Description of a Neotropical gall inducer on Araceae: Arastichus , gen. nov. (Hymenoptera, Eulophidae) and two new species

A new genus of a Neotropical gall inducing tetrastichine eulophid on Araceae is described and confirmed using Ultraconserved Elements (UCE) phylogenomic data. Arastichus Gates, Hanson, Jansen-González & Zhang, gen. nov. , includes two new species and one species transferred from Aprostocetus Westwood: A. capipunctata Gates, Hanson, Jansen-González & Zhang, sp. nov. , A. gallicola (Ferrière), comb. nov. , and A. gibernau , Gates, Hanson, Jansen-González & Zhang, sp. nov. both the and toruli; malar lateral ventral lateral oral Variation. Both sexes: setation and sculpture variable; sometimes with faint traces of submedian scutellar grooves; vertexal suture can be rounded or angulate. Females: 2.6–3.8mm, scutellum with brown coloration often incomplete laterally, complete medially and anteriorly/posteriorly on scutellar margins; ocellar triangle sometimes brown; pronotal setation ranges from 1–3 per side, adnotaular setation ranges from 1–3 per side with the occasional odd seta in the notaulus; ocellar triangle often with two small divergent setae. Males: 2.5–3.0mm, may have brownish infuscation of the pro- and mesofemur, meso- and metacoxa may be entirely brown. Specimens from Araras Zoo in Brazil consistently had two setae on the lateral lobes of mesoscutum, whereas other specimens had three. However given the lack of other consistent char-acteristics, we conservatively group them under A. gallicola . Variation in female and male genitalia was found. Females reared from T. bipinnatifidum showed two distinct ovipositor morphologies with variation due mostly to larger or smaller first and second valvifers. Females reared from T. solimoesense showed an intermediate size ovipositor. Males reared from T. solimoesense show a longitudinal submedian suture in the digiti that begins at the base of the digital tooth and does not reach the base of the digiti. Biology. Reared from Thaumatophyllum bipinnatifidum and T. solimoesense . Distribution. Brazil and Paraguay. Variation. Considerable variation is noted. Females: 3.5–5.2mm, pronotal setation ranges from 1–3 per side, adnotaular setation ranges from 1–2 per side. Males: 2.8–3.5mm, may have brownish infuscation of the pro- and mesofemur.


Introduction
The Chalcidoidea is a large and diverse superfamily with broad biological diversity (Heraty et al. 2013), and with estimates of over 500,000 species (Noyes 2019) on Earth. Although primarily entomophagous, many phytophagous forms are known, among these the gall inducers that deform plant tissue in order to complete their development. The biology of gall inducers and the evolution of gall induction in Chalcidoidea have been reviewed recently (LaSalle 2005). Gall induction has evolved in seven different families of Chalcidoidea, with at least 16 independent origins both from entomophagous and phytophagous ancestors (LaSalle 2005;Böhmová et al. 2022). This includes the Eulophidae, the largest and most diverse chalcidoid family including over 4,300 species in 332 genera (Noyes 2019;Rasplus et al. 2020). The diversity of gall inducing eulophids is highest in the Australian Opheliminae and the cosmopolitan Tetrastichinae, the latter is a large and diverse subfamily with 15 genera recorded as phytophagous species (Kim et al. 2004(Kim et al. , 2005Mendel et al. 2004;LaSalle 2005;Kim and LaSalle 2008;Rasplus et al. 2011;Fisher et al. 2014). Overall, knowledge of the specific biology of gall associated tetrastichines is minimal but falls into three categories: parasitoid of gall inducer, inquiline, or gall inducer. LaSalle (2005) divides gall-associated tetrastichines into two groups: (1) the Australian inducers that gall Myrtaceae, and (2) mostly Neotropical groups that are often larger and more heavily sclerotized. LaSalle (1994) records seven plant families serving as hosts for gall inducing Tetrastichinae: Araceae, Chenopodiaceae, Euphorbiaceae, Fabaceae, Myrtaceae, Myrsinaceae, and Solanaceae. Additionally, tetrastichines have also been recorded as gall inducers from Casuarinaceae (Fisher et al. 2014), Sapotaceae (Singh et al 2022), and Smilacaceae . In evaluating the gall-inducers associated with the Araceae (Alismatales) in general, focusing specifically on Hymenoptera, we note that very few taxa are known to be associated with this family, particularly as suspected phytophages (Table 1).

