Open and Disturbed Habitats Support Higher Diversity of Syrphidae

Journal of Insect Science RESEARCH

Open and Disturbed Habitats Support Higher Diversity of Syrphidae (Diptera)? A Case Study During Three Yr of Sampling in a Fragment of Araucaria Forest in Southern Brazil Jana Magaly Tesserolli de Souza,1 Renato Contin Marinoni,2 and Luciane Marinoni Departamento de Zoologia, Universidade Federal do Parana´, Caixa Postal 19020, CEP 81531-980, Curitiba, Parana´, Brazil 1Corresponding author, e-mail: [email protected] 2In memoriam

Subject Editor: Marc De Meyer

J. Insect Sci. 14(236): 2014; DOI: 10.1093/jisesa/ieu098

ABSTRACT. The diversity of hoverflies in five different habitats of the Parque Estadual de Vila Velha, Ponta Grossa, Parana´, Brazil, from September 1999 to August 2002 was studied. These five areas had been exposed to various types of anthropogenic disturbance at different times, which resulted in different stages of plant succession. In total, 2,841 specimens of 151 species of Syrphidae were collected using malaise traps. The highest species richness and abundance were found at the edge of the forest (Forest Edge), whereas the lowest was found in the most conserved areas. Evenness values increased along the succession gradient. The accumulation curve of new occurrences of syrphid species nearly approached an asymptote starting on the third month of the third year. Estimates of syrphid species richness using different methods indicated that between 155 and 288 species are found in the study area, according the Michaelis–Menten and the Chao2 estimators, respectively. The richness and abundance of representatives of the subfamily Syrphinae were higher in more open areas, where herbaceous plants predominate. Nineteen hoverfly species can be considered as environmental indicators, as they were collected exclusively or mainly in one of the defined habitat types. The degree of opening of the vegetation seems to be the factor determining the distribution of Syrphidae species, a conclusion based on the finding that grassy clearings in Araucaria Forests had more species. Hence, we conclude that, to become established, Syrphidae communities need conserved mosaic landscapes.

Key Words: indicator species, hoverﬂy, malaise trap, plant succession, species richness estimator

Insects provide easy, sensitive, and inexpensive ways to measure the limpos” associated with the Mixed Ombrophilous Forest (Araucaria effects of anthropogenic stress on the biodiversity and the environment Forest), which may form gallery forests or isolated woods across the (Kim 1993). Among them, the species of Syrphidae have great potential landscape, where the presence of the Araucaria angustifolia (Bertol.) as indicators of the environmental quality of forests (Thompson 1999, Kuntze (Parana-pine) is evident (Maack 1981, Veloso and Goés Filho Smith et al. 2008). This is mainly because they are easily found in 1982). nearly all terrestrial ecosystems, and their larvae have highly diverse The survey was carried out in an area of Araucaria Forest at the feeding habits (Sommaggio 1999). PEVV. Five sites, characterized by varying degrees of plant succession Entomofaunistic surveys are relevant for understanding how insect (or environmental conditions due to management) were chosen: 1) for- communities are structured and may broaden our knowledge on est edge (located between the arboreal vegetation and the open area), 2) Neotropical Syrphidae. Thus, this study is an extension of the one de- Araucaria (planted forest of A. angustifolia); 3) Phase 1 (native forest in veloped by Jorge et al. (2007) on the syrphid diversity at the Parque an initial to intermediate stage of plant succession); 4) Phase 2 (native Estadual de Vila Velha (PEVV), Parana´ State, Southern Brazil. Jorge forest in an intermediate to advanced stage of plant succession); and 5) et al. (2007) studied the species collected during the first year of the Phase 3 (native forest in the most well conserved stage). The informa- project “Levantamento da Fauna Entomolo´gica no Parque Estadual de tion on the history of the anthropogenic disturbance and floristic differ- Vila Velha, Ponta Grossa, Parana´ – PROVIVE” (Ganho and Marinoni ences among the areas provided the basis for their recognition as 2003). Our investigation, carried out over a period of 3 yr, has gathered different habitats across the landscape (Ganho and Marinoni 2003). more striking results. We were able to perform both spatial and tempo- Collection, Mounting, and Identification. Insects were collected ral assessments of the diversity of Syrphidae and to establish relation- using malaise traps (Townes 1972). This collecting method has been ships between community structure in each of the five surveyed areas particularly useful in quantitative studies of Syrphidae (Owen and and their conservation status. Because differences in species composi- Owen 1990, Sommaggio 1999), even though it is logistically costly, tion decrease with increasing mobility of the organisms studied, we which makes it necessary to reduce replications (Burgio and also evaluated whether syrphid species, known for their outstanding Sommaggio 2007). Even if a small number of malaise traps are used, capacity to fly, are distributed evenly across the environment. the information provided by long-term monitoring can be very useful Furthermore, this study revealed potential indicator species of because these traps can sample both singletons and the most abundant Araucaria Forest habitats. species (Fraser et al. 2008). A trap was placed in each site, and the material was removed weekly for 3 yr (September 1999 to August 2002), Materials and Methods totaling 153 samples per area. Characterization of the Biome. The vegetation formation Campos After sorting and mounting the specimens, we identified them based Gerais do Parana´ is located in the Eastern region of the second plateau on keys available in the literature (Thompson 1999, Borges and of Parana´ (Southern Brazil), immediately following the Devonian Pamplona 2003, Marinoni et al. 2007, Borges and Couri 2009)by ridge, including a broad north–south stretch of natural grassland forma- comparing them with reference material in the Coleçaõ Entomolo´gica tions. It is characterized as a phytogeographical zone of “campos Padre Jesus Santiago Moure of the Zoology Department at the

VC The Author 2014. Published by Oxford University Press on behalf of the Entomological Society of America. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected] 2 JOURNAL OF INSECT SCIENCE VOLUME 14

