Cryptic or underworked? Taxonomic revision of the Antistrophus rufus species complex (Cynipoidea, Aulacideini)

) Cryptic or underworked? Taxonomic


Introduction
Cryptic species are those that cannot readily be distinguished on the basis of phenotypic variation alone (Struck et al. 2018); they have created challenges in myriad areas of the life sciences, especially in agro-economic systems (Andrews et al. 2020;Hansen et al. 2021;MacLeod et al. 2021).One such system of emerging significance is that of the genus Silphium L., a genus of composite herbs in the sunflower tribe (Asteraceae: Heliantheae) that is currently being investigated for use as biofuels, oilseeds, and additional economic purposes (Kowalski et al. 2013;Jasinskas et al. 2014;Peni et al. 2020;Vilela et al. 2020;von Cossel et al. 2020;Țîței et al. 2021).Despite recent interest in cultivating and domesticating various Silphium species, as well as their prominence in threatened tallgrass prairie habitats, their associated arthropod fauna remains poorly studied (Henderson and Sauer 2010;Buffington et al. 2017).A major component of Silphium communities is herb gall wasps of the genus Antistrophus Walsh, 1869 (Fig. 1), which includes a known case of cryptic species (Tooker et al. 2004).
Six described species of Antistrophus induce galls in the disc flowers and stems of four Silphium species (Nastasi and Deans 2021).Perhaps the most intriguing Silphium gall wasps are three cryptic species composing the rufus species complex (A.jeanae Tooker and Hanks, A. meganae Tooker and Hanks, and A. rufus Gillette).These species are not only thought to be morphologically cryptic, but also induce galls that are inconspicuous and do not externally deform the host plant's tissues (often referred to as "cryptic galls" in the literature; e.g., Ronquist and Liljeblad 2001); each species induces these galls within stems of their respective host species of the genus Silphium (Tooker et al. 2004;Nastasi and Deans 2021).The complex was described by Tooker et al. (2004), who found each species to be distinct based on allozyme and ecological data, including use of host plant species (Tooker et al. 2004); A. jeanae is associated with S. perfoliatum L., A. meganae is associated with S. terebinthinaceum Jacq., and A. rufus is associated with S. laciniatum L. These species would be considered cryptic, or functionally so, based on the diagnostic morphological criteria which included averages of precise ratios of antennomere dimensions, lengths of dissected ovipositors, depths of galls within host stems, and masses of mature larvae (Tooker et al. 2004).These metrics are hard to replicate and do not fully diagnose the species, thus the original diagnoses are arguably insufficient to reliably arrive at a species identification without host plant data.While these are the only morphological diagnostic characters given by Tooker et al. (2004), the descriptions of these species, including Gillette's 1891 description of A. rufus, are fairly limited in the number of morphological characters examined.Due to the brief original descriptions of species composing the A. rufus complex, it seems likely that this is a case of the so-called "superficial description impediment", a taxonomic phenomenon in which species are not readily identifiable based on characterization in existing literature (Meier et al. 2021), rather than a true case of cryptic species.
The potential for a superficial description impediment in the A. rufus complex is supported by the minimal history of taxonomic investigation of herb gall wasps in North America; among the early diverging lineages of gall wasps are several tribes of non-oak herbaceous gallers (Aylacini sensu lato), but these remain more poorly understood than the species-rich oak gallers (Cynipini).While the herb gall wasp fauna of the Western Palearctic has been comparatively well studied (e.g., Nieves-Aldrey 1994;Nieves-Aldrey et al. 2008; Nieves-Aldrey 2012; Nieves-Aldrey 2022), those in North America remain poorly investigated.Eighteen valid native species are known (Nastasi and Deans 2021;Nastasi et al. 2024); only 14 published papers explore genus-or species-level taxonomy of these taxa, and many of these papers treat only a single species (Table 1).Of these works, only two (Nieves-Aldrey 1994; Tooker et al. 2004) have been published in the last 100 years.When combined with the known diagnostic characters for the rufus complex, the paucity of taxonomic work on this group further increases the likelihood of a taxonomic impediment rather than a true case of the cryptic species phenomenon.
In our analysis of the Antistrophus rufus complex, we included A. jeanae (gall inducer on S. perfoliatum), A. meganae (gall inducer on S. terebinthinaceum), and A. rufus (gall inducer on S. laciniatum) by sampling individuals reared from stems of these three Silphium host species.In the process we reveal in the stems of S. integrifolium the presence of an additional species that we newly describe, Antistrophus laurenae Nastasi sp.nov.; S. integrifolium is a new host plant for the rufus complex.We include in this analysis A. minor Gillette, 1891; we treat this species in the rufus complex based on its adherence to the diagnostic criteria for the rufus complex Table 1.Taxonomic treatments of North American herb gall wasps, 1869-present.Names later transferred to other genera marked with (*).Names later synonymized marked with (◊).Species that are nonnative to North America (or questionably so) are excluded.

