Introduced Pheidole of the world: taxonomy, biology and distribution

Abstract The objective of this study is to provide a detailed taxonomic resource for identifying and studying ants in the genus Pheidole that have established beyond their native ranges. There is an increasing need for systematists to study taxa of specific concern to 21st century environmental, food security and public health challenges. Systematics has an important role to play in both the theoretical and applied disciplines of invasion biology. Few invaders impact terrestrial ecosystems more than ants. Among the world’s 100 worst invasive species is the cosmopolitan and highly destructive Pheidole megacephala (Fabricius). Accurate identification of Pheidole megacephala is imperative for the success of screening, management and eradication programs designed to protect native ecosystems from the impacts of this destructive species. However, accurate identification of Pheidole species is difficult because of their taxonomic diversity, dimorphic worker caste and lack of taxonomic resources. Illustrated keys are included, along with the taxonomic history, taxonomic diagnoses, biological notes and risk statements for the 14 most invasive members of the genus. Global distribution maps based on over 14,000 specimen and literature records are presented for each species. These results of this work will facilitate identification of pest species, determination of climatic and habitat requirements, discovery of pest origins, horizon scanning and assessment of invasion pathways. The following new synonym is proposed, with the senior synonym listed first and the junior synonyms in parentheses: Pheidole indica Mayr (= Pheidole teneriffana Forel, and its synonyms Pheidole taina Aguayo and Pheidole voeltzkowii Forel). Pheidole navigans Forel, stat. rev., stat. n. is removed from synonymy and elevated to species rank. It is proposed that records of Pheidole moerens Forel outside of the Mesoamerica and the Caribbean refer instead to Pheidole navigans or other heterospecific taxa in the Pheidole flavens species complex. We propose that the names Pheidole anastasii Emery and Pheidole floridana Emery have been widely misapplied to North American outdoor records of Pheidole bilimeki Mayr. It is suggested that the synonymy of Pheidole lauta Wheeler be transferred from Pheidole floridana Emery to Pheidole bilimeki Mayr.


Introduction
The science of systematics has historically focused on the discovery, description and classification of life on earth with relatively little attention given to the ecological or economic impact of the study taxon. Today, there is an increasing need for systematists to study taxa of particular concern to 21 st century environmental, food security and public health challenges. Systematics has an important role to play in both the theoretical and applied disciplines of invasion biology (Gotzek et al. 2012;Le Roux and Wieczorek 2009). Although biological invasion is most closely associated with the recent range expansions of species during the Anthropocene, the dispersal of colonist taxa into novel ecosystems is as old as life itself. Study of contemporary invasion ecology and its evolutionary consequences informs a deeper understanding of prehistoric patterns and processes. Correspondingly, study of ecological and evolutionary impacts of ancient colonization events gives historical context to contemporary invasions. Systematics can help bridge this gap separating invasion biology from biodiversity science by advancing integrative theories such as taxon cycles Ricklefs and Bermingham 2002). Systematics is also critical to the success of invasive species management. Pest screening, early detection programs and monitoring efficacy all depend on accurate, species-level identifications. Obtaining high-confidence identifications is particularly difficult for hyper-diverse groups such as insects. The few taxonomic resources that exist for insect identification are too often dated, regional, poorly illustrated, and inaccessible to non-specialists.
Invasive species are among the most significant drivers of global change, and few invaders impact terrestrial ecosystems more than ants (Lach and Hooper-Bùi 2009). Of the approximately 15,000 described ants species, more than 100 have established populations outside of their native ranges (McGlynn 1999). Although invasive ants are economically costly in both urban and agricultural areas, the most serious consequences of their introduction may be ecological. Invasive ants can greatly modify ecosystems by reducing native ant diversity, displacing other arthropods, negatively impacting vertebrate populations, and disrupting ant-plant mutualisms (Suarez et al. 2009).
A small subset of introduced ants have become highly destructive invaders, five of which are currently listed among the world's 100 worst invasive species (Lowe et al. 2000). Unfortunately, detection of non-native ants is hampered by the taxonomic specialization required for accurate species identification of these tiny and overwhelmingly diverse insects. Most of the scientists responsible for identifying ant specimens for pest screening, early-detection programs and monitoring initiatives are not trained ant taxonomists. Although advances in specimen imaging, interactive keys and online resources are welcome developments, increasing the taxonomic capacity for identifying invasive, introduced and commonly intercepted ant species remains a challenge (Sarnat 2011).
Among the world's 100 worst invasive species (Lowe et al. 2000) is the cosmopolitan and highly destructive Pheidole megacephala (Fabricius), known commonly as the big-headed ant. Accurate identification of P. megacephala is imperative for the success of screening, management and eradication programs designed to protect native ecosystems from harmful impacts. Many non-specialists -and many myrmecologists -have confused other species of Pheidole with P. megacephala (Wetterer, 2012). This is not surprising. Pheidole (Formicidae: Myrmicinae) is arguably the most speciose monophyletic ant genus in the world, with over 1,000 extant valid species, 138 valid subspecies (Bolton 2014) and hundreds of undescribed species. Accurate identification of Pheidole species is especially difficult because of their taxonomic diversity, dimorphic worker caste, and lack of taxonomic resources. The term 'Pheidole sp.' is ubiquitous across ecological and biodiversity publications, including many that focus on tramp ants (e.g. Passera 1994). Recent revisions of Pheidole (Eguchi 2001b;2004b;Eguchi et al. 2007;Fischer and Fisher 2013;Fischer et al. 2012;Longino 2009;Longino and Cox 2009;Sarnat 2008;Wilson 2003) have advanced the taxonomy of this difficult group. Even at the pace of these past decades, it will be generations before modern identification resources will be available for the majority of known species.
Identification resources for diagnosing the world's 14 most invasive Pheidole species are presented here (Table 1). These resources include a fully illustrated key, specimen photographs and distribution maps, in addition to sections on taxonomic history, taxonomic diagnoses, biology, and risk statements. The results of this work will facilitate identification of pest species, determination of climatic and habitat requirements, discovery of pest origins, horizon scanning, and assessment of invasion pathways. This analysis of the world's introduced Pheidole indicates that two of the most widespread tramp species, P. indica Mayr and P. teneriffana Forel, each the subject of considerable research and attention, are actually conspecific. This underscores the importance of systematics in understanding biodiversity dynamics in the Anthropocene.

Defining invasiveness
Biological invasion is a process that is most simply described by successive stages of transport, introduction, establishment and spread (Vermeij 1996). Quantifying invasiveness is a difficult task, but one made easier by the unified framework for biological invasion proposed by Blackburn et al. (2011). We incorporate Pheidole taxa into this framework in an effort to advance comparative invasion biology, but acknowledge that these categories fail to convey the idiosyncrasies of each lineage's invasion history. Separating native range from introduced range is similarly problematic. It is likely that all the treated species are expanding their range within their native biogeographical region as the result of human activity and global environmental change. We therefore define introduced populations strictly as those occurring beyond the lineage's native biogeographic realm.
Following the Blackburn et al. (2011) framework, we assign four Pheidole lineages to category 'C0' (individuals released into the wild in location where introduced, but incapable of surviving for a significant period) ( Table 2). Pheidole noda fits this category, as it is recorded as being introduced only in European greenhouses and is not known to have established permanent populations outside its native range in Asia. The other three species we assigned to the 'C0' category all belong to the Mesoamerican P. punctatissima clade. All three have been reported as occurring indoors beyond their native bioregion. Of these, P. bilimeki is the most widespread and maintains established reproducing outdoor populations in the southern United States. While it can be argued that these northern populations resulted from human-mediated dispersal, they remain contiguous with putatively native Mesoamerican populations, thus we refrain from defining them as introduced.
Nine lineages are assigned to category 'D2' (individuals surviving in the wild in locations where introduced, with reproduction occurring, and population self-sustaining). Table 1. Species of Pheidole recorded as established outside of their native bioregion. The clade or species group to which each species belongs is listed and defined by the given authority. Clade names are informal designations and are given to convey broad evolutionary relationships among the species. Asterisks (*) note clade designations that are presumed based on morphological similarity.
Biological characteristics of introduced Pheidole species arranged by species name. Sizes (head width measured in mm) are the same as the observed values reported in the species accounts. Invasiveness codes refer to Blackburn et al. (2011). Asterisks indicate name used in reference is considered here to be either a junior synonym or misapplication. References listed: (1) Birkemoe and Aak 2008, (2) Longino and Cox 2009, (3) Naves 1985, (4) Wilson 2003, (5) Morrison 1996, (6) Passera 1994, (7) Reimer 1994, (8) Martínez 1996, (9) Chen et al. 2011, (10) Wilson and Taylor 1967, (11) Boer and Vierbergen 2008, (12) Sarnat and Economo 2012, (13) Longino 2014, (14) Martínez 1992, (15) Fischer and Fisher 2013, (16) Hölldobler and Wilson 1990, (17) Hoffmann 1998 Delabie et al. 1995, (19) Yamamoto et al. 2009, (20) Yamawo et al. 2012, (21) King and Tschinkel 2007, (22) Storz and Tschinkel 2004, (23) Man and Lee 2012, (24) Green and Gunawardana 2006, (25) Berry et al. 1997, (26) Harris et al. 2005, (27 The least invasive of these are likely the three Australian species (P. proxima, P. rugosula, P. vigilans) that have established persistent populations in New Zealand, but have not been reported from elsewhere (although P. vigilans is reported as introduced in Western Australia). Only P. rugosula has been listed in examined interception records, and that was a single New Zealand record from Australia (Ward et al. 2006). Pheidole parva, P. fervens and P. indica are all from the Indomalayan bioregion and have managed to establish reproducing outdoor populations beyond their native range. Although the propagule pressure of these species is relatively strong, as evidenced by their frequent interception at ports of entry (Table 3), the introduced populations of all three tend to be small and relatively localized. Pheidole obscurithorax and the P. flavens-complex (including P. flavens and P. navigans) are both Neotropical lineages that have established persistent and actively spreading populations in the southern United States. The former is documented as causing a greater ecological impact, but the latter is more widespread and appears to have greater propagule pressure, as evidenced by high numbers of interception records and establishment of at least temporary populations in California and several Pacific Islands.
Pheidole megacephala is the only species assigned to category 'E' (fully invasive species, with individuals dispersing, surviving and reproducing at multiple sites across a greater or lesser spectrum of habitats and extent of occurrence). The vast majority of introduced Pheidole specimen and occurrence records are attributed to P. megacephala. It is the most geographically widespread species in the entire genus and its impact on native ecosystems and agriculture are extensively documented.

General characteristics and characters associated with invasion success
All Pheidole species treated here have a dimorphic worker caste. Their colonies typically have hundreds to thousands of workers. They are all generalist foragers that feed on some combination of dead arthropods, living arthropods, seeds and human foodstuffs. The aforementioned characteristics are shared by nearly all of their congeners, however, and cannot be considered promoters of invasion success among Pheidole. There is a suite of biological characters that are broadly associated with introduced populations of invasive ants, including unicoloniality and omnivory (Holway et al. 2002). Unicoloniality -defined as the ability to form expansive and polygynous (multiple queened) supercolonies -has only been observed in P. megacephala (Table 1). Only three other species (P. fervens, P. indica and P. noda) are reported to be at least facultatively polygynous. Four species besides P. megacephala are reported to exhibit polydomous populations (P. anastasii, P. bilimeki, P. fervens and P. indica).
One interesting pattern deserving further study is the propensity of introduced Pheidole to use vegetation for either foraging or nesting (Table 1). Pheidole species, in general, are most strongly associated with the ground, and exploitation of the vegetative or arboreal strata is relatively uncommon. One potential reason that foraging and nesting in vegetation is overrepresented among the introduced species is that, if quarantine interception records are any indicator, human-mediated dispersal events are predominately associated with commercial trade of plants or plant material (Suarez et al. 2005;Ward et al. 2006). Furthermore, none of the introduced Pheidole species are strictly arboreal, and their capacity for occurring on vegetation reflects the type of broad habitat tolerances required for successful establishment.

Taxonomic patterns
Although referring to each of the included lineages as a discrete biological species is convenient, there are at least some instances -including the P. flavens species complex and P. megacephala complex -that defy such neat classification. A disproportionate number of synonyms and infraspecific names in the genus Pheidole belong to the lineages treated here, and this pattern holds true across the Formicidae. We offer several explanations for this pattern. The first is attributed to nomenclatural artifact. Taxonomists unfamiliar with distant faunas and working outside of a global context often described introduced populations as new species. The second explanation for the myriad names associated with invasive species reflects a truly biological pattern: invasive populations tend to be derived from geographically widespread and morphologically variable lineages. Geographically widespread species have greater propagule pressure because they are broadly exposed to opportunities for human-mediated dispersal (Theoharides and Dukes 2007). The taxonomic work undertaken during this study suggests the phenotypic diversity of many of these introduced lineages is only a thumbnail of a much broader morphological spectrum observed across their respective native ranges.

Morphological patterns
The 14 Pheidole species treated here do not adhere to a particular morphotype, especially when phylogenetic relationship is corrected for. Although none of the species occupy the extreme ends of the genera's size spectrum, they do range from small to large. None of these species exhibit aberrant or specialized morphology, such as spinescence.
(Although not treated in this review, the P. sexspinosa complex is a spinescent lineage that is considered a tramp ant around the Pacific and has likely increased its range with the help of human-mediated transport.)

Phylogenetic patterns
Invasive Pheidole species are not evenly dispersed across the phylogeny . Rather, a few lineages tend to be responsible for spawning successful invaders. In particular, the punctatissima clade, flavens clade, fervens clade have each given rise to multiple introduced species (Table 1). The most parsimonious explanation for this pattern is that at least some promoters of invasion success are plesiomorphic traits inherited from common ancestors. More generally, all of the known clades to which invasive Pheidole belong can broadly be considered tramp groups composed of species that exhibit relatively wide geographic ranges, few habitat constraints, and high infraspecific variability.

Biogeographical patterns
Strong biogeographical patterns among introduced Pheidole are difficult to find. One pattern shared by all introduced Pheidole is that they invariably occupy low elevation habitat. This is not surprising, as connectivity is much greater between lower elevation sites (e.g. coastlines and shipping ports) than among higher elevation sites (e.g. montane forests). The invasive Pheidole invariably come from tropical and subtropical lineages, but this pattern broadly reflects the richness patterns across the entire genus . The Neotropical, Afrotropical and Indomalayan regions have all produced Pheidole lineages that have invaded other bioregions. Australia is nominally home to three invasive Pheidole species, but the introduced populations of all three are restricted to the island of New Zealand (in addition to Lord Howe Island in one instance) and are not likely capable of invading another continental system. Although the common recipient of non-native Pheidole introductions, and ant introduction in general (McGlynn 1999), Oceania is the only tropical bioregion from which a successful invader has not evolved. Pheidole sexspinosa Mayr and possibly P. oceanica Mayr are native to Oceania and widely considered tramp species, but thus far there is no evidence that either has ever successfully established outside the Pacific.

Taxon selection
The taxa treated here represent all Pheidole species known to have been introduced outside of their native biogeographic region. These taxa span the spectrum from species that have become naturalized across the globe (such as the highly invasive P. megacephala) to species known only to have established temporary indoor populations beyond their native region (such as P. noda). We do not include species that are repeatedly intercepted by quarantine but are never recorded as establishing non-native populations. The species included here represent the vast majority of published Pheidole quarantine interceptions records (Table 3), and have proven the most capable among their congeners of establishing beyond their native range. In addition to the quarantine intercepts, there are many synanthropic tramp species of Pheidole that are likely expanding across their native bioregion with the inadvertent assistance of human exploration and commerce. This is particularly true in Oceania, where species such as P. oceanica, P. umbonata Mayr and P. sexspinosa are widespread across the entire region. However, we were unable to confirm any records of their introduction outside of Oceania. While excluded from our current study, we advise readers to be aware of these and similarly widespread species. Their expansive ranges increase the propagule pressure for anthropogenic dispersal, and their high tolerance for habitat disturbance pre-adapts them for establishing beach-head populations outside their native ranges.

