||Wide-bore (0.75mm) glass capillary columns were evaluated and used in the routine detection of arson accelerants using headspace sampling and gas chromatography. Non-polar (SPB-1), intermediate polar (SPB-35), and polar (Supelcowax-10) stationary phases were evaluated. The sensitivity and reproducibility were better for the wide-bore capillary columns than the comparative packed columns with better resolution. The non-polar SPB-1 column studied showed excellent separation of non-polar components in the petroleum products examined especially higher boiling accelerants such as kerosene and diesel fuel. The intermediate and polar stationary phases can also be used effectively with the wide-bore capillary column. A glass tee installed in the injection port allows one to split the sample onto two columns and gather additional data using columns with stationary phases of different polarity effectively. This configuration gives a quantitative split of the sample which is reproducible. A study of different parameters affecting the sampling of arson accelerants using static headspace analysis was conducted by determining the percent recoveries of the components of a standard hydrocarbon mixture gathered under different conditions. When disposable syringes are used at room temperature for sampling, discrimination can occur especially with higher boiling components (n-dodecane or higher). Pumping the syringe resulted in less reproducibility and caused discrimination in chromatographic patterns. Studies showed that a significant amount of the more volatile components in an accelerant can escape through the filament tape while the can is heating in the oven. Further tests showed that recoveries, discrimination in sampling, and loss of volatiles can be improved significantly through the use of a heated glass gas-tight syringe. The characteristics of various adsorbents were studied using the dynamic headspace technique. The studies showed Tenax-GC to be a superior adsorbent when compared to traps containing comparative amounts of other common adsorbents (Porapak Q, Chromosorb-102, Carbotrap, Charcoal). Tenax-GC traps showed quantitative desorption characteristics as well as good recovery of the test mixture used. The dynamic headspace technique was shown to be more sensitive than heated static headspace by a factor of 800-6500 depending on the components' vapor pressure.