280 



Scientific Proceedings^ Roi/al Duhlin Sociciij. 



In order to avoid this zero-error, deeper vessels [in and e) were iised for 

 containing the freezing mixture and for enclosing the test-tubes (see fig. 1). 

 The latter (« and h) were as before carried in a perforated cork bung (c), 

 which, however, was made to fit loosely down into the inner vessel, and 

 was supported tliere by a stout bi-ass wire {w) passing down through an 

 upper cork (rf) wliich fitted the neck of the inner vessel ie) closely. 



In this way the cold air-space en- 

 closed not only the tubes containing 

 the freezing liquids and the junc- 

 tions, but also included the whole 

 of the couple, which was thus 

 throughout its whole length brought 

 approximately to the temperature of 

 2o Galrfmcyncief? 'CnQ junctions. The rods (r and r') 

 supporting the arms of the couple 

 {J and J') were led througli holes 

 in the upper cork vertically over 

 the mouths of the tubes, so that tlie 

 fluids in the tubes could be kept 

 agitated during an experiment. In 

 order to equalize the effects of this 

 agitation, provision was made for the 

 connexion of the two upper ends of 

 the supporting rods by means of the 

 screw-clamp (s). This clamp served 

 as a handle which simultaneously 

 stirred both tubes. The small heat- 

 ing effects of stirring which have 

 been noted by other investigators 

 were in this way cut out. 



It has been elsewliere noted' that 

 Fio. 1. tlie shifting of the zero of such a 



galvanometer as must be used in these determinations is a cause of trouble, 

 and, unless specially guarded against, will give rise to inaccuracy. Two 

 kinds of shifting of the zero may be distinguished : one which occurs in the 

 galvanometer itself, even when the circuit is open, and the other which is 

 noticeable only when the circuit is completed. The first is due to the fatigue 

 of the metal-suspension, and may be observed whenever the galvanometer is 



Trans. Roy. Irish Acad., vol. xxii, sect. B, pt. iii, p. 149. 



