156 
DR. R. J. TILLYAHD ON THE 
(3) Es and M are slightly separated basally, but come together so as to be 
almost indistinguishable while passing the callus, and then separate distally. 
(i) Sc arises by itself fi'om the alar trunk ; considerably further down, a 
group of three separate trachea3, Ri, Rs, and M, arise close together. 
All the above conditions are occasional variants only from the dominant 
condition, which is that shown in text-fig. 6. It should be noted that 
trachea M, in this instar, usually arises quite separately from trachea 
Sc + Ri, and its course basally from alar trunk to callus diverges widely from 
that of the latter. The number of tracheae developed from Rs is increased 
to four by the forking of R4+5, and there is an indication of the addition of 
an anterior branch to R2 in the form of a pale band, which never, at any 
time, as far as I could see, develops more than the merest basal rudiment of 
a trachea, and usually possesses none at all. The number of branches of M 
TEXT-na. 7. 
Ameletus ornatus (Eaton). Tracheation of fore-wing in (re4-2)th or antepenultimate 
larval instar. For lettering, see p. 162. (x41.) 
is also increased to three by formation of the trjad Mi, Mj, M3_j,4 ; but the 
tracheal supply of the middle branch Mj is quite irregular, and varies 
greatly in the different larvpe examined. A separate costal trachea appears as 
a short basal anterior branch from Sc. The alar trunk continues complete 
and of fair calibre. 
Passing next to the (?t + 2)thor antepenultimate instar (text-fig. 7), we 
find the conditions not very different from those indicated for the previous 
instar, except for the presence of additional branches of R2, which now 
number five in all, and the establishment of the triad Mj, Mj, M3+4 on a firm 
basis, with a single trachea supplying M2 as shown. In text-fig. 7, I show 
a curious condition of trachea Cui near the base. The course of the vein Cuj 
is indicated hy the dotted line, but the trachea arches up out of its noinial 
course, and passes close under M. I have seen this condition several times 
