FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2562525905> ?p ?o ?g. }

• W2562525905 endingPage "364" @default.
• W2562525905 startingPage "352" @default.
• W2562525905 abstract "We recorded ground-foraging ant species in forest and savannah habitats along a 52-km-long road planned for upgrade in the buffer zone of the Moukalaba-Doudou National Park in south-west Gabon. Sixty stations were established with three sampling points on each side of the future road and baited with peanut butter to record the presence of invasive Wasmannia auropunctata (Roger, 1863). We documented 46 ant species including one genus and eight species not previously reported in Gabon, but no evidence of the presence of W. auropunctata. We also found species known to have an opportunistic behaviour such as Cardiocondyla emeryi (Forel, 1881), Tetramorium simillimum (Smith, 1851) and Trichomyrmex destructor (Jerdon, 1851). Species richness in forested stations was significantly higher than in savannah. Among the most common ant species in the area, we identified 13 associated with forests, eight associated with savannahs and one generalist. Four species were highly tolerant to human disturbance. Our study, even if biased towards stress-tolerant species, provides new insights about ant species associations with habitats and contributes to the establishment of a reference system to classify African ant species that could be used to monitor the success of restoration of areas impacted by human activities. Nous avons étudié les espèces de fourmis dusol deshabitats de forêts et de savane le long d'une route de 52 km qu'il est prévu d'améliorer dans la zone-tampon du Parc National de Moukalaba Doudou, dans le sud-ouest du Gabon. Nous avons établi 60 stations avec trois points d’échantillonnage de chaque côté de la future route quenous avons appâtés avec du beurre de cacahuète pour détecter la présence de l'espèce invasive Wasmannia auropunctata (Roger, 1863). Nous avons documenté la présence de 46 espèce de fourmis, y compris un genre et huit espèces encore jamais rapportés au Gabon, mais aucune preuve de la présence de W. auropunctata. Nous avons aussi trouvé des espèces connues pour avoir un comportement opportuniste comme Cardiocondyla emeryi (Forel, 1881), Tetramorium simillimum (Smith, 1851), et Trichomyrmex destructor (Jerdon, 1851). La richesse spécifique des stations forestières était significativement plus élevée qu'en savane. Parmi les espèces de fourmis les plus communes dans la région, nous en avons identifié 13 qui sont associées aux forêts, huit qui sont associées aux savanes et une généraliste. Quatre espèces étaient très tolérantes aux perturbations humaines. Notre étude, même si elle est biaisée en faveur des espèces tolérantes au stress, donne un nouvel aperçu des associations d'espèces de fourmis avec leurs habitats et contribue à la création d'un système de référence pour classifier les espèces de fourmis africaines qui peut servir pour vérifier le succès de la restauration de zones impactées par les activités humaines. Soil and litter arthropods are one of the most prominent components of ecological communities both in terms of biomass and diversity. Among them, ants (Hymenoptera, Formicidae) interact with or mediate many ecological processes, including the decomposition of organic matter, soil turnover and structure, nutrient cycling, plant protection, seed dispersal, and seed predation (Hölldobler & Wilson, 1990; Folgarait, 1998; Nakamura et al., 2006). Ants are sensitive to anthropogenic disturbances including logging, grazing, fires, mining, land conversion and fragmentation (Underwood & Fisher, 2006) and should be monitored closely in all development projects. In southwest Gabon, about 9,000 people living in the town of Gamba have been expecting for years to connect their isolated town to the national road system. In 2012, the government approved the paving of a 52-km road track and its construction started in 2014, with an opening date in 2015. As the new road cuts through the buffer zone of the Moukalaba-Doudou National Park, national and international scientists conducted a series of studies to document the distribution and abundance of several plant and animal species, including an assessment of the invasive ant Wasmannia auropunctata (Roger, 1863), a species known to have major negative effects where it is introduced (Agosti et al., 2000; Wetterer & Porter, 2003). This species has been documented in central (Walker, 2006; Mikissa et al., 2013) and southwest Gabon (Mikheyev et al., 2008; Mikissa et al., 2016) and has been linked to reducing arthropod diversity in sensitive habitats such as the Galápagos, and in continental mainland in Gabon (Walker, 2006; Ndoutoume-Ndong & Mikissa, 2007; Mikissa et al., 2008). W. auropunctata is documented in Gamba and Mayonami, both situated west of the new road (Dallmeier et al., 2006; Mikissa et al., 2016), and in Tchibanga (Marechal et al., 