Hybridisation between red deer (Cervus elaphus) and Japanese sika (C. nippon) on the Kintyre Peninsula, Scotland

Abstract

Hybridisation between introduced and endemic species causes conservation concerns, but also provides us with an opportunity to study the dynamics of gene flow between two species as they first meet. Japanese sika deer (Cervus nippon) were introduced to the British Isles at a number of locations at the beginning of the 20th century. In the intervening time, sika have spread and their range now extends across approximately 40% of Scotland, where they overlap with that of native red deer (C. elaphus), with which they hybridise. In this study we focus on the consequences of one particular introduction that took place at Carradale, on the Kintyre Peninsula in 1893. First, I assessed the current state of hybridisation using a sample of 735 red and sika deer samples collected in 2006/7 from forestry blocks throughout the Kintyre Peninsula. Genetic analysis was conducted with a panel of 22 highly differentiated microsatellite loci and one mtDNA marker. Population admixture analysis of the microsatellite data was conducted with the Bayesian clustering programme STRUCTURE. Over most of the study area, levels of introgression into red and sika deer were low and were consistent with a scenario of very occasional F1 hybridisation followed by backcrossing. There was, however, one forestry block where 43% of individuals could be defined as hybrids. Second, I developed a branching process model of introgression via backcrossing, to assess whether variation in introgression across microsatellite loci could be interpreted as a signature of selection, or could in fact be attributed to stochastic processes. If only a few hybridisation events have contributed to the hybridising population, the pattern of introgression, even with a large number of genetic markers, will be highly stochastic. This pattern of neutral variation in introgression can have high enough variance that it could be mistaken for selection. Therefore, even if strong selection is acting, it may not be possible to distinguish its effects from neutral variation. Third, I analysed trends in hybridisation and introgression over 15 years on the peninsula, through analysis of a dataset of 1513 red and sika deer samples at 20 microsatellite and a mtDNA marker. There was little evidence of change in the extent of hybridisation and introgression over time. MtDNA introgression was predominantly from red deer into sika. Recent introgression into sika on the peninsula can be explained by a very small number of F1 hybridisation events (~10) via analysis of the number of alleles that have introgressed from polymorphic red deer into the genetically homogenous sika population (a similar analysis cannot be conducted for introgression into red deer). Finally, I conducted a regression analysis of genetic hybrid scores against phenotypic traits to assess the effect of hybridisation on phenotype. Hybridisation has caused changes in the weight of sika-like deer and red-like females. Hybridisation has caused changes in incisor arcade breadth of both populations and jaw length (a proxy for skeletal size) in sika-like females. However, there is no evidence that hybridisation has caused changes in kidney fat (a measure of condition) or pregnancy rates in either population. In conclusion, even a small number of F1 hybridisation events can lead to extensive introgression and the timing and spatial distribution of these events is likely to have a large impact on the structure of a recently hybridising population - stochastic factors dominate both the distribution of hybrid individuals and the distribution of the genes that introgress following a hybridisation event. In red deer and sika deer, increasing phenotypic similarities of the two populations caused by hybridisation are likely to facilitate further breakdown between the two species. It is possible that breakdown in assortative mating between the two species could occur across their range.