Habitat loss and fragmentation is one of the main factors causing species decline. Theoretical studies have proposed that a critical threshold occurs below which a small change in the amount of habitat can cause an abrupt change in population persistence. Empirical support for the existence of threshold responses is, however, sparse and studies identifying the ecological processes driving such thresholds are even more limited. Changes in the size and isolation of habitat patches can alter the availability of resources, predation risk and competition between individuals. This can subsequently cause changes in the behaviour and spatial structure of the population, and ultimately survival.  These processes may drive potential thresholds.
Intensification of agriculture has resulted in an increase in cultivated areas and the reduction and fragmentation of natural habitat. Within these managed landscapes, arable fields are, for most of the year, largely inhabitable for many species, and the grassy strips between them (field margins) are commonly the only remaining areas of semi-natural habitat. Field voles (Microtus agrestis) are habitat specialists inhabiting grassy areas, and within intensively farmed landscapes are found primarily in field margins. This provided an ideal system to test the threshold concept and determine the individual and population processes which may be driving such thresholds.  Field voles were trapped in fourteen field margins of varying width over two years and their demographic parameters were measured.
Results/Conclusions

I found strong support for a threshold in margin width, below which vole abundance was extremely low. Narrow margins below the threshold were characterised by greater average distance moved and degree of social exclusiveness between breeding individuals.  A higher percentage of male transients (individuals passing through the habitat) than females, along with the male bias in the narrow margins, suggest that narrow margins are used primarily as a movement corridor by a high proportion of nomadic individuals. However, despite these behavioural responses and a higher level of predator activity in wide margins, no effect was observed on the age structure or apparent survival of the population. I suggest that individuals were able to compensate for the lack of habitat through these alterations in their behaviour sufficiently to maintain their survival. Although the landscape was highly fragmented and areas of continuous habitat were sparse, the network of linking margins may have prevented individuals having to move across barren areas where predation risk was potentially high.