2014) and Mayumba (Walsh et al., 2004) both to the east of the new road. Typical spread of W. auropunctata occurs by transportation of earth material for human activities such as the construction of linear infrastructure like roads or railway (Walsh et al., 2004). Once established, the colonies expand locally, at a rate that can be augmented by the presence of waterways (Walker, 2006). The paving of the road in the seasonally flooded plains south of Moukalaba-Doudou National Park therefore represents a high risk of introduction or spread of W. auropunctata. In addition to assessing the presence of W. auropunctata, the ant data set was used to establish a baseline of the ant species composition before the road construction. This baseline was important as a reference to monitor future restoration success of areas impacted by the road building such as camp grounds, quarries and storage areas. Ants are good indicators of anthropogenic disturbance and habitat type. They are an abundant terrestrial taxon, have relatively high species richness and include many specialists in high trophic levels (Majer, 1983; Underwood & Fisher, 2006). In addition, ants respond relatively quickly to environmental changes (Nakamura et al., 2006), their community structure is affected by changes in diversity and biomass of other groups such as plants (Andersen & Majer, 2004) or other invertebrates (Kalif et al., 2001), and sampling and identifying them is cost-effective compared with other taxa (Lawton et al., 1998; Andersen & Majer, 2004). Unfortunately, very little is known about central African ant communities (Underwood & Fisher, 2006; Folgarait, 1998; but see Watt, Stork & Bolton, 2002), so this assessment was an opportunity to look for associations of ant species with habitat types and human disturbance levels, similar to those established for ants of tropical Australia (Andersen, 1997; King, Andersen & Cutter, 1998; Hoffmann & Andersen, 2003). In this study, we report our findings related to the following research questions: Are W. auropunctata and other tramp species (cosmopolitan species that can affect ecosystems when introduced) present in the area before the road is built? What environmental variables determine ant species assemblages in the study area? What are the habitat preferences and tolerance to anthropogenic disturbance of the ant species of the area? Anthropogenic disturbance is defined here as a human-induced change in environmental conditions that affects the local ecosystem, and includes deforestation, pollution of soil and water, and construction of infrastructures such as roads or camps. Such changes can cause ecological stress for some ant species and represent an advantage for others; therefore, our use of the word disturbance does not imply a reduction in production or biomass of ants, like in other definitions (e.g. Grime, 2006). To identify ant species both easy to monitor and that could become indicators of specific habitats and disturbance levels, we focused our sampling on common, ground-foraging species at each site (Andersen et al., 2002). The study site (3°07'32” S, 10°27'57” E) is in a mosaic of grassland, secondary forests and wetlands of approximately 175 km² along the Atlantic coast of Gabon (Vanthomme et al., 2015). Annual rainfall averages 2000 mm, with a dry season lasting from May to September and a wet season from October to April (Thibault, Fisher & Goodman, 2004). The study site is the major access point by land to the town of Gamba (Alonso et al., 2014); and people have used an unpaved road going from Loubomo to Mougagara villages across the study site for decades. Travel in the wet season is challenging since many areas are flooded. As a result, estimated traffic ranges from less than two vehicles a day in the wet season to 12-14 in the dry season (Vanthomme et al., 2015). The study site is a 52-km-long band centred on the Loubomo–Mougagara road (Fig. 1). The Mouambi River and Nyanga swamps in the Moukalaba-Doudou National Park border most of the study site to the northeast, and the eastern end of the Loubomo–Mougagara road line terminates in the N6 national road. The study area is bordered to the south and west by the Atlantic Ocean, and the Boumé Boumé River divides it into two zones. The Ouanga Plains zone (northwest, 96 km²) is a littoral lowland savannah-forest mosaic flooded most of the year. It is a fauna reserve under the protection of the Gabonese Ministry of Waters and Forests and is entirely situated in the buffer zone of the park. Two permanent (Moulondo and Mougagara) and three temporary fishing camps are installed in the Ouanga Plains where fewer than 10 people live throughout the year. The Panga zone (south-east, 79 km²) is a littoral highland savannah-forest mosaic with very steep hills. The Panga zone is partly situated in the buffer zone of Moukalaba-Doudou National Park and includes the Panga village (~80 people). An old laterite road, partially