Geologic Context

The source of the plant assemblage discussed in this report is the CCP deposit, consisting of USNM locality numbers 41005, 41006, 41007, 42292, 42305, and 42306, in Foard County, north-central Texas (fig. 1). Each locality number represents a replicate collection from the same shale bed. The CCP collections are here treated as a single collection. The CCP deposit is located toward the northern end of the N 5°E-trending Clear Fork Group outcrop belt across north-central Texas (Hentz and Brown 1987; Chaney et al. 2009; Mamay et al. 2009). Within the Clear Fork Group, the fossiliferous bed occurs in the informal middle unit (Nelson et al. 2001, 2013). These strata are of Leonardian age, equivalent to the earliest interval of the Kungurian age (Wardlaw 2005), based on the general correlation and revised timescale of Gradstein et al. (2012; fig. 1). The fossiliferous bed at the CCP site consists of finely laminated reddish claystones infilling what was probably an abandoned oxbow channel (Chaney et al. 2009). Deposition took place on the eastern coastal plain of the Midland Basin (Hentz 1988).

The CCP Flora and Comparisons to Relevant Early Permian Floras

The four most abundant taxa in the CCP flora, in decreasing rank order (table 1), are the unaffiliated platysperm species seed (e.g., Sharov 1973), the conifer Walchia piniformis (Looy and Duijnstee 2013), the possible cycadophyte Taeniopteris spp. (e.g., Gillespie and Pfefferkorn 1986), and the peltasperm Auritifolia waggoneri (Chaney et al. 2009). Although we are considering Taeniopteris spp. as a single species, it likely represents multiple species at CCP, based on variable features such as the secondary vein branching angle from the midrib and degree of bifurcation of the secondary veins. Five numerically uncommon elements are, in decreasing rank order, an indeterminate broadleaf category that probably includes multiple species such as Rhachiphyllum and a mixoneuroid odontopteroid, the gigantopterid Evolsonia texana (Mamay 1989), the peltasperms Sandrewia texana (Mamay 1975) and Supaia thinnfeldioides (White 1929), and the probable cycadophyte Taeniopteris sp. nov. The two rare taxa are Callipteris sp. 1 and the sphenopsid Sphenophyllum thonii. Indeterminate axes are common and unassignable to laminar foliage. Overall, the CCP flora is dominated by peltasperms, represented by the four genera Auritifolia, Supaia, Callipteris, and Sandrewia and possibly the gigantopterid Evolsonia.

The diverse spectrum of early Kungurian peltasperm taxa, prominent in the CCP flora of the lower Clear Form Group, contrasts significantly with the older (Artinskian) Taint flora of the Waggoner Ranch Formation, dominated by Taeniopteris sp., an indeterminate broad-leaved seed plant, and a diverse spectrum of other plants including the peltasperm Comia and the gigantopterids Zeilleropteris and the Gigantopteridium-Cathaysiopteris species complex (Beck and Labandeira 1998; fig. 1). The CCP flora also contrasts with the even older (Sakmarian) CBB of the Nocona Formation, dominated by the peltasperm Autunia sp. cf. Autunia conferta and, subordinately, the cordaite Cordaites sp., the conifer W. piniformis, and the probable noeggerathialean Russellites taeniata (Labandeira and Allen 2007).

Euramerican Terrestrial Arthropods of the Mid-Early Permian

Very few insect or mite fossils have been recovered from any of the numerous Early Permian redbed deposits, and CCP is no exception. Nevertheless, the Elmo insect fauna from the Wellington Formation, of early Artinskian age, in central Kansas provides a well-preserved and diverse assemblage of insects consisting of 152 described species (Lubkin and Engel 2005; Beckemeyer and Hall 2007). The relevance of this fauna to the insect damage at the Taint locality has been established (Beck and Labandeira 1998). Approximately 60 herbivorous species are represented in the Elmo insect fauna, and it is likely that their successor taxa were culprits for particular types of insect damage at the somewhat younger CCP locality.

Qualitative Analyses

The initial recognition of herbivory at CCP in the laboratory is based on several explicit criteria (Labandeira 1998; Scott and Titchener 1999; Labandeira et al. 2002). Due to the Paleozoic age of the CCP flora, lower-level taxonomic attribution to extant herbivorous taxa was rarely possible, such as is possible for floras from the late Mesozoic to Recent (Opler 1973; Waggoner and Poteet 1996; Wilf et al. 2000). This rarity of taxonomic uniformitarianism (Dodd and Stanton 1990) typically necessitated use of intrinsic attributes of the flora related to general structural, behavioral, and ecological knowledge of insect feeding patterns on modern vegetation (Labandeira 2002). Most important are insect modification of plant tissues during feeding, such as those of mandibulate external feeders (Boys 1989; Chapman and Joern 1990) and the various responses of host plants to insect-induced damage (Johnson and Lyon 1991; Tovar et al. 1995).

In addition to the rare cases of taxonomic uniformitarianism in Paleozoic floras, relevant features that indicate consumption of live tissue can be grouped into four criteria. First is the presence of thick, raised rims of reaction tissue, such as parenchymatous callus, that are produced by the host plant as a response to consumption or other modification by insect herbivores (Meyer 1987). Second is the presence of atypical and specific histological features associated with the external chewing or internal consumption of foliage. Examples of such features on foliage include veinal stringers or necrotized tissue flaps left by the inability of insect mouthparts to ingest completely physically resistant vascular tissue (Keen 1952; Weintraub et al. 1994; Araya et al. 2000). For example, the presence of contiguous cuspules formed along a cut leaf margin within larger cuspate excisions can reveal the trajectory of insect head movement and mouthpart action during feeding on live tissues (Gangwere 1966; Kazakova 1985). A third line of evidence is the recurring stereotypy of tissue removal patterns, based on shape, size, and juxtaposition of those areas with consumed tissue, as well as vein avoidance and the repeated occurrence of herbivory at particular leaf regions (Bodnaryk 1992; Heron 2003). Fourth is the preferential presence of plant damage on particular host plant tissues, organs, and species, indicating herbivore targeting rather than the more random patterns that would be expected from physical breakage (MacKerron 1976; Wilson 1980; Vincent 1990) or detritivory (Brues 1924; Mitter et al. 1988).

After the distinction of herbivory from detritivory or physical battering was established, the spectrum of insect-mediated damage was qualitatively categorized into eight major types of insect folivory (Labandeira and Allen 2007). This was followed by a categorization of the insect damage into a system of explicitly defined DTs (Wilf and Labandeira 1999; Labandeira et al. 2002, 2007; Labandeira 2006a; Blois et al. 2013). Additionally, evidence of fungal modification of herbivorized tissues was noted.

Quantitative Analyses

First, we calculated the proportion of foliar elements that displayed herbivory, expressed as a percentage of the total number of examined foliar elements (Schmidt and Zotz 2000). Second, we calculated the herbivory index, expressed as the percentage of surface area of foliar elements removed by herbivores as a fraction of the total surface area (Landsberg 1989; Williams and Abbott 1991; O’Neal et al. 2002; Bradshaw et al. 2007). Both indices were expressed as values for each measured host plant species as well as for the bulk flora (for a modern example, see Mazia et al. 2004).

For image capture of foliar surface areas, each specimen was initially digitized under incandescent light with a Canon PowerShot G7 camera with a 1∶2.8 to 1∶4.8 7.4–44.4-mm-wide zoom lens or a Canon EOS 50D camera with a Canon EF-S 60 mm f/2.8 macrolens, using when necessary a MT-24EX twin strobe flash. After image capture, each foliar element and its insect-damaged zones, if any, were digitized using Adobe Photoshop Elements 10. A conservative approach was taken, as in previous studies (Adami-Rodrigues et al. 2004a), whereby foliar element silhouettes had to be justified anatomically or estimated based on previous knowledge of leaf shape outline and morphological variability (Adami-Rodrigues et al. 2004a, their fig. 3-3). Areas encompassed by total foliar and insect-herbivorized outlines were calculated in Media Cybernetics Image Pro Plus. For detailed resolution of insect herbivory damage to foliar surfaces, specimens were photographed under higher magnification using an Olympus SZX12 microscope with an Olympus DP25 camera. Images were white balanced, focused, and captured using Olympus DP2-BSW imaging software.

