Cerapachys

Worker.-Small ants with peculiar, long, subcylindrical body; the head excavated behind, with prominent, depressed posterior corners and very short clypeus, with which the closely approximated frontal carinae are fused. The latter are erect, leaving the articulations of the antennae exposed. The antennal fovea is bounded externally by a distinct carina. Mandibles with distinct, obscurely denticulate apical border. Antennae stout, 9- to 12-jointed, the scape incrassated distally, the terminal funicular joint large, swollen, oval or glandiform, at least as long as the three preceding joints together, thus forming a one-jointed club. Eyes small, sometimes wanting. Thorax with the premesonotal and mesoepinotal sutures absent or indistinct. Petiole and postpetiole not marginate on the sides, the latter strongly constricted off from the gaster which is largely formed by its first segment.

Female scarcely larger than the worker and very similar, sometimes apterous and ergatoid. Fore wings when present with a discoidal and a single cubital cell.

Male with the clypeus and frontal carinae much as in the female. Antennae filiform, 13-jointed; basal funicular joints short. Mesonotum without Mayrian furrows. Wing venation like that of the female.

Ants in this genus are subterranean predators on other ants and they are not commonly encountered in California. Workers of C. augustae Wheeler have been recorded from three localities in southern coastal California (specimens in LACM and UCDC), while males of C. davisi M. R. Smith have been collected at light in desert locations in Imperial and San Bernardino Counties (Snelling & George 1979).

Species identification: male key in Mackay and Mackay (2002). Additional references: Brown (1975), Smith (1942b), Snelling and George (1979), Wheeler (1902e, 1903j).

Canindeyú (ALWC, INBP).

Body elongate; head narrowed before and behind the eyes; eyes ovate, lateral, placed about the middle of the head; antenna short, incrassate; mandibles triangular, obsoletely toothed within. Thorax oblong-quadrate, strangulated in the middle. Abdomen: oblong, with a deep strangulation between the first and second segments.

> Cerapachys F. Smith, 1857: 74, [[ worker ]]. Type: Cerapachys antennatus , monobasic.

> Cerapachys : Emery, 1902: 24, subgeneric classification and relationships, new spp. described. ----- 1911: 8, diagnosis and subdivision into subgenera, species list.

> Cerapachys : Arnold, 1915: 11 - 17, S African spp., key and diagnoses.

> Cerapachys : Mann, 1921: 408, key to workers of Fijian spp.

> Cerapachys : Borgmeier, 1957: 107, discussion of subgenera, with doubts expressed about their validity; New World spp. listed; Cerapachys neuter in gender after Follett, 1955: 10.

> Cerapachys : Wilson, 1959: 39 - 57, discussion, description, list of spp. and key to spp., Melanesia.

> Cerapachys : Kempf, 1972: 76, list of neotropical species; Syscia and Parasyscia synonymized on p. 7.

> Ceratopachys Schultz , 1906, Spolia Hymenopt., p. 155, invalid emendation of Cerapachys .

> Syscia Roger, 1861: 19. Type: Syscia typhla, monobasic. Synonymized with Cerapachys , Kempf, 1972: 7.

> Cerapachys subgenus Syscia : Wilson, 1959: 39, 44 - 45, 51 - 52, 54 - 55, Melanesian species, with keys.

> Syscia : Wilson and Taylor, 1967: 33, Polynesian sp., species synonymy.

> Ooceraea Roger , 1862: 248. Type: Ooceraea fragosa , monobasic. Synonym of Cerapachys , teste Brown, 1973: 183.

> Cerapachys subgenus Ooceraea : Emery, 1902: 24. ----- 1911: 10, [[ worker ]], diagnosis, species list.

> Lioponera Mayr, 1878: 666. Type: Lioponera longitarsus , monobasic. Synonym of Cerapachys , teste Brown, 1973: 181.

> Lioponera : Emery, 1911: 11 - 12, [[ worker ]] [[ queen ]] [[ male ]].

> Lioponera : Donisthorpe, 1939: 252 - 257, review and species list.

-> Parasyscia Emery, 1882: 235. Type: Parasyscia piochardi , monobasic. Synonymized with Cerapachys in Kempf, 1972: 7.

> Cerapachys subgenus Parasyscia : Forel, 1892: 343.

> Cerapachys subgenus Parasyscia : Emery, 1902: 22 - 24. ----- 1911: 9, ??, diagnosis, species list.

