During the above changes the nuclei of the vitelline spheres pass to the surface and rapidly divide. Eventually, together with part of the protoplasm of the vitelline spheres, they appear to give rise to a layer of hypoblastic cells. This layer encloses the remains of the vitelline spheres, which become the yolk. [FIG. 159. TWO EMBRYOS OF CLEPSINE IN WHICH THE GERMINAL STREAKS HAVE PARTIALLY MET ALONG THE VENTRAL LINE. (After Robin.) _gs._ germinal, _i.e._ mesoblastic streaks. The area covered by epiblast is shaded. The so-called neuroblasts at the end of the germinal streaks are shewn in B.] At the front end of the germinal streaks, in a position corresponding with that of the four original epiblast cells, two depressions appear which coalesce to form the single oral invagination; in the centre of which are formed the mouth and pharynx by a second epiblastic invagination. The most important point in connection with the above history is the fate of what have been called the germinal streaks. According to Whitman they are composed of two kinds of cells, viz. four rows of smaller superficial cells, which he calls neuroblasts, and, in the later stages at any rate, a row of deeper large cells, which he calls mesoblasts. As to the eventual fate of these cells he states that the neuroblasts uniting together in the median line form the rudiment of the ventral ganglionic chain, while the mesoblasts equally coalesce and give rise to the mesoblast. Such a mode of origin for a ventral ganglionic chain is, so far as I know, without a parallel in the whole animal kingdom; and whatever evidence Whitman may have that the cells in question really do give rise to the nervous system he has not thought fit to produce it in his paper. He figures a section with the eight neuroblastic cells in the middle ventral line, and in the next stage described the nervous system is divided up into ganglia! The first stage, in which the so-called nervous system has the form of a single row of eight cells, is quite unlike any rudiment of the nervous system such as is usually met with in the Chætopoda, and not a single stage between this and a ganglionated cord is described or figured. Whitman, whose views seem to have been influenced by a peculiar, and in my opinion erroneous, theory of Rauber's about the relation of the neural groove of Vertebrata to the blastopore, does not seem to be aware that his determination of the fate of his neuroblasts requires any special support.