Electrical potential drop (EPD) and compliance techniques are compared as techniques for crack length measurement in determining the crack growth resistance-curve (R-curve) of two Carbon/Carbon (C/C) composites and two carbon-bonded oxide-graphite refractories. The two C/C composites differ in the strength of the fibre/matrix interaction, resulting from the use of untreated and surface treated carbon fibres. The refractories differ in the volume fraction of graphite flakes. R-curve measurements on the C/C composites were made on specimens with chevron notches whilst straight-through notches were used for carbon bonded refractories. In the EPD method, the instantaneous crack length was determined from the instantaneous electrical potential across the notch plane, which was recorded in line with load and displacement data, and experimental calibration data. In the compliance method, the instantaneous crack length was determined analytically using the instantaneous load and displacement data. From the EPD technique smaller crack lengths were calculated than from the compliance technique in the regions of fracture where the composites had well developed process zones, and for the whole region in the refractories. The EPD technique underestimates the actual crack length, due to current conduction in the wake zone by bridging fibres/grains, and as a result the R-curves are different from those reported by the compliance technique, which are considered to be more reliable. The compliance-based results are used to establish the effects of fibre surface functionality and graphite flake content on crack growth resistance in the two systems. (C) 2005 Springer Science + Business Media, Inc.