1 04 



KANSAS IfN'lVERsri'V (.)l' AR IKKLV. 



Fig. 4 represents two sets of measurements. The center of 

 each star represents a measurement with tlie pump rod, discon- 

 nected 20 feet below the wheel; that is, in the set represented by 

 the stars no water was pumped — no useful work was done. The 

 mean line CD of this set is seen to be nearly parallel to AB and 

 not far distant from it. From this we see that if this mill is run- 

 ning at a given rate and not doing any useful work, it will run at 

 the same rate and do its useful work if the wind velocity be in- 

 creased three feet per second. Or, stating this in another way, if 

 the mill is doing no useful work at the rate of 25 strokes of the 

 upper part of pump rod in 58 seconds, it will do its useful work in 

 the same velocity at the rate of 20 strokes per 58 seconds. 



We see from this one reason for the great variation in the work 

 windmills do in a given wind (Table II, last four columns). The 

 heavily loaded one will do much more work than the lighter loaded 



ICOO 



/o /3 Zi? ^"^ 30 



Fiy:. II. l:Jft. sled iiiill. :.'<).:.' ft.-n)S per si rokc. 



one in a wind strong enough to run it. Fcpiation (lo) expresses 



the relation between the ratio and angle x which the fans make 



c 



with the direction of the wind for maximum power. If the angle 



is given, as it is for any given mill, we can find from this e(]uation 



the value of for an\- small area of the fans, and from this, how 

 c 



heavy to load the mill. If, then, this ratio be kept constant, as c 

 increases, by increasing the load, the maximum power may be ob- 

 tained. One device for doing this is to automatically increase the 

 length of stroke as the velocity of the wind increases. 



In Fig. 5 the circles varj^ considerably from the mean line. 

 This is due to the fact that the airmeter was not used in these meas- 