Occurrence and specimen records
Our biogeographic data are taken from the Global Ant Biodiversity Informatics (GABI) project, a database consolidating literature, museum, and biodiversity database records on ant species distributions (Suppl. material 1). Each literature record for an occurrence outside the putative native range was examined by reviewing the primary reference and evaluating it for veracity and accuracy. Specimen records included in the GABI database were similarly evaluated. Literature records considered to be derivative (e.g. checklists referring to a previously published record) and records from online checklists were excluded unless the primary reference or specimen record was confirmed. A confirmed literature record in the context of this study means only that the valid name or a synonym was verified as appearing in the text, and does not imply that the species identification was accurate. We verified a total of 14,162 occurrence records.
Locality references of literature records were converted to coordinates where possible. First, all variables that described the location of a record were merged into a single string that contained all descriptive information about a location. Unique values were extracted from these strings resulting in a total of 3,803 unique locations. Locations that referred to large areas such countries or states (n=221) were removed from these 3,803 locations. All of the remaining unique locations were then converted to coordinates using the Bing geocoding API (Microsoft 2015). From these unique locations 1,265 were geocoded with a tolerable precision (+/-10 km) for a global distribution study. These geocoded locations were manually checked for inconsistencies. Of the remaining 2,538 locations the coordinates given in the database were assigned as the correct coordinates when available (n=1,349). This resulted in a total of 968 locations that lacked coordinates. These records contained problematic locality names and were geocoded manually by correcting the locality name and using the GEOLocate web application (Rios and Bart 2010). Of these 968 records 83 were incorrect and could not be geocoded, and 210 did not refer to a point location (country or state).

Illustrations
Original specimen images taken by the authors were taken using the Auto-Montage software package (Syncroscopy) in combination with a JVC KY-F7U digital camera mounted on a Leica MZ16 dissecting scope, and the software package Helicon Focus in combination with a Leica DFC450 digital camera mounted on a Leica M205C dissecting scope. Vector artwork used to illustrate character states referred to in the taxonomic key and species diagnoses were made in Adobe Illustrator by tracing specimen photographs. All specimen images are available from Antweb.org and can be searched for using the specimen identifier. All vector illustrations are available from the 'Introduced Pheidole taxonomic characters' media gallery on Antkey.org (http:// antkey.org/en/gallery?f[0]=im_field_smg_galleries%3A33508).

List of abbreviations of museum collections
The abbreviations follow Evenhuis (2009) and are used in the text in place of the full museum collection name. Type material from these collections examined by the authors is noted in the species accounts.

Measurements
Measurements reported here include those taken and reported by various researchers. Original measurements taken by the authors were made with a stereo microscope at 40× magnification using a dual-axis stage micrometer wired to digital readouts. Morphometric measurements were recorded in thousandths of millimetres, but are reported here to the nearest hundredth as a range from minimum to maximum across all measured specimens. Specimens for measurements were chosen to reflect potential morphological variation across the full geographic range. The number of specimens from which measurements were taken for a given caste is referred to by n. Measurements for Neotropical Pheidole include data supplied by John Longino. Measurements for Old World Pheidole include values reported in previously published studies (Eguchi 2001a;2004b;Eguchi et al. 2007;Fischer and Fisher 2013).

EL
Eye Length (mm): Maximum diameter of eye measured in profile view.

HL
Head Length (mm): Maximum distance from the midpoint of the anterior clypeal margin to the midpoint of the posterior margin of the head, measured in full-face view; in majors, measured from midpoint of tangent between anteriormost position of clypeus to midpoint of tangent between posteriormost projection of the vertex.

HW
Head Width (mm): Measured at widest point of the head, in full-face view behind eye level.

SL
Scape Length (mm): Maximum scape length, excluding basal condyle and neck.

Identification keys to introduced species of Pheidole
Readers are warned that there are hundreds of native Pheidole species that are not treated in the following keys. The keys are most useful for diagnosing Pheidole specimens intercepted at quarantine facilities, collected from regions with depauperate native Pheidole diversity (such as small islands), and in highly disturbed habitats such as urban areas. Even in urban areas, however, there remains considerable likelihood that native Pheidole species occur that are not treated here, and readers are cautioned to use these keys judiciously.

Lucid3 Key
An interactive and fully illustrated Lucid3 key that includes all Pheidole species treated in this study is available from the website Antkey.org (Sarnat and Suarez 2012) at the following URL: http://antkey.org/en/content/key. To use this key for introduced Phei-dole identification, users are advised to first filter by the genus Pheidole, then proceed by using the 'best' and 'next best' functions. Users are referred to documentation and video tutorials on the webpage for additional instructions and best practices.

Combined major and minor workers
1 Major + minor Postpetiole swollen relative to petiole (Fig. 1, Fig. 2 Major + minor Postpetiole with a posterodorsal (Fig. 1a) and anteroventral ( Fig. 1b) bulge. Promesonotum in profile forming a single dome (Fig. 4, major; Fig. 42 minor), lacking a distinct mound or prominence on the posterior slope. Major Head heart-shaped (Fig. 6); posterodorsal surface smooth, glossy and entirely lacking strong rugoreticulate sculpture (Fig. 9). Minor Antennal scapes surpass posterior head margin by approximately same length as eye (Fig. 40) 2b). Promesonotum in profile with two convexities, the large anterior dome in addition to a distinct mound or prominence on the posterior slope (Fig. 5,major;Fig. 43,minor). Major Head subquadrate (Fig. 7); dorsal surface covered in strong longitudinal rugae that form a reticulated network laterally and posteriorly (Fig. 8). Minor Antennal scapes surpass posterior head margin by approximately twice eye length (Fig. 39)  Major + minor Promesonotum in profile with two convexities, the large anterior dome in addition to a distinct mound or prominence on the posterior slope (Fig. 5,major;Fig. 43,minor). Relatively large species with long limbs (HW major > 1.10 mm, HW minor > 0.50 mm). Major Head with strong rugoreticulate sculpture at least on posterolateral lobes (Fig. 8). Minor Head glossy (Fig. 36); sculpture restricted to at most a few arcuate carinae between eye and antennal insertion. Antennal scapes with erect hairs (Fig. 56); scapes surpass posterior head margin by at least a distance equal to or greater than eye length (Fig. 39) (Fig. 4), lacking a distinct mound or prominence on the posterior slope (sometimes with a weak protuberance or inconspicuous mound). Size and relative limb length variable. Major Head with variable sculpture patterns including glossy ( Major Head almost entirely covered by network of intersecting rugae ( Fig.  12a), lacking long, well-organized and parallel longitudinal rugae on the frons (Fig. 12b). Frontal carinae indistinct, quickly becoming integrated into dense rugoreticulum that covers the entire face. Antennal scrobes entirely lacking. Antennal insertions surrounded by deeply excavated pits (Fig. 12c). Head often a lighter reddish brown than the mesosoma. Minor Posterior head margin strongly convex in full-face view such that the head outline forms a single unbroken curve from eye to eye (Fig. 44). Petiole and postpetiole strongly sculptured laterally (Fig. 47). Antennal scapes extremely long, surpassing posterior head margin by more than 2× eye length (37) ...............obscurithorax -Major Head rugoreticulate on posterolateral lobes and laterad of frontal carinae (Fig. 13a), but frons dominated by long, well-organized and parallel longitudinal rugae (Fig. 13b). Antennal scrobes indistinct to moderately impressed, but frontal carinae always forming a border capable of accepting the antennal scape (Fig. 13c). Antennal insertions not surrounded by deeply excavated pits. Head usually a similar shade as the mesosoma. Minor Posterior head margin weakly convex to flat in full-face view (Fig. 45). Petiole and postpetiole glossy to very weakly sculptured laterally (Fig. 48 Major Frontal carinae relatively longer, extend 4/5 distance of head before terminating (Fig. 14). Promesonotal prominence flatter, less pronounced (Fig. 63a). Propodeal spine weaker, narrower at base, weakly downcurved at apex (Fig. 63b). Minor Promesonotal prominence more flat (Fig. 49a). Metanotal depression deeper (Fig. 49b). Eye relatively small, eye length distinctly less than length of antennal segment 10 (Fig. 65) (Fig. 15). Promesonotal prominence rounder, more pronounced (Fig. 64a). Propodeal spine stouter, broader at base, relatively straight ( Fig.  64b). Minor Promesonotal prominence more convex (Fig. 50a). Metanotal depression shallower (Fig. 50b). Eye relatively large, eye length subequal to length of antennal segment 10 ( Fig. 66)  Major Petiolar node strongly punctate (Fig. 16). Metapleuron with moderate rugulae and some weak punctation (Fig. 16). Hypostomal bridge with a small median tooth in addition to a pair of larger inner teeth (Fig. 18). Smaller (HW < 1.0 mm). Minor Antennal scapes reach but do not surpass posterior head margin (Fig. 41). Mesopleuron entirely punctate (Fig. 52a). Propodeal spines moderately produced and spiniform (Fig. 52b). Petiole distinctly sculptured except for apical portion of node. Smaller (HW < 0.48 mm) .................. proxima -Major Petiolar node mostly glossy (Fig. 17), not covered by punctate sculpture. Metapleuron almost completely glossy with strongly reduced carinulae and lacking punctation (Fig. 17). Hypostomal bridge with two well-developed inner teeth but lacking a median tooth (Fig. 19). Larger (HW > 1.2 mm). Minor Antennal scapes surpass posterior head margin by approximate distance of eye length (Fig. 40). Mesopleuron entirely glossy (Fig. 51a). Pro-podeal spines weakly produced and dentiform (Fig. 51b) Major Promesonotal dorsum glossy with thin but distinct subparallel striae running oblique to the longitudinal midline (Fig. 20). Head with distinct parallel rugae extending from frontal lobes posterior to apices of frontal carinae. Shorter lengths of rugae present across entire posterior region of head and extending to posterior margin in full-face view (Fig. 24). Minor Head with well-defined, long segments of rugae running longitudinally from below the eyes to the posterior head margin (Fig. 38). Frontal carinae distinct and reaching towards the posterior head margin, although they may occasionally be interrupted (Fig. 38). Punctate ground sculpture present on lateral surfaces of head and just mesad of the frontal carinae, but median portion of head with a large glossy section (Fig. 38)  Major Posterolateral lobes, including posterior head margin, covered in rugoreticulum (Fig. 26). Promesonotum in dorsal view transverse with strongly projecting shoulders (Fig. 28). Promesonotal dorsum rugoreticulate with distinct long longitudinal striae in addition to shorter sections of transverse and intersecting striae (Fig. 22). Minor Posterior portion of head with many short to medium length segments of striae distinctly interlaced among punctate ground sculpture (Fig. 59). Antennal scapes with many erect hairs ( Major + minor Gaster with at least anterior 1/3 of first tergite matte (Fig. 33). Major Postpetiole relatively broad; distinctly more than 2× petiolar width in dorsal view (Fig. 31). Promesonotal dorsum usually foveolate and never with distinct transverse striae. Head often entirely foveolate (Fig. 11), but portions of posterolateral lobes can be glossy. Posterolateral lobes never with distinct rugae. Minor Hairs on mesosoma stout, stiff, of equal length and arranged in pairs (Fig. 53). Antennal scapes lack many erect to suberect hairs (Fig. 55). Postpetiole broad in dorsal view, distinctly broader than petiole (Fig. 62)    Posterolateral lobes lacking sculpture (including foveolate ground sculpture, carinae and rugae) posterior to maximum extent of antennal scapes in repose (major worker) 10 Posterolateral lobes rugose or rugulose (major worker)

N Ilustration Description 12
Head almost entirely covered by network of intersecting rugae (a), lacking long, wellorganized and parallel longitudinal rugae on the frons (b). Frontal carinae indistinct, quickly becoming integrated into dense rugoreticulum that covers the entire face. Antennal scrobes entirely lacking. Antennal insertions surrounded by deeply excavated pits (c). Diagnostic characters of P. obscurithorax major workers among introduced Pheidole 13 Head rugoreticulate on posterolateral lobes and laterad of frontal carinae (a), but frons dominated by long, well-organized and parallel longitudinal rugae (b). Antennal scrobes indistinct to moderately impressed, but frontal carinae always forming a border capable of accepting the antennal scape (c). Antennal insertions not surrounded by deeply excavated pits. Illustration applies to P. indica and P. fervens 14 Frontal carinae relatively longer, extend 4/5 distance of head before terminating (major worker). Diagnostic character separating P. fervens from P. indica

15
Frontal carinae relatively shorter, extend 3/4 distance of head before terminating (major worker). Diagnostic character separating P. indica from P. fervens 16 Petiolar node strongly punctate and metapleuron with moderate rugulae and some weak punctation (major worker). Diagnostic character separating P. proxima from P. vigilans 17 Petiolar node mostly glossy, metapleuron almost completely glossy with strongly reduced carinulae and lacking punctation (major worker). Diagnostic character separating P. vigilans from P. proxima 18 Hypostomal bridge with a small median tooth in addition to a pair of larger inner teeth (major worker). Diagnostic character separating P. proxima from P. vigilans

19
Hypostomal bridge with two well-developed inner teeth but lacking a median tooth (major worker) Diagnostic character separating P. vigilans from P. proxima 20 Promesonotal dorsum glossy with thin but distinct subparallel striae running oblique to the longitudinal midline (major worker). Diagnostic character separating P. rugosula from other introduced Pheidole 21 Promesonotal dorsum with distinct transverse striae (major worker). Character present among some species of the P. flavens complex, including P. navigans

22
Promesonotal dorsum rugoreticulate with distinct long longitudinal striae in addition to shorter sections of transverse and intersecting striae (major worker). Illustration refers to P. parva 23 Promesonotal dorsum glossy, lacking foveolate ground sculpture or striae (major worker). Character useful for separating P. vigilans and P. proxima from P. rugosula.

24
Head with distinct parallel rugae extend from frontal lobes posterior to apices of frontal carinae. Shorter lengths of rugae present across entire posterior region of head and extending to posterior margin in full-face view (major worker). Diagnostic character useful for separating P. rugosula from other introduced Pheidole, especially those introduced in New Zealand

N Ilustration Description 25
Posterolateral lobes variously sculptured, but posterior head margin always free of distinct rugae or rugoreticulum (major worker). Illustration refers to P. flavens, P. navigans and other members of the P. flavens complex 26 Posterolateral lobes, including posterior head margin, strongly rugoreticulate (major worker). Illustration refers to P. parva and character is useful for separating that species from those of the P. flavens complex and the P. punctatissima clade 27 Posterolateral lobes variously sculptured, but posterior head margin always free of rugoreticulum (major worker)

28
Promesonotum in dorsal view strongly transverse with strongly projecting shoulders (major worker). Illustration refers to P. parva and character is useful for separating that species from those of the P. flavens complex and the P. punctatissima clade 29 Promesonotum in dorsal view less transverse with weakly projecting shoulders in dorsal view (major worker)

30
Postpetiole relatively narrow, distinctly less than 2× petiolar width in dorsal view (major worker). Character useful for separating members of the P. flavens complex, including P. flavens and P. navigans, from those of the P. punctatissima clade 31 Postpetiole relatively broad, distinctly more than 2× petiolar width in dorsal view (major worker). Character useful for separating members of the P. punctatissima clade from those of the P. flavens complex, including P. flavens and P. navigans 32 Gaster with entire first tergite glossy (major and minor worker). Character useful for separating members of the P. flavens complex, including P. flavens and P. navigans, from those of the P. punctatissima clade 33 Gaster with at least anterior 1/3 of first tergite matte (major and minor worker). Character useful for separating members of the P. punctatissima clade from those of the P. flavens complex, including P. flavens and P. navigans 34 Head bicolored with the yellowish posterior two-thirds contrasting with the darker brown anterior third and rest of body (major worker). Diagnostic character for separating P. punctatissima from all other introduced Pheidole

35
Head uniform in color (major worker). Character used to separate P. anastasii and P. bilimeki from P. punctatissima

36
Head predominantly glossy, lacking punctation and or rugae above eye level (minor worker). Character used to separate P. indica, P. fervens, P. obscurithorax, P. proxima and P. vigilans from all other introduced Pheidole.

