He made a similar, although less important, series of investigations upon some of the great extinct Crustacea; but, perhaps, his most important palæontological work was done later, after he had been convinced by Darwin of the fact of evolution. In 1855 he had expressed the opinion that the study of fossils was hopeless if one sought in it confirmation of the doctrine of evolution; but five-and-twenty years' continuous work completely reversed his opinion, and in 1881, addressing the British Association at York he declared that "if zoölogists and embryologists had not put forward the theory, it would have been necessary for palæontologists to invent it." In three special groups of animals his study of fossils enabled him to assist in bridging over the gaps between surviving groups of creatures by study of creatures long extinct. He began to study the structure of the Labyrinthodonts, a group of extinct monsters which received their name from the peculiar structure of their teeth. He published elaborate descriptions of Anthracosaurus from the coal-measures of Northumberland, of Loxomma from the lower carboniferous of Scotland, and of several small forms from the coal-measures of Kilkenny, in Ireland, as well as describing skulls from Africa and a number of fragmentary bones from different localities. But in all this work it was the morphology of the creatures that interested him, and the light which their structure threw upon the structure of each other and of their nearest allies. He shewed that these monsters stood on the borderland between fishes, amphibia, and reptiles, and he added much to our knowledge of the true structure of these great groups. Next, he turned to the extinct reptiles of the Mesozoic age. It was generally believed that the Pterodactyls, or flying reptiles, were the nearest allies of birds, but Huxley insisted that the resemblances between the wings were simply such superficial resemblances as necessarily exist in organs adapted to the same purpose. About the same time, Cope in America, and Phillips and Huxley, in England, from study of the bones of the Dinosaurs, another great group of extinct reptiles, declared that these were the nearest in structure to birds. In association with the upright posture, the ilium or great haunch-bone of birds extends far forwards in front of the articulation of the thigh-bone, so that the pelvis in this region has a T-shape, the ilium forming the cross-bar of the T, and the femur or thigh-bone the downward limb. Huxley shewed that a large number of the Dinosaurs had this and other peculiarities of the bird's pelvis, and separated these into a group which he called the "Ornithoscelida," seeing in them the closest representatives of the probable reptilian ancestors of birds. While further work and the discovery of a still greater number of extinct reptiles has made it less probable that these were the actual ancestors of birds, Huxley's work in this, as in the many other cases