degraded, connects Panga and the village of Loubomo. We used systematic sampling to determine the ant diversity present in the study area. We stratified the area crossed by the road by main habitat type including forest (road in a forested area or less than 10 m away from a forest edge; total length 15.3 km) or savannah (road in open savannah and more than 10 m away from a forest edge; total length 45.3 km). Using a Geographic Information System (ArcGIS 10.1), we determined 30 points in each stratum regularly spaced along the road (separated by 0.5 km in forest and by 1.4 km in savannah). Due to the unequal length of the road in the Panga and Ouanga zones, we had 43 stations in Ouanga, and 17 in Panga. We established the stations in May 2012 (early dry season) in homogeneous habitat (forest or savannah). Each station had three sampling points (triplet) on each side of the road (n = 6 points per station). Within a triplet, the sampling points were located at one, three and five metres away from the road or the forest edge. If a forest station was close to a forest edge, we placed the two triplets of sampling points on the forested roadside, separated by 10 m from each other. At each sampling point, we placed on the ground a white card (10 × 15 cm) and a kebab stake coated with peanut butter on approximately 10 cm of its length to attract W. auropunctata (Williams & Whelan, 1992; Wetterer & Porter, 2003; Causton, Sevilla & Porter, 2005) and other ants species. We sampled ants found on the white card at least 1 h after the station was baited (maximum 107 min after baiting) and preserved them in a labelled vial filled with ethanol. Sampled specimens were determined by FHG and Georg Fischer (Okinawa Institute of Science and Technology) using Bolton (1973, 1980, 1982, 1987, 1994, 2007), Brown (1976), Lapolla, Hawkes & Fisher (2011), Rigato (2002), Schmidt & Shattuck (2014) and Ward et al. (2014). In some cases (e.g. Pheidole), specimens were directly compared with type material located in museum collections. Voucher specimens will be deposited by FHG at the Iziko museum of South Africa in Cape Town, and specimens currently stored at the Smithsonian Conservation Biology Institute in Gamba will be deposited at the Ecole Nationale des Eaux et Forêts in Cap Estérias, Gabon, with JBM. We gathered the following metadata at each station: date, duration of the bait, type of soil (sand, limestone, sandstone or laterite), lower vegetation type (leaf litter, short grass, and tall grass and weeds), exposure (sun, shade), wetness of the soil (inundated, wet, dry), width of the road (m) and presence of human disturbances within 20 m (e.g. camps, garbage). We used EstimateS software (Colwell, 2013) to compute rarefaction and extrapolation species accumulation curves per number of sampling stations using the estimated number of species S (Colwell et al., 2012) with unconditional variance (Shen, Chao & Lin, 2003). We estimated ground-foraging ant species richness for the entire study site, for the two zones (Ouanga and Panga), and for each main habitat (forest and savannah), using the first-order Jackknife asymptotic estimator (Chazdon et al., 1998). We tested for differences in diversity between zones and land cover by checking that the 84% CIs for S were not overlapping at the number of sampling stations of the largest sample (Colwell et al., 2012). We computed the Chao et al. (2000) estimator of shared number of species between zones and main habitats. We assessed associations between ant species assemblages (species presence–absence variables), and environmental, confounding (potentially affecting ant detectability), and disturbance variables (hereafter ‘independent variables’) with a canonical correspondence analysis (CCA; Legendre & Legendre, 2012). The environmental variables considered were main habitat type (forest or savannah), lower vegetation type (leaf litter, short grass, and tall grass and weeds), exposure (shade or sun) and soil type (sand or other). In addition, we generated land cover maps by automatic classification of radar imagery of the study site (25 m resolution) and assessed the land cover type (dry savannah, inundated savannah, dry forest, young forest, partially inundated forest or inundated shrubs) at each station from them (Vanthomme et al., 2015). Confounding variables were Julian date, duration of the bait (in minutes) and humidity index (from 0 for all sampling points on dry soil to 2 for all sampling points on wet or inundated soil). The disturbance variable was an index ranging from 1 for sampling stations on a narrow dirt road with no trash, to 4 for sampling stations on a large dirt road with large amounts of trash and human-caused forest openness. The data set is provided in supporting information (S1). CCA is a nonlinear ordination technique designed for direct analysis of the relationships between multivariate ecological data sets such as our species assemblages and