SOMK MKKCUUY STAM'AIMiS OF RESISTANCE, ETC. 107 



had a nominal value of 1 ohm, while at C. their values were considerably 

 iliH'-Tent to the combination value of the bridge coils. 



Each of the tul>es U and M was somewhat scratched near one end. Fearing that 

 these scratch marks might develop into fractures, a length of about 1 centim. was 

 broken off each tube and the new ends ground as before. The resistance of these 

 two tubes was thus made equal to 1 ohm at about 30 C., their values at C. being 

 approximately - U7 ohm. 



The two tubes were erected as j>er Method II. already descrilxjd, and during the cycle 

 of temperature no change resulted in the purity of the mercury. Evidence of this 

 statement is furnished by observations taken during the cycle; also, the initial 

 resistance at the ice-point was identical with that observed when the cycle of 

 temperature had been completed. 



Originally it was intended to immerse the tubes in petroleum, but had this been 

 done, final observations at C. would have been impossible. Water was therefore 

 used throughout, the insulation resistance being carefully measured at each tempera- 

 ture. The lowest value obtained for this was a little greater than 3 megohms. 



The capacity of the l>ath was 8 gallons, and the lagging of cork dust was retained. 

 The liquid was well stirred by means of a vane secured at one end of the l>ath, and 

 in communication with the op{>osite end by means of a long brass tul>e. A small 

 electric motor supplied the power. The efficiency of the stirring was tested by 

 reading the temperature at different parts of the bath. The result was very 

 satisfactory. 



To secure different steady temperatures, several automatic devices seemed possible. 

 The constancy desired (maximum variation of 0'005 C.) was, however, beyond the 

 limits of any of them. Eventually, capital results were obtained by allowing cold 

 (or warm, according to the desired temperature) drops of water to fall into the bath. 

 The water was contained in a funnel supported above the box, and the stream could 

 be regulated by means of a tap. The approximate temperature was indicated by an 

 open-scale thermometer supported vertically. This was read with a cathetometer 

 telescope, a change of temperature of 0'005 C. being easily discerned. With a little 

 practice, temperatures between C. and 24 C. could be maintained for several hours 

 without the maximum variation exceeding this amount. Indeed, for the second set 

 of olservations, a temperature approximating to 5 C. was maintained for 4 hours, 

 the greatest variation in temperature being that indicated. The mercury within the 

 tul>e responded to changes of temperature very rapidly, however, and a few observa- 

 tions showed that it was only necessary to maintain a steady temperature for half an 

 hour or even less. 



It AVUS ili-irahle that the thermometer employed should l>e capable of being read 

 to 0'01 C. with ease and without the employment of a telescope. One of the 

 thermometers presented to the Laboratory by the widow of the late Mr. SWORN, and 

 juimbered 2221, answered these requirements. The stem of this thermometer is 



P 2 