Collection and identification
Mature infrutescences of Thaumatophyllum bipinnatifidum and Philodendron radiatum were cut from the plant in the laboratory, or mass reared in bags hung on clothes-line. As the spathe was still closed in most of the infrutescences, careful incisions with a knife were used to expose the fruits beneath. A few fruits were dissected under a stereomicroscope to ensure they had galls with pupae or adults inside. The selected infrutescences were then put in individual organdy bags (40 cm × 30 cm) for wasp emergence. Emerging wasps were collected and stored in 70% EtOH.
Ethanol-preserved specimens were dehydrated through increasing concentrations of ethanol and transferred to hexamethyldisilazane (HMDS) (Heraty and Hawks 1998) before point-mounting. A Nikon SMZ1500 stereomicroscope with 10× oculars (Nikon C-W10X/22) and a Chiu Technical Corporation Lumina 1 FO-150 fiber optic light source were used for point-mounted specimen observation. Mylar film was placed over the ends of the light source to reduce glare from the specimen. Scanning electron microscope (SEM) images were taken with a Hitachi TM3000 (Tungsten source). Body parts of a disarticulated specimen were affixed to 0.1 mm minuten pins with Loctite Ultra Gel super glue. These were then adhered to a 12.7×3.2 mm Leica/Cambridge aluminum SEM stub by a carbon adhesive tab (Electron Microscopy Sciences, #77825-12). Stub-mounted specimens were sputter coated with gold-palladium using a Cressington Scientific 108 Auto from at least three different angles to ensure complete coverage (~20-30nm coating). One set of wings was removed and slide-mounted in polyvinyl alcohol prior to imaging; wings were photographed with a Olympus SC-100 digital camera attached to a Olympus BX43 light microscope and processed using Monitoriella elongata Hedqvist P. radiatum Infante et al. 1995Shimbori et al. 2011 analySIS getIT 5.2 (Olympus Soft Imaging Solutions). The habitus image was captured using an EntoVision Imaging Suite, which includes a firewire JVC KY-75 3CCD digital camera mounted on a Leica M16 zoom lens via a Leica z-step microscope stand. The program Cartograph 5.6.0 (Microvision Instruments, France) was used to merge an image series into a single in-focus, composite image. Lighting was achieved using techniques summarized in Buffington et al. (2005), Kerr et al. (2008), and Buffington and Gates (2008).When possible, male and female genitalia were extracted, cleared with KOH 10% and temporarily mounted in glycerin for imaging. Genitalia were photographed using the same setup as for wings (indicated above). These and SEM images were used for the elaboration of schemes of each genitalia using GIMP 2.8.10. Morphological terminology follows Gibson (1997), while the surface sculpture follows Harris (1979). Several measurements were taken, including: body length, in lateral view from the anterior projection of the face to the tip of the metasoma; head width through an imaginary line connecting the farthest lateral projection of the eyes; head height through an imaginary line from the vertex to the clypeal margin bisecting both the median ocellus and the distance between the toruli; malar space, in lateral view between the ventral margin of the eye and lateral margin of the oral fossa; eye height in anterior view; vertex bristle in anterior view; mesoscutum and scutellum, in dorsal view through imaginary, median transverse and longitudinal lines; marginal vein, the length coincident with the leading fore wing edge to the base of the stigmal vein; stigmal vein, the length between its base on the marginal vein (M) and its apex; postmarginal vein (PMV), the length from the base of the stigmal vein (S) to its apex on the leading fore wing edge. Metasomal sclerites were measured dorsally along the midline.

Molecular analysis
One specimen each of A. gallicola and A. capipunctata were extracted, amplified, and sequenced at the Laboratories of Analytical Biology (LAB) at the Smithsonian Institution's National Museum of Natural History (NMNH, Washington, DC, USA). A modified Ultraconserved Elements (UCE) protocol was used (Faircloth et al. 2012;Branstetter et al. 2017) along with the HymV2P probe set to enrich the UCE loci, (Branstetter et al. 2017 (Faircloth, 2015) was used for UCE processing. SPAdes v3.14.0 (Bankevich et al. 2012) was used to align the contigs, sequences were aligned using MAFFT v7.490 (Katoh and Toh 2008), and trimmed using Gblocks v0.91b (Castresana 2000) with the following settings: b1 = 0.5, b2 = 0.5, b3 = 12, b4 = 7. A 50% complete matrix was used for downstream phylogenomic analysis. Additionally, fragments of legacy markers (COI, 28S, and CytB) were extracted from the UCE contigs using PHYLUCE. Trimmed reads for the newly generated sequences in this study are available from the National Center for Biotechnology Sequence Read Archive (SRA; BioProject ID PRJNA827143), and Sanger markers are available on GenBank (Suppl. material 1).