independent variables. Only species present in at least five per cent of the sites were considered because rare species typically have a minor influence on results of multivariate statistics and can be perceived as outliers in ordinations (Gauch, 1982). The independent variables entering the CCA were selected using a backward stepwise selection (Blanchet, Legendre & Borcard, 2008). We added the human disturbance score to the independent variables used in the final model to assess its effect on species assemblages. Significance tests for the general model relating assemblage structure to independent variables, for each independent variable, and axes were based on 10,000 permutation tests (Oksanen et al., 2013). We performed automated partitioning of species around medoids using all significant CCA axes with PAM algorithm (Reynolds et al., 1992). We determined the number of clusters (K) by running successive PAM partitionings with increasing Ks (3 to 20), and selecting the partitioning with highest average silhouette width (Rousseeuw, 1987; Kaufman & Rousseeuw, 2005). We used indicator species analysis (de Cáceres et al., 2012) to identify combinations of at most two ant species (‘species groups’) indicators of groups of stations (‘site groups’) of forest or savannah land covers and, within these land covers, high disturbance (disturbance variable ≥ 3) and humid habitats (i.e. inundated savannah, partially inundated forest or inundated shrubs). The association of species groups to the target site group was assessed with the generalized indicator value index (gIndVal) and its 95% CI obtained by a randomization procedure (10,000 permutations). The index is the product of specificity and fidelity, where the former is the probability that a station belongs to the target station group given that the species group was found, and the latter is the probability of finding the species group in stations belonging to the site group. The gIndVal index ranges from 0 to 1 and reaches the maximum when all occurrences of a species group are found in a single site group (high specificity) and when the species group occurs in all stations in the site group (high fidelity). Following de Cáceres et al. (2012), we identified ‘indicator groups’ as species groups with: (1) lower bound of 95% CI of specificity higher than 0.6 for an expected rate of false positive in future assignments lower than 40%; and (2) fidelity higher than 0.25 to discard indicator groups that are powerful but occur in less than 25% of the stations. In the results, we report first the species that contribute most to the total indicator value (TIV: sum of indicator values of all indicator groups) alone or in association with others. We computed CCA modelling, indicator analysis and statistical tests in R 2.13 (R Development Core Team, 2013) with statistical software packages Vegan (Oksanen et al., 2013), Cluster (Maechler et al., 2004) and Indicspecies (de Caceres & Legendre, 2009). R scripts are provided in supporting information (S2). We documented 46 morphospecies at the 60 stations studied; 63% (n = 29) were assigned a known species name and the rest were given morphospecies numbers until further analysis can verify the species identifications (Table 1). We did not find W. auropunctata in the studied area. Panga and Ouanga areas had different ant assemblages, with 15 species and morphospecies specific to Panga, 11 specific to Ouanga and 20 found in both zones. The Jackknife estimate for total number of species was 60.8 (SD = 4.4). Species richness in forested stations (Jackknife = 51.5) and Panga (Jackknife = 49.1) was significantly higher than in savannah (Jackknife = 26.8) and Ouanga (Jackknife = 38.8), respectively (Fig. 2). More species were shared between Ouanga and Panga zones (Chao = 25.4) than between forest and savannah stations (Chao = 17.9). The shape of the accumulation curve suggested sampling effort was sufficient in all strata except in Panga (Fig. 2). Because sampling effort appeared insufficient in Panga, we could not be confident in the observed associations between species assemblages and environmental variables in our canonical correlation and indicator analyses for this stratum. Therefore, further analyses were done only within the Ouanga, forest, and savannah strata. In Ouanga, the following 11 species were present in less than five per cent of sites and thus were not included in the canonical correspondence analysis: Bothroponera FHG01, Camponotus FHG03, Crematogaster FHG01, C. FHG02, Lepisiota FHG01, Nylanderia umbella (Lapolla, Hawkes & Fisher, 2011), Pheidole caffra senilifrons (Wheeler, 1922), P. GF04, Plagiolepis FHG01, Polyrhachis latispina (Emery, 1925), and Solenopsis punctaticeps (Mayr, 1865). The final model after stepwise selection included main habitat, lower vegetation, Julian date and humidity. These, plus human disturbance, were used as independent variables in the final model. The relationship