20 ILLINOIS BIOLOGICAL MONOGRAPHS [20 



antiquity. While it does not extend back to that distant age when all 

 the segments bore the same setal plan, it furnishes a connecting link 

 between that period and the present day. I have no hesitancy in denomi- 

 nating, as Dyar does, a seta as only "subprimary" when it is con- 

 stantly wanting in the first stage, however invariable may be its presence 

 in the second. For the same reason I can not consider a subprimary 

 seta of a specialized group as homologous with a primary one of a 

 generalized group. Such an homology is entirely inconsistent with the 

 recapitulation theory. 



In a word, the arrangement of the setae in the larvae of the Lepi- 

 doptera gives us every reason to believe and no cause to deny the hypo- 

 thesis Weismann expressed in 1876: "New characters first appear in 

 the last stage of individual development ; these move back gradually 

 into the earlier stages and so crowd out the older characters until the 

 latter finally disappear." 



VARIATION 



In many cases the presence of secondary or tufted setae is con- 

 fusing in determining the location of the primary ones. The acquisition 

 of additional scattered setae in one species or genus is a very common 

 occurrence and caution should be used in giving this character as bound- 

 ing any group. The condition in Drepanidae, Thyatiridae, and Geomet- 

 ridae furnishes good examples. 



When several primary setae are united by being surrounded by a 

 chitinized plate they have a tendency to vary in number. This is true 

 of the Pi group on the proleg-bearing segments of the abdomen. Pi 

 itself consisted originally of two primary setae but tau is usually asso- 

 ciated with them. In many Noctuidae the three are borne on a chitinized 

 leg plate. In notodontians, arctians, and other groups specialized from 

 the ancestral noctuid type, this plate bears many setae, none of which 

 can be homologized with the primary ones (cf. Figs. 31 and 33). 



III. Application of the Evidence 



As stated above, two setae on different segments of the same or 

 different larvae are homologous, regardless of their position at the 

 present time, when they have descended from the same or homologous 

 organs of a generalized ancestor. It follows from the definition that no 

 single case of homology can be absolutely proven. No combination of 

 circumstances is sufficient to show, beyond the shadow of a doubt, that 

 two setae on a modern larva are descendants from the same or homol- 

 ogous structures of some extinct ancestor. Nature's directions are, 

 however, sufficiently clear to remove the problems completely from the 



