In order to reduce the carbonate content, the alkaline KI solution was made fresh every few days from two stock solutions. Carbonate free NaOH solution was made by dissolving 100 grams of NaOH in 100 c.c. in a glass-stoppered bottle and pipetting off after the car bonate had settled. One part of this was mixed with one part of 20 per cent KI solution before being used. NaOH made from metallic sodium contained a trace of nitrite and NaOH purified by alcohol con tained but little more. This nitrite causes no error if the titration is quickly made immediately after adding the acid, but the contents of the flask slowly turn blue for hours after the end-point has been reached. If the acidity is greatly increased, however, the nitrite causes an appreciable error in the titration.
The solution contained a trace of Mn(OH)3, but this was removed by decantation. The thiosulphate was dissolved in CO2-free water and kept in an automatic burette with soda-lime tubes, and stand ardized with pure iodine. The starch solution was allowed to settle and only the clear upper portion used. When dextrins appeared in it a fresh solution was prepared.
Although pH determinations and calculations were made in all experiments, it was found that the oxygen absorbed could be deter mined much more accurately than the CO2 given out, and the CO2 determinations are not listed with the experiments, but are collected together in the form of respiratory quotients. The respiratory quo tients were 0.7, 0.74, 0.76, 0.84, 0.85, 0.86, 0.88, 0.9, 0.91, 0.92, 0.97, 0.99, 1, 1.02, 1.03, 1.1, 1.15, 1.2. It was impossible to tell whether the variation is due entirely to errors in the determinations or whether the respiratory quotient varied. It is improbable, however, that respiratory quotients of 1.2 existed for even short periods of time, and these at least may be considered due to technical errors. Accord ing to Mayer (1914), Cassiopea lives on animal food exclusively and does not absorb carbohydrates from its symbiotic algae, since it starves as rapidly in the light as in the dark. It may, however, get some carbo hydrates from its animal food, or from glycoproteins during starvation. In this connection it may be of interest to note that Cassiopea secretes a mucin-like substance. Since the error in estimating production may be 30 per cent in half-hour experiments with small cassiopeas, it is convenient to assume that the respiratory quotient is constant and is about 0.95, which is also the average found by Vernon for the hardier
species of jelly-fish, on which he made most of his determinations.
The oxygen consumption is about doubled when the temperature is raised from 20° to 30°, and from results on other animals is probably an exponential function of the difference in temperature. Harvey (see Mayer, 1917) found the velocity of the nerve-impulse in Cassiopea to be a linear function of the temperature and to increase about 64 per cent on raising the temperature from 23° to 33°. I found the activity of the ganglia (rhopalia) in inducing pulsations of the umbrella to be about doubled with rise from 20° to 30° in temperature. The data are given in table 12.
These experiments show the necessity of accurate temperature con trol, and in all of the other experiments in this paper the temperature was measured to the nearest tenth of a degree and maintained to within 0.2° of 30°.
The sea surface is at about the optimum pH for metabolism (usually 8.1 to 8.3), but the variation in metabolism with variation of pH within the range studied is very slight, as shown by table 13 (diameter of cassiopea = 10 cm.).
These variations in oxidation may be due to experimental errors and variation in 02 concentration, except the first two, which show a slight lowering of oxidation when the pH is re duced to about 7.5.
In comparing the rise in temperature in the calorimeter with that calculated from the 02 consumption, the assumptions were made that the respiratory quotient was 0.95, and that a certain mixture of pro teins, fats, and carbohydrates was burned, giving 6 gram-calories per cubic centimeter of as in table 14.
In these experiments it was assumed that there was no loss of heat, although some heat must have passed into the thermometer bulb and glass lining of the thermos bottle. The specific heat of the sea-water and the cassiopea was taken as unity because the determinations were not accurate enough to warrant the application of small corrections. The experiments had to continue for 2 hours in order to obtain an accu rately measurable rise in temperature, and the tedium of keeping the tank at the same temperature as the calorimeter necessitated the substitution of indirect calorimetry in the remainder of the experiments.