Author and year
Description of taxonomic work Riley and  Described the genus Aulacidea for some species placed in Aylax Beutenmüller 1908 Described Aylax chrysothamni * Beutenmüller 1910a Reviewed North American species of Aylax sensu lato Beutenmüller 1910b Reviewed North American species of Aulacidea Kinsey 1920 Described Aulacidea abdita and A. annulata McCracken and Egbert 1922 Described Aylax microseris* Nieves-Aldrey 1994 Revised genera of "Aylacini" including those in North America Tooker et al. 2004 Described Antistrophus rufus species complex including A. jeanae and A. meganae we propose here along with that of Tooker et al. (2004).Antistrophus minor induces inconspicuous, externally imperceptible galls like those of A. rufus in the same host plant species (S. laciniatum).While some authors have discussed these species as potential synonyms (e.g., Beutenmüller 1910a), no nomenclatural actions have been published.Antistrophus minor was excluded from Tooker et al.'s (2004) description of the A. rufus complex, but because its status as a valid species remained unclear, we opted to treat A. minor here.

Gall collection and rearing
In late autumn and winter of 2020, 2021, and 2022, we collected entire, senesced stems of Silphium integrifolium, S. laciniatum, S. perfoliatum, and S. terebinthinaceum from sites in Illinois, Indiana, Iowa, and Ohio (details below).We identified plants in the field based on external morphology, as the Silphium species we studied are easily diagnosable in our sampling region by characters of the leaves and stems that are observable even in senescent plants (e.g., Brock and Weakley 2020).At each site, we confirmed via stem dissection that externally inconspicuous stem galls were present in each host plant species.We then collected additional stems, cut them into evenly sized pieces, labeled them, and stored them in plastic zip bags, which we aerated by poking holes in them using insect pins.We returned stems to lab facilities at The Pennsylvania State University (University Park, Pennsylvania, USA), where we stored bagged stem samples in a barn (without climate controls) to expose occupants to natural environmental conditions.From April through September, we checked stems every two to three days for emerging insects, which we collected into vials containing ethanol and stored at -20 °C.For morphological study, we air-dried and mounted selected individuals.Nastasi (2023) provides further recommendations for rearing cynipid galls on herbaceous plants.
Using Darwin Core biodiversity data standards (Wieczorek et al. 2012), we digitized label data of all specimens that we examined.Digitized specimen data for all individuals we examined, including plant sample numbers and exact emergence dates for newly reared material, are available in Suppl.material 1: table 1.

Morphological character selection, description, and examination
To assess morphological boundaries of species belonging to the rufus complex, we selected 50 morphological characters (Suppl.material 1: table 2) from contemporary works on taxonomy of herb gall wasps (i.e., Nieves-Aldrey 1994; Tooker et al. 2004;Melika 2006;Ronquist et al. 2015;Azmaz and Katilmiş 2020;Buffington et al. 2020;Nieves-Aldrey 2022;Tavakoli et al. 2022).We selected characters based on perceived potential for species-level diagnosis as well as ubiquity across the taxonomic treatments that we considered.We matched anatomical terms for adult wasps to concepts in the Hymenoptera Anatomy Ontology (Yoder et al. 2010) and provide a URI table (see Seltmann et al. 2012) outlining morphological terminology (Suppl.material 1: table 3).Our terminology relating to cuticular surface sculpture follows Harris (1979).We used the following additional abbreviation that is absent from the Hymenoptera Anatomy Ontology: -DLO (diameter of lateral ocellus) for the largest possible diameter of either lateral ocellus.
We evaluated each character for five females and five males of each species ("primary morphological exemplars;" Suppl.material 1: table 1); we putatively identified these individuals based on the associated host plant species and characters presented in the original descriptions and scored the primary type and some secondary types for each species.We selected specimens based on geographic origin and used material from various counties across four states to ensure that we accounted for different populations and intraspecies variability.In addition to "target" specimens that we used in the primary morphological analysis, we examined an additional 20 females and 20 males of each species ("secondary morphological exemplars;" Suppl.material 1: table 1) to confirm the characters given in descriptions and obtain replicate measurements of body length for each species.We provide raw morphological data (Suppl.material 1: table 4), and a summary of characters, states, and corresponding diagnostic utility (Suppl.material 1: table 5).We report morphological characters in taxon treatments as character-state pairs matching the findings of the morphological "test".
We performed morphological observations and measurements of mounted specimens with an Olympus SZX16 stereo microscope (Olympus Life Science, Tokyo, Japan) fitted with an optical micrometer.We measured antennae at a resolution of 0.005 (1/200) millimeters using 10× magnification in combination with the 2× objective.Other measurements were taken at an appropriate magnification using the 1× objective.We used a gooseneck illuminator fitted with mylar strips to diffuse light, which was especially helpful for discerning minute patterns involving sculpture and other aspects of the cuticular surface.
Terms relating to gall phenotypes follow Deans et al. (2023).

