In an earlier attempt to find a reference framework to classify the kinds of migratory flight adopted by different kinds of insects, I used the term 'boundary layer' to describe a hypothetical layer of air near the ground within which it was supposed that insects were able to control their movements relative to the ground because their flight speed exceeded wind speed; outside the boundary layer they would, perforce, move downwind (Taylor 1958). Evidence for such a layer was found by analysing existing data on the profile of insect density near the ground, although the sampling techniques then in use were questionable (Taylor 1960). The boundary layer concept has apparently proved useful (see Southwood 1966; Johnson 1969); it has helped interpretation of experimental work, e.g. on insect density in wind shadows (Lewis 1967) and has also provided a basis for classifying migratory flight of aphids (Taylor & Palmer 1972). Diurnal flight periodicity and body size were suggested as major determinants of the migratory potential of insects (Taylor 1958). For example, it was suggested that migration tends to be minimal in the small Nematocera that are behaviourally adapted to fly within a shallow boundary layer at night when atmospheric lift is minimal and, as a consequence, are restricted to travelling the short distances their own powers of flight can sustain. In contrast, the small day-flyers are typified by those aphids that have developed a migratory response in which initial flight upwards towards the light, at times when maximum upward air transport is available, frees them from the limitations of the boundary layer and enables them to enlarge their migratory ambit by using the wind as a vehicle. Large day-flyers, like butterflies, could migrate directionally within their boundary layers, whilst large nocturnal migrants, like moths, might migrate downwind (Lewis & Taylor 1965). Until now, however, despite its common acceptance, the boundary layer has remained more conceptual than experimental; it is a term of convenience with few of its implications investigated. I have, therefore, measured accurately a series of twenty-four vertical density profiles for different insect taxa and examined them in relation to periodicity of flight and body size, to investigate the more obvious deductions from the hypothesis. No other accurately measured low level density x height profiles in exposed situations exist, except those for Oscinella frit L. (Calnaido, French & Taylor 1965), Noctuidae and Geometridae (Taylor & French 1974), and some three-point profiles for Lepidoptera (Taylor & Carter 1961). Service (1971a, b, 1973) has produced profiles, for three groups of Nematocera, inside woodland.