Systematics of Western North American Dioryctria (Lepidoptera: Pyralidae)

Species of Dioryctria Zeller (Lepidoptera: Pyralidae) are notoriously difficult to identify due to variable forewing morphology, few species specific genitalic features, and broad sympatric distributions. Accurate species delimitation and identification are essential for effective monitoring and control of this conifer pest. I explored several aspects of species delimitation in Dioryctria using a combination of molecular, morphological, ecological, and behavioural characters. I compared mtDNA variation, adult morphology, larval host association and pheromone attraction in a complex of Dioryctria species and identified eight distinct mitochondrial DNA (mtDNA) lineages that corresponded to eight Dioryctria species. Morphology, pheromone attraction, and larval host association also delimited these species and provided useful diagnostic characters. To evaluate how patterns of molecular evolution affect species delimitations I examined mitochondrial cytochrome c oxidase I and II (COI-COII) nucleotide substitution patterns within and between sister species of Lepidoptera and Diptera. I found heterogenous accumulation of maximum divergence and phylogenetic signal, overlap between intra- and interspecific divergence, and no optimally informative 600 by location (length chosen to assess the information content of the DNA barcoding). I used a combination of independent molecular loci, forewing morphology, geography, and larval host association to test the traditional delimitation of two sympatric Dioryctria species, D. pseudotsugella and D. reniculelloides. I found restricted gene flow between an eastern D. reniculelloides clade and a western D. pseudotsugella clade. Two independent molecular loci, morphological variation, and larval host association supported these two clades. However, diagnosis of these two species remains difficult variation and future examinations are needed to elucidate which barriers to gene flow are maintaining this sympatric distribution. To improve species delimitations I recommend that researchers (1) combine multiple lines of evidence to validate, and cross validate species boundaries, (2) broadly sample biological and geographic variation within and between closely related species, and (3) maximize mtDNA sequence length to increase the probability of sampling regions of high divergence, minimize stochastic variance in estimating total COI-COII divergence, and incorporate regions of informative phylogenetic signal.