Imaging and drawings
We took images of point-mounted adult wasps using a Macroscopic Solutions 'microkit' (Tolland, CT) imaging system.Additional images were captured from mounted wasps using Olympus SZX16 microscope (Olympus Life Science, Tokyo, Japan).We stacked images using Zerene Stacker LLC (Richland, WA), edited them using Adobe Photoshop (Adobe Inc.), and prepared plates using Adobe Illustrator (Adobe Inc.).

Nomenclature of host plants
For taxonomy of Silphium host plants we follow Brock and Weakley (2020).While taxonomy of Silphium is widely considered unsettled, the four species considered here (Silphium integrifolium, S. laciniatum, S. perfoliatum, and S. terebinthinaceum) are robust in the geographic area that we considered (M.Brock, in litt.).

Distribution of gall wasps and host plants
We generated maps showing confirmed and potential distribution of each gall wasp species using MapChart ( figures are licensed via CC BY-SA 4.0).We defined potential distribution of each gall wasp species as the known native range of its host plant; distribution of Silphium species follows Kartesz (2015).

Molecular phylogenetics
We sequenced and analyzed the cytochrome c oxidase subunit I (COI) gene to test the species concepts suggested by our morphological analysis and host plant data.We sequenced the COI gene of three individuals of each species of the Antistrophus rufus complex.We also sequenced single specimens of three additional Antistrophus species: A. silphii Gillette, 1891 reared from a terminal stem gall on Silphium integrifolium Michx.(Fig. 1D), A. laciniatus Gillette, 1891 reared from a flower gall on Silphium laciniatum (Fig. 1F), and A. microseris (McCracken & Egbert, 1922) reared from a stem gall on Microseris douglasii (DC.)Sch.Bip.We also included a single sequence of Isocolus leuzeae Nieves-Aldrey, 2003 from GenBank (DQ012643) to serve as outgroups to the A. rufus complex.
We performed DNA extraction using an E.Z.N.A. Microelute Genomic DNA Kit (Omega Biotek Inc., Norcross, Georgia, USA) following kit protocols and eluting in 30 µL of buffer (15 µL eluted in two steps).We extracted DNA from wasps either entirely destructively, by grinding the entire body, or minimally destructively, by incubating the entire insect.We amplified the COI gene using primer pairs LEPR and LEPF (Hebert et al. 2004) or LCO1490 and HCO2198 (Folmer et al. 1994), following Hebert et al.'s PCR conditions for LEP primers and following those of Pang et al. (2020) for Folmer primers.Sequencing of PCR products was performed using The Huck Institute's Genomics Core Facility (The Pennsylvania State University, University Park, PA).We edited and aligned sequence data using Geneious (Biomatters Ltd., Auckland, New Zealand, http://www.geneious.com/)using the Geneious alignment function.Aligned sequence ends were trimmed to exclude primer sequences.We calculated genetic distances in Geneious using the Tamura-Nei distance model and 1,000 bootstrap replications.We determined the number of variable and parsimony informative sites in our alignment using AMAS v1.0 (Borowiec, 2016).
We estimated phylogenetic relationships using MrBayes v3.2.7a (Ronquist et al. 2012) using the following parameters: ngen=15000000; samplefreq=1000; nchains=4; nruns=3.We used two partitions for the first+second and third codon positions, respectively, and selected nucleotide substitution models for the two partitions using PartitionFinder 2 (Lanfear et al. 2017) using a greedy search (Lanfear et al. 2012) and specifying for models available in MrBayes; PartitionFinder 2 recommended GTR+G for the first+second positions and HKY+G for the third position.We specified I. leuzeae as the outgroup and used a burnin value of 3750000 corresponding to 25% of samples.We visualized our consensus tree using FigTree v1.4.4 (https://tree.bio.ed.ac.uk/software/figtree/).
All new sequences generated for this study were deposited in GenBank, accessions PP739172-PP739189.We provide collection data for sequenced specimens (Suppl.material 1: table 6) and DNA barcode divergence data (Suppl.material 1: table 7).We deposited DNA vouchers and additional material from gall samples and/or rearing events at PSUC.

