Family Formicidae, subfamily Ponerinae, tribe Odontomachini (Hölldobler & Wilson 1990). Odontomachus is a senior synonym of Pedetes Dalla Torre, Champsomyrmex Emery, and Myrtoteras Matsumura (Bolton 1995); these junior synonyms have not been used for more than 25 years, and there is no current confusion about these names.

Diagnosis (modified from Hölldobler and Wilson 1990)

Mandibles slender, elongate, attached near middle of anterior margin of head, abruptly bent inward at apex, widened apices expanded with three teeth arranged in a vertical series; third abdominal segment not differentiated by a constriction from the rest of the abdomen. Nuchal carina (the ridge delimiting the occiput) V-shaped, narrowed toward its median into a mid-dorsal groove; apophyseal lines present as convergent lines from the vertex of the head up to the nuchal carina.

The shape of the nuchal carina and the presence of apophyseal lines distinguish Odontomachus from the somewhat similar genus Anochetus. A simple character for separating Odontomachus from Anochetus in the U.S. is the shape of the petiole, ending in a dorsal spine or simple cone in Odontomachus, and in a pair of spines or a pair of angles in Anochetusmayri Emery, the only representative of its genus known from the U.S. (Deyrup 2002). Anochetus mayri, which is native to the West Indies and exotic in Florida, is smaller (4 to 5 mm in length) than the Odontomachus species considered here. Anochetus kempfi Brown (West Indies) is within the size range of local Odontomachus, but very slender, and with a two-spined petiole.

Diagnosis of worker (Fig. 1).

Distinguished from other U.S. Odontomachus by the following combination of character states: conspicuous striae present on basalar lobe (oval sclerite at dorsal posterior corner of mesopleuron); posterior side of petiole without transverse striae; appressed hair on first gastral tergite sparse, short, spaces between hairs often as wide as the length of hairs.

Description of holotype worker

Features visible in lateral view described from left side. Measurements in mm. Total length (length of head excluding mandibles + length of mesosoma + length of petiole + length of gaster): 7.48; head length: 2.12; width of head at rear margins of eyes: 1.80; width of head at widest part of occipital lobes: 1.62; length of left mandible: 1.20; maximum width of eye: 0.30; maximum width of clypeal area: 0.30; length of mesosoma: 2.67; length of petiole: 0.52; length of gaster: 2.17. Head: fine striae diverging from frontal lobes, covering frontal aspect of head, disappearing before occipital area, covering only the upper quarter of the extraocular furrow; posterior lateral corners, occipital area, underside of head smooth and shining. Mesosoma: pronotum with roughly circular concentric striae, without longitudinal striae reaching the hind margin; mesonotum and propodeum with transverse striae; striae present on basalar lobe; mesopleuron smooth, shining, with longitudinal striae along dorsal and ventral margins. Petiole: apically spinose; in profile anterior face convex, posterior face bisinuate; posterior face in posterior view smooth and shining, without hairs. Gaster: shining, no surface sculpture except for minute punctures from which hairs emerge; first tergite sparsely covered with short, appressed, pale hairs, spaces between hairs usually as wide as length of hairs; first tergite with sparse, scattered, suberect long hairs, including an uneven subapical row. Color: head, antennae, mesosoma, petiole dark reddish brown, contrasting with blackish-brown gaster; legs dark yellow, contrasting with mesosoma.

Diagnosis of queen

Queens of the four U.S. species were examined. Diagnosis as in worker.

Description of a paratype queen

Measurements in mm. Total length (length of head excluding mandibles + length of mesosoma + length of petiole + length of gaster): 8.37; head length: 2.15; width of head at rear margins of eyes: 1.77; width of head at widest part of occipital lobes: 1.70; length of left mandible: 1.20; maximum width of eye: 0.27; maximum width of clypeal area 0.25; length of mesosoma: 2.75; length of petiole: 0.52; length of gaster: 2.95. Structural character states and color as in worker, except for occurrence of ocelli and expansion of the mesosomal dorsum (pronotum, mesonotum, scutellum) associated with flight; pronotum transversely striate, mesonotum longitudinally striate.

