47 



Weather Bureau from the naval radio station at Arlington at 10 

 a. m. and 10 p. m. Occasionally the weather experienced by the 

 patrol vessel did not accord with the Weather Bureau's data when 

 plotted on a base map. In such cases it was necessary to collect 

 reports from as many ships as possible and well scattered within a 

 radius of 500 miles of our position. This work often revealed the 

 presence of disturbances that had formed or developed over the sea, 

 the position of these storms being unknown to the Weather Bureau, 

 at least when it had sent out its routine data. 



Twice daily, as in former years, at 8 a. m. and 8 p. m., a report 

 was dispatched to the United States Weather Bureau, Washington, 

 D. C, and at the end of each cruise a more detailed report was for- 

 warded b}^ mail to the Washington weather officials. 



ICE FORECASTING BY MEANS OF THE WEATHER 



Last year's annual report (Bull. No. 15, pp. 45-48) contained an 

 account of a scientific investigation carried on at Harvard University 



Fig. 13. — The anomaly of atmospheric pressure over the northwestern 

 North Atlantic for the month of October, 1926. Isobars drawn every 

 2 millibars. Conditions when reflected the following spring spell less 

 ice than normal 



and later at the British Meteorological Office, London, into the pos- 

 sible relationship, between the varying amounts of Arctic ice from 

 year to year and meteorological events occurring in the northern 

 regions some months previously. It appears logical to believe that 

 the prevailing direction of winds over the Labrador-Greenland region, 

 when expressed in terms of departures from normal, and considered 

 in monthly periods, would be reflected sometime later in the varia- 

 tions from normal in the amounts of ice. It has been found best 

 throughout the investigation to work with differences from means, 

 .and this fact should be kept in mind by the reader. 