> Phyracaces Emery, 1902: 23. Type: Cerapachys mayri , by original designation. ----- 1911: 10 - 11, pl. 1, fig. 3, [[ worker ]] [[ queen ]], diagnosis, species list. Synonym of Lioponera , teste Brown and Taylor, 1970: 957 - 958; synonym of Cerapachys , teste Brown, 1973: 181.

> Phyracaces : Arnold, 1915: 17 - 19, S. African species.

> Phyracaces : Wheeler, 1918: 215 - 216, discussion of affinities, etc.; 220 - 223, generic characters, distribution, ethology; 239 - 263, figs. 7 - 17, descriptions, review and key to species of Australia. ----- 1922: 22.

> Phyracaces : Clark, 1923: 73, 78 - 89, normal and ergatoid queens, larvae, pupae; 7 spp. described from SW Australia. ----- 1924: 75 - 89, pl. 6 - 7,

[[ queen ]] [[ male ]] biology, 8 Australian species described. ----- 1930: 3 - 6, 3 Australian spp. described. ----- 1934: 22 - 27, 6 Australian spp. described. ----- 1941: 71, 74 - 76, 2 Australian spp. described.

> Phyracaces : Wilson, 1959: 55 - 56, keys and lists, New Guinea and New Caledonia spp.

> Cerapachys subgenus Phyracaces : Forel, 1902: 405, 407. ----- 1915: 18 - 21, 3 Australian spp. described.

> Cerapachys subgenus Cysias Emery, 1902: 24. Type: Ooceraeapapuana , by original designation. Synonymized with Cerapachys subgenus Syscia by Emery, 1911: 10.

> Procerapachys Wheeler, 1914: 27 - 28. Type: Procerapachys annosus , by original designation. (Fossil in Baltic Amber.) New synonym.

> Chrysapace Crawley , 1924: 380 - 383, " apterous female, " recte [[ worker ]] or ergatoid [[ queen ]]. Type: Chrysapace jacobsoni , by original description, monobasic. Synonym of Cerapachys , teste Brown, 1973: 179.

> Cerapachys subgenus Chrysapace : Wheeler, 1924: 225.

> Neophyracaces Clark, 1941: 71, 76. Type: Phyracaces clarus , by original designation. Synonym of Cerapachys , teste Brown, 1973: 183.

There has been a good deal of confusion about the gender of the name Cerapachys . Actually, words in Greek for " horn " containing the stem equivalent of ker- occur in all 3 genders, masculine, feminine and neuter. Keras is the neuter form, and heros is the masculine form. When Smith named the genus, he seems to have been a bit careless in using an " a " instead of an " o " for the fourth letter, but his original employment of the name as masculine is clear from the masculine ending given to the name of the type species (antennatus) in adjectival form. Thus, there seems to be little room for argument, other theories (Borgmeier 1957) notwithstanding, and we should regard Cerapachys as masculine.

Worker: With characters of tribe. Postpetiolar segment (true abdominal segment III) strongly constricted from body of gaster and varying greatly from species to species in size relative to petiole and to true abdominal segment IV, which always is the largest gastric segment. In the more extreme forms of the fragosus and edentatus groups, the petiole and postpetiole form 2 nodes that are small in comparison to segment IV, which covers and forms most of the gaster as in most Myrmicinae . (Compare figs. 91 and 95). No strong constriction between principal gastric segments.

Antennae 9 - 12 segmented, often with a swollen apical segment. Palpi segmented 4,3 to 2,2. Compound eyes varying from large and multifacetted to completely absent. Ocelli present in workers in a minority of species.

Tibial spurs on middle legs; tarsal claws simple or with a submedian tooth.

Queen: Usually winged, but sometimes wingless and ergatoid, always with compound eyes present so far as known, though they may be very small. Characters otherwise as in the worker of the same species, but body usually a little broader, segment for segment. Wings, when present, like those of male.

Male: See characterization under tribe Cerapachyini (p. 15). So far as known, the male of Cerapachys always has 13 - merous antennae and apical spurs on the midtibiae, and the mandibles are triangular, though often with acute apex and concave masticatory border. A sampling of genital capsules is shown in figs. 123 - 126, and subgenital plates in figs. 115, 116, 118, and 122.

