Artificial thermoremanent magnetization imposed on natural basalt samples was used to check the reliability of Shaw's palaeointensity method <Shaw, 1974; Kono, 1978; Rolph and Shaw, 1985> and to give an indication of the accuracy of paleointensity determinations. First, thermal remanence was induced in all samples (50 in total) in a known field (50 ¿ 0.1¿T). This remanence was regarded as the original natural remanent magnetization (NRM). The samples were divided in to three subsets (A, B, and C) and intensity experiments were performed. Each subset of samples used different thermal remanent magnetization (TRM) acquisition parameters. Rock-magnetic properties on 14 representative subsamples were also measured both before and after the TRM acquisition steps to monitor possible magnetochemical alteration. It has been demonstrated that intensity estimates by the Shaw method give values close to the expected value (50 ¿ 5¿T) in all three sets of basalt samples. The modified method with Rolph and Shaw <1985> correction, i.e., using ratios of anhysteretic remanent magnetization given before the TRM step (ARM1) and after (ARM2), is proved to be a powerful approach to correct the TRM capacity changes that result from magnetochemical alteration. This is especially so for sample set C, where the samples were heated up to 700¿C and then were held at this temperature for 40 min during TRM acquisition. After correction, 45% of set C samples gave acceptable results. However, some samples with strong magnetochemical alteration sometimes invalidated the ARM1/ARM2 ratio correction and gave wrong intensity estimates. Monitoring rock magnetic properties is therefore very important when applying the Shaw method.