336 DISCOVERY REPORTS 



In specimen K 2, which has a float-length of 1 2 mm. and is very well preserved, it appeared that there 

 were lateral groups present in the first group of cormidium I. I therefore made camera-lucida sketches 

 of both cormidium II and of group i of cormidium I of K 2. These are reproduced in Text-fig. 16. 

 They show that group i has developed, beside a bud of basal leeward groups, at least two lateral 

 groups, i A and i B . This condition is very exceptional. 



GON 



Text-fig. 17. Physalia physalis. Oro-lateral view of the base of cormidium II, specimen number Lanzarote 2, to show 

 positions from which complex, secondary, basal branches were budded. Branch A was cut off and is shown in PI. XVII, fig. 2. 

 Branch B was cut off. Only the bases of branches C and D are shown. These four branches gave rise to 54 tripartite groups. 

 A scar between the bases of tentacles 1 and 2 marks the position from which the gonodendron of group 1 was detached. 

 The bases of gonodendra 2 and 3 are indicated, groups 4-1 1 are not shown; they would curl round to the left behind. 

 C lt 4 , 4 = number of end-sections of branch C on three separate branchlets. 



In the course of re-examination I confirmed that, in all other specimens, these branches of 

 cormidium I originate from the compound joint pedicels of group 2 et seq., and that group 1 is a 

 simple, isolated tripartite group (see page 329). In K 2 on the other hand, group 1, though isolated, is 

 exceptional in having rudiments of leeward groups and lateral groups. In all the other thirteen 

 specimens of about the same age sent me by Miss Dodge this isolation of group 1 of cormidium I can 

 be clearly seen. It was this isolation which led me to suspect the presence of two cormidia anterior to 

 cormidium II bearing the main tentacle. 



