130 FRIDTJOF NANSEN. 



During my work with our oceanographical observations made 

 dm'ing the Fram Expedition where I had tried in vain to attain 

 what I would consider sufficient accuracy, I had fully realized 

 that it was imperative that the highest possible degree of ac- 

 curacy should be introduced into the determinations of the tem- 

 perature, salinity, and density of the sea-water at all depths if 

 we shall ever attain a fairly complete and reliable conception of 

 the circulation of the Ocean, its causes and periodical variations. 

 I consider it therefore to be one of the first and most important 

 aims of oceanographers to improve materially the methods of 

 oceanographical research, and then with improved methods to 

 collect new material of observations from all parts of the Ocean. 



The results of our expedition of last summer have, I be- 

 lieve, fully justified this view. It appears that the variations 

 in the temperature, and still more in the salinity, of the deeper 

 strata of the sea, at depths greater than 600 or 800 m., are so 

 extremely slight that the highest possible degree of accuracy 

 would be necessary to demonstrate them; and it even seems 

 doubtful whether we possess at this moment methods sufficiently 

 accurate to trace the variations in the salinity of the deep-water 

 in the JN^orwegian Sea. We may, at any rate, say with cer- 

 tainty that it would be worthless to make observations of the 

 temperature and sahnity at greater depths in this sea with 

 no more accurate methods than have hitherto generally been 

 employed, for the errors of observation are much greater than 

 the natural variations. Nevertheless the slight differences in den- 

 sity, etc. which may exist in the deeper strata of the Ocean are 

 certainly very important for the circulation of the whole hydro- 

 sphere, and it is therefore very desirable that they should be 

 exactly determined. But our investigations prove that the varia- 

 tions in the salinity even at lesser depths are very gradual, and 

 generally so small that they too necessitate a very high degree 

 of accuracy. 



