602 



SCIENCE. 



[N. S. Vol. XIX. No. 485. 



stream, and they had to be raised and 

 strengthened to meet the added load. These 

 additions were made as the need of them 

 developed. These needs were forecast from 

 time to time as nearly as possible. It was 

 a difficult problem. The 'potential high 

 water' of floods to come has been the sub- 

 ject of much study and discussion. The 

 nearest approach to a standard has been 

 that the levee should be three feet above 

 the highest previous flood line in that lo- 

 cality. In order to ascertain this line it 

 has been the practice to record extreme 

 high-water levels by marks on trees at in- 

 tervals of a few miles along the bank. 

 These records show such irregularity of 

 behavior in the great floods as to make it 

 necessary to fit the grades of the levee to 

 local conditions. One elenient of disturb- 

 ance and uncertainty yet remains in the 

 completion of the closure of the St. Francis 

 front, where about 60 miles of embankment 

 remain to be built. That great basin, 6,700 

 square miles in area, has exercised in the 

 past a profound influence on the channel 

 along and below it. In its natural state it 

 received a vast volume of overflow which it 

 returned to the main stream again at and 

 about the mouth of the St. Francis, imme- 

 diately above Helena. This abstraction of 

 water from the river along the iipper and 

 central portions of the front of that basin 

 Aveakened the stream and so tended to 

 shrink the channel in those parts ; while its 

 return at the foot of the basin augmented 

 the stream, with a resulting tendency to 

 enlarge the channel. Consequently, when 

 we come to confine the flood discharge by 

 levees along the St. Francis front, the 

 water, in passing down, finds a channel of 

 more ample dimensions in the neighbor- 

 hood of the foot of the basin than that 

 which it finds above. It is to be expected, 

 therefore, that a great flood will reach a 

 higher elevation in those parts of the chan- 

 nel where, for ages past, the flood volume 



has been depleted by overflow than in those 

 parts where, during the same ages, the 

 flood volume has been augmented by re- 

 turn of the overflow. 



Still other circumstances enter in to com- 

 plicate the problem, so that, upon the 

 whole, only the actual confinement of a 

 series of great floods without a break from 

 Cairo to the sea will give us the ultimate 

 high-water profile. 



But we are making rapid progress toward 

 that information. The great floods of 1897 

 and 1903 carried more water to the sea be- 

 tween banks than any of their predecessors. 

 They have left records of gauge readings 

 and discharge measurements which afford 

 a great field of interesting study into which 

 I can not enter within the time and space 

 at my command. Those limitations will 

 permit only the briefest summary of what 

 the levees accomplished and what they 

 failed to accomplish toward the protection 

 of the lands behind them. They were 

 floods of the first rank and may be taken 

 as typical of what may be expected to occur 

 at intervals of a few years in the future. 



The flood of 1897 made 38 crevasses, hav- 

 ing an aggregate width of about 8 miles; 

 the flood of 1903 made 9 crevasses, having 

 an aggregate width of about 3 miles. The 

 levees in place in 1903, if no crevasses had 

 breached them, would have protected about 

 26,000 square miles from overflow. Of that 

 area a total of about 3,000 square miles was 

 overflowed in conseqiience of the crevasses 

 which took place, which is less than one 

 eighth of the entire area which the existing 

 levees could and would have protected if 

 they had all been high enough and had held 

 their places. In the phrase of the target 

 shooters, they accomplished 87-^ per cent, 

 of success out of a possible 100. 



These experiences indicate that with com- 

 plete restraint of the floods by levees we 

 shall have, as an immediate result, some 

 further elevation of the maximum flood 



