The dinosaurs of the Jurassic era (206–144 million years ago) were the tyrants of their world. But, concurrent with those dinosaurs, a mammal‐like reptile gained arboreal proficiency, permitting exploitation of the insect fauna that was co‐evolving and diversifying with the flowering plants. This key step would eventually lead to the evolution of the monotremes, marsupials and placental mammals.
Marsupials were early descendents of that last common ancestor and were established between 180 and 130 million years ago. The marsupial radiation separated from the placental mammals, and can provide us clues to the details of retinal evolution.
Around 180 million years ago, Pangea, the last supercontinent, began separating into Laurasia and Gondwanaland. That separation coincided with the last common ancestor to the mammalian clades. This continental drift would eventually isolate marsupials to Gondwanaland, especially to Australia, Antarctica and South America. Once isolated on South America and Australia, perhaps with minimal competition from placental mammals, radiation continued, resulting in different orders of marsupials. Seven such orders are extant, with three represented in the New World. With the eventual connection of South America to North America through the Isthmus of Panama, one of the later‐evolving marsupials successfully wandered north to establish a marsupium on North America.
The North American opossum (Didelphis virginiana) is that continent's only extant marsupial. This house‐cat‐sized omnivore is a more lately evolved marsupial and shows an unusual attempt at inner retinal nutrition, suggesting that the retinal vessels have evolved multiple times.
Most marsupials have an avascular retina much akin to that of their reptilian ancestors. Reptiles generally have a conus, which is a collection of vessels protruding from the optic nerve head and is probably responsible for inner retinal nutrition. The brush‐tailed possum has an avascular retina on the right cover and belongs to the order Diprotodontia (which includes the kangaroos and koalas), but the North American opossum has a retinal vascular tree on the left cover that superficially resembles that of other mammals. Don't be fooled, though! D virginiana and some other marsupials have a different vascular tree, and illustrate the potential variability and limitations of evolution in retinal vascularisation.
The retinal vasculature of D virginiana is peculiar, by our standards, because the arterial and venous segments of retinal vessels, including capillaries of the smallest calibre, occur in pairs . These vessels do not form the usual anastomic channels as would be found in mammals with vascularised retinas. Rather, the vessels travel in pairs and lead to hairpin‐end loops. The paired vessels, with the arteriolar limb usually on the vitread aspect, penetrate the retina and branch to form three distinct layers of capillaries—the most superficial lying in the nerve fibre layer, the middle one situated in the inner nuclear layer and the deepest extending to the external limiting membrane, which is considerably deeper than in normal mammalian holangiotic retinas (McMenamin PG,Krause WJ. Cell Tissue Res 1993;271:461). There is a minimal overlap of the area supplied from adjacent vascular loops. In adults,the choriocapillaris is mostly absent, and it cannot readily supply nutrients to the retina as efficiently as a more robust vascular plexus might. So, this paired vascular arrangement in the retina probably evolved to provide nutrients to the inner retina, although the forces leading the evolution are unknown.
Flat mount of retina, showing paired vessels.
D virginiana has other unusual ocular features, including a retinal tapetum lucidum found in the superior half of its retina. It is the only marsupial or mammal to have one. There are other mammals with this reflective structure, to be sure (BJO, April 2005), but these are choroidal tapeta such as those found in cats. Retinal tapeta are otherwise sauropsidian or piscine. In tapetal regions, the pigment epithelial cells enlarge, become relatively free of the more typical melanosomes and are filled with reflective, cholesterol‐containing spheres. This unusually thickened retinal pigment epithelium may interfere with diffusion from the choroidal circulation to the photoreceptors, and thus the deeply penetrating retinal vessels may be required to compensate.
Retinal vascularisation came relatively late to the orders of marsupials that have retinal vessels, because most marsupials such as the brush‐tailed possum do not have them. As D virginiana does have retinal vessels, albeit unusual ones by our standards, this suggests that the imperative for such vascularisation was not a unique feature of placental mammalian evolution.
The marsupial radiations were arboreal and at some point became nocturnal. This led to other major changes in visual function and morphology. D virginiana has a nocturnal eye with a large spherical lens dedicated to light gathering. The corneal diameter is nearly 90% of the diameter of the eye, again to increase light gathering potential. Surprisingly, D virginiana has a distinct area centralis (but not a true fovea) rather than the visual streak typically observed in many of the Australasian marsupials. The opossum has an almost pure rod retina with few cones, perhaps only 8000/mm2. There are single cones and double cones with oil droplets, although placental mammals do not have either double cones or oil droplets. Much like Tachyglossus (Echidna; BJO, February 2005), the retinal appearance and mosaic is typically reptilian. But, the reptilian retina is classically diurnal, not nocturnal. Rods are usually a relatively minor component in retinas of most sauropsids, but are abundant and differentiated in placental mammals and also in marsupials and monotremes. This suggests that the move to nocturnality came well after that last common ancestor, because some marsupials remain diurnal or at least arrhythmic (BJO, April 2005).
So, the two possums, although of different orders, tell us much about retinal evolution in the mammalian clades—an evolutionary tale.
Footnotes
Photographs by the author

