New host record for the enigmatic Neotropical mantidfly genus Anchieta Navás, 1909 (Neuroptera, Mantispidae), a mimic of wasps and stingless bees

Species of Symphrasinae (Neuroptera: Mantispidae) are ectoparasitoids of larvae and pupae of holometabolous insects, primarily of Hymenoptera in their larval stages. Herein we present the third case of an association between the mantidfly genus Anchieta Navás, 1909 with the order Hymenoptera. The hymenopteran species attacked by the as of yet undescribed species of Anchieta is Montezumia dimidiata Saussure, 1852 (Vespidae: Eumeninae), a predacious wasp that constructs mud nests. The association was observed in Peruvian Amazonia (near Tarapoto, San Martín), after rearing the mantidflies from a wasp nest. The biology and mimicry pattern with stingless bees of the reared Anchieta species is discussed.

Both direct observation (Dejan & Canard, 1990) and circumstantial evidence (Linsley et al., 1980;Hook et al., 2010) suggest that symphrasine females deposits their eggs near the host nest entrance, or near individual cells within the host nest. The newly hatched first instar mantispids then migrates into the cells while it is being provisioned, later the larva becomes attached to the surface of the host and remains attached until the host dies. The mantispid larvae feed on the host until completion of development when a mobile pupa chews its way out of the cocoon and emerges as a mobile pharate adult.
Anchieta is the smallest genus of Symphrasinae, including eight little-known species found from Panama to Southern Brazil, with most of the species distributed across the Amazon rainforest (Ardila-Camacho et al., 2018;Oswald, 2020). The genus is remarkable among the symphrasine genera, as all of the species mimic different groups of Hymenoptera (i.e., Apidae (Meliponini), Vespidae, and Braconidae), and is distinguished from other genera of the subfamily by having a prominent, blunt process on the fore trochanter, and a straight anterior radial cell of forewing (Ardila-Camacho et al., 2018). The only previously reported host of Anchieta was the mud dauber Trypoxylon (Trypargilum) aestivale Richards, 1934 (Hymenoptera: Crabronidae) (Buys, 2008). From the mud nest of this wasp species, a single specimen of A. fumosella (Westwood, 1867) was reared out and presumably it had fed on the last instar larva or on the pupa of the wasp (Buys, 2008). Recently, further unspecified associations with wasps or bees were reported by Araújo et al. (2021).
Based on specimens of a new Anchieta species -referred here as Anchieta sp. nov. -which attacked mud nests of M. dimidiata, the purpose of the present paper is to provide the only second known and identified host record for the genus, as well as scattered observations on the biology and the mimicking of Anchieta.

MATERIAL AND METHODS
Irregular surveys for nests of Hymenoptera have been performed in San Martin since 2002 by the first author, including near "Urku Estudios Amazonicos" educational center close to the Boca Toma of Rio Shilcayo, Tarapoto, Peru (06.4595°S, 76.3512°W, 410 m a.s.l.). The locality is adjacent to the local water reservoir of Tarapoto, within 100 m of the Shilcayo river, and represents an interesting transition area between lowland rainforest and lower montane rain forest, or cloud forest, in Peru. The high precipitation associated with the abrupt elevational gradient in the region of the "Cordillera Escalera" result in great environmental heterogeneity and a unique biological diversity (e.g., Rasmussen & Skov, 2006;Rasmussen, 2009;Rasmussen & Gonzalez, 2009). On this location a single adult female of the potter wasp Montezumia dimidiata Saussure, 1852 (Vespidae: Eumeninae) was observed resting on a mud nest plastered beneath a roof tile (Fig. 1A). The nest was collected the same day on July 10 th , 2012 and left in a jar awaiting emergence (Figs. 1B, 1C). All observations hereafter were made as insects emerged from the jar left at room temperature.
Species identification of the emerging Anchieta was made by dissecting and clearing the abdomen, following the standard procedures with 10% Potassium Hydroxide solution (KOH). The external morphology and genital sclerites were compared with the types of all known species of Anchieta. All these structures were examined using a Zeiss Dicovery V8 stereomicroscope. Specimens were deposited at Museo de Historia Natural de Lima, Peru (MUSM) and Museum für Naturkunde, der Humboldt-Universität, Berlin, Germany (ZMB).