General Patterns of Herbivory

There are 52 distinctive DTs in the CCP deposit, organized into the following FFGs: external foliage feeding, with three subgroups of (1) margin feeding, (2) hole feeding, and (3) surface feeding, and (4) piercing and sucking, (5) oviposition, (6) galling, (7) seed predation, and (8) wood boring. Four of these FFGs—the three of external foliage feeding and piercing and sucking—exophytically target plants, as they are made by insect culprits positioned outside of the targeted plant organ while the insect’s mouthparts are engaged in consuming tissues (Labandeira 1997). The remaining four—oviposition, galling, seed predation, and wood boring—are endophytic, wherein the insect body, in most cases a nymph or a larva, occurs within the tissue being consumed (Coulson and Witter 1984). Fungal damage (Sinclair et al. 1987), frequently difficult to characterize, can occur on leaf surfaces, as epiphyllous fungi, or within leaf tissues (Labandeira and Prevec 2014), typically associated with infection of preexisting insect damage such as galls (Constantino et al. 2009).

Major examples of insect and perhaps mite herbivore damage encountered at CCP are documented in figures 2–15. This damage is presented as three themes. First, damage is displayed on the various plant hosts. Figures 2–5 illustrate herbivory on Taeniopteris sp.; figure 6 documents herbivory on Evolsonia texana; figure 7 includes herbivory on the four seed plants, a mixoneuroid odontopteroid, a possible Rhachiphyllum sp., Sandrewia texana, and Walchia piniformis; figures 8–12 consist of herbivory on Auritifolia waggoneri; figure 13 illustrates a variety of seed predation; figure 14 documents the herbivore community structure on the two most herbivorized plants, A. waggoneri (top) and Taeniopteris sp. (bottom); and figure 15 shows scanning electron micrographs of submicroscopic damage. Within each plant-host sequence, there is a second theme that focuses on a particular class of damage, ranging from external foliage feeding at the beginning of the series of figures, such as margin and surface feeding, to the internal foliage consumption in piecing and sucking, and ending in galling. A third theme illustrates three to seven particular specimens with insect damage for each figure, which are separated from other such figures by bold black borders. Within each leaf specimen ensemble are magnified areas that display enlargements of selected damage within the leaf, separated from each other by thin black lines.

External Foliage Feeding: Margin Feeding

Definition. Margin feeding is the consumption of laminate foliage along edges, such that there is histological evidence for the removal of all tissues of the leaf blade. The consumption of foliar tissues can be divided into single excisions of leaf regions such as the leaf margins some distance from (DT12 if shallow or DT15 if trenched) or adjacent to (DT14) a primary vein or the removal of the distal apical tip (DT13).

DTs represented. DT12, DT13, DT14, and DT15.

USNM figured specimens. CCP (USNM loc. 41006, 41007, 42292, 42305, 42306). For DT12: figure 2A–2C (USNM-559816); figure 2D, 2E (USNM-559817); figure 2F–2H (USNM-559818); figure 3A (USNM-559821); figure 5A (USNM-539375); figure 5L (USNM-559832); figure 5S (USNM-559834); figure 14K (USNM-559852), figure 14M (USNM-539375), figure 14N (USNM-559837), and figure 15D (USNM-530923). For DT13: figure 2A, 2B (USNM-559816), and figure 5A (USNM-539375). For DT14: figure 2D (USNM-559817), figure 3A (USNM-559821), and figure 5L (USNM-559832). For DT15: figure 2D, 2E (USNM-559817); figure 2F, 2H (USNM-559818); figure 3A (USNM-559821); figure 5A (USNM-539375); figure 5L (USNM-559832); and figure 8A (USNM-559840).

Plant hosts. Cycadales: Taeniopteris spp. (DT12, DT13, DT14, DT15); indeterminate: indeterminate broadleaf (DT12); ?peltaspermales: A. waggoneri (DT12, DT14, DT15); and Gigantopteridales: E. texana (DT12, DT15).

Remarks. The spectrum of margin feeding at CCP is unexceptional. The DT spectrum is typical of a generalized feeding syndrome and lacks stereotyped patterns of damage that would indicate more specialized feeding modes. All margin-feeding damage at CCP is consistent with Mesozoic to modern external foliage feeding by mandibulate insects (Coulson and Witter 1984; Johnson and Lyon 1991; Tovar et al. 1995; Labandeira et al. 2007).

Distribution in Paleozoic habitats. Margin feeding is one of the earliest compression-based FFGs in the fossil record and becomes common during the Paleozoic compared to almost all other feeding modes. The earliest-known occurrence of margin feeding is DT13, where it co-occurs with limited hole feeding on the liverwort Metzgeriothallus sharonae, from the latest Middle Devonian of New York State (Labandeira et al. 2014). The next earliest-known occurrence of margin feeding is DT12 on Triphyllopteris austrina, an early seed plant referable to the Lyginopteridaceae, from the Serpukhovian of Australia (Iannuzzi and Labandeira 2008). Although initially rare, margin feeding, especially DT12, commonly occurs on a range of seed plant taxa throughout the later Moscovian to the latest Permian. In the Late Carboniferous of Euramerica, DT12 occurs sporadically on Macroneuropteris scheuchzeri and other medullosan foliage (Müller 1982; Scott and Taylor 1983; Castro 1997; Jarzembowski 2012). Later and throughout the Permian, a range of margin-feeding DTs occurs predominantly on ?peltasperm, taeniopteroid, and gigantopterid foliage (Geyer and Kelber 1987; Beck and Labandeira 1998; Labandeira 1998, 2006b). During the later Permian, margin feeding, particularly DT12 but also DT13, DT14, and DT15, is found in Gondwana almost exclusively on Glossopteris and other common glossopterid foliage morphotypes (McLoughlin 1994a, 1994b; Prevec et al. 2009; Cariglino 2011; Srivastava and Agnihotri 2011; Slater et al. 2012). In Cathaysia, where there is a dearth of studies, margin feeding has been documented sparingly on gigantopterid specimens (Glasspool et al. 2003).

External Foliage Feeding: Hole Feeding

Definition. Hole feeding is the consumption of the entire thickness of a leaf blade that is circumferentially enveloped by unaltered leaf tissue. Hole-feeding DTs are defined by a combination of hole size and shape, whereas damage from more complex modes of hole feeding is defined by shape.

DTs represented. DT1, DT2, DT3, DT4, DT5, and DT8.

USNM figured specimens. CCP (USNM loc. 41005, 41007, 42292, 42305, 42306). For DT1: figure 6B (USNM-559836). For DT2: figure 3A (USNM-559821) and figure 5L, 5O (USNM-559832). For DT3: figure 3Q (USNM-559825), figure 6A (USNM-559835), and figure 6B (USNM-559836). For DT4: figure 5E, 5F (USNM-559830); figure 7A, 7C (USNM-530921); and figure 14K (USNM-559852). For DT8: figure 2D (USNM-559817) and figure 2F, 2H (USNM-559818).

Plant hosts. Cycadophyta: Taeniopteris spp. (DT1, DT2, DT3, DT4, DT7); ?peltaspermales: A. waggoneri (DT2, DT3, DT4, DT5, DT7); Gigantopteridales: E. texana (DT1, DT2, DT3, DT4); and indeterminate: unaffiliated broadleaf (DT4).

Remarks. Like margin feeding, the range of hole feeding at CCP presents a standard repertoire of complete tissue excision away from leaf margins. This pattern of damage usually indicates a generalized feeding mode typically inconsistent with herbivore specialization on particular plant-host taxa (Currano et al. 2010). All hole-feeding damage at CCP likely occurred, as in modern plant-insect associations, by external foliage-feeding mandibulate insects (Coulson and Witter 1984; Johnson and Lyon 1991; Tovar et al. 1995).

