654 GEKARDA STIASNY-WIJNHOFF 



invaginating stomodaeum with a few glandular cells only 

 develops into a gastric tube that opens only a short distance 

 behind the beginning of the enteron, as in Balaenanemertes 

 m u s c u 1 o c a u d a t u s or Nectonemertes. With the greater 

 development of the stomodaeum the different parts become 

 better differentiated, in the first place gastric cavity and 

 pylorus, and this can be obtained in different ways, as shown 

 by Siboganemertes (Text-fig. 11) and Drepanophorus (Text- 

 fig. 8), or by A. m armor at us (Joubin) and A. mar- 

 moratus (Burger) ; the Pelagica provide us with a whole 

 series of stages in this development. Thirdly an oesophagus is 

 differentiated, which all Pelagica lack and often even other 

 groups. So Siboganemertes with its diverging structure of the 

 digestive system is not as primitive as the pelagic forms ; but 

 it cannot be included in the Drepanophoridae either, as the 

 development goes in a different direction. 



The vascular system shows a very primitive type, as 

 in Siboganemertes there are no anastomozing vessels with the 

 exception of the cephalic loops. The cephalic vessels bend 

 into the nervous ring of the brain in the ordinary way and 

 a dorsal vessel is present, as far as sections were made ; but it 

 is not in contact with the rhynchocoelomic cavity. The absence 

 of metamerically situated vascular loops is known in Uniporus, 

 in Keptantia, and in all Pelagica. The occasional presence of a 

 double anal loop in an abnormal individual of Pendonemer- 

 tes levinseni (Text-fig. 21, h) and the existence of a blind 

 dorsal median vessel in the tail arising from the anal loop, makes 

 it at first sight rather plausible that the reduction series as given 

 by Brink ma nn on p. 163 of his monograph (4) gives a true 

 account of the facts. But when we know that in primitive 

 Anopla the anastomozing vessels are absent, that they fail in 

 Uniporus (in which genus even the anal loop should be absent) 

 (3), and that they fail in Siboganemertes, we become sceptical 

 to the explanation of their absence in Pelagica. Moreover the 

 reconstruction of the tail of Pendonemertes on p. 20 (Text- 

 fig. 21, b) and the scheme on p. 163 (Text-fig. 21, c) are rather 

 different, and it seems not at all certain that the hinder vessel 



