210 Direction of the Wind at Wallingford, Conn. 

 The following figure shows the form of the vane, 

 ^ e 



X 



y 



and its dimensions are as follows : — 



Length, </ to / =r 10 ft. 8 in. ; o to ^ = 5 ft. 3 in. 

 Breadth, ab =: 10 in. ; cf? = 3 in. : efz=igh=i\^ inch. 

 Thickness, g to k =: 1^ inch. ; A; to ^ tapers to \ inch. 



The vane is of hard wood, and balanced by a leaden ball at gh. It 

 was elevated about 50 feet from the ground, and supported by a mast 

 erected on the back part of the house ; while the shaft passed verti- 

 cally through the roof down to a closet conveniently situated for ob- 

 servation. A vertical cylinder, 3 ft. 8 in. high and 2f inches in diam- 

 eter, was firmly attached to the shaft, so as to follow its slightest mo- 

 tion. Near the cylinder was placed a seven-day clock, whose weight 

 (a loaded box descending in a groove) carried a pencil through a ver^ 

 tical height of 42 inches in seven days, being at the rate of one-fourth 

 inch per hour. The pencil was pressed by a spring against a paper 

 pasted upon the the cylinder, and when the cylinder was stationary 

 described a vertical line upon the paper. The vertical motion of the 

 pencil combined with the movement of the cylinder when the wind 

 w^as iiTcgular, traced a zig-zag line upon the paper. The direction of 

 the meridian was determined by setting up a series of stakes in the 

 range of the pole star. These observations wei'e made with the naked 

 eve, and no care was taken to select the instant when the pole star 

 was on the meridian. The error arising from this latter source might 

 amount to two degrees. The vane having been set in the meridian, 

 the points of the compass were marked upon the cylinder, and thus 

 the directions denoted by the zig-zag line could be readily determined. 



The«Dbservations on the direction of the wind commenced July 1st, 

 1857, and were continuous to June 9th, 1862. From Jiily, 1857, to 

 January, 1859, only the eight cardinal points were employed by Dr. 

 Harrison in the copy of his record from which the following reductions 

 were made; but from January, 1859, to June, 1862, sixteen points 

 were employed. To determine the mean direction of the vane for 

 each hour during each month of the year, the number of hours that 

 each direction occurred during the month for the first hour, for the 

 second hour, etc., was counted, and the ?um of the corresponding 



