166 
GENETICS: W. H. TALIAFERRO 
Proc. N. A. S. 
Erdmann 8 believes that she has demonstrated the same type of phenomena 
in Paramecium after endomixis. While the increase in variability is 
probably due to some nuclear phenomena during the life cycle of the tryp- 
anosome in the flea, we have no method of judging, from the present data 
on the subject, whether it is a sexual or a reorganization process. Minchin 
and Thomson 4 have suggested that the peculiar effects produced by 
passage through the invertebrate host are connected with the transfor- 
mation of the trypanosome into a crithidial stage and the reverse process, 
both of which always take place in the life cycle in the invertebrate host. 
It is conceivable that passage of a "pure line" through the inverte- 
brate host might cause significant changes in the mean size as well as 
in the variability. All of our experiments have shown this not to be 
true. It is true that after the "pure line" is passed through a flea there 
is often a difference in the mean which is significant from a statistical 
standpoint, but as these differences are never greater than the maximum 
difference given above for growing the "pure line" in different rats, we 
cannot ascribe the difference to any peculiar effect of the flea. It is 
probably due simply to the fact that the trypanosomes are measured in 
different rats. 
Size and Variability in "wild" 1 Injections Occurring in Nature. — A num- 
ber of naturally infected rats were collected from around Baltimore and 
Washington. Up to the present time measurements have been made of 
specimens from ten of these infections. While these "wild" infections do 
show differences in their means and a higher coefficient of variation on the 
average than the pure lines, one is struck with the constancy of the means 
and the low variability. The longest mean length (rat 221) obtained is 
32.503 ±.060 and the shortest (rat 411)29.093 ±.062 with a difference of 
3.410 =*= .087. It can be seen that the difference between the longest 
"wild" infection and the shortest is comparatively great and that the 
difference is much greater than is obtained by growing a "pure line" in 
different rats. The least variable "wild" infection (rat 413) showed a 
coefficient of variation of only 2. 125 ±.101. This is slightly lower than 
any of the coefficients of the "pure lines" in the laboratory. The most 
variable "wild" infection (rat 65) showed a coefficient variation of 4.583 
± .223 The difference between the least and the most variable "wild" in- 
fection is 2.45 ± 24. The remainder of the "wild" infection measure fall 
between the figures given here, both as regards their means and coefficients 
of variation. 
From these results we must conclude that while the "wild" infections 
occurring in nature show comparatively small differences in size, these 
differences are greater than can be explained by the fact that the in- 
fections are occurring in different rats. We must also conclude that most 
"wild" infections consist either of very few "pure lines" or a larger number 
