418 



Prof. H. Marshall Ward. 



abundant in the river. Four varieties were isolated and cultivated. It 

 is the same as that previously described by P. Frankland as B. viscosus, 

 and by Schmolck as B. fluorescens nivalis. Some of the varieties 

 liquefied very feebly or not at all, suggesting that B. fluorescens non- 

 liquefaciens may be only a form of the present species, a conclusion 

 borne out by comparing the two side by side, on, and in, various 

 media. 



Group IV comprises the non-liquefying forms of the type 

 B. fluorescens non-liquefaciens, extremely common in the river, and of 

 which eight were isolated and kept in culture. 



These ei^ht varieties differ in details, but the differences are not 

 constant, and their behaviour suggests the identity of several so- 

 called species of green fluorescing bacteria, hitherto regarded as 

 distinct, e.g., the Bacillus of Lesage, that of Adametz, Lustig's 

 B. aquatilis fluorescens, Flugge's B. fluorescens putidus, and certain forms 

 described by Frick and Zimmermann. 



The fact that some of these varieties were pathogenic and some 

 not, points to the possibility of there being two species, but as 

 experience shows that pathogenicity varies as well as other 

 characters, this could not be relied on, any more than the differences 

 in size, colour, &c. 



Group V embraces a large series of forms comprising the typical 

 B. coli communis and a number of varietal and allied forms. In all six- 

 teen varieties were studied in detail, some of them very closely and for 

 long periods. They are common in the river, especially in summer, 

 thouo-h some of the forms show variations so marked that few 

 observers would be likely to suspect their close relationships. For 

 instance, varieties occurred incapable of coagulating milk; others 

 which formed no gas-bubbles in gelatine ; others not pathogenic for 

 guinea-pigs. Some other forms, suspiciously like B. coli in other 

 respects, but forming capsules, or liquefying gelatine to a greater or 

 less extent, are placed in different groups in deference to usually 

 received opinion, though I am convinced of their close relationship. 



High-power cultures show considerable differences in size of 

 rodlets, and facts were obtained of value in understanding the very 

 great differences observed in the macroscopic characters. 



Group VI comprises a series of forms centering around the type of 

 Hauser's Proteus vulgaris, and of which eleven forms were isolated 

 and cultivated. 



One of the most striking features of this type is its variability as 

 to liquefying pow er ; another is the concurrent variations in motile 

 power of the plate-colonies. I recognised the types named P. 

 mirabilis and P. proteus by Hauser, and during cultures extending 

 over three years have convinced myself that not only are these and 

 his P. Tienkeri merely varieties of the same form differing in liquefying 



