Mitochondrial paraphyly between arthropod species is not uncommon, and has been speculated to largely be the result of incomplete lineage sorting (ILS) of ancestral variation within the common ancestor of both species, with hybridisation playing only a minor role. However, in the absence of comparable nuclear genetic data, the relative roles of ILS and hybridisation in explaining mitochondrial DNA (mtDNA) paraphyly remain unclear. Hybridisation itself is a multifaceted gateway to paraphyly, which may lead to paraphyly across both the nuclear and mitochondrial genomes, or paraphyly that is largely restricted to the mitochondrial genome. These different outcomes will depend upon the frequency of hybridisation, its demographic context, and the extent to which mtDNA is subject to direct selection, indirect selection, or neutral processes. Here we describe extensive mtDNA paraphyly between two species of iron-clad beetle (Zopheridae) and evaluate competing explanations for its origin. We first test between hypotheses of ILS and hybridisation, revealing strong nuclear genetic differentiation between species, but with the complete replacement of Tarphius simplex mtDNA through the introgression of at least five mtDNA haplotypes from T. canariensis. We then contrast explanations of direct selection, indirect selection, or genetic drift for observed patterns of mtDNA introgression. Our results highlight how introgression can lead to complex patterns of mtDNA paraphyly across arthropod species, while simultaneously revealing the challenges for understanding the selective or neutral drivers that underpin such patterns.