30 TAXONOMY 



pathogenic species with every desirable gradation of virulence, but they 

 are not new varieties. They are mere laboratory stocks, which revert at 

 once to the original type as soon as the original conditions of life are 

 restored *. It has not been hitherto possible to entirely suppress a single 

 biological character in any species. The supposed conversion of the anthrax 

 bacillus into the harmless B. subtilis was never realized, and, like most 

 reports of a similar character, arose from imperfection in the technique of 

 pure culture. 



Although there is no longer any doubt that the various kinds of bacteria 

 constitute definite species and genera, it would be futile to minimise the 

 great difficulty that confronts us as soon as we try to arrange them in a 

 system of classification. The morphological similarity of the spherical 

 forms (cocci) and the close resemblance to each other of the rod-shaped 

 bacilli make a classification based on purely morphological data quite 

 impossible. Physiological characters have therefore been made use of as 

 auxiliaries — for instance : growth on various media and need for oxygen 

 or particular forms of nourishment ; specific products such as pigments, 

 granulose, sulphur, fluorescence, phosphorescence ; specific chemical changes 

 such as fermentation, putrefaction, disease. A reliable description of the 

 species of bacteria is a task for the future, and will only be attained by the 

 united labours of pathologists, physiologists, chemists, and botanists (19). 

 Much has been done by arranging them in groups according to their most 

 prominent physiological characters. Examples of such groups are the 

 saprogenic or putrefactive, the zymogenic or fermentative, the chromogenic 

 or pigment-producing, the phosphorescent, the thermogenic, the nitrifying, 

 the iron bacteria, and the sulphur and purple bacteria. Such a classification 

 is not without value, but it is not justifiable to make generic names on this 

 principle, and use them as though the groups they indicate were equivalent 

 to true morphological genera. Such names as Photobacterium, Nitro- 

 bacter, Nitrosomonas, Granulobacter (for butyric bacteria with the granulose 

 reaction), Halibacterium (for marine forms), Gotiococcus, and Proteus (for 

 several putrefactive bacteria) are names that should not be admitted in 

 a systematic classification. As working or trivial names they are useful 

 enough, but have no claim to equality with genera based on morphological 

 characters, which must, after all, form the ground-work of all systems of 

 classification. An attempt must be made, despite the dearth of material, 

 to obtain clear morphological definitions of the genera at least, even if it 

 be impossible at present for species. This was Cohn's guiding principle 

 (20), and must be adhered to. The very factor to which the progress of 

 bacteriology is mainly due, namely the number and variety of its students, 

 has been a hindrance to the proper growth of a classification. Side by 



The various ' races* of brewery yeasts have a different value to these ' attenuated' bacteria. 



■3 1 



