110 ANNUAL RECORD OF SCIENCE AND INDUSTRY. 



the direction of the path of the locomotive, was two and 

 eight -tenths kilometers per hour; and the velocities ob- 

 served by timing the train as it passed each mile - post 

 agreed with the velocities recorded by the anemometer to 

 within two per cent. A more definite method of investiga- 

 tion, however, is that of Combes, which consists in attach- 

 ing the anemometer to a long arm, which latter is made to 

 revolve in a circle at a known rate, in a space where strong 

 currents of air do not occur. Velocities up to forty kilome- 

 ters per hour were attained in this way by Dohrandt ; and 

 from the investigation of six different Robinson's anemom- 

 eters, and of two of Casella's instruments, and of a number 

 of electric recording anemometers, it was concluded that the 

 arithmetical mean of the indications of an anemometer, when 

 the arm which carried it revolved in both a positive and a 

 negative direction, might be adopted as the result that would 

 be given if the instrument had simply moved in a straight 

 line. The determination of the effect of the moving: ane- 

 mometer upon the air surrounding it an effect which it is 

 known consists essentially in carrying the air with it, and 

 thereby diminishing the relative velocity of the air and the 

 instrument was first made by placing a Casella and a Wolt- 

 man's anemometer on the same level with the rotating ane- 

 mometer, but fixed in their positions, and so near to the cir- 

 cle described by the latter that they could feel the influence 

 of the wind dragged along by the Combes' apparatus, and 

 be thereby set in motion. As the result of this portion of 

 his investigation, Dohrandt finds that the wind due to the 

 dragging influence of the moving anemometer is very ap- 

 proximately proportional to the velocity of the latter itself. 

 As a final result in reference to the velocity of the anemom- 

 eters, Dohrandt gives a formula applicable to each one indi- 

 vidually, by means of "which the true velocity may be com- 

 puted from the indications of the dials of the instruments. 

 Having thus some half-dozen well-investigated instruments 

 as standards, it becomes easy to compare numerous others 

 with these. The result of such comparison showed that the 

 errors determined in this secondary manner were of the same 

 nature and expressed by the same formula as those deter- 

 mined bv the original investigations. In General, in relation 

 to the velocity of the anemometers, Dohrandt finds that the 