Universidade Federal do Parana´ (DZUP-UFPR) and using the help of Dr. Christian F. Thompson of the National Museum of Natural History, Table 1. Total abundance of Syrphidae and its subfamilies resulting Smithsonian Institution. As a result of our investigations, some changes from a survey using malaise traps in five areas of the PEVV, Ponta in the definition of complexes of cryptic species, as well as species with Grossa, Parana´, shown separately for each of the 3 yr and for the high intraspecific variability, were made, resulting in slight differences entire dataset (September 1999 to August 2002) between the current data and those presented by Jorge et al. (2007). Eristalinae Microdontinae Syrphinae Syrphidae These differences, however, have not resulted in a noticeable effect in 1999–2000 Forest edge 16 5 663 684 the comparisons of species richness and abundance and, therefore, have Araucaria 11 42 109 162 not altered the interpretations made in that study. Phase 1 26 33 191 250 The material has been deposited in the Coleçaõ Entomolo´gica Phase 2 26 26 51 103 Phase 3 12 44 90 146 Padre Jesus Santiago Moure (DZUP-UFPR). All the data are at the Total 91 150 1,104 1,345 database of the Taxonline - Rede Paranaense de Coleço˜es Biolo´gicas 2000–2001 Forest edge 4 10 459 473 (www.taxonline.ufpr.br) and the species list is in Appendix 1. Araucaria 8 25 39 72 Data Analysis Phase 1 6 14 48 68 Abundance and Species Richness. Abundance and species rich- Phase 2 5 26 22 53 Phase 3 4 13 15 32 ness were compared among the areas and the subfamilies of Syrphidae Total 27 88 583 698 (Eristalinae, Microdontinae, and Syrphinae) using generalized linear 2001–2002 Forest edge 4 4 534 542 models with a Poisson error distribution. Analyses and graphs were car- Araucaria 5 28 49 82 ried out in STATISTICA software 7.1 (StatSoft 2005). Phase 1 6 5 45 56 Evenness Index. The Simpson’s index (SD) provides the probabil- Phase 2 7 20 33 60 Phase 3 4 14 40 58 ity that two randomly selected individuals from an infinitely large com- Total 26 71 701 798 munity belong to the same species. It is one of the most popular and 1999–2002 Forest edge 24 (1) 19 (1) 1,656 (98) 1,699 robust diversity measures, because it captures the variance in the distri- Araucaria 24 (8) 95 (30) 197 (62) 316 bution of species abundance. It is heavily weighed by the most abun- Phase 1 38 (10) 52 (14) 284 (76) 374 Phase 2 38 (18) 72 (33) 106 (49) 216 dant species in the sample and less sensitive to species richness Phase 3 20 (8) 71 (30) 145 (61) 236 (Magurran 2004). The complement of this index (1 SD ¼ SE) can be Total 144 (5) 309 (11) 2,388 (84) 2,841 considered a way of expressing the Simpson’s evenness index. The Values in parentheses correspond to relative frequency in percentage. dominance index of Berger and Parker (BPD) is defined by the ratio of the number of individuals in the most abundant species by the total number of captured individuals. The reciprocal of this index (1/BPD ¼ BPE) is the Berger and Parker evenness index. This index, in and low humidity; 3) Phase 2 þ Phase 3, forest areas in their intermedi- addition to being simple and informative, is considered as one of the ate and advanced succession stages, better-conserved than (2), with less most satisfactory measures of diversity (Magurran 2004). Because of canopy openings, lower luminosity, and higher humidity; 4) the sensitivity of the first index to species abundance distribution, and Araucaria þ Phase 1 þ Phase 2 þ Phase 3, that encompassed all the of the second one to species richness, we used both evenness indexes. woody vegetation, not considering successional level and interior envi- Estimate of Species Richness. To assess syrphid richness estimates ronmental conditions. of the areas, the following methods were used: Bootstrap, Chao 1, From the definition of these habitat types, the species that better Chao 2, first-order Jackknife, second-order Jackknife, and characterize each of them were determined using an indicator value Michaelis–Menten. According to Palmer (1990), nonparametric esti- index (Dufreˆne and Legendre 1997), whose significance was obtained mators obtain better results than their alternatives. These calculations after 999 permutations. This analysis was performed using R 2.13.1 were conducted using the software EstimateS Version 8.0 (Colwell (R Development Core Team 2011) with the package labdsv (Roberts 2006), with 50 randomizations with abundance classes equal to 10. 2010). The indicator value index considered: 1) the specificity of a Calculations were conducted using both the classical and the corrected given species to the area and 2) its fidelity to the same area over the 3 Chao, always recording in the table the highest value obtained by either sampled years. When analyses were performed with the same sample method, which in our case resulted from the classic Chao. The estimates units, but combined in five, three, or two groups, we found that the spe- were first calculated with the data from the first annual cycle (52 wk) cificity and fidelity components are more easily achieved when a for each area separately and for PEVV as a whole. Then, the same pro- smaller number of sample units are considered for each group. cedure was followed using the combined 2-yr effort, and finally, using Therefore, for the only area that was not grouped with any other (Forest the data for the entire survey (three annual cycles, for a total of 153 wk). edge), the significance was taken from the preliminary analysis that Community Structure. The nonmetric multidimensional scaling considered the classification into five areas, without any grouping. This (NMDS) was used to assess the general dissimilarity pattern in the com- procedure minimized possible biases in the significance of the specific munity structure among the five areas surveyed. This analysis was per- indicator values. Only the taxa identified to species level were consid- formed using both total abundance of Syrphidae and abundance of each ered in this analysis. Only statistically significant values (a ¼ 0.05) subfamily separately. Data were log transformed prior to analyses, and were reported, implying that the Syrphidae species that presented such the Morisita–Horn coefficient was used to generate dissimilarity matri- values would have a preference for a particular type of habitat in the ces. NMDS was performed using R 2.13.1 (R Development Core Team Araucaria Forest. 2011) with the package MASS (Venables and Ripley 2002), and graphs were generated in STATISTICA 7.1 (StatSoft 2005). Results Indicator Species. Based on floristic differences previously noticed Abundance. At the end of the 3 sampling years, a total of 2,841 syr- (Ganho and Marinoni 2003), evenness values, number of shared spe- phids had been collected in all five areas, as follows: 1,345 in the first cies, and community structure analyses, some groups of areas were year, 698 in the second, and 798 in the third. When the data of all years defined before conducting the analysis on the indicator species: 1) were analyzed together, the Forest edge had the largest number of syr- Edge, an environment with open vegetation, corresponding to a contact phids, followed by Phase 1, Araucaria, Phase 3, and Phase 2, respec- zone between forest and gaps; 2) Araucaria þ Phase 1, forest areas in tively (Table 1). There was a significant effect of the interaction their initial and intermediate stages of succession, corresponding to a between area and subfamily on Syrphidae abundance (Wald disturbed habitat with significant canopy opening, intense luminosity Stat. ¼ 383.06, df ¼ 8, P < 0.001). In this case, Syrphinae was more 2014 SOUZA ET AL.: DIVERSITY OF SYRPHIDAE IN DIFFERENT HABITATS 3

