THE TELEOST RETINA 585 



have been dealt with in the ecological chapters. Here, we can only 

 generalize about the retinae of the more ordinary teleosts, taking a 

 little space to mention a few of the more outstanding departures from 

 standard conditions. 



The retina in teleosts varies more in thickness than in other vertebrate 

 classes, from less than 100[1 to more than 500|1. Much of this variation 

 is caused by variation in the number of conductive elements per num- 

 ber of visual cells; but an unusual proportion of the thickness is usually 

 occupied by the visual-cell + pigment-epithelial layers, owing to the need 

 for scope for the extensive photomechanical changes characteristic of 

 the group. The other layers exhibit a neatness and 'purity' {i.e., an ab- 

 sence of ectopic elements) which we shall not see elsewhere until we 

 reach the lizards, birds, and mammals. The horizontal cells usually have 

 small bodies and slender (conductive?) processes, but occasionally — 

 and not only in physostomes (e.g., Esox) but even in the 'highest' tele- 

 osts {e.g., Stizostedion, a percid) one encounters massive, stellate hori- 

 zontals with broad bodies and short, thick, anastamosing processes — 

 exactly like those of some of the lowest fishes. Where the cones are pre- 

 dominant, the piling up of conductive and integrative elements results 

 in a thick inner nuclear layer and a compact ganglion-cell layer. At the 

 other extreme is the situation in such a light-shunning fish as the bull- 

 head, Ameiunis nebulosus, where the rods are large and the cones few 

 and small (see Fig. 63, p. 147). Here, the outer nuclear layer contains 

 only two rows of nuclei, the inner nuclear layer but one; the ganglion 

 cells are widely scattered, and only 2/9 of the thickness of the whole 

 retina lies between the external and internal limiting membranes. 



The cells of the pigment epithelium are usually long, with most of 

 the length contributed by their processes, which are few but thick, and 

 reach nearly (or quite) to the external limiting membrane (Fig. 20e, 

 p. 44). The migratory fuscin is usually in the form of needle-shaped 

 granules, the stationary pigment in round granules. Guanin may also 

 be present in large amounts, as in those minnows and perches which 

 have evolved occlusible tapeta lucida, and also in many deep-sea fishes 

 and in the Mormyridae, Elopidae, and Thunnidse, some anchovies, some 

 mackerels, the louvar, and one serranid {Polyprion) .* 



*These were not included in Table VII, pp. 240-1, as having effective tapeta, since the 

 presence of pigment as well as guanin in some of them (together with the absence of any 

 pronounced photomechanical changes) makes it questionable, without further study, how 

 effective the guanin may be as a mirror. 