In the discussion that follows, I shall show that the synonymy of Cerapachys extends very widely — much more widely, in fact, than I would have believed before I began this study. Nevertheless, all but one of the generic or subgeneric names here listed as synonyms of Cerapachys were based on species originally included in Cerapachys , or early assigned to Cerapachys s. lat. by Emery or Forel. Thus in a sense we are returning to an earlier generic concept.

The 4 subgenera of Cerapachys were based primarily on the number of antennal segments, thus: Cerapachys s. str., 12 segments; Parasyscia , 11; Ooceraea , 10; and Syscia , 9. The subgenus Cysias had already been synonymized under Syscia by Emery (1911: 10).

The series Cerapachys s. str. — Syscia formed a rough morphocline, not only in the decline of antennal segment number from 12 to 9, but also in the loss of eyes in the worker, and in the reduction of the postpetiolar segment and relative increase in dominance of the succeeding (first gastric) segment. The most extreme result of these trends is seen in such Syscia species as edentatus and biroi , in which the petiole and postpetiole are similar in size, and the succeeding segment, true abdominal tergum IV, is enlarged to cover most of the gastric dorsum (fig. 95). This arrangement is formally like that of the subfamily Myrmicinae , and clearly represents a convergence to the myrmicine condition. It also renders very difficult the use of " waist " characters for the separation of formicid subfamilies in classifications and keys, especially when one has to distinguish certain army ants with 2 - segmented waists from cerapachyines and myrmicines.

While the general morphocline in species-groups of Cerapachys involves reductions in both eye size (ommatidial count) and antennomere number, the reductions do not occur with complete concordance. Undescribed species with 12 antennomeres and dot-like eyes in the worker have been found in Africa and Asia, and in these same two continents we have large-eyed species (e. g., nitidulus ) with 11 antennomeres. C. kodecorumnew species from southeastern Kalimantan (Borneo) usually has 11 segments in the worker, but one worker from the same series has 7 or even only 6 abnormally thick segments in the funiculus. Who is to say really what is normal and abnormal in antennal segmentation in this genus?

In some 11 - segmented species, the basalmost ring segments are often exceedingly short and indistinct, especially the first segment after the pedicel. This segment may even be largely hidden inside the pedicel, and may be partly fused with the succeeding segment (funiculus III). I believe that subgenus Ooceraea , based on 2 supposedly decamerous species, is actually cryptically 11 - merous. At least, the type and other specimens of 0. fragosus [65] seem to me to be obscurely 11 - merous, but when the ambiguities of counting and of determining which segments are fused or partly fused become this great, antennal segment number has weakened taxonomic value. Borgmeier (1957: 107), and after him Kempf (1972: 7), apparently hold the same opinion. C. fragosus and relatives from Asia, described and undescribed, are so much like the 9 - segmented C. typhlus group (" subgenus Syscia ") in body form, sculpture, and postpetiolar-gastric proportions that it seems absurd to recognize Syscia merely on the flimsy antennal segment character.

The main characters supposed to separate Phyracaces and Lioponera from Cerapachys are the proportions of the segments near the end of the antenna, and the shape of the petiolar node. In Cerapachys , the apical antennomere is usually very long and thick, even egg-shaped, and can be said to form a club of a single segment. While there can be no denying that many species fit the Cerapachys antennal pattern, others do so less well. The antennatus group (including antennatus, type species of Cerapachys ), for example, shows wide variation in this character, tending to bridge the gap between the Cerapachys condition, with a very thick apical segment, and the Phyracaces condition, in which the apical segment is little or not at all thicker than the penultimate segment.

The 5 - merous club cited as a character of Lioponera seems to me completely ambiguous, as already discussed above. We also have such " bridging " species as C. crawleyi [24] and C. lividus [36]. Figures 87 - 90 illustrate the CerapachysLioponera-Phyracaces morphocline for antennal clubbing.

The character involving the petiolar node really boils down to whether the node is margined on the sides above. Phyracaces species usually have strong dorsolateral margins on petiole and trunk, and the margination often extends to the postpetiole and even, in a few Australian species, to the head behind the eyes. Thus one finds a morphocline for strength and extent of margination along the body axis within Phyracaces , and this morphocline stretches of course into Lioponera (C longitarsus group), which has only the petiolar node laterally marginate. Here again the antennatus complex of forms provides a bridge between Cerapachys and Phyracaces , because in this complex, the petiole can be more or less Phyracaces-like or Cerapachyslike in different-sized individuals of the same nest series. I have pointed out in the discussion of this group [25] that one species was originally described as " Phyracaces vandermeermohri " which illustrates the ambiguity of the generic lines here. In addition to the antennatus group, we have the annectant forms crawleyi [24], pruinosus [60], and lividus [36]. Had the last-named species been earlier described from Australia instead of Madagascar, I believe it would have been put into Phyracaces rather than Cerapachys .

