MECHANICAL rEEPARATIOX OF SOILS. 43 



marked out his feeding-drains, and inserted pipes with elbow-joints at their 

 outfall to receive them, he begins by opening those at D, farthest from the 

 outfall, filling in each as fast as the pipes can be Liid. This is necessary, 

 as frosts might occur, and cause considerable damage to the open works. 



109. In the soils we have indicated, the main drain M 2 would require to 

 be both deep and strongly constructed, probably fom* feet, laid with 9 or 12-inch 

 pipes, with jimction-sockets and elbow-joints to receive the lateral drains. The 

 lateral main drains, Ml, M 3, M 4, should be three feet deci?, increasing to four 

 feet near to the junction, entering the main drain with a fall of six inches in 

 the last six feet, the pipes four or six inches in diameter ; but each the result 

 of close calculation of the water to be carried off, 



no. The feeders /^^ to the main drain M 2, should be 3-inch pipes, begin- 

 ning at three feet, increasing to three and a half, as in the former case ; while 

 for the feeders gff, of the lateral main di-ains, 2-inch pipes will probably suffice. 

 Each of the feeders should be from fifteen to twenty feet apart for garden 

 purposes. These should enter the main drain obliquely, and not at right 

 angles, as in the engraving. 



111. Where house sewage and the overflow of cisterns have to be provided 

 for, a barrel-drain will probably be requisite ; and its size will be a matter of 

 calculation before construction. 



112. As in the case of the lateral main drains, so in the feeders ; the fall at 

 the junction should not be less than 3 or 4 inches ; they may be from 15 to 

 20 feet apart, according to the nature of the soil, and their minimum dej)th 

 30 inches : this will allow for the deej^est trenching without disturbing 

 the drain. 



113. The best pipes for these feeders are 2-inch pipes, which vaiy in dif- 

 ferent localities from 11 to 15 feet in length, and are from 3O5. to 355. j^er 1,000. 

 They are sometimes laid with collars, that is, short pieces of piping sufficiently 

 large to receive the ends of two pipes ; thus keeping them firmlyin their place. 

 These collars are sometimes perforated on the upper surface to admit the water. 

 We have sometimes seen the pipes bound together by bands of tempered clay 

 at the junction of the pipes, which answei'ed very well ; but in this case the 

 upper sides of the pipes were perforated with holes for the reception of the 

 water, so that the solid junction of the i^ipes was unobjectionable on that 

 account. It is not usual, howevei*, to do more than lay the pipes in a straight 

 line when the workman is left to himself ; and few will do that unless very 

 closely watched. In the junction with the main drains the union should be 

 carefully made, either by clay or cement, where iiermanent drainage is 

 expected. 



1 14. The rationale of di-ainage is veiy happily explained by our contributor 

 Mr. D. S, Fish, Lady Cullum's gardener at Hardwicke. ''Drainage," he 

 says, " as popularly understood, means the art of laying land diy. This, 

 however, is a very imperfect definition, either of its theoretical principles or 

 practical results. Paradoxical as it may appear, drainage is almost as useful 

 in keeping land moist as in laying it diy. Its proper function is to maintain 



