METEOROLOGY AND ALLIED SUBJECTS. 253 



fallen into tlie rain-gauge are diverted from their paths, and some of 

 them fall to the leeward. Jevons concluded that measurements of rain- 

 fall made with gauges that are high above the ground, and exposed to 

 the wind, are entirely useless. The observations made by the rainfall 

 committee of the British Association, as communicated in their report 

 in 1870, confirmed Jevons's explanation. They found that the greatest 

 rainfall was measured in gauges that were so sunken within pits in the 

 earth that the mouth of the receiv^er was on a level with the earth's sur- 

 face, and entirely protected from violent wind currents, and their recom- 

 mendation of the so-called pit-gauge has generally been considered our 

 best knowledge on the subject. But another form for the gauge was 

 suggested bj' Jevons in 1801, in which the mouth of the receiver is sur- 

 rounded by a large horizontal metallic disk. This form has been modi- 

 tied by Messrs. Kipher and Woodward, by the introduction of cells, and 

 finally simj^lified into a simple upright tube, surrounded by a protecting 

 screen in the general shape of a filter, whose broad lip protects the 

 mouth of the receiver from gusts of ^nd. Two rain-gauges, one with- 

 out and the other with the protecting lips, were exposed side by side 

 during a summer and spring, and the unprotected gauge collected 3 per 

 cent, less than the protected one. In the experiments of the British 

 Association an unprotected gauge, in a similar position, collected five 

 l)er cent, less than the pit-gauge. Again, sixteen gauges with protect- 

 ing lips were placed in various locations on the roof of the university 

 building in Saint Louis, a hundred feet above the earth. 



The result of these experiments is to show that the so-called correction, 

 for the altitude of the rain-gauge reduces to nothing when the gauge is 

 properly j)rotected against the wind, and that under this condition the 

 rain gauge may be safely established at any altitude whatever. {Z. 0. 

 G. i¥., XIV, p. 250.) 



Koppen has attempted to apply the results of Dohrandt's investiga- 

 tions into the accuracy of the Eobinson anemometer to the reduction of 

 the observations made at the stations of the German Manne Observa- 

 tory. According to Dohrandt the true velocity (w) of the wind can be 

 derived from the velocity {a) of the centers of the hemispherical cups 

 of the Eobinson anemometer by the formula tr=K+B a, where K and 

 B are constants peculiar to each instrument. In the absence of any 

 special determination, K may be assumed equal to 1.0 meter per second, 

 which is the mean of the values determined -by Dohrandt for the ane- 

 mometers and auemograiihs investigated by him. 



The constant B, which, according to Eobinson, should be fqual to 3.0, 

 can be more accurately' computed by Dohrandt's empirical formula 



"R2 T?* 



B =3.0133-53.7367 ^^^ + 1033.81 ^^ 



r 7'^ 



where E is the radius of the hemispherical cups and r the distance of 

 the centers of the cups from the axis of rotation, both expressed in 



