578 HIGHER FISHES 



families. In other words, it appears that all of these special teleostean 

 structures had been evolved before the great schism came; and, though 

 in general the physostomes are anatomically a bit 'primitive' as com- 

 pared with physoclists, there is no majestic progress to be seen in passing 

 through the families of the one division to the families of the other. 



The margin of the circumocular sulcus usually forms a narrow circular 

 lid-fold, lapping onto the eyeball. Where the eye is retractile, temporary 

 'lids' may largely cover the eye (sometimes moved by a special dermal 

 sphincter muscle, like an orbicularis) ; and 'adipose lids' (^.v.) are com- 

 mon in swift swimmers. The orbit is usually roomy unless the eyeball is 

 very large or tubular. Cushioning venous sinuses are developed to 

 greater or lesser degree, but other orbital structures are very variable 

 and our knowledge of them lacks synthesis. A tenacular ligament often 

 holds the eyeball in the orbit; this has no genetic relation to the selachian 

 optic pedicel, for in some rays both structures are present side by side. 



The oculomotor muscles are usually normal in number and arrange- 

 ment (see Fig. 165, p. 565). They are often long, and are carried 

 through canals in the bones of the skull — an anterior canal holding the 

 two obliques, and a posterior one the four recti. There are no special 

 retractor- or levator-bulbi muscles. 



The eyeball is almost always flattened anteriorly, with its axial length 

 its shortest diameter, and with its horizontal diameter tending to be its 

 greatest dimension in swift forms, but more nearly equal to the vertical 

 diameter in slow-swimming and small-eyed species. Since the corneal 

 surface is eliminated optically, there is no need for it to be smooth; and 

 it is often irregular, concentrically ridged, etc. 



The sclera is very variable in its morphology. Primitively, it must have 

 contained a complete cup of hyaline cartilage as in all lower fishes. It 

 does contain at least some cartilage except in gymnotid eels, pearl-fishes, 

 and a few others (where it is entirely tendinous) and in the tetras (where 

 it is entirely bony) . But in no instance is the cartilage-cup intact fundally 

 — one might put it that the floor of the original cup, over the whole back 

 of the eyeball, has been replaced by fibrous tissue. This fibrous window 

 is often so large that the cartilage is restricted to a broad equatorial, or 

 narrow just-post-limbal, ring (e.g., pipefishes, many salmonids). The 

 cartilage may also be widely distributed, not as one piece but in the form 

 of little islands in a fibrous continuum (elephant-fishes) . Typically there 

 are thin plates of bone temporally and nasally, which may develop either 

 anterior to the cartilage or external to it (the cartilage beneath them 



