Cephalotes Species Groups

The following is based on de Andrade, M. L.; Baroni Urbani, C. 1999. Diversity and adaptation in the ant genus Cephalotes, past and present. Stuttgarter Beitrage zur Naturkunde Series B (Geolgie and Palaontologie). 271:1-889.

Antwiki refers to the clades listed here, in some contexts, as species groups. This is done to provide consistency in organizing these with other cogeneric sets of species.

angustus clade

A relatively large clade comprising 10 species characterised by combinations of characters present homoplastically also in other clades and secondarily lost in a few ingroup species. Two of these characters appear as synapomorphic in our phylogenetic reconstruction: the propodeum with 4 or 5 pairs of lateral denticles and the petiole dorsally armed. The combination of these two traits in characterising this clade is suggestive of what we understand as one of the major evolutionary trends in Cephalotes: improving its passive defence system. All the members of the clade are South American, particularly abundant in the tropical zone, and extending up to Colombia as the northernmost limit.

atratus clade

This clade corresponds to the former genera Cephalotes and Eucryptocerus. It is characterised by two synapomorphies among the characters we tentatively retained for our data matrix: (1) the bispinose vertexal border, a character which, in this order of magnitude, is uniquely derived among the members of this group but appears to be paralleled by some (certainly homoplastic) denticles to be found among other species like Cephalotes borgmeieri, Cephalotes betoi, etc.; (2) the presence of a pair of dorsal pronotal spines, another character derived uniquely for the members of this clade but which appears to have been ( ?secondarily) lost among very small atratus and opacus workers. The absence of a soldier caste, which we stressed as a significant plesiomorphic trait of the hamulus clade, is still recognisable among some members of the atratus clade: there is at least one species certainly without morphologically differentiated soldiers (atratus) but the worker caste, in this case, is slightly polymorphic in size. Other species, like oculatus and placidus, are still known from too few specimens as to permit definitive conclusions about their polymorphism. The members of this clade are among the largest Cephalotes. Only some species of the clypeatus clade approach their size.

basalis clade

A relatively large clade containing nine Recent and one fossil species. It is characterized by the following synapomorphies: (1) worker propodeum with a pair of spines shorter than the basal face, a trait secondarily lost among the more specialized species inca and basalis; (2) worker with incrassate fore femora, a character invariant within the clade and similar but probably not homologous to the condition characteristic of Procryptocerus, and, (3) soldier without vertexal denticles. In our reconstruction this character state results as a secondary loss of the clade which should have been re-gained among the most specialised species basalis and cordiventris.

bimaculatus clade

Cephalotes bimaculatus

A clade apparently characterised by only one, probably phylogenetically not very significant, synapomorphy: the bicoloured head of the soldier. This is a relatively uncommon character in Cephalotes present nonetheless homoplastically in all terminal species of our cladogram (clades bruchi to angustus). Although we can easily imagine that this trait must have some important mimetic function, its probably simple adaptive significance is the main reason to question its evolutionary significance. The main arguments, however, to maintain bimaculatus in a separate clade lie on its lack of important synapomorphies characterising all the species included in both its ingroup clades: wheeleri and texanus. The most important of these synapomorphies are the petiole and postpetiole with very thin and long lateral expansions (with small expansions in bimaculatus). From the members of the wheeleri clade, in addition, bimaculatus differs by missing all 6 synapomorphies characteristic of the clade (q. v.) and from the members of the texanus clade (q. v.) by missing 3 synapomorphies.

bruchi clade

Cephalotes bruchi

A clade comprising only one species which had been already considered as “remarkably distinctive” by Kempf (1958). In our analysis it is characterised by two synapomorphies: the cephalic disc of the soldier and of the gyne incomplete. The development of the cephalic disc among soldiers and gynes is not parallel in Cephalotes. This results clearly from the different distribution of this basically identical character in the two castes. For this reason we felt preferable to code the cephalic morphology of gynes and of soldiers as two separate characters. The resulting independent evolution of the disc among different Cephalotes castes may not apply to bruchi. In our cladogram, the unknown ancestor of bruchi should have had both soldiers and gynes with a complete disc. For this species, in fact, the reduction of the cephalic disc appears to be a secondary loss and it is likely to have occurred simultaneously among soldiers and gynes. This isolate case of involution of the disc morphology strengthens the segregate position of bruchi in a separate clade.

