12 /x) 0° 0-25 0-13 0-05 0-92 45° 77 370 1-48 0-93 0-86 1-39 90° 2-70 077 o-o6 0-07 135° 4-10 1-83 o-i8 180" 072 0-19 225° 0-28 078 0-31 0-22 0-25 1-16 270"^ 0-39 0-14 0-12 315° 0-07 360° 072 0-19 DEPOSITION OF OTHER SPORES ON INCLINED PLANE SURFACES Lycopodium was studied in greatest detail because it is easily handled and has relatively large spores. Less detailed tests were also made with the smaller spores oi Lycoperdon {Calvatid) giganteum^ with Erysiphe graminis conidia, and with spores of Ustilago perefinans (Table IX). In general it 71 THE MICROBIOLOGY OF THE ATMOSPHERE may be said that impaction efficiency on a microscope slide held at right- angles to the wind is low, but might be as high as 25 per cent with grass pollen or rust uredospores in winds of 9 metres per sec. EFFECT OF THICKNESS OF SLIDE Deposition on a horizontal or inclined surface is evidently complex and may be disturbed by edge shadowing. Tests were therefore made with both thicker and thinner slides, and with thick plates, 10 cm. wide, having a double-bevelled edge. With a horizontal plate 6-4 mm. thick, the edge-effect was present, and at medium wind-speeds edge-shadow became pronounced. At i-i metres per sec. nearly the whole surface was in the shadow of the leading edge, and at 3-2 and 5-5 metres per sec. there was almost no deposit on the slide. At 9*5 metres per sec, however, there was some turbulent deposition on both upper and lower surfaces, both in the turbulent and streamlined wind-tunnel. When the leading edge of the plate, 6-4 mm. thick, was sharpened with a double bevel to form a 45° edge facing the wind (as used by Landahl & Herrmann, 1949), a very different effect was observed. In a wind of 0-5 metres per sec. there was a very even deposit over the whole upper surface, but at i • i metres per sec. an edge-shadow developed and spread across the whole surface as the wind-speed increased, until at 5-7 metre per sec. deposition was negligible. There was no recovery by turbulent deposition at higher wind-speeds, and at 9-5 metres per sec. the thick, bevelled trap under-estimated spore concentration by a factor of 200 times. Thin horizontal surfaces, on the contrary, gave efficiencies much nearer to the expected values for gravity sedimentation. However, edge- shadow and turbulent deposition occurred — even with an edge o-oi6 cm. thick (a microscope cover-glass). A double-edged 'wafer' safety razor- blade gave uniform deposits on the upper surface, which were close to expected values— except at 9-5 metres per sec, where the deposit was three times that expected. ORIENTATION OF SPORES Lycopodiuni spores showed different orientations in different parts of the deposit. Gregory (1951) stated that, on the stagnation zone upwind of a vertical cylinder, the spores lie with the rounded distal surface uppermost, and that spores settling in air under the influence of gravity come to rest in the same position. Further observation shows that this contention was incorrect, and that in the stagnation zone or its equivalent the spore lies with the rounded distal surface touching the sticky cylinder. Evidently the spore becomes orientated with the point trailing as it moves through the air. Orientation with the point upwind is therefore characteristic of Lycopodiuni in the stagnation zone. 72 DEPOSITION PROCESSES Microscopic observation of deposits showed that, as was expected, when the glass slide was horizontal the stagnation zone was on the edge, and when the slide was vertical this zone was in the middle of the slide. At intermediate presentation angles, seen most clearly at angles near 90°, the stagnation zone shifted, at 115° occupying zone B, and reaching zone A at about 120° at the higher wind-speeds. At lower wind-speeds, orien- tation was less definite. Deposition on 9 cm. Diameter Petri Dish The Petri dish trap, extensively used in aerobiological mould surveys, was tested horizontally after pouring with 1 5 cc. of 2 per cent water agar (tests showed that deposition and retention on this medium were similar to those on glycerine jelly). Mean deposition efficiencies (per cent A.D.) for I cm. square zones on the agar surface are given by Gregory & Sted- man (1953). At all wind-speeds, narrow edge drifts occurred behind the rim of the leading edge and in front of the rim of the trailing edge. Efficiency at 0-5 metres per sec. was low, but at i-i and 1-7 metres per sec. efficiency was high, apparently because of the large contribution to the total made by the front and back edge drifts. At 3-2 metres per sec. and above, efficiency fell off substantially below expectation — apparently because the sampling surface was almost entirely shadowed by the i cm. high rim of the dish. Efficiencies were somewhat higher at 9-5 metres per sec. Effects produced by the rim of the Petri dish were nearly eliminated by placing the dish at the bottom of a metal cylinder 13 cm. deep and 11-5 cm. in diameter, sunk below a horizontal flat surface consisting of a square cardboard platform cutting the central axis of the wind-tunnel. The cardboard fitted flush with the mouth of the cylinder and extended 1 1 cm. up- and downwind. The Petri dish was also tested in a vertical position. At 9-4 and 5-5 metres per sec, the deposit was four times as great in the central zone of 2*5 cm. radius as it was in the peripheral centimetre around the rim. At, and below, 3-2 metres per sec. the difference was reversed, with nearly 75 per cent more spores just insids the rim of the dish than elsewhere. Retention and Blow-off From Clean Surfaces Experiments showed that there is no appreciable loss of Lycopodium spores from the deposit on the surface of a slide with a sticky coating of glycerine jelly at any of the wind-speeds tested. Blow-off from a non- sticky glass surface, however, depended on the wind-speed and the angle of incidence of the wind. Clean microscope slides were placed in a spore- cloud at 0-5 metres per sec. to obtain a deposit, and were then placed 73 THE MICROBIOLOGY OF THE ATMOSPHERE successively in winds of increasing speeds. Tests were made at angles of 0°, 45°, and 90°. The percentage of the original deposit that was retained after i minute at each wind-speed was estimated by counting. At the highest wind-speed, 9-5 metres per sec, slight traces of grease on the slide greatly increased retention when the slide was horizontal, and, unless the surface was carefully cleaned before use, erratic results were obtained under these conditions. With the slide vertical (90°), blow-off was nearly linear with wind- speed, 98 per cent being retained at i-i metres per sec, and 60 per cent at 9-5 metres per sec. Blow-off was least at 45°, and at this angle retention was 100 per cent at wind-speeds up to 5-5 metres per sec, and 95 per cent was retained even at 9-5 metres per sec. By contrast, blow-off was greatest with the surface horizontal (0°), when 77 per cent was retained at 1-7 metres per sec, and only 26 per cent at 9-5 metres per sec. These results illustrate the way the laminar boundary layer acts as a dust trap (p. 25), but the actual values probably have little application to plant surfaces. Deposition and Retention on Potato and Bean Leaves The tests described in the preceding sections were all on artificial surfaces, approaching ideal conditions, and gave information on principles of particle deposition from air. This is useful in devising apparatus for sampling airborne particles. We now need to ask how relevant this work is to spore deposition on plant surfaces, especially on leaves which, though rough, are not particularly sticky, and which flap in the wind. To imitate natural conditions, shoots of potato {Solanimi tuberosum) with rough leaves, and broad bean {Vicia faha) with smooth leaves, were placed in the turbulent wind-tunnel and exposed to clouds of Lycopodium spores in the usual manner. The petiole of the leaf was clamped and the leaflets allowed to flap freely, trailing the leaf-tip downwind. After ex- posure, the leaf surfaces were examined under the microscope and the deposit w^as counted on zones across the tip, middle, and base of the lamina (upper and lower surfaces), the deposition efficiency being calcu- lated (Table X). Considerable differences from deposition on a rigid, sticky horizontal slide are apparent. Turbulent deposition either fliiled to develop at the higher wind-speeds, or the spores were shaken off again in the wind. Deposit on the undersides of the leaves was small at all wind-speeds, but on the upper side it was up to 5 per cent of area dose at winds of 0-5 to i-o metres per sec — both on potato leaves and on the still smoother broad-bean leaves. At low wind-speeds the deposit w^as similar to that expected from sedimentation under the influence of gravity; but potato leaves had more spores at the base, and broad-bean leaflets had more at the tip. 74 DEPOSITION PROCESSES TABLE X EFFICIENCY (E %) OF DEPOSITION OF LycopodiwH SPORES ON UPPER AND LOWER SURFACES OF POTATO AND BROAD-BEAN LEAFLETS IN TURBULENT WIND-TUNNEL (Gregory & Stedman, unpublished) Wind velocity (metres/sec.) Part of leaf 9-4 5-3 3-0 1-6 1-3 II 0-6 Potato leaflet Tip (upper) 0-0 o-i6 0-17 0-68 — 0-42 0-0 (under) 0-0 o-o6 0-04 019 — 0-89 0-0 Middle o-o6 0-24 0-30 1-34 — 162 3-20 0-0 0-0 0-02 0-28 — 00 0-0 Base 0-0 0-12 0-47 I-IO — 87 5:8 0-04 0-0 0-0 I -06 — 0-0 00 Broad-bean leaflet Tip (upper) o-o6 0-0 0-15 — 1:4 2-82 5-20 (under) 0-0 00 0-0 — 0-0 0-02 0-0 Middle 0-0 0-0 0-46 — 2-0 3-52 60 0-0 0-0 0-0 — 00 o-o6 0-0 Base 0-0 0-0 0-35 — 2-2 I-3I 2-0 0-0 0-0 0-02 — 0-03 0-05 0-0 Horizontal slide 0-42 0-32 0-05 o-oq — 0-29 0^ 0-36 0-05 0-02 o-oi — 0-004 — Theoretical 0-20 0-34 0-57 1-05 1-35 1-6 3-45 (sedimentation) 0-0 0-0 o-o 0-0 0-0 0-0 o-o There is reasonable agreement between observation and the theory of impaction, but other forms of deposition are impure and complex. Effi- ciency of a collecting surface as an impactor spore-trap increases as the wind-speed and particle size increase, and as the size of the collecting surface decreases. Deposition on horizontal trap surfaces and leaves can be predicted in terms of sedimentation under gravity at low wind-speeds, but edge effects can easily predominate as wind-speed increases, and at 5 to 10 metres per sec. turbulence may result in deposition upwards against gravity. 75 H ^OODS HOLE, MASS.^ VII NATURAL DEPOSITION Having become airborne and been transported by the wind, the spore must quit the turbulent layers of the atmosphere and re-cross the boundary layer before coming to rest in the still layer of air on a solid or liquid surface, which may or may not prove favourable to growth. Some char- acters of spores seem to have been evolved in response to problems of take-off, while others may well have been evolved as adaptations for deposition. Little was known about deposition processes until the development of wind-tunnel techniques — originally for research in aerodynamics — made it possible to experiment on the behaviour of spores in controlled winds in the laboratory. The principal methods of spore deposition in nature can now be tentatively suggested: impaction, sedimentation, boundary-layer exchange, turbulent deposition, rain-washing, and electro- static deposition. Measurement of Deposition The relation between x, the cloud concentration, and 'd', the surface deposition, has been little studied, although it is a relation of considerable biological importance — for instance in pollination, and in epidemics of plant diseases and their control by protectant dusts and sprays. With wind flowing across a smooth surface, it would be possible to calculate deposition directly from a knowledge of concentration, wind-speed, and terminal velocity. However, it is not clear what factors dominate the situation under the complex conditions obtaining in nature, and so the problem must be approached experimentally. With some insight, obtained from Chapter VI, into deposition under relatively simple conditions in a wind-tunnel, we can better understand the complex factors of deposition in nature. In wind-tunnel studies, and in calibrating spore-traps, it is convenient to use percentage efficiency of deposition, E = (Trap dose/Area dose) X 100. For field conditions, two expressions have been used for deposition on the ground. The 'deposition coefficient', p = d/n (where 'd' = num- ber of spores deposited per sq. cm. of surface, and 'n' = number sus- pended per cc. of air), measures the thickness of the slice of cloud cleared while travelling over unit length of ground surface (Gregory, 1945). 76 NATURAL DEPOSITION Under a given set of conditions, p is assumed to depend only on concentra- tion, though we do not yet know how it is affected by wind-speed, tur- bulence and other factors. The 'velocity of deposition', introduced by Chamberlain (1956), was defined as amount deposited per sq. cm. of surface per second ^ volumetric concentration per cc. above surface If p is assumed to be independent of wind-speed, it will be apparent that, in our notation, P = Vg/u. MEASUREMENT OF DEPOSITION COEFFICIENT, 'p'. The first attempt to evaluate p by Gregory (1945) was based on experi- ments by Stepanov (1935), whose results were tested against Sutton's (1932) eddy-diffusion theory. Sutton's formulae had been developed for calculating the concentra- tion of a cloud of particles whose deposition was negligible — the number of particles in the cloud, Q^q, remaining constant throughout the diffusion. In our problem, although the effect of gravity on dispersion has been neglected, the quantity of spores remaining in suspension is steadily diminishing owing to a relatively large deposition from that part of the cloud which is in contact with the ground, so that Q^^, the total quantity remaining in the cloud when its centre has moved to a distance x, is less than the original Q^q. It has been showTi* that Q^x"^ill decrease exponenti- ally with increasing distance, according to the equation : ,(l-*m) Clx=Q.oexp. 2px* Virr) C(I - |m)_ Values of Q^^ ^^d d for two values of the parameter m have been calculated, and it is now possible to test the theory against Stepanov's results. In each of his experiments the total number of spores liberated differed, so that the data had first to be put on a comparable basis by equating the mean deposition, d, at 5 metres, to 100 per cent, and then expressing the deposition observed at greater distances by relative percentages. The logarithms of the observed relative depositions were plotted against the logarithms of the distances in centimetres. The ex- pected depositions when p = 0-05, 0-025, and zero, respectively, were also plotted, and the line calculated for p = 0-05 was seen to approach most clearly to the observed values (Gregory, 1945). A deposition coefficient of p = 0-05 means that, in travelling across i sq. cm. of surface, the entire cloud would deposit a quantity of spores approximately equivalent to the number contained in a slice half-a-millimetre thick through the axial plane of the cloud. This value of p was estimated from experiments in winds of about i metre per sec. * By Margaret F. Gregory {see Appendix in Gregory, 1945). 77 THE MICROBIOLOGY OF THE ATMOSPHERE The problem was next taken up experimentally by Chamberlain (1956), who made Lycopodhim spores radioactive by steeping them in a solution of iodine-131 in carbon tetrachloride. When dry, they were liberated at a height of i metre above a grass field. The concentration of the spore-cloud was measured 20 metres downwind by using sticky cylinders 0-65 cm. in diameter at 30, 60, and 90 cm, above ground-level. The wind-speed at each height was measured and, from the known impaction efficiency of cylinders, the dosage at each height could be estimated (dosage did not vary much with height at this distance). The radioactivity of the grass turf just in front of each sampling position was measured to give an estimate of deposition. 'Velocity of deposition', Vg, and deposition coefficient, p, were then calculated from these two measurements (Table XI), the results giving some support to the idea that p is independent of wind-speed. TABLE XI DEPOSITION ON GRASS OF LyCOpodiuTll SPORES ACTIVATED WITH IODINE-131 (data of Chamberlain's, 1956) AT 20 metres from point where liberated AT I metre above GROUND-LEVEL Wind velocity at I metre above ground-level Vg P cm./sec. cm./sec. 815 2-07 0-0025 310 1-07 00035 160 0-50 0-0031 Further information was obtained by Gregory, Longhurst & Sreer- amulu {unpublished) from experiments in 1956-57 at the Imperial College Field Station, Ascot, England. Spores of Lycopodium^ and the much smaller spores of a bracket fungus, Ganoderma applanatum, were liber- ated a short distance above ground-level in a field of short, rough grass. Cloud concentration and deposition, measured at a number of positions simultaneously, enabled p or Vg to be estimated at various distances up to 10 metres from the point source (Table XII). (Two additional recent estimations at i metre from the source, with Lycopodium spores on a smooth lawn at Rothamsted Experimental Station, gave p = 0-05 and 0-09, and Vg = 3'i and 5-3 cm. per sec, respectively.) A remarkable phenomenon, evident from Table XII, is that both Vg and p vary with distance, decreasing with distance from the source. Much more experimental work is needed before the relation between X and d can be established, but meanwhile it appears that, at 10 metres or so from the source near ground-level, Vg nearly equals the terminal velocity of the particle, whereas closer to the source the value may be up 78 NATURAL DEPOSITION TABLE XII DEPOSITION OF SPORES ON GROUND, FROM SIMULTANEOUS MEASUREMENTS OF X AND d BY VISUAL COUNTS UNDER MICROSCOPE. (Gregory, Longhurst & Sreeramulu, unpublished) Distance from source Mean wind- 2-5 m. S'O m. lo-om. 20-0 m. speed p Vg p vg p Vg p Vf Height of liberation Lycopodium spores i-o metre 1-98 0-83 0-62 0-40 0-25 metre 1-28 i-io 1-09 Mean: Ganoderma spores 0-25 metre 1-70 0-91 076 o-6i Mean: o-oi 0-02 0-44 0-05 0-07 0-13 0-12 0-022 0-015 0-014 0-017 cm. /sec. 2-3 1-2 17-6 6-9 7-2 137 8-16 1-97 I-I3 0-83 1-31 0-02 0-05 0-09 0-05 0-03 O-OI 0-04 0-042 1-5 m. 0-006 0-006 o-oio 0-007 cm./sec. 3-6 3-8 5-5 2-1 4-4 1-5 4-5 3-64 0-56 0-43 0-61 0-S3 cm./sec. 0-02 0-06 0-05 0-02 0-02 0-02 0-02 0-029 2-5 m. 0-003 0-005 0-003 0-006 0-004 3-4 5-1 2-9 i-o 2-7 1-7 1-8 2-8 0-44 0-46 0-24 0-34 0-37 0-03 0-03 5-0 m. 0-0008 0-004 o-ooi 0-005 0-003 cm./sec. 1-7 1-7 0-14 0-34 0-08 0-29 0-21 Summary Lycopodium c. 32 \x. dia. va = 1-76 cm./sec. Ganoderma 20 ^ 181 490 13 2-707 0-072 Examples of deposits obtained with the Hirst trap are shown in Plate I (Frontispiece), and some results of its use are described in the next chapter. (x) The Portable Volumetric Spore-trap (Gregory, 1954) samples 10 litres of air per minute, w ith suction obtained by turning a light, sliding- vane pump by hand. The apparatus weighs about 10-12 lb. Its advantages lOI THE MICROBIOLOGY OF THE ATMOSPHERE are cheapness and portability; time-discrimination is accurate to seconds if desired; it is suitable for work in the less accessible places and for short- time samples (5-10 min.). However, it is fatiguing to use, and being unsuitable for continuous operation, it cannot conveniently be used to STAGE NO. JET SIZE JET VELOCITY STAGE 1 0.0465" DIA. 3.54 FT/SEC STAGE 2 0.0360" DIA. 5.89 FT/SEC STAGE 3 0.0280" DIA. 9.74 FT/SEC STAGE 4 0.0210" DIA. 17.31 FT/SEC STAGE 5 0.0135" DIA. 41.92 FT/SEC STAGE 6 0.0100" DIA. 76.40 FT/SEC ^^-^^^^^-^-^^^-^ W^/////////>>^^ Fig. 18. — Diagram of six-stage Andersen sampler. (Reproduced by permission of Dr. Ariel A. Andersen from the journal of Bacteriology, 76, 1958.) trace diurnal or weather effects on the air-spora. Brook (1959) has devel- oped essentially similar equipment for sampling the air-spora of pastures in New Zealand. A desirable development of both the Hirst trap and the Portable trap would be versions which presented the catch in a form suitable for culture as an alternative to visual microscopic identification. (xi) The Andersen Sampler can be regarded as a development of the so-called 'sieve device' of duBuy & Crisp (1944). In principle this sampler 102 AIR SAMPLING TECHNIQUE (Andersen, 1958) resembles the Cascade Impactor; after entering a circular orifice, air is drawn in succession through a series of six circular plates, each perforated with 400 holes through which spores are impacted directly on to sterile medium in Petri dishes (Fig. 18). Succeeding plates in the series have progressively smaller holes; the largest particles are deposited in the first dish and the smallest in the last; different media can be used for the different size-fractions. Air is sampled at the rate of 28-3 litres per min.; wall losses are claimed to be negligible, and retention is said to be 100 per cent, even for single bacterial cells. In our work the Andersen sampler has proved very convenient and has given good results with bacteria, actinomycetes and moulds, though for particles larger than about 8-10 /x the addition of some form of pre- impinger seems desirable for avoiding losses on the front of the first plate. Reference should be made here to the use of the animal lung as a sampling device. Lurie & Way (1957) injected macerated lung tissue of various animals intrapcritoneally into mice, from whose livers and spleens two dermatophytes {Trichophyton mentagrophytes and Microsporon gypsemn) were subsequently isolated in culture. (xii) Whirling arm. The principle of moving an object through air on a rotating arm has been developed in aerodynamics laboratories as an alternative to the wind-tunnel. As a device for air sampling it has been developed in the United States — beginning w ith the 'airw hip' of Durham (1947), who used a 36 in. aluminium rod to swing a forward-facing, sticky glass slide in a circle at 100 r.p.m. Near a stand of flowering ragweed {Ambrosia spp.), Durham recorded a maximum pollen concentration of about 10 million per cubic metre of air. A high-speed whirling arm trap was developed by Perkins (1957) as the so-called 'Rotorod sampler'. It has been used in plant pathological studies by Asai (i960), and has been somewhat modified by Harrington et a I. (1959). The Rotorod sampler consists of a length of i/i6th in.-square cross-section brass rod, bent to form a vertical U-shaped collector with arms 6 cm. high, 8 cm. apart, and fixed to the shaft of a miniature electric motor running at 2,520 r.p.m. over the range 9 to 15 volts from dry batteries. The sticky arms effectively sample air at 120 litres per min. Rotating at a peripheral speed of about 10 metres per sec, high collecting efficiency would be expected for pollens and spores down to about 12 ju. diameter; for smaller spores, efficiency would fall below 50 per cent. This sampler has proved reliable in use, light in weight, and self-contained; the batteries are cheap, and will run continuously for two or three days. ADHESIVES The choice of adhesives for samplers is limited. Petroleum jelly and glycerine jelly have many advantages. Glycerine jelly has the best optical properties for visual examination, but it is hygroscopic and dissolves in damp weather. Petroleum jelly (e.g. 'Vaseline') is reliable for coating 103 THE MICROBIOLOGY OF THE ATMOSPHERE Hirst trap slides, but addition of paraffin wax (i2| per cent) is often necessary to harden it. The coating must be kept as soft as possible, how- ever, and for extremes of temperature Pady & Kelly (1949) introduced silicone grease for coating trap surfaces. When spores have to be picked off and transferred to culture media, pectin jelly is recommended by Prof. A. J. P. Oort (personal communication). Thermal Precipitation A hot body placed in a dust-laden atmosphere produces a dust-free space around itself (Watson, 1936). This well-known phenomenon has been used in the thermal precipitator, in which a dust-laden airstream flows slowly past a wire heated electrically to ioo°C. above the ambient temperature, depositing dust particles on glass slips for examination. It has been little used for aerobiological work, but is highly efficient for sub-microscopic particles and larger ones up to 5 /x diameter. It is most suitable for use when the particles are in high concentrations, as the volume of air sampled is only about 7 cc. per min. (see Green & Lane, 1957)- Electrostatic Precipitation The movement of charged particles in an electrical field is widely used in industry to extract dust from air because the pressure drop im- posed by the requisite apparatus is small, even w ith high rates of air-flow. Berry (194 1 ) realized that an efficient sampling method could be developed on this principle, and the General Electric electrostatic air-sampler was devised by Luckiesh et al. (1946). Petri dishes of culture medium are placed on flat metal plates (electrodes) which are oppositely charged to 7,000 volts from a half-wave rectifier. Air enters through the apex of a fairly flat inverted metal cone extending to near the edge of each dish, and each cone carries a charge opposite to that of the electrode under its dish. A small pump draws air at 14 litres per min. over each dish, and particles move in the electrostatic field and are deposited on the agar surface. When Escherichia coli in aqueous suspension was atomized into a room, the dish on the positive electrode collected nearly ten times as many cells as that on the negative. With naturally-occurring airborne bacteria, 30 per cent more were deposited over the negative electrode. Presumably each dish collects a proportion of the uncharged particles by gravity and impaction, as well as collecting the charged particles moving towards it. The positions and dimensions of the upper electrodes have been decided by empirical tests and may need modifying for mould spores and pollen. For naturally airborne bacteria the concentration, based on the sum of the counts on the two dishes, was from 2 to 3 per cent higher than simul- taneous tests with the 'duplex radial-jet air sampler'. How far particles 104 AIR SAMPLING TECHNIQUE are charged by ions after entering the apparatus is not yet known. The apparatus was used in aerobiological studies by Kelly et al. (1951), and later by Pady and his colleagues in Kansas. Other electrostatic sampling methods include those of Rack (1959), and O'Connell et al. (i960). Comparison of Methods Under simple conditions it is not difficult to define an absolute standard for air-sampling. With non-aggregated spores of one species liberated in a wind-tunnel, isokinetic sampling through a feathered orifice facing up- wind (using a suitable membrane filter, a Cascade Impactor, or a liquid impinger, with precautions against overloading) should give a reliable visual estimate of the number of particles in a measured volume of air. The Cascade Impactor tends to reveal any spore clumps intact, and, if this feature is undesirable, the liquid inpinger should be used to break the aggregates. The more varied the population in species, size, and state of aggregation, the harder it becomes to devise equipment to measure microbial concentration in the air. Particles over lOju, in diameter must be sampled directly and cannot be ducted around corners on the way to the apparatus without heavy wall-losses. Air hygiene in bacteriology has been mainly a study of the air within buildings, and its equipment has therefore been developed for sampling still or slowly moving air. Aerodynamic effects have been neglected, despite the fact that bacteria are carried on 'rafts' or spray droplets of greatly varying size, so that efficiency of retention has been achieved but efficiency of collection has been neglected. Most devices, such as the slit sampler and the electrostatic sampler, avoid this difficulty by pointing the orifice upwards — but this makes them unsuitable for use in the open air. In outdoor aerobiology the sizes of pollen grains and fungus spores {see Appendix I), and the variability of wind-speeds, has focused atten- tion on collection efficiency. Results from the various 'surface' traps, depending on natural deposition processes, are usually difficult to translate into volumetric results. Over short periods while wind-velocity is constant out-of-doors, or in a wind-tunnel, a vertical strip or cylinder can be used to estimate concentration provided the wind-speed and the deposition efficiencies of the particles concerned are known. If the particle's terminal velocity is known, fairly close estimates can also be made by using theor- etical formulae (e.g. of C. N. Davies & Peetz, 1956). Most data from sur- face traps, such as gravity slide and Petri dish counts, cannot be translated into concentration but merely measure surface deposition. Only the vast diffisrences in natural concentrations, that occur at different times and places, make it possible to infer changes from deposition records. Never- theless Hyde (1959^) shows that in general and over a long period, gravity 105 THE MICROBIOLOGY OF THE ATMOSPHERE slide sampling indicated the same quantitative composition and seasonal variation of the pollen cloud over South Wales as did the Hirst trap. As an exception, the Hirst trap revealed that the abundance of nettle {Urtica) pollen had been greatly underestimated by the gravity slide method. Continuous records provided by the Hirst trap have proved highly illuminating in mycology, plant pathology, and allergy — even though the results are still limited to those obtainable by visual identification. Outdoor bacteriology awaits the development of convenient, continuous sampling equipment for cultures {see Miquel & Benoist, 1890), and this would lead also to further precision in knowledge of airborne fungi. Air sampling, indoors and out, can have either of two aims: (i) to attain the broadest knowledge of the whole range of organisms in the air, which requires the most complete and undistorted sampling methods possible; or (2) to obtain detailed knowledge about a single species or group, which may require highly selective methods. Air sampling has been successful in revealing the diversity of organisms forming the air-spora, in defining conditions for the outbreak of epidemics of some plant diseases, and in measuring dispersal gradients of spore concentration. So far it has proved less useful as a routine measure in forecasting outbreaks of crop disease, because existing methods are mostly insensitive to small concentrations of inoculum in the air (Table XVII). TABLE XVII ESTIMATED DETECTION THRESHOLDS OF CONCENTRATION (SPORES PER CU. METRE OF air) OF HIRST TRAP AND STICKY MICROSCOPE SLIDES INCLINTID AT 45°, ASSUMING EXPOSURE FOR ONE HOUR AND COMPLETE COUNT OF 28 SQ. MM. OF SPORE DEPOSIT. (Hirst & Stcdman, 1961.) Wind-speed (metres/sec.) 05 i-i 175 3-2 55 95 Microscope slide inclined at 4^° Lycopodium 941 Erysiphe 2,000 Ustilago 28,000 Hirst trap Lycopodium 2 Ustilago 2 As pointed out by Hirst (1959): 'No trap is Hkely to detect spores as sensitively as an acre of a susceptible crop in weather favourable to in- fection. Thus epidemics may be started by spore concentrations which traps will not reveal, so we must define the value of "nil catches". With volumetric traps this can be done by calculating the "detection threshold", or concentration at which one spore should appear in the area scanned for 106 260 150 40 20 8 1,500 860 340 70 10 13,000 20,000 1,400 640 490 2 2 3 3 — 2 3 2 2 — AIR SAMPLING TECHNIQUE each sample. In our routine scanning of hourly samples from the Hirst trap the detection theshold is less than lo spores per cubic metre of air. This high value explains why spore traps are of little practical use in forecasting epidemics of potato blight which start from minute local sources, but are valuable for apple scab or black rust, with which initial spore concentrations may be high because of the sudden liberation of accumulated spores or the arrival of a spore-laden air-mass.' Before using an air sampler, its performance should be explored experimentally. Consistency of performance alone is an unsatisfactory criterion, because a trap may be consistently misleading. 