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



the agreement between this and the dry thermometer in 

 the screen was e.xcellent. The corrected temperatures, 

 as well as the values of relative humidity, are recorded 

 in Table I b; temperatures found by means of the swing- 

 thermometer are printed in italics. Sometimes the wet- 

 bulb thermometer gave a slightly higher corrected reading 

 than the dry thermometer of the psychrometer; in these 

 cases the relative humidity is reckoned as 100 and placed 

 within brackets (lOOJ. 



We had two aneroid barometers and a barograph 

 on board. They were often read, but as they did not give 

 quite trustworthy records the results are omitted in the 

 Tables. Sufficient information on the distribution of 

 pressure may be deduced from the Hoffmeyer charts 

 (edited by The Danish Meteorological Institute and the 

 Deutsche Seewarte in cooperation). Parts of these charts 

 are reproduced on pp. 77 ^ 79-', ST-', and 83*. 



Observations of wind directions were made along 

 with the temperature observations. The directions recorded 

 in Table lb are magnetic, jus! as they were noted in the 

 journals'). The wind velocity was estimated according to 

 the "half Beaufort" (0—6), a scale which has been preferred 

 by many Norwegian sailors. 



The cloudiness was estimated according to the scale 

 0-10 (0=-blue sky, 10=overcast). Fog and rain were noted 

 in the journals. In the column headed "Remarks" in 

 Tables I and I b, the ciphers indicate the number of the 

 hydrographic stations; F means fog, and R rain. 



The work in the small steamer with only few observers 

 on board had to be concentrated upon the essential 

 investigations, and but slight attention could be paid to 

 meteorological investigations during the expedition. To 

 begin with it seemed doubtful whether it would be worth 

 while to publish the observations. Closer examination, 

 however, has shown that the observations give much more 



interesting results than might be expected and it is strongly 

 to be recommended that such observations every hour or 

 every second hour be more generally made than is now 

 the case. 



5. Statistical Treatment of the Observations. 



The occasional variations in the meteorological elements 

 tend to obscure the general features. The former may, 

 however, to a great extent be eliminated by means of 

 statistical methods. 



When we calculate continuous means for 24 hours 

 the daily period and many casual variations disappear, 

 and the more general variations stand out. If we have 

 observations made every hour and calculate the means 



for all observations between O' and 23' (incl.), between 1 



and on the following day, between 2 ' and 1 ' on the next, 



and so on, we get the means centred at about 1 1 ' 30"^, 1 2 ^ 



30 , 13 30 , etc. These means may be termed ni,, . 



In order to have them referred to full hours we may take 

 the averages of every two succeeding values, so getting 



the means for 12', 13', etc. These means may be termed 

 ^2^2- % taking the difference fo—m^^^j between the 



single observations (o) and the means referred to the hour 

 of observation /'«24?) ^'^ S^' values corresponding to the 

 casual and short-period {e. g. the daily) variations. In 

 the case of observations which were only made every second 

 hour, we get similar values by taking continuous means 

 of 12 successive observations, and calculate rn and 



o-OTj2 2 for ^ve-ry second hour (cf. W.allen, 1910 and 1913, 



Helland-Hansen and Nansen, 1917 and 1920]. 



Our material of observations from the sea-surface 



') The directions may be reduced In true directions by applying the following vahici of magnetic variation: 



