Abstract
Diadromy is a life history strategy involving regular migrations between freshwater and the ocean. These migrations can maximise fitness across life stages, while also reducing isolation, evolutionary divergence, and speciation rates by maintaining genetic connectivity. However, loss of the marine stage has occurred in many diadromous fish species, causing them to become isolated permanently or semi-permanently in freshwaters. Fish that become landlocked may undergo physiological, morphological and behavioural changes since they spend at least some parts of their life cycle in a different habitat with differing salinities, diets, parasites, and predators. However, long-term isolation can also enhance genetic diversity among populations and formation of new species. All freshwater fish species in New Zealand are either diadromous or have evolved from a diadromous ancestor. Common smelt, <i>Retropinna retropinna<i/> and Stokell’s smelt, <i>Stokellia anisodon<i/> are two diadromous fish species native to New Zealand. Common smelt is a widespread migratory fish species in New Zealand that can also form landlocked populations, while Stokell’s smelt is limited to several rivers in the Canterbury region. Although landlocking is widespread on Chatham Island, these populations have not been studied in depth, and there is some ongoing taxonomic uncertainty about smelts in New Zealand. In this study, I aimed to understand the phylogeographic structure of smelts throughout New Zealand and their phylogenetic relationship with other retropinnids, the evolutionary history of landlocking on Chatham Island, and the effects of landlocking on morphological and genetic differentiation of common smelt. As common smelt is a highly dispersed low gradient fish, I hypothesized that this fish would not show a strong phylogeographic structure across New Zealand compared to other native fishes. I also hypothesized that landlocked common smelt populations would display a higher degree of genetic structuring and lower levels of genetic diversity than their migratory counterparts because of the expected restriction or elimination of gene flow among landlocked smelt populations. I predicted that landlocking would pre-date human arrival on Chatham Island, indicating natural colonisation rather anthropogenic translocation into lakes. Lastly, I hypothesized that landlocking would result in morphological traits differences between landlocked and diadromous populations and that morphological changes would align with ecological interactions and functional morphology. Consistent with this hypothesis, a comparatively weak phylogeographic structure was observed in New Zealand, though a strong structure was found for Chatham Island. The phylogenetic relationship confirmed that Stokell’s smelt are more closely related to New Zealand <i>Retropinna<i/> than either is to Australian <i>Retropinna<i/>. Interestingly, I discovered an enigmatic group of smelt in a distinct mitochondrial clade, whose status should be resolved using nuclear markers and morphological data. I found significant morphological and genetic variation between the two life history forms of common
smelt. Landlocked populations showed higher genetic structuring and lower within-population genetic diversity than migratory populations. My analysis also suggested that at least three independent landlocking events occurred on Chatham Island during the Pleistocene period. Additionally, I found that landlocked populations consistently had longer and deeper head size, bigger eyes, wider mouth gape, and deeper body depth and caudal fin base than diadromous populations. Overall, this study provides insights into the evolutionary history of smelt and the morphological and genetic effects of landlocking, with implications for the management and conservation of native diadromous fish species in New Zealand.