CCELENTERATA. 171 



remarkable and unexpected way. The law is expressed by Lacaze Duthiers 

 as follows. "The appearance of the new chambers is not, as has been 

 believed, a consequence of the production of a single chamber between each 

 of the twelve already existing chambers, but of the birth of two new 

 chambers in each of the six elements (chambers) of the smaller cycle." The 

 result of this law is that a pair of tentacles of the third cycle is placed in 

 every alternate space, between a large and a small tentacle, of the two 

 already existing cycles, which may conveniently be called the first and 

 second cycles (fig. 80 B). 



The twenty-four tentacles formed in the above manner are obviously at 

 first very irregularly arranged (fig. 80 B), but they soon acquire a regular 

 arrangement in three graduated cycles of 6, 6 and 12. The first cycle of the 

 six largest tentacles is the large cycle of the previous stage, but the two 

 other cycles are heterogeneous in their origin, each of them being composed 

 partly of the twelve tentacles last formed, and partly of the six tentacles 

 of the second cycle of the previous stage. 



The further law of multiplication has been thus expressed by Lacaze 

 Duthiers: "The number of chambers and still later that of the corre- 

 sponding tentacles is carried from 24 48 and from 48 96 by the birth of 

 a pair of elements in each of the 12 or 24 chambers, above which are placed 

 the smallest tentacles which together constitute the fourth or fifth cycle. 

 Since, after the formation of each fresh cycle, the arrangement of the 

 tentacles again becomes symmetrical, it is obvious that all the equal sized 

 cycles except the first are formed of tentacles entirely heterogeneous as to 

 age." 



The fixation of the free swimming larva takes place during the period 

 when the tentacles are increasing from 12 to 24. 



The general formation of the chambers in Bunodes and Sagartia is 

 nearly the same as in Actinia. 



In the two types of Actinozoa with an embolic gastrula stage the laws 

 as to the formation of the tentacles do not appear to be the same as those 

 regulating the forms observed by Lacaze Duthiers. 



In Cerianthus four tentacles are formed simultaneously at the period 

 when only four chambers are present. In Arachnitis (Edwarsia) the suc- 

 cession of the tentacles is stated (A. Agassiz, 166) to resemble that in 

 Cerianthus. There are originally four tentacles, and at one extremity of the 

 long axis of the mouth are the oldest tentacles, while at the other tentacles 

 are constantly added in pairs. An odd tentacle is always found at the 

 extremity of the mouth opposite the oldest tentacles. 



In the other species with an embolic gastrula eight tentacles would seem 

 to appear simultaneously at the period when eight chambers are present ; 

 though on this point Kowalevsky's description is not very clear. The 

 presence of such a stage would seem to indicate a close affinity to the 

 Alcyonidas. 



Amongst the sclerodermatous Actinozoa, except Caryophyllium, the 

 embryo closely resembles that of the delaminate Malacodermata. The first 