Distribution in Paleozoic habitats. Hole feeding appears on the same hosts as margin feeding in the early compression fossil record of plants, in particular, DT2 on the liverwort M. sharonae from the latest Middle Devonian of New York State (Labandeira et al. 2014). However, throughout the Paleozoic, hole feeding is significantly less common than margin feeding. Hole feeding increases suddenly during the Late Carboniferous, principally on medullosan foliage different from those exhibiting margin feeding, suggesting colonization by a different assemblage of insect herbivores from those responsible for margin feeding. The greatest occurrence of hole feeding is on the mid–Early Permian gigantopterid taxa of Gigantopteridium, Cathaysiopteris, and Zeilleropteris at the Taint locality, in north-central Texas (Beck and Labandeira 1998), where it occurs with extensive margin feeding and surface feeding on the same leaves. Hole feeding on Permian foliage is skewed toward smaller-size holes, indicating that exophytic orthopteroid herbivores of the hole-feeding guild were more diminutive than their Pennsylvanian forbearers. Studies of modern insects based on head-capsule widths (Calvo and Molina 2008) and mandible size (Hochuli 2001) on a range of modern ectophytic herbivores indicate a positive relationship between insect body size and feeding hole size, although there are other factors that also affect hole size (Dalin and Björkman 2003).

External Foliage Feeding: Surface Feeding

Definition. Surface feeding consists of abrasion of surface tissues and the stripping of one or more tissue layers, such as epidermis or hypodermis, from the leaf surface without removal of the entire leaf blade.

DTs represented. DT25, DT27, DT29, DT30, DT31, DT97, DT103, DT130, and DT263.

USNM figured specimens. CCP (USNM loc. 41005, 41006, 42292, 42305, 42306). For DT25: figure 3A (USNM-559821). For DT27: figure 2A, 2B (USNM-559816). For DT29: figure 5L (USNM-559832); figure 6H, 6J (USNM-559837); figure 6K, 6P (USNM-538922); and figures 11K, 11M, 14E (USNM-559853). For DT30: figure 5L (USNM-559832); figure 6B, 6D–6F (USNM-559836); and figure 7A, 7B (USNM-530921). For DT31: figure 6B, 6C, 6G (USNM-559836), and figure 7A, 7C (USNM-530921). For DT97: figure 5L (USNM-559832). For DT103: figure 14K (USNM-559852). For DT130: figures 8A, 8B, 14F (USNM-559840). For DT263: figure 2F, 2G (USNM-559817); figure 2I–2K (USNM-559819); and figures 2L, 2M, and 14H (USNM-559820).

Plant hosts. Cycadophyta: Taeniopteris spp. (DT25, DT27, DT29, DT30, DT31, DT97, DT103, DT263); indeterminate: unaffiliated broadleaf (DT29, DT97); peltaspermales: A. waggoneri (DT29, DT30, DT31, DT130); and Gigantopteridales: E. texana (DT27, DT29, DT30, DT31, DT103).

Remarks. The pattern of surface feeding combines elements that would be expected for generalized herbivory (DT25, DT27, DT29–DT31) and more specialized patterns of herbivory (DT97, DT103, DT130). Modern examples of damage associated with DT103 and DT130 frequently are made by host-specific surface feeders that abrade the upper layers of tissues, particularly certain polyphagan beetle clades (Lin et al. 1990; Heron 2003), which originated during the later Mesozoic.

An intriguing interaction is new DT263, which represents a distinctively elongate, interrupted patch of surface abrasion on a Taeniopteris sp. leaf surface located medially between the leaf margin and the midrib (fig. 2F, 2G, 2I–2M). The leaf damage of this interaction is identical to that of a leaf folder or leaf roller in which an herbivorous nymph, larva, or adult insect rolls or folds a leaf around its body to secure protection from predators or parasitoids while accessing a reliable source of nutrition from the epidermal and subjacent layers of the enclosed leaf (Frost 1959). Leaf folding and leaf rolling typically require specialized features such as labial silk glands for tying leaves and a distinctive mouthpart structure for abrading surface plant tissue (Frost 1959; Fritz and Nobel 2008), and foliar enclosure enhances survivability in leaf-rolling compared to non-leaf-rolling insects (Fukui et al. 2002). Evidence for leaf folding or leaf rolling has not been demonstrated for any Paleozoic or preangiospermous Mesozoic flora. If confirmed, this interaction would be the earliest occurrence for leaf rolling or folding, a pattern that has been documented in some cases in Cenozoic surface feeding and skeletonization associations associated with DT19, DT22, DT27, and possibly DT130 (Labandeira et al. 2007).

New damage type DT263. Surface feeding consisting of linear arrays of polylobately abraded tissue parallel to primary vein and leaf margin, commonly occurring in linear clusters, 2–6 mm long and up to 2 cm wide, texturally distinct from surrounding foliage tissue (fig. 2F, 2G, USNM-559817; fig. 2I–2J, USNM-559819; and fig. 2L, 2M, USNM-559820).

Distribution in Paleozoic habitats. Surface feeding, typically consisting of DT29, DT30, and DT31, is the least common of external foliage feeding during the Paleozoic. The earliest occurrence is on the liverwort M. sharonae in New York State (Labandeira et al. 2014), where it is associated with small-scale margin and hole feeding made by microarthropods. There are no well-documented examples of surface feeding during the Late Devonian to Late Carboniferous, and the next occurrence is on the gigantopterid taxa of Gigantopteridium, Cathaysiopteris, and Zeilleropteris at the Early Permian Taint locality of north-central Texas, where it prominently co-occurs with extensive margin and hole feeding on the same plant hosts. Surface feeding is present very sporadically on glossopterid hosts from the Karoo Basin of South Africa, where it is documented at the Late Permian Kwa Yaya assemblage of Kwa Zulu-Natal, one of which is the surface-abrasion mode of DT103 (C. Labandeira and R. Prevec, unpublished observation, 2012). Surface abrasion frequently is mistaken for necrotic tissue resulting from fungal secondary invasion of lesions induced by arthropod herbivores. Paleozoic examples include fungal damage associated with DT97 on Late Carboniferous M. scheuchzeri (Stull et al. 2013) and secondary fungal colonization of hole feeding (DT2), margin feeding (DT13), and oviposition (DT101) on the K2 glossopterid morphotype series from the Late Permian Kwa Yaya assemblage of Kwa Zulu-Natal (Labandeira and Prevec 2014).

A particularly interesting case is DT97 (fig. 5L), which consists of thickened tissue in a U-shape pattern that generally parallels the secondary venation of medullosan neuropteroid foliage (Müller 1982). This DT likely represents a precursor interaction that gave rise to a gall that occurred on the medullosan host Odontopteris readi at Padgett, an Asselian-age locality in north-central Texas (Stull et al. 2013). Notably, the precursor of these galls is DT97, a widespread interaction that occurred throughout the late Middle and Late Pennsylvanian of Euramerica (Müller 1982; Labandeira 2006b) on medullosan seed ferns, particularly M. scheuchzeri. The presence of DT97 at CCP likely represents an antecedent of the earlier Permian Padgett gall that survived in a later Permian flora.

Piercing and Sucking

Definition. Piercing and sucking is the puncturing into plant tissue through the use of stylet-like mouthpart elements to access fluid tissues such as phloem, xylem, mesophyll, or pollen protoplasts. Two major categories of piercing-and-sucking damage are recognized in the compression-impression fossil record: individual puncture marks into plant tissues and ovoidal to ellipsoidal marks caused by the adpression and modification of plant surfaces from sessile scale insect bodies while piercing and sucking.

DTs represented. DT46, DT47, DT48, DT77, DT138, DT157, DT183, and DT244.

USNM figured specimens. CCP (USNM loc. 41005, 41006, 42292, 42305, 42306). For DT46: figure 3A, 3B (USNM-559821); figure 14G (USNM-559853); figure 14I (USNM-539333); figure 14M (USNM-539375); and figure 15B (USNM-539324). For DT47: figure 3C, 3D (USNM-539324). For DT48: figure 6H, 6I (USNM-559837); figure 14E (USNM-539333); and figure 6V, 6W (USNM-539430). DT77 is not illustrated. For DT138: figures 3J–3M, 14I (USNM-539333), and figure 14H (USNM-559820); DT157 is not illustrated. For DT183: figures 3E–3G, 14L (USNM-559822), and figure 8I–8M (USNM-559842). For DT244: figure 7J–7M (USNM-559839) and figure 7N–7P (USNM-530911).

Plant hosts. Cycadophyta: Taeniopteris spp. (DT46, DT47, DT48, DT77, DT138, DT157, DT183); indeterminate: unaffiliated broadleaf (DT46, DT47, DT77); ?peltaspermales: A. waggoneri (DT46, DT47, DT48, DT77, DT157, DT183); peltaspermales: Supaia thinnfelioides (DT46, DT47); Gigantopteridales: E. texana (DT48, DT77); and coniferales: W. piniformis (DT244).

