Four new neotropical Trioza species associated with Loranthaceae (Santalales) and comments on mistletoe inhabiting psyllids (Hemiptera, Psylloidea)

Four new Trioza species associated with mistletoes are described, diagnosed and illustrated from Brazil and Chile. They are monophagous on the Loranthaceae Struthanthus uraguensis (Trioza struthanthi sp. n.), Tripodanthus acutifolius (Trioza tripodanthi sp. n.) and narrowly oligophagous on Tristerix spp. (Trioza tristericis sp. n.). For a fourth species (Trioza vagata sp. n.) host information is unavailable but its association with mistletoes is likely. The four species form a putative monophyletic group, together with three North American species (Trioza acuminata Tuthill, T. incidata Tuthill and T. phorodendrae Tuthill), based on the apically incised paramere, the highly modified valvulae of the female ovipositor and the host association with mistletoes. T. acuminata stat. n. from California is raised to species rank (originally described as subspecies of T. phorodendrae). The Trioza struthanthi-group is diagnosed and a key for the identification of its constituent members is provided. New host data are provided for Calophya sp. from Brazil as well as for Notophorina fusca Burckhardt and Zonopelma australis Burckhardt from Chile. Host plant and biogeographical patterns of mistletoe feeding psyllids around the world are briefly discussed.


Introduction
Psyllids or jumping plant lice are plant sap sucking insects, which are generally highly host specific, i.e. they complete their development on only one plant species or a few species of the same plant genus.In addition, related psyllid species tend to develop on related plant species, making them an interesting group for evolutionary studies (Burckhardt et al. 2014).Several studies showed that these host plant patterns are not primarily the result of cospeciation, as there is no large scale congruence of the phylogenies of the two groups, but rather host shifts within the particular host group (Ouvrard et al. 2015).Mistletoes are a particularly fascinating group to study host patterns as they themselves, by their hemiparasitic life style, depend on a host (Lázaro-Gonzales et al. 2017).
Mistletoes are members of the order Santalales which contains 13 families, 151 genera and 1992 species of hemiparasitic and non-parasitic plants (Stevens 2001).Psyllid associations have been reported from the three families Loranthaceae, Misodendraceae and Santalaceae (including Viscaceae) (Table 1).Taylor (2016), in a study of Australian Acizzia species from mistletoes of the genus Amyema, suggested that the switch of psyllids from ancestral hosts (probably Fabaceae) to mistletoes may have been mediated by their close proximity (i.e.within the same canopy).
Phoradendron sp.[as Phoradendron "pubescens" sic!] USA (California) Tuthill (1943) last three decades shows that the apparent paucity in South America is an artefact of poor knowledge on the psyllid fauna of this continent (Burckhardt and Queiroz 2012).
Here we describe four new Trioza species from Brazil and Chile associated with Loranthaceae (confirmed for three, likely for one species), report host plants (Misodendraceae and Santalaceae) for another three species from Brazil (Calophya sp.) and Chile (Notophorina fusca Burckhardt and Zonopelma australis Burckhardt) and discuss the host patterns of Psylloidea developing on Santalales.

