PARTULA OTAHEITANA. 175 
which bear red young only. Therefore the excess over 77 or 21 comprises indi- 
viduals which are RR in composition themselves and which have mated with DR 
yellow adults, but whose young are too few to include both yellow and red individuals 
in the expected numbers. Clearly such red adults could not have mated with DD 
yellow adults, for in that case their young would be all yellow (RR x DD = DR only). 
Hence we must correct (2) above by transferring to the middle class the 21 red 
adults mated with DR, but which do not provide yellow young, with this result: 
(9) Red adults: yellow young only, 16; yellow and red young, 28; red young only, 77=121 
But we have found that adults bearing red young only, amounting to three- 
fourteenths of all in that class, really belong to the middle class. Presumably, 
therefore, some of the adults bearing yellow young only would contain red young 
also if their offspring were sufficiently numerous to represent their genetic consti- 
tution and that of the other parent. Such individuals would have to be RR them- 
selves, and they would have had to mate with DR in order to bear yellow and red 
young. The only basis for estimating their relative numbers is the three-fourteenths 
fraction of the red class that had mated with DR, and bore young of one color only. 
Proceeding on this basis, we may subtract three-fourteenths, or 4, from the first 
class and add that number also to the middle class, when (g) becomes: 
(10) Red adults: yellow young only, 12; yellow and red young, 32; red young only, 77=121 
Now these three classes stand for the other mates of the red parents whose 
young have provided the data for the analysis. Hence we have arrived at a deter- 
mination of the relative numbers of DD and DR in the yellow part of the whole 
population, unattainable on inspection, which is, specifically, DD : DR ::3:8. 
In order to test the truth of the above analysis and its results, we may return 
to the yellow adults of (7) to see if their embryonic contents correspond to expecta- 
tion should the proportionate numbers of DD, DR, and RR be as derived from the 
entirely independent data relating to red parents. Such yellow parents (74) are 
either DD or DR, in the proportion of 3 to 8, or DD 20: DR 54. 
The 20 DD yellow parents will have mated with DD, DR, and RR in the pro- 
portions indicated in (10); but all of their young in all cases would be yellow, for 
the RR combination is impossible in their offspring. 
The 54 DR yellow parents will also have mated with DD, DR, and RR in the 
proportions of 12:32:7, respectively. Specifically, 
DRXDD= 5.5 cases yellow young only. 
DRXDR=14 cases yellow and red young. 
DRXRR=34.5 cases yellow and red young. 
Combining the figures, we have: 
(11) Yellow adults: yellow young only, 25.5; yellow and red young, 48.5 =74 
Before we may compare the expected with the actual figures of (7), certain 
combinations and corrections of the latter must be made. A yellow adult, being 
DD or DR by assumption, could not bear red young only, whatever its mate; the 19 
individuals so classified must therefore be DRX DR, whose offspring are not abun- 
