Introduction

Beetles in the genus Anthia Weber are some of the largest and most conspicuous representatives of the family Carabidae in sub-Saharan Africa (Scholtz and Holm 1985; Picker, Griffiths and Weaving 2002). Adults of Anthia species (Figure 1) and those of the closely related genus Termophilum Basilewsky (Figure 2) are boldly patterned in black and white or yellow stripes and/or spots. These beetles are armed with potent chemical defenses and are able to spray highly concentrated acidic secretions over a distance of a meter or more, often directed at the head and eyes of an attacker (Scott et al. 1975; Huey and Pianka 1977). Sympatric species of Anthia and Termophilum often have similar color patterns, a fact that has been interpreted as a possible example of Müllerian mimicry (Marshall and Poulton 1902; Huey and Pianka 1977). The genus Anthia and its relatives offer remarkable opportunities for studies of chemical ecology, as well as aposematic color patterns and the evolution of mimetic coloration.

Unfortunately, studies of the ecology, evolution, and behavior of these beetles have long been hampered by the lack of reliable, illustrated identification materials for species of Anthia and related genera. The two older revisions by Péringuey (1896) and Obst (1901), and the catalogue of species by Csiki (1929) often conflict in the placement of individual taxa, and these historic works generally lack good habitus images or line drawings that could be used by non-specialists to identify species in this group. Fortunately, recent revisionary work by (Schmidt (2001, 2002) and Schmidt and Gruschwitz (2002) has helped to clear up many longstanding nomenclatural issues in this group and provide a firm foundation for the development of reliable identification materials.

The present treatment was written to provide identification materials for species of Anthia (sensu Basilewsky 1950) from the present-day Republic of South Africa (RSA). This paper forms part of a series of studies on southern African Carabidae and Cicindelidae, with a particular focus on taxa associated with the Kruger National Park in the Republic of South Africa. Previous contributions in this series include (Mawdsley and Sithole (2008, 2009), (Mawdsley (2009, 2011), and Mawdsley et al. (2011). The goal of this series of publications is to provide high-quality identification materials for groups of ground beetles and tiger beetles which are of potential interest to conservation biologists, environmental scientists, and park and natural area managers.

Unlike many parts of sub-Saharan Africa, the Republic of South Africa has been well sampled for Anthia and related genera and there is a wealth of museum material available for study. Even with this wealth of material, there has been considerable confusion in the literature regarding the identification and appropriate names to assign to these species. At the level of generic names, many authors follow Péringuey (1896), Obst (1901), and Csiki (1929) in applying the genus name Anthia to a larger group of approximately 66 species from sub-Saharan Africa, southern Asia, and India. In this paper, we follow Basilewsky (1950), (Schmidt (2001, 2002), Schmidt and Gruschwitz (2002), and (Lorenz (2005a, 2005b) in restricting the use of the genus name Anthia Weber to the group of approximately fourteen species exhibiting sexual dimorphism in the structure of the pronotum and/or mandibles.

There has also been considerable confusion regarding the number of species of Anthia in southern Africa. Csiki (1929) represents the most conservative estimate, recognizing just two species, Anthia maxillosa (which he called Anthia fabricii) and Anthia thoracica. Péringuey (1896) recognized an additional five species for a total of seven. In this paper, we recognize four species that can be readily separated on the basis of attributes of pronotal and elytral vestiture and surface sculpture, as well as by the structure of the male genitalia. Our conclusions parallel those of (Schmidt (2001, 2002) and Schmidt and Gruschwitz (2002) as well as the late Pierre Basilewsky, who studied this genus for many years. According to Basilewsky’s identification labels on museum specimens that we examined, Basilewsky recognized the same four species from South Africa that we recognize here.

Materials and methods

We examined specimens of species of Anthia, Termophilum, and allied genera in the collections of the following museums: Field Museum of Natural History, Chicago, Illinois (FMNH); Kruger National Park Museum (Scientific Services), Skukuza, South Africa (KNPC); South African National Collection of Insects, Pretoria, South Africa (SANC); National Museum of Natural History, Smithsonian Institution, Washington, D.C. (NMNH); Transvaal Museum, Pretoria, South Africa (TMSA).

