XIII THE GEOLOGICAL EVIDENCES OF EVOLUTION 405 



from the Middle Silurian sandstone of France, If this 

 occurrence of a true hexapod insect from the Middle Silurian 

 be really established, taken in connection with the well- 

 defined Coleoptera from the Carboniferous, the origin of the 

 entire group of terrestrial arthropoda is necessarily thrown 

 back into the Cambrian epoch, if not earlier. And this cannot 

 be considered improbable in view of the highly differentiated 

 land plants — ferns, equisetums, and lyco^^ods — in the Middle or 

 Lower Silurian, and even a conifer (Cordaites Robbii) in the 

 Upper Silurian ; while the beds of graphite in the Laurentian 

 were probably formed from terrestrial vegetation. 



On the whole, then, we may affirm that, although the 

 geological record of the insect life of the earth is exceptionally 

 imperfect, it yet decidedly supports the evolution hypothesis. 

 The most specialised order, Lepidoptera, is the most recent, 

 only dating back to the Oolite ; the Hymenoptera, Diptera, 

 and Homoptera go as far as the Lias ; while the Orthoptera 

 and Neuroptera extend to the Trias. The recent discovery of 

 Coleoptera in the Carboniferous shows, however, that the 

 preceding limits are not absolute, and will probably soon be 

 overpassed. Only the more generalised ancestral forms of 

 A^dnged insects have been traced back to Silurian time, and 

 along with them the less highly organised scorpions ; facts 

 which serve to show us the extreme imperfection of our 

 knowledge, and indicate possibilities of a world of terrestrial 

 life in the remotest Palaeozoic times. 



Geological Succession of Vertebra ta. 



The lowest forms of vertebrates are the fishes, and these appear 

 first in the geological record in the Upper Silurian formation. 

 The most ancient known fish is a Pteraspis, one of the buck- 

 lered ganoids or plated fishes — by no means a very low type 

 — allied to the sturgeon (Accipenser) and alligator - gar 

 (Lepidosteus), but, as a grouj^, now nearly extinct. Almost 

 equally ancient are the sharks, which under A^arious forms 

 still abound in our seas. We cannot suppose these to be nearly 

 the earliest fishes, especially as the two lowest orders, now 

 represented by the Amphioxus or lancelet and the lampreys, 

 have not yet been found fossil. The ganoids were greatly 

 developed in the Devonian era, and continued till the 