coffeae clade

A relatively large clade containing 11 species, 7 of which are fossil. It is characterized by two synapomorphies, both secondarily lost in some specialised species: postpetiole with broad lateral expansions and first gastral sternite simply reticulate (the synapomorphic state is the secondary loss of striation). There is, however, another potential synapomorphy for these species not resulting from our list of characters and worth to be mentioned in this context: the presence of abundant standing hairs on the gaster well visible in the Recent setulifer, peruviensis, trichophorus and coffeae and in the fossils alveolatus, sucinus and obscurus. The general habitus of all these species, however, is very homogeneous and different from that of the other Cephalotes species. The species belonging to this clade are in part fossil, either Mexican or Dominican, or confined to Central and Northern South America. The workers of three fossil species (dieteri, sucinus, and integerrimus) share also a darker spot in the center of the first gastric tergite but this character is difficult to appreciate in the amber specimens even within different individuals of the same species.

crenaticeps clade

This small clade possess no known strong synapomorphies. The sole character resulting uniquely shared by descent among its two members and resulting from our analysis are the propodeal spines of the gyne, diverging backwards. This character is rare but not unique within Cephalotes and we are rather sceptical about its phylogenetic importance. See also the introduction to the patei clade. To increase the insecurity about the distinction between this clade and the emeryi clade, one could add that Cephalotes crenaticeps and Cephalotes emeryi (belonging to the emeryi clade) are rather similar in pilosity and for having the gaster anteriorly without lobes.

clypeatus clade

A small (three species), very distinctive clade characterised by the following four synapomorphies: (1) vertexal angles of the worker triangular, (2) worker CI > 155, (3) worker pronotal spines fused with a lamella, (4) propodeal spines fused with a lamella. The triangular vertexal angles appear, homoplastically, also among workers of some of the species of the basalis clade. The high CI value, outside this clade, appears only in the basalmost species of the next closely related clade (the multispinosus clade): the Mexican fossil Cephalotes poinari. The three representatives of this clade had been formerly considered as a separate genus: Zacryptocerus s. str. If, on one hand, the synapomorphies listed above would be sufficient for the characterization at generic level of these three species, such a decision is rendered impossible by the examination of our cladogram since it would render paraphyletic any other generic distinction of its most related species.

depressus clade

A medium-sized clade containing eight Recent species with a broad tropical distribution. It is characterised by the following three synapomorphies: (1) worker propodeum with a pair of spines followed by a tooth (char. 43, state 1), (2) worker petiole with indistinct anterior and posterior faces (char. 50, state 2), and, (3) aedeagus internally concave (char. 126, state 2). Synapomorphies (1) and (3) appear very robust and support entirely the separation of this clade. The shape of the worker petiole appears less convincing, and we suppose that its resulting synapomorphic value may be attributed either to our imprecise coding of it or to an artifact of character optimisation. Members of this clade are the first compact group of species the soldiers of which show either an in incipient or a complete cephalic disc. All remaining 79 more terminal species of the genus, except the problematic Cephalotes bruchi the soldier of which has an incipient disc only, have completely developed cephalic discs. Kempf (1951) intuitively recognised a "pavonii group" roughly similar to our depressus clade. The main difference between our and Kempf’s groupings of species (besides the trivial addition of new species in this paper) is our exclusion from the depressus clade of Cephalotes multispinosus and Cephalotes biguttatus, two species that we rank within our multispinosus clade, a clade very close but paraphyletic to the depressus clade.

emeryi clade

This clade contains two Dominican fossil species and a Recent one from Curacao. The exact relationships among the three and with other Cephalotes species may be subject to future changes since all of them are only known from workers. Its members are unique within a large, terminal Cephalotes clade comprising 52 species, for the secondary loss of the coloration of the frontal carinae. This character is likely to have an obvious adaptive and a doubtful phylogenetic meaning. See also the introduction to the patei clade.

fiebrigi clade

A relatively large clade comprising 11 species confined to the southernmost portion of the distribution of Cephalotes and extending up to Mato Grosso to the North. The worker propodeum with three or four lateral denticles results synapomorphic for the clade, although both morphologies can be homoplastically present also among members of other species groups. The sole known four males of the clade, moreover, share also very long gonostyli. All these characters separate the members of the clade from their closest relatives with which they share the solid membranaceous border of the gastral expansions.