Material and methods
Material was examined or is cited from following institutions: BMNH-Natural History Museum, London, UK; MHNG-Muséum d'histoire naturelle, Genève, Switzerland; MMBC-Moravian Museum, Brno, Czech Republic; MZSP-Museu de Zoologia, Universidade de São Paulo, SP, Brazil; NHMB-Naturhistorisches Museum, Basel, Switzerland; UFPR-Coleção Entomológica Padre Jesus Santiago Moure, Centro Politécnico, Universidade Federal do Paraná, Curitiba, PR, Brazil; USNM-United States National Museum collections, Beltsville, MD, USA.Females are selected as holotypes of the four new species which is somewhat unusual in psyllid taxonomy.The four species have highly modified valvulae in the female ovipositor, a feature which is unique within the large family Triozidae, and each species is well diagnosed by the shape of the female terminalia and the valvulae.The male terminalia are also diagnostic but for T. tripodanthi sp.n.only two specimens are available, one specimens is slide mounted, the other one has abnormally developed, asymmetrical genal processes.Neither specimen is ideal as holotype.
The morphological terminology is mostly that of Hollis (1984,2004) but see also Figs 11,13,32.Measurements were taken as follows: adult body length from dry mounted or ethanol preserved specimens measuring the distance between fore margin of head and tip of forewings when folded over body; body length of immatures from specimens preserved in ethanol; the other measurements were taken from slide mounted specimens.The adult body length is given as range, mean and standard deviation, the other measurements and the ratios as range.
The plant nomenclature accords with The Plant List (2013).
Trioza acuminata Tuthill, 1943, was described as subspecies of T. phorodendrae Tuthill, 1939 based on a single male from California.According to Tuthill (1943) the genal processes of T. acuminata are distinctly longer than in T. phorodendrae suggesting that the two taxa are distinct.Until additional evidence becomes available, the two taxa are treated as different species and we elevate the former to species rank as Trioza acuminata Tuthill, 1943, stat. n.
Description.Adult (Fig. 1).Colouration.Brown to black, intersegmental membranes red.Tips of genal processes dirty yellowish; eyes grey, ocelli yellowish to grey.Antennal segments 1 brown and 2 yellow, flagellum brown, gradually becoming darker towards apex.Tibiae entirely and basitarsi partially yellow.Veins of forewing light brown, membrane yellowish with indistinct brown patch along clavus (cell cu 2 ), transparent.Hindwing transparent, colourless.Male terminalia brown.Younger specimens almost entirely dirty yellow, getting gradually darker with age.
Structure.Body length ♂ 2.3-2.9 mm (2.60±0.17mm), ♀ 2.7-3.1 mm (2.85±0.11mm) (10 ♂, 10 ♀).Genal processes 0.8 times as long as vertex along midline, irregularly tapering to subacute apex (Fig. 2).Antenna 1.4-1.6 times as long as head width; segments 4 and 6 distinctly inflated apically.Forewing (Fig. 9) narrowly lanceolate, subacute apically, 4.2-4.7 times as long as head width, 2.6-2.8 times as long as wide; surface spinules present in all cells, leaving spinule-free stripes along the veins and reduced at the base of cells, only few or no spinules in cell c+sc, forming indistinct transverse rows (Fig. 10).Metatibia 0.8-0.9times as long as head width, genual tooth prominent.-Terminalia as in Figs 17-19, 29-32.Male: setae on proctiger covering a wide area in apical two thirds arranged in several indistinct transverse rows.Paramere about as long as proctiger, in profile, about twice as long as broad; outer and inner lobe of about the same length; outer lobe with sclerotised tooth subapically, inner lobe with sclerotised, forward directed point; inner surface with long setae, in basal half with a group of thick bristles.Distal segment of aedeagus weakly expanded apically.Female: dorsal outline of proctiger in basal two thirds almost straight, then abruptly curved down, forming angular bump; apical process bearing five evenly spaced, dorsal teeth.Subgenital plate, in ventral view, truncate apically, with a small group of setae apically which is well separated from other setae.Oblique apex of ventral valvula smooth, lacking teeth.-Measurements see Table 2.
Fifth instar immature (Fig. 42).Colouration.Head and thorax yellow, wing pads and abdomen light ochreous, the latter sometimes greenish.Eyes reddish-grey.Antennae brown.Ventral body surface yellow, tips of tarsi brown.
Etymology.The species is named after its host plant genus Struthanthus; struthanthi is a noun in the genitive case.
Distribution.Brazil (Minas Gerais, Paraná, Rio Grande do Sul, Santa Catarina).Diagnosis.Member of the Trioza struthanthi-group.Body of adult orange with conspicuously black genal processes.Genal processes 0.7 times as long as vertex along mid-line, strongly tapering near base, then tubular with blunt apex.Forewing subacute apically, 2.8 times as long as wide.

Number of measured specimens
Paramere about three times as long as broad, straight.Dorsal margin of female proctiger distal of circumanal ring evenly curved down to process; valvula ventralis with three apico-ventral teeth.Fifth instar immature with following numbers of marginal sectasetae (one side only): head 29-31, forewing bud 86-88, hindwing bud 13-15, precaudal abdominal margin 1-2, caudal plate 79; distal portion of sectasetae on forewing bud 1.5 times as long as wide.
Etymology.The species is named after its host plant genus Tripodanthus; tripodanthi is a noun in the genitive case.

Host plant. Tripodanthus acutifolius (Ruiz & Pav.) Tiegh. (Loranthaceae).
Discussion.Trioza tripodanthi differs from the other six members of the T. struthanthi-group in light body colour with the very conspicuous dark genal processes and details of the male and female terminalia.The fifth instar immatures of T. tripodanthi differ from those of T. struthanthi in the smaller number of marginal sectasetae and from those of T. tristericis in the shorter and broader marginal sectasetae.The immatures of T. vagata and of the North American species of the T. struthanthi-group are unknown.