Remarks. The damage categorized as DT46, DT47, DT48, and DT138 tends to be rather nondistinctive, although DT138 does indicate the tracking of particular tissues, such as veinal vascular phloem or intercostal mesophyll (fig. 3J–3M). However, the most evolutionarily and ecologically interesting of the piercing-and-sucking interactions are the distinctive scale marks indicated by DT77, DT157, DT183, and DT244, some of which indicate a general preference for particular seed plant hosts. One of these associations is new DT244, a distinctive ovoidal to ellipsoidal scale mark up to ca. 2 mm long by 1.2 mm wide that occurs only on the abaxial surfaces of the lanceolate needles of the early conifer W. piniformis (fig. 7J–7P). This interaction likely represents the attachment scar of a sessile wooly conifer aphid of the Adelgidae, a basal clade of phloem-feeding, sternorrhynchan Hemiptera that currently infest pinaceous conifers in the Northern Hemisphere, such as eastern hemlock Tsuga canadensis (McClure 1991; Young et al. 1995). The association between adelgids or their precursor lineage and conifers is considered ancient (Von Dohlen and Moran 2000), and evidence indicates that modern associations extend deep into the Mesozoic (Heie 1967; Shaposhnikov 1989). This host-specific association suggests promise for identification of a CCP association with an extant insect herbivore clade or more likely its Permian precursor lineage. Consistent with this assignment is an adelgid-like gall that occurs on the same plant host, W. piniformis, from the earlier Sakmarian-age CBB locality (Labandeira and Allen 2007, their figs. 6-8, 6-9). Although many modern adelgids are piercing-and-sucking hemipterans that leave ovoidal attachment marks on needles of the Pinaceae, others form bud galls that mimic vegetative buds but occur on anomalous sites on conifer branchlet systems, as in CBB damage.

New damage type DT244. Ovoidal to ellipsoidal structures with longitudinally striate to pustulose surfaces parallel to anterior-posterior scale insect axis; margins irregular and sometimes without a discernible biological border; 1.0–0.5 mm long by 0.5–0.10 mm broad; located at needle fascicle bases.

Distribution in Paleozoic habitats. The Paleozoic fossil record of piercing and sucking encompasses earlier occurrences that are three-dimensionally preserved in silica and carbonate as well as mostly later presence in compression-impression floras. The earliest occurrence of piercing and sucking is in the Rhynie Chert, Early Devonian Dryden Flags Formation, Scotland, where small stylet trajectories with accompanying disrupted tissue occur in silica-permineralized stems of early vascular plants (Kevan et al. 1975). The same type of preservation, with cone-shaped patterns of damage originating as multiple probes from a surface entry point, also occurs on silicified stems of early vascular plants but in younger Early Devonian deposits from Gaspé Peninsula, Québec, Canada (Banks and Colthart 1993). In the latest Middle Devonian Plattekill Formation, near Albany in New York State, distinctive clusters of ovoidal stylet punctures with cratered centers (DT273) are found on the liverwort M. sharonae (Labandeira et al. 2014). Three-dimensionally preserved carbonate permineralizations of stylet tracks penetrating the stem and rachis cortical tissues of ferns (Etapteris, Psaronius) are known from Late Carboniferous Euramerican coal-ball deposits (Scott and Taylor 1983; Labandeira and Phillips 1996b). The remaining record of Paleozoic piercing and sucking essentially consists of Permian compression-impression plants preserving single occurrences of cratered or uncratered ellipsoidal or circular puncture marks (DT46, DT47, DT48) on a range of seed plant foliage or, rarely, sphenopsid stems. During the Late Permian, linear rows of puncture marks are present on glossopterid foliage, which indicates specificity for mesophyll or vascular tissue (Prevec et al. 2009) and the targeting of pollen protoplasts by punch-and-sucking microarthropods (Wang et al. 2009). Plant-feeding arthropods that could have created this damage include the paleodictyopteroids, hemipterans, thysanopterans, and possibly some terrestrial mite lineages.

Oviposition

Definition. Ovipositional scars consist of a prominent lenticular or ovoidal rim of callus tissue surrounding a central flattened region in which there is disruption of epidermal tissue at one end, marked in certain cases by the presence of a slit or, rarely, evidence of an insect egg. Although oviposition is not a feeding mode, it does represent the effects of slicing plant tissue with a swordlike, egg-laying abdominal device, the ovipositor.

DTs represented. DT54, DT72, DT76, DT100, DT101, DT108, DT175, DT245, and DT246.

USNM figured specimens. CCP (USNM loc. 41005, 41006, 42292, 42305, 42306). For DT54: figure 4G, 4H, 4K (USNM-559827). DT72 is not illustrated. For DT76: figure 3J (USNM-539333); figure 3N, 3O (USNM-559824); figure 3Q, 3R (USNM-559825); figure 4A, 4C, 4D (USNM-559826); figure 4G, 4J (USNM-559827); figure 5A (USNM-539375); figure 6K, 6N, 6P, 6Q (USNM-538922); figure 7D–7G (USNM-536472); figure 9A–9D (USNM-559844); figure 9E–9G (USNM-559845); figure 11I, 11J (USNM-528610); figure 12O, 12P (USNM-559855); figure 14A (USNM-596951); figure 14B (USNM-559843); figure 14D (USNM-596952); figure 14E (USNM-559853); figure 14J (USNM-559820); and figure 14M (USNM-539375). DT100 is not illustrated. For DT101: figure 4I (USNM-559827); figure 6V, 6X (USNM-539430); and figure 8N, 8O (USNM-559843). DT108 is not illustrated. For DT245: figure 6B (USNM-559836) and figure 4L–4O (USNM-559828). DT175 is not illustrated. For DT246: figure 4A, 4B, 4E, 4F (USNM-559826); figures 4P–4S, 8S (USNM-559828); figures 8P, 8T, 14J (USNM-559829); figure 14A (USNM-596951); figure 14C (USNM-559854); figure 14D (USNM-596952); and figure 14N (USNM-559857).

Plant hosts. Cycadophyta: Taeniopteris spp. (DT54, DT76, DT100, DT101, DT245, DT246); ?Cycadophyta: Taeniopteris sp. (DT76, DT246); indeterminate: unaffiliated broadleaf (DT76); peltaspermales: S. texana (DT72, DT101, DT246); ?peltaspermales: A. waggoneri (DT54, DT76, DT100, DT101, DT108, DT175, DT245, DT246): peltaspermales: Callipteris sp. (DT76); and Gigantopteridales: E. texana (DT76, DT101, DT175, DT245, DT246).

Remarks. Of the broad spectrum of oviposition at CCP, three interactions stand out in importance. The distinctive ovipositional pattern of DT54 is typical of a highly stereotyped pattern that consists of arcs of eggs that occasionally crisscross. The arcs of egg sets are inserted in leaf tissue from the tip of a swinging abdomen that is repositioned to a new thoracic pivot point during the same ovipositional event (fig. 3G–3K). This pattern of associated arcuate sets of oviposited egg sets is typical for the Cenozoic (Sarzetti et al. 2009) but occurs rarely in the later Mesozoic and has not been previously encountered in the Paleozoic. It is possible that a paleodictyopteroid lineage bore an external, laterally compressed, sawtooth ovipositor and associated egg-laying behaviors (Labandeira 2006b) that evolved convergently with other piercing ovipositors, such as those of odonatans, paleodictyopteroids, and orthopterans and, after the Paleozoic, sawfly hymenopterans.

A second interesting association is DT175 (not illustrated), in which elongate egg insertions and associated scar tissue are deployed end to end in a series of three or more ovipositional events. This DT previously had an earliest occurrence in the late Middle Jurassic of China (C. C. Labandeira and J.-H. Ding, unpublished observation, 2012). An odonatopteran, or post-Paleozoic, dragonfly lineage is suggested as the culprit for the mid-Mesozoic occurrences, but it is unknown as to what Early Permian dragonfly, paleodictyopteroid, or other group of insects bearing a robust, external ovipositor was responsible for this damage.

