Background/Question/Methods   Extreme habitats such as salt flats, alvars, hydrothermal vents and serpentinite soils are areas of intense environmental filtering often leading to a unique set of species adaptations suited to the environmental resources at hand. In woodlands and forests of the mid-Atlantic piedmont there are geologic juxtapositions resulting in nutrient poor and shallow ultramafic or serpentinite soils adjacent to well developed and nutrient rich mafic soils. Tree species on these resource bases are likely to employ different mechanisms to cope with the droughty, nutrient poor edaphic soils compared with the more normal mafic soils. Such mechanisms might include early stomatal closure and different allocation to root versus shoot strategies. To understand the geoecologic processes controlling local woodland tree species community assembly, a greenhouse experiment was used to explore resource gradients and provenance as variables effecting typical representative upland oak (Quercus) species based on their carbon allocation strategies and stomatal control. Accordingly, we carried out a year long reciprocal transplant experiment using seedlings of five oak species (Quercus marilandica, Q. stellata, Q. prinus, Q. michauxii and Q. alba.). Acorns were hand collected from Maryland ultramafic sites and mafic sites. There were two soil types (ultramafic with Ca:Mg = 0.2 and mafic with Ca:Mg = 5.3) and three soil depths: 15 cm, 30 cm and 60 cm with four replicates for each of the oak species amounting to 240 growth tubes.

Results/Conclusions   Overall mortality was low (3.75%) and occurred only on serpentinite soils. Local adaptation appears muted except for Q. marilandica and Q alba, which had greater growth in their respective native soils. Seedling growth rates were higher on mafic rather than ultramafic soils in all treatments. In the xeric species of Quercus marilandica and Q. stellata, at least 70% of seedling biomass was allocated belowground in serpentinite soils with lower percentages on mafic soils. Mean stomatal conductance was markedly similar in all treatments during the initial growing season (May) varying between 60 and 105 mmol/(m²·s). In the later part of the season (August) the differences between species became pronounced with Q. marilandica having the highest stomatal conductance and Q. michauxii the lowest. Mean stomatal conductance and stem growth were positively correlated with soil depth in all treatments. These results suggest that the edaphic factor is a strong environmental constraint forcing unique physiological differences among these seedlings. Such morphological and physiological responses allow for a better understanding of the mechanisms controlling plant soil-water-stress relations.