Biology
The M. dimidiata nest was smaller but otherwise similar to two nests of the same species described from Colombia by Evans (1973), including the observation that individual cells were not obvious due to irregular plastering of mud covering the surface of the nest. The collected nest contained a total of five cells, three basal cells in parallel and adhered to the surface of the tile, followed by two additional cells in a second outer row. Of the three basal cells, the first was empty, possible damaged during the collection of the nest, and the second cell was the only open and had a small 4 mm long turret at the entrance (Fig. 1B). No provision, egg, or larvae were observed in that cell, and it is possible that this cell was either ready for provisioning by the founding female wasp observed resting on the nest or, that the female outside was instead a recently emerged wasp from this very cell. However, the presence of a turret suggests that the female outside was the founding female, although it is unclear why the last cell to be provisioned would be the most basal cell in her construction. The last of the basal row cells is the one where all five mantispid cocoons were encountered together (Figs. 1C, 1E). The second row of cells both contained wasps, with one female M. dimidiata wasp later emerged while the wasp in the last cell was found dead. Cells measured internally about 9 by 24 mm. No provision for the wasps were encountered in this nest, but Evans (1973) reported microlepidoptera species as prey for M. dimidiata.
No egg chorions were found on the smooth inner walls of the cell and it is not clear when or where the eggs of the five mantispids were deposited, but the nest was collected on July 10 th , 2012, and contained by then five cocoons with content. The adult mantispids emerged July 19 th (female, Fig. 1D), July 22 nd (male), July 28 th (male, Fig. 2A), September 6 th (male pupae, died), September 8 th (male, Figs. 2C-2G). This is 9, 12, 18, 58, and 60 days after the encounter and collection of the nest. The emerging mantispids are identified as Anchieta, but does not correspond to any of the eight known species of the genus and will be described as a new species by Ardila-Camacho in a forthcoming taxonomic revision of the genus, so here we treat it as Anchieta sp. nov.
The mantispid cocoons (Fig. 2B) were only lightly attached to the cell wall and to each other, made of loosely woven silken threads, which upon maturing turned darker yellow until reaching the coloration of the pupae, including two dark spots presumably marking the eyes. The size of the cocoons were 6 by 3 mm (the four males) and a single larger, 8 by 4 mm (the single female), made from a darker (or older) thread.
Two of the emerging Anchieta sp. nov. were observed first actively moving around as exarate pupae or pharate adults (Fig. 2C-2E), but within hours of leaving the cocoon, shed the exuviae and then rested for hours (Fig. 2E) until the cuticle and wings had completely hardened and darkened (compare the coloration of the hind leg from the same individual in Figs. 2E and 2G). One of the exarate pupae did not survive and died be- fore molting, although it can be observed that a complete adult is trapped beneath the exuviae of the pupae. While active and moving exarate pupae were not observed for the specimens, the exuviae were always found away from the nest cell and the cocoon where they had emerged, suggesting that they would leave the pupae with the exuviae still attached. The actual shedding of the exuvia took less than ten minutes on the single instance when it was observed, that is from the individual began moving and until the exuvia was removed.