The ambiguity of the margination as a generic character is also demonstrated in the descriptions of such species as Phyracaces pygmaeus and P. aberrans [33, 47] (Clark, 1934: 25 - 27) and P. braytoni (Weber, 1949: 3). I have examined the type of P. braytoni [34], and find it to be a member of the same species-group as Lioponera longitarsus , type species of that genus, and P. pygmaeus is actually a synonym of longitarsus . R. W. Taylor recognized independently (personal communication) that Lioponera was nothing more than an indistinct species-group within Phyracaces , and we accordingly tacitly synonymized the latter in our contribution to " The Insects of Australia " (Brown and Taylor, 1970). The name Lioponera , being older, or course then took priority over Phyracaces . The members of the longitarsus group are, so far as we know, arboreal or subarboreal dwellers in hollow twigs and perhaps other tubular cavities in wood or bark, a microhabitat in consonance with the workers' slender, cylindrical body build and large compound eyes. I assume that they prey on ants of other species occupying similar habitats, but there is no real information available on their feeding habits. Both Lioponera and Phyracaces axe treated as synonyms of Cerapachys in Brown (1973: 181).

Clark (1941) raised a genus Neophyracaces for a group of Australian species in which the workers normally possess ocelli. Most of these species are relatively large in size and prevailingly bright orange or reddish ferruginous in color, and they seem especially well adapted to xeric conditions. Otherwise, they conform to the " typical " Phyracaces pattern, with strongly developed margination of the trunk, petiole, and even the postpetiole. A number of Phyracaces species in Australia are like them, except that the workers lack ocelli. Elsewhere in the collective genus Cerapachys — as in the complexes of C. antennatus [25] and C. fragosus [65] — the appearance of ocelli is an allometric character within as well as between colony series. Some of the large, ocellate individuals may in fact be functional reproductives (ergatoids), even in species known to possess dealate queens, but at present we have no direct information on this matter. In view of our scanty and largely ambiguous knowledge of ocellar occurrence and function in Cerapachys s. lat., it does not seem to me that Clark's Neophyracaces is worthy of recognition as more than a species-group within Cerapachys . It would be interesting to know whether and to what extent the ocellate workers of this group also function as reproductives.

C. crawleyi illustrates one method of dealing with annectant species: make it the type of a new genus. This species was originally described by Crawley as Chrysapace jacobsoni ( Chrysapace , like Phyracaces , is an anagram of Cerapachys ). While the species [24] is aberrant in its own right, it fits fairly comfortably in either Cerapachys or Phyracaces as they have been constituted in recent years, and Wheeler (1924) doubted that it was outside of Cerapachys . However, C crawleyi has one very primitive character in addition to its striking sculpture: the middle and hind tibiae each have a large and a small apical spur. The evidently related species C. sauteri has not been examined for the spur character, but if the tibiae in this poorly known Taiwanese species also have 2 spurs each, it may be necessary to resurrect Chrysapace as a genus.

We still have to account for the genus Simopone . Although described by Forel in 1891 as a genus, Simopone was placed as a subgenus of Cerapachys by its author in the following year, but Emery's 1911 treatment of it as a genus set apart in a special tribe with Cylindromyrmex tended to obscure the early relationship. Most of the species subsequently described in Simopone tended to strengthen the concept of a separate genus; the 11 - merous antennae, large, flat eyes, presence of ocelli, and above all, the separated frontal carinae, often framing demiscrobes for the antennae, tended to mark off a presumably arboreal group of species with its own distinctive habitus. However, the central African species grandis [76] is not so typical, and the fact that it combines traits of Cerapachys and Phyracaces with those of more characteristic Simopone species was obscured by a mediocre original description and by the paucity and isolation of material available for study in collections. The discovery and analysis of another specimen of grandis [76] permits us now to say that it is such a strong link between Simopone and Cerapachys s. lat. that the generic distinction comes into doubt.

