The development of plants, like the development of animals, is essentially a progress from the general to the special (p. 242). Botanists have not been troubled by any recapitulation theory, and in founding their big groups, Acotyledons, Monocotyledons, and Dicotyledons, upon embryological characters, they were guided by true principles, which ought indeed to be followed in zoology. If we knew the development of all kinds of animals sufficiently well, then the best way to classify them would be according to the characters they show in their early development, for it is in early development that they show the characters of the type in their most generalised form. As it is, we have in our ignorance to establish the big groups by the study of adult structure, but we find, on putting together all we know of comparative embryology, that a classification of animals according to the mode of their development gives, as is only natural, the same four groups as does the study of adult structure. The four types of development are thus:-- (1) The double-symmetrical, which is found in Vertebrates. It is called the double-symmetrical, because in Vertebrates development takes place from a central axis (notochord) in two directions, upwards and downwards, in such a way that two tubes are formed, one above and one below the axis. (2) The second type is the symmetrical, which is shown by Annulates. A primitive streak is formed on the ventral surface of the yolk; development proceeds symmetrically on both sides of the streak. (3) Radiate development is probably typical of the radiate structural type. (4) In the massive type, the development seems to be a spiral one. Common to most modes is a separation of the germ into animal and plastic layers, a separation which seems to be conditioned largely by the presence of yolk. A classification based upon embryological characters ought to be applied even to the lesser groups and would here prove itself of service. Embryology, for instance, fully supports de Blainville's separation of Batrachia from true reptiles,[176] for reptiles develop an amnion and Batrachia do not.