Table 2. Species richness of Syrphidae and its subfamilies resulting from a survey using malaise traps in ﬁve areas of the PEVV, Ponta Grossa, Parana´, shown separately for each of the 3 yr and for the entire dataset (September 1999 to August 2002) Eristalinae Microdontinae Syrphinae Syrphidae 1999–2000 Forest edge 8 4 43 55 Araucaria 5 13 18 36 Phase 1 14 12 24 50 Phase 2 7 7 16 30 Phase 3 8 11 12 31 Total 26 18 57 101 2000–2001 Forest edge 4 8 42 54 (25) Araucaria 6 9 14 29 (11) Phase 1 3 7 20 30 (6) Phase 2 2 8 7 17 (4) Phase 3 4 6 7 17 (3) Total 13 18 48 79 (25) 2001–2002 Forest edge 3 4 43 50 (18) Araucaria 4 12 14 30 (3) Fig. 1. Number of individuals of Syrphidae (mean 6 SE) in each Phase 1 3 5 20 28 (7) subfamily and each area of the PEVV, Ponta Grossa, Parana´, collected Phase 2 4 6 9 19 (6) with malaise traps between September 1999 and August 2002. Phase 3 3 5 13 21 (6) Total 9 16 52 77 (25) 1999–2002 Forest edge 12 15 71 98 abundant than the other subfamilies in the Forest edge. This area also Araucaria 9 16 25 50 Phase 1 14 14 35 63 had a higher abundance of Syrphinae than the forested areas (Fig. 1). Phase 2 8 8 24 40 Syrphinae had 69% of its total abundance in Forest edge, followed by Phase 3 11 12 17 40 only 12% in Phase 1. The abundance of the other subfamilies was simi- Total 35 28 88 151 lar across all areas, and the abundance of all Syrphidae subfamilies in Values in parentheses correspond to new records of species in each area forested areas was similar (Fig. 1). However, in the forested areas, and period. Eristalinae and Microdontinae were relatively more frequent (Table 1). Species Richness. In total, 151 species of Syrphidae were collected, belonging to 27 genera. The largest number of species was recorded from the Forest edge. In the forested areas, richness was highest in Phase 1, and the fewest number of species were recorded in Phase 3(Table 2). When subfamilies are considered separately, the largest number of species collected belonged in Syrphinae, representing 65% of the total species. Furthermore, there was a significant effect of the interaction between area and subfamily on Syrphidae species richness (Wald Stat. ¼ 42.22, df ¼ 8, P < 0.001). Syrphinae also had the highest species richness among all subfamilies in Forest edge and in Phase 1. It also surpassed Eristalinae in species richness in the Araucaria area (Fig. 2). The Forest edge had the largest number of Syrphinae species, whereas species richness did not vary among areas for the other subfamilies (Fig. 2). Abundant Species and Rare Species. After 3 yr of sampling, the six most abundant species accounted for 51% of all syrphids collected, namely Toxomerus polygraphicus (Hull, 1940), Syrphus phaeostigma Wiedemann, 1830, Allograpta neotropica Curran, 1936, Toxomerus procrastinatus Metz, 2001, Toxomerus croesus (Hull, 1940), and Ocyptamus funebris Macquart, 1834 (Appendix 1). Ocyptamus Fig. 2. Richness of Syrphidae species (mean 6 SE) in each subfamily arabella Hull 1947 was recorded for the first time in the state of Parana´. and each area of the PEVV, Ponta Grossa, Parana´, collected with All of these species belong to the subfamily Syrphinae. malaise traps between September 1999 and August 2002. In Microdontinae, the following species were noteworthy: Mixogaster polistes Hull, 1954, Microdon aureopilis Marinoni, 2004 Simpson’s index, was higher for the forested areas. In contrast, even- and Microdon aurifex Wiedemann, 1830, which occurred in all areas ness as calculated by the Berger and Parker’s index increased with the (Appendix 1). In the case of Eristalinae, the most abundant species level of plant succession (Table 3). included Copestylum macquarti (Curran, 1926), Chalcosyrphus Species Accumulation Curve. The cumulative number of species armatipes (Curran, 1941), and Copestylum selectum (Curran, 1939) throughout the survey in all five areas indicates that more than half of (Appendix 1). the species (52%) had already been collected in the first 5 mo of sam- The survey resulted in a large number of species represented by a pling, and at the end of the first year, 67% of the species had been few specimens and a few species represented by many individuals. sampled. In each of the following years, the number of species sampled From the 151 species collected, 110 were represented by fewer than 10 increased by 16% (Fig. 3). The curve did not reach an asymptote before specimens, which correspond to 73% of all species. the third year of sampling, i.e., there were new records of species for Evenness Index. With respect to both indexes applied, the Forest nearly every month during the first 2 yr. This indicates that the 3-yr edge was the least uniform area, particularly due to the presence sampling effort was important to provide more robust results to the sur- of dominant species such as T. polygraphicus (495 individuals) vey of syrphid fauna of PEVV. In the last months of the third year, the and S. phaeostigma (397 individuals). Evenness, according to the curve tended toward stability (Fig. 3). 4 JOURNAL OF INSECT SCIENCE VOLUME 14

