Ovipositor and mouthparts in a fossil insect support a novel ecological role for early orthopterans in 300 million years old forests.

A high portion of the earliest known insect fauna is composed of the so-called 'lobeattid insects', whose systematic affinities and role as foliage feeders remain debated. We investigated hundreds of samples of a new lobeattid species from the Xiaheyan locality using a combination of photographic techniques, including reflectance transforming imaging, geometric morphometrics, and biomechanics to document its morphology, and infer its phylogenetic position and ecological role. Ctenoptilus frequens sp. nov. possessed a sword-shaped ovipositor with valves interlocked by two ball-and-socket mechanisms, lacked jumping hind-legs, and certain wing venation features. This combination of characters unambiguously supports lobeattids as stem relatives of all living Orthoptera (crickets, grasshoppers, katydids). Given the herein presented and other remains, it follows that this group experienced an early diversification and, additionally, occurred in high individual numbers. The ovipositor shape indicates that ground was the preferred substrate for eggs. Visible mouthparts made it possible to assess the efficiency of the mandibular food uptake system in comparison to a wide array of extant species. The new species was likely omnivorous which explains the paucity of external damage on contemporaneous plant foliage.

Introduction

The earliest known insect fauna in the Pennsylvanian, ca. 307 million years ago, was composed by species displaying mixtures of inherited (plesiomorphic) and derived (apomorphic) conditions, such as the griffenflies (stem relatives of dragon- and damselflies), but also by highly specialized groups, such as the gracile and sap-feeding megasecopterans, belonging to the extinct taxon Rostropalaeoptera. A prominent portion of this fauna were the so-called ‘lobeattid insects’. They have been recovered from all major Pennsylvanian outcrops, where some species can abound (Béthoux, 2005c; Béthoux, 2008; Béthoux and Nel, 2005a). Indeed, at the Xiaheyan locality, China, for which quantitative data are available, they collectively account for more than half of all insect occurrences (Trümper et al., 2020). Additionally, another extinct group, the Cnemidolestodea, composed of derived relatives of lobeattid insects, was likewise ubiquitously distributed during the Pennsylvanian until the onset of the Permian (Béthoux, 2005b).

The phylogenetic affinities of lobeattid insects are debated. They have been regarded as stem relatives of either Orthoptera (crickets, grasshoppers, katydids; Béthoux and Nel, 2002; Béthoux and Nel, 2005a) or of several other lineages within the diverse Polyneoptera (Aristov, 2014; Rasnitsyn, 2007). A core point of the debate is the presumed wing venation ground pattern of insects, which, however, will remain elusive until Mississipian or even earlier fossil wings are discovered. Ecological preferences of lobeattid insects are also poorly known. Traditionally, they have been regarded as foliage feeders (Labandeira, 1998) but, given their abundance, this is in contrast to the paucity of documented external foliage damage during that time.

The Xiaheyan locality is unique in several respects (Trümper et al., 2020), including the amount of insect material it contains. Over the past decade, a collection of several thousand specimens was unearthed, allowing for highly detailed analyses of, for example, ovipositor and mouthparts morphology of extinct insect lineages (Pecharová et al., 2015b). These character systems are investigated herein in a new lobeattid species, based on hundreds of remains, using reflectance transforming imaging (RTI) together with more traditional approaches. Dietary preferences were inferred using a comparative morphometric and biomechanical analysis of gnathal edge shape based on an extensive dataset of extant polyneopteran species, with a focus on Orthoptera. Together, this investigation provides information regarding the phylogenetic affinities of loebattid insects and on their preferred mode of egg laying and dietary niche.

Archaeorthoptera Béthoux and Nel, 2002

Ctenoptilidae Aristov, 2014

Ctenoptilus Lameere, 1917

Ctenoptilus frequens Chen et al., 2020

LSID (Life Science Identifier). F0D67EC6-1C1A-4A8E-A8C0-31641FD057E3

Results

Systematic palaeontology

Etymology

Based on ‘frequens’ (‘frequent’ in Latin), referring to the abundance of the species at the Xiaheyan locality. Holotype. Specimen CNU-NX1-326 (female individual; Figure 1).

Figure 1.

Ctenoptilus frequens sp. nov., holotype (CNU-NX1-326).

(A) Habitus drawing and (B) habitus photograph (composite); (C–D) details of head and right foreleg (location as indicated in B), (C) color-coded interpretative drawing and (D) photograph (composite); and (E–F) details of ovipositor (location as indicated in B), (E) drawing and (F) photograph (composite). Color-coding and associated abbreviations: red, lacina (la); dark blue-purple, mandible (md); green, pharynx (pha). Other indications, head: ce, composite eye; f, frons; co, coronal cleavage line; fc, frontal cleavage line. Wing morphology abbreviations: LFW, left forewing; LHW, left hind wing; RFW, right forewing; RHW, right hind wing; ScP, posterior subcosta; RA, anterior radius; RP, posterior radius; M, media; CuA, anterior cubitus; CuPa, anterior branch of posterior cubitus; CuPb, posterior branch of posterior cubitus; AA, anterior analis. Photograph (composite). Color-coding and associated abbreviations: red, lacina (la); dark blue-purple, mandible (md); green, pharynx (pha). Other indications, head: ce, composite eye; f, frons; co, coronal cleavage line; fc, frontal cleavage line. Wing morphology abbreviations: LFW, left forewing; LHW, left hind wing; RFW, right forewing; RHW, right hind wing; ScP, posterior subcosta; RA, anterior radius; RP, posterior radius; M, media; CuA, anterior cubitus; CuPa, anterior branch of posterior cubitus; CuPb, posterior branch of posterior cubitus; AA, anterior analis.

Referred material. See Appendix 1, Section 2.1.2.

Locality and horizon

Xiaheyan Village, Zhongwei City, Yanghugou Formation (Ningxia Hui Autonomous Region, China); latest Bashkirian (latest Duckmantian) to middle Moscovian (Bolsovian), early Pennsylvanian (Trümper et al., 2020).

Differential diagnosis

The species is largely similar to Ctenoptilus elongatus (Brongniart, 1893), in particular in its wing venation (Appendix 1, Section 2.1.2). However, it differs from it in its smaller size (deduced from forewing length) and its prothorax longer than wide (as opposed to quadrangular).

General description. See Appendix 1, Section 2.1.2.

Specimens description

See Appendix 1, Section 2.1 and Appendix 1—figures 2–8; details of ovipositor, see Figure 2; details of head, see Figure 4.

Appendix 1—figure 2.

Ctenoptilus frequens sp. nov., specimens composed of fore- and hind wings in connection with body remains.

(A–B) Specimen CNU-NX1-752; habitus, left forewing as positive imprint and right forewing and hind wings as negative imprints, (A) drawing and (B) photograph (composite). (C–D) Specimen CNU-NX1-738; habitus, right hind wing as positive imprints and left forewing as negative imprints, (C) drawing and (D) photograph (composite; slightly shifted vertically with respect to drawing).

Appendix 1—figure 5.

Ctenoptilus frequens sp. nov., specimens composed of forewings, isolated or by pair, and forewing and ovipositor.

