Review of the inquilinous fauna associated with insect galls in Brazilian restingas

Data on the arthropod fauna associated with insect galls have been reported by several inventories in five phytogeographic domains in Brazil: (1) Amazon Forest (Maia, 2011; Carvalho and Motta, 2018), (2) Atlantic Forest (Maia, 2001, 2013, 2014; Maia et al., 2008, 2014; Bregonci et al., 2010; Rodrigues et al., 2014; Maia and Souza, 2013; Carvalho-Fernandes et al., 2016; Maia and Carvalho-Fernandes, 2016; Maia and Mascarenhas, 2017; Ansaloni et al., 2018; Flor et al., 2018; Maia and Siqueira, 2020), (3) Caatinga (Carvalho-Fernandes et al., 2012; Costa et al., 2014; Brito et al., 2018), (4) Cerrado (Fernandes et al., 1988; Urso-Guimarães et al., 2003; Maia and Fernandes, 2004; Araújo et al., 2007; Santos et al., 2010; Araújo et al., 2011; Saito and Urso-Guimarães, 2012; Santos et al., 2012, 2018; Bergamini et al., 2017; Lima and Calado, 2018; Silva et al., 2018 a, b; Vieira et al., 2018; Abstract Several records of associated fauna, including parasitoids, inquilines, predators, and successors, have been reported by insect gall inventories in Brazilian restingas. Although most guilds are well established, inquilines have frequently been misinterpreted. In this paper, the inquilinous fauna of insect galls is revised based on five criteria: food habit; coexistence with the inducer; modification of gall tissues or production of new tissues; phylogenetic relationship with the inducer; and mobility. Gall inventories dated from 1988 to 2019 were examined, totaling 16 publications, eight of them with inquiline records. This guild was reported in 53 gall morphotypes in 44 plant species and four morphospecies distributed among 36 genera of 24 host families for a total of 65 records. Most inquilines were repositioned into the cecidophage guild and others into the kleptoparasite guild, resulting in a large reduction in the frequency of inquilines (from 65 to five records), and in first reports of cecidophages and kleptoparasites, with 46 and 13 records, respectively. Cecidophage was the most diverse guild with insects of five orders (Diptera, Coleoptera, Lepidoptera, Hemiptera, and Thysanoptera) while kleptoparasites were represented only by two orders (Diptera and Hymenoptera) and inquiline solely by Hymenoptera. Other results indicate that Leptothorax sp. (Formicidae) could be a successor and not an inquiline.


Introduction
Data on the arthropod fauna associated with insect galls have been reported by several inventories in five phytogeographic domains in Brazil: (1) Amazon Forest (Maia, 2011;Carvalho and Motta, 2018), (2) Atlantic Forest (Maia, 2001(Maia, , 2013(Maia, , 2014Maia et al., 2008Maia et al., , 2014Bregonci et al., 2010;Rodrigues et al., 2014;Maia and Souza, 2013;Carvalho-Fernandes et al., 2016;Maia and Carvalho-Fernandes, 2016;Maia and Mascarenhas, 2017;Ansaloni et al., 2018;Flor et al., 2018;Maia and Siqueira, 2020), (3) Caatinga (Carvalho-Fernandes et al., 2012;Costa et al., 2014;Brito et al., 2018), (4) Cerrado (Fernandes et al., 1988;Urso-Guimarães et al., 2003;Maia and Fernandes, 2004;Araújo et al., 2007;Santos et al., 2010;Ribeiro et al., 2019), and (5) Pantanal (Ascendino and Maia, 2018). Several gall inventories of Atlantic Forest areas were carried out in restingas, mainly in the state of Rio de Janeiro (Maia, 2001;Rodrigues et al., 2014;Carvalho-Fernandes et al., 2016), but also in Espírito Santo (Bregonci et al., 2010), and São Paulo (Maia et al., 2008). This fauna has been classified as parasitoids, inquilines, gall modifiers, successors, and predators. Nevertheless, other guilds are known, such as cecidophages, symbionts and kleptoparasites, but they were not been cited in these publications, probably due to terminological problems. Mani (1964) defined most of these guilds. According to this author, successors are organisms that use the gall after the inducer leaves. They are mostly mites, spiders, thrips, beetles, ants and springtails, which apparently use gall tissues for food and shelter, but dead or decaying gall tissue is probably attacked by fungi that also serve as food. Inquilines and cecidophages use galls simultaneously with the galler. Both are phytophagous and consume gall tissues, but only the former promote the production of new tissues (they have been referred to as "gall modifiers" in some inventories). Kleptoparasites steal the resource (in this case the gall) from other organism, causing the galler death or its expulsion from the gall (Mound and Morris, 2000). Predators invade galls and feed on their inducers, killing them immediately, such as some species of Carabidae and Coccinellidae, for example (Ramamurthy, 2007). Parasitoids are organisms that live in close association with the host at the host's expense, which results in its death when the parasitoids complete their development (Godfray, 1994).
Guilds of predators, parasitoids, and successors are well delimited in the literature, while inquilines are a major problem as historically all organisms associated with galls that were not predators or parasitoids were placed in this guild (Redfern and Askew, 1992). Mani (1964), Godfray (1994), and Mound and Morris (2000) provided tools for distinguishing among these guilds, but none of them used multiple non-superposed parameters. Luz and Mendonça-Júnior (2017) proposed five interaction parameters to conceptualize and differentiate inquilines, cecidophages and kleptoparasites from each other: food habit; coexistence with the inducer; modification of gall tissues or production of new tissues; phylogenetic relationship with the inducer; and mobility. According to these authors, inquilines are exclusively phytophagous, coexist with the gall-inducer, modify gall tissues or stimulate production of new tissues, have a close phylogenetic relationship with the inducer and are sedentary. Cecidophages are exclusively phytophagous, coexist with the gall-inducer, but do not modify gall tissues or stimulate production of new tissues, do not have a close phylogenetic relationship with the inducer and have high mobility, while kleptoparasites are omnivorous, do not coexist with the inducer, do not modify gall tissues or stimulate production of new tissues, can have a close phylogenetic relationship with the inducer and have low mobility. Based on these criteria, the inquilinous fauna associated with insect galls in Brazilian restingas is reviewed herein. This paper also aims to answer the following questions: (1) Which are the most frequent and richest taxa of inquilines, cecidophages, and kleptoparasites? (2) Which gall-inducing taxa host the richest fauna of inquilines, cecidophages, and kleptoparasites? (3) In how many gall morphotypes and plant species, genera and families have these guilds been recorded? (4) Which plant families, genera and species shelter the richest guilds of inquilines, cecidophages, and kleptoparasites? (5) What are the most frequent features of host galls? (6) What is known about the taxonomy of the inquilines, cecidophages, and kleptoparasites of gall? and (7) What is necessary for the correct positioning of the associated fauna in the different guilds?