Diagnosis of male

Distinguished from other U.S. Odontomachus by the following combination of character states: ocelli very large, wider than the distance between the lateral ocelli and the eyes (Fig. 2A); body color medium brown, antennae yellowish.

Description of a paratype male

Measurements in mm. Total length (length of head excluding mandibles + length of mesosoma + length of petiole + length of gaster): 6.66; head length: 1.07; width of head at widest part, including eyes:1.45; length of mesosoma: 2.25; length of forewing: 4.95; length of petiole: 0.47; length of gaster: 2.57. Head: in frontal view, eyes longer than the distance between them dorsally; median ocellus wider than the distance between a lateral ocellus and the margin of the eye; clypeus in profile not strongly protuberent. Mesosoma: pronotum, mesopleural area above and below episternal suture feebly shining, not striate; mesonotum finely striate, transversely on anterior quarter, remainder longitudinally; scutellum convex, shining, lacking a median carina; propodeum without a raised carina, feebly shining, with weak, fine striae in the following patterns: a median series of concentric ovals, posterior portion with transverse bisinuate lines, obliquely longitudinal lines laterally ventral to spiracle; propodeum in profile long and low, without a declivitous posterior portion; gaster shining, tergites without surface sculpture except for fine, hair-bearing punctures, evenly covered with long, fine, sub-appressed hairs.

