THE ACTION OP FLOWIlfG STEEAM8. 737 



spreads over it, while the narrow channels are distinctly depressed 

 below the general surface. Taking the slope of the latter to repre- 

 sent the average slope of the stream, this alternate elevation and 

 depression of the stream-bed with reference to it accords perfectly 

 with my hypothesis. 



Besides my own observations, I have studied the maps produced 

 by the Survey of India, and find that wherever the circumstances 

 are favourable this alternate expansion and contraction of the stream- 

 bed is very noticeable ; nor does it seem to be confined to small 

 streams only, but may also be detected, though, as might be expected, 

 obscurely, in such large rivers as the Ganges and the Jumna. 



Turning now to the Ganges-canal levels, if the maps in the 

 Ganges-canal atlas are examined and the position of contour lines, 

 drawn at a vertical distance of five feet apart on the Patthri and Rani 

 Eaos, be represented by dots, the crowding of these together where 

 the channel broadens is palpable, but the original map does not 

 indicate where the channel is well and where ill defined. Portu- 

 nately the text comes to our assistance and we are informed that 

 in the case of the Patthri Rao the canal crosses it where it had 

 no defined channel ; in the case of Rani Rao the canal was taken 

 across a well-defined channel, but we are informed that 2000 feet 

 below the canal this stream ceased to have a defined channel and was 

 spread abroad over the country ; and it is evident at a glance that 

 the gradient is there higher than where the channel is well defined, 

 precisely as should be the case if my hypothesis were true. 



These two torrents are carried over the canal by " superpassages," 

 that is bridges built so as to form a continuous channel for the 

 torrents. In the case of the Patthri Rao the course of the stream 

 was diverted, and it is consequently of no use for our present purpose. 

 But for the Rani Rao, the superpassage was built so as to replace 

 that part of the defined channel which was removed in excavating 

 the canal. Within a few years of its completion, it is recorded 

 that the region where the channel was no longer well defined, 

 instead of commencing 2000 feet below the superpassage, had 

 extended right up to it, and presently sand began to accumulate on 

 the bridge. This accumulation of sand for many years gave much 

 trouble to the engineers in charge, but it has now ceased. Whether 

 this is, as the engineers believe, altogether due to their works, or 

 whether it is due to the defined channel below the fan having 

 worked up to the bridge, or whether, as is most probable, it is due 

 to a combination of the two, the records are not sufficient to deter- 

 mine. 



Thus the surveys for and records of the Ganges Canal show the 

 alternation of reach and fan, the greater gradient of the latter and 

 the gradual progress of both up stream as demanded by the hypo- 

 thesis formulated above. The small stretches of high and low gra- 

 dient are not shown in these records, being probably of too small 

 extent and too difficult to detect without special search, but they 

 are indicated by the later records of the Patthri Rao. Here the 

 stream, after crossing the superpassage, is artificially confined in a 



