37° C. VS 20° C COUNT 293 
temperature (37° C.), the determination of the numbers of blood-heat 
bacteria is important. The counts on the two media usually differ as 
each is favorable to the growth of its own group of organisms.” The 
same authors note that the presence of liquefying bacteria on the gelatin 
plates may make it necessary to count the plates at the end of twenty- 
four hours, or at the end of a shorter time than is directed in the com- 
monly accepted procedure. Most bacteriologists apparently believe 
that the agar count at 37° C. indicates more nearly the bacteria of sewage 
origin and that it consequently is more important in ascertaining the 
potability of water. The results of Savage’s experiment show, however, 
that erroneous conclusions will be drawn if the count at that tempera- 
ture is considered to furnish indication of possible sewage pollution, 
because soil may contribute large numbers of harmless bacteria that 
grow at 37° C. 
The advantages and disadvantages of each medium are well known 
to all bacteriologists. Agar is better for field work in warm weather 
because it can be transported with less trouble either in test tubes or 
in Petri dishes; great care must be used to prevent liquefaction if gelatin 
is used. As gelatin has to be shipped in scoled containers in warm 
weather it can seldom be used in the field. A less elaborate incubator is 
required for agar because no special precauticn has to be taken to keep 
the temperature below a certain maximum. In some filter plants in 
which agar is used to control the operation agar plates are left at room 
temperature for two days before counting. This could not be done with 
gelatin plates, for which constant low temperature must be maintained. 
The count on agar is made at the end of twenty-four hours whereas the 
count on gelatin is made at the end of forty-eight hours. Spreaders 
often develop on agar plates but any inconvenience from these is offset 
by liquefiers on gelatin plates. The two media themselves differ in 
their properties; gelatin is good for bacteria, but no bacteria have been 
described that can metabolize agar. Both media are subject to wide 
variations in chemical content, but agar is more easily purified. 
Recent work by Cummings (1916) indicates that in the summer the 
agar and gelatin counts are nearly equal. In February the gelatin 
count, on Potomac River water, became sixteen times as great as the 
agar count. The gelatin count seems to follow the turbidity readings. 
Where possible as Tanner (1916) has pointed out, both media should be 
used where it is convenient. Race (1916) has stated that in water 
analysis no one medium will give a count that will possess any constant 
ration to the excremental bacteria as shown by B. coli content. The 
37° C. count is said to be nearest to the B. colt content. The highest 