between species assemblages and independent variables was significant overall (F = 2.63, P = 0.005). CCA axis one (F = 6.69, P = 0.005), two (F = 2.90, P = 0.005) and three (F = 1.90, P = 0.017) were significant and together explained 87.6% of total variation of the distribution of species among sites. Main habitat (F = 6.4, P < 0.001), humidity (F = 2.74, P < 0.001) and lower vegetation (F = 1.71, P = 0.043) were significantly associated with ant assemblage structure. The CCA showed that axis one was mainly discriminating main habitats and vegetation types (savannahs with grass with positive scores and forest with leaf litter with negative scores); axis two was mainly discriminating humidity and lower vegetation types (wet leaf litter having positive scores, and dry grass negative scores); and axis three was mainly discriminating Julian date and level of disturbance (sites sampled later in the study and with lower levels of disturbance with positive scores, and sites sampled earlier in the study and with higher level of disturbance with negative scores). Partitioning around medoids (PAM) supported four groups of species (Fig. 3). Two groups were associated with savannah land covers: group 1 was associated with dry savannah with grass (high positive scores on axis one and high negative scores on axis two) and included seven species: Brachyponera sennaarensis (Mayr, 1862), Cardiocondyla FHG01, Monomorium exiguum (Forel, 1894), Nylanderia boltoni (Lapolla, Hawkes & Fisher, 2011), Pheidole megacephala melancholica (Santschi, 1912), Tetramorium sericeiventre (Emery, 1877), and Trichomyrmex destructor (Jerdon, 1851); group 4 was associated with wet savannah (high positive scores on axis one and two, and negative scores on axis three) and included two species: Monomorium jacksoni (Bolton, 1987) and Pheidole GF05. The two other groups were associated with forest land covers: group 3 was associated with forest with leaf litter (high negative scores on axis one) and included nine species: Crematogaster FHG03, Monomorium captator (Santschi, 1936), M. inquietum (Santschi, 1926), Odontomachus troglodytes (Santschi, 1914), Pheidole GF06, P. GF07, Tapinoma lugubre (Santschi, 1917), Tetramorium setigerum (Mayr, 1901), and T. zambezium (Santschi, 1939); group 2 was associated with disturbed forest (negative scores on axis one and high negative scores on axe three) and included two species: Cardiocondyla emeryi (Forel, 1881) and Nylanderia lepida (Santschi, 1915). In Ouanga, 13 indicator groups were associated with forest habitat. Species indicator of forest included Pheidole GF07, which contributed 26% of TIV, alone (gIndVal = 0.61) and in association with Monomorium captator (gIndVal = 0.39), Odontomachus troglodytes (gIndVal = 0.39), Tapinoma lugubre (gIndVal = 0.28) and Tetramorium zambezium (gIndVal = 0.28); M. captator (25% of TIV), alone (gIndVal = 0.60) and in association with the previous, O. troglodytes (gIndVal = 0.33), Pheidole megacephala melancholica (gIndVal = 0.28) and T. zambezium (gIndVal = 0.28); O. troglodytes (17% of TIV), alone (gIndVal = 0.50) and in association with the previous; Tetramorium zambezium alone (15% of TIV and gIndVal = 0.44); Tapinoma lugubre (9% of TIV), alone (gIndVal = 0.28) and in association with previous; and Nylanderia lepida alone (6% of TIV and gIndVal = 0.28). We found only one indicator group of disturbed forest: Camponotus FHG03 associated with Myrmicaria FHG01 (gIndVal = 0.33). No valid indicator group of wet forest was found. In Ouanga, only Pheidole GF05 alone (gIndVal = 0.32) was a valid indicator group of savannah land cover. Due to the limited number of highly disturbed savannah stations at our study site, no indicator group of disturbed savannah was found. No valid indicator group of wet savannah was found. We summarized our findings for the 46 ant species reported in the study (Tables 1 and 2) and compared the associations found in the CCA and indicator analyses. We report in the tables each species’ preferred habitat (from the associations of its PAM group with CCA variables in Ouanga), and if it was found to be an indicator of land cover type or high level of disturbance in the indicator analysis. If the association and indicator status did not overlap, we reported the species as ‘generalist’. If for a particular species no association or valid indicator status could be detected and the species was rare (occurred in less than five per cent of stations) in one of the strata or in the entire study area, we reported it as ‘rare in our data set’ and supposed the occurrence of the species was too low to infer its associations with habitat types or disturbance levels. We documented ground-foraging ant community assemblages of forest, savannah and humid habitats along the buffer zone of the Moukalaba-Doudou National Park and showed that main habitat, humidity and lower vegetation are important variables for ant assemblages in southwest Gabon. The