Significant differences in above-ground net primary productivity (ANPP) have been identified along a soil moisture gradient in mature longleaf pine-wiregrass ecosystems [Can. J. For. Res. 29 (1999) 743]. To determine if differences in ANPP were apparent at the individual tree level, patterns of mature longleaf pine ( Pinus palustris Miller) shoot and needle phenology were characterized on xeric and wet/mesic sites in southwest Georgia. Towers provided access to the crowns of the mature trees. Needle abscission patterns were monitored over one growing season (1999); shoot and needle growth patterns were monitored over three growing seasons (1999–2001). Environmental variables, including soil moisture content, predawn needle water potential and gap fraction, as well as several crown structural characteristics, were also measured. Crowns at the two sites had similar average branch characteristics, with the exception of average branch angle, which was significantly greater at the wet/mesic site. The starting dates for shoot elongation and linear shoot growth were similar at the two sites during all 3 years; cessation of shoot growth occurred significantly earlier at the wet/mesic site in 1999. Duration of shoot elongation was longer at the xeric site; however, shoot growth rate was greater at the wet/mesic site. In contrast, during all 3 years, start of needle elongation and start of linear needle growth occurred significantly earlier at the wet/mesic site; cessation of needle growth occurred significantly earlier at the wet/mesic site in 2000. Duration of needle growth was similar at both sites, but needle growth rate was greater at the wet/mesic site in 2000 and 2001. No consistent differences in shoot and needle phenology with crown position were observed, except shoots were longer in the top crown third at the xeric site during 1999 and 2001. Average needle lifespan was determined to be 15 months. Mature longleaf pine exhibited a determinant pattern (one seasonal flush) of shoot and needle growth during the first two growing seasons, and an indeterminate pattern (multiple seasonal flushes) during the third growing season. Soil moisture content was not strongly correlated to patterns of needle abscission; however, predawn needle water potential explained 17% of the variation in early growing season needle demography.