100 Annals of the Carnegie Museum. 



or bilobation. The nucleus also is more basad, and the radial 

 system is not reduced to a U, though there is a certain tendency 

 for two of the radii to connect and form this figure. The abundant 

 longitudinal circuli in the interradial region distinguish Crenuchine 

 scales from those of Characidium blennoides. So far as the scales 

 go, the Crenuchines seem more primitive, or less specialized, than 

 the fishes with which I have compared them. 



Iguanodectin.e. 

 The only species is Iguafiodectes tenuis Cope. The scales are trans- 

 versely oval, with the broad nuclear area approximately central; 

 circuli basal and lateral only, the latter widely spaced; no radii. 

 This is evidently not far from the condition found in Cheirodon 

 insignis. The fish itself is superficially just like some Menidia. 



Bryconin.e. 



Brycon Jalcatus Miiller and Troschel (Plate XXVI, fig. 6). The large 

 subquadrate scales are about ten mm. long and twelve mm. broad, 

 the exposed part strongly silvery, the rest dull. The silvery part 

 (about the apical third, or a trifie less) is ornamented with numeious 

 (about thirty) radii, which are more or less curved, as the figure 

 shows. This radial field is variably crossed by irregular growth- 

 ridges, which are interrupted at the radii. The other part of the 

 scale is densely covered with the finest possible circuli, except in the 

 broad nuclear field, where the circuli are broken up and form a 

 minutely labyrinthine pattern. The nuclear and adjacent regions 

 present fine irregular cracks, which do not seem to represent de- 

 generate radii. The basal circuli are entirely transverse. This is 

 a very distinct scale. 



Holobrycon pesu (Miiller and Troschel). As in B. Jalcatus, the 

 exposed part (about one-third) is silvery, the rest dull. The scales 

 are about as broad as long (about five mm.), and have evident 

 laterobasal angles. The apical field has a very variable number of 

 radii (six to twelve or more), which extend over about half the length 

 of the scale. The radia' field has very distinct though widely-spaced 

 circuli, which converge mesad, but are not far from longitudinal. 

 The other part of the scale has very fine circuli in the manner 

 of B. Jalcatus, but the nuclear modified area is small and rather 

 different, the broken circuli being reduced more nearly to minute 

 spots. This is evidently close to B. Jalcatus, the latter being the 

 more specialized of the two. 



