A summary of relationships obtained from the literature between log(male) and log(female) size from a wide range of taxa with predominantly FBSSD or with mixed MBSSD and FBSSD in which data were provided separately for FBSSD and MBSSD species.

aThe three values for N are: the number of species in the study (approximate number of species in the taxonomic or functional group considered); number of species in sample with FBSSD.
bWe expand upon our reasons for any disagreement in the text.
cTaxa are classified in [2] as female-biased if at least 80% of species in the sample had FBSSD.
dEstimates of the slope and its 95% C.I. were measured off figure 2 of [2].
eThe conclusion in [2] that Rensch's rule was a general phenomenon based on a meta-analysis of the taxa listed here as well as taxa with MBSSD and mixed SSD, and so we list them as supporting Rensch's rule in each case; whereas we consider each taxon separately.
fSample sizes given are for all species in which females are larger than males. However, in the analysis we consider only those species in which the female is ≤2 x the size of the male, which is the conventional delimitation of ‘extreme’ SSD in spiders [40]. When females are much larger than males in this group, the correlation between male and female size breaks down, and it is not clear what the appropriate technique for examining scaling in SSD is in such a situation. The sample sizes for the analyses using data from [6] and [40] were therefore 11 and 476 species respectively.
gWe analyse data for all species-level data given in the appendix to [41]; note that the analyses presented [41] are at the subfamily level, and use data for 209 species of waterstrider. He found evidence for Rensch's rule in 8 of 9 subfamilies with at least 10 species; we too support Rensch's rule for species with FBSSD in the two largest subfamilies, Gerrinae (n = 59, slope = 0.89±0.028, P = 0.0003) and Halobatinae (n = 21, slope = 0.78±0.081, P = 0.0144).
hWe focus on thorax length from the several morphological measures given in [34].
iWe have combined data from the three sources cited in the table. For the small number of species occurring in more than one dataset, we took the arithmetic mean size (across datasets) for male and female size. Including source as a factor in the analysis did not substantially affect the estimate of the slope (common slope estimated as 0.86±0.028, significantly <1 with P<0.00001).
jWe reject Rensch's rule in this case as the P value for the test for a slope significantly different from 1 is 0.0553, but accept that this is only marginally non-significant.
kThis slope is an unpublished result mentioned in [34], and is the slope of log(male) on log(female) size, i.e. a situation in which Rensch's rule predicts a slope >1.
lWe focus on body mass from the several morphological measures given in [17].