GIBSON— CEPHALOCHORD A : " AMPHIOXIDES." 245 



had been derived. On this view, the larval phase of Branchiostoma presents a true 

 recapitulation, and its metamorphosis— now rightly so-called— represents the transition 

 from an ancestral to a modern phylogenetic stage, a transition which occurred 

 historically when the pelagic ancestor became adapted to life in the sand. Tlie generally 

 accepted view that the larval peculiarities of Branchiostoma are in the main cocnogenetic 

 and not primitive, is, according to Goldschmidt, an unsatisfactory explanation of the 

 I'acts. lie further regards Amjihioxides as standing much nearer to the ancestral line 

 of the higher craniates than its more specialized relatives. 



These views are, of course, unaffected hy the subsequent admission that Amphioxides 

 may be only the larva of some more normal Cephalochordate : its organisation still 

 represents for Goldschmidt, albeit in a recapitulated larval form, the primitive type. 



Van Wijhe (1906) and WiUey (1906) have already criticised Goldschmidt's conclusions 

 on several important points. Since this report deals with Amphioxides in particular, and 

 brings forward a few new facts which weigh against them, I may perhaps venture on a 

 more detailed discussion of the questions he has raised. 



It will be as well to state at once in what characters Goldschmidt considers the 

 primitive nature of Amphioxides to be displayed, and how he conceives the evolution of 

 '' Ampkioxus" from an AmphioxidesAike ancestor to have occurred. The most impor- 

 tant supposed primitive characters are : — 



(1) The possession of a single, median series of gill-slits, symmetrically disposed 

 along the mid-ventral line, eumetamerically arranged {in A. pelagicus), and separated 

 from one another by gill-bars which show an obvious bi-lateral structure in their 

 branchial epithelium and musculature. 



(2) In close connection with the last, the absence of an atrial cavity. The gill-slits 

 open directly to the exterior between metapleural folds which hang down freely on 

 either side. 



(3) The separation of the pharynx by means of symmetrical lateral folds into a dorsal 

 "pars nutritoria" and a ventral "pars respii-atoria." 



(-i) The possession of an asymmetrical — left-sided — mouth. 



(5) The asymmetrical development, in the wall of the pharynx opposite to the mouth, 

 of the " endostyle " and club-shaped gland. 



The primitiveness of the second character is obviously indisputable, and we may leave 

 it out of consideration for the moment. Of the rest, by far the most important is the 

 tirst; the separation of the pharynx into dorsal and ventral portions, and tlie asym- 

 metrical position of the mouth and endostyle, are all, according to Goldschmidt, adapta- 

 tions to secure non-collision of the nutritive and respiratory functions, necessitated by 

 the appearance in an originally unjierforated pharynx of a mid-ventral series of gill-slits 

 extending to its extreme anterior end. A dorsal position for the mouth is rendered 

 impossible by the anterior extension of the notochord, a ventral by that of the gill-slits; 

 it therefore breaks through laterally into the dorsal section of tlie pharynx, wliicii has 

 become partially separated from the ventral by the development of lateral folds, food- 

 particles being thus prevented from falling through the gill-slits ; anteriorly, where the 

 separation is less complete, the asymmetrical endostyle is developed, and, aided by the 



SEOONrt SERIES.— ZOOLOGY, VOL. XIII. 33 