Material and Methods
A survey of scientific papers involving Brazilian restingas published from 1988 to 2019 was carried out by consulting the database "Web of Science" using "gall" or "galha" and "restinga" as keywords. Sixteen publications about galls in Brazilian restingas were found and examined. Eight of the publications provided data on inquilinous fauna, which were retrieved, compiled and analyzed, following the criteria established by Luz and Mendonça-Júnior (2017). The new positioning of associated fauna among guilds was based on literature data, as well as on unpublished biological information obtained from laboratory works, including gall arthropod rearing, gall dissection and observation of gall tissues and gall dweller behavior. These laboratory works were carried out during inventories of which I have participated, namely Maia, 2001;Maia et al., 2008;Bregonci et al., 2010;and Rodrigues et al. 2014. Finally, the positioning of the associated fauna into parasitoid and predator guilds was not revised, since the conceptions of these guilds are well understood.

.2.1. Cecidophage guild
Although Coleoptera, Sciaridae, Tephritidae, Chloropidae, Muscomorpha, Clinodiplosis sp. (Cecidomyiidae), Hemiptera, Lepidoptera, and Thysanoptera were represented among the taxa considered inquilines, biological observations indicated these to be cecidophages since they feed on galls without modifying them or stimulating production of new tissues, coexisted with the inducer and had high mobility.
Other laboratory observations showed that immature stages of dipterans, coleopterans and lepidopterans occurred in galls, but not the adults, since they left galls immediately after their emergence. Different from these insects, adult hemipterans and thysanopterans were frequently observed in galls together with their eggs and nymphs. Pupal exuviae of coleopterans, lepidopterans, tephritids and chloropids were found in galls, while those of sciarids remained attached to gall openings, in the same way as cecidomyiid exuviae.
Field and laboratory observations showed that caterpillars, such as that of Stenoma annosa, for example, could be voracious, feeding and destroying several galls (Butler, 1877). Furthermore, their excrement is accumulated in the internal chamber(s).
with more than one cecidophage record. The first hosted aphidids, coccids, tingids (Hemiptera), and curculionids (Coleoptera) in only one gall morphotype; the second hosted curculionids and lepidopterans in one morphotype as well; the third hosted curculionids and lepidopterans in two morphotypes; and the last, hosted lepidopterans in two gall morphotypes.
Concerning plant organs, cecidophages were obtained from galls on leaves, stems, buds, tendrils, flowers, and fruits, with leaf galls being most frequent (61%). They occurred in galls of several shapes, but mainly in globoid galls (31%). Although they were obtained from green, brown, yellow, red, and purple galls, 67% occurred in green galls, and 95% occurred in glabrous galls.

