The influence of a deep river valley system on the distribution of dispersing Aedes mosquitos was studied during 1958–59 in the vicinity of Edmonton, Alberta, Canada. Mosquito breeding within an area extending two to three miles beyond the city limits was prevented by anti-larval measures, and almost the only remaining breeding places were situated on the plains several miles to the north-west and south-east, yet considerable numbers of mosquitos appeared every spring in that part of the city included in the valley-ravine system of the North Saskatchewan River. Previous work had suggested that mosquitos dispersing across the plains were collected by the valley-ravine system and entered the city by passing down the main valley. The present paper attempts to account for this directed dispersal in terms of the behaviour of Aedes mosquitos in response to certain physical factors in the environment. The species mainly concerned were Aedes cataphylla Dyar, A. fitchii (Felt & Young), A. excrucians (Wlk.) and A. stimulans (Wlk.).

The initiation of dispersal flight downwind across the plains at twilight was observed. This took place soon after emergence and was not prevented by moderate winds. It was shown experimentally that neither high wind speeds nor turbulence inhibited the flight of newly emerged Aedes mosquitos, but that individuals several days old were reluctant to take flight in turbulent conditions. Since wind speeds on the plains were generally higher than those in the valleys, these reactions would tend to confine older mosquitos to the valleys while not preventing dispersal of newly emerged mosquitos across the plains.

The direction of the initial dispersal flights from the breeding places was observed in mosquitos escaping from rearing cages. Take-off was characteristically into the wind, but re-orientation downwind took place when the mosquitos began to be blown backwards in flight. It was shown experimentally, by releasing mosquitos in the field when the wind was very light, that the preferred direction of flight uninfluenced by wind was towards the lowest part of the horizon, both on the plains and in the valley. This response would tend to cause mosquitos on the plain to fly into ravines and valleys, and thence down-stream within them. Dispersing mosquitos would thus tend to be held by any valleys or ravines in their path.

Swarming was never observed near the pools from which the mosquitos emerged, but was mainly seen in the valley or on the edge of the valley and ravines. Swarming of A. cataphylla took place during the evenings after the initial dispersal. The swarms were predominantly male but mating occurred during swarming. A cohesive factor seemed to be involved in the behaviour of the swarms, and further dispersal also seemed to occur gregariously; the movement of large male swarms down the river valley was observed.

By the application of optomotor theories of insect flight to previously published data on the flight speed and behaviour of A. aegypti (L.) and A. punctor (Kby.), the relationships between wind speed and permissible heights of flight of these species upwind and downwind were calculated. These relationships were consistent with the observations on dispersal of Aedes mosquitos in the Edmonton area, and could largely explain the mechanism by means of which deep valleys attracted and held dispersing mosquitos.

In contrast to the open plains, the wooded river valley provides a more favourable environment for adult mosquitos. The supply of nectar is greater, saturation deficiency and wind speeds are lower and less variable, and shaded resting sites are more abundant.

Observations on mosquitos in chambers with either vertical or horizontal humidity gradients suggested that the behaviour of Aedes mosquitos is not greatly influenced by humidity.