286 Direction of the Wind, N'ew Haven, Conn. 



In order to determine the mean direction of the wind for each of 

 the hours of observation for each month of the year, tlie number ot 

 times that each direction occurred during the month for each of the 

 hours of observation was counted, and the sum of tlie corresponding- 

 numbers for the entire period of years was taken. These numbers 

 are given in Table I, pages 270-273, and Table II, pages 274-285. 



The mean direction of the wind for each of the hours of observa- 

 tion was obtained by solving a traverse in Avhich the number of times 

 that each wind was recorded was regarded as the distance traveled. 

 The mean direction of the wind thus obtained for each of the hours 

 of observation, and for each month of the year, is given in Table III. 

 The angles are reckoned from the North point around the circle 

 through the West and South. 



Table III also shows the ratio of the wind's ])rogressive motion in 

 its mean direction to the total distance traveled, for each hour of 

 observation. These numbers were obtained as follows : Having com- 

 puted the mean dii-ection of the wand for each hour, the absolute 

 length of the line indicating its direction was computed trigonometri- 

 cally, and the number representing this line was divided by the 

 number of observations for that hour, without regard to direction. 

 When these ratios are large it shows that the direction of the wind 

 was comparatively steady ; when the ratios are small it shows that 

 the direction of the wind was extremely variable. 



If we compare the direction of the wind as deduced from the first 

 series of observations with the direction as deduced from the second 

 series, we shall find considerable discrepancies. Table IV, Part 1, 

 shows the result of such a comparison. From the month of October 

 to the month of March inclusive, the directions in the first series are 

 more westerly in every instance, except for March at 6 p. m. During 

 the i-emaining six months of the year the two series present still 

 greater discrepancies. These diffei'ences are larger than was antici- 

 pated, and are not easily explained. The observations of the second 

 series were principally made at a station about half a mile nearer the 

 harbor than the first series ; but this circumstance does not seem 

 suflicient to account for so large diflerences as appear in the results. 



It is suspected that in the first series of observations the direction 

 of the wind recorded for M, and E. was not designed to give the 

 direction noticed at any fixed hour, but rather the prevalent direction 

 for the forenoon and afternoon. Such a result, deduced, as it probably 

 was, not from several recorded observations, but from casual obser- 

 vations of the vane loosely preserved in the memory, could not claim 