107 IX THE AIR-SPORA NEAR THE EARTH'S SURFACE Ultimately we hope to attain an undistorted picture of the ambient outdoor air-spora. All air-sampling methods are more or less selective. This chapter deals with the concentrations of microbes in suspension in air near the ground — that is, within the laminar and turbulent boundary layers ordinarily inhabited by man, animals, and plants. The account is based on the limited amount of information obtained by volumetric air sampling with reasonably efficient apparatus. No attempt will be made to summarize the extensive results from gravity-slide and Petri dish traps, as these are already covered by excellent summaries by Feinberg et al. (1946), Maunsell (1954), Werff (1958), and others, though the results of long-term sampling with such surface traps will be drawn upon for supplementary information when required. Composition of the Air-Spora Some 1,200 species of bacteria and actinomycetes are recognized. Other spore-producing plants include perhaps 40,000 species of fungi, numerous mosses, liverworts, ferns and their 'allies', and more than 100,000 species of pollen-producing flowering plants of which about 10 per cent are wind-pollinated. (Of the Protozoa able to enter the air-spora, our informa- tion is very meagre and unsystematic.) A taxonomist, having in mind the twenty-five volumes of Saccardo's Sylloge Fungoriim or the many volumes and supplements of Index Kewensis, may wonder what useful statements can possibly be made about the air-spora where most of the fungus or other plant bodies whose characters could aid identification are lacking. Fortunately, as a cursory microscopic examination of the deposit from an impactor trap shows, the potentially airborne organisms are not all equally common in the air. One sample is normally dominated by one or two types of spore, with several other types in fair abundance and many more encountered in ones and twos only. The frequency distribution of individuals of different species in an air-spora resembles the series of the logarithmic and log- normal types discussed by Fisher et al. (1943), and by Williams (1947, i960). Investigation will doubtless show that different air-sporas have different 'diversities'. 108 THE AIR-SPORA NEAR THE EARTH'S SURFACE In practice the problem of recording an air-spora requires the recogni- tion of a number of categories for the organisms most commonly present in the sample, including a miscellaneous group, which may ultimately contain from lo to 15 per cent of the total. Pollens of flowering plants can often be identified to the species level, and so can a few fungal spores — especially of the Urcdineae and some other plant pathogens (Appendix I, p. 207). In samples of the outdoor air, bacteria can seldom even be recognized visually as such, let alone identified, and the only sampling devices suitable for their study involve making cultures. In practice the categories adopted are of varying degrees of arbitrariness, and the names are applied to them for convenience of reference ; but the different cate- gories behave so differently that it would be intolerable to have no way of referring to them. Taxonomic Groups Needing Study in the Air-Spora We now have some knowledge of the occurrence of bacteria, fungi, and pollens as components of the outdoor air-spora, but there are some groups whose presence is obvious enough yet about which we have scarcely any quantitative information. Thus I know of no systematic records having been kept to assess the concentration of actinomycetes, and of moss and liverwort spores, in the atmosphere. (i) Protozoa. For these we have the estimate by Miquel (1883) of an average of o-i to 0-2 airborne protozoan 'eggs' per cubic metre at the Observatoire Montsouris, Paris. Using a Pasteur-type filter, Puschkarew (191 3) sampled air near ground-level on the right bank of the Neckar downstream from Heidelberg. In forty-nine tests, on different occasions and at different times of the day, his catch works out at an average of 2-5 protozoan cysts per cubic metre of air. His cultures included species of Amoeba, Bodo, Monas, Calpoda, etc. Curiously enough, he concluded that this concentration was too small to account for the observed almost world-wide uniformity of species of freshwater protozoa, and that other dispersal routes must be important (as no doubt they are). This study awaits convenient techniques. (ii) Algae. Microscopic terrestrial and freshwater algae occur in the air, but have been little studied. A few samples were taken on the roof of buildings at Leiden by Overeem (1937), using the 'standard aeroscope' and Rettger bubbler. At least forty algae were obtained from a total of 20 cubic metres of air, including: Chlorococctwi, Chlorella, Pleurococcus, Stkhococcus, and Navicula. The occurrence of blue-green algae resembling species of Gloeocapsa or other members of the Chroococcaceae was recorded by Gregory et al (1955) from continuous sampling with a Hirst trap at Thorney Island in Chichester Harbour, England, from 30 June to 13 July, 1954. Concen- trations averaged 1 10 colonies per cubic metre of air (averaging 8 cells per 109 THE MICROBIOLOGY OF THE ATMOSPHERE colony). Diurnal periodicity showed maximum numbers near midnight (210 colonies per cubic metre) and a minimum in the morning (30 col- onies). Similar, but fewer, colonies were found regularly in London and Rothamsted. They show^ed no pronounced seasonal trend, according to Hamilton (1959), who also reported the rare occurrence of diatoms and desmids. Evidently microscopic algae are widely prevalent in the at- mosphere in numbers varying from a few to a few hundred per cubic metre, and occasionally they may be heavily deposited on the ground (D.S.I.R., 1931). (iii) Ferns. For ferns the reports are few. At Rothamsted — with no large areas of bracken within several kilometres, and only small quantities within I km. — spores of the Pteridimn type occurred frequently in warm, dry weather from late July to mid September. They averaged 4 per cubic metre, with a maximum concentration of 36 per cubic metre (Gregory & Hirst, 1957; Hamilton, 1959). Miquel's Work on Bacteria and Moulds Recognizing the paucity of information on airborne microbes, Pierre Miquel made daily counts in Paris during the last quarter of the nineteenth century (cf. p. 9). Miquel's 'contribution to the microscopic flora of the air' is probably the most sustained series of volumetric measurements of the microbial population of the outdoor air ever attempted. Daily observations in the Pare Montsouris, about 5 km. south of the centre of Paris, served him as a standard for comparison with the polluted air in the densely populated city. The bacteria of the outdoor air were classified in the following percentages as: Micrococcus 66, Bacillus 25, Bacterium 6, Vibrio 1-2. Miquel (1899) shows a seasonal variation in total bacterial and mould concentrations (Table XVIH). Most of the samples were taken with a form of the Pasteur trap {see p. 5), using a sterile plug of powdered an- hydrous sodium sulphate as a filter. This was dissolved after exposure and inoculated to flasks of filtered saline beef extract. At the Pare Mont- souris, bacteria were nearly three times as numerous in summer as in winter, but moulds fluctuated rather less. Near the Hotel de Ville in the centre of Paris, bacteria showed a similar seasonal variation but were 2| times as many as in the Pare; moulds were 10 times as numerous but showed little seasonal variation. At first Miquel argued that, as only one-tenth could have been blown in from the country to the centre of the city, the rest must have come from houses. But after the year 1881 he noted a steady annual decline and he attributed this to improved street cleaning and washing to lay dust which, we may suppose, consisted largely of soil enriched with horse droppings. Data are also given for a narrow, unhygienic street, and for one of the main sewers of Paris. The air of sewers was no more highly contaminated than the outside air, and was often surprisingly pure (Chapter XH). no THE AIR-SPORA NEAR THE EARTH S SURFACE TABLE XVIII MEANS OF MONTHLY MEAN NUMBERS OF BACTERIA AND MOULDS PER CUBIC METRE OF OUTDOOR AIR IN PARIS (Miquel, 1 899), IN CULTURE IN NEUTRAL BEEF BROTH NearH [otel de Passage Pare Montsouris Ville, _ place Saint- Pierre Main sewer (16- and 9-year Saint-Gervais 1897- -1898 Blvd. Sebastopol Month means, respectively) (1888- 1897) (Mean) (1891- -1897) Bacteria Moulds Bacteria Moulds Bacteria Moulds Bacteria Moulds January 198 160 3,840 1,555 6,610 1,665 2,670 4,535 February 148 no 3,475 1,375 3,265 1,790 3,095 1,965 March 209 155 4,995 1,290 2,790 1,630 2,555 2,485 April 362 140 8,260 2,445 11,710 1,885 3,875 6,290 May 295 230 8,725 1,560 4,910 1,650 3,845 1,865 June 355 222 10,830 1,835 5,015 2,630 2,705 2,360 July 464 205 12,040 2,590 5,930 4,235 4,460 3,490 August 450 270 10,300 2,450 4,265 2,770 4,645 3,195 September 395 215 9,920 2,435 5,545 1,735 3,630 1,845 October 260 228 7,160 2,445 7,900 2,165 3,965 4,135 November 195 240 5,845 2,175 4,735 2,270 3,800 5,210 December 167 166 4,365 2,005 4,015 1,390 6,750 2,560 Average 290 195 7,480 2,015 5,555 2,150 3,835 3,330 EFFECT OF RAIN The numbers of bacteria in summer averaged several hundred per cubic metre, and were reduced in a few hours by rain to a mere 20-30 per cubic metre, but they increased again as the ground dried out. Sur- prisingly enough, bacterial numbers often increased after snowfall. The numbers of bacteria in the air increased with increasing wind-speed and remained high during a drought, unless it was prolonged. To Miquel it was clear that rain had a complex action: air that contained many bacteria after a fine, dry spell of weather was rapidly purified by rain, but often during a spell of humid weather the fall of rain would contaminate the air more than it purified it — possibly because raindrops collected bacteria in their fall towards the ground and, by evaporating before reaching the soil, added their collection to the air near ground-level. We may also suspect that bacteria were put into the air by rain-splash. The first rain after drought might contain 200,000 bacteria per litre instead of the average number of 3,380 per litre (see Chapter XI). Slowly Miquel came to the conclusion that the source of most outdoor airborne bacteria is the surface of the ground, whence they are picked up with dry soil particles by wind — a conclusion still acceptable 80 years later. Moulds reacted differently to the fall of rain. With the onset of rain, the air was at first purified ; but when rainy periods lasted for some days, the numbers of mould spores in the air at Montsouris often increased remarkably, even reaching 95,000 to 120,000 per cubic metre. In dry III THE MICROBIOLOGY OF THE ATMOSPHERE weather coloured spores abounded. Re-invasion of the air after rain was mostly by colourless organisms, which Miquel diagnosed, probably incorrectly, as immature spores. Rain and warmth increased the atmos- pheric fungus spore content, though this might decrease during high winds because their extra lifting power did not compensate for their power of desiccation and, hence, killing. DIURNAL PERIODICITY (i) Diurnal periodicity in numbers of bacteria outdoors could be studied only on selected days of the year because, by Miquel's methods, for hourly studies from 600 to 700 culture flasks had to be handled in one day. The fullest data are probably those of 1884, which showed continual change from hour to hour of the day in relation to changes in meteorological factors that have not yet been unravelled. NOON Fig. 19. — Diurnal periodicity of total numbers of bacteria in air at the Observatoire Montsouris, Paris, based on hourly readings between March 1882 and September 1884 (Miquel, 1886). Bacterial numbers showed diurnal periodicities differing between Montsouris and the centre of Paris. At Montsouris there were two daily maxima at o8-oo hours and 20-00 hours, and two minima at 02-00 hours and 14-00 hours, respectively (Fig. 19). In the centre of Paris, however, there tended to be a single maximum at 14-00 hours and a minimum at 02-00 hours during most of the year, but in autumn the double peaks tended to occur in central Paris as at Montsouris. This diurnal variation was shown to hold irrespective of wind direc- tion — an effect that was possibly in part attributable to mechanical causes such as traffic and the sweeping of streets. Furthermore the peaks occurred 112 THE AIR-SPORA NEAR THE EARTH's SURFACE also in rainy weather so long as not more than 2-3 mm. of rain fell in 24 hours. Miquel also showed that the outside changes soon penetrated into the rooms of buildings unless they were exceptionally well sealed. (ii) Diurnal periodicity of ttwuld spore numbers resembled that of bacteria. Trapping by impaction on a moving slide which gave him hourly readings, Miquel found that moulds had two maxima at o8-oo and 20-00 hours. These maxima were independent of wind velocity and fluctuated much more than did bacterial counts. Miquel then tried 15-minute sampling periods and discovered the important principle that hourly values are merely a smoothing of still more rapid fluctuations; he says (transl.) 'what I wish to establish by all these examples is the variability of the nature of the organisms living in the atmosphere'. RELATIVE NUMBERS OF BACTERIA AND MOULDS Miquel had started with the aim of describing the cryptogamic flora of the atmosphere, and in the earlier years of his work he reported much larger numbers of moulds than bacteria. In 1879 Miquel was assessing mould spores visually by a continuously operated, aspirated aeroscope at 2 metres above ground-level at the centre of a lawn in the Pare. He caught microbes at 100 times the rate of the non- aspirated aeroscopes of Maddox and Cunningham, and he concluded that less than 10 per cent of the organism seen visually would grow in culture. The numbers of germs (principally mould spores as showTi by his draw- ings) in the air at the Pare Montsouris during continuous sampling in 1878 averaged 28,500 per cubic metre. In rainy periods in June they rose to 100,000 or even 200,000 per cubic metre. In winter the numbers were as low as 1,000 per cubic metre during snow, though they might be 14,000 per cubic metre when the wind came from over the centre of Paris. Numbers increased again in spring ; they remained high in summer and diminished again in autumn. By the 1890's Miquel had lost interest in airborne moulds, and for sampling air he constantly recommends the use of sugar-free media which discourage moulds but enhance bacterial counts. Henceforth the mould counts which he reported fell to the level of the bacteria and, by deliber- ately using a selective medium, he could forget the rich fungus spora that had embarrassed him in the earlier years. The values already given (Table XVIII) are for media which favour bacteria but repress moulds, and are certainly underestimates for the latter. Knowledge of the broad features of the bacterial flora of the outdoor air near the ground remains to this day substantially as Miquel left it at the beginning of this century. Further, comparable measurements in this century include those by: Forbes (1924), Wells & Wells (1936), Buch- binder et al. (1945), and Colebrook & Cawston (1948). On the whole the topic has been neglected and, significantly, the American Association for H 113 THE MICROBIOLOGY OF THE ATMOSPHERE the Advancement of Science's book on ^Aerobiology^ (Moulton, 1942) has no chapter on bacteria in outdoor air over land. Recent Study of Fungi and Pollen The pollen and fungus components of the outdoor air-spora have attracted much attention in this century, and the development and ex- tensive use of volumetric sampling equipment for the purpose has been highly illuminating. Quantitative visual counting of spores from 3 /x in diameter and upwards, confirms Miquel's impression of recurrent 'tides' of spore concentration ; but different groups of organisms are now known to have separate 'tidal waves', and the concentration and composition of the air-spora varies enormously with place, season, time of day, weather, and human activity. THE AIR-SPORA AT 2 METRES ABOVE GROUND-LEVEL Continuous records in a mixed agricultural environment were obtained during the summer of 1952 by Gregory & Hirst (1957), using the Hirst automatic volumetric spore-trap. The mean spore concentration at 2 metres above the ground over the period i June to 25 October was 12,500 spores per cubic metre. These were grouped visually into twenty-five categories. The commonest spore-type was Chidosporiimi (probably mainly C. herbarum)^ which accounted for 47 per cent of the total. The second commonest were classified as hyaline basidiospores and made up 31 per cent of the season's catch; most of these were probably spores of species of Sporobolomyces, with Tilletiopsis adding another 0-56 per cent. Coloured basidiospores of mushrooms and toadstools (agarics, boleti, and bracket fungi) amounted to 3-3 per cent of the season's total. Pollen made up i per cent of the total. Conidia of pow^dery mildews (Erysiphaceae), 'brand spores' of Ustilago species, and conidia of Altemaria, amounted to between I and 2 per cent each. Ten other recognizable categories contributed between 0-03 and 0-5 per cent each. All other particles recognizable as spores of micro-organisms were put in the 'unclassified' category, totalling 8 per cent of the season's catch, and included many organisms which, although abundant in soils, form only an insignificant fraction of the summer outdoor air-spora (for example, PeniciUiimi, Aspergillus, and various Mucoraceae). Bacteria and actinomycetes are not revealed by this trap method, which is efficient only for particles over 3 /^ in diameter. In Britain the attempts to get a relatively undistorted picture of the outdoor air-spora have demonstrated beyond doubt that Cladosporium and Sporobolomyces predominate, followed by the hyaline and coloured basidiospores of the mushrooms and toadstools. Fewer in number, but not necessarily less in total volume, are the pollens, Alternaria, ascospores, and the large-spored plant-pathogenic fungi. Under ordinary conditions, splash-borne types seem not to amount to more than a few per cent of the 114 THE AIR-SPORA NEAR THE EARTHS SURFACE total alr-spora. Many other t^'pes are also found, but they are infrequent, except in special localities or under special circumstances. The importance of basidiospores from mushrooms, toadstools, bracket fungi, and especially mirror-yeasts, as components of the air-spora, is a recent discovery (Gregory & Hirst, 1952, 1957). It is remarkable that even the existence of the Sporobolomycetes was unrecognized until 1930. Basidiospores are not efficiently caught by surface traps, and con- firmation of their numbers (which were doubted at first) has been slow in forthcoming. However, Hyde & Adams (i960) report that at Cardiff, over the whole year of 1958, the basidiospore types collectively amounted to 1,059 ou'^ of the average fungus spore content of 2,164 P^r cubic metre of air. Furthermore, estimating volume instead of number, they showed that basidiospores came second only to grass pollen. Daily estimates with a slit sampler for one year at Manhattan, Kansas, gave the numbers of basidiospores as 24-3 per cent of all spores caught — second only to Cladosporium (Kramer et al.^ ig^ga). THE AIR-SPORA AT OTHER HEIGHTS NEAR THE GROUNT) In general the spore concentration increases at positions nearer the ground than the standard sampling height of 2 metres, and decreases at greater heights. Using a Hirst trap at 24 metres in a lattice tower at Rothamsted, the average spore concentration was 81-5 per cent of that at 2 metres, though some spores characteristic of the night air-spora were actually commoner at the higher level (Gregory & Hirst, 1957). Tests at heights below 2 metres with a portable suction trap in the New Forest, England, showed a general decrease with height (Table XIX). The difference was greatest at night, when the total spore concentrations were lowest. By way of exception, Cladosporium numbers were reversed at I3'00 hours: this is taken to mean that sources of Cladosporium were not present close to the trap and that the air nearest the ground \^•as being depleted of this organism in passage over the Earth's surface. TABLE XIX TOTAL NUMBER OF POLLEN GRAINS AND SPORES PER CUBIC METRE IN OAK- BIRCH WOOD, NEW FOREST, ENGLAND, 23 JULY 1 953 (Gregory, 1 954). Height above ground-level 7 cm. 30 cm. 120 cm. 05-00 hours G.M.T. 20,600 19,000 7^250 13-00 hours G.M.T. 31,300 24,200 20,300 In his studies of the 'phyllosphere' of cereal leaves, Last (1955) sampled air among wheat plants at 11, 46, and 80 cm. above ground-level and found 237,000, 170,000, and 41,000 spores of Sporobolomyces per cubic metre, respectively. 115 THE MICROBIOLOGY OF THE ATMOSPHERE The air-spora at greater heights than in these examples is deak with in Chapter X. TABLE XX DIURNAL PERIODICITY IN THE AIR-SPORA ON LAND (APPROXIMATE TIME OF MAXIMUM CONCENTRATION IN THE LOWER ATMOSPHERE AS RECORDED BY VOLUMETRIC SAMPLERS) Organism Local time (hr.) References Bacteria 6 and i8 (I) Fungi (Phycomycetes) Peronospora tabacina 6-10 (9) Phytophthora infestans II (2) (Ascomycetes) Filiform ascospores 19-3 (2), (3), (4) Fusiform ascospores I (2), (4) Erysiphe (conidial) 10, 13-15 (2), (3), (4), (5) (Basidiomycetes) Sporobolomyces 3-5 (2), (3), (4) TiUetiopsis 3-6 (3), (4) Rusts 'Uredospores' 12-16 (2), (3), (5) Piiccinia polysora 9 (dry season)! (6) 13 (wet season)/ Smuts Tilletia 14 and 20 (5) Ustilago 10, 12-16 (2), (3), (4), (5) Hymenomycetes Coniop/iora 4 (5) Ganoderma 22-3 (3), (5) 'coloured basidiospores' 1-3,5 (2), (3), (4), (5) 'hyaline basidiospores' 23-3 (3), (4) (Fungi Imperfecti) Altenmria and Stemphylium 10-12-15 (2) (3), (4), (5) Botrytis 12-15 (3), (5) Cladosporium 10, 11-15, 17 (2), (3), (4), (6) Epicoccum 10-18 (3), (5) Helmintlwsporium 14 (19 in London) (3), (5) Nigrospora spherica 1 1 (dry season) ) (6) 13 (wet season)/ Penicillium H (3) Periconia 14 (5) Piricularia oryzae 1-3 (10) Polythrincium trifolii 10-12 (16 in London) (2), (3), (5) PuUularia 13-17 (3) Torula herbanim 11-13 (3) Algae 'Gloeocapsa' type 22-23 (4), (8) 116 TABLE XX — cojitinued Higher Plants Total grass and weec Grass pollen Bettda (birch) Corylus (hazel) Fraxinus (ash) pollen 15-16 19 9-15 13 11-13 Pinus (pine) Platamis (plane) 15-17 13-15 Qiiercus (oak) Tilia (lime) Artemisia 13, 15, 17 II, 13, 17 9 other Compositae Chenopodium Plantago Rumex 13 II, 13, 19 13, 17, 23 II, 15 Urtica 15, 17 THE AIR-SPORA NEAR THE EARTH'S SURFACE (2), (5) (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) (3) Key to references (i) Miquel (1886), (2) Hirst (1953), (3) Hamilton (1959), (4) Gregory & Sreeramulu (1958), (5) Sreeramulu (1959), (6) Cammack (1955), (7) Gregory & Stedman (1958), (8) Gregory et al. (1955), (9) Waggoner & Taylor (1958), (10) Panzer et al. (1957). DIURNAL PERIODICITY OF THE FINE-WEATHER SPOR.'V Concentrations of spores of a single species or a group of related species often show diurnal rhythms comparable with those observed by Miquel {see above). This effect was studied in detail by Hirst (1953), whose mode of presenting the results has proved a convenient model for sub- sequent workers. Hourly or two-hourly concentrations are obtained on as many days as possible. At regular intervals throughout the day, the mean spore concentration is plotted as a percentage of the maximum. Geometric means are preferred to arithmetic means, and separate curves can be drawn for different weather types. Examples are shown in Fig. 20. The time of day when various organisms reach maximum concentration in the Hirst trap and similar volumetric samplers is listed in Table XX. Ill-defined peaks or multiple peaks may perhaps be due to failure to discriminate in visual counts between two morphologically similar spore- types belonging to different organisms. Several patterns of diurnal periodicity are now obvious. (i) The bacterial pattern found by Miquel at Montsouris, with two maxima and two minima, remains an unexplained phenomenon worthy of re-examination. (ii) The nocturnal pattern contains a group of organisms appearing in highest concentration at some hour between sunset and dawn. This group comprises all of the ballistospore forms so far studied (notably Sporobolomyces, Tilletiopsis, basidiospores of hymenomycetes), and also certain fusiform ascospores. In cooler weather, however, coloured bas- idiospores may reach a maximum in the afternoon (Gregory & Stedman, 1958). 117 Pollen [L] THE MICROBIOLOGY OF THE ATMOSPHERE Alternaria 9 ^ Cladosporiuw ^ --0-- Noon Phytophthora /\> 1 infestans [L] 75 - /' \ - Polythrincium 50 _ /' '^ _ trifolii ^ / '^ \ 25 ~--^^->^ 1 ^^ -— ^ Noon Group A filiforms Group B fusiforms --0-- Group C hyaline unidentified 75 50 25 / Noon \ Coloured basidiospores [U] 75 1 \ 1 X- Hyaline basidiospores 50 f \ . V^-~^ ^^^ 25 - / V - -'' '-~-. .____] .--'- / / / / / -I Fig. 20. — Mean diurnal periodicity curves of thirteen spore-groups expressed as percentage of the peak geometric mean concentration. (From Hirst trap records at Rothamsted Experi- mental Station, summer 1952.) (The symbols 'L' and 'U' refer to corrections applied to the data, based on wind-tunnel tests of the Hirst trap with Lycnpodiiim and Ustilago spores, respectively.) (Reproduced by permission of J. M. Hirst from the Transactions of the British Mycological Society, 1953.) (iii) The forenoon pattern. The hours after dawn, o6-oo to lo-oo hours, bring an interlude, but at about lo-oo or i roo hours the forenoon pattern is characterized by peak concentrations of a few crop-pathogenic fungi, e.g. Polythrincium trifolii and Phytophthora infestans. 118 THE AIR-SPORA NEAR THE EARTHS SURFACE (iv) The afternoon pattern develops from noon to i6-oo hours and contains the majority of the day-time spore forms, notably: Cladosporium, Alternaria^ conidia of Erysiphe, many other Fungi Imperfecti, uredospores of rusts, and brand spores of smuts. (v) The evening pattern. The early part of the evening, from lyoo to 21-00 hours, forms another interlude with few well-defined maxima, until the nocturnal spora starts at about 22-00 hours, lasting until dawn. All these diurnal periodicities are based on mean records for a number of days. Weather on a particular day may disturb the normal rhythm : for example, Sporobolomyces may persist to 10-00 hours or later; anthesis of grasses may be suppressed in dull weather, when pollen concentration may remain low. In the dry season in Nigeria, certain typically afternoon types may occur in the forenoon (Cammack, 1955). One other feature is shown by many organisms, namely, the rise from a low value to the maximum is often steep, whereas the fall is relatively slow; with others, such as Sporobolomyces., the reverse is often true. Panzer et al. (1957) refer Leptosphaeria, Epicoccimi., Piricularia., Chidosporiimi, Diplodia, and Ophiobolus, to the 'night spora' with maxima between 18-00 and 09-00 hours, and Nigrospora, Penicillium, Alternaria, Curvularia, and Tricoconis to the day spora, with maxima between o8-oo and 17-00 hours. Some discrepancies with findings in Britain, such as the occurrence of Chidosporium and Epicoccum at night, suggest that local conditions favouring a high concentration in air may be found by day in one climate and at night in another climate. The causes of these diurnal fluctuations are complex. Hirst suggests that those spores which are commonest in the forenoon depend on hygro- scopic changes during drying to liberate their spores ; commonest in the afternoon are often those pollens and spores which are passively dispersed by shaking and wind erosion from dry surfaces, and conditions favour their liberation in the afternoon. Some nocturnal forms, such as Sporo- bolomyces., Tilletiopsis, and some ascospores, depend on dew\ High night- concentrations need not always be interpreted as resulting from an increased number of spores being discharged at this time. We do not yet know whether the basidiospore diurnal rhythm (Fig. 20) reflects a diurnal spore-production rhythm, or whether spore emission is relatively constant during the 24 hours, being diluted through the considerable height of the turbulent boundary layer by day, but remaining concentrated near the ground during the night — when eflfects of frictional and thermal turbulence are usually small. SEASONAL CHANGES Seasons aflfect the air-spora profoundly. Cladosporium and Alternaria show pronounced seasonal periodicity, as of course do the pollens and spores of mosses, pteridophytes, and plant-pathogenic fungi. On the other hand PenicilUum may show little seasonal change, whilst in cities 119 °F. °C. 40-44 4-4-67 45-49 7-2-9-4 50-54 I00-I2-2 55-59 I2-8-I5-0 60-64 I5-6-I7-8 65-69 I5-6-I7-8 70-74 2I-I-23-3 THE MICROBIOLOGY OF THE ATMOSPHERE it may even be more plentiful in winter than in summer (Maunsell, 1958; Hamilton, 1959). To the allergist the winter is a period of allergen-free outdoor air. Spring brings the deciduous tree-pollens, followed by those of conifers in early summer and, more important, by grass pollens char- acterizing the 'hay-fever' season. Late summer brings the moulds and what are known as the 'weed pollens' (including the notorious ragweeds, Ambrosia spp., in North America), extending into early autumn. Late autumn, like the winter, is relatively free from allergens. Hamilton (1959) recorded the temperatures at which various spore types were in the air in maximum numbers (Table XXI). TABLE XXI TEMPERATURES AT WHICH HIGHEST CONCENTRATIONS WERE RECORDED. (Hamilton, 1959.) Spore Category Leptosphaeria. Ventiiria. Nolanea, '' Penicilliuni' . Yellow basidiospores. Fusiform (thin) ascospores, Sporobolomyces, Tilktiopsis. Coniophora, Etitomophthora, Fusiform (fat) ascospores, Lactarius. Cladosporiiim, Dicoccum, Erysiphe, Helicomyces, Periconia, Ustilago. ']S~19 23-9-26-I Ahernaria, Chaetomium, Filiform ascospores, Ganoderma, P/iytophthora, Polythrincium, Pullularia, Sordaria^ T/iele- phora, Torula, uredospores. 80-84 267-28-9 Brown basidiospores, Yellow basidiospores, Botrytis, Epi~ cocciim, Fusiform (thin) ascospores, Helmint/iosporium, Macrosporiwn, Psilocybe. 85-89 29-4-3 1 7 Hyaline basidiospores. EFFECT OF LOCALITY The air-spora near the ground tends to be dominated by local sources, and these components of local origin are seen against a background of others from many distant sources. Volumetric sampling shows that some species are practically ubiquitous, whereas others are more or less con- fined to certain localities. Valuable surveys of airborne pollen are given by Hyde (1952, 1956, 1959), and of fungus spores by Werff (1958). Species of Cladosporimn belong to the ubiquitous group. They dom- inate the day-time spora in temperate regions and in the moist tropics (Hirst, 1953; Cammack, 1955; Gregory & Hirst, 1957; Pady, 1957; Hamilton, 1959; Kramer et al. 1959). This phenomenon agrees with surveys by gravity sampling methods in Britain and New Zealand (Dye & Vernon, 1952; Menna, 1955; Richards, 1954^). The dominance of various types of basidiospores (ballistospores) at night is probably also a widespread phenomenon whose magnitude is now becoming apparent. 