PATTERN OF BUDDING 329 



two or three centimetres show this division clearly. For study of the cormidia, I found it best to 

 dissect off the whole of the budding-tract from the float and to examine the inner side first of all. In 

 a well preserved, relaxed specimen a series of depressed openings, about a dozen in number, can be 

 seen. Five or more in the oral zone are separated by the basal internode from the main series of seven 

 in the aboral zone (PI. XIII, fig. 1). The isolated first group of Cormidium I* opens separately into the 

 pericystic space, as can be seen in specimen Lanzarote Juv B (PI. XIII, fig. 4), so that although 

 there are definitely only six cormidia in the aboral zone of this specimen, there are seven main 

 openings into the pericystic space of the float, including two such openings in Cormidium I: 

 they are the openings of the peduncles of the chief tentacles of each cormidium. Once the position of 

 these peduncles has been found (PI. XIII, fig. 1) it is possible, even in large specimens, to cut off 

 each cormidial complex entire. The openings are clearly shown in the plate. Before cutting off the 

 cormidia, it is essential to note their orientation, both with regard to the oral (protozooid) end of the 

 float and to the leeward side, otherwise analysis of the pattern of branching becomes very difficult. 



Steche (19 10) and Okada (1932) numbered the cormidia of the main aboral zone starting from the 

 aboral end; I have followed this method, both for the oral and aboral zones. In PI. XI, figs. 4, 5, the 

 ventral budding-tract removed from a specimen of float-length 10-11 cm. is shown with the cormidia 

 clearly numbered. In this specimen there are at least six oral and six aboral groups ; the seventh main 

 aboral group is not developed. Seven is the maximum number of primary main-zone cormidia that 

 I have found. 



The branching-system of typical cormidia (III and VI of the main aboral zone, and 1 and 2 of the 

 oral zone) of a specimen, number 25 from Lanzarote, were very carefully analysed. The results are 

 shown in schematic drawings (Text-fig. 12), which not only give the correct proportions of the cor- 

 midia, but the exact number of groups on the related branches. Tables 10 and 11 (p. 338) give similar 

 accurate analyses of the groups and branches of all the cormidia of both main and oral zones of 

 specimen 25. This is the first time that such an analysis has been attempted. 



Comparison of the four drawings of Text-fig. 12 will at once make clear the main pattern of arrange- 

 ment of groups in cormidia both from the oral and the main zone. 



Main aboral zone (Pis. XII-XVIII) 



In the main zone, the definitive number of cormidia cannot be determined in early stages; in 

 specimens of from 1 to 2 cm. float-length there are often only five ; in specimens of from 3 to 4 cm. 

 float-length there may be six; even in a specimen with float-length of 17 cm., taken by R.R.S. 

 'Discovery II ' in 3 22' S., 32 25' W., there are still only six cormidia in the main zone, although (as 

 already mentioned) seven is the number usually found in the largest specimens. As in the oral zone, 

 the smallest and presumably the youngest (VI or VII) lies at the oral end of the main zone. 



Careful examination of seven of the best preserved of my collection of younger specimens, measuring 

 from 1-7 to 2-1 cm. in float-length, served to illustrate the early essential pattern of growth in the 

 cormidia of the main zone (Table 10). There are three chief components of- each cormidium: (1) a 

 reduced group at the oral end consisting only of a gastrozooid with a peculiar gonodendron at its base, 

 (2) a first tripartite group, or first lateral group, consisting of gastrozooid, tentacle and ampulla, and a 

 gonodendron at the base of the gastrozooid, and (3) a further series of similar lateral tripartite groups, 

 each growing from the base of its predecessor. This is the essential arrangement which visual analysis 

 brings to light and which I have been able to record photographically. It is obscured only too soon 

 by the development of (4) secondary series of branches (leeward, oral and aboral) from the bases of 

 some of the lateral groups, bearing more tripartite groups (Text-fig. 12). 



* Arabic numerals will be used for the cormidia of the oral zone and Roman for those of the aboral zone. 



