58 



PEOCEEDINGS OP THE AMERICAN ACADEMY 



TABLE OF WEIGHTS. 

 First Series. 



The balance here used is only sensitive under such a load to the 

 tenth of a milligramme, and hence the constancy of the results obtained 

 is very striking. There can be no question that the mean of the sev- 

 eral weighings is accurate to the full limit of the sensibility of the 

 instrument used. Moreover, during the course of the observations it 

 was also evident that the slight variations observed (only j\ of a 

 milligramme between the extreme limits) were owing to the want of 

 exactness in the measure of temperature of the balance-case. "VYe 

 used a common thermometer reading only to whole Centigrade degrees ; 

 although, as afterwards appeared, a diflference of only ■j'^g of a degree 

 caused a variation of -^^ of a milligramme in the weight, and a differ- 

 ence of a whole degree corresponded to the extreme difference between 

 the observations. In the Second Series (made with the same tube, 

 but differently mounted) we used a standard thermometer (Geisler's 

 make) reading to tenths of a degree, and the results are all that could 

 be expected with the instruments emi)loyed. The observations were 

 made, nevertheless, under the most unfavorable conditions, during 

 exceedingly hot weather, when the temperature was rapidly changing ; 