37
Head, including the area mesad of the frontal carinae, entirely covered by reticulated network of punctures, giving it a dull appearance; if rugae are present they are generally short segments and mostly restricted to posterior portion of head (minor worker). Character used to separate P. anastasii, P. bilimeki, P. flavens, P. navigans and P. parva from all other introduced Pheidole

N Ilustration Description 38
Head with well-defined, long segments of rugae running longitudinally from below the eyes to the posterior head margin. Frontal carinae distinct and reaching towards the posterior head margin, although they may occasionally be interrupted. Punctate ground sculpture present on lateral surfaces of head and just mesad of the frontal carinae, but median portion of head with a large glossy section. Diagnostic characters separating P. rugosula from all other introduced Pheidole 39 Antennal scapes extremely long, surpassing posterior head margin by more than 2× eye length (minor worker). Diagnostic character separating P. obscurithorax from P. fervens and P. indica

40
Antennal scapes surpass posterior head margin by approximately same length as eye (minor worker)

41
Antennal scapes relatively short, either failing to surpass posterior head margin, or surpassing it by less than the distance of eye length (minor worker)

42
Promesonotum in profile forming a single dome, lacking a distinct mound or prominence on the posterior slope (minor worker)

43
Promesonotum in profile with two convexities, the large anterior dome (a) in addition to a distinct prominence on the posterior slope (b) (minor worker)

44
Posterior margin strongly convex in full-face view such that the head outline forms a single unbroken curve from eye to eye (minor worker). Diagnostic character for separating P. obscurithorax from P. fervens and P. indica

45
Posterior head margin weakly convex to flat in full-face view (minor worker). Diagnostic character for separating P. fervens and P. indica from P. obscurithorax

46
Posterior head margin weakly concave in full-face view (minor worker) Petiole and postpetiole strongly sculptured laterally. Diagnostic character for separating P. obscurithorax from P. fervens and P. indica

48
Petiole and postpetiole glossy to very weakly sculptured laterally (minor worker). Diagnostic character for separating P. fervens and P. indica from P. obscurithorax 49 Promesonotal prominence relatively flat (a); metanotal depression relatively deep (b) (minor worker). Diagnostic character for separating P. fervens from P. indica N Ilustration Description 50 Promesonotal prominence relatively convex (a); metanotal depression relatively shallow (minor worker). Diagnostic character for separating P. indica from P. fervens

57
Posterior head margin relatively broad (minor worker). Diagnostic character for separating P. bilimeki from P. anastasii and P. punctatissima

58
Posterior head margin more narrow (minor worker). Diagnostic character for separating P. anastasii and P. punctatissima from P. bilimeki

59
Posterior portion of head with many short to medium length segments of striae distinctly interlaced among punctate ground sculpture (minor worker). Diagnostic character for separating P. parva from P. flavens and P. navigans

60
Posterior portion of head lacking many short to medium length segments of striae distinctly interlaced among punctate ground sculpture (minor worker). Diagnostic character for separating P. flavens and P. navigans from P. parva

61
Postpetiole narrow in dorsal view, only slightly broader than petiole (minor worker). Diagnostic character for separating P. flavens and P. navigans and from P. anastasii, P. bilimeki and P. punctatissima

64
Promesonotal prominence convex and pronounced (a); propodeal spine relatively stout, broad at base, straight (b) (major worker). Diagnostic character for separating P. indica from P. fervens 65 Eye relatively small, eye length distinctly less than length of antennal segment 10 (minor worker). Diagnostic character for separating P. fervens from P. indica

66
Eye relatively large, eye length subequal to length of antennal segment 10 (minor worker). Diagnostic character for separating P. indica from P. fervens

67
The carinae between eye and mandible are branching and reticulated (major worker). Diagnostic character for separating P. fervens from P. oceanica

68
The carinae between eye and mandible are parallel and not reticulated (major worker). Diagnostic character for separating P. oceanica from P. fervens

69
The length of propodeal spine equal to or less than the diameter of propodeal spiracle (minor worker). Diagnostic character for separating P. fervens from P. oceanica

70
The length of propodeal spine greater than the diameter of propodeal spiracle (minor worker). Diagnostic character for separating P. oceanica from P. fervens

71
Antennal scrobe distinct and narrow, shallow but capable of receiving the entire antennal scape in repose (a); bordered by strong, unbroken frontal carina mesially (b); depression marked by a continuous smooth surface entirely (or nearly entirely) uninterrupted by rugulae (major worker). Diagnostic character for separating P. navigans from P. flavens 72 Antennal scrobe broad, ill-defined, incapable of receiving the entire antennal scape in repose (a); bordered by relatively weak and interrupted frontal carina mesially (b); depression opaque and strongly punctate. Diagnostic character for separating P. flavens from P. navigans  7); often entirely punctate (Fig. 11), but portions of posterolateral lobes can be glossy. Posterolateral lobes never with distinct rugae. Promesonotum in profile forming a single dome (Fig. 4). Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively broad; distinctly more than 2× petiolar width in dorsal view (Fig. 31). First gastral tergite with anterior third to entire surface matte. Minor HW 0.38-0.50, HL 0.44-0.59, SL 0.44-0.58, CI 82-90, SI 106-120 (n=49, Longino pers. comm.). Head dull, entirely covered by reticulated network of punctures (Fig. 37). Posterior head margin relatively narrow and rounded (Fig. 58). Antennal scapes lack standing hairs (Fig. 55); scapes surpass posterior head margin by a distance equal to or greater than eye (Fig. 40). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Hairs on mesosoma stout, stiff, of equal length and arranged in pairs (Fig. 53). Postpetiole narrow in dorsal view, only slightly broader than petiole. Gaster with at least anterior 1/3 of first tergite matte (Fig. 33). Identification, taxonomy and systematics. Pheidole anastasii, P. bilimeki and P. punctatissima all belong to the P. punctatissima clade . These ants are all relatively small species characterized by densely punctate ground sculpture that gives them a dull, matte appearance. Among species treated here, the P. punctatissima clade species are most easily confused with those of the closely related P. flavens complex. Major and minor workers are most reliably diagnosed from those of the P. flavens complex by the relatively broad postpetiole (Fig. 31, major; Fig. 62, minor) and the matte anterior portion of the gaster (Fig. 33) in addition to other characters listed in the key. The minor workers can also be confused with those of Asian native P. parva, but can be distinguished by the more uniform and stout mesosomal hairs (Fig.  53), and by the antennal scapes which lack erect hairs (Fig. 55) and tend to surpass the posterior head margin by a distance equal to or greater than eye (Fig. 40). In the Neotropics, there are many native species that closely resemble P. anastasii (Wilson 2003), and identification of the minor worker subcaste is especially challenging.

Pheidole anastasii Emery
Among introduced members of the clade, the major workers of P. punctatissima are immediately distinguished from those of both P. anastasii and P. bilimeki by the bicolored head (Fig. 33). The minor workers of P. punctatissima tend to have narrower posterior head margins and longer antennal scapes than those of P. anastasii and P. bilimeki. Separating P. anastasii from P. bilimeki is particularly difficult. They are most reliably distinguished by ecological characteristics, with the former preferring to nest arboreally and the latter preferring to nest under stones or in dead wood. The morphological characters separating these two species are highly variable, but the major workers of P. anastasii tend more often towards yellow (versus tending towards brown in P. bilimeki) and can have relatively wider heads (HW 0.74-1.16 mm vs. 0.71-1.07 mm). The minor workers of P. anastasii tend to have more narrow heads posteriorly then P. bilimeki (Fig. 58 vs. Fig. 57) and relatively longer scapes (SI 103-125 vs. 95-108). See Longino and Cox (2009) for additional details.
Adding to the already confusing taxonomy separating P. anastasii and P. bilimeki is the widespread application of the name P. floridana Emery to populations across the southern United States. The first record of P. floridana from Florida was the type series described by Emery from Coconut Grove (Miami area) in 1895. Smith (1930) recorded P. floridana in his original list of Florida ants, and added P. anastasii three years later (1933), stating only "This species [P. anastasii], which was originally described from Costa Rica, is recorded here for Florida on the basis of information secured from Dr. Wheeler…I have seen the same species in greenhouses in the District of Columbia, New Jersey, and Illinois." The previous year (1932) Wheeler, who had received type material of P. floridana from Emery (Wheeler 1908c), included P. floridana and P. anastasii in his own list of Florida ants. Naves (1985) in his study of Florida Pheidole, also recognized both species and distinguished P. anastasii from P. floridana by the matte base of the gaster in the former and the glossy gaster in the latter. Indeed, the type specimens of P. floridana from Coconut Grove are consistent with this characterization (CASENT0904424, CASENT0904425). Naves wrote that the Miami area was the only place where he was able to locate P. floridana. Pheidole anastasii, in contrast, was reported by Naves as widely distributed across the state. Deyrup et al. (1988), lamenting the taxonomic confusion surrounding P. floridana, P. flavens and P. anastasii in Florida, stated, "Traditionally (Creighton 1950;Smith 1979) the name P. floridana has been applied to a widespread upland species that has a distinctive matte area on the base of the first gastral tergite and very evenly rugose head…This is the species we report from the Keys [Florida]." Subsequent reviews of Florida ants have thus excluded P. anastasii from their lists (Deyrup 2003;Deyrup et al. 2000;Moreau et al. 2014). Wilson (2003) followed Deyrup in treating all outdoor populations from the United States as P. floridana, but conceded that his concept of P. floridana could represent a northern geographic variant of P. bilimeki or an endemic species modified by intergradation with a P. bilimeki immigrant population.
With respect to all outdoor North American records, we follow Wheeler (1932), Smith (1933), and Naves (1985) in treating the localized glossy-gaster P. floridana as distinct from the widespread matte-gaster species referred to as P. anastasii by the aforementioned authors. However, the relatively short scapes and posteriorly broad heads of the minor workers, together with the habitat and nesting preferences of the matte-gaster species suggests the name P. bilimeki Mayr more accurately applies to this widespread taxon than does P. anastasii Emery. The issue is discussed in further detail under the P. bilimeki section.
Biology. Pheidole anastasii, named by Emery on behalf of Sig. Anastasio Alfaro, is a Neotropical species that is occasionally found indoors beyond its native range. Although at least some arboreal colonies appear to be polydomous, P. anastasii is a low-impact adventive that has thus far shown little capacity for becoming a significant invader. The biology of P. anastasii, especially across its native range in Costa Rica and in comparison to P. bilimeki was reviewed by Longino and Cox (2009). The species was noted as being among the most abundant ants in the low arboreal forest understory of La Selva Biological Station (Costa Rica). Although tolerant of disturbance, P. anastasii requires some vegetation cover and does not occur in open areas. All collections reviewed by Longino and Cox were from wet forest habitats. Most were from below 500 m elevation, but several ranged to a maximum of 1200 m. The propensity for the species to be inadvertently transported to greenhouses across the world is predicted by its arboreal foraging and nesting habits. Longino and Cox (2009) observed the species nests in almost any kind of cavity or sheltered space, including live stems, and that workers often build galleries and tunnels with carton or earthen construction. The species was reported to occur in lowland second growth, evergreen forest, coffee plantation, limestone, ravine, mixed hardwood-pine forest, wet forest, on karst, and cloud forest. It was also reported to nest in dead sticks and branches on or above the forest floor, under bark flaps on tree trunks, beneath epiphytes and under stones.
Distribution. Pheidole anastasii is a Neotropical native that ranges from Mexico to southern Central America or northern South America. We consider many of the outdoor records of P. anastasii from the southern United States to refer instead to P. bilimeki (see discussion above). There are, however confirmed records of the species from heated indoor locations -especially greenhouses. In North America there are records from hothouses in Washington D.C. and New York (Longino and Cox 2009), and also from Massachusetts. In Europe, the Netherlands occurrences reported as P. anastasii by Boer and Vierbergen (2008) refer to P. bilimeki (Boer, pers. comm.). The records from Denmark and Norway might also refer to P. bilimeki, but until specimens can be examined we follow the authors' use of P. anastasii (Birkemoe and Aak 2008;Lomholdt 1986).
Risk statement. Pheidole anastasii is a synanthropic species with a high tolerance for habitat disturbance. It is occasionally found in human habitations and in greenhouses. There is little indication that is causes significant impact to agricultural systems or native ecosystems. The species is a quarantine risk, and is thought to be transported with fresh plant material.  7); often entirely punctate (Fig. 11), but portions of posterolateral lobes can be glossy. Posterolateral lobes never with distinct rugulae. Promesonotum in profile forming a single dome (Fig. 4). Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively broad; distinctly more than 2× petiolar width in dorsal view (Fig. 31). First gastral tergite with anterior third to entire surface matte. Minor HW 0.42-0.52, HL 0.47-0.59, SL 0.40-0.54, CI 83-93, SI 88-108 (n=38, Longino pers. comm.). Head, including the area mesad of the frontal carinae, entirely covered by reticulated network of punctures, giving it a dull appearance (Fig. 37). Posterior head margin relatively broad and flat (Fig. 57). Antennal scapes lack standing hairs (Fig. 55); surpass posterior head margin by a distance equal to or greater than eye (Fig. 40). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Hairs on mesosoma stout, stiff, of equal length and arranged in pairs (Fig. 53). Postpetiole narrow in dorsal view, only slightly broader than petiole. Gaster with at least anterior 1/3 of first tergite matte (Fig. 33). Identification, taxonomy and systematics. Pheidole bilimeki is a member of the Neotropical P. punctatissima clade, together with P. anastasii and P. punctatissima . Among species treated here, it is easily confused with the aforementioned and members of the P. flavens complex. Minor workers can also be confused with those of P. parva. See section under P. anastasii for identification notes. In the southeastern United States, P. bilimeki is often confused with P. floridana Emery, which is discussed in more detail below. In the Neotropics, there are many native species that closely resemble P. bilimeki (Wilson 2003).

