232 ME. J. T. CUNNINGHAM ON 



type and included diversity in the shells of Cephalopoda, etc., it 

 is impossible to suppose that crystals or sphsero-crystals should 

 aggregate themselves in modes of corresponding diversity of type 

 and detail merely because they were formed in a colloid medium. 

 Moreover, we know that the form of the shell is determined by 

 the mantle, the border of which secretes an organic layer of 

 conchiolin, and this has the specific form of the shell before 

 calcification takes place in it at all. Lastly, no reason has been 

 given why the form or aggregation of crystals should produce the 

 characteristic parallel or concentric markings on molluscan shells, 

 which correspond to the edge at which growth takes place, each 

 of which has, in fact, at some previous moment in the growth of 

 the shell been its extreme edge. These markings are in fact 

 evidence that the growth of the shell is not perfectly continuous 

 but intermittent, although we do not know fully the causes of 

 this rhythmical periodicity in the growth except in the annual 

 markings. That the forms of spicules such as the three-, four-, 

 or six-rayed spicules of sponges, may be determined by the form 

 and aggregation of crystals, seems probable enough, but this is 

 quite a different question from that of the form and markings of 

 molluscan shells, which are determined by the extent and the 

 physiological activities of the shell-secreting epithelium. 



In text-figs. 3 & 4 I have given diagrams which show on an 

 enlarged scale the mode in which growth takes place in a 

 molluscan shell, and in one of the imitation shells of paraffin- 

 wax, respectively. In the mollusc the edge of the mantle 

 secretes conchiolin only, the periostracum ; and for reasons, not 

 so far as I know discovered, the growth of this edge is not 

 uniform and continuous, but is stationary for a time, and then 

 starts again not quite in the same direction, but the mantle 

 leaves the extreme edge of the periostracum projecting and 

 secretes a new band starting from the lower surface of the 

 preceding band. In certain cases, as is well known, the edge of 

 a band at certain intervals may be fringed with long spines or 

 processes, as in Cardium echinatum. When the extreme border 

 of the mantle has extended to a certain distance beyond its former 

 limit, the next internal band of the mantle comes into contact 

 with the band of periostracum just secreted, and forms on the 

 inner surface of this an addition to the prismatic layer of the 

 shell. When the next growth-movement of the mantle takes 

 place, the prismatic-forming region passes out to the new border 

 of periostracum, and the prismatic band just formed becomes 

 covered by the region of the mantle which secretes the nacreous 

 layer. The successive processes of secretion are shown in text- 

 fig. 3 A, where c is the extreme band of periostracum in process 

 of secretion, b the preceding band with the prismatic layer added 

 to it, a the band preceding b, to which a single layer of nacreous 

 shell-substance has been added. It is to be noted that in the 

 true shell, as also in the wax imitation, between the more con- 

 spicuous lines of growth, which alone are indicated in the 



