Raymond Pearl 
257 
that consequently we are dealing with a spurious correlation which appears 
because of local differentiation within the culture. If the two individuals lying 
nearest in the field of view to each conjugant pair measured be taken as a 
pair and a correlation table formed for such pairs, the coefficient for the homo- 
gamic correlation is sensibly zero. Yet such pairs came at the same instant 
fx'om the same environment that the conjugated pairs did. Therefore the observed 
correlations for conjugated pairs cannot be due to these factors. 
We may now proceed to look at tliis fact of a high degree of homogamy in 
conjugant pairs in another way. To say that there is a high degree of homogamy 
TABLE XVIII. 
Frequency Distributions of the Difference in Length hettveen the Members of 
(a) Observed Conjugant Pairs, and (b) Random Pairs of Non-Gonjugants. 
Conjugant Pairs 
Difference 
Frequency 
in 
microns 
Series A 
Series C 

Conjugated 
Kandom 
Conjugated 
Eandom 
0— 1-9 
11 
10 
12 
4 
2— 3-9 
19 
4 
22 
8 
4-— 5-9 
14 
11 
17 
12 
6— 7-9 
16 
9 
13 
12 
8— 9-9 
6 
6 
11 
4 
10—11 -9 
10 
9 
5 
8 
12—13-9 
5 
6 
5 
5 
14—15-9 
4 
10 
4 
4 
16-17-9 
7 
8 
3 
3 
18—19-9 
4 
6 
2 
9 
20—21 -9 
6 
5 
2 
4 
22—23-9 
1 
6 
1 
7 
24—25-9 
1 
5 
2 
3 
26—27-9 
0 
0 
1 
2 
28—29-9 
0 
2 
0 
3 
30—31 -9 
1 
0 
0 
6 
32—33-9 
1 
0 
1 
34—35-9 
4 
0 
1 
36—37-9 
1 
1 
0 
38—39-9 
2 
0 
40—41-9 

1 
42—43-9 
1 
44—45-9 
2 
46—47-9 
0 
48—49-9 
0 
50—51-9 
1 
Totals 
105 
105 
101 
101 
Kandom Pairs of 
Non-Conjugants 
Difference 
Frequency 
in 
microns 
Series A 
Series C 
0— 3-9 
9 
12 
4— 7-9 
16 
12 
8—11-9 
15 
13 
12—15-9 
14 
6 
16—19-9 
11 
11 
20—23-9 
13 
6 
24—27-9 
10 
12 
28—31 -9 
7 
7 
32—35-9 
2 
3 
36—39-9 
3 
4 
40—43-9 
0 
8 
44—47-9 
3 
3 
48—51 -9 
0 
1 
52—55-9 
1 
0 
56-59-9 
0 
2 
60—63-9 
1 
0 
64—67-9 
1 
105 
101 
1 