Pheidole bilimeki Mayr
We propose the synonymy of P. lauta Wheeler to be transferred from P. floridana to P. bilimeki. In his original description Wheeler (1908c) wrote, "…the worker has the base of the gaster opaque whereas this is shining in the specimen of floridana given me by Prof. Emery." The description and the photographs we have examined of the type specimens all agree with the concept of P. bilimeki used here and in Longino and Cox (2009).
Should P. floridana therefore be synonymized under P. bilimeki? Wilson (2003) offered that the former might represent the northernmost population of the latter, and recent phylogenetic analyses Moreau 2008) show these two as sibling taxa. Based on the results of her analysis, Moreau (2008) found that her samples of P. bilimeki (Costa Rica, RA0162) and putative P. floridana (Florida, RA0331) were each other's closest relatives, and that this pair was sister to P. anastasii (Costa Rica). The result is also supported by Economo et al. (2015), which found a shallow divergence separating P. bilimeki from putative P. floridana, especially compared to the deep divergence separating these sister taxa from P. anastasii. Moreau (2008) concluded that in order for P. anastasii to be a valid member of P. bilimeki, as proposed by Wilson (2003), P. floridana would also have to be accepted as a synonym of P. bilimeki.
We suggest that this conundrum stems from the common misapplication of the name P. floridana (a shiny gaster species) to collections of what are in fact the North American population of P. bilimeki (a matte gaster species). Naves (1985) came to a similar conclusion in his revision of the Pheidole of Florida, "P. floridana seems to be confined to southeast Florida in the Miami area. This is the only place where I was able to locate this species. Due to its close relationship to P. anastasii the latter has been misidentified as P. floridana many times, thus, mistakenly extending the supposed range of P. floridana. P. anastasii is actually the species widely distributed in Florida, while floridana is absent or at least must be rare in most of the state." One explanation for the confusing phylogenetic results is that RA0331 actually refers to P. bilimeki Mayr, and that true members of P. floridana Emery from the Miami area were not included in the aforementioned phylogenetic analyses. The samples of RA0331 were collected in central Florida from Polk County, well outside the Miami area from which the P. floridana Emery is known (Naves 1985). Deyrup, who collected and identified the specimens of RA0331, has previously (2003;1988;1989) applied the name P. floridana to matte gaster specimens that earlier authors (Naves 1985;Smith 1933;Wheeler 1932) would have considered P. anastasii Emery, and that we consider P. bilimeki Mayr. To properly ascertain the taxonomic status of P. floridana Mayr we suggest a future phylogenetic analysis that includes specimens matching the type material of P. floridana, preferably from the Miami area. If there is evidence supporting the conspecificity of samples matching our concept of P. bilimeki, then the validity of P. floridana Emery must be revaluated. If, rather, the P. floridana samples are heterospecific with respect to P. bilimeki, then there are at least two hypotheses that could explain this result. One is that P. floridana is endemic to Florida. The second, perhaps more compelling albeit ironic explanation, would propose the Miami population of P. floridana is conspecific with a Neotropical species inadvertently introduced to Florida. Miami is a major shipping port and was the gateway for many introduced ants over the past two centuries (Deyrup et al. 2000).
Biology. The taxonomic confusion surrounding whether published accounts refer to our proposed concept of P. bilimeki, or instead to either P. floridana or P. anastasii, makes it difficult to ascertain the natural history of the species. The following account given by Longino and Cox (2009), however, refers definitively to P. bilimeki. They report that P. bilimeki is a common species in open, recently or frequently disturbed habitats. In Costa Rica it occurs in lowland dry forest, lowland wet forest, and montane habitats to about 1500 m elevation. It is a common ant of roadsides, nesting under stones or in dead fence posts. It is a frequent pest ant in houses and is a common ant at baits in second growth dry forest vegetation in seasonally dry Guanacaste Province. It can also be abundant and dominant in large disturbances deep within primary forest reserves. We tentatively treat the account given by Wilson (2003) for P. floridana as referring to the North American population of P. bilimeki. That account stated that winged reproductives have been found in nests during September and October, and that the species occurs in a variety of woodland habitats, nests in soil, litter, and rotten wood, and in both xeric and mesic situations. It also noted the observation of Stefan Cover that colonies are monogynous, may contain 1000 or more ants, and are sometimes polydomous. Cover observed that the species is omnivorous, but does not appear to harvest seeds (but see Naves 1985). Naves (1985) discussed the biology of P. bilimeki (as P. anastasii) in Florida. He found the species most often nesting under the bark at the base of pines or along the roots, but occasionally found it nesting in the soil. The colonies he observed supported over 600 workers with a 5:1 ratio of minors to majors. Mature colonies were monogynous, although in laboratory conditions colonies that lost their original queen would accept other conspecific queens. Several colonies were discovered with two or three founding females, but laboratory experiments found that one would kill the others before the rearing of the first brood. Naves also recorded that the species feeds on seeds, fruits, and scavenges on small dead arthropods and is predaceous on small live arthropods.
Distribution. Pheidole bilimeki is a Neotropical native that ranges from northern South America to southern North America and across the Caribbean. The records included here from the southern United States have previously been treated as P. anastasii and P. floridana (see discussion). Pheidole bilimeki was not reported from Florida until 1932 (Wheeler). While it is possible that the penetration of P. bilimeki into the southern United States represents a recent dispersal event, even one that has been anthropogenically facilitated, there are several reasons for considering P. bilimeki as native to the region. Firstly, the range of North American populations appear contiguous with those of Mexico and the Caribbean, and gene flow among them is probable. Secondly, populations from Florida are known to host two parasites, a mermithid that parasitizes workers, and a hymenopteran parasite species of the genus Orasema (Naves 1985). Pheidole bilimeki has been recorded from greenhouses in Illinois and Ohio in North America. The species has also been found indoors and greenhouses across Europe, including the Netherlands (Boer and Vierbergen 2008), Germany (Forel 1908), Great Britain (Forel 1908), Ireland (Stelfox 1927), and Switzerland (Forel 1908). The only occurrence of P. bilimeki in Jamaica is reported by Wilson (2003). Although the species might occur there, it is also possible that Wilson was referring to P. jamaicensis Wheeler. The single Mauritius occurrence is of a single minor worker examined by Donisthorpe (1946), but this specimen more likely refers to the superficially similar P. parva which is widespread across the island and its neighbors in the Indian Ocean.
Risk statement. Pheidole bilimeki is a synanthropic species with a high tolerance for habitat disturbance. It is occasionally found indoors, especially in greenhouses. There is little indication that is causes significant impact to agricultural systems or native ecosystems.  chi 2001a;Fischer and Fisher 2013). Head square to subquadrate (Fig. 7); rugoreticulate on posterolateral lobes and laterad of frontal carinae (Fig. 13a), but frons dominated by long, well-organized and parallel longitudinal rugae (Fig. 13b). Antennal scrobes indistinct to moderately impressed, but frontal carinae always forming a border capable of accepting the antennal scape (Fig. 13c). Frontal carinae relatively longer, extend 4/5 distance of head before terminating (Fig. 14). Promesonotum in profile with two convexities (Fig. 5), the large anterior dome in addition to a distinct mound or prominence on the posterior slope. Postpetiole not swollen relative to petiole (Fig. 3). Minor HW 0.52-0.63, HL 0.66-0.73, SL 0. 77-0.87, CI 79-88, SI 133-154 (n=16, Eguchi 2001a;Fischer and Fisher 2013). Head predominantly glossy (Fig. 36), lacking punctation or rugulae above eye level. Posterior head margin weakly convex to flat in full-face view (Fig. 45). Antennal scapes long (e.g. Fig. 39), but not surpassing the posterior head margin by more than 2× eye length. Promesonotum in profile with two convexities, the large anterior dome (Fig. 43a) in addition to a distinct prominence on the posterior slope (Fig. 43b). Promesonotal prominence relatively flat (Fig. 49a). Metanotal depression relatively deep (Fig. 49b). Petiole and postpetiole glossy to very weakly sculptured laterally (Fig. 48). Postpetiole not swollen relative to petiole (Fig. 3). Identification, taxonomy and systematics. Pheidole fervens is a medium to large sized species with long limbs. It belongs to the P. fervens clade along with its Australasian congeners P. cariniceps, P. hospes, P. impressiceps, and P. oceanica . The major workers have strong cephalic rugulae that become reticulated towards the posterior of the head and the minor workers have completely glossy heads with very long antennal scapes. Majors and minors of the species can be separated from those of P. megacephala and P. noda by the postpetiole which is not swollen compared to the petiole (Fig. 3), and the promesonotum which has the large anterior dome in addition to a distinct prominence on the posterior slope (Fig. 5, major; Fig. 43, minor). The minors of P. fervens can also be separated from those of P. megacephala by their larger size and longer antennal scapes (Fig. 39). The majors are easily distinguished from P. megacephala by the very sculptured head (Fig. 13).

Pheidole fervens F. Smith
Among species treated here, P. fervens is most easily confused with its close relative, P. indica, and the characters used to separate these two are subtle. For both subcastes, the promesonotal prominence is flatter in P. fervens (Fig. 49a, minor; Fig. 63a, major) compared to that of P. indica (Fig. 50a, minor; Fig. 64a, major). The eyes of P. fervens minors (Fig. 65) are relatively smaller than those of P. indica minors (Fig. 66), especially in comparison to antennal segment 10. The propodeal spines of P. fervens are weaker, narrower, and more downcurved in majors of P. fervens (Fig. 63b) compared to those of P. indica (Fig. 64b). Readers are referred to Eguchi (2004b; for characters used to separate P. fervens and P. indica from their Asian congeners. In the Pacific Island region P. fervens is often confused with the nearly identical P. oceanica, which is native to that region. The carinae between eye and mandible are branching and reticulated in the majors of P. fervens (Fig. 67), versus parallel and not reticulated in those of P. oceanica (Fig. 68). This character was erroneously reversed in the key provided in Sarnat and Economo (2012). The minors are more difficult to separate, but in P. fervens the length of propodeal spine is equal to or less than the diameter of propodeal spiracle (Fig. 69), whereas in P. oceanica it is greater (Fig. 70).
Biology. For such a ubiquitous species across its native and introduced range, very little is known about the biology of Pheidole fervens. It is a synanthropic species with a high tolerance for disturbance (Eguchi 2004b;Fischer and Fisher 2013;Martínez 1996), but can also thrive under some degree of canopy cover (Morrison 1996;Sarnat and Economo 2012). In Fiji, where it is likely a recent colonizer, it was collected most frequently in human dominated landscapes between 0-800 m, although several collections were also made from primary forest at low elevations. In Hawaii, where it is definitely an introduced species, it is more abundant locally in wet regions than P. megacephala (Gruner et al. 2003) and occurs in the hot lowlands only below 900 m (Reimer 1994). In the Philippines, P. fervens is found in irrigated lowlands (rice fields) where it is characterized as dominant species capable of displacing S. geminata in the dry season (Way et al. 1998). In Japan it occurs in open land grading to forest edge (Harada et al. 2009;Ogata 1981). Pheidole fervens recruits in large numbers to bait and forages both on the ground and on vegetation . Baiting experiments on Pacific Islands found that P. fervens can act as a numerically and behaviorally dominant species capable of excluding other invasive ant species (including Anoplolepis gracilipes, Nylanderia bourbonica, and Tetramorium bicarinatum) from baits (Morrison 1996). Although foragers can be slow to discover food resources, once found they can recruit in large numbers and displace competing species (Morrison 1996). Experiments in China suggest that P. fervens can provide some degree of biotic resistance to the Red Imported Fire Ant (Solenopsis invicta) by acting in groups to dismember the limbs of individual fire ants (Chen et al. 2011). Martínez (1996) suggested the California population of P. fervens was polydomous, and Passera (1994) suggested the Hawaii population is unicolonial and polygynous, but detailed colonylevel studies of the species are required to verify these claims. Wittenborn and Jeschke (2011) attributed their assertion that P. fervens practices dependent colony founding to Harris et al. (2005a), but we were unable to find any reference to colony foundation in that report and cannot substantiate their evidence.
Distribution. We consider Pheidole fervens as native to a broad expanse of the Indo-Malay region spanning from India east to the Philippines and south to the islands west of New Guinea. This is a broad and admittedly arbitrary boundary, but a more precise circumscription of the native range requires a population-level analysis outside the scope of the present study. In particular, it is difficult to ascertain the extent of its range into the Pacific Island region prior to the Anthropocene. The only known occurrence of P. fervens from New Guinea was a single record from the westernmost part of the island (Emery 1887b). East of New Guinea, however, the species is established on nearly all islands of the Pacific, including those which were uninhabited by any ant prior to human arrival. Although it is quite possible that P. fervens reached some of these islands without human assistance -especially those between Taiwan and mainland Japan -we treat these as introduced populations. And although established on Mauritius, the species is rarely encountered there and is currently known from only two localities (Fischer and Fisher 2013). The only record of introduction in North America is a California population that established nests in cracks of roads and along the sides of buildings in a two-block area of downtown Los Angeles (Martínez 1996). Pheidole fervens has been collected from greenhouses in the Netherlands (Boer and Vierbergen 2008), and is frequently intercepted by quarantine inspections (Ward et al. 2006).
Risk statement. Pheidole fervens can be a dominant species where it is locally abundant. Although few studies have measured the effect of P. fervens on native ecosystems, we predict that it could negatively impact native arthropods. We were unable to find documentation on the effect of P. fervens on agricultural systems, but it can be among the most abundant ant species in irrigated lowland crop systems such as rice fields. Pheidole fervens can also be an indoor nuisance species (Wilson and Taylor 1967), but is not a risk for structural damage. According to New Zealand records, the species is among the most commonly intercepted ants in that country (Ward et al. 2006). Sixty-nine percent of the interceptions were in freight from Fiji (> 92% from the Pacific Islands). Interceptions were mostly in fresh produce (69%) and cut flowers (8%). Pheidole fervens was also intercepted multiple times in air passengers' luggage and shipping containers. The species could become more globally widespread in the future.  Naves, 1985: 62, figs. 21, 45, 57 (s.w Wilson (2003) is now known to be polyphyletic Moreau 2008). Readers are referred to the P. flavens-complex for additional discussion of identification, taxonomy and systematics. The taxonomy of P. flavens and its close relatives remains in a state of confusion. It is beyond the scope of the present study to resolve this issue, but we contribute the following discussion as a step towards that goal.

Pheidole flavens Roger
Pheidole flavens was originally described by Roger from Cuba, but the type material is considered to be lost. Wilson (2003) designated a neotype from Cuba and synonymized a total of eight nominal taxa with P. flavens. Of these, P. greggi Naves (Florida) and perhaps P. flavens st. spei Santschi (Mexico) are most similar to the Cuban neotype. They, together with the types of P. moerens subsp. creola, are the only specimens examined thus far that have clearly reticulated rugulae posterior to the scrobes of major workers. Naves ' (1985: fig. 55) concept of P. flavens Roger, at least as evidenced by his figures and descriptions, more closely matches our concept P. navigans, a species that is spreading across the southeastern United States. The syntype major of Pheidole flavens var. vincentensis Forel differs substantially from the neotype in that the head is completely glossy between the rugulae, which are themselves entirely longitudinal and do not extend far beyond the maximum extent of the antennal scapes in repose. These characters make it at least superficially more similar to P. moerens and P. navigans. Pheidole flavens r. gracilior and P. navigans were both described by Forel from workers intercepted at a Hamburg quarantine facility, which is testament to the dispersive ability of this complex. The syntype major of the latter species and that of P. floridana subsp. aechmeae Wheeler, also described from Mexico, are quite similar. Pheidole exigua var. tuberculata Mayr has the strongly convex head and promesonotal dome of P. exigua Mayr, and also exhibits tuberculate angles on the mesonotal declivity. Type specimens of P. flavens var. haytiana Forel were not examined for this study.
The only material from outside Central America and the Caribbean that we were able to confirm as matching the Wilson's neotype was from Florida. The Florida populations referred to here as P. flavens and P. navigans are almost certainly heterospecific. We suspect that Nearctic records of P. flavens outside of Florida such as those reported from Louisiana (Colby and Prowell 2006;Dash and Hooper-Bùi 2008) refer to either P. bilimeki or the species we are treating as P. navigans in the southeastern USA.
Biology. The biology of Pheidole flavens, as currently conceived, was reviewed by Wilson (2003) with contributing observations by Jack Longino. The species prefers rotting wood, but also nest beneath the bark of trees, in dead knots on tree trunks, in sod on rocks, in the soil beneath stones, and in epiphyte masses. In the Caribbean it is recorded from forests and thickets from sea level to 900 m, and in Costa Rica it occurs in both wet and dry forests below 1000 m. The nest galleries are diffuse and irregular. Mature colonies are large containing up to thousands of workers. Workers collect small arthropods and will recruit to sugar baits.
Distribution. Pheidole flavens is among the most widespread and abundant species of its genus in the New World, although this range might be representative of multiple cryptic species. As currently conceived, however, we consider P. flavens native from southern Mexico east through the Caribbean and south to Uruguay and northern Argentina. It is difficult to know whether the disjunction separating the western and eastern regions of South America is accurate or a sampling artifact. The Florida population is believed to have derived from an accidental introduction by commerce (Deyrup et al. 2000;Wilson 2003).

Risk statement.
Pheidole flavens (or at least it's very close relatives) are easily transported long distances, and are known to hitchhike with fresh plant material (Wilson 2003). However, the species is not known to cause significant impact to agricultural systems or native ecosystems, and is not considered a house pest (Hedges 1998;Klotz et al. 1995). Fig. 88e The P. flavens-complex is defined here to include P. flavens Roger, P. moerens Wheeler, P. navigans Forel, and their respective junior synonyms. A clear understanding of the phylogenetic relationship among the aforementioned taxa that are invading regions beyond the Neotropics remains a challenge for future studies (Sarnat et al. 2014).

