Table 2.
Total number of species |
Expected number of taxa with SS/SI |
||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Barraclough et al. | Isaac et al. |
Barraclough et al. | Isaac et al. |
||||||||
Dav | Wik | EBe | LBe | Dav | Wik | EBe | LBe | ||||
W | –443 | –509 | –506 | –501 | –503 | –339 | –373 | –365 | –331 | –339 | |
SC vs. SS | Z | –3·216 | –3·694 | –3·673 | –3·636 | –3·651 | –2·462 | –2·708 | –2·650 | –2·404 | –2·462 |
P | 0·002 | <0·001 | <0·001 | 0·001 | 0·001 | 0·019 | 0·010 | 0·012 | 0·022 | 0·019 | |
W | –17 | –23 | –21 | –25 | –25 | –5 | –19 | –19 | –19 | –19 | |
SC vs. | Z | –0·892 | –1·198 | –1·096 | –1·300 | –1·300 | –0·280 | –0·994 | –0·994 | –0·994 | –0·994 |
Homomorphic SI | P | 0·268 | 0·195 | 0·219 | 0·171 | 0·171 | 0·384 | 0·243 | 0·243 | 0·243 | 0·243 |
W | –30 | –32 | –32 | –32 | –32 | –24 | –24 | –24 | –24 | –24 | |
SC vs. | Z | –2·135 | –2·275 | –2·275 | –2·275 | –2·275 | –1·715 | –1·715 | –1·715 | –1·715 | –1·715 |
Polymorphic | P exact | 0·025 | 0·025 | 0·025 | 0·025 | 0·025 | 0·060 | 0·060 | 0·060 | 0·060 | 0·060 |
W | –119 | –129 | –123 | –119 | –121 | –87 | –85 | –79 | –63 | –63 | |
SC vs. | Z | –2·602 | –2·820 | –2·689 | –2·602 | –2·646 | –1·905 | –1·862 | –1·731 | –1·383 | –1·383 |
SS unclass | P | 0·014 | 0·007 | 0·011 | 0·014 | 0·012 | 0·065 | 0·071 | 0·089 | 0·153 | 0·153 |
W, Sum of ranks; Z, value; P, associated probability; P exact, exact probability for Wilcoxon signed rank test when n < 10) used to test the hypothesis that both species richness and DRs have been the same in self-compatible and sister clades. Significant values are highlighted in bold. Comparisons were based on the total number of taxa in clades with SC against both: (1) the total number of taxa in sister group or (2) the estimated number of species in the sister group expected to have SS (see text for details). Two contrast methods were used to analyse the data: that given in Barraclough et al. (1996) based on comparisons of species richness and that given in Isaac et al. (2003) based on DR {[ln(xi)/ln(yi)] and [log(xi) – log(yi)]/ti, respectively} using the estimates of Davies et al. (2004) (Dav) and Wikström et al. (2001) (Wik) and exponential and log-normal calibrations of Bell et al. (2010) (EBe and LBe, respectively) for the divergence age of the sister group (ti).