﻿Absolute Expansion of Mercury. 681 



of the bulb itself was measured. It is still, however, liable 

 to the third objection. In fact a difference of little more 

 than 2 per cent, between the radial and axial expansions of 

 the bulb would suffice to bring his results into agreement 

 with ours for the absolute expansion. 



Expansion of Water. — By testing four of the same bulbs 

 filled with water, Messrs. Thiesen, Scheel, and Sell obtained 

 a comparison between the expansion of water and mercury 

 in which the expansion of the glass was eliminated. As- 

 suming the expansion of mercury between 0° and 100° 0. to 

 be '016245, they d duced the expansion of water between 

 the same limits. But the converse method does not appear 

 to be satisfactory for deducing the expansion of mercury, 

 because the errors of observation or manipulation (arising 

 chiefly from the water results at 100° C.) covered a range of 

 no less than 28 units. Even the mean results for each bulb 

 quoted by Scheel and Heuse can hardly be said to agree 

 satisfactorily for this purpose, since they differ by 13 units, 

 which would be an excessive error for a mercury weight 

 thermometer, apart from the uncertainty of the expansion of 

 the bulb. 



The absolute determinations of the expansion of water by 

 Thiesen over the range 0° to 100° C. were made with a single 

 pair of columns 2 metres high and nearly 2 metres apart. 

 The temperatures of the columns were taken at two points 

 with mercury thermometers. The variation of the results at 

 J 00° C. covered a range of 64 units, corresponding to a 

 variation of only 0*13 mm. in the manometer. In our 

 mercury apparatus six pairs of columns were employed, giving 

 an effective height of 11*6 metres, and the variation of the 

 results for the fundamental interval covered a range of only 

 5 units, corresponding to O06 mm. Thiesen's mean results 

 for water exceeded those found bv Chappuis with the verve 

 dur bulb by 22 units at 100° C, and by 11 units at 40° Q. 

 Considering the extreme difficulty of the determination of 

 the absolute expansion of water, and the additional uncer- 

 tainty of the weight thermometer comparisons, it does not 

 appear that so much weight should be attached to this indirect 

 method of deducing the expansion of mercury, as the ingenious 

 mode of presentation adopted by Messrs. Scheel and Heuse 

 might seem at first sight to indicate. 



Expansion of Silica. — We have ourselves supplied the only 

 serious objection adduced by Messrs Scheel and Heuse which 

 rests on the close agreement of Harlow's silica bulbs with 

 the verve dur bulb of Chappuis for the fundamental interval, 

 when the cubical coefficient of expansion is assumed to be 