Type localities and associated information, as appear on specimen labels.--Holotype Worker. FL: Highlands Co., Archbold Biological Station, 15-IV-1996, M. Deyrup, recently burned mature scrub, Red Hill, nest in sand. Paratype dealate queen used for description of queen (designated on label). Same locality, collector as holotype, 18-I-1983, Florida scrub habitat with Ceratiola ericoides, E. side of Tract 7. Paratype male used for description of male (designated on label). Same locality, collector as holotype, 8-XI-1990, at window, main building. Paratype workers and queens (all from Florida). The following specimens all have the same locality and collector as the holotype: 18 workers: same data as holotype (nest series with holotype); 2 workers: 30-IX-1982, sandhill area with Quercus laevis, SE Tract, at bait; 4 workers: 3-X-1982, in patch of Hypericum, road 19E, at bait; 2 workers: 19-IX-1982, mature stand of Pinus clausa, SE Tract, trail 10, at bait; 3 workers: 30-XIII-1985, mature sand pine scrub habitat, Red Hill, trail 1; 1 worker: 28-X-1982; 1 worker: 7-VI-1984, sandhill habitat, Tract 19E; 1 dealate queen with associated worker: 13-X-1982, in leaf litter of Bejaria racemosa, Tract 7, road 18; 1 worker: 24-XI-1982, in dry leaf litter at base of oak, sandhill habitat, Tract 19; 2 workers: 28-VI-1985, in leaf litter of Carya floridana, sandhill habitat, Tract 19E; 1 worker: 6-IV-1983, mature Pinus clausa scrub, in pan trap below Townes trap; 1 worker: 5-IX-1985, mature Pinus clausa scrub, NE firelane of NE Tract; 1 worker: 14-XI-1985, mature sand pine scrub, Red Hill, SE Tract; 1 worker: 13-VI-1985, sand pine scrub habitat; 1 worker: 17-X-1988; 1 worker: 3-X-1982, in tussock of Andropogon; 1 worker: 6-VII-1984, mature sand pine scrub habitat, NE firelane of NE Tract; 1 worker: 25-VII-1984, mature sand pine scrub habitat, NE firelane of NE Tract; 1 worker: 28-VI-1985, Tract 19E, sandhill habitat. Paratype workers and queens not from holotype locality: 3 workers: Highlands Co., Lake Placid, 18-IX-1987, P. Martin, sand pine scrub habitat in former YMCA Camp 2 mi. S. of town on Highway US 27; 1 worker: same locality, collector, habitat as previous, 30-X-1987; 3 workers: same locality, habitat as previous, 27-X-1987, J. Cronin, 1 worker: Highlands Co., Lake Placid, 15-X-1986, M. Deyrup, Holmes Avenue scrub site east of town; 3 workers: Highlands Co., Sebring, 11-IX-1987, P. Martin, former Flamingo Villas Devel., 3.7 mi. SE of Sebring, sand pine scrub; 2 workers: same locality, habitat as previous, 10-XI-1987, J. Cronin; 1 dealate queen with associated worker: same locality, habitat as previous, 17-IX-1990, M. Deyrup; 1 worker: Polk Co., TNC Tiger Creek Preserve, 5-X-1989, M. Deyrup, recently burned scrub with Quercus laevis; 1 worker: Polk Co., Lake Wales, 26-I-1988, P. Martin, Flaming Arrow Boy Scout Camp, east of town, sand pine scrub; 1 dealate queen: same site, habitat as previous, 24-XI-1987, J. Cronin; 12 workers: Orange Co., Walt Disney World, S16, T24S, R27E, 5-VII-1996, Z. Prusak, sand pine/oak scrub zone MW-5 (unburned) #88; 2 workers: Orange Co., Walt Disney World, 22-VII-1996, Z. Prusak, MW-5 Cons. Area, unburned zone, sand pine scrub; 2 dealate queens, 4 workers: Citrus Co., 12 mi. NW of Brooksville, 11-V-2002, M. Deyrup & J. Mosley, Withlacoochee State Forest, Sugar Mill Tract, north of 480, sandhill with oaks; 4 workers: Citrus Co., 5 mi. W. of Inverness, 14-IX-1991, M. Deyrup, sandhill habitat, Withlacoochee State Forest; 2 workers: same locality, collector, habitat as previous, 14-XI-1991; 4 workers: Citrus Co., Withlacoochee State Forest, 14-XI-1991, M. Deyrup, 3.2 km south on forest road that begins 7.2 km west of jct. State Road 44 and County Road 581, open sandhill habitat, sparse ground cover of Aristida beyrichiana, scattered mature Pinus palustris and Quercus laevis; 5 workers: Citrus Co., Pine-Oak Estates, 1-IV-1993, M. Deyrup, sandhill habitat along Rt. 488, 3 mi. NE jct. with Rt. 495. Paratype males. The following males are from the Archbold Biological Station, collected by M. Deyrup in Townes flight traps on trails through mature sand pine scrub; dates as follows: 2: 19-VI-1985; 1: 8-VII-1985; 2: 29-VI-1985; 1: 13-VII-1985; 1: 14-VII-1985; 1: 10-VI-1985; 1: 26-VI-1985; 2: 5-VII-1985; 4: 1-VII-1985; 1: 9-X-1987; 1: 28-XII-1983; 1: 16-VII-1986; 1: 24-VII-1985; 1: 21-VII-1983; 1: 3-VII-1983; 1: 27-VI-1983; 2: 28-VI-1983; 1: 21-VI-1983; 1: 20-XI-1985; 1: 8-VII-1983; 1: 23-VIII-1984. Additional paratype males: 7: Orange Co., Walt Disney World, S16, T24S, R27E, 1-5-VII-1996, Z. Prusak, sand pine/oak scrub zone MW-5 (unburned), Malaise trap; 2: Orange Co., Walt Disney World, 8-13-VIII-1996, Z. Prusak, MW-5 Cons. Area unburned zone, sand pine scrub, Malaise trap.

Deposition of types

Holotype, 3 paratype dealate queens, 32 paratype workers, 12 paratype males: Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts; 6 paratype workers, 4 paratype males: The Natural History Museum, London; 7 paratype workers, 4 paratype males: Los Angeles County Museum, Los Angeles, California; 10 paratype workers, 4 paratype males: Florida State Collection of Arthropods, Gainesville, Florida; remaining type material in the arthropod collection of the Archbold Biological Station, Lake Placid, Florida.