120 THE AIR-SPORA NEAR THE EARTH S SURFACE At Rothamsted, Gregory & Hirst (1952) found that between early August and late September the concentration of coloured basidiospores seldom fell below 1,000 per cubic metre. From June to October, 1952, coloured basidiospores formed 3-3 per cent and hyaline basidiospores 46-5 per cent of the total air-spora in the size-range above about 4^, and their abundance at Rothamsted in 1954 was confirmed by Hamilton (1959). It soon became clear that the hyaline basidiospores were mostly from colonies of Sporobolomyces occurring on leaves, and that their abundance varied greatly in different places. Last (1955), when sampling air within a stand of wheat, found large differences in concentration of Sporobolo- myces between manured and unmanured plots in the same field (Broad- balk field, Rothamsted). The highest concentration of sporobolomycete spores so far reported (up to i million per cubic metre) was found near Chichester Harbour, England (Gregory & Sreeramulu, 1958). CLADOSPORIUM 100 RELATIVE NOS. 50 ROTHAMSTED G-~ ■0---0- LONDON 01 15 2 3 HRS. 6.M.T. Fig. 21. — Diurnal periodicity of Cladosporium at London (South Kensington) and Harpen- den (Rothamsted), based on Hirst trap catches from May to September, 1954. (Reproduced by permission of Elizabeth D. Hamilton from Acta Allergologica, 1959.) In any one place it is difficult to disentangle the respective contributions of local and distant sources. It might be expected that spore concentrations would be lower in a city than in the near-by countryside. This was indeed found by Hamilton (1959), who compared continuous records from two Hirst traps, one at Rothamsted (2 metres above ground) and the other in London (16 metres above ground on a roof in South Kensington). The counts of total pollen were higher in London because of a large excess of Platamis (plane) pollen, but grass pollen grains were 50 per cent more numerous in the country. Fungus spores outnumbered pollen grains by 121 THE MICROBIOLOGY OF THE ATMOSPHERE 75 to I, and although counts in London were less than half those at Rothamsted, they still averaged 6,500 per cubic metre during the 1954 season. With some species, diurnal changes in concentration tended to be less pronounced in London than in the country (Fig 21); this may per- haps have indicated that spores of the species trapped in London came mostly from distant sources. Results of extensive surveys on a roof at Manhattan, Kansas, are based on daily sampling at approximately 09-00 local time. A Pady-Rittis slit sampler was used for visual identification and the G. E. Electrostatic sampler for making cultures on Rose-Bengal agar. The maximum numbers of fungus spores recorded at one time was 100,000 per cubic metre in visual traps and 20,000 per cubic metre in culture (Pady, 1957; Rogerson, 1958; Kramer et a I. 1959). The main constituents were identified as follows : Visual In Culture per cent per cent Cladosporiuin 409 44-5 Basidiospores 243 — Non-sporulating — 176 Alternaria 34 12-6 Yeasts 7-3 8-4 PenicilUiim 6-1 Aspergillus — 5-4 Smuts 5-9 — 2-celled hyaline 4-4 — Fiisariwn 29 — i-celled hyaline 1-4 . — Cercospora i-o — EFFECT OF WEATHER Atmospheric spore concentration fluctuates according to meteorological conditions. It also fluctuates for biological reasons such as growth and differentiation of the spore-producing organisms. Studies by Hirst (1953) show that the pollens, and spores of Cladosporium, Erysiphe, Alternaria, smuts, and rusts (which together form the main components of the day- time 'dry-air' spora), are mostly removed by prolonged rain which, however, soon puts into the air a characteristic damp-air spora. Fluctuation is a property of the fine-weather air-spora, but some types depend on rain to get into the air, and occur in high concentration only after measurable rainfall. Keitt & Jones (1926) showed that liberation of ascospores of the apple scab fungus {Venturia inaeqiialis) is correlated with rain. Hirst et al. (1955) trapped no ascospores of this fungus during dry weather in orchards, and during the first hour after the onset of rain they found only a few ascospores ; yet high concentrations occurred in the second and third hours. Rain at night led to lower concentrations than an equal amount of rain falling by day. In general, perithecia of most ascomycetes possessing them must be wetted before ascospores are 122 THE AIR-SPORA NEAR THE EARTH S SURFACE ejected. Spores of Ophiobolus graminis, the wheat take-all pathogen, do not occur in the air during dry weather, but they reached a concentration of 3,700 per cubic metre in air over wheat stubble within 2 hours of the fall of 1-3 mm. of rain; a few ascospores were liberated by as little as 0-25 mm. of rain (Gregory & Stedman, 1958). Ascospores of some species are evidently discharged when the ground is wet with dew, and these types appear as part of the nocturnal air-spora. Little is known yet about the composition of the damp-air spora, or about the occurrence of spores taking-off in rain-splash, and their study awaits improved technique. Hamilton (1957) studied correlations at two centres (London and Rothamsted) between spore concentration of twent\'-eight visual types and the weather. Her main positive findings are as follows. Rainfall had no effect on the atmospheric concentrations of hyaline basidiospores (including those of Nolanea, Lactarius^ Tilktiopsis^ and possibly Sporo- bolomyces). The concentrations of pollen and of most t}'pes of spore decreased with rain, but all ascospore types and Helicomyces increased with rain. In half of the types studied, concentrations were significantly increased by increases in temperature, dew-point, or relative humidity. The only significant decreases were in grass pollen (and possibly Ustilago spores) with increased relative humidity, and in Nolanea with increased dew-point. Sunshine had no significant effect except for positive corre- lation with Ustilago and algal groups i^Gloeocapsa'). Increased wind significantly decreased the concentration of Alternaria, some basidio- spores (including Ganoderma, Tilletiopsis, and Sporobolomyces)^ Botrytis, Cladosporium, Entomophthora, PuHularia, uredospores, insect fragments, and Urtica pollen. By contrast, plant hairs and algal groups (^Gloeocapsa^) were increased by increasing wind-speed — possibly because both are re- leased by friction. Gustiness was associated with increases in Alternaria^ filiform ascospores, and Ustilago. Conidia of Cladosporium., one of the best studied of the spore t}''pes, show an interesting anomaly in relation to weather. Hamilton (1959) found an appreciable decrease in their number during the hours when rain was falling, but Ainsworth (1952), Hirst (1953), and Gregory (1954) demonstrated a transient increase in concentration of Cladosporium spores when rain started to fall. So far this phenomenon remains unexplained. BIOTIC FACTORS Human activity can also play a part in affecting atmospheric spore concentration. Mowing and tedding of grass can produce a great and immediate local increase in Cladosporium and Epicoccum spores, and (with an apparent delay of 2 hours) of grass pollen (Sreeramulu, 1958). Threshing of grain produces a local spore source (Heald & George, 191 8). The role of overhead irrigation must not be overlooked, and spraying 123 THE MICROBIOLOGY OF THE ATMOSPHERE with insecticides and fungicides has been claimed to spread some fungal diseases. Marine Air The oceans, forming three-quarters of the Earth's surface, act as a vast source, putting a mainly bacterial microbial population into the atmosphere. Compared with air over land, the concentration in surface layers over sea is usually very small. Processes by which marine organisms become airborne include : spray droplets from the breaking of waves on shore or at sea; foam blown off white-caps; and bursting of bubbles produced by white-caps, rain or snow (Blanchard & Woodcock, 1957). These processes, however, also facilitate removal of suspended particles from sea-air by the large liquid surface whose relatively constant tem- perature determines continued up-and-down movement in the lower layers of air. Much of the older work is reviewed, and new data added, by ZoBell (1946). A critical appraisal of the whole subject of aerobiology comes from Jacobs (1951), who calculated, on the basis of salt-concentration of the air and the bacterial concentration of sea water (which seldom exceeds 500 per cc), that the number of marine bacteria in air near the sea surface averages about 5 per cubic metre. The microbial exploration of marine air was pioneered by Miquel (1885, 1886) with the help of a sea captain, Mons. Moreau, during seven voyages. For visual examination of crytogamic spores, an aspirator was worked by suction provided by an engine condenser — an arrangement which sampled 700 litres per 24 hours. Bacteria were estimated by drawing air through glass-wool plugs in tubes at the rate 1,000 litres per 24 hours, washing the plugs, and inoculating aliquots into flasks of liquid beef extract. A total of 113 cubic metres sampled in the seven voyages averaged i bacterium per cubic metre, or o-6 per cubic metre if samples taken within 100 km. of land were excluded. Visual counts over the ocean usually showed a few hundred cryto- gamic spores and many pollen grains per cubic metre (i/30th of the number usual on land), but on one occasion a total of 3,700 per cubic metre were found at a distance of 30 km. from land in a wind off the coast of Senegambia; this comprised a spora very different from that found by Miquel in Paris. Near to continents the winds coming from land always brought impure air, but the sea rapidly purified it and so a broad stretch of water provides an effective obstacle to the spread of contagious epi- demic diseases. In normal weather, bacteria from sea water were not put into the air, but in rough weather Miquel found that the sea air contained a few marine bacteria. The air in a ship's saloon always contained incomparably more microbes than sea air, but its purity increased rapidly in the early days of the voyage until it reached an equilibrium between purification by 124 THE AIR-SPORA NEAR THE EARTH S SURFACE ventilation and contamination by vital activity on board — at a level of perhaps i per cent of that in dwellings in Paris. Nevertheless, Miquel concluded, a ship travels in an atmosphere of self-contamination with bacteria, moulds, and starch grains. On a voyage to the Caribbean, B. Fischer (1886) found very few terrestrial microbes in ocean air — except near major land-masses, where large numbers of bacteria appeared, apparently derived from the soil. Flemming (1908) sampled air on a voyage from Hamburg to Rio de Janeiro and Santos. Of his numerous 20-litre samples taken more than 200 km. from land, two-thirds were sterile — but even at this distance he averaged thirty-four viable spores per cubic metre. These were mostly of moulds and yeasts, though bacteria increased in proportion nearer to land. Although over the sea the air is extremely pure in comparison with air over land, most investigators on board ship have found bacteria, yeasts and mould spores wherever tests have been made. Bisby (1935) exposed Petri dishes on a voyage from Montreal to England and isolated bacteria, Botrytis cinerea^ and Phoma hibernka, all near the coast of Ireland. The microbes of marine air have been studied at the Scripps Insti- tution of Oceanography, California, by ZoBell & Mathews (1936) and ZoBell (1942). They claimed that less than 5 per cent of bacteria in sea water will grow in freshwater nutrient media, and a still smaller percen- tage of freshwater bacteria will grow on sea-water media. Petri dishes of nutrient media made up with distilled water (FW) or sea water (SW) were exposed horizontally at distances of up to 1,600 metres inland from a sea wall during a sea breeze of 5-8 m.p.h. (2-6 metres per sec). The ^SW count decreased and the 'FW count increased with the distance inland, the 'SW'/'FW ratio decreasing steadily from 10-20 at the sea wall, to i-o at 400 metres, and 0-5 at 1,600 metres inland. The number of mould spores usuallv increased with increasing distance from the sea. In a land breeze, littoral spray puts into the air salt-water bacteria which can be detected at up to 8 km. out to sea, after which the ratio 'SW'/'FW goes down to i-o owing to the predominance of terrestrial bacteria in the air for distances of 160 km. out to sea in fine weather. Exceptionally at 880 metres height on Mt. Woodson, 32 km. inland, plates exposed in a sea breeze following rain gave a ratio of 'SW/'FW = 2-06, which was interpreted as indicating a predominance of marine bacteria in the air in a region where soil bacteria normally predominate. It has been calculated that an average of 127 cubic miles of sea water is put into the Earth's atmosphere each year in the form of splash droplets, and this would provide an average of only about one marine bacterium per square centimetre of the Earth's surface per year — a small quantity compared with the deposition rate from the land air-spora (ZoBell, 1942). Although all workers agree that marine air contains extremely few bacteria, ZoBell points out that the use of sea-water media might be 125 THE MICROBIOLOGY OF THE ATMOSPHERE expected to increase the counts of earlier worliers by from lo to 20 times. On these media gram-negative bacilH predominate, there are few cocci, no vibrio or spirilla forms, and fewer than half were spore-formers. A pink yeast has been reported from ocean air by several workers. This spora contrasts strongly with the gram-positive rods, spore-formers, cocci, Bacillaceae, and Micrococcaceae which, with mould spores, are normally abundant in air over land (ZoBell, 1942). Rittenberg (1939) sampled in (presumably horizontal) Petri dishes at 21 metres above the deck of the vessel E. W. Scripps off the Pacific Coast of California. Contrasts between sea-water and fresh-water media were not so clear as in previous tests, and the numbers varied widely at different stations; but, on the average, moulds decreased in numbers with increasing distance from land (Table XXII). TABLE XXII NUMBERS OF MICROBES AND DISTANCE FROM LAND. SUMMARY OF 25 PETRI DISH EXPOSURES AT 21 METRES ABOVE DECK (Rittenberg, 1 939). Distance Average number of colonies per hour of from land exposure (on 4 Petri dishes of each inoculum). km. Sea-water Medium Tap-water Medium Bacteria Bacteria (and yeasts) Moulds (and yeasts) Moulds 0-20 45 115 20 200 20-300 48 79 13 69 300-800 71 20 39 36 Detailed examination of 100 bacterial and yeast cultures taken at random by Rittenberg showed that 32 per cent were yeasts, 30 per cent cocci, 15 per cent gram-negative rods, and the remaining 23 per cent gram-positive spore-forming rods. They included: Bacillus subtilis^ B. flavus^ B. megatherium^ B. mjcoides, B. tufnescens, B. cohaerens, B. laterosporus, Flavobacterium aquatilis^ Achromobacter liquifaciens^ Staphylo- coccus aureus, S. albus, S. citreus, Micrococcus Jlavus, M. candidus, and Sarcina jlava* One hundred mould cultures included: Cladosporium (Hormodendrum), 22 per cent; Penicillium, 18 per cent; Alternaria — Macrosporium — Stetuphylium, 11 per cent; Cephalosporium, 7 per cent. Others identified included : Plenozythia, Catenularia, Spicaria or Paecilo- myces, and Trichoderma. To avoid the usual embarrassment of the micro- biologist, all sampling in free air seems to have been done with sugar-free media; otherwise, moulds might have been many times more abundant. Rittenberg points out that this marine airborne flora is unlike that of sea water itself, but resembles the air-spora over land. * But additional information by ZoBell (1942) indicates that these identifiable species were all from fresh-water plates. 126 THE AIR-SPORA NEAR THE EARTH S SURFACE Although the air-spora over the sea is clearly largely of land origin, Petri dish sedimentation tests are difficult to interpret and cultural and volumetric work is needed on the contribution from the ocean itself — both in the zone of littoral influence studied by ZoBell, and far from shore. The mode by which the ocean purifies the air flowing over it, and the fate of airborne spores trapped by the ocean, are still obscure. Earlier workers, including Miquel and B. Fischer, found marine air almost free from pollen, but this freedom is now seen to be only relative. Erdtman (1937) operated a vacuum-cleaner filter trap at the masthead of the M.S. Drottningholm during a voyage from Gothenburg to New York extending from 29 May to 7 June, 1937. Compared with the average of 180 pollen grains per cubic metre recorded during spring at Vasteras (no km. west of Stockholm), he found only o-i8 per cubic metre in the North Sea, and 0-007 P^^ cubic metre in mid-ocean, with an increase again on approaching North America. Temporary higher concentrations ('pollen rains') occurred three times: oi Pinus (0-13 per cubic metre) in the North Sea; of Almis viridis (0*045 P^^ cubic metre) and Cyperaceae (o-oo6 per cubic metre) at a distance of 250 to 600 km. off" Newfoundland; and of combined grasses, Plant ago^ and Rumex (totalling o-i per cubic metre) at 220 to 300 km. from Nova Scotia and Massachusetts. During strong western and north western winds, about mid way between Iceland and Ireland, Erdtman caught tree pollens {Alnus, Betula^ Coryliis^ Jiiniperus, Myricci^ Picea, Pinus, Popiilus, Qiiercus^ Salix, Tilia, Ulmus) and herb pollens (Chenopodiaceae, Cruciferae, Cyperaceae, Ericaceae, Gramineae, Plantago^ Umbelliferae, and Urtica)^ as well as spores of Dryopteris and Lyco podium clavatiim. Erdtman's volumetric sampling firmly establishes the occurrence of pollen in small but measurable concentration near the surface of the sea right across the Atlantic, and there is no reason to doubt that the land air- spora extends to all parts of the globe. Confirmatory evidence comes from Transatlantic sampling by Dyakowska (1947) and Polunin (cf. 1955). Bishop Rock Lighthouse stands on a low rock at the southwestern extremity of the Scilly Isles, which are a group of small, rocky islands with few trees (mainly Ulmus and Pinus). Gravit}-' slide sampling on the lighthouse platform 38 metres above sea-level (Hyde, 1956) showed mainly pollen of Betula^ Qiiercus, and Fraxinus, with some Phtus. The total tree- pollen deposit was quite large (2,800 per sq. cm. per year, compared with 2,000 at Aberdeen and Brecknock Beacons, and 10,000 at Cambridge). The proportion of tree pollen of Bishop Rock was 27 per cent, and this is t}''pical of country areas in Britain (in to\Mis it may reach 50 per cent). It is remarkable that the greatest deposition of grass pollen recorded for any centre in Britain during Hyde's gravit}^ slide survey was 1,679 grains per 5 sq. cm. at Bishop Rock on 29 June 1953. Whether this resulted from high concentration, or from high efficiency of turbulent deposition in strong winds, is not yet clear. 127 THE MICROBIOLOGY OF THE ATMOSPHERE Sreeramulu (1958^) used a Hirst trap at 70 ft. above sea-level on a voyage in the Mediterranean in October and November, 1956. At 5 to 50 miles from land he found an average of 56-4 fungus spores and i-6 pollen grains per cubic metre. In Malta Harbour the concentrations v^ere 121 and 1 2 per cubic metre, respectively. At sea, Cladosporium predominated with 16 spores per cubic metre, followed by smuts with 5 per cubic metre, and coloured basidiospores with 7 per cubic metre. Also of interest was the occurrence of spores oi Helminthosporium^ Alternaria, Torula herbarum^ Nigrospora^ Curvularia, and Epicoccum^ as well as hyphal fragments. Pollen in marine air must come from land plants; the mould spora is more characteristic of above-ground sources than of the soil; the bacteria, however, may well come largely from sea water and soil. The Air of Polar Regions The air of polar regions seems to be still purer than that over the sea. Levin (1899), who aspirated air through powdered-sugar filters, obtained only three bacterial colonies and a few moulds in a total of 20 cubic metres of air sampled at various points in Spitsbergen (Svalbard). During 2 years on an island near Graham Land, Antarctica, Ekelof (1907) exposed Petri dishes at intervals; 40 per cent of them grew bacteria, which he thought came from the soil. On the average, one colony arose per 2-hours exposure. Pirie (191 2) exposed Petri dishes in the 'crow's nest' of the Scotia in the Weddell Sea, Antarctica, during the summer of 1903, for as long as 20 hours, and also on a glacier at Scotia Bay during winter; they all remained sterile. E. Hesse (191 4) exposed Petri dishes while at sea south of Spitsbergen and also found the air to be almost sterile. Darling & Siple (1941) exposed jars and dishes of media in remote places in Marie Byrd Land, Antarctica, and from their isolations identi- fied: Achromobacter delicatulum^ A. liquidum^ Bacillus albolactis, B.fusi- fonnis, B. mesentericus, B. subtilis, and B. tumescens. They concluded that, although some bacteria had been brought to Antarctica by man and migrating animals, the vast majority must have come as atmospheric dust in subsiding air. Recent work in the Arctic has demonstrated a fair range but sparse 'population' of microbes to be present in the air in summer near ground- and sea-level in various parts of those vast regions. Polunin (1954, 1955) organized the exposure of sticky slides at several points ranging eastwards from Point Barrow, Alaska, to Spitsbergen, in 1950, and found a consider- able variety of pollen grains and 'probable moss spores' at each station. Remarkably enough the pollen grains caught most plentifully through most of that summer in Spitsbergen were of Pinus — several hundreds of kilo- metres from their nearest possible source. In 1954 Polunin (1955^?) was responsible for the exposure of sticky slides (cf. Polunin, i960) off the 128 THE AIR-SPORA NEAR THE EARTHS SURFACE north coast of Ellesmere Island and, in 1955, for the exposure of sticky shdes (Barghoorn, i960) and Petri dishes of nutrient media (Polunin et al. i960; Prince & Bakanauskas, i960) on Ice-Island T-3 when it was floating in the North Polar Basin at about 83°N. Here again the air-spora was found to be very sparse compared with middle latitudes, but it included pollen grains which, in some instances, were indicative of long- range transportation (cf. Barghoorn, i960, p. 91, despite p. 88). From the T-3 exposures a small number of slow-growing fungus isolates (all identified as PeniciUiiim viridicatum) and more of Actinomycetes were obtained, but none of bacteria. The Origin of the Air-Spora There is no reason to doubt the conclusion of Miquel and of Proctor (1935) that most bacteria of the air originate from the soil, or from the oceans (ZoBell, 1946). But it is doubtful whether the soil makes any sub- stantial contribution to the fungus spore content of the atmosphere, as has been argued by some writers. It seems more likely that this air-spora is derived predominantly either from moulds, plant parasites, and other fungi growing on vegetation, or from surface-growing fungi equipped with explosive mechanisms which project their spores into the freely- moving turbulent air layer. In the soil, bacteria, Penicillia, and Aspergilli predominate; but Cladosporium predominates in the air, seconded by basidiospores. Similarity between the soil- and air-sporas results mainly from the soil being the ultimate 'sink' to which most of the spores of the air are destined. Much of the air-spora comes from wild vegetation. Industry pollutes the atmosphere mainly with inorganic particles and gases. It is not generally appreciated, however, that agricultural practices may pollute the air with plant pathogens and with respiratory allergens on a large scale. Even such small operations as mowing grass may cause a local increase in the Cladosporium content of the air by a factor of 20 times, as I have found in recent tests (and cf. Sreeramulu, 1958). For microbes to get into the air in the high concentrations observed at peak seasons, a take-off mechanism is necessary. However, the most unlikely organisms occasionally get into the air. Siang (1949) isolated one colony of the aquatic phycomycete, Hypochytrium catenoides, from air on a roof at McGill University, Montreal, Canada. Probably almost every kind of microbe would be found if sampling were continued long enough. From recent work on the air-spora near the ground, we learn that its composition and concentration often fluctuate enormously — sometimes w ithin quite short time-intervals. The significance of this for plant path- ology and plant breeding is obvious. Some constituents, such as the grass pollens, are important in respiratory allergy (the volume of air inhaled by the human lung is of the order of i cubic metre per hour). In the course I 129 THE MICROBIOLOGY OF THE ATMOSPHERE of 24 hours we inhale perhaps 50 micrograms of a mixture of microbes. Ahhough some constituents of this mixture are harmful to allergic subjects, this dose might also bring in useful quantities of active organic compounds. Chauvin & Lavie (1956) found antibiotics in Salix and maize pollen, and their presence in fungus spores, too, would not be surprising {see Whinfield, 1947). We may wonder whether the reputed beneficial effects of country air for human health are attributable not only to its freedom from smoke and fumes, but also to positive gains from the air- spora. Finally, to balance loss with gain, as the Earth's surface is the ultimate receptacle of almost all spores liberated, it can be estimated that the soil must receive a dose of fertilizer from the air-spora equivalent to perhaps 2 kg. of nitrogen per acre per annum — an amount that is negligible on fertile land, but enough to aid plants colonizing barren places. 130 Woods HOLE Mass.* X THE UPPER-AIR SPORA The air-spora near the ground is dominated by fluctuations in its im- mediate local sources. In the upper air, however, the effects of local sources are smoothed out and attention can be focussed on organisms under- going long-distance transport. Concentrations in the upper air are sparse, the necessity of keeping samples free from contamination is paramount, and sterile technique for enumerating the microscopically small particles becomes exacting when they are exceedingly dilute. Vertical Diffusion Whereas spores and pollen grains are heavier than air, and tend to fall under the influence of gravity, atmospheric turbulence and convection tend to work in the opposite direction. As a result, the atmosphere is in a sense a spore suspension that generally decreases in concentration from ground-level up to the base of the stratosphere. Eddy diffusion will bring spores to the top of the outer frictional turbulence layer: above this, convection will operate and, in the upper part of the troposphere, we would expect to find mostly components of the day-time spora. When it first became possible to explore the air overhead, it was a matter of surprise to find how far up microbes could go. As methods have been developed for exploring greater and greater heights, we can begin to form a picture of the changes in concentration with height and of the circulation of spores of micro-organisms over the surface of the globe. Evidence that concentration decreases with increasing height comes from two distinct sources of information which have often been confused : (i) observations at a standard height above local ground-level at a chain of stations differing widely in altitude above sea-level; and (2) observations at widely different altitudes above local ground-level at a single station. GROUND STATIONS AT DIFFERENT ALTITUDES AB0\T: SEA-LEVEL Observ-ations in this category are extremely fragmentary and have the flavour of holiday tasks on fine days in summer. Samples at various altitudes are taken at successive times as the climber reaches a suitable station— as in Pasteur's visit to the Mer de Glace, where the relative purity of mountain air was convincingly demonstrated {see p. 4). Using a volumetric method, Miquel (1884, p. 524) confirmed this conclusion. 131 THE MICROBIOLOGY OF THE ATMOSPHERE In July and August, while bacterial concentrations of 55,000 per cubic metre were current in the air of the Rue de Rivoli in Paris, and 7,600 per cubic metre in the Pare Montsouris, Miquel found only eight bacteria (and numerous moulds) per cubic metre at a height of i metre above the surface of a field at Lake Thun (570 m. above sea-level), and none at stations between 2,000 metres altitude and the summit of the Eiger at nearly 4,000 metres. Miquel attributed this purity to the effect of reduced atmospheric pressure doubling the volume of air and diluting its dust load, to the rarefied air less easily holding particles in suspension, and to the absence of local sources of contamination — especially in the regions of perpetual snow. In similar volumetric data from the Dauphine Alps recorded by Bonnier et al. (191 1), bacteria decreased with height more rapidly than moulds. No one has yet compared concentrations at different heights above ground-level over plateaux with those over mountains, or over flat and convex surfaces at the same altitude. The purity of the air in regions of perpetual snow is understandable, but it is surprising that air at one or two metres above ground-level in mountain valleys should also contain so few microbes. Geiger (1950) envisages mountain slopes as covered with a skin of air having the usual characteristics of air near the ground but easily removed by wind and convection — except where protected by vegetation. Convex surfaces generally have a more extreme climate than flat surfaces, and concave surfaces are still more equable. THE ROLE OF TURBULENCE The role of turbulence in diffusing spore-clouds vertically was first emphasized by Schmidt (191 8, 1925), although the theory was developed for heat transfer by Taylor (191 5) with information derived from tem- perature records over the Great Banks of Newfoundland. Schmidt argued that, when a stable state of diffusion by eddies has been reached, the number of particles falling under the influence of gravity across any horizontal boundary is compensated for by the number of particles moved upwards by diffusion, and so the concentration of particles in the air should decrease exponentially with increasing height according to the equation VsZ X = Xo exp. - -j- where xo = concentration at height z = 0, Vs = terminal velocity of fall, A = Schmidt's 'Austausch' or intermixing coefficient which is assumed to be invariable with height. The total spore content of the column of air standing above i sq. cm. 132 THE UPPER-AIR SPORA would be A =^ xo^l'^s- Concentrations should give a straight line when plotted against the logarithm of the height. The method of approach to the problem suffers from two defects in practice. The coefficient for diffusion (Schmidt's 'A' or Taylor's 'K') is not invariable with height, and it is doubtful whether a steady state is ever reached with the great diurnal changes occurring when the source consists of living organisms. C. G. Johnson & Penman (195 1) supposed that the vertical distribu- tion of aphids at any one time is determined by the net effect of upward transport by turbulence and downward transport by the combined action of gravity and biological impulse — the mean clearance rate. If x is the concentration at height z, and co is the 'mean clearance rate', they deduced that a graph of log x against log z should yield a straight line. Attempts have been made to fit empirical curves to observational data on vertical gradients. Wolfenbarger (1946, 1959) used regression equa- tions of the t}'pe: Y = a + b log x + c/x. C. G. Johnson (1957) fitted records of insect-trap catches with: f(z) = C(z + Ze)""^, where f(z) is concentration at height z, C is a scale factor depending on population size, A is an index of the diffusion process and the profile, and Ze is a parameter whose significance probably depends on the rate of exchange of insects between the air and the ground. Particles entering the air near ground-level become mixed throughout the layer of frictional turbulence so long as the wind blows. Convection provides a local intermittent mechanism which distributes spores from the ground layer throughout the troposphere. Observed vertical concen- tration gradients sometimes fit theoretical lines quite adequately, and they may well describe long-term averages. But theoretical treatments of this problem are often unsatisfactory, especially in failing to predict concen- trations in the first few hundred feet. The ideal situation is seldom realized, because conditions change too rapidly for a stable state to be attained. Wind velocity increases with height, and layers of air at different heights in a vertical column at any one time will previously have been over different places at different times. The thickness of the turbulent layer of air is always changing. Biological factors in a diurnal cycle put vastly differing numbers of organisms into the air at different times, and vertical concentration is continually building- up or decaying. Temperature inversions will affect vertical diffusion; according to Jacobs (195 1), 'The presence of a stable layer at the surface will prevent or retard the introduction of surface organisms into the upper atmosphere but will, at the same time, maintain higher concentrations of organisms in the surface layers ; the presence of a discontinuity surface in the upper air will limit vertical transport in either direction, resulting in the concentration of organisms above or below such a surface'. Con- centration will often start to decrease from the active surface of a crop and not from true ground-level. 