210 



PROCEEDINGS OP THE AMERICAN ACADEMY 



have been corrected for the variation of the zero point. At the close 

 of the investigation the zero points were again tested and found to be 

 unchanged. 



The barometer used was a large standard instrument of the familiar 

 mountain form, made by Greene of New York. Hanging at the side 

 of it was a smaller instrument of the same maker, whose indications 

 agreed with the first within the tenth of a millimeter. Here again, as 

 in the case of the thermometer, there can be no question as to the ac- 

 curacy of the differences of tension observed within moderate limits ; 

 but how far the absolute heights can be compared with those observed 

 by Regnault and others is also a problem that cannot be readily solved, 

 and one not within the means of a single experimenter to determine. 

 All heights of barometer given hereafter have been reduced to 0° by 

 the tables of Guyot.* 



Tare of the Empty Balloon by Rf.gnault's Method. 

 In the method of experimenting devised by Regnault the tare of the 

 empty glass globe used, although not explicitly given, was implicitly 



determined. His arrangement for filling 

 the globe, shown by Figure 1, has already 

 been referred to, and his general method 

 of experimenting consisted in taking the 

 tare of the globe when filled with the same 

 gas at the temperature of melting ice under 

 as different tensions as possible. The ten- 

 sions were found by measuring with a 

 cathetometer the difference of height, a ft 

 (see Fig. 1 ) , under each condition, and the 

 tares were found by means of the balance 

 by the system of compensation already de- 

 scribed. The difference of tare ofi'ves the 

 weight of gas which the globe contains at 

 0° at the tension H — A, and from this 

 may readily be deduced the weight at any 

 other tension ; as, for example, at the stand- 

 ard tension of 760 mm. The calculated 

 weight at either of the tensions H or h, deducted from the corre- 

 sponding tare, gives what we have called the tare of the empty globe ; 

 but obviously with Regnault's method this value is not needed in the 

 calculations of specific gravity or density. 



Fig. 1. 



* Published b}' tlie Smithsonian Institution, Washington. 



