vasodilation (and even the hypoxic regulation of blood flow to fatiguing muscle depends upon, for instance, tissue vasoactive substances and hypoxia sensors within smooth muscle, among other things). Time to fatigue during exercise is far too complicated a parameter to allow one to isolate or disentangle (from other maximally-activated vasoregulators) the contributions of the RBC to blood flow. Hypoxic vasodilation occurs in the systemic vessels, yet Isbell et al. 1 carried out only an in vitro assay of responses in pulmonary arteries, in which hypoxia normally causes vasoconstriction through a mechanism that partly depends on RBCs6. Assessment of pulmonary vascular responses to oxygen tension (pO2) or to RBCs cannot substitute for study of systemic hypoxic vasodilation, and the behaviors of pulmonary and systemic vessels are mechanistically different (for example, in their responsiveness to SNO, ATP and thiols and in the role of RBC deformation versus hypoxia in regulating tone). Furthermore, the data of Isbell et al. 1 suggest that relaxations of intact pulmonary arteries induced by humanized mouse RBCs are markedly diminished as compared with those of normal human RBCs and endogenous SNO-Hb7–10 (~ 8% versus 30–70%) 8. Moreover, the relaxations induced by the mutant RBCs were blocked by nitric oxide synthase (NOS) inhibition, implying a dependence on endothelium-derived NO and not on an NO-based signal derived from RBCs2, 8–10. By contrast, hypoxic vasodilation by native human RBCs is unaffected by NOS inhibition or endothelium removal and is greatly attenuated by SNO-Hb depletion8–10. Moreover, this dilation by RBCs is cyclic GMP dependent (in both intact and de-endothelialized vessels), consistent with NO-dependent relaxation8, 10. The small relaxations reported by Isbell et al. 1, probably mediated by ATP released from hypoxic or lysed RBCs, are thus characteristic of residual vasorelaxation elicited by RBCs that no longer dispense NO bioactivity, either because they have been depleted of SNO-Hb8, 9 or because export of NO bioactivity is impaired7. Indeed, the amount of SNO-Hb measured by Isbell et al. 1 in ‘wild-type’humanized cells (~ 200 nM, objections over their analytical methods notwithstanding11) should be sufficient to elicit substantial relaxation; as it did not, it would be important to know, at a minimum, if humanized SNO-Hb can bind and transnitrosylate mouse band 3 protein (appropriately coupled to pO2), which is required for hypoxic vasorelaxation2, 7. Central ideas in Isbell et al. 1 are constructed from misapprehension of previous findings and fundamental principles. It has not been asserted that SNO-Hb deficiency causes pulmonary hypertension, but rather that SNO-Hb deficiency aggravates hypoxemia-mediated pulmonary hypertension8—a condition that was not assessed by Isbell et al. 1 Moreover, the primary function of such vasoactivity in the lung is not to regulate pulmonary arterial pressure but to promote ventilation-perfusion matching, which was also not assessed. Furthermore, key controls necessary for the assessment of hypoxic vasodilation are missing, including measurements of vasodilation across an oxygen gradient (to evaluate allostery2), membrane