Inquiline guild
Only eulophids (Hymenoptera) were considered inquilines in this revision, since their larvae were sedentary, had a close phylogenetic relationship with the gall-inducer, and whenever they were present, the gall-inducers died.
Inquilines were found mainly in bud galls, but also in leaf galls. They occurred in conical, ovoid, and cylindrical galls, with the first two being more frequent. They were obtained mainly from green galls, but also from brown galls, and all occurred in glabrous galls. Inquilines were identified only to the family level.

Sucessors
Leptothorax sp. (Formicidae) was reported as an inquiline of cylindrical galls on Eugenia hiemalis Camb. (Myrtaceae) by Maia (2001), but no gall modification was related to the presence of this ant. Furthermore, Leptothorax sp. could not be considered cecidophagous since it did not feed on gall tissues. Maia (2001) reported Leptothorax sp. as a successor in similar galls on Eugenia astringens Cambess. (Myrtaceae). These two cases are probably the same.

Kleptoparasites
Larvae of species of Contarinia, Resseliella, and Trotteria (Cecidomyiidae) were recorded in galls of other cecidomyiids, indicating a close phylogenetic relationship with the inducer since all belong to the same family. Whenever they were observed, they were the unique dwellers of galls, so the criterion of non-coexistence was fulfilled. Furthermore, these larvae showed low mobility and did not modify or stimulate production of new gall tissues. Therefore, they can be considered as kleptoparasites. Nevertheless, the criterion of food habit was not met, since their larvae were phytophagous and not omnivorous.

Gall-inducing taxa and cecidophage, inquiline and kleptoparasite guilds
Cecidophages, inquilines and kleptoparasites were recorded in 46 gall morphotypes induced by cecidomyiids (Diptera) and in two morphotypes induced by hemipterans. Cecidomyiid galls comprised all three of these guilds, while those of hemipterans sheltered only cecidophages. Some kleptoparasites and cecidophages were obtained from five gall morphotypes (9% of the total) whose inducers are still unknown.
Cecidophages, inquilines and kleptoparasites were associated with 16 species and 15 morphospecies of gallinducing cecidomyiids of 18 genera. Among these, galls induced by species of Lopesia Rübsaamen, 1908, hosted the greatest variety of guilds (cecidophages, inquilines, and kleptoparasites) and associated taxa (Coleoptera, Eulophidae, Cecidomyiidae, and Lepidoptera), followed by galls induced by species of Stephomyia Tavares, 1916, the galls of which sheltered inquilines (Eulophidae) and kleptoparasites (Trotteria sp.). Galls induced by species of Clinodiplosis Kieffer, 1894, hosted only cecidophages, as did those induced by species of Neolasioptera Felt, 1908, Asphondylia Loew, 1850, and Bruggmannia Tavares, 1906. Membracids, sciarids and curculionids were recorded in galls of Clinodiplosis, while thysanopterans and lepidopterans were recorded in galls of Neolasioptera. Representatives of Diptera (Cecidomyiidae and Trypanea sp.: Tephritidae) and Thysanoptera were obtained from galls of Asphondylia, whereas representatives of Hemiptera (Aphididae) and Lepidoptera were obtained from galls of Bruggmannia. The other gall midge genera sheltered a single guild and a single insect taxon.
Gall-inducing hemipterans were identified only to the level of order. Their galls hosted cecidomyiids, one of them identified to genus (Clinodiplosis sp.) and the other to family.
The new composition of these guilds allows the questions proposed at the beginning of this paper to be answered: 1) The most represented taxa among cecidophages were Lepidoptera and Coleoptera. Inquilines represented only Eulophidae (Hymenoptera) and kleptoparasites mainly Cecidomyiidae (Diptera). The richest taxa were not indicated, as most records were in suprageneric categories, without the discrimination of morphospecies; 2) Cecidomyiidae is the gall-inducing taxon with the richest fauna of inquilines, cecidophages and kleptoparasites; 3) Inquilines were reported in four gall morphotypes, four plant species, four genera and three families; cecidophages in 39 gall morphotypes, 35 plant species, 28 genera, and 20 families; and kleptoparasites in 12 gall morphotypes, eight plant families, 11 genera, and 12 species; 4) Asteraceae and Myrtaceae, Mikania (Asteraceae), Andira (Fabaceae), and Guapira (Nyctaginaceae), Mikania cf. biformis, Andira fraxinifolia (Fabaceae), Guapira opposita (Nyctaginaceae), Piptocarpha cf. cinerea, and Neomitranthes obscura (Myrtaceae) were the plant taxa with the richest cecidophage guild. No plant taxa were highlighted as sheltering the richest kleptoparasite guild or inquiline guild; 5) Cecidophages, inquilines and kleptoparasites were recorded mainly in leaf, green, and glabrous galls. The first occurred more frequently in globoid galls, the second in conical and ovoid galls, and the last in conical galls; 6) The taxonomic knowledge of these guilds remains poor, since most were identified to suprageneric categories, and only four to species level; 7) Biological data are necessary for the correct positioning of associated fauna into guilds.

