Identifying the genes underlying phenotypic variation in natural populations can provide novel insight into the evolutionary process. The candidate gene approach has been applied to studies of a number of traits in various species, in an attempt to elucidate their genetic basis. Here, we test the application of the candidate gene approach to identify the loci involved in variation in gastrointestinal parasite burden, a complex trait likely to be controlled by many loci, in a wild population of Soay sheep. A comprehensive literature review, Gene Ontology databases, and comparative genomics resources between cattle and sheep were used to generate a list of candidate genes. In a pilot study, these candidates, along with 50 random genes, were then sequenced in two pools of Soay sheep; one with low gastrointestinal nematode burden and the other high, using a NimbleGen sequence capture experiment. Further candidates were identified from single nucleotide polymorphisms (SNPs) that were highly differentiated between high- and low-resistance sheep breeds. A panel of 192 candidate and control SNPs were then typed in 960 individual Soay sheep to examine whether they individually explained variation in parasite burden, as measured as faecal egg count, as well as two immune measures (Teladorsagia circumcincta-specific antibodies and antinuclear antibodies). The cumulative effect of the candidate and control SNPs were estimated by fitting genetic relationship matrices (GRMs) as random effects in animal models of the three traits. No more significant SNPs were identified in the pilot sequencing experiment and association study than expected by chance. Furthermore, no significant difference was found between the proportions of candidate or control SNPs that were found to be significantly associated with parasite burden/immune measures. No significant effect of the candidate or control gene GRMs was found. There is thus little support for the candidate gene approach to the identification of loci explaining variation in parasitological and immunological traits in this population. However, a number of SNPs explained significant variation in multiple traits and significant correlations were found between the proportions of variance explained by individual SNPs across multiple traits. The significant SNPs identified in this study may still, therefore, merit further investigation.