4.2 Impacts of provenance but mixed evidence for local adaptation
In comparing the growth and survival of upland and lowland provenances between the DRO and Thiaki we had hoped to isolate the impacts of temperature - and given the proximity of maternal populations to planting sites (Figure S1) the prevalence of local adaptation in our study species. Although we found that the lowland provenance often outperformed the upland provenance at the lowland site, we rarely found the reciprocal to be true i.e. upland provenances did not outperform lowland provenances in the upland site, indicating a partial lack of local adaptation sensu Kawecki and Ebert (2004). Despite this study only including three sites and thus having relatively low power to detect local adaptation (Blanquart et al . 2013), many examples supporting our results challenge the notion that this is widespread (Leimu & Fischer 2008, Bucharova et al. 2016, Mushagalusaet al. 2020). Our findings resemble those observed in the reciprocal transplant of Quercus oleoides populations conducted by Deacon & Cavender-Bares (2015). In their study, the lowland provenance also exhibited higher mean fitness at all sites compared to the upland provenance. Given that faster growth rates are typically observed in lowland environments (Bradford et al. 2014, Baumanet al. 2022), it is possible that lowland populations of rainforest trees generally have evolved higher growth rates than upland populations, potentially at the cost of stress tolerance or other fitness metrics (Willi & Van Buskirk 2022). Another possibility is that the lowland provenance seedlings in our study were sourced from more intact populations with higher reproductive population size and therefore genetic diversity, which has been linked to higher fitness (Reed & Frankham 2003). Confirming this requires further work assessing within-species genetic diversity and differentiation across populations. However, it is possible that upland provenances may possess greater stress tolerance, and a full life-cycle assessment is needed to determine any true differences in population fitness. Furthermore, in Deacon & Cavender-Bares (2015) study, the advantage of the lowland provenance was only temporary with provenance differences no longer evident by the third year, highlighting the importance of continued monitoring of such studies when assessing patterns of local adaptation (Preece et al. 2023).