HYDROGRAPHY. ,g 



ent this theoretical quantity. As s (-£) = i -\- a&; s (-£) can be found from log « and log CI °/oo. Log a 

 can be found in the table on the other side, and log C7°/oo, is known from the titrations. 



The table referring to log a, is computed for the two arguments: temperature and amount of 

 chlorine, and such that log a is found for every tenth of a degree from — 2° + io°, as well as for the 

 amounts of chlorine 17, 18, 19 and 20°/oo. The calculation of the table has been made on the basis of 

 the above-mentioned results from Ekman's investigations on the expansion of the different water- 

 samples by heating. Still is to be observed that only two of these samples have been used as a 

 basis for the calculation, namely B and Z>, as they seem to be about similar in composition, while 

 that of C deviates considerably from them. The amount of chlorine is stated to be: in B 14.254 %o, 

 in D i9.374°/oo, while the specific gravity s(^j is: for B 1.02084, for D 1.02831. Furthermore the vo- 

 lume of the two water-samples can be found at all whole degrees (according to the mercury thermo- 

 meter), as the volume at o° degree is considered as equal to 1. 



By means of these tables, I have computed the table for the two water samples, as mentioned 

 above , as we have j (-) = s (~j j V(-) , in which V(°) represents the volume of distilled water at o°, 

 divided by the volume of the same weight of distilled water at 4 . This quantity is according to 

 Landolt and Bdrnsteirts tables page 39 considered to be 1.000127. We get then jW = s (t)/ v {~z) 

 where v means the same for sea-water as V for distilled water. 



s (j) is a constant quantity for each water-sample. From the 7-ciphered log of this is subtrac- 

 ted log v (~) for all whole degrees. The numbers corresponding to the difference of logarithms were 

 then J (^) , written with 6 decimales. As according to the above we have the equation, a—( s (j) — I )/^> 

 we find thus log a for the two water-samples at all whole degrees. Log a is for the same water- 

 sample such a function of the temperature, that the second difference is very nearly constant If the 

 temperature be the same, log « will grow with the amount of chlorine, but the addition is very small, 

 and has a minimum between 3 and 4 , where, for 1 %o of chlorine it is only o or 1 on the fifth decimal. 



From the table which applies to the water-samples B and Z>, log a is found for water with 

 the amounts of chlorine 17, 18, 19 and 20%o by simple linear interpolation at every whole degree. 



In this table is now interpolated to log a for every o°.i according to the formula: f{a-\-x) 

 = /a~\~xA' -f x(x—\)j\.2 J", as A' and A" respectively mean the first and second difference. Higher 

 differences are as aforesaid zero. This table is used, because the function log a, as it will be seen, 

 only varies very little with the amount of chlorine, for which reason only the stated amounts of chlo- 

 rine are needed as arguments to use the table with facility. It will furthermore be seen from the 

 computation of the table, that Ekmatis observations have been used with the greatest exactness they 

 could give, so that we certainly may regard all possible errors in *(-£) as arising from inexactness 

 in the determinations of temperature and chlorine made on board. 



We might of course do as Dr. G. Schott has done, (Wissenschaftliche Ergebnisse einer For- 

 schungsreise zur See; Petermanns Mittheilungen XXIII Erganzunsgband, 1892—93, pag. 18) and make 

 a table about the correction that is to be applied to j (2|) to get s (|) , but when the amount of chlo- 

 rine is stated, and not -?(—), then we should first be obliged to determine this quantity, which, of 

 course, would make the calculation much more difficult In the last column in the tables of the sta- 

 tions are tabulated the numbers of the plates, on which the observations are represented. 