(A–B) Specimen CNU-NX1-748; right forewing, negative imprint, (A) drawing and (B) photograph (composite, flipped horizontally, light-mirrored). (C–D) Specimen CNU-NX1-732; right forewing, positive imprint, (C) drawing and (D) photograph (composite). (E–F) Specimen CNU-NX1-757; right forewing, negative imprint, (E) drawing and (F) photograph (composite, flipped horizontally, light-mirrored). (G–H) Specimen CNU-NX1-758; left forewing, negative imprint, (G) drawing and (H) photograph (composite). (I–J) Specimen CNU-NX1-744; right forewing, negative imprint, (I) drawing and (J) photograph (composite, flipped horizontally, light-mirrored). (K–L) Specimen CNU-NX1-751; forewing pair, both as negative imprints, and apical fragment of a hind wing, (K) drawing and (L) photograph (composite). (M–Q) Specimen CNU-NX1-743; (M–N) habitus, right forewing, positive imprint, (M) drawing and (N) photograph (composite); and (O–Q) details of ovipositor (location as indicated in N), polarity unknown, (O) drawing and (P–Q) photographs, (P) with color-coded interpretative drawing and (Q) without (composite, flipped horizontally).

Figure 2.

External ovipositor in Ctenoptilus frequens sp. nov. in lateral view.

(A–C) Specimen CNU-NX1-749, (A) overview of the ovipositor with overlaid indications of the ovipositor parts (see also) overview of the ovipositor with overlaid indications of the ovipositor parts (see also Appendix 1—figure 7A–C) and (B, C) details of basal part of the same ovipositor as in A. (B) composite photograph and (C) reflectance transforming imaging (RTI) extract in normals visualization; (D–F) specimen CNU-NX1-742, (D) overview of the ovipositor with overlaid indications of the ovipositor parts (see also) overview of the ovipositor with overlaid indications of the ovipositor parts (see also Appendix 1—figure 8B and C) and (E, F) details of basal part of the same ovipositor as in D; (E) composite photograph and (F) RTI extract in normals visualization. Olistheter (‘olis’) configurations at different parts of each respective ovipositor are shown as insets. Abbreviations: Gonostylus IX (gs9); gonapophysis IX (gp9); gonapophysis VIII (gp8).

Appendix 1—figure 7.

Ctenoptilus frequens sp. nov., specimens composed of body remains including well-preserved head, legs, and/or ovipositor.

(A–E) Specimen CNU-NX1-749; (A) photograph of habitus (composite), left forewing as positive imprint; (B–C) details of ovipositor (location as indicated in A; to be compared with main document Figure 2C), polarity unclear, (B) drawing and (C) photograph (composite); and (D–E) details of head (location as indicated in A), (D) color-coded interpretative drawing and (E) photograph (composite). (F–H) Specimen CNU-NX1-756; (F) photograph of habitus (composite); and (F–H) details of head (location as indicated in F), imprint polarity unclear, (G) color-coded interpretative drawing (F) photograph (composite). (I–K) Specimen CNU-NX1-754; (I) photograph of habitus (composite; frame delimiting head indicating the location of main document Figure 3A–B), (J–K) details of distal portions of fore-legs and a mid-leg (location as indicated in I), (J) drawing and (K) photograph (composite).

Appendix 1—figure 8.

Ctenoptilus frequens sp. nov., specimen CNU-NX1-742.

(A) Photograph of habitus (composite), right forewing as positive imprint, flipped horizontally, and (B–C) details of ovipositor (location indicated in A; to be compared with main document). Photograph of habitus (composite), right forewing as positive imprint, flipped horizontally, and (B–C) details of ovipositor (location indicated in A; to be compared with main document Figure 2D), (B) drawing and (C) photograph (light-mirrored).

Ovipositor morphology

The external genitalia in insects consist primarily of a pair of mesal extensions, the so-called gonopods, or ovipositor blades, and a pair of lateral projections, the so-called gonostyli, or ovipositor sheaths on abdominal segments 8 and 9. These sclerotized elements are collectively referred to as ‘valves’. The studied fossils possess three pairs of valves in their ovipositor, each strongly sclerotized (Figure 2, and Appendix 1—figure 7B, C, 8B and C). Especially the valve margins are still visible in the anterior area (‘base’), including the dorsal margin of the gonostylus IX (gs9), the ventral margin of the gonapophysis IX (gp9), and the dorsal and ventral margins of gonapophysis VIII (gp8). All observed ovipositors, but in particular the one of specimen CNU-NX1-742 (Figure 2D–F, and Appendix 1—figure 8B and C), display, from the second third of their length onwards, a thin longitudinal line much sharper and more developed than other visible linear structures in the area. This is the primary olistheter (olis1), a tongue-like structure which commonly interlocks gp9 and gp8 in extant insects having more or less well-developed external ovipositors (Figure 3; Klass, 2008). In the distal half of the ovipositor, the linear structure occurring between the dorsal edge of gs9 and olis1 is interpreted as the dorsal margin of gp9.

Figure 3.

The evolution of major ovipositor configurations across Orthoptera.

(A) External ovipositor of external ovipositor of Ceuthophilus sp. (Orthoptera: Rhaphidophoridae; extant species) in laterial view (left side, flipped horizontally, left gonostylus IX [gs9] removed). (B) Same as above, but annotated. The three black vertical lines labelled ‘a’, ‘b’, ‘c’ indicate the position of the three schematic sections shown in C. (C) Schematic ovipositor cross-sections in Grylloblattodea, Ctenoptilus frequens sp. nov., and several extant Orthoptera possessing well-developed ovipositors (not to scale; (see Appendix 1, Section 2.2). Ovipositor configurations are mapped onto the phylogenomic inference carried out by Song et al., 2020. Pale cross-section along the stem of Grylloidea is hypothetical; sections delineated by brackets represent conditions along the antero-posterior axis. Color-coding and associated abbreviations: light blue, gonostylus IX (gs9; light green, gonapophysis IX (gp9); red, gonapophysis VIII (gp8); royal blue, secondary olistheter (olis2); light orange, tertiary olistheter (olis3); purple, ‘lateral basivalvular sclerite’ (specific to Caelifera). Other indications: olis1, primary olistheter; int./ext., internal/external, respectively; dors./ventr., dorsal/ventral, respectively; and ant./post., anterior/posterior, respectively.

Together with the position of the antero-basal apophysis (=outgrowth) of this valve, the anterior margin of gp9 can then be traced. The extent of olis1 indicates that gp9 reaches the ovipositor apex, which is corroborated by the length of its inferred dorsal margin, well visible in specimen CNU-NX1-749 (Figure 2A, and Appendix 1—figure 7B and C). This specimen also shows that gp8 bears ventrally oriented teeth, more prominent and densely distributed near the apex, as in many extant orthopterans. The location of the dorsal margin of gp9 could not be observed with confidence near the base, which might be due to a lower degree of sclerotization.

