CLASSIFICATION OF DAIRY BACTERIA. 97 
different stock cultures may be forced, by modifying conditions, 
but rather what are the physiological characters of the types 
that actually exist under the conditions of nature and which 
may be found in normal dairy products. Most of the types 
described in the following pages, therefore, are actual cultures — 
obtained from dairy materials, and in no case do they represent 
modifications of such forms, except such modifications as natur- 
ally occur under the normal cultures in the laboratory. A few 
cultures sent us by others form an exception to this rule. 
To our mind the aim of classification of such a group of 
bacteria at the present time should be as follows: 
1. To recognize the groups of bacteria in the terms above 
used, and then to describe these groups in such a way that 
they may be with tolerable ease recognized by others working 
upon the same subject. 
2. To recognize the kinds of variation possible within the 
groups. By this is meant the study of as many natural cultures 
of the members of the group as possible, noting what variations 
appear in the different cultures, as isolated from milk, and thus 
determining to what extent the group characters are modified 
in the different forms of bacteria as isolated from actual milk 
products. 
3. To find, if possible, the natural limits of such groups. 
By this, of course, is meant to determine the limits within 
which the variations may occur, and yet the type in question 
may be legitimately regarded as belonging to the group in 
which it is placed. If no such limits could be found, natur- 
ally the whole question of classification of bacteria would be 
pretty nearly hopeless. It may be that as information accumu- 
lates, it will be found that all of these groups so run into each 
other as to make it impossible to logically separate them. This 
is, of course, what would be naturally assumed upon the 
general theory of evolution of types, but whether such is the 
case can only be determined after an immense amount of data 
has been accumulated. At the present time, with the data at 
hand, it seems to be possible to arrange these forms into fairly 
distinct groups which, though connected more or less with 
intermediate types, are nevertheless recognized with as much 
certainty as the different types of animals and plants may be 
recognized. 
