Six in one: cryptic species and a new host record for Olixon Cameron (Rhopalosomatidae, Hymenoptera) revealed by DNA barcoding

Olixon testaceum is a widely distributed species of brachypterous parasitoid wasp (Vespoidea: Rhopaloso-matidae) occurring in Meso-and South America, but little is known of its biology. Here, the first known host of O. ? testaceum is identified as the cricket Anaxipha sp. (Grylloidea: Trigonidiidae) through DNA barcoding of six Olixon larvae and their hosts. Barcoding results also indicated substantial genetic diversity within nominal O. testaceum specimens. The number of species and statistical significance of these groups were tested using Maximum Likelihood phylogenies, distance-based cluster analyses, and coalescence models. All analyses revealed at least six distinct lineages, which suggests six or more cryptic species within O. ? testaceum . Combined with what is currently known about Rhopalosoma host use, these results indicate that rhopalosomatids may be generalist rather than specialist parasitoids, and further confirm the benefits of open global collaboration and DNA barcoding in advancing taxonomic knowledge.


Introduction
In recent decades, new discoveries have greatly increased our knowledge of the diversity, systematics, and behavior of Olixon Cameron, 1887-an historically understudied genus of cricket-assassin wasps (Vespoidea: Rhopalosomatidae) (Townes 1977) (Fig. 1).These unusual brachypterous wasps are rarely seen alive but are now being collected in substantial numbers in pitfall and Malaise traps around the world (Mayhew and Dytham 2008;Lohrmann et al. 2012).These specimens have led to new species descriptions (e.g., Lohrmann and Ohl 2007;Krogmann 2009;Lohrmann et al. 2012;Bulbol et al. 2023), new distributional records (e.g., Ramsdell and Taylor 2006;Wood and Maupin 2007), and behavioral notes (Lohrmann et al. 2014), as well as discussions of Olixon biogeography and diversification (Krogmann 2009).However, knowledge of their biology remains limited.Here, we report the cricket Anaxipha sp.(Grylloidea: Trigonidiidae) as the first confirmed host of O. ?testaceum and discuss evidence that six cryptic species are included within the nominal species O. testaceum in Meso-and South America.
Despite their brachypterous wings which would seem to limit long-range dispersal, Olixon testaceum, as currently understood, is among the most widespread of all rhopalosomatids, occurring throughout Meso-and South America, from Argentina to Arizona (Lohrmann et al. 2012).Specimens are most often collected in Malaise traps and can be found in diverse habitats including rain forests, dry forests, prairies, and cultivated landscapes.The broad distribution of O. testaceum coupled with the recent discovery of sympatric cryptic species of Rhopalosoma nearcticum in the USA (Miller et al. 2019) made O. testaceum an excellent candidate for a species delimitation study using sequence records from the Barcode of Life Database (BOLD; www.boldsystems.org).
An initial search of publicly available rhopalosomatid records revealed 221 sequences from specimens identified as O. cf.testaceum using the key to species in Lohrmann et al. (2012).Remarkably, four sequences from unidentified rhopalosomatid larvae were also included among these records (Fig. 2).Most of the O. cf.testaceum and all the larvae originated from the ongoing BioAlfa inventory of the Area de Conservación Guanacaste (ACG) in northwestern Costa Rica in Sector Santa Rosa (Janzen 1986;Janzen and Hallwachs 2016;Janzen et al. 2009).Given the biological significance of new host records for Rhopalosomatidae, our objectives were to 1) identify the unknown larvae to species by placing them within a phylogeny of Rhopalosomatidae, 2) identify their host species by searching for associated specimens within BOLD and by generating new barcode sequences as needed, and 3) explore the genetic diversity of O. testaceum for evidence of cryptic species.

