288 EXPEKIMENT STATION RECOED. 



per acre is best, irrigating 3 or 4 rows at once. Water should be applied 

 directly to each check from the ditch and the water surface in the ditch should 

 be at least 1 ft. above the highest check. 



Derivation of run-off from rainfall data, J. D. Justin {Proc. Amer. Soc. 

 Civ. Engin., 39 {1913), No. 6, pp. 1211-1228, pi. 1, figs. 20).— In an attempt to 

 develop a rational method of deriving run-off from rainfall data on various 

 watersheds the author found that on the watersheds examined the relation may 

 be expressed by the formula G=KR^, in which G is the annual run-off in 

 inches, R the annual rainfall in inches, and K is a constant depending on the 

 variations in the relations between rainfall and run-off from one watershed to 

 another, depending on slope and mean annual temperature. Comparisons of 

 numerous watersheds also indicated that the relation of run-off to rainfall may 

 be expressed by the general formula C=0.934 S^-^^^ R\ in which T is the mean 



T 

 annual temperature and S the slope of the watershed (the difference between 

 the highest and lowest elevations divided by the square root of the area). 



Numrons rainfall and run-off curves and tables of data are presented to sub- 

 stantiate this view, and tables are given to aid in the solution of the formula. 

 The author believes this formula to be applicable to watersheds in the eastern 

 United States and suggests its use, in that part of the country, where run-off 

 data are meager or lacking. 



Seepage losses from earth canals, E. A. Mokitz {Engin. News, 70 {1913), 

 No. 9, pp. 402-405, figs. 2). — This article considers the so-called nonpreventable 

 losses from seepage or percolation through the bed or banks of earth canals, 

 and presents figures said to represent the average results obtained from observa- 

 tions on several hundred miles of canals on 8 different reclamation projects 

 to express the losses in terms of depth in feet in ^4 hours through the wetted 

 perimeter of the canal prism as follows: Cement gravel and hardpan with 

 sandy loam 0.34 ft., clay and clay loam 0.41 ft., sandy loam 0.66 ft., volcanic 

 ash 0.68 ft., volcanic ash with some sand 0.98 ft., sand and volcanic ash or clay 

 1.20 ft., sandy soil with some rock 168 ft, and sandy and gravelly soil 2.20 ft. 

 It is concluded that the limits within which seepage losses should be considered 

 in earth canal design may be generally defined as 0.5 ft. and 2.5 ft. per 24 hours 

 over the wetted area of the canal 



From a mathematical consideration of seepage losses the following equation is 



derived: i'=0.2 cXy^r, in which s is the seepage loss in second feet per mile 

 of canal, Q the canal discharge, V the mean velocity of flow, and c an experi- 

 mental coefficient equivalent to the depth of water in feet lost over the wetted 

 area in 24 hours. 



A diagram platted from results obtained from this equation, using values of c 

 varying between the above prescribed limits, is given. This shows the effect 

 of variations in velocity on these results and the advantages of using as high 

 velocities as possible. The magnitude of the error claimed to be involved in 

 stating the seepage loss in percentage of the flow is also illustrated. 



The development of balancing devices for centrifugal pumps, A. V. 

 Mueller {Engin. Netvs, 10 {1913), No. 11, pp. 490-494, figs. 2i).— Several de- 

 vices for caring for the axial thrust of the shaft in centrifugal pumps are 

 described and their design analyzed. Both partial and complete balance devices 

 are dealt with, and it is stated in conclusion that entirely automatic devices 

 have the advantage over incompletely balanced devices in that they rid the 

 pump of such members as marine thrust collars, ball bearings, etc., which are a 

 constant source of trouble. 