The third type of ovipositional damage, DT245, is a new DT and occurs on A. waggoneri and Taeniopteris spp., the two dominant host plants. Each ovipositional mark is a strongly rounded, teardrop-shaped lesion, with sharply curved angulate ends on both sides of the mark, indicating the effectiveness of the ovipositor saw. DT245 is miniscule compared to other ovipositional damage and must have been made by an exceptionally small ovipositor-bearing insect capable of delivering relatively rectilinear insertions of eggs on foliar surfaces (fig. 4P–4S). There is indication that the linear files of egg insertions had the same pattern of subparallel to somewhat overlapping sets as demonstrated in DT54 above. Oviposition represented by DT245 may be represented in an SEM (fig. 15C) of USNM-539324, which shows the molds and casts of two small egglike structures.

New damage type DT245. Minute oviposition scars, between 0.3 and 1.0 mm in length, occurring in large numbers, parallel to venation across the entire leaf surface, including the midvein and lamina. Shape ranges from circular to teardrop shaped to lenticular; may have an outer rim.

New damage type DT246. Wide teardrop-shaped oviposition scars oriented parallel to leaf venation, with a length-to-width ratio between 2∶1 and 1∶1. In many cases the scars vary greatly in size on a single specimen and can occur singly or in clusters. Scars are at least 1.0 mm in length and range in shape from lenticular to short ellipsoidal to almost circular, with a round to gently undulatory outer margin.

Distribution in Paleozoic habitats. Oviposition DTs frequently are associated with particular organs and tissues of host plants, indicating moderate to high levels of plant targeting by insects bearing slicing ovipositors. Likely insect lineages bearing plant-piercing ovipositors during the Late Paleozoic include paleodictyopteroids, odonatopterans, hemipteroids, and orthopterans (Labandeira 2006a), responsible for the rich fossil record of oviposition during the Paleozoic. Among the oldest examples of oviposition are distinctive scars on the surfaces of calamitalean stems from the Late Pennsylvanian (Gzehlian) of France (Béthoux et al. 2004). However, Paleozoic oviposition does not become common until the Late Permian of Gondwana, particularly on glossopterid leaves in which both midribs and blades sporadically display extensive lenticular-shaped lesions (Prevec et al. 2009; McLoughlin 2011). The earlier but limited Late Carboniferous occurrences of oviposition mostly on sphenopsid axes can be contrasted with Late Permian expansion of oviposition on seed plant foliage. The ovipositional pattern of the Late Carboniferous also can be compared with analogous damage at CCP, reflecting targeting of the midribs and blades of peltasperm and taeniopteroid foliage, for example, by the insect that caused the considerable ovipositional scarring of DT245.

Galling

Definition. Galls are compact, differentiated, anomalous tissues occurring in organs of plant hosts that are developmentally induced by the hormonal control of a gall inducer such as an insect or mite. Insect gallers develop in a typically spheroidal chamber within the gall that is enveloped by nutritive or similar consumable tissue, in turn surrounded by a protective commonly hardened or woody tissue. Galls are a distinctive plant-insect association in which the host plant not only responds to an endophytic intruder by creating reaction tissue but also is developmentally directed by the endophytic galler to produce anomalous protective and nutritionally rich tissue.

DTs represented. DT32, DT33, DT34, DT80, DT120, DT122, DT247, DT259, DT260, and DT262.

USNM figured specimens. CCP (USNM loc. 41005, 41006, 42292, 42305, 42306). For DT32: figure 6K–6M, 6O, 6P (USNM-538922); figure 6R–6U (USNM-559838); and figure 14E (USNM-559853). For DT33: figure 4L (USNM-559828). For DT34: figure 6B (USNM-559836). For DT80: figure 5P–5R (USNM-559833); figure 10L, 10M (USNM-559850); and figure 15A (USNM-539324). For DT120: figure 5A–5D (USNM-539375); figure 5G–5K (USNM-559831); figure 5L–5N (USNM-559832); figure 7H, 7I (USNM-530930); figure 8C–8H (USNM-559841); figure 8N, 8P–8T (USNM-559843); figure 9E, 9H, 9I (USNM-559845); figure 9J–9M (USNM-559846); figure 9P–9R (USNM-530932); figure 10G–10K (USNM-559849); figure 11E–11H (USNM-528672); figure 14A (USNM-596951); figure 14C (USNM-559854); and figure 14D (USNM-596952). DT122 is not illustrated. For DT247: figure 10L, 10N (USNM-559850). For DT259: figure 11A–11D (USNM-559851) and figure 11E–11H (USNM-528672). For DT260: figure 10A–10F (USNM-559848); figure 12A–12H (USNM-559854); figure 12I–12N (USNM-528206); and figure 12O–12R (USNM-559855). For DT262: figure 9N, 9O (USNM-559847).

Plant hosts. Cycadophyta: Taeniopteris spp. (DT32, DT33, DT34, DT80, DT120, DT122, DT120, DT247, DT262); ?Cycadophyta: Taeniopteris sp. (DT120); indeterminate: unaffiliated broadleaf (DT32, DT33, DT120, DT247, DT260); peltaspermales: S. texana (DT120); ?peltaspermales: A. waggoneri (DT32, DT33, DT34, DT80, DT120, DT247, DT259, DT260, DT262); ?peltaspermales: S. thinnfeloiodes (DT32, DT120); and Gigantopteridales: E. texana (DT32, DT33, DT34, DT80, DT120, and DT247).

Remarks. Although the two dominant foliar taxa at CCP are heavily galled, they share most gall DTs. Shared gall types DT32, DT33, DT34, DT80, DT120, and DT122 occur from the Permian to the Holocene (Labandeira et al. 2007) and form on the foliage of a wide range of host plant taxa. The gall DT120 is the most abundant gall at CCP, occurring on multiple hosts and in many cases at high densities on the foliar surface. The more distinctive galls, DT247, DT259, DT260, and DT262, have unique features that include shape and size, chamber number, inner gall tissue development, ornamentation of the wall surface, wall thickness, and response of the host foliar tissue to gall presence. These distinctive galls also tend to occur on only one host plant, usually A. waggoneri. The identified gall types indicate that CCP has a more diverse gall than any other known flora from the Permian of north-central Texas.

New damage type DT247. Formless, circular to polylobate flattened surface galls with multiple chambers and a prominent reaction rim; highly variable in size and shape. Surfaces are distinguished from surrounding foliar material by minute desiccation cracks and locally by spongy tissue. See figure 10L–10O (USNM-559850).

New damage type DT259. Dark, spheroidal, single-chambered galls; inner structure homogenous, surrounded by a broadly botryoidal but muted surface sculpting the outer wall surface; eccentric markings at edge; thin, dark, smooth reaction rim; gall 0.5–1.5 mm in diameter. Outer margin spheroidal in shape, occurring singly or in loose clusters. See figure 11A–11D.

New damage type DT260. Circular to ellipsoidal galls occurring on midrib and blade, one wall layer distinctively pustulose, enveloped by a thick outer sclerenchymatous wall. Outer margin spheroidal but frequently irregular and polylobate. Very variable in size, from 0.6 to 11 mm in maximum dimension. Typically single chambered, possibly compound in mature specimens, occurring singly or in loose clusters. See figure 12A–12H (USNM-559854) for best material; see also figure 12I–12N (USNM-528206); figure 12O, 12Q, 12R (USNM-559855); figure 11E–11H (USNM-528672); figure 10A–10F (USNM-559848).

New damage type DT262. Massive, woody, single-chambered midrib gall consisting of a thick to hemispherical cylindrical to spheroidal structure ranging from 1 to more commonly 5 mm across, with distinctive, thick ridges originating from the base and extending to the distal gall region; in some cases, galls are linked together but invading blade tissue. See figure 9N, 9O (USNM-559847); figure 9P, 9Q (USNM-530932).

Distribution in Paleozoic habitats. The fossil record of galls commences in the Early Pennsylvanian (Labandeira 1998), where the apical strobilus of the calamite Paracalamostachys was galled in a developmental mode similar to that of bud galls in higher plants (van Amerom 1973). Well-documented Late Pennsylvanian galls occur within the inner parenchyma of Stipitopteris petiolar fronds borne by the tree fern Psaronius chasei Morgan (Labandeira and Phillips 1996a; Redfern 2011). These petiolar galls were occupied by an early holometabolous larval insect that was surrounded by tufts of proliferating tissue, coprolites, and other types of frass. During the Early Permian, the best-documented galls are anomalous strobilus-like proliferations appearing as anomalous buds on walchian conifers (Florin 1945; Labandeira and Allen 2007).

