SAEGENT : THE OPTIC KEFLEX APPARATUS OF VERTEBRATES. 203 



tions than elsewhere. As a result of the lateral pressure, the torus has 

 become flattened laterally (Fig- 60), so as to appear in transverse sections 

 dorso-ventrally elongated, and the median longitudinal fissure, dividing 

 the torus into two lateral lobes, has almost disappeared. 



Auerbach ('88, p. 376) failed to recognize the torus in the Salmonidae. 

 In describing the sixty-day trout larva, he says, " Der Torus longitu- 

 dinalis fehlt." In his figure, however, he has obviously shown the torus, 

 describing it as " der in das Tectum eingebettete Zellstraiig." That 

 which he designates as the " Anlage des Torus longitudinalis " is the 

 inner and ventral edge of the tectum, where it meets the mesal edge of 

 the tectum of the opposite side below the torus (Figs. E, F). 



The cells of the optic reflex apparatus lie wholly within the torus 

 longitudinalis, of which they constitute the great mass (Fig. 60). They 

 are spindle-shaped and apparently bipolar, their long axis extending 

 vertically. The nuclei of the cells are relatively large, making up the 

 greater mass of the cell body. The cytoplasm is small in amount, 

 stains light!}', and is not sharply outlined, except at the two opposite 

 poles of the cell, where the neurites pass off. 



Tlie neurites from the dorsal ends of the cells are fine delicate fibres 

 without medullary sheatlis. Proceeding dorsad, these fibres become 

 aggregated into fascicles, forming two fibre-tracts, one on either side of 

 the median plane, the tractus toro-tectalis. At the upper level of the 

 torus they turn laterad and enter the tectum on either side (Fig. 60, 

 trt. tor. tct.). Where they leave the torus the bundles are compact, but 

 within the tectum they become more diffuse and break up in the super- 

 ficial fibre-zone of the tectum. These tracts were first seeii in teleosts 

 by C. L. Herrick ('91). Tlieir appearance led him to designate them 

 as "the gelatinous tracts of the torus," and to assign to them a neu- 

 rogiiar nature. 



The ventral neurites, the true axons, enter the mesocoele to form 

 Eeissner's fibre in two ways. The axons of those cells which lie in the 

 ventral and posterior parts of the torus, near the median plane, enter 

 the ventricle from the median ventral fissure of the torus, — tractus toro- 

 fibrae Reissneris ^J'os^mo;- (Plate 7, Fig. 51 ; Plate 8, Fig. 60, /os. Reis.). 

 In Cristivomer, where the tectum is less developed and the mesencephalic 

 roof less compact, most of the constituent axons of Reissner's fibre enter 

 the ventricle from the median fissure. These axons in the larvae are 

 not aggregated into bundles, but make their way singly into the mesocoele, 

 where tliey coalesce to form larger trunks, all of which eventually unite to 

 form Reissner's fibre. In Salmo and Salvelinus the portion of the meso- 