Identification of additional museum material
After developing diagnostic characters for each studied species of Antistrophus, we identified additional material of the A. rufus species complex from collections we referenced, with the goal of checking previous determinations of identified specimens and identifying undetermined specimens for the first time.

Morphological evaluation
Of the 50 characters that we assessed, 13 had utility as diagnostic characters for at least one species of the rufus complex.Two of the diagnostic characters, both of which concerned dimensions of the first two flagellomeres, were used previously by Tooker et al. (2004) in their treatment of the complex; our observations of those characters were generally consistent with those of Tooker et al. (2004).
While the specimens that we reared from S. laciniatum, S. perfoliatum, and S. terebinthinaceum all matched existing species concepts, the wasps associated with S. integrifolium appeared to compose an independent species, and our morphological observations suggest significant differences between these individuals and those of the described members of the rufus complex.Based on these differences, and the results of our molecular phylogenetic analysis, we describe below the species associated with S. integrifolium as Antistrophus laurenae Nastasi, sp.nov.

Molecular phylogenetics
Our trimmed alignment was 623 bp in length and contained 216 variable sites and 187 parsimony informative sites.Our Bayesian phylogenetic analysis of the 19 sequences (Fig. 2) provided clear evidence of distinct species matching those found by our morphological data, as each of the species had very short branch lengths within species (maximum COI divergence of 1.4% in A. meganae) and very long branch lengths between them.COI species divergences between the closest species pair (A.meganae and A. minor) averaged 10.6% (Suppl.material 1: table 7), well beyond the often contested 2% 'suggestion' for distinguishing species.Each species was supported by a posterior probability of 1, and A. laurenae and A. jeanae were strongly supported as sister species, but the remainder of the between-species relationships are less robust.Each recognized species has a distinct host plant except for A. rufus and A. minor, which are both associated with Silphium laciniatum but were not supported as sister species.Our analysis also retrieved the Antistrophus rufus complex as monophyletic within the sampled species of Antistrophus with a posterior probability of 1.
Taxonomy of the Antistrophus rufus species complex

Antistrophus rufus species complex
Diagnosis.Overall, the A. rufus complex is best diagnosed by the following combination of characters: head and mesosoma mostly reddish-brown in color (Figs 3A-D); facial radiating striae complete (Fig. 3A); compound eye longer than malar space in anterior view (Fig. 3A); F2 much longer than F1 (e.g., Fig. 4B); mesopleuron sculpture reticulate with fine intermediate striae (Fig. 3B); notauli incomplete, indistinct in anterior third of mesoscutum (Fig. 3C); female fore wing without apparent marginal setae (Fig. 3D); galls inconspicuous, induced in stems of Silphium species (e.g., Fig. 4F).Comparative diagnostic characters for the Antistrophus rufus species complex relative to other Antistrophus are presented in Fig. 3 and Table 2 below.Table 2. Diagnostic characters of the Antistrophus rufus complex.Morphological data for species outside the A. rufus complex are from ongoing revisionary work on Aulacideini (Nastasi et al., unpublished data).Biological data are from Nastasi and Deans (2021).* = Gall and host unconfirmed for A. bicolor Gillette, 1891; see Nastasi and Deans (2021).† = Coloration of fresh material unknown; based on description by Beutenmüller (1908).Remarks.The Antistrophus rufus complex includes all species of Antistrophus known to induce inconspicuous galls in stems of Silphium species.While other species of Antistrophus induce galls on Silphium, they induce perceptible galls on the apical stems or in the flowerheads (Gillette 1891;Nastasi and Deans 2021) and are generally distinct morphologically (Table 2).Other species of Antistrophus induce galls on different host genera.A complete revision of Antistrophus and other North American Aulacideini is being undertaken by the authors (Nastasi et al., unpublished data).

