With one or two exceptions the yolk cells secreted by the vitellarium retain their vitality till they are swallowed by the embryo, after the development of its mouth. The few not so swallowed become disintegrated. They are granular nucleated cells, and, as was first shewn by von Siebold, are remarkable for exhibiting spontaneous amoeboid movements. Very important light on the nature of the vitellarium is afforded by the structure of the generative organs in Prorhyncus and Macrostomum. In Prorhyncus there is no separate vitellarium, but the lower part of the ovarian tube functionally and morphologically replaces it. The ovum becomes surrounded by yolk cells, which according to Hallez (No. 10) retain their vitality for a long time. According to Ed. van Beneden yolk-spherules are formed in the protoplasm of the ovum itself, in addition to and independently of the surrounding yolk cells. In Convoluta paradoxa a special vitellarium is stated to be absent; though a deposit of yolk is formed round the ovum (Claparède). In Macrostomum again the yolk-glands are at most represented by a lower specialized part of the ovarian tube. The ova in passing down become filled with yolk-spherules. According to Ed. van Beneden these spherules are formed in the protoplasm of the ovum itself; but this is explicitly denied by Hallez, who finds that they are formed from the lining cells of the ovarian tube, which, instead of retaining their vitality as in Prorhyncus, break up and form a granular mass which is absorbed by the protoplasm of the ovum. In Prostomum caledonicum (Ed. van Beneden) the generative organs are formed on the same plan as in other Rhabdocoela, but the cells which form the yolk-gland give rise to yolk particles which enter the ovum, instead of to a layer of yolk cells surrounding the ovum. Amongst the marine dendrocoelous Turbellarians the ova are formed in separate sacks widely distributed in the parenchyma of the body between the alimentary diverticula. In these the ova undergo their complete development, without the intervention of yolk-glands. The ovaries of the Nemertines more nearly resemble those of the marine Dendrocoela than those of the Rhabdocoela. They consist of a series of sacks situated on the two sides of the body between the prolongations of the digestive canal. The eggs are developed in these sacks in a perfectly normal manner, and in many cases become filled with yolk-spherules which arise as differentiations of the protoplasm of the ovum. The protecting membranes of the ova have not been accurately studied. In some cases[21] two membranes are present, an internal and an external. The former, immediately investing the vitellus, is very delicate: the external one is thicker and hyaline. [21] Amphiporus lactiflorius and Nemertes gracilis. McIntosh. _Monograph on British Nemertines._ Ray Society.