Pheidole flavens-complex
Diagnosis among introduced Pheidole. Color variable. Major Head subquadrate (Fig. 7). Longitudinal carinae extend from anterior frons margin a variable distance beyond frontal carinae, but never reach posterior head margin (Fig. 25). Rugae of posterolateral lobes variable from mostly absent, to predominantly longitudinal, to distinctly reticulated. Posterior head margin always free of distinct rugae (Fig. 25) or rugoreticulum (Fig. 27). Microsculpture of posterolateral lobes variable from glossy to moderately punctate. Hypostoma with stout median and submedian teeth (Fig.  19). Promesonotal dorsum usually with distinct transverse striae (Fig. 21), but sometimes lacking distinct striae. Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope. Promesonotum not strongly transverse with strongly projecting sides in dorsal view (Fig. 29). Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively narrow in dorsal view; distinctly less than 2× petiolar width. Gaster with entire first tergite glossy (Fig. 32). Minor Head covered in punctate microsculpture, giving it a dull appearance. Posterior portion of head lacking many short to medium length segments of striae distinctly interlaced among punctate ground sculpture (Fig. 60). Antennal scapes often, but not always, surpass posterior head margin; if they do it is usually by a distance less than eye length. Antennal scapes with standing hairs present (Fig. 56). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Hairs on mesosoma fine and flexuous, not arranged in pairs (Fig.  54). Pronotal humeri not angular. Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively narrow (Fig. 30); distinctly less than 2× petiolar width in dorsal view. Gaster with entire first tergite glossy (Fig. 32). Identification, taxonomy and systematics. Members of the P. flavens-complex are small species ranging from yellowish to dark reddish brown. The head and mesosoma of the minor workers are covered by densely punctate ground sculpture. The head of the major worker tends to be shinier with the posterior margin always free of sculpture. Among the species treated here, those of the flavens complex are most easily confused with those of the closely related and often sympatric P. punctatissima clade (P. anastasii, P. bilimeki, P. punctatissima). The postpetiole is narrower in the P. flavens complex (Fig. 30, major; Fig. 61, minor) relative to those of the P. punctatissima clade (Fig. 31, major; Fig. 62, minor). The gaster is completely glossy in both subcastes of the P. flavens complex (Fig. 32), while at least the basal portion of the first gastral tergite is matte in those of the P. punctatissima clade. In the P. flavens complex the minors have finer hairs of variable lengths on the mesosoma (Fig. 54) and the antennal scapes have many erect to suberect hairs (Fig. 56). In contrast the P. punctatissima clade have thicker mesosoma hairs of equal length (Fig. 53) and lack erect antennal scape hairs (Fig. 55). The minors of the P. flavens complex are very difficult to distinguish from those of Pheidole parva. They can be separated by the lack of interlacing striae on the posterior head margin (Fig. 60, P. flavens complex vs. Fig. 59, P. parva). See key for additional characters.
The P. flavens group as defined by Wilson (2003) is now known to be polyphyletic (Moreau 2008), and unpublished data analyzed by the authors suggests that the P. flavens complex as defined by Wilson also lacks monophyly. Pheidole exigua is morphologically quite similar to the aforementioned taxa, and future attempts to define the flavens complex clade should include it in analyses, along with P. glomericeps and possibly other species not initially considered by Wilson. The most recent phylogeny of Pheidole includes eight taxa that form a well-supported P. flavens clade . The clade consists of taxa that have been determined by various ant taxonomists as P. moerens, P. flavens, P. glomericeps; several morphospecies including Pheidole sp. JTL-177 and a P. flavens-complex taxon recently established on Vanuatu (Sarnat et al. 2014); and also the species we refer to as P. navigans Forel. Although these taxa represent only a fragment of the diversity attributed to the P. flavens complex, the analysis demonstrates the taxonomic confusion of the clade. For example, the P. moerens sample from the Dominican Republic is most closely related to the taxa recently discovered on Vanuatu in the South Pacific. There is strong support for these taxa being more closely related to Pheidole JTL-177 (Venezuela) and two taxa from Central America determined as P. flavens (collection codes PSW16014 and JTL4928) then to P. navigans from Alabama (collection code PSW15833) and Venezuela (collection code PSW16167). The Alabama and Venezuela P. navigans are actually most closely related to a taxon determined as P. glomericeps (collection code Wa-D-01-2-16).  . 7); rugoreticulate on posterolateral lobes and laterad of frontal carinae (Fig. 13a), but frons dominated by long, well-organized and parallel longitudinal rugae (Fig. 13b). Frontal carinae extend 3/4 distance of head before terminating (Fig.  15). Antennal scrobes indistinct to moderately impressed, but frontal carinae always forming a border capable of accepting the antennal scape (Fig. 13c). Hypostoma with weakly produced median tooth and submedian teeth. Promesonotum in profile with two convexities (Fig. 5), the large anterior dome in addition to a distinct mound or prominence on the posterior slope. Postpetiole not swollen relative to petiole (Fig. 3).

Pheidole indica Mayr
Minor HW 0.50-0.65, HL 0.60-0.74, SL 0.64-0.81, CI 72-90, SI 120-149 (n=20). Head predominantly glossy (Fig. 36), lacking punctation and or rugae above eye level. Posterior head margin weakly convex to flat in full-face view (Fig. 45). Antennal scapes long (e.g. Fig. 39), but not surpassing the posterior head margin by more than 2× eye length. Promesonotum in profile with two convexities, the large anterior dome (Fig.  43a) in addition to a distinct prominence on the posterior slope (Fig. 43b). Promesonotal prominence relatively convex (Fig. 50a). Metanotal depression relatively shallow (Fig. 50b). Petiole and postpetiole glossy to very weakly sculptured laterally (Fig.  48). Postpetiole not swollen relative to petiole (Fig. 3). Identification, taxonomy and systematics. Pheidole indica is a medium to large reddish brown species with relatively long limbs. It belongs to the P. fervens clade along with its Australasian congeners P. cariniceps, P. fervens, P. hospes, P. impressiceps, and P. oceanica (Economo et al. 2015, unpublished data). The major and minor workers are distinguished from those of P. megacephala by the lack of a swollen postpetiole (Fig. 3). The majors are also easily separated from those of P. megacephala by the strongly sculptured head (Fig. 13). The minors can be confused with those of P. megacephala because both have glossy heads. However, the minors of P. fervens can be separated from those of P. megacephala by the relatively longer antennal scapes (Fig. 39 vs. Fig. 40) and the presence of a promesonotal prominence (Fig. 43 vs. Fig. 42). Pheidole indica is broadly sympatric with P. noda and P. fervens. It is easily separated from the former by the lack of a swollen postpetiole (Fig. 3 vs. Fig. 2). Separation from P. fervens is quite difficult, and readers are referred to corresponding section under that species for distinguishing characters. Readers are referred to Eguchi (2004b; for characters used to separate P. indica and P. fervens from their Asian congeners. Pheidole indica was originally described from India. Eguchi (2004b) synonymized several other Asian congeners under P. indica and discussed taxonomic differences used to distinguish it from P. fervens and other morphologically similar species. We synonymize P. teneriffana under P. indica based on morphological analysis of the type specimens and genetic analysis of previously determined specimens (unpublished data). Forel, in his original description of P. teneriffana, noted the similarity between it and P. striativentris [= indica].
The biogeographical origin of P. teneriffana has been a minor mystery of the past century, as revealed by the recent review of the species by Wetterer (2011). There appeared to be general consensus that P. teneriffana was native to at least some portion of North Africa, Arabia, the Middle East or the Mediterranean. Santschi (1918), suggested the upper Nile area (South Sudan). Wilson (2003) suggested North Africa and potentially the Canary Islands. Collingwood et al. (2004) suggested it was native throughout northern Africa and observed it to be, "spreading over a wide front in the Middle East, Arabia and the Mediterranean countries." Wetterer (2011) found the distribution of P. teneriffana enigmatic, "Curiously, most Old World records of P. teneriffana are subtropical, but all New World records are tropical, except one from California…If P. teneriffana is truly native across North Africa, it is remarkable how few records I found from any North African country other than Egypt." Biology. In Asia P. indica is known to nest in soil or under stones in open and dry habitats (Eguchi 2004b). It is among the most widespread Pheidole species in Asia. In the Caribbean Wetterer (2011) found P. indica [as P. teneriffana] almost exclusively on beaches and at highly disturbed urban sites, particularly in waterfront areas. In northern Africa, Santschi (1908) noted the tramp-like distribution of what he treated as P. teneriffana, "This species, described by Forel on samples from the Canary Islands, was sent to me from Cairo. I discovered it most recently in Sousse [Tunisia], in the park, near the port. As it does not exist in the interior, I think it is one species cosmopolitan tendencies. It nests in the ground and under stones." Santschi (1934) later reported the species from Alexandria, Egypt, and noted that P. teneriffana was rarely reported far from seaports. Collingwood et al. (1997) reported that in the United Arab Emirates, P. indica [as P. teneriffana] was populous in irrigated gardens and along the coast where it appeared to be spreading rapidly, possibly to the detriment of local species. The species has also been reported from urban areas of the Balearic Islands where it is common in the gardens and trees and on sidewalks near the harbor (Gómez and Espadaler 2006). Fischer and Fisher (2013) reported P. indica [as P. teneriffana] from the Malagasy region. It was collected on the Comoros, Mauritius, the Seychelles, and from coastal towns in Madagascar, usually from under stones, ground nests, or foraging on the ground or lower vegetation in urban or garden habitats at elevations below 300 m. It was also found on Mayotte in native littoral and secondary forest below 10 m.
Perhaps the most detailed study of P. indica in the New World comes from the account of Martínez (1992) who reported a vigorous population, represented by a putatively single polydomous colony spanning several hectares, that was discovered in Long Beach, California in 1989. Martínez (1992) reportedly observed 23 inseminated queens from a single colony that was changing nest sites (although no details are given for how he knew the queens were inseminated). He described the colony nests as low mounds on the soil, along curbs or sidewalks, at the edges of lawns, in cracks in pavement, and at the bases of trees. New colonies were started by budding. Workers foraged night and day unless temperature exceeded 26 °C, taking seeds and scavenging dead or dying insects. They were observed feeding on sweet or greasy foods, but were not seen tending aphids. Martínez (1992) observed the species attacking native ants, including Pogonomyrmex californicus (Buckley). More remarkably, he reportedly observed P. indica destroying colonies and taking over nest sites of Linepithema humile. Despite the purported success of these battles, P. indica must have lost the larger war against L. humile, as the eventual extirpation of the Californian population was attributed to the Argentine ant (Gulmahamad 1999).
Distribution. We treat all occurrence records from the regions of Indomalaya west of the Korean Peninsula as native. The Korean and Japanese populations are considered introduced (Choi and Bang 1993;Choi et al. 1993a;Choi et al. 1993b;Terayama 1992), and additional portions of the range in Asia might also have resulted from anthropogenic transport. Pheidole indica has been introduced to scattered localities across the globe, although the vast majority of these records were attributed to its junior synonym, P. teneriffana. Introduced populations have been reported from the Mediterranean, northern Africa, the Malagasy region, Western Australia, Peru, the Caribbean, and southern California.
Risk statement. Pheidole indica is not considered to be a major pest to either agriculture or native ecosystems. Although the species is tolerant of disturbed and urban areas, we found no reports of it infesting structures. Few studies have measured the effect of P. indica on ecosystem health, but we predict that it could negatively impact native arthropods. The species is continuing to spread across the globe and further studies are required to test its ecological and agricultural impact outside its native range.  Fischer and Fisher 2013). Head heart-shaped (Fig. 6); posterior 1/3 of dorsal surface smooth, glossy and entirely lacking rugoreticulate sculpture. Hypostoma lacking distinct median and submedian teeth. Promesonotum in profile forming a single dome (Fig.  4), lacking a distinct mound or prominence on the posterior slope. Postpetiole with a posterodorsal (Fig. 1a) and anteroventral (Fig. 1b) bulge. Minor HW 0.50-0.61, HL 0.57-0.68, SL 0. 61-0.72, CI 86-92, SI 114-122 (n=20, Fischer and Fisher 2013). Head predominantly glossy (Fig. 36), lacking punctation and or rugae above eye level. Antennal scapes surpass posterior head margin by approximately same length as eye (Fig. 40). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Postpetiole with a posterodorsal (Fig.  1a) and anteroventral (Fig. 1b) bulge. Identification, taxonomy and systematics. Pheidole megacephala is a medium sized species of variable color that is most easily recognized outside of its native range by the heart-shaped head and bulging postpetiole. It belongs to a diverse and taxonomically confusing clade of morphologically similar taxa centered in the Afrotropical and Malagasy regions. Both major and minor workers are distinguished from all other introduced Pheidole by the swollen shape of the postpetiole (Fig. 1). Pheidole noda also has a swollen postpetiole, but whereas the postpetiole of P. megacephala is characterized by a posterodorsal and anteroventral bulge, that of P. noda is formed as a high dorsally bulging dome that is tallest at its midpoint.
Pheidole megacephala has often been confused for P. pallidula Nylander in Europe, especially in the Mediterranean region. The introduced populations of P. megacephala can be distinguished from P. pallidula by the following characters. For both major and minor workers the postpetiole of P. megacephala has a posterodorsal (Fig. 1a) and anteroventral (Fig. 1b) bulge, while that of Pallidula is not swollen relative to petiole (Fig. 3). The propodeal spines of both subcastes are distinct in P. megacephala but are strongly reduced in P. pallidula. Additionally, the major worker of P. megacephala has a heart shaped head that broadens significantly posterior to eye-level (Fig. 6) while the head of Pallidula is more rectangular (more approximate to Fig. 7).
Accurate identification within the Afrotropics is more problematic. While for Madagascar previously described subspecies have been synonymized with P. megacephala (Fischer and Fisher 2013), the taxonomy of the megacephala group in Africa remains rather chaotic with a number of unrevised subspecies, most of which remain insufficiently characterized. In a taxonomic overview of the group, Emery (1915) studied type and non-type material of P. megacephala-related species, yet for several subspecies he was not able to define clear species limits from the multitude of different, yet highly similar, phenotypes. We suspect that some of those names are probably due to intraspecific variation within P. megacephala and P. punctulata Mayr. Other, morphologically unique taxa like P. megacephala nkomoana Forel are clearly valid biological species. However, without a comprehensive taxonomic treatment supported by a robust phylogeny, the following species characterizations may be subject to future taxonomic changes.
Within the megacephala group, minor workers are difficult to separate morphologically and thus have only limited use for species identification, but the majors tend to be more distinct in their morphologies and can be separated by differences in head and body shape and sculpture, and in size and pilosity, although the limits are often unclear and characters are sometimes distributed along a continuum rather than being separated into distinct, clear-cut states.
Major workers of P. megacephala melancholica Santschi are characterized by presence of weak punctures on the majority of the head, including the sides in lateral view, promesonotum with punctures and irregular transverse rugulae, and moderately abundant short and stout standing hairs on head and body, whereas major workers of P. megacephala entirely lack punctures on the posterior 1/3 of the head, have a mostly smooth and glossy promesonotum, and often possess longer, more flexuous standing hairs, which often branch at the tips. Pheidole megacephala nkomoana majors are characterized by a weakly defined antennal scrobe and relatively long frontal carinae that reach about ¾ towards the posterior head margin, two well-defined submedian hypostomal teeth, a weak prominence on the promesonotal dome, and very long, flexuous standing hairs on the dorsal promesonotum. Also the spines tend to be shorter than in P. megacephala, in length almost equal to the diameter of the propodeal spiracle. Both subspecies have been described from and collected in western African forests. Another closely related species to P. megacephala is P. punctulata. It is very widespread in sub-Saharan Africa and usually found in dry forests and grassland habitats. Morphologically close to P. megacephala, its major workers can be distinguished by their often enlarged and strongly heart-shaped heads, the presence of a softly or superficially punctuated sculpture on parts of the head dorsum, promesonotum, postpetiole and gaster, and relatively uniform, short and stout, erect hairs covering the body. Minor workers tend to be slightly larger and more robust than in megacephala, often with a few oblique carinae present between the eyes and the mandibles and reaching the posterior eye level, the hairs similar as in major workers and usually more abundant than in P. megacephala.
Morphologically very similar to P. punctulata are P. megacephala ilgi Forel, megacephala impressifrons Wasmann, and megacephala rotundata Forel. Like P. punctulata, they are usually found in drier forest and grassland habitats and their workers seem to be highly polymorphic, which means that in addition to normal major workers, colonies are capable of producing so-called supermajors. These supermajors possess a very strongly heart-shaped head, which can be disproportionately big compared to the size of the mandibles and the rest of their bodies. As Emery (1915) stated for P. megacephala rotundata, on first glace they look quite distinct from P. punctulata, but at closer examination of series with different major worker sizes it seems impossible to define species limits. From our own observations it seems likely that these subspecies are a result of sampling bias and phenotypic variation within P. punctulata, rather than historic speciation events (Fischer et al., in preparation). Incomplete sampling can also be a problem when only smaller major and minor workers are collected, which are often very similar to those of P. megacephala, with very similar head sculpture and general morphology.
In the Malagasy region, P. megacephala can be confused with three other species: Pheidole punctulata spinosa Forel, which, on average, has longer spines, a slightly higher propodeum and a more extensively smooth and glossy posterior portion of the head in the larger major workers. Pheidole megatron, which was described from the Comoros and is possibly present in the Northwest of Madagascar as well, is characterized by major workers with a less heart-shaped, and slightly more rectangular head shape, and sometimes sculpture and rugulae present on the posterior head portion (see Fischer and Fisher 2013). Finally, P. decepticon, described from Mayotte and distributed over several of the smaller Southwest Indian ocean islands, is characterized by possessing a denser, more prominent and longer pilosity as well as slightly smaller, less rounded ventral bulges on the postpetiole in both minor and major workers (see Fischer and Fisher 2013). It is however possible that P. decepticon is a geographic variation of and conspecific with P. punctulata spinosa.
Biology. Pheidole megacephala is listed among the top five invasive ants (Lowe et al. 2000). Although this species prefers humid and disturbed habitats where it is usually found in very high abundances (Burwell et al. 2012;Hoffmann et al. 1999;Wilson 2003), it can generally be found in a large variety of landscapes, from coastal habitats to human settlements and plantations in lower elevations, degraded dry forest, to mid-elevation rainforest or even montane forest -in Papua New Guinea up to 2150 meters altitude (Fischer and Fisher 2013). The distribution range and activity of P. megacephala appears to be somewhat limited by susceptibility to desiccation and higher temperatures. Thus, colonies are often found in more humid microhabitats, and workers tend to forage inside the leaf-litter and at night, or even build covered trails (Greenslade 1972, personal observations). However, some studies reported that on smaller islands or after successful introduction in a new area, P. megacephala expanded its range and invaded into the forest interiors where it attacked and displaced other introduced and natively occurring ant species (Burwell et al. 2012;Hoffmann 1998). In a citrus orchard in Tanzania for example, P. megacephala was able to partly displace highly territorial and competitive Oecophylla weaver ants (Seguni et al. 2011). Pheidole megacephala is an especially common and abundant nuisance and pest on islands, which are generally more strongly impacted by invasions of alien species.
Part of the success of P. megacephala as a pantropic pest species is its generalist behavior. Like many other Pheidole species its diet is broadly omnivorous with a large proportion of its food probably acquired by scavenging on the ground. Pheidole megacephala is also a good predator with an efficient nest mate recruitment that enables the species to dominate baits and to retrieve prey too large for single workers to carry (Dejean et al. 2008;Dejean et al. 2007). Devastating effects on the abundance and diversity of native invertebrates, in northern Australia for example, are well documented (Hoffmann 1998;Hoffmann et al. 1999;Hoffmann and Parr 2008). Pheidole megacephala has also been documented to negatively impact agricultural systems. Workers tend plant and crop-damaging scale insects for honeydew (Campbell 1994;Gaigher et al. 2011;González Hernández et al. 1999;Greenslade 1972;Petty and Tustin 1993;Reimer et al. 1993), protect plants with extrafloral nectaries from phytophagous insects and possibly collect seeds (Hoffmann 1998). A recent study experimentally evaluating the performance in interference competition found that P. megacephala ranked lowest among seven of the world's worst most destructive invasive ant species (Bertelsmeier et al. 2015). The authors, citing Dejean et al. (2008) suggested that P. megacephala does not dominate invaded ant communities through direct physical interactions (interference competition) but by raiding their colonies.
Nesting sites are variable and can occur in any crack and crevice that is large enough for them to enter, including soil, inside rotting logs, under rocks, in houses or in tree bark. As in several other invasive ant species, colonies are polygynous, and dependently founded via budding, with nests in large areas often forming supercolonies (Hoffmann 1998) that aggressively fight other ants or outcompete them by depleting their prey and other resources (Dejean et al. 2008;Fournier et al. 2009;Hoffmann et al. 1999;Vanderwoude et al. 2000).
Distribution. Pheidole megacephala is a cosmopolitan species that has established across the globe as a household and agricultural pest throughout the tropics. Wetterer (2012) provided a detailed review of the worldwide spread of P. megacephala, and cites Wheeler's statement (Wheeler 1922a) that it is most likely of Afrotropical or Malagasy origin, the only two regions with a diversity of related species ("subspecies and varieties"). Theoretically it is possible that a common ancestor of P. megacephala and the Malagasy endemics P. punctulata spinosa, P. megatron and P. decepticon arrived on the islands in prehistoric times, diversified there, and that P. megacephala was later transported to all other regions including Africa only after the arrival of humans. But the distribution of P. megacephala on Madagascar strongly resembles the distributions of other invasive species on the island -e.g. those of Monomorium floricola, M. pharaonis, Tapinoma melanocephalum, Technomyrmex albipes, Trichomyrmex destructor. While P. punctulata spinosa has established a broad distribution range across the island's variable habitats and elevations, P. megacephala, like the other invasives, is found mostly along the coast, in low elevation and disturbed habitats or near human settlements.
Similar to Wheeler's observation, our argument for the "out-of-Africa" hypothesis is an overall much higher complexity in different morphotypes and species-level diversity in African megacephala group taxa and the presence of both, very closely, but also more distantly related taxa (e.g Pheidole aurivillii Mayr). For these reasons and for the purposes of this study, we consider all records from Africa to represent the native range of Pheidole megacephala. However, a further resolution will require a comprehensive phylogeographic study of the species and its allied taxa, especially from the poorly studied and sampled African region.
Populations of P. megacephala recorded from the southwestern extent of the Arabian Peninsula are treated as native as this region is commonly considered as belonging to the Afrotropics. However, recent studies on generic distributions of global ant diversity that find little support for including any portion of the Arabian Peninsula in the Afrotropics (unpublished data). Until robust phylogeographic data is available for P. megacephala, this decision must be considered tentative and open to future revision.
We do agree with Wetterer's (2012) conclusions that records of P. megacephala from Mediterranean Europe northward are either temporary indoor records or misidentifications of P. pallidula. Outside of Africa, the Malagasy region and the range of P. pallidula (western Palearctic), P. megacephala is easily recognized as it does not cooccur with species of similar morphology. We therefore consider all records reviewed from outside the aforementioned regions as confirmed unless otherwise stated.
Dubious records. The following records are considered dubious mostly because there is reason to believe they represent misidentifications of P. pallidula. However, it is possible that some of the following literature records were based on accurate identifications, but that P. megacephala was since extirpated from the referenced localities. This latter possibility is plausible especially for the Mediterranean region where L. humile has established a stronghold. For example, (Heer 1852) described Oecophthora pusilla (=P. megacephala) as ubiquitous on the island of Madeira, "In the town of Funchal there is probably not a single house that does not harbor millions of the tiny creatures…" Less than a century later Wheeler (Wheeler 1927b) reported, "Now it is an interesting fact that the Argentine ant, soon after its arrival in Madeira, completely replaced the Pheidole as a house ant." Similar instances of well-established populations of introduced ant species becoming locally extirpated have been documented (Moreau et al. 2014;Wetterer 2006).
ALGERIA: The material referred to by André (1883) P. megacephala is distinguished by that author from P. pallidula only by the difference in size of the propodeal spine, and was otherwise observed to be identical. Considering the other characters separating these two species discussed earlier, we tentatively consider this record to be a misidentification of P. pallidula. CROATIA: The material listed from this country (Petrov and Collingwood 1992;Petrov and Legakis 1996) is considered to refer to P. pallidula according to Bračko (2006). EGYPT: Egypt is the type locality of the nomen oblitum Formica edax Forskål. Emery (1892) wrote that edax is undoubtedly a small Pheidole, and possibly refers to P. megacephala. Dalla Torre (1892) was also uncertain as to which species (or even genus) the name edax referred to. Given the uncertainty of these two authors, the occurrence of P. pallidula in Egypt and the unconfirmed single literature record of Bakr et al. (2007), it is difficult to know when P. megacephala was first reported from Egypt. FRANCE: Bignell (1901) reported the ant species listed in his study of Corsica were identified by Saunders, who is known to have confused P. pallidula for P. megacephala. As P. pallidula was not listed in the publication, we consider the record to either be a misidentification of that species or from an extirpated population. GREECE: The only primary references to an outdoor occurrence we could confirm are Collingwood (1993) and Borowiec and Salata (2012). The former authors reported P. megacephala was found only once during their study of five Greek islands on the threshold of a small hotel in Pigadhia on Karpathos. The second study reported finding the species on a road in Crete. The record from Macedonia in (Karaman 2011) is from material identified by Petrov. We tentatively follow (Bračko 2006) as treating this as a misidentification of P. pallidula. ITALY: Piedmont is the type locality for Myrmica trinodis Losana which was synonymized with megacephala by Roger in 1863. Losana also lists a M. megacephala Latrielle in the same publication. Latrielle never described any species by the name megacephala, however. Losana might have instead been referring to Messor megacephala Leech (= Messor barbarous Mayr). Regardless, the original description of M. trinodis states that the species was collected from outdoor gardens. There is some reason to suspect this name might refer instead to P. pallidula, as the only verifiable occurrences of megacephala in Italy since are for specimens collected from plant nurseries, greenhouses and cargo hangars used for holding imported plants, fruits and vegetables (Jucker et al. 2008;Limonta and Colombo 2003). MOROCCO: Saunders (1888) appears to be the only primary reference for P. megacephala occurring in Morocco, but it is likely that the author was referring to misidentified material of P. pallidula (Wetterer 2012). This view is further evidenced by Cagniant and Espadaler (1993) who were unable to find the species in their survey. SPAIN: We consider the following records from the Balearic Islands and Gibraltar to refer to P. pallidula (Saunders 1888;Saunders 1904;Walker 1889). USA: The specimens reported in Fischer and Fisher (2013) from Arizona were from a quarantine collection intercepted from Florida, and there is no reason to believe the species has ever established in Arizona. Wetterer (2012) cited a specimen record of P. megacephala from Catalina Island (California). If the identification proves accurate, it is the only known record from that island and the population has since been extirpated (perhaps by L. humile). However, a population (CASENT0248690) has been discovered recently in southern California (Orange Co.). Although P. megacephala is listed in the Missouri Ants web page (2015), we cannot verify the entry with any specimen or literature record.
Risk statement. Pheidole megacephala is known as a major agricultural and ecological pest species (Williams 1994) and its widespread pantropic distribution and often very close association with humans make it a high-risk invasive species with a serious potential for ecological, agricultural and economic damage. In Ward et al. (2006) it has been the most intercepted exotic ant species (890 out of 4355 interception records between 1955 and 2005) arriving with trade products in New Zealand. Many aspects of its biology indicate that it is highly adaptable and thus able to survive outside of its preferred habitat, by finding suitable microhabitats for nesting and by killing or outcompeting native species. Although mutualistic relationships with scale insects and other crop pests are dominant in agricultural systems with introduced P. megacephala, positive side-effects on plant fitness have been observed as well (Bach 1991).  Smith 1967, Wojcik 1975, Glancey 1976, Naves 1985, Deyrup 1988, Deyrup 2000, Dash 2008, MacGown 2010, Guénard 2012 (Fig. 7). Longitudinal carinae of the frons extend to approximately an eye's length distance from the posterior head margin (Fig. 25). Rugae of posterolateral lobes predominantly longitudinal. Posterior head margin always free of distinct rugae (Fig. 25) or rugoreticulum (Fig. 27). Microsculpture of posterolateral lobes glossy to weakly punctate. Antennal scrobe distinct and narrow, shallow but capable of receiving the entire antennal scape in repose (Fig. 71a); bordered by strong, unbroken frontal carina mesially (Fig. 71b); depression marked by a continuous smooth surface entirely (or nearly entirely) uninterrupted by rugulae. Hypostoma with stout median and submedian teeth. Promesonotal dorsum with distinct transverse striae (Fig. 21). Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope. Promesonotum not strongly transverse with strongly projecting sides in dorsal view (Fig. 29). Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively narrow in dorsal view; distinctly less than 2× petiolar width (Fig. 30). Gaster with entire first tergite glossy (Fig. 32). Minor HW 0.40-0.45, HL 0.45-0.50, SL 0.40-0.44, CI 86-92, SI 96-102 (n=8). Head covered in punctate microsculpture, giving it a dull appearance (Fig. 37). Antennal scapes reach or weakly surpass posterior head margin; if they do it is usually by a distance less than eye length. Antennal scapes with standing hairs present (Fig. 56). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Hairs on mesosoma fine and flexuous, not arranged in pairs. Pronotal humeri not angular. Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively narrow (Fig. 30); distinctly less than 2× petiolar width in dorsal view. Gaster with entire first tergite glossy (Fig. 32). Identification, taxonomy and systematics. Pheidole navigans is a small, shortlimbed, reddish brown species that belongs to the P. flavens complex. See discussion under corresponding section of P. flavens complex for how to distinguish this species from introduced Pheidole outside the complex. Within the complex, minor workers are impossible to distinguish based on known characters. Major workers can be separated from those of P. flavens by the combination of predominantly longitudinal rugae on the posterolateral lobes, the more distinct and narrow antennal scrobe bordered mesially by strong, unbroken frontal carina, and the more continuously glossy scrobe depression.

