Januakt 15, 1904.] 



SCIENCE. 



113 



X Reversion 



117 



f34, yellow 

 f7Y, yellow i ,„ ^ 

 ' -^ (42, yellow 



15, wild 

 5, white 



148, gray 



— There were three oflFspring-; one was uni- 

 formly black (a color not found in the an- 

 cestry for at least three generations) and two 

 were white. Black rats were got by Crampe 

 (1877, p. 394) by crossing mongrels between 

 wild and white-and-blaek rats ; and black mice 

 by von Guaita by crossing piebald dancing 

 mice with albinos. 



VIII. Pielald 



(f6, white \ 

 [7, yellow/ 



X Gray-yellow 



(( 8, yellow \ 



yellow / ' 



—Of the six offspring four (nos. 77, 78, 79, 

 81) had the back colored yellow-red. In 

 matching the color with the color wheel red, 

 orange and yellow were found to constitute 

 60 per cent, to 70 per cent, of the color. In 

 no. 82 the dorsal pelage was yellow-chocolate, 

 with the color formula, N51, E28, Y12, W9;* 

 and in no. 80 it was chocolate, with the color 

 formula, N76, E13, Y4, W7. Four had some 

 white on the belly or flanks. In this cross a 

 new color — chocolate — arose, which I interpret 

 to mean that some of the primitive gray was 

 added to the yellow; that is, there was a par- 

 tial reversion. Otherwise the yellow is dom- 

 inant. 



IX. Gray and White 



/ r 37, yellow (6, white X ''', yellow) \ 

 y'^ I 7, yellow ) 



(f 4, white \ 

 ^H 5, white)- 



— There were eight offspring, three albinos and 

 five gray. Four of the latter had white 

 patches on the sides, legs or middle of the 

 belly. This is like the case described by 

 Castle (1903, p. 543). 



C. SUMMARY OF .RESULTS. 



When the parents are of the same color, 

 especially if they are pure bred, there is a 



* N, nigrum, black; E, red; Y, yellow; W, 

 white. The numbers are percentages. 



Strong tendency for the offspring to be of the 

 same color as the parents. In the case of 

 albinos this tendency is so strong that the off- 

 spring are probably always albinic; if both 

 parents are black the tendency to black off- 

 spring is likewise very strong; if chocolate, 

 the result is more variable ; if yellow, still more 

 intermediate. The results indicate that there 

 are different degrees in the strength of inherit- 

 ance of different colors. 



When the parents are of dissimilar color 

 the offspring show different kinds of inherit- 

 ance in the different cases. When gray and 

 white are crossed the offspring are gray with 

 a little white, and this white is the more re- 

 duced the purer bred the albinic parent; the 

 gray is dominant. Likewise in the cross of 

 gray and black, black is quite shut out, and 

 the same is true of gray and piebald rats 

 (Crampe, 1877, p. 394). The wild, gray color 

 is strongly prepotent. Melanism and albinism 

 act quite similarly in crossing. Both are in 

 the nature of ' sports.' Perhaps ' purity of 

 the germ cells ' is the mechanism of isolation 

 for which we have been so long looking, by 

 which mutations are preserved from the 

 ' swamping effects of intercrossing.'* 



When gray and chocolate are crossed the 

 gray is incompletely dominant. When black 

 and white are crossed, typical reversions ap- 

 pear; neither color is dominant. When black 

 is crossed with yellow the result is highly 

 variable. If white and yellow be mixed in 

 various proportions for several generations the 

 progeny takes on various shades of yellow and 

 may acquire the wholly new color of chocolate. 

 Similarly, black may result from a mixed an- 

 cestry in which there is no black. The study 

 of the cropping out of new colors certainly 

 forms an enticing subject for further inquiry. 



D. BEARING OP THE RESULTS ON MENDEl's LAW. 



The enthusiasm kindled by the discovery of 

 a new law leads us to go to extremes, to as- 



* Since the above was written I have received 

 Castle's paper in Science for December 11, 1903, 

 in which he states that the long-haired character 

 in the Angora guinea-pig is recessive. This char- 

 acter is probably a mutation, and as such behaves 

 in accordance with the above hypothesis. 



