MUSEUM OF COMPARATIVE ZOOLOGY. 129 



I have frequently found in sections of scales of different ages, killed 

 by various reagents, a narrow marginal zone appearing somewliat 

 different from the rest of the section (Plate I. Fig. 3, and Plate III. 

 Figs. 17, 20). This often appears strikingly like a different layer 

 of material, and is sharply bounded from the deeper-lying layers. I 

 believe that this appearance is due to tlje action of reagents used in 

 killing or staining {in toto), or both, and not to any natural difference 

 between the parts of the scale. No such difference is found in material 

 killed in alcohol. Klaatsch refrains from stating anything concern- 

 ing the methods he used or the manner in which his fish had been 

 killed. The impression produced by the study of his paper is that he 

 was misled by some artificial condition such as I have just mentioned 

 and figured, 



I have already stated that the spines frequently leave no trace of 

 their existence in the adult scale. To what their disappearance is due 

 I am unable to state positively. Before the outer layer begins to be 

 formed they have almost completely disappeared from all the central 

 part of the scale, but usually a few still I'emain close to the posterior 

 edges. Hertwig ('79, p. 7, Taf. II. Figg. 1, 2, 3, 10) has described and 

 figured certain little knob-like elevations projecting up into the ganoin 

 layer, and he believes them to be remnants of spines which have been 

 lost. In this conclusion he agrees with that already expressed by Reiss- 

 ner ('59, p. 260). I have carefully looked through series of sections 

 from two diff'erent adult "gars" for such structures, but without finding 

 them. Hence it would appear that their occurrence cannot be considered 

 a constant feature. In scales from one young fish 46 cm. long, however, 

 1 found some peculiar structures similar to those figured by Hertwig. 

 One of these is shown in Plate TI. Figure 14 and is without doubt, as 

 Hertwig maintained, the base of a lost spine. It had been eiitirely 

 buried in the ganoin. There were also present a number of other 

 much smaller bodies lying between the top of the scale and the over- 

 lying tissue in the space from which the ganoin had been dissolved. 

 These were widely scattered, and many were little larger than an odon- 

 toblast cell ; they were of an ovoidal or spheroidal form, and appear 

 to be the same in composition as the one figured (Plate II. Fig. 14), 

 and I am led to the conclusion that they are also remnants of the 

 bases of lost spines, which were probably in process of absorption when 

 the secretion of the ganoin began and buried them. Hence I be- 

 lieve that the obliteration of spines is in general due to absorption, 

 though it is hardly conceivable that the distal part of the spine dis- 



VOL. XXIV — NO. 5. 2 



