32 



TRANSPORTATION OF DEBRIS BY RUNNING WATER, 



dunes closely and shares their transformations. 1 

 (See fig. 11.) Usually each antidune occupied 

 the full width of the experiment trough; and in 

 natural streams, so far as I have observed, they 

 either reach from side to side of the channel or 

 else form well-defined rows in the direction of 

 the current. Not only is a row of antidunes a 

 rhythm in itself, but it goes through a rhythmic 

 fluctuation in activity, either oscillating about 

 a mean condition or else developing paroxys- 

 mally on a plane stream bed and then slowly 

 declining. Paroxysmal increase starts at the 

 downstream end of a row and travels upstream, 

 gaining in force for a time, and the climax is 

 accompanied by a combing of wave crests. 

 Where the debris is very coarse, as on the out- 

 wash plains of glaciers, a din of clashing bowl- 

 ders is added to the roar of the water. 2 



Of the phases of process in the laboratory 

 Mr. Murphy writes : 



Their [the dunes'] form is continually changing as they 

 move forward; they divide and again unite, the parts 

 traveling at different rates, and new ones form on top of 



rl.O 



the older ones. The grains roll up the gentle slope, fall 

 over the crest, are covered by other grains, and rest until 

 the dune again passes over them an d they are again uncov- 

 ered. Thus the time during which they are in motion is 

 small compared to the time during which they rest. As 

 the velocity of the current increases, the rate of feeding 

 being correspondingly increased, the size of the dunes and 

 their rate of movement increase. Thus we find that when 

 the discharge is 0.363 ft. 3 /sec., load 11 gin. /sec., and slope 

 0.32 per cent, the dunes are 7 to 9 inches long and one-half 

 inch high and move at the rate of 0.56 foot a minute, but 

 when the discharge is 0.734 ft. 3 /see. and the load 30 

 gm./sec., the dunes are 13 to 15 inches long, three-fourths 

 of an inch high and move at the rate of 1.5 feet a minute. 

 As the velocity of the current increases some of the grains 

 leap as well as roll, and some, instead of dropping over 

 the crest of a dune and resting, leap to the next dune. 

 The dune grows less distinct in form and finally at a criti- 

 cal velocity it disappears, dune motion ceases, and the 

 sand surface becomes comparatively even. This condi- 

 tion of even surface flow continues as the slope increases 

 until at another critical velocity antidune movement 

 begins. A profile of the sand surface for this kind of 

 motion is shown in figure 11. For this experiment the 

 trough width is 1.32 feet, discharge 0.734 ft. 3 /sec., load 

 213 gm./sec. and the sand slope 1.23 per cent. These sand 

 waves are from 2 to 3 feet in length from crest to crest, they 

 extend the width of the trough, and some of them are 0.5 



6 '9 



FIGURE 11. Longitudinal section illustrating the antidune mode of traction. 



trough. 



12 15 



The numbers show distance in feet from the head of the experiment 



foot in height from crest to trough of the wave. They 

 travel slowly upstream, some of the sand being scoured 

 from the downstream face in the vicinity of Y (fig. 11) 

 and deposited on the upstream face at X. Some of these 

 waves remain for two minutes or longer, but most of them 

 not longer than one minute. A whitecap forms on the 

 surface of the water when the larger waves disappear. 

 Sometimes two or more will disappear at once and leave 

 the surface without waves for a distance of 10 feet or more. 

 Only a portion of the sand transported takes part in the 

 formation of these sand waves. The velocity in a wave 

 trough is greater than near the crest. The sand grains 

 flow nearly parallel to the bed as they pass through the 



1 Antidunes, though less common than dunes, are by no means rare 

 under natural conditions. They are described by Vaiighan Cornish 

 (Geog. Jour. (London), vol. 13, p. 624, 1899; Scottish Geog. Mag., vol. 17, 

 pp. 1-2, 1901), and are mentioned by John S. Owens (Geog. Jour., vol. 31, 

 p. 424, 1908). 



2 The sequence of bed characters dune, smooth, antidune was 

 observed by John S. Owens in studies with natural currents in 1907, and 

 the characters were correlated with velocities. With depths of 3 to 6 

 inches and a bed of sand, he noted sand ripples [dunes) when the ve- 

 locities, measured by floats, were from 0.85 to 2.5 ft./sec., and the 

 appearance of antidunes at a velocity of about 3 ft./sec. (Geog. Jour. 

 (London), vol. 31, pp. 416, 424, 1908). Sainjon and Partiot, study- 

 ing the movement of debris in the Loire, had previously observed 

 that whereas with low velocities the entire bottom load was transported 

 through the progress of dunes, with higher velocities the del>ris was swept 

 along from crest to crest and the dunes were reduced In height (Annales 

 des ponts et chaussees, 5th ser., vol. 1, pp. 270-272, 1871). 



trough, but at the crest they have an up and down motion 

 as well as a forward motion. On the crest of the larger 

 waves their forward motion is small compared with their 

 vertical motion. * * * 



There is a sand movement by rotation or whirls that 

 aids transportation. These whirls have been observed 

 during dune motion only for smaller sizes of sand . They 

 are of short duration, lasting usually less than one minute, 

 but in this time one of these may scour a hole 1 to 3 inches 

 deep and 4 to 10 inches in length. They usually start 

 near the side of the trough, the axis inclining downstream 

 and toward the center, making an angle of 30 to 60 with 

 the side and a small angle at the bottom. These whirls 

 are 3 to 5 inches in diameter and the sand grains are thrown 

 violently up as well as downstream by them. This move- 

 ment aids transportation by its lifting action, some of the 

 grains being carried in suspension for a short distance by it. 



The change in the appearance of a loaded stream as the 

 load is increased, the discharge remaining constant, is 

 very striking. For no load the water surface is even and 

 smooth. As fine sand is fed into the water at a slow rate, 

 small sand dunes will form on the bottom and many little 

 waves will form on the surface. As the rate of feeding is 

 increased, the slope and velocity increasing, these waves 

 become larger and fewer and have the shape of an inverted 

 canoe. These canoes are side by side, the number 

 depending on the trough width and size of waves. When 

 the width was 1.0 foot two sets formed side by side; when 