The species accumulation curves for the forested areas indicate that edge, the number of species found continued to increase throughout the >55% of the species had been captured in the first 5 mo. The patterns 3 yr and tended to stabilize during the last 5 mo (Fig. 3). for the rate at which new species are found in the most conserved areas Species Richness Estimators. To determine which index shows (Phases 2 and 3) were similar to one another during the 3 yr. More than best results as an indicator of the likely species richness at a given envi- 70% of the species in those areas had been collected during the first ronment (Table 4), we tested several indexes in this study. year of the survey, contrasting with the Forest edge, where only 57% of The number of species collected was always lower than the number the species had been collected during the same period. In the Forest of species predicted by all indexes. This is true for each area considered individually, as well as for all areas combined, and for each year or for all years combined. The closest estimate for PEVV resulted from the Table 3. Number of species (S), abundance of individuals (N), the Michaelis–Menten index, which was only 3% higher than the number Simpson’s evenness index (SE), and the Berger and Parker’s even- of species collected at the end of the third year. Estimates by the Chao 1 ness index (BPE) of Syrphidae resulting from a survey using malaise and Chao 2 indexes were the highest, 83 and 91% higher than the num- traps in ﬁve areas of the PEVV, Ponta Grossa, Parana´ (September ber of species collected. With increased sampling effort, the values esti- 1999 to August 2002) mated with the Michaelis–Menten index differed less and less from the S N SE BPE number of species collected: 14% difference in the first year, 6% in the Forest edge 98 1,686 0.88 3.43 second, and 3% in the third. In contrast, the difference between the esti- Araucaria 50 315 0.93 4.57 mated and observed number of species became increasingly larger each Phase 1 63 373 0.94 5.74 year when the remaining indexes were used. For instance, the Chao 2 Phase 2 40 216 0.93 6.17 estimate was 73% higher than the number of collected species in the Phase 3 40 236 0.94 7.38 first year, 72% in the second, and 91% in the third.

160 150 140 130 120 110 100 90 80 70 60

Number of species of Number 50 40 30 20 10 0 sep/99 nov/99 jan/00 mar/00 may/00 jul/00 sep/00 nov/00 jan/00 mar/01 may/01 jul/01 sep/01 nov/01 jan/02 mar/02 may/02 jul/02

Forest edge Araucaria Phase 1 Phase 2 Phase 3 Total Fig. 3. Cumulative numbers of species of Syrphidae captured using malaise traps in all ﬁve areas surveyed in the PEVV, Ponta Grossa, Parana´ (September 1999 to August 2002).

Table 4. Observed and estimated species richness of Syrphidae based on the estimators Chao 1, Chao 2, Jackknife 1, Jackknife 2, Bootstrap, and Michaelis–Menten, separately for each area and for the entire dataset in PEVV, based on weekly data obtained using malaise traps in the PEVV, Ponta Grossa, Parana´ (September 1999 to August 2002) Sobs. Chao 1 Chao 2 Jackknife 1 Jackknife 2 Bootstrap Michaelis–Menten 1999–2000 Forest edge 55 97 6 24 108 6 31 81 6 6996 3666 165 Araucaria 36 69 6 26 64 6 21 51 6 4616 1426 153 Phase 1 50 98 6 29 95 6 25 75 6 5926 3606 169 Phase 2 30 42 6 9476 13 43 6 3516 1366 143 Phase 3 31 43 6 9436 9436 4496 2366 142 PEVV 101 168 6 30 175 6 33 141 6 7 169 6 3 118 6 1 115 1999–2001 Forest edge 80 124 6 20 124 6 20 115 6 6 135 6 3956 189 Araucaria 47 77 6 19 76 6 18 67 6 4806 2566 062 Phase 1 56 128 6 47 97 6 24 80 6 5976 2666 172 Phase 2 34 43 6 7466 9476 3536 1406 044 Phase 3 34 44 6 8446 8466 4516 2406 144 PEVV 126 208 6 32 217 6 35 179 6 8 215 6 2 148 6 1 134 1999–2002 Forest edge 98 203 6 45 203 6 45 146 6 7 182 6 2 118 6 1 105 Araucaria 50 91 6 29 91 6 29 68 6 4826 1586 161 Phase 1 63 119 6 31 102 6 20 91 6 6 109 6 2756 178 Phase 2 40 56 6 11 64 6 16 57 6 4686 1476 049 Phase 3 40 48 6 6486 6536 4556 1466 050 PEVV 151 276 6 45 288 6 49 218 6 9 267 6 2 179 6 1 156 Sobs. = observed number of species. 2014 SOUZA ET AL.: DIVERSITY OF SYRPHIDAE IN DIFFERENT HABITATS 5

Exclusive and Shared Species. From all identified species, 12% different from the forested areas, reflecting their unique floristic and were common to all areas studied, whereas 62% were exclusive to only environmental characteristics. The strongest structural similarity one of the five areas. Among the exclusive species, only six were repre- occurred between the forested areas where the canopy opening had sented by >10 specimens. The Forest edge had the largest number of more gaps (Araucaria and Phase 1), and the most conserved areas, with exclusive species, which represented more than half of the local rich- high humidity, formed a third group, not very cohesive. When the ness. In contrast, the most conserved forest areas (Phases 2 and 3) had Syrphidae subfamilies were ordered separately, a different arrangement the smallest number (<20%). of the areas was observed for the species of Eristalinae and The number of species shared between areas ranged between 23 and Microdontinae. In the case of Eristalinae, the forested areas nearly over- 37 but did not exceed half of the number of species found in each of the lapped on the diagram, revealing high similarity. Conversely, when areas (Table 5). The highest percentage share of species was between Microdontinae were considered not only the Forest edge but also the Phases 2 and 3 (48%)—the most conserved and geographically distant best-conserved area (Phase 3) showed a quite distinct community struc- from each other—and the lowest was between these areas and the ture when compared with the other areas. Forest edge (20%). Considering that light availability and vegetation Indicator Species. From the 151 Syrphidae species identified, 19 characteristics of Phase 1 and Araucaria areas are more similar to those were considered environmental indicators and were collected exclu- in the Forest edge, Phase 1 and Araucaria were the areas that shared sively or mainly in a particular habitat type. Twelve of them were asso- more species with the Forest edge (Table 5). ciated with the open vegetation, and seven with the forest vegetation Community Structure. The ordination analysis based on (Table 6). The other species sampled were not good indicators of any Syrphidae species occurrence and their relative abundance in the differ- particular habitat. ent sample areas (Fig. 4) indicated that the Forest edge is strongly Discussion The outcome of our 3-yr survey of the insects of the PEVV corrobo- rates the interpretations based on the data of the first year provided by Jorge et al. (2007). Our extended study, however, has not only provided Table 5. Total number and percentage of Syrphidae species shared more consistent data but has also advanced the knowledge on the family between pairs of areas, resulting from malaise trap sampling, in Syrphidae. The highest species richness and abundance were found in ﬁve areas of the PEVV, Ponta Grossa, Parana´, between September 1999 and August 2002 the subfamily Syrphinae, followed by Microdontinae and Eristalinae. The richness in hoverfly species and abundance of individuals were Forest edge Araucaria Phase 1 Phase 2 Phase 3 associated with the floristic characteristics and the level of conservation Forest edge 98 — — — — of the different areas. The more open, initial, or disturbed areas had Araucaria 33 (29%) 50 — — — more species and more individuals. In addition, the species distribution Phase 1 37 (30%) 34 (43%) 63 — — within the different habitat types in Araucaria Forest was heterogene- Phase 2 23 (20%) 27 (43%) 30 (41%) 40 — Phase 3 23 (20%) 26 (41%) 28 (37%) 26 (48%) 40 ous, and the community structure in the Forest edge was clearly distinct when compared with the other areas sampled.

