THE TELEOST RETINA 



587 



among the most primitive of living teleosts makes it seem fairly reason- 

 able that the typical identical-twin cones of teleosts have indeed been 

 derived from double elements like those of Amia. Against this view, 

 however, must be placed the presence of double cones in Fundulus (see 

 Fig. 24f, p. 59). Fundulus being one of the cyprinodonts, which (though 

 they are soft-rayed physoclists) probably deserve a place near the perches 

 at the top of the acanthopterygian heap, its double cones may well have 

 been manufactured from ancestral typical twins. And if this has been 



CD 



O 



CO 



Po 



03 



Fig. 170 — Visual-cell patterns of holosteans and teleosts. 



a, units of visual-cell mosaics in representative teleosts. Redrawn from Eigenmann and 

 Schafer. Only the single and twin cones are shown — the (much smaller) rods fill the spaces 

 around and among them; from above downward: commonest pattern, as seen in Perca; 

 pattern in Salmo (without the central single, this would represent Blennius); pattern in 

 Scorpana porcus. 



b, single cone, double cone, and rod of a holostean, Amia calva. x 1000. 



c, single cone, twin cone, and rod of a teleost, Sli;osleJtori vitretnn. xlOOO. Drawn from .1 

 preparation of George A. Moore. 



