3. Capillary vessels.
Plate II, fig. 9, represents two stellate pigment-cells, which have coalesced at a. In that instance two cells had been gene- rated at some distance from one another, their bodies may still be distinguished as two spots somewhat thicker than the rest of the structure. These cells became elongated on different sides into hollow processes, which, like the cavities of the bodies of the cells, are filled with pigment. Two processes of the two cells came into contact at a, and then coalesced, the separation at the point of union appears to have been absorbed also at the same time, so that the cavities of the two cells communicate immediately with one another; at all events there is no apparent interruption to the pigment, which forms the contents of the cells and their prolongations. (See page 78.) Now, if we imagine several such stellate cells to be developed on a large surface at similar distances from one another, and the several prolongations issuing from each separate cell to coalesce with those issuing from the other cells, in the manner represented in the figure at a, the result will be a network of canals ex- tending over the entire surface, and all communicating with each other. The size of the meshes of the network is determined by the distance of the cells from each other, and by the number of the prolongations issuing from each cell. Such, then, appears to be the process by which the capillary vessels are formed.
The observations, on which this mode of formation of the capillary vessels is based, were made partly on the tails of very young tadpoles, and partly on the germinal membrane of the hen’s egg. They are as follows:
1. The capillary vessels, in the tail both of the fully-deve- loped and young tadpoles, are seen to be surrounded by a thin, but distinctly perceptible membrane, which does not exhibit any fibrous arrangement. (See pl. IV, fig. 11.) The variety in the thickness of this membrane in different im- stances sufficiently explains why we cannot distinguish it in all capillary vessels, just as we cannot detect the cell-membrane even in the blood-corpuscles, although there can be no doubt of its existence. Where the capillary vessels exhibit a fibrous