Results
The 50% UCE matrix consisted of 567 loci, with A. capipunctata and A. gallicola having 1715 and 1802 UCE loci recovered, respectively. The topology recovered was largely identical to that of Rasplus et al. (2020). The new genus Arastichus was recovered within the subfamily Tetrastichinae with strong support (Fig. 2). Arastichus is within the Aprostocetus group sensu Rasplus et al. (2020) , and the sister to Neohyperteles DeSantis with strong support (Fig. 2). Type species. Arastichus gallicola (Ferrière).
Arastichus gallicola was first described by Ferrière (1924) as Trichaporus gallicola, which was then transferred to Exurus Philippi by Costa Lima (1959a). LaSalle (1994) synonymized Exurus with Aprostocetus, through its type species E. colliguayae Philippi. He was hesitant about the status of A. gallicola (Fig. 3) as he was not able to examine any type specimens, but commented that it is quite distinct and warranted its own genus. As Ferrière did not designate a holotype, we hereby designate the top left specimen (female) on the pin with three other specimens as the lectotype (Fig. 3). The degree of morphological variation seen in Aprostocetus makes it difficult to characterize consistently using few characters; however, according to LaSalle (1994), most species have the SMV with ≥ 3 seta, propodeal spiracle partially covered by overhanging lobe of callus, and one cercal setae distinctly longest and sinuate or curved. Although Arastichus shares these diagnostics, several additional apomorphies set it apart from Aprostocetus (as noted in the diagnosis above).
Etymology. Name from the host plant family, Araceae. Gender masculine. Biology. Ferrière (1924) first described Arastichus gallicola (as Trichoporus gallicola) and defined the species as gall inducer on pistilate flowers of Philodendron selloum (now a synonym of Thaumatophyllum (Philodendron) bipinnatifidum (Mayo 1991) (Fig. 3). Gibernau et al. (2002) described the galls of A. gallicola on flowers of T. solimoesense and reported it as a seed predator. Recently, a more detailed study of the developmental biology of A. gallicola discards seed predation and supports the idea that this species is a gall inducer specialized on ovaries of T. bipinnatifidum (SJG, unpublished).
Female wasps of A. gallicola oviposit during the period of anthesis which lasts 24-48 hours, when the inflorescence spathe is open and leaves the hundreds of pistilate flowers accessible to pollinators and female Arastichus (Gibernau et al. 2002). Once anthesis ends the spathe closes and the space between the spathe and the inflorescence fills with a liquid, often trapping and killing the female wasps inside.
Time of development can vary from one to four months in Arastichus gallicola. Once the infrutescence attains maturity, the spathe develops an encircling dehiscent line at its base and falls, uncovering the orange fruits and galls. Exposure of galls to light and outer atmosphere might trigger adult wasp emergence from the galls, which is done by chewing through each gall wall. A single wasp develops per gall with up to six galls developing in a single fruit. It is possible to find infrutescences and/or fruits containing only seeds, combinations of seeds and galls, or only galls (Fig. 1).
Although we have detailed information about gall induction only in A. gallicola, it is possible that the other two species of Arastichus are also gall inducers rather than seed predators. Examination of collected material for A. gibernau and A. capipunctata indicates that the biology of these species should not be very different from that of A. gallicola.
The eurytomid Prodecatoma philodendri is associated with the galls of Arastichus gallicola and A. gibernau. Ferrière (1924) reported that Prodecatoma were phytophagous, and oviposits from the outside when the spathe is closed and Arastichus galls are in the process of formation (Gibernau et al. 2002;SJG pers. obs.). When examining the cavities from which Prodecatoma adults emerge, a series of tunnels communicate with adjacent Arastichus galls. These attacked galls contained dismembered body parts of Arastichus pupae, indicating that Prodecatoma larvae might consume several of them along with some gall tissue (Gibernau et al. 2002;SJG pers. obs.); this is in line with the fact that the adult Prodecatoma is about 4-5 times larger than the Arastichus adult. Thus, taken all together, P. philodendri is likely entomophytophagous, a common mode of feeding within Eurytomidae.