Antistrophus jeanae Tooker & Hanks, 2004
Fig. 4 Antistrophus jeanae Tooker & Hanks, 2004in Tooker et al. 2004: 132    Diagnosis.A. jeanae is most similar to A. laurenae sp.nov.but is best distinguished by the dimensions of F2 in females, which in A. jeanae is 3.8× as long as wide (Fig. 4B) but 3.3× as long as wide in A. laurenae (Fig. 5B).The sculpture of the mesoscutellar disc (Fig. 4F) is also useful; the mesoscutellar disc is more or less rugose-reticulate throughout in A. jeanae but only has rugose-reticulate sculpture toward the outer margins in A. laurenae.
Distribution.Tooker et al. (2004) reported this species only from several prairie sites in Illinois (USA).Nastasi and Deans (2021) did not report additional localities.However, specimens we examined revealed records from three additional state records: Indiana, Iowa, and Wisconsin.A verifiable iNaturalist observation (https://www.inaturalist.org/observations/114414672)also records this species from Ohio; we have since examined adult gall wasps reared from the plant material from the same site, which confirmed their suspected identity (in litt.).Lastly, a specimen identified during this study (USNMENT 01822302; see complete specimen data in Suppl.material 1: table 1) confirms the occurrence of this species in Missouri (Columbia, Boone County) (see discussion).Known and potential distribution are summarized in Fig. 9.   Description.Female (Fig. 5A)-Body length: 1.8-2.6 mm (x -= 2.2 mm; n = 25; holotype = 2.6 mm).Color: Antenna color: red brown throughout, at most slightly darker distally than proximally.Head color: vertex and occiput dark red brown, mandi-bles red brown basally to darker red brown apically, rest of head red brown throughout.Mesosoma color: pronotum and propodeum red brown laterally to dark red brown medially, mesoscutum dark red brown, scutellum dark red brown, and mesopleuron dark red brown dorsally and ventrally but red brown medially.Wing membrane color: hyaline throughout.Wing vein color: light brown.Leg color: red brown throughout, except for apical tarsomere which is conspicuously darker.Metasoma color: red brown to dark red brown.Antennae (Fig. 5B): Antennomere count: 13.F1 length: 2.2× as long as wide.F2 length: 3.3× as long as wide.F2:F1 length ratio: 1.4.Placodeal sensilla on F2: absent; sensilla present only on F3 and following antennomeres.Head (Fig. 5C): Upper frons sculpture: reticulate.Gena posterolateral projection in anterior view: distinctly projecting past compound eyes.Facial radiating striae: distinct and complete, reaching compound eyes.Supraclypeal area sculpture: reticulate.Supraclypeal area projection: slightly projecting.OOL vs POL: OOL distinctly longer.OOL vs LOL: OOL greater than twice LOL.POL vs LOL: POL greater than twice LOL.LOL vs DLO: LOL longer.Vertex sculpture: reticulate throughout.Clypeus sculpture: reticulate.Mesosoma (Figs 5D-F): Pronotum pilosity: densely pilose along anterior margin and with only sparse setae elsewhere.Pronotum excluding pronotal plate sculpture: reticulate.Pronotal plate sculpture: reticulate.Mesopleuron excluding speculum sculpture: reticulate with fine intermediate striae.Speculum sculpture: reticulate.Mesopleuron pilosity: ventral margin and mesopleural triangle densely pilose and bare elsewhere.Mesoscutum pilosity: sparsely pilose.Mesoscutum sculpture: reticulate.Apparent length of anterior parallel lines: reaching one third across mesoscutum.Morphology of anterior parallel lines: narrow, distinct throughout perceptible length.Apparent length of parapsidal grooves: reaching halfway across mesoscutum.Morphology of parapsidal grooves: narrow, distinct throughout perceptible length.Morphology of median mesoscutal impression: apparent as a shallow impression extending across most of mesoscutum.Notauli completeness: incomplete, distinct posteriorly to indistinct in anterior third.Morphology of notauli: appearing as wide indentations with sloping edges throughout distinct portions.Metapleural sulcus: meeting posterior mesopleuron at about one third of its height.Lateral propodeal carinae: distinct and subparallel.Metapleuron sculpture: reticulate.Mesoscutellar foveae distinction: distinct, relatively deep anteriorly to shallower and inconspicuously delimited posteriorly.Mesoscutellar foveae sculpture: reticulate.Mesoscutellar disc sculpture: rugosereticulate, primarily reticulate, but with distinct rugose-reticulate sculpture seemingly restricted to outer margins.Mesoscutellar foveae length: long, reaching about one third across mesoscutellar disc.Mesoscutellar foveae shape: subquadrate, about as wide as long, separated by a narrow linear carina.Mesoscutellar disc shape: subcircular, about as wide as long.Wings: Marginal cell length: 3.1× as long as wide.Fore wing distal fringe of marginal setae: absent.Metasoma: Punctation of metasomal terga: T3 punctate only in posterior third and with T4 and following punctate throughout.
Etymology.Named for Lauren Ahlert, a biology teacher at Wayne Valley High School in Wayne, New Jersey, USA, who has served as a tremendous source of inspiration and passion for the author of this species.
Distribution.Material examined in this study reveals that this species occurs in Illinois, Iowa, and Wisconsin (USA; see Suppl.material 1: table 1).A verified iNaturalist record (https://www.inaturalist.org/observations/136446787)also suggests that this species occurs in Mississippi.Known and potential distribution are summarized in Fig. 9.

