GONODENDRA 349 



terminal (final) sections and budding sub-terminal ones. The whole process was studied in the 

 developmental stages now described and illustrated on Pis. XXI-XXIII. 



I examined first the youngest stages found at the bases of the smallest gonodendra (PL XXII , figs. 1-3). 

 These genital clusters arise opposite the basal jelly-polyps as innumerable small, laterally flattened, 

 subdivided lobes, the pro-buds, arranged in many planes, and often look at first sight like trefoils. 

 They bud and rebud several times to form the ultimate branchlets of the gonodendron (PI. XXI, 

 fig. 1), as distinct from its main branches, which are formed by the elongated peduncles of the 

 gonozooids and first palpons, as described on page 347. The pro-buds of the genital clusters grow rapidly 

 in length in an apical direction and produce sub-terminal buds on one side, so that the apex becomes 

 bent to the other side. Each of these little lobes thus comes to appear bi-lobed (Text-fig. 27 A and b). 

 These two stages have not been specially identified, and are hypothetical. One half, really the faster 

 growing terminal end, can soon be recognized as the rudiment of a palpon and a jelly-polyp ; the sub- 

 terminal part is a new pro-bud, B (Text-fig. 27 b). As a result of further budding, pro-bud B develops 

 secondary dichotomies, and the new pro-buds B lt B 2 , (Text-fig. 27 d) each become transformed into 

 replicas (Text-fig. 27 e) of the previous stage, shown in Text-fig. 27 c. The terminal part of each 

 successive pro-bud grows faster than the sub-terminal part to form a new terminal palpon and jelly- 

 polyp, while the sub-terminal section can either end by becoming the palpon and nectophore, or go 

 on developing and dividing into another terminal and sub-terminal section. We are now at stage D 

 (Text-fig. 27) where there is a major dichotomy between the left 'B '-side bearing pro-buds B 1 and B 2 

 and the right 'A '-side bearing the terminal palpon and pro-bud A 2 . The two halves are essentially 

 alike, but one is farther advanced in growth than the other, for on the 'A '-side we can soon 

 recognize the rudiment of another palpon and jelly-polyp pro-bud A 2 1A (Text-fig. 27 f). As a result 

 of repeated sub-terminal budding and dichotomy we get a series of terminal sections — palpon and 

 jelly-polyp— alternating with sub-terminal pro-buds. These developmental processes which I have 

 postulated would account for the origin of the final branchlets of the whole gonodendron (Text-fig. 26). 

 Having followed the development of one of the branchlets as far as the last dichotomy, we can see 

 what is the destiny of the final sub-terminal pro-buds:* they divide for the last time to form 

 the rudiments of the sub-terminal nectophore and palpon of the sub-terminal section (PI. XXI, 



fig- 3)- 



In some cases a final sub-terminal pro-bud is not developed : that is, the final palpon and nectophore 



are not budded off, then the branchlet ends in a terminal section only, consisting of a palpon and 

 jelly-polyp. PI. XXI , fig. 7 shows a pair of final branchlets, the right with two terminal sections and the 

 left with two terminal and a final sub-terminal section. Normally there is a matching pair of final 

 branchlets with sub-terminal sections. 



I must now describe in greater detail how the middle growth-stages of the genital clusters 

 (Schneider's ' Genitaltrauben ') develop into the final branches of the fully grown gonodendra. I have 

 not determined whether there is a definite number of new buddings and dichotomies which might fix 

 the amount of branching. But if we take the hypothetical stage shown in Text-fig. 27 F as the starting- 

 point of the next stage, it can be seen at once, on comparison with PI. XXII, fig. 4, that the whole of 

 the right side of Text-fig. 27 F as well as the whole of the left side— the major dichotomy is between 

 pro-buds A 2 and B 1 — can develop into something similar to what is shown in PL XXII, fig. 5. In both 

 figures the larger palpon (p.st) and the jelly-polyp (jp) form a terminal section; and the sub-terminal 

 pro-bud (pr. A 2 2 ) above it in Text-fig. 27F has given rise in PL XXII, fig. 5 to a smaller terminal section 

 and a new sub-terminal one. From the latter arise (1) the nectophore (nect), and (2) the palpon of a 



* Such as A, 2 , Text-fig. 27F. PI. XXII, fig. 4 is the equivalent of the 'B'-side of Text-fig. 27F. 



