Purpose:To assess the effect of sample site (earlobe vs toe) and incremental exercise protocol (continuous vs discontinuous) on training zone prescription in rowing.Methods:Twenty-six rowers performed two incremental exercise tests on an ergometer: (1) a five-step discontinuous test with 4-min stages and 30-W increment, with blood samples taken from the earlobe and toe at the start of the 1-min break between steps; (2) a continuous test, with 2-min stages and 30-W increment, with blood samples taken from the right first toe at the end of each stage. Blood was analyzed for lactate concentration.Results:At a lactate concentration of 2 mmol·L−1, the mean (95% CI) power output was 8.1 (± 15.4) W greater for the continuous protocol, the random error between the methods (1.96 × SD of differences) was ± 58.8 W, and there was no evidence of any relationship between power output and error between methods. At a lactate concentration of 4 mmol·L−1, the mean (95% CI) power output was 24.2 (± 17.0) W greater for the continuous protocol, and the random error was ± 64.8 W. At 4 mmol·L−1, systematic bias between methods increased with high power outputs.Conclusions:The continuous protocol with toe sampling led to higher power outputs for a given lactate concentration compared with the discontinuous protocol with earlobe sampling. This was partly due to the choice of sample site and largely due to the choice of protocol. This bias, and also random variability, makes direct comparison of these tests inappropriate.