Further trouble for this distinction comes in the form of a new species, Simopone conciliatrix [77], Simopone-like in habitus, but with 12 antennomeres and other details that put it in the category of annectants between Simopone and Cerapachys . Of course, there exist species of Cerapachys with 11 antennomeres (subgenus Parasyscia ), but, as already discussed, they mostly have the eyes reduced or absent, rather than enlarged as in Simopone . Thus the 11 - merous condition in Simopone and Parasyscia has apparently been considered (by anyone who may have thought about it at all) as convergent. The new 12 - merous species now ties Simopone back to the more primitive line of Cerapachys , which (according to Williston's Rule) must have had 12 antennal segments. Williston's Rule may also be invoked in the matter of the maxillary and labial palpi, which in C. grandis have 6 and 4 segments respectively, but in C. conciliatrix only 3 and 2. We see here the expression of an interesting tendency in ants as a family to maintain high palpal counts among some epigaeic, and especially arboreal, foragers. In these same lines, however, antennal segments may either be reduced in number, or stay at the primitive number 12.

In the circumstances, my instincts have been to place Simopone in the synonymy of Cerapachys as a primitive arboreal group of the latter genus, and this is the way the situation may well be viewed by future revisers. There does, however, remain one character by which the two groups can still be unequivocally separated in at least the workerqueen castes, and that is the tibial spur of the middle leg — present in all Cerapachys I have examined, and absent in Simopone . Since the middle-leg spur character is concordant in a rough way with the traditional characters of Simopone , it seems best to continue to recognize the generic separation, a course that also avoids some awkward specieslevel homonymy that would result if the Simopone species were thrown into Cerapachys .

bionomics: Discussed previously under the tribe.

distribution: Cerapachys as here constituted is by far the largest genus in the tribe, and its geographical range is virtually coextensive with that of tribe Cerapachyini . The genus is much better represented in the Old World than the New, and the majority of species, both described and undescribed, are in the Indo-Australian region. Forms with laterally marginate petiole (formerly Phyracaces , Neophyracaces , and Lioponera ) are restricted to the Old World, and have radiated especially extensively in Australia, where they occur in semidesert as well as wet and dry forest habitats. The 9 - segmented group ( Syscia ) is widespread in northern and eastern Australia as one or two species [63], and also has endemic species in New Guinea, the Solomons and Fiji; s. biroi of this group has become established in Hawaii and even in the West Indies, following probable overseas transport by human commerce. The 9 - segmented species are good colonizers.

It is interesting to note that the " more typical " Cerapachys — the species with rounded petiole and 12 or 11 antennal segments related to C. dohertyi and C. cribrinodis — have not penetrated continental Australia, though they have spread through Melanesia as far eastward as Fiji. This absence may be related to the extraordinary radiation in Australia of Sphinctomyrmex in the adaptive zones that elsewhere are mainly occupied by Cerapachys .

In the New World, where Sphinctomyrmex is known only from a single rare, localized species, Cerapachys is represented sparsely by a few 11 - segmented species ranging from Sonoran North America southward to Panama; one 11 - segmented species apparently isolated in southeastern Brasil; and one 12 - segmented, minute-eyed species from Trinidad that possibly could be a historic immigrant like C. biroi , known from the same island. The 11 - merous species from North and Central America appear to be endemic, and number perhaps as many as 5 or 6, counting undescribed samples; these form a tightly knit group of blind or minute-eyed forms related to C. augustae . It seems likely that Cerapachys is limited in the New World by competition with myrmecotherous ecitonines such as Neivamyrmex .

In Africa, Cerapachys is widely and fairly well represented by a diversity of groups of species with 11 or 12 antennal segments (mainly 12) in habitats ranging from rain forest to arid karroo veld, montane grassland, and Saharan oases and wadis. Although at least one species ( C. piochardi ) occurs in the Middle East, no cerapachyine is known to reach Europe.

Cerapachys ranges to the southern shores of Australia and South Africa, and is well represented on Madagascar, but is still unknown from south of Brasil in South America. In Asia, the genus reaches north to the Himalayan wall, central China, and southern Japan.

Cerapachys (common names include "raider ant" and "ant-raiding ant") is a genus of ants in the subfamily Dorylinae. Species are mainly myrmecophagous ants which raid the nests of other ants for prey. The genus is distributed widely throughout the Indomalayan region. The genus was revised by BoroWiec (2016) who split a number of previously synonymized genera out of Cerapachys, leaving only 5 species in the genus.