Pheidole navigans
Although the type locality of Pheidole navigans is Germany, the species was originally described by Forel from specimens intercepted during quarantine inspection of orchids originating from Veracruz, Mexico. We revive this name from synonymy and elevate it to species rank so that it can be applied to a putative species that has recently established in the southeastern United States and Hawaii. This ant has most often been referred to as P. moerens since it was first reported from Alabama nearly fifty years ago by M.R. Smith (1967). However, the examination of type specimen photographs (MCZ-ENT00009137) suggests that these introduced populations are heterospecific with P. moerens Wheeler.
Whether the introduced populations are actually conspecific with P. navigans Forel will require a thorough revision of this taxonomically vexing species complex. Of all the type material we have examined, however, that of P. navigans bears the closest resemblance in gross morphology. Thus we propose P. navigans Forel be used in place of P. moerens for referring to the aforementioned introduced populations. Future systematic study of this species should also examine P. floridana subsp. aechmeae (currently synonymized under P. flavens, but also recorded from Veracruz, Mexico) and P. flavens var. mediorubra Santschi (described from Loreto, Argentina and currently treated as a synonym of P. alacris Santschi).
The major workers of Pheidole navigans differ from those of P. moerens in the following respects. They exhibit a distinct and narrow antennal scrobe capable of receiving the entire antennal scape in repose. The scrobe is bordered by a strong, unbroken frontal carina mesially, and the depression is marked by a continuous smooth surface entirely (or nearly entirely) uninterrupted by rugulae. The rugulae of the frons extend to approximately an eye's length distance from the posterior head margin. The anterior portion of the promesonotum is crossed by long and distinct transverse striae.
The examined major workers of P. navigans from Alabama (CASENT0106664) and Venezuela (CASENT0248831), along with those from Florida and Hawaii, and a specimen imaged from Paraguay (CASENT0178020), share a notably consistent morphology for being spread across such as wide range. The characteristics shared among these majors include the following. Frontal carinae strongly produced, forming the mesad border of a shallow but well-demarcated antennal scrobe capable of accommodating the entire scape in repose. Antennal scrobe weakly foveolate. Cephalic carinulae mostly longitudinal with very little reticulation posterior to the eye. Cephalic carinulae extending up to, but not beyond the medial excision ('V') of the posterior head margin. Promesonotal dome with a relatively low profile, mesonotal declivity short and relatively gradual. In dorsal view, promesonotum weakly punctate, anterior portion with distinct transverse carinulae. Although we tentatively treat the specimen from California (CASENT0005742) as P. navigans, it differs morphologically from the aforementioned specimens and bears closer resemblance to P. exigua var. tuberculata Mayr (currently synonymized under P. flavens).
The similarity of these northern hemisphere specimens to the one from Paraguay raises the possibility that these putatively conspecific populations originated in South America. Indeed, the Paraguay specimen was collected in the Reserva Natural del Bosque Mbaracayú near the Río Paraná -a region infamous for serving as a cradle of ant invasion (Suarez and Tsutsui 2008).
Biology. In Florida, Naves (1985) reported Pheidole navigans (as P. moerens) nesting under boards, at base of oak trees and fence posts, along roots, under palm leaves, inside wall crevices, and rarely in the ground. The chambers are built with small soil or debris particles and have small openings. Most nuptial flights occur in July. The species was found to practice dependent nest founding, but became monogynous before the first brood was reared. Mature colonies can support over 100 majors and over 500 workers. They feed on seeds and scavenge and prey on small dead or live arthropods, and forage very close to the nesting sites. Deyrup (2000) also provided observations of this species (as P. moerens) from Florida, adding that it occurs in both disturbed areas and mesic or moist woods, also nests in hollow twigs, nuts and in leaf litter, and is occasionally arboreal.
Distribution. The precise native range of P. navigans is unknown, but it is certainly of Neotropical origin. The record of the species from the Paraná region of South America suggests it could be South America. We tentatively treat both known South American records (Paraguay and Venezuela) as native, and the Mexican record as introduced, but other scenarios are equally possible. Pheidole navigans was first reported as introduced in the United States by M.R. Smith (1967) under the name P. moerens. The name P. moerens has since been applied to North American records from Alabama (Glancey et al. 1976;Smith 1967), California (Garrison 1996;Martínez 1997), Florida (Deyrup et al. 1988;Deyrup et al. 2000;Wojcik et al. 1975), Louisiana (Dash and Hooper-Bùi 2008), Mississippi (MacGown and Hill 2010), North Carolina (Guénard et al. 2012) and Texas (Wilson 2003). We tentatively treat all of these records as P. navigans, but the California and Texas records could also belong to another species in the flavens complex. In the Pacific, P. navigans is established in Hawaii (Gruner et al. 2003). We cannot confirm whether the P. moerens records from Cocos Island (Solomon and Mikheyev 2005) or the indoor records from a butterfly house in the northwestern United States (collection code KRW26Feb99) refers to P. navigans or another member of the flavens complex.
Risk statement. The species most often referred to as P. moerens in the southeastern United States, and treated here as P. navigans, has been expanding its range since it was first reported in Alabama in 1967. However, this species is not considered a major pest and is only occasionally reported to enter houses (Deyrup et al. 2000). In Louisiana P. navigans is considered a pest (Dash and Hooper-Bùi 2008). Pheidole navigans could become more regionally and possibly globally widespread in the future.

