46 THE VOYAGE OF H.M.S. CHALLENGER. 



[In obtaining the first of these numbers, I assumed from Despretz that the density 

 of water at 1 atm. and — 9°C. is 0"9984.] Of course it would be vain to attempt similar 

 calculations for the stronger solutions, as the indicated maximum density points are so 

 widely outside the limits of my experiments. But the example just given seems to 

 show that if fresh water be made, by pressure, to have its maximum density point the 

 same as that of a common-salt solution under atmospheric pressure, the densities of 

 the two will be nearly the same at that point, and will remain nearly alike as tempera- 

 ture changes. 



NOTE. 



In all that precedes it has been tacitly assumed : — 



1. That the pressure is the same outside and inside the piezometer. 



2. That the pressure measured by the gauge is that to which the contents of 



the piezometer were exposed. 



3. That the pressure was uniform throughout the contents. 



None of these is strictly true, so that cause must be shown for omitting any 

 consequent correction. 



The third may be dismissed at once, as the height of the piezometer bulb is only 

 a few inches. 



The difference of levels between the upper end of the gauge and the bulbs of the 

 piezometers, when in the pressure -chamber, was about three feet, so that on this 

 account the pressure applied was less than that in the gauge by one-tenth of an 

 atmosphere. But as differences of pressure alone were taken from the gauge, this 

 '■ause merely shifts (to a small extent) the range through which the compression was 

 measured. But the rise of mercury in the piezometer stem made a reduction of the 

 range of pressure as measured, which for 3 tons pressure might amount to about 

 0'5 atm. The error thus introduced was, at the utmost, of the order - l of the com- 

 pressibility measured. Thus the second cause, also, produces only negligible effects. 



I preferred to settle the first question by experiment rather than by calculation, 

 as the obtaining of the data for calculation would have required cutting up of the 

 piezometer bulbs. The - 5 atm. spoken of above represented, in extreme cases, the 

 excess of external over internal pressure in the piezometers. By direct experiment on 

 two of the instruments themselves, it was found that their internal volume was 

 diminished at most '00002 of the whole by 0"6 atm. of external pressure. This would 

 involve as a correction the adding of O'l per cent, only to the results at 3 tons, so that 

 it also is well within the limits of error of the measurements above. 