Late Carboniferous galls were involved in developmental control of particular meristematic tissues within axial organs and differ from the overwhelmingly greater number and broader spectrum of Permian galls that occurred on foliar tissues from a variety of floras worldwide. From a temporal perspective, the shift from axial to foliar galls continues at a meager pace throughout the remaining Permian (Stull et al. 2013). Galls are absent from the Taint flora and rare at CBB. Galls exhibit a broader spectrum of shapes and morphologies at CCP, providing a contrast to the simple and highly stereotyped morphologies found in earlier floras. During the Late Permian, galls are particularly noticeable on Gondwanan glossopterid foliage (Prevec et al. 2009).

Seed Predation

Definition. Seed predation is the consumption of live megagametophytic and in some cases embryonic tissues in dispersed or plant-attached seeds and invariably results in the death of an entire plant disseminule.

DTs represented. DT73, DT74, and DT257.

USNM figured specimens. CCP (USNM loc. 41006, 42306). For DT73: figure 13A, 13B (USNM-539326), and figure 13L, 13M (USNM-559859). For DT74: figure 13A–13D (USNM-539326) and figure 13J, 13K (USNM-559858). For DT257: figure 13E–13G (USNM-559856) and figure 13H, 13I (USNM-559857).

Plant hosts. Unknown Spermatophyta seeds: platysperm sp. 1 (DT73, DT74); platysperm sp. 2 (DT74, DT257); platysperm sp. 3 (DT73), collectively designated herein as platysperm sp. seed.

Remarks. In the absence of ovulate fructifications with organic connection to vegetative material at CCP, it is impossible to assign platysperm sp. 1–3 to particular respective seed plant lineages. These three platysperm seed morphotypes may represent multiple species, including a walchian conifer with a diminutive micropylar region and other seed plant taxa with more prominent micropyles. For this reason, we are considering all three platysperm seed morphotypes as one categorical unit in this analysis until further study can define the limits of each morphospecies or establish whole-plant taxon affiliations.

Generic winged platysperm seeds occur in many floras throughout the Permian and Triassic and have been assigned to a variety of major lineages, including gigantopterids, cordaites, corystosperms, and ginkgophytes (Anderson and Anderson 2003; Taylor et al. 2009). A notable feature of the seed predation at CCP is the elevated incidence of DT73 and DT74 when compared to other Permian floras. Also present is a different form of piercing and sucking in DT257, which represents a scale insect that is specific to platysperm sp. seed and likely fed on the seed’s megagametophytic and wing tissues. The scale attachment scars appear frequently lodged between the sulcus formed by the seed wing and main body. The closest modern analogue of this association is scale insect predation on the winged seeds of pinaceous conifers (Turgeon et al. 1994).

Although three of the occurrences of DT257 (fig. 13F–13H) are on the main body of platysperm type 2 seeds, the other seven occurrences of seed predation, consisting of DT73, DT74, and DT257, occur on the wings of platysperm type 1, 2, and 3 seeds. One explanation for this pattern is that a host plant cue was missed for targeting the megagametophytic tissues of the seed main body. An alternative and more likely possibility is that the scale insect seed predator had a more eclectic spectrum of tissue specificities and consumed wing as well as megagametophytic tissues. Seed predator consumption of photosynthetic wing tissue currently occurs in a variety of plant hosts by a variety of seed predators. These associations include several species of Abies (fir) consumed by the fir seed gall midge Dasyneura abiesemia (Diptera: Cecidomyiidae; Keen 1958, fig. 13), Tsuga (hemlock) attacked by the elongate hemlock scale Fiorinia externa (Hemiptera: Diaspididae; McClure 1979; Johnson and Lyon 1991, fig. 45C), and the maple samara leaf miner Etainia sericopeza (Lepidoptera: Nepticulidae; Puplesis 1994, fig. 809). Based on these patterns of herbivory in modern counterparts, it is highly likely that platysperm types 1–3 bore photosynthetic tissues in their wings.

New damage type DT257. Circular to broadly ovoidal scale scars, 0.8–1.5 mm maximum dimension; surface texture smoother or spongier than surrounding seed test. Occurring on seeds, on the central body and wings.

Distribution in Paleozoic habitats. While seed predation is a recognizable FFG based on a very distinctive trophic interaction, the means of predation are similar to other such groups, such as piercing and sucking in the case of paleodictyopteroid and hemipteran insects (Shcherbakov et al. 2009) or the creation of larval galleries by seed weevils (Janzen 1971) during the Mesozoic. In many cases, seed predation requires mouthpart structures to puncture or chew through hard or thick integumental structures that enclose seeds, such as an outer sclerotesta or inner fibrous pericarp.

The earliest documented seed predation originates during the Early Pennsylvanian, consisting of small round perforations into the Euramerican medullosan seed Trigonobalanus (Labandeira 2007). The same style of seed predation occurs in Late Pennsylvanian platysperm cordaitalean seeds from the Chunya locality of the Tunguska Basin, Russia (Sharov 1973). Identical damage on winged cordaitalean seeds also has been described from the Russky Island site of the Prospelovo Formation of the Pechora Basin, Russia, on Samaropsis danilovii Suchov (Shcherbakov 2008). This locality, about the same age as CCP, bears a similar type of small circular perforation that targeted megagametophytic tissues of the central body. This type of damage has been attributed to the stout, probing beaks of some Palaeodictyopteroidea (Sharov 1973), in which piercing-and-sucking action flushed out the contents of megagametophytic tissues in a manner similar to seed bugs (Lygaeidae) of modern Hemiptera (Ralph 1976).

Borings

Definition. Borings are tunnels and galleries produced by arthropods in indurated tissues such as wood and sclerenchyma or engravings in softer tissues such as cambia or pith parenchyma wherein a discrete network is present that may contain frass, fungi, or infilled sediment.

DT represented. DT243.

USNM figured specimens. CCP (USNM loc. 42305). For DT243: figure 13N–13P (USNM-559860).

Plant hosts. Incertae sedis: an indeterminate axis, petiole, or isolated midrib of a seed plant leaf.

Remarks. Eight simple borings of relatively small diameter compared to other Paleozoic occurrences (Weaver et al. 1997; Noll et al. 2004; Naugolnykh and Ponomarenko 2010) were found on a single specimen of a CCP woody element. Although the woody axis lacked any attached branches or foliage, its overall robustness and heavily vascularized structure are suggestive of a rachis of A. waggoneri. Wood borings on such small branches are especially rare in the fossil record, attributable to hydrodynamic qualities of small axes that disallow their incorporation into fossil leaf layers or, alternatively, are too small to enter the permineralized record.

New damage type DT243. Small (ca. 1–1.4 mm diameter) borings perpendicular to wood surface infilled with fine-grained matrix.

Distribution in Paleozoic habitats. Borings are rare in plant compressions and impressions but are commonly encountered in permineralized woods, where three-dimensional preservation allows for observation of tunnel detail. Several features are important in the determination of tunnel culprits, including the elucidation of three-dimensional tunnel and gallery network geometry, frass infilling contents, and behavioral targeting of nutritious tissues such as cambia (Tovar et al. 1995; Labandeira 2002). Generally, Permian-age borings are uninformative regarding the identities of both the targeted host plant and the wood-boring arthropod.

Borings have a sporadic fossil record during the Pennsylvanian, occurring principally in the central tissues of medullosan trunks (Labandeira 1998), Psaronius tree fern rachises (Rößler 2000; Labandeira and Phillips 2002), and calamite stems (Rößler 2006). Much of this damage, however, involves nonwoody pith or other medullary tissues that are amenable to tunneling by insects. These tunnels lack structures typical of those boring into harder substrates such as wood. Nevertheless, there is a detectable shift beginning during the Early Permian toward rare borings present in the more indurated tissues of calamites (Rößler 2006) and especially seed plant taxa, particularly the wood of coniferophytes and glossopterids (Zavada and Mentis 1992; Weaver et al. 1997; Noll et al. 2004; Rößler 2006). Throughout the Permian, borings in mostly arborescent gymnosperms are present, where the larvae of archostematan beetles have been cited as offending culprits (Naugolnykh and Ponomarenko 2010).

Fungal Damage

Definition. Fungal damage consists of the pathogenic invasion of live plant tissue through a breach in a plant’s integument, commonly introduced through lesions produced by insect herbivory. Pathogenic features resembling fungi also may be caused by physical conditions such as absence of essential nutrients or abnormal temperature conditions.

