In vivo radioactive tracer and microsphere studies have differing conclusions as to the magnitude of the gravitational effect on the distribution of pulmonary blood flow. We hypothesized that some of the apparent vertical perfusion gradient in vivo is due to compression of dependent lung increasing local lung density and therefore perfusion/volume. To test this, six normal subjects underwent functional magnetic resonance imaging with arterial spin labeling during breath holding at functional residual capacity, and perfusion quantified in nonoverlapping 15 mm sagittal slices covering most of the right lung. Lung proton density was measured in the same slices using a short echo 2D-Fast Low-Angle SHot (FLASH) sequence. Mean perfusion was 1.7 +/- 0.6 ml x min(-1) x cm(-3) and was related to vertical height above the dependent lung (slope = -3%/cm, P < 0.0001). Lung density averaged 0.34 +/- 0.08 g/cm3 and was also related to vertical height (slope = -4.9%/cm, P < 0.0001). By contrast, when perfusion was normalized for regional lung density, the slope of the height-perfusion relationship was not significantly different from zero (P = 0.2). This suggests that in vivo variations in regional lung density affect the interpretation of vertical gradients in pulmonary blood flow and is consistent with a simple conceptual model: the lung behaves like a Slinky (Slinky is a registered trademark of Poof-Slinky Incorporated), a deformable spring distorting under its own weight. The greater density of lung tissue in the dependent regions of the lung is analogous to a greater number of coils in the dependent portion of the vertically oriented spring. This implies that measurements of perfusion in vivo will be influenced by density distributions and will differ from excised lungs where density gradients are reduced by processing.