326 L. H. HYMAN AND A. W. BELLAMY. 



between the three phenomena in question is the object of the 

 present paper. 



II. METHODS. 



The metabolic gradients have been tested by various methods, 

 chiefly the susceptibility to various toxic solutions and to dyes, and 

 the capacity to reduce potassium permanganate. Wherever practi- 

 cal, direct determinations of the respiratory rate at different levels 

 have also been made. For a general discussion of metabolic gradi- 

 ents and of the susceptibility method see Child ('20) and of the 

 permanganate method Child ('190). 



The potential differences have been tested by galvanometers of 

 the D'Arsonval type, put out by the Leeds and Northrup Com- 

 pany. The one which has been used for most of the work is a 

 small portable galvanometer having a resistance of 310 ohms and 

 a sensitivity of 73 megohms (i volt through a resistance of 73 

 megohms gives a swing of I division of the scale). During the 

 summer of 1919, while being used by Hyman at Friday Harbor, 

 the suspension of this instrument broke and another instrument 

 having a spring suspension was borrowed ; the sensitivity of this 

 instrument was not recorded, but it was slightly more sensitive 

 than ours. Part of the data on the medusas and Nereis and all of 

 the data on ctenophores was obtained with this second instrument. 

 Non-polarizable electrodes of the zinc-zinc sulphate type were 

 used, their tips packed with kaolin and terminating in rolls of hard 

 filter paper, frequently changed, soaked in the medium in which 

 the animal to be tested lives. It is necessary to renew these elec- 

 trodes frequently, as they are apt to develop potential differences. 

 These can be eliminated by running a compensating current 

 through them as was done in part of the work. Usually the dif- 

 ference between the electrodes was eliminated by reversing the 

 animal on the electrodes. In the case of sessile animals the elec- 

 trodes were clamped in position and the animal placed across their 

 tips ; in motile animals the electrodes were held in the hands by 

 means of burette clamps and placed on the animal in the desired 

 position. It is not an easy matter to work with active animals. 

 Usually at least two readings were taken on each individual for 

 each position of the electrodes. The readings recorded in the 



