Second, in Experiment 1 with Shiraz there were two treatments for which a change in management practices occurred during the experiment: FERT and IRR-FERT (irrigation and fertilization started in 2011). We showed before that this led to a change in ranking of treatments according to water and nitrogen stresses (Table 2). Moreover, FERT and IRR-FERT treatments are shown with different symbols in Fig. 1C and D (gray circles for FERT and open circles for IRR-FERT) and their trajectories fall within the general trends of the relationships. Another methodological concern would be the non-independency of the two seasons 1–2 sequences (i.e. 2010–2011and 2011–2012). Might water or nitrogen stress in 2010 have impacted yield components in 2012? Given the timing of the morphogenesis of inflorescences, the only way of water and nitrogen stress in 2010 to impact bud fertility and berry number per bunch in 2012 would be through the dynamics of carbon and nitrogen reserves between 2010 and 2011: a low reserve status around flowering in 2011 might have an impact on inflorescence formation in 2012, due to a high stress in 2010 during reserves replenishment. However, values of LNC and predawn at veraison along with YAN at harvest in 2010 (Table 1) indicate mild levels of water and nitrogen stress that are not expected to disturb reserves replenishment (Holzapfel et al., 2010). Therefore, no effect of stress in 2010 on yield components in 2012 was expected in the present study.