Pheidole noda F. Smith Figs 81, 88i
noda. Pheidole nodus Smith, F. 1874: 407 (s.) JAPAN, Hyogo. Forel 1900: 268 (w.); Wheeler, W.M. 1906: 309 (q.);Ogata 1982: 196 (m.); Wheeler, G.C. & Wheeler, J. 1953: 75 (l. (Fig. 7). Head rugoreticulate on posterolateral lobes and laterad of frontal carinae (Fig. 13a), but frons dominated by long, well-organized and parallel longitudinal rugae (Fig. 13b). Antennal scrobes indistinct to moderately impressed, but frontal carinae always forming a border capable of accepting the antennal scape (Fig. 13c). Promesonotum in profile with two convexities (Fig. 5), the large anterior dome in addition to a distinct mound or prominence on the posterior slope. Postpetiole forming a high dorsally bulging dome that is tallest at midpoint (Fig. 2a); ventral margin flat to very weakly convex (Fig. 2b). Minor HW 0. 57-0.66, HL 0.71-0.82, SL 0.91-1.07, CI 80-82, SI 157-162 (n=5, Eguchi 2008). Head predominantly glossy ( Fig. 36), lacking punctation and or rugae above eye level. Posterior head margin strongly convex (Fig. 44). Antennal scapes long (e.g. Fig. 39), but not surpassing the posterior head margin by more than 2× eye length. Promesonotum in profile with two convexities, the large anterior dome (Fig. 43a) in addition to a distinct prominence on the posterior slope (Fig. 43b). Petiole and postpetiole glossy to very weakly sculptured laterally (Fig. 48). Postpetiole forming a high dorsally bulging dome that is tallest at midpoint; ventral margin flat to very weakly convex (Fig. 2). Identification, taxonomy and systematics. Pheidole noda is a large, long-limbed, dark colored species most easily recognized by its distinctly enlarged dome-like postpetiole. The species belongs to a clade of large-bodied species that has diversified across Indomalaya . Although both P. noda and P. megacephala are considered to have an enlarged postpetiole, they are very different in shape. That of the former is dome-like (Fig. 2) and that of the latter has an anteroventral bulge in addition to the posterodorsal bulge (Fig. 1). The majors of P. noda are easily separated from those of P. megacephala by the strongly sculptured face (Fig. 8 vs. Fig. 9). The minors both have glossy faces, but those of P. noda are larger with relatively longer antennal scapes (Fig. 39 vs. Fig. 40). Pheidole noda is occasionally confused with other Asian tramp Pheidole, including P. fervens and P. indica, but both major and minor workers are easily separated from these by the enlarged postpetiole. Readers are referred to Eguchi (2008) for characters used to separate P. noda from its other Asian congeners.
Biology. Despite being a relatively common species across its native range, little is known about the biology of P. noda. The species is apparently easy to keep in laboratory settings, and Yamamoto et al. (2009) reported that they kept a colony with five dealated queens, suggesting dependent colony foundation or polygyny. The authors also noted that in Japan it nests in the ground but also forages in vegetation. Pheidole noda was the most frequent visitor to extrafloral nectaries of Mallotus japonicus in an experiment conducted in Japan (Yamawo et al. 2012). Eguchi (2008) observed that P. noda occurs from open lands to relatively developed forests, and nests in the soil, under shelters on the ground, and in rotting logs. Eguchi (2004a) noted that the species takes seeds of sesame and amaranth put on the ground, and majors serve as repletes. During a recent survey in Yunnan, China, the species was found to occur in rubber tree plantations and rainforest between 550 and 1219 m (Liu et al. 2015).
Distribution. Pheidole noda is considered native across mainland Asia, occurring from western India east to Japan. Forel (1903) reported the species from the Andaman Islands but it was not recovered during a more recent survey of the islands (Mohanraj et al. 2010). There is geographic disjunction between the mainland Asia population and the populations from the southern islands of Indonesia. The majors of the Indonesian taxon, originally described as Pheidole treubi Forel, were considered a distinct population by (Eguchi 2001b), but conspecific with P. noda. Although not included on the map, if verified, the records from the Russian Far East (Kupianskaia 1990) would be the most northern extent of the native range. The dispersive capacity of P. noda is demonstrated by its colonization of Volcano Island (Nishino-shima Island), which is 22 ha in size and located 1,000 km south of mainland Japan. The island erupted in 1973, virtually eradicating all life. Pheidole noda was the only ant species discovered during the 1983 survey, and was one of only two discovered during the 2004 survey (the other being Tetramorium bicarinatum).
The only confirmed record of P. noda occurring outside of its putative native range is from a glasshouse in Italy (Limonta and Colombo 2003), where it was found together with P. megacephala and Tetramorium bicarinatum on nursery plants imported from Asia. The species was also found on plant material imported from Asia and intercepted at quarantine facilities in Washington and Hawaii.
Risk statement. Pheidole noda is not considered an agricultural, ecological or structural pest species, although it is often associated with disturbed habitats. The species is also not known to have established outdoors beyond its native range. However, perhaps because it can be easily maintained in artificial nests, colonies with laying queens listed as Pheidole noda and Pheidole cf. noda are available for sale from businesses advertising on the internet. The shipment of this species outside its native range to hobbyists increases its chances of accidental release into non-native habitats.  (Fig. 7); almost entirely covered by a network of intersecting rugae (Fig. 12a), lacking long, well-organized and parallel longitudinal rugae on the frons (Fig. 12b). Frontal carinae indistinct, quickly becoming integrated into dense rugoreticulum that covers the entire face. Antennal scrobes entirely lacking. Antennal insertions surrounded by deeply excavated pits (Fig. 12c). Head often a lighter reddish brown than the mesosoma. Promesonotum in profile with two convexities (Fig. 5), the large anterior dome in addition to a distinct mound or prominence on the posterior slope. Postpetiole not swollen relative to petiole (Fig. 3). Minor HW 0. . Head predominantly glossy (Fig. 36), lacking punctation and or rugae above eye level. Posterior margin strongly convex in full-face view such that the head outline forms a single unbroken curve from eye to eye (Fig. 44). Antennal scapes extremely long, surpassing posterior head margin by more than 2× eye length (Fig. 39). Promesonotum in profile with two convexities, the large anterior dome (Fig. 43a) in addition to a distinct prominence on the posterior slope (Fig. 43b). Mesopleuron mostly sculptured. Postpetiole not swollen relative to petiole (Fig. 3). Petiole and postpetiole strongly sculptured laterally (Fig. 47).

Identification, taxonomy and systematics.
Pheidole obscurithorax is a member of the New World (and polyphyletic, see Moreau 2008) P. fallax species group defined by Wilson (2003). It is a large dark species over 6 mm in body length. The species is easily distinguished from P. megacephala by the much larger body size and relatively reduced postpetiole, in addition to the strongly sculptured head of the major worker (Fig. 12), and the much longer antennal scapes of the minor. It is separated from other New World species treated here, including those of the P. punctatissima clade and P. flavens complex, by the much larger size, prominence on the posterior slope of the promesonotum (Fig. 5, major; Fig. 50, minor), densely rugoreticulate face of the major (Fig. 12), and smooth head and long antennal scapes of the minor. The Old World species P. fervens, P. indica, and P. noda all have majors with strongly sculptured head and minors with smooth heads, and the reader is referred to the key for characters used to separate these from P. obscurithorax.
Biology. In its introduced range of the southeastern United States, P. obscurithorax is characterized by its large size, large nest mounds, very active foraging and fast recruitment to bait such as cookie crumbs (King and Tschinkel 2007). It nests in soil in open areas, where it produces conspicuous nests, each generally with a single large opening often covered by a leaf or other collected material (Storz and Tschinkel 2004). The species is an omnivorous scavenger of dead arthropods (possibly including dead fire ants), and less frequently of plant material such as flower petals (Storz and Tschinkel 2004). Studies in its introduced range found evidence that P. obscurithorax is monogynous and is spreading by natural dispersal of winged females in addition to human-mediated long-distance dispersal (King and Tschinkel 2007). The species was most often found associated with disturbed habitats such as lawns and roadsides, but there are also records of it occurring in natural areas such as hardwood forests (Wilson 2003). However, its steady expansion across the southeastern United States and cooccurrence with S. invicta suggest it is an important species to monitor.
Distribution. Pheidole obscurithorax is presumed native to the South American region of Argentina, Paraguay and southern Brazil that includes the Paraguay, La Plata and Parana Rivers. This flood-prone area is the cradle of many other well-known invasive ants including fire ants (Solenopsis invicta Buren and S. richteri Forel), the Argentine ant (Linepithema humile), and many lesser-known species that were anthropogenically introduced (King and Tschinkel 2007;Storz and Tschinkel 2004;Suarez and Tsutsui 2008;Wilson 2003). Most of these species, including P. obscurithorax, were first introduced to North America via the Mobile, Alabama shipping port pathway. Pheidole obscurithorax was introduced to Mobile, Alabama around 1950 (Naves 1985) and subsequently expanded its range to include Florida, Georgia, Mississippi and Texas (Storz and Tschinkel 2004;Wilson 2003). Additional occurence records, including the first record for Bolivia, were published (Wetterer et al. 2015) just as this manuscript was going to press, and were not included in the present study.
Risk statement. Pheidole obscurithorax is not currently considered a pest in its introduced range, as it does not sting and is not known to infest dwellings or structures (King and Tschinkel 2007). However, the species is an aggressive predator (Deyrup et al. 2000) and may have the potential to become a pest or to negatively impact native species if its populations continue to grow and spread. Pheidole obscurithorax is thought to spread across the southeastern United States by mated queens (not colony fragments) that are being transported in substrates such as potted plants. It is possible that P. obscurithorax could become more widespread regionally and globally in the future. Fig. 83, Fig. 88K parva. Pheidole parva Mayr 1865: 98, pl. 4, fig. 28 SL 0.41-0.45, CI 85-92, SI 45-51 (n=11, Eguchi et al. 2007). Head subquadrate (Fig. 7). Posterolateral lobes, including posterior head margin, covered in rugoreticulum (Fig. 26). Antennal scrobes indistinct to moderately impressed, but frontal carinae always forming a border capable of accepting the antennal scape (Fig. 13c). Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope. Promesonotum in dorsal view transverse with strongly projecting shoulders (Fig. 28). Promesonotal dorsum rugoreticulate with distinct long longitudinal striae in addition to shorter sections of transverse and intersecting striae (Fig. 22). Postpetiole not swollen relative to petiole (Fig. 3). Minor HW 0. 39-0.50, HL 0.43-0.54, SL 0.38-0.46, CI 88-94, SI 84-102 (n=17, Eguchi et al. 2007). Posterior portion of head with many short to medium length segments of striae distinctly interlaced among punctate ground sculpture (Fig. 59). Antennal scapes with erect to suberect hairs (Fig.  56); scapes do not surpass posterior head margin (Fig. 41). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Pronotal humeri angular (Fig. 28). Hairs on mesosoma fine, flexuous, of unequal length and not arranged in pairs (Fig. 54). Postpetiole not swollen relative to petiole (Fig. 3); postpetiole narrow in dorsal view, only slightly broader than petiole (Fig. 61). Identification, taxonomy and systematics. Pheidole parva is a very small and inconspicuous species that is thus far reported only from Asia, a few localities in Arabia, and the islands of the Indian Ocean and the Pacific Ocean. It belongs to an Old World clade scattered across Indomalaya and into Oceania, and was treated as part of the P. rinae complex by Eguchi et al. (2007). The minor workers are completely covered in punctate sculpture and are difficult to differentiate from those of the Neotropical P. flavens complex. The similarity is so close that an introduced population of P. parva from the Seychelles was described by Forel, on the basis of the minor worker, as Pheidole flavens var. farquharensis. The similarity is entirely convergent, as these lineages are distantly related. Pheidole parva minors can be separated from those of the P. flavens complex most reliably by the interrupted striae that are interlaced among the punctate ground sculpture of the posterior head (Fig. 59 vs. Fig. 60). This character can also be viewed in the dorsal view. Pheidole parva minors can be separated from those of the P. punctatissima clade treated here by the glossy gaster (Fig. 32 vs. Fig. 33) and finer mesosomal hairs of unequal length (Fig. 54 vs. Fig. 53). The major workers are characterized by a defined and moderately depressed antennal scrobe and a thick network of reticulated rugulae on the posterior lobes. This pattern is most similar to that of the broadly sympatric P. fervens and P. indica, but P. parva is much smaller than those species (HW < 0.95 mm vs HW > 1.10 mm) and lacks the distinct prominence on the posterior slope of the promesonotal dorsum (Fig. 4 vs. Fig. 5). The majors of P. parva can be separated from those of the P. flavens and P. punctatissima group species treated here by the much stronger and more reticulated carinae which reach the posterior margin (Fig. 26 vs. Fig. 25 and Fig. 27) in addition to other characters given in the key. Readers are referred to Eguchi (2008; for characters separating P. parva from its Asian congeners.

Pheidole parva Mayr
Biology. Little is known about the biology of P. parva, but it does appear to be expanding its range and is worth monitoring in the future as it exhibits a high tolerance for disturbance. Eguchi (2008) observed that the species seems to inhabit open lands and forest edges, and has probably expanded its range in some part as the result of human commerce. Pheidole parva was one of the most commonly collected ants in a myrmecological study of agricultural fields in Vietnam and Okinawa (Anh et al. 2010;Suwabe et al. 2009). A recent study of 18 structure invading pest ants of healthcare facilities in Singapore found P. parva the most frequently encountered species (Man and Lee 2012). Pheidole parva and Pheidole megacephala were the two most common ant species encountered and together accounted for over 50% of the total collection (25.9% and 25.2%, respectively). In Mauritius and the Seychelles P. parva can be locally abundant and can be found in soil and leaf litter, under stones or root mats, in rotten logs, foraging on or nesting in the ground, as well as in lower vegetation and even under the bark of live trees (Fischer and Fisher 2013). It was collected there in parks, gardens, mangrove, coastal scrub, degraded dry forest, littoral and mixed forest, and rainforest, in elevations between 1-445 m. It was collected inland on the Arabian Peninsula from date tree orchards, banana plantations and under potted plants between 675-735 m elevation (Fischer and Fisher 2013).
Distribution. Pheidole parva is considered here as native to the Indo-Malay region. The species is recorded from the Asian mainland from India east to China. We consider the records from Indonesia, Borneo, the Philippines and Taiwan to be native, but much of this distribution could represent a more recent anthropogenic expansion. We consider the records from the Okinawa and Kagoshima prefectures of Japan to be introduced along with the records from Palau to represent introduced populations, but it is difficult to know whether the species arrived in these islands before, with or after the arrival of humans. The species is introduced in the Seychelles, Mauritius, Saudi Arabia and the United Arab Emirates (Fischer and Fisher 2013). Pheidole parva was also collected from hothouses in Austria and Germany.
Risk statement. Pheidole parva is not currently considered to be a significant pest species, and no impacts on agricultural systems or native ecosystems have been documented as of yet. The species is known to invade structures, however, and its prevalence in Singapore health care facilities (Man and Lee 2012) suggests it could become a more widespread nuisance pest in the future. Live colonies have been reported from various ships (Fischer and Fisher 2013) and should be screened for during quarantine inspections. Fig. 84 (Fig. 7). Posterolateral lobes lacking sculpture (including foveolate ground sculpture, carinae and rugae) posterior to maximum extent of antennal scapes in repose (Fig. 9). Head glossy, lacking foveolate ground sculpture. Hypostomal bridge with a small median tooth in addition to a pair of larger inner teeth (Fig. 18). Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope.