This morphology implies that, at the base, dorsal to the anterior margin of gp8, only gs9 and gp9 occur. Therefore, the sharp and heavily sclerotized longitudinal line, located slightly dorsally with respect to the ventral margin of gs9, can only be an olistheter interlocking these two valves. This second olistheter (olis2) reaches olis1 but its development beyond this point could not be inferred with the available material. The occurrence of a mechanism locking gs9 onto gp9 is further supported by the fact that these valves remained connected to each other in the specimen CNU-NX1-742 even though it endured heavy decay (head and ovipositor detached from thorax and abdomen, respectively; Appendix 1—figure 8).

Mandibular mechanical advantage

The head and mouthpart morphology could be investigated in more detail in six specimens (see Appendix 1) while we could study the mechanical advantage (MA; see Section 1.5 of Appendix 1) of their mandibles in four of the six (viz. CNU-NX1-326, −747,–754, –764). The MA is defined as the inlever to outlever ratio and thus indicates the percentage of force transmitted to the food item (i.e. the effectivity of the lever system). Therefore, the MA allows for a size-independent comparison of the relative efficiencies of force transmission to the food item. Low MA values usually indicate quick biting with low force transmission typical for predators, while high MA values indicate comparatively slow biting with higher force transmission typical for non-predatory species.

Calculation of the MA along the entire gnathal edge revealed characteristic MA curve progressions for the studied taxa (Appendix 1, Section 2.3, and Appendix 1—figure 9). Compared to the studied fossils, extant Dermaptera, Embioptera, and Phasmatodea showed comparatively high MAs with an almost linear curve progression towards more distal parts of the mandibular incisivi whereas Plecoptera, Zoraptera, and Grylloblattodea were located at the lower end of the MA range with a gently exponential decrease towards the distal incisivi. The analysed extant Orthoptera occupy a comparatively wide functional space, with lineages at the higher and lower ends of the MA range. The composite fossil mandible representation (CFMR) of Ct. frequens (see Materials and methods) is located in the centre of the observed range of MAs for Orthoptera (Figure 4).

Appendix 1—figure 9.

Outlines of the mandibular gnathal edges for all studied taxa.

Figure 4.

Head morphology (A–D) in Ctenoptilus frequens sp. nov. and (E) mandibular mandibular mechanical advantage in Ct. frequens sp. nov. and a selection of polyneopteran species.

(A–B) Specimen CNU-NX1-754, (A) color-coded interpretative drawing, and (B) photograph (composite) (as located on Appendix 1—figure 7I); (C–D) Specimen CNU-NX1-764, (C) color-coded interpretative drawing, and (D) photograph (composite). (E) Principal component analysis of the mandibular mechanical advantage. Color-coding: (A–D) red, lacina (la); salmon, cardinal and stipital sclerites (ca and st, respectively); dark blue-purple, mandible (md); yellow, tentorium, including anterior tentorial arm (ata), posterior tentorial arm (pta), and corpotentorium (ct). Other indications: co, coronal cleavage line; fc, frontal cleavage line.

A polynomial function of the fifth order resulted in the best relative fit on the MA curves according to the Akaike information criterion (AIC) value (–661.3, see Materials and methods). The five common coefficients were subjected to a principal component analysis (PCA, Figure 4E), and, because phylogenetic signal was detected (K = 1.03316; p = 0.0001), also analysed using a phylogenetic principal component analysis (pPCA) (Appendix 1, Section 2.3, and Appendix 1—figure 10). The first four principal components (PCs) accounted for 96.8% (PCA)/96% (pPCA) of the variation in MA (Appendix 1—table 2).

Appendix 1—figure 10.

Progression of mechanical advantage curves for the studied taxa.

x-axis = % tooth row; y-axis = MA (mechanical advantage).

Appendix 1—table 2.

Importance and factor loadings of the principal component analyses of the polynomial regressions of the mechanical advantages (MAs).

			Principal component analysis
	PC1	PC2	PC3	PC4	PC5	PC6
Standard deviation	0.380	0.158	0.083	0.061	0.039	0.025
Proportion of variance	0.794	0.136	0.038	0.021	0.008	0.003
Cumulative proportion	0.794	0.930	0.968	0.988	0.997	1.000
Factor loadings:
	PC1	PC2	PC3	PC4	PC5	PC6
Intercept	−0.013	0.288	0.150	0.907	0.258	0.074
Regression coefficient 1	−0.948	−0.250	0.173	0.046	−0.048	0.065
Regression coefficient 2	0.300	−0.754	0.497	0.090	0.186	0.229
Regression coefficient 3	−−.037	0.509	0.528	−0.372	0.290	0.489
Regression coefficient 4	−0.057	−0.164	−0.650	−0.015	0.436	0.597
Regression coefficient 5	0.079	0.028	−0.010	0.171	−0.789	0.585
			Phylogenetic principal component analysis
	PC1	PC2	PC3	PC4	PC5	PC6
Standard deviation	0.745	0.368	0.172	0.131	0.099	0.055
Proportion of variance	0.740	0.180	0.040	0.023	0.013	0.004
Cumulative proportion	0.740	0.921	0.960	0.983	0.996	1.000
Factor loadings:
	PC1	PC2	PC3	PC4	PC5	PC6
Intercept	1.000	−0.021	−0.003	0.003	0.000	0.000
Regression coefficient 1	−0.088	−0.995	−0.048	0.009	0.005	0.004
Regression coefficient 2	0.054	−0.228	0.940	−0.247	−0.008	−0.024
Regression coefficient 3	0.076	0.026	−0.412	−0.905	0.067	−0.006
Regression coefficient 4	−0.023	0.067	0.096	0.101	0.985	0.078
Regression coefficient 5	0.077	0.085	0.180	−0.112	−0.241	0.940

In both PCAs, PC1 mainly codes for the vertical position of the MA curve, that is, the effectivity of the force transmission along the whole toothrow, while PC2 mainly codes for the curvature, that is, whether there is an almost linear or a gently exponential decrease in the effectivity of force transmission. Due to the narrow distribution of species along PC3, it was not possible to associate a clear biomechanical pattern to this PC.

The CFMR of Ct. frequens is located at the centre of the first three PCs (Figure 4E). Omnivorous Orthoptera and all herbivore taxa, with the exception of Apotrechus, are located along the width of PC1, while there is a tendency for the carnivorous taxa within the sampling to be spread along PC2.

Discussion

Phylogenetic implications

Our analysis of material of Ct. frequens provides unequivocal evidence that olis2 occurs in this species. Therefore, the new species was an orthopteran. The ovipositor configuration in Ct. elongatus furthermore conforms that observed in extant cave crickets (Raphidophoridae) in which olis2 occurs in addition to olis1 and interlocks gs9 and gp9 (Figure 3A–C; Appendix 1, Section 2.2). Indeed, this structure is present in ensiferan (‘sword-bearing’) Orthoptera possessing a developed ovipositor and is absent in caeliferan (‘chisel-bearing’) Orthoptera (Cappe de Baillon, 1920; Cappe de Baillon, 1922; Kluge, 2016; and see below). It follows that the new species is either more closely related to Ensifera than to Caelifera (owing to the possession of olis2), or it is a stem-orthopteran and olis2 was secondarily lost in Caelifera.