133 the microbiology of the atmosphere Early Studies of the Upper Air Measurements of microbial concentrations at heights above ground- level were first attempted from towers and tall buildings by Miquel (1883), Carnelley et al. (1887), and, more recently, by Kelly and others (cf, p. 143). Miquel found that the bacterial content of the air at the level of the Lanterne of the Pantheon in Paris was only i/20th of that in the street below. Probing upwards into the atmosphere for microscopic life started dramatically when the Manchester physician Blackley (1873) used two kites in series to lift sticky microscope slides to a height of 300 metres and caught from 15 to 20 times as much pollen as on slides similarly orientated at 1-4 metres above the ground. Kites were also successfully used in India by Mehta (1952) to catch spores of the cereal rusts, Puccinia graminis, P. triticina^ and P. glumarum^ and small balloons were used for the same purpose by Chatterjee (1931). SAMPLING from BALLOONS Cristiani (1893) obtained bacteria and a few moulds by volumetric sampling from a balloon at up to 1,300 metres above Geneva (at a total of 1,700 metres above sea-level). He was obviously puzzled by his results, which he regarded as inconclusive, and he attributed most of his catch to contamination from the surface of the balloon and its rigging — remaining convinced that the upper air is extremely pure. The credit for first demonstrating the existence of a microbial popula- tion in the upper air should probably go to the mycologist Harz (1904), who sampled during a balloon ascent over southern Bavaria on a sunny morning in March. At altitudes of between 1,500 and 2,300 metres he aspirated air through a Miquel-type filter of powdered sodium sulphate by suction obtained with a horse's stomach-pump; culturing the catch in nutrient gelatine, he found a few moulds, and bacterial concentrations ranging from 179,000 to 2,870,000 per cubic metre. At 1,800 to 2,000 metres there was a zone with 16 times the concentration at 1,500 metres and 5 times that at 2,300 metres. These phenomenally large bacterial concentrations were associated with a large temperature lapse and strong convection from hot dry soil. Moulds were identified as: PenicilUum glaucum, P. cinereum, P. atro-viride^ Sporidesjiiium sp., Acreinonimn alter- 7tans, Mucor racemosus, M. mucedo^ Oospora ochracea, 0. ferruginea, Perkonia arta^ Hormodendrum {Cladosporium) penicillioides, Arthrococciis lactis, Aspergillus niger, and a sterile mycelium. During ascents from Berlin with both captive and free balloons, Flemming (1908) used trapping methods similar to those of Harz. He found viable microbes up to 4,000 metres, averaging 370 per cubic metre above 500 metres, and 12,900 per cubic metre lower down. Sterile 134 THE UPPER-AIR SPORA samples were rare. Concentrations were not uniform but increased strik- ingly at the level of the cloud base. Species identified included : Alicro- coccus radicatus, M. albus, M. Jiubilus, M. aerogenes, Bacillus ubiquitus, B. aurescens, B. aureo-Jlavus, B. terrestris, B. aerophilus, B. submesenter- oides, B. mycoides, and PeniciUium crustaceum — all spore-formers. Above 2,500 altitude he found Bacillus terrestris, B. aerophilus, and Sarcina lutea, and, at 4,000 metres. Micrococcus citreus, M. luteus, and PeniciUium crustaceum. Flemming commented on the frequency of pigment-formers among the bacteria and yeasts of the upper air. From catches made during balloon flights over southern Germany, Hahn (1909) concluded that, on the average, bacterial and dust counts run parallel with each other and decrease with height because of sedimentation. He claimed that the air above a certain height was germ-free, and that this zone was lower in winter than in summer. THE STRATOSPHERE We would expect the stratosphere to be almost devoid of organic particles, because of the apparent inadequacy of mechanisms able to carry them above the top of the troposphere. At great heights the intensity of radiation would be unfavourable to survival. However, well-documented evidence is worth more than theory and, for the present, we must admit that we know nothing of the possibilities of life in the stratosphere. The only attempt to sample the stratosphere known to me was made with a balloon. Rogers & Meier (1936, 1936^) devised a sampler to be opened and closed by an aneroid between 21,000 and 11,000 metres during the descent of the Balloon 'Explorer H'. They obtained five bacterial cultures, all of which were different species of Bacillus., and five fungi {Rhizopus sp., Aspergillus fiiger, A.fumigatus, PeniciUium cychpium, and Macrosporiutn tenuis) ; in all, these were equivalent to approximately 0-14 viable organisms per cubic metre. SAMPLING FROM AEROPLANES Exploration for microbes in the upper air from aeroplanes was started in 1921 when Stakman et al. (igi^) exposed Vaseline-coated slides over the Mississippi Valley to trap cereal rust spores. Flights from Texas as far north as Minnesota and at altitudes up to 3,300 metres yielded numerous pollen grains and fungus spores, among which Alternaria (often in chains) were most numerous, followed by Puccinia, Helminthosporium, Clado- sporium^ Cephalothecium., Ustilago^ Tilletia^ and Scolecotrichum. Among rust spores the uredo forms predominated, but some teleutospores and aecidiospores were also caught. Spores became relatively scarce at alti- tudes above 3,000 metres. At 5,400 metres (the highest altitude tested), two uredospores of Puccinia triticina were caught. Alternaria from alti- tudes of 1,000 to 3,000 metres germinated readily, as also did uredospores from 2,300 metres. 135 THE MICROBIOLOGY OF THE ATMOSPHERE In the Slimmer and autumn of 1923, Mischustin (1926) exposed Petri dishes of nutrient agar in flights from Moscow. On rather slender evi- dence obtained from tests in a wind-tunnel, he concluded that he was sampling 20 litres of air per minute — probably a large underestimate. The plane was first flown above the layer to be sampled to free it from ground dust, and then lowered to the required height. At 500 metres the num- bers of bacteria increased in windy weather to 7,000 or 8,000 per cubic metre, from Mischustin's normal of 2,000 to 3,000 at this altitude. Micrococcus and Sarcina decreased greatly in calm weather, but the number of bacterial rods and of moulds increased. At 1,000-2,000 metres, numbers were small. The proportion of spore-forming bacteria, of moulds, and of actinomycetes, was greatest at the greater heights. The concentration of organisms over the city of Moscow at 2,000 metres included an average of 650 bacteria per cubic metre, and was from four to five times as great as that over the surrounding countryside. Pollen was found at up to 5,800 metres over the Mississippi, with the greatest concentrations often at from 300 to 1,100 metres (Scheppegrell, 1924, 1925). Craigie & Popp (1928) caught wheat-rust spores at up to 3,000 metres over the Canadian prairies. In flights from Cambridge, England, Weston (1929) found that fungi and bacteria were abundant up to 3,000 metres, but relatively scarce above this altitude. Air within clouds tended to contain more bacteria and fungi than did air above or below clouds — a phenomenon noted by other workers, including Heise & Heise (1948). On flights up to 2,200 metres over the arid lands of southern Arizona, Browne (1930) isolated 'white and grey' bacterial colonies, Aspergillus, Penicillium., Alternaria, and yeasts; but no spores of wheat rusts were observed on slide spore-traps. Cotter (1931) studied dispersal of wheat rust in flights near Lake Michigan, trapping on oiled microscope slides. Uredospores were not more numerous during rain than in fine weather; fewer were caught over Lake Michigan than over near-by land, and more were caught over areas abounding in barberry (the alternate host of the parasite). MacQuiddy (1935) exposed Petri dishes and slides in flights up to 2,100 metres over Omaha, Nebraska. Pollen was abundant up to 900 metres, and bacteria and mould spores began to decrease between 1,200 and 1,500 metres. MacLachlan (1935) made flights in early May over Massachusetts, to trace spores of the juniper rust {Gymno sporangium biseptatum). Petri dishes exposed over the side of the 'plane for i minute gave viable spores up to 600 metres (the maximum height tested), although numbers and viability decreased steadily with height. (i) Aerobiological work of F. C. Meier. Fred C. Meier of the United States Department of Agriculture planned an extensive investigation of the upper-air spora. Unfortunately, when only preliminary abstracts of his work had been published, he was lost on a flight over the Pacific Ocean (Haskell & Barss, 1939). 136 THE UPPER-AIR SPORA Sticky slides were exposed by Colonel Charles A. Lindbergh in special containers at about i,ooo metres altitude during a flight between Maine and Denmark. Material trapped over Davis Strait and East Greenland included algae, fragments of insects' wings, diatoms, and possibly sponge spicules, volcanic ash, and glass. Fungus spores were tentatively identified as belonging to: Macrosporium^ Cladosporium, Leptosphaeria^ Myco- sp/iaerella, Trichotlieciiim^ Helicosporium, Uromyces, Camarosporium, and VentKria. Some of these were abundant over Maine and Labrador but diminished over Davis Strait, the ice-cap of Greenland, and Denmark Strait (Meier, 1935, 1935^7). Flights over the United States at from 150 metres to 5,500 metres showed a varied spore 'population' which usually decreased in both numbers and varietN'' above 2,400 metres. \'iable spores of Pestahzzia were caught above Washington at 5,500 metres on 22 March 1932. Other genera recognized included: Acremonklla, Alternaria (Macrosporium), Aspergillus, ChaeTomium, Cladosporium, Coniothyriimi, Dejnatium, Epi- coccuniy Fumago, Fusarium, Helniinthosporium, Penicillium, Sclerotinia, StachybotrySy Stemphylium, and Trichoderma (Meier et al, 1933)- After flights over the Caribbean Sea, Meier (1936) came to feel that trade-winds might be important in disseminating microbes. Viable spores were found at 800 to 1,200 km. from land; but after rain squalls, Petri dishes sometimes remained sterile after exposure at 60-240 m. altitude a few kilometres to the leeward of islands — so demonstrating that showers remove spores from surface winds. Sugar-beet pollen was trapped on agar plates during flights over a small sugar-beet seed-growing area of 900 acres in New Mexico. Viable beet pollen, mixed with pine pollen and fungus spores, occurred up to 1,500 metres, which was the greatest height tested and the level of the dust horizon (Meier & Artschwager, 1938). (ii) Vertical gradients. During an epidemic of wheat rust in Manitoba in July and August 1930, Peturson (1931) trapped spores at different altitudes in eight aeroplane ascents. The average numbers of spores caught per square inch of trap surface (presumably with comparable exposure times) were: 305 metres, 10,050 spores; 1,520 metres, 1,180 spores; 3,050 metres, 28 spores; and 4,260 metres, 11 spores. By substi- tution in Schmidt's equation {see p. 132) we find A = 5-8 X lo"*, if Vs = I cm. per sec. Hubert (1932) trapped spores during two flights at the time of an epidemic of yellow rust of wheat at Halle in Germany. During the second flight, for which data are more extensive, the numbers of spores trapped per square centimetre per minute of exposure at various heights were : 30 metres and less, 1,418 spores; 400 metres, 683 spores; 600 metres, 336 spores; and 800 metres, 82 spores. Substitution in Schmidt's equation gives A = 1-5 X 10^. Similar values for A were indicated when tree pollen was trapped over 137 THE MICROBIOLOGY OF THE ATMOSPHERE German forests in spring during a series of aeroplane flights by day and by night (Rempe, 1937). In general, with light to moderate windy weather and with cumulus clouds at about 2,000 metres, the pollen concentration decreased only slightly up to 1,000 metres, and the maximum number of grains might occur as high as 200 or even 500 metres. This was regarded as a sign of a complete inversion of air masses. A similar distribution also occurred under high pressure conditions without clouds but with strong- thermal turbulence. By way of contrast, conditions associated with a stratified cloud-layer and high wind velocities showed a marked decrease of pollen with height. In night flights, the maximum number of grains was often reached at a height of about 200 metres, i.e. above the temperature inversion which often develops at night. At night the numbers trapped usually decreased with increasing height much more than by day. The total numbers trapped at all heights were also fewer by night than by day. From Rempe's data the mean numbers of pollen grains trapped per 1-275 ^^- ^^i. of trap surface per 20 minutes for all flights which extended up to 1,500 metres were : Altitude (metres) : 10-40 200 500 1,000 1,500 Day flights 904 849 852 581 267 Night flights 577 560 283 85 45 These records include pollen grains of various species; but, taking Vs = 3 cm. per sec. as a moderate value for the speed of fall of pollen grains, it can be shown that, for altitudes above the zone affected by strong thermal turbulence and temperature inversions, Schmidt's in- terchange coefficient A = 2-6 X 10^ for day flights, and r6 x 10^ for night flights. This provides further evidence of the appropriateness of considering the spore or pollen cloud as a suspension in air. Although at heights of about 1,000 metres and upwards the average distribution agreed well with that expected from terminal velocity balanced by eddy diflfusion, near ground-level the suspension tended to become more uniform than predicted, owing to the intermittent stirring of the lower layers by strong mechanical and diurnal thermal turbulence. (iii) Sampling the upper air over the United States. In the upper convective layer. Walker (1935) exposed Petri dishes of blood agar and, after sampling an estimated 2,400 cubic metres of air, he concluded that the atmosphere in that layer was sterile (two cultures of Staphylo- coccus aureus were reasonably enough ignored as contaminants). However, Proctor & Parker (1942) suggested that Walker's agar surfaces may have been frozen and non-adhesive, because their own researches at the Massachusetts Institute of Technology showed that the upper air over the United States was far from sterile. In trapping from aeroplanes. Proctor & Parker used filters of lens paper supported on wire gauze in brass tubes connected with the free 138 THE UPPER-AIR SPORA air and sampling about 28 litres per minute. The catches were examined both microscopically and by culturing. Bacteria averaged 12 per cubic metre on all flights, and 9 per cubic metre at 6,100 metres or higher. There was sometimes evidence of a zone of greater concentration at a height of several thousand metres. Moulds were usually less numerous than bacteria, but the nutrient agar on which filter washings were plated was recognized as unfavourable to mould growth. Bacteria were mostly spore-formers and those identified included species of Bacillus, Achro- mobacter, and Micrococcus. Among moulds, Aspergillus and Penicillium predominated, occurring with some other Fungi Imperfecti including Cladosporium {Hormodcndrum) and Fusarium, as well as Mucoraceae, Actinomycetes and, occasionally, yeasts. Pollen was found on only three flights (Proctor, 1934, 1935). The highest mould count obtained in the M.I.T, studies occurred at an altitude of 200 to 300 metres in May over a wooded area, where 22 bacteria and 260 moulds per cubic metre were recorded. Particularly large counts of bacteria and moulds occurred during a dust storm which apparently came from Nebraska and South Dakota — the same dust- storm during which Soule (1934) recorded mass invasion of his labora- tories in Michigan by Bacillus megatherium. During this dust-storm, at an altitude of 1,500-3,300 metres over the Boston area, bacteria totalled 140, moulds 44, and dust particles 2,800, per cubic metre, respectively. However, during the whole survey, dust particles were over 100 times as numerous as viable microbes — suggesting that much of the dust came from industry and combustion rather than from the soil (Proctor & Parker, ^938)- ... Petri dishes of nutrient agar w ere exposed during flights at from 300 to 3,250 metres over Nashville, Tennessee, during winter by Wolf (1934). On this medium bacteria outnumbered moulds, the bacilli contributing 37-7 per cent, non-spore-forming rods 24-6 per cent, and cocci the remaining 37-7 per cent, of the total bacterial count. The bacteria were very similar to those found by Proctor (though with a smaller percentage of spore-formers) and further study supports the general conclusion that aerial bacteria are of types commonly found in soil and water, are generally unable to ferment common sugars with the production of gas, and are unable to produce indole. From these flights by Wolf, Actinomyces griseolus was isolated twice, at 700 and 1,400 metres, and A. phaeochromogenus once, at 620 metres. A pink yeast was found at 1,750 and 3,050 metres. Fungi isolated, with their percentage frequencies, included: Fusarium, 29; Alternaria, 22; Clado- sporium {Hor?nodejidru?ii), 20; Verticillium, 5; Aspergillus, 3; Penicillium, 1-6; and among others w^ere Acladium, Brachysporium, Cephalothecium, Chaetomium, Hebninthosporium, Macrosporium, Mucor, Oospora, Pleno- zythia, and Scopulariopsis. The large numbers of Fusarium spores and small numbers of Aspergillus and Penicillium spores are remarkable. 139 THE MICROBIOLOGY OF THE ATMOSPHERE The average of all samples gave a concentration of 7-5 cultivable organisms per cubic metre and varied from none at 850 metres altitude in De- cember to a maximum of 42 per cubic metre at 460 metres in October. In general the concentration decreased with increasing height, but on 25 January there was a zone of high concentration at 900 to 1,200 metres altitude. SPORES OF GREEN PLANTS IN THE LOWER TROPOSPHERE Using a glass-wool filtration apparatus, Overeem (1936, 1937) sampled from aircraft over the Netherlands on six occasions extending from July to October at heights of 100, 500, 1,000 and 2,000 metres. Filter washings were inoculated to Pringsheim's culture solution for green plants and kept in the light. From a total of about 28 cubic metres of air she obtained the following cultures. Algae: Chlorococcum sp., 9; Phormidium luridum var. nigrescens, Chlorella vulgaris^ Pleurococcus vulgaris^ and Stichococcus bacilhiris, 3 each; Aphanocapsa sp., 2; Actin- astrum sp., Stichococcus jninor^ and Hormidium fiaccidum^ i each. Moss: Funaria hygrometrica^ 2 (from 500 and 1,000 metres). Fern: i (unidenti- fied, from 500 metres). Total numbers at the various altitudes were in the ratios 5 : 10 : 3 : 3 at 100, 500, 1,000 and 2,000 metres, respectively. This work is of particular interest as one of the few demonstrations that spores of green plants invade the troposphere in fiiir numbers and variety. McGiLL University Studies FLIGHTS OVER THE ARCTIC Extensive exploration of the upper air by aeroplane was initiated by Polunin at McGill University, Montreal, in 1947, and continued until 1 95 1 with the co-operation of the Royal Canadian Air Force and the United States Air Force. Flights during 1947-49 ^^'^^^ primarily directed to the study of arctic conditions. In the summer of 1947 flights w^ere over the Northwest Territories northwards to Cape Bathurst, then north-east from Cambridge Bay to Victoria Island to beyond the region of the north magnetic pole and back, and finally south-west from Cambridge Bay to Yellowknife and to Ed- monton, Alberta. Petri dishes with nutrient medium, and also sticky slides, were exposed from his planes by hand, mostly at about 1,500 metres altitude (Polunin et a I., 1947, 1948). There were small but meas- urable concentrations of fungus spores, and the composition of the air- spora appeared to depend on the origin and sometimes on the trajectory of the air mass rather than upon the locality of sampling (Polunin, 1951, 1951a, and cf. 1954). The bacteria were identified as: gram-positive rods, about 40 per cent (two thirds of which morphologically resembled Corynebacterium), Micrococcus (23 per cent), Achromobacter or Flavo- bacterium (17 per cent), spore-formers (4 per cent), and Sarcina (3 per cent). Fungi identified in culture included Cladosporium (over 40 per cent 140 THE UPPER-AIR SPORA of the total), Sporormia^Pulhdaria, VerticiUium, Penicilliuni, yeasts, Phyllo- sticta, Leptosphaeria, Alternaria, Stemphylium, Chaetomium, Pestallozia, and Streptomjces {?-a.Ay et al., 1948; Pady, 1951; Polunin, 1951, i95irt). The sticky slides exposed during flights over the Northwest Terri- tories showed small concentrations of angiosperm and g^'mnosperm pollen, spores of pteridophytes and bryophytes, Alternaria, and Helniinthosporiutn sativum, totalling about i per cubic metre. Uredospores of the cereal rusts Piiccinia graminis and P. glumarum occurred in small numbers (except in the most northerly flight, though a few were found north of the Arctic Circle); their concentration rose to about 12 per cubic metre over northern Alberta, w here there was also a smut concentration of about 6 per cubic metre (Pady et al, 1950; Polunin, 195 1«). In further flights over the Arctic the McGill workers attempted more elaborate sampling methods to eliminate possible contamination from within the aircraft, which could have increased the counts in the first two or three exposures of the earlier flights. In September 1948, Polunin flew over the Geographical North Pole in a B-29 aircraft fitted with a breech loading tube to hold a Petri dish projecting 30 cm. forward of the nose. Before exposure the interior of the Petri dish was coated with a silicone grease; and after returning to base, the dish was poured with molten agar and incubated. Immediately over the North Pole in late summer at 920 metres neither bacteria nor fungi were caught. However, at greater heights over the Pole and at other high latitudes, some Petri dishes caught nothing, while others, exposed at altitudes up to 6,770 metres, grew a few colonies of bacteria or moulds; no Actinomycetes were found (Polunin, 1951; Polunin & Kelly, 1952). Thus microbes appear to be present, though irregularly distri- buted, even over the Poles. During a flight over the Geographical North Pole under winter con- ditions in March 1949, the McGill workers used three kinds of samplers: (i) siHconed slides exposed in the tube forward of the nose of the plane ; (2) an electrostatic sampler installed in a box through which a slow stream of air was passed ; and (3) a filter tube packed with glass-wool and lens-paper. The electrostatic sampler and filters indicated a viable concentration of 26 bacteria + yeasts and i-6 fungi per cubic metre in some ver}' high latitudes. The authors concluded that the air over the Pole and its environs is nearly sterile and that it is of very mixed origin. Here again there was evidence that the origin of an air-mass is more im- portant than the localit}' of sampling (Polunin, 195 1, 195 1«, 1954; Polunin & Kelly, 1952). The results of further arctic and sub-arctic flights are reported in detail by Pady & Kelly (1953) and Pady & Kapica (1953). On one from Winnipeg via Churchill to Baker Lake in the Northwest Territories at an altitude of about 1,000 metres, using the G.E. Electrostatic sampler, cultures averaged: bacteria 10, and fungi 25, per cubic metre. Bacteria 141 THE MICROBIOLOGY OF THE ATMOSPHERE were predominantly cocci and spore-formers (though in a local flight over Churchill gram-positive pleomorphic rods predominated). Fungi were mainly Cladosporium and Alternaria, but included Penicillium^ Papularia^ and Stemphylium. In the summer of 1950, northern air was sampled daily by McGill workers for 21 days with the G.E. Electrostatic sampler and the slit sampler on a roof 17 metres above ground at Churchill near the tree-line on Hudson Bay. This survey formed the standard of reference for two flights to Resolute Bay, Cornwallis Island, some 1,600 km. to the north, on 1-3 August 1950, at an altitude of around 3,000 metres. At Churchill the catch consisted of: gram-positive pleomorphic rods (46 per cent), gram-negative rods (20 per cent), spore-forming rods (18 per cent), and cocci (15 per cent). In the two flights to Resolute Bay in the Arctic during this period, 51 per cent of the bacteria caught were spore-forming rods (Pady & Kelly, 1953). Fungi were assessed both in culture, and visually on silicone-coated slides: the numbers per cubic metre, with the visual counts in parenthesis, were: Cladosporium, 17 (132), with its maximum in an air mass of tropical origin; Alternaria, 0-7 (2-1); Stemphylium, i-i (i-8); rusts (9-2); smuts (Ustilago) (86); yeasts, 3-5 (304); and Penicillium, 2-1. Of the fungus cultures, 57 per cent were non-sporulating. In addition, Fusarium was reported as common on slides but rare in culture, and Septoria was sometimes abundant on slides. An interesting list of fungi which were caught only infrequently includes: Pullularia, Actinomycetes, Botrytis, Aspergillus, Verticillium, several ascomycetes, and a single culture of Cunninghamella — one of the rare isolations of a mucoraceous fungus from the upper air. In addition there were numerous moss spores and pollen grains, which together averaged 20 per cubic metre (Pady & Kapica, 1953). On the flights to Resolute Bay at 3,000 metres, the fungi were essentially the same as at ground-level at Churchill; but they were in much lower concentrations averaging 12 per cubic metre (125 per cubic metre if determined visually) and were principally yeasts, Cladosporium, and Ustilago. Pollen grains averaged 16 per cubic metre; and in warm air on the southern part of the flight moss spores averaged 47 per cubic metre. It was thought that the southern parts of these flights lay through an old continental tropical air mass, which had moved into the Arctic where most of the spores had died; north of this was cold polar air containing very few spores. The general conclusion reached was that the air-spora over the Arctic comes mainly from the agricultural regions of the south (Pady & Kelly 1953; Pady & Kapica, 1953). However, the larger numbers caught at ground-level at Churchill, and the numerous moss spores, suggest that the tundra also made an important contribution. MICROBIOLOGY OF AIR MASSES OVER NORTHERN CANADA The earlier McGill studies suggested that, in the upper air above the level of pronounced concentration gradients, microbial concentration 142 THE UPPER-AIR SPORA depends mainly on the history of the air mass. This was clearly shown in a series often flights over northern Canada (Kelly & Pady, 1953; Pady & Kapica, 1953) between September 1948 and August 1949. The history of air masses encountered during sampling and the positions of fronts were correlated with the results of sampling. The electrostatic sampler, loaded alternately with Petri dishes and siliconed slides, gave the most consistent results. (However, as sampling was non-isokinetic, pollen and large spores may have been underestimated.) On these flights bacteria varied in concentration less than fungi. At the end of December many samples were blank and the air was almost sterile. Fungi were much more plentiful in June, July, and August, than in the rest of the year, but bacteria were most numerous in spring and autumn. Kelly and Pady suggest, reasonably enough, that the bacteria come mainly from soil which is exposed and cultivated in spring and autumn, giving the opportunity for wind erosion; but their suggestion that the fungi also come from soil seems to be contradicted by the predominance of fungi in summer, and all evidence points to the fungi coming mainly from vegetation and debris above ground-level. All the bacteria isolated and examined in detail were regarded as typical soil forms; they were classified as : aerobic spore-formers (37-9 per cent of the total recorded catch), gram-positive pleomorphic rods (23-8 per cent). Micrococcus (i8-8 per cent), gram-negative rods {Flavobacterium^ Achromobacter^ or Pseudomonas, 4-8 per cent), and Sarcina (4-6 per cent). Fungi obtained in culture were : Cladosporium (73 per cent), Alternaria (7 per cent), PenicilUum (2-9 per cent), Streptomyces (2-9 per cent), Stemphylium (1-5 per cent), Aspergillus (0-7 per cent), yeasts (0-7 per cent), and other fungi (11 per cent). Many more fungi could be counted visually on silicone-coated slides than could be grown in culture — an effect which was exaggerated by the numerous smut spores that were obtained in one flight over the prairies in October 1948, which yielded: smut spores 52-4 per cent, Cladosporium 32-4 per cent, Alternaria 3-3 per cent, Helminthosporium 0-3 per cent, and rusts o-i per cent. AIR MASSES OVER MONTREAL In a further survey of microbes associated with different air masses, the McGill workers used the electrostatic sampler and the slit sampler between 10-00 and 13-00 hours on 1 13 days between September 1950 and December 1951, at the top of a building high in Montreal (Kelly & Pady, 1954; Pady & Kapica, 1956). Ten t^'pes of air mass were recognized, classified on the basis of exposure to agricultural land ; but as most examples of any one type occurred at one time of the year, the effects of differences in origin of the air mass may be confounded with seasonal effects. Further, at the altitude of 130 metres, sampling at midday is likely to be done in the frictional turbulence layer and to be dominated by local ground sources 143 THE MICROBIOLOGY OF THE ATMOSPHERE which are themselves affected by the temperature and humidity of the air mass. This survey is therefore perhaps best regarded as a valuable contri- bution to knowledge of the local air-spora near the ground; judging from knowledge about the upper-air, the amount contributed by the air mass is likely to have been small. Cladosporium and yeasts were the chief constituents of all the air masses (even of fresh polar air), and on our interpretation the abundance of PenkilUum is not surprising for samples taken in a large city. Alternaria and Fusarium were commoner in tropical air. Smut spores occurred in all air masses and at all seasons. Basidiospores of agarics were suspected but not positively identified. Fungi were most numerous in July and August, when 625 cultivable (8,610 visible) spores were recorded per cubic metre, and least numerous from December to February, when 36 (28) were recorded per cubic metre. Bacteria were present in greatest numbers in polar air during spring and autumn, rising from fewer than 70 per cubic metre during March to 710 per cubic metre in June, then decreasing to the end of August and rising to a second maximum during November. In air of maritime origin the trend was irregular. AIR MASSES OVER THE NORTH ATLANTIC OCEAN In two flights from Montreal to London, England, at altitudes ranging between 2,700 and 3,000 metres, the McGill workers were able to study the relation between microbial concentration and air mass (Pady & Kelly, 1954; Pady & Kapica, 1955). Over the ocean polar air had generally fewer bacteria and fungi than tropical air (Table XXIII). Over Quebec Province, in one air mass which was classified by meteorologists as of polar origin, and which gave few bacteria or fungi in culture, very many fungus spores \^'ere caught on a silicone-coated slide in the slit sampler. The authors interpreted this as evidence of a load of non-viable organisms which could only have originated in the tropics, and suggested that the air had been carried into the Arctic, thence eastwards, and finally southwards, during which passage most of the suspended micro-organisms lost their viability. However, another explanation seems possible on careful examination of the data. The visible total on the silicone-coated slides, amounting to 18,700 per cubic metre, was made up largely of yeasts (9,900 per cubic metre) and yellow-brown spores (7,500 per cubic metre). As 'about 50 per cent of the latter had an apiculus and were considered to be basidiospores', the other 50 per cent were probably also basidiospores lying in the alternative position (in which the apiculus would be invisible). The flight may well have been through one or more thermals arising from coniferous forests of Labrador and Quebec Province, by which a polar air mass was becoming charged with the air-spora of the ground layer. 