THE SKULL Ol' (^lU.M.KHA. 139 



side, after issuing from the anterior opening of tlie etlniioidHl 

 canal, enters this space and there bi-eaks up into numerous 

 terminal branches. The two layers of the membrane extend 

 into the supramaxiliary fold, and apparently end in its venti-o- 

 anterior edge. The tubules of the outer buccal group of ampulla? 

 open on the external surface oral to the supramaxiliary fold, and 

 they and the related sacs lie internal to the membrane that 

 lodges the other tubvdes, but, as already stated, fibrous sub- 

 dern)al tissues are found in the lips and the haso-labial folds 

 that seem to correspond to this layer of the corium, but they do 

 not form a definite membrane. 



In my work on Jhistelas (Allis, 1901) no attention was given 

 to the relations of the anipullse to this tibrous membrane, but I 

 now find, on ]-e-examining m^^ sections of this fish, that the con- 

 ditions there are strictly similar to those in Chimcera. In the 

 work on Micstelus I came to the conclusion that each ampullary 

 pore of the adult fish indicated, approximately, the place of 

 origin of the related ampullary organ, the long ampullary tubule 

 of certain of these organs being formed by an exctedingly rapid 

 growth of a primarily short tube, that tube being stretched out 

 between the two relatively fixed points represented by the 

 surface pore and the point where the sensoiy nerve enters the 

 organ. This has been since confirmed by Coggi (1902), and is 

 further confirmed by the conditions that 1 have since found in 

 Chlamydoselachus, the ampullje of that fish all having short tubes, 

 and the ampullary sacs all lying immediately beneatli the related 

 surface-pores. This marked difierence in the positions of the 

 ampullary sacs in this fish and those in 2Iustelus and Chimcera 

 evidently needs explanation, and it would seem as if it must be 

 in some way related to the amount of cranial flexure at the time 

 the ampullfe are developed. When the cranial flexure is at its 

 gi-eatest, those portions of the external surface of the head 

 on which the ampullary pores are found in the adult must lie 

 anterior or ventral to the curved anterior end of the central 

 nervous system, and hence in the region of the future rostrum. 

 If the ampullary sacs and the related nerves were well developed 

 at this time, it would seem as if the tendency would be to 

 hold the sacs there when the cranial flexure was later reduced 

 and the brain drawn relatively backward. The dermal tissues 

 Avould, on the contrary, probabl}^ i-etain their relative relations 

 to the underlying parts of the bi-ain, and hence also be drawn 

 Irackward ; and if the ampulla? had already penetrated the fibrous 

 layer of the corium and continued to lie in it, their shoit tubes 

 would be drawn out into long tubules lying in the fibrous laver 

 of the corium, as is actually the case in Jhisielvs and Chimara. 

 But if the ampullary sacs were not well developed when the 

 cranial flexure was at its greatest, their tubules could not be 

 stretched, and it would be the related nerve strands that would 

 be lengthened, as in Cldamydoselachus. This would not, howevei-, 

 explain Avhy these organs penetrate the fibrous layer of the 

 corium without whollv perforating it, nor whv these organs alone . 



10* 