Fig. 4. NMDS ordination of abundance data of hoverflies collected with malaise traps in five areas sampled in the PEVV, Ponta Grossa, Parana´, between September 1999 and August 2002. The first graph is based on all species of Syrphidae (NMDS stress < 0.001), the second refers only to species of Eristalinae subfamily (NMDS stress ¼ 0.008), the third to Microdontinae (NMDS stress ¼ 0.003), and the last to Syrphinae (NMDS stress < 0.001). 6 JOURNAL OF INSECT SCIENCE VOLUME 14

Table 6. Specific indicator values (IndVal) of syrphids captured using malaise traps in the PEVV, Ponta Grossa, Parana´ (September 1999 to August 2002) that had statistical significance for groups of habitats in the Araucaria Forest Species Subfamily Condition IndVal P Allograpta neotropica Curran, 1936 Syrphinae Cl 0.90 0.020 Allograpta similis Curran, 1925 Syrphinae Cl 1.00 0.010 Copestylum macquarti (Curran, 1926) Eristalinae Fo 0.83 0.040 Microdon aureopilis Marinoni, 2004 Microdontinae Fo 0.83 0.041 Microdon mourei Thompson, 2004 Microdontinae Ad 0.62 0.035 Mixogaster polistes Hull, 1954 Microdontinae In 0.63 0.027 Ocyptamus antiphates (Walker, 1849) Syrphinae Cl 1.00 0.020 Ocyptamus clarapex (Wiedemann, 1830) Syrphinae Cl 1.00 0.010 Ocyptamus funebris Macquart, 1834 Syrphinae Cl 0.68 0.020 Ocyptamus gastrostactus (Wiedemann, 1830) Syrphinae Cl 0.96 0.010 Paramicrodon flukei (Curran, 1936) Microdontinae In 0.63 0.036 Pseudodoros clavatus (Fabricius, 1794) Syrphinae Cl 0.97 0.010 Syrphus phaeostigma Wiedemann, 1830 Syrphinae Cl 0.63 0.040 Toxomerus basalis (Walker, 1836) Syrphinae Cl 0.69 0.040 Toxomerus croesus (Hull, 1940) Syrphinae Ad 0.58 0.027 T. croesus (Hull, 1940) Syrphinae Fo 0.93 0.001 Toxomerus pictus (Macquart, 1842) Syrphinae Cl 1.00 0.010 Toxomerus polygraphicus (Hull, 1940) Syrphinae Cl 0.99 0.010 Toxomerus politus (Say, 1823) Syrphinae In 0.67 0.040 Toxomerus watsoni (Curran, 1930) Syrphinae Cl 0.99 0.010 Cl, open vegetation; In, initial and intermediate forest succession areas; Ad, intermediate and advanced forest succession areas; Fo, forested environments.

