BEHAVIOUR 373 



erection of the crest took place when air was blown against the float with bellows. Wilson (1947) found 

 that splashing seawater on the float was also effective, although according to Bigelow only rain-water 

 elicits the response. In the Canary Islands it was observed that in specimens brought into the calm 

 of the laboratory, the crest usually collapsed and the float became flaccid. In specimens exposed to 

 the wind, on the other hand, the crest became erect. The erect condition of the crest is shown in 

 PI. XXVI, fig. 1. 



I 



t-J^i*/ 



Text-fig. 1. Diagram of Physalia, seen from above, to show the sailing posture. M = direction of movement, W = wind 

 direction, a = apical pore end of float, b = bulge, c = oral end, 5 = sail (crest). The arrows around the bulge (b) represent 

 the forces of drag caused by the tentacles. 



The way in which the wind acts has not been determined. Nerve-cells equipped with sense-hairs 

 (to be described below) occur in the ectoderm of the float, and may be concerned in the reception of 

 wind stimuli. Alternatively the buffeting and agitation caused by the wind or spray, or a combination 

 of such factors may evoke direct responses in the float musculature, leading to crest erection. It 

 should be explained that erection of the crest occurs automatically whenever the float musculature 

 achieves a certain degree of tension. Any stimulus evoking an overall tightening of the float muscula- 

 ture will therefore normally lead to crest erection. The mechanism is surprisingly simple. The float 

 consists of two containers : an outer muscular case, the codon (strictly : ' pneumatocodon ') and an inner 

 air bladder, the saccus ('pneumatosaccus'). The saccus lies free in the coelenteron, being surrounded 

 by the codon and attached to it only in the region of the apical pore. The saccus is an inert, pliable 

 structure whose shape is determined by that of the enclosing codon. Electric shocks applied to its wall 

 were found to evoke no visible response, either local or general. Along its upper side it bears a row of 

 thin- walled outpushings which become inflated whenever the air in the saccus as a whole is sufficiently 

 compressed. At other times, the outpushings are collapsed and inconspicuous. This can be demon- 

 strated experimentally by removing a saccus intact from a specimen, ligaturing its apical connection 

 with the codon as a precaution against loss of air, and placing it under pressure by means of weights. 

 In PI. XXVI, fig. 2 a saccus is shown under pressure, with the outpushings inflated. In PI. XXVI, 

 fig. 3 the same saccus is shown without the weights. In normal circumstances, the necessary 

 pressure is engendered by contraction of the enclosing codon. The dorsal processes of the saccus fit 

 into pockets (Text-fig. 2) in the roof of the codon causing them to fill out, and stand erect. The 

 arrangement is reminiscent of a pneumatic tyre, with inflatable inner tube. The enclosed air functions, 

 indeed, as a 'pneumatic skeleton', comparable in some ways to the 'hydrostatic skeleton' of actinians 

 (Chapman, 1949). 