Etymology

relictus, past participle of relinquo: left behind, referring to the distribution of the species on relict patches of Florida scrub and sandhill vegetation on high sand ridges in south-central Florida.

Relationship between relictus and other species

Workers of relictus and clarus are morphologically similar, except for the striate basalar lobe (Fig. 2K) and consistently spinose petiole of relictus. Until we compared the males of the two species, we had interpreted these differences as small divergences between widely separated populations of a single species, especially since clarus shows considerable variation through the Southwest. The differences between males are the only clear evidence at present that clarus and relictus are distinct. The extraordinarily large eyes and ocelli of relictus (Fig. 2A) suggest that there is some feature of relictus flight behavior that is different from the flight behavior of clarus; any major difference may be likely to confer reproductive isolation. A possibly relevant feature of relictus flight behavior is that male activity, as monitored by Malaise traps, seems to be concentrated around moonlit nights (Deyrup et al. 1985). Another major structural difference is the longer, less declivitous propodeum of relictus; all four species occurring in the U.S. show conspicuous differences in the shape and sculpture of the propodeal area (Figs. 2E-H).

It is tempting to hypothesize that relictus is closely related to clarus, representing an eastern offshoot of a western lineage adapted to dry habitats. Some other animals, such as the Florida sand roach (Arenivaga floridensis Caudell), the Florida scrub-jay (Aphelocoma coerulescens) and the gopher tortoise (Gopherus polyphemus) seem to be examples of western lineages in relic desert-like habitats in Florida (Deyrup 1989). These animals could have spread east, along with a rich fauna of now extinct savanna-dwelling vertebrates, along an arid corridor in southern North America during the late Pliocene through mid-Pleistocene (Webb 1990). In Deyrup’s 1989 and 1990 papers relictus, under the name of clarus, is specifically mentioned as an example of such a western lineage. Male relictus, however, do not support this hypothesis. They share their large eyes and ocelli, non-carinate scutellum and low propodeum with the other native southeastern species, brunneus (Figs. 2C, G). When the taxonomic status and distribution of southeastern brunneus is better understood, it may be possible to propose a new hypothesis on the derivation of relictus.

All four of the U.S. species of Odontomachus are combined with 17 New World species and two Old World species in the haematodus species group, distinguished by the reduction of the segments of the labial palps from four to three (Brown 1976).

Habitat and distribution of relictus

This species is a subterranean nester, and found only in areas of deep, unconsolidated, silica sand. These areas may be covered with Florida scrub vegetation, consisting of scattered pines, small oaks and other small trees and shrubs. Sometimes there are areas of bare sand, especially in the first few years following a fire. Alternatively, areas where relictus occurs may be sandhill vegetation, consisting of scattered pines above a low layer of grasses and forbs. For descriptions of these habitats, see Myers (1990). Nest entrances are not marked by a mound, but by scattered pellets of sand. Digging into a nest may produce a few workers, sometimes with brood, but no large aggregation of workers.

Odontomachus relictus is known from the Lake Wales Ridge, the southern Brooksville Ridge and the Orlando Ridge. It has not been found in scrub or sandhill habitats on the Atlantic Coastal Ridge, the Northern Brooksville Ridge, or the sandy uplands of northern Florida. The inland south-central sand ridges of the Florida Peninsula are over a million years old, and are known to have many plants and animals found nowhere else (Deyrup 1989; Deyrup 1990). Some of these species appear to be remnant populations of species that were once more widespread, others are probably true (autochthonous) endemics. Up to the discovery of the distinct status of O. relictus, it seemed that this ant was an example of a series of remnant populations; now it appears that this species could just as easily be a true endemic of south-central Florida.

