Description of the Data and file structure

We proposed a new paradigm for evaluating the evolution of branches through changes in characters on continuous spatiotemporal scales, for better interpreting the impact of biotic/abiotic drivers on the evolutionary radiation. Four homologous traits of scarabs under various selection pressures were chosen: the mandible and hindwing are related to the feeding and migration of scarabs and have been corroborated to be associated with the development of diversity, and the pronotum and elytron, regarded as important parts of the scarabs' body, are closely related to the muscular system associated with digging and movement ability. It reveals a causal link between morphological changes and selective pressures: consistent deformation signals for all tested characters on timeline, which provided strong support for the evolutionary hypothesis of relationship between scarabs and biotic/abiotic drivers; the evolutionary strategies under niche differentiation, which were manifested in the responsiveness degree of functional morphological characters with different selection pressure.

-This dataset includes 9,332 specimens of 5,900 species for the testing of the pronotum and elytron, 250 specimens of 216 species for the mandible, and 263 specimens of 238 species for the hindwing (Supplementary Table 4). Most of the specimens were deposited in the Institute of Zoology, Chinese Academy of Sciences, and the Natural History Museum London. Additional photographs of species were taken from the literatures. The specimens were examined and dissected using a LEICA MZ 12.5 dissecting microscope, and all the photographs were taken using an Olympus EM5 (60mm) camera. Standard dorsal images were selected for this study. To facilitate accurate representation, images were only used when the testing characters were not covered or blurry, and the images possess adequate resolution (the smallest one was 90 pixels).

-Three curves were extracted and resampled into 50/25/50 equally spaced semi-landmarks (SLMs) from the left contours of the mandible/pronotum/elytron through MorphoJ (Version: 1.06a), for quantitative analysing the morphology, respectively (Supplementary Data 2). The first curve was taken from the outer contour of mandible covered by the base of the left to the base of the right, a silhouette of dorsal view of mandible was used to avoid the partial asymmetry of the left and right mandible in three-dimensional space which could lead to instability in the results; the second curve was collected from the middle of the anterior margin of the pronotum and end up at the middle of the posterior margin of the pronotum; the third curve started from the anterior margin of the left elytron and stopped at the end of elytron. 16 landmarks were taken from the right hindwing through MorphoJ for quantifying the structure of wing vein nodes (Supplementary Data 2), in order of numbering of landmark points: the base of ScA, the intersection of the RA3+4 vein with the leading edge; the end of RA3+4; the end of RA3; the end of RA4; the base of RA1+2; the base of MP; the base of RP; the end of RP; the end of MP; the base of CuA; the end of CuA; the base of AA; the end of AA; the base of AP; the end of AP83.
For the preprocessing of the mandible/pronotum/elytron dataset, all SLMs were digitized with tps-Dig (Version: 2.05). The format of data files used for morphological analysis were achieved by converting SLMs into LMs in text files for the subsequent analysis: the curve number and point number for each sample were deleted, then landmark numbers were replaced by point numbers.

-19 subfamilies from eight families of Scarabaeoidea were selected for reconstructing the ancestral feeding types of Scarabaeidae: 1) six subfamilies in Scarabaeidae were included in the inner group; 2) 13 subfamilies from seven families of Scarabaeoidea were included in the outgroups. Test groups were divided according to the main test members or the typical representative feeding types of the test groups: 1) the omnivory: Aphodiinae-Rhyparini; Geotrupidae, Glaresidae, Hybosoridae, Trogidae; 2) the coprophagy: Aphodiinae-Aphodiini, Scarabaeinae; 3) the phytophagy: Cetoniinae, Dynastinae, Glaphyridae, Lucanidae, Melolonthinae, Passalidae, Rutelinae5. A phylogenetic relationship of Scarabaeoidea was revised by the published tree of 89 genes, then the feeding types of ancestor nodes were reconstructed through the feeding types of living taxa in Mesquite (Version: 2.72)(Supplementary Table 2).

About the supplements:

[1] Supplementary Data 1: Deformation rate (DR) and sequential growth rate (SGR) of test characters;
Description of Supplementary Data 1:
This document contains 11 files, among which the files (No. 1-8) represent the original step graph of the DR of the test mandible/pronotum/elytron/hindwing, respectively (see Supplementary Table 3 for values); the file No. 9 represents the Illustration of the DR step graph of the test characters (the Illustration has been inserted into the article titled 'Evolutionary radiation strategy revealed in the Scarabaeidae with evidence of continuous spatiotemporal morphology and phylogenesis', which has been accepted on 2024-04-10); the file No. 10 represents the original curve of the SGR; the file No. 10 represents the Illustration of the SGR curve (see Supplementary Table 3 for values).
The files (No. 1-8 and No. 10) contain all of the information necessary to support not only the illustrations (the Figure 3 and Figure 4 in the article 'Evolutionary radiation strategy revealed in the Scarabaeidae with evidence of continuous spatiotemporal morphology and phylogenesis') but also the research findings.

