3ii 



FLOATATION 



From this figure it will be seen that the crest is of considerable thickness, and that there are depres- 

 sions where the lower edges of the septa meet the sides of the float at the sites of the original fusion of 

 pockets. For this reason the sides of the crest have not got plane surfaces. 



The method of formation of the transverse septa which hang down inside the crest of the float can 

 be deduced by seriating growth-stages. In a young specimen, No. Juv. 5 (Text-fig. 2), which has a 

 float-length of 19 mm. (artificially stretched some 3 mm. by injection of fluids into the pericystic 

 space), the upper surface of the pneumatocyst itself, now withdrawn from the crest, shows the 

 beginnings of six dome-shaped pockets. At the apex of each pocket are the collapsed rudiments of 

 later-developed twin pockets, themselves showing the beginnings of subdivision. Development of 

 the crest is less marked at the ends than in the centre of the float. 



POCKETS OF CODON 



GG 

 E££3 GAS GLAND 



1223 PROJECTIONS OF CYST 



Text-fig. 2. Physalia physalis. Young left-handed specimen, number 5, x 8. Float-length 19 mm. To show origin of the 

 crest as a fusion of crest-pockets. A, B from to windward, C, from above. PZ = protozooid at oral end, GG = gas-gland, 

 S = septum, 1 = before fusion, 2 = after fusion. 



Part of the upper wall of the float is pushed out by the expanding muscular projections of the 

 pneumatocyst, and the anterior wall of each pocket gradually fuses with the posterior wall of the 

 adjacent pocket, forming a septum (Text-fig. 2), which grows deeper as the fusion-process proceeds 

 with age. The secondary, tertiary and subsequent septa develop in a similar way as the finger-like 

 projections of the pneumatocyst divide and push out smaller pockets at a higher level (PI. X, fig. 7). 

 The outer longitudinal muscles seem not to be involved in the fusion-process. A section of a septum 

 shows a core of structureless lamella covered with gastrodermis. Leuckart (1851) stated that the septa 

 were formed by the inner (gastrodermic) layer of circular muscles. 



To obtain accurate data on the relationship between float-length, saccus-capacity and area of gas- 