Methods
Specimens from each sequence cluster in this study were identified to genus using Townes' (1977) key to rhopalosomatid genera and the specimen photos available through BOLD.The subsequent assignment of Olixon specimens to the nominal taxon O. testaceum was based on the presence of the following combination of characters that distinguishes this species: a more or less uniform testaceous to pale brown coloration with the exception of a dark marking on metasomal segment II, the presence of a malar sulcus, a short temple, and a strong and complete carina between the posterolateral processes of the propodeum (Lohrmann et al. 2012).
The publicly available records (n = 221) were combined with additional private sequences of rhopalosomatids made available by JS, PH, DJ, and WH to ensure maximum coverage, including another two larvae (for a total of six).Specimens were predominantly collected via weekly Malaise trap samples (n = 398) between 2012 and 2020.All but 10 specimens (sourced from GenBank) were sequenced at the Centre for Biodiversity Genomics, and most sequences (n = 309) were generated by a Sequel (Pacific Biosciences) high-throughput sequencer, while 95 were analyzed using Sanger sequencing (https://ccdb.ca/resources/).All sequences, specimen images, and collection data are available in the dataset "DS-RHOP" on BOLD.
An initial Maximum Likelihood (ML) phylogeny of Rhopalosomatidae was created using RAxML (v.8.2.12) (Stamatakis 2014) through the CIPRES Science Gateway (Miller et al. 2010).Alignments were partitioned by codon position and analyzed using the GTRCAT model of nucleotide substitution.Statistical significance was analyzed using 1000 bootstrap (BS) pseudoreplicates.This tree (not shown) included 424 specimens, of which 221 represented O. testaceum and six were larvae.All six larvae were located within clades of nominal O. testaceum, five in one clade and one in another.In total, the specimens of O. testaceum formed six distinct clades, one of which contained 206 of the 221 specimens.As this group contained very little genetic diversity (average genetic distance = 0.0099), subsequent analyses reduced the number of specimens for this clade was reduced from 206 to five representatives.
To narrow the focus to only larvae and potential cryptic species of O. testaceum, a new ML phylogeny was created using all six larvae, specimens from each apparent O. testaceum clade, all other Olixon specimens available (five in total), and four outgroup specimens (two each for Liosphex Townes, 1977 andRhopalosoma Cresson, 1865).Sequences were aligned using MAFFT v.7.450 (Katoh et al. 2002;Katoh and Standley 2013) in Geneious Prime® 2020.0.2 (Biomatters, Auckland, NZ).The presence of pseudogenes in the alignment was checked via translation to amino acids.No stop codons were present.The ML phylogeny was reconstructed as described above.Specimen IDs, collection information, and tentative species groups are found in Table 1.Intra-and interspecific genetic distances (K2P) were calculated in MEGA 11 (Tamura et al. 2021).
To test for potential cryptic species diversity within O. testaceum, both distance-based cluster analyses and phylogenetically informed tests of species delimitation were employed.Assemble Species by Automatic Partitioning (ASAP) (Puillandre et al. 2021) is a distance-based method that tests various species hypotheses using the intra-and intergenetic distance scores for putative species to calculate a custom ASAP score for each hypothesis (lower scores = more statistical robustness).The online version of ASAP (https:// bioinfo.mnhn.fr/abi/public/asap,last accessed Dec. 12, 2022) was used under default settings with genetic distances calculated using the K2P model of molecular evolution.
Additional distance-based cluster analyses were carried out in R (Paradis et al. 2005;R Core Team 2017;Paradis and Schliep 2018).The "dist.DNA" function with model = "TN93" was used to calculate a corrected distance matrix from the imported nucleotide alignment.TN93 was used as it most closely approximates the more complex GTR model used in the ML analyses which is unavailable using dist.DNA.Hierarchical clustering, which iteratively combines taxa with minimal dissimilarity to create a dendrogram of potentially statistically significant clusters, was performed within the function "parPvclust" using the correlation method and average linkage agglomeration (Suzuki and Shimodaira 2006).The statistical significance of the clusters was confirmed by 1000 bootstrap replicates and the calculation of Approximately Unbiased p-values (AU) and Bootstrap Probability (BP) values as suggested by Suzuki and Shimodaira (2006).The distance matrix was also analyzed using partitioning around medoids, which creates and scores clusters by collapsing the distance of intra-cluster points to a hypothetical centroid.Statistical significance is inferred from average silhouette width of clusters (>0.5 is considered significant) (Kaufman and Rousseeuw 1987).(The "pam" function (Schubert and Rousseeuw 2019) and variable k values ranging from 2 to 8 were used.Results were visualized using "fviz_silhouette" (Kassambara and Mundt 2020).
Phylogenetically informed species delimitation methods were also used.The Multi-rate Poisson Tree Process (mPTP) introduced by Kapli et al. (2017) uses maximum likelihood and is based on a single-locus coalescent-based method.The online version of mPTP (https://mptp.h-its.org/#/tree,last accessed Dec. 12, 2022) was used with default parameters.For a Bayesian analysis, the Bayesian Poisson Tree Process (bPTP) program of Zhang et al. (2013) (https://species.h-its.org/ptp/,last accessed Dec. 12, 2022) was used.The rooted phylogeny was uploaded and analyzed for 250,000 MCMC generations, thinning was set to 150, and the first 25% were discarded as burn-in.Within the program, a simple heuristic search determined the most likely number of species represented on the tree according to the most supported partition.Convergence was verified by visually checking the likelihood plot.
High throughput sequencing (HTS) has the advantage of generating sequences of biota associated with the target specimen.This property enables the identification of potential host-parasitoid interactions, predator-prey relationships, pathogen infections, etc.If a specimen yields more than one sequence contig, the additional contig(s) can be examined for biologically relevant associations.To identify the host for each larva, we compared any additional sequence information generated by the HTS to BOLD and generated a new barcode record when a potential orthopteran host sequence was found.

