Correlations between the amounts of Xe 129 and Xe 128 released in stepwise heating of neutron-irradiated meteorites are used to determine the initial ratio I 129/I 127 and hence a relative formation time for the various samples. In such an experiment, we obtain the formation times in millions of years (relative to the L4 chondrite Bjurböle) of nine specimens: -3.9 ± 0.7 for Karoonda (C4 chondrite); 3.1 ± 0.6 and 10.5 ± 0.7 for the matrix and chondrules, respectively, of Chainpur (LL3); 7.5 ± 1.0 for St. Severin (LL6); 3.9 ± 2.9 and -2.3 ± 1.0 for the matrix and chondrules, respectively, of Allegan (H5); 3.6 ± 0.7 for Peña Blanca Spring and 20.8 ± 9.5 for Bishopville (aubrites); 3.8 ± 0.7 for a silicate inclusion of the iron meteorite El Taco (Campo del Cielo). No assumptions are made about the amount of Xe 129 in the trapped gas. The correlations for previously published data have also been recalculated in the same way, with no assumptions about trapped Xe 129. For a group of chondrites previously reported to be isochronous with a mean simultaneity of 2.5 million years, the recalculation confirms this mean simultaneity, but significant differences in formation times are resolved. An iodine-xenon age of 53 ± 9 million years (after Bjurböle) is obtained from unpublished data for the achondrite Lafayette. The reliability of iodine-xenon ages of individual meteorites is considered; in particular, the ages of Bishopville and Lafayette are less reliable than those of most other meteorites studied, especially in view of the anomalous ages reported. The relevance of iodine-xenon dating to theories of nucleosynthesis, early solar system chronology, and theories of meteorite parent-body formation is discussed.