Further evidence for the phylogenetic placement of Ct. frequens is based on the lack of jump-related specializations in the hind-leg. Such specializations define the taxon Saltatoria within Orthoptera, and therefore Ct. frequens can be confidently excluded from crown-Orthoptera. This conclusion is furthermore corroborated by wing vein characteristics: Ct. frequens lacked a forked CuPa vein before its fusion with the CuA vein. Such a forked CuPa vein is typical for Panorthoptera, which includes crown-Orthoptera and their nearest stem relatives (Béthoux and Nel, 2002). Given this evidence, based on the configuration of several body parts, Ct. frequens, and its various Pennsylvanian relatives collectively referred to as ‘lobeattid insects’ are stem relatives of Orthoptera (Figure 3C). The absence of olis2 in Caelifera therefore is the consequence of a secondary loss.

Evolution of ovipositor morphology

Based only on extant species, the evolution of the external ovipositor in crown-Orthoptera was ambiguous due to the organizational diversity of its substructures (Cappe de Baillon, 1920; Cappe de Baillon, 1922; Kluge, 2016; Thompson, 1986; Walker, 1919; Appendix 1, Supplemental Text, Section 2.2). Comparison has traditionally been made between Grylloblattodea (rock-crawlers) and Orthoptera (Walker, 1919) even though the two groups are not closely related (Wipfler et al., 2019). In both groups the ovipositor displays an elongate gs9 and a ball-and-socket locking mechanism, the so-called primary olistheter (olis1), interlocking gp9 onto gp8 (Figure 2G). This olis1 occurs widely among insects (Klass, 2008). Orthoptera possess a variety of additional olistheters, including one interlocking gs9 onto gp9 (royal blue in Figures 2 and 3; olis2), commonly present in ensiferans possessing a well-developed ovipositor, as exemplified by Rhaphidophoridae (cave crickets; Figure 2E, igure 3A and B, and see sections labelled ‘a–c’ on Figure 3C), and Gryllacrididae (raspy and king crickets) and Anostostomatidae (king crickets) (Figure 2 G3C). The occurrence of an olis2 is diagnostic of ensiferan (‘sword-bearing’) Orthoptera (Kluge, 2016; and see below).

Even though it is unclear how far posteriorly olis2 extends in Ct. frequens, the asserted phylogenetic placement of this species provides new insights on the evolution of ovipositor interlocking mechanisms in Orthoptera (Figure 3). The one in Ct. frequens is best comparable to the one of Rhaphidophoridae, the main difference concerning the rachis (‘ball’ as in ‘ball-and-socket’), which is limited to a short protrusion in these insects, while the aulax (‘socket’ as in ‘ball-and-socket’) extends further posteriorly. In addition, gs9 extends more ventrally, concealing gp8 for some distance. Compared to Gryllacrididae the only notable difference in Ct. frequens is the ventral extension of gs9 in the former. In Anostostomatidae, the ventral margin of gs9 enters a socket in gp8, regarded as composing the premises of a third olistheter (olis3). The most parsimonious hypothesis is that this new structure ultimately replaces olis2 in Tettigoniidae and thereby allows a coupling of gs9 with gp8.

Grylloidea (true crickets) and Ct. frequens are separated by more severe morphological differences. A gp9 is not present in all Grylloidea and, if present, it occurs at the ovipositor base and is reduced compared to, for example, Rhaphidophoridae. Gs9 and gp8 are connected by an olistheter and we suggest that it might represent a variant of olis2, assuming a hypothetical case (shaded scheme in Figure 3C) in which olis2 interlocks gs9, gp9, and gp8 altogether. The reduction of gp9 would then mean that only olis2 connects gs9 and gp8. The alternative is a convergent acquisition of an olis3, as in Tettigoniidae.

Unlike other orthopterans displaying a well-developed external ovipositor, Caelifera use valves for digging a tunnel to accommodate their entire abdomen and, additionally, dig egg pods (Fedorov, 1927; Stauffer and Whitman, 1997; Uvarov, 1966). The shoving operation to move forward is accomplished by powerful, rhythmic, dorso-ventral openings and closings of two sets of valves (Thompson, 1986), gs9 and gp8+ gp9, the two latter ones being interlocked via olis1. Even though gp9 is often reduced, it plays an important role in the closing of the ovipositor via muscles attached to it (Thompson, 1986). Obviously, an olistheter interlocking gs9 and gp8 (i.e. olis2) would impede such movements. Given the ovipositor configuration and phylogenetic placement of Ct. frequens, it follows that the olis2 was lost in Caelifera, a likely consequence of their highly derived oviposition technique.

The evolutionary scenario resulting from our findings in Ct. frequens addresses a long-standing debate on the respective position of the two main lineages of Orthoptera, Ensifera and Caelifera. On the basis of early, fossil Saltatoria/Orthoptera displaying elongate ovipositors, palaeontologists already assumed that caeliferans derived from ensiferans (Sharov, 1968). However, the placement of the corresponding fossils remained contentious, leaving it possible that both, Ensifera and Caelifera, derived from an earlier, unspecialized assemblage (Ander, 1939). The discovery of an elongate ovipositor in the stem-orthopteran Ct. frequens provides a definitive demonstration that caeliferans derived from ensiferans. Because rock-crawlers can also be understood as possessing an elongate ovipositor, which would render the term ‘Ensifera’ ambiguous, it is proposed to coin a new taxon name, Neoclavifera, to encompass species bearing an olis2, that is, all extant orthopterans and their stem relatives as currently known (Figure 3C; Appendix 1, Section 2.1.1).

Another important input on the early evolution of orthopterans regards the abundance of lobeattids. Indeed, these insects are emerging as the main component of the Pennsylvanian insect fauna. They have been reported in high numbers from all major Pennsylvanian deposits (Béthoux, 2005c; Béthoux, 2008; Béthoux and Nel, 2005a; and Appendix 1, Section 2.1), such as Miamia bronsoni at Mazon Creek (Béthoux, 2008). At Xiaheyan, they collectively account for more than half of all insect occurrences (Trümper et al., 2020). Besides a high abundance, lobeattids and other stem-orthopterans compose a species-rich group at Xiaheyan, where they represent about a third of all insect species currently known to occur at this locality (Appendix 1, Section 3, taxon Archaeorthoptera). Orthoptera, which represent the bulk of extant polyneopteran insect diversity, therefore must have diversified early during their evolution.

Ovipositor shape and use

Extant Orthoptera resort to a wide diversity of substrates where to lay eggs, including ground, decaying leaves or wood, and stems or leaves of living plants (Cappe de Baillon, 1920; Cappe de Baillon, 1922; Ingrisch and Rentz, 2009; Rentz, 1991). This operation aims at ensuring a degree of moisture conditions suitable for eggs to fully develop, and providing protection, for example against predation. Ground is the preferred substrate of the majority of Orthoptera, including Caelifera (Agarwala, 1952; Stauffer and Whitman, 1997; Uvarov, 1966; and see above). Within this group, the epiphytic and endophytic habits of several, inner lineages represent derived conditions (Braker, 1989; Ramme, 1926). This habit translates into finely serrated ovipositor valves, including gs9.