grandinosus clade

This clade contains a small set of seven species characterised by a unique but important synapomorphy: the hind femora of the soldier with a dorsal crest. Only the nominal species grandinosus has a broad geographical distribution; other members of the clade appear to be restricted to the Brazilian-Peruvian rain forests and two species are fossil in Dominican amber. The Recent species of this clade are also among the minority of Cephalotes with light colours. The coloration, of course, can not be assessed with certainty for the fossils.

hamulus clade

This clade corresponds to the former genus or subgenus Hypocryptocerus and contains all the less specialised species of the genus Cephalotes. It is characterised by only three synapomorphies: (1) the vertexal angles broad, round and with crenulate margin, (2) the outer face of tibiae with longitudinal, irregular rugae, a character shared homoplastically also with the far relative Cephalotes rohweri (a member of the phylogenetically distant wheeleri clade) and with some species of the sister genus Procryptocerus, and, (3) petiolar spines present but shorter than 1/3 of the petiolar length, a character re-appearing in a number of distantly related, more specialised Cephalotes species. We regard this as an additional demonstration of the inappropriateness of considering Hypocryptocerus as a valid genus. We shall name the species of this group as hamulus clade for uniformity with other, comparable, phyletic lines we recognize within Cephalotes. Since the name Hypocryptocerus already exists in the literature, there are, however, no serious reasons against the use of this name as a subgeneric name for those preferring this formal solution.

Practically, if one disposes of abundant Cephalotes nest series, he should be able to recognise the members of the hamulus clade at glance by their possibly unique possession of one very important plesiomorphy: the absence of soldiers and the monomorphism of the worker caste, a character state unknown in other Cephalotes (though it may be present in the poorly known species Cephalotes placidus and Cephalotes oculatus in the atratus clade), but which is the rule among the species of the sister genus Procryptocerus.

Within the hamulus clade we recognise eight valid species and there is a possible, insufficiently known, ninth species. All these species are presumably endemic on the Island of Hispaniola and two of them are known only from copal samples of probable historical age. One could expect to encounter them again in the Recent fauna of the island.

A circumstance apparently contrasting with the antiquity of the clade is that its members are very similar each other and are likely to be the product of relatively young speciation events, a hypothesis supported also by palaeogeographic evidence. The main diagnostic characters among them appear to be confined to integumental sculpture, pilosity and coloration. As a result of this situation, the internal phylogeny of the clade is far from being properly understood and, due to our incapacity to detect sufficient, plausible synapomorphies, the internal phylogeny of this clade should be considered as tentative.

laminatus clade

A medium size clade containing seven Recent species characterised by the following synapomorphies: vertexal angles of the worker with a truncate lamella and gyne with four coloured spots on the gaster, a character which appears to have been secondarily lost in christopherseni; in this species the four coloured gastral spots are still visible as an undifferentiated macula in some specimens. Most species of the clade appear to be restricted to the S American rain forest with the remarkable exception represented by minutus which is one of the commonest and most broadly distributed species of the genus. This clade corresponds roughly to Kempf (1951) "spinosus group" with his 1974 addition of duckei.

multispinosus clade

This clade is characterised by only one synapomorphy: the reduced propodeal lamella. It comprises a Mexican and two Dominican fossils and three Recent Mesoamerican species. The outgroup of the clade is the clypeatus clade. It is remarkable that the first species of the multispinosus clade (poinari) still exhibits the general habitus of the members of the clypeatus clade though missing three of the four synapomorphies of it. The fourth synapomorphy of the clypeatus clade (CI > 155) is present also in poinari but it results homoplastic with the one of the clypeatus clade under any customary assumption on character evolution. Kempf (1951) supposed that biguttatus and "multispinosus" (= goniodontus) represent a distinct subgroup of his pavonii-group. Our cladogram suggests a paraphyletic position for biguttatus and multispinosus as compared with the pavonii clade. This forces us to regard these two species plus one Recent and three fossil ones closely related to them as a separate clade sharing with the species of the pavonii clade and with Cephalotes umbraculatus (the unique member of its own clade) the angle between the basal and declivous faces of the propodeum ≥ 110°.

pallens clade

A relatively large clade containing ten recent species. It roughly corresponds to the former subgenus Cyathomyrmex but Cephalotes setulifer has been excluded from it.

