228 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 1912. 



gases of the flame. We removed the gases with the pump, and, by- 

 means of carbon cooled with liquid air, absorbed the oxygen introduced 

 at first to displace the air, which might have contaminated the gases 

 with a trace of neon and helium. The residue measured 0.042 cubic 

 millimeter; its spectrum was that of pure helium, and its weight 

 was eight millionths of a milligram. Adding this to the weight 

 already indicated by the balance, we obtain thirty-five millionths, 

 which differs by only three millionths from the figure calculated upon 

 the hypothesis that every atom of niton, on dismtegrating into radium 

 D, lets three atoms of helium escape — i. e., thi-ee a particles. 



Our main object was to find the true atomic weight of niton. To do 

 this we introduced, in five experiments, quantities of it varying 

 between 0.075 cubic millimeter and 0.0566 cubic millimeter into 

 density tubes as I have already described. These are the volumes of 

 the niton actually removed from the tubes with the pump. It is self- 

 evident that corrections had to be made for the losses of niton due to 

 its partial spontaneous decomposition into solid products and for the 

 part that had penetrated the walls, either under its own form or under 

 that of helium. The weights we found range between 572 and 739 

 millionths of a milligram. The atomic weights were 227, 226, 225, 

 220, and 218; mean 223. Startmg from the determinations of 

 Madame Curie and Sir Edward Thorpe of the atomic weight of 

 radium and subtracting that of helium, 4, we obtained the atomic 

 weight of niton, 222.4, a sufficiently satisfactory agreement. 



Let me describe to you briefly another experiment by Messrs. 

 Cuthbertson and Porter, of the Physical Institute, University College. 

 The object is to find the index of refraction of niton by employing less 

 than one-tenth of a cubic millimeter of this gas. At the end of a sealed 

 capillary tube they polished two surfaces parallel to the axis of 

 the tube and perforated the tube with a small hole perpendicular to 

 these surfaces. By cementing two plates of polished glass to the 

 surfaces a small chamber was constructed with a capacity of about 

 1 cubic millimeter. They platinized two small circles on the polished 

 glass so that the light could pass through the platinized part and also 

 be reflected by the metallic surfaces of the plates. This thermometer 

 tube fonneil the upper part of an apparatus such as that which I have 

 described to you, into which purified niton could at pleasure be intro- 

 duced. Looking at the green light of a mercury lamp through the 

 hole, fringes in the form of concentric circles could be seen, the lengths 

 of whose radii change with every change of pressure. By measuring 

 the number of bands which pass tlirough a fixed point for a Imown 

 change of pressure the refraction of the gas can be calculated. 



Several measurements were more or less successfully made. The 

 difficulty which made it impossible to continue the investigation was 

 entirely unforeseen. This was the corrosion of the platinum by the 



