cut transversely, while in Pl. 6, fig. 1 the section is mainly horizontal. It is pointed out by Gatenby and Roth (in press) that in the gryllid Nemobius, the number of plates can be accurately counted. For example, there are almost always seventeen in the acroblast in the spermatid of this animal, and ten in the primary spermatocyte dictyosomes.

It is well known from light microscopy that vacuoles or canals, presumably fluid-filled, are usually associated with the Golgi apparatus. That this is so was confirmed by Dalton and Felix (1953) in their well-known study of the cells of the epididymis. The position of the vacuoles with reference to the lamellae is still a matter of opinion. It is now universally admitted that each lamella consists of three layers, upper and lower electron opaque layers enclosing a third inner, electron translucent layer. The inner layer, according to some electron microscopists, is to be regarded as a hollow space, which can be expanded at intervals by vacuoles (VE) as shown in Text-fig. 14. On the other hand there seems to be considerable evidence from electron microscopy, that free vacuoles (VF) can lie both between and beside the piles of lamellae.

This brings one to the question of the significance of the pile of lamellae as a functional unit in the animal cell. It probably will be readily agreed that the pile of lamellae is some kind of filter (Text-fig. 13). That this is so appears very likely from a study of the pile of oriented lamellae with which the formation of the acrosome is associated. In the case of the Gryllidae, the acroblast has a cover on the obverse side, consisting of two presumed sealing layers (C1, C2 in Text-fig. 12), and the acrosome makes its appearance in a nidus formed on the reverse side by the bent over lamellae (Text-fig. 12, A). On the other hand, as will have been noted, the acroblast of Lumbricus (Pl. 7, fig. 5) has no large vacuoles associated with it, and except for the appearance of the acrosome on one side, there is no elaborate cap or cover as occurs in gryllids. In the case of the Lumbricus acroblast, it thus seems difficult to understand how this pile of lamellae could function as a filter. In all cases, however, it seems that the upper and lower layers of the single lamellar plate are formed by oriented molecules, but the relationship of piles of such plates to the ground cytoplasm is today quite obscure. That each plate is independent of the others in the pile can be ascertained by inspection of such micrographs as those in Pl. 6, fig. 2, and Pl. 7, fig. 7, where, in the elongating spermatids, the pile of plates spills over in such a way as to show this independence. One of the clearest demonstrations of the independence of each lamella or saccule of the Golgi apparatus of protozoan cells has been given by Noirot Timothée (1957) in the case of ophryoscolecid infusorians. She has found chains of such lamellae, sometimes more than a dozen in number, all more widely separated than in lumbricids, but still parallel to one another.

It appears to the present author that in the years gone by, studies of the remarkable behaviour of Golgi bodies and mitochondria have tended to divert attention from the apparent fact that the ground cytoplasm is the more important part of the cytosome, and not those bodies floating in it. Presumably the materials which ultimately become segregated within the acrosome (lipolytic and/or proleolytic enzymes) originate in the ground protoplasm and the lamellae of the Golgi apparatus in some way not understood filter and concentrate it.

The resemblance between the lamellar pile of the Golgi apparatus, and the voltaic pile will be apparent to all, except that separate alternate layers have not been found in the former. There is no evidence that the lamellar pile of the Golgi apparatus is either a hydrostatic or natant organ, because the acrosome of Lumbricus is brought up to the head of the elongate nucleus not by the Golgi apparatus but by a special protoplasmic bead (or carrier). So far as is known, the lumbricid dictyosome is the only one described which is free of internal or peripheral large vacuoles; this case makes it difficult to sustain the osmotic hypothesis of Golgi