234 



The Journal of Heredity 



h can be shown, however, that there is 

 a real phvsiolo_L,Mcal correlation from 

 other data. The evidence from the data 

 of Holmes and Loomis will serve as an 

 example. 



This (lata were olitained from students 

 at the L'niversity of Wisconsin. In 71 

 families fairly complete records of hair 

 and eve color were obtained for all 

 brothers and sisters, parents, and grand- 

 parents. The authors' conclusions have 

 already been noted briefly. The data are 

 worthy of further analysis. For the 

 present jnirpose eye color is conve- 

 niently divitled into two classes — dark 

 and light. The division is made at the 

 point at which a bluish or greenish tint 

 is obviously present. I'.lack. brown, 

 light brown, and hazel are called dark, 

 while green brown, gray, and blue are 

 called light. Hair color is classified as 

 dark, light, or red. Ijy dark is meant 

 onlv black in adults, but both black and 

 dark brown in children. This compen- 

 sates, in part, for the great change with 

 age. The category red is made as wide 

 as possible, including auburn, chestnut, 

 and a few yellows, as well as dark and 

 light reds. Even thus it includes only 

 59 individuals or 8.87c- The tables are 

 based on 364 adults (parents and grand- 

 parents), and 305 children. Among 

 the adults 136 are dark haired. 200 

 light haired, and 28 red haired. Among 

 the children these classes number 81. 

 193. and 31, respectively. The adult 

 males do not differ appreciably from 

 the females in hair color, but have 

 slightly lighter eyes (68% light in 

 males. 58% in the females). 



'r.\riLi; I. — Per Cenl of Lis,hl Eyes Associated 

 with Each Hair Color 



Hair 



Dark 

 LiKht . 

 Red.. 



Adults Children 



41 

 76 



42 

 64 



6S 



Dark-haired persons clearly have 

 light eves much less frecpiently than do 

 light (jr red-haired persons. That thi> 

 is not due to assortative mating such 

 as would be the case if the students 

 came in part from Italian, in part from 



Scandinavian families, is shown by 

 Table 2. in which the percentage of 

 hus])ands and wives of those with the 

 different hair and eye colors are given 

 for each hair and eye color. 



T.\BLE 2. — Association of Hair and Eye Colors 

 in Husbands and Wives by Per Cent. 



Persons with dark hair showed some 

 preference above the average for their 

 own color, but also for red hair and for 

 light eyes. Persons with light hair pre- 

 ferred their own hair color, but showed 

 no marked preference in eye color. 

 Those with red hair show a very 

 marked preference for black hair and 

 also for black eyes. Dark-eyed persons 

 showed little preference in hair color, 

 but chose light-eyed mates considerably 

 in excess of the average. The converse 

 of the last statement is true of light 

 eyed persons. While the numbers are 

 too small in most of these cases to at- 

 tribute much significance to them, it is 

 at least clear that there is no assorta- 

 tive mating which can account for the 

 correlation between hair and eye color 

 found in imlividuals. There can be no 

 (juestion that light hair is connected 

 j)hvsiologically with light eyes. In the 

 case of red hair, the difference between 

 72% light eves among 5^) red-haired 

 persf)ns and 41% light eyes among 227 

 dark-haired persons is significant and 

 not due to assortative mating. 



This conclusion that hair, eye, and 

 also doubtless skin color are physiolo- 

 gicallv connected, finds abundant analo- 

 gies in the lower maiumals. Indeed we 

 have found it to be the rule that dilu- 

 tion f;ictors of any kinfl (as opposed 

 to pattern factors) produce tluir elTect 