Field observations on adults of Anthia thoracica and other Anthiini were conducted during month-long visits by the senior author to the Kruger National Park in 2007, 2008, 2009, and 2010. Systematic field surveys for Anthiini and other diurnally active Carabidae were conducted in areas where adults of Anthiini had been collected historically in the park, or where adult beetles had been observed recently by park staff. Our surveys focused primarily on the Skukuza Ranger District in the central portion of the park, with field trips north to Satara, Letaba, Olifants, and Shingwedzi and south to Pretoriuskop. In surveying these areas we employed a variety of techniques, including driving surveys, walking surveys, and pitfall trapping (Mawdsley et al. 2011). Driving surveys were quickly identified as the most productive of these survey techniques for large Anthiini and consequently were widely applied throughout the southern area of Kruger National Park. In this survey approach, a party of four searchers drove slowly in a car along secondary sand or gravel roads and stopped whenever large carabid beetles were observed running on or across the road, or in vegetation along the side of the road. When a beetle was observed, the car was stopped and one or more persons left the vehicle in order to capture the beetle. In walking surveys, a group of four searchers walked slowly along segments of sand and gravel roads, looking for Carabidae running among dead leaves, vegetation, or on bare ground. We also deployed pitfall traps of different diameters in areas where Anthiini were observed, although we found through experience that our small-diameter traps were less productive at capturing large-bodied Carabidae and were frequently raided by baboons, mongoose, and other predatory vertebrates. We conducted field surveys under all weather conditions, from sunny days with no cloud cover, to sunny days with afternoon thunder showers, to overcast, rainy days. Ambient air temperatures during surveys varied from 18C to 35C.

The driving technique proved to be a particularly productive method for collecting Anthiini, as well as other large Carabidae such as species of Tefflus Leach (Mawdsley et al. 2011). In addition to adults of Anthia thoracica, this method resulted in capture of adults of Termophilum burchelli (Hope), Termophilum homoplatum (Lequien), Termophilum massilicatum (Guérin), and Cypholoba graphipteroides (Guérin). The driving technique has the advantage of being able to cover a large geographic area within a relatively limited amount of time. The disadvantage to this technique is that many smaller-bodied species of Carabidae and Cicindelidae are overlooked, particularly those that are cryptically colored. However, we found that these taxa (particularly Carabidae of the genus Graphipterus Latreille and Cicindelidae of the genera Dromica Dejean and Lophyra Motschulsky) were easily detected during walking surveys.

For studies of beetle biology in captivity, individual adult beetles were captured by hand and placed singly into large 4-liter plastic holding containers containing a shallow layer of sand and gravel in the bottom. Each container was provided with a small ball of cotton soaked in water, to provide a water source for the adult beetles. To examine prey preferences, we provided each captive beetle with a variety of potential food items, primarily insects and other arthropods which were collected at lights at night in the Skukuza research camp. We recorded acceptance/rejection of each potential food item and the associated order and family of each prey item offered to the beetles.

Voucher specimens of beetles collected in Kruger National Park are deposited in the KNPC, NMNH, and TMSA collections.

Attributes of the abdominal ventral sterna are referred to using the numbering system generally accepted in Carabid studies, i.e., the sternum divided medially by the hind coxae is sternum II (the first being hidden) and the last visible is sternum VII (Liu et al. 2011).

