Soon after the publication of the Philosophie anatomique (1818) Geoffroy went further in his search for unity, and maintained that the structure of insects and Crustacea could be reduced to the vertebrate type.
He proposed to replace Cuvier's classification of the animal kingdom into the four large groups, Vertebrata, Mollusca, Articulata, and Radiata by the following classification:—[90]
| Vertébrés |  | Hauts-Vertébrés (Vertebrata, Cuv.). |
| Dermo-Vertébrés (Articulata, Cuv.). | ||
| .Invertébrés. | | Mollusques (Mollusca, Cuv.) |
| Rayonnés (Radiata, Cuv.). |
The idea upon which is based the comparison of Articulates with Vertebrates is that each skeletal segment of Articulates is a vertebra. In the Hauts-vertébrés the vertebræ are internal; in the Dermo-vertébrés they are external. "Every animal lives either outside or inside its vertebral column."[91] The essence of a vertebra is not its form, nor its function, but its composition from four elementary pieces which unite round a central space (Isis, loc. cit., p. 532). Serres had shown that in the higher animals every vertebra is formed from four centres of ossification, that the body of the vertebra is at first tubular, and that afterwards it becomes filled up. In lobsters and crabs each segment is composed of four elementary pieces, as may be seen most easily in young ones. "Accordingly each segment corresponds to a true vertebra in composition: there is the same number of 'materials,' the same order in the course of ossification, the same kind of articulation, the same annular arrangement, the same empty space in the middle" (p. 534). The only difference is that in Articulates the central space is very great and contains all the organs of the body, whereas in the higher Vertebrates the body of the vertebra becomes completely filled up. In the thoracic region of Crustacea it is not the whole segment with part of the carapace which corresponds to a vertebra, but merely the part round the ventral nerve-cord (endophragmal skeleton).
Fig. 2.
"Vertebra" of a Pleuronectid.
(After Geoffroy.)
If the skeleton of the segment in Articulates corresponds to the body of a vertebra and is here external, then the appendages of the Articulate must correspond to ribs (p. 538). The full development of this thought is found in a Memoir of 1822, "Sur la vertèbre."[92] He takes as the typical vertebra that of a Pleuronectid, probably the turbot. His original figure is reproduced (Fig. 2).
He includes as part of the vertebra not only the neural (e′, e″) and hæmal (o′, o″) arches, but also, above and below these, the radialia (a″, u′) and the fin-rays (a′, u″). (Neither the radialia nor the fin-rays are, by the way, in the same transverse plane as the body of the vertebra). Every vertebra, he considers, contains these nine pieces—the cycleal (or body), the two perials (e′, e″) and the two epials (a′, a″) above, the two paraals (o′, o″) and the two cataals (u′, u″) below. The epials and the cataals are in reality paired bones which in fish mount one on top of the other to support the median fins. In the cranial region—the skull is formed of modified vertebræ—the epials and perials open out so as to form the walls and roof of the brain; in the thoracic region the paraals and cataals reach their maximum of development and perform the same service for the thoracic organs, the paraals becoming vertebral, and the cataals sternal, ribs.
We have seen that in Arthropods the body of the vertebra (cycleal) forms the open ring of the segment, which lies immediately under the skin, the vertebral tube coinciding with the epidermal tube. The homologues of the other eight pieces of the vertebra must accordingly be sought in the external appendages. At first sight there seems here a contradiction of the principle of connections, for the appendages in Arthropods are lateral, whereas the paired bones of the vertebra are dorsal and ventral. But there is in reality no contradiction, for "what our law of connections absolutely requires is that all organs, whether internal or external, should stand to one another in the same relations; but it is all one whether the box (coffre) that encloses them lies with this or that side on the ground. What similarities in the organisation of man and the digitate mammals, and yet what differences between their attitudes when standing! The same holds true as regards the normal attitudes of the pleuronectids and the other fishes" (p. 107).
The exact way in which Geoffroy homologised the parts of the appendages in Arthropods with the paired pieces of the typical vertebra is best shown by the reproduction of his figure of an abdominal segment of the lobster (Fig. 3), in which the parts homologous with those represented in the figure of the typical vertebra ([Fig. 2]) are indicated by the same letters. The ingenuity of the comparison is astonishing.