species–habitat associations found in our study mostly agree with those reported in the literature: species preferring forest include Nylanderia lepida (AntWeb, 2015); species preferring savannahs and open areas include Brachyponera sennaarensis, Trichomyrmex destructor (AntWeb, 2015), Tapinoma lugubre, Monomorium captator (Braet & Taylor, 2008) and Tetramorium sericeiventre (Braet & Taylor, 2008; Fischer, 2011); and generalist species include Pheidole megacephala melancholica (Braet & Taylor, 2008; Mikissa et al., 2008; Fischer, 2011; AntWeb, 2015). In addition, we report for the first time Monomorium inquietum's association with forested habitats. Some species had records in the literature of specimens collected from a range of habitats including savannah, forest and disturbed areas (Braet & Taylor, 2008; Mikissa et al., 2008; Fischer, 2011; AntWeb, 2015; AntWiki, 2015), but appear to have stronger habitat preferences at our study site. Odontomachus troglodytes, Tetramorium zambezium, T. setigerum and Cardiocondyla emeryi were associated with forested habitats, and Monomorium exiguum and Nylanderia boltoni were associated with savannah habitats at our site. Only for Monomorium jacksoni does the species–habitat association found in our study contradict the current records for the species. In previous studies in Cameroon and Gabon, Monomorium jacksoni was only found in forested habitat (AntWeb, 2015), whereas it was found exclusively in savannahs in our study. These unexpected associations may result from the intricate mix of habitats in our research area that could have helped species adapt and thrive in habitats that are not preferred in other sites, or may indicate the existence of cryptic species for these taxa. The Ouanga area is a mosaic of habitats growing on sandy and low nutrient soils and is seasonally flooded with brackish water. This challenging environment and the fact that we sampled along the road is probably why most ground-foraging species we recorded are known to be widespread (AntWeb, 2015; AntWiki, 2015) such as: Odontomachus troglodytes, found across sub-Saharan Africa; Brachyponera sennaarensis, Monomorium exiguum, Nylanderia boltoni, Oecophylla longinoda (Latreille, 1802), Pheidole aurivillii (Mayr, 1896), Polyrhachis latispina, Solenopsis punctaticeps, Technomyrmex andrei (Emery, 1899), and Tetramorium aculeatum (Mayr, 1866), which have been reported from the entire Afrotropical region; Tetramorium sericeiventre which was also found in the Malagasy (Madagascar) and Palearctic regions; Camponotus mayri (Forel, 1879), Pheidole njassae (Viehmeyer, 1914), Tapinoma lugubre, Tetramorium setigerum and T. zambezium, which have been reported in the eastern Afrotropical region; and Monomorium captator, M. inquietum, M. jacksoni, Nylanderia lepida, N. umbella, Paraparatrechina albipes (Emery, 1899), Pheidole caffra senilifrons, Tetramorium minimum (Bolton, 1976) and T. ataxium (Bolton, 1980), which have been reported from the western Afrotropical region. From our study, only Camponotus FHG03, Myrmicaria FHG01, Cardiocondyla emeryi and Nylanderia lepida can truly be considered stress tolerant, as they were associated with the most disturbed stations. This study provides baseline data about the common ground-foraging ant fauna before the road to Gamba is upgraded and paved. Of the 17 morphospecies of ants that we could not identify in our study, most belong to the genera Camponotus, Crematogaster and Pheidole, which are very species rich and in need of taxonomic revisions (Hita Garcia et al., 2009). To a lesser extent, the genera Lepisiota, Myrmicaria and Plagiolepis are less diverse and also need taxonomic revisions. Two of the three species of the genus Cardiocondyla, which was revised for the Afrotropical region (Rigato, 2002), seem to be undescribed. Based on Fisher (2004), Ndoutoume-Ndong & Mikissa (2007), Braet & Taylor (2008), Mikissa et al. (2008), Yanoviak, Fisher & Alonso (2008), Mikissa et al. (2013), AntWeb (2015) and AntWiki (2015), 383 species of ants are recorded in Gabon. To our knowledge, our study includes one genus (Trichomyrmex) and eight species that are new records for Gabon: Brachyponera sennaarensis, Monomorium inquietum, Nylanderia umbella, Pheidole aurivillii, P. njassae, Solenopsis punctaticeps, Tetramorium zambezium and Trichomyrmex destructor. In addition, Pheidole caffra was only reported in the Pongara National Park in Gabon (Braet & Taylor, 2008), but the specimen is reported to look more like the type from South Africa than the P. caffra senilifrons subspecies described in Democratic Republic of Congo and found at our site. Finally, Monomorium jacksoni, Tetramorium ataxium and T. setigerum have been reported from Gabon only in the AntWeb (2015) online repository of specimens, but were never reported in previous literature. The choice of our sampling method intended to maximize the likelihood of documenting the presence of W. auropunctata and