Taxonomic note on Apristus promontorii Péringuey

Obst (1901) erroneously included an “Anthia promontorii Péringuey (1898)” in his list of species of the genus Anthia, with a comment that he had not actually examined the original description of the species. Had Obst consulted Péringuey’s original description (Péringuey 1898:329), it would have been readily apparent that the species in question was described in the genus Apristus Chaudoir (Coleoptera: Carabidae: Lebiini) rather than Anthia Weber. Furthermore the species could not possibly belong to the tribe Anthiini, as its maximum dimensions were given as 3.5 mm in length and 1.5 mm in width, measurements that are far smaller than those of any known species of Anthiini. Following Obst, Csiki (1929:381) listed Anthia promontorii Péringuey as a species incertae sedis under the genus Anthia in the Coleopterorum Catalogus. This placement has been followed by other authors and cataloguers, although clearly incorrect. The name Anthia promontorii should be treated as a synonym of Apristus promontorii.

Comparative distribution of Anthia species in the Republic of South Africa

Figures 32–35 illustrate the distribution of species of Anthia in the Republic of South Africa based on museum specimen records, literature records, and our own recent collections. The individual species show markedly different patterns of distribution. Anthia thoracica has been most frequently collected in the northern and eastern portions of the country (Figure 32), particularly in Gauteng, Limpopo, and Mpumalanga Provinces. Anthia maxillosa has been most frequently collected in the west (Figure 33), particularly the Western Cape and Northern Cape Provinces. Anthia cinctipennis has a distribution pattern somewhat similar to that of Anthia thoracica (Figure 34), with a high concentration of records in Gauteng, Limpopo, and Mpumalanga Provinces, while Anthia circumscripta appears to be rare everywhere in the Republic of South Africa (Figure 35), with very few literature or specimen records. As noted above, Anthia circumscripta has been collected in much greater numbers in Botswana and Namibia, and forms a characteristic part of the Anthia fauna in both the Kalahari and Namib deserts.

Association of Anthia species with terrestrial ecoregions and veld types in the Republic of South Africa

Schmidt (2001, 2002) and Schmidt and Gruschwitz (2002) were the first to note that species of Anthiini appear to be closely associated with particular vegetation communities in southern Africa. We extend this analysis here using a new map and classification of terrestrial ecoregions that have been developed for continental Africa and Madagascar by Burgess et al. (2004). Comparisons between the map presented by Burgess et al. (2004) and the species distribution maps presented in this paper (Figures 32-35) clearly show the associations between the distributions of individual species of Anthia and the terrestrial ecoregions recognized by Burgess et al. (2004) in southern Africa. To examine the association of Anthia species with vegetation communities, we also compared our Anthia distribution maps against an older but nonetheless widely-used vegetation map which shows the distribution of various veld types in South Africa (Acocks 1988). For this comparison, we used Acocks’ Map 2, which is intended to show actual (as opposed to historic or potential future) vegetation communities in the Republic of South Africa.

Adults of species of Anthia are widely distributed throughout the Republic of South Africa (Figures 32–35) and have been collected in twelve of the seventeen terrestrial ecoregions recognized by Burgess et al. (2004). The five ecoregions with no records of Anthia species are: Kalahari Acacia Woodlands; KwaZulu-Cape Coastal Forest Mosaic; Knysna-Amatole Montane Forests; Montane Fynbos and Renosterveld; and Southern African Mangroves. In some cases (e.g. Southern African Mangroves) this absence of records may represent a true absence of Anthia beetles from the ecoregion, while in other cases (e.g. Kalahari Acacia Woodlands) the absence of records is likely due to the relatively small amount of this particular ecoregion that occurs in the Republic of South Africa and the absence of surveys for Anthia species in those particular small habitat patches.

Adults of Anthia thoracica have been collected in eight of the seventeen terrestrial ecoregions recognized by Burgess et al. (2004) in the Republic of South Africa: Highveld grasslands (36.1% of collecting sites); Zambezian and Mopane Woodlands (19.7% of collecting sites); Southern Africa bushveld (13.1% of collecting sites); Nama Karoo (11.5% of collecting sites); Albany Thickets (8.2% of collecting sites); Maputaland Coastal Forest Mosaic (4.9% of collecting sites); Lowland Fynbos and Renosterveld (4.9% of collecting sites); and Drakensburg Alti-Montane Grasslands and Woodlands (1.6% of collecting sites).