Syrphidae larvae are more strongly affected by particular environ- that, in forested areas, syrphids are most abundant in the canopy, which mental conditions than are the adults (Ricarte et al. 2011). Given the is where flowers are more widely available. lack of efficient protocols for sampling the entire hoverfly larval com- The Araucaria area harbored a great number of species. When a for- munity (Ricarte et al. 2011) and the importance of the highly diverse est is planted, as in the case of the Araucaria area, the forest structure larval habits in the evaluation of the environmental conditions of a site has different strata and grows more rapidly than in natural circumstan- (Sommaggio 1999), it is necessary to consider the ecological require- ces (Dekoninck et al. 2005). In addition, despite the height of the can- ments of both adults and larvae when assessing hoverfly communities opy, the interior of the Araucaria Forest receives more sunlight than the collected as adults (Meyer et al. 2009). most conserved areas in our study, which had lower syrphid species The highest species richness and abundance of Syrphidae in the richness. Although the complexity of the vertical structure increases Forest edge and in the areas at the initial to intermediate stages of plant with successional age and potentially provides physical habitat for spe- succession can be a result of the higher productivity in these areas, cies, the lower abundance of herbs and shrubs provides fewer food which can be associated with a higher incidence of light and a greater sources for adults (Humphrey et al. 1999). availability of flowers—the preferred food source of Syrphinae and Despite not significant, the number of Eristalinae species was higher Eristalinae adults (Vockeroth and Thompson 1987). This was expected in more open and less humid areas (Phase 1 and Forest edge). On the based on the literature, given that open and sunny areas are generally other hand, the highest relative frequency of adults in the area at the considered suitable habitats for syrphids (Owen 1991, Gittings et al. intermediate to advanced stages of forest succession (Phase 2) could 2006). Moreover, grassy clearings may promote the entry of insects reflect the reproductive characteristics of hoverflies. The lower inci- from the surrounding areas (Taki et al. 2010), providing conditions for dence of light in the interior and the moister, leaf-litter-rich soils of that migrant, pioneer, and eurytopic species to establish temporary popula- area provide a suitable microenvironment for the development of larvae tions. Hence, the high species richness in such areas might also be the with saprophagous feeding habits. In these conditions, the adults would result of a large number of vagrant species. be caught as soon as they emerge from pupa. It has already been shown On the other hand, areas in the most advanced stage of conservation that, although hoverfly species richness is affected by resource hetero- had the lowest abundance and species richness. A general decrease in geneity, hoverfly density depends on resource accessibility, such as the these community descriptors with increasing forest structural retention availability of larval microhabitats (Meyer et al. 2009). had been previously reported (Deans et al. 2007). Similarly, a study on The relative frequency of Microdontinae was higher in each of the the effects of plant succession and the type of forest management on forested areas than in the Forest edge. However, species richness was Syrphidae communities in Belgium documented a decrease in the total significantly higher in the three most open areas. It is still not possible diversity in advanced forest succession stages (Dekoninck et al. 2005). to understand the reasons for these results because the biology of Because the Syrphinae were prevalent in the survey, the results Microdontinae is not well documented. All we know is that larvae of a obtained for Syrphidae richness and abundance largely reflect these few species are associated with ants (Duffield 1981, Elmes et al. 1999, parameters for that subfamily. The area in initial to intermediate stage Hironaga and Maruyama 2004). of forest succession (Phase 1) had elevated species richness and abun- The estimates generated by Bootstrap and Jackknife 1 richness esti- dance of Syrphinae, though lower than in the Forest edge, a pattern con- mators were closer to what was posited by Owen and Owen (1990), i.e., sistent with data obtained for hoverflies in general. Our results confirm if in 3 yr one can uncover 75% of the syrphid fauna present at a given an evaluation of different habitats in China, where the highest diversity location, one could expect that the surveyed forest, whose richness was of flower-visiting hoverflies occurred in the secondary forest habitats 151 species, might in fact harbor 200 species, which is an intermediate that were only slightly disturbed (Du et al. 2009). The richness of value between the Bootstrap (179) and Jackknife 1 (218) estimates. Syrphinae decreased with an increase in the level of plant succession, Granted that the number of new occurrences of species decreases with which results in flowers and short plants becoming scarce in the under- the increase in sampling effort, one can assume that the index of story. Despite the scarcity of studies on the subject, one can assume Michaelis–Menten is also an adequate estimator, as revealed by the 2014 SOUZA ET AL.: DIVERSITY OF SYRPHIDAE IN DIFFERENT HABITATS 7 decrease in the percent difference at each year of survey. On the other conserve mosaic landscapes for the development of different stages of hand, the indexes Chao 1, Chao 2, and Jackknife 2 yielded much higher the life cycle of hoverfly species and thus support highly diverse estimates than the actual number of species collected. Moreover, the communities. increase in sampling effort over the 3 yr resulted in a progressively larger estimate, assuming the existence of an infinite number of species Final Considerations in a finite environment. In the fauna of Alticini (Linzmeier et al. 2006), This study is the outcome of the longest survey of the syrphid the lowest estimates were also obtained by the Bootstrap and fauna in Brazil using malaise traps. In the case of taxa with highly Michaelis–Menten methods, whereas the highest estimates in each area diverse life histories as Syrphidae species, only intensive and extensive and in the entire dataset were obtained using Jackknife 2. studies can adequately survey its richness and allow the understanding The percentage of shared species in this study was higher than that of the role of them in nature. The species accumulation curve, with an found for Coleoptera (4%) and Alticini (7%) (Marinoni and Ganho asymptote only on the third month of the third year, demonstrates the 2006, Linzmeier and Ribeiro-Costa 2009). In fact, some