ambulacrals just mentioned, and sometimes by the second set of ambulacrals also; the ambulacra end in the five oculars or terminals, which meet in a ring around the anal area and have no podial pores, but one of them serves as a madreporite; within this ring is a star-shaped area filled with minute irregular plates, none of which can safely be selected as the homologues of the so-called basals or genitals of later forms; within the ring of ambulacrals around the mouth are five somewhat pointed plates, which Jaekel regards as teeth, but which can scarcely be homologous with the interradially placed teeth of later echinoids, since they are radial in position; small spines are present, especially around the podial pores. The position of the pores near the centre of the ambulacrals in _Bothriocidaris_ need not be regarded as primitive, since other early Palaeozoic genera, not to mention the young of living forms, show that the podia originally passed out between the plates, and were only gradually surrounded by their substance; thus the original structure of the echinoid ambulacra differed from that of the early asteroid in the position of the radial vessels and nerves, which here lie beneath the plates instead of outside them. To this point we shall recur; palaeontology, though it suggests a clue, does not furnish an actual link either between Echinoidea and Asteroidea, or between those classes and Pelmatozoa. [Illustration: FIG. 5.--_Bothriocidaris globulus._ A, from the side; B, the plates around the aboral pole. (After Jaekel.) The short spines which were attached to the tubercles are not drawn.] The argument from embryology leads further back. First, as already mentioned, it outlines the general features of the _Dipleurula_; secondly, it indicates the way in which this free-moving form became fixed, and how its internal organs were modified in consequence; but when we seek, thirdly, for light on the relations of the classes, we find the features of the adult coming in so rapidly that such intermediate stages as may have existed are either squeezed out or profoundly modified. The difficulty of rearing the larvae in an aquarium towards the close of the metamorphosis may account for the slight information available concerning the stages that immediately follow the embryonic. Another difficulty is due to the fact that the types studied, and especially the crinoid _Antedon_, are highly specialized, so that some of the embryonic features are not really primitive as regards the class, but only as regards each particular genus. Thus inferences from embryonic development need to be checked by palaeontology, and supplemented by comparison of the anatomy of other living genera.