Fig 1.
Fore wing venation shape and landmark placement.
(A) Wing vein terminology used throughout the study, shown on an idealised ichneumonid fore wing. Vein 2m-cu marked orange; on this vein, a curve with six semilandmarks was added. (B) Placement of 21 fixed landmarks.
Table 1.
An overview of the datasets and analyses.
Fig 2.
(A) PCA plot with landmarks placed by two different persons and comparison of illustrations and photos. Green triangles show the data of the same species three times, twice from an illustration, measured by two different persons, and once from a photo. Grey triangles represent species from the same genus, one species measured by two persons from an illustration and a photo of a different species, in order to additionally show some of the intra-generic variation. In those cases where our taxon sampling allowed it (Dusona, Rhimphoctona, etc.), both aspects were combined. The genus ID in the figure is composed of the first three letters of the genus (Dia–Diadegma) and the abbreviations, as well as the included taxa are listed in S2 Table. (B) Data quality assessment in fossils. PCA showing shape variation within and among the five newly described fossil species, “m” is representing the part (main part) and “c” the counterpart of the fossils, whereas the black outline is marking the right fore wings, the ones without black outlines the left fore wings.
Fig 3.
Correlation of the centroid size and the wing shape in ichneumonid subfamilies.
(A) Regression analysis based on the raw data, where centroid size is plotted against the regression score, with linear regressions slopes shown by subfamily. (B) Centroid size against regression score pooled by subfamilies, with the centroid size distribution of each subfamily centred with its mean on zero. The black wings at the bottom of each plot represent shape at low and high centroid size and centroid size range, respectively. The underlying grey wing represents the mean shape of all taxa.
Table 2.
Result of the procrustes ANOVA of shape with centroid size and subfamily affiliation.
Fig 4.
Means of Procrustes distances between species pairs among and within subfamilies.
The means of the pairwise Procrustes distances for members of each subfamily pair are given in the left portion of the graph, sorted in ascending order. Whiskers indicate the standard error for each subfamily pair. On the right, the mean intra-subfamily pairwise distances are shown as barplots, with dashed lines extending to the left portion of the graph to allow direct comparison of mean intra- and inter-subfamily distances. The smallest distance between subfamiles is observed between Tryphoninae and Ctenopelmatinae and is lower than the three largest observed intra-subfamilial distances (Ophioninae, Cremastinae and Metopiinae). The largest distance is the one between Ophioninae and Tersilochinae and is about twice as high as the distances within Ophioninae. Horizontal lines and abbreviations of subfamilies are coloured according to the colour code above.
Fig 5.
Canonical variation analysis with seven extant subfamilies.
Examples of the shape at the maximum and minimum of the first four canonical axes. (A) Plot of the first two CV axes. Metopius Panzer, 1806 species are marked with a black contour. (B) Plot of the third and fourth CV axis. If subfamilies are more or less separate from others, we coloured the max/min shape of the wing accordingly. Grey wing outlines represent the general shape if several subfamilies overlap. The outlines on the wing illustrations are interpretative and oriented on the specimen that has the minimum/maximum CV score on the respective axis.
Table 3.
Cross-validated classification results in % from their CV scores.
Fig 6.
Fore wing mean shapes of subfamilies and fore wings of the five new fossil species.
Mean shapes of all nine subfamilies, where Metopiinae were split into two groups: the genus Metopius with a very distinct fore wing venation, and the rest of Metopiinae. The mean shape of areolets are represented open (eg. Tersilochinae) or closed (eg. Campopleginae) based on the majority of included species. Group specific venation shapes are coloured in the subfamily colour code when mentioned in the results or discussion. Names of the new fossil specimens are underlined with the colour of the respective subfamily they were placed in.
Fig 7.
PCA analyses of fore wing shape including extant and fossil taxa.
(A) PC1 against PC2 of seven subfamilies, including fossil Pimplinae and the newly described fossils. All 21 fixed and six semilandmarks were used, and the landmarks that were missing in the fossil specimens were estimated here. (B–F) PC1 and PC2 of the fossil species together with their candidate subfamilies. Landmarks of the reference subfamilies were adjusted to fit the respective fossil specimen (Table 1).
Table 4.
Cross-validated classification results and probabilities of the fossil assignment for classification.
Fig 8.
Holotype of Crusopimpla weltii sp. nov. (FUR #11550).
(A) Photo of the fossil part. (B) Interpretative drawing, part and counterpart were used as templates. Scale bar: 2mm.
Fig 9.
Holotype of Ebriosa flava gen. et sp. nov. (FUR #11112).
(A) Photo of the fossil part. (B) Interpretative drawing, fossil part and counterpart were used as templates. (C) Strongly bowed M&Rs vein of the left fore wing, which is the name giving character. (D) Fore wing of the paratype #13812. Scale bars: (B) 2mm, (C) 1mm.
Fig 10.
Holotype of Entypoma? duergari sp. nov. (FUR #11264).
(A) Photo of the fossil part. (B) Interpretative drawing, fossil part and counterpart were used as templates. Scale bar: (B) 2mm.
Fig 11.
Holotype of Lathrolestes? zlatorog sp. nov. (FUR #10651).
(A) Photo of the fossil part. (B) Interpretative drawing, fossil part and counterpart were used as templates. Scale bar: (B) 5mm.
Fig 12.
Holotype of Triclistus bibori sp. nov. (FUR #13809).
(A) Photo of the fossil part. (B) Interpretative drawing, fossil part and counterpart were used as templates. (C) Photo and drawn outlines of the fore tarsus, with the tarsomeres 2–4 very short and stout. (D) Head with visible groove between antennae. Scale bars: (B) 4mm, (C) 1mm, (D) 0.5mm.