142. With respect to the transition of the spindle-shaped cells into fibrils, since there is a gap in the observations of Boll, and since those of Schmidt are subsequent to the disappearance of the cells, and in both cases all trace of nucleus has disappeared, I suggest that we have here an analogy with what Weismann has recorded of the metamorphoses of insects. In the very remarkable memoir of that investigator[172] it is shown that the metamorphoses do not take place by a gradual modification of the existing organs and tissues, but by a _resolution_ of these into their elements, and a _reconstruction_ of their elements into tissues and organs. The muscles, nerves, tracheæ, and alimentary canal, undergo what may be called a fatty degeneration, and pass thence into a mere blastema. _It is out of these ruins of the old tissues that the new tissues are reconstructed._ On the fourth day the body of the pupa is filled with a fluid mass--a plasma composed of blood and dissolved tissues. The subsequent development is thus in all essential respects a repetition of that which originally took place in the ovum.[173] Two points are especially noticeable: First, that in this resolved mass of granules and fat globules there quickly appear large globular masses which develop a fine membrane, and subsequently nuclei. A glance at the figure 51 of Weismann’s plates reveals the close resemblance to the earliest stages of nerve-cells; and the whole process recalls the regeneration of nerves and nerve-centres after their fatty degeneration. Secondly, the nerves reappear in their proper places in the new muscles, and this at a time when the nerve-centres are still unformed; so that the _whole peripheral system is completely rebuilt in absolute independence of the central system_. The idea, therefore, that nerve-fibres are the products of ganglia must be relinquished. This idea is further discountenanced by Boll’s observations, which show that the fibre-cells are from the first different from the ganglionic cells; and by the observations of Foster and Balfour, that “fibres are present in the white substance on the third day of incubation”; _whereas cell processes do not appear until the eighth day_. Foster and Balfour are inclined to believe “that even on the seventh day it is not possible to trace any connection between the cells and fibres.” In the later stages, the connection is perhaps established.[174] 143. We may, I think, conclude from all this that in the higher vertebrates the white substance of brain and cord is not the direct product of the gray substance; in other words, that here nerve-fibres, even if subsequently in connection with the ganglionic cells, have an independent origin. They may grow towards and blend with cell processes; they are not prolongations of those processes. They may be identical in structure and property, as one muscle is identical with another, but one is not the parent of the other.