Diagnosis. Member of the Trioza struthanthi-group.
Body of adult orange to brown with orange to ochreous genal processes.Genal processes 0.8 times as long as vertex along mid-line, relatively evenly tapering to subacute apex.Forewing narrowly rounded, 2.7 times as long as wide.Paramere about three times as long as broad, weakly curved.Dorsal margin of female proctiger between apex of circumanal ring and base of apical process almost straight with very small bump near the middle; valvula ventralis with several small apico-ventral teeth.Fifth instar immature with distal portion of sectasetae on forewing bud 2.6 times as long as wide.
Description.Adult (Fig. 5).Colouration.Orange to brown, intersegmental membranes orange.Genal processes orange to ochreous, tips often dirty whitish; eyes and ocelli grey; head ventrally yellow.Antennal segments 1 and 2 yellowish orange, segments 4-7 pale yellowish with apices of segments 4 and 6 brown, segments 8-10 brown to black.Meso and metathorax with indistinctly delimited longitudinal dark stripe.Thoracic pleura and venter yellow.Legs yellow, femora greyish brown.Veins of forewing light brown, membrane weakly yellowish.Abdomen almost black; male terminalia light brown, female terminalia orange to brown, apex almost black.Younger specimens dirty yellow or orange, getting gradually darker with age.
Structure.Body length ♂ 2.7-2.9 mm (2.76±0.07mm), ♀ 2.6-3.2mm (2.97±0.18mm) (8 ♂, 10 ♀).Genal processes 0.8 times as long as vertex along mid-line, relatively evenly tapering to subacute apex (Fig. 6).Antenna 1.4-1.6 times as long as head width; segments 4 and 6 not inflated apically.Forewing (Fig. 13) narrowly lanceolate, narrowly rounded apically, 4.5-4.6 times as long as head width, 2.7 times as long as wide; surface spinules strongly reduced, present at apex of cell c+sc, base of r 1 , apex of r 2 , a few scattered spinules in m 1 , m 2 and cu 1 , covering most of cu 2 , arranged in squares or rhombs (Fig. 14).Metatibia 0.8-0.9times as long as head width, genual tooth small.-Terminalia as in Figs 23-25, 36-38.Male: setae on male proctiger covering a wide area in apical two thirds irregularly arranged.Paramere slightly longer than proctiger, in profile, about three times as long as broad, weakly curved; outer lobe distinctly shorter than inner lobe; outer lobe digitiform, without sclerotised subapical tooth; inner lobe with sclerotised, forward directed point; inner surface with long setae, those in basal half not conspicuously thicker than those in apical half.Distal segment of aedeagus with long, gradually expanded apical dilatation.Female: dorsal outline of proctiger between apex of circumanal ring and base of apical process almost straight with very small bump near the middle; apical process bearing several small dorsal teeth near apex.Subgenital plate, in ventral view, shallowly incised apically, with a longitudinal row of setae apically which is well separated from other setae.Oblique apex of ventral valvula with several small teeth.-Measurements see Table 2.
Fifth instar immature.Colouration.Head, thorax and abdomen orange or light brown, wing pads slightly lighter.Eyes reddish-grey.Antennae reddish in basal half, dark brown apically.Tips of tarsi brown.
Structure.Only one damaged specimen available.Distal portion of sectasetae on forewing bud 2.6 times as wide (Fig. 46).

