BIGELOW: EXPLORATIONS IX THE GULF OF MAINE. 



77 



Knudsen's (1901) tables and from Ekinan's (1910) tables of sea water 

 under pressure. Such calculations are approximately correct arith- 

 matically, but notice must be called to the fact that the probable 

 limits of error are the sum of the two observational errors, first for 

 salinity, i. e., =*= .02 of salinity (p. 40), second for temperature, which 

 is =fc .3° F, approximately .15° C. Now the sum of these errors has a 

 considerable efTect on the calculated densities, and for this reason 

 the fifth decimal point is disregarded in the table. Of course a much 

 higher degree of accuracy could be, and is, obtained with improved 

 instruments, for example, during the North Atlantic cruise of the 

 Michael Sars in 1910 (:\Iurray and Hjort, 1912). But it would be 

 misleading to claim better results with our instruments. 



Fig. 34. — Curve of density ui situ at Piatt's Bank (Station 23); Jeffrey's 

 Bank (Station 25) and German Bank (Station 29). 



The correction for pressure has often been disregarded, especially 

 in shallow water; but it can easily be applied from Ekman's tables. 

 In depths less than fifty fathoms it is of little practical importance, 

 but by the time 100 fathoms is reached it is by no means negligible. 

 For example, at Station 28, 120 fathoms, the density at the tempera- 

 ture in situ without pressure correction, is 27.02; with pressure correc- 

 tion, 28.03. In the accompam-ing table the pressure correction for 

 depths less than fifty fathoms is calculated by the use of Ekman's 

 table IV alone, which is sufficiently accurate for our present purpose. 



The most important thing which the table and curves (fig. 3-1-36) 

 show is that there was a steadj^ increase of density at every station from 

 the surface down to the bottom, which, as we now know, is the normal 



