326 
these reach sufficient height to break the 
water surface, the following spring they 
will continue to grow and will provide the 
protection necessary to anopheline larvae. 
Only submergence for a period exceeding a 
year will kill the majority of willow stumps. 
As a result, special investigation of the con¬ 
trol of willow stumpage has been under¬ 
taken. It appears that hack-girdling 
(frilling) of the cut stump, just above the 
surface of the ground and the application 
of an arboricide, is the most efficient method 
of killing them. It is important that the 
hack-girdling shall completely and con¬ 
tinuously encircle the stump, cutting 
through the bark and cambium layers. The 
hacked portion provides a trough into 
which the arboricide may be applied by 
means of a knapsack sprayer. Thorough 
application of the poison is necessary. The 
original treatment utilized by the Tennessee 
Valley Authority consisted of an aqueous 
solution of sodium arsenite (approximately 
12.5 per cent). Owing to the dangers to 
laborers and also to livestock, a 30° Baume 
diesel oil has been substituted and has been 
found to be fairly satisfactory. 
Marginal drainage is a most important 
adjunct to reservoir preparation. All pools 
located between maximum and minimum 
lake levels must be connected to the main 
body of the lake to prevent isolated breed¬ 
ing pools upon fluctuation of the water 
level. Such work must be completed prior 
to filling the reservoir. 
An important consideration with im¬ 
pounded water projects is the initial winter¬ 
time filling of the reservoir. It has been 
demonstrated repeatedly that if the im- 
poundage is created during the growing 
season in the Southeastern States, a situa¬ 
tion conducive to the propagation of A. 
quadrimaculatus will follow. When water 
rises slowly into terrestrial vegetation, it 
picks up fine debris (flotage) which makes 
a rich medium for larvae. Therefore, a new 
reservoir should be filled sometime after the 
close of one growing season and before the 
initiation of the following one. Preferably 
each project should have provision for a 
flood surcharge. This provides an oppor¬ 
tunity for filling above the normal summer¬ 
time operating level and drawing down to 
this point just at the beginning of the breed¬ 
ing season. This results in the stranding 
of drift and flotage which would otherwise 
offer protection to larvae. 
Water Level Fluctuation 
Following adequate reservoir prepara¬ 
tion, the next most important step in ma¬ 
laria control on impounded waters is the 
proper management of the water level. 
Carter (1914) was apparently the first to 
record the beneficial effect of the variation 
of pool level as an anti-larval measure. His 
first observations were made on power 
plants in which a weekly variation of ele¬ 
vation resulted from the closing down of 
mills over the week-end. Carter, LePrince 
and Griffitts (1916) and Griffitts (1926) 
further studied the employment of this 
measure for anopheline control. Since 
these basic studies by the workers of the 
U. S. Public Health Service, certain power 
lakes in central Alabama, after clearing, 
have successfully utilized fluctuation for 
several years as the sole measure of control 
of anopheline larvae. In such cases a rather 
wide cyclical (weekly) fluctuation has been 
possible. Van Dine (1922) observed that 
the growth of vegetation along the margin 
of an impounded bayou was discouraged by 
a variable water level. The Alabama State 
Board of Health regulations governing the 
impounding of waters (1927) point to the 
advantageous use of a surcharge to strand 
flotage and drift and state that “during 
the mosquito breeding season, pronounced 
fluctuations of the water level may be, or 
may be made to be, a controlling factor in 
breeding. . . . Changes in water level 
tend to prevent rank growth of aquatic 
plants and aid fish enemies ( Gambusia) in 
reaching the mosquito larvae or flush the 
larvae within reach of the fish.” 
Hinman (1938) reported upon the bio¬ 
logical effects of fluctuation of water level 
on anopheline breeding. The continuation 
of these studies forms the basis for the fol¬ 
lowing discussion. Pool level fluctuation 
for malaria control may be employed along 
