Environmental and genetic variation in phenotype
In this study we set out to explain the nature, origin, and maintenance of phenotypic variation in L. minor in the field. Phenotype varied widely among sites, with mean frond area varying by a factor of two (Fig. 2A), and mean root length by a factor of more than eight (Fig. 2B). This variation was overwhelmingly the result of phenotypic plasticity. Although there were persistent differences in phenotype among sites in the common garden assay, the reduction of variation in frond area by 93% and in root length by 96% (Fig. 3, Table 2) reveals that among site phenotypic variation is almost exclusively environmental. This is consistent with previous work that has shown a large degree of plasticity in these traits, (Vasseur and Aarssen 1992, Cedergreen and Madsen 2002) and the absence of local adaptation (Vámos and van Moorsel 2022). Both among and within sites, the environmental contribution to phenotypic variation was larger for root length than frond area, which is also consistent with previous work reporting root length as L. minor ’s most plastic trait (Vasseur and Aarssen 1992). Phenotypic variation in L. minor in the field is largely explained as a plastic response to the abiotic environment, shifting its phenotype to levels of resource availability. 35% of among site variation in frond area is explained by light availability, with plants producing larger fronds in more heavily shaded environments. The production of larger leaves in low light environments is a standard ecophysiological response in plants (Meziane and Shipley 1999, 2001), that influences fitness through photosynthesis, transpiration and thermoregulation (Anten et al. 1995, Hirose et al. 1997). Similarly, 46% of among site variation in root length is explained by nutrient availability with a dramatic increase for plants growing in sites with low levels of dissolved N and P. This is consistent with previous experimental work that has documented a plastic increase in root length in L. minor in response to nutrient limitation (Cedergreen and Madsen 2002). Although L. minor can uptake inorganic nutrients through both the root and the frond (Landolt 1986, Cedergreen and Madsen 2002), this balance shifts depending on both nutrient availability (Cedergreen and Madsen 2002), and irradiance (Cedergreen and Madsen 2004) with the production of longer roots resulting in an increase in root N uptake and NO3 reduction. Variation in frond area and root length in L. minor can be conceptualised as a simplified root-shoot ratio (Cedergreen and Madsen 2002). A well-studied trait in land plants (Brouwer 1962, Poorter and Nagel 2000), L. minorseems to respond to resource limitation by investing more biomass into increasing the surface area of the tissue responsible for the uptake of the limiting resource.
In addition to among-site phenotypic variation, we observed significant phenotypic variation within sites. Whereas frond area varied substantially both among sites and among microsites within sites, the majority of variation in root length was at the among site level. Given the largely environmental origin of this variation, it is perhaps uprising that frond area would vary within sites due to the patch-like variation in light availability caused by fine-scale shading from macrophytes and riparian plants (Bell et al. 1991). In contrast, water nutrient availability is likely much more homogenous within sites due to mixing and diffusion resulting in most variation in root length manifesting among and not within sites. For both frond area and root length, the proportion of phenotypic variation with a genetic origin was much higher within sites (26% and 21%) than among sites (7% and 4%). The larger contribution of environmental variation to phenotype among sites can be explained by the greater environmental variation at the higher geographical resolution. However, we observed a surprisingly large amount of within site genetic variation. Environmental variation aside, the absolute amount of genetic variation in frond area was twice as large within sites than among sites, and equal within and among sites for root length. Whereas among site genetic variation is easily explained by adaptation to local conditions or genetic drift given limited gene flow, the large amount of within site genetic diversity is surprising, especially in the absence of sexual reproduction.
In the common garden assay, the contribution of replicate flask to overall phenotypic variation was significant and second only to residual variation. This is perhaps surprising since replicate flasks consisted of clones, descending from the same ancestor sampled from the field. However, replicate flasks confounded several sources of variation including flasks effect, chamber effect (from the blocked design), and birth order effects from the original parental frond which have been shown to persist over several generations (Barks and Laird 2015, 2016, Mejbel and Simons 2018). Removing this variation from the residuals enabled us to detect the higher-level effects of microsite and site.