TESTS OF MACHINERY TOR POISONING COTTON WORMS. 41 



placed in the usual position the mule must travel on a row to have the 

 two wheels straddle it properly, and this not practical. I obviated the 

 objection by a pair of rough shafts set to out- side, one shaft coming 



from the center of the cart and the other standing outside of The wheel : 



thus the mule is held nearly in front of one wheel and midway between a 

 pair of rows. In practice it is shown that the slight side-draft caused 



by this arrangement amounts to almost nothing. And it should be re- 

 membered th<at it is common in the Xorth to use side shafts on sleighs, 

 buggies. &c. The apparatus is easily pulled by one mule, which should 

 travel preferably in front of the right wheel. 



The personal labor required is such that the pumper may also drive 

 and keep an eye on the machine behind. It is most convenient when 

 the stirrer, pump, and barrel are placed near the left side of the cart with 

 the lever or pump-handle standing crosswise. The operator then only 

 looks to the right and the left instead of having to look backward. Tie- 

 mule, accustomed to working in cotton, follows between the rows as a 

 rule without being guided, and the pumper is free to turn him at the 

 ends. In this manner one man can use the apparatus alone without 

 working any harder than he should. Still, it is generally preferable to 

 have a boy attendant to assist at times, and where a larger pump with 

 a very wide system of pipes to supply a large number of rows is engaged 

 it will be well to have two men to take turns at pumping in cases where 

 a pump motor is not provided. With such labor, the adjustable under- 

 spraying machines which I have devised and shall notice below were 

 operated, the best hand being a mulatto who worked for fifty cents a 

 day, which is a common ]mce for cotton-field hands. Thus the labor 

 cost is small, and one or two men with a machine can do much more and 

 better spraying than a large gang does by ordinary methods. 



The rapidity depends altogether on the width of the pipe-system, or 

 number and width of rows supplied, or the size of the cotton, of the noz- 

 zle-discharges and of the pump, on the velocity or pressure applied, and 

 on the speed of the mule. It may vary with any one of these details. I 

 had only one pump, this rather small, and could not try the effects 

 different sizes. There was not much diversity in the cotton and it was 

 below medium size. At Selma, I labored under difficulties from bad 

 workmanship, from lack of available mechanics tor constructing the 

 devices, on which account there were bad joints in the apparatuses, which 

 leaked some, and which would loosen and at times come apart when 

 high pressure was applied, thus causing stoppages in the work. Under 

 the circumstances the question of time and scope could not be very sat- 

 isfactorily tested 5 but as to the very tine small sprays for small cotton 

 the following conclusions result from this experience : 



Taking nearly average sized cotton, and the parts of the apparatu 

 medium capacity, an acre and a half was poisoned in one-half an hour. 

 2-4 feet wide being poisoned at a single drive, and the rows were quite 

 short. Twice this rapidity can be attained. 