As for ‘ensiferan’ Orthoptera, they generally possess a pointed and elongate ovipositor used to insert eggs in various substrates. In Grylloidea (including true crickets), females insert eggs in the ground using a needle-like ovipositor, or deposit them in subterranean chambers or burrows adults may inhabit, in which case the ovipositor is usually reduced (Cappe de Baillon, 1922; Loher and Dambach, 1989; Otte and Alexander, 1983). However, within Grylloidea, three groups, the Trigonidiinae (sword-tail crickets), the Aphonoidini, and the Oecanthinae (tree crickets), evolved oviposition in plants. In the former, which lay eggs in soft plant material, gs9 displays serration in its distal third, along its dorsal edge (Kim, 2013; Otte and Perez-Gelabert, 2009). In contrast, both Aphonoidini and Oecanthinae lay eggs in more robust plant material, translating into apices of gs9 provided with strongly sclerotized sets of teeth and hooks (Loher and Dambach, 1989). In Oecanthinae, in which oviposition functioning was studied in most detail, the alternate back and forth movements of gp8 induce apices of gs9 to alternately approximate and diverge (Dambach and Igelmund, 1983), and therefore act as a shoving tool.

The Rhaphidophoridae commonly lay eggs into the ground, or, alternatively, into rotten leaves or wood (Hubbell, 1936). In the latter case, the ovipositor is often curved. Interestingly, Ceuthophilus spp. use the ovipositor tip, somewhat truncated, to rake ground surface above oviposition holes (Hubbell and Norton, 1978), presumably to hide them. Anostostomatidae lay eggs in the ground or on walls of subterranean chambers (Monteith and Field, 2001; Stringer, 2001). These preferences also apply to both Gryllacrididae (Hale and Rentz, 2001; Morton and Rentz, 1983) and Stenopelmatidae (Davis, 1927; not represented in Figure 3C), in which the ovipositor, if well developed, is long, narrow, and rectilinear to curved (Cadena-Castañeda, 2019; Ingrisch, 2018).

Although most Tettigoniidae (katydids) lay eggs in the ground, a variety of plant tissues, including galls, are also targeted by members of this very diverse family (Cappe de Baillon, 1920; Gwynne, 2001; Rentz, 2010). As above, shape and serration relate, to a large extent, to the preferred substrate. A needle-shaped ovipositor generally indicates preference for ground, a sickle-shaped one for plant tissues. Curved ovipositors indicate preference for decaying wood, and more strongly falcate ones, which are usually also laterally flattened (as opposed to sub-cylindrical), preference for either bark crevices or leaf tissues. Katydids laying eggs in hollow grass stems or leaf sheaths possess straight to slightly falcate, flattened, and unarmed ovipositors. Marked serration on the dorsal side of the ovipositor indicates preference for plant tissues.

Given the relation of ovipositor shape and substrate in extant species, Ct. frequens, with its needle-shaped ovipositor including ventrally oriented teeth, likely oviposited in the ground (Figure 5). It is therefore unlikely that Pennsylvanian stem-orthopterans were responsible for endophytic oviposition traces documented for this epoch (Béthoux et al., 2004; Laaß and Hauschke, 2019). More likely candidates for these endophytic egg laying are the extinct Rostropalaeoptera (Béthoux et al., 2004; Pecharová et al., 2015a).

Figure 5.

Reconstruction of a female of Ctenoptilus frequens sp. nov. laying eggs.

Courtesy of Xiaoran Zuo.

Dietary preferences

Unlike in an extant tropical forest, a limited proportion of Pennsylvanian plant foliage experienced external damage, in particular generalized feeding types such as margin and hole feeding. Although such damages were reported from multiple localities, they are so rare that their occurrence was considered worth being reported (Correia et al., 2020; Iannuzzi and Labandeira, 2008; Laaß and Hauschke, 2019; Scott and Taylor, 1983). Quantitative data from Pennsylvanian localities indicate that generalized external damages were indeed rare, and concentrated on pteridosperms (‘seed ferns’; Donovan and Lucas, 2021; Xu et al., 2018). Such damages have been traditionally assigned to Orthoptera and their purported stem relatives (Labandeira, 1998). Indeed, investigation of mouthparts morphology in a subset of these insects suggested that, at least for the representatives belonging to the Panorthoptera/Saltatoria (Figure 3C), these insects were herbivores (Labandeira, 2019). However, there is an inconsistency between the paucity of damage on Pennsylvanian plant foliage on the one hand, and the abundance of lobeattid insects on the other. If these insects were all external foliage feeders, evidence of such damage would be more prevalent.

Given the reconstruction of the mandibular gnathal edge and its position in PC space in relation to other Orthoptera and Polyneoptera (Figure 4E; Appendix 1, Section 2.3), Ct. frequens was likely an omnivore species – not a solely herbivorous or carnivorous one. The new species is the second most common insect species at Xiaheyan, where it occurs in all fossiliferous layers at a rate of ca. 10%. This implies that a significant portion of Pennsylvanian neopteran insects were opportunistic, omnivorous species, which reconciles the paucity of foliage damage with the abundance of stem-Orthoptera.

Materials and methods

Fossil material

The studied specimens are housed at the Key Laboratory of Insect Evolution and Environmental Changes, College of Life Sciences, Capital Normal University, Beijing, China (CNU). All specimens were collected from the locality near Xiaheyan village, where insect carcasses deposited in an interdeltaic bay (Trümper et al., 2020).

The adopted morphological terminology is detailed in Appendix 1, Section 1.1. Documentation methodology is detailed in Appendix 1, Section 1.2.1. General habitus was investigated based on a selection of 23 specimens (including the holotype; Appendix 1, Section 2.1.2). Ovipositor morphology was investigated based on four specimens (Appendix 1, Section 1.2.2). Head and mouthparts morphology was investigated based on six specimens (Appendix, 1 Section 1.2.3).

To ensure an exhaustive documentation of ovipositor, head and mouthparts morphology, we also computed RTI files for details of several specimens. RTI files are interactive photographs in the sense that light orientation can be modified at will. The approach, originally developed in the field of archaeology (see Earl et al., 2010 and references therein), has also been applied to a variety of sub-planar fossil items (Béthoux et al., 2016; Hammer et al., 2002; Jäger et al., 2018; Klug et al., 2019; among others).

We computed RTI files based on sets of photographs obtained using a custom-made light dome as described elsewhere (Béthoux et al., 2016), driving a Canon EOS 5D Mark III digital camera coupled to a Canon MP-E 65 mm macro lens. Sets of photographs were optimized for focus using Adobe Photoshop CC 2015.5. RTI computing was then performed using the RTIbuilder software (Cultural Heritage Imaging, San Francisco, CA) using the HSH fitter (a black reflecting hemisphere placed next to the area of interested provided reference). Several snapshots were extracted using the RTIviewer software (Cultural Heritage Imaging, San Francisco, CA), including those in ‘normals visualization’ mode, which provides a color-coded image according to the direction of the normal at each pixel (i.e. the direction of the vector perpendicular to the tangent at each pixel; see Figure 2C and F). This allows to quantify subtle height differences in fossilized structures.

