We have read with attention the letter from Drs. Poole and Musch (Poole and Musch 2008) and we thank them for giving us the opportunity to explain in further details our methodology and interpretations. To clarify the ideas developed in our paper, we would like to go back to some of the definitions related to the key parameters we have used. In the original study (Daussin et al. 2007), maximal arteriovenous oxygen difference D aÀ" v O 2 max ð Þwas determined, according to the Fick equation, by dividing maximal oxygen consumption _ VO 2 max À Á by maximal cardiac output _ Q max À Á and is often considered to indirectly reflect peripheral oxygen extraction through the following equation: O 2 extraction (%) = D aÀ" v O 2 max arterial O 2 content  100 (Mortensen et al. 2005). As a first step, Poole and Musch questioned the validity of our results, especially _ Q max and D aÀ" v O 2 max ; emphasizing the possibility that our values may have been unrealistic. After analysis of the literature on the question, it appears that some studies measured higher values of _ Q max for similar _ VO 2 max and then lower D aÀ" v O 2 max compared to our study (Hagberg et al. 1985;McGuire et al. 2001). For example, Stickland et al. (2004) measured, with the ''gold standard'' direct Fick method, a _ Q of 22.2 L min -1 (19.5 L min -1 in our study) for a _ VO 2 of 2.49 L min -1 (2.51 L min -1 in our study), thereby giving a DavO 2 around 11.2 mL dL -1 (12.7 mL dL -1 in our study).During the study, the subjects trained at 61% of maximal power output (P max ) in the continuous training (CT) modality and they alternated 4 min at low intensity (49% of P max ) with 1 min at 90% of P max during interval training (IT). In our design, CT involved a constant intensity and displayed a constant heart rate. Therefore, we assumed that both _ Q and muscle blood flow would have been relatively constant during the training session and may not have been high enough to elicit significant enhancement of _ Q max and systemic O 2 delivery.We agree that we observed a small elevation in _ Q max and compared to the substantial increase of _ VO 2 max : However, the ratio between D _ Q _ Q max À _ Q rest À Á and D _ VO 2 _ VO 2 max À _ VO 2rest À Á was 6.8 before training and went down to 6.1 after training, both being values similar to those documented in the literature (Taivassalo et al. 2003) and there are in accordance with remarks of Poole and Musch (Poole and Musch 2008).Regarding the error in the determination of _ Q determination and _ VO 2 max pointed out by Poole and Musch, we would like, first, to remind the reader that the multiplication of the mean values is different from the mean of multiplied values. Regarding the training induced improvement of _ VO 2 max ; we have carefully checked all the individual data and we notice that for one subject an error occurred. Taken the new value into account, we get an increase of 21.5% corresponding to a _ VO 2 max of 32.2 mL min -1 kg -1 . In exact details, this improvement in _ VO 2 max was brought about by a...