270 A Biometrical Study of Conjugation in Paramecium 
a co7ijugant type. Furthei'more, this conjugant type is differentiated from the 
general population, not only in physiological but in morphological characteristics. 
Now, it seems to me that the fundamentally important question is whether 
this differentiated conjugant type is relatively more constant as tve pass from culture 
to culture or race to race than is the type of the general population. We know that 
under varying environmental influences the type of the general population in the 
case of Paramecia (and Protozoa generally) can be greatly modified, as a direct 
result of environmental action*. Is the conjugant type modified to as great a 
degree as is the type of the general population by such direct environmental 
influences ? The question is clearly one of inter- rather than intra-racial variability, 
and as such requires a much larger amount of material than is now available 
for a conclusive answer. However, the results from the present material, as will 
be shown, point very definitely to the conclusion that the "conjugant type" is 
much less variable, that is, fluctuates less, as we pass from one environmental 
condition to another, than does the type of the general population. This fact will 
have been noticed, I think, from the tables given in Sections III., IV., and V. of 
the paper, but to make certain of the matter a direct consideration of the 
problem will be undertaken here. 
In order to test the question as to whether the "conjugant" and "general 
population " types are equally changed with changing environmental conditions 
the biometrical procedure is to measure the amount of variation in the two types 
as we pass from one culture or sample to another. Now, in the present case, we 
have only four series of conjugants {A, G, D, and B) to put over against five 
series of non-conjugants {A, C, D, E, and B). The Ann Arbor series {AA and F) 
cannot fairly be included here because unfortunately no non-conjugants were 
measured from the cultures from which these came. With such small series as 
four in one case and five in another, it is quite clear that if any conclusions at 
all are to be drawn as to a difference between the series, the differences between 
the constants must be large. As a matter of fact the results of a comparison of 
conjugant and non-conjugant types in respect to length of body, are as follows -f- : 
Mean of conjugant means = 172'408. 
Mean of non-conjugant means = 203"177. 
Standard deviation of conjugant means = 4'4.59. 
Standard deviation of non-conjugant means = 10'513. 
In other words we see that as we pass from one culture to another, and from 
one state of a culture to another, the non-conjugant type fluctuates more than 
twice as much as does the conjugant type. It might be objected that we should 
* Cf. for example, Yasuda, A.: Jour. Coll. Sci., Imp. Univ. TokyO, Vol. xiii. pp. 101—140, and 
Pearl, E., and Dunbar, F. J., loc. cit. 
t It should be stated that in calculating these inter-racial constants the different series were 
weighted roughly in proportion to the number of individuals included in each series. It seemed hardly 
fair to allow the same weight to the very short series B and D as to the longer A, C and E series. So 
then the series were assigned weights as follows: A = C = S, B = D = 1, E = 'i. As a matter of fact 
it makes very little difference in the final results if all the series are given equal weights. 
