This paper reports results from the mode II testing of adhesively-bonded carbon-fibre-reinforced composite substrates using the end-loaded split (ELS) method. Two toughened, structural epoxy adhesives were employed (a general purpose grade epoxy-paste adhesive, and an aerospace grade epoxy-film adhesive). Linear Elastic Fracture Mechanics was employed to determine values of the mode II adhesive fracture energy, G IIC for the joints via various forms of corrected beam theory. The concept of an effective crack length is invoked and this is then used to calculate values of G IIC. The corrected beam theory analyses worked consistently for the joints bonded with the epoxy-paste adhesive, but discrepancies were encountered when analysing the results of joints bonded with the epoxy-film adhesive. During these experiments, a microcracked region ahead of the main crack was observed, which led to difficulties in defining the true crack length. The effective crack length approach provides an insight into the likely errors encountered when attempting to measure mode II crack growth experimentally.