112 Cockerell — Scales of Some Malacopterygian Fishes. 



ferences in structure, however, are considerable, and I do not wish 

 to suggest tliat the scales should count against a positive opinion of 

 experts in fish-anatomy that the suggested relationship is impossible. 

 A chance to examine the other three living Osteoglossid genera is 

 of course eagerly awaited. 



Nijtopteridpe. 



Notopierus afer Gthr. Elongate (oblong) scales, with rounded corners; 

 nuclear area far apicad ; circuli extremely tine, the apical transverse ; 

 apical radii numerous but rudimentary, wholly submarginal ; basal 

 radii about 22, very long, very well developed, the middle ones inclined 

 to be wavy; basal margin not at all scalloped. Boulenger compares 

 this family with the Hiodontida, and the scales are of the same gen- 

 eral type, although readily separable. They do not at all resemble 

 those of the Mormyrids or Osteoglossids. In general appearance, tlie 

 Notopterids are very unlike the Hiodontids. 



There is a distinct, even close, resemblance between the scales of 

 Nutopterus and Gadus. This is singular, because the (ladidfe are sup- 

 posed to have Ijeen derived from some ]\Iacruroid form, and the 31acru- 

 ridie are superficially much like the Notopteridte. 

 Macrurid scaler I have never seen. 



rantodontidrc. 



Pantodon huchhohi Peters. Scales subcircular, with rounded but evident 

 laterobasal angles; nuclear area a little apicad of the center; circuli 

 very well developed, the apical transverse. With a microscope it is 

 seen that the lateral and basal circuli are moniliform, the inner ones 

 actually broken into separable minute elements, while the apical cir- 

 culi are thickened near the nuclear area, much broken beyond, and 

 lacking in the submarginal field. Kadii few and very strong, more 

 basal than apical, the latter more or less branched. In the radii, and 

 general form, this scale is like that of the Alestini and the South Ameri. 

 can Characinids of the type of Erythrinus, Huplias, Pyrrltulina, 

 Chalceus, Clialcinus, etc. It is also like one of the groups of Barhu-s. 

 Boulenger states that the Pantodontidie are closely allied to the Osteo- 

 glossidie; the scales are very dissimilar, and yet agree well in the 

 character of the circuli, which in both are moniliform or resolved into 



granules. 



I'liractuhi'inUUc. 



I'liravtolwmus ansorgii Boulenger. Scales also of the Alestiform type, 

 with very strong laterobasal angles and few very strong railii; the 

 truncate base is broad and crenulated. The circuli are dense and 

 strong basally and laterally, but apically are very remarkable, every 

 third or fourth being strong and conspicuous, while those between 

 are evanescent, the whole being entirely longitudinal, the strong 

 ones ending in short triangular teeth on the margin; between these 

 strong circuli,, in the sul)apical field, are scattered ^^niall circular j)its. 