Adults of Anthia maxillosa have been collected in eight of the seventeen terrestrial ecoregions recognized by Burgess et al. (2004) in the Republic of South Africa: Succulent Karoo (48.3% of collecting sites); Nama Karoo (25.9% of collecting sites); Lowland Fynbos and Renosterveld (6.9% of collecting sites); Kalahari Xeric Savanna (5.2% of collecting sites); Highveld Grasslands (5.2% of collecting sites); Albany Thickets (3.4% of collecting sites); Southern Africa Bushveld (3.4% of collecting sites); and Drakensburg Montane Grasslands, Woodlands, and Forests (1.7% of collecting sites).

Adults of Anthia cinctipennis have been collected in eight of the seventeen terrestrial ecoregions recognized by Burgess et al. (2004) in the Republic of South Africa: Highveld grasslands (36.2% of collecting sites); Southern Africa Bushveld (25.5% of collecting sites); Nama Karoo (14.9% of collecting sites); Zambezian and Mopane Woodlands (10.6% of collecting sites); Kalahari Xeric Savanna (6.4% of collecting sites); Succulent Karoo (2.1% of collecting sites); Maputaland-Pondoland Bushland and Thickets (2.1% of collecting sites); and Drakensburg Montane Grasslands, Woodlands, and Forests (2.1% of collecting sites).

Adults of Anthia circumscripta have been collected in three of the seventeen terrestrial ecoregions recognized by Burgess et al. (2004) in the Republic of South Africa: Nama Karoo (37.5% of collecting sites); Highveld Grasslands (37.5% of collecting sites); and Drakensburg Montane Grasslands, Woodlands, and Forests (25% of collecting sites). In adjacent countries (Namibia, Botswana, Zimbabwe), this species has also been collected in the Namib Desert, Kalahari Xeric Savanna, and Southern Africa Bushveld ecosystems.

To compare these findings across the four Anthia species and the seventeen terrestrial ecoregions in the Republic of South Africa, we used the Spearman rank correlation coefficient (Lowry 2011). We use the Spearman rank correlation coefficient rather than standard regression because we do not know whether the numbers of collecting sites are distributed normally across the set of ecoregions (Lowry 2011). The Spearman rank correlation coefficient tests the significance of correlation between two lists of values. A test statistic close to 1 or -1 indicates that the relative rankings expressed in the two lists of values are markedly correlated. To obtain lists of values for each of the four species of Anthia, we calculated the percentage of the total number of collecting sites in each of the seventeen terrestrial ecoregions listed by Burgess et al. (2004) for the Republic of South Africa (Table 1). Next, the Spearman rank correlation coefficient was calculated for each of the possible comparisons of the lists of values in each of the species columns in Table 1, using an online calculator developed by Wessa (2011). These coefficients are reported in Table 2. Inspection of Table 2 shows clearly that for all four species, the relative percentages of collecting sites in each ecoregion are not significantly similar at the 95% level (test statistic between .95 and 1.0 or between -.95 and -1.0). From these findings, it can be inferred that these four species appear to be distributed across the set of southern African ecoregions in statistically dissimilar ways. These results make intuitive sense, given the obvious visual differences in the species distribution maps for these taxa (Figures 32-35).

Adults of Anthia species are likewise broadly distributed across the various veld types or large-scale vegetation communities recognized by Acocks (1988) in southern Africa, with collection records of species of Anthia from eight of the ten veld types recorded for the Republic of South Africa. The two veld types with no records of Anthia species are Scrubby mixed Grassveld and Forest and Scrubforest. It is possible that the lack of records of Anthia species from these two veld types is due simply to lack of sampling activities in these two communities, both of which are relatively restricted in geographic extent (Acocks 1988).

Adults of Anthia thoracica have been collected in six of the ten veld types: Bushveld (51.6% of collecting sites); Karoo (11.3% of collecting sites); Sweet Grassveld (9.7% of collecting sites); Mixed Grassveld (9.7% of collecting sites); Sour Grassveld (9.7% of collecting sites); and Fynbos (8.1% of collecting sites).

