122 The Res^ni'atorn Sj/steni of Mallo 2 )h(uja 
more reasonable to conclude that this divergence implies a separate 
origin. 
Packard (1874) gives the primitive number of spiracles in winged 
insects as twelve pairs. Neglecting the Anoplura and Copeognatha, 
although they are pretty certainly entitled to the same interpretation, 
it will be seen that seven of these survive in most Mallophaga, the 
posterior two pairs of the thorax and the anterior two pairs of the abdo¬ 
men having disappeared, vdth, presumably, that of the ninth abdominal 
segment. This allows a generalisation which should be of frequent 
use in determining the number and disposition of the abdominal seg¬ 
ments in Mallophaga, namely that the sixth pair of abdominal spiracles 
occurs on what is morphologically the eighth segment, though in the 
Boopidae, Ricinidae, and Gyroptis of the Amblycera, and in the whole 
of the IscHNOCERA, this segment is apparently the seventh. Again 
certain post-embryonic stages in both Ischnocera and Amblycera 
show ten apparent abdominal segments, ‘which become reduced to an 
apparent nine in the adult. In particular species of Lipenms from 
petrels illustrate this condition. Kellogg (1896) has described as species 
two forms which he calls L. diversus and L. limitatus. The former has 
nine segments, the latter ten. But the latter is the young of the former. 
Similarly Taschenberg (1882) has described L. fidiginosus with nine 
segments, and L. testaceus with ten. The latter is not, in this case, the 
young of the former, but is the immature form of a species of the same 
type, for which the name will stand. In these, as in other examples 
that might be quoted, the last ecdysis changes the three terminal 
segments of the abdomen into an apparent two. The segments in 
question come after the last which bears spiracles, so we can thus state 
that the maximum number of segments in the abdomen of Mallophaga 
is eleven, though this number is not found in any adult. This statement, 
of course, involves the assumption that an anterior segment of the abdo¬ 
men in the Ischnocera has become incorporated with the thorax, or 
suppressed. We may assume this by analogy with what has been 
‘ pointed out for the Amblycera above; but, in addition, the condition 
in Trichodectes offers very strong support. In T. longicornis, for example, 
I find interpolated between metathorax and first abdominal tergite a 
rudimentary tergite which does not reach the lateral margins, but which 
bears a transverse band bordered posteriorly by a row of six hairs. 
In T. divaricatus a similar rudimentary tergite is very surely present, 
but is uncoloured and without hairs. In T. setosvs, on the other hand, 
there is no trace of this tergite. 
