Trabecular thickness and the trabecular star volume were estimated in 49 normal individuals (20 males and 29 females) using an unbiased sampling procedure consisting of anisotropic, vertical sections, and a corresponding anisotropic test system. Eight-microns-thick undecalcified stained vertical sections were obtained from iliac crest specimens. Implying a trabecular plate model mean trabecular thickness (Tb.Th., microns) was estimated by different methods: (a) the ratio between bone volume and bone surface (Tb.Th.ratio); (b) the mean of the orthogonal intercepts multiplied with pi/4 (Tb.Th.l0); and (c) the mean of random linear intercepts divided by 2 (Tb.Th.l1). Trabecular star volume (Vtr*, mm3), a true three-dimensional size parameter independent on any model, was estimated using linear intercepts in random directions. Significant (2p less than 0.001) correlations were found between the different methods with coefficients of correlation ranging from 0.71 to 0.89. The Tb.Th. ratio was overall higher (155 +/- 31 microns, (X, SD), 2p less than 0.001) than the other linear estimates, Tb.Th.l0: 144 +/- 26 microns and Tb.Th.l1: 135 +/- 27 microns. Estimates based on orthogonal intercepts (l0) were slightly higher (2p less than 0.001) than those from random linear intercepts (l1). Among all the individuals no significant age-related changes were seen in the measured parameters. However, postmenopausal women (greater than 50 years of age) had more thick trabeculae than younger women (2p = 0.03). Intra-individual distribution analysis disclosed that this finding could not be explained by preferential perforation and loss of thinner trabeculae during bone remodeling. Therefore, normal postmenopausal women do increase trabecular thickness with age.(ABSTRACT TRUNCATED AT 250 WORDS)