The restricted range and habitat of O. relictus might raise questions about its conservation status. About twenty-five years ago, its prospects would have seemed poor. At that time there were only two protected populations, one on the Lake Wales Ridge (Archbold Biological Station), the other on the southern Brooksville Ridge (Withlacoochee State Forest). Upland areas were being converted rapidly to housing and agriculture, and it seemed that few scrub and sandhill areas would remain within the range of O. relictus. Since that time, development and habitat destruction have occurred at an unprecedented rate, but the establishment of ecological preserves also has been remarkably fast, especially on the Lake Wales Ridge. This species now appears to be adequately protected, unless it is subjected to some widespread environmental threat that pervades natural areas. Odontomachus relictus is a good example of a species that seemed destined for the endangered species list, with all the trouble and expense implied in such listing, but was preemptively protected by more general conservation programs.

The following keys distinguish between the four species known from the U.S. Brief comments on distribution, nomenclature and natural history follow the keys.

Notes on Species

Odontomachus brunneus. This species apparently occurs throughout Florida, although there are no records from the three westernmost counties. It is also known from southern Georgia and Alabama. We have seen specimens from low coastal areas in Alabama, and there is no obvious reason why it should not occur in coastal Mississippi, Louisiana and Texas, although it is not reported from any of those states. A report of brunneus from Cuba (Fontenla 1997) might refer to some other species, as we have seen specimens that would key to brunneus from the Dominican Republic, but are probably closer to O. insularis Guerin. Its distribution in Central and South America is unclear, since this species was combined with O. ruginodis in Brown’s revision of the genus (1976). Associating the various brunneus-like forms with their males in Central and South America and the West Indies would be an interesting and useful project for local myrmecologists, and might easily yield distributional surprises or new species. Workers can easily be distinguished from other U.S. species by the densely hairy gaster.

Southeastern records of insularis in the Formicidae section of the Catalog of the Hymenoptera (D. R. Smith 1979) refer to brunneus. Although the catalog appeared several years after Brown’s revision, the cut-off date for changes in the Formicidae section was mid-1975. Smith’s treatment of North American Odontomachus differs from those of Creighton (1950) and M. R. Smith (1951) in elevating to species level three subspecies of O. haematodus (Linnaeus). This was backed by no taxonomic references, and certainly was not intended to compete with the earlier, but unavailable, revision by Brown.

Odontomachus brunneus occurs in both well-drained and poorly drained habitats; nests may be in soil or in rotten wood. This species, along with many others, was studied by Van Pelt (1958) at the Welaka Reserve (now Welaka State Forest) in Putnam Co., Florida. Van Pelt found many colonies, which occurred in all the terrestrial habitats in the area, including flatwoods, mesic forest, swamp forest, upland scrub and sandhill. Nests were in various microhabitats, including deep leaf litter, fallen logs, at the bases of trees, and open or sparsely covered sandy areas. At the Archbold Biological Station, brunneus occurs in moist habitats, including flatwoods, bayheads, the edges of seasonal ponds, and elevated tussocks or fallen pines within seasonal ponds. It has not been found in the more elevated upland areas of the Station, which are occupied by relictus. This distribution gives the impression that there is some competitive displacement based on differential adaptation to moisture conditions, but the evidence remains circumstantial. It may be relevant that in parts of its range devoid of relictus, where brunneus occurs in dry, upland areas individuals never achieve the large size and dark color seen in some specimens from wet areas. Nobody knows, however, whether the smaller, paler individuals represent stressed individuals in suboptimal conditions, or whether they represent an adaptive phenotypic response in a robust population.

