Quantity of Water Consumed in Irrigation. 33 



being compared, the points suggested further on (see par. 

 32) should be noted in every case. And, further, when the 

 field-channels are unusually long, it would be better to con- 

 sider them separately. 



IV. — Loss by Evaporation and Percolation. 



15. The remainder of the consumption (averaging, in the 

 case of the Ganges canal for the year 1875-76, 51,000 cubic 

 feet per acre) represents loss by evaporation and perco- 

 lation in the various channels from the point where the 

 discharge was gauged to the point where the water was 

 issued for individual fields. The total quantity lost daily 

 over the various channels included on this occasion was 192 

 million cubic feet (see App. B), which will probably be found 

 a near approximation to the loss each year on this canal. The 

 author particularly wishes to urge that this loss, instead of 

 being referred to as so much per cent, of the supply or so 

 many cubic feet per second per mile, as usual hitherto, 

 should be expressed by the depth spread over the whole 

 wetted area of the bed and slopes of the various water- 

 courses. In this form, results obtained on one canal could 

 be applied, as the author will endeavour to show, to other 

 canals, not only in the same country, but also to canals in 

 other countries. The chief points for consideration are 

 noted further on (see pars. 29 and 33). 



16. The author has calculated that the loss daily by 

 percolation from the Ganges canal and its distributaries in the 

 cold season of 1875-76 was from 4 J- inches to 7 J inches in 

 depth over the whole wetted surface (see App. B). The 

 lesser depth supposes the loss from the canal itself to be in- 

 cluded in the 102,000 cubic feet of water consumed per 

 acre ; the greater depth supposes the measurements of 

 discharge to have been made at the distributary heads, thus 

 taking no account of the loss in the canal itself. The latter 

 supposition appears to the author more likely to be correct 

 than the former (see par..-7). Assuming, then, the greater 

 depth as correct, and that the loss in the canal was 

 equal in depth to that in the distributaries and minor 

 watercourses, the total loss during the season would 

 suffice to fill a trench having a width equal to that 

 of the wetted perimeter and an average depth of over 

 90 feet, extending the whole length of canal and dis- 



