The micro-organisms of the soil

In 1886 Adametz,[1] investigating the biochemical changes occurring in soils, isolated several species of fungi. It was, however, only with the work of Oudemans and Koning,[17] in 1902 when forty-five species were isolated and described, the majority as new to science, that the real study of the fungus flora of the soil commenced. There is now no doubt that fungi form a large and very important section of the permanent soil population, and certain forms are found only in the soil. Indeed, Takahashi[22] has reversed the earlier ideas by suggesting that fungus spores in the air are derived from soil forms. The majority of investigations on this subject fall, perhaps, into one or more of three classes: (a) purely systematic studies such as those of Oudemans and Koning,[17] Dale,[5] Jensen,[9] Waksman,[25a] Hagem,[8c] Lendner,[12] and others, which consist in the isolation and identification of species from various soils: (b) physiological researches, such as those of Hagem[8c] on the Mucorineæ of Norway, or the many investigations on the biochemical changes in soils produced by fungi, such as those of Muntz and Coudon,[15] McLean and Wilson,[15] Kopeloff,[11] Goddard,[7] McBeth and Scales,[14] and others: (c) quantitative studies, such as those of Remy,[20] Fischer,[6] Ramann,[18] Waksman,[25c] and Takahashi,[22] which involve numerical estimates of the fungus flora in soils.

Qualitative Study.

With very rare exceptions soil fungi cannot be examined in situ, and the necessary basis of any qualitative research is the isolation of the organisms in pure culture. Most soil forms belong to the Fungi imperfecti, and often show considerable plasticity on artificial media. This makes it very difficult to determine them by comparison with type herbarium specimens or published morphological diagnoses. In consequence many soil fungi have not infrequently been given new specific names, as humicola, terricola, and so forth, which is very unsatisfactory, and means that the determinations have little significance.

Furthermore, most artificial media are slight variations on a few common and simple themes, and are very selective, permitting the growth of a moiety only of the fungi present. In addition, many fungi grow so slowly that they are overwhelmed by the more rapidly germinating or spreading forms, or on the other hand, they may be eliminated by the metabolic products of different adjacent colonies. The extremely selective nature of the technique commonly used is shown if one tabulates systematically all the fungi which have been recorded or described in soil investigations. Of Phycomycetes there are fifty-six species of eleven genera; of Ascomycetes twelve species of eight genera; and of Fungi imperfecti, including Actinomycetes but not sterile Mycelia, 197 species of sixty-two genera. Rusts and Smuts one might not expect, but that of the multitudes of Basidiomycetes growing in wood and meadow not one should have been recorded is indeed startling. It was at first thought that many imperfect fungi might be conidial stages of Basidiomycetes, but much search among forms isolated at Rothamsted has, up to the present, failed to reveal clamp connections in the hyphæ.

Since various species of soil fungi have different optimum temperature, humidity and other conditions[3] one would not expect to find an even geographic distribution. Very little is yet known of this aspect, but Rhizopus nigricans, Mucor racemosus, Zygorrhynchus vuilleminii, Aspergillus niger, Trichoderma koningi, Cladosporium herbarum, and many species of Aspergillus, Penicillium, Fusarium, Alternaria, and Cephalosporium have been commonly found throughout North America and Europe wherever soils have been examined. Species of Aspergillus, however, would appear to be more common in the soils of south temperate regions and species of Penicillium, Mucor, Trichoderma, and Fusarium more abundant in northern soils.

It is well known that in many plant and animal communities there occurs a definite rhythm, various species following each other in a regular sequence as dominants in the population. Although it is not yet possible to make any definite statement there would seem indications that this may also be true of the soil fungi.

Much work has been done on the distribution of species at different depths in the soil, but the results are still confusing. Thus, examining eighteen species, Goddard[7] found no difference in relative distribution down to 5¹⁄₂ inches. Werkenthin[26] found identical species from 1-4 inches, and then an absence of fungi from 5-7 inches, which latter was the greatest depth he examined. Waksman[25] found little difference in the first six inches, but very few species below 8 inches except Zygorrhynchus vuilleminii, which extended down to 30 inches and was often the only species occurring below 12 inches. Taylor[23] has reported species of Fusarium at practically every depth to 24 inches. Rathbun[19] found Aspergillus niger, Rhizopus nigricans, and species of Fusarium and Mucor down to 34 inches, and Oospora lactis, Trichoderma koningi, Zygorrhynchus vuilleminii and species of Penicillium, Spicaria and Saccharomyces as deep as 44 inches. Eleven species were isolated from the alimentary canal of grubs and worms, and Rathbun concluded that soil fungi may be spread by these organisms.

