These larvae are at first very different from the adult animal, and the necessary changes to be passed through are more or less great according to the species. It is not our intention to describe these changes in detail.[329] The larva increases in size, the permanent chaetae make their appearance in regular order, and the body exhibits segmentation, the new segments always appearing just in front of the anal segment. The internal organs gradually develop, and the prostomial and parapodial appendages grow out in their turn. In the Sabelliformia the multifilamentous "gills" arise by the continued branching of an at first simple process (the palp) arising from the latero-ventral surface of each side of the preoral lobe.[330] These gradually encroach dorsally and ventrally till the prostomium is more or less encircled; meanwhile the peristomium grows forwards so as to conceal the prostomium, which no longer increases at the same rate as does the rest of the body. Although most worms appear to discharge their ova directly into the sea and take no further care of them, some make provision for their offspring either by laying the eggs in a jelly, which will serve as food for the young larvae—_Aricia_, _Ophelia_, _Protula_, _Phyllodoce_—or by attaching them to their body. In certain Polynoids the eggs are attached by means of a secretion to the back, under the elytra, where they undergo development up to a certain stage. In _Exogone_ and some other Syllids they are attached to the ventral cirri, or in _Grubea limbata_, all over the back. In the female _Autolytus_ (_Sacconereis_) a ventrally-placed brood sac is formed by the hardening of a {276}secretion; the eggs develop into embryos inside the brood sac, and then become free, with head appendages and three pairs of parapodia. Enormous numbers of such embryos may occur; for instance, some 300 were counted in a brood sac of _Autolytus ebiensis_. In the case of tubicolous worms, the eggs are frequently attached to the tube, either inside or outside. In _Spirorbis_ and _Salmacina_ the operculum serves as a brood pouch. Only a very few species are known to be viviparous, viz. _Syllis vivipara_ Kr., _Cirratulus chrysoderma_ Clap., _Marphysa sanguinea_ Mont., and _Nereis diversicolor_ Müll.