( 637 ) 



the connections bnt the individual parts are drawn to scale, it can 

 be compared with PI. I of Conim. n". 69. For those parts which 

 correspond the same letters are retained, wiiere an alteration has 

 been made the letters are accented, while new parts are characte- 

 rised by new letters. A detailed description is lience nmiecessary, 

 bnt it may be noted that tlie screw head a^ is chaniied, Ihal a 

 closed nut screwed on at t\^^ has been added by whicli the pressure 

 can be suddenly released if necessary, and tliat a cock Cj^ has been 

 introduced, to enable tlie level glass to be sliiil oil" if required. 



The compression tube A\ is desigued for use iu tlie first [)lace 

 with piezometers of tlie kind described in § 2 but of larger dimen- 

 sions. The nse of this tube A\ is then the same as the original 

 A^ (comp. §§ 3 and 4) and it may be introduced (Urectly in place of 

 this into the system shown on PL I of Comui. u". 69. 



In the second place this compression tube serxes to hold glass 

 tubes with a stem b^ — 65 (cf. PL II fig. 2 Comm. n". 69) onto which 

 other apparatus can be screwed in place of the simple nut and 

 capillary shown there. In PL I fig. 2 is shown a tiiree way cock 

 with two steel capillaries ^// and ƒ//' which is employed as follows. 



One of the capillaries <// is connected directly- with the small 

 measuring piezometer of the type of ƒ fig. 2 PL II Comm. N". 69 

 the other c//' with a volumenometer, so that when k^^^ is shut and 

 ^'lao and /i'i2 ^I'G open a known quantity of gas can be brougiit 

 into the compression cylinder from the volumenometer. On the other 

 hand when k^^^ is open and k^^„ sluit this gas can be compressed 

 into the piezometer where the temperature and pressure are measured 

 as before. The large glass tube with stem and the piezometerreservoir, 

 form in this way a piezometer of variable xolume (constant quantity) 

 and the difference with the former method consists in the measure- 

 ment of the normal volume in a volumenometer instead of in the 

 piezometer itself. The volume of the large glass tul)e iji this 

 method is not required to any high accuracy and the small ir tube 

 at the bottom may be omitted {c/^ Comm. n". 50 PL 1 fig. 4 June '99 

 and b^ Comm. n". 69 PL II March 'OJ). Tlie accuracy is now 

 really that obtainable with the volumenometer (cf. § 20) in so far 

 as the determination of the normal volume is concerned. The spaces 

 ^001' ^'130' ^"120 of the small three way steel cock must be also 

 accurately calibrated. Care is taken also that the pins really shut 

 properly into the sockets xvhicli makes the whole absolutely trust- 

 worthy up to at 100 At. 



Although we wish to confine ourselves to the method of variable 

 volume (constant mass) a second measurement with the volumeno- 



