Robinson — On the Cup Anemometer. ■ 433 



be placed on his estimation of the wind's effect ; and any attempt to 

 combine the results for each verst going and returning seems hopeless. 

 The space, for instance, traversed by the cups, in passing No. 2, are 

 546-11 me. and 319-25 me. ; the Fs' are 21-22 k, and 24-38 i?;; the 



_ 1-953 and 3-342; for No. 10, they are 403-3, 386-5; 36-58, 35-98, 



V 



and 2-645, 2-741. 



M. Dohi-andt only considers that part of the wind which acts in 

 the mean direction of the rail, but its rectangular component cannot 

 be omitted, for it can turn the Anemometer both going and coming. 

 And there is uncertainty about the real velocity with which the air, 

 in such trials, passes the instrument ; a body moving through a fluid 

 carries before and behind it a mass of quiescent fluid, but at the side 

 of a ship it is well known that the water moves astern, in consequence 

 of its displacement in front. 



The same is probably the case on the tender of a train, especially 

 between embankments, in which case the V will be greater than what 

 is given in the Table. 



His determinations of the ratio differ, as might be expected from 

 the preceding remarks, considerably from mine, always in defect, but 

 unequally in tlie different instruments. It must, however, be observed, 



V 



that we mean different things by the term ratio. My m u being 



v + u 



V- a 



a variable, depending on the fiiction and v ;^-' his is , «beingacon- 



l 



fitant.f I must also remark that his equation V= a + hv is strong evi- 

 dence that the circumstances under which his experiments were made 

 are abnormal. In this instrument, when a pennanent state of rotation 

 is established, the mean impelling force must be equal to the mean 

 resistance. Now, with all the defects of this branch of Hydrody- 

 namics, it is certain that the equation expressing this equality must 

 contain V^ v^ and Vv ; and one in which they do not appear cannot 

 give the V which corresponds to a v produced by the action of the real 

 wind. 



* I wish to correct a mistake in ray paper on the Anemometer (1. c.) I stated that 

 m is independent of the size of the instrument, forgetting that b', the co-efRcient of 

 that part of the resistance which is caused by centril\igal force, is in part inversely 

 as the arm of the Anemometer. Therefore m must be larger in small instruments 

 than in large ones. On the other hand, the divisor of /is less. 



t This constant differs considerably, and more than can he explained by mere 

 friction in his instruments. The two Casellas, which are quite 'similar, are special 

 examples of this. In those cases where he extends the interpolation to v'^ ; the a 

 differs from that of the simple equation so much that we can scarcely suppose either 

 to be a true friction co-efficient. It is to be regretted that it did not occm- to him 

 to determine the friction of each Anemometer, and the V at which they began to 

 move ; a knowledge of these would have made his results more useful. 



