406 MOSQUITOES OF NORTH AMERICA 
“While, as has been stated, all fishes have some measure of usefulness, if 
only in the way of deterrent effect, there are only a few species likely to be found 
in waters in which mosquitoes breed. The most important of these are the gold 
fish (introduced), several species of Fundulus (the killifishes), and allied 
genera, three or four species of sunfish, and the roach or shiner and perhaps one 
or two other small cyprinoids. In addition there are a few sluggish and soli- 
tary species like the mud-minnow (Umbra) and the pirate perch (Aphredo- 
derus). The sticklebacks have been mentioned in this connection, but the 
Atlantic Coast species, and probably the entire family, are undoubtedly useless 
for the purpose, being bottom feeders, living in the shallow tide pools and gutters, 
hidden among plants, or under logs and sticks at the bottom, where they find an 
abundance of other food. 
“Tn the salt marshes there are myriads of killifishes running in and out and 
over them with each tide, while countless numbers of other and smaller genera, 
such as Cyprinodon and Lucania remain here at all stages of the tide. So 
numerous and active are all of these, that there is no possibility of the develop- 
ment of a mosquito where they have access. 
“ Of the killifishes two species, Heteroditus and Diaphanous, ascend to the 
furthest reaches of tide flow, but it is a question as to whether they would 
prove desirable for the purpose of stocking land-locked waters, since they are a 
good deal like the English sparrow, aggressive toward the more peaceable and 
desirable kinds. Even Cyprinodon, which would at first thought be a valuable 
small species in this respect, is viciously aggressive toward goldfish and no doubt 
all other cyprinoids. It is so characteristic of all the cyprinodonts, that they can 
only be kept by themselves in aquaria. They are the wolves or jackals of the 
smaller species. 
“The writer has come to the conclusion, after many experiments in both 
tanks and ponds, that a combination of the goldfish, roach, and top-minnow 
would probably prove to be more generally effective in preventing mosquito 
breeding than any other. The goldfish is somewhat lethargic in habit, and is 
also omnivorous, but there is no doubt that it will devour any mosquito larve 
that may come in its way, or that may attract its attention. The one great 
objection is that they grow too large, and the larger will eat the smaller of them. 
That is one of the drawbacks to goldfish breeding. There is no danger of over- 
population, but there is of the reverse. Whether or not it is the same with the 
roach, they are never excessively numerous, although no doubt the most abundant 
and most widely distributed of the Cyprinide. They are largely the prey of 
predaceous fishes, and never approach to the numbers of the killifishes. But at 
all events they are not lethargic like the goldfish, being on the contrary one of 
the most active of the family, and equally at home in flowing or stagnant water. 
The roach is always in motion, back and forth, and around and about, on a never- 
ending patrol. 
“ The top-minnows would supply the deficiences of the other two species, and 
in combination they should very thoroughly populate any waters not already 
stocked with predaceous kinds, and exercise an effective control. One of the 
great difficulties in the case is that there are dozens of kinds of insect larve 
besides those of the mosquito, and other forms of life as well, which are natural 
and possibly preferred food of the fishes, thus requiring an enormous population 
to devour them all. 
“The larvee of gnats, midges, ephemera, and other flies and insects which 
breed in the water, as well as the many small crustceans, afford a menu of 
delicacies that would stagger a gourmand. The above combination of mosquito 
destroyers might be supplemented by two small species of sunfish, Enneacanthus 
obesus and FH. gloriosus, which live among plants and would be a check on larvee 
