Studies of social wasp diversity in Brazil : Over 30 years of research , advancements and priorities

In the last decade, neotropical social wasps have stood out as role models in studies on ecology, biology and animal behavior (Prezoto et al., 2011; Prezoto & Souza, 2015). This growing interest in the group is due to the acknowledgement of the wasps’ role in the trophic balance of ecosystems, since they can contribute both as pollinators (during the collection of nectar and pollen) and as predators (during their search for the animal protein used in the nourishment of their larvae); thus, wasps show potential as possible agriculture pest control agents (Hunt, 2007; Prezoto et al., 2008; Elisei et al., 2010; Clemente et al., 2013; Clemente et al., 2012, Barbosa et al., 2014). Furthermore, some species Abstract The first records of social wasps in Brazil were made during expeditions focused on the taxonomy and distribution of the species throughout the country. From the 1970s the essence of publications on the diversity of social wasps has been changing, with studies focusing on specific areas and incorporating the use of sampling methodologies and analysis of results through ecological indexes. Since then, the neotropical social wasps have gained more prominence due to the acknowledgement of their decisive role in the trophic balance of ecosystems, which has been increasing the interest in studying these insects. Therefore, we aimed to make a detailed analysis of the social wasp diversity studies published in Brazil over the past 33 years, looking to build knowledge on the research history of the group. For the literature review, selected publications must have attended to the following criteria: including keywords addressing the matter and being indexed in databases within the defined period. We found 78 publications, most of them (70.52%) published in scientific journals. Diversity studies featured in publications in a regular basis from the year 2005 on, and the years 2010, 2012 and 2014 were the most productive; there was also a concentration of studies in the BA, MG and SP states. There were 11 different collection methods used, from which the Active Search and Attractive Trap methods stood out as most common; however, we found no pattern regarding study duration or collection methodology. The contribution of this analysis is to extend the current status of knowledge of social wasps research, as well as to guide and encourage future studies in unexplored areas. Sociobiology An international journal on social insects

Studies of social wasp diversity in Brazil: Over 30 years of research, advancements and priorities

Introduction
In the last decade, neotropical social wasps have stood out as role models in studies on ecology, biology and animal behavior (Prezoto et al., 2011;Prezoto & Souza, 2015).This growing interest in the group is due to the acknowledgement of the wasps' role in the trophic balance of ecosystems, since they can contribute both as pollinators (during the collection of nectar and pollen) and as predators (during their search for the animal protein used in the nourishment of their larvae); thus, wasps show potential as possible agriculture pest control agents (Hunt, 2007;Prezoto et al., 2008;Elisei et al., 2010;Clemente et al., 2013;Clemente et al., 2012, Barbosa et al., 2014).Furthermore, some species

Abstract
The first records of social wasps in Brazil were made during expeditions focused on the taxonomy and distribution of the species throughout the country.From the 1970s the essence of publications on the diversity of social wasps has been changing, with studies focusing on specific areas and incorporating the use of sampling methodologies and analysis of results through ecological indexes.Since then, the neotropical social wasps have gained more prominence due to the acknowledgement of their decisive role in the trophic balance of ecosystems, which has been increasing the interest in studying these insects.Therefore, we aimed to make a detailed analysis of the social wasp diversity studies published in Brazil over the past 33 years, looking to build knowledge on the research history of the group.For the literature review, selected publications must have attended to the following criteria: including keywords addressing the matter and being indexed in databases within the defined period.We found 78 publications, most of them (70.52%)published in scientific journals.Diversity studies featured in publications in a regular basis from the year 2005 on, and the years 2010, 2012 and 2014 were the most productive; there was also a concentration of studies in the BA, MG and SP states.There were 11 different collection methods used, from which the Active Search and Attractive Trap methods stood out as most common; however, we found no pattern regarding study duration or collection methodology.The contribution of this analysis is to extend the current status of knowledge of social wasps research, as well as to guide and encourage future studies in unexplored areas.
The first records on social wasps in Brazil were made during expeditions focused on taxonomy and species distribution by Von Ihering (1904), Ducke (1904Ducke ( , 1905Ducke ( , 1907Ducke ( , 1918)), Zikán (1949Zikán ( , 1951) ) and Araújo (1944Araújo ( , 1946Araújo ( , 1960)).In 1978, Richards publishes the book "The Social Wasps of the Americas, Excluding the Vespinae", which comprehends an extensive review on the neotropical species, with details on their distribution, morphology and biology; this publication instantly became a milestone for posterior studies on the group, and references to it can still be found in recent publications (e.g.Melo et al., 2015;Jacques et al., 2015).
Universidade Federal de Juiz de Fora, Juiz de Fora-MG, Brazil REVIEW After the publishing of Professor Vilma Maule Rodrigues's paper in 1982 on the wasps in the Horto Florestal Navarro de Andrade garden in the city of Rio Claro/SP, Brazil, there was a major change in the essence of publications regarding social wasps; researchers began to focus their sampling effort on a chosen locality, and, as years went by, applied sampling methodologies and data analysis through ecological indexes.
It is estimated that we know less than 10% of the Brazilian insect species (Lewinsohn & Prado, 2005) and although diversity studies are essential to the conservation of species, particularly for social wasps, these efforts must be carried out as to enhance the existing knowledge, thus allowing the comparison between studies and providing information to guide future investigations on the matter.
Therefore, we aimed to analyze in detail the publications on social wasps diversity in Brazil for the last 30 years, aiming to increase the knowledge on these studies through a discussion on the advancements and research priorities regarding the methods applied and the attained results.