On an unmanured grass plot at Rothamsted twenty species were isolated from a depth of 1 inch, nineteen from 6 inches, and eleven from 12 inches, whereas on the unmanured plot of Broadbalk wheat field twenty-six species were obtained from 1 inch, seven from 6 inches, and five from 12 inches. There appeared to be no conspicuous differences between the floras of the two plots save that in the Broadbalk plot there were fewer Mucorales, and Zygorrhynchus mœlleri and Absidia cylindrospora were absent. In the grass plot samples about one-half the forms occurring at the lower levels were isolated also from the upper levels, but in the Broadbalk sample the five forms isolated from 12 inches, and five out of seven of those at 6 inches occurred only at those levels, i.e. each of the three levels appeared to have a specific flora. The difference in depth distribution in these two cases may be due to the fact that in the Broadbalk plot the stiff clay subsoil occurs at 5-7 inches, whereas in the grass plot the depth of soil is greater than 12 inches. Much further work needs to be done on this aspect before any definite conclusion can be reached.

Much scattered information is available concerning the effect of soil type, manuring, treatment, cropping, and so forth upon the fungus content, but no clear issue as yet emerges from the results. Hagem[8] found that cultivated soils vary greatly from forest soils in the species of Mucor present, and that certain species seem to be associated in similar environments. Thus in pinewoods Mucor ramannianus is usually found, together with M. strictus, M. flavus, and M. sylvaticus, and with this “M. Ramannianus Society,” M. racemosus, M. hiemalis, and Absidia orchidis, are frequently associated. The differences found by Hagem between the species of Mucor from forest and cultivated land could not, however, be confirmed by Werkenthin.[26]

Dale,[5] examining sandy, chalky, peaty and black earth soils, found specific differences, although many of the species were common to all. A soil which had been manured continuously for thirty-eight years with ammonium sulphate alone, contained twenty-two species, whereas the same soil with the addition of lime only had thirteen species. Both Goddard[7] and Werkenthin,[26] in their investigations, found a constant and characteristic fungus flora regardless of soil type, tillage, or manuring. Waksman’s[25] studies of forest soils showed few species of Mucor but many of Penicillium and Trichoderma[2]; orchard soil contained no species of Trichoderma, very few of Penicillium, but a large number of species of Mucor; species of Trichoderma were common in acid soils, whilst cultivated garden soil contained all forms. The examination of very differently manured plots on the Broadbalk wheat field at Rothamsted has not shown any striking differences in the fungus flora, all the more important groups of species being represented in every plot, but significant minor differences are present. Thus, plot 13, manured with double ammonium salts, superphosphate and sulphate of potash, is especially rich in “species” of Trichoderma, whereas the unmanured plot contains large numbers of species of green Penicillium, Trichoderma, and a species of Botrytis (pyramidalis?).

The effect of the crop upon the fungus flora is seen in cases where the same crop is grown year after year as in certain flax areas, where species of Fusarium accumulate in the soil and tend to produce “flax sickness.”[13]

Quantitative Study.

