344 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS. 



still more complicated. Primitively it formed a portion of the second 

 region, e", and its histological elements had the same arrangement, 

 i. e., were inverted. Later, when this third portion of the retina curled 

 over backwards, revolving through an angle of 180 (compare Figs. 5 

 and 6, Plate XLVII, with Fig. 1, Plate XLVIII), it became secondarily 

 uninverted. By the final reversal in the position of the whole eye this 

 third region became again physiologically inverted. It is not easy to 

 conceive the advantage derived from this complicated series of changes. 

 One thing, however, seems to be most clearly indicated, namely, that 

 there is no fundamental difference between inverted and non-inverted 

 eyes, and that one sort can very readily and apparently for slight reason 

 pass into the other sort. It is therefore to be expected that closely 

 related animals may possess eyes differing in this regard. 



A further comparison of the horseshoe-shaped eye of the solitary 

 form with the immature eye of the chain individual should be made. 

 That the two eyes are fundamentally similar is shown by the fact that 

 the eye of the chain individual passes through a horseshoe-shaped onto- 

 genetic stage. After it has assumed its disk-like form and before it has 

 become reversed, its posterior portion (e") corresponds to the posterior 

 curved part of the horseshoe eye, and its anterior portion (e r ) to the two 

 anterior limbs of the horseshoe. The primitive condition for the horse- 

 shoe eye is that of inversion (as shown above) ; secondarily it becomes 

 non-inverted. In the immature disk-like eye of the chain individual 

 (Plate XLVIII, Fig. 1) the posterior portion (e") has an inverted retina, 

 i. e., is primitive in this respect, while the retina in the anterior portion 

 (e') is from the first non-inverted ; that is, the ontogenetic development 

 of the second region of the eye (e") retains more of the phylogenetic 

 history than does the first region of the retina (e'). We must explain 

 this as a coenogenetic abbreviation of the ontogeny. 



SECTION II. The Innervation of the Ciliated Funnel. 



The question of the innervation of the ciliated funnel in the ascidi- 

 ans has been much discussed, but its actual innervation has never been 

 shown. The nerve supply of the funnel in pyrosoma, doliolum and 

 salpa has not, so far as I can learn, been demonstrated. 1 The fact of a 

 definite nerve supply is essential to prove that it functions as a sense 



1 Joliet (7) argues plausibly that there is probably such a nerve supply in pyro- 

 soma, but he does not show that it is present. 