Method and Data Search Criteria
On this study we followed the protocol suggested by the PRISMA method for systematic studies and metaanalysis (Moher et al., 2009) adapted by Moher et al. (2015).The methodological approach included the development of the selection criteria, the definition of search strategies, the evaluation of the studies' quality and the extraction of relevant data.
The criteria for selection and inclusion of publication were: publications approaching the matter; publications indexed on the Google Scholar, Scientific Electronic Library Online (SCIELO), Scopus and Web of Science databases; papers published in journals within the period limited between January 1982 and October 2015.The key words used to search publications were 'social wasps' and 'diversity'.Publications such as monographies, theses and books were added through cross-referencing.
We recorded the following data from each publication: study area, focus, duration, sampling methods and identified social wasp species.Based on this information, we generated: (1) a map of the distribution of the publications by state and (2) a table of social wasp species and the methods used to sample them, also sorted by state.

Data analysis
In order to assess social wasp species richness, we generated species rarefaction curves (sensu Gotelli & Colwell, 2001) in the software EstimateS 9 (Colwell, 2013) with 5000 randomizations.This software generates 5000-species accumulation curves by randomizing the order of samples; this way, each point along the curve represents the mean of the accumulated richness for the 5000 curves and is associated to a standard deviation value.Each publication was considered a sample, therefore resulting in 76 samples; two studies were not considered since their authors did not identify organisms to species-level.
We calculated the Constancy Index suggested by Bodenheimer (1955) in order to assess social wasp species constancy recorded in studies in the Brazilian territory.To perform the calculations, once again, each publication was considered a sample.Species present in more than 50% of the samples were considered constant; the ones present in 25% to 50% of the samples were considered accessory; species present in less than 25% of the samples were considered accidental.
The first publication on social wasp diversity in Brazil dates from 1982 (Rodrigues and Machado, 1982); from 1985 to 2002, 12 more studies werepublished on an uneven frequency.From 2005 on, however, some publication regularity started to appear, the years of 2010, 2012 and 2014 being the most productive (with eight, eight and 12 publications respectively) (Fig 1, Table 1).In this chronological sense, it stands out that most papers (n= 45; 57.69%) were published in the last six years (from 2010 to 2016) (Fig 1).
The distribution of publications throughout the Brazilian states showed a concentration in the states of Minas Gerais (n= 24), São Paulo (n= 13) and Bahia (n= 10), which together make up more than half of all studies (Fig 2 ), while nine states still haven't had any published studies on their social wasp fauna.This asymmetry regarding publications may be explained by the fact that Minas Gerais, São Paulo and Bahia host some of the core social wasp research groups in Brazil, present in universities, research institutes and technology centers; these groups perform important roles not only by carrying out studies on the group, but also by developing human resources which would organize new research groups dispersed in other Brazilian states.
However, the increased amount of studies in the states of Minas Gerais e São Paulo does not grant the southeastern region the status of most studied in the country, since the states of Espírito Santo and Rio de Janeiro still have, on their territories, 10.5% and 18.6% (respectively) of the original fragments of the Atlantic Rainforest, which is considered one of the most endangered of the Brazilian biomes (SOS Mata Atlântica, 2013); therefore, it is surprising that there aren't any publications on their social wasp fauna.
Data on the duration of studies was present on 73 (93.58%) publications and ranged from a few days to 144 months, being 12 months the most usual duration (n= 15).In 28 publications the duration was superior to 12 months (from 13 to 144 months), while in other studies (n= 30) the duration was less than 12 months (from 7 consecutive days to 11 months) (Table 1).It is evident that there is no uniformity in the duration of the sampling period, and there is a necessity to create a pattern for study duration in order to enable data comparison between studies.
One of the consequences of the variable duration of studies can be observed when a Species Accumulation Curve is generated (Fig 3).Taking the 24 months-long study performed by Barbosa (2015) as role model, we can relate the curve's behavior to the potential of species to be sampled through time.The Species Accumulation Curve is asymptotic and grows in a decreasing rate, since for each sampling event the potential for finding new species decreases.On the first six months of study this rate is very high, which shows a possibility of sampling a greater number of species in the area.This rate is noticeably lower between six and 12 months, and is minimum after this period (between 12 and 24 months).This curve pattern shows that short-term studies tend to underestimate the number of species in an area, thus reassuring the precision of long-term studies.For the social wasps, it is evident that studies with 12 or more months of samplings have a better estimation of the species diversity when compared to the expected value.
Regarding and Fogging Technique (n= 1) (Table 1).Furthermore, most publications (55.12%, n= 43) used more than one sampling method, which became a trend after 2006; before that, the use of a single sampling method prevailed on most studies.This trend agrees with many studies (Silveira, 2002;Souza & Prezoto, 2006;Togni et al., 2014) that highlight the importance of conciliating methods in order to better record the fauna of wasps in an area.