Workers of brunneus sometimes emerge to forage on cloudy days, but are generally nocturnal. The formidable jaws of brunneus are not used as assertively as one might expect, and there is fragmentary evidence that brunneus is sensitive to chemical defenses. Prey are approached tentatively, and the ant recoils immediately after striking the prey (Brown 1976). There may be a delay before the prey is picked up and carried away; Brown (1976) suggested that these ants react to chemical defenses, which are allowed to dissipate before the prey is retrieved. Alex Wild, while a student at the Archbold Biological Station, twice observed brunneus retreating hastily when confronted by aroused workers of Dorymyrmex bureni (Trager) (unpublished natural history notes on file at the Archbold Biological Station). Dorymyrmex bureni is much smaller than O. brunneus, but can release large quantities of defensive chemicals that are pungent to the human nose. Van Pelt (1958) reported accumulations of brunneus head capsules in the nests of Formica archboldi Smith, and suggested the possibility that brunneus is a regular part of the diet of F. archboldi. If this is the case, it is more likely that the brunneus are subdued by chemical means than by mandible-to-mandible combat.

During this study specimens were examined from the following areas: FL: Alachua, Baker, Bay, Bradford, Brevard, Broward, Citrus, Clay, Collier, Columbia, Dade, De Soto, Dixie, Duval, Franklin, Gadsden, Gilchrist, Glades, Hamilton, Hernando, Highlands, Hillsborough, Indian River, Jackson, Jefferson, Lake, Lee, Leon, Levy, Liberty, Madison, Marion, Martin, Monroe, Nassau, Okeechobee, Orange, Osceola, Palm Beach, Pasco, Polk, Putnam, Sarasota, St. Lucie, Sumter, Taylor, Volusia, Wakulla, Walton Counties; GA: Clinch County; AL: Baldwin, Houston Counties.

Odontomachus clarus. This species is known from northern Mexico and from Texas, New Mexico and Arizona. It is the only Odontomachus known from the southwestern U.S. and northern border of Mexico, so specimens may be identified tentatively by their source alone. The similar species ruginodis will probably be transported to the southwestern U.S.; ruginodis has conspicuous striae on the posterior face of the petiole (Fig. 2I). It is also possible that the species we consider clarus includes unrecognized cryptic species. If this is the case, the occurrence of additional species might be detected by finding more than one form of male within the range of clarus. We have seen males associated with workers of clarus from two widely separated sites: Cochise Co., AZ, and Jeff Davis Co., TX. In our experience, it is difficult to find workers of clarus with associated males. Examination of specimens from light traps or flight traps might be a convenient way to establish whether there is more than one western species. Records of clarus from the West Indies (Smith 1979) refer to some other species, perhaps ruginodis. All references to clarus in Florida (Deyrup et al. 1985; Deyrup 1989; Deyrup et al. 1989; Deyrup 1990; Sivinski et al. 1998) should be referred to relictus, as discussed above.

The species names coninodis Wheeler and desertorum Wheeler, listed in the 1979 catalog, were synonymized under clarus by Brown (1976). Brown reported that “coninodis,” which has a blunt petiolar spine, occurs at the higher elevations in Arizona, in isolated, low mountain ranges, surrounded by lower areas occupied by typical clarus with an elongate petiolar spine. The distribution of the two forms is unlike that of a normal pair of geographic subspecies, and Brown characterized the short-spined forms as “depauperate ecotypes or ecophenotypes.” As in the case (mentioned above) of the smaller, paler brunneus found in dry sites, it seems premature to apply the pejorative “depauperate” to a condition that, for all we know, could be a superb adaptive response.

In Arizona this species is found, usually in small numbers, under rocks and grass tussocks, in both dry and mesic sites. In western Texas it shows a preference for more mesic sites and fine soils; nests are usually found under rocks or logs (Cokendolpher & Francke1990).

During this study specimens were examined from the following sites (we provide more detailed site information for clarus than for brunneus or ruginodis because of evidence of geographic variability in clarus). AZ: Cochise Co. (Chiricahua Mts: Cave Creek, Texas, and Idlewild Canyons), Pima Co. (Tucson), Santa Cruz Co. (Patagonia Mts., Pajarito Mts.); TX: Bosque Co. (Meridian), Brewster Co. (Big Bend National Park: Rio Grande Village), Denton Co. (no locality), Jeff Davis Co. (Davis Mts.), Travis Co. (Bull Creek, McNeil); MEXICO: Chihuahua (Riva Palacio, Guerero, Conchos), Coahuilla (25 km E. of Saltillo), Cuernovaca (no locality), Guanjuato (highway 57 km 57), Hidalgo (San Miguel), Jalisco (Guadelajara), Mexico (Pedrigales), Nuevo Leon ((Monterrey), Queretaro (3 mi. W. of Queretaro).

