The CNP (15°32′31″N, 88°15′49″W) encompasses approximately 23,440 ha, with an elevational range of 500–2242 m a.s.l. (ICF 2015) (Fig. 1). The park is divided between an inner core zone (7690 ha) where settlement and resource extraction are prohibited and an encircling buffer zone (15,750 ha) where some of these practices are allowed. A closed canopy forest dominates the core zone, with a diverse community of broadleaf evergreen tree species present in the mid- to upper elevational ranges (1300–1800 m a.s.l.), interspersed with pine forest occurring mostly on the drier, eastern slopes of the park. Secondary forest, at various levels of succession, is also present below 1300 m a.s.l. as a result of commercial logging during the mid-20^th century (Martin et al. 2021). At higher elevations (>1800 m a.s.l.), the upper montane rainforest, characterised by dense growth of mosses, liverworts, ferns, and a high abundance of bromeliads, is present as a result of cool temperatures and comparatively higher rainfall. At the uppermost peak elevations (>2000 m a.s.l.), a combination of soil erosion and a lack of decomposition results in stunted but densely interwoven vegetation known as elfin forest (Martin et al. 2021).

Figure 1. 

Map of Cusuco National Park. Buffer and core zone boundaries are shown (red and green, respectively) with camps (red dots) and corresponding transects with sampling subsites (black lines). Cuscuo elevation data are derived from Burdgis.org (accessed 16 September 2021). Continental relief map derived from SimpleMappr, Shorthouse, 2010.

Ant-species observations from two different projects were pooled into a single dataset. Respective projects and methodologies are described below. The sampling techniques used for each project are summarised in Table 1 and Fig. 2. Sampling was completed mainly in the core zone of the park, between the elevations of 1170 and 2030 m a.s.l.

Figure 2. 

Distribution of sampling effort across the elevation gradient of Cusuco National Park. For the leaf litter arthropods of Mesoamerica (LLAMA) and MyrmEcoDex (MED) collections. The size of the circle corresponds to the number of samples at a specific elevation. Leaf litter sampling includes Mini-Winkler, Maxi-Winkler, and Berlese traps.

1. Leaf-litter arthropods of Mesoamerica (LLAMA)

Project LLAMA, funded by the U.S. National Science Foundation, sampled leaf litter-dwelling arthropods across Mesoamerica, from southern Mexico to Nicaragua, with a focus on ants and weevils (Curculionidae) (Longino et al. 2014). Specimens were collected in CNP from 29 May to 3 June 2010, during the transition from the dry season to the short wet season. Most collections were made between 1210 and 1360 m a.s.l., but additional non-quantitative samples were collected between 1580 and 2030 m a.s.l. (Fig. 2). A few additional leaf litter samples were collected earlier by R. Anderson on 24 August 1994, and included in this dataset.

Some identifications could not be confirmed and are marked as cf. (Latin: confer) or by a summation in the species epithet. These specimens appear similar to the named species, but verification was not possible. Verification requires more specimens and comparison with morphologically similar species.

An assessment of biogeographic distribution status was made for all recorded species in this study using Antmaps (Janicki et al. 2016; Guénard et al. 2017). The following categories were applied: regionally restricted (to Honduras), exotic (to Honduras, i.e., non-native), and globally invasive species (showing wide global occurrence patterns). Species not previously reported in Honduras were also noted.

Species accumulation curves were made to provide insight into sample completion and method efficiency. By assessing species richness cumulatively per additional sample, we show the intensity of individual sampling techniques, respectively, and the potential for collecting additional species with additional sampling. Accumulated species richness was also compared with sampling coverage of the community, which is the probability that an individual of the entire ant community belongs to a species that has been sampled before. As sampling techniques each address a different subset of the total ant community, respective subset communities are considered. Species presence-absence matrices using unique sample codes as individual sampling units were built per collection methodology. Non-species-level identified specimens and OUMNH material were excluded as a result of low taxonomic resolution and a lack of collection codes, respectively. Final accumulation curves and summary statistics were generated with the iNEXT R package using the first Hill number (species richness) (Chao et al. 2014; Hsieh et al. 2022).

To improve the accessibility of this work and the Honduran ant fauna in general, a dichotomous identification key was constructed for ant genera for the whole of Honduras (Suppl. material 2). Genera occurring in Honduras were determined using records from Antweb (www.AntWeb.org). The identification key was constructed manually by combining multiple works and identification keys on relevant taxa (Ward 1985; Hölldobler and Wilson 1990; Shattuck 1992; Bolton 1995; Longino 2007; Wild 2007; Donoso 2012; Fayle et al. 2014; Schmidt and Shattuck 2014; Baccaro et al. 2015; Borowiec 2016; Ward and Fisher 2016; Ward et al. 2016; Williams and Lapolla 2016; Solomon et al. 2019; Prebus 2021b; Camacho et al. 2022; www.Antwiki.org), with respective citations provided (provided as Suppl. material 2).

