Polyphenols in litter from tropical montane forests across a wide range in soil fertility

In nutrient-poor ecosystems high polyphenol concentrations in plant litter have been proposed to influence soil nutrient availability in benefit of the plants. We addressed the question whether litter polyphenol concentrations vary across a soil chronosequence of almost identical geology, climate and plant species composition, but of a wide range in nitrogen (N) and phosphorus ( P) availability in the Hawaiian Islands. Concentrations of total phenolics (TPh) and proanthocyanidins (PA) in leaf litter of the dominant tree species Metrosideros polymorpha were higher at the oldest, P-limited site compared to the youngest, N-limited site, with intermediate values at the two relatively fertile sites co-limited by N and P. Polyphenol concentrations in fine root litter differed considerably from those observed in leaf litter and varied differently across the soil age gradient. Long-term fertilization did not significantly alter polyphenol concentrations in Metrosideros litter at either site. Moreover, green leaves and leaf litter of Metrosideros showed similar relative differences among sites when compared between natural populations and plants from the same populations but grown in a common garden. These results suggest that polyphenol concentrations inherently vary among populations of the dominant tree species in Hawaiian montane forests possibly indicating an adaptation to ecosystem properties such as substrate age related differences in soil fertility. The combined above- and below-ground input rate of TPh ranged from 62.4 to 170.8g/m(2)/yr and was significantly higher at the P-limited than at the N-limited site. Root-derived polyphenols contributed a much higher absolute and relative amount of phenolic input at the N-limited than at the P-limited site. The differences in amount, quality, and pathways of input might suggest specific interactions with soil processes and nutrient cycling among the Hawaiian rainforests studied here.