MOSQUITO BIOLOGY 37 



do their work. At New Brunswick the swarm arrived May 12, and ex- 

 tended onward. At Newark the invasion was almost unprecedented and 

 compelled the closing of stores in some cases. This swarm not only 

 reached Paterson, but covered the Garret Mountains, where they were 

 swarming on June 11. It was reported further that it had extended to 

 Bernardsville, well up in the hills, and far enough back to be beyond 

 the reach of ordinary migrations. In the Pines, cantator reached Laha- 

 way May 17, weeks ahead of the 1903 record. 



There is one feature in which cantator differs from the other migra- 

 tory forms, i.e., both sexes fly together for some distance and occasion- 

 ally females with developed ovaries are found far from any point where 

 they can deposit them with any hope of development. The males are 

 seen for a day or two only, but they arrive with the females and are 

 evidently able to stand a flight of several miles at least. In Newark and 

 Elizabeth both sexes may be taken on first arrival in equal numbers ; 

 at New Brunswick the males are numerous enough to "dance" in small 

 swarms just at dusk. From Lahaway no males have been sent and my 

 records are not sufficiently complete to give the limit of male flight. 



While, as has been noted, occasional females with developed ovaries 

 are found at some distance from shore, yet this is a feature only in the 

 early arrivals ; i.e., they occur during the week or two after a flight 

 and disappear, none being found during the later days of the brood's 

 stay. Whether they make their way back to the marshes to oviposit, or 

 whether they lay their eggs in the most likely place to be found, is not 

 definitely known. As the percentage of such females is larger in Newark 

 and Elizabeth than in New Brunswick, it is quite probable that many 

 find their way back, and as an occasional brood of the larvae is found 

 above the marsh line, it is possible that some may lay eggs which in 

 most instances fail to develop. 



In New Jersey cantator dominates the Newark, Elizabeth, and Rari- 

 tan meadows early in the season. If the year is favorable, the early start 

 will carry this dominance to midsummer or even through the summer. 

 At the Barnegat shore cantator shares with sollicitans the early honors, 

 but becomes steadily less as the season advances, leaving sollicitans in 

 almost sole possession. At Atlantic City and Cape May I found no 

 cantator during the period when they were swarming farther north, 

 though the species does occur there in small numbers throughout the 

 summer. Records show that they must have bred on the Mullica River 

 marshes in greater number than sollicitans during the present year ; but 

 that is probably the southern limit of their dominance, and in ordinary 

 seasons it does not extend so far. Just why they should be more plentiful 



