Advantages of agar.-At present agar seems to be more promising for this purpose than gelatin. No agar medium has been found as yet on which the colonies of the different non-spore-formers can be distinguished from each other; but the asparaginate-glycerin agar (see page 109) is of considerable value in distinguishing the different kinds of Actinomycetes. Unfortunately, however, the types of Actinomycetes that produce the most easily recognized colonies on this medium are generally the less common ones. Two of the three most abundant types produce non-characteristic colonies. There are media upon which all three of these most abundant types produce characteristic growth; but no such medium has been found that gives a high enough count from soil to be satisfactory for plating purposes. Work is now in progress with the object of finding such a medium. When a satisfactory medium for the Actinomycetes is obtained, a similar attempt should be made to find a medium adapted to the non-spore-formers.

Technic at present recommended.-The best procedure at present available for making a qualitative study requires the use of two media for plating. From gelatin the following points can be recognized: number of colonies of B. mycoides, B. megatherium, B. cereus and Ps. fluorescens; and also the relative numbers of Actinomycetes and lower bacteria. On asparaginate-glycerin agar it is possible to count the colonies of several types of Actinomycetes. If cultures of the unknown types of Actinomycetes are to be isolated for further study, it is advisable to use the asparaginate-glycerin agar for their selection. If cultures are to be isolated from the punctiform colonies of lower bacteria, they may be selected from either medium.

A long incubation at comparatively low temperatures is so very important in qualitative work that it needs special emphasis. Investigators who have used incubation periods of four days or less in quantitative work have claimed that as it was impossible to obtain a complete count, results obtained from the plates would have only a comparative value at best and that an incubation of three or four days would give as good comparative figures as a longer one. This argument may be justified when plating is done for quantitative purposes alone, but not when qualitative results are desired. Colonies of the more slowly growing organisms just begin to appear in four days. Counts made on the fourth day would include only a very small proportion of the Actinomycetes, and almost none of the large group of slow-growing non-sporeformers. Naturally a qualitative study of soil bacteria based on plates incubated for such a short time would give a very erroneous conception of the actual soil flora.

An example of an error caused in this way can be obtained by comparing the work of Chester37 and that of Hiltner and Störmer.38 37 Loc. cit., see footnotes 14, 2, and 3.

38 Loc. cit., footnote 4.

Hiltner and Störmer kept their plates until no more new colonies appeared (sometimes as long as ten days); while Chester counted his plates on the fourth day. As stated in the preceding paper (see p. 99), the number of punctiform colonies and of Actinomyces colonies obtained by Hiltner and Störmer agrees very closely with the results obtained in the present work. Chester evidently obtained different results, although it is hard to compare the two pieces of work as the results of each are described in entirely different ways. The chief reason for thinking that Chester's results differed from those of the present investigation in this respect is because only two of his "species" described in Report 1139 belong to the group which has been found the most abundant in the present work (slowly liquefying or non-liquefying non-spore-forming short rods). Similarly in Report 1440, where he has described the three predominating micro-organisms in one sample of soil, only the least abundant one of the three is a non-spore-former, and that one is motile so that it does not belong to the group which has been found most abundant in all the soils studied in the course of the present work. The natural inference is that Chester's short period of incubation caused him to overlook the most numerous group of bacteria in soil.

THE STUDY OF PURE CULTURES.

The classification card of the Society of American Bacteriologists.— Including the work of the present series of bulletins and that published in Cornell Bul. 338,41 over a thousand cultures have been isolated for study. In the Cornell work all of the cultures were studied by means of the methods called for by the classification card of the Society of American Bacteriologists. To some extent the same method of study has been used in the later work, and as the classification card represents the best attempt ever made to standardize descriptions of bacteria, a brief description of it is necessary here. A more detailed description of the card was published some time ago by Harding,42 and more recently by Rahn and Harding.

43

This classification card is designed as a means of recording in a uniform manner the important characteristics of different cultures of bacteria. One card is to be filled out for each culture studied. Under each heading on the card is a series of descriptive terms of which the one applying to the culture in question is to be underlined. The characteristics included on the card (in the order of the amount of stress laid upon them) are: (1) Important morphological character

39 Loc. cit., footnote 2.

40 Loc. cit., footnote 1a. 41 Loc. cit., footnote 20. 42 Harding, H. A.

The constancy of certain physiological characteristics in the classification of bacteria. N. Y. Agr. Exp. Sta., Tech. Bul. 13. 1910.

Rahn, O., and Harding, H. A. Die Bemühungen zur einheitlichen Beschreibung der Bakterien in Amerika. Centbl. Bakt., II Abt., 42:385-393. 1914.

istics such as shape, spore-production, presence and arrangement of flagella; (2) physiological characteristics, such as liquefaction of gelatin, production of acid, reduction of nitrates, and so forth; (3) less important morphological characteristics, such as size, structure, and grouping; (4) cultural characteristics, such as form of growth in broth, on agar and so forth.

