30 MODERN CLASSIFICATION OF INSECTS. 
particularly by the possession of a row of thin plates on each side 
of the abdomen, ordinarily united in pairs by their bases, and which 
are a species of external false branchiz or gills, in which the tra- 
chee are extended, and ramify ; thus serving as organs of respiration, 
as well as assisting in locomotion by their constant undulatory mo- 
tion: the abdomen in the larva is terminated by three sete, which 
is the case, not only in those species which have three filaments 
in the imago, but also in those with only two tails. I have observed 
that these setae acquire a greater length at each moulting. 
Those species which reside in burrows seldom quit their retreats ; 
whilst the smaller species, which live at large in the water, are much 
more active, and have the body of a firmer consistence than the 
others. The pupa differs only from the Jarva in having the rudi- 
mental wing-covers more conspicuous at the sides of the meso- and 
meta-thorax. 
The investigation of the preparatory stages of the different species 
of these insects, will be sufficient to prove the necessity of their 
separation into even more genera than have hitherto been proposed 
for them. The pupe of several species are represented by De 
Geer and some other authors: the larva, however, are not figured, 
but we may consider them as similar in character to the pupz, from 
which they differ only in the absence of rudimental wing-covers. In 
the species to which the generic name has been restricted by recent 
authors (E. vulgata, &c.), the pupa (my jig. 61. 5., and De Geer, 
tom. xxi. tab. 16.), is distinguished by a transverse-quadrate prothorax 
as broad as the head, a very gibbous meso-thorax, a head of rather 
small size, with two short horns in front, and two long, acute, slightly 
recurved mandibles, originating at the sides of the mouth, and being 
as long as the head (fig. 61.6. head sideways). Considering the rudi- 
mental nature of the mouth of the imago, it is surprising that no 
one has hitherto described the real structure of the mouth, in the 
preparatory states. Reaumur has attempted it, but his figures are 
so rude and insufficient, that no idea can be gleaned as to their 
true structure; Swammerdam, also, passes them over undescribed. 
In the pupa of E. vulgata, the upper lip is of moderate size, with 
the anterior angles rounded off, and ciliated ; it is flat, and quite 
membranous (/ig.61.7.) ; the mandibles (/ig.61.8.) are horny, armed 
with several teeth within, near the base (fig. 61. 9.), which is dilated 
into a flattened molary plate ; whilst the upper angle of the mandible 
