OF WINDMILLS. 399 



REMARKS. 



1st. The most material consequence to be derived from the 

 above table is the great diminution of the effective force of 

 the wind, as the velocity of the sail increases ; which shew.-, 

 that the sail-cloth should be placed as* near the centre as pos- 

 sible, only observing that the wind must have a free escapement ; 

 for a square foot of sail, moving with half the velocity of the 

 wind, appears to have three times the effectrVe power as when 

 moving with double the wind's velocity; tfor the power of the 

 lever when time is considered must be out of the calculation ; this 

 also agrees with Mr. Smeaton's experiments, who found, that 

 by enlarging the breadth of his sails, he gained more than by 

 increasing the radius. Probably the extremity of the sail should 

 not exceed the velocity of the wind; and as this will increase the 

 angle of weather, the wind willhavea more free escapement, and its 

 reflections be less liable to impede the following sail : the angle of 

 reflection is easily seen from the relative angle of incidence, DSR. 



2dly. That Mr. Hall Gower's hypothesis is highly disad- 

 vantageous; for by approximating to 90° at the centre it has 

 the least power, where it should have the most. 



3dly. It appears evident from theory, and all Mr. Smeaton's 

 experiments, that the greater the angle of weather the slower 

 will be the motion ; therefore if by any simple contrivance the 

 angles of weatlier could be occasionally altered, it would be the 

 best mode of making the revolutions more uniform, and even 

 of stopping them altogether: I am now making an experiment 

 at large on this method. 



4thly. Although the forces appear greatest in the first co- 

 lumn, from taking RF 3 xRGh-SN 3 as the measure, yet if the 



UF-xRG 



measure had been taken according to Maclaurin, 



SNsxSR 



then the second column had shewn the greatest forces, and 

 the third column, if Bb was the true measure — but on no hy- 

 pothesis could Gower have any competition. 



N. 13. RFxRG is a maximum when WCxcC is a maximum, 

 and RF 3x RG is a maximum, when cW x cC is a maxium — 

 the first when Wc is 4 of Wv, the last when Wc is i. of'Wv; 

 the greatest right-angled triangle in the segment VBW. 



G g 