144 THE UPPER-AIR SPORA TABLE XXIII ANALYSIS OF PADY & KELLY's (1954) DATA ON TWO RETURN FLIGHTS OVER THE NORTH ATLANTIC, SHOWING CONCENTRATIONS PER CUBIC METRE OVER LAND, AND MEAN CONCENTRATION OF BACTERIA ANT) FUNGI IN AIR- MASSES 0\TR OCEAN Bacteria Fungi Air mass Month Position E S E Culture Visual Polar June Over Quebec Province — — — — 240 Polar Aug. Over Quebec Province 140 7-0 53-0 92-0 — Tropical Aug. Over Quebec Province 28-0 no 57-0 1600 940 Polar Aug. Over Quebec Province 6-4 39 i8-o 32-0 12,700 Polar Aug. Over Labrador 3-5 7-0 21-0 70-0 — Polar June Over Ocean 59 7-5 3-9 190 26-0 Aug. Over Ocean 8-2 4-6 2-8 i6-o 56-0 Tropical June Over Ocean 6-8 6-9 35-0 140-0 2o8-o Aug. Over Ocean 5-2 15-0 23-0 194-0 67-0 Tropical June Over England 1-3 2-0 170-0 317-0 — Tropical Aug. Over England 137 53-0 52-0 215-0 580-0 E = electrostatic sampler S = slit samp ler — = not investigated The bacteria obtained on these flights were classified as: June August per cent per cent Micrococcus & Sarcina 41-4 13-2 Gram-negative rods 4-3 20-7 Gram-positive pleomorphic rods 20-4 37-0 Aerobic spore-formers 33-2 29-0 The fungi identified occurred in the following percentages (mean of the two flights): Cladosporium^ 82-3; Alternaria, 2-6; Pulhilaria, 2-3; yeasts, 2-1; Penicillium, i-6; Botrytis, 1-5; Stemphylium, i-i; non-sporu- lating colonies, 3-2 per cent. Of these, Alternaria^ yeasts, Botrytis^ and Penicillium were noted as more abundant in tropical air, whereas Stem- phylinm, PuUularia, Fusarimn, and Papularia were more abundant in polar air. Sporormia was found several times, always in polar air. Many other fungi occurred in small numbers. Among the many interesting results that stand out clearly from these flights is the discovery that viable bacteria and fungi occur at an altitude of 3,000 metres in air masses all the way across the North Atlantic, though the bacteria were so few that some samples of about 2 cubic metres appeared to be entirely devoid of them. There was, however, no gradual diminution with the distance from land. Cladosporium is the dominant fungus over the oceans, as it is also over land, but it probably loses viability as the air mass travels. K 145 THE MICROBIOLOGY OF THE ATMOSPHERE VERTICAL GRADIENT OVER THE OCEAN The McGill workers showed conclusively that the upper air contains an appreciable spore-load in all the places which they examined. Even in the high-arctic winter, the evidence proves that samples of a few cubic metres of air may or may not be sterile. Cladosporium appears to dominate the upper-air spora, often together with many yeasts and bacteria. Pos- sibly the purest air is to be found near sea-level in mid-ocean. The few observations suggest that samples taken on board ship are collected in a purified layer of the atmosphere, and that higher up over the ocean surface the spore concentration is greater than at sea-level. The troposphere is always more or less contaminated with micro-organisms. From Erdtman's (1937) results a ship in the North Atlantic in spring would be in a region of about one pollen grain per 100 cubic metres, whereas Pady & Kapica (1955), at 3,000 metres over the same ocean, recorded up to 25 pollen grains (with moss spores) per single cubic metre. On a flight from New Zealand to Australia, at about 1,000 metres above the Tasman Sea, Newman (1948) exposed sticky slides behind a leading wire in the hope of improving the trapping efficiency by breaking the stagnant layer. At a position 1,100 km. off Australia, he estimated pollen grains at 0-73 per cubic metre, and fungus spores at 0-70 per cubic metre; at 340 km. from Australia he found pollen to be 8-75 and fungus spores 1 6-8 per cubic metre — which is about 100 times as numerous as the concentration of pollen grains and fern spores at ship's mast-level on the North Atlantic crossing recorded by Erdtman (cf. p. 127). Newman's values for fungus spores are somewhat similar to those of the McGill University workers for the upper air over the Atlantic. It seems likely that Erdtman's samples, taken on board ship, were from a zone of surface air which had been largely cleaned by rain-wash, sedimentation, and contact with the ocean, and only partially replenished from the stock in the air mass overhead, where aircraft samples were taken. Summary Knowledge of upper-air microbiology is based on occasional samples and is affected by place, season, weather, air mass, and so on. There are no continuous records; but there are some hints that a 'biological zone' occurs at middle height, which can probably be explained in terms of temperature inversions, air masses, and precipitation. Molisch (1920) introduced the concept o^ aero plank ton to denote the microbial complex referred to in this book as the air-spora. It has been argued that the word 'plankton' suggests organisms based on the air during at least a vigorous phase of growth, whereas the air-spora is only airborne temporarily, even though adapted to wind transport as a means of dissemination. Clearly this argument is valid for pollens and plant spores; 146 THE UPPER-AIR SPORA but Is there, in addition, a vegetative air-inhabiting plankton ? We can- not yet give this apparently improbable hypothesis a decisively nega- tive answer. Evidence in favour of it has been stated by R. C. McLean (1935, 1943), who wTote : ' "Dust to dust" seems to be the only cycle envisaged. Yet the experiments of Trillat and others show at least the possibility that the air may be a vegetative habitat and the large pro- portion of non-spore-formers present . . . needs more than a conven- tional explanation.' Proctor & Parker (1942) noted that one third of the bacteria collected from the upper air could grow at o°C., and survive 48 hours exposure at — 26°C. If there is a truly indigenous aeroplankton, its habitat must be exacting in the extreme, and tolerable only by specialized bacteria, yeasts, or act- inomycetes. Frequent drying must reduce the population to inactivit}% though metabolism could be resumed in a cloud of water droplets when gaseous nitrogen and carbon compounds could be absorbed and used. In constant danger of being removed from the air by rain or snow or by contact with the ground, the risk of removal would be increased by any attempt to parasitize organic particles brought up by convection from below. However, radioactive dust can persist for several weeks in the troposphere, and this is a long period on a microbial time-scale. The aerial environment is not obviously beyond the range of exploitation by micro- organisms, for the rate of loss by death or deposition might not be greater than for bacteria in the sea, and there would be freedom from predators. If anywhere, such an aeroplankton might be expected to ride clouds on the ascending side of a tropical convection 'cell' over the Equator. Although the origin of the upper-air spora from the soil has been assumed by most investigators, the circumstantial evidence suggests a wider range of sources. The bacteria are probably mainly soil forms with a small proportion from sea water. But hyphal fragments, especially conidiophores of Alternaria and Cladosporium, which are commonly reported from the upper-air (Pady & Kapica, 1953, p. 321), evidently come from the ground vegetation-layer rather than from the soil. The numerous yeasts, coloured basidiospores of toadstools, and smut spores, evidently originate above the soil surface. It is hard to believe that wind could burrow into soil, picking out the few spores of Cladosporium and Alternaria, yet leaving behind most of the far more numerous PcnicilUimi, Trichoderma, Aspergillus and Mucoraceae spores — not to mention clay particles ! 147 XI DEPOSITION IN RAIN, SNOW, AND HAIL Airborne microbes can be deposited direct or they may be washed out of the air in raindrops, hailstones, or snow-flakes. Trillat & Fouassier (19 14), from their laboratory experiments with artificial fogs condensing on a suspension of pathogenic bacteria in small vessels, thought that air- borne microbes act as condensation nuclei. Condensation nuclei are now thought to be small hygroscopic particles, and it seems more likely that droplets already formed collect spores by impaction {see Chapter VII). McCully et al. (1956) estimate that, over all land areas of the globe, from 35 to 50 per cent of the total atmospheric dust load is washed out each day. Here we will consider the results of rainwash in nature and the spore content of precipitation water. Over the last 300 years about a score of people are known to have sought microbes in precipitation water. Collecting the sample has some pitfalls, however; the vessel must obviously be clean, but the danger of contamina- tion by rain-splashed soil has not always been anticipated though, with current knowledge of the magnitude of splash and its part in soil erosion, the danger is now clear (Laws, 1940). Much of the early work summarized below, however, is clearly trustworthy. Animalcules in rain-water deHghted Leeuwenhoek (1676, in Dobell, 1932). Rain was collected in a clean porcelain dish set on a wooden tub to avoid earth being splashed by rain. Minute organisms were searched for in vain until after the rain-water had stood for some days, by which time it would also have been contaminated by dry deposition — so we do not know whether or not Leeuwenhoek found microbes of precipitation water. Rain The only systematic study of precipitation micribiology comes from Miquel (1884, p. 597; 1886, p. 530) at the Pare Montsouris, Paris. Miquel caught his rain in a metal funnel fixed at 1.7 metres above ground-level on a pillar, well away from trees and buildings. Rain falling into the funnel was collected in a platinum crucible with a cover, both funnel and crucible having been heated to redness just before sampling. The sample was then sown, drop by drop, in 50 to 100 flasks of beef broth. Miquel also designed apparatus which placed raindrops on a moving band of nutritive paper. After 6 days' incubation, the paper was dried and kept as a record of 148 DEPOSITION IN RAIN, SNOW, AND HAIL bacterial and mould colonies. The largest catches of bacteria occurred In the warmer months, when numbers varied from o-ooo8 to 8-3 per ml., with a general mean of 4-3 per ml., but these figures excluded the first rain after several dry days when 200 bacteria per ml. might be recorded. During prolonged rainfall the numbers fluctuated instead of contin- uing to diminish, suggesting to Miquel that the rain clouds themselves had a characteristic bacterial content, with the percentage composition of: Micrococcus (60), Bacillus (25), and Bacterium (15). Moulds fluctuated in the same manner as bacteria and averaged 4 per ml. Miquel estimated the annual precipitation of bacteria and moulds at Montsouris at over 4 million per square metre — a figure that was obviously too low as he excluded the contribution of the first rain after dry days. The pharmaceutical use of rain-water induced Lindner (1899), in Germany, to collect 28 samples of rain in a clean porcelain dish on a bleaching ground near his house. Samples were then added to sterile hay-infusions, albumen, milk, or blood serum. His liquid cultures gave a regular succession of bacteria, flagellates, and monads, in the first day or two and, later on, stalked Vorticella-like ciliates, Paramaecium, Stylonychia, and Volvox. Once he got two amoeboid forms, but never gregarines or coccidiens. Lack of precautions against splash-contamina- tion appears to leave the interpretation of his data in doubt. In this century various workers have cultured microbes from rain. Minervini (1900) collected numerous rain samples on board ship in the North Atlantic. Bacteria were abundant, half the samples yielded pink yeasts, and a quarter of them Penicillium. He also obtained Aspergillus glaucus, A. niger, Monilia Candida^ and many other moulds. Busse (1926) recorded pine pollen in rain. Rain-water collected over the ocean at considerable distances off" shore by ZoBell (1946, p. 179) averaged i to 10 bacteria per ml, with few or no mould fungi. Rain-water collected on land at the Scripps Institution of Oceanography, California, contained from 10 to 150 microbes per ml. As usual, the highest counts were obtained during the first rain and were associated with a predominance of mould spores. Protozoa in rain were studied at Heidelberg by Puschkarew (1913), who collected ten samples of rain-water in a sterile funnel, and added nutrient solutions. At the start of rain he found large numbers of fungi and bacteria, and the numerous protozoa included a new species. Amoeba polyphagus^ with species of Bodo^ Manas, Calpoda, and other genera. Twice in the month of November, rain was collected in sterile flasks on a roof at Leiden by Overeem (1937) and inoculated into flasks of a nutrient medium favourable to growth of green plants in light. In a total of 221 cc. of rain-water she obtained the following cultures. Algae: Stichococcus minor (8), S. bacillaris (5), Chlorococcum sp. (7), Pleurococcus vulgaris (4), Chlorella vulgaris (2), Hormidium flaccidum (2), and Navicula 149 THE MICROBIOLOGY OF THE ATMOSPHERE minuscuhi (i). Myxomycete : Physarum nutans (i). Moss: Brachythechim rutcibulum (i). Another worker who made a re\^ arding study of autotrophic plants in rain-water was Pettersson (1940), working at the Zoological Station at Tviirminne, Finland, in the summer of 1936. Glass funnels (176 sq. cm. in area) were lined with filter-paper, sterilized and taken, covered, to the trapping site. After exposure, the filter-paper was sprayed with a nutrient solution, and the funnel was covered with a glass lid and left to stand in a light place for a few days. Developing organisms were picked off and transferred to new culture vessels to continue their growth. The originality of the method lies in the medium being unfavourable for the development of bacteria and fungi because the cellulose of the filter-paper was the only carbon source provided. Pettersson, like Over- eem, was therefore able to explore a novel part of the air-spora. Snow-traps were also used, consisting of shallow glass dishes 15-20 cm. in diameter, with a thick bed of blotting paper and an upper layer of filter-paper. These two methods gave an unexpectedly rich harvest. A sample of snow taken at Pikis (Piikio) from the start of snowfall on i March 1936, gave thirt)'-six lichen thallus fragments and a moss gemma. On the next day, 2 hours after the start of another snowfall, a sample corresponding to 625 ml. of water yielded nineteen lichens and two Chlorococcum colonies. A third sample of 805 ml. of water, taken 4I hours later, yielded six lichens, and three mosses which were identified after 6 months' growth as Brachythecium velutiniun^ Hypnum cupressiforme, and Pylaisia polyantha {see also Pettersson, 1936). Pettersson's rain-trap yielded a wealth of information from the fourteen samples investigated, for details of which the original paper must be consulted. The interest was taxonomic and qualitative rather than quantitative. For the early samples the funnel was placed on a low rock, 2-5 metres above sea-level, on open grassy soil. Some of the organisms caught may possibly have come by splash from the ground, but not many can have done so because the largest catch of mosses belonged to a genus hitherto unrecorded in Finland {see Chapter XIV). The precaution of raising the funnel on a wooden base i metre high was adopted in later tests. In a total of 1,373 ^- of rain collected, Pettersson obtained 1,200 conifer pollen grains, 300 liverwort spores (all of Marchantia polymorpha except for one of Metzgeria), Myxomycete spores {Stemonitis fusca three times and Arcyria denndata), and numerous algae. Blue-green algae were scarce, being represented only by Nostoc commune and Gloeocapsa sp. in separate samples. Green algae were abundant in almost every sample, those identified including: Chlamydomonas nivalis^ Chlorella vulgaris, Chlorococcum humicolum, Cystococcus pseudostichococcus, Prasiola stipitata, Roy a sp., and Tetraedron punctulatum. Some of these, the author suggests, may have originated from lichen soredia. Lichen spores and soredia were 150 DEPOSITION IN RAIN, SNOW, AND HAIL not the main source of lichens in the traps, for the hchcns mostly originated as thallus fragments and were evidently of fairly local origin. The 2,000 moss plants cultured from the spores caught in Pettersson's traps included specimens of: Aloina brevirostris, A. rigida^ Amblystegium serpens^ Brachythecium velutiniwi, Bryum spp., B, argentemn, B. pa liens, Ceratodon piirpureus, Funaria hygrometrica, Leptobryum pyriforme, Mnio- bryum carnemn, Pohlia cnida, P. nutans, and Pylaisia polyantha. Observations on micro-organisms in rain were made at Rothamsted Experimental Station in 195 1 by Gregory, Hirst, and Last {see Hirst, 1959) while they were comparing various spore-trapping techniques. Two conical glass funnels 20 cm. in diameter were exposed on a wooden structure at a height of 2 metres above ground-level. One funnel was open to rain (rain-trap), while the other (dry-trap) was protected by a flat asbestos-cement disk held 25 cm. above the mouth of the funnel — to keep off rain but still allow^ dry deposition. Washings from both funnels were collected daily and the fungus spores separated by sedimentation onto a glass cover-slip. On dry days the fully exposed rain-trap consistently caught fewer microbes than the dry-trap ; but, as might be expected, this was reversed during rain — especially in the first rain after dry weather (Table XXIV). TABLE XXIV GEOMETRIC MEANS OF RATIOS OF CATCHES BY R.AIN-TRAP TO DRY-TRAP, 2 METRES ABOVE GROUND, ROTHAMSTED, JUNT-SEPTEMBER 1951 (HirSt, 1959). Ratio for dry days Ratio for all rainy days Ratio of single rainy days to the first of a succession Smuts (mainly Ustilago) 0-6 3-8 60 Cladosporium 0-8 1-3 1-8 Alternaria 0-8 37 6-9 Pollens < 20 fi 0-8 1-4 2-3 Pollens > 20 ^ 0-8 1-5 2-4 Rain falling during one thunderstorm was studied in detail (Gregory, 1952; Hirst, 1959), and a detailed account of changes in the air-spora during this period, observed with the aid of the Hirst automatic volumetric spore-trap, has already been published (Hirst, 1953, pp. 382-5). A y-day spell of warm, dry weather ended in a thunderstorm at 13-25 hours on 22 July 1 95 1. The rain-trap was cleaned immediately before the rain started, and the first i mm. of rain which fell in the first half-hour of the storm was collected separately from the succeeding 375 mm., which contained manv fewer spores (Table XXV). As Hirst (1959) remarks in discussing this series of observations: 'Spores released during rain are presumably removed from the air as 151 THE MICROBIOLOGY OF THE ATMOSPHERE readily as spores already there when rain starts to fall, so that concen- trations of airborne spores measured during rain represent, not the total released, but the excess of those released over those removed. Rain- scrubbing seems an ideal method of deposition for air-dispersed soil fungi. For foliage pathogens its biological significance is far from clear. Many spores may be lost in "run-off" unless they can attach themselves to the leaf surface or penetrate into crevices they would be unlikely to reach when deposited from dry air.' TABLE XXV SPORES BROUGHT DOWN BY THUNDER RAIN TERMINATING 7-DAY DRY SPELL, ROTHAMSTED, 22 JULY 1 95 1 (Gregory, Hirst & Last, unpublished). Number of spores per ml. of rain in I St 0-95 mm. in succeeding 375 mm. " of rain falling I3-25-I3-55 hr. of 13-55- rain falling -08-25 (23 July) Smuts (mainly Ustilago) 455 55 Cladosporium 1770 205 Ahernaria 370 20 Erysiphe 280 10 Small pollen grains 270 10 Large pollen grains (over 20 /x diameter) 120 5 In contrast Asai (i960), who introduced the useful method of filtering rain through membrane filters under reduced pressure, failed to obtain uredospores of Puccmia grammis, although they were known to be in suspension in the air at the time the rain samples were collected. Spores in raindrops appear to play a part in some processes of plant infection. Dry wind-blown spores of barley loose-smut {Ustilago nuda) rarely infect the ears of susceptible barley varieties but, when drops containing spores in suspension fall on flowers, the spores are brought into direct contact with the ovary, and infection follows (Malik & Batts, i960). Snow Janowsky (1888), Pettersson (1940), and others have found a few organisms in falling snow. Only Gazert (191 2) gave a negative report from the Antarctic on the microbial content of fresh-fallen snow in Kaiser Wilhelm II Land. A. L. McLean (191 8), on the other hand, reported numerous organisms in snow and ice in Adelie Land ; but it is not certain whether they were brought down with the snow or deposited dry in fine weather from the atmospheric dust which settles over the Antarctic. However, on three occasions McLean caught falling snow in a sterile 152 DEPOSITION IN RAIN, SNOW, AND HAIL basin : 'elaborate precautions having been taken to prevent contamination, the thawed-out samples showed under a cover-slip cocci, motile bacilli, and, invariably, zoogloea masses of bacteria in moderate numbers. Diplococci, and occasionally cocci, were observed to be invested by a pale capsule. ... A glucose agar slope culture of falling snow showed a few small greyish colonies.' Atkinson isolated a motile bacterium believed to have been carried to the Antarctic by upper-air currents and brought do\Mi by the snow (Scott, 1913). Most of the arctic and antarctic snow samples were taken from fallen snow, and organisms could therefore possibly have reached the snow by a process of dry deposition (e.g. Salimovskaja-Rodina, 1936; Darling & Siple, 1941) and are considered in Chapter IX. Hail Large numbers of microbes were recorded by Bujwid (1888), who collected hailstones in Warsaw in the month of May, washed them in sterile water, and, plating out the melt- water, found 21,000 bacteria per ml. They included Bacillus fluorescens liquefackns, B.f. ptitridus, and B. janthinus. From these numbers Bujwid concluded that surface waters must have been carried aloft and frozen. During a hailstorm in St. Petersburg, windows were broken by hail- stones the size of walnuts. Foutin (1889) washed some of these and, on plating-out the melt-water, obtained 628-729 bacteria per ml., but no fungi or yeasts. In July storms at Guelph, Ontario, Harrison (1898) collected hail- stones, washed them in i in 500 mercuric chloride solution and, after rinsing, plated-out the melt-water. One storm gave 955 colonies per stone, of mixed bacteria and moulds, including ''Penicillhim glaucum\ Miicor sp., Aspergillus sp.. Bacillus fluorescens liquefaciens^ B.f. non-liquefaciens, and Proteus vulgaris. A later storm averaged 1,125 colonies per ml. but these included fewer moulds than the first. Harrison concluded that the bacteria must have come from surface water, but that the moulds were picked up from the air. Belli (1901) obtained 140 organisms per ml. of hail melt-water, of which eight were Aspergillus or Penicillium and the remainder bacteria. Hail has also been sampled by Dubois (191 8). The organisms in precipitation water remain almost unstudied. The little we know from existing records is tantalizing. Precipitation water is non-sterile, whether on land, over the oceans, or about the poles. A wide variety of organisms has been recovered from such waters — including bacteria, fungi (moulds, yeasts, and plant pathogens), algae, liverworts, mosses, pollens, and protozoa. Microbes are found in rain, hail, and snow, when collected as it falls — before the possibility of ground contamination. The highest counts are recorded from hail and, at present, these are 153 THE MICROBIOLOGY OF THE ATMOSPHERE perhaps the most reliable records — because hailstones can be surfiice- sterilized. The first rain after a dry spell is heavily contaminated and, even during prolonged wet weather, the spore numbers in rain remain substantial. A spore liberated near ground-level has a high probability of being deposited dry; but wash-out by rain, hail, or snow, probably mosj: often terminates the journey of spores reaching the tail-end of the dispersal gradient. Microbial sampling of precipitation is still in the naive stage. Methods have not been tested, and we still do not know how a collecting vessel should be placed to avoid contamination from soil and vegetation. Conceivably, spores may undergo re-concentration ^^■ithin a cloud. Rising convection bubbles may bring new spores to the top of the cloud, where they can be collected and washed down in raindrops to the base of the cloud. Here the drops might evaporate, allowing the spores to be carried up again — perhaps eventually to be brought down to earth in hailstones. The abundance of microbes in hail, and the reports of a 'biological zone' at several thousand metres, supports the suggestion that convective clouds may be spore-concentrators. Exploration of organisms in precipitation needs an experimental study of methods of sampling from ground, ships, and aircraft. Systematic sampling could then be attempted with some prospect of learning what part such precipitation plays in terrestrial microbial circulation. 154 XII THE AIR-SPORA OF ENCLOSED SPACES A SMALL but important fraction of the atmosphere is walled-in and pro- vides microbes with an environment different from the outdoor world. Indoor air-hygiene is an aspect of medical science with a voluminous literature which can be approached through such works as: Aerobiology (Moulton, 1942), Studies in Air Hygiene (Bourdillon et ciL, 1948), Airborne Contagion and Air Hygiene (Wells, 1955), and Mould Fungi and Bronchial Asthma (Werff, 1958). The brief treatment given here of 'intra-mural aerobiology' presents an ecological instead of a medical viewpoint. Outdoor air moves as wind flowing bodily over surfaces, and a point near the ground is immersed in a continually flowing stream of fresh air. Rooms, on the other hand, are ventilated, and fresh air is assumed to mix thoroughly with the existing air instead of displacing it bodily. By one 'air-change' ('ventilation turn-over') is meant the introduction of a volume of fresh air equal to the volume of the room; an equal volume of mixed stale and fresh air is displaced during the process, leaving a mixture of stale and fresh air in the room. Unless continually renewed, any micro- bial concentration in the air of an enclosed space will tend to diminish with time as a result of ventilation and deposition. Concentration of viable organisms will also decrease with time — following the natural death-rate, or because of any disinfection that may have been applied. Die-away of Concentration Die-away of concentration is a phenomenon seen most clearly under intra-mural conditions, because out-of-doors a concentration is carried away bodily by wind. So far it has not been feasible to trace concentration changes out-of-doors in one air mass during its travels. Decrease of concentration with time is caused by: (i) exchange with outside air (i.e. ventilation); (2) deposition on walls, ceiling, and floor, by various processes including sedimentation; and (3) reduction in the viable count through death. Ventilation does not immediately sweep away the whole microbial load, but progressively dilutes it exponentially. Then 'n' air changes will reduce concentration in the ratio: i/e". Decreases in concentration due to deposition, death, or disinfection, may also follow a logarithmic law, and these can then be expressed in units of equivalent ventilation turn-overs for ease of comparison. 155 THE MICROBIOLOGY OF THE ATMOSPHERE Ways of expressing rates of removal or death of bacteria are discussed by Bourdillon et al. (1948), and are based on the constant, 'K', in the equation : N = NoC"^'^^', where No is the number present at time T = (e = base of Naperian logarithms) ; K, the 'die-away', is the rate of removal of bacteria by all processes during the period, and may be sub- divided. Thus Ki) is the death-rate, K^ is rate of removal by ventilation only and is identical with the ventilation rate in air changes per hour, rooo 100 - MINUTES Fig. 22. — Exponential form of the die-away of bacteria-carrying particles from the air of a room. Line A: In an observation military canteen after the occupants had left suddenly. Line B : Observations on the die-away following a group of sneezes in a small room. (From Lidwell (1948), reproduced from M.R.C. Special Report No. 262, Studies in Air Hygiene, by permission of the Controller of H.M. Stationery Office.) and Kj.. is the rate of removal by sedimentation. In an example of die- away rates of bacteria from all causes in a bedroom with open windows during fine weather at midsummer K, was equivalent to 6- 1 air changes per hour after the occupants settled down to sleep at 23-00 hours; K = 4-9 after they went down to breakfast at 07-55 hours; and K = 6-8 after the making of the beds at 08-50 hours (Lidwell, 1948, p. 253). Another example is shown in Fig. 22. The case of die-away with stirred settle- ment has been discussed by C. N. Davics (1947). Spore Movement in Convection Currents Convection currents alone, in an enclosed space without access of outside air, are often sufficiently active to diffuse fungus spores evenly 156 THE AIR-SPORA OF ENCLOSED SPACES through the whole volume of air. With fruit-bodies of basidiomycetes enclosed in chambers, Falck (1904) found that vertical tiers of horizontal paper shelves became covered with spore deposit in a remarkably uniform manner. By contrast, suspending the pileus of an agaric in a small glass vessel often resulted in 'curious and fantastically' irregular spore deposits on a piece of paper placed underneath. These were inter- preted by Buller (1909) as due to the convection currents in the vessel being of a velocity comparable with the terminal velocit}' of the spores. The heat from a lamp was sufficient to alter a previously established convection system. WARMER AIR OUTSIDE COLDER AIR OUTSIDE ■7 //////// y^ Fig. 23. — Diagram showing changes of circulation in a room according to relative temperature of walls and of inside air. Even without any ventilation, air circulates in a room because of thermal convection. Heating of air by rock surfaces in a mine may result in a flow along an adit and up a shaft. Heating of glasshouses in sunlight also leads to strong convection currents. Within a building the temperature of the air may be less changeable than that outside, and this may lead to characteristic air-movement patterns. Warmer walls will generate an up draught, colder walls a down draught — each being balanced by opposite currents in the centre of the room (Fig. 23) and often moving fast enough to counteract sedimentation under the influence of gravity. Circulation of air within a house is complex, but there is evidence of a fairly rapid exchange of air and of its suspended spores throughout a house. C. M. Christensen (1950) experimented with spores of Hormodendriim resimie, a mould that is peculiar for its ability to grow on a coal-tar creosote medium, and therefore suitable for use as a 'marker' spore in dispersal experiments. Spores were liberated in a room on the lowest floor of a house while all doors to the central hall-way were left open. Within a few minutes, spores were found deposited on Petri dishes in rooms communicating with the hall-way, but situated one, two and three storeys higher. 157 the microbiology of the atmosphere Intra-mural Sources Microbes in indoor air may come from the outdoor air-spora by ventilation, or they may originate within the enclosure — in which case they are probably limited in variety but may occur in high concentration. Defective timber attacked by fungi may be an important source of spores in dwelling houses. A. W. Frankland & Hay (1951) showed that some asthmatics are sensitive to the spores of the dry-rot fungus {Merulius lacrymans)^ and spore concentrations ranging from 1,630 to 360,000 spores per cubic metre have been recorded in buildings with active fructifications of this fungus (Gregory et al^ 1953). Timber in mines is particularly liable to fungal decay and may also have superficial moulds growing on it. Extensive growth of Sporotrichum beurmanni (the pathogen of human sporotrichosis) was found on fresh timber of mines in Transvaal by Brown et al. (1947). The fungus was isolated from the air, and ventilating currents of I metre per sec. could detach spores from wood provided its moisture content was less than 80 per cent. Processes by which infectious diseases are transmitted through the air have been matters of vigorous controversy in medicine, and the answers given have influenced social habits and prophylactic measures. Cornet (1889) held that pulmonary tuberculosis is normally acquired by inhaling dust of dried sputum, but Fliigge (1897) believed that infection was from germs expelled from the mouth and nose when coughing. G. S. Wilson & Miles (1955) conclude that both processes occur, dust infection being commoner in drier countries whereas droplet infection is the rule in moister climates and in crowded places. Air exhaled from the lungs in normal breathing is optically clean and almost sterile; but in coughing and sneezing, large numbers of droplets of mucus and saliva are propelled with explosive violence into the atmosphere. Jennison (1942) obtained photographic evidence of 20,000 droplets being put into the air from a single sneeze. The largest number observed was 40,000, and a weak, stifled sneeze gave only 4,600 droplets. A cough produced a few hundred droplets and the enunciation of consonants was also productive. Sneeze droplets, ranging in diameter from a lower limit of 5-10 /Lt, and with 20-40 per cent smaller than 50 /x, could evaporate instantaneously to 'droplet nuclei'. The concept of droplet nuclei, developed by Wells (e.g. 1955), has proved fruitful. 