Environmental distance and latitude did not drive germination patterns
Environmental distance between source population and common garden sites did not explain the observed seed-type and scale-dependent germination patterns. Although regional climate models do show that southern sites tend to be warmer and drier (PRISM Climate Group 2020), it is possible that the environmental covariates we used to estimate environmental distance were not meaningful, masking a potential environmental source effect on germination that can be difficult to detect (Hereford 2009). Other common garden studies have found that germination success is influenced by parental environment (Gallagher & Wagenius 2016). More source populations and environmental data could possibly reveal such a trend if one exists.
We expected to see higher germination of nonlocal seeds sourced from southern compared to northern sites due to recent warming patterns in the region (PRISM Climate Group 2020) but were surprised to find the reverse pattern in our study. A categorical comparison of locally sourced seeds to those sourced to the north and south of the HD common garden revealed that northern seeds performed significantly better than both local and southerly sourced seeds. This pattern was not as easily detected in a latitude-based germination model; as latitude increased, chasmogamous germination also increased slightly, while cleistogamous germination slightly decreased which produced a significant latitude by seed type interaction. It is likely that the highly reduced germination probability of cleistogamous seeds at only our most northerly site created this negative latitudinal trend for cleistogamous seeds. This suggests there may be a northern limit for seed sourcing success at Willamette Valley planting locations.
These results could be linked to our observation that seed weight slightly increased with latitude, although this trend is not significant. It is unlikely that the opposing effects of latitude on chasmogamous and cleistogamous seeds observed in our continuous latitude model are biologically meaningful. Perhaps sourcing from more sites would allow us to examine the impact of latitude on germination more accurately. Indeed, the latitudinal range for D. californicaextends from British Columbia to southern California (Darris & Gonzalves n.d.). Of course, our study only measured the germination probability within one growing season, which makes the detection of climate-driven latitudinal fitness patterns unlikely (Gaillart et al. 2018, DeMarche et al. 2019). Other local adaptation studies have found significant latitudinal responses (van Boheemen et al. 2019, Zhang et al. 2019), although they may not necessarily be tied to climate (DeMarche et al. 2021).