312 DISCOVERY REPORTS 



gland it would be necessary and far more easy to work on living material. However I have arrived at 

 some rough conclusions after measuring material fixed in formalin. The relationship between float- 

 length and capacity of a number of specimens was noted and a graph made.* Judging by the way in 

 which the capacity-curve flattens out at the higher values, it might be supposed that the maximum 

 float-length would not be more than about 23 cm. (9 in.), though Quatrefages's famous specimen 

 (1854) measured 25 by 10 cm., or nearly 10 in. in length. I am indebted to Mr B. S. Kisch for a 

 record of a 12-in. specimen washed ashore on 15 September 1957, at Erromardi, Bay of Biscay, at 

 about the same time as the Channel swarm noted on page 315, but I have no data as to capacity in this 

 specimen. When measuring the float-capacity of fixed specimens I withdrew and pumped back air 

 several times with a hypodermic needle and syringe, but this method cannot be as satisfactory as 

 working on living specimens. Even the measurement of float-length of fixed specimens cannot be 

 made satisfactorily owing to frequent distortion due to preservation. 



The relationship between the area of the gas-gland and the capacity of the air-sac appears to be 



roughly as shown in Table 2 : 



Table 2 



Area of gas-gland (mm. 2 ) 



17 



300 



1000 



1500 



Capacity of air-sac (c.c.) 

 0-5 

 5° 

 150 

 250-300 



Text-fig. 3. Physalia physalis. Gas-glands of three larvae seen in optical section through the float-wall. 

 A, x 32, B, x 70, C, x 45. ect" = secondary ectoderm. 



In the early stages the gas-gland and the air-sac itself resemble those of physonect siphonophores. 

 The secondary ectoderm forms a very deep layer which projects well into the air-sac itself as a solid 

 plug (Text-fig. 3). A section of an 1 1 mm. -long specimen at this stage was published by Okada (1935). 



It is remarkable that a simple epithelium can build up a gas-pressure. But Physalia is not alone in 

 its ability to do this, for although in most fishes with a closed swim-bladder this build-up against a 

 pressure-gradient is achieved by a counter-current capillary system — the rete mirabile — it has been 

 shown by Sundnes, Enns and Scholander (1958), in at least three species of salmonid fish, that the 

 pressure build-up of oxygen and nitrogen does not take place in the blood but is a function of the 

 epithelium lining the swim-bladder. 



* The data and graph, unfortunately, have since been mislaid. 



