Report to the Water Research Committee. 



419 



power, but that they vary in several directions towards powerfully 

 liquefying forms of the B. termo type, on the one hand, and yellow 

 chroinogenic forms, like Zimmermann's B. ocliraceus and P. Frank- 

 land's B. arborescens, on the other. 



A careful and prolonged study of the plate-colonies and " wander- 

 ing islets " was made under high powers, throwing light not only on 

 the movements of these colonies over the surface of the gelatine, but 

 also on the structure of all such colonies as are composed of tresses, 

 zooglcea-clumps, and contoured groups of filaments, motile or not. 



Drawings were made illustrating the movements and changes in 

 form of the "wandering islets" at intervals, and showing the indi- 

 vidual movements of the bacilli on which these depend. Evidence is 

 also given showing that the differences in size, vigour of movement, 

 and liquefaction, and chromogenic power, depeud on the vigour of 

 the organism as it comes from the river, as w r ell as on variations in 

 conditions of culture. 



Group YII includes seven forms evidently closely allied to the last 

 group, if really distinct. They are remarkable for their more or less 

 pronounced yellow pigment formation, and are termed the yellow 

 Proteus type. 



I have identified some of the varieties here collected together as 

 Zimmermann's B. radiatus and B. ocliraceus and P. Frankland's B. 

 arborescens, and there can be little doubt that other yellow chromo- 

 genic forms described will have to be included. Among these, a 

 number of "non-liquefying" yellow forms may also have to be 

 counted when they have been sufficiently studied in detail. 



Group VIII includes a form which varied so much that I had to 

 make numerous separation cultures at various periods during the 

 two years it was kept in culture, to satisfy myself there was no 

 admixture. Its most characteristic features were the star-like 

 character of the plate-colonies and the presence of a bluish irides- 

 cence when held up, but a yellow pigment was also formed in varying 

 degrees of intensity. In other characters it comes near Group IX, 

 on the one hand, and the yellow Proteus type of Group YII on the 

 other, but its peculiarities were so puzzling that I kept it apart. 



Group IX comprises a series of golden-yellow liquefying forms, of 

 which five varieties were studied. The type is that of Zimmermann's 

 B. julvus and B. subflavus, and, while they differ in liquefying power, 

 all form deep chrome-yellow pigment, like yolk of egg. The group 

 is clearly allied to the yellow Proteus group, and seems to connect 

 P. Frankland's B. aauatilis and the true Proteus type with B. 

 arborescens and B. ochraceus, the differences depending largely on the 

 formation of zooglcea or not. 



Group X consists of a series of very common non-liquefying forms, 

 of which three varieties were studied in detail. The type appears 



