500 



THE AMERICAN NATURALIST [Vol. LV 



in both cases. Broadly speaking, the wild type long- 

 winged flies have from two to three times as great an 

 expectation of life at any age as do the flies of the Quin- 

 tuple stock. Since all of these flies liv^ed under substan- 

 tially identical environmental conditions, as has been set 

 forth earlier, it follows that the basis of the great differ- 

 ence in expectation of life between these two groups, as 

 <'X('ni|)lihcd in l^'igs. 1 and 2, is hrrcditnry and not en- 

 rironnien/al. 



•'). It is apparent that on the whoh' the graduations 

 givTn by equations of the tvpc oT (i) are very satisfac- 

 tory, and as good as coiiM iv.-i-on.-ihlx he expected on ex- 

 perience bases of the niagniindc of those here dealt with. 

 T^ndoubtedly the curvcN would be slightly more smooth if 

 we had larger experience, especially in Tables IV and V", 

 where we are dealing with less than a thousand flies in 

 each case, but in the main the curves fit the observations 

 very well. 



4. The death rates (q,) generally increase steadily with 

 advancing age. An exception to this rule is the slight 

 dip b.-tweeii a-ev L^") iN i„ th,. >hor1 -xviiiu'e.l table. 



in :: and [ \\u- 1),n.n,,lnl„ I lin... nr.' cnnipaivd 



N\ith III. human / lin- taken iVnm (JlnxrrV il^.O l!»l() 

 r. S. Lifr Tal.h'^. In n,-,!..,- 1o niak.' a ,iu>t coniparl^on. 

 the human /, line is displacr.l to the left in the diaurams 

 nntil auv lift.Mm of human lif.' eoineides with auv one 

 of the tiy eurw. This drops out tin' infant and ehil.lhood 



nndei'slood that in the ])resent instance, this is a some- 

 what aihitrarv and purely graphic procedure. Whethei- 

 the point which , >o,lhi eorr(-])ond< on the human cui've 

 to the beginning of Ww I) , n.n pl/ihi iniau.. .-urNc i- exactly 

 15 years or- ^'^ oi- 14. or Minn- otiiei' near h\ \alue, i- a 

 matter for furthm- rcseareh. In a hroad way, however, 

 it i. clear that tin- tu.. linr. nui^t Im- lak. n h. .-..rrr^pond 

 at s.>m.'thiHu' like this pnint in the human curve. 



Fi'om the^e dlauram^, it i^ appar.-nt that, after leas in- 

 out the infant mortality component of the human curve. 