'Droplet nuclei' are the particles formed from the smallest droplets, which evaporate before falling to the ground and so remain sus- pended in air. They consist of the solid residue of the evaporated droplet, together with any bacteria or virus particles, and may be coated with semi- dried-up mucus which tends to preserve activity and viability. Few droplets are actually propelled more than 2 or 3 ft. ; but, when evaporated, the resulting droplet nuclei, with any bacterial cells or virus particles, would remain in suspension almost indefinitely. The droplet nuclei have 158 THE AIR-SPORA OF ENCLOSED SPACES no trajectory but move with the slightest air currents, and are emitted in large numbers. Thus akhough most airborne bacteria seem to be carried on rafts of dust particles which settle rapidly, they appear to be relatively innocuous saprophytes; the pathogens are present only in special en- vironments, being carried in much smaller and more insidious droplet nuclei which are small enough to be capable of entering, and being retained by, the alveoli of the lungs. The Air of Different Environments dwelling houses In spite of ventilation, Penicillium dominates the air inside most houses, in contrast to Cladosporium outside, and bacteria tend to be more abundant indoors in winter than in summer. Microbial concentration indoors varies greatly with mechanical and human activity. Carnelley et al. (1887), using Hesse's tubes in schools and mills in Dundee, Scotland, observed that, in densely-populated rooms, stirring up dust increased the total air-load and increased the ratio of bacteria to moulds. When air in rooms is left undisturbed the bacteria (or particles to which they are attached) settle out rapidly, but the moulds do so much more slowly. Maunsell (1954, 1954'^') used the slit sampler in bedrooms and found that shaking beds, brushing carpets, and any building repairs, increased the mould-spore content of the air up to 17 times, but that it rapidly returned to normal when activity ceased. Other studies of the air of dwelling houses are discussed by Miquel (1879, 1883), Rostrup (1909), Winslow & Browne (1914), Flensborg & Samsoe-Jensen (1948), Nilsby (1949), Wallace et al. (1950), and Swaebly & Christensen (1952), among others. Tests with the portable volumetric spore-trap (Gregory, unpublished) show that the airborne dust in inhabited rooms is commonly dominated by what appear to be fragments of human skin in the form of minute, flattened scales from the stratum corneum of the epidermis. Concen- trations of several thousand of these potential bacterial 'rafts' per cubic metre are common indoors, and 390,000 per cubic metre have been noted after bed-making. These epidermal scales probably carry a large propor- tion of the airborne bacteria of indoor air. HOSPITALS In studies of hospital air over a period of 15 months, Miquel (1883) found a mean value of 11,100 bacteria per cubic metre in the crowded wards of the Hopital La Pitie, Paris, the counts varying from 5,100 in June to 23,100 in December. The general improvement in hospital hygiene since that time is illustrated for example by Colebrook & Cawston 159 THE MICROBIOLOGY OF THE ATMOSPHERE (1948) for a Birmingham hospital, where they found from 210 bacteria and moulds per cubic metre under quiet conditions, to 2,800 per cubic metre with bed-making in progress (one very high count of 22,000 per cubic metre, including many moulds, was obtained with the ward windows closed). Recommendations for the maximum tolerable number of particles carrying bacteria in operating theatres are 700 per cubic metre for minor operations, and down to 70 or even 15 per cubic metre for dressing burns and for operations on the central nervous system (Bourdillon et al., 1948/'; and see Bourdillon & Colebrook, 1946). FACTORIES AND WORKSHOPS, SCHOOLS, PUBLIC BUILDINGS Anthrax is one of the few bacterial diseases which is clearly spread by airborne dust to workers who handle wool and hair contaminated by infected animals. Inhalation of airborne spores of Bacillus anthracis sometimes produces fatal infection of the lung. Bacterial counts in a variety of English factories and offices were reported on by Bourdillon et al. [i^^^a). SUBWAYS, MINES, AND CAVES London's underground railways have been investigated by Andrewes (1902) and by Forbes (1924), and the New York Subway by Soper (1908). Studies in caves are few, but include those by Lurie & Way (1957) and by Mason-Williams & Benson-Evans (1958). SEWERS Miquel (1880), in Paris, gave special attention to the sewer in the Rue de Rivoli near its junction with the large collector of the Boulevard Sebastopol. He found a steady load of from 800 to 900 bacteria per cubic metre. Pollens were absent, and cryptogamic spores were only | to j as numerous as in outdoor air at the same time. The contamination of the air in the near-by Rue de Rivoli was lower in winter but higher in summer than it was in the sewer. Comparable results were reported from London in sewers under the Palace of Westminster (Carnelley & Haldane, 1887). AIR-SPORA OF FARM BUILDINGS High microbial concentrations often occur in farm buildings, such as cowsheds, where hay is being fed to animals, or in barns where thrashing or cleaning is in progress. Inhalation may produce symptoms of the still- little-understood farmer's lung and thrasher's lung, as well as various diseases of farm animals. Milk also needs protection from contamination by barn air {see Ruehle, 191 5; Ruehle & Kulp, 191 5). 160 THE AIR-SPORA OF ENCLOSED SPACES AIR-SPORA OF GLASSHOUSES The air of glasshouses has received Httle attention, in spite of the fact that workers and crop plants in it may be exposed to high concentrations of micro-organisms. Glasshouses may act as important spore emitters by means of convection through open ventilators (Hirst, 1959). SHIPS Air in holds and living quarters on board ship was studied early by Miquel (1886). A modern study of air in ships — including submarines, which are relatively clean — is reported on by Ellis & Raymond (1948). 161 XIII DEPOSITION GRADIENTS AND ISOLATION Chapter V described efforts to formulate changes in concentration of the spore-cloud while it diffuses and travels downwind. We must now discuss the more complex phenomenon of deposition gradients — the decrease in number of spores deposited with increasing distance from the source. The infection of a plant by an airborne spore is itself a complex process of which physical transport is an important part. Infection may fail at any one of a chain of stages; and, for a deeper understanding of the whole process, it is necessary to understand the parts. Spore diffusion and deposition are stages selected for special attention in this chapter. It is impossible to predict, from knowledge of the characteristics of spore deposition gradients, how many infections will be acquired by a plant at a given distance from a source of known strength, because con- ditions may be unsuitable for infection; but there is a possibility of being able to predict an upper limit, and to use this knowledge in choosing safe isolation distances. This chapter deals with gradients measured up to distances ranging from a few metres to a few kilometres from the source ; long-distance dispersal is discussed in Chapter XIV. The discussion assumes the simplified conditions described on page 47; but, even so, deposition and infection gradients have complications. Factors Complicating Infection Gradients (i) Deposition coefficient. With sources at or near ground-level, the diffusing cloud of spores, unlike a gas or smoke, is robbed by heavy deposition close to the source. Concentration and area-dose at a point downwind therefore depend on two factors — the diffusion history and the deposition history which the relevant part of the cloud has experienced. In view of the evidence given in Chapter VI I, velocity of deposition can provisionally be taken as proportional to terminal velocity, and the importance of deposition will increase with spore size. (ii) Viability. It is assumed, in the absence of experimental evidence to the contrary, that viability is not affecting gradients over the short distances discussed here, though the future may show this assumption to be an over-simplification (p. 190). (iii) Available sites. Deposition gradients do not necessarily give rise to observable infection gradients. For instance, a fruit-body ofGanodernm 162 DEPOSITION GRADIENTS AND ISOLATION applanatum will emit spores continuously and copiously for months, and the spores will be diffused and deposited; yet the occurrence ofGanoderma fruit-bodies remains erratic throughout a forest, their occurrence being limited by what may be expressed in the ignorance-blanketing phrase — 'availability of sites'. An infection gradient can only develop when sites such as trap surfaces, nutrient medium, susceptible host-plants, ripe stigmas, or burnt soil, are freely available. The deposition gradient is a regular phenomenon: an infection gradient follows when the deposition gradient is superimposed on unoccupied sites. Ecologically, an infection gradient is a stage in succession, not a characteristic of a balanced state. (iv) Multiple infections. As long as the number of available sites is large, the slope of the infection gradient will be parallel to that of the deposition gradient. But as soon as available sites begin to be used up, the infection gradient will be flattened, the flattening beginning nearest the source, and the relation between the two gradients under simple conditions is given by the multiple-infection transformation (Gregory, '948)- When disease incidence is recorded as the number or percentage of plants attacked, irrespective of whether the plant has one or many lesions, aphid punctures, etc., the percentage will have to be suitably transformed before the formula for deposition can be applied. The need for the transformation may be illustrated by considering a hypothetical example of lOO potato plants, uniformly susceptible and exposed to infection by Phytophthora infestans from a distant source. The first spore that causes an infection must obviously infect i per cent of the plants. A second infecting spore, so long as it falls at random among the lOO plants, will have one chance in loo of alighting on the one plant already infected, instead of infecting a second plant. As the percentage of infected plants increases, the probability that each additional infection falls on a plant already infected (thus producing no increase in the percentage of plants infected) increases greatly. When 99 per cent of the plants are infected, another infection will have only one chance in 100 of falling on the single remaining healthy plant. Different parts of the percentage range, therefore, correspond to very different spore densities per unit area, and the transformation can be neglected only in the lower- percentage categories. Thompson (1924) applied the Poisson distribution to the problem of multiple infection and showed that if N = number of hosts available, and y = average number of hosts infected after the deposition at random among the hosts of x parasites, then y = N (i - e-'^/^), Table XXVI gives x calculated for values of y varying from i to 99-9 per cent; it shows that whereas only one infection is required to bring about an increase from i to 2 per cent, the increase from 98 to 99 per cent 163 THE MICROBIOLOGY OF THE ATMOSPHERE requires sixty-nine infections (that is, 460 minus 391 according to the table). TABLE XXVI MULTIPLE- -INFECTION TRANSFORMATION : PERCENTAGES TO INFECTIONS (< :alculated FROM 5 -FIGURE LOG tables) y% X y% X y% X y% X y% X y% X I I -00 23 26-14 44 57-98 65 105-0 83 177-2 93-5 273-3 2 2-02 24 27-44 45 59-78 66 107-9 83-5 180-2 94 281-3 3 3-05 25 28-77 46 61-62 67 110-9 84 183-3 94-5 290-0 4 4-08 26 30-11 47 63-49 68 113-9 84-5 186-4 95 299-6 5 5-13 27 31-47 48 65-39 69 117-1 85 189-7 95-5 310-1 6 6-19 28 32-85 49 67-33 70 120-4 85-5 193-1 96 321-9 7 7-26 29 34-25 50 69-31 71 123-8 86 196-6 96-5 335-2 8 8-34 30 35-67 51 71-33 72 127-3 86-5 200-2 97 350-7 9 9-43 31 37-11 52 73-40 73 130-9 87 204-0 97-5 368-9 10 10-54 32 38-57 53 75-50 74 134-7 87-5 207-9 98 391-2 II 11-65 33 40-05 54 77-65 75 138-6 88 212-0 98-5 420-0 12 12-78 34 41-55 55 79-85 76 142-7 88-5 216-3 99 460-s 13 13-93 35 43-08 56 82-10 77 147-0 89 220-7 99-1 471-0 14 15-08 36 44-63 57 84-40 78 151-4 89-5 225-4 99-2 482-8 15 16-25 37 46-20 58 86-75 79 156-1 90 230-3 99-3 496-2 16 17-44 38 47-80 59 89-16 80 160-9 90-5 235-4 99-4 511-6 17 18-63 39 49-43 60 91-63 8o-5 163-5 91 240-8 99-5 529-8 18 19-85 40 51-08 61 94-16 81 1 66- 1 91-5 246-5 99-6 552-1 19 21-07 41 52-76 62 96-76 81-5 168-7 92 252-6 99-7 580-9 20 22-31 42 54-47 63 99-43 82 171-5 92-5 259-0 99-8 621-5 21 23-57 43 56-21 64 102-2 82-5 174-3 93 265-9 99-9 690-8 22 24-85 With acknowledgements to Drs. S. B. Fracker and H. A. Brischle of the United States Department of Agriculture. If the distribution of infections is at random among the hosts, a straight hne will be obtained when the logarithm of the percentage that remain uninfected (100 minus the percentage infected) is plotted against the number of infections. The slope of the line is given by 'b' = — 0-00434. Blackman (1942), in ecological studies of flowering-plant communities, found that the occurrence of plants on quadrats may depart from the expected random distribution and yet still give a reasonably straight line when plotted as before. The slope of the observed line, however, differs from that of the expected line. Blackman's 'correction factor', the ratio of the slope of the expected line to the slope of the observed line (K = bexpccted/boLservcd), thcn givcs a uscful measure of the departure from the random arrangement. Various other mathematically plausible formulations of this deviation have been attempted and are reviewed by Fracker & Brischle (1944). Non-random distribution may result from various factors such as repulsion 164 DEPOSITION GRADIENTS ANT) ISOLATION between individuals ('under-dispersion' as understood by ecologists), and K will then be less than unity. Aggregation ('over-dispersion'), on the other hand, leads to K being greater than unit}". Aggregation may be due to such causes as local spread of infection, progeny remaining near parent, or to local differences in susceptibiHt}'. Plank (1946) has given a useful test for detecting aggregation in the field. (v) Infection efficiency. Gradients may be observed either by directly counting the numbers of spores deposited on equal areas at different distances, or by counting some consequent effect such as colonies, leaf- spots, or diseased plants. Usually these gradients will be the same when numbers are plotted against distance, but the infection gradient will be much lower than the deposition gradient. Even with a nearly loo-per-cent viable spore suspension, the general experience in inoculation tests with plant pathogens is that only a small proportion of the spores deposited will give rise to a lesion — even when conditions for infection are as favour- able as possible ; in unfavourable conditions the formation of lesions falls to zero. The proportion of spores successfully infecting is termed 'infec- tion efficiency' by Gaumann (1950, p. 1 57), and values recorded by various workers under highly favourable experimental conditions include: Phytophthora infestans 6-5 per cent, Alternaria solani 17 per cent, and Septoria ly coper sici 0-2 per cent (all on tomato leaves, cf. McCallan & Wellman, 1943); Botrytis sp. on Viciafaba 5 per cent (F. T. Last, un- published); Peronosporu tabacina approximately i per cent (Waggoner & Taylor, 1958). Rust fungi show relatively higher efficiencies. Thus Petersen (1959) observed penetration by 30 per cent of uredospores of Puccinia graminis tritici on wheat leaves; but, at the high concentrations tested, over 100 uredospores were required to produce one sporulating uredosorus {see Durrell & Parker, 1920). \\'ith the same fungus, Rowell & Olien (1957) obtained as many as eleven uredosori per 100 spores applied. McCallan (1944) evidently obtained about 10 per cent efficiency from uredospores of Puccinia antirrhini. R. H. Cammack (personal communication) ob- tained 15 to 23 per cent efficiency when inoculating Puccinia polysora to susceptible maize. Considering its importance in plant patholog}', it is remarkable how little attention has been given to infection efficiency. Its value must vary with dispersal conditions, but commonly the height of the infection- gradient curve above the x-axis will be only about one-hundredth that of the deposition curve, though the slopes of the two curves would be similar. (vi) Secondary spread. From turbulence theory we can predict only primary dispersal gradients. As Waggoner (1952) points out: 'Because proximity of a source is relatively more important than strength of the source, spatial distribution of diseased plants becomes more uniform as secondary infection progresses.' The infection gradient will therefore be flattened if observed long enough after deposition for secondary spread 165 THE MICROBIOLOGY OF THE ATMOSPHERE to occur around the primary lesions. Examples of this effect are found in Pape & Rademacher (1934), Zogg (1949, Figs. 10 and 11), Waggoner (1952, Fig. 2), and Cammack (1958). (vii) Sampling. At low levels of infection, the size of samples must be increased, as otherwise the sampling error becomes large {see Finney, 1947)- Empirical Methods As noted in Chapter V, gradients can be represented by either an empirical or a theoretical model. In the empirical method we make the curve to fit the data, but in the theoretical method we test the fit of the data to the curve. Frampton et al. (1942) concluded that incidence of some insect- transmitted virus diseases decreased logarithmically with distance. Zentmeyer et al. (1944) studied the spread of the Dutch elm-disease pathogen, CeratostomeUa tilmi, which is transmitted from tree to tree by the elm-bark beetle, Scolytus niulthtriatus. Their data, to distances of about 84 metres, indicated that the probability of infection decreased with the logarithm of the distance. Most subsequent curve-fitters agree that such decrease is logarithmic. Wolfenbarger (1946, 1959; see also Wadley & Wolfenbarger, 1944), in valuable surveys of literature on the dispersal of bacteria, spores, seeds, pollen, and insects, concluded that the observed data could be fitted by one of the two following equations : E = a + b (log x), or E = a + b (log x) + c (i/x), where E = the expected value, x = distance from source, and a, b, and c are parameters derived from the observed data. Values of the para- meters a, b, and c, obtained by Wolfenbarger, showed enormous variation between the numerous sets of published results, and it is not possible to make any kind of generalization using this method, or to use the para- meters, as given, to predict gradients. E. E. Wilson & Baker (1946, 1946^^) made field observations in California on the dispersion pattern of apricot brown-rot {Sclcrotmia laxa on Primus armeniaca), and they also liberated Lycopodium spores experi- mentally. They fitted the following equations to their data : (i) y = loo/x^ for the gradient of aerial spore concentration (con- centration at distance X2 and at subsequent distances being expressed as a percentage of the concentration at x^). (2) y = 100 (i + a)2 / (x + a)^, for the gradient of infection by airborne spores, where a is a parameter for the experiment. Dispersal by insect vectors is usually fitted by a logarithmic expression. Bateman (1947^) found that the proportion (F) of contamination of 166 DEPOSITION GRADIENTS AND ISOLATION seed crops by insect cross-pollination at distance (D) was fitted by the expressions : F = ye-'^^s or D ' where y = contamination at zero distance, and k expresses the rate of decrease of contamination with distance. Gregory & Read (1949) concluded that data for insect-borne viruses could be fitted well by the empirical expression : log I = a + bx, where I = number of infective punctures at a distance x from the source after a given time, and a and b are constants for any one given set of field conditions. With the empirical method an equation canusuallybeobtained, contain- ing at most three parameters, which gives a good fit to any particular set of data. However, it is not easy to compare the results obtained by different workers. The empirical method is difficult to use because point, strip, and area sources are not distinguished, the multiple-infection transforma- tion has not been applied even w^hen appropriate, and the parameters as calculated from the data correspond to no obvious natural phenomena; they may even conceal different units of measurement varying from centi- metres to nautical miles! Progress with the empirical method requires attention to these matters and, especially, the adoption of a standard metric unit of distance. Diffusion and Deposition Theories The more difficult, but potentially more useful, theoretical approach is derived from the diffusion phenomena described in Chapter V. w. Schmidt's theory Schmidt (19 18, 19 19) used his diffusion theory to calculate Q_x/Q^o> the fraction of the eddy-diffusing spore-cloud which remained in the air at distance x. To do this he assumed that any part of the cloud whose diffusion path would have brought it down to ground-level, would have been removed from the cloud by deposition. He represented the terminal velocity of the particle as equivalent in effect to tilting the ground, and he gives a table from which the 'dispersal limit' under average conditions of wind-speed and turbulence could be read for particles with different terminal velocities. Dispersal hmit (V) was defined as the distance exceeded by only i per cent of the particles liberated. Schmidt's theory was developed further by Rombakis (1947), who brought the fall velocity of the particles into the differential equation and 167 THE MICROBIOLOGY OF THE ATMOSPHERE replaced the arbitrary 99 per cent of Schmidt's dispersal Kmit by the concept of 'probable line of flight'. Rombakis postulated that a point P at height z and time t will be a point on the probable line of flight when it is statistically equally probable for a particle to occur above or below P. At the 'probable flight range' (a distance about one-tenth of Schmidt's 'V'), 50 per cent of the particles liberated will have been deposited; 'probable flight height', and 'flight duration', are similarly defined. Rombakis also reached the interesting conclusion that the 'probable final velocity' of a particle is half its terminal velocity in still air. These concepts were applied to the dispersal of fungus spores by Schrodter (1954), who used Falck's (1927) calculations of terminal velocities to predict probable flight ranges, altitudes, and durations for various spore- sizes, wind-speeds, and values of the turbulence coefficient. Examples from Schrodter's extensive calculations are given in Table XXVII. In his later, valuable review of the topic Schrodter (i960) uses Rombakis's method for estimating distance of dispersal, and Sutton's equations for the concentration of the spore-cloud. TABLE XXVII 'probable flight range' (Schrodter, 1954) based on rombakis's MODIFICATION OF SCHMIDT's THEORY Probable Interchange maximum Probable Probable flight range (km.) in coefficient, A flight flight winds of velocity (gm./cm. sec.) altitude (metres) duration 2 (metres/sec.) 6 10 small spores (5 < 3 aO o-i 5-41 17 hr. 124 km. 371 km. 618 km T.V.* 0-035 lo-o 541 72 days 12,400 37,100 61,800 cm. /sec. 50-0 2,705 I year 62,000 185,500 309,000 large spores (22 X id jj) o-i 0-19 I min. 0-2 km. 0-5 km. 0-8 km T.V.* 0-975 lO-O 19 2-2 hr. 16 48 80 cm./sec. 50-0 95 II hr. 80 240 400 T.V. = terminal velocity. DEVELOPMENT OF SUTTON's THEORY Sutton's equations predict concentration when there is no loss by deposition. To adapt them for particles which deposit appreciably during travel, it is necessary to calculate Q^^? ^^he number of spores remaining in suspension after the cloud has travelled a distance x, from the equation : A\-hn) -] X h (P- 77). Clx = CLo exp. 2pX^ V(7r)C(l -hn)_ Expected depositions at various distances from point, line, strip and area sources can then be calculated (Gregory, 1945, and unpublished). 168 DEPOSITION GRADIENTS AND ISOLATION (i) Point source. D, the /c/^z/ number of spores deposited on an anniilus I cm. wide at a distance x from a point source, is given by : Vin) Cx-"' (ii) And d, the mean number of spores deposited per square centimetre on an annulus i cm. wide at a distance x from a point source, is given by: 277tCx*(- + ^* (iii) And d„., the number of spores deposited per square centimetre at a distance x do\Miwind from a point source, is given by: _ p2Q^x " ttCV"' (iv) Line source, d,,,., the number of spores deposited per square centimetre at a distance x downw ind from a line source, is given by : . _ P2Q.X d... = ,. !^:t^^ — .x'-'"'" *aw W V(^)Cx^'" di.v is thus numerically equal to D. (v) Strip source, d^.^, the number of spores deposited per square centimetre at a distance x do\Miwind from a strip source of width w, is given by integration as : P4ax \/{tt)C (2 — w) Gregory (1945, p. 69) gave a set of calculated values for D, d, d,,. and di„., assuming Q^o = io^°, p = 0-05, C ~ o-6 (cm.)% ;;/ = 1-75 and 1-24. With further knowledge of these parameters, revised calculations are now given (pp. 1 7 1-6). Meanwhile, in a series of papers on pollen contamination of seed crops, Bateman (1947, 1947':?, 1947'^) gave many examples of dispersal gradients, adopting and simplifying the Gregory formulae and using a regression method for testing the adequacy of the diffusion theories of W. Schmidt and of Bosanquet & Pearson (1936). Bateman's method also makes it possible to estimate p/C and Q_„, provided all measurements are expressed in the same units. Regression tests showed that pollen dispersal was best fitted by the 1945 formulae of Gregory, but the data were inade- quate for choosing between the values 1-76 and 1-24 for the parameter m of Sutton's equations. Combining field inoculation studies with eddy-diffusion models, Waggoner (1952) made an important contribution by adapting the for- mulae: (i) to allow for non-isotropic turbulence; and (2) to incorporate 169 THE MICROBIOLOGY OF THE ATMOSPHERE the ratio of spores deposited to percentage leaves (or leaflets) diseased. Waggoner used the findings of E. E. Wilson & Baker (1946) and of Scrase (1930), to put the variance of concentration of vertical distribution (ct^^) as equal to 4/9 of the variances in the x and y planes. This led to the formula (in our symbols) : X = -^^^exp.[-x-(r2 + 9z2/4)], where r^ = x^ + y^- From observed gradients of late-blight [Phytophthora infestans) around artificially inoculated potato plants in the field, Waggoner took k as the ratio of spores deposited per square centimetre to the proportion of leaflets diseased, and estimated the parameters p (deposition rate) and Qp/k from the equation ^ o-i35p(Qp/k) / ^ o u where D = proportion of leaflets diseased. (In Waggoner's tests, D was comparatively small and did not need the multiple-infection transformation.) In his experimental potato plots in Iowa in 1949 and 1950, respectively, Waggoner found p = o- 12 and 0-15, and Qp/k = 18 X 10'' and 7-8 X 10^. Subsequently for Perojiospora tabacma Qo was estimated at approximately 10^ spores per sq. cm. of lesion per day (Waggoner & Taylor, 1958). For assessing concentrations of radioactive clouds, Chamberlain (1956, pp. 20-27) at Harwell developed expressions combining the eddy- diffusion equations (of Sutton, 1947) and the allowance for loss by deposi- tion (of Gregory, 1945). Chamberlain also modified the equations to allow for elevation of the source above ground-level, illustrating the fact that elevating the source greatly reduces the loss by deposition. Re-calculation of the Deposition Gradient With new information and further development of the statistical theory, deposition gradients can now be calculated for a wider range of conditions than was possible earlier. The values worked out here are offered as giving a useful indication of trends, but much further work still remains to be done. Chamberlain's modifications have been adopted, in preference to Waggoner's, because the latter uses the parameter k, which can vary over a wide range down to zero — depending as it does on leaflet size in a particular crop, and on how favourable conditions were for infection. Two sets of graphs are provided. Fig. 24 gives Q_x for m ~ i"]Si C — 0-6 (metres)^, and a variety of heights and distances from the source.* * The parameter 'C is now given in units of (metres)*, in place of (cm.)* used in Gregory, 1945. 170 DEPOSITION GRADIENTS AND ISOLATION Another set, Figs. 25-27, gives values of D, d, d,,., dj,,,, and d^^ on a logarithmic scale for sources of various geometrical form, assuming no loss from the cloud by deposition. To predict a gradient, the deposition values are first read-off for various distances from the appropriate line on Figs. 25 to 27, and allowance can then be made for loss from the cloud at each distance by reading-off Q^^ as a fraction of Q^„ in Fig. 24. X metres 100 1.000 loooo 1000,00 Fig. 24.— Fraction of spore-cloud remaining airborne (allowing for loss of spores from spore-cloud by deposition to ground), expressed as Q.x/Q.o- Calculated from Chamberlain (1956): for ?n = 1-75; Cz = 0-21 (metres)J; height of source above ground, h = o-o, o-i, i, 10, and ICO metres; distances from source, x = i metre to 100 km. CALCULATION OF Q_x Except in experimental spore-liberation tests, the quantity liberated, Q_o, is usually unknown. However, cloud concentration and deposition are proportional to Qp so, although the height of the gradient line will depend on the source-strength, its slope will be unaffected by the value of Elevation of source has been allowed for by using the equation number 52 of Chamberlain (1956), to calculate values of Q^^ for heights h = o, o-i, I, 10 and 100 metres. Fig. 24 shows that elevating the source decreases deposition on ground near the source, and in using Figs. 25-27 for elevated sources it is important to neglect those parts of the curves before deposition starts. This decreased deposition near an elevated source was confirmed experimentally by Colwell (1951), who liberated P-32 radioactive Phius pollen at 3-5 metres above ground-level into a wind averaging 8-i metres 171 THE MICROBIOLOGY OF THE ATMOSPHERE per sec. Colwell sampled simultaneously with vacuum cleaners and Petri dishes on the ground, and estimated the number of pollen grains with the help of a Geiger-Miiller counter. Maximum deposition in this experiment was obtained at 5-8 metres horizontal distance from the source. For Sutton's parameters, the values adopted here are : Cx = Cy = 0-21 (metres)^, Q = 0-12 (metres)*, and m = 1-75. As a measure of deposition, p has been chosen in preference to Chamber- lain's Vg/u, because the wind-speed under which deposition occurred is usually not known in the open air. However, the curves, which are calcu- lated for p = 0-05, o-oi, and o-ooi in Figs. 25, 26, and 27, respectively, can be used with Chamberlain's velocity of deposition if the wind-speed is known, because provisionally it is taken that p = Vg/u. More field experiments are needed before we can decide whether deposition over rough ground depends on time or on distance — a contrast analagous to the rival diffusion theories of W. Schmidt and Sutton. From Chapter VII it seems that p depends on the terminal velocity of the par- ticle, and numerically it has a value of about one-fiftieth to one-hundredth of the terminal velocit)' expressed in centimetres per second. No allowance has been made in these calculations for the hitherto unexplained larger values of p which have been observed within a few metres of a source near ground-level. Some plant pathogens have large, readily-impacting spores. When liberated at, or near, ground-level, although a small proportion may travel vast distances, most of these spores probably do not move very far from the source. Fig. 24 suggests that for 'impactor' spores with a deposition coefficient of p == 0-05, and liberated at 10 cm. above ground-level, 94 per cent would be deposited within 100 metres of the source.