226 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 1912. 



done by adding small pieces of silica or by volatilizing a trace of it 

 with the blowpipe. This operation requires about an hour. 



My hearers will now understand that we have a balance capable of 

 indicating a difference in weight of about two or three niillionths of 

 a milligram. 



In order to conceive of that, let us consider first that the bulb 

 contained a volmne of air which, at normal pressure and at 0° C, 

 weighed 0.027 milHgram. Each milHmeter of the pressure gauge 

 corresponds, consequently, to °j%l'^ , or 0.0000355 millogram = 3.5 

 hundred-thousandths, but each interval of ten divisions of the scale 

 on which the hght of the Nernst lamp is reflected from the mirror 

 of the balance corresponds to 1 milUmeter of pressure. Conse- 

 ciuently, one division of the scale registers three-millionths of a 

 milligram. As it is easy to read to the tenth of a division, we should 

 be able to determine a variation of three ten-miUionths of a milligram. 

 As a matter of fact the sensitiveness is not so great. As I have already 

 said, we may take it at two to three milUonths. 



Having now explained the construction and operation of the 

 balance, I should like to say a few words as regards our experiments 

 with niton, the density of which we have determined. 



Let me remxind you that we contrived a method by which a gas 

 can be imprisoned in a capillary tube. Before making experiments 

 with the precious niton, we tested the balance by weighing xenon, 

 and as I possess 100 cubic centimeters of it, it v/as easy to afford 

 half a cubic milhmeter. 



Having measured 0.0977 cubic millimeter of xenon in a capillary 

 tube, we sohdified it at the top of the tube with a little cone of moist 

 blotting paper filled with hquid air. It was sealed by turning a 

 little flame on the tube above the cone, and placed in a small ' 'bucket" 

 hung on the balance. It was counterpoised and the pressure at 

 which the spot of hght was at zero of the scale was observed. To 

 open the little tube without losing sphnters of glass we removed 

 with platinum-forceps the ''bucket" which it exactly fitted, and 

 we pushed it into the bucket so that the pointed end was broken. 

 We then replaced the bucket and tube on the balance. To remove 

 the xenon from the tube, a vacuum was made in the box several 

 times, and by means of the httle air-bulb by changing the pressure 

 we ascertained the loss in weight corresponding to the xenon. The 

 change of pressure was 17.1 millimeters (70-52.9), which corre- 

 sponds to 608 millionths of a milligram. 



But that does not give the tine weight of the xenon, for the tube 

 is filled with air at a pressure of 52.9 miiUmeters, and as the volume 

 of the small tube is laiown, its v/eight can be estimated. It is 46 

 milhonths, so the total weight is the sum of the two, or 654 mil- 

 lionths of a milhgram. There is even yet a correction to be made. 