Sampling through Bait Traps was the second most used method (present in 35 studies), right after Active Search, and was also the method that varied most on its way of application.There were different kinds of baits used, usually made of various fruit juices or sardine-based protein broths; the amount, disposition, duration and confection of traps also varied a lot (e.g.Santos, 1996;Souza & Prezoto, 2006;Clemente, 2009;Locher et al., 2014).This methodological diversity observed for bait traps is mainly due to the lack of a study that tests the best layout for this method; such possibility would generate data to optimize the distance between traps, the setting height for them, the kind of bait used (natural or industrialized juice), the setting duration on field, the container size, and so on.This standardization would bring direct benefits for a more fitting comparison of sampling efforts in future studies, and also to optimize time and money costs to set the traps on the field.
Almost half of the sampling methods used (45.45%,n= 5) recorded exclusive species (recorded by means of a single method).Among those, the methods that recorded the most exclusive species were: Active Search (n= 23) and Malaise Trap (n= 15) (Table 2).Curiously, sampling through Light Traps, recorded for only three studies and characterized as effective for capturing species with night habits, stood out for sampling eight species in the Apoica genus, which is known for its night activity, but also for other 63 species belonging to 13 genera (Agelaia, Angiopolybia, Asteloeca, Brachygastra, Clypearia, Leipomeles, Mischocyttarus, Parachartaegus, Polistes, Polybia, Protonectarina, Pseudopolybia and Synoeca), which are all active during the day (Table 2).
A possible explanation for the capture of that many day species may be due to the Light Traps being controlled by photosensors, which make the traps trigger by the end of the afternoon, a time in which many foragers of wasp species active during the day are still returning to their nests.Social wasps, in the same way as bees, have positive phototropism, which makes them attracted to the luminosity in the trap and therefore captured by it.
Only six studies focused on the difference of the setting height for the sampling methods (Silveira, 2002;De Souza et al., 2011;Somavilla 2012;Somavilla et al., 2014b;Clemente, 2015;Barbosa, 2015), usually adopting two different heights, the canopy (close to 5 meters high) and the understory (chestheight, approximately 1,5 meters high).Two of these studies (De Souza et al., 2011;Barbosa, 2015) recorded greater species richness for the traps set in canopy height and also exclusive species for each setting height.Therefore, these studies show the importance of sampling the different levels of the vegetal mosaic in the environment.
Regarding the use of diversity indexes, we observed that Mechi (2005) was the first study to apply a diversity index; on this particular case, the author used the Shannon-Wiener Index (H') while studying the social wasp fauna in Estação Ecológica Jataí, São Paulo state; the second study to use a diversity index was published by Souza and Prezoto (2006).Most publications (77.27%, n= 34) applied at least one diversity index, which shows the emergence of a trend to use this kind of test in order to discuss the results found since 2006.
Based on the studies that properly identified the social wasp species, 235 species were recorded, belonging to 19 different genera; of these, the most representative ones were the Mischocyttarus (n= 68), Polybia (n= 44) and Polistes (n= 25) genera (Table 2).
This impressive number of species recorded for Amazonas surely does not yet represent the region's mega diversity, since there were only four studies carried out on this state; further investigations should lead to a significant increase of recorded species.On the other hand, the small amount of species recorded on the single study carried out in Goiás (Santos, 1996) shows the particular characteristics of its methodology, since the study which took place at an orchard and not at the state's typical biome environments.
While representatives of the Polistinae are found throughout the whole world, its greatest diversity is achieved in tropical regions (specially the Neotropical region); its worldwide fauna is made of 26 genera and more than 1000 species (Carpenter and Andena, 2013).Some authors (e.g.Fox, 1889;Richard, 1978;Carpenter, 1991;Carpenter and Marques, 2001;Carpenter and Andena, 2013) estimate that Brazil holds 22 genera and 346 species of social wasps.Therefore, based on the properly identified species in the 76 publications hereby listed (Table 1), we observe that the 233 species recorded correspond to 77.74% and 68.62% of the total estimated species.
By generating a Species Accumulation Curve based on the studies and recorded species (Fig 4), the estimators Jack 1 and Jack 2 estimated, respectively, 301.08 and 341.38 species for Brazil, a lower amount than the described in the literature; however, we believe this percentage to be a little higher when we add, to the diversity studies, publications on natural history, biology and ecology of social wasps.
Considering the potential of social wasps as role models for studies in biology, behavior and ecology due to their importance as ecological service providers in the ecosystem, we must highlight the value of studies that investigate the ecology of these species in detail, aiming to further understand this group of organisms.The small amount of studies on social wasp ecology may be a consequence of the sometimes exaggerated behavior of human societies when concerning wasps (by associating wasps to the risk of accidents provoked by their stings) or even of disregard (by believing that these species have no value).Therefore, the analysis presented here may guide and subside future research on social wasp diversity and its ecological relations on the different Brazilian biomes.