* But 'penetrators', with p = o-ooi, would be deposited only to the extent of 6 per cent within 100 metres under similar conditions. In spite of neglecting terminal velocity during the diffusion process — a feature criticized by Schrodter(i96o) — this method predicts a much more rapid loss of material from the cloud than does Rombakis's method: perhaps because the ground-level concentration of the cloud is restored by downward diffusion much more efficiently than by sedimentation. On our theory, 50 per cent of large spores liberated at 10 cm. above ground would be deposited within 10 metres, but, according to Schrodter's calculations, this fraction would travel at least 200 metres (Table XXVII). Such rapid decrease in the value of Q^^ near a ground source emitting large spores is supported by experimental evidence which has already been summarized in Table XIV (Chapter VII). Yet some * My mistake (Gregory, 1952; see correction, 1958) in stating that 99-9 per cent of those liberated at ground-level would be deposited within 100 metres, was due to misreading Qo = 10'^ as 10^" in my own table (Gregory, 1945, Table 21) — not to the misunderstanding suggested by Schrodter (i960, p. 178). 172 o DEPOSITION GRADIENTS AND ISOLATION lO lOO I.OQO lO.OOO Distance in metres Figs. 25 to 27. — Dilution of spore-cloud by eddy diffusion. Calculated for ;h = 1-75 (also m = 1-24 and 2-0), expressed as logarithms of: d, dw, diw, D (and also for daw with width of 100 metres). Fig. 25. — Deposition coefficient, p = 0-05. THE MICROBIOLOGY OF THE ATMOSPHERE lO lOO I.OOO lO.OO O D=d Iw OO IpOO lO.OOO J L Distance in metres Fig. 26. — Deposition coefficient, p = o-oi. 174 o o o _ (D J tXO O DEPOSITION GRADIENTS AND ISOLATION lO lOO I.OOO lO.OOO 1 1 1 1 p = OOOI . \^ ^S\> k v^" 1 %\v\ A lOO lOOO lOOOO I d d. Distance in metres - 2 - 3 - 4 V 5 -6 D=d, - 8 Fig. 27. — Deposition coefficient, p = o-ooi. 175 '^ / WOODS HOLE, i THE MICROBIOLOGY OF THE ATMOSPHERE workers have considered that these large estimates for deposition near the source are incompatible with the facts about dispersal in the upper-air, and over long distances, which are discussed in Chapters X and XIV. This dilemma is considered further below (pp. i8o and 197). The effect of values of m lower than 1-75 is to increase deposition near the source and to decrease it at greater distances. With liberation at ground-level, the distance at which deposition per unit area is equal with either m = 1-75 or 1-24, lies between x = 10 and x = 100 metres. For other heights of liberation above ground-level, calculations are not yet complete. Application of Gradients The results given graphically in Figs. 24 to 27 can be used in various ways, some of which are illustrated in the following examples. (i) Deposition at a point downwind. Assuming that a million uredo- spores are liberated at a point one metre above ground-level under standard atmospheric conditions, what number would be deposited per square centimetre of ground at a point 100 metres downwind? Choose Fig. 25, because p = 0-05 is the relevant coefficient of deposition for uredospores (rather than Figs. 26 and 27, which refer to smaller spores). Choose the group of three lines marked 'd,,.' (deposition per square centimetre downwind of a point source). Choose the middle of these three lines, as m = 1-75 under standard conditions. Read-off the value of 'logarithm (deposit -^ Qp)' for the distance of 100 metres. This value is approximately = ^-4. As in this example Qp = lo*', log Qp = 6-o, so we now have log d^^ = 6-o + 5-2, therefore log d^^ = 1-4, and d,,, = 25 spores per square centimetre. But this calculation assumes no depletion of the cloud, and deposition must now^ be allowed for by replacing Q^^ by Q^^- From Fig. 24, choose the group of curves marked p = 0-05 and, as the height of liberation is i metre, choose the line for h = i, and read off the value for Q^x/Q_o ^t 100 metres. This value is approximately 0-25, indicating that only about a quarter of the spores liberated are still in suspension at the distance of 100 metres from the source. We therefore have the value of the corrected d^ = 0-25 X 25 = 6. The answer to the problem is that we predict a deposit of six uredospores per square centimetre under the conditions postulated. (ii) Fitting theoretical curve to observed data can best be illustrated by a published example that gives actual distances and percentages of leaves infected. Relative distances are useless because the slope of the gradient- line is characteristic of an absolute distance, and also with relative percentages we cannot apply the multiple-infection transformation if the data require it. 176 DEPOSITION GRADIENTS AND ISOLATION Waggoner (1952) gives two sets of data on the spread of potato late- blight {Phytophthora infestans) at Clear Lake, Iowa. 'Point source' foci were established in field plots by artificial inoculation. Hollow curves were plotted, showing the proportion of leaflets diseased at various dis- tances from the source on the ninth day after inoculation in 1949 and the eighteenth day in 1950. Reading-off observed values from Waggoner's graphs gives the values for 1949 and 1950 as plotted in Fig. 28; these agree well with the slope of the predicted gradient (d) for dispersal from a point source. (The highest incidence was between 7 and 8 per cent, so the multiple-infection transformation is unnecessary.) ^Q. - 3 - 4 10 - 5 1000 source 100 metres from Fig. 28. — Infection gradients of potato late-blight {Phytophthora infestans) observed by Waggoner (1952) at Clear Lake, Iowa, in 1949 and 1950; compared with d = theoretical line for deposition downwind from a point source, assuming m = 1-75, Cy = od (metres)i, p = 0-05, and allowing for Q_x. (iii) Fixing limits for isolation. The formula for d (average deposition at distance x in all directions around a point source) may be chosen when spread of a disease within a field, or gene dispersal, is being considered. For safe decision on isolating a healthy crop from contamination, the use of dw may be preferable as it predicts maximum concentration do\\Tiwind. Suppose we wish to estimate the upper limit of the percentage of those plants infected with Ustilago tritici to be expected at 50 metres distance from a lo-metre-square plot of smutted wheat acting as a source — knowing from past local experience that seed saved from the source plot M 177 THE MICROBIOLOGY OF THE ATMOSPHERE itself is likely to produce only 5 per cent of diseased plants. We choose the curve for deposition downwind of an area source, with liberation height h=i metre, under standard conditions (w = 1-75), and p = o-i. Assume that infection at i metre distant is also 5 per cent. From Fig. 24 we find that Q_x is o-88 Q^„ at 50 metres distance, and from Fig. 26, d^ at 50 metres is only about 28 per cent, of the value at i metre. The expected maximum level of infection at 50 metres is therefore 28/100 X 0-88 X 5 per cent = 0-12 per cent. Under similar conditions, Oort (1940) recorded about half this decrease at 50 metres. At i km. down- wind of a ID X 10 metre plot, it would be more appropriate to adopt the curve for a point source and estimate the level at 1-5 X 10"^^ per cent. (iv) Relative contributions of near and distant sources. A susceptible plant in a field is often exposed simultaneously to infection from near-by and distant sources. The problem arises as to the relative importance of a few diseased plants within the crop as compared with possible massive sources of infection in neighbouring fields. Quantitative answers have to be guessed to set limits of tolerance for issue of health certificates, and the data from gradients can be used to improve such guesses. A simple case would be to inquire how many diseased plants at 1,000 metres distance would give the same amount of deposition as that on a plot of 100 metres radius with a single diseased plant at its centre, assuming that distribution by wind is uniform in all directions. From Fig. 25, over the range x = to x = 1,000 metres (for m = i*75, p = 0-05, and h = o-i metre), the total deposition is Q^^ — Q^^ (for x = 100 metres) = 1-0 0^0 — o-o6 0^0 = 0-94 Q_o. From Fig. 27, the average deposition from one plant at a distance x = 1,000 metres is 4 x io~^^ Q_o spores per sq. cm. The circle of radius 100 metres around the single plant contains 3-14 X 10^ sq. cm., so a plant at 1,000 metres contributes: (4 x 10^^ Q^„) X (3-14 X 10^) = 12-5 X 10^ Q_o spores. The number of plants i km. away required to equal the contribution of one plant within the 100 metre radius circular plot is therefore approximately: 0-94 Q_o/i2-5 X 10^ Q^o =- 7,500 plants. Characteristics of Gradients* Some observed gradients fall off much more steeply than the line calculated for m = 1-75. For example, the only gradients recorded for aecidiospores oi Puccinia graminis (A. G. Johnson & Dickson, 1919; and Lambert, 1929) follow closely the expected gradient for a strip source with m = 1-24, suggesting that these spores are dispersed when tur- bulence is small. Data on potato late-blight {Phytophthora infestans) contributed by Limassct (1939) and Bonde & Schultz (1943), and for Peronospora destructor on onions by Newhall (1938), also suggest dis- persal under low-turbulence conditions. * Other properties of gradients are discussed by Plank (1960). 178 DEPOSITION GRADIENTS AND ISOLATION Further study will show whether certain groups of fungi are usually dispersed only under low-turbulence conditions. If it is confirmed that they are, a much smaller isolation distance could safely be permitted between disease sources and healthy crops than is permissible with patho- gens that are dispersed in normally turbulent air, because with a small degree of turbulence the contaminated area is narrower and more intense. Meanwhile it is safer for isolation purposes to choose for prediction the line from the appropriate set which has been calculated for normal turbulence, m = i-JS- The rapid decrease in deposition with increasing distance from the source helps to explain the observations of Schmitt et al. (1959), who showed that many small rust infection foci distributed over a wheat field are more destructive to the crop than the same number gathered together at a single focus. The fact that the slope of the curve is a characteristic of a particular distance can be used in the field to locate an unknown source. If lesions (or spores) are uniformly scattered over an area, we can assume that the source is so distant that the place of observation is on the tail of one (or many) dispersal gradients. Where counts increase rapidly, there is evidence that a source is being approached. Stakman & Hamilton (1939) located a large rust-infected barberry bush in western Minnesota, by first identifying unusual rust races in the vicinit}' and then tracing the telial stage of the rust on grasses leading up to the bush. Topographical Modification of Gradients Diflfusion has been treated so far under nearly ideal conditions, but the literature contains information on the effects of topographical features which modify the gradient. Pollen of wind-pollinated plants is distributed in typical gradients, as shown, for example, by Roemer (1932), Jensen & B^gh (1942), Jones & Newell (1946), and Bateman (1947a). During strong winds in open fields, Jensen & B(/gh's catches of pollens of rye, ryegrass, cocksfoot, timothy, sugar-beet and mangold, on sticky microscope slides, showed a steady decrease with distance up to i ,200 metres from the source field. But they found that hedgerows and plantations protected ryegrass, cocksfoot, and mangolds, in proportion to the height of the obstacle. In tests near to the source, a protection corresponding to an isolation distance of about 200 metres of open ground was obtained behind hedges, even so far do\Miwind as 5 to 10 times their height (cf. Rider, 1952; Schrodter, 1952; U.S. Weather Bureau, 1955; Cabom, 1957). Gene Dispersion The physical mechanisms of wind dispersal have a bearing on popu- lation genetics. Theories of gene dispersion in populations at first assumed 179 THE MICROBIOLOGY OF THE ATMOSPHERE that there is a random scatter around the source. In such a distribution the gradient in any direction would have the form of one-half of the normal frequency curve. S. Wright (1943, 1946) studied genetic effects of isolation distance, and his methods were applied by J. W. Wright (1953) to compare dispersion distances of pollens of various forest trees with a view to delimiting a 'neighbourhood' for race formation. Simple sticky-slide traps were exposed at various distances around isolated trees, and pollen counts were used to find the standard deviation of the scatter. Observed values for the standard deviation were as follows: ash, 17-46 metres; Douglas-fir, 18 metres; poplar and elm, 300 metres or more; spruce, 40 metres; Atlas cedar, 73 metres; Lebanon cedar, 43 metres; and pinyon (pine), 17 metres. Dispersal data were well fitted by the Gregory formulae, but not by theories which assume that the trajectory of each grain can be calculated from the rate and distance of fall, and the wind velocity. Bateman (1950) questioned whether gene dispersal is statistically 'normal', and showed by his regression method that many observed distributions — including those of fungus spores, passively borne insects, pollen, and wind-dispersed seeds — were highly leptokurtic, i.e. the peak and tails of the distribution are exaggerated at the expense of the shoulders. Compared with a normal frequency distribution having the same standard deviation ('same over-all degree of inbreeding'), the leptokurtosis charac- teristic of passive airborne dispersal produces more breeding between close relatives and simultaneously more breeding between distant relatives (see also Parker-Rhodes, 195 1). 180 XIV LONG-DISTANCE DISPERSAL Dispersal of microbes over long distances is an ever-present, world-wide phenomenon. Its experimental study is almost non-existent; but there is circumstantial and observational evidence of its magnitude, some of which has been reviewed by J. J. Christensen (1942). CONTROXTRSY ON THE IMPORTANCE OF THE AIR-SPORA The discovery of air dispersal in any group of plants has usually been followed by controversy between opposing specialists, some maximizing and some minimizing the significance of the phenomenon. Exaggeration is dangerous, and the object of this book is to present evidence from which balanced conclusions can be derived. Views current on air hygiene affect the design and planning of hospitals. The air dispersal of human pathogens was confidently accepted in the golden age of bacteriology, but was gradually discounted after the experi- ments of Fliigge which focused attention on the limited scatter of drop- lets from the mouth and nose (cf. Chapter XII). The balance has now been restored by Wells (1955) and others who have stressed the floatabilit}' of 'droplet nuclei'. In plant patholog}' the two schools of thought have been prominent simultaneously. Butler (191 7) held that spores of plant pathogens could be transported for short distances by air or by rain-splash, but claimed that 'the distance to which spores may be carried in the air has often been exaggerated in the past, and is much less than might be expected'. The discontinuous spread of plant pathogens, he considered, was likely to occur on seeds, plants, and horticultural produce; but 'infection by spores carried through the air from remote centres is not a contingency which needs to be taken seriously into account'. Naumov (1934) held that human activity accounted for most long-distance transport of fungus pathogens and that, in the absence of host plants, fungi were dispersed with extreme slowness. Endothia parasitica in North America could not cross, in a period of 1 years, a 45-60 km.-wide tract that was free from chestnut trees; the spread of a disease of palm trees was estimated at only 4-5 km. per annum. The literature contains many instances of fungus pathogens failing to infect hosts at distances of a few metres (e.g. H. W. Long, 1914), or to 181 THE MICROBIOLOGY OF THE ATMOSPHERE colonize apparently favourable environments. Not all of these necessarily represent the failure of transport : genetic differences in host populations, or the pre-establishment of competitors, may explain many puzzling failures. Diverse views have also been held about Bryophytes and Pterido- phytes; however, most authors accept the possibility of their dispersal over great distances, although the minimizing view has been held by some {see Pettersson, 1940, p. 22). Probably these differences in viewpoint will be resolved by quanti- tative studies. The maximizing view, if held too strongly, may lead to fatalism and to the neglect of local hygiene and of reasonable precautions when transporting plants from place to place. However, evidence presented in earlier chapters stresses the overriding importance of local sources. The minimizing view, if applied to certain diseases, may lead to over- reliance on local hygiene and neglect of protection by chemical and genetic measures. It may also increase the danger of introduction of a vigorous organism by neglect of quarantine precautions. Long-distance dispersal will be discussed under two headings: (i) diffusion theories extrapolating from short-range experiments; and (2) observations on distant dispersal of inorganic and radioactive particles, of rust fungi, and of other micro-organisms. The problem is complicated by the curvature of air-mass trajectories, by the possibility that spores are re-concentrated within cumulus and other clouds, by unpredictable removal in precipitation, and by loss of viability. Theoretical Discussion Presumably the tropopause limits the vertical expansion of the spore- cloud, and an extensive temperature inversion at a lower level may have the same effect. After the spore-cloud has travelled perhaps 20 km. or less, diffusion will become two-dimensional, and concentration may be expected to decrease more slowly with increasing distance than it would when the spore-cloud was nearer the source, where its diffusion was three- dimensional. In the limiting case of a land-mass acting as a long strip or area source, the decrease of concentration at 20 km. or more out to sea can be expected to depend only on depletion of Q^^ (the fraction of the cloud remaining in suspension). Sutton's theory appears satisfactorily to describe dispersal of microbes in air up to the limit of distance studied experimentally, but there is some doubt about its use for distances greater than i km. or heights above 30 metres. This is emphasized by Pasquill (1956), who dispersed fluorescent dusts and sampled, both on the ground and by aeroplanes, at distances of up to 64 km. and at heights of up to 1220 metres. In these tests the height of the cloud was much less than the width : the width did not increase uniformly with distance and the angle subtended at the 182 LONG-DISTANCE DISPERSAL release point by the cloud at the greatest distance was only about half that subtended by the cloud at i or 2 km. (Charnock, 1956). At distances greater than those considered in Chapter XIII, the difference between turbulence in a vertical and horizontal direction evidently becomes im- portant. Further diffusion is limited by the tropopause, if not by the top of the turbulent layers of the atmosphere. Samples taken after the accidental emission of about 20,000 curies of iodine-131 from a stack 122 metres high at Windscale atomic pile on 10 October 1957, have provided detailed records of ground contamination by the main plume over distances up to 290 km. (Booker, 1958). These records provide some evidence over longer distances for which microbial data are lacking. Plotted on a log.-log. scale, the radioactivity'- deposition curves, whether measured on herbage or in milk, are relatively flat up to 15 km. — a fact that is consonant with the height of emission. From about 28 km. onwards the slope is similar to that of d„. for m = 1-75. There seems no reason, therefore, why we should not use our formulae for distances of several hundreds of kilometres — bearing in mind that, as in the Windscale accident, the trajectory of the cloud is not likely to be in a straight line over the Earth's surface. Another view of long-distance dispersal of 'crowd diseases' of crop plants comes from Plank (1948, 1949, i949(^i', i960), who used gradients to define the novel concept of a 'horizon of infection' around a field, from beyond which the amount of infection received is negligible. Crowd diseases are defined as : 'diseases which neither spread far in considerable amount nor persist long in the soil'. They can be controlled by mixed cropping and by isolation. Plank put forward the generalization that 'if disease entering fields can easily be controlled by isolation, it can also be controlled by making the fields larger and proportionally fewer' ; this is thought to be true, \\hether infection enters from uncultivated plants outside the field or moves from field to field. For airborne spores travelling over distances between fields, Plank empirically estimates the probability' that a spore will settle at a distance 'x' from its source as p = k/x", where 'k' and 'n' are constants. From published data n is 2 or more for distances over 30 metres, and approaches 4 as the distance from the source increases. With a number of uniform fields scattered evenly over a large area. Plank sho^^■s that, if Q^is the number of spores received by one field from immediately neighbouring fields, the total number of spores received from all other fields, however distant, would be : If n = 2 or less, the series is divergent and there is no horizon. If n is more than 2, the series is convergent and, if n = 3 or more, a useful hor- izon exists (Table XXVIII). The data suggest that, for potato late-blight 183 THE MICROBIOLOGY OF THE ATMOSPHERE {Phytophthora infestans) and onion downy mildew {Peronospora destructor)^ n approaches 4. If n is small, local hygiene will soon be defeated by the flow of infection from outside; but if n = 3 or more, 'a considerable reduction of infection should be possible by a group of neighbours without it being neutralized almost immediately from elsewhere'. TABLE XXVIII EFFECT OF GRADIENT ON DISTANCE OF HORIZON OF INFECTION (Plank, 1949) Percentage coming from Neighbouring fields More than 3 fields away More than 5 fields away Distance of horizon (No. of fields away) n = 2-5 38 40 32 >5o n = 3 61 17 II 5 n = 4 83 3 Observations I 2 RE-COLONIZATION OF KRAKATOA The volcanic island of Krakatoa lies between Java and Sumatra and is almost encircled by land. The eruption of August 1883 blew away the mountainous two-thirds of the island, leaving a hole in the sea-bed 300 metres deep and covering the remainder of the island with lava and ashes. A column of fine dust rose to a height estimated at 27 km., and was carried westwards by the prevailing wind. Eventually this dust circled the Earth repeatedly, spreading over the whole tropical and temperate zones (Symons, 1888). Although this world circulation of dust is relevant to long-distance dispersal of microbes, the story of re-colonization of the island after the destruction of living things, which is usually considered to have been complete, is of even greater significance. Three years after the eruption, the only flowering plants found by Treub (1888) were two species of Compositae and two grasses. There were also eleven species of ferns, and the volcanic deposit was colonized by a film of blue-green algae (Cyanophyceae). All these were thought to have been carried in by wind — the nearest land being 40 km, away. Records of subsequent visits are summarized by Ridley (1930), who concluded that, of the 144 species of flowering plants then reported, 24 per cent were wind-distributed, 42 per cent were sea-borne on floating tree-trunks, and most of the remainder had probably been carried by birds. In addition there w^ere forty-eight species of Pteridophytes and nineteen of Bryophytes — all potentially wind-borne. Boedijn (1940, and see Leeuwen, 1936), who believed that most of the fungi present had been carried to Krakatoa by wind, was impressed also by the paucity of lichens, of which he found only thirteen species (o-i per cent of the world's list), compared with sixty-one species of Pteridophytes and 263 of flowering 184 LONG-DISTANCE DISPERSAL plants. Moreover, none of these lichens inhabited rocks : all were epiphytes which probably arrived on driftwood. Evidently lichens are poorly equipped for wind transport in comparison with Myxomycetes, which were represented by twenty-eight species (7 per cent of the world's list). Only two rusts were recorded (0-02 per cent of those known), but these obligate parasites must needs wait for their flowering-plant hosts. QUANTITATIVE STUDIES Ample qualitative evidence exists of mass transport of microbes by wind over large distances, but there are few quantitative data. The work of Zogg (1949) on one of the maize rusts in the Upper Rhine Valley is of exceptional interest, both for the distance studied and for the complex topography of the area. In that part of Switzerland, Pticcinia sorghi overwinters in its aecidial stage on the wild Oxalis stricta^ which grows in the level area where the Rhine flows into Lake Constance. In early summer, near-by maize becomes infected and a uredospore focus is established from which the fungus is spread by wind up the narrow Rhine Valley. At various dates in the 1945 and 1947 seasons, Zogg measured the incidence of infection for 66 km. (up as far as Chur). Three striking facts about the gradients observed are: (i) the general decrease in the number of uredosori per plant with increasing distance from the source; (2) the flattening of the gradient as a result of secondary spread later in the season; and (3) the great irregularity of the gradient, because the incidence of the rust de- creased locally wherever the valley widens, and increased again where it narrows : a feature which Zogg attributes, no doubt rightly, to the nozzle eflfect of the valley increasing spore concentration, though other ecocH- matic factors may play a part. In the terminology adopted in the present book, both area dose (A.D.) and efficiency of deposition (E) would be increased where the valley narrows. The diffusion theory developed in Chapter XIII referred to dispersal over a level plain, and these modifications, imposed on a spore-cloud by dispersal up an alpine valley, are therefore particularly interesting. In comparison with the distance travelled, the focus of infection near Lake Constance w^ould appear as a point source, though some flattening of the gradient would be expected near the source. As the winds blow charac- teristically up or down the valley, it seems appropriate to consider d,^ (deposition downwind of source). Plotting Zogg's data on a log.-log. scale, we find that the linear regression (calculated as log. y = 5-356 — I -8 1 8 log. x) is compatible for the slope of d.