Adults of Anthia maxillosa have been collected in seven of the ten veld types: Karoo (63.1% of collecting sites); Desert (14.3% of collecting sites); Bushveld (10.7% of collecting sites); Fynbos (7.1% of collecting sites); Succulent Karoo (2.4% of collecting sites); Sweet Grassveld (1.2% of collecting sites); and Sour Grassveld (1.2% of collecting sites).

Adults of Anthia cinctipennis have been collected in six of the ten veld types: Bushveld (45.8% of collecting sites); Sour Grassveld (29.2% of collecting sites); Karoo (18.8% of collecting sites); Desert (2.1% of collecting sites); Sweet Grassveld (2.1% of collecting sites); Mixed Grassveld (2.1% of collecting sites).

Adults of Anthia circumscripta have been collected in just three of the ten veld types: Sour Grassveld (50%); Karoo (37.5%); and Bushveld (12.5%).

We again applied the Spearman rank correlation coefficient (Lowry 2011) to examine the relative abundance of collecting sites of the four species of Anthia across the ten veld types. Table 3 shows the percentage of collecting sites for each species in each of the veld types. Table 4 shows the Spearman rank correlation coefficient for each pairwise comparison of columns in Table 3, calculated using the web calculator developed by Wessa (2011). Here again the pairwise comparisons between columns are not significantly similar at the 95% level. This findings suggests that the four species of Anthia are distributed across the veld type of South Africa in statistically dissimilar ways.

Seasonality of Anthia species in the Republic of South Africa

The histograms in Figures 36-39 summarize dates of collection of museum specimens of Anthia species across all years. These charts suggest that individual species of the genus Anthia exhibit marked seasonality in the timing of adult emergence and in the timing of adult activity patterns, as noted previously by Schmidt (2001) and Schmidt and Gruschwitz (2002). In general, emergence and activity patterns in adults of Anthiini appear to be coordinated with the onset of seasonal rains (Schmidt 2001; Schmidt and Gruschwitz 2002). As shown in Figure 36, adults of Anthia thoracica were most frequently collected between October and March, corresponding to the seasonal monsoonal rains in the eastern portion of the country (du Toit et al. 2003), where Anthia thoracica has been most frequently collected (Figure 32). In contrast, most collections of Anthia maxillosa have occurred between August and October (Figure 37), corresponding to the period of winter rains in the Northern and Western Cape provinces (Mares 2002) where this species has been most frequently collected (Figure 33). Anthia cinctipennis and Anthia circumscripta exhibit similar bimodal activity patterns in the Republic of South Africa (Figures 38 and 39), with populations in the eastern portion of the country active during the summer monsoon rains (September-January and November-March, respectively; Figures 38 and 39), while Kalahari Desert populations of both species are active during July.

Investigations of possible allometry in secondary sexual characteristics of males of Anthia species

Males of species of Anthia in the Republic of South Africa often have enlarged and elongated mandibles as well as distinctive flanges on the base of the pronotum (Figures 3–8, 12–15, 18–20, 23–25). Both mandibles and flanges come in large, small, and intermediate sizes, and in general, larger-bodied males appear to have larger mandibles and pronotal flanges than smaller males. We were interested in determining whether mandible and pronotal flange size of these beetles scale isometrically (mandible and flange size increases at the same rate as body size) or allometrically (mandible and flange size increase at a rate disproportionately greater than body size). We measured three variables, right mandible length, pronotal length (from apex to base of flange), and elytral width (a surrogate measure of body size, measured at the widest point with the elytra fully closed in a natural position) for all male specimens of these four Anthia species in the NMNH collection. For Anthia thoracica, we observed only a very weak correlation between elytral width and mandible length (R² = 0.61) and between elytral width and pronotal length (R² = 0.63) using standard linear regression. Both R² values improved with the use of a polynomial (x²) function in the regression, to 0.67 and 0.75, respectively. The improved fit with the polynomial function and the generally low R² values from the standard linear regression suggests that both mandible size and pronotal flange size scale non-linearly with respect to body size in this species. In the other three species, mandible length and pronotal length were very poorly correlated with elytral width (R² values ranging from 0.04 to 0.27), suggesting that mandible and pronotal flange size is not correlated with elytral width.