Discussion
In this review, the cecidophage guild was easily determined based on the five criteria proposed by Luz and Mendonça-Júnior (2017). On the other hand, some conceptual problems were faced in determining inquilines and kleptoparasites. According to these authors, inquilines have a close phylogenetic relationship with the gall-inducing species. This criterion was not met, but the other four (food habit, coexistence with the inducer, modification of gall tissues or production of new tissues, and mobility) were fulfilled. Regarding food habits, kleptoparasites are defined as omnivorous, which is a criterion not met by gall midges, although they met the other criteria. The use of these five criteria is only possible when biological data are known, which demands long-term studies.
The presence of cecidophages, inquilines and kleptoparasites can indirectly cause the death of the inducer, which can be important for population dynamics of gallinducing species. However, this effect was little discussed in Luz and Mendonça-Júnior (2017). Nevertheless, these guilds should be considered since their presence can be an important mortality factor, as indicated by Maia, 2001. Cecidophages were represented mainly by lepidopterans and coleopterans. These two insect orders were also indicated as the most represented among the secondary fauna inhabiting galls of cynipids (Hymenoptera) (Giannetti et al., 2019).
Cecidophage and kleptoparasite guilds were obtained mainly from leaf, green, and glabrous galls. These are the predominant features of insect galls in Brazilian restingas (Maia, 2001;Maia et al., 2008;Rodrigues et al., 2014), suggesting that guild frequency is related to resource availability. Nevertheless, other gall features were predominant for inquilines, but the number of records was too low to make generalizations. Most records were for cecidomyiid galls, the most diverse, abundant and frequent gall-inducing taxon in restingas (Maia, 2001;Maia et al., 2008;Rodrigues et al., 2014).
Asteraceae and Myrtaceae hosted the greatest richness and frequency of cecidophages, which was expected since they are the plant families with the greatest gall richness in restingas (Maia, 2013). Both families are well represented in this ecosystem (Lourenço and Barbosa, 2012;Melo-Júnior and Boeger, 2018). No plant family stood out as exhibiting the greatest number of kleptoparasite or inquiline records since both guilds were similarly distributed among different families. However, their records are also few, so new and broad studies are likely to modify this scenario.
The plant genera and species with the highest number of cecidophages were not necessarily those that hosted the greatest number of galls, since only Mikania and Guapira opposita (Nyctaginaceae) have been been cited as super host taxa (Maia and Oliveira, 2010;Maia, 2013;Rodrigues et al., 2014). Thus, cecidophage richness appears not to be related to gall richness. In fact, some galls can be more attractive than others, probably due to their own morphological and chemical features.
The high frequency of cecidophagy found here reveals the importance of gall-inducing insects as ecosystem engineers in restingas. Inducers do not merely provide habitat for specialists, but can also influence the structure of communities that do not directly interact with galls, as Wetzel et al. (2016) has shown. They can have significant impacts on the herbivore community not only by changing plant morphology, but also by altering host quality and modifying plant-induced responses to subsequent herbivory (Uesugi et al., 2016).
The taxonomic knowledge of cecidophage, kleptoparasite, and inquiline guilds remains still poor. The scarcity of identified species, for example, does not allow discussions to be made about their specificity.
A more complete review of guild richness, including also parasitoids, predators and symbionts, can contribute to revealing the importance of associated faunas for gall systems. However, in order to know the composition of each guild, specific identification is essential, as well as the correct categorization of inquilines, kleptoparasites and cecidophages, which depends on taxonomical and biological data, respectively.

Conclusions
Although cecidophages were not previously recorded in insect gall inventories in Brazilian restingas, they are actually frequent. Kleptoparasites are also present. Both of these guilds are formally reported here for the first time in this ecosystem. On the other hand, the frequency and diversity of inquilines are low, differing from literature data. These new records resulted from recategorization among guilds based on literature data and biological observations, following criteria proposed by Luz and Mendonça-Júnior (2017).
Although the kleptoparasites found did not fulfill the food habit criterion, and the inquilines did not satisfy the phylogenetic relationship criterion, both guilds could be easily established based on the other four criteria. In my opinion, future studies about insect galls and associated fauna should adopt these criteria to avoid misinterpretation and improve knowledge about these guilds in Brazil.