79 
plete immobility of those segments, Papilioninae stands alone, con¬ 
stituting an apparently separate family that made no further progress 
in this direction. 
The first of these lines of development, the Parnassian, seems the 
simplest; an early result is Thais (figs. 3-5), which has a truly Papilionid 
pupa, but one with certain peculiar features which it will be more in¬ 
structive to return to, when we come to consider the origin of 
Nymphalidae. 
In the second line of development, the Pierid, we find a change 
which does not occur, so far as I know, in the pupa of any of the 
Heterocera. By the development of certain dorsal tubercles at the 
margins of the segments, (possibly due to the fusion of the anterior 
trapezoidal tubercles of the larva), antero-posterior movement is lost, 
only lateral movement is possible ; so that whilst a Papilionid pupa can 
move its tail in any direction, that of a Pierid (figs. 6-11) can only do 
so from side to side. The peculiar method of suspension adopted by 
the Papilionid pupa, restricts antero-posterior movement, perhaps even 
renders it dangerous; accordingly, even in Papilio, this is less free than 
lateral movement, and we can understand how easily it might be lost, 
and the Pierid form originated. In the Pierids there is also a change 
in the egg ; the hemispherical, smooth form gives place to a taller, 
ribbed structure. 
Among the Pieridae further changes in the pupa occur, resulting in 
a further loss of motility of segments; in Pieris (figs. 9-10-11), and in 
the Bhodocerinae, we find only the 5th segment movable, whilst in 
Euchloe the pupa is solid and immovable. 
There is, I think, something to be said from a larval, and even from 
an imaginal point of view, in favour of the notion suggested by the 
pupse, that there is at least as great a gap between Aporia and Pieris, 
as between the latter and the Bhodocerinae. 
The third line of development from the primeval, Papilio-like form, 
is in the direction of the Lycaanids. Here my material is so scanty 
that I have found very little trace of intermediate forms. There is, 
however, I think, much to be said for the possibility of the Lycaanids 
having been derived from the Hesperids, merely grazing, as it were, 
the Papilionids on the way. 
The Lycaanids at least acquired the same method of pupal suspension 
as the Papilionids ; and further, they lost the “ Micro ” habit of the 
Hesperids of hybernating as larvae, and acquired the habit of doing so 
in the egg or pupa-stage; the egg, likewise, is much farther from a 
Hesperid egg than is that of a Papilionid. Parnassius, at first a very 
puzzling form, comes in here to give us some little assistance. Its pupa 
is very much like that of the Lycaanids, from which it differs chiefly in 
lying free in a cocoon, instead of being suspended by a girth ; this con¬ 
dition is, however, attained by not a few Lycaanids. The egg possesses 
certain Lycaanid features, in particular the depression at the top around 
the micropyle; hybernation also takes place in the egg-stage, a 
circumstance very rare in butterflies, except amongst the Lycaanids; 
but I have ascertained that in Parnassius the young larva is developed 
in the autumn, and passes the winter coiled up within the egg-shell. 
The Parnassids are a small family and, so far as we know, did not 
develop such a variety of forms as the parallel Lycaanids. Partiv, 
probably, for this reason, and partly because Thais, although apparently 
