NEW METHOD OF DETERMINING THE WIND'S VELOCITY. 557 



varying direction of the wind. A short distance below the swivel a hori- 

 zontal arm, fifteen inches in length, is firmly attached to the shaft, and just 

 below this another one is similarly attached. These two arms are equal in 

 length and parallel to each other. To the extremity of the upper arm are 

 fastened two fine iron wires, one, seventeen feet long, supporting a plummet, 

 and the other, six and a half feet long, supporting the copper sphere men- 

 tioned above. This sphere is 8.5 centimetres in diameter, and weighs 135.92 

 grammes. To the extremity of the lower arm is attached the scale, con- 

 sisting of two lath-like pieces of wood, about four feet long, fastened paral- 

 lel to each other and about one-half inch apart. Between these, constitut- 

 ing a guide, the wires are suspended, and on the front one the graduation is 

 made. The plumb line serves to determine the zero point, and is of no fur- 

 ther use. By means of the swivel above and the socket below, the appar- 

 atus is easily turned, so that when the ball is deflected by the wind its vi- 

 brations may take place in the space between the two laths. The observa- 

 tion consists in registering the'araount of this deflection from the zero point, 

 or point in which the perpendicular line cuts the scale. It is well known 

 that the force of the wind is not constant for any great length of time. It 

 is hence necessary to take a great many observations, at short intervals, in 

 order to obtain a correct mean. In my work I noticed the deflection of the 

 ball through fifteen minutes, making a record every fifteen seconds, which 

 gave me sixty observations for the quarter of an hour. The following table 

 taken from my note book willillustrate : 



June 19, 40 m. past 6, def.=20 cen. | June Id, 41 ni. pai^t 6, def-=]6 cen. 



40i " a, " 18 " ! 41} '• (i, " 18 '- 



40 .i ••' o, " 20 " t 41^ " G, " 19 " 



40f " 6, " 17 " i &c. 



Cup anemometer marked 15.24 per. h. 



I observed always the record of the anemometer for the same fifteen 

 minutes, and at the close of my work, which was continued on several days 

 to obtain mean deflections corresponding to different velocities, I had a 

 number of such records as the above, each consisting of the position of the 

 ball at sixty different periods. 



By a well known principle of mechanics I found the force or j)ressure of 

 the wind necessary to produce in the ball of known weight the observed 

 deflections, and for convenience made the followiiig table : 

 For deflection of 1 centirr.elre, force = .73 gramme. 

 " 2 " " 1.46 grammes. 



" " " 3 " " 2.19 



" ii a ^ u "2 92 " 



u u ago u u 4o_g 



Having this table of pressure, the next question is to find the wind's 

 velocity corresponding to each of them. There are several empirical for- 

 mulas, which might be applied to the solution of a problem such as this, but 

 one proposed by Weisbach seems to be the most reliable. It is for the ac- 

 tion of an unlimited stream, either of water or air, and by giving suitable 

 values to some of its factors it affords an easy solution to the problem in 



