90 Proceedings of the Biological Society of Washington. 



or denticulation can only be seen with a high power of the microscope; 

 apical field, from a little apicad of the nucleus, thin, granular, without 

 any circuli or radii. 



Lepidoblennius haplodartylus. New South Wales (Brit. Mas.). Scales 

 very small, the largest little over half a mm. diameter; broader than 

 long, with the same apical field without radii or circuli as in Clinus; 

 radii 9 to 12, basal, broadly cutting the very widely spaced circuli. This 

 is simply a very degenerate scale of the Clinus type, looking entirely 

 different because of its shape and few, widely spaced circuli, which are 

 only about nine in number. The ciculi are without beading. 



Stichaeidze. 



Dictyosoma temminckii. Japan (Brit. Mus.). Scales excessively mi- 

 nute, elongate-oval, about 800 microus long and hardly half as wide, the 

 ends broadly rounded; circuli coarse, extending all around the scale; 

 no distinct radii, but one or two irregular and imperfect ones in the basal 

 region. The circuli show exceedingly minute and imperfectly developed 



beading. 



Vivaria subbi/urcata. Scales of the same general type as those of 

 Dictyosoma, but less degenerate, having several well defined basal radii 

 and an apical thin field without radii or circuli. 



Pholididse. 



Pliolis gunnellus. Scales minute, more or less circular to oval ; radii 

 lateral as well as basal. Scales of the same general character as those of 

 Anarrliichas. 



All the above are really degenerate scales except those of Clinus, which 

 should indicate better than any of the others the affinities of the suborder. 

 The description of the scales of Clinus given above will nearly apply to 

 those of Helicolenus dactylopterus in the Scorpaenida?. In Hvlicolenus 

 the radii are not so close together, and the apical field is ornamented with 

 spots, representing broken up fragments of circuli, but the essential char- 

 acters are practically as in Clinus. Other Scorpaenids are ctenoid, 

 affording an excellent transition to the typical ctenoid scale of the Percoid 

 fishes. Mr. Regan writes ( in litt. ) : "That the Blennioids are modified 

 Percoids seems almost indisputable;" and the scale characters of the 

 Blenniiformes and especially Cliniformes support this view. Goodrich 

 places the Scorpaeniformes next to the Blennioids, and while they may 

 not occupy a strictly intermediate position between the Blennioids and the 

 Percoids, they at least have the intermediate type of scales. It must also 

 be said that there is nothing in the scales of the series described above to 

 suggest that they are not related. In spite of superficial differences they 

 conform to a single general type. 



When we come to Regan's division 3, the Ophidiiformes, the case is 

 very different. The scales are entirely of the Gadoid type. The very 

 numerous radii extend all around the scale, and there is no apical area 

 without circuli or radii. There is nothing whatever to indicate any 