As it is not possible to count the soil fungi in situ, any estimation of the numbers present in a soil must be arrived at by indirect means. The method adopted is to make as fine a suspension as possible of a known quantity of soil sample in a known amount of water, dilute this to ¹⁄₅₀₀₀, ¹⁄₁₀₀₀₀, and so forth by regular gradations, incubate cubic centimetres of the final dilution on artificial media in petri dishes, and count the colonies of fungi developing in each plate. Using the average figures from a series of duplicate plates, the number of “individual” fungi in a gram of the original soil sample may then be calculated. The very few students who have made quantitative estimations have obtained very unsatisfactory results. In bacterial or protozoal estimations, the shaking of the soil suspension separates the unicellular individuals, so that in the final platings each individual from the soil theoretically gives rise to one colony on the medium. In the case of fungi, the organisms may be in the form of unicellular or multicellular spores or larger or smaller masses of unicellular or multicellular mycelium differing for each particular species or phase of development within the single species. The organisms may be sterile in the soil or form fruiting bodies, consisting of few or myriads of locally or widely distributed spores. In the process of shaking the soil-suspension fungi of different organisation or of differing developmental stages may be broken up and moieties fragmented in totally different ways or to very different degrees. With protozoa and bacteria the relation of soil individual to plate colony is direct; with fungi we do not know what is the soil “individual” nor whether it is the same for different fungi; nor can we yet profitably discuss any significant numerical relationship of plate colonies to soil organisms. Thus Conn[4] has pointed out that the plate count of a fungus indicates only the ability to produce reproductive bodies and found that the spores of one colony of Aspergillus, if distributed evenly through a kilogram of soil, could produce the average plate counts obtained by Waksman. Abundant vegetative growth may, in some species, reduce or inhibit spore formation, so that of two species the one giving a lower count might really be much the more important and plentiful in the soil. Further, the colonies developing in the final plates represent only a selected few of the fungi present in the soil sample, the Basidiomycetes, and no doubt many other forms, being absent. In addition, different media differ among themselves in the average number of colonies developing on the plates, each medium giving, as it were, its own point of view. Thus, in one experiment carried out at Rothamsted by Miss Jewson, using the same soil suspension, twenty plates of Coon’s Agar gave 357 colonies, of Cook’s Agar 246, of Czapek’s Agar 215, and of Prune Agar 366. Thus if one only used Coon’s Agar and Prune Agar one would obtain a total of 723 colonies, whereas the same suspension on Cook’s Agar and Czapek’s Agar would give only 461, and the calculated numbers of fungi per gram of soil would be totally different. Further, if a single medium be taken, it is found that slight alterations in the degree of acidity may make very considerable differences in the final numbers. Thus Coon’s Agar acidified to a hydrogen ion concentration of 5·0 gave as the results of four series the following average numbers of colonies per plate, 17, 23·75, 18, 23. When, however, the medium was acidified to a PH of 4·0 to 4·3, corresponding averages from three series were 38, 46·3, and 44·8; i.e. the final estimations of numbers of fungi in the soil was about twice as great. Again, the degree of dilution of soil suspension used in plating may also be a very serious factor. Thus, if a series of dilutions be made of ¹⁄_80,000, ¹⁄_40,000, ¹⁄_20,000, ¹⁄_10,000, ¹⁄_5,000 and ¹⁄_2,500, the average plate numbers should be in the proportions of 1, 2, 4, 8, 16, and 32 respectively. In an actual experiment, the following average plate numbers were obtained, 15·4, 32·8, 59·1, 104·0, 150, 224·5, which show a very decided reduction in the higher numbers. If, however, dilutions of a suspension of spores of a single species be made, this reduction does not occur.

These are but three of the very numerous factors involved in the technique of quantitative estimation, and every single factor may be the source of errors of similar magnitude, minute fluctuations in the operations leading to the final platings having very considerable effect upon the numbers of colonies that develop.

By critically evaluating each particular factor in the method, and making statistical correction, it has, however, been found possible to obtain series of duplicate plates comparing very favourably and thus to extract certain figures which, whilst not possessing any final value, have yet a certain general and comparative worth. Thus, 20·0, 18·2, and 16·8 were obtained as the averages of six plates each, of a soil suspension divided into three parts, and the individual plate numbers in all three series were within the range of normal distribution. The meaning of these numerical estimates in relation to fungi per gram of soil sample is, however, entirely hypothetical, and to have value quantitative comparison should only be made between single species or groups of species closely related physiologically, and where the technique is standardised.

No comparative estimations have been made of the number of fungi in the soils of different regions. There are, however, certain figures which show that decided seasonal differences exist. Thus, correcting and averaging certain of Waksman’s results[25] the following numbers of fungi per gram of soil at 4 inches deep are obtained; September, 768,000; October, 522,000; November, 310,000; January, 182,000. At Rothamsted results have been obtained which would appear to mark a clear seasonal rhythm, corresponding in the time of its maxima in Autumn and Spring with the periodicities known for many other ecological communities (Fig. 19).