The members of this clade are characterised by the three following synapomorphies:

1. Frontal carinae of the workers broadly incised over the eyes.

2. Frontal carinae concolour with the rest of the head.

3. Soldier and gyne with the disc covering the mandibles.

An additional probable synapomorphy for the species of this clade is a character not considered in our data matrix: the capacity of the soldiers to secrete camouflage material. We have evidence of soldiers with the head more or less covered by such material for all the species of the clade. We did not include the secretion of camouflage material in our list of characters because of the difficulty to obtain evidence of its absence for a number of species. This trait, however, is homoplastically present also outside the pallens clade at least in Cephalotes setulifer, Cephalotes coffeae, and, probably, Cephalotes patei.

These three species, however, cannot be considered members of the pallens clade since they lack at least two important sets of synapomorphies, i.e. most of those already listed above as characterising the pallens clade (Cephalotes setulifer soldiers and gynes, represent the sole exception since they have a disc covering the mandibles) and the following important synapomorphies of a broader basalmost clade including our pallens, grandinosus and pinelii clades: antennal club two-jointed, secondary loss or great reduction of the scapular angles in dorsal view, and median depression of the disc. Forcing either of these species to belong to the pallens clade would cost respectively 9 additional evolutionary steps for setulifer, 8 for coffeae and 7 for patei. Forcing all three species together to join the pallens clade would cost a minimum of 16 additional evolutionary steps. These results have been obtained after only one heuristic search for each hypothesis and with a different constraint per search.

The species we recognise as belonging to the pallens clade represent a very homogeneous set and are often difficult to distinguish from each other, particularly in the worker caste. This situation is worsened by the fact that these ants appear to exhibit the greatest variability within Cephalotes in sculpturation and in body size. In this sense, one of the oldest available names for a species of this clade, Smith's (1876) varians, was premonitory. Kempf (1958 a) already tried to split the species and subspecies described in Cyathomyrmex. His efforts have been unsuccessful because of the limited number of soldiers associated with workers available to him.

patei clade

Cephalotes patei

A small clade containing only the homonymous species patei. This and the following two clades (emeryi and crenaticeps) appear weakly or insufficiently defined from the results of our cladistic analysis. We could try to add additional synapomorphies for each of them or to group the three clades in one but none of these solutions (i.e. artificially constructed synapomorphies or lumping all species in a single clade) would be really convincing. Patei exemplifies well this situation: it bears obvious similarities with emeryi and crenaticeps for the short pronotal lamellae of the worker, and, at the same time, it remains rather distinct within the genus by the stout (not flattened) mesosoma, by the flat head, and by the very broad, spatulate hairs of the gyne and of the soldier. None of these characters are considered in our data matrix since we regard their apomorphic value as doubtful. No other apomorphies for this species result from our analysis. We did not dare dissecting the sole known male preserved in the USNM but the description of its genitalia represent the most concrete foreseeable source of information to better assess the position of this species

pinelii clade

This is a medium sized clade containing nine species, including one fossil in Dominican amber. Both its known unequivocal synapomorphies are chromatic, i.e. two coloured spots or strips on the gaster of the worker and four coloured spots on the gaster of the gynes. However weak these characters may appear, they are likely to have a clear, probably aposematic, adaptive meaning. Similar colour patterns are known also in members of the muitispinosus, bimaculatus, coffeae and angustus clades. The pinelii clade, however, is distantly related to all of them for possessing membranaceous expansions on the mesosoma (absent in all members of the aforementioned four clades) and for the antennal club two-jointed (three-jointed in the muitispinosus, coffeae and angustus clades). The pinelii clade as we define it in this paper is much narrower than the "pinelii complex" previously recognised by Emery (1922) and by Kempf (1952). As already noted by Kempf (l. c.) most of these species occasionally nest in epiphytes.

prodigiosus clade

A small clade containing only two Argentinean species characterised by having soldiers with the cephalic disc concave and with irregular foveae bearing clavate hairs not surpassing the border of the foveae. The gynes of both species are also unique in the shape of the disc: concave in the two anterior thirds and flat posteriorly. Both species appear to be rather rare: fewer than 20 specimens are known.

pusillus clade

The two species constituting this clade have been considered as only one (pusillus) by Kempf (1951). Pusillus was regarded by Kempf (1.c.) as a member of his "spinosus species group" which is roughly equivalent to our laminatus clade and with which it shares several characters. In our analysis pusillus and columbicus result separate and paraphyletic respect to the laminatus clade. The main reason for this is the presence of fine reticulation under the head, a character present in Cephalotes solidus and in all the members of the laminatus clade but absent in pusillus and columbicus. Another character separating pusillus and columbicus from the members of the laminatus clade (absence of angulate femora in the former vs. presence in the latter) is actually present only in the basal species of the laminatus clade and absent in the internal ones. The synapomorphies of the pusillus clade resulting from our phylogenetic reconstruction are the following:

1. Frontal carinae of the soldier not reaching the vertex. This character is present also in some ingroup species of the laminatus clade.

