100 A CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



for achieving this very end. Such cultures permit us to determine which among 

 the vast diversity of germs present in a rich inoculum can successfully compete 

 with the others under the specific environmental conditions, determined and im- 

 posed by the investigator, so that gradually they become the predominant micro- 

 organisms in the culture. This proced^ire, chiefly initiated by Beijerinck and 

 Winogradsky (see their Collected Works, published in 1921 and 1949, respec- 

 tively) is pre-eminently suited to determine by direct experiment what particular 

 features of the environment are responsible for the abundant or exclusive devel- 

 opment of special types and, by inference, to clarify the "natural" conditions for 

 their existence. Furthermore, the results provide the information necessary for 

 studies on the behavior of pure cultures under such conditions. And last but not 

 least, they can be used to isolate at will from natural sources representatives of 

 those types whose ecological relationships have been sufficiently established. This, 

 in turn, makes it possible to conduct comparative studies with several strains 

 isolated from different localities in order to elucidate the normal range of varia- 

 tion displayed by the "wild types." Amplified with investigations on the competi- 

 tive value of observed differences in characteristics the accumulated knowledge 

 promises to be far more significant for reaching a satisfactory solution of tax- 

 onomic problems than are the results of those "standard tests" which at present 

 are the chief basis of our methods of differentiation, and which are generally 

 performed under conditions and with media utterly at variance with the "natural" 

 ones. (See, in this connection, e. g., van Niel, 1949b; Winogradsky, 1952.) 



But however much the approach outlined above may contribute to a better 

 understanding of the microorganisms in question, we should not anticipate that 

 it will solve the "species problem," and this for the reasons already mentioned. 

 Once this is recognized, the question arises whether a more promising attack can 

 be suggested. In this connection I believe that Winogradsky's latest publication 

 (1952) has opened up prospects for sound developments. In essence he proposes 

 the establishment of "biotypes," rather than species, genera, etc., for those groups 

 of bacteria that are easily recognizable and accessible and that represent special 

 and distinctive patterns of characteristics which can be related to the normal role 

 of the organisms in nature. Around these "biotypes" are to be grouped the 

 numerous "satellites," comprising the strains that differ from the "types" only 

 with respect to some secondary details, these to be indicated simply by numbers. 

 Abandoning all attempts at further classification, Winogradsky concludes (1952, 

 pp. 130-131) : 



. . . je ne pense pas que ce travail [i.e., to reconstruct present systems of classifica- 

 tion along these lines] puisse etre entrepris avant longtemps; je crois neanmoins, que 

 mes suggestions se montreront utile du jour ou les bact^riologistes, fatigues par I'aspect 

 touffu de la systematique bact^rienne, songeraient a la reformer en faveur d'un mode 

 plus simple et, a mon avis, plus rationnel. 



II se peut que certains microbiologistes soient cheques par I'idee de supprimer la 

 classification Linneenne dans le cas des bacteries, habitues qu'ils sont de s'en servir 

 pour toute classification. 



Or, tout travail 6tabli selon les regies de cette classification devrait etre base avec 

 quelque precision sur le principe philogenetique, qu'il est impossible d'appliquer aux 

 bacteries. II serait done plus correct de nous borner a I'appliquer au regne animal et 

 au regne v^g^tal, 6u il est bien k sa place, sans chercher a englober dans sa sphere les 

 formes plus el^mentaires de la vie. 