DT represented. DT58.

USNM figured specimens. CCP (USNM loc. 41006, 42292, 42306). For DT58: figure 5C (USNM-539375), figure 7B (USNM-530921), figure 12R (USNM-559855), and figure 14M (USNM-539375).

Plant hosts. ?Cycadales: Taeniopteris spp.; indeterminate: unaffiliated broad-leaved seed plant; and ?peltaspermales: A. waggoneri.

Remarks. As currently conceived, DT58 consists of generic unassigned fungal damage. A recent examination of the compression-impression fossil record indicates that many instances of pathogen-mediated plant disease, including fungal damage, are significantly underrecognized (Labandeira and Prevec 2014). Nevertheless, several examples of fungal necroses exist at CCP, principally through fungal invasion from open foliar wounds (figs. 5C, 7B, 12R). Examples include entry at the ruptured tissue at the base of a DT120 gall (fig. 5C) or after surface abrasion caused by the maker of DT31 (fig. 7B) and the apparent colonization of foliar tissue without any evident connection to a wound entry caused by an herbivore (fig. 12Q). The pattern of pathogenic fungal infection at CCP reflects a general pattern that also occurs on other well-documented Permian floras (Prevec et al. 2009; McLoughlin 2011; Labandeira and Prevec 2014).

Distribution in Paleozoic habitats. Much of the fossil record of plant-insect associations is linked with secondary pathogenic infection, particularly fungi (Labandeira and Prevec 2014). For the late Paleozoic, fungal damage is present as zones of texturally distinctive necrotic tissue that surround inner tissues exposed by external foliage feeding (Labandeira 2006b; Stull et al. 2013) and oviposition marks (Prevec et al. 2009). Fungal associations also have been documented for Permian woods (Dieguez and López 2005) whose necroses resemble the rots and other pathogenic signs of extant woods (Sinclair et al. 1987). The overall lack of documentation presumably is attributable to an absence of appreciation of fungal damage or to difficulties in assignment to a particular pathogen sign to a causative organism, such as a virus, bacterium, fungus, or nematode (Labandeira and Prevec 2014).

Patterns of Herbivory Revealed by DT Occurrences

Of the 2140 plant specimens examined at CCP, 649, or 30.33%, displayed one or more instances of herbivory (table 1). A total of 1390 instances of herbivory were observed throughout the CCP flora: 1346 on broadleaf taxa, 8 on conifers, 2 on axes, and 34 on seeds (table 2). Multiple DTs and FFGs were frequently found on individual specimens. Of the instances of herbivory recorded for the CCP flora, those allocated to the galling FFG constituted one-third of all DT occurrences; those of the oviposition FFG had one-third of all DT occurrences; and those of the external foliage feeding FFG, represented by the hole-feeding, margin-feeding, and surface-feeding subgroups, provided one-fifth of all DT occurrences. Minor levels of herbivory were present for the piercing-and-sucking FFG, responsible for one-tenth of all DT occurrences, and the remainder FFGs constituted 3.0% of all DT occurrences. Clearly, oviposition, galling, and external foliage feeding were the dominant FFGs at CCP. External foliage feeders targeted Taeniopteris spp., galling preferentially occurred on A. waggoneri, and oviposition was distributed evenly between these two most herbivorized taxa.

Of the 12 taxa represented in the CCP flora, four—W. piniformis, Taeniopteris spp., A. waggoneri, and platysperm sp. seed—constituted 90.09% of the flora (table 1), based on the numbers of specimens, and 91.67% of the herbivory. Of these, the two dominant herbivorized taxa were Taeniopteris spp. and A. waggoneri, which disproportionately constituted an elevated 37.44% and 47.92%, respectively, of the herbivorized specimens in the flora, clearly the two most herbivorized taxa (table 1). Perhaps unsurprisingly, A. waggoneri, the fourth most abundant host but the most abundant monophyletic broadleaf taxon, represented almost half of all herbivore-inflicted damage. By contrast, the conifer W. piniformis, the second most abundant taxon in the CCP flora, expressed a very low incidence value of 1.23%, likely a consequence of minimal exposure of surface area and structurally defended tissues that would be unavailable to herbivores. Similarly, the platysperm sp. seed, the most abundant taxon at CCP, bore a higher but still modest incidence of herbivory of 5.08%. Although a respectable value for seed predation of the Paleozoic and Mesozoic, this value likely was due to the poor development of the seed predation feeding guild before and during the Early Permian, particularly as megagametophytic tissues would be accessible and nutritious sources for consumption, unlike the more refractory and digestively more difficult foliar tissues of W. piniformis.

The remaining eight taxa, mostly broad-leaved medullosan, callipterid, gigantopterid, and probable cycadophyte seed plants, constituted 9.01% of the flora and a collective herbivory of 8.33%, commensurate with their collective abundance. However, there was significant variability of herbivory within this group, with taxon-specific values ranging from a high of 3.24% for the gigantopterid E. texana, with 32 specimens, to the absence of herbivory on the sphenopsid S. thonii, with three specimens. The taxon-specific herbivory values for these considerably less abundant eight taxa may indicate that arthropod herbivores were tracking overall taxa abundance, though only among broad-leaved foliage, as this pattern breaks down when the comparatively low values of herbivory are considered for the two most abundant taxa, the non-broad-leaved W. piniformis and platysperm sp. seed.

Patterns of Herbivory Revealed by Surface Area Removal by Herbivores

An assessment of the CCP flora also was made based on the amount of leaf surface area removed by herbivores. The total surface area examined was 39,367.32 cm², and the amount of herbivorized surface area was 622.26 cm², a removal of 1.58% for the total CCP flora (table 1). Although oviposition is not a type of feeding and rather represents the response of egg insertion into plant tissue, it is treated as herbivory in all analyses presented here because of its conspicuous fossil record and its feeding-like use of plant tissue as a resource. Individual leaves measured up to 481.99 cm², the largest specimen, and the smallest specimens were seeds and seed fragments that ranged in size from 0.03 to 1.50 cm². We note that the taxon with the most surface area removed by herbivory was A. waggoneri, with a value of 3.08%; this single taxon represented an astounding 80.17% of all surface area removed at CCP. The second most herbivorized taxon was Taeniopteris spp., with 1.36% of its surface area removed, but only representing 15.97% of all surface area removed at CCP. The rank order of A. waggoneri (first) and Taeniopteris spp. (second) matches the same rank order of taxa for the proportion of herbivorized specimens (table 1; fig. 14); however, A. waggoneri has a much greater representation of herbivorized surface area than Taeniopteris spp. These two dominantly herbivorized taxa account for 96.14% of the removed surface area, a value greater than the 85.36% of herbivorized specimens in the flora that these two taxa contribute.

The remaining CCP taxa constituted only 3.86% of the herbivorized surface area removed. Of these, the gigantopterid E. texana was the most common, with 3.42% (table 1), representing 88.6% of the remaining herbivorized surface area. The remaining taxa either lacked or had only trace amounts of foliar tissue removed.

Patterns of Arthropod Herbivory at CCP

Much of the herbivory at the CCP flora was generalized, particularly for Taeniopteris spp., Auritifolia waggoneri, and Evolsonia texana, the three most herbivorized taxa. Taeniopteris spp. and A. waggoneri exhibit approximately equal amounts of generalized and specialized DTs (table 2). Evolsonia texana, third in rank order, is instructive regarding specialized herbivory. Of the 44 occurrences of Evolsonia damage, approximately half (21 DTs) represent generalized herbivory, principally, various forms of external foliage feeding, whereas the other half (23 DTs) represent more specialized DTs such as piercing and sucking but especially oviposition and other modes of galling. Similar values were present for Sandrewia texana, fourth ranked in herbivore attack, consisting of 10 DTs assigned to generalized consumption and 11 DTs assigned to more specialized feeding modes. The other taxa provided a paucity of data—too few specimens, too few DTs, or both—for interpretation.

