Capillary diameter and geometry in cardiac and skeletal muscle studied by means of corrosion casts.

Studies of microvascular geometry made from microscope observations of tissues in vivo or after perfusion with a silastic elastomer or india ink are restricted to a two-dimensional field of view. Microvascular corrosion casts, however, if of sufficient rigidity and structural integrity, can yield three-dimensional information when examined under the scanning electron microscope. We have used modified Batson's No. 17 anatomical casting compound (having a shrinkage less than 1% on setting) to prepare casts of the microvasculature of the heart and skeletal muscles in anesthetized rats. In casts from the L. ventricle the capillary network appeared to parallel the arrangement of the muscle fibers, but showed many capillary loops and anastomoses. In skeletal muscles (gastrocnemius and gracilis) held at full extension, in situ, the casts showed long straight capillaries with fewer branchings than in the heart. In shortened skeletal muscle the capillaries exhibited an undulatory configuration. Capillary diameters (mean +/- SD) were 5.14 +/- 1.42 micrometers (N = 202), 5.04 +/- 1.45 micrometers (N = 294) and 4.84 +/- 1.97 micrometers (N = 335) in L. ventricle, gastrocnemius, and gracilis muscles (both shortened), respectively. The mean values for capillary diameter in these three tissues did not differ significantly. Combining our data with those of L. Henquell, P. L. LaCelle, and C. R. Honig on erythrocyte deformability in the rat (Microvasc. Res. 12, 259-274 (1976)) suggests that even when the capillary bed is fully distended the smallest capillaries, amounting to 1-2% of the total number, must be channels for plasma flow alone. In cross-sectional views of the casts from contracted skeletal muscle the capillaries appeared to form a tightly meshed network of convoluted vessels around the fibers, such that in some regions a large fraction of the surface of each fiber was in contact with blood. The Krogh cylinder geometry appears not to be appropriate for modeling O2 transport in maximally shortened skeletal muscle; a more appropriate model may be that of a cylindrical muscle fiber supplied, at any point down its length, by a uniform peripheral O2 supply.