108 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



The first problem was to ascertain the acreage or area drained by 

 these ravines, then to divert the drainage in such a way as to divide it 

 proportionally through the various existing ravines. Then came 

 the huge bund treatment which was to check the " run off." Lakes 

 were thus formed during the rains, and each reservoir or catching area 

 had to be provided with a safety outlet or escape so as not to imperil 

 the whole scheme, each catching area becoming a lake after its super- 

 imposed one had been filled, and this procedure had to be followed 

 till the overflow of the entire system was made to run into the river. 



Large quantities of rainfall were so retained, the gradual silting up 

 of the beds of these ravines or lakes went on with remarkable rapidity. 

 Similar operations were also conducted within the Government Grass 

 Farm Lands, and in less than five years the topography of the place 

 had changed. 



Permanent lakes combining beauty with utility were then made, 

 and these now collect most of the rainfall, the upper lake discharging 

 into those on the lower levels, which beds have been so designed as to 

 make use of every drop of water for irrigation of low-lying grounds. 

 When these lakes are emptied, they are fed from the canal, and are a 

 component part of one of the most effective Park irrigation schemes 

 met with in India. 



The difference of levels between the feeding minors and the lakes 

 is utilized to work several sets of hydraulic rams that force water 

 to the high ground, enabling over 75 acres of highland to be irrigated. 



A similar treatment of ravines was also conducted some twenty-five 

 years ago at the Allahabad Grass Farm. Hence, we have object-lessons 

 that should be generally applied throughout Upper India. 



On the Southern Ridge the training of ravines should in the 

 first instance aim at retaining the " run off " from all tablelands, and 

 this can be easily done by providing against the heaviest recorded 

 rainfall that may have taken place over a certain number of years. 

 The heavy downpour experienced during the year of the last Durbar, 

 which amounted to about 8 to 9 inches in 36 hours, may be taken as a 

 basis ; this, with the area draining on the tableland or particular catch- 

 ing area, would determine the height and relative position of the pro- 

 tective cross bund or bunds. It will be understood that the above will 

 lead to most of the upland " run off " being disposed of at once and 

 retained at its highest possible level, relieving the ravine beneath. The 

 diagram on p. 105 illustrates the above. It will be seen that the 

 second cross ravine bund has a comparatively small strain and so 

 on till the position of the last cross ravine bund, which will be found 

 necessary, is determined. 



In this particular case these cross ravine training bunds are made 

 of dry stone walling, but it is a question whether it would not be 

 more economical to make them in lime masonry and convex against 

 the flow of water (this is the distinction between stone or masonry 

 bunds and earthen ones). A comparatively small section should in 

 some cases suffice to resist the inrush of water, but stone being abun- 



