VISUAL SYSTEM: STATE OF THE ART 19 



et al. (1974) have described the external surface of the cornea in Mustelus 

 and Raja, finding structures very similar to those of mammals but quite 

 different from those of teleosts. The surface is covered with 0.2-jum wide 

 microprojections which probably represent microvilli. In mammals, anal- 

 ogous microprojections are thought to aid in holding the tear film in place, 

 but Harding et al. questioned this function for sharks. They noted, however, 

 that if ocular secretions were viscous enough, these projections could provide 

 some mechanical support and thus a stable optical surface. Certainly, the 

 microprojections increase surface area, thereby aiding in diffusion processes 

 probably necessary for aquatic corneas. 



The epithelium comprises about 50% (0.101 mm) of the corneal thick- 

 ness. In comparison, the human corneal epithelium is only 10% of the corneal 

 thickness (Maurice 1969). The basal layer of epithelial cells is separated from 

 Bowman's layer by a thin basal lamina similar to that of man. It is in this 

 layer that the sutural fibers are anchored. Zigman et al. (1965) indicate that 

 the corneal epithelium of the dogfish (Mustelus) contains approximately 12 

 layers of DNA-rich cells. 



Bowman's layer is comprised of a randomly oriented collagen fiber felt- 

 work, again similar to that of humans. The fourth layer, known as the 

 stroma or substantia propria, is about 0.07 mm thick and contains 25 

 lamellae. The lamellae are composed of extremely regular, dense collagen 

 fibers embedded in a gelatinous matrix similar to aqueous humor. The lamel- 

 lar ribbons are more nearly parallel than those of mammals and do not 

 interweave. This regularity would appear to be the most important optical 

 property of the cornea. 



The transparency of the cornea has been explained by Maurice (1957) on 

 the basis of an "interference theory." Maurice rejects the idea that the 

 cornea has a uniform index of refraction. Rather, the stromal fibers are 

 arranged in a regular semicrystal lattice with spacing of less than a single 

 wavelength of light and thus behave as a diffraction grating. The overall 

 effect is to suppress, by destructive interference, diffuse scattering of light 

 and to favor forward transmission. The main problem with this hypothesis is 

 that the relatively thick, randomly arranged fibers of Bowman's layer in 

 Squalus were shown by slit lamp examination (Goldman and Benedek 1967) 

 to scatter less light than the stroma. Thus, a regular lattice structure is not a 

 necessary condition for corneal transparency. In response to this criticism, 

 Maurice (1969) speculated that the degree of disorder tolerated by an inter- 

 ference mechanism might be considerable if the fibril axes were nearly equi- 

 distant. 



In man the cornea is the principal refracting element of the eye and 

 accounts for 75% of normal refraction (Maurice 1969). However, in aquatic 

 animals the refractive role of the cornea is severely reduced because seawater 

 and cornea have very similar indices of refraction. Consequently, most of the 

 refraction takes place in the large ocular lens. 



Goldman and Benedek (1967) describe the cellular elements of the dog- 

 fish stroma— keratocytes— as "quite similar to keratocytes found in normal 

 human and rabbit corneas." (p. 591). Keratocytes, which are modified 



