THE VOYAGE OF H.M.S. CHALLENGER. 
66 
C. ilumn II. gives the specific gravities of “ Standard Water” at the temperatures 
' nii'il in Column I., calculated by my equation. Column III. states the correction to be 
•• added " to the value ( Jolumn [I.to obtain Thorpe and Rucker’s value. Column IY. gives 
i! it Inns «»f (p(t) calculated from my constants ( i.e ., the logarithms of the ratios 
Column V. gives the correction to be applied to my logarithm <p(t) for obtaining 
• UH/ 
Thorpe and Rucker’s. Their specific gravities and values cp(t) I calculated from their 
Table XV* 
This table shows fully what discrepancies we must be prepared for if a series of 
Bpeci ities is being reduced to or from 15°*56, at one time by means of my constants, 
ml at another by Thorpe and Rucker’s. In regard to (p(t) we agree very satisfac- 
t -rilv. The greatest difference is 0 - 0017, corresponding to 17-^-4300 of the value of 
the computed correction, and consequently to in general less than about 0 ’008 units in 
the 8, which i* meant to be calculated to only two decimals, or to ±0'005. Our agreement 
iii regard to the specific gravities of “Standard Water” is less perfect; the differ- 
nces for from 0° to 2 especially going rather beyond what can be tolerated as an 
• tb e- of observational errors. Fresh experiments must show which of us has come nearest 
th- truth. Meanwhile the means of our results (i.e., the entry for <S t in Column II. plus 
half the entry in Column 111.) would presumably reduce the maximum uncertainty to 
something like rfcO’03. 
I prop"'' now to pass to an important point in connection with my experiments, which 
I ha vi hitherto kept in the background in order not to break the continuity of my 
exposition. 
Reduction of the Specific Gravities to the Vacuum. 
Whrn I tarted my specific gravity work, I did not expect to reach such a degree of preci- 
"ii as would justify the application of the vacuum-correction to every individual weighing ; 
unmanly calculated correction for the final results, I felt sure in my mind, would fully 
in- "t th*- i'i - juirements of the ease. But, in order to be able to found this calculation upon 
■mi- tb ng bi tter than mere assumption, I made periodic observations of the temperature 
ui of tin air in the laboratory. The data which I thus collected enable me now 
• - di termine limit-values for the corrections to which my values for the specific gravities, 
n ii t li<' above tables, are liable. Supposing the plunger to have been weighed (1) 
m iir, (2) in pure water of t , (3) in sea-water of t° , and the results to have been as 
follows, uncorrected : — 
Weight 
Loss of Weight as against (1). 
(1) In air, 
. (W) 
0 
(2) In pure water, 
W 0 
I* 
(3) In the sca-watcr, 
. w, 
Li 
• P-r tho information of any pen*. i, who EOftJ repeat i I I . I will state Chat .'ill tile specific gravities and 
' : (0 »vr< originally computed to 3 decimal* and 5 decimals respectively. 
