The next step, as we trace our human phylogeny to its origin, leads us further back into the lower Vertebrata, into that obscure Palæozoic age the immeasurable length of which (much greater than that of the Mesozoic) may, according to one of the newest geological calculations, have comprised about one thousand millions of years.[16]

The first important fact we have to face here is the complete absence of mammalian remains. Instead of these we find in the later Palæozoic period, the Permian, air-breathing reptiles as the earliest representatives of Amniota. They belong to the most primitive order of that class, the Tocosauria; and besides them there were the Theromorpha, which approach the Mammalia in a remarkable manner. These reptiles in turn were preceded, in the Carboniferous period, by true Amphibia, most of them belonging to the armour-clad Stegocephali. These interesting Progonamphibia were the oldest Tetrapoda, the first vertebrates which had adapted themselves to the terrestrial mode of life; in them the swimming fin of fishes and Dipneusta was transformed into the pentadactyle extremities characteristic of quadrupeds.

To appreciate the high importance of this metamorphosis, we need only compare the skeleton of our own human limbs with that of the living Amphibia. We find in the latter the same characteristic composition as in man: the same shoulder and pelvic girdle; the same single bone, the humerus or the femur, followed by the same pair of bones in the forearm and leg; then the same skeletal elements composing the wrist and the ankle regions; and, lastly, the same five fingers and toes.

The arrangement of these bones, peculiar and often complicated, but everywhere essentially the same in all the Tetrapoda, is a striking evidence that man is a descendant from the oldest pentadactyle Amphibia of the Carboniferous period. In man the pentadactyle type has been better preserved by constant heredity than in many other Mammalia, notably the Ungulata.

The oldest Carboniferous Amphibia, the armour-clad Stegocephali, and especially the remarkable Branchiosauri discovered by Credner, are now regarded by all competent zoologists as the indubitable common ancestral group of all Tetrapoda, comprising both Amphibia and Amniota. But whence this most remote group of Tetrapoda? That difficult question is answered by the marvellous progress of modern palæontology, and the answer is in complete harmony with the older results arrived at by comparative anatomy and ontogeny. Thirty-four years ago Carl Gegenbaur,[17] the great living master of comparative anatomy, had demonstrated in a series of works how the skeletal parts of the various classes of Vertebrata, especially the skull and the limbs, still represent a continuous scale of phyletic gradations. Apart from the Cyclostomes, there are the fishes, and among them the Elasmobranchi (sharks and rays), which have best preserved the original structure in all its essential parts of organization. Closely connected with the Elasmobranchi are the Crossopterygii, and with these the Dipneusta or Dipnoi. Among the latter the highest importance attaches to the ancient Australian Ceratodus. Its organization and development is now, at last, becoming well known. This transitional group of Dipnoi, 'fishes with lungs' but without pentadactyle limbs, is the morphological bridge which joins the Ganoids and the oldest Amphibia. With this chain of successive groups of Vertebrata, constructed anatomically, the palæontological facts agree most satisfactorily. Selachians and Ganoids existed in the Silurian times, Dipnoi in the Devonian, Amphibia in the Carboniferous, Reptilia in the Permian, Mammalia in the Trias. These are historical facts of first rank. They connote in the most convincing manner that remarkable ascending scale in the series of vertebrates for our knowledge of which we are indebted to the works of Cuvier and Blainville, Meckel, Johannes Mueller and Gegenbaur, Owen and Huxley. The historical succession of the classes and orders of the Vertebrata in the course of untold millions of years is definitely fixed by the concordance of those leading works, and this invaluable acquisition is much more important for the foundation of our human pedigree than would be a complete series of all possible skeletons of Primates.

Greater and more frequent difficulties arise if we penetrate further into the most remote part of the human phylogeny, and attempt to derive the vertebrate stem from an older stem of invertebrate ancestors. None of those had a skeleton which could be petrified; and the same remark applies to the lowest classes of Vertebrata—to the Cyclostomes and the Acrania. Palæontology, therefore, can tell us nothing about them; and we are limited to the other two great documents of phylogeny—the results of comparative anatomy and ontogeny. The value of their evidence is, however, so great that every competent zoologist can perceive the most important features of the most remote portion of our phylogeny.

Here the first place belongs to the invaluable results which modern comparative ontogeny has gained by the aid of the biogenetic law or the theory of recapitulation. The foundation-stones of vertebrate embryology had been laid by the works of Von Baer, Bischoff,[18] Remak, and Koelliker;[19] but the clearest light was thrown upon it by the famous discoveries of Kowalevsky[20] in 1866. He proved the identity of the first developmental stages of Amphioxus and the Ascidians, and thereby confirmed the divination of Goodsir, who had already announced the close affinity of Vertebrates and Tunicates. The acknowledgment of this affinity has proved of increasing importance, and has abolished the erroneous hypothesis that the Vertebrata may have arisen from Annelids or from other Articulata. Meanwhile, from 1860 to 1872, I myself had been studying the development of the Spongiæ, Medusæ, Siphonophora, and other Cœlenterata. Their comparison led me to the statements embodied in the 'Gastræatheorie,' the first abstract of which was published in 1872 in my monograph of the Calcispongiæ.

These ideas were carried on and expanded during the subsequent ten years by the help of many excellent embryologists—first of all by E. Ray Lankester and Francis Balfour. The most fruitful result of these widely extended researches was the conclusion that the first stages of embryonic development are essentially the same in all the different Metazoa, and that we may derive from these facts certain views on the common descent of all from one ancestral form. The unicellular egg