In 1982-1983, test sections were laid at three Texas locations using five asphalt sources and two grades. Not all combinations were used at all locations and two roads failed early, but results are reported on 16 sections. These roads were cored in 1984, 1985, and 1987. Viscosities, penetrations, voids, gel permeation chromatography (GPC), and Fourier transform infrared (FT-IR) analyses were run. Voids and aging are strongly correlated whether the aging is measured by a viscosity aging index, growth in carbonyl peak, or GPC. Although GPC, carbonyl peak, and sulfoxide generally increase with age, sulfoxide and other aging indexes are poorly correlated. Growth in sulfoxide correlates weakly with sulfur content. In general, the number of voids reveals large ranges in asphalt performance. Asphalts showing nearly the same aging index at low numbers of voids might differ several-fold at high numbers of voids. When comparison was possible, higher viscosity grades resulted in greater aging. Results also were obtained for several south Texas highways, ranging in age from 5 to 19 years. Although original tank asphalt, hot-mix, and early cores were not available for study, recovered asphalt properties from highway-aged cores followed trends similar to those for the test sections with respect to viscosity, penetration, FT-IR, and GPC measurements.