First record of the red imported fire ant, Solenopsis invicta Buren, from Hispaniola

The red imported fire ant, Solenopsis invicta Buren, is one of the most damaging pest species in the world. A native to the floodplains of subtropical South America, it has been spread to North America and more recently to Australia and Asia. In order to effectively control, manage, and mitigate its impact and spread, it is critical to understand its current distribution. While S. invicta has long been known from the smaller Caribbean islands, it has not been documented from the two largest islands (Hispaniola and Cuba). We present the first record of S. invicta from Hispaniola and discuss its presence in relation to its native congener, S. geminata.


Introduction
Invasive pests are a major threat to human health, environment, and biodiversity (Mazza et al. 2014;Bellard et al. 2016). Early detection and immediate control and management of invasive species are the most efficient, effective, and cost-effective prevention measures (Mehta et al. 2007;Ujiyama and Tsuji 2018;Polasky 2010). The red imported fire ant, Solenopsis invicta Buren, 1972, is one of the most invasive and damaging pests in the world (Lowe et al. 2004). Native to periodically flooded grasslands of subtropical South America (Ascunce et al. 2011), it has been introduced to the USA in the mid to late 1930's (Tschinkel 2006). From its initial entry point in Mobile, Alabama, it has spread throughout the entire Southeast, covering a total of over 148 million hectares in 14 states and Puerto Rico (USDA APHIS 2020). Negative impacts of S. invicta infestation are wide-ranging and severe: they damage electrical equipment, damage crops by feeding on seeds, seedlings, and developing fruit, and are medically relevant pests of people, pets, livestock, and wildlife (Tschinkel 2006). Although there is some debate over the exact nature and degree of the ecological impact in undisturbed ecosystems (King and Tschinkel 2013a, b;Stuble et al. 2011Stuble et al. , 2013, the negative impact in human-dominated environments is uncontroversial. Estimates of the exact economic costs of fire ant damage and control for the southeastern USA alone vary from more than half a billion to several billion dollars per year (Williams et al. 2001).
More recently, S. invicta has also spread far beyond the invasive range in North America. Earlier this century, it has been reported from mainland China, Taiwan, India, Philippines, Australia, and New Zealand (from where it was subsequently successfully eradicated) (Wetterer 2013). Interestingly, most of these populations all seem to have derived from the invasive US population (Ascunce et al. 2011), but China now appears to serve as a bridgehead for its further spread in Asia (Wylie et al. 2020) and a few incursions in Australia derived from the native range in Argentina (Wylie et al. 2016). Solenopsis invicta has been known from the Caribbean since the 1980's (Wetterer and Davis 2010;Wetterer 2013;Wetterer et al. 2014). To date, S. invicta has been reported from the Lucayan Archipelago, the Lesser Antilles, and most islands of the Greater Antilles (Puerto Rico, Jamaica, the Cayman Islands, and the Virgin Islands).
Interestingly, S. invicta has not yet been described from the Hispaniola, the second largest of the Caribbean islands with the longest history of European settlement, despite active entomological biodiversity research being conducted in the Dominican Republic (Lubertazzi 2019). The reasons for this could be manifold. Although unlikely, the species may not (yet) be established on the island for simple stochastic reasons, since unintended transport and establishment are probabilistic processes. If the introduction was recent, S. invicta may still be rare and highly localized, making detection unlikely. Further, a sampling bias of general biodiversity collections towards pristine habitats would also bias against the discovery of S. invicta, since this species prefers heavily disturbed habitats (Tschinkel 2006). Finally, S. invicta is easily confused with the native tropical fire ant, S. geminata (Fabricius, 1804), and can readily be mistaken for its congener if one does not utilize taxonomic keys including the South American native.
Recently, Dr. Santiago Martinez was presented with a medical case of a Santo Domingo patient displaying recurrent episodes of angioedema with skin lesions consisting of pustules typical of S. invicta stings, leading us to investigate its presence on Hispaniola.