Taeniopteris spp. and A. waggoneri are the most herbivorized plant hosts (table 1). This conclusion is supported by (1) the percentage of herbivorized plant host specimens, (2) the percentage of removed foliar surface area, and (3) the elevated number of DTs on each host plant (table 3). Oviposition and subordinately external foliage feeding dominated herbivory on Taeniopteris, whereas galling and subordinate oviposition were the most important sources for A. waggoneri. These and other host-specialized DTs are confined to one or the other of the two hosts. Surface feeding of DT103 and DT263 and galling of DT122 and DT120 are restricted to Taeniopteris spp., whereas piercing and sucking of DT138, oviposition of DT175, and galling of DT259 occur solely on A. waggoneri (table 3). Each of these plants supports a component community that consists of the source plant and all directly or indirectly trophically dependent organisms that derive energy from a resource provided by the plant (Root 1973). These interactions appear to be examples of host specialization that are overshadowed by more extensive co-occurrence of other DTs on both Taeniopteris spp. and A. waggoneri. The component arthropod herbivore community of Taeniopteris spp. prominently consists of external foliage feeding (DT4, DT12, DT103, DT263) and oviposition (DT76, DT245, DT246) but also piercing and sucking (DT46 and DT183) and galling (DT263; fig. 14H–N; table 3). By contrast, the component community of A. waggoneri records the predominance of galling (DT80, DT120, DT246, DT260) and oviposition (DT76, DT101) but also external foliage feeding (DT29, DT130) and various types of piercing and sucking (fig. 14A–G; table 3).

These general trends for greater palatability of certain seed plants are relevant for the greater levels of herbivory for certain hosts seen at CCP. Several lines of evidence from CCP plants suggest a constitutive antiherbivore defense syndrome. (Constitutive defenses are structural or chemical antiherbivore defenses that are intrinsic to basic plant architecture; by contrast, induced defenses are acquired by the plant host as a direct response to previous insect herbivory; Karban and Baldwin 1997.) The constitutive defenses for the two most herbivorized taxa of Taeniopteris sp. and A. waggoneri were (1) thickened, rigid cuticles; (2) thick foliage; (3) the perfusion of the lamina with robust, closely spaced, and hardened primary to tertiary veins that are expressed at the surface; and (4) prominent secondary reaction rims along necrotic zones (e.g., figs. 2B, 3D, 9A, 14, 15). There was no evidence for spines, stiffened trichomes, or glandular structures such as those found by Pott et al. (2012) on bennettitalean fronds. All of the physical constitutive defenses at CCP would have minimized herbivory. However, these metabolically costly structural features are most prominent in the two most heavily herbivorized plant hosts, indicating that a host plant–herbivore arms race was present at CCP. This suggests an escalated relationship (Vermeij 1987) between key plant hosts and their arthropod herbivores.

The absence of skeletonization at the CCP site is noteworthy. Skeletonization, like hole feeding, is the consumption of the entire thickness of intercostal leaf tissue but with one or more orders of veins remaining, forming a fine to coarse lacelike network (Coulson and Witter 1984). Skeletonization commonly occurs along a primary or secondary vein that serves as a structural barrier confining feeding damage to a portion of the leaf, such as angularly constrained tissue formed at junctures of robust primary and secondary veins (Heron 2003). The robustness of the remaining veins forming the skeletonized lattice can provide evidence for the mouthpart morphology and efficacy of chewing by the suspect herbivore. These relationships have been based on the modern relationship between the mode of plant damage and mandible type (Gangwere 1966; Bernays and Janzen 1988; Labandeira 1997). This type of feeding damage becomes gradually more abundant during the preangiospermous Mesozoic plant record. One notable exception and an early example of more intense skeletonization is the Late Triassic fern Dictyophyllum nathorstii Zeiller, which has extensive skeletonization (Feng et al. 2014). Skeletonization becomes a significant type of damage for angiosperm-dominated floras of the Late Cretaceous to Neogene (Labandeira 1998, 2006b).

One possible explanation for the overall paucity of Permian skeletonization is the absence of derived lepidopteran and coleopteran lineages that originated during the Early Cretaceous to Paleogene and whose larvae are dominant skeletonizers today (Powell et al. 1998; Hunt et al. 2007). Another factor explaining the dearth of skeletonization at CCP is that the site harbors a high proportion of host plant taxa that displayed robustly thickened laminae with considerable interstitial tissue development, as well as massive petioles and primary veins, supporting fibers, abundant bundle sheaths, and sclerenchymatous, cuticular, and other structural tissues (fig. 15; Gillespie and Pfefferkorn 1986; Chaney et al. 2009; Looy 2013). These structural features would have impeded access of herbivores to leaf tissues, particularly for Taeniopteris and Auritifolia. However, some modern insect taxa select for consumption those leaves with robust leaf structural traits that otherwise would be interpreted as herbivore deterrents (Peeters 2002). Robust frond construction, including massive bundle sheaths, thickened pinnules, and possible semisucculence, have been described for the Artinskian peltasperm Glenopteris splendens Sellards from the Wellington Formation of Kansas (Krings et al. 2005), also indicating the possibility of strong mechanical herbivore deterrence. Although these histological foliar features provide several features of leaves consistent with host plant repelling of insect herbivory, it remains unknown whether or how chemical defense was used in CCP plants.

Evolution of Insect Herbivore Component Communities during the Early Permian

Although additional assemblages are needed during the approximate 11-million-year interval of the Early Permian redbed sequence of north-central Texas, certain patterns already can be detected based on the current study of CCP (Kungurian) and on earlier studies of the Taint (Artinskian; Beck and Labandeira 1998; Labandeira 2012) and CBB (Sakmarian) floras (Labandeira and Allen 2007). A conspicuous pattern is that there is a broader range of FFGs and DTs at CCP than those of Taint or CBB. A second noticeable pattern is the approximate 10-fold increase in foliar surface area removed by herbivores and twofold increase in the proportion of herbivorized leaves between the CBB and Taint assemblages (table 4). This dramatic increase is followed by stabilization of both measures of herbivory between Taint and CCP.

Another trend is the consistent and elevated levels of herbivory on seed plant taxa at all localities, favoring highly herbivorized peltasperm or gigantopterid hosts. One explanation for this preference is the greater conspicuousness, or apparency (Feeny 1976), of seed plant foliage in the local environment. This increased herbivory matches the marked diversification of these plant groups in the southwestern United States during this time (Mamay 1989; DiMichele et al. 2005; Mamay et al. 2009). Alternatively, and more likely, is the greater palatability or availability of seed plants to a broad spectrum of local phytophagous insects (Labandeira and Currano 2013). The trend toward the preferential herbivory of broad-leaved seed plant leaves over other plant organs and vascular plants is indicated by the virtual absence of herbivory on non–seed plants at CBB, Taint, and CCP. In addition, this pattern already was present in Middle Pennsylvanian floras, during which medullosan foliage such as Macroneuropteris was preferentially consumed over non–seed plant taxa, especially the more abundant marattialean tree fern foliage of Pecopteris (Trout et al. 2000).

These data from the redbed sequence of north-central Texas (Chaney and DiMichele 2003) and an earlier study (Trout et al. 2000) strongly support the hypothesis that seed plant taxa were especially selected for consumption by local arthropod herbivores during the Early Permian (Beck and Labandeira 1998; Labandeira 2002, 2006a, 2006b, 2012; Labandeira and Allen 2007; Stull et al. 2013). This selectivity for seed plants also occurs in Permian floras of western Europe (Geyer and Kelber 1987), Cathaysia (Glasspool et al. 2003), and extensive glossopterid-dominated floras across Gondwana in South Africa (Prevec et al. 2009), India (Srivastava and Agnihotri 2011), Australia (McLoughlin 1994a, 1994b; Beattie 2007), Antarctica (Slater et al. 2012), and South America (Adami-Rodrigues and Iannuzzi 2001; Adami-Rodrigues et al. 2004a, 2004b; Cariglino 2011). A common observation from these studies is that the preferential targeting of seed plants was made by multiple FFGs, particularly external foliage feeders (Labandeira 2006a, 2006b; Labandeira and Currano 2013), piercers and suckers (Wang et al. 2009), ovipositing insects (Prevec et al. 2009; McLoughlin 2011), and gallers (McLoughlin 2011; Stull et al. 2013). The dominance of seed plant herbivory during the Permian is an extension of a pattern that began during the Late Mississippian (Labandeira 2007; Iannuzzi and Labandeira 2008), is sporadically documented throughout the Pennsylvanian (Scott and Taylor 1983; Castro 1997; Scott and Titchener 1999; Jarzembowski 2012), and achieves dominance during the Early Permian.