Results
Thirty-four sequences were used to generate the ML tree (Fig. 3).All sequences are publicly available in BOLD (dataset "DS-RHOP") and identification codes are listed in Table 1.Within Olixon, two major lineages were recovered, one including 25 specimens of O. testaceum and the other primarily composed of O. banksii specimens from the USA.The single specimen of Olixon from Australia was recovered as sister to O. banksii.Six clades were well-resolved (BS = 100) within O. testaceum (informal species IDs are included in Table 1).All rhopalosomatid larvae from Costa Rica were recovered as members of the group "O.?testaceum sp.6", thus confirming Anaxipha (Trigonidiidae) as a common host for Olixon.All novel or newly associated host specimens of Olixon are summarized in Table 2.
One larva (BIOU70270-H11) was collected in Suriname.It was found without a host and no associated specimens are confirmed at this time.The other five larvae were collected in ACG, Costa Rica.Two (BIOUG55891-B08_parasite and BIOUG63752-H08) were still attached to their host and barcodes for all four specimens were recovered.Both hosts were identified as Anaxipha sp.(Trigonidiidae).Three specimens (BIOUG59841-H04, BIOUG59151-H08, and BIOUG58943-C02) were not attached to a host, likely reflecting their detachment upon exposure to the ethanol in the Malaise trap.Potential orthopteran hosts associated with these samples (i.e., those collected from the same site, date, and trap) were all identified via barcodes as trigonidiid crickets.
All statistical analyses confirmed the presence of at least six distinct lineages within this dataset (Fig. 3), a conclusion reinforced by their differing Barcode Index Numbers on BOLD which often correspond to species (Ratnasingham and Hebert 2013).Hierarchical clustering results and PAM scores (Fig. 4) were statistically significant (AU scores >95, average silhouette width = 0.87).The lowest ASAP-score belonged to the   partition assuming six hypothetical species (ASAP-score = 1.0; threshold distance = 3.2%).The mPTP suggested six unique lineages, while the bPTP results supported at least six, and possibly up to nine species within O. testaceum, although species 7-9 were recovered with very low support.Genetic distances between and within species (Table 3) show very little intraspecific variation per clade (average = .005)and substantial interspecific variation (average = .106).

