22 ETHEL M. ELDERTON 



We cannot for a longer period consider the eye and hair colour to remain even 

 approximately constant. With a shorter period we might fail to obtain sufficient 

 material for statistical purposes. 



III. — There are ten kinds of normal first cousins. Let A and B stand for the two 

 cousins, thus : — 



Tiro (A and B may be sons of two brothers. 

 Male - A and B may be sons of two sisters. 

 Cousins\A and B may be sons of a sister and brother. 



Two (A and B may be daughters of two brothers. 

 Female-, A and B may be daughters of two sisters. 

 ( husins\A and B may be daughters of a brother and sister. 

 Male I A and B may be son and daughter of two brothers. 

 and ) A and B may be son and daughter of two sisters. 

 Female \A may be the son of a sister and B the daughter of a brother. 

 Cousins [A may be the son of a brother and B the daughter of a sister. 

 In this classification in the last group of " male and female cousins," A is taken 

 as the male and B as the female cousin. But in the actual schedule provision is 

 made for the case where the observer has taken A for the female and B for the 

 male cousin. 



The observer, after entering his or her own name, should fill in the names of 

 the cousins A and B and their sex by putting a cross under male or female. Next, 

 under type of cousinship, a cross should be put in the last column against the special 

 type of the two cousins observed. This is very important, because we have reason 

 to believe from the grandparental and avuncular relationships that the degree of 

 resemblance varies a good deal with the type. 



IV. — Any individual cousin A may be dealt with in any number of cases, but 

 it is not desirable to compare one cousin A with more than four other cousins who 

 are brothers and sisters to each other, and of these, not more than two should be 

 of one sex. Subject to this limitation A may appear, or A's brothers and sisters, 

 in any number of cousinships. A fresh schedule should be used for each such cousin- 

 ship. It is not, however, necessary to fill in on these additional schedules all the 

 measurements and characters. The name and sex only of the repeated cousin, and 

 the type of cousinship, need to be inserted. A cross reference to the number of the 

 observer's series in which the cousin is fully recorded will then suffice. A blank 

 is left for this reference under the name and sex of cousin. For example, if P, Q, 

 R, S be children of four different brothers and sisters, we first fill up a schedule 

 of P and Q, then one of R referring to the schedule containing P ; and another 

 schedule referring to the schedule containing R and to the one containing Q, but 

 giving the type of Q and R cousinship. Then we measure S and refer to P, and 