Antistrophus meganae Tooker & Hanks, 2004
Fig. 6 Antistrophus meganae Tooker & Hanks, 2004in Tooker et al. 2004   Diagnosis.A. meganae is the only species of the rufus complex in which the mesoscutellar foveae (Fig. 6F) are long and ovate, reaching nearly halfway across the mesoscutellar disc.A. meganae also differs from most other members of the rufus complex by the POL (Fig. 6C), which is longer than the OOL in A. meganae and A. minor but shorter than the OOL in the other species.
Distribution.Tooker et al. (2004) reported this species only from several prairie sites in Illinois (USA).Nastasi and Deans (2021) did not report additional localities; however, the specimens we examined revealed novel records from Wisconsin (Suppl.material 1: table 1).We also sequenced DNA barcodes for individuals from Ohio (Suppl.material 1: table 6), and a series of specimens identified using the key represents a new state record from Michigan (see Discussion; Suppl.material 1: table 1).Known and potential distribution are summarized in Fig. 9.
Antistrophus minor Gillette, 1891 Fig. 7 Antistrophus minor Gillette, 1891: 195.♀, ♂ (type locality: unknown location in Illinois, USA).Aulax gillettei (Gillette, 1891): Kieffer 1902: 93.Frison [1927]   Diagnosis.A. minor is the only described species of Antistrophus in which the mesoscutellar disc is strongly ovate and distinctly wider than long (Fig. 7F).Additionally, A. minor is easily separable from A. rufus (which also induces inconspicuous galls in stems of the S. laciniatum) and A. meganae by the shape and length of the mesoscutellar foveae; in A. minor, the mesoscutellar foveae are subquadrate and reach about one third across the mesoscutellar disc (Fig. 7F) but are shorter and subrectangular in A. rufus (Fig. 8F) and longer and ovate in A. meganae (Fig. 6F). A. minor and A. meganae are also the only species of the rufus complex in which the POL is longer than the OOL.See additional comments in the diagnosis to A. rufus.
Distribution.Antistrophus minor was described from adults reared alongside A. rufus in Illinois.We examined specimens providing new state records from Iowa and Wisconsin.Known and potential distribution are summarized in Fig. 9.It appears that A. minor is likely sympatric with A. rufus throughout the distribution of S. laciniatum, although further rearing of Silphium species will be needed to better understand the role of geography in the distribution of Antistrophus, especially regarding these two species.
Remarks.See remarks under Antistrophus rufus for notes on generic nomenclature and distinction of A. minor from A. rufus.