Fig 1 .
Fig 1. Number of publications per year, in national and international journals, on Brazilian social wasp diversity, in the period between 1982 to 2015.

Fig 2 .
Fig 2. Geographical representation of the number of publications on social wasp diversity per Brazilian state from 1982 to 2015.

Fig 3 .
Fig 3. Species Accumulation Curve model based on Barbosa (2015), with 24 months of sampling of social wasps.The vertical lines separate the sampling intervals in six, 12 and 18 months.

Fig 4 .
Fig 4. Rarefaction Curves for species richness estimators and Species Accumulation Curves generated through social wasp diversity studies in Brazil from 1982 to 2015, made from 5.000 randomizations on the sample order (see details in Data Analysis).
Nº of species found; III -Nº of genera found; IV -Nº of swarming species and V -Nº of independent species.(Continuation)

Table 2 .
.51 Social wasp species, studied environment, sampling methods and publication references sampling species diversity in Brazil from 1982 to 2015: Studied Environment: AG -Agroecosystem;

Table 2 .
Social wasp species, studied environment, sampling methods and publication references sampling species diversity in Brazil from 1982 to 2015: Studied Environment: AG -Agroecosystem; .51

Table 2 .
Social wasp species, studied environment, sampling methods and publication references sampling species diversity in Brazil from 1982 to 2015: Studied Environment: AG -Agroecosystem; .51

Table 2 .
Social wasp species, studied environment, sampling methods and publication references sampling species diversity in Brazil from 1982 to 2015: Studied Environment: AG -Agroecosystem;