Page:Quarterly Journal of the Geological Society of London, vol. 26.djvu/870

 in connexion with these columns, which seem to divide into their component parts as they approach them, two or three being thus joined to each column ; but in all our specimens this part of the fish is injured. The lateral plates are narrower than those lower down the body, and apparently gradually subside into the columns, without any distinguishing point.

These lateral plates at this point incline forwards ; the few next lower down are much inclined backwards ; so that there is here a large space over the abdominal region devoid of all such appendages. In all the four specimens this is the case, as well as in the German specimens before referred to ; and it is therefore probably the natural condition, and is not owing to accidental circumstances, though it is possible enough that pressure may have had something to do in forcing the plates asunder.

The ventral plates (m) extend from the root of the ventral fins to almost the commencement of the anal ; and from the lower part of the posterior plate there is a stout rod or process (o) extending upwards and backwards as far as the outer extremities of the lateral plates. Below, this process is wide and flattened ; above, it is cylindrical ; and at the point where this change takes place it is obtusely angulated backwards, the whole being a little arched in the same direction. There are apparently two of these great abdominal rods or processes, one belonging to each side, giving support probably to the lower posterior boundary of the abdominal walls and limiting in this direction the naked space above alluded to as extending over the lateral walls of the abdominal chamber. These two stout rods are closely pressed together, and in all the specimens are much confounded with each other.

Above the median line, in front, the outer members of the series of transverse plates and rods are likewise considerably modified. The change takes place gradually, just behind the elevated portion of the dorsal fin. The lateral plates (fig. 3, g), however, are only slightly modified ; but instead of being united to rods are continued as broad flat plates (i), which, inclining forwards, become angulated in front, and then for a short distance (j) incline backwards and are united to a series of hour-glass-shaped plates (k) that lean in the opposite direction, are extended to the dorsal margin, immediately below the anterior portion of the dorsal fin, and form part of an extended series that reaches to the root of the tail. The broad plate inclining forwards and in continuation of the lateral plates, is undoubtedly a modification of the rod of the posterior part of the body ; its margins are abrupt and produced into slightly elevated ridges ; and it contracts a little towards its outer extremity, and is slightly thickened at this point. The next portion, which inclines backwards, appears to represent the sigmoidal ridge or plate. The margins of this portion of the series are in contact, and seem to be slightly imbricated ; and the outer extremities are turned a little forwards to become united to the external hour-glass-shaped plates.

In the transverse centre of these modified sigmoidal plates there is a ridge (j'), similar in appearance to that on the median lozenge-