ATLANT. DEEP-SEA EXPED. 1910. VOL. i) PHYSICAL OCEANOGRAPHY AND METEOROLOGY 



17 



used for the surface-observations in Table 1. Surface- 

 observations belongiiiji to vertical series are entered in 

 botli tables, but the times recorded are not identical. 



The next two cohunns of Table II give the depth of 

 observation in metres (m) and English fathoms (f). Time 

 did not permit us to test the meter-wheels before the 

 Expedition started. Towards the end of the cruise, how- 

 ever, the meter-wheels that had been used were carefully 

 tested. It was then discovered that all of them were 

 correct except one, which unfortunately was just the wheel 

 mostly used for tiie hydrographic work. The correction 

 amounted to about 8-5 ",o, and the indication had to be 

 reduced by this amount in order to give the correct length 

 of wire let out. This means that a reading of e. g. 200 m. 

 corresponds to a length of 183 m. or 100 English fathoms. 

 In all cases where this meter-wheel has been used, we 

 have corrected the readings by taking half the reading as 

 the actual length of wire in English fathoms and converting 

 this length into metres. The final error will not amount 

 to as much as 1 "/o of the depth provided that the line 

 hung vertically. Much attention was paid to the attaining 

 of a vertical position by the wire. We succeeded practically 

 always in avoiding any deflection of the wire exceeding 

 10° from the plumb-line when working with the water- 

 bottle, as the "Michael Sars" could fairly easily be 

 manoeuvred and brought to the right bearing and speed 

 in relation to wind and current. 



16. Standard Depths. 



Owing to the above-mentioned defect in the meter- 

 wheel the depths of observation are in most cases 

 unsuitable. For comparison with other observations and 

 for the discussion of the results it is very useful to have 

 the various data referred to certain integer-numbers of 

 metres. Which depths ought to be selected, depends upon 

 the vertical variations. For the North Atlantic as well as 

 for oceanic regions generally it is desirable that the data 

 should be related to the following depths: Surface, 10, 

 25, 50, 75, 100, 150 metres; every 100 metres from 200 

 to 1000 m.; every 200 from 1000 to 2000 m.; 2500 m., 

 3000 m. and for greater depths every 1000 m. These 

 depths will be called standard depths [cf. Bjerknes, 1910, 

 p. 6]. When taking observations at these depths the 

 vertical variations will in all probability be represented 

 with sufficient accuracy. In coastal waters or in other 

 conditions in which the variations in the upper strata (as 

 e. g. within the Polar currents) are comparatively large, 

 it may be desirable to have smaller intervals of depth 

 for the upper 100 metres; in such cases the data ought 

 to be stated for every 10th meter down to 100 metres, 

 in many cases also for 5 metres below the surface. For 



our observations from the "Michael Sars" Expedition the 

 former series of standard depths will suffice in by far the 

 largest number of cases. Within the precincts of the 

 Labrador current the vertical variations may be very great. 

 They will be seen from the station curves and the vertical 

 sections. The observations have been converted into data 

 for the above mentioned standard depths by graphical 

 interpolations and the new values found are published in 

 Table III. 



In general, it would be an advantage to have data 

 recorded for these standard depths in hydrographical tables. 



17. Graphical Interpolations. 



For the interpolations for standard depths the serial 

 observations from all the "Michael Sars" stations have 

 been plotted out on mm-paper, with the depths along the 

 ordinate and temperature, salinity and density along 

 the abscissa. 1 mm on the paper represented 5 m. of 

 depth, 0-05° C, 0-01 "/oo of salinity, and 001 a,. 



When 2 out of the 3 quantities, temperature, salinity 

 and density, are known, the third can be computed. The 

 3 curves must, therefore, correspond at all points in such 

 a way that any one curve is determined by the two others. 



It is necessary to check the result by the graphical 

 interpolation. For instance, we must make sure that the 

 density found from the density-curve is exactly the 

 density defined by the salinity and temperature found 

 from the other curves. In many — probably in most — 

 cases it will be found that the demand for adequacy is 

 not fulfilled by the first drawing of the curves, though 

 the errors will be small and easily corrected when rather 

 numerous points are given from observations both of 

 temperature and of salinity. If there is a great difference 

 in depth between the observations it is more difficult to 

 attain adequacy. 



The water-layers within a comparatively uniform region 

 of the ocean generally exhibit a density-curve of definite 

 and almost constant shape. The station curves on pp. 

 65'''— 72* show many examples of this. If, then, the 

 temperature observations are so numerous that the tem- 

 perature-curve can be constructed with sufficient accuracy, 

 any missing salinities may be found in a fairly trustworthy 

 manner by means of this temperature-curve and the probable 

 density-curve, if there are but few observations of tem- 

 perature and salinity it is impossible to arrive at a satis- 

 factory result unless the general characteristics of the 

 water-layers are known from other stations within the 

 same region. In that case the final results may be quite 

 good, but the mutual checking and correction of the 

 curves will take considerable time [cf. Helland-Hansen, 

 1916]. 



