(2 
because of the chemical composition of the soil, and partly because it is so 
thin. The only method for the prevention of erosion of the sides of the val- 
ley is forestration. 
The form of the valley gives a deep narrow impounding basin, which is 
well protected from wind. This results in the complete thermal stratifica- 
tion of the water early in the summer. The leaves from the wooded slopes 
of the valley are blown into the water producing a very considerable 
amount of oxydizable material. The result is that during the latter part 
of the summer the oxygen is exhausted from the lower levels of the water. 
This oxygenless region increases in thickness until the autumn “turnover” 
and reaches the level of the intake early in August. As a result of this 
the water is very unpalatable during most of August and a part of Septem- 
ber. Of course a better flavor could be obtained by placing the intake above 
the thermocline, but the water is carried directly to the taps without icing 
and hence it is a marked advantage to take it from the cooler water below 
the thermocline. ; 
The apparatus consists of an ordinary compression pump such as is sold 
by the trade for use in garages. To this is coupled a ®, inch air line which 
is carried over the dam and down to a point near the bottom of the reser- 
voir, and just below the intake. This point is 9.75 M. below the top of the 
dam. At this end was attached a crow-foot formed of three 84 inch pipes 
7 feet long, and jointed at right angles to the line and to each other. The 
ends of these pipes are capped and small holes (144 inch) are drilled on the 
under side, These holes are 1 inch apart, placed alternately in two rows 
(Fig. 1). The pump is operated by a belt connecting it to the flywheel of 
one of the oil engines which operates the water pump. The pump usually 
runs eight hours in each twenty-four. 
The pressure at the air outlet is slightly more than one additional atmos- 
phere. The temperature was about 17 ©., while at the surface it varied 
from 22 C. to 26 C. It was 22 C. on September- 25, 1919. The increased 
pressure and the lower temperature would of course make the rate of gas 
absorbtion more rapid per unit of surface than it would be at the surface 
of the reservoir. By discharging the air through the small openings, small 
bubbles were produced which increased the surface per unit of air volume. 
It was hoped that this operation would introduce enough oxygen to main- 
tain a potable quality of water below the thermocline. The quality of the 
water was improved markedly, but the improvement was due, in part at 
least, to another set of factors which had net been considered when the 
apparatus was put into operation. These were the convection currents 
caused by the rising bubbles. The friction of these bubbles set up a yertical 
current that was approximately 100 sq. ft. in cross section. The rate of 
flow of this upward current was not measured but it was sufficient to raise 
the level of the surface some 2 or 3 em. 
The amount of dissolved oxygen was determined at various levels at the 
intake and at the deepest point in a cross section 150 feet above the dain. 
In the upper 8 M. at the upper station (the water was 1.75 M. below the 
top of the dam) there was more oxygen at every level than at the intake. 
