In the first place we turn to the 'records' of those rocks, in whose stratified layers[12] are entombed remains, often fragmentary and obscure, of the insects of past ages of the earth's history. Compared with the thousands of extinct types of hard-shelled marine animals, such as the Mollusca, fossil insects are few, as could only be expected, seeing that insects are terrestrial and aerial creatures with slight chance of preservation in sediments formed under water. Yet a number of insect remains are now known to naturalists, who are, in this connection, more particularly indebted to the researches of S.H. Scudder (1885), C. Brongniart (1894), and A. Handlirsch (1906). [12] See Table of Geological Systems, p. 123. We are now considering insects from the standpoint of their life-histories, and the individual life-story of an insect of which we possess but a few fragments of wings or body, entombed in a rock formed possibly before the period of the Coal Measures, can only be a matter of inference. Still it may safely be inferred that when the structure of these remains clearly indicates affinity to some existing order or family, the life-history of the extinct creature must have resembled, on the whole, that of its nearest living allies. And all the fossil insects known can be either referred to existing orders, or shown to indicate definite relationship to some existing group. Passing over some doubtful remains of Silurian age, we find in rocks usually regarded as Devonian[13] the most ancient fossils that can be certainly referred to the insects, while from beds of the succeeding Carboniferous period, a number of insect remains have been disinterred. These Palaeozoic insects were frequently of large size, and they show distinct affinities with our recent may-flies, dragon-flies, stone-flies, and cockroaches. In the Permian period, the latest of the divisions of the Palaeozoic, lived Eugereon, an insect with hemipteroid jaws and orthopteroid wings. All these insects must have been exopterygote in their life-history, if we may trust the indications of affinity furnished by their structure. In the Mesozoic period, however, insects with complete transformations must have been fairly abundant. Rocks of Triassic age have yielded beetles and lacewing-flies, while from among Jurassic fossils specimens have been described as representing most of our existing orders, including Lepidoptera, Hymenoptera and Diptera. In Cainozoic rocks fossil insects of nearly six thousand species have been found, which are easily referable to existing families and often to existing genera. We may conclude then, imperfect though our knowledge of extinct insects is, that some of the most complex of insect life-stories were being worked out before the dawn of the Cainozoic era. Some instructive hints as to differences in the rate of change among different insect groups may be drawn from the study of parasites. For example, V.L. Kellogg (1913) points out that an