Etymology.
The species is named after its host plant genus Tristerix; tristericis is a noun in the genitive case.
Discussion. Adult Trioza tristericis differ from the other six species of the T. struthanthi-group, apart from details of the male and female terminalia, as follows: from the North American species in the narrower forewing and from the South American species in the orange to brown body colour with concolorous genal processes.The fifth instar immatures of T. tristericis differ from those of T. struthanthi and T. tripodanthi in the longer and narrower marginal sectasetae.The immatures of T. vagata and the North American species of the T. struthanthi-group are unknown.
Diagnosis.Member of the Trioza struthanthi-group.Body of adult yellowish or orange to light brown.Genal processes 0.9 times as long as vertex along mid-line, irregularly tapering to subacute apex.Forewing subacute apically, 2.8-2.9 times as long as wide.Paramere about three times as long as broad, straight.Dorsal margin of female proctiger evenly curved; valvula ventralis lacking apico-ventral teeth.
Description.Adult (Fig. 7).Colouration.Light yellow to orange or light brown, intersegmental membranes whitish.Genal processes ochreous to brown, tips often dirty whitish; eyes dark reddish grey.Antennal segments 1-7 pale yellow with apex of segment 6 brown, segments 8-10 brown to black.Thoracic pleura and venter slightly paler than dorsum.Metacoxa brown.Veins of forewing yellow to ochreous, membrane colourless or weakly yellowish.Tip of paramere and process of female proctiger dark brown to almost black.Younger specimens paler, getting gradually darker with age.
Structure.Body length ♂ 2.6-2.7 mm (2.65±0.06mm), ♀ 2.8-3.1 mm (2.97±0.15mm) (4 ♂, 3 ♀).Genal processes 0.9 times as long as vertex along mid-line, irregularly tapering to subacute apex (Fig. 8).Antenna 1.4-1.5 times as long as head width; segments 4 and 6 distinctly inflated apically.Forewing (Fig. 15) narrowly lanceolate, subacute apically, 4.7-4.9times as long as head width, 2.8-2.9 times as long as wide; surface spinules present in most cells, leaving broad spinule-free stripes along the veins, forming indistinct transverse rows, in cell c+sc restricted to apex and in cell r 1 to base, almost completely absent from cells m 1 and m 2 (Fig. 16).Metatibia 0.8 times as long as head width, genual tooth prominent.-Terminalia as in Figs 26-28, 39-41.Male: setae on male proctiger covering a narrow stripe along posterior margin arranged in two irregular longitudinal rows as well as on apex.Paramere distinctly shorter than proctiger, in profile, about three times as long as broad; outer lobe shorter than inner one; outer lobe with sclerotised tooth subapically, inner lobe with sclerotised, forward directed point; inner surface with few long setae, in middle third with a group of thick bristles.Distal segment of aedeagus strongly expanded apically.Female: dorsal outline of proctiger in basal two thirds almost straight to weakly, then evenly curved down; apical process bearing many uneven dorsal and lateral teeth.Subgenital plate, in ventral view, truncate apically bordered on either side by pointed lobes, with moderately long setae in apical two thirds except for a longitudinal stripe in the middle which is almost bare and at apex where the setae are very long.Oblique apex of ventral valvula smooth, lacking teeth.-Measurements see Table 2.
Fifth instar immature unknown.
Etymology.From Latin vagare = to wander, to roam, for its discovery away from its supposed host, a mistletoe; vagata is the feminine form of the participle perfect passive.
Host plant.Adults have been collected on Clusia sp.(Clusiaceae) which is an unlikely host.
Discussion.Trioza vagata differs from the other six members of the T. struthanthi-group in the very narrow forewing (2.8-2.9 times as long as wide) which lacks surface spinules in most of cell r 2 and in the digitiform paramere.The female terminalia are similar to those of T. tripodanthi but the dorsal margin of the proctiger is more curved and the oblique apex of the valvula ventralis lacks teeth.

New host records
For the following three species new host records are provided here.
Comment.Zonopelma australis was described from Southern Argentina and Far Southern Chile (XII Region) but without host data (Burckhardt 1987a).Here we provide host data and add new localities from Chile (Regions X-XII).Zonopelma contains a second species (Z.myzodendri Burckhardt) which occurs in southern Chile on Misodendrum punctulatum.A third species from Paraguay, viz.Zonopelma borealis Burckhardt, develops on mimosoid Fabaceae.Extensive material collected on Mimosa spp. in Brazil is closely related to Z. borealis but not congeneric with Z. australis, the type species of Zonopelma (Burckhardt and Queiroz, unpublished data).