other tramp species, and to be able to compare savannah and forest habitats. Baiting is also an appropriate technique for sampling dominant species and monitoring ecosystem changes that could be induced by the road upgrade (Hölldobler & Wilson, 1990; Underwood & Fisher, 2006). Nevertheless, baiting biased our sample towards ground-foraging species attracted to peanut butter (Underwood & Fisher, 2006); therefore, the resulting assemblages of species should only be considered a fraction of the overall ant diversity in the area. With peanut butter baits on the ground, we also had very low probability of sampling subterranean species (e.g. Solenopsis punctaticeps), leaf litter species (e.g. Technomyrmex andrei, Tetramorium ataxium and T. simillimum [Smith, 1851]) and canopy species (e.g., Oecophylla longinoda, Polyrhachis latispina and Tetramorium aculeatum), even if they were abundant locally. This sampling bias is probably the main reason why 14 identified species were too rare in our data set to infer any association with habitat types or a high disturbance level. Although we did not find W. auropunctata, other species considered as tramp species worldwide were present in the study area, including Cardiocondyla emeryi, Tetramorium simillimum and Trichomyrmex destructor. All three species are opportunistic and typical of disturbed habitats, demonstrating that the area was already impacted in the past, when the traffic was still relatively low. The road building and likely increase in vehicular traffic that will result from the upgrade, in particular for trucks that were not able to access Gamba town by land on the unpaved road, therefore represent critical risks of introduction of W. auropunctata into the research area, which is situated in the buffer zone of Moukalaba-Doudou National Park. To reduce this risk, we recommend to the managers of the construction company to clean all vehicles and containers with high-pressure water before they enter the construction sites, and in particular, to clean vehicles stored in surrounding villages that are already infested with W. auropunctata. In addition, we recommend four ant assessments per year in the construction site, with particular emphasis on the camps and the newly constructed roadsides, to be able to detect and eradicate W. auropunctata immediately if it is introduced. Beyond identifying W. auropunctata in priority, the monitoring should provide information on the other opportunist species encountered in Ouanga (Cardiocondyla emeryi, Tetramorium simillimum and Trichomyrmex destructor), as they may colonize new areas with the extension of the road. The monitoring should reproduce the baiting technique and sampling done in this baseline study and consider pitfalls, leaf litter sampling with Winkler extraction sacks, and hand collection if time and resources allows, to have a better appreciation of the total ant diversity of the area (Underwood & Fisher, 2006). Monitoring should include preparation of specimens for identification and preservation of individuals for barcoding to help resolve current taxonomic uncertainties for the ants of Gabon (e.g. Ng'endo, Osiemo & Brandl, 2013). The monitoring of ant communities in response to the road upgrade in each main habitat in Ouanga could thus help develop a general classification system for African ants that could be used to monitor the restoration of areas impacted by human activities. We thank the Centre National de la Recherche Scientifique et Technologique, the Agence Nationale des Parcs Nationaux and the Direction de l'Aménagement des Aires Protégées of Gabon for authorizing our study (permits AR0016/12, AR0004/13, AE130010 and 071MEF). We also thank Georg Fischer from the Okinawa Institute of Science and Technology, Japan, for his identification efforts with the Pheidole species, Adrian Drewett for his support to conduct the study, and Shell Gabon and the Smithsonian Conservation Biology Institute for financial support. This is contribution 143 of the Gabon Biodiversity Program. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article." @default.
• W2562525905 created "2017-01-06" @default.
• W2562525905 creator A5035478065 @default.
• W2562525905 creator A5041044989 @default.
• W2562525905 creator A5045543682 @default.
• W2562525905 creator A5046323804 @default.
• W2562525905 creator A5064046933 @default.
• W2562525905 creator A5077140124 @default.
• W2562525905 creator A5089432210 @default.
• W2562525905 date "2016-12-29" @default.
• W2562525905 modified "2023-10-17" @default.
• W2562525905 title "Associations of ant species to land cover types and human disturbance in south‐west Gabon" @default.
• W2562525905 cites W1560374492 @default.
• W2562525905 cites W1627736016 @default.
• W2562525905 cites W1796398712 @default.
• W2562525905 cites W1871192594 @default.
• W2562525905 cites W1984455462 @default.
• W2562525905 cites W1987971958 @default.
• W2562525905 cites W1988619277 @default.
