SPECIES OF ANOPHELES—PREFERENTIAL BREEDING PLACES. 
From larvae secured in the pond A. quadrimaculatus predomi¬ 
nated—349 A. quadrimaculatus to 42 A. punctipermis is the record 
for the first visit to Parr Shoals. Stevens Creek gave about the 
same proportion, A. punctipermis being rather more numerous. A 
large amount of breeding was found in running streams and in 
marshes and pools adjacent to them, but these were not searched as 
widely as we did later at the Coosa Pool. The record from running 
streams and from marshes and pools adjacent to them is: A. quadri¬ 
maculatus , 2; A. punctipermis , 280. It is probable enough that these 
two A. quadrimaculatus were accidental contaminations. At Stevens 
Creek the branches had just been washed out and were not examined. 
The few developing from this catch— “eight only — were A. punctip- 
ennis. No A. crucians were seen at either of these two places, 
although they were abundant around Hartsville, a town about 40 
miles distant. 
A thing that impressed us here was that A. quadrimaculatus not 
only was found in the pond but larvae of this species was not found 
in the neighborhood of the pond in places where one would naturally 
expect to find them. It was a fairly flat country and showed a num¬ 
ber of small marshes, of little pools of clear water in grass, ditches, 
potholes, etc., in which one expects to find this species and in which 
we did find them at Eau Clair, about 20 miles distant. The same 
thing occurred at the examination of the Coosa pool. Although 
there were no check examinations there in which we found A. quadri¬ 
maculatus as at Eau Clair, yet finding this species in considerable 
number in houses at Shelby —5 or 6 miles from the pond—and in 
other houses in the country If to 2\ miles distant indicates that they 
do breed in this section. 
Though not extensive, the examination of Pool 17 on the Black 
Warrior, so far as it went, evidenced the same thing. It looks as 
if the pond were a preferential breeding place for A. quadrimacu- 
latus , so that when a pond is available this species may breed nowhere 
else; that places which produced this form before the pond was made 
do so no longer. If this be true, to determine the influence of the 
pond on the production in any district of A. quadrima'culatus , we 
must deduct from the production of the pond the number formerly 
produced by former breeding places outside of the pond and now no 
longer producing, as well as those covered by the pond. 
While this hypothesis will explain the facts observed, and hence 
is an allowable deduction from them, it is not a necessary one. To 
determine this question, it is proposed that the sites to be occupied 
by a large pond and its environment be surveyed before the pond is 
made as well as afterwards. If the site selected be in a district in 
