igi2. No. 12. 



THE SEA WEST OF SPITSBERGEN. 



37 



/oa 



2oa 



300 



wo 



500 



600 



700 



800 



900 



JOOO 



1100 



noo 



mo 



mo 



1500 



temperatures of these 

 Stations are hardly 

 trustworthy, as their 

 vertical curves de- 

 monstrate (see Fig. 

 31). Those of Stats. 

 A 4 and A 6, in particular, 

 have most peculiar shapes, 

 descending more or less verti- 

 cally from 100 or 200 metres. 

 These shapes look very im- 

 probable, and are entirely 

 different from those of other 

 curves, in a similarly warm 

 part of this current; and temperatures such 

 as 3.50 C. in 500 metres at Stat. A 3, 



2.05 C. in 850 metres at Stat. A4, and 1.30° C. 

 in 850 metres at Stat. A 6, are highly improb- 

 able in this region. The}' are much higher 

 than any temperatures hitherto observed in 

 similar depths of the Spitsbergen Atlantic Cur- 

 rent. They also give impossible values of the 

 density of the water in these depths. All tem- 

 peratures were taken with a Pettersson Insula- 

 ted Water-Bottle of the old model, in which a thermo- 

 meter was inserted after the bottle came up. The in- 

 sulation of this water-bottle was not sufficient for taking 

 temperatures from such great depths, especiall}' if the 

 bottle was not hauled up very rapidl}'. We know from 

 experience that ver}- great errors may arise in this 

 manner. There is also a possibility that the bottle may 

 have been closed by accident on its way down, before 

 it reached the desired depth. Accidents such as these 

 might often happen with the old Pettersson bottle, according to our experi- 

 ence. But whatever the reason may be, Arrhenius' temperature-readings 

 cannot at any rate be considered trustworthy for depths greater than 

 300 metres. 



By using the vertical temperature-curves of his four stations represented 

 in Fig. 31, and trying to correct for the most obvious errors, we have 

 found a mean temperature for 100, 200, 300, and 400 metres, of 2.1^ C. 



