328 DISCOVERY REPORTS 



of siphonophores (thirty) which he included in his monograph, by making idealized coloured 

 drawings, even when he was not sure of the details. It is greatly to his credit that he undertook the 

 formidable task of making coloured sketches of living animals in the field, a task which would be 

 quite impossible for me. 



In 1932 G. H. Parker made some interesting experiments at Key West, Florida, on the neuro- 

 muscular activities of the tentacles. He also recorded the measured length of an extended tentacle as 

 a little over 9 m., and stated that the tentacles can shorten to one-seventieth of their maximum length. 



Polymorphism 



In Physalia there are three. types of medusoid individuals all found on the gonodendra : (1) the reduced 

 male and female gonophores, (2) the peculiar asexual nectophores, and (3) the jelly-polyps, which are 

 probably vestigial nectophores. 



Of polypoid individuals there are four types, as well as the unique protozooid with its small monili- 

 form tentacle: (1) the feeding gastrozooids which become separated from their tentacles, (2) the 

 gastrozooids of group 1 of cormidia II-VII of the main zone and of all the cormidia of the oral zone, 

 (3) the secondary and tertiary gonozooids, which like those of (2) give rise by budding to the branches 

 of the gonodendra; none of these have tentacles, and (4) the gonopalpons, which have no tentacles. 



Pattern of Budding in the Cormidia 



In spite of the large number of papers on Physalia, there is no good account of its gross morphology, 

 especially of the cormidial groups of appendages and their development. 



Anyone who tries to decide on the number and structure of these groups of buds will not be sur- 

 prised to learn, I think, that so far no one has had sufficient patience to pursue the subject to a satis- 

 factory conclusion. Without properly relaxed specimens, it was in the past a hopeless task; even today 

 with better material to study, my conclusions must be of a tentative nature. In order to reach more 

 definite results in the future, efforts should be made to anaesthetize and preserve in a thoroughly 

 relaxed state the oldest as well as the very youngest stages of development. From a morphological 

 point of view, the potential value of well-preserved, young specimens increases greatly for every 

 millimetre below ten in float-length. 



Previous workers on Physalia have tended, I feel sure, to simplify the problem by ignoring all except 

 the largest branches of the cormidia. Accordingly, having had a great number of living specimens to 

 deal with at Arrecife, on my return in 1955 I decided to make a completely fresh study of the structure 

 and developmental sequence of the cormidia. My analysis is based upon detailed examination of some 

 two dozen younger and three dozen older specimens, including those taken by R.R.S. 'Discovery', 

 and has since been confirmed by examination of the lovely series of larvae, so kindly sent me from 

 Miami by Miss Eleanor Dodge. Three years of laborious observation, dissection and photography 

 have established a basic pattern of budding in the cormidia, a pattern which would have been less 

 clear but for this opportunity of studying the early larval stages. The complexity of this pattern is 

 illustrated by the series of growth stages shown on Pis. VII-IX. In some of the larvae examined 

 (PI. VII) very early indications of budding can be seen; others already carry several buds. The juvenile 

 specimens shown on PI. VIII vary in length between 12 mm. and 80 mm., and exhibit ever increasing 

 budding, which culminates in the branching mass of cormidia typical of the adult Physalia (PI. IX). 

 The cormidia are budded from a narrow tract of tissue lying along the under and windward side of 

 the float extending from the protozooid (PI. VIII, fig. 2) to a point half-way to the pore. This tract is 

 divided by a gap — Haeckel's 'basal internode' — into two (PI. XI, fig. 5), an oral zone and a main, 

 aboral zone. The gap is particularly well marked in young stages. Specimens with float-lengths of 



