Non-crystalline minerals are formed through weathering processes from volcanic materials and strongly bind phosphorus (P), which may reduce the P availability for forest trees on volcanic soils. On the other hand, volcanic materials are a direct source of P and enhance soil total P. It has not been clarified how P availability and P cycling in forest ecosystems relate to non-crystalline minerals in volcanic soils. The objective of this study is to clarify the differences of soil nutrient availability and foliar nutrients of trees by comparing volcanic vs. non-volcanic soils. We selected eight secondary forests where Quercus serrata was dominant in Japan as study sites. We collected surface soils from three layers (0-5, 5-15, 15-30 cm depth) from each study site and canopy fresh leaves of Q. serrata. We measured the concentration of oxalate-extractable Al and Fe (Al_o and Fe_o), P fractions, and exchangeable-cations in soils and N and P in leaves. Both Al_o and Fe_o and soil nutrients (P fraction and exchangeable cations) were significantly higher in volcanic soils than in non-volcanic soils. Al_o + 1/2Fe_o was significantly correlated to NaOH-P (indicator of P bound to Al and Fe). Both plant available-P and NaOH-P were significantly positively correlated with foliar N and P concentrations. Estimated maximum heights of trees tended to be higher in volcanic soils. Our results indicated that the addition of volcanic ash directly increased soil P availability via increasing the pool of P and flux of P from recalcitrant and organic forms of P to available inorganic forms of P in soils.

We surveyed species composition in relation to substrates, and microsite characteristics of temperate conifer species in an old growth Chamaecyparis obtusa forest on Mt. Shiraga, Kochi, Japan. We established nine study quadrates on a gentle slope where masses of rock or dwarf bamboos are dominant on forest floor, and we measured the diameter at breast height of all trees that were more than 1.5 m in height and the height of all conifer seedlings (< 1.5 m in height) while recording their standing substrates: stump, fallen log, rock, mound, and ground. According to detrended correspondence analysis, the nine quadrates were roughly divided into a group dominated by Chamaecyparis obtusa and Tsuga sieboldii, and a group of mixed broad-leaved tree stands, implying that the former establish on a block slope and the later establish on the ground. Chamaecyparis obtusa seedlings significantly established on fallen logs and old stumps while larger trees (> 1.5 m in height) were found on stumps. Thirty percent of large Chamaecyparis obtuse trees were established at an elevated position above the ground, and half of them had stilt-like roots. Tsuga sieboldii seedlings were established on rocks and stumps, while the larger trees were found on stumps. Both seedlings and larger trees of Pinus parviflora established on stumps though there were few of these. Abies firma seedlings were more likely to be found on the ground, but intermediate-sized trees were missing, showing their limited regeneration. Thus the stumps and logs play an important role as microsites for establishing conifers, except Abies firma, in an environment in which masses of rock or dwarf bamboo are dominant on the forest floor.