232 A Biometrical Study of Conjugation in Paramecium 
The facts regarding variation in length for the four series A, G, D and E are 
shown graphically in Diagram I. (pp. 234, 235). This will perhaps bring out 
more forcibly the facts of Table V. In the diagram we have the frequency 
distributions plotted to such scales that the areas of all are equal. The units 
of abscissae are microns of body length, and the ordinates give percentage 
frequencies. Beginning at the top the polygons are arranged in order of collection. 
Polygons for conjugants are drawn with continuous lines, those for non-conjugants 
with dotted lines. 
The differences between conjugants and non-conjugants are made ver}^ evident 
by these diagrams. The}^ also bring out clearly another point, namely the 
progressive change in the means as we pass from the beginning of the epidemic 
(Series ^) to the end (Series E). The mean length of the conjugants increased, 
during the period from August 15 to August 30, 13"637 microns or 8"2 per cent, 
of the original mean. In the same way, during the period from August 15 to 
September 6 the mean length of the non-conjugants increased 24<"494 microns, or 
" ordinary" Paramecia from a culture in which conjugation is not occurring. Now the worth of this 
contention can best be judged by an examination of the following table, which has been formed by 
rearranging the values given in Table II. (p. 226). 
Coefficients of Variation for Paramecium. 
A 
Conjugants 
B 
Non-conjugants in 
cultures where 
conjugation was 
occurring 
C 
" Ordinary " Paramecia 
of selected ancestry from 
cultures in which con- 
jugation never occurred 
6-668 (210) 
8-185 (210) 
8-736(500) 
7-439 (-202) 
9-123 (202) 
8-292(500) 
8-529 (132) 
8-772(500) 
9-345 (500) 
9-030 (500) 
7-369 (500) 
Av. = 7-053 
Av. =8-612 
Av. = 8-672 
Now the "ordinary" Paramecia of this table (column C) were all of selected ancestry, i.e., were 
the resultants of the continued division of one single individual. Therefore we should expect them 
to show if anything reduced variability as compared with Paramecia of mixed ancestry. Further their 
entire cultural history, beginning with the original single individual, was known and in it conjugation 
did not occur. But our non-conjugants (column B) are not, as Mr Lister presumes that they are, 
more variable than are these " ordinary " individuals. They are on the whole about equally variable, as 
we should expect them to be. Finally it is clear from column A and other data given in the body of the 
paper that the variability of conjugants is of a totally diJferent order from that of non-conjugants 
or " ordinary" Paramecia in fission generations. In the light of these figures based on the careful and 
painstaking measurements of large numbers of individuals I leave it to the reader to judge of the 
significance of Mr Lister's " grave source of error " in my results. B. P. 
