16 
ployed as arguments. Helland-Hansen and Sandstrom in ‘‘ Report 
on the Norwegian Fishery Investigations, Volume II, No. 4, Bergen, 
1903,” first provided a way to avoid such a ponderous, unwieldly 
work by calculating, as an initial step, the values of specific volume 
at frequent depths, and covering the normal range of change in com- 
pressibility in an ocean of 0° C., and a salinity of 35 per mille. A 
correction called the anomaly of specific volume is then added to this 
first figure, representing the specific volume of any charactered water, 
but under a similar pressure. According to these arrangements all 
corrections are embodied in a total of four small handy tables. The 
details of this ingenious method of tabulation are also described in 
Bjerknes’, ‘‘Dynamic Meteorology and Hydrography,’’ Carnegie 
Institution Publications, 1910-11. Later table groupings have been 
made and published by Hesselberg and Sverdrup, ‘‘Beitrag zur 
Berechnung der Druckund Massenyerteilung im Meere,” Bergens 
Museums Aarbok, 1914-15. 
GIVEN TEMPERATURE AND SALINITY—A GRAPHIC METHOD TO 
FIND DENSITY 
The specific volume in situ, as determined by the foregoing tables, 
is based upon an initial given density, usually found by means of 
Knudsen’s Hydrographical Tables with addendum. There is con- 
siderable labor attached to interpolating when there are perhaps 
several hundred observational records of temperature and salinity 
which require conversion into density form. The Geo-Physical 
Institute, Bergen, where the writer spent some time, finds it con- 
venient to facilitate such work by the construction of a graph based 
upon the three arguments of temperature, salinity, and density, within 
the range which prevails for the first two in the temperate zones. 
The method possesses such great advantages over the use of the 
tables that it is set forth here for the benefit of future investigators 
who may have to deal with a large number of field observations. 
The construction of the graph is based upon the three formule of 
Knudsen: 
(1) s=0.30+1.805 C7. 
(2) 6.= —0.069+ 1.4708 Cl —0.001570 C? + 0.0000398 CE. 
(8) 6: =24+ (6, + 0.1324) [1—A,+ By; (6, —0.1324)]. 
(For do, dt, t, At, and By, see Martin Knudsen’s, ‘‘ Hydrographical Ta- 
bles,” Copenhagen, 1901.) Density values are plotted as abcisse, 
salinity values as ordinates, and isotherm curves, determined in ac- 
cordance with the fixed relation existing between the three variables, 
run diagonally across the graph., In order to determine the latter 
with a sufficient degree of accuracy, it is necessary to fix definitely a 