Comparative analyses

The phylogeny adopted for comparative analyses is based on the most comprehensive account to date (Song et al., 2020), which is largely consistent with previous analyses (Song et al., 2015; Zhou et al., 2017), except for the position of the Rhaphidophoridae, either regarded as sister group of the remaining Tettigoniidea or of a subset of it. The same applies to the Schizodactylidae (splay-footed crickets), which lack a developed ovipositor.

Fossil ovipositor morphology was compared to original material of extant species and to literature data (Appendix 1, Sections 1.3.1, 2.2). Multiple interpretations of the fossil ovipositor morphology were considered. Among these, the favoured interpretation is the only one consistent with observations made on all specimens.

The MA of the mandibles, that is, the inlever to outlever ratio, indicates the effectivity of force transmission from the muscles to the food item (Appendix 1—figure 1). Apart from force transmission, the MA can also indicate the dietary niche and feeding habits (Blanke, 2019; Sakamoto, 2010; Westneat, 2004). The MA was extracted from 43 extant polyneopteran species (Appendix 1—figure 9) including 31 orthopterans and one CFMR of the newly described fossil species (Appendix 1, Sections 1.3.2, 1.4, Appendix 1—table 1). The CFMR was derived from a Procrustes superimposition (R package ‘geomorph’ v.3.0.5; Adams et al., 2013) of four fossil specimens which showed low levels of overall distortion and a mandible orientation suitable for extraction of individual MAs (Appendix 1—figure 9). For comparison of species and inference of the dietary niche, a PCA and, due to the detection of significant phylogenetic signal, a pPCA (R package ‘phytools’ v.0.6–44; Revell, 2012) were performed (for results of the pPCA, see Appendix 1—figure 9, Appendix 1—table 2).

Appendix 1—figure 1.

Workflow for the extraction of the mandibular mechanical advantage based on 3D models.

Appendix 1—table 1.

Food preference of polyneopteran species included in the mechanical advantage (MA) principal component analyses.

Order	Species	Food preference
Dermaptera	Diplatys flavicollis	Omnivore
Dermaptera	Forficula auricularia	Omnivore
Dermaptera	Hemimerus sp.	Carnivore
Embioptera	Aposthonia japonica	Herbivore
Embioptera	Embia ramburi	Herbivore
Embioptera	Metoligotoma sp.	Herbivore
Grylloblattodea	Grylloblatta bifratrilecta	Omnivore
Orthoptera	Acheta domesticus	Omnivore
Orthoptera	Comicus calcaris	Omnivore
Orthoptera	Conocephalus sp.	Omnivore
Orthoptera	Myrmecophilus sp.	Omnivore
Orthoptera	Gryllus bimaculatus	Omnivore
Orthoptera	Cyphoderris sp.	Omnivore
Orthoptera	Hemideina crassidens	Omnivore
Orthoptera	Meconema meridionale	Carnivore
Orthoptera	Papuaistus sp.	Omnivore
Orthoptera	Prosopogryllacris sp.	Omnivore
Orthoptera	Stenobothrus lineatus	Herbivore
Orthoptera	Stenopelmatus sp.	Omnivore
Orthoptera	Pholidoptera griseoaptera	Omnivore
Orthoptera	Tettigonia viridissima	Omnivore
Orthoptera	Tridactylus sp.	Herbivore
Orthoptera	Troglophilus neglectus	Omnivore
Orthoptera	Xya variegata	Fetritivore
Phasmatodea	Agathemera sp.	Herbivore
Phasmatodea	Peruphasma schultei	Herbivore
Plecoptera	Eusthenia lacustris	Carnivore
Plecoptera	Perla marginata	Carnivore
Zoraptera	Zorotypus caudelli	Herbivore
Orthoptera	Ctenoptilus frequens CFMR	Tested
Orthoptera	Ametrosomus sp.	Omnivore
Orthoptera	Ametrus tibialis	Omnivore
Orthoptera	Apotrechus illawarra	Omnivore
Orthoptera	Bothriogryllacris brevicauda	Omnivore
Orthoptera	Chauliogryllacris grahami	Omnivore
Orthoptera	Cnemotettix bifascicatus	Omnivore
Orthoptera	Cooraboorama canberrae	Omnivore
Orthoptera	Kinemania ambulans	Omnivore
Orthoptera	Mooracra sp.	Omnivore
Orthoptera	Nullanullia maitlia	Omnivore
Orthoptera	Nunkeria brochis	Omnivore
Orthoptera	Paragryllacris combusta	Omnivore
Orthoptera	Pararemus sp.	Omnivore
Orthoptera	Wirritina brevipes	Omnivore

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

Using the novel Reflectance Transforming Imaging techniques combined with geometric morphometrics, Chen and colleagues investigated a new lobeattid insect from the Pennsylvanian (Carboniferous) of North China, and demonstrated that the new form possessed a particular mechanism interlocking its elongate ovipositor parts thereby allowed to lay eggs into the ground. The results shed new lights into the debates on Orthoptera, currently representing the bulk of polyneopteran insect diversity, and their early diversification.

Decision letter after peer review:

Thank you for resubmitting your work entitled "Ovipositor and mouthparts in a fossil insect support a novel ecological role for early orthopterans in Pennsylvanian forests" for consideration by eLife. Your article has been reviewed by 3 reviewers, including Min Zhu as the Reviewing Editor and Reviewer #1, and the evaluation has been by Patricia Wittkopp as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Joachim Huag (Reviewer #2).

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

1) The phylogenetic part should be strengthened to a larger extent, in order to clearly demonstrate the phylogenetic signals of relevant anatomical traits.

2) "Reflectance Transforming Imaging (RTI)" should be explained in the "Methods" section, as it is a new technique to show the anatomical details of insect fossils.

3) Although some explanation for the PCs are given in the text, the factor loadings for each PC should be provided in the supplement. Another file that should be made available, best in the supplement, would be all outlines of the investigated mandibles.

Reviewer #1:

Using the novel Reflectance Transforming Imaging techniques combined with geometric morphometrics, Chen and colleagues investigated a new lobeattid insect from the Pennsylvanian (Carboniferous) of North China. They demonstrated that the new form possessed a particular mechanism interlocking its elongate ovipositor parts thereby allowed to lay eggs into the ground. The main claims of the manuscript are supported by the data in overall, and the Reflectance Transforming Imaging techniques used here are informative for broader readers, in addition to entomologists and evolutionary biologists. However, the phylogenetic part of this manuscript should be strengthened to a larger extent, in order to clearly demonstrate the phylogenetic signals of relevant anatomical traits.

Line 37: add "living" in front of "Orthoptera".

Line 39: "abundance" to be clarified. Species and/or individual (specimen) abundance?

Line 48,52: "the earliest known insect faunas" vs "this past insect fauna"?

Line 59: "this abundant fraction of the earliest insect faunas" redundant, replaced by "lobeatid insects"?

