'4 



Analysis of the 



The theoretical curves are checked fairly well considering that a small percentual error in 

 measurement appears greatly magnified in the error values. It is thought that these curves bring 

 out the fact that for any anemometer a constant additive error can be obtained by correct pro- 

 portioning of the arms to the cups for any given bearing. 



Although for the small cups (^=.031) the dissimilarity ratio d/D for the point h = k is 0.1 

 and this would mean an unusually slender anemometer; for the large cups (d=.i2j) the curve 

 shows that the point h = k would fall in the neighborhood of d/D= .6, and this would be a reason- 

 able type of anemometer to use. 



The ratio d/D of the cup system which would appear to prescribe to the law V— Vi= V a , for 

 the Standard Weather Bureau bearing and for cups d—.\oo, is in the neighborhood of d/D = .j. 

 This value was obtained in the manner outlined above from the values of h and k from Marvin's 

 data (Table la). 



From the extremely scant data on the beaded hemispherical cup (d=.ioo) the ratio d/D for 

 h = k appears to lie around a value of .5, this being estimated from the data in Table III at .7% 

 turbulence which is low enough according to considerations in the footnote on p. 9, to approximate 

 tree air conditions. 



