THE HISTORY OF BALANOGLOSSUS AND TORNARIA. 429 



mouth. The eyes, during these changes, have gradually lost their prominence, 

 becoming somewhat indistinct, and are at last completely absorbed in the walls of 

 the proboscis in somewhat older stages. The cavity of the body is formed by two 

 extremely thin walls extending from the base of the collar to the intestine, developed 

 probably by the extension of the problematic bodies of the stomach, though 1 have 

 never succeeded, owing to the opaqueness of the outer walls, in actually tracing this 



transition. 



When seen from below or above, two very distinct vessels — one dorsal, the 

 other ventral — can be traced, extending from the base of the collar to the intestine. 

 The vessels are apparently disconnected, being pointed at the two extremities. A 

 circular canal is formed round the oesophagus by the two forks of the water-system 

 which saddled the upper part of the anterior portion of the stomach in Tornaria. 

 It is connected with the heart, and opens outwards through the dorsal pore, but 

 seems to have no connecting link as yet with the outer vessels, though in older 

 stages this connection apparently exists. 



The gills, at first circular openings leading out from the oesophagus, become 

 gradually elliptical, then the walls nearest the middle line send out a loop, and 

 form the first trace of the complicated folds of the gills : they do not open exter- 

 nally, — at least in the earlier stages raised directly from Tornaria I could not trace 

 the opening so clearly seen in the Mediterranean species by Metschnikoff; it was 

 only in much more advanced stages of Balanoglossus that the outward opening 

 was discovered. Metschnikoff figures the gills as two large, funnel-shaped bodies 

 opening on each side of a deep dorsal furrow, which is also wanting in our young 

 Balanoglossus. I only succeeded in finding the skeleton observed by Metschnikoff 

 a minute granular plate in very small specimens of Balanoglossus soon after 

 their transformation from Tornaria. The only trace of the skeleton at the base 

 of the proboscis of To 



as 



•naria consists in an accumulation of granules similar to the 

 granular chord lying between the chitine supports of the gills in adult specimens. 



The oldest stage I succeeded in raising directly from Tornaria is a good deal 

 older than any described by Metschnikoff, so that, unless there is an error in his 

 observations, the formation of the gills in the European Balanoglossus must take 

 place very slowly, as his young Balanoglossus only showed one pair of gills in 

 the oldest stage he figures; our species commencing at once with four pairs of 

 rudimentary gills developed almost simultaneously already during the Tornaria stage. 



The smallest specimens of Balanoglossus dug up in the sand, although considerably 

 larger than those raised directly from Tornaria, are yet sufficiently different from 



vol. ix. 59 



