798 



AMERICAN MEN OF SCIENCE 



"three 

 salary 



aft^r a few years, this young man receives an ap- 

 pointment without limit of time, a somewhat 

 higher salary should be given him, with a small 

 advance each year for, say, three years. If this 

 instructor so commends himself that the university 

 desires his further eerAice, he should receive, as 

 assistant professor, a salary which will enaible him 

 to support a wife and two or throe children com- 

 fortably, but without luxury or costly pleasures. 



The scientific man receives his doctorate 

 at the average age of twenty-seven j'^ears 

 and is then eligible for an instructorship 

 with a salary for an ''unmarried man"; 

 after "a few years" and then 

 years" more he is to receive a 

 which will enable him to support "two or 

 three children. ' ' President Eliot also says : 



The recent tendency of sons of well-to-do, and 

 even rich, families, to go into the ministry, the 

 medical profession, academic life, and the public 

 service, is one in which all patriots may well re- 

 joice. ... It is a good deal safer to give a life 

 oflSce to e married man on whom marriage has 

 proved to have a good effect, than to a single man 

 who may shortly be married with uncertain results. 



There might well be inscribed at Har- 

 vard and at other universities the words 

 which President Eliot wrote for the 

 Water-Gate of the Chicago World's Fair, 

 changing one word, so that it would read : 



TO THE 

 BRAVE WOMEN 



WHO IN 



UNIVERSITIES 



AMID STRANGE 



DANGERS AND 



HEAVY TOIL 



REARED FAMILIES 



AND MADE HOMES 



The vital statistics of the United States 

 are entirely inadequate. "Where registra- 

 tions of deaths and births exist, they are 

 imperfect, and the changing population, 

 its age composition and the amount of im- 

 migration render them difficult to inter- 

 pret. The only information concerning 

 birth rates is given by the proportion of 

 children as determined by the census,^" 



10 In January, 1917, was issued the first bulletin 

 from the Bureau of the Census giving birth rates 

 for ten states. 



but even this is unreliable. It might be 

 supposed that it would be possible to 

 determine the number of children h-^ 

 counting them, but this is not the case. 

 The children reported in the census of 

 1850 were fewer than the survivors (with 

 the comparatively small excess of immi- 

 grants) counted ten years later. There 

 are alwaj's more children given as two 

 years old than as one in 1880 as many 

 as 170,000 more. Nor can we have com- 

 plete confidence in the compilations of the 

 experts of the census. Thus in the case 

 under consideration they give^^ figures 



'^.* 



Pig. 1. Ratio of white adults of self-supporting 

 age to white children under sdxtcen years in the 

 ITnited States, according to the Bureau of the Cen- 

 sus. 



11 "A Century of Population and Growth," Bu- 

 reau of the Census, 1909, pp. 80 and 103. An an- 

 swer to a letter addressed by the writer to the di- 

 rector of the census partly explains the way this 

 error was committed but the explanation was 

 marked "confidential"! lit is, however, no viola- 

 tion of this confidence to state, as the information 

 is available from official reports, that figures were 

 not at han<l prior to 1830 and that these were 

 guessed it appears very awkwardly so as to 

 give a regular curve. 



