432 



SCIENCE. 



[N. S. Vol. V. No. 115. 



position, form and development resemble 

 the lobi inferiori of fishes. In fishes, am- 

 phibia and reptiles two strands of nerve 

 fibres associated with the median eye, and 

 springing from a point just in front of the 

 posterior commissures, extend around the 

 sides of the fore-brain and terminate in the 

 neighborhood of the lobi inferiores. Thus 

 the anatomical relations in both Vertebrates 

 and Arachnids are essentially alike. (2) The 

 lateral eyes of Limulus and the i-elated fos- 

 sil forms, owing to more rapid growth of the 

 haemal margins of the eye, are kidney- 

 shaped, with the helum directed toward the 

 neural side. This gives the most advan- 

 tageous and economical arrangement of the 

 ommatidia on the convex surface of the 

 carapace of such animals. If such an eye 

 is infolded and forms a part of the brain, as 

 our theory demands, it will not only be 

 turned inside out, but upside down. The 

 most rapidly growing edge will then, in a 

 Vertebrate, be on the neural side and the 

 retina will be kidney-shaped. Under such 

 conditions, as there is no obvious hindrance 

 to continued growth in that manner, the kid- 

 ney shape will be accentuated, thus bringing 

 the hsemel margins together and forming the 

 characteristic choroid fissure of Vertebrates. 

 (3) Such a view implies that the ances- 

 tral optic ganglion of Vertebrates is not a 

 part of the retina, as is often assumed to be 

 the case, but a series of ganglionic lobes 

 similar to those belonging to the compound 

 eyes of arthropods. 



In insects where the ganglion is beauti- 

 fully developed it usually consists of three 

 great lobes : (1) a central one the largest, 

 and shaped like a thick hemispherical shell, 

 in fact, having much the same shape as the 

 compound eye itself; (2) a thick semicir- 

 cular band extending along its whole distal 

 margin (the retinal ganglion); and (3) a 

 neailj' spherical ganglion on its proximal 

 side. The latter is united to the base of 

 the fore-brain by a thick stalk, and each 



ganglion is united with its neighbor by de- 

 cussating bundles of fibres. The medullai-y 

 portion of each ganglion is mainly on the 

 ha;mal side, the ganglion cells on the dorsal. 

 In Limulus we have just such a set of optic 

 ganglia, and in the embryos they project 

 far away from the fore-brain and at right 

 angles to it, as in nearly all other arthro- 

 pods ; but they gradually' move backwards 

 toward the wide dorsal line till in the adult 

 crab they lie jammed close together on the 

 hfemal side of the fore-brain near the me- 

 dian line. 



Now if the migration of the optic ganglia 

 of Limulus should continue in the same di- 

 rection they would cross the median line 

 and, following the path of least resistance, 

 move in opposite directions towards the 

 open space just behind the cerebral hemi- 

 spheres. 



As the proximal end of the ganglion 

 stalk is fixed to the sides of the fore-brain, 

 the ganglion would be bent double so that 

 the stalk and optic nerve would lie parallel 

 and side by side. The whole ganglion 

 would now form the roof and sides of the 

 mid-brain, and would reverse the direction 

 of its curvature to fit its new position, thus 

 effectivelj' disguising its true character. 

 The ganglion hy its change in position is 

 partly inverted, turning the ganglion cell 

 laj'er toward the ventricle and the medul- 

 lary portion toward the outside, just as the 

 theory demands. In this position the reti- 

 nal or most distal ganglion becomes the 

 torus longus, the hemispherical one the 

 tectum opticum, the third one the collicu- 

 lus; the stalk or proximal end of the gan- 

 glion becomes the brachia, and probably 

 such other tracts as unite the various lobes 

 with the thalamencephalon ; the crossed 

 nerves form the chiasma and the optic 

 tracts, the fibres in both cases entering 

 what is morphologically the distal end of 

 the series of ganglia, i. e., the torus longus 

 and its vicinity. (4) The commissural 



