18 REPORT—1896. 
the vast period between the Upper Carboniferous rocks and the present day the cock- 
roaches have gained a rather different wing venation, and have succeeded in laying 
their eggs in a manner rather more specialised than that of insects in general; the 
stick insects and leaf insects have lost or reduced their wings, the grasshoppers 
have shortened their antenne. These, however, are the insects which most closely 
resemble the existing species ; let us turn to the forms which exhibit the greatest 
differences. Many species have retained in the adult state characters which are 
now confined to the larval stage of existence, such as the presence of tracheal gills 
on the sides of the abdomen. In some, the two membranes of the wing were not 
firmly fixed together, so that the blood could circulate freely between them. On 
the other hand, they are not very firmly fixed together in existing insects. Another 
important point was the condition of the three thoracic segments, which were quite 
distinct and separate, instead of being fused as they are now in the imago stage. 
This external difference probably also extended to the nervous system, so that the 
thoracic ganglia were separate instead of concentrated. The most interesting 
distinction, however, was the possession by many species of a pair of prothoracie 
appendages much resembling miniature wings, and which especially suggest the 
appearance assumed by the anterior pair (tegmina) in existing Phasmide. There 
is some evidence in favour of the view that they were articulated, and they exhibit 
what appears to be a trace of venation. Brongniart concludes that in still earlier 
strata, insects with six wings will be discovered, or rather insects with six of the 
tracheal gills sufficiently developed to serve as parachutes. Of these, the two 
posterior pair developed into the wings as we know them, while the anterior pair 
degenerated, some of the Carboniferous insects presenting us with a stage in 
which degeneration had taken place but was not complete. 
One very important character was, as I have already pointed out, the enormous 
size reached by insects in this distant period. This was true of the whole known 
fauna as compared with existing species, but it was especially the case with the 
Protodonata, some of these giant dragon-flies measuring over two feet in the 
expanse of the wings. 
As regards the habits of life and metamorphoses, Brongniart concludes that 
some species of Protoephemeridz, Protoperlide, &c., obtained their food in an 
aquatic larval stage, and did not require it when mature. He concludes that 
the Protodonata fed on other animals, like our dragon-flies; that the Paleeacridiides 
were herbivorous like our locusts and grasshoppers. the Protolocustide herbivorous 
and animal feeders like our green grasshoppers, the Paleoblattidze omnivorous 
like our cockroaches. The Homoptera, too, had elongated sucking mouth-parts 
like the existing species. It is known that in Carboniferous times there was a lake 
with rivers entering it, at Commentry. From their great resemblance to living 
forms of known habits, it is probable that the majority of these insects lived near 
the water and their larve in it. 
When we look at this most important piece of research as a whole, we cannot 
fail to be struck with the small advance in insect structure which has taken place 
since Carboniferous times. All the great questions of metamorphosis, and of the 
structures peculiar to insects, appear to have been very much in the position 
in which they are to-day. It is indeed probable enough that the Orders which 
zoologists have always recognised as comparatively modern and specialised, such 
as the Lepidoptera, Coleoptera, and Hymenoptera, had not come into existence. 
But as regards the emergence of the Class from a single primitive group, as regards 
its approximation towards the Myriapods, which lived at the same time, and of 
both towards their ancestor Peripatus, we learn absolutely nothing. All we can 
say is that there is evidence for the evolution of the most modern and specialised 
members of the Class, and some slight progressive evolution in the rest. Such evo- 
lution is of importance as giving us some vague conception of the rate at which the 
process travels in this division of the Arthropoda. If we look upon development as 
a series of paths which, by successively uniting, at length meet in a common point, 
then some conception of the position of that distant centre may be gained by 
measuring the angle of divergence and finding the number of unions which occur 
in a given length. In this case, the amount of approximation and union shown in 
