136 THE BEGINNINGS OF LIFE. 



living matter ( c micrococci ') may develop under dif- 

 ferent conditions and especially in different media 

 either into Torul* ('cryptococci'), or into Bacteria 

 ( c arthrococci '), and that both these forms may subse- 

 quently grow into perfect Fungi. We do not agree 

 with him, however, in his view that c micrococci ' are 

 invariably derived from fungi, and that they constitute 

 some of the normal reproductive units of these organ- 

 isms. We believe, and we think we have proved, 

 that Bacteria are frequently evolved de novo > 3 and we 

 further believe that when they do arise from pre- 

 existing fungi, they are either thrown off accidentally 

 as mere living particles, or on the other hand that they 

 result from the heterogenetic breaking up of the proto- 

 plasmic contents of a previously normal fungus-spore 

 or sporangium 1 . So that when derived from pre- 

 existing fungi in any stage of development, they are 

 accidental and occasional, rather than normal and con- 

 stant products. 



We believe that Bacteria and Torul* merely represent 

 two of the most prevalent forms which specks of new- 

 born living matter are prone to assume ; and in confir- 

 mation of this view, we may state that all intermediate 

 shapes are frequently to be seen between the most 

 typical Bacteria and the most typical Torul<e -, just as all 

 intermediate conditions are to be seen between the 

 smallest Bacteria of some highly fermentable infusion, 



1 Just as Bacteria may result from retrogressive changes taking place 

 in an Amceba which has reached the term of its existence (see p. 222). 



