In 1998 Cho et al. published a paper titled "Explosive invasion of plant mitochondria by a Group I intron." They included a single representative for six of the eight families of the order Zingiberales, hypothesizing three independent acquisitions of the intron in the mitochondrial coxI gene. The Zingiberales are a distinctive yet diverse order of broad-leaved monocots including bananas, heliconias, and gingers, with ~2000 species divided into 92 genera. The order and its two largest families (Marantaceae and Zingiberaceae) have molecular phylogenies available allowing the distribution of this intron to be mapped, and hypotheses regarding evolutionary history tested. A simple PCR screen was used to determine presence/absence of the intron for over 200 samples representing all families and most genera of the order. Targeted sequencing was used to refine estimates of the number of independent gains. This study confirms 1) the presence of the intron in three of the eight families in the order: Musaceae, Zingiberaceae, and Marantaceae, and 2) a minimum of three independent acquisitions of the intron. The broad sampling of this study suggests a single gain in the Musaceae, at least six gains and two losses for Zingiberaceae, and two gains for Marantaceae. Evidence for loss of the intron comes from base changes in two co-conversion tracts, regions in the exonic sequences flanking the intron. Not all taxa hypothesized to have lost the intron retain an intron "footprint" in the co-conversion tract suggesting that this footprint is rapidly lost and not a reliable indicator of historic intron presence.

Key words: coxI intron, Marantaceae, mitochondria, phylogeny, Zingiberaceae, Zingiberales