[2] Supplementary Data 2: Morphological dataset of test characters;
Description of Supplementary Data 2:
Three curves were extracted and resampled into 50/25/50 equally spaced semi-landmarks (SLMs) from the left contours of the mandible/pronotum/elytron through MorphoJ (Version: 1.06a), for quantitative analysing the morphology, respectively. The first curve was taken from the outer contour of mandible covered by the base of the left to the base of the right, a silhouette of dorsal view of mandible was used to avoid the partial asymmetry of the left and right mandible in three-dimensional space which could lead to instability in the results (the specific data see the file 'SLMs in mandible test' in Supplementary Data 2); the second curve was collected from the middle of the anterior margin of the pronotum and end up at the middle of the posterior margin of the pronotum (the specific data see the file 'SLMs in pronotum test' in Supplementary Data 2); the third curve started from the anterior margin of the left elytron and stopped at the end of elytron (the specific data see the file 'SLMs in elytron test' in Supplementary Data 2). 16 landmarks were taken from the right hindwing through MorphoJ for quantifying the structure of wing vein nodes, in order of numbering of landmark points: the base of ScA, the intersection of the RA3+4 vein with the leading edge; the end of RA3+4; the end of RA3; the end of RA4; the base of RA1+2; the base of MP; the base of RP; the end of RP; the end of MP; the base of CuA; the end of CuA; the base of AA; the end of AA; the base of AP; the end of AP83 (the specific data see the file 'LMs in hindwing test' in Supplementary Data 2).
The order information (column variables) of all sampling data in the .tps file is consistent with the sample sequence information (column variables) of the corresponding test characters in Supplementary Table 4, which is convenient for readers to directly access the sample information from the Supplementary Table 4 and use it for subsequent grouping and analysis.

[3] Supplementary Data 3: Morphological dataset of ancestral nodes;
Description of Supplementary Data 3:
I quantified morphological changes across evolutionary nodes. The average shapes of the existing groups' test characters that were treated as the terminal taxa in the phylogenetic combined analysis were computed in MorphoJ. Then, the landmarks of the test characters were entered into Mesquite as a continuous matrix and linked to the topology of the phylogenetic tree. The ancestral forms of all nodes were reconstructed using the traces of all characters and the landmark drawings from the modules.

[4] Supplementary Table 1: Dispersion degree between test groups through confusion matrix analysis;
Description of Supplementary Table 1:
In Mathematica (Version: 12.1.0.0)97,100, the degree of dispersion between test groups was quantified based on confusion matrix analysis (CMA).

[5] Supplementary Table 2: Feeding types of ancestor nodes, phylogenetic topology map of Scarabaeoidea;
Description of Supplementary Table 2:
Reconstruction of the ancestral feeding types of Scarabaeidae: 19 subfamilies from eight families of Scarabaeoidea were selected for reconstructing the ancestral feeding types of Scarabaeidae: 1) six subfamilies in Scarabaeidae were included in the inner group; 2) 13 subfamilies from seven families of Scarabaeoidea were included in the outgroups. Test groups were divided according to the main test members or the typical representative feeding types of the test groups: 1) the omnivory: Aphodiinae-Rhyparini; Geotrupidae, Glaresidae, Hybosoridae, Trogidae; 2) the coprophagy: Aphodiinae-Aphodiini, Scarabaeinae; 3) the phytophagy: Cetoniinae, Dynastinae, Glaphyridae, Lucanidae, Melolonthinae, Passalidae, Rutelinae. A phylogenetic relationship of Scarabaeoidea was revised by the published tree of 89 genes, then the feeding types of ancestor nodes were reconstructed through the feeding types of living taxa in Mesquite (Version: 2.72).

[6] Supplementary Table 3: Mahalanobis/Euclidean distance of test characters, deformation rate of each lineage based on test characters (Mahalanobis/Euclidean distance), sequential growth rate of each lineage based on test characters (Mahalanobis/Euclidean distance);
Description of Supplementary Table 3:
-I quantified morphological changes across evolutionary nodes. The average shapes of the existing groups' test characters that were treated as the terminal taxa in the phylogenetic combined analysis were computed in MorphoJ. Then, the landmarks of the test characters were entered into Mesquite as a continuous matrix and linked to the topology of the phylogenetic tree. The ancestral forms of all nodes were reconstructed using the traces of all characters and the landmark drawings from the modules, and the Mahalanobis distance and Euclidean distance between each pair of test groups (including all the estimated ancestral nodes and terminal existing groups) were calculated based on canonical variate analysis (CVA) in MorphoJ and Mathematica, respectively (about the Mahalanobis distances and Euclidean distances, see the Sheet 1 and Sheet 2 of this file, please);
-I proposed two parameters to interpret the diversity mechanism of biological evolution at the spatiotemporal scale: 1) the deformation rate (DR) was obtained by dividing the Mahalanobis distance and Euclidean distance of characters to the mean differentiation time of mean time between nodes; the sequential growth rate (SGR) was obtained by dividing the difference value between each DR by the previous DR (about the DR of Mahalanobis distances and Euclidean distances, see the Sheet 3-6 of this file, please);

[7] Supplementary Table 4: List of samples.
Description of Supplementary Table 4:
This study was based on three datasets for increasing the morphological information obtained from each test taxa to be more representative , which included 250 specimens of 216 species for the mandible test (Supplementary Table 4-Sheet 1), 9,331 specimens of 6,403 species for the testing of the pronotum and elytron (Supplementary Table 4-Sheet 2 and Sheet 3), and 263 specimens of 255 species for the hindwing test (Supplementary Table 4-Sheet 4).