• W2562525905 cites W1993046026 @default.
• W2562525905 cites W2014750767 @default.
• W2562525905 cites W2016383111 @default.
• W2562525905 cites W2028596084 @default.
• W2562525905 cites W2033156942 @default.
• W2562525905 cites W2034380476 @default.
• W2562525905 cites W2037469231 @default.
• W2562525905 cites W2054130559 @default.
• W2562525905 cites W2062631198 @default.
• W2562525905 cites W2069797419 @default.
• W2562525905 cites W2075949648 @default.
• W2562525905 cites W2077501362 @default.
• W2562525905 cites W2109715154 @default.
• W2562525905 cites W2114451958 @default.
• W2562525905 cites W2116601594 @default.
• W2562525905 cites W2116979541 @default.
• W2562525905 cites W2122051300 @default.
• W2562525905 cites W2131387099 @default.
• W2562525905 cites W2135804567 @default.
• W2562525905 cites W2141663179 @default.
• W2562525905 cites W2144163278 @default.
• W2562525905 cites W2156470345 @default.
• W2562525905 cites W2165619450 @default.
• W2562525905 cites W2170664409 @default.
• W2562525905 cites W2170854726 @default.
• W2562525905 cites W2177087227 @default.
• W2562525905 cites W2178964686 @default.
• W2562525905 cites W2297901388 @default.
• W2562525905 cites W2397180013 @default.
• W2562525905 cites W2751607293 @default.
• W2562525905 cites W4205785327 @default.
• W2562525905 cites W4251859964 @default.
• W2562525905 doi "https://doi.org/10.1111/aje.12362" @default.
• W2562525905 hasPublicationYear "2016" @default.
• W2562525905 type Work @default.
• W2562525905 sameAs 2562525905 @default.
• W2562525905 citedByCount "5" @default.
• W2562525905 countsByYear W25625259052018 @default.
• W2562525905 countsByYear W25625259052021 @default.
• W2562525905 countsByYear W25625259052022 @default.
• W2562525905 countsByYear W25625259052023 @default.
• W2562525905 crossrefType "journal-article" @default.
• W2562525905 hasAuthorship W2562525905A5035478065 @default.
• W2562525905 hasAuthorship W2562525905A5041044989 @default.
• W2562525905 hasAuthorship W2562525905A5045543682 @default.
• W2562525905 hasAuthorship W2562525905A5046323804 @default.
• W2562525905 hasAuthorship W2562525905A5064046933 @default.
• W2562525905 hasAuthorship W2562525905A5077140124 @default.
• W2562525905 hasAuthorship W2562525905A5089432210 @default.
• W2562525905 hasBestOaLocation W25625259051 @default.
• W2562525905 hasConcept C127413603 @default.
• W2562525905 hasConcept C151730666 @default.
• W2562525905 hasConcept C18903297 @default.
• W2562525905 hasConcept C205649164 @default.
• W2562525905 hasConcept C2777601987 @default.
• W2562525905 hasConcept C2780428219 @default.
• W2562525905 hasConcept C2780648208 @default.
• W2562525905 hasConcept C4792198 @default.
• W2562525905 hasConcept C78519656 @default.
• W2562525905 hasConcept C86803240 @default.
• W2562525905 hasConcept C97137747 @default.
• W2562525905 hasConcept C97467695 @default.
• W2562525905 hasConceptScore W2562525905C127413603 @default.
• W2562525905 hasConceptScore W2562525905C151730666 @default.
• W2562525905 hasConceptScore W2562525905C18903297 @default.
• W2562525905 hasConceptScore W2562525905C205649164 @default.
• W2562525905 hasConceptScore W2562525905C2777601987 @default.
• W2562525905 hasConceptScore W2562525905C2780428219 @default.
• W2562525905 hasConceptScore W2562525905C2780648208 @default.
• W2562525905 hasConceptScore W2562525905C4792198 @default.
• W2562525905 hasConceptScore W2562525905C78519656 @default.
• W2562525905 hasConceptScore W2562525905C86803240 @default.
• W2562525905 hasConceptScore W2562525905C97137747 @default.
• W2562525905 hasConceptScore W2562525905C97467695 @default.
• W2562525905 hasFunder F4320322892 @default.
• W2562525905 hasFunder F4320324026 @default.
• W2562525905 hasFunder F4320332805 @default.
• W2562525905 hasIssue "3" @default.
• W2562525905 hasLocation W25625259051 @default.

• next