10 KANSAS UNIVERSITY SCIENCE BULLETIN. 



In a second attempt, a single cell was drawn into a cap- 

 illary tube so fine that budding in two directions only was 

 possible, with the result that a single chain of cells was 

 formed in the tube. It was found, however, that the bud- 

 ding of older cells interpolated new cells in the chain, and it 

 was therefore impossible to keep track of successive genera- 

 tions without keeping the tube under observation night and 

 day. So this attempt was, for the time, also abandoned. 



The above experiments indicate that repeated selection 

 is necessary in this yeast if an abnormal standard of 

 size is to be kept up ; in this matter the yeast resembles 

 higher plants, where quantitive variations do not often per- 

 sist unless kept up by continuous selection. It is true that 

 we may conceive of a mutation in the direction of size among 

 lower plants as well as in the higher, but the variations ob- 

 served in the above experiments did not seem to have that 

 character. 



In the second series of experiments, conducted on varia- 

 tions in the form of cells, selections of abnormal cells were 

 continued for some weeks before any variations of a perma- 

 nent character were obtained. In November, 1903, a cell 

 which showed a narrow, mycelium-like outgrowth, was iso- 

 lated from a hanging drop of glucose broth culture, the cells 

 of which were the offspring of a single normal cell, isolated 

 the previous day. Growth after isolation was slow, but 

 after one or two days the extension and branching of hypha- 

 like outgrowths produced a mass resembling a small myce- 

 lium. This showed little yeast character until after two or 

 three days, when it produced at the tips of branches chains 

 of yeast-cells, which began to reproduce by budding after 

 the manner of yeasts. But the majority of these yeast cells 

 were of a character quite different from the typical form of 

 Saccharomyces anomalus. There was a tendency to assume 

 elongated forms, to put out hypha-like prolongations which 

 sometimes branch, and to adhere in groups, characteristics 

 not found in the normal type when grown under like condi- 

 tions. (See photomicrographs. In plate I, figure 1 repre- 

 sents an old wort culture of the parent stock, figure 2 a new 

 race grown under the same conditions, and figure 3 a ten 



