286 
ages are not developed in all Hymenoptera and Lepidoptera, and 
not at all in the Diptera, this would show that the power of inherit- 
ance of these ancestral traits had already in the first two of these or- 
ders begun to wane ; that their evolution had begun ona higher plane 
than the polypodous one of the Coleoptera, Orthoptera and ameta- ' 
bolous orders, and that the power was on the verge of extinction. 
Hence their appearance in certain forms and the cessation of their 
development in others may be accounted for, just as the scattered 
and sparse distribution of certain animals, and the reduction in the 
number of individuals is a preliminary step to their entire extinc- 
tion. 
The lack of these structures in dipterous embryos appears to con- 
firm the view that they are the most extremely modified of all 
insects. It should be borne in mind that such observations are ex- 
ceedingly difficult to make, the parts are so delicate and faintly 
developed, and yet when we take into account the fact that so 
_ skillful an observer as Kowalevsky detected them, who was the pio- 
neer in these studies, and who probably had no expectation of dis- 
covering such structures, and whose mind was free from any theory 
in the matter, it seems scarcely probable that he would have figured 
them unless he had actually seen them. 
Returning to the Lepidoptera and to Lagoa, with its rudimentary 
abdominal legs of the second and seventh segments of the hind 
body, we feel warranted in the present state of the subject in con- 
cluding that they may represent a persistent condition of two pairs 
of these deciduous abdominal legs. They are certainly of some 
use to the creature, and thus have survived because they, in a par- 
tial way to be sure, have been of service. The others have evi- 
dently disappeared from disuse. And it would thus seem that the 
Lepidoptera and Hymenoptera have descended, like other insects, 
from polypodous ancestors. 
If these conclusions are correct, then Lagoa, in respect to its ab- 
dominal legs, even if we do not take into account other characters, 
is a survivor of an ancient and very generalized type, and repre- 
sents, as no other known caterpillar, the polypodous ancestor of all 
Lepidoptera. 
The other alternative is that, as Graber once claimed, the abdom- 
inal legs of caterpillars are not primitive, but secondary and adap- 
tive structures. Of course these questions can only be settled by 
further researches. And it is possible that the similar abdominal 
