RAPPORT 1905—06 ANNEXE B — XXX — 
different concentrations characterized by the values o, = 9.67, a, = 18.99, 6, = 28.15, 
was compared with that of distilled water. This was done in the Mediterranean 
at 1500 m. and 2600 m. depth (temperature: respectively 12°.90 and 13°.10) and 
in the Atlantic at 2600 m. and 4500 m. depth (temperature 3°.14 and 2°.53 re- 
spectively). 
It is the pleasant duty of the C.L. to acknowledge the generous assistance 
of S. A.S. Prince Apert of Monaco, who put his yacht “Princesse Alice” at the 
disposition of the C.L. for all these measurements. 
It was found that the relationship betwen o, and the compressibility y: of 
sea-water at a given temperature is, with a high degree of accuracy, a linear one. 
If values of » be calculated by linear interpolation between the values for distilled 
water (a, = — 0.13) and sea-water of o) — 98, the value for y a, = 14 comes out 
only 0.00000006 too high, which is 0.0015 of the whole. Within the actual limits 
of concentration found in deep ocean-water, a, = 27 and o, = 31, the relationship 
can therefore be regarded as absolutely lmear, the maximum error likely to be 
introduced by linear interpolation being only about 1/so of the above, or 0.00003 
of the total compressibility. In accordance with the original programme the com- 
pressibility of sea-water under different circumstances was calculated from the 
difference of compressibility of sea-water and distilled water, the compressibility 
of the latter being taken from Amagat’s tables. The accuracy of the present 
measurements seems to be quite sufficient for the purpose, to judge from the 
inner agreement between the several values obtained; the probable errors of the 
final results being less than 1/2000 of the total compressibility. There is, however, 
a remarkable disagreement between the results now obtained and those given by 
Tair. The disagreement is of a magnitude sufficient to affect the 5th decimal 
place, on calculating the difference of density between different places at the same 
level in the deep ocean. It is therefore regarded as desirable to make also a 
control series of determinations by a direct method with artificial pressure, and to 
put off the publication of results until the whole work has in this way been 
finished. Steps have been taken to begin the new series of determinations. A 
method essentially the same as was used by Richarps and Strutt has been chosen, 
though with the difference that the pressure will be defined by the apparent com- 
pression of mercury (or of distilled water) at 0°. For this purpose an arrrange- 
ment has been designed consisting of a RicHarps and StuLL piezometer surrounded 
by melting ice, in connection with a differential mercury manometer. 
V. The absorbtion coefficients of the asmospheric gases in sea-water: 
The work so far as the Oxygen and Nitrogen are concerned has been com- 
pleted in accordance with the programme as laid down in former reports. As there 
stated, Estreicher’s Method modified, has been employed. 