^ as predicted by our theory (Chapter XIII). We take this to indicate that in the valley diffusion by atmospheric turbulence proceeds much as elsewhere, but that deposition is subject to pronounced fluctuations associated with the width of the valley. Observations over similar distances above the sea come from Hessel- man (1919), who traced the transport of tree pollens from the Scandinavian 185 THE MICROBIOLOGY OF THE ATMOSPHERE forests across the Gulf of Bothnia. Pollen was trapped in open Petri dishes on two lightships, 'Vastra Banken' and 'Finngrundct', situated respectively 30 and 55 km. from land, from 16 May to 25 June 1918 (Table XXIX). TABLE XXIX POLLEN TRAPPED ON LIGHTSHIPS IN GULF OF BOTHNIA, 1 6 MAY TO 25 JUNE 191 8 (Hesselman, 1919) No. of grains per sq. cm. ollen type Terminal velocity cm. /sec. Vastra Banken (30 km. from land) Finngrundet (55 km. from land) F/VB X 100 per cent Spruce Pine Birch 8.7 2-5 2-4 696,100 239,000 681,100 408,900 106,900 364,900 58.6 44-4 53-5 Others — 4.300 1,200 279 Pollen can be carried over much greater distances than this. In the month of June, Dyakowska (1948) found pine and spruce pollen falling on the coasts of Greenland, 600 to 1,000 km. from the nearest pine and spruce trees. Still further north, in Spitsbergen, Polunin (1955) noted a deposition of pine and spruce pollen which must have been equivalent to about 200 grains per square metre per day in July and August (Chapter IX). The long-distance record is probably that noted by Hafsten (195 1), who reported Nothofagus pollen in peat on the island of Tristan da Cunha, 4,500 km. from the nearest source in South America. An instance of long-distance transport of moss spores is reported by Pettersson (1940), whose investigations of plant spores in rain-water were described in Chapter XI. At Tviirminne, Finland, during 22-23 July 1936 (when there were persistent rains and light, mainly easterly winds), 104 cc. of rain-water were collected during 15 hours. This sample proved particularly rich in the spores of bryophytes, yielding 300 plants of Marchantia polymorpha and Metzgeria sp. Most remarkable, however, was the occurrence of spores which yielded 278 plants of Aloina brevirostris and 2 which were identified as A. rigida. These are small, annual or bi- ennial mosses of dry calcareous soils, and belong to a genus hitherto un- recorded in Finland. There are a few records oi A. brevirostris in Eastern Europe, but their presence in such large numbers in rain over Finland suggested that the spores must have come from a rich and extensive source area. This, Pettersson suggested, must lie in Siberia, in the region of the River Yenisei — a distance of at least 2,000 kilometres east of Tvarminne. It is true that this conclusion has been questioned by Persson (1944), who thinks that they must have come from a nearer source such as European Russia or possibly Sweden; and by Bergeron (1944), who examined air- mass trajectories for the day in question and reached a similar conclusion 186 LONG-DISTANCE DISPERSAL to that of Persson. Yet Pettersson found that Aloina spores were being deposited in rain to the number of ten thousand per square metre and, akhough their origin is unkno\\Ti, they must clearly have travelled a very long way. As an example of the diffusion of components of the air-spora over long distances, we have quantitative information about the deposition of uredospores of stem rust of wheat {Puccinia graminis tritici) contributed by Stakman & Hamilton (1939). In the early summer of 1938 they studied the deposition of uredospores of this pathogen at various points to the north of ripening winter- wheat fields in the southern United States, which were acting as a vast source of uredospores. At this time of year it could be assumed that spores were not being produced locally in the spring- wheat area in the northern United States, but that infection would follow the arrival of the spore-cloud borne on southerly winds. Table XXX, compiled from Stakman & Hamilton's data, indicates the amount of deposition first in the source area and then at various points farther north. TABLE XXX STAKMAN & Hamilton's (1939) data for long-distance dissemination OF Puccinia graminis (deposition on ground, 24-25 May 1938) Place Approx. distance from source Uredosp per ores per sq. ft 48 hours Dallas, Texas (source area) I 29,216 Oklahoma 300 km. 6,288 Falls City, Nebraska 560 km. 7,680 Beatrice, Nebraska 840 km. 1,968 Madison, Wisconsin 970 km. 192 Another series, taken during 13-14 June, gave the following numbers of uredospores deposited per square foot in 48 hours: Kansas, 336,000; Nebraska, 54,336; Iowa, 21,360; South Dakota, 12,624; Minnesota, 32,256; and North Dakota, 1,344. This long-distance transport of cereal rusts is not merely an occasional risk. On the contrary, it is clear from work done over a vast area that an annual double transcontinental migration through the atmosphere is an essential condition for the development of the rust epidemics which regularly attack cereal crops in North America. Moreover, wind dispersal is relatively unselective, and what happens to rust fungi no doubt happens also to countless other organisms whose spores travel on a global scale. It has been possible to demonstrate this phenomenon for cereal rusts because of the concerted study of a disease of a major food-crop (wheat) by a generation of scientists in plant pathology laboratories scattered over North America, and also because of the strange life-cycle of these rusts, which makes it possible to obtain clear evidence. The evidence derived 187 THE MICROBIOLOGY OF THE ATMOSPHERE from spore-trapping, field records of outbreaks, geographical distribution of the physiological races, and meteorological data, has recently been summarized by Stakman & Harrar (1957, pp. 221-32). To simpUfy a complex story, Puccinia graminis and P. rubigo-vera {= P. triticina), for various reasons, do not survive the cold winters of the northern part of the continent or the hot and dry summers of the southern parts. Spring- sown wheat in the northern United States and Canada receives rust-spore showers annually from rusted autumn-sown wheat in Mexico and Texas. In some years it comes by a succession of short jumps, with intervening stops for local multiplication, whereas in other years infection spreads suddenly from the south — for distances of a thousand or more kilometres when atmospheric pressure distribution produces suitable winds. Simi- larly, winter-wheat in the south becomes infected during the autumn by spore-showers from the north. Large-scale movement east and west across the North American Continent is relatively infrequent. Yellow rust of wheat {Puccinia ghwiariim) seems to be spread only much more locally than P. graminis or P. rubigo-vera — perhaps because its uredospores are more easily killed by exposure. A somewhat similar annual flow of airborne cereal-rust spores has been demonstrated in other parts of the world. For example, India has an annual flow from the hills to the plains: Mehta (1940, 1952) found that no local sources of rust infection survived the long hot summers of the Indian plains, yet rusts on wheat and barley caused heavy annual loss there. The explanation lies in the over-summering of rusts on cereal crops and self- sown plants at 2,000 or more metres in the hills, whence inoculum is carried by winds to start infection foci here and there on the plains in early winter. Upper-air currents and katabatic winds both play a part in the dissemination. The early dissemination of rust from inoculum coming from central Nepal and the Nilgiris and Palni Hills to the Indo-Gangetic plain, is the cause of annual devestating outbreaks. Russian work, mainly by L. F. Rusakov and A. A. Shitikova-Russa- kova (summarized by Chester, 1946), suggested transport of rust uredo- spores over hundreds of kilometres from the west across the Sea of Azov, and also from Manchuria up the Amur Valley in Eastern Siberia. On the other hand, the Irkutsk wheat-growing area west of Lake Baikal seems, for practical purposes, to be isolated from the rest of the world by deserts, mountains, and tundra. Puccinia graminis does not occur in the Irkutsk region, and P. rubigo-vera survives because it can form aecidia on Isopy- rumfumarioides, a common weed of arable land which serves as an alternate host. The wheat-growing areas of Australia and the Argentine also seem to be autonomous to the extent that airborne spores from outside do not affect either the annual rust epidemic cycles or the population of rust races present. Chester (1946, p. 164) concludes: Tn Australia and Argentina, as in the Lake Baikal region of Siberia, we evidently have regions which are so separated from other wheat areas by natural barriers, mountains, LONG-DISTANCE DISPERSAL oceans, deserts, and vast distances, that for all practical purposes they can be considered as totally isolated from the rest of the world, insofar as the introduction of wind-borne rust is concerned.' For the cereal rusts we thus have a picture of free interchange of uredospores over long distances, sometimes in the form of an annual immigration or even a return trip each year ; but isolation of several thou- sand kilometres limits this dispersal process. No other kind of microbe has been so fully traced in its atmospheric dispersal as have the cereal rusts. Rust uredospores are relatively large, and we would expect some smaller- spored organisms to be at least as well equipped for long-distance dis- persal — e.g. the coloured spores of agarics and myxomycetes. The width of the Atlantic Ocean, combined with the prevailing wind pattern in tropical latitudes, has formed a natural barrier to one of the maize rusts. From 1879 onwards (Cummins, 1941), Puccinia polysora has been collected on Zea mays and Tripsacian in the eastern and southern United States, in Central America, and in the Caribbean Islands; in 1940 Stakman found it on maize in Peru. In 1949 the fungus suddenly appeared in Africa, causing a widespread and severe disease on maize in Sierre Leone, spreading rapidly, and reaching most other parts of West Africa by 1 951; Congo, and East Africa from the Sudan to Tanganyika, were invaded in 1952; Southern Rhodesia, Portuguese South Africa, Mada- gascar, Mauritius, and Reunion all in 1953 ; and then the islands of Agalega and Rodriguez in 1955. Simultaneously another focus developed in Malaya in 1950, reaching Siam, the Phillipines, and Christmas Island in the Indian Ocean in 1956 (Wood & Lipscomb, 1956; Cammack, 1958). Evidently, until about 1949, the fungus lived quietly on tolerant American varieties of maize, isolated from vast areas of highly susceptible maize in Africa by the 5,600 km. of the Atlantic Ocean which, with its trade-winds, formed an impassable barrier. Cammack (1959) considered possible modes of immigration of Puc- cinia polysora into Africa and rejected wind transport in favour of intro- duction by aircraft on seed-corn or corn-on-the-cob, much of which was flown to West Africa from America during the last world war and post- war years. Once the parasite was established in Africa, the natural barriers there were evidently insufficient to stop its spread by wind over the whole continent, so that it reached Madagascar in only four years. Airborne pathogens present serious disease-control problems which are quite unlike those of such soil-bome diseases as potato wart. Other examples where a plant pathogen has spread rapidly after introduction by man into an isolated area include: potato late-blight {Phytophthora infest ans)^ which was introduced into Europe in the 1840's, and to Australia and South Africa in 1909; hollyhock rust [Puccinia malvacearum), which spread over western and central Europe between 1 869 and 1 874 (Gaumann, 1950, p. 141); and antirrhinum rust (P. rt;z//rr^/;//), which recently spread over New Zealand (Close, 1958). 189 THE MICROBIOLOGY OF THE ATMOSPHERE For the extreme limit of dispersal we have as yet no direct quantitative microbial evidence, but recent nuclear test explosions throw light on the problem (Libby, 1956; Stewart & Crooks, 1958). Clouds of radioactive dust, thrown up from bombs of less than one megaton, rise to 10-12 km., but tend to stay in the troposphere without penetrating to the stratosphere. In the troposphere the dust-cloud, consisting of particles mostly less than I ft in equivalent diameter, diffuses vertically and horizontally as it travels eastwards on the prevailing winds of temperate latitudes. The cloud circles the Earth every 4 or 5 weeks, but most of the particles are removed from the atmosphere in a month or two. A region of stable air over the tropics acts as a barrier to its spread between the northern and southern hemispheres. A thermonuclear bomb of the megaton range, exploded near the ground, puts a large proportion of its sub-microscopic dust into the stratosphere, where it behaves quite differently from the radioactive dust in the troposphere. The cloud spreads uniformly over all latitudes and is removed much more slowly, depletion to 50 per cent of the original quantity being variously estimated as taking from 5 to 10 years. Dust from successive tests thus accumulates in the stratosphere. As no natural mechanism has been suggested which would convey micro-organisms into the stratosphere on a similar scale, the results of thermonuclear tests are of less immediate interest to aerobiology than are tests of lower power or so-called 'normal bombs'. Possibly a hint at the ultimate horizontal dispersal-limit of micro- organisms in air is to be found in a study of the geographical distribution of species. In a world survey of the fungus, Schizophyllum commune^ the numerous incompatibilit}'' factors appeared to be randomly distributed (Raper et al., 1958), though the same may not be true oiCoprinus. On the whole, fungi are believed to show a wider natural geographical distribution than flowering plants. Europe and North America share more species of H^menomycetes (mushrooms, toadstools, and their allies) than of flowering plants. The species of fungi in the tropics and the south-temperate regions differ considerably from those of the north-temperate regions (Bisby, 1943). Many saprophytic species of fungi, especially soil moulds, tend to be cosmopolitan. How far wind has operated in their transport, and how far other means, especially man, are responsible for the world-distribution of these micro-organisms, is unknown; but the limits of the rust fungi give us some evidence. Viability Although, under ideal conditions, dispersal of airborne microbes is a limitless process, phenomena discussed in this chapter indicate that practical limits exist. One factor or another may reduce to negligible proportions the amount of inoculum transmitted between distant places. 190 LONG-DISTANCE DISPERSAL The source of supply at the place of origin may be too weak or, because of dilution and deposition, the spore-cloud may arrive at a concentration below the threshold for detection. Or again, when the spore-cloud arrives, few of the spores may be viable, or perhaps all may be dead. The dispersion equations given in Chapter XIII ignored the death- rate of spores in transit, and treated merely their diffusion as particles — regardless of whether they were living or dead. The justification for this is that the physical processes controlling dispersal likewise fail to dis- criminate between living and dead spores. Further, although a plant pathologist or plant breeder is concerned only with living cells, to an allergist a spore or pollen grain is equally significant whether alive or dead. None of the gradients measured so far contains any hint that loss of viability is affecting the dispersal process: it may very well do so, but the distances covered in any quantitative study are still too short, and our methods too coarse, to allow us to detect a decrease in viability along a gradient. Yet, over longer distances, loss of viability may well dominate the gradient. Some organisms may lose viability while travelling quite short distances in air; but this effect has apparently not yet been disen- tangled from effects of changes in concentration caused by diffusion and deposition. Uredospores of some cereal rusts as, for example, Piiccinia graminis and P. rubigo-vera, can evidently remain viable after travelling many hundreds of kilometres and reaching the top of the troposphere, though we suspect that others, such as those of P. glumarum and P. polysora, are less robust. Aecidiospores of rusts in general are believed to travel shorter distances in the viable state, and the basidiospores (sporidia) of the white- pine blister-rust {Cronartium ribicola) may well have their viable range restricted to a few hundred metres. Intensive study of this phenomenon in the open air is needed before we can know the quantitative significance of death for the dispersal gradient. PHYSIOLOGICAL STUDIES OF VIABILITY The immense literature on microbial survival and death-rates throws much light on factors influencing percentage viability in a population, under almost all imaginable conditions — except during suspension in air. Of changes in the viability of organisms during air-dispersal we have only circumstantial evidence, except for bacteria which are small enough to allow their longevity to be studied while suspended as aerosols in the ro- tating steel drum devised by Goldberg et al. (1958). Using this method, Webb (1959, 1959^) concluded that bacterial cells suspended in air die as a direct result of loss of bonded water from their protein. Loss of via- bility takes place in two stages : a rapid killing in the first second of time, followed by a slow death-rate which might be delayed by some bacterio- static substances whose presence might actually increase survival. 191 THE MICROBIOLOGY OF THE ATMOSPHERE The death-rate of a microbial pure culture normally proceeds expon- entially with time, the same fraction of surviving individuals dying in each successive equal time-interval — a process analogous to radioactive decay. But natural populations are often heterogeneous, and these mixtures of individuals or species with different propensities for hfe may deviate from the exponential die-away curve. Measurements of survival-time are often expressed as the time taken for most or all of the organisms to die under a particular set of conditions. Yarwood & Sylvester (1959) rightly point out that this limit is difficult to measure accurately, and that a more useful concept is the half-life of a population (as already used for decay of radioactivity). Apart from being easier to measure accurately than the end-point, the half-life is more logical than an arbitrarily selected value such as 90 per cent or 99 per cent death of the population, because it is the time at which all the individuals in a population which were alive at the start have an equal chance of being alive or dead. As an example, Yarwood & Sylvester give the half-life of basidiospores of Cronartium ribicola as 5 hours. It is not always easy to determine the status of a par- ticular spore or cell. If it can be grown it is clearly viable, but failure to grow may merely reflect failure to provide suitable conditions. Standard texts on microbial physiology deal with the effects of external conditions on the longevity of micro-organisms. Most of the experimental work is in the laboratory, with organisms in a liquid or at a solid/gas interface, and it is evident that the main factors of the aerial environment which affect survival (not always in the direction expected) are : humidity, temperature, and the visible or ultra-violet radiation. For instance, in laboratory tests, Whisler (1940) found the common airborne Sarcina lutea to be one hundred times more resistant to ultra-violet radiation than was the intestinal Escherichia coli. At first sight, conditions in the atmosphere might appear to be very unfavourable for the survival of isolated microbial cells (or even of resting spores) which, because of their minute size, have a high surface/volume ratio giving great exposure to external conditions. Endospores of bacteria are highly resistant to unfavourable environments; but this is not true of all plant spores, many of which are more delicate structures than the parent. Desiccation is a hazard already mentioned; it is greatest in day-time and in air-layers near the ground. At higher altitudes, and throughout the atmosphere at night, conditions are less favourable for evaporation, and spores may even be found germinating in clouds — a phenomenon occasionally reported for the uredospores of rust fungi. We are still not clear how to relate meteorological observations to conditions for viability. Evaporation in still air is a function of the absolute dr^Tiess of the air, but in moving air evaporation may be more nearly related to relative humidity. Possibly it is best to regard an airborne spore as 'still' in relation to the air in which it is suspended. A complicating factor is that, because of 192 LONG-DISTANCE DISPERSAL heat radiation, a spore is seldom at the same temperature as the air in which it is suspended; though, because of its small heat-capacity, the difference is not likely to be great. Improvements in the technique of freeze-drying show that damage to organisms is greatly affected by the temperature and speed at which they are desiccated. Repeated wetting and drying often lowers viability. As physiological experiments show, many of the regular components of the air-spora are resistant to desiccation. Less hardy organisms are probably better able to survive when they are high above the Earth's surface. Most micro-organisms will survive in a resting condition longer at the temperatures found in the upper air than they will at ground-level; temperatures in the atmosphere are preservative rather than lethal for most of the air-spora {see Meier, 1936^). Radiation is a much more serious hazard, and on this there is a rapidly increasing literature. The most quickly lethal radiations in the atmosphere are in the ultra-violet region, and these are largely absorbed by the air before they reach the ground. Ascent to the upper air therefore brings the risk of greatly increased dosage of ultra-violet radiation — except in the shelter of clouds. Pigmented spores, and bacteria carried on larger rafts, may also be effectively screened from radiation. The interactions of humidity, temperature, and radiation are not well known, and it may be that low temperature and desiccation partly protect a spore against radiation damage. In addition, when a spore returns to ground-level, the recently studied phenomenon of photo-reactivation by visible light of organisms which have been exposed to ultra-violet, may perhaps reduce the damaging effects of radiation received at high altitudes. Photo-reacti- vation is defined by Jagger (1958) as: 'the reversal with near-ultraviolet or visible light of ultraviolet radiation damage to a biological structure.' Visible light reverses lethal effects of ultra-violet radiation in many microbes — including bacteria and actinomycetes, fungi and yeasts, and also protozoa — but the survivors of high-altitude passage may be expected to show an increased mutation rate on their return to Earth. The suitability of the atmosphere for the survival of 'aerial-plankton' is summed up by Gislen (1948) as follows: 'While the lower cloudy air- strata — let us say under 3,000 to 4,000 metres — form a suitable medium for the transport of micro-organisms, the higher layers are very inhospit- able to them, not so much because of the low temperature, drought and barometric pressure, as because of destructive radiation.' N 193 XV AEROBIOLOGY The literature of aerobiology is scattered, but extensive reference lists are given in the following works: Cunningham (1873), Heald (1913), Gardner (191 8), Sernander (1927), Stepanov (1935), Rempe (1937), Pettersson (1940), Craigie (1941), Moulton (1942), duBuy et al. (1945), Gregory (1945), Stakman & Christensen (1946), Wolfenbarger (1946, 1959), Jacobs (1951), Maunsell (1954), Wells (1955), Werff (1958), and Hirst (1959). It now remains to review the conclusions of each of the preceding chapters in the light of the whole, to consider their implications, and to attempt a synthesis. The Phenomena Airborne microbes, whether carried singly, in groups, or on 'rafts', are heavier than air. In still air they fall under the influence of gravity, with constant terminal velocities ranging from 0-05 to 150 cm. per sec, according to their size and density. This falling would lead to their sedi- mentation out of the air if other forces did not oppose gravity. Two atmos- pheric processes tend to prevent sedimentation through the quiet bound- ary layer : turbulent diffusion in wind carries the spore-cloud horizontally, at the same time diffusing it both horizontally and vertically; and thermal convection can carry a spore-cloud to great heights in the troposphere. Modes by which spores cross the boundary layer of air near the Earth's surface and reach the turbulent wind layer, are therefore of prime importance in the dispersal of microbes. Energy for 'take-off' may be supplied by the organism itself or may come from external sources supported by a wide variety of mechanisms, such as wind or rain- splash, and factors controlling the take-off mechanism also control the occurrence of spores in the air. The more an organism is specialized towards one dispersal route, the more unfit it becomes for dispersal by another route. For many purposes, knowledge of the 'take-off' mechanism is important. Rain-splash dispersal results in a local scatter because the larger splash droplets, which are less easily carried by the wind, pick up more spores than do smaller droplets. Frictional turbulence will diffuse the spore-cloud to the top of the 194 AEROBIOLOGY dust horizon — convection will take it higher. The concentration of spore- clouds approximates to the expected logarithmic decrease with height, but equilibrium is never reached in the atmosphere. Often there is a zone of increased concentration at an altitude of two or three thousand metres. Dilution of the spore-cloud as it travels horizontally downwind is the result of eddy diffusion. Of the various meteorologists and physicists who have formulated eddy diffusion, O. G. Sutton has provided the most use- ful method. Experimental values obtained in microbiological work for the parameters in Sutton's diffusion equations are useful in predicting spore concentrations at various distances from the source. We await still better methods; but they may be long in coming, for the problem appears to be particularly intractable. Concentration of the spore-cloud decreases not only by diffusion but also because particles are lost by various deposition processes — of which impaction, turbulent deposition, and sedimentation under the influence of gravity, are the most important. Wind speed, the area and orientation of the surface, and the size of the particle, all have great effects on im- paction efficiency, both in the wind-tunnel and out-of-doors. Deposition from the atmosphere under natural conditions outdoors has been studied experimentally. Under uniform conditions, particle size has an important effect on the amount deposited from a spore-cloud of given concentration at ground-level. When liberation is near ground-level, loss by deposition is great. We cannot yet choose between two theories of deposition to the ground, one of which depends on the wind-speed and the other on the distance traversed by the cloud. Near the source, unex- plained high deposition values have been observed. Elevating the point of liberation reduces the amount of deposition near the source. Topo- graphical features which affect wind-speed also affect the number of spores deposited, and may play a part in development of a plant disease epidemic — complicated by the fact that they may simultaneously modify the ecoclimate in a direction favouring or inhibiting infection. Estimation of the microbial content of the air is particularly difficult because, although microscopic, the particles are often large enough to demand attention to the aerod}Tiamic design of the sampling equipment. Choice of the correct sampling equipment must be determined by the range of organisms to be sampled. Throughout this book emphasis has been placed on methods, because methods commonly determine results. Single bacterial cells in aerosols are small enough to be handled in the manner of a gas, without regard to their inertia; but larger organisms (and bacteria on 'rafts') impact on surfaces, stick on corners, slip out of streamlines, and settle under the influence of gravity. These aerodynamic effects must be allowed for in the design of apparatus for reasonably accurate sampling. The basic study of the air-spora must be by visual methods under the microscope. Crude as visual identification is, it is based on the only common 195 THE MICROBIOLOGY OF THE ATMOSPHERE property shared by all airborne microbes — visibility — and no other method can reveal the whole range of the air-spora and disclose what numbers and kinds are in the air awaiting identification. More precise methods can then be applied to the study of smaller groups. The air-spora is very imperfectly explored, but sampling has already shown that bacteria, algae, yeasts, spores of fungi, mosses and ferns, pollens, and protozoa, commonly occur in the air. The air-spora near the ground is extremely variable, both from time to time and from place to place. It changes with season, and with weather — the concentration of constituent types often changing several thousand-fold in the course of an hour or two. It changes regularly in composition and concentration throughout the day and night. Visual methods indicate that fungus spores are normally in the majority over other components of the air-spora — probably outnumbering bacteria. However, we lack methods for the complete enumeration of the bacteria in outdoor air, and improved tech- niques may reveal the presence of far more bacteria than we can recognize at present. We now know that basidiospores (ballistospores) form an important part of the air-spora, often carrying electrical charges, and probably out- numbering even Cladosporiiwi (which is the dominant mould almost everywhere). The tardy recognition of the abundance of basidiospores is partly explained by their inefficient collection by standard samphng methods, and partly by the unfamiliarity of many microbiologists with the spores of the higher fungi (cf. Plate 6). Damp night air has its character- istic spora, but at dawn the night-spora disappears suddenly; where it goes we do not know. Rain removes the dry-air spora and substitutes a different one. Vegetation is the main source of the air-spora, but most bacteria come from blown soil or splashed water. The source of the yeasts, which are sometimes recorded in large numbers, is still obscure — unless they prove to be Sporobolomycetes. Air near the ground at times contains from tens of thousands to hundreds of thousands of micro-organisms per cubic metre. The air-spora near the ground is commonly dominated by con- tributions from local and intermediate sources; but local effects are smoothed out at high altitudes, over the oceans, and in polar regions. Spore concentration over the land usually decreases with altitude, though not always regularly. Often a large proportion of the air-spora must occur at altitudes higher than lo metres. Far out at sea, a few metres above sea-level, the microbial content of the air is usually small ; but, at an altitude of a few thousand metres over the ocean, the air often contains a few bacteria and from tens to hundreds of fungus spores per cubic metre. Current information suggests that, over the oceans, concentrations are greater in the upper air than near sea-level. Clearing the lowest zone of the atmosphere from microbes is most evident over the ocean. For above land the spore-cloud near the ground is not only fed from above by 196 AEROBIOLOGY diffusion and, to a minor extent, by gravity, but it may also be fed from below, from new sources of soil and vegetation. Even over the Arctic in winter, the air is only relatively sterile; samples of a few cubic metres may or may not contain viable spores. Wind transport of microbes is a process that is evidently going on continuously on a w^orld-wide scale. Many unsolved problems remain. Is there a biotic zone at the height of a few^ thousand metres ? Or is the lower air cleared by rain, with microbes re-concentrating transiently at the base of a cloud ? Is there a true aerial plankton in the sense of a population permanently living and reproducing at great heights, as suggested by R. C. McLean (1935, 1943) ? This would seem improbable — unless a microbial population is permanently balanced over tropical regions by rising air-currents and descending cloud droplets. Do air masses retain a characteristic polar or tropical spora, or do they rapidly receive and give up the spora of the land over w hich they pass ? Sooner or later, if they have not been already deposited by other means, airborne microbes are removed from the atmosphere by collection in rain, snow\ or hail. If the carrying droplet later evaporates below the cloud base, there could possibly be an actual local increase in concentra- tion high above ground. Rain, hail, and snow bring down a large microbial flora and are factors making for non-uniform deposition, in contrast with the more uniform diffusion which results from turbulence. Spores deposited on the ground, or on vegetation near to a source, show a pronounced gradient (following the concentration gradient of the wind-blown spore-cloud). This can be estimated by making assumptions based on observed data, and can be used in predicting the danger of contamination by foreign pollens, plant pathogens, and so on. In practice the ideal gradient is often modified by topography — sometimes a decreased wind velocity decreases spore deposition. In the past it may have been too easily assumed that the dispersion of an organism around its origin is like a normal frequency distribution. This may be true when dispersion is due to motile surface animals, actively flying insects, and possibly even rain-splash. But wind-dispersal is not 'normal' (in a statistical sense) around the point of origin; by contrast, it is a hollow curve. This throws light on a paradoxical situation: in the w ind-dispersal of microbes liberated near ground-level, the gradient near the source is very steep, and relatively short isolation-distances give good protection. Pollens and plant pathogens with large dry spores liberated at ground-level may have 90 per cent of their spores deposited within about 100 metres — or perhaps more than 90 per cent if allowance is made for the unexplained large values of p observed within a metre or two of a ground source. Yet spores can travel immense distances, and spores found over the ocean clearly represent the tails of the distributions of all the sources on the up-wind continent. The equation for Q^^. has the curious property that the farther a spore has travelled, and the longer it has survived deposition, the farther it is likely to travel. 197 THE MICROBIOLOGY OF THE ATMOSPHERE Although most spores are deposited near their source, some are readily transported to great distances. Transport over long distances plays a regular role in some crop disease epidemics, and presumably in the movement of many other organisms. Mountain ranges, oceans, and deserts, may all be effective barriers to dispersal. Although conditions in the upper air, especially in cloud, may not be unfavourable for survival, loss of viability rather than failure of the transport mechanism limits the colonization range of many organisms. Enhanced variability may be expected among microbes that survive exposure to ultra-violet radiation. Because of the overwhelming importance of unsuspected near-by sources, it is often difficult to be sure that an organism observed has come from a great distance. Implications of Aerobiology BIOLOGICAL warfare It seems that microbiological weapons have not been used on any large scale by man against man. The example of myxomatosis in rabbits should convince sceptics of what might happen when all conditions are suitable for epidemic spread. The topic is shrouded in official secrecy, but the little information already released suggests that, if deliberate dissemination of pathogens (or toxins) were ever attempted, contamination of the air might be one of the dangers to be anticipated. Rosebury et al. (1947), from their comprehensive analysis of the principles of bacterial warfare, consider that the airborne group of pathogens contains the most important infective agents for war use {see also Rosebury, 1947, 1949)- The published studies on air-sampling equipment and epidemiology that come from official defence laboratories are small compensation for this threat by man to his own health and agriculture. IS