Materials and methods
Sample collection: We searched for fire ants from August 26-31, 2019 in urban and agricultural areas of National Province, Puerto Plata Province, and Hato Mayor Province (Figure 1). A total of 18 fire ant nest series were collected from nests or foraging trails (Table 1). When possible, we always collected major workers to aid in morphological species identification (Trager 1991;Pitts et al. 2018). Vouchers are deposited in the entomological collection of the National Museum of Natural History (USNM), Smithsonian  Institution, Washington DC, USA (USNMENT01129693-USNMENT01129710) and at the Museo Nacional de Historia Natural (MNHNSD), Santo Domingo, Dominican Republic (USNMENT01126782-USNMENT01126799). Species identification: We used the keys in Pitts et al. (2018) and Trager (1991) to identify samples to species. For each nest series, we used the three largest workers (ideally majors) to work through the taxonomic keys. We further compared our samples to extensive fire ant collections at the NMNH.
Since species identification of fire ants based solely on morphology is error prone even for experienced systematists due to the paucity of fixed character differences between species and rampant intra-and interspecific variability, we also conducted molecular genetic-based species identification using DNA barcoding (Hebert et al. 2003) of specimens hypothesized to be S. invicta.
DNA barcoding: Genomic DNA was successfully isolated from the four samples provisionally identified as S. invicta at the Smithsonian Laboratory of Analytical Biology (LAB) using the Autogen Mouse Tail Tissue Prep. Only a single ant was extracted per collection event to avoid subsampling the same colony. Amplification and sequencing of the DNA barcode followed the standard protocol for this locus as applied to ants (Smith et al. 2005). Briefly, the barcoding fragment of the cytochrome c oxidase subunit I (COI) gene were amplified using the Folmer and jg primer pairs (Folmer et al. 1994;Geller et al. 2013) using the protocol of Smith et al. (2005). Amplification success was verified on an agarose gel. PCR products were purified using ExoSAP-IT (ThermoFisher Scientific, USA) and Sanger sequenced in both directions with the PCR primers (with BigDye Terminator v3.1 (Applied Biosystems, USA)) and run out on an ABI 3730xl DNA Analyzer.
After automated quality control and manual inspection of sequencing traces using Geneious Prime 2019.2.3 (Biomatters 2019), all reads from a specimen were assembled to generate full length COI barcodes (GenBank accession number MT909544). Barcodes were used for species identification by searching NCBI GenBank nr database (National Center for Biotechnology Information 2020) using megablast (Morgulis et al. 2008) in the BLAST+ 2.10.0 package (Zhang et al. 2000;Camacho et al. 2008). We also downloaded the top 100 COI sequences returned from the blast search and constructed a maximum likelihood tree using automated partition selection scheme (greedy strategy, Lanfear et al. 2012) and model selection (Modelfinder, Kalyaanamoorthy et al. 2017) implemented in IQ-TREE 1.16.10 (Nguyen et al. 2015). We implemented ten independent runs and estimated branch support using ultrafast bootstraps (1,000 replicates, Minh et al. 2013;Hoang et al. 2018) and two approximate likelihood-based measures (aLRT) (Shimodaira-Hasegawa-aLRT [SH-aLRT] with 1,000 replicates and the Bayesian-like transformation aLRT [aBayes] of Anisimova et al. 2011, which is particularly robust for branch-support of very short branches).

Figure 2.
Images of head and side aspects of minor workers of Solenopsis geminata (top) and S. invicta (bottom). The two species are readily distinguished by the huge, cordate head in major workers in S. geminata (not shown). However, even without access to major workers, the two species can readily be differentiated using minor workers (shown here). Further characters are the presence of a medial tooth in S. invicta and a biphasic coloration of the head and legs in S. geminata.

Results
Of the 18 fire ant samples collected, we identified 14 as the Tropical Fire Ant, Solenopsis geminata, a wide-spread fire ant from Central and South America and also presumably native to the Caribbean (Wetterer 2011;Gotzek et al. 2015). The remaining four samples we identified as the red imported fire ant, S. invicta Buren. These four samples were all collected from nests at a single site, a coffee plantation in central Hato Mayor Province (Figure 1). The plantation itself was established in the mid 1980s and is surrounded by other types crop agriculture (e.g., cacao). Key features we used to differentiate S. invicta from its congener were lack of megacephaly in even the largest workers, presence of a median tooth on the clypeal margin, and homogeneous reddish-brown color of the head, thorax, and legs (see Figure 2). While not all nest series of S. geminata contained majors, all lacked the distinguishing median clypeal tooth. Additionally, in our samples all S. geminata were the bi-color morphs typical of the Caribbean cluster (Gotzek et al. 2015), a color pattern never seen in S. invicta (D. Gotzek pers. observation). The DNA barcodes confirmed our morphological species identification. All four ostensible S. invicta samples shared a single COI haplotype. BLAST searches using this haplotype of the NCBI nucleotide collection database returned a S. invicta sequence from the USA as their top hit (Genbank accession number HQ215538) with very high support (E values = 0.0, percent identity > 99.2%, coverage 100%). Further, a maximum likelihood tree of the top 100 COI blast results not only placed our newly generated S. invicta sequence in the midst of their conspecific sequences with very high support, but shows it is identical to other S. invicta COI sequences from the USA (Figure 3).