The numbers of fungi at various depths in the soil show very clearly marked differences. The distribution in the top 4-6 inches depending probably upon the depth of soil, is more or less equal, but there is a very rapid falling off in numbers, especially between 5-9 inches, until at 20-30 inches fungi are either very few in number or absent. Thus Takahashi[22] found 590,000 fungi per gram at a depth of 2 cms. and only 160,000 at 8 cms.

TABLE XIII.—INFLUENCE OF SOIL TREATMENT UPON THE NUMBERS OF FUNGI AS DETERMINED BY THE PLATE METHOD—(AFTER WAKSMAN).

The type of soil and its treatment exercise a great influence over the number of fungi present. Fischer[6] found that farmyard manure increased the number of fungi in uncultivated “Hochmoor,” cultivated “Grunlandmoor,” and a clay soil by two, three, and five times respectively. Waksman’s results[25] indicate that the more fertile soils contain more fungi, both in number and species, than the less fertile ones, and if one averages his results, the following figures are obtained: garden soil, 525,000 per gram; orchard soil, 250,000; meadow soil, 750,000; and forest soil, 151,000. Recently Waksman[25e] has found that manure and acid fertilisers increase the numbers of fungi in the soil, whereas the addition of lime decreases them (Table XIII.).

Jones and Murdock[10] examined surface and sub-surface samples of forty-six soils representing seventeen soil types in eastern Ontario. Molds were fairly uniform in numbers in all soils except a sandy clay loam and sandy clay shale, in which they were absent.

It has also frequently been pointed out that acid and water-logged soils are richer in fungus content than normal agricultural soils. On the other hand, Brown and Halversen[2] found, examining six plots receiving different treatment and studied through a complete year, that the numbers of fungi were unaffected by moisture, temperature, or soil treatment. Against this, however, must be set the work of Coleman[3] who studied the activities of fungi in sterile soils and found such factors as temperature, aeration and food supply to exercise a deciding control.

Investigations at Rothamsted show that Broadbalk plot 13, receiving double ammonium salts, superphosphate and sulphate of potash and yielding 31 bushels per acre, and plot 2, receiving farmyard manure and yielding 35·2 bushels, contain approximately equal numbers of fungi. This figure is about half as high again as that for plot 3, which is unmanured and yields 12·6 bushels, plot 10, with double ammonium salts alone and yielding 20 bushels, and plot 11, with double ammonium salts and superphosphate and yielding 22·9 bushels per acre. A primary factor, however, in all considerations such as these is the equality of distribution of fungi laterally in any particular soil. There are probably few soils so homogeneous as the Broadbalk plots at Rothamsted, and on plot 2 (farmyard manure since 1852) samples taken from the lower and upper ends and the middle region gave average numbers of colonies per plate of 24, 23, and 25 respectively. On the other hand, soil samples taken only a few yards apart in the middle region of the plot gave average plate counts of 33·7 and 56·8.

Conclusion.

Surveying generally the field covered in this chapter, one can only be impressed with the fragmentary character of our knowledge and with the fact that, owing to the selective nature of the technique, the data we possess, if assumed to be representative, give an entirely partial and erroneous picture of the soil fungi. From the qualitative aspect, the chief impediment is the impossibility of obtaining reliable specific determinations of very many of the soil fungi. Lists of doubtfully-named forms from particular soils or geographic regions are useless or a positive evil, and there is imperative need for the systematising of selected genera by physiological criteria, such as has been partially done for Penicillium, Fusarium, and Aspergillus. Furthermore, until a standardised and non-selective technique has been devised, or a number of standardised selective methods for particular groups, comparative investigations into specific distribution can give little of value. This latter criticism is also very applicable if regard be paid to the quantitative aspect of soil work, for progress here largely depends upon the elaboration of a standardised fractionation technique. Every single factor in these methods needs exact analysis, for each gives opportunity for great error, and each error is magnified many thousand times in the final results. Much has been done in this direction at Rothamsted, but more remains to do. Finally, working with single species in sterilised soil under standardised conditions, there is fundamental work to be done on the relation of plate colony to soil “individual.”