MUSEUM OF COMPARATIVE ZOOLOGY. 124 
largely animal (not mineral) matter, a substance which closely resembles 
the matrix of the scale itself. But I know of no case of true enamel in 
which the secretion is of such a mixed nature. 
Whether the bony material of the spine or the enamel tip begins to 
be formed first I have no means of determining. I believe that the two 
formations begin nearly simultaneously, although there is some reason 
for thinking that the enamel is not secreted until a part of the bony 
spine has been formed, for in a few cases I find peculiar abnormally 
shaped tips (Plate I. Fig. 6, Plate III. Fig. 18, and Plate IV. Fig. 26) 
which are more readily explained upon the former assumption. These 
all oceur on individuals raised in aquaria, and I believe them to be due 
entirely to artificial influences, — perhaps to handling. If the enamel 
were already present over the tip of the spine, I cannot believe that its 
point could be so distorted, whereas the matrix of the bony part of the 
spine is probably somewhat plastic when first secreted. It is also con- 
coivable that, before the commencement of secretion, some distortion of 
the papilla permanently altered its form and caused the abnormal shape 
of the spine. "This seems to me less probable, however, in view of the 
particular forms which the points of the spines have taken. 
The material of the spine is in lamelle, though these are not as dis- 
tinct as in the scales. The insoluble part of the point (Plate II. Fig. 
12, b) is doubtless made up of the first formed lamelle of the spine 
which have separated from the rest of the cone on treating with caustic 
potash. In Figure 25 (Plate IV.) is shown the splitting between lamella 
caused by an oblique cut through the basal part of a spine. No selero- 
blastic cells, however, are enclosed between the layers, and the material 
of the spine therefore differs from that of the scale at its base in that 
particular. But the absence of enclosed cells is not surprising, in view 
of the thinness of the walls of the spine. In reaction toward acids there 
is no difference between spine and scale. 
The number of spines upon a scale is wholly indefinite, and their 
arrangement for the most part not subject to any precise law. They 
are more numerous near the free (i. e. posterior) margins, and here 
show a tendency toward an arrangement in lines parallel with the 
edges of the scale (Plate IV. Fig. 31). I have counted over thirty 
spines on the posterior margins of a scale whose central area was 
nearly destitute of them. Though some spines are formed near the 
centre of the scale, by far the greater number arise near the posterior 
edges. With the growth of the scale, however, the margin advances, 
leaving them farther and farther from the edge. 
