Seasonal variability in basal isoprene emission factor (&mgr;g C g(-1) h(-1) or nmol m(-2) s(-1), leaf temperature at 30 degrees C and photosynthetically active radiation (PAR) at 1000 &mgr;mol m(-2) s(-1)) was studied during the 1998 growing season at Duke Forest in the North Carolina Piedmont. Emissions from eight upper-canopy white oak (Quercus alba L.) leaves were measured periodically from the onset of isoprene emission on Day of Year (DOY) 119 (April 29) to leaf senescence in late October (DOY 299). Emissions from four leaves were measured under basal conditions with a controlled-environment cuvette system equipped with 10-ml gas-tight syringes and a reduction gas detector. Emissions from the other four leaves were measured under ambient conditions with the same system. Emission rates from the four leaves measured under ambient conditions were adjusted to basal conditions based on the PAR and leaf temperature algorithms of Guenther et al. (1993). The seasonal onset of isoprene emission was in agreement with previous studies where cumulative degree days from the date of the last spring frost were used to estimate bud break, leaf expansion, and increase in basal emission factor (EF). Between DOY 141 (May 21) and 240 (August 28), mean meteorological conditions 6 to 18 h prior to the EF measurements (ambient PAR and temperature) explained up to 78% of the variability in mean basal EF between measurement periods. Summertime mean isoprene emission potential was reached on DOY 141 (May 21) and was maintained until DOY 240 (August 28), when isoprene emission began to decline monotonically as leaf senescence approached. The mean value for leaves measured under ambient conditions and adjusted to basal conditions for DOY 141-240 was 75.6 &mgr;g C g(-1) h(-1) (74.2-79.1), whereas the mean value for leaves measured under basal conditions was 72.9 &mgr;g C g(-1) h(-1) (64.7-88.9). Between DOY 141 and 240, daily mean isoprene EFs varied from 54 to 96 &mgr;g C g(-1) h(-1) (27 to 49 nmol m(-2) s(-1)). In agreement with previous work at this and other sites, basal isoprene emission rates of fully exposed leaves at the crown apex of this tree were about 20% higher than those of the selected leaves. The length of the period prior to measurement of isoprene emission, during which meteorology was correlated with basal EF, appeared to be related to the timing and periodicity of meteorological change, and probably explains quantitative differences in the length of this period among studies. The empirical equation that we derived for this effect explained variability in midday EFs at the study site, but its general applicability remains to be tested. Strong diurnal changes in EF (as high as a factor of 2) are implied in this study, and should be examined further.