It is difficult to estimate the taxonomic breadth of the relationship between Arastichus and Araceae. Philodendron is traditionally subdivided in three subgenera: Meconostigma, Philodendron and Pteromischum, but members of Meconostigma have been recently recognized as a distinct genus Thaumatophyllum Schott (Sakuragui et al. 2018). Arastichus has been found in species belonging to Thaumatophyllum (T. bipinnatifidum, T. solimoesense), and in the subgenus Philodendron (P. radiatum). SJG has collected what seem to be female Arastichus body parts from inside closed spathes of P. cordatum and P. curvilobum in Brazil. Further studies and more extensive collecting are needed to determine the degree of species-specificity and to determine whether Arastichus is present in the subgenus Pteromischum as well. Mesoscutum bilobed at posterior margin (Fig. 23). Face with numerous large punctures (Figs 21, 22). Female body (excluding legs and lower face) completely brown (Fig. 6) Posterior corner of metapleuron with circular fossa that is at least half as wide as propodeal spiracle (Fig. 25, arrow). Vertexal suture rounded where it curves down along the inner eye margin (Fig. 24)  Posterior corner of metapleuron without a noticeable fossa, or with an elongate depression (Fig. 4). Vertexal suture angulate or rounded where it reaches the inner eye margin (Fig. 10)  Description. Holotype female. Body length 2.9 mm. Color: Brown except for the following yellow: scape, pedicel, lower face, prepectus, legs (except metacoxa brown), wing veins white to brown (Fig. 6).
Etymology. Named for the distinctive punctate head. Biology. Reared from Philodendron radiatum.
Male. Overall morphology as in female (Fig. 5). Body length 2.5 mm. Color: Dark brown except the following golden: base of scape, ventral mouthparts, acropleuron, coxae apically, legs, metatibia in apical 1/4. Antennal ratio of scape (minus radicle):pedicel: A1:F1:F2:F3:F4:F5:F6:club as 6.9:1.4:1:4.3:4.7:4.7:4.3:4.1:3.6:2.9; scape with distinct, white ventral plaque in apical ½ (Fig. 5), funicular segments wide at base and narrowing off towards apex, with whorl of setae extending ~1.5x length of the funicular segment to which it is attached, MPS sparse and located at midlength; clava with basal whorl and apical setae, MPS located at apex (Fig. 5). Genitalia: phallobase twice as long as broad, digitus slender without projection on anterior margin, aedeagus slender, with apex pointed; digiti with or without a submedian longitudinal suture from the base of the digital tooth but not reaching the base of the digiti (Fig. 27). Variation. Both sexes: setation and sculpture variable; sometimes with faint traces of submedian scutellar grooves; vertexal suture can be rounded or angulate. Females: 2.6-3.8mm, scutellum with brown coloration often incomplete laterally, complete medially and anteriorly/posteriorly on scutellar margins; ocellar triangle sometimes brown; pronotal setation ranges from 1-3 per side, adnotaular setation ranges from 1-3 per side with the occasional odd seta in the notaulus; ocellar triangle often with two small divergent setae. Males: 2.5-3.0mm, may have brownish infuscation of the pro-and mesofemur, meso-and metacoxa may be entirely brown. Specimens from Araras Zoo in Brazil consistently had two setae on the lateral lobes of mesoscutum, whereas other specimens had three. However given the lack of other consistent characteristics, we conservatively group them under A. gallicola. Variation in female and male genitalia was found. Females reared from T. bipinnatifidum showed two distinct ovipositor morphologies with variation due mostly to larger or smaller first and second valvifers. Diagnosis. Arastichus gibernau is morphologically similar to A. gallicola, but the posterior corner of metapleuron of A. gibernau has a noticeable fossa, or with an elongate depression ((Figs 8,9). Additionally, the vertexal suture is always rounded where it reaches the inner eye margin in A. gibernau (Fig. 24), whereas in A. gallicola this suture is angulate or rounded (Fig. 10).
Etymology. Named in honor of Dr. Marc Gibernau for providing a very large sample of specimens of this species for our research.

Discussion
Gall induction have evolved multiple times within Tetrastichinae, and to date is known from 10 different host plant families (LaSalle 1994;Fisher et al. 2014;Gates et al. 2020;Singh et al 2022) around the world. However, given the diversity and the lack of taxonomic attention in recent years, the true number of tetrastichine gall inducers is likely much higher. Hopefully with the advances of phylogenomic techniques such as UCEs and broader taxonomic sampling of species-rich regions such as the Neotropics, we can gain a better understanding of the true diversity of gall induction within Eulophidae and Chalcidoidea as a whole.