In capillary gel electrophoresis (CE), changing run conditions such as temperature can result in minor variations in the size determination of an allele. These effects are caused by secondary structure differences that can occur between the amplified sample and the internal standard. The type of method chosen to generate the sizing curve in STR analysis can influence the relationship between estimated allele size and temperature. To better understand the effects of temperature and sizing method on the reproducibility of DNA migration, two fluorescently labeled allelic ladders, CTTv and Y-PLEX 6 were analyzed using the ABI Prism 310 Genetic Analyzer. The default method on the Genetic Analyzer utilizes an electrophoretic temperature of 60 degrees C and a Local Southern method to generate a sizing curve from the fragment migration times of the internal lane standard. In this work, electrophoresis was conducted at 35-70 degrees C using the commercially available POP 4 buffer at pH 8 and two sizing methods, Global Southern and Local Southern, were compared. The slopes of the regression line between estimated allele size and temperature, using either sizing method, were measured in order to demonstrate the temperature sensitivity of migration time and the importance of the operator-chosen method. Our results indicate that the Global Southern method is a better choice in situations where temperature fluctuations can occur. In addition, the temperature dependence of the DNA size estimates using the POP 4 system were compared to results obtained using an experimental buffer consisting of 3% hydroxyethylcellulose at pH 11. These results demonstrate that secondary structure effects are minimized at an elevated pH, increasing the precision of size estimates obtained.