390 PSYCHE. 
and third form seem to be modifications 
of each other, for, in the longer and 
more clavate hairs of the second form, 
the granular matter in the hair extended 
without interruption into the larva, 
and these hairs often burst and give out 
viscid fluid. The larger hairs of the 
second form are often upon the sides of 
the conical warts which bear at their 
summits hairs of the third form. Hairs 
of similar nature, but smaller and with- 
out viscid fluid, clothe the larva of Oxy- 
ptilus periscelidactylus and Pterophorus 
monodactylus. 
Mimeseoptilus phaeodactylus feeds on 
Ononis repens, M. mictodactylus on Saxi- 
fraga granulata, Lioptilus sericidactylus 
on Vernonia noveboracensis and Aciptilus 
lobidactylus on Solidago ?canadensis: 
thus, as will be noticed, all the species 
of plerophoridae mentioned above feed 
on plants clothed with glandular or long 
hairs.* The glandular hairs of the larva 
protect it, probably, by causing it to re- 
semble the surface of the plants on which 
it feeds, a kind of resemblance not rare 
in insects; but it is also probable that 
the secretion of the glandular hairs also 
protects the larva, to some extent, from 
the attacks of ichneumons and of other 
parasites, for I obtained no parasites 
from over fifty specimens of <Aciptilus 
*Miss Murtfeldt writes meas follows: ‘I had not 
thought to mention it in connection with my description 
of Leioptilus sericidactylus but there 7s a very close 
imitation in the dermal clothing of the larva to that of 
the young leaves of Vernonia, on which the spring and 
early summer broods feed.” ...‘‘Later in the season, 
when feeding chiefly on the flowers, the larva acquires a 
purplish tinge which, with the particles of the flowers 
that adhere to its glandular hairs, is a sufficient disguise 
from any but the eye practised in its detection.” 
lobidactylus, reared from larvae taken 
when nearly full-grown, while I have al- 
ways obtained, in rearing a much less 
number of Oxyptilus periscelidactylus 
under similar conditions, several para- 
sites. It is noticeable that the viscid 
secretion upon the larvae of Aciptilus 
lobidactylus is alkaline to litmus paper, 
while the fluid from the tubercles of the 
larvae of Attacus cecropia is acid. 
Turning from the study of insect-hairs 
which furnish more or less temporary ef- 
ferent ducts for glands, there are forms 
of hairs—leaving out of account, as un- 
proved, Weismann’s curious, but not 
improbable, view* that scales of lepi- 
doptera may be sometimes ducts for an 
odorous fluid secreted by cells at their 
bases—which are regularly and perma- 
nently the outlets for glands at their 
bases. An interesting example of this 
kind of open hair duct is furnished by . 
the hairs upon the foot of the common 
house-fly (Musca domestica), through 
which hairs, as recently shown by De- 
witz,’° a sticky secretion is poured, the 
fly being enabled by the adhesion of this 
secretion to rest upon smooth vertical 
surfaces. Leydig,® in 1859, describes 
and figures glands in the tarsi of several 
coleoptera; West,” in 1862, describes 
and figures the tarsal hairs of the fly and 
of many other insects; and Dewitz,’® in 
1882, discusses the function in locomo- 
tion of the viscid secretion which is dis- 
charged by the glands of the foot through 
the tarsal hairs, both in flies and in 
coleoptera. 
Defensive glands of another kind, the 
ducts of which open into spines, are the 