taxa showed importance of long periods of collecting for robust surveys of syrphid distinct values of differential diversity, even though the environments faunas. compared were the same (Lawton et al. 1998). The highest proportion Our data on the richness and abundance of Syrphidae corroborated of common Syrphidae species is probably due to their remarkable the hypothesis that more well-preserved forests tend to maintain a ability to fly (Gilbert 1983) and to their diverse eating habits during smaller number of species with more equitable abundances, when com- their life cycle, which leads to greater mobility among the areas pared with areas that are impacted or are in the early stages of plant surveyed. succession. Our results support the idea that syrphids are appropriate ecological In most syrphid species with known biology, adults are flower visi- bioindicators (Sommaggio 1999). The values of specific indicators tors and play an important role as pollinators. The larvae have feeding revealed that the 12 indicator species of the Forest edge can be consid- habits that usually characterize them as predators, saprophagous or phy- ered primarily species of grassy clearings or disturbed areas. All of tophagous. As a consequence of the different behaviors displayed by them belong to the subfamily Syrphinae. Related species whose biol- adults and larvae, a given species might, throughout its life cycle, ogy is known (Rojo et al. 2003) are generally aphid predators, and we occupy environments that are completely at odds with one another— think that it is safe to assume that our species have similar habits. The for instance, open areas with high incidence of sunlight and flower seven indicator species of forested habitats can be considered primarily availability versus shaded areas with low incidence of sunlight and rich forest species, three of which occurred exclusively, or mainly, in the in decomposing matter. Therefore, the interpretation of the results based initial and disturbed areas, and two occurred in the advanced and most on the capture of adults using malaise traps is complex, leading to conserved areas. Four of those species belong to the subfamily different hypotheses that need to be tested by further biological and Microdontinae, three to Syrphinae, and one to Eristalinae. As they are ecological studies on the Neotropical Syrphidae fauna. It would be allocated to different groups and have different larval morphologies, necessary also to identify environmental indicator species in different their larval feeding habits are varied, ranging from predation to sap- vegetation types. rophagy (Vockeroth and Thompson 1987). However, their ecological requirements appear to be present especially in forests. A similar analy- Acknowledgments sis performed in Belgium also indicated that the major component This article is a part of J.M.T.S. Master dissertation at the determining the distribution of Syrphidae species is the degree of vege- Programa de Po´s-Graduaçaõ em Cieˆncias Biolo´gicas/Entomologia of tation cover (Dekoninck et al. 2005). Natural or anthropogenic small Universidade Federal do Parana´, contribution number 1795 of the disturbances, which create a gradient of canopy openness in Araucaria Zoology Department. We are grateful to Coordenaçaõ de Forest and increase their environmental heterogeneity, may increase the Aperfeiçoamento de Pessoal de Nı´vel Superior (CAPES) and to entomological diversity (Marinoni and Ganho 2006, Linzmeier and Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico Ribeiro-Costa 2009). Although several studies have demonstrated that hoverflies are (CNPq) for providing scholarships for the authors. We thank F. C. negatively affected by a reduction in landscape diversity, few studies Thompson for helping in the identification of Syrphidae; J. M. Silva have used this taxonomic group to evaluate environmental health for assistance in the field; C. M. Jorge and D. K. Rezende for aid with (Sommaggio 1999). This may be due to the lack of data on the habits and sample sorting; A. A. Padial for statistical discussions; and L. habitats of Brachycera fly species, especially in the Neotropical region, Iannuzzi and N. G. Ganho for the critical reading of the manuscript where the lack of this kind of information hinders ecological research on and excellent suggestions. Coordenaçaõ de Aperfeiçoamento de the group (Beaulieu and Wheeler 2005). Although it is risky to establish Pessoal de Nı´vel Superior and Conselho Nacional de habitat associations based only on five areas, the general patterns found Desenvolvimento Cientı´fico e Tecnolo´gico. in this study have allowed us to make a hypothesis for the distribution of the syrphid fauna of Araucaria Forest. Only more knowledge on the References Cited biology of the species may confirm such patterns. Beaulieu, F., and T. A. Wheeler. 2005. Diptera diversity in a homogeneous Considering that the habits of foraging, mating, oviposition, and habitat: Brachycera associated with sedge meadows (Cyperaceae: Carex) in diapause in Syrphidae may require habitats quite different from one Quebec, Canada. Proc. Entomol. Soc. Wash. 107: 176–189. another, environmental heterogeneity is fundamental (Meyer et al. Borges, Z. M., and M. S. Couri. 2009. Revision of Toxomerus Macquart, 2009). Additionally, there are hoverfly species that are more closely 1855 (Diptera: Syrphidae) from Brazil with synonymic notes, identifica- associated with specific vegetation types. Many of them are associated tion key to the species and description of three new species. Zootaxa 2179: 1–72. with forest environments (Branquart and Hemptinne 2000) with a high Borges, Z. M., and D. M. Pamplona. 2003. Revision of the Neotropical diversity and abundance of microhabitats, as well as refuges in adverse Xanthandrus Verral (Diptera, Syrphidae). Revista Brasileira de Entomologia conditions (Ricarte et al. 2011). Other species colonize open and 47: 155–167. anthropogenic habitats (Branquart and Hemptinne 2000), of which Branquart, E., and J. L. Hemptinne. 2000. Selectivity in the exploitation of grassy clearings provide additional resources, increasing Syrphidae floral resources by hoverflies (Diptera: Syrphinae). Ecography 23: 732–742. diversity (Ricarte et al. 2011), as in the disturbed areas of this study. Burgio, G., and D. Sommaggio. 2007. Syrphids as landscape bioindicators in Italian agroecosystems. Agric. Ecosyst. Environ. 120: 416–422. Taking into consideration the availability of exclusive species in both Colwell, R. K. 2006. EstimateS: statistical estimation of species richness and forested and open areas, the results of this study confirm those of shared species from samples, version 8.0.0. User’s guide and application. Meyer et al. (2009) and Ricarte et al. (2011), highlighting the need to (http://viceroy.eeb.uconn.edu/estimates). 8 JOURNAL OF INSECT SCIENCE VOLUME 14

