306 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1925 



I will not attempt to deal with these various fjeneia in any detail, 

 but it may be of interest just to ^rive some indication of the gen- 

 eral nature of the principles underlying one of the systems of classi- 

 fication, which differs, however, in some respects from that of Hansen. 

 According to this, yeasts may be divided into five main groups. The 

 first group includes the Schizosaccharomycetes^ characterized as I 

 have indicated above by their method of reproducing by transverse 

 division. The ascs often result from isogamic conjugation, and as a 

 rule from four to eight ascospores are formed. 



The second group, or Zygosaccharomycetes^ consists of yeasts in 

 which conjugation of a more or less well-marked character is con- 

 nected with the formation of the asc. This may, perhaps, be a very 

 primitive group, and one from which, possibly, the parthenogenic 

 yeasts have been in process of time derived. This group has been 

 divided into a number of genera, depending on the precise character 

 of the conjugation process. 



The third group comprises all those yeasts which reproduce 

 mainh' by budding, and in which the formation of the asc is not, 

 so far as is known, preceded by any process of conjugation. This 

 group, which has been subdivided into several genera, comprises the 

 true Saccharomycetes — that is to say, the alcohol-producing yeasts 

 of industry. 



The fourth group, which includes the genera Pichia and Willia, 

 comprises yeasts which do not exhibit any trace of sexuality, and 

 which usually develop in the form of a film or skin upon the 

 surface of the liquids in which they happen to be growing. These 

 organisms do not, as a rule, produce any appreciable (Quantity of 

 alcohol, but some of them give rise to the production of fruity 

 ethers. Finally, there is a fifth group, including tw^o or three 

 genera, consisting of organisms w^hose connection with the Saccha- 

 romycetes is somewhat doubtful. 



From what I have said it will have been gathered that the 

 division of the yeasts into a number of more or less Avell-defined 

 genera has been based almost entirely upon differences in their 

 morphological and physiological characters. For the further dif- 

 ferentiation into species it was found necessary, in many cases, to 

 adopt other methods of investigation, such as the behavior of the 

 yeasts toward certain selected carbohydrates, and observations on 

 the optimum conditions required for the formation of ascospores 

 and of films. 



With regard to the first point, on which I have already touched, 

 it has, for example, been found possible to subdivide the genus 

 Saccharomyces — which is included in the third of the main groups 



