The only way of removing the complication arising from thermal radiation lies in varying the temperature condition of the plant, by direct contact with water at different temperatures. This procedure will also remove uncertainty regarding the body of the plant assuming the temperature of surrounding non-conducting air. The disturbing effect of sudden variation of temperature is also obviated by a more uniform regulation of rise of temperature. The inner cylinder containing the plant is filled with water; heat from gradually warmed water in the outer cylinder is conducted across the inner cylinder made of thin copper and raises the temperature of the water contained in the inner cylinder with great uniformity. A clock-hand goes round once in a minute; the experimenter, keeping his hand on the stop-cock, adjusts the rate of rise of water in the inner cylinder, so that there is a rise, say, of one-tenth of a degree every 6 seconds or of one degree every minute. The mass of water acts as a governor, and prevents any sudden fluctuations of temperature. The adoption of this particular device eliminated the erratic changes in the rate of growth that had hitherto proved so baffling. The elongation recorded by the Crescograph will now be made up of (1) physical expansion, (2) expansion brought about by absorption of water, and (3) the pure acceleration of growth. The disentanglement of these different elements presented many difficulties. I was, however, able to find out the relative values of the first two factors in reference to the elongation of growth. This was done by carrying out a preliminary experiment with a specimen of plant in which growth had been completed. It was raised through 20°C in temperature, records being taken both at the beginning and at the end. This was for obtaining a measure of the physical change due to temperature, and also of the change brought about by absorption of water. I should state here that for the method of continuous record of growth which I contemplated, the record had to be taken for about 18 minutes. The magnification had to be lowered to 250 times to keep the record within the plate. With this magnification, the fully grown specimen did not show in the record a change even of 1 mm. in length in 18 minutes, while the growing plant under similar circumstances exhibited an elongation of 100 mm. or more. In records taken with low magnification, the effect of physical change is quite negligible. DETERMINATION OF THE CARDINAL POINTS OF GROWTH. The cardinal points of growth are not the same in different plants; they are modified in the same species by the climate to which the plants are habituated; the results obtained in the tropics may thus be different from those obtained in colder climates. At the time of the experiment, the prevailing temperature at Calcutta in day time was about 30°C.