South African Anthiini as vectors for Cyaneolytta larvae (Coleoptera: Meloidae)

Many large-bodied Carabidae in southern Africa have been documented as serving as vectors for the phoretic larvae of species in the blister beetle genus Cyaneolytta Péringuey (Bologna et al. 1990; Di Iulio et al. 2003). In the Republic of South Africa, larvae of these beetles have been collected in association with adults of all four species of Anthia as well as several sympatric species of Termophilum (Di Iulio et al. 2003). The nature of these phoretic associations remains unclear (Di Iulio et al. 2003) although a number of possible explanations have been advanced by Bologna et al. (1990).

Observations on the biology and behavior of Anthia thoracica (Thunberg) in the Kruger National Park

Anthia thoracica is one of two species in the genus Anthia recorded from Kruger National Park, a large and well-known conservation area which is located in eastern Limpopo and Mpumalanga Provinces, South Africa (du Toit et al. 2003). Recent fieldwork in the southern portion of this park by the senior author has provided us with opportunities to examine certain aspects of the biology of Anthia thoracica under both field and laboratory conditions. Protocols for field surveys for Anthia thoracica and other Anthiini, as well as methods for laboratory maintenance of captive Anthiini, are described above under the Materials and Methods section.

Occurrences within Kruger National Park: Specimens of Anthia thoracica have been collected at the following localities within Kruger National Park: Malelane, Numbi Gate, Pumbe Sands, Shingwedzi, Skukuza, and Stolsnek. In addition, the senior author and associates have recently encountered this species on sand and gravel roads in the vicinity of the Pretoriuskop and Skukuza rest camps, and in the N’waswitshaka Research Camp at Skukuza.

Associated ecological communities: Gertenbach (1983) characterized a series of 35 “landscapes” in Kruger National Park which are defined on the basis of plant communities, climate, geology, soils, vegetation, and vertebrate species composition. Adults of Anthia thoracica have been collected in the following five landscapes within the Park: Lowveld sourveld bushveld of Pretoriuskop (Gertenbach Landscape 1; Pretoriuskop; Figure 40). Malelane mountain bushveld (Gertenbach Landscape 2; Malelane, Numbi Gate, Stolsnek). Thickets of the Sabie and Crocodile Rivers (Gertenbach Landscape 4; N’waswitshaka Research Camp, Skukuza; Figure 41). Combretum spp./Colophospermum mopane (Bentham) Leonard (Fabaceae) rugged veld (Gertenbach Landscape 22; Shingwedzi). Pumbe sandveld (Gertenbach Landscape 30; Pumbe Sands). Common geomorphological features in these landscapes include eroding granite, gneiss, or basalt koppies and hills. Extensive sand deposits, sometimes as deep as 6 meters, are also found in many of these landscapes. The silver cluster-leaf tree (Terminalia sericea Burchell, Combretaceae; Figure 40) is a major component of the vegetation in several of these landscapes (although less common in the vicinity of Skukuza; Gertenbach 1983).

Co-occurrence with allied species: We collected adults of Anthia thoracica in association with adults of of Termophilum burchelli (Hope), Termophilum homoplatum (Lequien), Termophilum massilicatum (Guérin), and Cypholoba graphipteroides (Guérin) (all Carabidae: Anthiini). Adults of Termophilum homoplatum and Termophilum massilicatum were found primarily in the open savanna areas extending northward area from Skukuza to Satara (Figure 42), while adults of Termophilum burchelli and Cypholoba graphipteroides were found in denser thickets and woodlands from Pretoriuskop in the south along the banks of the Sabie River to Skukuza (Figures 40-41). The related species Anthia cinctipennis Lequien has also been collected in Kruger National Park, with records from Skukuza, Satara, and the Nyandu Sandveld (Figure 42).