Discussion
Trioza, in its present definition, is a large, artificial genus of worldwide distribution (Hollis 1984, Burckhardt and Ouvrard 2012, Ouvrard 2017).Several, probably monophyletic species-groups have been defined whose relationships to each other, however, remain (Hollis 1984, Burckhardt 1988, Brown and Hodkinson 1988).The incised paramere and the highly modified valvulae of the female ovipositor, along with the association with Santalales, strongly support the monophyly of the Trioza struthanthi-group.Tuthill (1939) noted a resemblance of T. phorodendrae, a member of the T. struthanthi-group, to T. mexicana Crawford, though without giving details.Both species bear an apical process on the female proctiger, but T. mexicana lacks the apically incised paramere and the highly modified valvulae of the T. struthanthi-group.The host of T. mexicana is unknown.A close phylogenetic relationship of the two species is, therefore, not supported.At present, no detailed synapomorphies are known linking the T. struthanthi-group to other members of Trioza and its phylogenetic relationships remain obscure.
Psyllids are generally highly host specific, i.e. they can complete their development only on a single (monophagous) or on several plant species of the same genus (narrowly oligophagous), family or order (widely oligophagous).Polyphagy is very rare among psyllids.Host data in the literature are, unfortunately, blurred by reports of plants on which adult psyllids have been observed or collected but on which they would be unable to complete their development (Burckhardt et al. 2014).Of the 34 psyllid species listed in Table 1, which summarises all psyllid species from Santalaceae reported in the literature or represented in the collections of the MHNG and NHMB, hosts are confirmed with immatures (or skins) for 18 and are likely for another eight species.For three species there are no host records but an association with Santalales is likely, and for five species the host records are improbable or uncertain.Mathur (1975) described four species, collected during a survey of sandalwood in northern India.All four species were described from adults only.Macrohomotoma maculata Mathur, 1975 andMycopsylla indica Mathur, 1975 are members of the Homotomidae, a family which is restricted to hosts within the Moraceae (Hollis and Broomfield 1989).Santalum album is, therefore, a very unlikely host.In the case of Diaphorina venata Mathur, 1975 andPsylla santali Mathur, 1975 there is no evidence for or against sandalwood being the host and the information is, therefore, uncertain.Equally uncertain is the record of Diaphorina verbera Kandasamy, 1986 described from two adults from southern India also on sandalwood.In terms of host specificity, 15 of the 26 species with confirmed or likely hosts are monophagous, 9 narrowly oligophagous and 2 widely oligophagous.
Taylor (2016) suggested that the switch of the Australian Acizzia species from their ancestral hosts, probably Fabaceae, to mistletoes may have been mediated by the close proximity of these plants.This may be also true for Notophorina fusca and Zonopelma from south temperate and subantarctic rain forests in southern Chile developing on Misodendrum, whose closest relatives live in the same habitats and develop mostly on Myrtaceae (Notophorina fusca-group) or on Euphorbiaceae and Rosaceae (Sphinia), respectively.The phylogenetic relationships of the other taxa developing on Santalales are unknown so that their ancestral hosts cannot be inferred.In Acizzia, the Trioza struthanthi-group, Zonopelma as well as the species pairs of Calophya and Afrotropical Psylla, probably all monophyletic, a single shift to Santalales followed by a radiation within the host group is the most likely explanation of the observed host patterns.For the other groups, there is no evidence for a radiation after the host switch.
Psyllids associated with Santalales are relatively species rich in the New World and the Australian biogeographical realm.In the Afrotropical and Palaearctic regions only three, resp.one psyllid species are known from mistletoes.There are no confirmed or likely records from the Oriental region.
Five genera with 28 species of Santalales are known from southern Brazil (Dettke and Waechter 2014a, b) but only three associated psyllid species (Calophya sp., T. struthanthi and T. tripodanthi) could be found, each on a different host genus and species (Table 1), despite intensive field work.A fourth species (Trioza vagata) was collected in a single location and without reliable host information.A similar pattern is found in Chile with 13 genera and 26 species of Santalales (Marticorena and Quezada 1985) but only one associated psyllid species.This suggests that there was only a very limited radiation after the shift to Santalales wich is in contrast to other groups of phytophygous insects.Butterflies of the family Pieridae, e.g., colonised Santalales three times, probably from ancestral Fabaceae, and then successfully radiated on this plant or-der: about 440 of a total of about 1100 species of Pieridae develop on Santalales (Braby and Trueman 2006).

Conclusions
The present study documents that the psyllid fauna associated with mistletoes from the New World is much richer than previously estimated.On the other hand, a critical review of published data shows that the host records from the Oriental region are unlikely or, at least, questionable.
Santalales were colonised by psyllids at least nine times independently.In Acizzia and Freysuila the shift was probably from ancestral Fabaceae, in Calophya from Sapindales and in Notophorina from Myrtaceae.For the remaining groups, the plant taxon from which the psyllids colonised mistletoes remains unknown due to unsolved phylogenetic relationships of the psyllid taxa.Santalales are only sparingly used by psyllids as hosts and there seems to be no major radiation within this host taxon despite multiple colonisation events unlike some other groups of phytophygous insects, such as Pieridae which extensively exploits this plant order (Braby and Trueman 2006).
More targeted field work is necessary to confirm the host patterns described here and to examine the doubtful host records.More phylogenetic data is required to reconstruct ancestral host plants from which the psyllids colonised the mistletoes.

Table 2 .
Measurements in mm of adult Trioza species.