In an effort to obtain an improved mode III fracture toughness test suitable for a testing standard, mechanics analysis, experimental testing, and finite element analysis (FEA) have been conducted. Of particular concern are the merits of one-point and two-point edge crack torsion (ECT) test methods, the influence of specimen geometry that overhangs beyond load/support points, and the influence of crack length on the compliance and energy release rate. Shear stress distributions at the crack front are determined to examine the uniformity of mode III loading and mode II influence. The shear stress distributions in the one-point and two-point tests are virtually identical, indicating that either of the two tests could be used interchangeably. Based on the uniformity of the mode III shear stress distribution along the crack front, it was found that the ECT specimen should have minimum overhang. Longer crack lengths tend to produce nonuniform shear stress distributions. A modified two-point ECT test fixture was developed to allow testing of specimens with a range of dimensions. This development enabled experimental verification of the results from the FEA overhang series. The specimens with a minimum overhang produced consistent mode III toughness data. The most reliable way to reduce data is through the original compliance calibration method. A modified ECT specimen was developed with a staggered crack front to produce uniform mode III crack growth. Finite element analysis of the modified ECT specimen shows a uniform mode III stress distribution along the crack front with little mode II interaction.