Species/genus richness among groups
Species/genus richness values from the four feeding types and 37 taxonomic groups (11 families and 26 subfamilies) are illustrated in Fig. 1 (see also Suppl. 1). Phytophagous species exhibited both the highest species richness (3651) and genus richness (575) values while non-feeders were found to exhibit the lowest species richness (29) with only 2 genera. The omnivorous and coprophagous groups exhibited similar species richnesses (370 and 369, respectively), while the genus richness of the former was higher than the later (62/36). In the family category, for our sample Scarabaeidae and Lucanidae both exhibited very high species richness (2776/1305) and genus richness (534/81) values primarily because of their large and diverse sub-groups; Glaphyridae, Glaresidae and Pleocomidae exhibited very low species (29/29/26) and genus (2/1/1) richnesses (with the species richnesses of the Ochodaeiade and Passalidae being higher than those of all three families mentioned above; 36/33), as well as higher per capita genus richness values (9/14). In addition, the Diphyllostomatidae was found to exhibit the lowest species richness value (3) with only one genus included in the test sample. Similar patterns were also found in subfamily category test, with most subfamilies exhibiting high species richness and relatively high genus richness values (Fig. 1).
Morphological Variations Of Pronotum And Elytron
The first three principal components (PCs) of the pronotum and elytron from 5148 species accounted for 87.90% and 93.36%, of observed shape variation respectively (Fig. 2, 3). Shape models were calculated at equally spaced intervals along these PC axes to document the deformation of test characters, with the overlaying of these shape models used to reveal the comparative along-axis shape-change trends (Fig. 2, 3). Variation summaries for the other principal components from this analysis are provided in Supplement 2.
Along the positive direction of the first PC axis, pronotum shape changed from a trapezoidal to inverted trapezoidal outline with the producing anterior angle, and retracting anterior margin, making the anterior angle sharper, while the posterior angle diminished and became less distinct. Along the positive direction of the second PC axis, the pronotum became stretched longitudinally and diminished horizontally with both anterior and posterior angles diminishing and becoming less distinct and the entire pronotum becoming markedly more oval. Along the third PC axis, the entire pronotum tends toward an inverted triangle with the anterior angle extending outward horizontally, the posterior angle shrinking inwardly, and the posterior margin stretching longitudinally.
Along positive direction of its first PC axis, the elytron becomes laterally narrower with the anterior angle contracting inward. The shrinking smaller scutel causes the anterior margin to become shorter with the lateral margin, that protruded outward originally, contracting inward. Along the positive direction of the second elytron PC axis the anterior angle becomes larger and the posterior margin protrudes. The entire elytron stretches longitudinally and shrinks horizontally. Along the positive direction of third elytron PC axis, the smaller scutel flattens the front margin. At the same time, the posterior margin shrinks inward to make its trace a smooth curve.
Morphological Diversity Among Groups
Morphological diversity of both the pronotum and elytron among the 11 families and 26 subfamilies reflects the influence of phylogenetic factors. The Scarabaeidae exhibited both the highest morphological diversity (Fig. 1) and species richness (Suppl. 1). The Diphyllostomatidae, a small group for which only 3 species were included in this study, was found to have the lowest relative morphological diversity (Fig. 1). However, the relationship between morphological diversity and species richness was not always consistent. The Lucanidae exhibited high species-richness values despite its fairly low morphological diversity (Fig. 1). The Hybosoridae, with a comparatively small number of test species (57), was found to have the second highest pronotum diversity.
In the test of 30 subfamilies (Fig. 1) (Suppl. 1), results similar to the family test were obtained. Morphological diversities of certain subfamilies were also found to lack a strong association with species richness. The Aesalinae were found to have the highest pronotum diversity, despite their low species richness (47) in our sample. Similarly, with 263 species included in our sample the, Melolonthinae was found to have fairly low elytron diversity.
In addition, morphological diversity was also found to vary with the character of features tested. The changing trends of characters and species richness values were not always highly associated (Fig. 1). For example, elytron diversity of the Lucanidae was low despite its high pronotum diversity. In the subfamily test, the Aesalinae were found to exhibit a high pronotum diversity, but a fairly low elytron diversity.
