For small, flightless terrestrial organisms, the world is littered with barriers to dispersal. Deserts, oceans and mountain ranges present apparently insuperable limits to migration and spread for animals that are only a few millimetres in size and cannot fly. But from work by S. J. Coulson and colleagues1, published in Functional Ecology, it seems that some soil arthropods can survive long enough in sea water to permit them to travel to distant lands. Assumptions about the causes of biogeographical distribution patterns among such organisms now need to be reconsidered.

Disjunct distributions are relatively common in soil arthropods, as in the case of the springtail Tetracanthella arctica (order Collembola), a dark blue, flightless insect, just 1.5 mm long, that lives mainly among mosses on the soil surface and feeds on plant detritus and fungi. It is found in isolated locations through the Arctic, ranging from Spitzbergen, Iceland and Greenland to northern Canada, but also occurs in alpine habitats in the Pyrenees, the Tatra Mountains and the Carpathians2. Its isolation in the European mountain sites is best explained by the survival of relict populations after the end of the last glaciation: a presumed extensive distribution in glacial times would have contracted and become fragmented as warmer conditions spread and forest became abundant. But the scattered distribution of this springtail in the Arctic regions could be explained in several ways, including carriage on ice floes, or by larger, more mobile animals (particularly birds), or on drifting chunks of sea-borne turf or driftwood.

On 14 November 1963, an undersea volcanic eruption off the south coast of Iceland led to the formation of a new island, Surtsey, which provided a unique opportunity for observational studies in transoceanic colonization by terrestrial organisms. In 1975, Sturla Fridriksson3 published a summary of this work and recorded that six species of Collembola had reached Surtsey within 10 years of its formation. Tetracanthella arctica was not among them, despite being present on some of the Westman Islands only 15 km away. Such small soil organisms as the springtails, however, are evidently capable of quite rapid colonization over distances of several kilometres.

Coulson and his colleagues1 adopted an experimental approach to the question of dispersal potential. They subjected five species of Collembola and one mite to a variety of seawater conditions to test whether these terrestrial organisms can survive in the sea, even in the absence of driftwood or other materials. Many springtails have hydrophobic, unwettable cuticles that enable them to survive on water surfaces. Subjecting such species to 16 days on the surface of agitated sea water showed that they could indeed survive this long. Three out of the five species tested (including T. arctica) displayed over 80% survival. In a separate experiment, T. arctica and the mite Camisia anomia were submerged by pushing them below the surface with a paintbrush. They continued to be active, and 75% of the mites survived over a 15-day period. Only 12% of the springtails did so. Yet prolonged survival of even a small proportion of these organisms — which reproduce parthenogenetically, without the need for males — could permit their wide dispersal and subsequent establishment.

The implications for studies of springtail dispersal are considerable. Coulson and Birkemoe4 previously demonstrated that some species can survive at a temperature of −22 °C for over four years, so the incorporation of sea-transported springtails into pack ice is entirely possible and could lead to their far-flung travel in the Arctic. Even without the aid of ice, the movement of floating or sunken springtails between land masses such as Norway and Svalbard is quite possible if they can survive for over two weeks in the open ocean.

Distribution patterns of soil invertebrates are of biogeographical interest in a variety of situations. Work in the tropics, for example, has indicated5 that termite biogeography and biodiversity might be a reliable indicator of tropical rainforest persistence through the Quaternary, the past 1.6 million years or so. Such hypotheses depend upon the concept of relict, non-mobile populations of an organism, which in turn is underpinned by the assumption of poor dispersal capacity. For termites, the fact that only specialist individuals in a population can start new colonies further limits the mobility of these species, but the revelation of unsuspected methods of travel in a seemingly static animal should always stimulate reflection.