30 



SCIENCE 



[N. S. Vol. XLIV. No. 1123 



betterment is emphasized as well as its direct 

 value in the advancement of knowledge, and 

 strong claims are made for its support. 



It certainly is the case that animal psychol- 

 ogy in this country has, in the past decade, 

 done very solid and instructive work with very 

 little financial support from universities or re- 

 search funds. The experiments here reported 

 represent a gift of the time of one man of sci- 

 ence and a gift of material resources from an- 

 other. They are typical of the scientific de- 

 votion and self help which the public can 

 profitably reward by any means in its power, 

 and which any individual honors himself by 

 supporting. Edward L. Thorndike 



Teachers College, 

 Columbia University 



RETROGRESSION IN AMERICAN 



LONGEVITY AT ADVANCED 



AGES 



It is generally suspected among a limited 

 group of scientific men that although we seem 

 to be improving in matters of health we are 

 doing so in spite of adverse conditions at the 

 more advanced ages. 



We have certainly improved on the whole, 

 for the area in the United States from which 

 acceptable records in mortality statistics are 

 received annually (the registration area) has 

 doubled in the number of states included, 

 within the past decade (1900-1910), although 

 it is still no more than half of the total num- 

 ber of the states of the Union, to the shame of 

 such great states as Illinois, Iowa, Kansas, 

 Nebraska, etc. That mortality conditions have 

 improved in the neighborhood of the age of 

 birth and in fact, at all the earlier ages, is so 

 well established that it needs no comment. 

 Also, the general death rate in this country 

 has decreased more in the past decade (2.6) 

 than in the previous two decades taken to- 

 gether (2.2). 



But all this improvement is too deceiving; 

 it covers up the fact that in some respects we 

 are worse off now than we were twenty years 

 or more ago. Stated concretely we expect to 

 show in this paper that individuals between 

 the ages of about 50 and 75 do not, on the 



average, live as long now as they did twenty 

 years ago ; and the extent of this retrogression 

 is increasing. We shall refer to this period or 

 interval of ages as the Period of Retrogres- 

 sion. 



We hope to point out also slight indications 

 of tendencies to " come back " at the still more 

 advanced ages, say from 75 on. That the indi- 

 viduals at these extreme ages are "coming 

 back " seems pretty firmly indicated by the 

 results of this investigation, but not only is the 

 " come back " small but it is also manifested 

 at ages where statistical data are faulty ; hence, 

 we recommend that these indications be held 

 in abeyance until they are more clearly veri- 

 fied by other investigations of similar nature. 



The English statistician and actuary, Mr. 

 George King, has explained a short method of 

 constructing abridged mortality tables wherein 

 only representative portions of the tables and 

 the corresponding death rates and expecta- 

 tions of life are given. We have utilized this 

 method to construct six abridged mortality 

 tables based upon the mortality statistics of 

 each of the sexes, and for the three single 

 years 1890, 1900, and 1910. The year 1880 

 was not included because the population data 

 and mortality statistics for that year which 

 were reliable do not cover exactly the same area. 

 The mortality statistics of all years previous 

 to 1880 are worthless for our purpose. The 

 essential purpose of this paper is to compare 

 and discuss the results obtained through the 

 construction of these mortality tables. 



The fact that each mortality table is con- 

 structed from data covering but a single year 

 absolutely prohibits the use of such tables ex- 

 cept to point out general conclusions such as 

 are indicated in this paper. Our attitude in 

 this matter should not be forgotten. 



As the explanation of the method of con- 

 struction of the tables is technical and has no 

 special bearing upon the interpretation of the 

 final results, we shall merely refer the reader 

 who desires further information to Mr. King's 

 explanation — which, however, will bear much 

 simplification — in the Registrar General's re- 

 port for 1914. 



The statistical data for the year 1890 com- 



