Dear Colleagues, This list has proven to be a productive (often excessively) source of statistical advice. I submit the following quote from Kobe et al discussing the lack of independence of regressing green leaf nutrient content against senescent leaf nutrient content. The authors do a good job of circumventing this problem but elude to another related problem of ratio analysis;
"Thus, this modeling approach characterizes resorption efficiency as a mathematical relationship between nutrient concentrations in green and senesced leaves and avoids limitations of statistically analyzing ratios (e.g., not accounting for separate variances of the numerator and denominator, ratio responses driven by either or both the numerator and denominator, potential lack of independence)." With the rise of ecological stoichiometery and nutrient ratios I'm wondering how to get around this problem of separate variances in numerators and denominators. How does this affect the statistical analysis of N:P ratios of different biomass pools? Thanks, David M Bryant Ph D [EMAIL PROTECTED] 978-356-1928