Antistrophus rufus Gillette, 1891
Fig. 8 Antistrophus rufus Gillette, 1891: 195.♀, ♂ (type locality: unknown location in Illinois, USA).Aulax rufa (Gillette, 1891)  Diagnosis.A. rufus is easily recognized amongst members of the rufus complex by the shape and dimensions of the mesoscutellar foveae (Fig. 8F), which are short and subrectangular in A. rufus but longer and subquadrate or ovate in the other species.The narrow, well-defined notauli (Fig. 8E) are also characteristic, as the other species all have wider and less-defined notauli.The smallest A. rufus specimens may exhibit slightly wider notauli, but the short, rectangular scutellar foveae are always apparent and readily distinguish this species from others. A. rufus also has the shortest F2 relative to its width in females (2.8× as long as wide; Fig. 8B)).
A. rufus are commonly reared alongside A. minor from inconspicuous, externally imperceptible galls in stems of Silphium laciniatum L.; these species are separated by the characters given above and those in the diagnosis of A. minor.
Biology.Antistrophus rufus induces inconspicuous, externally imperceptible galls in stems of Silphium laciniatum (Fig. 8H) (Tooker et al. 2004;Nastasi and Deans 2021).Records associating A. rufus with other host plants most likely represent distinct species including the other species addressed in the present work.
Distribution.Antistrophus rufus was described from material collected in Illinois (USA) and has since been reported from Kansas (Tooker et al. 2004;Nastasi and Deans 2021).Additional material that we examined revealed additional records from Iowa and Wisconsin, and reiterated occurrence of this species in Kansas (Suppl.material 1: table 1).Known and potential distribution are summarized in Fig. 9.
Remarks.Kieffer (1902) synonymized Antistrophus with Aylax (=Aulax) Hartig, 1840.Some authors (e.g., Beutenmüller 1910) follow Kieffer's synonymy, but others appear to have rejected or otherwise omitted this change from many works treating the group (e.g., Weld 1951;Nieves-Aldrey 1994).This generic synonymy and the resulting species synonyms were accidentally omitted from Nastasi and Deans' (2021) catalogue, and we are unable to locate a nomenclatural act reinstating Antistrophus as a valid genus.The genus Aylax has now long been regarded to include only two species of gall wasps inducing galls on Papaver (Papaveraceae) (Nieves-Aldrey 1994; Ronquist et al. 2015).Additionally, the two genera are now placed in different tribes: Aylax is now placed in the tribe Aylacini, while Antistrophus is placed in the Aulacideini.We regard here Antistrophus as a valid genus independent of Aylax, although Antistrophus represents a heterogeneous assemblage as currently defined based on preliminary data from an ongoing revisionary study (Nastasi et al., unpublished data).
Lastly, two specimens labeled as paralectotypes in the USNM collection (USN-MENT 01823001; 01823000) do not appear to be Antistrophus rufus and more closely resemble A. minor.They are both in relatively poor condition; at this time, their species identities cannot be substantiated due to damage to multiple body parts.

Key to the species of the Antistrophus rufus complex
Our assessment of diagnostic characters allowed for development of a key to the species of the rufus complex.A key to all species of Antistrophus is beyond the scope of this work but will be prepared as part of ongoing revisionary studies of North American herb gall wasps (Nastasi et al., unpublished data).For best results using this key, we recommend using light diffusion in combination with high magnification (see Methods).
Before running a specimen through the below key, one should confirm that it is indeed an Antistrophus species using the keys in Nastasi et al. (2024) and/or Buffington et al. (2020) where appropriate, and then confirming that the specimen belongs to the A. rufus complex using the diagnostic criteria given in the diagnosis to the A. rufus complex and in Table 2

