Relationships of hydraulic traits and wood morphology in Douglas-fir branches from wet vs. dry regions and sites

It is well known that water transport in xylem depends on its anatomy, but recent research shows that water transport through different parts of the growth ring are affected by the water potential in different ways. The goal of the project was to learn the extent to which wood morphology is related to several hydraulic properties, making use of samples and an existing dataset on hydraulic properties of 2-year old branches from 11-year old trees of Douglas-fir (Pseudotsuga menziesii subsp. menziessi Mirb. (Franco)) from several provenances that were grown in a site with low or high annual precipitation. The dataset provided the specific conductivity (K[subscript]s) of samples that were well hydrated, the percent loss of conductivity of those samples after being subjected to -4 MPa (PLC@4), the percent of water mass that was lost between hydration and -4 MPa, the density of the sample, and the sample itself that had been stained after the -4 MPa treatment. The pattern of stain showed the flow paths that were still operative at -4 MPa. There were 36 samples each from an Oregon Coast Range provenance and a low elevation Siskiyou provenance. Half of the samples from each provenance were grown in a Coast Range site, and half near Medford (for 18 trees in each of four treatment combinations). The specific hypotheses were the following.\ud 1) PLC@4 is positively correlated with Ks and with the percentage of water lost between 0 MPa and -4 MPa, and is negatively correlated with wood density.\ud 2) PLC@4 is positively correlated with the percentage of the growth ring that is from certain locations (the first-formed earlywood of the inner growth ring, etc.).\ud 3) K[subscript]s is positively correlated with the percentage of the growth ring that is from certain locations (the earlywood of both rings, and the outer ring).\ud Using image analysis, I determined the percentage of the growth ring in each of four locations, and then with the help of a statistician, analyzed the data to test the hypotheses. The data supported hypothesis 1 and 3 but not 2. The most interesting result was that the percentage of the cross section in a particular growth ring (e.g. earlywood, latewood), was not significantly correlated to drought vulnerability. This result suggests that the anatomy of the individual samples may be better predictors of drought vulnerability than the proportion of cross-sectional area alone