260 A Biometrlcal Stndij of Conjugation in Paramecium 
the absolute values of the means are low. From these it would appear that, in 
proportion to its size, the difference is sensibly equally variable in the conjugant 
and in the random pairs. It is oerhaps doubtful, however, whether any significance 
should be attached to the coefficient of variation in this case, considering the 
character we are dealing with. I think it is quite clear that in comparing the 
conjugated pairs and random pairs of conjugants with respect to the variability of 
the difference, we should use the standard deviation as the measure, since we are 
dealing with identically the same individuals in the two cases. The noteworthy 
thing in this case is that random pairings, whether of conjugants or non-con- 
jugants, give higher mean differences between the members of the pair, and 
these differences more widely scattered about their mean value, than do normal 
actually conjugated pairs. 
The results of this section may be summarized as follows : In a series of 
samples taken from different cultures and from the same culture at different times 
in the history of a conjugation epidemic, it has been fo und that there exists a high 
degree of correlation between tlie lengths of the two members of conjugating pairs. 
There is a sensible, though not a high, degree of correlation between the members 
of conjugant pairs with respect to the other characters studied, viz., breadth and 
index. These high homogamic correlations have been shown to be due to something 
other than random pairing in a homogeneous population of low variability. An 
examination of the distribution of the differences in length between the members 
of conjugant pairs has shown tliat, as woidd be expected, the mean difference is 
much lower and there is less "scatter" about the mean in the case of the conjugating 
pairs, as compared toith random pairings of either conjugants or non-conjugants. 
VII. Discussion and Interpretation of Results. 
All discussion of resiilts, either in the way of accounting for the phenomena or 
pointing out their significance, has been deferred to this point for the reason that 
with all the data in hand the relationship of the diffei'ent sets of facts to each 
other comes out more clearly. 
We may consider first the fact that the conjugant population is differentiated 
from the general non-conjugant population in the same culture at the same time. 
Maupas* maintained that there were several morphological stigmata by which 
Infusoria in the depressed condition preceding conjugation were marked. One 
of these morphological characteristics was reduced size. That pre-conjugants are 
smaller in size than ordinary individuals has been noted by a number of other 
workers. Thus, for example, Gruberf says that " meist kleinere durch rasch 
aufeinander folgende Theilungen enstandene Individuen sich conjugiren," though 
he proceeds in the next sentence to point out what is quite true, that it is " nicht 
* Arch. d. zool. exper. et gener. (2) t. vi. pp. 165 — 277 and Ibid. t. vii. pp. 149 — 517. 
t Bericht der Naturf. Ges., zu Freiburg. ijB. Bd. ii. p. 51. 
