$ 330. THE INSECTA. 411 
while those of the others are usually fused together. During the pupa- 
state, a remarkable change takes place. The commissures between the 
first and second, and the third and fourth ganglia, are gradually shortened. 
The ganglia are thereby gradually approximated, and, in the end, are fused’ 
together, forming the two thoracic ganglia of the adult, which send off 
nerves to the legs, and to the muscles of the wings. At the same time, the 
fifth and sixth ganglia entirely disappear or are fused into one.‘ 
With the Hymenoptera, the ventral cord is composed of seven to eight 
ganglia connected by double commissures. The first of these, smaller 
than the second, is, like it, produced by the fusion of several ganglia ; and 
both are situated in the thorax. Of the remaining five or six abdominal 
ganglia, the last two are closely approximated, or fused into one.“ Here, 
as with the Lepidoptera, the number of ganglia in the ventral cord of the 
larvae, is eleven, as has been specially shown in the false caterpillars of the 
Tenthredinidae. 
With the Orthoptera, and Neuroptera, the nervous system is nearly 
always composed, in their various states, of three thoracic and six to seven 
abdominal ganglia connected by double commissures and forming a chain 
as long as the body.“ 
With the Coleoptera, the number and disposition of the ventral ganglia 
present the widest variations of all. The longitudinal commissures, always 
double, are shortened or even wholly wanting at certain points. 
The 
ganglionic chain is, therefore, more or less abbreviated, and sometimes the 
ganglia are almost fused into a single mass. 
In this respect this system 
here presents two principal types, the limits between which, however, have 
18 For the nervous system of the larvae of Va- 
nessa urticae and Bombyx mori, see the figures of 
Swammerdamm,|oc. cit. p. 387, 230, Taf. XXVIII. 
fig. 3, and Taf. XXXIV. fig. 7 ; also for that of the 
larva and imago of Cossus ligniperda, the works of 
Lyonet, Traité, &c., p. 190, Pl. IX., and in the 
Mém. du Maus. loc. cit. p. 191, Pl. LI. (17). For 
that of Gastropacha pini, pupa and imago, 
see Suckow, Anat. physiol. Untersuch. p. 40, 
Taf. VII. fig. 37, 38 ; but see especially the excel- 
lent description of that of the larva, pupa, and 
imago of Sphinx ligustri, for which we must 
thank Newport. Philos. Trans. 1832, p. 383, Pl. 
XII. XTII.; also, 1834, p. 389, Pl. XILL.—-XVIIL., 
and Cyclop. &c. loc. cit. p. 943, fig. 406, 414, 415. 
14 This metamorphosis of the nervous system 
was first observed by Herold (Entwickelungsgesch. 
d. Schmetterlinge, loc. cit. Taf. II.) with Pontia 
brassicae, and has since been confirmed by New- 
port with Sphinx ligustri and Vanessa urticae ; 
see Philos. Trans. 1834, Pl. XV. XVI. fig. 20-30, 
and Cyclop. loc. cit. p. 962, fig. 420-423. 
- 18 See Swammerdamm, Bib. der Nat. p. 207, 
Taf. XXII. fig. 6 (Apis mellifica) ; Treviranus, 
Biologie, ¥. Taf. I. (Bombus muscorum), and 
Brandt and Ratzeburg, Medizin. Zool. II. p. 203, 
Taf. XXV. fig. 31 (Apis mellifica). For the dis- 
position of the ventral chain of the Lepidoptera in 
general, see, moreover, ZL. Dufour, Recherch. sur 
les Orthopt., Hymenopt. &c. p. 381. According to 
this last-mentioned naturalist, the number of ventral 
ganglia is five with Vespa, Scolia, and with most of 
the Apidae and Andrenidae ; six with Odynerus, 
+a. 
Bembecidae, with Larra, and Tiplica ; four with 
Tripoxylon, and three with Eucera. 
16 The ventral chain of the Tenthredinidae, 
Apidae, Vespidae and other Hymenoptera, under- 
goes, undoubtedly, with the pupae, a metamorpho- 
sis similar to that occuring with the Lepidoptera. 
W Swammerdamm, loc. cit. p. 108, Taf. XIV. 
(a pupa of Ephemera) ; Marcel de Serres, Mém. 
du Mus. IV. 1818, Pl. VIII. (1.) fig. 1 (Acridium); 
J. Miller, Nov. Act. Nat. Cur. XIV. Tab, IX. 
fig. 4, and XII. p. 568, Tab. L. fig. 1 (Acridium 
and Bacteria); Newport, Cyclop. II. p. 950, fig. 
409, 410 (Forficula and Locusta) ; finally, L. Du- 
our, Ann. d. Sc. Nat. XIII. 1828, p. 361, Pl. 
XII. fig. 4 (Rorfcula), Recherch. sur les Or- 
thopt. &c. p. 281, Pl. Il. fig. 7 (Oedipoda), and p. 
561, Pl. XI. fig. 160 (Libellula). According to 
L. Dufour, there are seven ventral ganglia with 
Libellula and Ephemera, while there are only six 
with Perla and Phryganea. But Pictet (Re- 
cherch. pour servir a Phist. et 4 Panat. des Phry- 
ganides. Pi. II. fig. 33-36) and Burmeister 
(Handb. &c. I. p. 895, 898) assign to these In- 
secta, in both their larva and their perfect state, 
eight ventral ganglia. There are even nine of 
these ganglia with the Ephemeridae, according to 
Burmeister (loc. cit. p. 763). In the very chub- 
bed larvae of Myrmeleon there are eight contigu- 
ous ventral ganglia beside two thoracic ones (Cu- 
vier, Lecons, &c., III. p. 841). Loew (Germar’s 
Zeitsch. LV. p. 424) remarks that the proper Neu- 
roptera are distinguished by the separation of their 
last two abdominal ganglia, while, with all the Or- 
Sphex, Pompilus, Chrysis, the I:h . 
* [ § 830, note 17.) See also for a description 
and figures of the nervous system of Spectrum 
. 
thoptera, they are fused together.* 
femoratum, Leidy, Proceed. Acad. Sc. Philad. 
1846, III. p. 83. — Ep. 