Discussion
We present evidence documenting the highly invasive pest species, Solenopsis invicta, on the Hispaniola for the first time. Solenopsis invicta has been continually expanding its invasive range since first discovered around Mobile, AL in 1942 (Wilson and Eads 1949;Tschinkel 2006). From the USA, this species has been moved to Australia and Asia (Ascunce et al. 2011) and is also well known from other Caribbean islands, especially the Lesser Antilles (Davis et al. 2001;Wetterer 2013). Since the Dominican Republic has the largest economy in the Caribbean and its most important trading partner is the USA (Central Intelligence Agency 2020), it is not surprising to also find S. invicta on Hispaniola. It is a little bit more difficult to explain why the species has not been detected earlier. Either it is a recent arrival, rare, only found in disturbed habitats (its discovery would be biased against during biodiversity assessments, which often focus on undisturbed habitats), or simply misidentified.
We note that based on our admittedly brief observation, S. invicta is not very widespread, nor common or dominant, especially not compared to its congener, the Tropical Fire Ant. The latter species is widespread, common, and dominant, especially in urban centers. Dominicans attest that this is one of the most common species on the island (D. Gotzek pers. comm.), although it is likely that the two species are conflated in local folk taxonomy. Solenopsis geminata is a well-known element of the Caribbean fauna since the mid-19 th century (Wetterer 2011). Creighton (1930) describes it as very abundant in the Antilles and it readily occupies urban and agroecosystems (Perfecto 1991;Holway et al. 2002;Perfecto and Vandermeer 2011) and is often ecologically dominant (George and Narendran 1987). Wilson (2005) argued it being one of the infamous Hispaniolan plague ants of the early 16 th century.
Two not mutually exclusive reasons can explain the scarcity of S. invicta in the Dominican Republic. First, it is possible that S. invicta is a recent arrival to Hispaniola and is not yet widespread enough, nor has yet built up sufficient population density to be common enough to be readily collected. This is possible, but not likely. The Dominican Republic has the largest GDP of Caribbean economies and has strong trading relationships in the Caribbean and further abroad. The USA in particular is its primary trade partner (41.4% imports from USA in 2017), from where it imports food and fabrics (Central Intelligence Agency 2020), giving ample opportunity for S. invicta to invade. Other Caribbean countries, especially the West Indies, with much smaller economies, have been known to have S. invicta since the 1980's (Wetterer 2013;Wetterer et al. 2014). Since trade is a major conduit for fire ants (Ascunce et al. 2011;Gotzek et al. 2015), it is unlikely that the S. invicta is a late arrival to the Dominican Republic.
Second, S. invicta may not be able to outcompete S. geminata in the tropics, as it has done along the Gulf Coast in the USA (Porter et al. 1997). Even though Morrison et al. (2004) predict the potential range of S. invicta to encompass much of the tropics (including the Caribbean), their model is based exclusively on rainfall, an environmental factor well known to restrict S. invicta in the USA. However, other climactic and especially biological factors are likely to limit its range elsewhere, effectively limiting its ecological dominance to the subtropics. In this scenario, it may well behave more like a tramp ant than an invasive pest species in the Caribbean. Now that S. invicta has been documented on Hispaniola, researchers and biosecurity officers need to be careful to distinguish between the native S. geminata and invasive S. invicta. While the bi-color morphs are typical of the Caribbean cluster S. geminata (Gotzek et al. 2015) and are never seen in S. invicta (D. Gotzek pers. observation), we strongly advise against using coloration as a sole or primary discriminatory character during specimen identification. This is especially true since S. geminata also has a red phase (S. geminata rufa), which is well documented from the Caribbean (Creighton 1930) and thus renders them easily confused with S. invicta. Megacephaly in the largest workers is also a character never seen in S. invicta (Trager 1991;Pitts et al. 2018), but like coloration it is also an unreliable character since major workers or not always collected. Our recommendation is thus using the median tooth on the clypeal margin (present in S. invicta, absent in S. geminata) as the key character to differentiate S. invicta from its congener. To guard against mis-identifications due to possible wearing down of the median tooth in some individuals, we consider it prudent to always inspect a series of workers from a given collection / nest series (which is a general best practices recommendation when identifying taxonomically challenging taxa).
While further studies are required to determine S. invicta's exact distribution and population densities on Hispaniola, its ecological dominance (especially in relation to S. geminata), and whether it poses an ecological, agricultural, medical, or economic threat, our findings warrant an immediate Risk Assessment and an Emergency Response Plan. Multiple case studies in New Zealand and Australia have shown that with early discovery and quick decisive action, eradication of S. invicta incursions are possible (Hoffmann et al. 2016).