208 Annals Entomological Society of America [Vol. VIII, 
as will be shown, disappears and is replaced by a second one 
which is Semper’s ‘‘Grundmembran.’”’ The accepted name, 
middle membrane, is so well established that the introduction 
of a new term, such as middle lamella or middle layer, would be 
futile; the old term will be adopted both for the earlier and the 
later layer although we cannot see in Platyphylax that there is 
a true membrane present in the median part of the developing 
wing. 
In many sections at this age, and later, there can be found 
running through this middle layer small darker dots and short 
lines which would correspond to the middle membrane of other 
observers (Fig. 5). These are more or less distinct in different 
sections but do not show continuously for any great distance 1n 
any section and we are unable to find any regular structure 
strictly homologous to a membrane. The strands of proto- 
plasm which connect this middle layer to the outer layers will 
be spoken of as the perpendicular strands. 
All the stages so far described can be found in the developing 
wing while it is still within the peripodial cavity, in fact many 
internal rudiments show stages more advanced than these. We 
also find in these stages as well as later ones a number of dividing 
nuclei, the mitotic figures were nearly all found near the outer 
surface of the developing wing and away from that portion 
where the nuclei are most crowded together. 
The middle membrane can be recognized when the perpen- 
dicular strands and the vacuoles between: them first become 
clearly differentiated as layers of the wing. With the growth 
of the wing changes take place in the relative thickness of the 
different layers, the two original hypodermal layers decrease in 
thickness and the layers just inside of them, composed of the 
perpendicular strands and vacuoles, increase in thickness to 
finally, as will be seen later, occupy by far the largest part of the 
wing. The perpendicular strands are not straight but branch 
and divide and are generally curved for part of their length, such 
irregularities are more noticeable in the older stages when the 
strands are longer. Each strand appears to pass from the mid- 
dle membrane to a nucleus in the hypodermal layer (Fig. 6). 
In sections such a connection is not always discernable but no 
doubt holds true for a great majority of the strands. Mayer (5) 
says: ‘‘Each of the hypodermis cells gives rise to one, and only 
