Vapour- Densities of some Carhon Compounds. 499 



stoj^cock li into F. The nu'rourv was drawn helow the level 

 of the point 1 in question : the Fleuss pump was disconnected 

 from F, and bv turning the stopcock 

 rapidly the mercury rose until it nearly 

 touched the point. Adjustment was made 

 by help of a sinole lens. During calibra- 

 tion the whole tube was surrounded with 

 water of known temperature. The average 

 error was about 1/30,000 of the total 

 volume. 



After calibration the next step was to 

 determine the vertical distance between 

 points 1 to 8 and a point (fig. 1) attached 

 to the outside of the tube below the jacket. 

 The volume-tube was set up opposite the 

 scale B. and made vertical with a plumb- 

 line, and the distances read with a telescope 

 to within 0'02 mm. In order to correct 

 these differences the linear expansion of 

 the glass was required; as a sufficient 

 approximation the cubical expansion of 

 this sample of glass (0*000029) was deter- 

 mined, and one-third taken as the linear 

 expansion. The cubical expansion was, of 

 course, also required to correct the read 

 volumes. 



The manometer consisted of a glass 

 U-tube B, 100 cms. long and 13 mms. in 

 diameter. It was filled with mercury 

 and boiled out in vacuo. This gives no trouble provided the 

 tube is carefully freed from dust by washing out with bichrome 

 and sulphuric acid, and subsequently with distilled water^ 

 before drying it. The side-tube h is provided at the highest 

 point with a stopcock through which any bubbles of air 

 can be displaced. The short limb of the U-tube is connected 

 by means of the side-tube / and thick-walled rubber tubing 

 with the mercury reservoir C. Final adjustment of the 

 mercury surface to the points in the volume-tube A is made by 

 means of a squeezer, consisting of two pieces of wood and a 

 screw. The manometer was water-jacketed, a delicate ther- 

 mometer being inclosed in the jacket. The scale was obtained 

 from Zeiss. It was tested by means of a travelling microscope, 

 reading to 0*001 mm., and was found to be consistent to 

 0*03 ; a table of corrections was made by which readings 

 were corrected to within 0*01 mm. 



2 K 2 



