380 The National Geographic Magazine 



crease in the width is a greater hin- 

 drance to bending than an increase in 

 the thickness in the proportion of the 

 square of the increment, doubled width 

 being four times as effective for resist- 

 ance as doubled depth. 



Strictly, in the case of the turning 

 river, it is a matter of momentum. 

 Could we have a single thread of water 

 flowing along a curved or crooked 

 channel, its turns would be made with 

 ease, however sharp ; but actual streams 

 are alwa3's made of many threads of 

 water side by side, and when the outer 

 thread seeks to rebound from the outer 

 bank at a turn, the. momentum of the 

 inner threads drives it along in a course 

 that is kept straighter in proportion as 

 there are more inside threads of current. 

 A column of men, marching, turns as 

 easily with twenty files abreast as with 

 one, but only because the inside man 

 has been trained to stand still and 

 merely rotates 011 his axis until his 

 companions have got around the corner. 

 The inside threads of river current have 

 no volition, no training to stop and wait 

 for their neighbors outside. Their mo- 

 mentum carries them along and their 

 number makes the turn longer, the 

 meander belt wider. 



In view of these considerations let us 

 return to our table of meander belts and 

 consider the stream width in each case. 

 The table is here reproduced, omitting 

 fall and discharge, and introducing two 

 new columns, headed w for width of 



stream in feet, and — , or ratio of mean- 



w 



der belt to width of stream. 



River. 



mb 



IV 



mb 



w 



Place. 



Matfield 



450 

 46S8 

 55.ooo 



4,000 

 8,448 

 9./6S 

 9,000 

 26,400 



30 

 243 



2,830 

 260 

 660 

 528 

 800 



1,400 



15.0 

 59.3 

 19.4 

 15.4 

 12.8 

 18.5 

 11-2 

 18.9 









Mississippi. ... 



Greenville, Miss. 







Greenbrier. ,.. 

 Shenandoah .. 

 Tennessee 



Alderson, W. \*a. 

 Near Potomac River 

 Chattanooga, Tenn. 



Here is a good suggestion of a con- 

 stant value in the column — . But 



w 



these rivers were selected under con- 

 straint when I was trying to utilize the 

 best meanders that occurred near points 

 where measurements of volume had 

 been made. We are now able to select 

 the best flood plains, and on these the 

 best-developed meanders on any Ameri- 

 can or European maps that are accessi- 

 ble. The list follows as Table A. No 

 stream has been excluded because its 

 ratio was discordant. Many were re- 

 jected because of too small a value of 



the quantity —z-, which gives the ratio 



of measures at meanders across and 

 along the general river course ; d rep- 

 resents the distance from meander to 

 meander along the river axis. Starred 

 rivers have cuts-off near. 



The mean meander ratio is 17.6. 

 Study would doubtless remove some of 

 the discordancies. The Rhine near 

 Speyer and Worms, as the accompany- 

 ing map shows, is a corrected stream, 

 flowing in an artificial channel. The 

 width of the meander belt may be meas- 

 ured from the old course, which still 

 subsists, but the width of the river, 

 measured on the artificial channel, 

 which is confined between walls, is 

 probably less than that of the uncor- 

 rected stream that made the meanders. 

 This must tend to give an excessive 

 value to the meander ratio. 



It is interesting to observe in the case 

 of the Rhine that the flood-plain width 

 is not far from the width of the maxi- 

 mum meander belt. 



The small ratio for the Mississippi at 

 Baton Rouge is of interest with the 

 opener character of the meanders in the 

 lower course, where each arm of the 

 river seems to point away from the next 

 arm upstream instead of swinging around 

 toward it. A little farther and the 

 river stops swinging, to rush headlong 