Mimicking
An interesting observation on this new species was the immediate confusion of its identity by the first author upon emergence. Having collected the nest of a known species of potter wasp and expecting the emergence of this wasp, the surprise was that the first emerging insect had a close resemblance to the social stingless bee, Ptilotrigona lurida (Smith, 1854), both with respect to size, coloration of wings and body, and down to the imitation of the expanded hind legs, resembling the pollen basket The last, and fifth, Anchieta sp. nov. to emerge, a male. This demonstrates that the pharate specimen is active and moving as a pupa (C). The specimen is removing the exuvia from the prepupa. Notice that the wings are still soft and bent (C and D). The same specimen is now waiting for the wings to strengthen few minutes after having shred the exuvia (F). The same specimen with hardened wing. Notice that the wing did not display perfectly (G).
of corbiculate bees (Engel & Rasmussen 2021). Dissected individuals are compared side by side in Figs. 3A-3B. The ovoid, light orange forefemur with black markings on the outer surface, plus the short and narrow hind wing resemble the compound eye and the overall shape of the hindwing of the bee, respectively. The first author has caught these bees commonly in the area, but only in forested parts, such as the type locality for Anchieta sp. nov. It has previously been reported that some species of Anchieta mimic stingless bees (Penny, 1982;Penny & Costa, 1984;Hogue, 1993), social wasps (e.g., A. fumosella in Buys, 2008) or braconids equipped with repugnant glands (e.g., A. fasciatella (Westwood, 1867) in Ardila-Camacho & García, 2015). Aggressive social paper wasps are common models for mantispids throughout the world (Batra, 1972;Boyden, 1983;Beck, 2005;Snyman et al., 2020), and individual species have been shown to form distinct color morph mimics that traces the local wasp fauna through polymorphism (Batra, 1972;Opler, 1981). While a Vespidae wasp model is observed in a single species (i.e., yellow and black morph of A. fumosella), most species of Anchieta appear to closely resemble either Ptilotrigona Moure, 1951 (e.g., A. apiculasaeva Thouvenot, 2009, A. bella (Westwood, 1867), A. eurydella (Westwood, 1867), A. partheniella (Westwood, 1867), A. remipes (Gerstaecker, 1888), and Anchieta sp. nov.) or Trigona Jurine, 1807 (e.g., dark morph of A. fumosella, and A. notha (Erichson, 1830)) stingless bees. For instance, A. notha co-occur with Trigona spinipes (Fabricius, 1793), while a Colombian specimen of Anchieta eurydella from Amacayacu (Amazonas, Colombia) was sweeped flying with a group of stingless bees, including Trigona dallatorreana Friese, 1900, Plebeia sp. and Nogueirapis butteli (Friese, 1900) with T. dallatorreana possibly the model in a protective Batesian mimicry system. These species of Ptilotrigona and Trigona form large nests and are amongst the most aggressive of all stingless bees, in part due to relatively large-sized and strong mandibles with pointed teeth (Kerr, 1951;Wille & Michener, 1973;Rasmussen & Camargo, 2008). The bees will attack in swarm near the nest but are unaggressive away from the nest. This is rather intriguing, as in order to provide an efficient model phenotype in a mimicry system, they also need to be noxious to predators away from the nest. However, stingless bees often carry sticky resins on the legs, have distasteful and foul-smelling mandibular gland substances (Kerr, 1951;Smith & Roubik, 1983;Roubik, 1989), and plant-originated terpenes on the bees' cuticles (Lehmberg et al., 2008;Leonhardt et al., 2009) which makes them noxious and less prone to predation, as well as an efficient model for mimicry systems. In the case of Anchieta, the biological mimicry system is obviously Batesian, with the stingless bees being the model phenotype and Anchieta the mimic.
While stingless bees are one possible model, other bees such as e.g., Paratetrapedia Moure, 1941 and Tetrapedia Klug, 1810 (Hymenoptera: Apidae) repeat the same coloration pattern (Kerr, 1951;Aguiar & Melo, 2011), and might form a mimicry complex, and a Müllerian mimicry model to the stingless bees. Paratetrapedia females have a sting and are able to defend themselves and their hind legs are expanded due to dense pubescence. However, suggestive of the stingless bees being the model phenotype for all mimics is the fact that stingless bees are the most abundant of the species. To provide a classical workable mimicry system imitators are always less numerous in individuals (Wallace, Figure 3. Anchieta sp. nov. and Ptilotrigona lurida (Smith, 1854) next to each other. The former with the abdomen inflated following the dried up condition (A and B).

ACKNOWLEDGMENTS
Daniel Vecco and the rest of the staff at Urku are thanked for encouraging research on the area. In addition, Michael Ohl inspired the collaboration here and identified the specimen of Anchieta eurydella, Josef Gusenleitner and Bolívar Rafael Garcete-Barrett identified the Montezumia dimidiata wasp, Antonio Aguiar, Hélcio Gil, Menno Reemer, Keith Bayless and Doug Yanega helped with mimicry references, and Victor Gonzalez helped with literature for which we are grateful. AAC acknowledges Posgrado en Ciencias Biológicas, UNAM, for its support during his studies, and Consejo Nacional de Ciencia y Tecnología (CONACyT) for the financial support through a doctoral scholarship. The second author also wants to express sincere thanks to Atilano Contreras-Ramos of Intituto de Biología, UNAM for his support during all stages of his doctorate. Specimens were collected under permit Nr. 0356-2012-AG-DGFFS-DGEFFS.