NEW SPECIES OF EDWARDSIA FROM NEW GUINEA. 527 



In 2 the megacoele has been divided peripherally into two microcoeles by 

 the formation of the micromesentery a. The primary tentacle of the original 

 undivided megacoele is marked /, and the new tentacle i has grown out from 

 the more dorsal microccele. In 3 the megacoele has been subdivi<led into 

 throe microcoeles by the formation of the micromesentery h ; the primary 

 tentacle remains in the middle, and the new tentacle 3 has sprouted from the 

 ventral microccele. This is the condition found in all the dorso-lateral, 

 lateral, and ventro-lateral megacoeles in E. rakaiya'. 



In 4 a new micromesentery c has been formed on the dorsal side of the 

 primary tentacle, and in connection with it the new tentacle ^ has been 

 formed. In the next stage (text-fig. 2, 5) a new micromesentery d has 

 been formed on the ventral side of the primary tentacle, and the new 

 tentacle 5 has been formed in connection with it. This is the condition 

 of the right dorso-lateral megacoele of specimen B and of both dorso-lateral 

 megacoeles of specimen C'. The formation of two new micromesenteries, 

 with their corresponding tentacles, one on the dorsal side and one on the 

 ventral side of the ])rimary tentacles, will produce the arrangement found in 

 the left dorso-lateral megacoele of specimen B. 



My examples A and C of E. carnea were not so fully expanded as 

 example B, and their peristomes being crumpled it was not possible to 

 determine the lengths of Ihe micromesenteries with the same accuracy 

 in sections ; but, so far as I was able to observe, they showed the same 

 arrangement as has been described above. 



In the ventro-lateral megacoeles the order of appearance of the micro- 

 mesenteries and tentacles is the reverse of that described for the dorso- 

 lateral megacoeles, and in the lateral chambers it appears that the first 

 micromesentery may appear on the dorsal side of the primary tentacle, 

 as in E. claparedii, or on the ventral side, as in E. n-iUeycwa and heaii- 

 tempsii. 



It will be observed that the inferences drawn from the study of the 

 length of the micromesenteries in a single species confirm in every respect 

 the inferences drawn from the relative lengths and insertions of the tentacles 

 in a number of other species of the genus Edivardsia. 



Summing up the preceding argument, the lav/ of the succession of the 

 mesenteries and tentacles in the Edward sia? may be stated as follows : — 



After the establishment of eight macromesenteries dividing the coelenteron 

 into eight megacoeles, each of the latter is produced in the peristomial reoion 

 into a single primary tentacle. 



The two directive megacoeles are never subdivided, and never prolonged 

 into more than one tentacle. 



In the capitulum the peripheral portions of the remaining megacoeles 

 are subdivided into microcoeles by the successive formation of micro- 

 mesenteries. 



