504 MODERN CLASSIFICATION OF INSECTS. 
nearly to the end of the order with the Muscidz, whereas the structure 
of the antenne of these two groups is totally at variance with such an 
arrangement. In the Familles Naturelles, we find the preference given 
to the nature of the metamorphoses, and hence the order (after the 
removal of the Pupipar) is divided into four primary families, thus 
arranged: 1. Nemocera, 2. Tanystoma, 3. Notacantha, and 4. Athe- 
ricera; the 3d, composed of the Stratiomide, agreeing with the 4th 
in having a coarctate pupa, whilst the 1st and 2d have the pupa not 
enclosed within the skin of the larva. 
Still, however, there is a diversity in the nature of the metamor- 
phoses of the third and fourth groups ; the larva skin of the Stratiomide 
nearly retaining its form, whilst that of Musca and the other Athericera 
is completely altered, assuming that of an oval inarticulate mass. “ La 
nature,” observes Latreille, “ d’aprés ses vues particuliéres, et qui 
nous sont inconnues, varie sa marche et ses combinaisons.” ( Consid. 
gén. p. 85.) And hence it is, that whilst Culex and Chironomus are 
so intimately allied in general structure and metamorphoses, as to be 
united into the same natural family by some writers *, the structure 
of their mouths is totally different, that of Culex being exceedingly 
developed, even to a greater extent than in the Tabanide, with 
which, however, it agrees in the number of the pieces of which it is 
composed ; an arrangement, therefore, founded on the structure of the 
mouth, in which Tabanus and Culex would be united, must be arti- 
ficial. How far the circumstance of some of these insects possessing 
a coarctate pupa ought to be considered as superseding other consi- 
derations, originating in the characters of the perfect state, may per- 
haps be questioned, when we consider that a coarctate pupa is in fact 
nothing else than an imperfect pupa enveloped in the larva skin 4, 
far more analogous to the true zzcomplete pupa of the bee and the 
beetle, than it is to the obtected pupa of the Lepidoptera. 
If, again, independently of metamorphoses or trophi, we look to 
the antenne, and other characters of the imago, for affording the 
grounds of a natural arrangement, we still find it difficult to decide 
whether the Tabanide or the Notacantha are nearest related to the 
Tipulide, the former group by Macquart, and the latter by Latreille, 
being made to occupy this station. By the arrangement of Macquart, 
* Latreille (Genera, tom. iv. p. 245.) unites them into a section under the name 
of Tipulariz aquatice. 
+ In some species of beetles, the incomplete pupa is inclosed in the skin of the 
larva (Anthrenus, Chilocorus), thus being, in fact, coarctate pup. 
