Raymond Pearl 
231 
We first notice the large differences between the means. Averaging for the 
four series the conjugants are 14"9 per cent, shorter than the non-conjugants living 
in the same culture at the same time. The absolute variability, measured by 
the standard deviation, averages in Series A, G and D (all from the same culture, 
it will be remembered) 30'8 per cent, lower in conjugants than in non-conjugants. 
Of course we should expect the standard deviations for conjugants to give lower 
values than for non-conjiigants, because the conjugants have the lower mean 
values. Not all the difference, however, is due to this factor, as a glance at the 
coefficients of variation at once shows. From these it is seen that (averaging Series 
A, C and D) in proportion to size the conjugants are 18"9 per cent, less variable 
than the non-conjugants. Series B is in apparent contradiction to this conclusion 
regarding variability, since it gives a negative difference in both variability 
columns ; that is, in this series, the conjugants are the more variable both 
absolutely and relatively. The probable errors show, however, that neither for 
standard deviations nor coefficients of variation are the differences significant. 
This Series B, for reasons stated above (p. 219), is a very short series, comprising 
only 12 pairs of conjugants, and a corresponding number of non-conjugants. As a 
matter of fact, within the large errors incident to such small random samples, 
conjugants and non-conjugants are equally variable in Series B. The means for 
this series are in good accord with the others, showing as they do a difference of 
15"8 per cent, in favour of the non-conjugants. 
I think then, that there can be but one conclusion from the facts set forth in 
Table V., viz. that the Paramecia wJiich are conjugating at a given time are 
markedly differentiated from the non-conjugant individuals living in the same 
culture at the same time in both type and vainahility. For the character length of 
body this differentiation is in the direction of lowered mean and variability for 
conjugants as compared with non-conjugants*. 
* Note added Nov. 10. Mr Lister (loc. cit.) in criticism of this conclusion says : "Now the 
non-conjugant population of the latter species [Paramecium caudntum] measured by Dr Pearl to ascertain 
the range {sic) of their variability -would include, not only ordinary individuals, but all stages of 
individuals in process of differentiation as gametes. The non-conjugants are a heterogeneous popu- 
lation ; the conjugants are, on the other hand, approximately homogeneous. This appears to me 
another and grave source of error in his results on the degree of differentiation and variability of the 
conjugants." My conclusions from the data presented in this paper are that conjugants are smaller in 
mean dimensions and less variable than non-conjugants. The first of these conclusions Mr Lister 
agrees to, because by searching the literature he has found that Maupas observed the same thing. If he 
had searched the literature a little more diligently he vpould have found that practically every one who 
has ever worked on conjugation in Paramecium has noted this fact. Finally if Mr Lister had himself 
ever carefully observed a culture of Paramecium in which conjugation was going on he would not have 
needed to go to the literature to learn that conjugant Paramecia are smaller than non-conjugant. To 
the second of my conclusions, viz., that conjugants exhibit lower variability than do individuals in 
fission generations Mr Lister objects that the non-conjugant populations measured were heterogeneous. 
In his opinion then the difference between the variation constants for conjugants and non-conjugants 
arises because the non-conjugant populations exhibit a spuriously high variability resulting from 
heterogeneity. Or in other words his contention is that a random sample of non-conjugant individuals 
from a culture in which conjugation is occurring will show higher variability than a random sample of 
30—2 