Discussion
Our findings suggest that the species belonging to the A. rufus complex are not truly cryptic, and are readily identifiable with appropriate diagnostic characters.Other species complexes of herb gall wasps might benefit from a similar investigation of morphological species boundaries, especially species of Aulacidea Ashmead inducing galls on Lactuca L. or other Antistrophus inducing externally perceptible galls on terminal stems and in flowers of Silphium.Such analyses would supplement the original descriptions and would help determine whether true cryptic species occur in Aulacideini.
While the taxonomic impediment impacts a tremendous number of species across all taxa, especially in groups such as "little brown beetles," microlepidoptera, and many groups of parasitic wasps, our study of the A. rufus complex demonstrates that such issues can be disentangled with dedicated revisionary approaches.While no number of characters can ensure discovery of novel diagnostic characters, attempts to evaluate larger numbers of characters, as attempted here, may help decipher boundaries of cryptic species that would otherwise require molecular characters or additional methods.Examining 50 characters proved sufficient for the rufus complex, but for other taxa, it could be useful to apply a "brute force" approach involving a much greater number of candidate morphological characters.In such a scenario, different groups of characters could be tested iteratively until robust species diagnoses are achieved.
After developing the key to the rufus complex that we present above, we keyed and identified over 50 additional specimens belonging to the rufus complex from the INHS and USNM collections (Suppl.material 1: table 1), including both already determined and previously unidentified material.One series of specimens deposited at USNM represents a new state record of A. meganae from Michigan; these specimens' labels suggesting they were collected off "elecampane" (presumably referring to Inula helenium L.), but this likely constitutes a misidentification of the proper host plant Silphium terebinthinaceum which is rather similar visually.We found a single specimen of A. jeanae amongst unsorted material in the USNM collection (USNMENT 01822302); this specimen was collected in Missouri and represents a new state record (see also remarks for A. jeanae).
Our examination indicates that A. minor is truly a species distinct from A. rufus despite previous suggestion that they may be synonymous.In various cynipid taxa, including Aulacideini, a single host plant species is galled by several distinct species of gall wasps (Nieves-Aldrey et al. 2004;Melika 2006;Nastasi and Deans 2021;Nastasi et al. 2024).However, few (if any) gall wasps share the same host plant while inhabiting the same plant organ and the same portion of gall phenospace (i.e., host plant organ combined with gall phenotype).While most of the A. rufus complex species each occupy a unique host plant species, the Antistrophus rufus complex appears to be somewhat unique among gall wasps, as two species both induce inconspicuous galls within the stems of the same host plant S. laciniatum.This suggests speciation mechanisms may be involved (e.g., geographic isolation or host switching) that may not follow the typically observed speciation in gall wasps, which normally is associated with shifts between host plants or gall morphotypes, at least in oak gall wasps (Cynipini) or their inquilines (Synergini sensu lato) (e.g., Ward et al. 2020Ward et al. , 2022Ward et al. , 2024)).
Regarding evolution of host plant associations, the topology of our Bayesian analysis provides suggestion of cospeciation of A. rufus complex species with their host plants, as there was sorting of A. rufus complex species by host plant lineage (Fig. 2).Clevinger and Panero's (2000) phylogenetic analysis of ETS and ITS markers suggested two major clades of Silphium; the rufus complex includes species associated with both sections.Their sect.Composita includes species with deep taproots and prominent basal rosettes, including S. laciniatum and S. terebinthinaceum, while their sect.Silphium is composed mostly of species with more fibrous roots and without prominent basal rosettes, including S. integrifolium and S. perfoliatum.In our phylogenetic analysis, species galling sect.Silphium (A. jeanae and A. laurenae) formed a well-supported clade sister to the three species galling sect.Composita (A. meganae, A. minor, and A. rufus), although the latter clade was less strongly supported (posterior probability = 0.89).
More generally, our data show that each studied Silphium species has distinct species of Antistrophus occupying their stems.While the present study focused on the "big four" species of Silphium ubiquitous in prairies in the Midwestern U.S., a recent taxonomic treatment (Brock and Weakley 2020) suggests there are as many as 23 Silphium species; it is likely that more species of the rufus complex or other Antistrophus will be revealed in these other species.The deep divergences in DNA barcodes among these gall wasps suggest a long history of isolation and may indicate that additional species of the A. rufus complex await discovery and description.While interspecific divergences in our DNA barcode data are much higher than the supposed 2% "barcode gap" between animal species, other studies of diverse insect groups often find divergence in excess of 10% between congeneric species (Pentisaari et al. 2014;Lin et al. 2015;Song et al. 2018).Locating additional undescribed members of the rufus complex could aid in resolving relationships amongst the known species, and future research on host plant affinity and co-phylogeny between Silphium and their gall wasps will help inform understanding of these processes.

Figure 1 .
Figure 1.Examples of Silphium-galling Antistrophus wasps and their galls A adult female A. jeanae Tooker & Hanks, 2004 B galls and larvae of A. jeanae in stem pith of S. perfoliatum L. C adult female A. silphii Gillette, 1891 D galls of A. silphii on terminal stem of S. integrifolium Michx E adult female A. laciniatus Gillette, 1891 F galls of A. laciniatus in disc flower of S. laciniatum L. adult gall wasps photographed by Antoine Guiguet.Galls photographed by Andrew R. Deans.

Figure 2 .
Figure 2. Bayesian phylogenetic tree of DNA barcodes for the Antistrophus rufus complex.Support values shown only for nodes with posterior probability <1.Host plant clade associations are indicated by colored boxes (Silphium sect.Silphium in orange and S. sect.Composita in green) A adult female Antistrophus jeanae Tooker & Hanks, 2004 B adult female Antistrophus rufus Gillette, 1891.

Figure 9 .
Figure9.Known and potential distributions for species of the Antistrophus rufus complex.Host plant distribution followsKartesz (2015).States in green reflect established distribution, while states in yellow reflect potential distribution based on records of the host plant.