Activity period: Specimen records from South Africa (Figure 36) indicate that collections of Anthia thoracica have occurred between September and May, with a peak in November-December. Emergence of adults of Anthia thoracica appears to be triggered by the onset of seasonal rains (Schmidt 2001), which in Kruger National Park typically occurs in November or December (du Toit et al. 2003). Adult beetles are usually encountered singly, walking rapidly on sand or gravel roads or in open areas of the veld. Our field observations of this species in the Park were conducted during the early rainy season in November and December, corresponding to the time of peak collections of Anthia species and other Anthiini as indicated by museum specimen labels. No adults of Anthia thoracica were observed during a visit to the Park during the late dry season in September, 2006.

Weather and climate: Adult activity patterns of Anthia thoracica appear to be markedly influenced by weather and climatic conditions (Schmidt 2001). As mentioned above, adult emergence is clearly associated with the onset of seasonal rains (Schmidt 2001). In the Kruger National Park, adults can be found diurnally walking on sand and gravel roads and in open areas of the veld immediately following a rainfall event. After a series of days without rain, adult activity patterns change and the foraging adults are only encountered crepuscularly and nocturnally. On subsequent overcast or rainy days, however, adults again become active diurnally. Similar activity patterns with respect to weather and climatic conditions have been noted in populations of the carabid Tefflus meyerlei delagorguei in Kruger National Park (Mawdsley et al. 2011).

Defensive behaviors: Adults of Anthia thoracica spray copious amounts of highly concentrated formic acid from their pygidial glands when disturbed (Scott et al. 1975; Schmidt 2001). As noted by Huey and Pianka (1977), the spray is often directed towards the head and eyes of the person disturbing the beetle. If picked up by a human, the beetle often directs the spray towards the hands of the person disturbing it. There appears to be a limited supply of acid available to the beetle; the adults of Anthia thoracica and Termophilum species which we kept in captivity did not spray again after the initial capture. It has been suggested to the authors that the formic acid may be acquired and concentrated by the beetle as a by-product from consuming ants (Hymenoptera: Formicidae). However, other species of Carabidae have the ability to synthesize formic acid in their pygidial glands and to spray this acid as a defensive compound (e.g. Rossini et al. 1997).

Mimicry: Anthia thoracica and the sympatric species Termophilum homoplatum (Lequien) share similar color patterns which include large round or ovate eyespots, a black, shining dorsal integument, and a narrow white linear band along the outer margin of the elytra (Figures 1, 2). In Anthia thoracica, the eyespots are on the pronotum while in Termophilum homoplatum the eyespots are on the base of the elytra. Both species occur in similar habitats at the same times of year and exhibit similar fast walking behaviors. Both species are chemically defended, with the capacity to spray similar combinations of highly concentrated formic acid and other acidic compounds from the pygidial glands (Scott, Hepburn, and Crewe 1975). Marshall and Poulton (1902) suggested that the similarity in coloration between these two species may be an example of mimicry. Since both species share a similar noxious defensive behavior, any mimetic interactions would likely be an example of Müllerian mimicry (Wickler 1968). Wickler (1968) emphasizes the importance of identifying potential agents of selection which could be responsible for driving the evolution of mimetic resemblances between species. In Kruger National Park, potential predators which forage for terrestrial arthropods in the areas where Anthiine beetles are encountered and which might serve as agents of selection for mimetic interactions between Anthia and Termophilum species include chacma baboon (Papio ursinus (Kerr)), vervet monkey (Chlorocebus pygerythrus Cuvier), banded mongoose (Mungos mungo Gmelin), black-backed jackal (Canis mesomelas (Schreber)), southern ground hornbill (Bucorvus leadbeateri (Vigors)), and the secretary bird (Sagittarius serpentarius (Miller)). We observed foraging adults of all of these vertebrate species in the areas that we surveyed for Anthiine beetles, with chacma baboon, vervet monkey, and banded mongoose present in significant numbers (n > 10 individuals) at many sites. The abundance of these potential predators suggests that there may be considerable selection pressures on large carabid beetles that are driving the evolution of chemical defenses, aposematic coloration, and Müllerian mimicry complexes. On balance, the hypothesis of mimicry between these two beetle species seems reasonable.

