830 



flange plate (fig-. 7 and 9;. The llani^e plate is pressed tiglit bv bolts 

 and nuts. 



We must discuss the piece J/, on which the glass measuring tube 

 is fastened, somewhat more fully. This fastening seemed an insuperable 

 difficulty for a long time. We have said up to now that the pressure 

 inside and outside the measuring tube was the same; this is true, 

 however, only by approximation. For the way R Q Af N is filled 

 ^vith mercury, the way RS^S^S^S^ with oil, and the difference of 

 height of rhe measuring tube and R certainly amounting to 1.5 m., 

 there prevails a pressure inside the measuring lube of about 2 atm. 

 less than outside it. The different kinds of cement w^hich we used 

 to fasten the measuring tube on ^'^(Caillktkt cement, with or without 

 shellac, sealing wax, also packing material put between, such as 

 ivory) were always cracked l>y ilie pressure, even though sometimes 

 only invisible cracks arose. On account of the excess of pressure 

 outside, oil then entered the tube from the outside, and rendered the 

 measurements impossible by contamination of the mercury. 



At last it was resolved to platinize the measuring tube over some 

 centimeters' distance on the bottom side, to coat this with copper, 

 and then to solder it to the copper tube k fastened on the steel piece 

 N (Fig. 10). Now it seemed that a soliïtion was found, but a new 

 difficulty i)resented itself. When pressure brought the meirury into 

 the tube, it could come in contact with the tin when it passed the 

 place of soldei'ing, and the amalgam formed contaminated the 

 measuring tube. To prevent this the tube was lengthened at the 

 bottom by a conical piece /, which litted in a conical part of the 

 steel pi[)e X. Though it was tried, besides, to improve this closure 

 with zapon lac, the place of soldering was not yet sufficiently 

 protected. We then drew out the measuring tube some centimeters 

 into a point m. When the gas is compressed in the tube, part of it will 

 be enclosed in the small space n outside the drawn-out point. We must 

 then for the present give np the thought of a tiuantitative transferrence 

 of a quantity of gas. Besides, care should always be taken that 

 during the measurement of isotherms the mercury does not get below 

 the drawn-out point in, because then the quantity of gas in the 

 measuring tube might change. But that the soldering place now 

 remains separated from the mercury by gas up to high pressures, 

 is at present an indispensable advantage. 



In the steel piece J/ there are four passages /;, and thro.ugh each 

 of them passes an insulated wire. The passages end at the top in a 

 conical widening. In this fits a conical ivoi-y ring r, and in this a 

 copper cone q. The wire insulated by the passage is soldered to the 



