COAT CHARACTERS IN GUINEA-PIGS AND RABBITS. 67 



The departure here from the expected equality of the two classes is 

 too great and too uniform to be referable to chance. It would seem 

 probable, rather, that fewer gametes are formed bearing the character 

 short-coat than are expected. This maybe due either to failure of con- 

 trasted characters to segregate at gamete formation in certain cases, or 

 possibly to partial latency of the character short-hair in gametes which 

 transmit active the character long-hair, the resultant being an interme- 

 diate condition. Further experiments are needed to show which of these 

 two ideas is correct, though at present the former seems more probable. 



Short-haired females obtained by mating a primary hybrid, Sh. (.), 

 with a pure short-haired animal have, in several cases, been mated 

 with the long-haired males, 2002 (PL i, fig. i) and 2060 (PI. 3, fig. 5). 

 It is expected that half of such females will contain recessive the charac- 

 ter long-coat, and that half will be free from it, z'. e., will produce only 

 short-haired young. The outcome shows that 9 out of 12 females 

 tested contain recessive the character long-coat, for they produce long- 

 coated young. This is another bit of evidence that segregation 

 occurred less often than expected in the gametes of the primary hybrids, 

 though it throws no doubt on the uniformly dominant character of 

 short-coat. Again, the nine extracted hybrids, which contain reces- 

 sive the character long-coat, should, when mated with long-haired 

 animals, produce short-haired and long-haired young in equal numbers 

 if segregation occurs in every case where it is expected. In the mat- 

 ings in question there have been produced 10 short-haired young and 16 

 long-haired ones further evidence that segregation is less frequent than 

 expected. It is true that these numbers are yet small, but their uniform 

 deviation in one direction from the expected result indicates that they are 

 significant. If so, we must recognize in this pair of characters full 

 Mendelian dominance, attended, however, with only partial Mendelian 

 segregation. 



CORRELATION AMONG COAT-CHARACTERS. 



In the preceding pages we have discussed separately the heredity of 

 three different pairs of alternative coat-characters which we have found 

 to conform more or less closely with Mendel's law of heredity. It 

 remains to inquire whether there is any correlation between one of 

 these pairs of characters and another, /. e., whether rough coat is 

 more often associated with white or with pigmented coat, with long 

 or with short coat, etc. An inquiry of this sort may be based upon ex- 

 periments made with the albino male 2002 (PI. i,fig. i), which possessed 

 the recessive members of two of the pairs of characters (viz, albinism 

 and long-coat), but the dominant member of the third pair (viz, rough 

 coat). It goes without saying that he was pure as regards albinism 



