This article describes the activity of an E-glass microfiber (104E) during chronic inhalation and intraperitoneal injection studies in rats. Results are compared with another microfiber of similar dissolution rate (k(dis)), code 100/475, and the more durable amosite asbestos, both of which we had previously used in similar experiments (Davis et al., 1996). Rats were exposed to aerosol concentrations of 1000 fibers (longer than 5 microm)/ml, as measured by optical microscopy, for 7 h/day, 5 days/wk. Subgroups of rats were followed for mean lung burden, early and late signs of fibrosis, and tumor incidence. At the end of 12 mo of exposure, the mean number of 104E fibers of all lengths in the lungs was approximately double that for amosite but two-thirds of that for 100/475. For fibers longer than 15 microm, the mean 104E burden was similar to that for the amosite and more than twice that of the 100/475. After a 12-mo recovery period, the retained lung burdens (of fibers of all lengths) were approximately 30% of those at 12 mo for both microfibers, and somewhat higher (approximately 44%) for amosite. Amosite and 100/475 fibers longer than 15 microm were more persistent in the lungs than 104E fibers. The chemical composition of 104E fibers did not appear to have been significantly altered by up to 24 mo of residence in lung tissue, whereas the composition of 100/475 was substantially altered over the same time period. From the inhalation study, out of the pathology subgroup of 43 animals exposed to 104E microfibers, 10 had lung tumors (7 carcinoma, 3 adenoma) and 2 had mesotheliomas, whereas in 42 rats exposed to amosite asbestos, there were 16 lung tumors (7 carcinoma, 9 adenoma) and 2 mesotheliomas. The 104E- and amosite-treated animals had similar levels of fibrosis. In contrast, 38 animals treated with 100/475 had little fibrosis, 4 lung tumors (adenomas), and no mesotheliomas. The greater pathogenicity of the 104E fibers, compared to 100/475 fibers, might be partly explained by the greater numbers of long fibers retained in the lung after 12 mo of inhalation. However, we speculate that modification of surface properties by extensive selective leaching of some glass components reduces the toxic potential of 100/475. In a parallel intraperitoneal injection study, 104E caused considerably more mesotheliomas (21 rats out of 24) than 100/475 (8 rats out of 24). In addition, 104E appeared to be more active than amosite asbestos, since mesotheliomas appeared much more quickly in the 104E-treated animals. In conclusion, we have shown that two microfiber types, 100/475 and 104E, of similar dissolution rates, had markedly different pathogenicity in rats. We believe that this contrast is only partly due to differences in numbers of long fibers and that differences in surface properties of the fibers, possibly due to proportionately greater leaching of 100/475 fibers, play an important role.