COAT CHARACTERS IN GUINEA-PIGS AND RABBITS. 21 



is small, as we expect on the theory of probabilities, but we find no 

 falling off in production of albinos as the amount of albino ancestry 

 decreases. 



Similarly throughout Table B we find the expected Mendelian equality 

 of albino and pigmented young approximated, irrespective of the num- 

 ber of albino grandparents and great-grandparents. Thus the proportion 

 of albinos is actually higher when there are only two than when there 

 are three albino grandparents, though on Darbishire's hypothesis we 

 should expect this relation to be reversed ; for the young produced in 

 cases where there are three albino grandparents aggregate 55 albino to 

 64 pigmented animals, an excess of 4.5 pigmented ; but the young pro- 

 duced in cases where there are only two albino grandparents (in the same 

 total number of young) aggregate 63 albino to 56 pigmented animals, an 

 excess of 3.5 albinos. The deviations from equality are in the two cases 

 opposite in character and almost equal. Undoubtedly it is merely a 

 chance outcome that they are opposite in nature to what Darbishire's 

 hypothesis demands. The tables as a whole, however, do give an 

 emphatic negative to Darbishire's position. They lend support to the 

 alternative (Mendelian) hypothesis, that any pigmented animal which 

 forms albino gametes forms approximately 50 per cent of such gametes. 



In what precedes I have made no mention of what has been called 

 Pearson's modification of Galton's law of ancestral heredity. Else- 

 where (Castle, : 03") I have discussed this briefly in applying a statis- 

 tical test to the laws of Galton and Mendel in the case of albinism in 

 mice. Galton's law I applied generation by generation to Von Guaita's 

 ('98, : oo) mouse records, as Galton ('97) himself had applied his law 

 to the Basset hound records. The test thus made showed the complete 

 failure of Galton's law as applied to the heredity of albinism. No such 

 detailed test was made in the case of Pearson's law, but the statement 

 was made : " Comparing Pearson's series with that of Galton we see 

 that the parental influence is reckoned as substantially the same by both 

 Galton and Pearson, but that Pearson assigns a much greater influence 

 to the more remote ancestors than does Galton." For which reason 

 it was concluded, " The discrepancies noted between observed and cal- 

 culated [in testing Galton's law] will remain and even be accentuated 

 if we replace Galton's series with one of those suggested by Pearson. 

 For the result will be unchanged [I should have said similar, rather 

 than unchanged] in Generation II, but the calculated numbers will in 

 most cases diverge still more from the observed ones, in the later gener- 

 ations, because Pearson attaches more weight to the remoter ancestors 

 than does Galton." To these conclusions Pearson (: 04) takes exception, 

 maintaining that neither of the two series which I took from his writ- 

 ings was quoted in a form comparable with that of Galton's series. 