Moving on to the ecological factors (Fig. 1), phytophagous species were found to exhibit both the highest morphological diversities and guild richness to comprise more than 70.92% of all included species (Suppl. 1). The non-feeding group, containing two families (only 29 species and 2 genera), was found to exhibit the second highest pronotum diversity. However, as with the previous analysis, the relation between morphological diversity and feeding-guild richness was not always consistent. Coprophagous taxa exhibit high guild richness values despite their fairly low morphological diversity. Relative to be far more numerous phytophagous species, omnivorous species were found to exhibit the highest elytron diversity. At the same time, we found the changing trends of characters tested and species richness values also displayed inconsistent associations. The phytophagous group was found to have the highest pronotum diversity, however, its elytron diversity was not unusually high.
Relationship Between Morphological Diversity And Species Richness
Based on ordinary least squares (OLS) regression analysis, an ANOVA (F-test) was used to test the significance of the regression between morphological diversity and species richness quantitatively, and the equations of the best-fit lines estimated to verify the linear relations (Fig. 4). Test results varied both by groups and traits (Suppl. 3).
Three level of relations were revealed by the phylogenetic factor examined in this study. Firstly, the correlation between morphological diversity and species richness among sub-/families was found always to be consistent, with two exceptions: the genus level test of pronotum on subfamily (p-value = 0.188) and the species level test of elytron on the family category (p-value = 0.089), the regressions of which were non-significant statistically. These results reflect the influence of phylogeny on the mediation of morphological shape in these two broadly construed regions of the phenotype. Correlation of the family category with pronotum shape was significant statistically (p-value = 0.005 for species; p-value = 0.029 for genera) with the subfamily category test being marginally non-significant (p-value = 0.049 for species). With regard to the elytra, the comparative result was the opposite in all tests returning highly significant statistical results (p-value = 0.043 for genera in family category; 0.006 in species and 0.020 for genera in subfamilies).
Application of the non-parametric Spearman correlation coefficient to our test parameters revealed variation in both groups and characters. For the family test, the correlation between test parameters was found to be higher than it was for the subfamily both for the pronotum test (r = 0.5321/0.7335 for genus/species level test in family category and 0.1755/0.3397 for genus/species level test in subfamily category test respectively) and for the elytron test (r = 0.4587/0.5968 in genus/species level test in family category and 0.3898/0.5565 in subfamily category). At the same time, linear modeling in the species-level test was found to be lower than it in the genus-level test for both character complexes.
In addition, morphological diversity and known global species richness values for these same categories (11 families and 26 subfamilies include 32000 species and 2453 genera), were calculated. Based on a covariance analysis no significant difference was found in the slopes and intercepts of fitting linear models between the global and sampling (p > > 0.05) (Table 1), and the correlation between morphological diversity and species/genus richness was same as in the sampling test. This suggests our analysis represented the correlation between species/genus richness and morphological diversity objectively. For the family test, this correlation was found to be higher than it in the subfamily in the pronotum (r = 0.6800/0.7130 for genus/species level test in family category and 0.04260/0.2460 for genus/species level test in subfamily category test respectively) and in the elytron (r = 0.7130/0.7710 in genus/species level test in family category and 0.6200/0.4020 in subfamily category). The species-level test was found to exhibit a higher correlation than in the genus level test in family category, on the contrary the species-level test was found to exhibit a lower correlation than in the genus level test in subfamily category (r = 0.4020/0.2460 in species level test and 0.6200/0.4260 in genus level test).
However, morphological diversities were found to be insignificantly correlated with species richness in all feeding types based on linear regression analyses (ecological factor) (p > > 0.05) (Suppl. 3, Fig. 4).
Table 1
Covariance analysis results based on phylogenetic factors
Characters | Categories | Intercept of fit line | Slope of fit line |
Elytron | Family | Species level test | 0.4050 | 0.7710 |
Genus level test | 0.5760 | 0.7130 |
Subfamily | Species level test | 0.1260 | 0.4020 |
Genus level test | 0.3190 | 0.6200 |
Pronotum | Family | Species level test | 0.5760 | 0.7130 |
Genus level test | 0.5440 | 0.6800 |
Subfamily | Species level test | 0.4790 | 0.2460 |
Genus level test | 0.7370 | 0.4260 |