Differential success in northwards range expansion between ecotypes of the marble gallwasp Andricus kollari: a tale of two lifecycles.

The Marble gallwasp Andricus kollari has a native range divided into two geographically separated lifecycles. In Eastern Europe and Turkey, the lifecycle involves a sexual generation on Turkey oak, Quercus cerris, while in Iberia and North Africa the sexual generation host is cork oak, Q. suber. Over the last 500 years, A. kollari has expanded its range into northern Europe, following human planting of Q. cerris from Italy and the Balkans. We ask: (i) what is the genetic relationship between eastern and western distributions of Andricus kollari? Can we determine which lifecycle is ancestral, and how long ago they diverged? (ii) To what extent have eastern and western native ranges contributed to northwards range expansion? (iii) Is there any evidence for hybridization between the two life cycle types? We present analyses of allozyme data for 13 polymorphic loci and of sequence variation for a 433 bp fragment of the mitochondrial cytochrome b gene. These show: (i) that four haplotype lineages (one in Spain, two in Hungary/Italy and one in Turkey) diverged more or less simultaneously between 1 and 2 million years ago, suggesting the existence of at least four refuges through recent ice age cycles. Our data cannot resolve which lifecycle type is ancestral. (ii) Populations north of putative refuges are divided into two sets. Populations in south-west France are allied to Spain, while all remaining populations in northern Europe have been colonized from Italy and the Balkans. (iii) The transition from one race to another in south-west France is marked by abrupt transitions in the frequency of refuge-specific private alleles and corresponds closely to the northern limit of the distribution of cork oak. Although hybrids were detected in north-west France, none were detected where the two lifecycles meet in south-western France. The biology of oak gallwasps predicts that any hybrid zone will be narrow, and limited to regions where Q. cerris and Q. suber meet. Our data suggest that eastern and western A. kollari are effectively separate species.