Vol. 70.] ANNIVERSARY ADDRESS OF THE PRESIDENT. lxxxix 



referred, but there are many examples of artificially dammed-up 

 waters which provide opportunities of measuring the amount of 

 material transported from a known area in a known time. Mill- 

 dams are the least satisfactory for the purpose, inasmuch as a part 

 of the stream only is diverted to the reservoir. In flood- time, when 

 practically all the work of the river is done, no record is preserved 

 of the material which is rolled or carried down the main stream. 



But there are many ornamental waters and reservoirs for various 

 purposes through which the whole stream passes, and in which the 

 entire burden of solid impurities is intercepted. Indeed, works are 

 constructed for the express purpose of clearing Avater of suspended 

 matter ; yet, although the amount of water passing and the rainfall 

 have been generally observed, the bulk of the solid contents of the 

 Avater has seldom been measured, possibly from motiA r es of discretion. 



Unpler this head, also, may be placed rivers controlled by weirs. 

 An example will be referred to later in a description of some 

 experiments on the Exe, where the amount of material removed 

 from the channel annually above a dam, in order to keep the channel 

 at a constant depth, has been recorded. Canalized riyers present 

 a different problem. In normal weather the water is held up at 

 a uniformly low gradient by a succession of weirs. Its poAver to 

 roll material along the bottom is obviously annulled, and its ability 

 to keep matter in suspension greatly diminished. It \A r ould appear 

 at first sight that the river must be paralysed as a denuding agent. 

 That this is not the case is clear enough during the passage of a 

 flood. Weirs at such times are thrown open, or may even be so 

 deeply submerged as to play the part of mere rugosities in the 

 channel. The poAver of a river at normal level to roll material is 

 probably negligible, whether canalized or free ; its poAver in flood- 

 time is apt to disregard control. At such times holes are dug in 

 one part or another of the bed, shoals are formed in others, but 

 always Avith the result of transporting material a stage farther doAvn 

 the valley. These effects are counteracted by dredging. Obviously 

 then, a canalized river should provide unusual opportunities for 

 ascertaining the rate at which material is transported by rolling. 

 The amount dredged betAveen each pair of weirs should, in the 

 course of years, give an accurate account of the activity of the river 

 in every part of its canalized course. 



The distance over which rivers are noAv rolling coarse material is, 

 I believe, not beyond the reach of actual observation. An exami- 

 nation of the stones in the recent gravels and alluvium of a river 



