1 912. Xo. 12. 



THE SEA WEST OF SPITSBERGEN. 



Fig. I o Temperature-Curves of July and September, due tO inacCUracieS of the salini- 



Om 



iQio, for Stat. 30. 



3'- '^ - 5' 



;?2Vll,)9L^ 



100- 



200^ 



Fig. 1 1. Curves of Density <ot> for Stats. 5 — 9- 



/«. 2£.l' 23t't 27-9i> rr^O ZJ-70 



300 



100 z 



100 



iOO 



^00 



500 



MO 



ties; but not more in the mentio- 

 ned depths, between 500 and 600 

 metres, than in other depths, and 

 the irregularities go in different 

 directions. This proves that there 

 cannot have been any essential difference as to accuracv, 

 between the temperatures taken with 

 the Xansen-thermometer and those of the 

 reversing thermometer. 



The curves of density of June (Figs. 

 II and 12) are entirely different in this 

 respect. Most of them exhibit a sudden break in 

 their courses at the boundary between their upper 

 parts, based on the temperature-readings of the 

 Xansen-thermometer, and the lower parts based on 

 the readings of the reversing thermometer, the va- 

 lues of their upper parts being obviously too great. 

 This becomes still more conspicuous bv a com- 

 parison between the curves of density of June and 

 September at Stations 3 and 4 (Figs. 16 and 17I. 

 While the curves of September of both stations 

 have very regular shapes along their whole length, the curves of 

 June exhibit striking irregularities at the boundary between the 

 upper parts based upon the readings of the Xansen-thermometer, 

 and the lower parts based upon the read- 

 Curves of Density yo o for j^^ ^j- ^j^^ reversing thermometer. The 



^tats. 10—13. ^ ° 



2Si-W 27-90 Z7-ßO 1170 »^PP^r parts of the curves indicate evident!}' 



much too hisfh values of densitv because 



the temperatures are too low. The curves 

 of July and September for Stat. 29 (Fig. 18) 

 exhibit no similar irregularities, and they follow each 

 other very closely ; evidently because the temperatures 

 are correct in their upper as well as their lower parts. 

 It is also noteworthy that the density curves for Stat. 30 

 (Fig. 19) and for Stat. 31 (Fig. 20) show no great differ- 

 ence in density between July (22nd and 23rd) and Sept- 

 ember 6th, 1 9 10. 



As heavier water-strata cannot as a rule rest upon 

 lighter water-strata, it is fairlv easv to see how much 



