56 



The Journal of Heredity 



3 (number single plus number mar- 

 ried) and the number of questionnaires 

 returned, as shown in Table 2. is due 

 to the fact that the information con- 

 cerning only the marriage of some de- 

 ceased graduates was furnished on 

 many questionnaires, and this infor- 

 mation has been incorporated in the 

 columns "Number single" and "Num- 

 ber married" of Table 3. Other appar- 

 ent discrepancies in the number of 

 persons listed in Table 3 are due to 

 the fact that the ages at graduation or 

 marriage of many persons were not 

 given. Information concerning many 

 individuals was fragmentary. 



It is evident that during the three 

 decades the per cent marrying, the 

 age at graduation, and the age at mar- 

 riage remained fairly constant for both 

 the men and the women. The low 

 average age at marriage for the 

 women in the classes of 1873-1879, as 

 well as the great fluctuations in the 

 percentages of childless married wom- 

 en graduates, may be due to the small 

 number of families concerned. 



Before indicating the conclusions 

 which logically follow from the analy- 

 sis of the data, it is necessary to con- 

 sider some of the data's limitations 

 and peculiarities. 



(1) The number of individuals dealt 

 with is small. Consequently, the con- 

 clusions arrived at are of general sig- 

 nificance only when considered in con- 

 nection with researches among gradu- 

 ates of other colleges, and among 

 other sections of society. 



(2) Sometimes it was not clear 

 whether the number of children indi- 

 cated included deceased children. The 

 data on children in such questionnaires 

 were not used in the following compu- 

 tations. 



(3) The age of the wife in many 

 cases was less than 45 years. Women 

 imder that age may have more chil- 

 dren. Only sixteen wives were less 

 than 40. Thirty-one ranged in age 

 from 40 to 44. As will be evident 

 later, two series of com])utations were 

 made : one including all families 



whose total number of children to 

 date was ascertained, and another 

 using the same data, though excluding 

 all families which may still produce 

 more children. 



(■i) The 554 living graduates form 

 a group which has survived thus far 

 the ravages of disease. They are 

 doubtless much healthier on the aver- 

 age than their 180 colleagues ( 24.5 

 per cent of the individuals of the 

 period studied) who are deceased. 

 Beir has shown that long-lived persons 

 have more children on the average than 

 short-lived ones. Therefore, the sur- 

 viving seventy-five per cent of the 

 graduates were doubtless more fecund 

 on the average than the whole group 

 of graduates combined. 



This view is confirmed in a limited 

 way by the fact that the mean family 

 size was 1.07 children in the families 

 of fourteen deceased male alumni. 

 The total number of children in each 

 family was known. Compare this fig- 

 ure with the families in Tables 5 and 6. 



About ninety-four per cent of the 

 married graduates (men and women), 

 whose children are used in our com- 

 pilations, are living. Hence the fecund- 

 ity of the married individuals who 

 returned the questionnaires has doubt- 

 less, like that of the living alumni in 

 general, been higher than for the 

 whole group of graduates, living and 

 dead. This fact should be kept in 

 mind. 



The deaths have naturally been re- 

 latively more numerous among the 

 earlier graduates (38 per cent for the 

 period 1870-1884) than among the 

 later (17 per cent for the period 1885- 

 LSt)9). The fact that the number of 

 children per capita is larger in the 

 earlier period than in the later (see 

 Tables 5 and 6) is doubtless due in 

 part to the more highly selected con- 

 dition of the earlier group. 



There is no reason, however, for 

 suspecting that those who returned the 

 questionnaires are any more or any 

 less vigorous physically than those, 

 still living, who did not return them. 