Prey species: Adults of Anthia thoracica in captivity showed few preferences regarding prey items and consumed a wide range of prey species, including representatives of the following orders and families: Coleoptera: Carabidae, Cerambycidae, Curculionidae, Scarabaeidae, Tenebrionidae. Hemiptera: Cicadidae, Cydnidae. Hymenoptera: Formicidae. Isoptera: Termitidae. Lepidoptera: Arctiidae, Geometridae, Noctuidae, Saturniidae. Orthoptera: Acrididae. These insects were captured at lights at night in the N’waswitshaka Research Camp and fed to the live Anthia thoracica.

Foraging behavior: Adults of Anthia thoracica exhibit a rapid walking behavior which appears to serve multiple functions: foraging for food, detection of mates, and dispersal of adults. In captivity, this behavior was often noticed in adults of Anthia thoracica which had not been fed for several hours.

Prey detection: Adults of Anthia thoracica were observed to move and vibrate their antennae in the presence of potential prey items, suggesting that chemical cues may form an important part in the detection and recognition of prey items.

Predatory behaviors: Adults of Anthia thoracica seized prey using their mandibles and rapidly crushed prey items with the mandibular bases, using the labial and maxillary palpi to hold and manipulate the prey item. Liquid contents of prey and soft tissues were consumed whole while heavily sclerotized parts and appendages (legs, wings) were discarded.

Drinking: Captive adults of Anthia thoracica were observed drinking from wet cotton balls which we provided in their containers; in drinking, the adult beetle stands perpendicular to the water source, the ventral surface of the head is pressed against the wet cotton and water is then taken directly into the buccal cavity.

Antennal cleaning: Adults of both sexes of Anthia thoracica have an antennal cleaning notch along the inner margin of the protibiae. Adults were observed drawing the antennae through this notch after feeding, after handling by humans, and after being introduced into a new captive holding chamber.

Concluding Remarks

We hope that the identification materials presented in this paper will be of assistance to entomologists, field biologists, national park managers, and others who want to identify specimens of Anthia species in the Republic of South Africa. We also hope that this paper helps to spark additional interest in these large and spectacular members of the South African insect fauna. Species of the genus Anthia and related genera are large, common, and conspicuous members of savanna and woodland ecosystems throughout southern and eastern Africa. Our understanding of the natural history of this group is limited at present and further investigations of the life history, immature stages, and biology of these beetles are clearly needed. The genus Anthia and its relatives provide excellent opportunities for studying the evolution of chemical defenses, aposematic coloration, and Müllerian mimicry. Anthia and its relatives also have attributes which suggest they may be excellent candidates for inclusion in monitoring programs that track ecosystem condition or ecological integrity in southern Africa. Adult activity patterns of Anthia species closely track a variety of environmental and climatological variables (Schmidt 2001), and the presence of adult beetles could be taken as indicative of the presence of certain favorable climatic conditions (Schmidt 2001). Adults (and presumably larvae) of Anthia species and relatives are voracious predators (Schmidt 2001) and thus the presence of the adult beetles is indicative of a suitable arthropod prey base for both larval and adult development. Our results and those of Schmidt and Gruschwitz (2002) suggest that individual species of Anthia show close associations with particular ecoregions or vegetation communities, and thus species of these beetles could potentially also serve as indicators of community condition. The survey protocols for adults of Anthia and related genera are relatively straightforward (see Methods above) and can be easily replicated using small teams of surveyors. Identification of species in the genus Anthia and related genera is based largely on external surface sculpture and setal patterns (Schmidt 2002), which could be mastered by non-specialists. By making these basic identification materials widely accessible, we hope to stimulate further interest in these fascinating members of the southern African carabid fauna.