Discussion
Our analyses indicate several genetic lineages fall within the morphological variation of "Olixon testaceum."However, since the holotype has not been barcoded, we are unable to link any particular genetic lineage with the "real" O. testaceum.To indicate this uncertainty and acknowledge that future morphological work is needed to formally describe these cryptic species, we refer to the species group collectively as "O.?testaceum." Although O. ?testaceum is one of the most widespread morphospecies of Rhopalosomatidae, very little is known about its biology (Lohrmann et al. 2012).Here, we have confirmed one new host record for the genus Olixon and added several new associated host records for the O. ?testaceum species group.At the family level, a new host record for O. ?testaceum adds a fourth confirmed host for rhopalosomatid species (Lohrmann et al. 2014;Miller et al. 2019).
Olixon is now known to parasitize species within two subfamilies of crickets: Trigonidiinae (herein, Perkins 1908) and Nemobiinae (O.banksii, Lohrmann et al. 2014).Similarly, the only other rhopalosomatid with known hosts, Rhopalosoma ?nearcticum, parasitizes both Trigonidiinae and Podoscirtinae crickets (Miller et al. 2019).The finding of two different subfamilies of crickets as hosts each for Olixon and Rhopalosoma supports the hypothesis that rhopalosomatids, at least at the genus level, are generalist rather than specialist parasitoids.
Furthermore, the identification of Anaxipha as a host of O. ?testaceum is quite remarkable as Anaxipha are also among the known hosts of the distantly related R. ?nearcticum (Miller et al. 2019).Olixon is hypothesized to be the basal branch of Rhopalosomatidae and sister to a clade comprising all recent macropterous forms (Guidotti 1999;Lohrmann et al. 2020) (Fig. 5).The basal position of Olixon coupled with the fact that these two rhopalosomatid genera (Olixon and Rhopalosoma) are not sister taxa but utilize species of the same genus of crickets as hosts may suggest Trigonidiinae as the ancestral host for Rhopalosomatidae.However, other groups of crickets (i.e., Nemobiinae and Podoscirtinae) are also used as hosts by rhopalosomatids and we still know relatively little about host use throughout the family.It is possible the shared trigonidiine host of Olixon and Rhopalosoma is an example of convergent evolution rather than a plesiomorphy.Unfortunately, hosts for Paniscomima Enderlein, 1904 and Liosphex are still unknown, but as seen here, large-scale Malaise trap sampling programs are excellent sources for rhopalosomatid adults, larvae, and hosts that may soon fill the gaps in our knowledge of rhopalosomatid biology and evolution of host use.
While a new host record and discovery of cryptic species is significant, there is still much work to be done.Future efforts relating to O. ?testaceum should investigate morphological or ecological differences that might further distinguish clades from one another.A previous study in the ACG (Hebert et al. 2004), discovered that the target species Telegonus (previously Astraptes) fulgerator (Walch, 1775) contained at least ten cryptic species with defining variation in morphology and host plant preference.Future efforts should investigate whether O. ?testaceum displays similar variation, especially considering its widespread range.Brief comparison of the six clades revealed by this study showed that O. ?testaceum sp. 1 from Honduras is easily characterized by its distinct wings which seem, in terms of the grade of their reduction, intermediate between O. melinsula Lohmann et al., 2012 and O. ?testaceum (Fig. 6).Formal species descriptions of these genetic clades should employ an integrative taxonomy approach (e.g.Padial et al. 2010)-adding morphological, behavioral, and host data to the genetic data presented here.
Such integrative taxonomic research may reveal more cryptic species within O. testaceum beyond those discovered here.In their revision of the New World Olixon, Lohrmann et al. (2012) investigated several hundred specimens assigned to O. testaceum originating from the southern United States (Arizona) to northern Argentina.The cryptic diversity of O. testaceum reported here is certainly an underestimation of this clade's true diversity since all but one of the specimens analyzed were from Honduras and Costa Rica.The holotype of Olixon testaceum was collected in Bugaba, Panama (Cameron 1887), but it is not clear whether its barcode would match any of the clades examined in this study.Furthermore, the slightly different color pattern of Saphobethylus pallidus Kieffer, 1911 from Teapa in Mexico, currently treated as a synonym of O. testaceum (e.g., Turner and Waterston 1917;Townes 1977;Lohrmann et al. 2012), supports the hypothesis that it too represents a distinct species.Rhopalosomatidae appears to be a hotspot for cryptic species diversity.Our ML tree included four specimens of nominal O. banksii specimens from Iowa, Texas, Virginia, and Oklahoma.Three specimens from IO, TX, and VA showed low intra-group variation (patristic distance = 0.032), while the average nearest-neighbor distance be-tween these three and a specimen from OK was significantly higher (patristic distance = 0.234).Another closely related species, O. melinsula, is currently known only from Texas to Florida (along the Gulf of Mexico) and southern Paraguay.This distribution pattern may well represent "two sibling species, so similar as to be indistinguishable at the moment" (Lohrmann et al. 2012).Both O. banksii and O. melinsula are promising groups for future investigations of cryptic species diversity among the cricket-assassin wasps.
The utility of DNA barcoding to identify and reveal cryptic species complexes is well established across a wide range of taxa and biomes (e.g., Hebert et al. 2003;Janzen et al. 2005;Sáez and Lozano 2005;Bickford et al. 2007).Many of the early case studies that applied barcoding to discover cryptic diversity exposed many undescribed species (e.g., Hebert et al. 2004;Smith et al. 2006;Smith et al. 2008).DNA barcoding offers solutions to many of the limitations of a morphologically dependent taxonomic system.Traditional methods for classification require much time and expertise (Stoeckle and Hebert 2008).As barcoding gains wide adoption by both taxonomists and ecologists (Valentini et al. 2009), open collaboration and accessibility to sequences are essential.In order to aid access, BOLD has consolidated this information into a public database with over 15 million specimen records for over 330,000 named species and more than a million putative species.
The taxonomic impediment is a significant and well-known problem in entomology (e.g., Agnarsson and Kunter 2007;Engel et al. 2021) and the results of our study highlight the importance of BOLD's open access policies and friendly collaboration among researchers for the future of insect systematics.In this case, exploring BOLD as part of an undergraduate class revealed unexpected and interesting discoveries in rhopalosomatid host use.New collaborators enthusiastically joined the project, freely sharing specimens, photographs, and their expertise in rhopalosomatid systematics and morphology, DNA barcoding, and species delimitation.We hope future researchers will be similarly generous and collaborative across disciplines and skill levels, allowing for a deeper scientific understanding of the diversity of life and inspiring the next generation of insect taxonomists.

Figure 3 .
Figure 3. Maximum Likelihood phylogeny of Olixon testaceum.Bootstrap support >75 is shown.Putative cryptic species of O. testaceum are labeled.Vertical bars indicate statistically significant species groups identified by Hierarchical clustering (HCL), Assemble Species by Automatic Partitioning (ASAP), Multi-rate Poisson Tree Process (mPTP), and Bayesian Poisson Tree Process (bPTP).Light orange bars for bPTP are not significant.

Table 1 .
Specimens used in phylogenetic and statistical analyses.Process and Voucher IDs from BOLD: www.boldsystems.org.