classification, however, is open to considerable criticism, as pointed out by Lehmann and Neumann,45 because it separates species that are known to be closely related, such as B. subtilis and the anthrax organism, or Bact. aerogenes and the colon organism. It is generally considered that the type of the genus Bacillus should be B. subtilis, but whether the genus should contain all peritrichic rods regardless of spore-production or all spore-producers regardless of motility is a matter of dispute.

In the present work, fortunately, it has not been necessary to take either side of this dispute, as the only peritrichic organisms found to be typical of soil are also spore-formers and so belong in the genus Bacillus without question. Lehmann and Neumann, who adopt the classification depending upon spore-production, disregard Migula's genus Pseudomonas; but there seems to be no reason for doing this even if Bacillus is to be characterized by the presence of spores. Rods with spores and polar flagella are rare, if they occur at all, so the genus Pseudomonas can be recognized without regard to the matter of spore-production. It has been distinguished from Bacterium in the present work; but because of the difficulty of determining the presence of flagella, their absence has never been considered of sufficient diagnostic value to distinguish between two cultures that agreed in everything else.

Presence of spores (represented by the first figure in the group number). This has proved a point of considerable diagnostic value. Migula, it is true, claimed that the ability to produce spores is easily lost, and that failure to produce them under ordinary laboratory conditions does not indicate that they are never produced. For this reason Migula lays less stress upon this characteristic than upon flagellation. Altho there is considerable truth in this contention of Migula's, it is coming to be felt that the production of spores is more nearly correlated with other important characteristics than is motility.

It has ordinarily proved fairly easy to determine spore-production. An ordinary dried microscopic preparation of a culture two days old, stained, without heating, by means of either fuchsin or methylene blue, is usually sufficient to show whether spores are present in any particular culture. Occasionally some large rod-shaped organism has been found that does not show spore-production when examined in this way. Some of these cultures, after continued observation for weeks or months, have at last been observed to produce spores and have been identified as B. simplex-which is never a vigorous spore-producer. It is believed that the others are also closely related to B. simplex, as non-spore-formers are seldom large rods (i. e. over 0.8 microns in diameter). Except in such a case as this, the

5 Lehmann, K. B., and Neumann, R. O. Atlas und Grundriss der Bakteriologie. 1912. Munich. See part II, p. 155.

presence or absence of spores in a microscopic preparation has always proved of great diagnostic value.

The shape of spores, when they are present, is of equally great diagnostic value. This was pointed out by Meyer and by Gottheil46 and has been used by Chester47 and by Ford48 as well as by other bacteriologists who have recently worked on the spore-forming organisms. Altho the spores present in any preparation vary greatly in shape and size, it is usually possible to recognize some one particular shape and size as typical of the organism in question; and for this reason much stress is laid on that point in the following paper on the spore-formers of soil.

Growth in the absence of oxygen (second figure of the group number). This is a feature of great diagnostic value; but it is not yet known what method to use to obtain the most consistent results. No organism is known to grow in the absence of free oxygen unless some ready source of combined oxygen, such as sugar or nitrate, is at hand. Some organisms which grow anaerobically under certain conditions may possibly be unable to obtain their oxygen from any of the compounds ordinarily used in bacteriological culture. The method used in this work has been to notice the presence or absence of closed-arm growth in fermentation tubes of various sugars. There are two serious objections to this method: First, oxygen may be dissolved in the medium and furnish aerobic conditions even in the closed arm; and secondly, when gas is produced, the bubbles as they form may carry the bacteria up with them and cause turbidity in the closed arm although no actual growth occurs there. The first objection was not considered very important, however, because it was largely obviated in the present work by the use of freshly sterilized media; while the second objection applies only to gasproducing organisms, none of which were studied in this work. For these reasons the fermentation tube method proved fairly satisfactory.

Liquefaction of gelatin (third figure of the group number). This has been determined in stab culture, keeping the tubes under observation for six weeks. It is undoubtedly a feature of considerable diagnostic value, altho there is some possibility that certain kinds of bacteria vary in their ability to liquefy gelatin. Fermentation of sugars (first three places after the decimal point in the group number). This is a matter that has given very inconsistent results. It is very seldom possible to obtain exactly the same fermentation reactions for the same culture upon repetition of the tests. This may be due to variation in the acid-producing power of the organisms, or possibly to imperfections in technic. The irregularity in results may be due in part to the fact that many 46 Loc. cit., footnotes 6 and 7.

47 Loc. cit., footnote 5.

48 Loc. cit., footnote 10.

of the organisms studied do not grow well in broth. For this reason, the technic was sometimes modified by using litmus-agar containing the sugar in question. Slant cultures in such agar often gave better results than broth cultures. In studying the spore-formers, the fermentation reactions were found to be of some value if taken in conjunction with other tests.

Nitrate reduction (fourth figure after decimal point in group number). This test has proved of considerable value in the separation of species, altho sometimes the results appear to be erratic. The erratic results, however, all seem to occur in the case of organisms that grow poorly in nitrate broth. In interpreting the test, therefore, it must be borne in mind that no growth or poor growth in nitrate broth does not necessarily mean that the organism in question is unable to reduce nitrates. This point has already been emphasized by Breed.49

Chromogenesis (fifth figure after decimal point). Chromogenesis has proved to be quite constant in the case of the organisms investigated. It is a particularly valuable aid in the classification of Actinomycetes; for altho these organisms are usually gray to brown on ordinary media, the colors produced upon special media are both striking and varied. Excluding the Actinomycetes, there are only two common types of chromogenesis found among soil bacteria. One is the production of fluorescence, the other of a lemon yellow pigment. A few strains have been found that produce an orange pigment. Fluorescence, when present, is of distinct diagnostic value; but the power of producing it is so easily lost that its absence is of no great significance. The production of yellow or orange pigment is constant so far as has been observed.