Across all sampling projects, a total of 5690 ants were collected in CNP, resulting in nine subfamilies comprising 60 genera and 162 species (Table 2). Appendix 1 provides a complete list of all ant species found in CNP, with the respective sampling method, recorded elevational distribution ranges, and habitat (data also in Suppl. material 1). The complete specimen data is made available as well (Suppl. material 4). Characteristic specimens from our collections are shown in Fig. 3.

Figure 3. 

Select ant species collected in Cusuco National Park. Top: Pheidole absurda (A major worker), Leptogenys imperatrix (B worker); Bottom: Labidus coecus (C major worker), Odontomachus yucatecus (D wingless female).

In addition to the first 127 species collected by LLAMA (3445 specimens), the MED collections resulted in an additional 41 species for CNP. The latter yielded a total of 78 species and 2155 specimens: 894 in 2018 and 1261 in 2019. Of these, 286 and 419 were mounted, respectively, and added to the RBINS collections (together with the remaining specimens in ethanol). The checklist also includes 90 mounted specimens from the OUMNH, collected during earlier field surveys in CNP and identified by MED.

Some ant species in CNP have notable distributions (Table 3). Seven species are regionally restricted to Honduras, and two species are exotic. Three species have global distributions (including the two exotic species), and five species were recorded for the first time in Honduras.

We report the first collection of male Pheidole balatro individuals (Fig. 4). Males of this species were previously unknown. Six male individuals were collected alongside minor and major workers of this species from a nest residing inside a bromeliad plant (Base Camp, 26 June 2019, 1613 m a.s.l.). The nest was basally located in the bromeliad between the leaves and was discovered by removing leaves from their basal attachment. All six specimens of P. balatro males were stored in the RBINS collections (three mounted, three preserved in ethanol; sample code: CNP-222).

Figure 4. 

Pheidole balatro male specimen (Specimen code: CNP-222-3-3) A lateral and B frontal view.

Sampling efforts across the elevation range varied between LLAMA and MED (Fig. 2). There were also differences in the collection techniques used between the two datasets. LLAMA focused on a narrower elevation range between ~1210 and 1360 m a.s.l. (with the exception of three higher samples at ~1580 m and ~2030 m a.s.l.), using thorough sampling in a standardised framework. A large part of the ant diversity in CNP (127 species) was thus recorded in a narrower elevation range using four sampling techniques: leaf litter extraction, bait trapping, Malaise trapping, and vegetation beating. Project MED considered a broader range of elevation from ~1170–1960 m a.s.l., employing less exhaustive sampling. A different set of sampling techniques was employed (pitfall traps and more hand collections) at wider elevation ranges, including higher elevations. This resulted in the collection of a different subset of the ant fauna (78 species total; 41 newly reported). Additional species were even collected at the same narrow elevation range just by changing techniques (e.g., Odontomachus yucatecus).

Leaf litter extraction obtained the greatest proportion of genera (26%) and species (31%) compared to other methods, with hand collection also comprising a large proportion of genera (20%) and species (21%) (Table 4). Baiting, Malaise, leaf litter extraction, and vegetation beating were predominantly conducted at relatively lower elevational ranges (Fig. 2). Pitfall traps and hand collection events were distributed more evenly across the elevation range in contrast, with a concentration of hand collection events at 1600 m a. s. l. To reduce interpretation bias on the efficiency of sampling techniques, Table 4 complements Fig. 2, illustrating sampling effort.

Fig. 5 presents species accumulation curves, showing the captured species richness and the sampling intensity (observed and extrapolated). Additional sampling for each individual sampling technique is expected to result in the capture of more species, with a total asymptotic richness estimating 257 (95% confidence interval: 204–365) species when all sampling efforts are combined. The maximum rate of increase in species richness seems not to have been reached yet for vegetation beating and Malaise trapping. However, for leaf litter extraction, hand collection, pitfall trapping, and baiting, it appears that the maximum rate has been reached.

Figure 5. 

Species accumulation curves according to the number of samples taken. Shaded regions represent 95% confidence intervals A total species richness for the number of samples taken B species richness for the number of samples taken for each technique.

