8 BULLETIN 3*76, U. S. DEPARTMENT OF AGRICULTURE. 



has persisted that there is less friction in wood pipe than in metal 

 pipe. It has often been insisted that new wood pipe not only has 

 a higher carrying capacity than new metal pipe but that the wood 

 pipe becomes smoother with age, while it is a well-known fact that 

 metal pipe becomes rougher. (See discussion, p. 72.) 



While the analysis of all the tests on wood pipe now available bears 

 out the above claims in a general way (excepting that wood pipe 

 is not shown to become smoother with age), yet the consideration 

 of tests on individual pipes led to hasty conclusions presently shown 

 to be greatly at variance with facts. The following ideas of hydrau- 

 licians have been extracted by the writer from all the literature on 

 the subject known to him: 



The experiments of Darcy and Bazin in 1857 and 1859 (Nos. 22 

 and 33) and of Clarke in 1884 (No. 49) were considered but little in 

 later discussions for the reason that they were made on rectangular 

 rather than on round pipe. Smith's test (No. 1), made in 1877, 

 has also not been considered in the discussion of wood pipe, as the 

 test was made on a bored pipe of very small caliber; yet these four 

 series supplied the data upon which Tutton based his formula. (See 

 p. 50.) 



Although none of the 81 tests considered by Kutter and his col- 

 league in developing the Kutter formula had been made on closed 

 channels rimning full, yet nearly all of the experimenters on wood- 

 stave pipe have determined for their tests the value of n in this 

 formula. Kutter's formula has undoubtedly been used a great deal 

 in estimating the capacity of wood pipes, but the writer will endeavor 

 to show (p. 56) the fallacy of employing a constant value of n in 

 this formula and the advantages lying in a formula of the exponential 

 type. 



The first experiment of pubUc record was mentioned by the late 

 J. D. Schuyler ^ in speak,ing of a test (No. 34) on the newly installed 

 30-inch pipe for Denver, but unfortunately he did not give sufficient 

 details by which the test might be weighed. Mr. Schuyler states 

 that " as low a coefficient of n as 0.0096 can be used." This appeared 

 reasonable, as the pipe was made of planed lumber and all lists of 

 proper values of n then pubMshed recommended a value of 0.009 for 

 such material. The earlier designers adopted a value of 0.010 "in 

 order to be conservative." 



The next tests were made by A. L. Adams ^ on the Astoria, Oreg., 

 18-inch pipe (No. 23) . Here, too, a low value of n was found, 0.00985, 

 which led Adams to observe "that the value of 0.010 for n used by 

 many engineers in deahng with stave pipe, is here found to be practi- 

 cally correct." The low value of the friction factor found in this 



1 Trans. Amer. Soc. Civ. Engin., 31 (1894), p. 144. 



2 Trans. Amer. Soc. Civ. Engin., 36 (1896), p. 26. 