Diastatic action on potato starch (sixth figure after the decimal point). This may be a constant characteristic, if a satisfactory technic for determining it were known. This determination was added to the group number by the committee that revised the card in 1907; but no official method for making the determination was given. As a result, this figure of the group number seems to be quite generally ignored. In the early part of the present work, some attempt was made to include this determination. The cultures were inoculated onto potato plugs and incubated for two weeks; then the material was crushed in water and a weak solution of iodine added, drop by drop, recording the diastatic action as strong" if no color change occurred, as weak if a reddish color appeared, or as absent if the color was blue. This method soon proved so obviously unsatisfactory that it was discontinued, and the determination has not been used in this classification; hence the question-mark in the group numbers given in the following

papers.

49 Breed, R. S. The standard method of determining nitrate reduction. Sci., N. S., 41:661. 1914.

Fermentation of glycerin (last figure of group number). This determination has proved of no value whatever. Results are more erratic than in the case of fermentation of sugars.

Production of indol. This test and the characteristics discussed in the two following paragraphs are not necessary in determining the group number; but they are so often included in descriptions of bacteria that a discussion of them here is not out of place.

In the first part of the present work, the determination of indolproduction was always made; but it proved so irregular that it was soon discontinued. The method of determining it was to inoculate peptone solution and incubate ten days. It is quite possible that the irregular results may have been due to irregularities in the composition of the peptone itself.

Action on milk. Cultures in plain milk and in litmus milk were made in the first part of the work. Like the indol test, this soon proved to give too inconsistent results to be of much service in classification.

Form of growth. To a considerable extent this is constant and is in some cases a point of considerable diagnostic value. The long chains formed by B. cereus and B. mycoides cause these organisms to grow in a very different manner from others that break up quickly into isolated rods. The parallel chains of B. mycoides, with their false branching, give rise to its characteristic rhizoid growth. These peculiarities of growth impress themselves to a certain extent upon colonies, liquid culture, and growth in stab and on streak. But the limits of variation of the different organisms must be known before this characteristic can be used for diagnosis.

The form of growth produced by Actinomycetes is strikingly constant. Certain characteristics such as consistency, lustre, wrinkling, and presence of aerial hyphae vary so little on the same medium as to be of great diagnostic value.

Reasons for the inconsistent results. It will be readily seen that such lack of consistency as has just been shown in the results of the various determinations lessens their value for diagnostic purposes. It is possible, even tho not probable, that the inconsistency may be due to actual variations in the physiology of the organisms studied. This possibility is taken into account in compiling the Standard Methods of the American Public Health Association, in which it is recommended that all bacteria be invigorated before study, so that they might all be in an equally vigorous condition. The method of invigoration called for is to inoculate into three successive daily transfers of broth, incubated at 37° C. This procedure was out of the question for nearly all the soil bacteria, because they grow poorly or not at all in broth, and the temperature recommended is an unfavorable one. In an attempt to improve the technic for soil organisms, they were invigorated on agar "slants" at 25° C. Under such conditions all the organisms studied grow vigorously;

but the change in technic did not increase the consistency of the results. The conclusion drawn from this is that the inconsistency in the results generally arose from imperfections in the technic in making the tests rather than from physiological variations of the bacteria.

CONCLUSIONS.

Methods for use in bacteriological investigations of soil are still in an experimental stage. Various methods have been suggested for different purposes, but none of them have been generally adopted. For the purpose of making a flora study of soil, it has proved necessary to develop many new methods.

The poured plate of agar or gelatin has been used as the basis of the present flora studies. Plate culture is useful for two purposes: It gives some idea as to the number of micro-organisms in the soil; and it serves as a basis for qualitative work.

When using the plate method for quantitative work, the important points in technic are the temperature used for incubation and the length of time the plates are incubated. Fairly low temperatures (18° C. for gelatin and not over 25° C. for agar) and at least 7 days incubation have been found to give the best results. The exact composition of the medium used (within certain limits) does not seem to have much influence upon the count obtained. The two media found most satisfactory have been 20 per ct. Gold Label gelatin (or 12 per ct. of either Bacto-gelatin or U. S. Glue Co. gelatin) dissolved in tap-water, and an asparaginate agar with glycerin. Very similar quantitative results, however, can be obtained with the use of other media.

When using the plate method as the basis of qualitative work, the composition of the medium is of much greater significance. No single medium has been found upon which all kinds of soil microorganisms can be recognized by means of their colonies. The sporeforming bacteria can be recognized fairly easily upon 15 to 18 per ct. Gold Label gelatin (in tap-water). Some kinds of Actinomycetes can be recognized by means of their colonies on asparaginate-glycerin agar; but unfortunately this medium fails to show any difference between two of the most common types. The different kinds of non-spore-formers cannot be recognized upon any of the media yet investigated with the exception of Ps. fluorescens, which is not one of the most abundant types.

For a complete qualitative analysis of the soil flora, at least three media may prove necessary in plating, one of them adapted to bring out the distinctive characteristics of the common soil types of each of the three groups of soil micro-organisms. By plating soil on three such media, it would then be possible to recognize at a glance the most abundant types. The types that could not be recognized by means of their colonies could be isolated and studied in pure culture.

Um