The relation between species richness and community sample coverage is presented in Fig. 6. As stated before, sample coverage concerns the probability that an individual of the entire ant community addressed per sampling technique belongs to a species that has been sampled before. High values show low probability (1 – sample coverage) of sampling additional species for each technique; low values show high probabilities to find species that are yet unaccounted for. The total sampling considered in this study shows high community sample coverage with a value of 0.978. There are, however, differences between individual sampling techniques. Leaf litter extraction, pitfall trapping, and baiting show a community sample coverage of >0.9 (0.977, 0.916, and 0.978, respectively), whereas this value is lower for hand collection, vegetation beating, and Malaise trapping (0.835, 0.815, and 0.44, respectively). The extrapolation shows the expected increase in species richness with greater sampling effort. The 95% confidence intervals are low for most techniques, except for Malaise trapping, which had a considerably lower effort compared with other sampling methods.

Figure 6. 

Species accumulation curves according to community sample coverage. Sample coverage is the probability for an individual in the entire ant community to belong to a species that has been sampled before. Shaded regions represent 95% confidence intervals A total species richness for the sample coverage B species richness for the sample coverage for each technique.

Our research confirms that CNP supports a high richness of ant species, on par with many other taxa in the park (Martin et al. 2021). We present the first checklist for the ants of CNP, which is also the first species list published for ants in Honduras in over a century. With the inclusion of a genus-level identification key for all Honduran ant taxa (Suppl. material 2), we hope this work will stimulate further ant research in Honduras.

Ant species composition showed patterns consistent with other ant faunal inventories (Yanoviak and Kaspari 2000; Patrick et al. 2012; Donoso 2017), with Myrmicinae, Formicinae,Ponerinae, and Dolichoderinae comprising over 80% of species collected (Table 2). Compared to other subfamilies, Dorylinae contributed a substantial number of genera (7), considering only 14 species were recorded.

Seven ant species recorded in CNP are spatially restricted to the region. According to current known occurrences, six species are regionally restricted to Honduras. Of these, four are restricted to CNP: Octostruma leptoceps, Stenamma cusuco, S. atribellum, and Temnothorax altinodus. Two other species restricted to Honduras belong to the genus Leptogenys, previously reported from La Ceiba (Leptogenys bifida) and Lombardia (Leptogenys honduriana). Pheidole cusuco was originally described by the CNP and has also been reported just across the Guatemalan border (Longino 2019). This indicates the importance of CNP as a potential refuge for these spatially restricted species, which has implications for their conservation. Conversely, two non-native ant species were recorded: Monomorium pharaonis (at the edge of the core zone) and Solenopsis invicta (in the buffer zone, a deforested area). Both species show globally invasive distribution patterns, as does Solenopsis geminata, which is native to the region (recorded in both the core and buffer zones). This suggests that the core zone of CNP remains largely unaffected by highly invasive ant species. Five species represent new records for Honduras: Pheidole natalie, Strumigenys cf. calamita, Solenopsis invicta, Solenopsis texana/carolinensis, and Pseudomyrmex pallens. The relatively high number of new country records identified in a single protected area indicates the understudied nature of the region, confirming the need for more field surveys.

The discovery of Pheidole balatro males is another noteworthy find. Given the large diversity of Pheidole ants, it is common practice not to describe newly discovered males through extensive morphological descriptions, which we have adopted here. The provided photograph and information in the results section could, however, provide a useful basis for comparison.

The genus Procryptocerus is understudied and requires greater taxonomic attention in order to better understand populations and species delimitations. Though we present P. batesi Forel, 1899, here, we were unable to exclude the possibility of P. mayri Forel, 1899, given that they are very morphologically similar.

Results suggest that the recorded species richness (162) appears to be an underrepresentation of the actual species richness in CNP (asymptotic estimate 257; 95% CI: 204–365) (Figs 5, 6). The observed and extrapolated patterns in Fig. 5 can be interpreted as the relative efficiency of techniques for capturing different ant species. Since a plateau of species richness is not reached for any of the sampling techniques, all show much potential to record additional species with additional sampling. However, the rate of increase in species richness appears to have reached a maximum for most all techniques, meaning that greater sampling will be required to keep finding additional species.

