B. Deeper layer of ectoderm, composed of ganglion cells, young cnidoblasts and muscle cells. A flattened ganglion cell, g, is figured with its nucleus and outrunners. Von Lendenfeld did not observe any anastomosis of these fibrils in Eucopella. There are ten ganglion cells in the proximal portion of the tentacles in this hydroid, but none in the distal. Between the ganglion cell and the edge of the covering cell-layer is a young cnidoblast. The muscle cells, m, form a continuous layer applied to the supporting lamina, and this continuous layer is absent only at the tip of the tentacle. They are here deeply placed, i. e. sub-epithelially, as they are in some other Hydroids (Plumularia, Eudendrium) not superficially as in Hydra and others. Each cell consists of a small quantity of protoplasm surrounding the nucleus, and a long contractile fibre pointed at each end. The pointed ends are shown turned up in the figure. There are many nuclei scattered about, and the filaments are all disposed longitudinally, as is the case in the tentacles of most Hydrozoa.

C. The supporting lamina between the ectoderm and endoderm, and parts of two endoderm cells. The lamina is very thin: not so the cell-wall of the endoderm cells, which is thick and resistent, and shown in the diagram as dark lines. The cells are arranged in a single linear series as they generally are in a solid tentacle. At the proximal end they abut against the gastric endoderm, from which they are in the first instance derived. Their protoplasm is much vacuolated, and surrounds the nuclei in a little mass, from which radiate towards the cell-walls irregular strands. A very thin layer covers the walls internally. The nuclei are disposed in a linear series in the axis of the tentacle. A large fat globule lies above the nucleus in the figure. This is sometimes broken up into 2-3 smaller drops.

FIG. 8. A portion of the head of a Guard-polype from a Plumularia, showing a sense-cell, supporting cell, and cnidoblast in relation with a ganglion cell; from Von Lendenfeld, Z. W. Z. xxxviii. 1883, P1. xviii. Fig. 3 (cf. A. N. H. (5) xii. 1883).

g. Ganglion cell, such as have been detected in the ectoderm of many Hydromedusae including Siphonophora. It has a nucleus, granular protoplasm and outrunners.

s'. Sense-cell. The body of the cell is slender, nucleated, and bears at its outer extremity a cilium, and at its inner it is prolonged into a slender filament. The body of the cell is sometimes very granular. The basal process has been traced into continuity with a ganglion cell. Cf. Z. W. Z. vol. cited, P1. xxix. Fig. 8.

s. Covering or supporting cell. It bears a cilium, and its outer extremity is broad. The body of the cell is stouter than in s', but the basal extremity is elongated. In many instances it is broad, and rests on the supporting lamina.

M. Muscle cell. The protoplasm with nucleus is on one side, the contractile filament (dark in the diagram) on the other.

C. Cnidoblast containing an undischarged nematocyst. The outer extremity of the cell bears on the side opposite the nucleus a pointed process, sensory in nature, the cnidocil. This cnidocil has in some instances been observed to have a complex structure and to contain an axial simple or basally trifid filament. The cell body contains a long oval nematocyst, granular protoplasm, and a nucleus. Its basal portion is attenuated, and is in relation with processes of the ganglion cell. Actual continuity of the two structures has been observed by Jickeli (op. cit ante, under Hydra), and by von Lendenfeld in Cyanea Anaskala (Z. W. Z. xxxvii. 1882, pp. 480, 513). This connection would explain the voluntary control which many of these animals appear to exert over the discharge of the nematocyst. It does not, however, exist in all instances.

There is some dispute over the nature of the basal process of the cnidoblast. In some instances it is hyaline, and Hamann believes that it is attached to the supporting lamina, and is purely a supporting structure itself. Chun has observed that in the case of some of the cnidoblasts in Physalia it is transversely striated and the protoplasm of the cell itself contains a network of striated filaments. Jickeli (M. J. viii. 1883, p. 393) appears to be of an opinion that the cnidoblasts possess muscular processes. Von Lendenfeld points out, in the paper from which this figure is taken, that the basal processes of cnidoblasts in Medusae and Anthozoa are granular; that in Crambessa mosaica he has observed the discharge of nemato-cysts contained in cnidoblasts within the jelly when acetic acid was applied to the olfactory epithelium; that in arenicolous Actiniae contact of sand will not cause discharge of the nematocysts, but that contact of the prey does so at once. These facts point to some nervous control.

The discharge of the thread is therefore brought about through pressure exercised on the sac of the nematocyst, by contraction of the cell body called out in response (1) to direct mechanical or chemical irritation of the cnidocil, or (2) to a nervous impulse generated by the will of the animal set in motion by special stimuli.

Von Lendenfeld has pointed out in the same paper the existence in certain Guard polypes of adhesive cells. These cells form a globule, which approaches and finally projects beyond the surface, and retains any body which comes into contact with it. As such cells occur in the same polyp side by side with cnidoblasts, or replacing these structures in later stages, Von Lendenfeld believes that there is an homology between the two. Adhesive cells occur to the exclusion of cnidoblasts in Ctenophora.

For the nervous system in Hydroid forms add to the works of Jickeli and Von Lendenfeld referred to, Chun on Siphonophora, Z. A. vi. 1883.

For cnidoblasts and nematocysts see also Chun, Z. A. vi. 1883, and Hamann, J. Z. xv. 1882.

FIG. 9. Two nematocysts, A. partially discharged, B. undischarged from a Millepora; after Moseley, Ph. Tr. 1877, P1. ii. Fig. 2 a and b, a is slightly reduced in this copy.

The nematocysts are removed from the cells or cnidoblasts in which they were developed, and where they usually remain until discharged and torn away by the struggles of the prey. The cyst consists of two membranes, an outer thick and an inner thin, both of chitinoid character. The thick membrane apparently does not close over the pole at which the filament is emitted. The thin one, on the contrary, forms a closed sac. The filament is developed as a hollow ingrowth of one pole of this sac, and as growth proceeds and the filament becomes longer than the greater diameter of the sac, it is disposed in coils shown in B. When the filament is discharged the discharge takes place by a process of eversion, the inner surface of the hollow ingrowth becoming the outer surface of the hollow filament. This fact can be readily made out in a filament, the discharge of which has been arrested. The sac itself contains a homogeneous fluid, and it is apparently the pressure of the surrounding cell protoplasm on the sac with its fluid contents that causes the eversion of the filament. But after discharge the sac is said to retain its form and dimensions unchanged.

The base of the filament as seen in A is dilated, and hence the clear rod-like space passing down the centre of B. The dilatation is armed near its distal end with three spines. This form is characteristic of the larger nematocysts of Hydrozoa, and is only found in that group. In some Anthozoa the basal dilatation is of great extent, and is armed with spines disposed in a spiral. Nematocysts are found in all Anthozoa and Hydrozoa, but in no other Coelenterata. The trichocysts of Infusoria are formed in the same manner, as are the structures found on the dorsal processes of Aeolidia among Mollusca. In the Turbellaria very similar structures are met with in the rhabdocysts or rod-cells.