2. Aedeagus with long spines hidden by the basal projection.

3. Internal face of the aedeagus ventrally concave.

We are inclined to recognise a significant phylogenetic value to characters 2 and 3. The crucial fact forcing the distinction between the pusillus clade and the laminatus clade, however, is the position of the insufficiently known species solidus (q. v.) - intermediate between the two clades in our Cephalotes phylogeny. This position may be subject to change after discovery of the solidus soldier, gyne and male.

solidus clade

Cephalotes solidus

This clade contains a single species known from five workers only. Several important male, gyne and soldier characters remain unknown and we may expect the phylogenetic position of this species to be subject to considerable change once soldiers, gynes and males will be discovered. The sole apomorphy of the clade resulting from our data matrix is the propodeum with basal and declivous faces not separate by a spine, a character state common to all most specialized species of the genus and regularly absent in the clades topologicaly closer to solidus. Other autapomorphies of solidus that we consider of lesser importance are listed under the species' treatment. Kempf (1974) considered solidus as a member of his (1958a) "angustus group", corresponding grosso modo to the three terminal clades of our phylogeny. All members of the "angustus group" (and other species as well) exhibit the same propodeal morphology as solidus and confirm, in this way, Kempf's intuitive placement of this species. In addition, most critical synapomorphies for the clade better corresponding to Kempf's "angustus group" in our phylogeny are soldier or gyne characters, i. e. characters unknown in C. solidus. Solidus, however, lacks a number of worker synapomorphies of different hierarchic level regularly present among most members of the "angustus group". The most significant ones are probably characters 39-42 state 1 (solidus = 0).

By means of the Constraints command of Paup we forced solidus to into a clade corresponding to Kempf’s "angustus group". All 14 resulting possible optimal trees are only two steps longer than our phylogeny of Fig. 24. The position of solidus is constant in all 14 trees: it results as the outgroup species of all the other members of Kempf's "angustus group".

texanus clade

A small clade (6 species) comparable under several respects to its sister clade, wheeleri. Its members, however, lack the six synapomorphies characteristic of the wheeleri clade and are grouped together by the following ones, absent in the latter: worker propodeum with three pairs of lateral denticles and, presumably, in spite of the limited evidence available, the internal border of the aedeagus narrowly concave. There is abundant presumptive evidence suggesting that most species of this clade should be obligatory inhabitants of epiphytes. These species range from southern United States up to Costa Rica to the south.

umbraculatus clade

Cephalotes umbraculatus

As already intuitively proposed by Emery (1922) and by Kempf (1958) this is a very distinctive clade containing only one species characterised by the following five autapomorphies: characters 8: 0->2, 50: 0->1, 108: 0->1, 118: 0->1, 126: 0->1. In spite of the resulting isolate status within the Cephalotes phylogeny, umbraculatus shows some characters clearly intermediate between its most adjacent clades as listed under the discussion of the species. Umbraculatus appears to be one of the commonest species of the genus in large parts of the Neotropical rain forests.

wheeleri clade

A small clade containing only 5, very distinctive, species, all restricted to a broad Sinaloan-Sonoran region from Arizona to the states of Jalisco and Morelos to the south and including Las Islas de Tres Marias. These species are easily recognizable and, in our cladogram, they are characterised by a set of six synapomorphies: for the workers, (1) the postoccipital suture extending over the ventral side of the head, (2) the first gastral tergite striate, (3) the loss of gastral coloured spots, (4) the presence of erect, truncate hairs, and, for the soldiers, (5) the antennal scrobes terminating below the eyes, and, for the gynes, (6) loss of the gastral coloured spots (a trait not necessarily correlated with the analogous one of workers). In spite of their compact phylogenetic robustness and unequivocal presence of advanced synapomorphies (soldiers and gynes with disc, etc.), the workers of the species of this clade bear a strong, superficial similarity with the basalmost Cephalotes clade: the hamulus clade.