CONDUITS, PORTAGE, ETC. 295 



tached, and is designated by the letters o wv c t, &c. The Y-forks, tyn, 

 are the same as have been described. Each of these is joined, at v, to 

 a side branch of the pipe on the side-bar. Such a branch is shown at 

 V J), Plate XLIX, Fig. 8. This branch is preferably made of a piece of 

 tube cut very diagonal and sweated on over a hole in the arm-pipe. 

 The diagonal cut gives it a much longer attachment, and the acute angle 

 can be more easily filled or braced, at v, than could a right angle. Even 

 where a right-angled fork is desired it had better be formed in this way 

 and then have a curve to bring its distal parts to a right-angled direc- 

 tion. The straight part shown in Fig. 8 can be very short, and thus 

 have only the length of a coupling, or it may extend as an entire long 

 section reaching to the next branch. In the latter case, the machine 

 \rill embrace fewer pieces and be simpler in construction. The sections 

 of pipe upon the frame are i)referably joined together by sheath-coup- 

 lings of hose with single wraps of wire or metal bands. The main 

 branch, o, of this pipe-system is connected by a hose to the spout of a 

 force-pump, operated as in Plates* XLYI aud XLYII. Any suitable 

 force-pump may be used 5 but that described (p. 282) and illustrated 

 (Plate XLYI) is probably the best. 



This machine is probably the broadest one ever put on wheels. It 

 poisons sixteen rows at one time, a strip about 60 feet or 20 yards wide. 

 Yet the whole concern is light, and as I tested it in the cotton, the pull- 

 ing was easy work for two horses. This is about the maximum size 

 that can be used practically, and experience may demonstrate that it is 

 better to use a narrower, more wieldy machine and drive at greater 

 speed. One of the smallest force-pumps of the trade with a f -inch or 

 |-inch spout supplies it with sufficient liquid and pressure without being 

 tiresome work for one man, and with my nozzles it required only 10 to 

 15 gallons to poison an acre, whereas, before their introduction, 40 

 gallons or more were required. By these devices the poison can be 

 more evenly and sparsely diffused, and one-third the amount of poison 

 thus applied will be equallj^ as effective as the whole amount admin- 

 istered with the other devices now known. With large cotton more 

 poison and more water should be distributed, so the amount of water 

 with these nozzles will probably vary from 10 to 25 gallons per acre, 

 and the poison in proportion. After the pipes are once cleared of rub- 

 bish, the strainer on the pumj) prevents clogging material from reach- 

 ing the nozzles and there is comparatively little difficulty from choking 

 of the outlets. When this does occur each nozzle is readily opened to 

 remove any obstructing object. At the outset it is desirable to pump 

 very rapidly until the pipes are filled, all the air is out, and a good 

 pressure established equally upon all the jets 5 then the spray will be 

 discharged regularly, filling the rows with a very fine mist, which poi- 

 sons all surfaces of the plant but especially the under-surfaces, where 

 the poison will be practically permanent to destroy existing worms or 

 prevent others thereafter. 



