226 SCIENCE PROGRESS 



When a satisfactory tube has been found, the length of the thread 

 is recorded for the position nearest to which the meniscus will 

 be formed, and the mercury is run out and weighed. The radius 

 of the tube can then be calculated from the equation r = ■\/mf7rpl, 

 where m is the mass of the mercury, p its density at the room 

 temperature, and / the length of the thread corrected for the 

 meniscus at each end by a quantity which can be obtained from 

 a table. By means of a very fine sharp file, a small ring is then 

 cut off the end of the capillary and is placed on the stage of a 

 microscope with a micrometer eyepiece, and the diameter is 

 measured in four directions at 45 degrees to find if it is truly 

 circular, the same standard as before being required. If the 

 cross-section is found to be satisfactory, a divided millimetre is 

 placed on the stage, and the micrometer readings are reduced 

 to millimetres. The diameter thus found can be compared with 

 that found by weighing the mercury, the preference being given 

 to the direct measurement in case there is any slight difference. 

 Ramsay and Shields {J.C.S., 1893, T. 1089) then proceeded as 

 follows : A short tube of thin-walled glass containing a spiral 

 of iron wire was sealed on to a short piece of glass rod i mm. in 

 diameter. The capillary was dipped in ether, inclined so as to 

 fill it with the liquid, and one end was sealed ; the open end was 

 then held so as to touch the glass rod, and a very small pointed 

 flame was directed on to the junction ; the glass was then sealed 

 on, and the vapour pressure of the ether blew a small bulb at 

 the bottom of the capillary and a minute hole in the side of the 

 bulb. The sealed end was then cut off, and a thin ring of glass 

 cut off for the determination of the radius. The capillary was 

 then placed in the wider tube and the liquid was added. The 

 wider tube was constricted at the top, and attached to the pump 

 in order to boil out all the air ; it was then sealed off at the 

 narrow portion. Thus Ramsay and Shields' determinations 

 were all made with the liquid in contact with its own vapour. 

 Renault and Guye (/. Chim. Phys., 1907, vol. v, p. 81), used the 

 same method, but carried out their determinations in the pre- 

 sence of air as being much simpler — the difference in the values 

 obtained was about 0*5 per cent. In carrying out a determina- 

 tion, a scale was attached to the front of the wider tube and this 

 was placed in a vapour-bath. The capillary could now be 

 lowered by means of the attraction of the electromagnet on the 

 iron wire spiral attached to its lower end. It was lowered so 

 as to allow the liquid to flow over the top, thus renewing the 

 surface before each reading and ensuring the wetting of the 

 glass above the meniscus ; the capillary tube was allowed 

 to rise until the meniscus came to a point 2 mm. below the 

 end of the tube, and the height was then read by means of a 

 cathetometer and the scale in front of the glass. As an example 



