We examined the timing of root elongation intensity (REI) - the production of newly elongating roots measured in length per unit soil volume per unit time - in a mature white oak-chestnut oak forest in Tennessee, USA, over a 4-year period. We used a network of minirhizotrons to (1) examine environmental control of soil moisture and temperature over REI, (2) evaluate the control of phenology over REI (3) develop a multivariate regression equation using the variables of soil temperature, soil water potential and phenology to predict REI, and (4) delineate soil water potentials that were optimum for growth. Fifteen minirhizotron tubes were installed on the upper slope of an upland oak stand on Walker Branch Watershed in Oak Ridge, Tennessee. Measurements of REI were examined over three growing seasons in the upper 30 cm of the soil profile. Composited sets of data on REI, soil water potential, soil temperature and phenology were analyzed to determine the strength of relationships between the latter three variables and REI. There was no statistically significant relationship between REI and soil temperature during the growing season (March-November). The highly significant (p < 0.0001) relationship between soil water potential and REI was best described as log linear, using log(1)0 (-Psi) vs. log(10) (REI). The correlation between a phenology index and REI was also highly significant (p < 0.0001), as the 24 highest REI levels occurred between late April and early August. Whereas soil water potential played a major role in controlling REI, our data indicate that at times phenologically related factors appeared to override environmental variables. The fact that REI generally peaked every year following completion of leaf expansion is congruent with the fixed shoot growth pattern of many temperate-zone deciduous tree species. In an attempt to explain the importance of phenology in controlling REI, we refer to the concept of 'phenological programming'. This concept contrasts with more environmentally determinant explanations of the timing of root growth. Those theories cannot explain either our data or that of others, wherein, during late summer and early autumn, REI does not return to early summer rates, even though soil moisture and temperature conditions are equally, if not more, favorable. The particular type of 'phenological programming' hypothesized in this study may be limited to mature deciduous trees exhibiting a fixed growth pattern of shoot expansion and growing in a temperate climate with a mid-summer drought period.
