The samples for liquid phase thermal reaction studies were prepared in Pyrex capillary tubing 2.5 mm. i.d. and about 15 cm. long. In a few experiments the tubes were made from standard 6 mm. i.d. Pyrex tubing of 1 mm. wall thickness. Both types of tube withstood the pressure of approximately 20 atmospheres exerted by the carbon tetrachloride at 220 `. The photochemical reaction cells consisted of 10 mm. i.d. Pyrex tubing, 5.5 cm. long, diffraction effects being minimized by the fact that the light passed through only liquid glass interfaces and not gas glass interfaces. These cells were used rather than square Pyrex tubing because of the tendency of the latter to shatter when thawing frozen carbon tetrachloride. The round cells were reproducibly positioned in the light beam which entered the thermostated mineral oil bath through a window. Two types of light source were used, a thousand watt projection lamp and an AH6 high pressure mercury arc. The light was filtered by the soft glass window of the thermostat thus ensuring that only light absorbed by the chlorine and not by the carbon tetrachloride could enter the reaction cell. Relative incident light intensities were measured with a thermopile potentiometer system. Changes of intensity on the cell were achieved by use of a wire screen and by varying the distance of the light source from the cell.

Following reaction the cells were scratched with a file and opened under a 20% aqueous sodium iodide solution. Carrier ** f was added and the aqueous and organic phases were separated (cells containing gaseous reactants were immersed in liquid air before opening under sodium iodide). After titration of the liberated ** f with ** f, aliquots of the aqueous and of the organic phase were counted in a solution type Geiger tube. In the liquid phase runs the amount of carbon tetrachloride in each reaction tube was determined by weighing the tube before opening and weighing the fragments after emptying. The fraction of exchange was determined as the ratio of the counts/minute observed in the carbon tetrachloride to the counts/minute calculated for the carbon tetrachloride fractions for equilibrium distribution of the activity between the chlorine and carbon tetrachloride, empirically determined correction being made for the difference in counting efficiency of ** f in ** f and ** f.