THE OBSERVATIONS. 



25 



as collected, the ordinary measurement of col- 

 lected load is presumably affected by a smaller 

 probable error. 



DETERMINATION OF SLOPE. 



The observations of slope were made with 

 surveyors' level and rod. The rod, made for 

 the purpose, without unnecessary length or 

 weight, was graduated to hundred ths of a foot 

 and read by eye estimate to thousandths. It 

 was held by an assistant while the observer 

 and recorder stood at the telescope. The po- 

 sitions were determined by a graduation of the 

 trough, which was marked at every foot. To 

 measure the water slope, heights of the sur- 

 face were taken at several points along the 

 trough. To measure the sand slope, heights 

 were taken at intervals of ither 2 or 4 feet, the 

 shorter interval being used with the shorter 

 trough. The water slope could not be meas- 

 ured when the surface was rough. When the 

 debris surface was rough, it was usually graded 

 before measurement by scraping from crests 

 into adjacent hollows. 



The observed heights were plotted on section 

 paper, with relatively large vertical scale, and 

 a straight line was drawn through or among 

 them. The line served the purpose of a pre- 

 liminary determination of slope, and the plots 

 were inspected for the detection of systematic 

 errors. As a result of this inspection a portion 

 of the profile was selected for the determina- 

 tion of slope, and from the observations on this 

 portion the slope was computed by least- 

 squares method. 



CONTRACTOR. 



As will be explained more fully in another 

 connection, the slope measurements were af- 

 fected (1) by systematic errors connected with 

 the conditions under which the water entered 

 and escaped from the trough, and (2) by acci- 

 dental errors arising from rhythm. One of the 

 measures used to diminish the systematic er- 

 rors was the contraction of the current at the 

 outfall end of the trough. The apparatus for 

 this purpose consisted of two boards as wide 

 as the depth of the trough and arranged as in 

 figure 3. Their attachment to the sides was 

 flexible, so that the degree of convergence and 

 the width of aperture at the outfall could be 

 modified at will. This apparatus will be called 



the outfall contractor. The theory and effi- 

 ciency of the contractor will be considered in 

 the discussion of the slope errors. 



FIGURE 3. The contractor. 



MEASUREMENT OF DEPTH. 



The depth of the current was measured ai 

 mid width and near midlength of the trough. 

 The determination was made by means of the 

 gage already described (p. 21), during the 

 period of tune for which the load was measured. 

 As the water surface was subject to rhythmic 

 fluctuation, a series of observations of its posi- 

 tion were made, and their mean was used. A 

 series of observations of the position of the de- 

 bris surface were sometimes made also, but 

 usually only a single observation, and the read- 

 ing obtained was subtracted from the mean of 

 readings on the water surface. The observa- 

 tions of the d6bris surface were subject to an 

 error which was regarded as more serious than 

 that of the observations of water surface be- 

 cause, being essentially systematic, it could 

 not be eliminated by repetition. The prssence 

 of the gage rod in the water modified the dis- 

 tribution of velocities, and this modification in- 

 cluded an increase of the current's velocity a 

 little below the end of the rod. As the bot- 

 tom was approached by the rod, the current 

 scoured a hollow in the bed immediately under 

 it; and if the rod were lowered to actual con- 

 tact, the reading would give an excessive esti- 

 mate of depth. What was attempted was to 

 lower the rod to a position as nearly as possible 

 at the level of undisturbed parts of the bod 

 surrounding the visible hollow. This was a 

 matter of judgment, but not of confident judg- 

 ment, because the actual bed was concealed by 

 a cloud of saltatory debris particles. It is 

 therefore recognized that the measurements of 

 depth are uncertain. 



Whenever the water profile as well as the 

 debris profile was surveyed, an independent 



