less. None had a valve which could be set so that the operator would know when 

 he started across the field how much solution would be applied per hundred feet. 

 Actual tests of the different types of tanks showed considerable variation in the flow 

 depending on how full the tank was. A 14-quart tank described below required 

 more than twice as long to deliver the last quart as it did the first. The one-gallon 

 copper tank first constructed by Stone (Mass. Agr. Expt. Sta. Circ. 21) — which 

 is still in the possession of the Station — was tested and showed a difference in time 

 of 15 per cent, between the first and last point. None of the tanks previously 

 described has any method of correcting this error. None of them has any provision 

 for eliminating the error due to variation in the speed of the operator. In short, 

 there were so many sources of error and variation in all of the machines described 

 up to the present that they were found to be utterly unsuitable for accurate research 

 studies and a new tank and method of regulation had to be devised. The one which 

 was finally evolved after experimenting with many different modifications is illus- 

 trated in Fig. 5 and described below.* 



The tank itself (Fig. 5A) is a 14-quart cubical galvanized iron box attached be- 

 tween the handles of the drill just back of the seed box. The bottom of the tank is 

 not flat but has a shght slope to a lowest point at the rear from which the solution 

 is conducted through a J^-inch pipe (E) to the stopcock (K) and to the union (F). 

 From there it is led through a ^-inch flexible block tin tube (G) and distributed on 

 the seed and soil just back of the seed spout and in front of the coverers. The valve 

 (K) is operated by a J^-inch iron rod (R) from the rear end of the handles. It has 

 nothing to do with the regulation of the flow but merely starts or stops the stream 

 at the ends of the row or wherever desired. 



The rate of flow from the tank is regulated by a series of brass disks with central 

 apertures of graded sizes. A disk is held in place between rubber washers in the 

 imion of the outlet pipe (Fig. 6). Any number of these simple brass disks may 

 be quickly cut out with tin shears from a sheet of brass. A hole is driUed through 

 the center of each disk and this is enlarged with a rat-tail file to the proper size for 

 the delivery of a gallon of solution in the previously calculated number of seconds 

 which the operator wishes to use. The number of seconds required to deliver a 

 gallon of solution is then stamped on the disk (Fig. 6). The size of the aperture 

 needed depends on the formula of apphcation which the operator wishes to use and 

 the rate at which he walks. The number of disks which the experimenter or grower 

 needs depends on the number of formulas he wishes to use, and the number of men, 

 of different speeds of walking, who expect to push his drill. In our own experi- 

 mental work seven different formulas of apphcation were being tested and the drill 

 was pushed at different times by three operators, the first one walking 4 feet per 

 second, the second one A]/2 feet per second and the third one 5 feet per second. Thus 

 before the experiments were completed there were 21 disks. The operator could 

 quickly find the proper disk by reference to a table attached to the side of the tank 

 (Table 1). 



TABLE 1. 



Indicating the Disks of the Onion Drill to be Used According to the Speed 



of Workman and Formula Desired 



♦The authors are glad to acknowledge the very helpful cooperation of Professor C. I. Gunness and 

 Professor J. L. Strahan of the t>epartrnent of Rural Engineering, whose aid in the solving of some of the 

 mechanical difficulties and the preparation of mechanical drawings, has been invaluable. 



21 



