Mi'scellancoua. 441 



The walls of the stomach are differentiated into two layers — an 

 external one of niuscnlar fibres, and an internal rauoous layer; this 

 latter produces five horny teeth, preceded sometimes by the api)ear- 

 ance of a single larval plate. The vitelline sacs, of which there are 

 two at first, unite into one in the Orthoconcha. This sac, which 

 opens into the dorsal part of the stomach, is absorbed and diminishes 

 rapidly in the Hj-aleaceae ; on the contrary, it is developed in the 

 Styliolaceac and the Creseideaj, where it seems to play, provisionally, 

 the part of the liver. In every case it diminishes in proportion as 

 the liver is developed. The liver is composed of small diverticula of 

 the wall of the stomach. The nutritivi' sacs have nothing to do with 

 the formation of this organ. 



Theotocysts are formed early, in the midst of a layer produced by 

 a doubling of the ectoderm still composed of large embryonic cells. 

 The otolith originates in the thickness of the wall of the vesicle, 

 and falls afterwards into its cavity. In the Limaceft and the Cepha- 

 lopoda the otocyst is formed by an invagination of the ectoderm 

 iJready composed of very small cylindrical cells. The size of the 

 embryonic cells of the generative layer seems to be in this case, as 

 in many others, the cause which determines the mode of formation 

 of an organ by invagination or by simple folding. 



The nervous system is composed of a cephalic nervous mass and of 

 a suboesophageal mass. The former is formed by a double invagina- 

 tion of the ectoderm of the cephalic region in the area circumscribed 

 by the velum ; the mode of formation of the second has not been 

 observed in the Ptei'opoda. 



The appearance of the shell is preceded by the formation of an 

 invagination of the ectoderm a little in front of the aboral pole. 

 This preconchylian invagination turns round ; and the first rudiment 

 of the shell appears on the projection thus formed. In exceptional 

 or abnormal cases this invagination does not turn round, or rather it 

 is reformed after having disappeared. Its existence is incompatible 

 with that of an external shell and vice verad. It is the point of de- 

 parture of the band which secretes the shell ring by ring, and which 

 becomes the margin of the mantle. The first part of the shell, that 

 which the larva inhabits, often differs from the poition which is 

 added later on ; it may persist, fall or break off; and it has furni-hed 

 me with characters which have enabled me to subdivide the sub- 

 order of the Thecosomatous Pteropoda. The existence of the pre- 

 conchylian invagination cannot be satisfactorily ex|)lained by purely 

 physiological causes ; it seems, then, to have hereditary causes, and 

 may morphologically be compared to the conchylian invagination of 

 the mollu.sks with internal shells, which invagination I have studied 

 in Sepiola and the Limax. The existence and signification of that 

 invagination in the Cejjhalophora, the Cephalopoda, and the Larael- 

 libranchiata have been gradually cleared up by Lereboullet, Semper, 

 Salensky, Ray Lankester, and myself. 



The sexual products originate at the expense of the endoderra. 

 Sexuality can only be attributed to one embryonic lamella. — Coinptes 

 7?rM</»s, January IS, 1875. p. 19(). 



Ann. ii- May. X. Htst. Sev. 4. Vol. xv. 31 



