142 
bulletin of the bureau of fisheries. 
Lucius ), and much less dense circuli. On the other hand, the scale of Dallia pectoralis does not essen¬ 
tially differ from the smaller pceciliid scales, though the sculpture is less regular. Latinucleate scales 
are common in the pceciliid material examined. 
The following rather unsatisfactory key is based on normal (not latinucleate) scales: 
Scales larger, more than 3 mm. across. 1 
Scales smaller, less than 3 mm. across.. • • • 4 
1. Apicolateral outline a regular even curve, or without angles; nucleus a short distance apicad of 
middle; basal radii about 17, only about the 10 midmost complete. 
Cyprinodon carpio Gunther. Tarpon Springs, Fla. (Evermann & Kendall). 
Apicolateral outline evidently bent or angled, the apical margin thus separated from the lateral.. 2 
2. Nucleus less than a third of length of scale from apex; apicolateral angles pronounced, but apical 
margin convex; basal radii about 13 to 16. 
Gambusia puncticulata Poey. San Antonio, Pinar del Rio, Cuba (Eigenmann & Riddle) 
Nucleus more than a third of length of scale from apex. 3 
3. Scale larger, very broad (lat. pr. 6 mm.), apical margin hardly elevated. 
Fundulus heteroclitus (Linnaeus), Woods Hole, Mass. 
Scale smaller, not so broad (lat. pr. 4 mm.) apical margin more elevated. 
Mollienisia latipinna Le Sueur, Gordon’s Pass, Fla. (Henshall). 
Differing from the above two in having the scale longer than broad. 
Fundulus majalis (Walbaum). 
4. Nucleus far above middle of scale; basal radii about 10 or n. 
Poecilia butleri; Jordan; fish about 40 mm. long; SalinaCruz, Mexico (C. C. Deam). 
Nucleus about middle of scale or even below. 5 
5. Scale subquadrate, with more or less evident anterolateral angles. 
Fundulus diaphanus (LeSueur), Osterville, Mass. 
Scale broad, not subquadrate, without anterobasal angles. 6 
6. Scale less than 1% mm. diameter. Lucania parva (Baird & Girard), Woods Hole, Mass. 
Scale more than mm. diameter. 
Gambusia affinis (Baird & Girard), Myakka River, Florida (J. A. Henshall). 
Dr. Max Ellis has kindly allowed me to examine scales of Anableps anableps (Linnaeus) the four¬ 
eyed fish, and A. microlepis Garman, which he collected at Georgetown, British Guiana. They are 
quite large, those of A. anableps much broader than long, with a width of about 8 mm., those of 
A. microlepis smaller, and not so broad. In all respects they are quite typical pceciliid scales, much 
like those of Cyprinodon, but with considerably denser lateral circuli. 
Order XENOMI. 
DALLIIDAJ. Alaska blackfish. 
Dallia pectoralis Bean, from Nushagak, Alaska (Albatross collections) has small scales about mm. 
diameter, with coarse circuli and numerous (about 15) basal radii, the basal margin finely scalloped. 
There is nothing in these scales which might not be expected in the Poeciliidae. Regan has recently 
proposed to include the Esocidae, Umbridae, and Dalliidae in Haplomi and to recognize a new order, 
Microcyprini, for the Poeciliidae and Amblyopsidae. So far as the scales go, the following arrangement 
would seem natural. 
1. Umbridae. 
2.A. Esocidae. 
B.a. Dalliidae. 
b. Poeciliidae, 
the last showing a close approach to the acanthopterygian type. 
