The ventral cords have begun to be formed as thickenings of the epiblast, and the limbs are established. _me.s._ mesoblastic somite; _vn._ ventral nerve-cord; _yk._ yolk.] A similar mode of formation of both the ventral cords and the supraoesophageal ganglia obtains in Insects (fig. 245). The ventral cords are however much less widely separated than in Spiders, and early unite in the median line. In the supraoesophageal ganglia the invaginated epiblast has in Lepidoptera (Hatschek) the form of a pit on the dorsal border of the antennæ. Hatschek states that there takes place an invagination of a median part of the skin between the two ventral cords, for the details of which I must refer the reader to Vol. II. p. 410. He has made more or less similar statements for the earthworm, but his observations in both instances are open to serious doubt. [FIG. 244. SECTION THROUGH THE PROCEPHALIC LOBES OF AN EMBRYO OF AGELENA LABYRINTHICA. _st._ stomadæum; _gr._ section through semicircular groove in procephalic lobe; _ce.s._ cephalic section of body cavity.] [FIG. 245. TWO TRANSVERSE SECTIONS THROUGH THE EMBRYO OF HYDROPHILUS. (After Kowalevsky.) A. Transverse section through an embryo in the region of one of the stigmata. B. Transverse section through an older embryo. _vn._ ventral nerve-cord; _am._ amnion and serous membrane; _me._ mesoblast; _me.s._ somatic mesoblast; _hy._ hypoblast (?); _yk._ yolk-cells (true hypoblast); _st._ stigma of trachea.] Full details as to the development of the nervous system in the Crustacea are still wanting; a fairly complete account of what is known on the subject is given in Vol. II. pp. 521-2. It appears that the ventral cord may either arise as an unpaired thickening of the epiblast (Isopoda), marked however by a shallow median furrow, or from two cords which eventually coalesce[154]. It is not certain how far the supraoesophageal ganglia are usually in the first instance continuous with the ventral cord. In Astacus, the early stages of which have been elaborately investigated by Reichenbach (No. 331), they are stated to be so; the supraoesophageal ganglia are moreover described by this author as having a somewhat complicated origin. Five elements enter into their composition. There is first formed a pair of pits on the procephalic lobes, which become very deep during the Nauplius stage, and are continuous with a pair of epiblastic ridges which pass round the mouth, and join the ventral cords just described. The walls of the pits are believed to form a part of the embryonic ganglia which gives rise to the retina as well as to the optic ganglia. The ridges form the remainder of the ganglia and the oesophageal commissures; while the fifth element is supplied by a median invagination in front of the mouth, which appears at a much later date than the other parts.