Local population success in heterogeneous habitats: Reciprocal transplant experiments completed on a desert spider

Abstract

Reciprocal transplant experiments were completed to test for selection against the mixing of behavioural phenotypes in a desert spider. Most Agelenopsis aperta populations experience low prey abundances and competition for websites that provide shelter from thermal extremes. These conditions favour aggressiveness towards both prey and conspecifics (an 'arid-land behavioural phenotype'). The spider also occupies narrow stretches of riparian habitat bordering spring-fed streams and rivers. Here it is released from competition for prey and foraging sites, but is subject to predation by birds. A less aggressive/more fearful behaviour is selected for in these riparian habitats (a 'riparian behavioural phenotype'). Previous work with this spider indicates that there is genetic differentiation between arid-land and riparian populations. However, the degree to which genetic differentiation is achieved may be limited by gene flow. Reciprocal sets of enclosures were established in: (1) a dry evergreen woodland site (arid-land phenotype) and (2) a neighbouring riparian site (riparian phenotype) in south-eastern Arizona. Equal numbers of field collected, early instar A. aperta were introduced into native and transplant enclosures in each habitat. After 6 months of site-imposed selection, survivorship was determined and growth estimates and behavioural trials completed on spiders remaining in the different enclosures. The same behavioural test was subsequently applied to lab-reared offspring of the spiders surviving the respective selection regimes. Riparian transplants showed both poor survival and retarded growth in the dry woodland habitat when compared with both arid-land and riparian natives. Arid-land transplants that survived, however, grew equally well in riparian habitat as did dry woodland and riparian natives. Behavioural assays conducted on test subjects after selection and on their offspring reared in a controlled laboratory environment indicate that phenotypes that were inappropriate to the respective habitats were selected against in the transplant experiments. The frequency distribution of transplant spider behaviour on a continuum from fearful to aggressive was intermediate between that exhibited by respective native riparian and dry woodland spiders. It is concluded that while arid-land and riparian behavioural ecotypes do exist, directional gene flow of arid-land phenotypes into riparian habitat limits population subdivision.