Figure 10. Hemodynamics in complete spectrum of retinal vessel sizes in population of vessels (in 19 normal C57BL6/J mice, across 123 vessels: 25 arterioles, 8 venules and 90 capillaries across all vessel generations in retina).

(A) Mean velocity vs lumen diameter in all vessels. Heterogeneity in velocity is observed across the spectrum (for arbitrary linear fits on velocity-diameter relationship, arteriole: R² = 0.31, venule: R² = 0.43, capillary: R² = 0.21). (B) Mean flow rate vs lumen diameter in all retinal vessels, from single-file-flow capillaries to the largest vessel near optic disk in mouse retina. Inclusion of diameter in flow calculation induces strong dependence of flow rate on diameter (for power fit, arterioles: exponent = 2.56, R² = 0.86, venules: exponent = 2.49, R² = 0.96). Murray’s theoretical model for blood vessels predicts a cubic relationship between flow rate and diameter. (C–D) Zoom-in of velocity and flow rate vs diameter for capillaries only (n = 90 capillaries). The linear vertical axis in B prevented the visualization of the near four orders of magnitude range of blood flow rates measured. Wide heterogeneity is observed, with weak correlation of capillary velocity and flow rate with lumen diameter (for arbitrary linear fit, velocity vs diameter: R² = 0.21, flow vs diameter: R² = 0.61). (E) Data from label-free measurements of single-cell flux in capillaries reported in our previous publication, reproduced here for 90 vessels. (F) Correlation of measured flux and flow rate in the same 90 single-file-flow capillaries (linear fit, R² = 0.79). The spread in the data represents variations in discharge hematocrit due to plasma skimming. Thus, in single-file-flow vessels (i.e. capillaries), measurement of velocity and flow alone gives an incomplete picture of nutrient delivery; cell flux gives a more accurate picture of the same.