The Design Institute for Physical Properties (DIPPR) maintains the DIPPR 801 database for the American Institute of Chemical Engineers. Autoignition temperature (AIT) is one of the properties included in the database and is the focus of this work including improvement of the overall state of AIT in the database. Phenomena related to AIT as well as the relevant literature are reviewed. Likewise, the database is presented to respond to significant misuse of the DIPPR 801 database in the literature. The database is evaluated, respecting AIT, as a whole to show where improvement is needed. An experimental study of minimum autoignition temperatures reveals unexpected behavior of pure n-alkanes not predicted by current current phenomenological understanding of autoignition processes. Measurements show an increase at C16 and a dramatic and previously unexplained step increase between C25 and C26. Experimental modifications are presented to compensate the effect of altitude. Measured values for several n-alkanes are reported and compared to the literature. Other ignition experiments and decomposition measurements using differential scanning calorimetry are also reported and examined to elucidate the unexpected trends. Explanations for these trends are proposed. Finally, the implications of this for trends in other chemical families are discussed. A comprehensive examination of AIT family trends reveals variation from the n-alkane family trend. Measured AIT values are presented and discussed. Evaluated AIT values are recommended for several single-group chemical families. Phenomenological explanations for observed differences are proposed and discussed along with the broader implications for these trends. Methods for predicting autoignition temperatures (AIT) have been historically inaccurate and are rarely based on the underlying physical phenomena leading to observed AIT. An improved method for predicting AIT based on the method by the late Dr. William H. Seaton is presented and discussed. The method of Seaton is described in detail. An evaluated data set is used to regress new parameters for the Seaton method parameters. Improvements to Seaton's model and underlying principles are presented and discussed. Finally, an improved AIT prediction method is presented and recommended.