Table 1.
Fossil taxa examined.
Fig 1.
Wing vein and forewing cell nomenclature used in the current study.
Fig 2.
Illustration of the generation of a ‘RoguePlot’.
Panel (a) shows three different topologies (I–III) as they might be found during a Bayesian MCMC analysis or based on a set of bootstrapped matrices. The frequency with which each topology is encountered is shown above. Fossil X attaches to one of three different branches in each tree. Two of these branches are also present in the majority-rule consensus tree shown in panel (b), while the third branch, respectively, is in conflict between the trees and thus appears as a polytomy in the consensus. Fossil attachment frequencies to branches present in the consensus are reflected by the colour of the branches parallel to the direction of the tree, while those to branches not present are shown on the perpendicular branch that corresponds to its most recent common ancestor. In the current example, the percentages of Fossil X to attach to a branch in the crown group of genus C sums to 93%, while it is 7% for it being a stem-group representative and 0% to be with either of the other genera A and B. Classifying it as belonging to genus C is thus well-supported in this fictitious example. In addition to showing fossil placements, RoguePlots can be used to illustrate the attachment points of any taxon in a phylogeny, such as a species that was identified as a rogue.
Table 2.
Marginal likelihoods of the different models of morphological evolution as obtained under stepping-stone sampling.
Fig 3.
Majority-rule consensus tree of the topologies of the extant taxa sampled during the Bayesian analysis of a morphological dataset of 222 discrete characters. Values near nodes represent posterior probabilities of clades. Informal subfamily groups (Xoridiformes, Ophioniformes, etc.), the subfamily Pimplinae and the Polysphicta genus group (which includes parasitoids of spiders) are indicated by bars on the right. Tips are coloured according to their subfamily placement, and either subfamily or tribal (in the case of Pimplinae) classification is also indicated by the last three letters in the label.
Fig 4.
Partial RoguePlots of four fossil species.
The plots include all consensus tree branches reaching more than 1% attachment probability. Xoridinae as outgroup and in some cases subfamilies in which the fossil has been placed previously are included regardless of attachment probabilities. Branches parallel to the direction of the phylogram are coloured according to the probability of direct attachment of the fossil to the branch, while perpendicular branches represent the sum on any branch not in the consensus tree, of which the corresponding node is the most recent common ancestor. The fossil drawings represent our interpretation of fossil morphology, with dotted line indicating uncertain interpretations. a) Carinibus molestus, holotype USNM 580881. b) Crusopimpla rediviva, holotype #2156. c) Crusopimpla tadushiensis, holotype #PIN 3364/277. d) Dolichomitus? saxeus, holotype #PIN 3364/31.
Fig 5.
Partial RoguePlots of four fossil species.
The plots include all consensus tree branches reaching more than 1% attachment probability. Xoridinae as outgroup and in some cases subfamilies in which the fossil has been placed previously are included regardless of attachment probabilities. Branches parallel to the direction of the phylogram are coloured according to the probability of direct attachment of the fossil to the branch, while perpendicular branches represent the sum on any branch not in the consensus tree, of which the corresponding node is the most recent common ancestor. The fossil drawings represent our interpretation of fossil morphology, with dotted line indicating uncertain interpretations. a) Eclytus? lutatus, holotype PALE-1418. b) Eopimpla grandis, holotype USNM 66581. c) Glypta transversalis, modified from original drawing (Plate X, Fig 25 in [47]). d) Ichninsum appendicrassum, holotype UCM 39378.
Fig 6.
Partial RoguePlots of four fossil species.
The plots include all consensus tree branches reaching more than 1% attachment probability. Xoridinae as outgroup is included regardless of attachment probabilities. Branches parallel to the direction of the phylogram are coloured according to the probability of direct attachment to the branch, while perpendicular branches represent the sum on any branch not in the consensus tree, of which the corresponding node is the most recent common ancestor. The fossil drawings represent our interpretation of fossil morphology, with dotted line indicating uncertain interpretations. a) Lithotorus cressoni, holotype PALE-4652. b) Mesoclistus? yamataroti, holotype UCM 62725. c) Mesornatus markovici, holotype SF MeI 15245. d) Phygadeuon? petrifactellus, holotypeUSNM 66580.
Fig 7.
Partial RoguePlots of four fossil species.
The plots include all consensus tree branches reaching more than 1% attachment probability. Xoridinae as outgroup is included regardless of attachment probabilities. Branches parallel to the direction of the phylogram are coloured according to the probability of direct attachment to the branch, while perpendicular branches represent the sum on any branch not in the consensus tree, of which the corresponding node is the most recent common ancestor. The fossil drawings represent our interpretation of fossil morphology, with dotted line indicating uncertain interpretations. a) Pimpla? eocenica, holotype USNM 66582. b) Plectiscidea? lanhami, holotype UCM 19167. c) Polyhelictes bipolarus, holotype MeI 16069. d) Rhyssella vera, holotype MeI 8814.
Fig 8.
Partial RoguePlots of four fossil species.
The plots include all consensus tree branches reaching more than 1% attachment probability. Xoridinae as outgroup is included regardless of attachment probabilities. Branches parallel to the direction of the phylogram are coloured according to the probability of direct attachment to the branch, while perpendicular branches represent the sum on any branch not in the consensus tree, of which the corresponding node is the most recent common ancestor. The fossil drawings represent our interpretation of fossil morphology, with dotted line indicating uncertain interpretations. a) Scambus fossilobus, holotype SF MeI 13431. b) Scambus? mandibularis, holotype USNM 501474. c) Scambus? parachuti, holotype 565885. d) Tilgidopsis haesitans, holotype USNM 66931.
Fig 9.
Partial RoguePlots of four fossil species.
The plots include all consensus tree branches reaching more than 1% attachment probability. Xoridinae as outgroup is included regardless of attachment probabilities. Branches parallel to the direction of the phylogram are coloured according to the probability of direct attachment to the branch, while perpendicular branches represent the sum on any branch not in the consensus tree, of which the corresponding node is the most recent common ancestor. The fossil drawings represent our interpretation of fossil morphology, with dotted line indicating uncertain interpretations. a) Trigonator macrocheirus, holotype SF MeI 17304. b) Tryphon amasidis, holotype UCM 15690. c) Xanthopimpla messelensis, holotype SF MeI 16988. d) Xanthopimpla praeclara, holotype SF MeI 17300.