LXXVI. REVIEW OF WATER CONSERVATION IN N.S.W. 
A most difficult case is that of a river which dries up 
altogether for long periods. In such circumstances it is 
best to construct the storages continuously in the bed of 
the channel, leaving them to be filled by the periodical 
floods. Only a limited supply of water can be retained by 
such means, and irrigation would generally have to be done 
by means of expensive pumping. Speaking broadly, the 
condition of the majority of our rivers is such that a com- 
bination of both these methods is required to give the best 
results. A most important point to be attended to in 
designing storages, is to secure the greatest possible depth. 
If a site includes large areas, which can be only lightly 
submerged, it is advisable to bank them off from the reser- 
voir, otherwise they will form mere evaporating pans. 
The subject of evaporation is of importance in a country 
where the maximum rate seems to be as high as in any 
part ofthe world. Experiments with small iron tanks have 
been carried on for many years, and a fair assumption from 
the published records would be that the maximum rate 
might be about 6 feet per annum. Unfortunately, there 
is no connection between the evaporation from a small 
experimental tank and the evaporation from a large sheet 
of water exposed to the accumulative action of a Sweeping 
hot wind. An opportunity lately occurred to obtain the 
actual rate from a large sheet of water at Cobar under 
natural conditions. The observations extended over 3 
years, and the result shewed that between 10 and 11 feet 
of water was lost from the reservoir every year. 
The importance of depth in a reservoir was also forcibly 
brought out in the following case:—It lately became 
necessary to select a site in a dry portion of the State 
where the evaporation had been assumed to be 6 feet per 
annum. Two locations were available, one with a depth 
of 35 feet, the other with a depth of 70 feet. The first 
