septum or basis of support. It is obvious, therefore, from the condi- 

 tions under which the partitions are successively secreted, that they 

 must adhere not only to the circumference of the valve, but to the pre- 

 ceding and succeeding septum at the part occupied by the adductor 

 muscle, and for an extent corresponding to its circumference. The 

 progressive change in^the position of this muscle by the absorption 

 of the posterior fibres, and the addition of others anteriorly, changes 

 in a corresponding degi'ee the relative position of these subcentral 

 confluent parts of the septa, and a beautiful undulated disposition of 

 the whole chambered part results. If the adductor muscle were a 

 tube instead of a solid mass, the central confluent part of the septa 

 would of course be perforated, and a siphon would result, the calca- 

 reous walls of which, from the proximity of the chambers, would no 

 doubt be continuous, as in many fossil Polyihalamous shells. 



A disposition to form chambers is manifested, but in a much less 

 degree, in the smaller flattened or superior valve of the Water Spon- 

 dylus. In the specimen here described there are three chambers, 

 with narrower intervals, and much thicker partitions than in the 

 lower valve. These partitions are confluent opposite the muscular 

 impression, as in the lower valve, and each partition expands from 

 this attachment in an infundibular manner, which reminds one of 

 the emboitement of the calcareous parts of the siphon in the Spirula. 



The secreting power of the lower lobe of the mantle in the Spon- 

 dylus is greater than in the upper ; and the layers of nacre which 

 are successively deposited on the cardinal margin push forward in a 

 corresponding degree the upper valve, leaving a heel or umbo be- 

 hind the hinge of the lower valve, which, from the inactivity of the 

 secreting surface of the upper lobe of the mantle, is not opposed by 

 a corresponding umbo in the upper valve. 



The laminae, which are deposited in a continuous series of super- 

 imposed layers at the hinge of the lower valve, are not continued in a 

 like state of superposition throughout ; they soon separate from each 

 other, and do not again unite except at the space corresponding to 

 the adductor muscle, and at the circumference of the valve. 



The interspaces of these successive layers of the growing Spon- 

 dylus cannot, firom the absence of a medium of intercommunication, 

 serve any purpose hydrostatically with reference to locomotion : it 

 is a singular fact, indeed, that the Spondylus, in which the chambered 

 structure is constant, and the Ostrea, and other bivalves, in w'hich it 

 is occasional, are cemented to extraneous bodies by the outer surface 

 of the shell, generally by the concave valve. So that the septa must 

 be regarded as mere dermal exuvia; still left adhering to the animal, 

 to which, as a motionless bivalve, they are no incumbrance. It is 

 highly probable that all the chambers are originally filled with fluid, 

 as more or less is found in the outer ones of the specimens brought 

 to this country. 



In the Testaceous Cephalopods a new structure is added, viz., the 

 siphon, whereby the exuvial layers of the old shell and the deserted 

 chambers are converted into a hydrostatic instrument, subservient to 

 the locomotion of the animal. The operation of the siphon and 