Pheidole proxima Mayr
Promesonotal dorsum glossy, lacking foveolate ground sculpture or striae. Pronotal striae in dorsal view mostly absent (Fig. 23). Metapleuron with moderate rugulae and some weak punctation (Fig. 16). Petiolar node strongly punctate (Fig. 16). Postpetiole not swollen relative to petiole (Fig. 3). Minor HW 0.46, HL 0.52, SL 0.40, CI 90, SI 86 (n=1). Head predominantly glossy (Fig. 36), lacking punctation and or rugae above eye level. Posterior head margin weakly convex (Fig. 45) to weakly concave (Fig. 46) in full-face view. Antennal scapes reach but do not surpass posterior head margin (Fig. 41). Mesopleuron entirely punctate (Fig. 52a). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Propodeal spines moderately produced and spiniform (Fig. 52b). Petiole distinctly sculptured except for apical portion of node. Postpetiole not swollen relative to petiole (Fig. 3). Identification, taxonomy and systematics. Pheidole proxima is a relatively small, brownish yellow, short-limbed species with a strongly shining integument. The phylogenetic placement of P. proxima is unknown, but it almost certainly clusters within an Old World clade that has radiated across Australia and New Guinea. The species is slightly smaller than P. megacephala, but both have workers with almost entirely glossy faces. The postpetiole of P. proxima is not swollen relative to the petiole (Fig. 3), as it is in P. megacephala (Fig. 1). The head of the major is subquadrate (Fig. 7), while that of P. megacephala is more heart-shaped (Fig. 6). The antennal scapes of the minor do not surpass the posterior head margin (Fig. 41), as they do in P. megacephala (Fig. 40). The other two Pheidole species established in New Zealand are P. rugosula and P. vigilans. The glossy face of P. proxima easily separates both worker castes of from those of P. rugosula. In addition to being significantly smaller (major HW < 1.0 mm, minor HW < 0.48 mm) than P. vigilans (major HW > 1.2 mm, minor HW > 0.52 mm), the major of P. proxima is more sculptured (Fig. 16 vs. Fig. 17), and the hypostomal bridge has a distinct median tooth (Fig. 18 vs. Fig. 19). The minors of P. proxima are separated from those of P. vigilans by the shorter scapes (Fig. 41 vs. Fig. 40), more sculptured mesopleuron (Fig. 52a vs. Fig. 51a), and more robust propodeal spines (Fig. 52b vs.  Fig. 51b). Additional taxonomy of these species is discussed in (Berry et al. 1997).
Comparison of the Pheidole proxima Mayr type series and images of the two subspecies suggests that all three taxa are heterospecific. There is some reason to believe, however, that the name P. proxima Mayr does not apply perfectly to the species recently introduced to New Zealand. The specimens examined from New Zealand conflict with Mayr's original description and type specimens on several points. The pronotal dorsum of the type major worker is transversely rugose whereas that of the New Zealand specimens are completely glossy. Although we were unable to examine minors from the type series, Mayr described the head of the minor worker as coriaceous and striate-rugose with scapes that barely exceed the posterior margin. In contrast the minor workers from New Zealand have heads that are completely glossy and scapes that do not exceed the posterior head margin. Forel, in his description of P. proxima subsp. bombalenis, describes the minor worker as identical to P. proxima Mayr with the exception of having longer propodeal spines. The specimen images of the P. bombalensis syntype minor show a strongly sculptured face, similar to the pattern described by Mayr. The major workers from the type series are larger than the New Zealand specimen we measured (HW 1.03-1.05 mm vs. HW 0.95 mm), have relatively narrower heads (CI 87-89 vs. CI 92), and relatively shorter antennal scapes (SI 42-46 vs. SI 52). While a more exhaustive survey of Australia's Pheidole may reveal the New Zealand population to be more closely related to another species from that fauna, we follow Berry et al. (1997) in using P. proxima Mayr. Biology. The only natural history published for P. proxima was recorded by Green and Gunawardana (2006) from their work with the New Zealand incursion. They reported that P. proxima produced large nests recognizable by tiny conical mounds of sandy or grainy material above the ground near the entrance. The size of the mounds varies with soil type, with mounds as small as 5 mm high by 200-300 mm in diameter. They are tolerant to disturbance and capable of invading structures. The minor and major workers are both active foragers and were observed recruiting to both sweet and savory baits in high numbers.
Distribution. Pheidole proxima Mayr is native to Queensland, Australia. The sparse records of the species are scattered from Cape York at the northernmost tip of the continent down to the Gold Coast. The species is introduced to New Zealand and was first detected during a 2004 survey of the Port of Napier following an incursion of Solenopsis invicta (Green and Gunawardana 2006). The species is now widespread across the North Island from the Napier-Hastings area to Auckland.
Risk statement. Pheidole proxima is at most considered a nuisance species in New Zealand on account of its ability to infest structures. However, very little is known about the species, including its impact on agricultural systems and native ecosystems. There is little reason to believe that it will become globally or regionally widespread.  Longino pers. comm.). Head bicolored with the yellowish posterior two-thirds contrasting with the darker brown anterior third and rest of body (Fig. 34). Head subquadrate (Fig. 7); often entirely foveolate (Fig. 11), but portions of posterolateral lobes can be glossy. Posterolateral lobes never with distinct rugae. Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope. Promesonotal dorsum usually foveolate and never with distinct transverse striae.

Pheidole punctatissima Mayr
Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole relatively broad; distinctly more than 2× petiolar width in dorsal view (Fig. 31). Gaster with at least anterior 1/3 of first tergite matte (Fig. 33). Minor HW 0.44-0.50, HL 0.54-0.59. SL 0. Longino pers. comm.). Head, including the area mesad of the frontal carinae, entirely covered by reticulated network of punctures (Fig. 37). Posterior head margin relatively narrow (Fig. 58). Antennal scapes lack standing hairs (Fig. 55); scapes surpass posterior head margin by a distance equal to or greater than eye (Fig. 40); scapes relatively long (SI 103-125). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Hairs on mesosoma stout, stiff, of equal length and arranged in pairs (Fig. 53). Postpetiole not swollen relative to petiole (Fig. 3). Postpetiole broad in dorsal view, distinctly broader than petiole (Fig. 62). Gaster with at least anterior 1/3 of first tergite matte (Fig. 33). Identification, taxonomy and systematics. Pheidole punctatissima is a small species with entirely punctate minor workers that are usually dark red brown to nearly black. The major workers are easily recognizable by the distinct bicolored head which is dark anteriorly and yellowish white posteriorly. Pheidole punctatissima is a member of the Neotropical P. punctatissima clade, together with P. anastasii and P. bilimeki . Among species treated here, it is easily confused with the aforementioned and with members of the P. flavens complex. Minor workers can also be confused with those of P. parva. Within the P. punctatissima clade, the major workers of P. punctatissima are immediately distinguished from those of both P. anastasii and P. bilimeki by their bicolored heads (Fig. 33). The minor workers of P. punctatissima tend to have relatively narrower posterior head margins and longer antennal scapes than those of P. anastasii and P. bilimeki, but separation can be difficult. See section under P. anastasii for identification notes.
Biology. Pheidole punctatissima is a weedy species that tends to be arboreal and prefers open, disturbed habitat (Longino and Cox 2009). It is most commonly found nesting in dead wood on the ground or in dead tree branches. Wilson (2003) reported winged reproductives were found in nests during April and July. Specimen records retrieved from Antweb.org indicate the species was collected from 10-2500 m elevation (570 m average). Pheidole punctatissima has also managed to establish indoors in several European countries. Colonies were found in Denmark infesting a hospital and in Norway inhabiting private homes and a nursing home (Birkemoe and Aak 2008). Birkemoe and Aak (2008) speculated that the species was inadvertently imported along with nursery plants.
Distribution. Pheidole punctatissima is considered here as broadly native to the Neotropics from southern Mexico to northern South America. We tentatively treat the Caribbean records as native but these might represent more recent human-mediated dispersal events. The records from southern Brazil, reported at least in part from 10 different urban centers (Lutinski et al. 2013), have not been verified with specimen examination. Should the records refer to P. punctatissima Mayr and not one of its many morphologically similar congeners we would consider this to be an introduced population. Indoor colonies were found in Denmark and Norway (Birkemoe and Aak 2008).
Risk statement. Pheidole punctatissima is considered a nuisance pest that can infest structures both in its native and introduced ranges (Longino and Cox 2009). The presence of this species in hospitals and nursing homes suggest it could be a potential nuisance.  (Fig. 7); with distinct parallel rugae extending from frontal lobes posterior to apices of frontal carinae. Shorter lengths of rugae present across entire posterior region of head and extending to posterior margin in full-face view (Fig. 24). Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope. Promesonotal dorsum glossy with thin but distinct subparallel striae running oblique to the longitudinal midline (Fig. 20). Pronotal striae in dorsal view mostly oblique (Fig. 20). Postpetiole not swollen relative to petiole (Fig. 3). Minor HW 0.45, HL 0.48, SL 0.41, CI 95, SI 91 (n=1). Head with well-defined, long segments of rugae running longitudinally from below the eyes to the posterior head margin (Fig. 38). Frontal carinae distinct and reaching towards the posterior head margin, although they may occasionally be interrupted (Fig. 38). Punctate ground sculpture present on lateral surfaces of head and just mesad of the frontal carinae, but median portion of head with a large glossy section (Fig. 38). Antennal scapes reach but do not surpass posterior head margin (Fig. 41). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Postpetiole not swollen relative to petiole (Fig. 3). Identification, taxonomy and systematics. Pheidole rugosula is a small, brownish yellow, short-limbed species with moderate head sculpturing that most likely belongs to the Australian-New Guinea clade that includes close relatives of P. variabilis Mayr. The head sculpturing of both the major (Fig. 10) and the minor (Fig. 38) is distinct among all other Pheidole species treated here. These characters easily separate P. rugosula from P. megacephala (Fig. 24, Fig. 36). These same characters, together with a more sculptured promesonotal dorsum (Fig. 20, major) and stout propodeal spines, can be used to separate P. rugosula from its two other congeners that are established in New Zealand, P. proxima and P. vigilans (which is also much larger, major HW > 1.20 mm). There is a bewildering diversity of native Australian (and to a lesser extent New Guinea) Pheidole that approach the morphology of P. rugosula. Additional characters for identifying New Zealand Pheidole species are provided in Berry et al. (1997). A significant revision of the P. variabilis group is required before P. rugosula can be reliably separated from these species.

Pheidole rugosula Forel
Biology. In New Zealand, P. rugosula is strongly associated with human disturbance and is the most commonly encountered of the four Pheidole species established in New Zealand (Berry et al. 1997). It has been recorded frequently from gardens, orchards, structures and urban areas. The species was reported nesting in compost, in the soil of vegetable gardens, in the soil of lawns, and near rubbish baskets (Berry et al. 1997;Harris et al. 2005b). It has also been recorded as scavenging dead arthropods, human food waste, nuts and seeds, and is often found associated with rotting fruit (Berry et al. 1997;Harris et al. 2005b). Other collection records suggest P. rugosula will forage arboreally. Berry et al. (1997) also mention that label data suggests the species was collected several times attacking ootheca of mantids, including those of the native mantid Orthodera novaezealandiae (Colenso).
Distribution. Pheidole rugosula is believed to be native to the New South Wales region of Australia. The only country where the species has established is New Zealand (Berry et al. 1997). Berry et al. (1997) published museum records of P. rugosula from New Zealand. The first known occurrence of P. rugosula in New Zealand is from Takapuna, where it was collected in 1958 and it had reached Auckland by 1963. Since then it has been found across the Auckland and Waikato regions.
Risk statement. Pheidole rugosula is considered to be a nuisance pest around urban areas in New Zealand (Harris et al. 2005b), where foragers are attracted to pet food left out and to windfall fruit. Although it occurs in native habitats in New Zealand, its impacts are unknown. Collection data indicating a P. rugosula attack of native mantids suggest it could have some negative impact on native biodiversity, however.  (Fig. 7); glossy, lacking foveolate ground sculpture. Posterolateral lobes lacking sculpture (including foveolate ground sculpture, carinae and rugae) posterior to maximum extent of antennal scapes in repose (Fig. 9). Hypostomal bridge with two well-developed inner teeth but lacking a median tooth (Fig. 19). Promesonotal dorsum glossy, lacking foveolate ground sculpture or striae. Promesonotum in profile forming a single dome (Fig. 4), lacking a distinct mound or prominence on the posterior slope. Metapleuron almost completely glossy with strongly reduced carinulae and lacking punctation (Fig. 17). Petiolar node mostly glossy (Fig. 17), not covered by punctate sculpture. Postpetiole not swollen relative to petiole (Fig. 3). Minor HW 0.55, HL 0.58, SI 0.55, CI 95, SI 101. Head predominantly glossy (Fig. 36), lacking punctation and or rugae above eye level. Antennal scapes surpass posterior head margin by approximate distance of eye length (Fig. 40). Mesopleuron entirely glossy (Fig. 51a). Promesonotum in profile forming a single dome (Fig. 42), lacking a distinct mound or prominence on the posterior slope. Propodeal spines weakly produced and dentiform (Fig. 51b). Petiole almost entirely glossy. Postpetiole not swollen relative to petiole (Fig. 3). Identification, taxonomy and systematics. Pheidole vigilans is a large, light colored, glossy species native to Australia and introduced in New Zealand. The species belongs to an Old World clade centered in Australia. The glossy head of the majors and minors give it a superficial appearance to P. megacephala, but it is substantially larger than that species. Additionally, the postpetiole of P. proxima is not swollen relative to the petiole (Fig. 3) as in P. megacephala (Fig. 1), and the head of the major is subquadrate (Fig. 7), while that of P. megacephala is more heart-shaped (Fig. 6). Readers are referred to the section under P. proxima and P. rugosula for a discussion of how to differentiate it from the other Pheidole species established in New Zealand. Additional taxonomy of these species is discussed in (Berry et al. 1997). Within Australia, there are many taxa similar to P. vigilans and its close relative P. ampla Forel. However, a revision of that fauna is required before it can be reliably diagnosed there.
Biology. Records show it has established in urban areas and been found with fruit, in gardens, indoors and nesting in failing pasture (Berry et al. 1997).
Distribution. Pheidole vigilans is considered endemic to the south eastern corner of Australia (Brown 1971). Heterick et al. (2013) reported P. vigilans as introduced to Perth in Western Australia. The species was first collected outside of Australia in Kerikeri, New Zealand in 1956, and remains the least frequently collected Pheidole species in New Zealand (Berry et al. 1997;Cumber 1959). Although Pheidole ampla subsp. norfolkensis Wheeler was originally described as endemic to the Norfolk Islands, Brown (1971) later proposed that the species was introduced to those islands.
Risk statement. Pheidole vigilans is not considered a pest in New Zealand, but it has been collected from urban areas and may be a minor garden nuisance (Harris et al. 2005c).

Dubious and erroneous records of introduced Pheidole species
The following species were reported by McGlynn (1999) as introduced in Hawaii based on Nishida (1996a): Pheidole barbata W.M. Wheeler, P. fervida F. Smith, P. hyatti Emery, P. punctatissima, and P. noda. These records refer to quarantine interceptions (Wheeler 1934b), and it is doubtful that any of the aforementioned species ever established in Hawaii. We propose that P. barbata, and P. hyatti be removed from future lists of tramp ants. The URL referred to in Nishida (1996a) is obsolete, but the data is available at Nishida (1996b).
Pheidole fervida is also reported by McGlynn (1999) as introduced in Tahiti based on putative specimens at the LACM. These records are unverified, however, and it is possible that the Tahiti specimens actually refer to P. fervens. While P. fervida, as it is currently recognized (e.g. Eguchi 2008) is reported as occurring widely across Asia (Guenard and Dunn 2012), we do not consider any of these records to represent recent introductions outside of its native range. However, the species does exhibit synanthropic habits and may yet prove itself as an important tramp ant.
Pheidole guineensis Fabricius is a West African species that has likely never established outside of its native range. The species most often misidentified in museum collections and in the literature as P. guineensis is Tetramorium bicarinatum (Nylander). Between 1862 (Mayr 1862;Roger 1862) and 1977(Bolton 1977 The latter species was universally and mistakenly referred to as Tetramorium guineensis Fabricius. Readers are referred to Bolton (1977) for a thorough explanation of the taxonomic history.
Pheidole micula Wheeler is a species native to the southwestern United States (Moody and Francke 1982;Wheeler and Wheeler 1973;Wilson 2003). Ward (2005) unintentionally reported the species as introduced in California (pers. comm. December 17, 2013).
Pheidole umbonata Mayr is reported by McGlynn (1999) as introduced to Polynesia and possibly New Caledonia. Although this synanthropic species is widely distributed across Oceania, and is very tolerant of disturbed habitats , we consider P. umbonata to be native throughout the Pacific. Morphological (Wilson and Taylor 1967) and molecular (unpublished) variation throughout the Pacific populations suggest that the species is not a recent introduction to any of the Polynesian, Micronesia or Melanesian countries where it is found. The New Caledonia record reported by Emery (1914) was determined by Wilson and Taylor (1967) as referring to a different species. Although we are unaware of any confirmed records from New Caledonia, it is quite possible the species does occur there.