NO. 4 TEMPERATURE VARIATIONS IN THE NORTH ATLANTIC IO3 



For each of our io° longitude fields in the course Channel to 

 New York and for each field of io° longitude and 2° latitude in 

 the region Portugal to the Azores, for the months January and 

 February of each year, and in the region Channel to New York, 

 also for the month of March, we obtained the mean direction of 

 the isobars (in the direction of the wind, according to the barometric 

 law of the wind). We have also folmd values for the aver- 

 age intensity of the air pressure gradients, which we obtained 

 by measuring the distance between the isobars and taking the 

 reciprocals of these values. As a unit, we have taken the thousandth 

 of the reciprocal value of the distance between the two-millimeter 

 interval isobars, measuring this distance on the charts to millimeters. 

 If, for example, the distance between two such isobars was 6 milli- 

 meters, then the gradient numbers according to our figures would 

 be 1000-^6=167. As a rule we have taken mean of the distances 

 between several isobars. By making progressive vector diagrams 

 for each month in which the direction of the vectors of that month 

 for each year are drawn according to the isobar angle, and the 

 lengths are given by the relative gradient numbers just described, we 

 have obtained average isobar directions for each of the months 

 January, February and March in each of the 10° longitude fields for 

 the eleven periods 1898 to 1908.'' This period is unfortunately not 

 identical with the eleven-year period 1900 to 1910, which we have 

 employed in the determination of the temperature normals. 



Next we have determined the anomalies of the isobar direction 

 for the dififerent months and years as deviations from the average 

 direction for these months. Deviations toward the south, that is 

 to say, when the isobars are directed southerly of their normal posi- 

 tions, we have designated as negative, and deviations toward the 

 north as positive. The product of the gradient number and the 

 sine of this anomaly angle is then used as a measure of the possible 

 influence of the wind on the surface temperature. In doing so, we 

 consider that the normal position of the surface isotherm is de- 

 pendent on the average direction of the isobar and that a deviation 

 from this must produce lateral displacements of the isotherms. The 

 sine of the deviation angle would be equal to the component of the 

 air motion at right angles to the average direction. 



This process obviously cannot pretend to any great degree of 

 accuracy. For example, it is not easy to know in advance which is 



^Unfortunately we received the isobar charts for 1909 and 1910 too late to 

 carry through this computation. 



