254 LORANDE LOSS WOODRUFF 
after a lapse of nearly three years showed a few survivors of nearly 
all the chief forms, mostly at the bottom among the algae and 
débris. 
2. Middle fauna 
It is impossible to determine any definite sequence of forms for 
the middle of the infusions—this region being, as already pointed 
out, amore or less neutral territory which is encroached upon from 
time to time by organisms from the top and bottom as conditions 
in these regions vary. 
3. Bottom fauna 
The bottom fauna also has not exhibited a definite succession 
similar to that of the top. A study of the data already presented 
shows that the protozoan forms under consideration, with the 
exception of many amoebae, are essentially surface dwellers and 
seldom resort to the bottom except during or after a period of great 
development at the top. However, there is no invariable correla- 
tion between a fall in numbers at the top and a rise in numbers of 
the same species at the bottom, and it seems clear that, in the 
majority of cases, when a species declines in one region, most of 
the animals encyst or die. The latter is certainly true for Para- 
maecium because many hundreds of passive and dying individuals, 
affording a feast for Coleps, may sometimes be seen among the 
débris at the bottom. Again, myriads of cysts of hypotrichous 
forms are frequently found at the bottom as the surface decline 
proceeds. Amoebae, among the protozoa under consideration, 
appear to give some evidence of migrating from the surface to the 
bottom which is their chief abode. The data on amoebae give 
the impression that some forms first appear in the infusions as 
amoebo-flagellates which gradually increase in size and before 
long are unable to assume the flagellated phase. The pesudo- 
podia of these are first of the guttula type but become more and 
more long and slender until many typical radiosa forms are present, 
and these in turn give place to typical large A. proteus. Only 
in certain infusions has it been possible to trace such a series, but 
