Abstract
Phylogenetic reconstructions are often plagued by difficulties in distinguishing phylogenetic signal (due to shared ancestry) from phylogenetic noise or homoplasy (due to character-state convergences or reversals). We use a new interpretive hypothesis, termed hemiplasy, to show how random lineage sorting might account for specific instances of seeming "phylogenetic discordance" among different chromosomal traits, or between karyotypic features and probable species phylogenies. We posit that hemiplasy is generally less likely for underdominant chromosomal polymorphisms (i.e., those with heterozygous disadvantage) than for neutral polymorphisms or especially for overdominant rearrangements (which should tend to be longer-lived), and we illustrate this concept by using examples from chiropterans and afrotherians. Chromosomal states are especially powerful in phylogenetic reconstructions because they offer strong signatures of common ancestry, but their evolutionary interpretations remain fully subject to the principles of cladistics and the potential complications of hemiplasy.