Line 72: "Reflectance Transforming Imaging (RTI)" should be explained in the "Methods" section, as it is a new technique to show the anatomical details of insect fossils.

Line 75: "decisive" seems too strong. Anyway it is a statistical result.

Note genus and species names in italics (in several figures).

Line 225: add "crown-group or crown" in front of "Orthoptera".

Line 264-275: The cladogram in Figure 2G is a simplified diagram? Where is the source of this diagram? Without any phylogenetic analysis in this study, it is very difficult to discuss the phylogenetic impacts or signals of anatomical details revealed by the new species, considering the substantial parallelism in insect evolution. This paragraph should be largely revised to corroborate your statements.

Line 333: art credit?

Reviewer #2:

In this study, the authors provide very interesting fossils of early orthopteran insects. The images are of very high quality, both photos and drawings. These fossils provide new and very important insights into the early diversification of winged insects. Especially the comparison to extant representatives is well performed, using geometric morphometric methodology for inferring ecological function of the mouthparts of the fossils. With their data, the authors are able to draw conclusions about the phylogenetic position, the feedings habits as well as the oviposition habits of the species.

I have only few suggestions to improve the manuscript, mostly focussing on explaining some aspects for the wider readership and some other minor aspects as, for example, getting rid of ranked taxonomy and providing some more insights into the extant comparison.

Title: As eLife is a non-specialists journal, I suggest to add a number for the age of the Pennsylvanian (for me as a biologist this does also not immediately relate to something).

line 29: I do not immediately know what "lobeattid" means. Given the readership of the journal, this needs explanation.

line 49: not sure whether a semantic or philosophical issue, but to my understanding "faunas" are not populated.

line 50: suggest to add some broader groups to the named ones, e.g. "griffenflies" (Odonatoptera) and "megasecopterans" (Palaeodictyopteroidea), again to provide the not-too-deep specialists some frame.

line 56: nothing causes more useless discussion than ranks of groups. I therefore suggest to avoid ranks altogether (there is ample literature on this), in this case the term "order". I think this is in line with the style of the systematic palaeontology section which was also kept largely rankless: I also suggest to get rid of the ranks in the supplementary information.

line 62: in the age of growing creatonism and alike, I would always go for the more careful wording; "ground pattern" is less problematic than "groundplan". This also applies to other instances in the text.

line 70: again "order" could be substituted by a rankless expression such as "lineage".

line 103: the term "gonapophyses" indicates that these structures are of sternal origin. As most researchers instead agree that they are derived parts of appendages, they should be better attributed as "gonopods".

line 331: there were other candidates for endophytic egglaying (e.g. palaeodictyopteroideans), that could be mentioned here.

line 378 "recent" is no longer considered a valid time, "extant" would be the alternative.

Figure legends: Species names should be put in italics.

I would like to see some more of the supplementary figures in the main paper, but I guess that is due to length restrictions of the journal.

Although some explanation for the PCs are given in the text, the factor loadings for each PC should be provided in the supplement. Another file that should be made available, best in the supplement, would be all outlines of the investigated mandibles.

Reviewer #3:

Authors investigated large samples of Palaeozoic lobeattids, stem-relatives of all Orthoptera from the Xiaheyan locality using Reflectance Transforming Imaging combined with geometric morphometrics in order to assess lobeattid morphology, infer its ecological role, and phylogenetic position. The analysis of their ovipositor shape indicates that ground was the preferred substrate for eggs and mouthparts indicate omnivory, which explains the paucity of external damage on contemporaneous plants foliage. These very interesting and important results were achieved with use of Reflectance Transforming Imaging (RTI) files for details of several specimens and statistical tools. The results enabled to place analysed fossils in context of their palaeoceological role in the Pennsylvanian habitats but also in phylogenetic context, providing a definitive demonstration that caeliferans are derived from ensiferans. This gives the clue to the debate of evolutionary history and debate on Orthoptera, currently representing the bulk of polyneopteran insect diversity and their early diversification.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

1) The phylogenetic part should be strengthened to a larger extent, in order to clearly demonstrate the phylogenetic signals of relevant anatomical traits.

We appreciate the opportunity to be able to place the anatomical traits in a phylogenetic context in more detail in the main text. To address this issue, we deeply reworded the first paragraph of the section ‘Ovipositor morphology and phylogeny’ within the Discussion section and modified the Results section.

Our phylogenetic reasoning is essentially based on the occurrence of a secondary olistheter, which is uniquely present in extant Ensifera. It is therefore clear that the new fossil is either a stem- or a crown-orthopteran. At the same time, the fossil species can be clearly excluded from the taxon Saltatoria because it lacks jumping hind legs. This aspect is further corroborated by the lack of a wing venation feature present in all extant Orthoptera and some of their closest stem-relatives. It is therefore clear that the new fossil species is a stem-orthopteran.

The phylogenetic affinities of the new fossil species are assessed based on an unambiguous combination of inherited and derived character states. Arguably, resorting to taxon names ‘Saltatoria’ and ‘Panorthoptera’ without explanation on their relation with ‘Orthoptera’ made our reasoning confusing. We hope the proposed new wording clarified this aspect.

The item composing our new Figure 3 (formerly Figure 2G) now better illustrates our reasoning for the placement of the new species. The phylogeny is based on a recently published account (Song et al., 2020), which indeed was not referenced adequately (it originally appeared in the section 1.4.2 of the supplement). This is the most comprehensive account on the phylogeny of Orthoptera to date, and the outcome of this account is largely consistent with other recent studies, as far as the taxon subset we are focusing on is concerned.

2) "Reflectance Transforming Imaging (RTI)" should be explained in the "Methods" section, as it is a new technique to show the anatomical details of insect fossils.

We fully agree that this method needs more explanation in order to easen its more widespread use in the palaeontological community. We added more explanations on page 12 of the manuscript within the Methods section which reads as follows:

“RTI files are interactive photographs in the sense that light orientation can be modified at will. The approach, originally developed in the field of archaeology (see Earl et al., 2010 and references therein), has also been applied to a variety of sub-planar fossil items (Béthoux et al., 2016; Hammer et al., 2002; Jäger et al., 2018; Klug et al., 2019; among others).

We computed RTI files based on sets of photographs obtained using a custom-made light dome as described elsewhere (Béthoux et al., 2016), driving a Canon EOS 5D Mark III digital camera coupled to a Canon MP-E 65 mm macro lens. Sets of photographs were optimized for focus using Adobe Photoshop CC 2015.5. RTI computing was then performed using the RTIbuilder software (Cultural Heritage Imaging, San Francisco, CA, USA) using the HSH fitter (a black reflecting hemisphere placed next to the area of interested provided reference). Several snapshots were extracted using the RTIviewer software (Cultural Heritage Imaging, San Francisco, CA, USA), including those in ‘normals visualization’ mode, which provides a color-coded image according to the direction of the normal at each pixel (i.e. the direction of the vector perpendicular to the tangent at each pixel; see Figure 2C and F). This allows to quantify subtle height differences in fossilized structures.”

