PARTULA OTAHEITANA. 215 
of 591 are: plain adults 76.3 per cent, banded adults 23.6 per cent, taken as 24 per 
cent. The gravid adults with distinguishable young are as follows: 
Plain adults: plain young 296; both kinds, 8; banded young, 13 =317 
Banded adults: plain young, 28; both kinds, 10; banded young, 39 =77 
Proceeding first on the assumption that banding is recessive, the proportions of 
the DD, DR, RR mates of the RR adults come out as 13:46:18, the corrections 
to be made amounting to 3 of the first and third figures as given above. When 
these relations are assumed in an analysis of the plain parents, the expected number 
of adults with plain young and adults with both kinds are 112 and 205, respectively, 
as compared with the empirical corrected figures of 137 and 180, a difference of 25 
out of 317, or 7.8 per cent. 
On the contrary assumption, when the proportions of DD: DR: RR are worked 
out from the plain series, and tested for the 77 banded adults, the departure of the 
observed facts from expectation amounts to 10.4 per cent, and is thus greater than 
in the alternative case. 
PapEIHA VALLEY. 
This association is the least favorable of the five, apparently on account of the 
relatively small numbers of the banded adults, which constitute only 8.5 per cent 
of a series of 163. Nevertheless, the agreements between expectation and observa- 
tion are remarkably close on the assumption that banding is recessive. The gravid 
parents, when classified as to their embryonic contents, are: 
Plain adults: plain young, 75; both kinds, 0; banded young, 3=78 
Banded adults: plain young, 3; both kinds, 1; banded young, 3 =7 
From the second line, the proportion of DD, DR, and RR mates come out as 
1:5:1. Using these figures for an analysis of the plain adults, the expected 
numbers of adults with plain young would be 22, while those with both kinds of 
young should be 56; the actual numbers, with the adjustments made, are 25 and 
53, a difference of only 3 out of 78, or 3.8 per cent. On the contrary assumption, 
the proportions of DD: DR: RR would be 3:4:71, which seem absurd on inspec- 
tion; the figures can not be tested for the other series on account of the small 
numbers involved. 
Uruural VALLEY. 
In a population made up of 79 per cent plain and 21 per cent banded, the 
parents with classified young are the following: 
Plain adults: plain young, 94; both kinds, 2; banded young, 9=105 
Banded adults: plain young, 6; both kinds, 0; banded young, 17=23 
Without retailing the steps in the process of analysis, simply the results 
may be stated. On the assumption that banding is recessive, the proportions of 
DD:DR:RR prove to be 2:9:12; the independent test for the plain parents with 
plain young and parents with both kinds as 26.5 and 78.5, respectively, in theory, 
as compared with 28 and 77 as deduced from the observations. The difference is 
I.5 in 105, or I.4 per cent, whereas on the contrary assumption the difference 
amounts to 26.1 per cent. 
