MARSUPIAL RETINA; PLACENTALIAN EYE 675 



however, is as reptilian as that of the monotreme. The visual-cell types 

 are the same ones as in Ornithorhynchus — single and double droplet- 

 bearing cones, and filamentous rods (which always outnumber the cones 

 very greatly, in contrast to Ornithorhynchus) . Only minor modifications 

 of the full monotreme pattern occur in marsupials. Thus in the Amer- 

 ican opossums some of the single cones lack oil-droplets (see Fig. 195c, 

 p. 670) ; and in all Australian marsupials so far examined by O'Day, 

 the double cones have oil-droplets in both their members — a curious 

 situation which occurs in American marsupials (and in some birds) only 

 as an occasional anomaly.* 



The rod nuclei in marsupials contain only one or two chunks of 

 chromatin — a differentiation, from the larger and open nuclei of the 

 cones, which is characteristic of the placentals but not of the mono- 

 tremes. It is not known whether the rod and cone foot-pieces are also 

 differentiated in marsupials. The cones of marsupials, like those of all 

 other mammals, lack paraboloids. This seems a point of some value in 

 defense of the monophyletic derivation of all the mammals from a single 

 reptilian stock. 



The single and double cones of marsupials and monotremes, from 

 their oil-droplets (and despite their loss of the paraboloids) , are clearly 

 the 'same' elements as the corresponding ones of the reptiles. The mono- 

 treme-marsupial rod is left to be homologized with the droplet-free 

 element of the sauropsidans (see Plate I). Its increased (nuclear) 

 differentiation in the marsupials, over that in the monotremes, coupled 

 with the persistence of the useless oil-droplets in both groups (these are 

 gone entirely in the placentals!), suggests that the ancestral monotremes 

 and the original marsupials were diurnal, and that the monotreme-mar- 

 supial line acquired its rods secondarily through transmutation and 

 perfected them within the confines of the line (cf. reptiles, birds). 



(B) Placentals 



The earliest placentals were 'insectivores' of the Deltatheridium type. 

 In the Mesozoic, these primitive forms diversified and established several 

 separate lines of ascent. The insectivore type itself persisted (giving off 

 the still extant Lipotyphla and, later, forms ancestral to the whales) and 

 culminated in the 'creodonts' — archaic carnivorous forms, of which the 



* According to Albarenque, Didelphis marsupialis and 'Azara (= D. azara?) have only 

 rods. This seems improbable in view of the extremely close relationship of these forms to 

 D. virginiana. 