3) Although some explanation for the PCs are given in the text, the factor loadings for each PC should be provided in the supplement. Another file that should be made available, best in the supplement, would be all outlines of the investigated mandibles.

We fully agree and added the factor loadings in the supplement in table S3. We also provide a new figure S9 in the supplement showing the outlines of all investigated mandibles. In addition we provide a new file “MA_outlines.blend” within the Dryad repository, which can be opened with the open source software Blender (www.blender.org) in order to obtain all coordinates of the MA outlines for reanalysis of the dataset.

Reviewer #1:

Using the novel Reflectance Transforming Imaging techniques combined with geometric morphometrics, Chen and colleagues investigated a new lobeattid insect from the Pennsylvanian (Carboniferous) of North China. They demonstrated that the new form possessed a particular mechanism interlocking its elongate ovipositor parts thereby allowed to lay eggs into the ground. The main claims of the manuscript are supported by the data in overall, and the Reflectance Transforming Imaging techniques used here are informative for broader readers, in addition to entomologists and evolutionary biologists. However, the phylogenetic part of this manuscript should be strengthened to a larger extent, in order to clearly demonstrate the phylogenetic signals of relevant anatomical traits.

The main suggestion of this review (improvement of the phylogenetic inference) is addressed above (1st main point).

Line 37: add "living" in front of "Orthoptera".

Changed.

Line 39: "abundance" to be clarified. Species and/or individual (specimen) abundance?

We refer to high individual numbers in this case and clarified the text accordingly.

Line 48,52: "the earliest known insect faunas" vs "this past insect fauna"?

We deleted “past insect” and tried to increase readability of this paragraph by several other modifications.

Line 59: "this abundant fraction of the earliest insect faunas" redundant, replaced by "lobeatid insects"?

Changed accordingly.

Line 72: "Reflectance Transforming Imaging (RTI)" should be explained in the "Methods" section, as it is a new technique to show the anatomical details of insect fossils.

We applied modifications as requested, please also refer to the second main point of the editors mentioned above.

Line 75: "decisive" seems too strong. Anyway it is a statistical result.

Note genus and species names in italics (in several figures).

Changed accordingly.

Line 225: add "crown-group or crown" in front of "Orthoptera".

Changed accordingly.

Line 264-275: The cladogram in Figure 2G is a simplified diagram? Where is the source of this diagram? Without any phylogenetic analysis in this study, it is very difficult to discuss the phylogenetic impacts or signals of anatomical details revealed by the new species, considering the substantial parallelism in insect evolution. This paragraph should be largely revised to corroborate your statements.

We applied modifications as requested, please also refer to the first main point of the editors mentioned above. In short, the cladogram is based on the most recent transcriptome based study of Song et al., (2020) and we mapped the character states of the relevant anatomical details on this phylogeny. Larger modifications were applied in the text sections referring to our phylogenetic reasoning.

Line 333: art credit?

Proper credits is now given. We also provide a letter from the author stipulating that we can reproduce this artwork.

Reviewer #2:

In this study, the authors provide very interesting fossils of early orthopteran insects. The images are of very high quality, both photos and drawings. These fossils provide new and very important insights into the early diversification of winged insects. Especially the comparison to extant representatives is well performed, using geometric morphometric methodology for inferring ecological function of the mouthparts of the fossils. With their data, the authors are able to draw conclusions about the phylogenetic position, the feedings habits as well as the oviposition habits of the species.

We thank the reviewer for his/her encouraging comments.

I have only few suggestions to improve the manuscript, mostly focussing on explaining some aspects for the wider readership and some other minor aspects as, for example, getting rid of ranked taxonomy and providing some more insights into the extant comparison.

Title: As eLife is a non-specialists journal, I suggest to add a number for the age of the Pennsylvanian (for me as a biologist this does also not immediately relate to something)

We added an age in the title and hope this is in line with the journal style.

line 29: I do not immediately know what "lobeattid" means. Given the readership of the journal, this needs explanation.

The sentence was reworded so that the term ‘lobeattid’, arguably cryptic, appears later, in a position where is it better connected with the points being addressed.

line 49: not sure whether a semantic or philosophical issue, but to my understanding "faunas" are not populated.

We changed the wording to “composed”.

line 50: suggest to add some broader groups to the named ones, e.g. "griffenflies" (Odonatoptera) and "megasecopterans" (Palaeodictyopteroidea), again to provide the not-too-deep specialists some frame.

We made modifications accordingly.

line 56: nothing causes more useless discussion than ranks of groups. I therefore suggest to avoid ranks altogether (there is ample literature on this), in this case the term "order". I think this is in line with the style of the systematic palaeontology section which was also kept largely rankless: I also suggest to get rid of the ranks in the supplementary information

We fully agree and rephrased all occurrences accordingly.

line 62: in the age of growing creatonism and alike, I would always go for the more careful wording; "ground pattern" is less problematic than "groundplan". This also applies to other instances in the text.

We changed this accordingly at all occurrences.

line 70: again "order" could be substituted by a rankless expression such as "lineage".

Changed.

line 103: the term "gonapophyses" indicates that these structures are of sternal origin. As most researchers instead agree that they are derived parts of appendages, they should be better attributed as "gonopods".

Changed.

line 331: there were other candidates for endophytic egglaying (e.g. palaeodictyopteroideans), that could be mentioned here.

Agreed and changed accordingly.

line 378 "recent" is no longer considered a valid time, "extant" would be the alternative.

Changed.

Figure legends: Species names should be put in italics.

Changed.

I would like to see some more of the supplementary figures in the main paper, but I guess that is due to length restrictions of the journal.

Relocating some of the supplementary figures to the main text would entail a relocation of the corresponding descriptive sections, which we believe the readers of eLife will have limited interest for. Nevertheless, we now divided the original Figure 2 into two, with items E-G being parts of a new figure (now Figure 3). This allowed to add extracts of RTI files of the two fossilized ovipositors to the new version of Figure 2 (items C and F).

Although some explanation for the PCs are given in the text, the factor loadings for each PC should be provided in the supplement. Another file that should be made available, best in the supplement, would be all outlines of the investigated mandibles.

Agreed. We provide now the requested data in the supplement. See also our comment to the respective comment of the editor.

Reviewer #3:

Authors investigated large samples of Palaeozoic lobeattids, stem-relatives of all Orthoptera from the Xiaheyan locality using Reflectance Transforming Imaging combined with geometric morphometrics in order to assess lobeattid morphology, infer its ecological role, and phylogenetic position. The analysis of their ovipositor shape indicates that ground was the preferred substrate for eggs and mouthparts indicate omnivory, which explains the paucity of external damage on contemporaneous plants foliage. These very interesting and important results were achieved with use of Reflectance Transforming Imaging (RTI) files for details of several specimens and statistical tools. The results enabled to place analysed fossils in context of their palaeoceological role in the Pennsylvanian habitats but also in phylogenetic context, providing a definitive demonstration that caeliferans are derived from ensiferans. This gives the clue to the debate of evolutionary history and debate on Orthoptera, currently representing the bulk of polyneopteran insect diversity and their early diversification.

We thank the reviewer for his/her assessment of our study.