On the Surface Temperature of North Sea and North Atlantic. 7 



tlic j\feteorological Institute of Christiania, and sent to me through 

 i*rot'essor HoUand llanseu. (b) Similar data from a number of 

 Danish lightships and lighthouses, as recorded year by year in the 

 Nautisk-Meteorologisk Aarbog. (c) Similar data from four Dutch 

 lightshipt>, sent me by my friend the late Professor Wind of Utrecht. 



The methods which we are accustomed to use in dealing with 

 our large collections of Temperature statistics, in order to bring out 

 tlie main lessons which they contain, have been described at various 

 times in the Eeports of our North Sea Investigations, and especially 

 in a paper by me on " Some IMethods and Eesults of Hydrographical 

 Investigation," published in 1907, in our second volume of Eeports. 

 It is not necessary, therefore, to deal at any length with the matter 

 here. 



What we are at present mainly concerned with is to show (1) the 

 iiiea^i conditions, as determined over a long series of years; and 

 (-) the main annual Huctuation, as it is expressed by the mean 

 temperature conditions which are characteristic of each of the 

 twelve months of the year. Besides these, there are many other 

 important phenomena of temperature which, in due time, deserve 

 consideration. For instance, there are minor but still regular 

 fluctuations, such as those between night and day ; there are inter- 

 esting phenomena connected with the relations between the 

 temperature of the sea and that of the air ; and of very great 

 interest and importance are the fluctuations from one year to 

 another, which are apparently irregular, but in which it seems by no 

 means impossible to discern some orderly sequence, underlying the 

 apparent irregularity. These, and many other phenomena besides, 

 are matters for special inquiry, after the mean, or mean annual, 

 phenomena have been sufficiently investigated and ascertained. 



We begin, then, by ascertaining for each station, and for the 

 whole period over which our observations extend, the Mean Annual, 

 and the twelve Mean Monthly temperature values; and these we 

 may set straightway down upon the chart, and draw in each case, 

 as best we can, the isotherms, or lines of equal mean temperature, 

 between the corresponding points. 



In many cases, and especially where we are dealing with a con- 

 tinuous line of observations, or of means drawn from observations 

 in small contiguous areas, we may with advantage employ a more 

 systematic method of interpolation. For instance, if we take the 

 Atlantic Temperature observations collected by the Meteorological 

 Office, we find these averaged out for small areas of the ocean, each 

 one measuring two degrees of longitude by two degrees of latitude ; 

 the mean temperature for each small area being determined to the 

 nearest whole degree Fahrenheit. If we take any one linear series 

 of these means across the ocean, along some one degree of latitude, 

 or some one degree of longitude, and plot the successive values on 

 squared paper, we get a broken line through which it is not 

 difficult (as a rule) to draw a smooth curve. From this 

 smooth curve we can read off the approximate mean tem- 

 perature at any point, or the point which approximately coincides 

 with any particular temperature ; we may accordingly use it to 

 determine the iwsition of our isotherms. Thus, in the following 