Odontomachus ruginodis. This species occurs sporadically through southern and central Florida, at least as far north as Orlando, and also in the West Indies. Its distribution in South and Central America is unclear because it has been confused with brunneus. Its ability to thrive in disturbed habitats should allow it to invade mainland Neotropical areas, if it is not already present. It is probable that this species will be distributed by commerce to disturbed areas in the Southwest. The conspicuous striae on the posterior face of the petiole (Fig. 2I) distinguish workers of this species from the similar clarus and relictus, but there are additional species with petiolar striae (e.g., O. bauri Emery) outside the U.S.

The name ruginodis was first applied by Wheeler (1905), and Wheeler was designated as the author of ruginodis in Deyrup et al. (1985) and Deyrup et al. (1989). The name ruginodis, however, was first used as a quadrinomial (Odontomachus haematodus insularis ruginodis), and therefore is not a valid name under the rules of nomenclature. The first use of ruginodis as a trinomial, or subspecies (Odontomachus haematodus ruginodis), was by M. R. Smith (1937). Since this is the first valid use of ruginodis for this species, M. R. Smith is the author of the name. This, along with hundreds of other tangles of nomenclature, was straightened out in Bolton’s 1995 catalog of ants.

In Florida, this species occurs in disturbed areas, including urban and suburban habitats. It occurs along the beaches in the tropical part of the state. It has not yet been found inland in natural habitats. In Puerto Rico it differs from another sympatric species (perhaps O. bauri) in its preference for open, sunny areas, especially river bottoms (Smith 1937).

The defensive mandible-snapping behavior of ruginodis was studied by Carlin and Gladstein (1989). When a nest is attacked by other ants, the ruginodis workers rush out, snapping at anything that seems a threat. Enemy ants may be dismembered or knocked out of the way by the mandibular strikes. If the mandibles hit a solid object, the ruginodis may itself be flung into the air for a distance of several centimeters. This does not seem to be an escape mechanism, as the worker, upon landing, immediately charges back into the fray. The nest entrance is usually guarded by a single worker, who stands with cocked mandibles near the entrance. If an intruder approaches within striking distance, the mandibles snap shut, responding to signals from the antennae and long sensory hairs at the bases of the mandibles. The heavy apices of the mandibles do not slice into the intruder, but knock it away a distance of about one to fourteen centimeters. Carlin and Gladstein call this the “bouncer defense.”

During this study specimens were examined from the following areas: FL: Brevard, Broward, Charlotte, Collier, Dade, Glades, Hendry, Highlands, Hillsborough, Indian River, Lee, Manatee, Martin, Monroe, Orange, Palm Beach, Pinellas, Orange, Polk, St. Lucie, Volusia Counties; BERMUDA; BAHAMAS: New Providence, San Salvador, Rum Cay, North Andros Islands; PUERTO RICO: Rio Grande.

Residual Problems

There are still some questions on the taxonomy and distribution of the Odontomachus species that occur in the U.S. The distribution of brunneus and ruginodis is unclear. Does brunneus occur in coastal wetlands around the Gulf of Mexico? Is brunneus as it appears in the southeastern North America the same species as the brunneus populations reported from the West Indies and the mainland Neotropics? Does ruginodis occur outside the West Indies and Florida? Another kind of question deals with the morphological divergences between males of different species. Do the kind of differences we have reported relate to differences in ecology and behavior of the species? Will Odontomachus males prove useful in distinguishing species throughout the Neotropics? The few males of insularis and bauri that we have seen show conspicuous species differences, but there could be species complexes that cannot be elucidated by male morphology.