Karl Pearson 
377 
is a member should be as close as to the parent itself — either equally represents 
the character of the pure line. This method, as far as I can see, might be very 
profitably used to test pure line theories, because the parent will generally have 
been reared in the same environment as its co-fraternity*. Table IV has then 
TABLE IV. 
Number of Tentacles of Member of Parents Fraternity^ and of 
Offspring of Parents. 
Number of Tentacles of Parents' Brethren. 
<o .5 
6 
9 
,0 
11 
Totals 
6 
4 
3824 
7681 
3143 
384 
70 
8 
15114 
15 
6370 
17869 
7980 
1587 
218 
25 
34064 
8 
9 
3322 
10258 
5919 
1102 
175 
22 
21807 
9 
6 
733 
2514 
1235 
338 
52 
8 
3886 
10 
36 
167 
93 
35 
6 
337 
11 
5 
91 
4 
11 
2 
154 
Totals 
34 
14290 
38580 
18415 
3457 
523 
63 
75362 
been formed giving the relation between offspring and each member of the parents' 
co-fraternity. We have : 
Mean 
Standard Deviation 
'Aunt or 
Uncle " 
(•170 
•816 
)hew or Niece " 
(75362) 
7-213 
•851 
091 
002 
Kegression 
•095 
± ^003 
It is cleai' that the correlation between parent and offspring is doable what it is 
between the parents' co-fraternity and the offspring, a result wholly inexplicable 
on the theory of the pure line. 
We can follow this idea further and remark that the degree of resemblance 
between every member of the pure line ought to be exactly the same. There 
ought to be no distinction at all between the correlation of individuals with their 
ascendants, their descendants and their collaterals in all grades. Now in dealing 
with man and with mammals generally, the biometric school has found the 
resemblance of brethren to be slightly greater than that of parent and offspring, 
parental correlations running from "4 to '5 and fraternal from -5 to -6. For 
* I do not see further why it should not be applied to test "pure lines " in the case of bisexual repro- 
duction, where it is met at once by the fact that the relationship between uncle or aunt and nephew or 
niece has only about one half the intensity of that between parent and child ; yet the parent and the 
uncle or aunt are both products of the same pair of "pure lines." If it be said, which I am ready to 
admit, that the gametes of one individual are not all alike, that on bisexual reproduction the zygote will 
not always be the same, and that the somatic characters will be related to these selected gametic 
characters, the obvious reply is that there is no evidence at present that, eitlier in budding or in 
parthenogenetic reproduction, there is always a transfer of precisely tlie same type of cells as basis for 
the new individual. 
t Parent excluded, i.e. " Uncle and Nephew." 
