Human bone blood flow, mean blood speed and the number of moving red blood cells were assessed (in arbitrary units), as a function of time, during one cardiac cycle. The measurements were obtained non-invasively on five volunteers by laser-Doppler flowmetry at large interoptode spacing. The investigated bones included: patella, clavicle, tibial diaphysis and tibial malleolus. As hypothesized, we found that in all bones the number of moving cells remains constant during cardiac cycles. Therefore, we concluded that the pulsatile nature of blood flow must be completely determined by the mean blood speed and not by changes in blood volume (vessels dilation). Based on these results, it is finally demonstrated using a mathematical model (derived from the radiative transport theory) that photoplethysmographic (PPG) pulsations observed by others in the literature, cannot be generated by oscillations in blood oxygen saturation, which is physiologically linked to blood speed. In fact, possible oxygen saturation changes during pulsations decrease the amplitude of PPG pulsations due to specific features of the PPG light source. It is shown that a variation in blood oxygen saturation of 3% may induce a negative change of ∼1% in the PPG signal. It is concluded that PPG pulsations are determined by periodic 'positive' changes of the reduced scattering coefficient of the tissue and/or the absorption coefficient at constant blood volume. No explicit experimental PPG measurements have been performed. As a by-product of this study, an estimation of the arterial pulse wave velocity obtained from the analysis of the blood flow pulsations give a value of 7.8 m s(-1) (95% confidence interval of the sample mean distribution: [6.7, 9.5] m s(-1)), which is perfectly compatible with data in the literature. We hope that this note will contribute to a better understanding of PPG signals and to further develop the domain of the vascular physiology of human bone.