Three eukaryotic lineages generally are believed to have plastids that are primary in origin; that is, descended directly from a cyanobacterial endosymbiont. The recovery of these plastids as a monophyletic group in most molecular phylogenetic analyses, along with similarities in genome content and protein targeting mechanisms, have been cited as strong evidence in support of the hypothesis of a single endosymbiotic origin of all plastids. Although these data indeed are consistent with a single plastid origin, they also are consistent with the proposition of multiple endosymbiotic origins. Each hypothesis requires certain evolutionary assumptions in order to be reconciled with all existing data; at present, it is unclear which of these assumptions most likely reflect the historical process that gave rise to plastid diversity. Here we examine similarities in gene content among representatives of the three primary plastid lineages, using as a control the genome of a mitochondrion that almost certainly originated as an independent endosymbiotic association. To minimize metabolic constraints on gene retention we focus on two data sets, ribosomal protein and transfer RNA genes, neither of which is tied directly to specific organellar functions. Analyses of all possible pair-wise comparisons among the three plastids and mitochondrion indicate that genomic similarities are most consistent with convergent evolution due to constraints on gene loss, rather than with hypothesized shared evolutionary histories. We find no evidence of phylogenetic signal in the pattern of gene loss overlying this convergence. In light of these results, we address other lines of evidence and arguments that have been raised in support of a single plastid origin.

Key words: convergent evolution, endosymbiosis, gene loss, mitochondrion, phylogenetic, plastid origins