THE PLACENTALIAN RETINA 685 



nuclear elements and in the stratification of the inner plexiform layer 

 (c/. Fig. 193a, p. 660). 



Most placentalian groups have duplex (rod-and-cone) retinae, but in 

 the lowest orders it seems to be the rule for only rods to be present. The 

 cones are the simplest imaginable — all single, without paraboloids, oil- 

 droplets, or myoid extensibility. There are no cones in the armadillo, 

 possibly none in any edentate. All of the bats have only rods. Among 

 the Lipotyphla, the hedgehogs are pure-rod according to most inves- 

 tigators, though Menner found a few cone-type nuclei. The tree-shrews 

 (Tupaia et al) should have many cones; but the shrews have few or 

 none.* There are probably many more pure-rod rodents besides the 

 guinea-pig, and no cones have been reported by modern investigators 

 for any prosimian below the true lemurs"*", or by any of the half-dozen 

 investigators who have studied various species of Aotus. 



The rodents characteristically have great numbers of excessively slen- 

 der rods, like those of the rodent-like opossums (see Fig. 23f and g, 

 p. 55). Slender rods are also the rule in nocturnal primates and carni- 

 vores, and in the fruit-bats. In all such forms the outer nuclear layer is 

 naturally very thick, with up to 16-17 rows of nuclei. The inner nuclear 

 layer ini placentals rarely contains more than four or five rows, except 

 in Tupaia and in diurnal squirrels, where it may be several times as thick 

 as the outer (c/. Fig. 72, p. 177). The ganglion cells usually form but 

 one layer (in which they are often widely separated), except in an area 

 centralis (where any) and in the neighborhood of the primate fovea. 



The more slender and numerous the rods, and the fewer the cones, 

 the more likely it is that the rod and cone nuclei will be found markedly 

 differentiated from each other in size, shape, and chromatin distribution 

 (see p. 57). 



The cones of most of the placentals which have many of them are 

 much like those of man as a rule (Figs. 19, 22f; pp. 43, 54). In flying- 

 squirrels and ungulates, however, their 'myoid' regions are more or less 

 elongated; and in diurnal squirrels (except prairie-dogs) there appear to 

 be two types of single cones, one bulky proximally and slender distally, 

 the other slender proximally and plumper distally. In prairie-dogs how- 

 ever the cones are all alike, very slender, and not thus pseudostratified. 



*MIle. Verrier found all the cells alike (and, from her drawings, rods) in Crocidura 

 mimula; but in C leucodon and C. aranea there are more cones than in mice, according 

 to Schwarz. 



tKolmer claims a few for Nycticebus tardigradus, but Detwiler found none in this loris. 



