30 MR. A. D. IMM3 ON THE GILL-KAKERS [May 3, 



gill-i-akers with his account, I have not been able to make out any 

 salient points of resemblance between the two structures. With 

 the exception of the rhombic plates and " fulcra " of the tail, 

 the scales are too degenerate to admit of a similar comparison 

 being extended to them. The plates and " fulcra " of the tail 

 are, however, tolerably well-developed structures. The former I 

 have examined after having thinned them, by rubbing down on 

 the surface of a fine hone, in the same way in which the gill- 

 rakers were treated. The matrix of a scale is colourless in thin 

 slices and is pervaded everywhere by lacunae which are similar to 

 those found in a gill-raker, but it does not contain any blood- 

 channels. The substance of the plate appears to be deposited 

 around a longitudinal core-like centre in the matrix. Adjacent 

 plates are united to one another by means of ligamentous connec-- 

 tions the fibres of which penetrate deeply into their matrix. These 

 fibres are comparable to what Hertwig calls the " Schuppen- 

 ligament " of the scales of Lepidosteus. 



In structure there is, therefore, a considerable likeness 

 between a rhombic plate and the basal portion of a gill-raker. 

 The matrix and its lacunae are identical in both cases; the 

 hollowed core or cavity in the base of the gill-raker might be 

 compared to the core of one of the plates, and to this may be 

 added the absence of blood-channels in both cases. The fibres 

 which connect a gill-raker to the branchial arch, and also which 

 bind adjacent ones together, are comparable with the ligaments 

 which unite neighbouring scales. For these reasons I think 

 that it is not improbable that the basal portion of each 

 gill-raker is the homologue of a ganoid scale — i. e., of one of 

 the rhombic plates which are found along the sides of the upper 

 lobe of the tail. The shaft or principal part of a gill-raker may 

 correspond to a greatly elongated spine, or to one of the evanescent 

 spines which are found in relation with each scale in the develop- 

 ing Lepidosteus, and which are regarded as representing the 

 spinovis portions of placoid scales. In Lepidosteus, as Nickerson 

 has pointed out, the basal plate, which is the essential part of 

 a ganoid scale, has come to be developed independently of 

 the vestigial spines, instead of being a continuation of the pi-ocess 

 by which the latter are produced ; and, in comparison with the 

 basal plate of the Selachian scale, it has increased greatly in size 

 and importance and has incorpoi/ated within itself fibres from the 

 dermis *. In a gill-raker, it would seem that we have a basal 

 plate which is similarly specialised, though not to so great an 

 extent, but that there has been no corresponding reduction in the 

 spinous portion, which, on the contrary, has become greatly 

 elongated. It has no trace, however, of a hard dentine layer, nor 

 of a coat of enamel or ganoin, unless the covering of mucous 

 miembrane is to be looked upon as the representative of the latter. 



# Bull. Mus. Comp. Zool., Harvard, vol. sxiv. 1893, pp. 115-140, 