Community sample coverage is relatively high overall, suggesting that our sampling is rather representative of the ant community in CNP. This holds true for the communities addressed by leaf litter extraction, pitfall trapping, and baiting (CSC >0.9). For these techniques, the probabilities of sampling additional species are relatively low. It seems that high community sample coverage only increases mildly with additional species. Species that have not been recorded yet at this point in the accumulation curve seem to play only a small role in the ant community and are likely to be rare. However, hand collection, vegetation beating, and Malaise trapping still seem to show underrepresentation of the community (CSC <0.9). The relative probabilities of sampling additional species are relatively higher when compared to the other three techniques. Hand collection and vegetation beating still show relatively high sample coverage (CSC = 0.8–0.9), with trends suggesting that some common species might still be added using these techniques. Malaise trapping shows a low sample coverage of <0.5 and thus might yield many more species, both common and rare. Sample coverage for hand collection might be lower because of less consistent sampling along the elevation range. For vegetation beating and Malaise trapping, this is probably due to lower sampling numbers (N = 4 and 16, respectively). It is interesting to consider that just four Malaise traps captured a similar number of ant genera as extensive pitfall trapping, although the collected genera are more typical of arboreal ant fauna, such as Procryptocerus, Pseudomyrmex, and Crematogaster, indicating an alternative faunal community sampled.

The elevational range addressed with each technique is to be considered. The species accumulation rate of leaf litter extraction, baiting, vegetation beating, and Malaise trapping is expected to increase when used along larger elevational ranges, especially when including higher altitudes, which will likely collect altitude specialist species and subclades (e.g., Stenamma).

Despite the substantial species list accumulated from the two projects, there remains high potential to add more species. This study confirmed the presence of 162 ant species; however, a total of 250+ species is predicted to be present in CNP. Knowledge gaps are presented below, which could be considered in order to obtain a more complete ant species inventory for the park. First of all, it is important to note that there is a lack of specific canopy and subterranean sampling. Though Malaise trapping and leaf litter extraction may sample a subset of those communities, more species are likely present and have yet to be collected.

By expanding methodological approaches and sampling along a broader elevational range, we increased the number of ant species recorded from CNP. However, there are still some elevation zones that were not sampled using the primary survey methods (Fig. 2), and an unequal sampling effort was used along the elevation gradient. To address these sampling gaps, we recommend employing a variety of sampling techniques along the full elevational gradient with appropriate replication in order to ensure the inclusion of less prevalent species.

The middle elevation ranges were sampled most intensely; however, the higher (mountain peaks) and lower elevation ranges (e.g., buffer zone) remain undersampled. Leaf litter extraction results in the highest number of genera and species recorded, followed by hand collection and pitfall trapping. Although hand collection and pitfall trapping were used along a broader elevational gradient than leaf litter extraction, the latter still shows a higher number of ant species captured. Sampling leaf litter at higher altitudes, in particular, should provide promising results. Baiting is still unexplored at higher altitudes. Additional Malaise trapping and vegetation beating are recommended in general, regardless of elevation range.

As a hotspot of biodiversity, numerous novel species have been previously described from CNP (Martin et al. 2021), including ants (Branstetter 2013; Longino 2013, 2019; Prebus 2021a). Further targeted surveys in CNP are expected to lead to the discovery of more ant species, especially at the higher, undersampled altitudes and buffer zones.

Sampling in a standardised framework would allow for a better understanding of species ecology and the taxonomy of the local ant fauna, which could then lead to improved knowledge of regional diversity and wider biogeographical patterns (especially for highly understudied groups such as those in the genera Apterostigma, Procryptocerus, and Temnothorax). The effect of anthropogenic habitat change could also be examined, given persistent habitat alterations across both core and buffer zones.

The authors have declared that no competing interests exist.

No ethical statement was reported.

LLAMA was funded by a National Science Foundation grant (DEB-0640015).

According to the author order and the credit categories: Conceptualization (FCDW, MJ, and MTH); Methodology (FCDW, DO, and MTH); Validation (FCDW, MGB, WD, and MTH); Formal analysis (FCDW and MTH); Investigation (FCDW, DO, DDG, JS, RVO, and MTH); Resources (TEM, WD, and MJ); Data curation (FCDW and MTH); Writing – original draft (FCDW and MTH); Writing – review and editing (FCDW, DO, MGB, DDG, WD, MJ, TEM, JS, RVO, and MTH); Visualisation (FCDW, WD, and MTH); Supervision (FCDW and MJ); Project administration (FCDW).

Frederik C. De Wint https://orcid.org/0000-0003-0819-3903

Michael G. Branstetter https://orcid.org/0000-0002-3734-6166

Wouter Dekoninck https://orcid.org/0000-0002-7065-7310

Merlijn Jocque https://orcid.org/0000-0002-7196-7476

Jennifer Sudworth https://orcid.org/0009-0005-2994-5205

Matthew T. Hamer https://orcid.org/0000-0001-6728-9046

All of the data that support the findings of this study are available in the main text or Supplementary Information.