N. H. SWELLENGREBEL AND C. STRICKLAND 
363 
TABLE I. (Series F.) 
lay after 
Number of 
Number of non- 
Day after 
Number of 
Number of non 
feeding 
infected fleas 
infected fleas 
feeding 
infected fleas 
infected fleas 
1 
3 
0 
10 
2 
2 
1 
2 
0 
11 
1 
2 
2 
5 
0 
12 
1 
2 
3 
2 
3 
13 
1 
1 
4 
3 
3 
14 
1 
3 
5 
2 
12 
15 
2 
2 
6 
2 
5 
16 
i 
3 
7 
3 
i 
18 
i 
1 
8 
1 
3 
9 
4 
3 
Total 
37 
46 
Because the possibility existed that the flagellates we found in the 
fleas of Series F were natural forms (in accordance with Patton’s 
views), we started 
a control series with 
60 fleas out 
of the same 
culture. 
These fleas 
were allowed 
to feed 
for 12 hours 
on an unin- 
fected rat 
and were 
treated afterwards in 
the same way 
as the fleas 
of Series 
F. Table 
II shows the 
results of this experiment. 
TABLE II. 
(Series 
G.) 
Day after 
Number of 
Number of non- 
Day after Number of 
Number of non- 
feeding 
infected fleas 
infected fleas 
feeding 
infected fleas 
infected fleas 
1 
0 
4 
8 
0 
4 
1 
0 
2 
9 
1 
2 
2 
0 
4 
10 
0 
12 
3 
1 
3 
11 
0 
7 
4 
0 
5 
12 
0 
2 
5 
0 
4 
13 
0 
2 
6 
0 
4 
7 
0 
3 
Total 
2 
58 
Here we see that 3'3°/o of the control fleas were infected, or (if we 
do not take into consideration the first two days) 4°/ 0 . This infection 
of the control fleas was probably due to an infected wild rat having 
broken into the flea box. But even if we consider the flagellates in 
the control fleas to be natural forms which have no connection with 
T. lewisi, we may safely conclude that the majority of the forms found 
in the fleas of Series F were developmental forms of T. lewisi because 
the rate of infection was much larger in Series F (36'9°/o) than in 
Series G (4°/ 0 ). This difference becomes more striking when we 
consider that generally a flea of Series F was only considered infected 