The group number.-As a means of simplifying the description of these characteristics and the classification of the organisms, those features that are considered of greatest diagnostic value are summed up in what is called the "group number." group number." In the group number, the symbol of the genus according to Migula is followed by a series of ten digits, each of which denotes the presence or absence of some characteristic. The group number is as follows:

Diastasic action on potato starch, absent

.0000001 Acid and gas from glycerine

.0000002 Acid without gas from glycerine

.0000003 No acid from glycerine

.0000004 No growth with glycerine

The genus according to the system of Migula is given its proper symbol which precedes the number thus:

BACILLUS COLI (Esch.) Mig. becomes B. 222.111102.

Value of the card in studying soil bacteria.-It was hoped at the outset that each type of soil micro-organism would have a constant and distinctive group number. If that were the case, a description of the different organisms would prove a very simple task. It was realized, however, even before the preliminary work given in Cornell Bul. No. 338 was ready for publication that the types distinguished by the tests of the classification card were not species— a fact which was thought to be particularly true in the case of the Actinomycetes. It was stated in that bulletin that the descriptions given must be considered to apply to mere physiological types; but it was not realized then how far from satisfactory these types were. Further study has shown that many of the tests included among those on the card and used for diagnosis in the earlier work give too inconsistent results to serve as a basis of distinction; while on the other hand, there are many distinctive characteristics of soil bacteria not mentioned on the classification card. The card has proved of considerable value in the study of those organisms that grow well in liquid media such as the spore-formers and Ps. fluorescens; but even with these bacteria different group numbers do not always indicate different species, so it is necessary to know which of the features give consistent and which inconsistent results. In the study of those organisms which grow poorly in liquid media, such as the Actinomycetes and the majority of the non-spore-formers, the card has proved almost valueless.

These facts have made it necessary to develop methods entirely different from those used in the first part of the work. Even in studying the spore-formers, it has been found advantageous to lay less stress on physiological characteristics and much more stress on features of morphology than are called for on the classification card. In studying the Actinomycetes, it has been found necessary to lay most stress upon the appearance of their growth and their color reactions on various special agar media not even mentioned on the classification card. In studying the non-spore-formers, the card has proved of but little value and no other system of classification has yet been worked out.

The changes in technic that have proved necessary in order to obtain a better classification than that given in Cornell Bulletin No. 338 have resulted in changing the number of types recognized. The types of spore-formers are changed less than either of the other two groups. About the same number of types of spore-formers are recognized now as were mentioned in that bulletin; but the ones that are recognized now are much more clearly defined. Many more types of Actinomycetes are recognized now than were when Cornell Bulletin No. 338 was published, although most of them are still too indefinitely defined to be called species. The nonspore-formers, on the other hand, of which over twenty types were listed in that bulletin, have now been gathered together into a few

groups. Each of these groups of non-spore-formers may well contain a large number of species; but in the absence of diagnostic features no smaller groups than these can be recognized.

Diagnostic value of the tests given on the card.-The value of the classification card depends upon the consistency of the results obtained in making the tests which it calls for. In the course of the present work considerable information has been obtained to show which of these tests give the most consistent results and are therefore of most value for diagnostic purposes. It has been found that the constancy of the results depends largely upon the methods used in making their determination. The methods generally recommended for making these tests are those given in the second edition of the Standard Methods for the Examination of Water and Sewage.44 In some cases these methods have been found as satisfactory as any investigated, but often quite different methods were found to give more consistent results.

The information on this subject that has been obtained in the course of this work may be of value to others wishing to use the classification card. It seems worth while, therefore, to take up one by one the various tests that must be made in determining the group number of an organism (together with one or two of the other important tests given on the classification card), and to discuss the methods used as well as the consistency of the results obtained.

Shape of organism. There is a fairly sharp line between the rods of the Bacteriaceae and the round forms of the Coccaceae. Although there are some organisms on the border line between these two forms, there is ordinarily no trouble, in the case of soil bacteria, in distinguishing between rods and cocci.

Size of organism. In a general way, size is of importance, but the limits of variation for any particular organism are quite large. The difference between the large spore-forming bacteria and the small non-spore-formers is striking enough to be of undoubted diagnostic value (even though an occasional spore-former is of very small size). When an organism produces spores, the size of the spores is of greater diagnostic value than the size of the vegetative forms. But for the fact that it is difficult for two different men to make measurements that agree, it would be possible to characterize several of the spore-formers almost entirely by means of the size of their spores.

Flagella. The importance of this point arises from the fact that the Society of American Bacteriologists uses the nomenclature of Migula, and in Migula's classification peritrichic rods are named Bacillus, rods with polar flagella Pseudomonas, and immotile rods Bacterium. If this classification is used, it is important to have some means of determining the arrangement of flagella. Migula's

44 Published by the American Public Health Association. 1912. Boston, Mass.