INTERPRETATION OF RESULTS 399 
The number of microscopic ficlds to be counted will depend some- 
what upon the kind of data that are desired. If this method is to be used 
simply for the purpose of dividing milk into grades, it will in most cases 
be unnecessary to do the actual counting, since a Grade A milk will show 
field after field without any bacteria at all, while a Grade C milk will 
show the field crowded with bacteria. In all doubtful cases, however, 
counting should be done, and there should never be less than thirty fields 
counted in order to have reliable results. Counting thirty fields is not 
so tedious a task as would seem to be, since, in ordinary milk, the num- 
ber of bacteria in each field is small, and the counting may be done 
very rapidly. 
Counting. Counting the bacteria in such a smear may be dore in 
two ways: 1. The number of groups of one or more bacteria present. 
2. The number of individuals. The second, cf course, is really the 
correct count of the number of bacteria, but the former will give a count 
much closer to that obtained by the plate count, since the colonies upon 
the plate represent groups of bacteria rather than individuals, each 
group growing into a single colony only. Extensive tests have shown 
that there is a fair correspondence between the number of groups re- 
ported by experienced observers and the number of colonies that may 
grow in plates made from the same milk, although there are occasionally 
discrepancies of considerable extent. These discrepancies are caused 
by variations in judgment as to what constitutes a group, variations in 
the extent to which groups break up in the dilution waters when the 
smears are made, and the presence of dead bacteria or of bacteria which 
do not grow on the plates. Some experience is needed by the micros- 
copist in determining Just what should be counted. In high-grade 
milks, an inexperienced person is apt to fail to recognize differences 
between bacteria and other minute objects. This results, as a rule, in 
an overcount by inexperienced men. In milk containing many readily 
recognizable bacteria in each field the inexperienced man is apt to over- 
look some of them, giving an undercount. These difficulties are over- 
come, however, by training and experience. 
Interpretation of Results. It must be recognized that the results 
obtained from the microscopic record give a closer approximation to the 
actual number of bacteria present in the milk than those obtained by 
the plate method, since the plate method will count as one, either a 
single bacterium or a group which may sometimes contain a hundred or 
even more individuals. Inasmuch, however, as the plate count has 
become a method of analysis that is well known and commonly applied, 
it becomes desirable to know as closely as possible what relations there 