Deans, A. M., S. M. Smith, J. R. Malcolm, W. J. Crins, and M. I. Bellocq. Marinoni, R. C., and N. G. Ganho. 2006. A diversidade diferencial beta de 2007. Hoverfly (Syrphidae) communities respond to varying structural reten- Coleoptera (Insecta) em uma paisagem antropizada do Bioma Arauca´ria. tion after harvesting in Canadian Peatland Black Spruce Forests. Environ. Revista Brasileira de Entomologia 50: 64–71. Entomol. 36: 308–318. Meyer, B., F. Jauker, and I. Steffan-Dewenter. 2009. Contrasting resource-de- Dekoninck, W., P. Pals, and P. Grootaert. 2005. Hoverfly communities on for- pendent responses of hoverfly richness and density to landscape structure. mer agricultural fields: a study of afforestation and planted forests stands in Basic Appl. Ecol. 10: 178–186. the Voeren region. Bulletin de L’Institut Royal des Sciences Naturelles de Owen, D. F., and J. Owen. 1990. Assessing insect species-richness at a single Belgique - Entomologie 75: 281–290. site. Environ. Conserv. 17: 362–364. Du, X. J., B. Z. Ren, Y. G. Wu, and L. W. Song. 2009. Differences of flower- Owen, J. 1991. Hoverflies, pp. 148-193. In J. Owen (ed.), The ecology of a gar- visiting hoverfly (Diptera: Syrphidae) communities in habitats with various den: the first fifteen years. Cambridge University Press, Cambridge. degrees of disturbance and altitude in Changbai Mountain, N. E. China. Acta Palmer, M. W. 1990. The estimation of species richness by extrapolation. Entomologica Sinica 52: 551–560. Ecology 71: 1195–1198. Duffield, R. M. 1981. Biology of Microdon fuscipennis (Diptera: Syrphidae) R Development Core Team. 2011. R: a language and environment for statisti- with interpretations of the reproductive strategies of Microdon species found cal computing. The R Foundation for Statistical Computing, Vienna. North of Mexico. Proc. Entomol. Soc. Wash. 83: 716–724. Ricarte, A., M. A. Marcos-Garcıá, and C. E. Moreno. 2011. Assessing the Dufreˆne, M., and P. Legendre. 1997. Species assemblages and indicator species: effects of vegetation type on hoverfly (Diptera: Syrphidae) diversity in a the need for a flexible asymmetrical approach. Ecol. Monogr. 67: 345–366. Mediterranean landscape: implications for conservation. J. Insect Conser. 15: Elmes, G. W., B. Barr, J. A. Thomas, and R. T. Clarke. 1999. Extreme host 865–877. specificity by Microdon mutabilitis (Diptera: Syrphidae), a social parasite of Roberts, D. W. 2010. LabDSV: ordination and multivariate analysis for ants. Proc. R. Soc. London B 266: 447–453. ecology. R package version 1.4-1. (http://cran.R-project.org/ Fraser, S.E.M., C. Dytham, and P. J. Mayhew. 2008. The effectiveness and package¼labdsv). optimal use of Malaise traps for monitoring parasitoid wasps. Insect Conserv. Rojo, S., F. S. Gilbert, M. A. Marcos-Garcıá, J. M. Nieto, and M. P. Mier. Divers. 1: 22–31. 2003. A world review of predatory hoverflies (Diptera, Syrphidae: Ganho, N. G., and R. C. Marinoni. 2003. Fauna de Coleoptera no Parque Syrphinae) and their prey. CIBIO Ediciones, Alicante. Estadual de Vila Velha, Ponta Grossa, Parana´, Brasil. Abundaˆncia e riqueza Smith, G. F., T. Gittings, M. Wilson, L. French, A. Oxbrough, S. das famı´lias capturadas atrave´s de armadilhas Malaise. Revista Brasileira de O’Donoghue, J. O’Halloran, D. L. Kelly, F.J.G. Mitchell, T. Kelly, et al. Zoologia 20: 727–736. 2008. Identifying practical indicators of biodiversity for stand-level manage- Gilbert, F. S. 1983. Hoverflies. The Richmond Publishing, Slough. ment of plantation forests. Biodivers. Conserv. 17: 991–1015. Gittings, T., J. O’Halloran, T. Kelly, and P. S. Giller. 2006. The contribution Sommaggio D. 1999. Syrphidae: can they be used as environmental bioindica- of open spaces to the maintenance of hoverfly (Diptera, Syrphidae) biodiver- tors?. Agric. Ecosyst. Environ. 74: 343–356. sity in Irish plantation forests. Forest Ecol. Manage. 237: 290–300. StatSoft. 2005. STATISTICA (data analysis software system), version 7.1 Hironaga, T., and M. Maruyama. 2004. The myrmecophilous hoverfly genus (www.statsoft.com). Microdon (Diptera, Syrphidae, Microdontinae) in Hokkaidoˆ, Japan, with de- Taki, H., T. Inoue, H. Tanaka, H. Makihara, M. Sueyoshi, M. Isono, and K. scriptions of four new species. Bull. Natl. Sci. Museum Tokyo A 30: 87–103. Okabe. 2010. Responses of community structure, diversity, and abundance Humphrey, J. W., C. Hawes, A. J. Peace, R. Ferris-Kaan, and M. R. Jukes. of understory plants and insect assemblages to thinning in plantations. Forest 1999. Relationships between insect diversity and habitat characteristics in Ecol. Manage. 259: 607–613. plantation forests. Forest Ecol. Manage. 113: 11–21. Thompson,F.C.1999.A key to the genera of the flower flies (Diptera: Jorge, C. M., L. Marinoni, and R. C. Marinoni. 2007. Diversidade de Syrphidae) of the Neotropical region including descriptions of new genera Syrphidae (Diptera) em cinco a´reas com situaço˜es florı´sticas distintas no and species and a glossary of taxonomic terms. Contrib. Entomol. Int. 3: Parque Estadual Vila Velha em Ponta Grossa, Parana´. Iheringia, Se´rie 321–378. Zoologia 97: 452–460. Townes H. 1972. A light-weight Malaise trap. Entomol. News 83: 239–247. Kim, K. C. 1993. Biodiversity, conservation and inventory: why insects matter. Veloso, H. P., and L. Goés-Filho. 1982. Fitogeografia brasileira. classificaçaõ Biodivers. Conserv. 2: 191–214. fisionoˆmico-ecolo´gica da vegetaçaõ neotropical (Boletim Tećnico - Se´rie Lawton, J. H., D. E. Bignell, B. Bolton, G. F. Bloemers, P. Eggleton, P. M. Vegetaçaõ). Projeto RADAMBRASIL, Salvador. Hammond, M. Hodda, R. D. Holt, D. S. Srivastava, and A. D. Watt. Venables, W. N., and B. D. Ripley. 2002. Modern applied statistics with S, 1998. Biodiversity inventories, indicator taxa and effects of habitat modifica- 4th ed. Springer, New York. tion in tropical forest. Nature 391: 72–76. Vockeroth, J. R., and F. C. Thompson. 1987. Syrphidae, pp. 713-743. In J. F. Linzmeier, A. M., and C. S. Ribeiro-Costa. 2009. Spatio-temporal dynamics McAlpine, B. V. Peterson, G. E. Shewell, H. J. Teskey, J. R. Vockeroth, of Alticini (Coleoptera, Chrysomelidae) in a fragment of Araucaria Forest in and D. M. Wood (eds.), Manual of Nearctic Diptera, vol. 2. Agriculture the state of Parana, Brazil. Revista Brasileira de Entomologia 53: 294–299. Canada, Research Branch, Monograph 28, Otawa. Linzmeier, A. M., C. S. Ribeiro-Costa, and R. C. Marinoni. 2006. Fauna de Alticini (Newman) (Coleoptera, Chrysomelidae, Galerucinae) em diferentes esta´- Received 11 June 2013; accepted 26 February 2014. gios sucessionais na Floresta com Arauca´ria do Parana´, Brasil: diversidade e esti- mativa de riqueza de espećies. Revista Brasileira de Entomologia 50: 101–109. Maack R. 1981. Geografia Fı´sica do Estado do Parana´. Jose´ Olympio Editora, Rio de Janeiro. Appendix 1 Magurran, A. E. 2004. Measuring biological diversity. Blackwell, Oxford. Species of Syrphidae collected with malaise traps in five areas sampled Marinoni, L., M. N. Morales, and I. Spaler. 2007. Chave de identificaçaõ ilus- in the PEVV, Ponta Grossa, Parana´, between September 1999 and trada para os geˆneros de Syrphinae (Diptera, Syrphidae) de ocorreˆncia no sul August 2002. do Brasil. Biota Neotropica 7: 143–158.