The influence of specimen size on the measured values of true mode II fracture toughness (K IIc) of rocks have been investigated both experimentally and theoretically. In the experimental study, the double-edge notched Brazilian disk (DNBD) test is utilised to perform fracture toughness experiments with various specimen radii ranging from 27 to 101 mm. Test results demonstrate a significant increase of K IIc with the specimen size. The reason for this trend is explained by a novel mathematical model developed based on the fracture process zone and the energy release rate (ERR) concepts. Comparing the predictions of the theoretical model with the experimental data reveals that the proposed theory can accurately predict the size effect phenomenon in the DNBD test data. To compare the results of K IIc with K Ic, similar experimental and theoretical analyses are also performed for mode I fracture toughness tests using the semi-circular bend specimens. Lastly, we present a set of correction factors as functions of specimen size, whereby making it possible to estimate the size independent value of K IIc from the test results obtained from laboratory size DNBD specimens.