Identifying traits that affect rates of speciation and extinction and, hence, explain differences in species diversity among clades is a major goal of evolutionary biology. Detecting such traits is especially difficult when they undergo frequent transitions between states. Self-incompatibility, the ability of hermaphrodites to enforce outcrossing, is frequently lost in flowering plants, enabling self-fertilization. We show, however, that in the nightshade plant family (Solanaceae), species with functional self-incompatibility diversify at a significantly higher rate than those without it. The apparent short-term advantages of potentially self-fertilizing individuals are therefore offset by strong species selection, which favors obligate outcrossing.

Breakdown of self-incompatibility occurs repeatedly in flowering plants with important evolutionary consequences. In plant families in which self-incompatibility is mediated by S-RNases, previous evidence suggests that polyploidy may often directly causeself-compatibility through the formation of diploid pollen grains. We use three approaches to examine relationships between selfincompatibility and ploidy. First, we test whether evolution of self-compatibility and polyploidy is correlated in the nightshade family (Solanaceae), and find the expected close association between polyploidy and self-compatibility. Second, we compare the rate of breakdown of self-incompatibility in the absence of polyploidy against the rate of breakdown that arises as a byproduct of polyploidization, and we find the former to be greater. Third, we apply a novel extension to these methods to show that the relative magnitudes of the macroevolutionary pathways leading to self-compatible polyploids are time dependent. Over small time intervals, the direct pathway from self-incompatible diploids is dominant, whereas the pathway through self-compatible diploids prevails over longer time scales. This pathway analysis is broadly applicable to models of character evolution in which sequential combinations of rates are compared. Finally, given the strong evidence for both irreversibility of the loss of self-incompatibility in the family and the significant association between self-compatibility and polyploidy, we argue that ancient polyploidy is highly unlikely to have occurred within the Solanaceae, contrary to previous claims based on genomic analyses. K E Y W O R D S :Angiosperms, breeding systems, comparative methods, polyploidy, self-incompatibility, statistical phylogenetics.Self-incompatibility (SI) is the common ability of hermaphrodite plants to recognize and reject their own pollen with a genetically based mechanism. Approximately one-half of all extant angiosperm species prevent self-fertilization by deploying SI (Brewbaker 1959;. Although many different mechanisms of SI evolved in flowering plants, a particular system, which uses RNases in the female component of self-rejection, appears to be ancient. It likely originated at least 90 million years ago (mya), and it may cause SI in dozens of extant eudicot families, including some of the most diverse . A notable property of SI systems is the high rate at which they transition to self-compatibility (SC), considered one of the most common evolutionary transitions in plants (Stebbins 1974). Consequently, many individuals, populations, and species do not express SI. Empirical data on the mechanisms that underlie losses of SI are sparse. Although the genetic causes of transitions from SI to SC vary widely (Stone 2002), a significant proportion of such transitions may involve polyploidization (Livermore and

We characterized the molecular allelic variation of RNases at the self-incompatibility (SI) locus of Solanum chilense Dun. We recovered 30 S-RNase allele sequences from 34 plants representing a broad geographic sample. This yielded a species-wide estimate of 35 (95% likelihood interval 31-40) S-alleles. We performed crosses to confirm the association with SI function of 10 of the putative S-RNase allele sequences. Results in all cases were consistent with the expectation that these sequences represent functional alleles under single-locus gametophytic SI. We used the allele sequences to conduct an analysis of selection, as measured by the excess of nonsynonymous changes per site, and found evidence for adaptive changes both within the traditionally defined hypervariable regions and downstream, near the 3 0 -end of the molecule.

No abstract