Diagnosis and relationships of the myrmicine ant genus Ishakidris gen.n. (Hymenoptera: Formicidae)

Abstract. The ant genus lshakidris from Sarawak is described as new. Its relationships with the Brasilian monotypic genus Phalacromyrmex Kempf and the Malagasy monotypic genus Pilotrochus Brown are discussed, and the Phalacromyrmex genus‐group is established to hold the three genera. The resemblances of lshakidris to the smithistrumiform dacetine genus Glamymmyrmex Wheeler and the agroecomyrmecine genus Tatuidris Brown & Kempf are discussed and the similarities are analysed as the results of convergence in the characters concerned.


Introduction
In 1960 Kempf described a remarkable Brasilian myrmicine ant genus, Phalacromyrmex, which bore a striking similarity in habitus t o the smithistrumiforin dacetines of the genus Glamyromyrmex Wheeler. He stated in the original description that 'this new genus seems at least superficially related with the ants of the tribe Dacetini', and observed that its habitus was reminiscent of both Glamyromyrmex and Codiomyrmex Wheeler. However, after listing the diagnostic characters of Phalacrornyrmex he ultimately decided that 'it seems best not to include the present genus in the Dacetini', but admitted that it appeared 'even farther removed from all other myrmicine tribes'.
Bolton ( 1 983) retained Phalacrornyrmex as a peripheral dacetine, including it in the synopsis of the dacetine generic distribution and enumeration of species, as there seemed no better placement for it. In the text it was pointed out that Phalacromyrmex, together with a related but undescribed Indo-Australian genus, did not fit in any of the recognized subtribes of the Dacetini, and speculated that Correspondence: M r U. Bolton,1) 'there is a very strong possibility that Phalacromyrmex and its undescribed relative may be convergent on the smithistrumiform dacetines fromsome other part of the Myrmicinae'.
The new genus, Ishakidris, is described below and its diagnostic characters listed. The description is followed by four separate sections analysing and discussing its real and apparent relationships with the genera Phalacromyrmex, Glamyromyrmex, Tatuidriy and Pilotrochus respectively. It is concluded that Ishakidris shares a genuine close phylogenetic relationship with Phalacromyrmex and that these two genera together are most probably distantly related to Pilotrochus. A genusgroup is established and defined to hold these three genera. The analysis of Glamyromyrmex and Tatuidris indicates that their apparent similarity to Ishakidris does not reflect a genuine relationship but is the result of convergence in a number of character-states.
Representatives of all the genera under discussion are components of the extensive tropical leaf litter and topsoil fauna of ants. All appear t o be uncommon or rare and most species are represented by only very few specimens in museum collections. Phalacromyrmex and Tatuidris are exclusively Neotropical, the former known from a single sample collected In Brasil, the latter from two samples collected in El Salvador and Mexico respectively. Pilotrochus is represented by a unique worker from east central Madagascar and a single worker also constitutes the known representation of Ishakidris, from the lowland rainforests of Sarawak. Glamyromyrmex is more widely distributed. With eighteen described species and a number of undescribed species it occurs in the tropical zones of all zoogeographical regions, but has a preponderance of species in the Neotropical and Afrotropical regions.
Ishakidris g e m .
DIAGNOSIS OF WORKER. Ants belonging to subfamily Myrmicinae. Mandibles massively constructed and in profile the convex apical (masticatory) margin almost at a right-angle to the long axis of the head. Mandibles serially dentate and opposable, with 12 teeth o n each apical margin and with the teeth alternately larger and smaller from base to apex; basal series of teeth larger than apical series and the mandible lacking a basal lamella. Labrum narrowly long-bilobed anteromedially. Palp formula 2,2 (in situ count). Outline shape of head in the form of a shield in full-face view (Fig. 2), the occipital corners acute and the margin between them concave except for a shallow median convexity which encloses a translucent area of cuticle. Sides of head shallowly convex and continuous with the frontal carinae, the latter massively expanded anteriorly and laterally, fused t o the sides of the clypeus and completely obscuring the lateral parts of the head from full-face view; the expanded portions of the frontal carinae semitranslucent. Clypeus broad and broadly inserted between the antennal articulations, the median portion of the anterior free margin evenly convex and prominent. Sites of the antennal insertions showing through the expanded flanges of the frontal carinae in fullface view as a pair of blister-like areas close to the anterior margin of the head and abutting the lateral clypeal margins. Antennae with 9 segments, the apical club of 2 segments. Head in profile with deep antennal scrobes, the eyes small and situated posteriorly within the scrobal area close to the ventrolateral margin. Ventrolateral margin of head delimited by a sharp longitudinal carina on each side which runs from the ventral base of the mandibular insertion to the posterior cuticular lug or flange at the occipital corner. Ventral surface of head with a broadly convex median bulge or tumulus. Alitrunk fused, without sutures, the pronotum marginate anteriorly and the sides marginate throughout their length. In profile the propodeal dorsum sloping downwards posteriorly t o the triangular laminate propodeal spines, which themselves are subtended by, and partially fused with, a broad infradental lamella on each side. Orifice of propodeal spiracle circular, situated close t o the posterior margin adjacent to the narrowest point of the propodeal infradental lamellae. Mesopleuron excavated anterodorsally just posterior t o the posteroventral corner of the pronotal laterotergite, the excavated area containing a semicircular patch of thin cuticle upon which (or embedded in which) are a series of radially arranged fine hairs or cuticular processes. From this presumably glandular organ a shallow open groove extends posteroventrally, ending in a slit-like impression at the margin of the mesopleuron. In dorsal view the pronotum much broader than the mesonotum and propodeum, the last two separated by a low but sharp transverse carina. Petiole curved-clavate in profile, with a long anterior peduncle and a long low node. Anterodorsal angle of node blunt but well developed, the dorsum of the node merging evenly with the shallowly sloping posterior face, without a developed posterodorsal angle. Petiolar spiracle small, situated at about the midlength of the peduncle. Ventral process of petiole a low laminar crest which runs the length of the segment. In ventral view the process seen to be double and narrowly V-shaped, with a laminar crest ventrolaterally on each side of the peduncle diverging posteriorly and meeting anteriorly close t o the insertion. First gastral tergite without a transverse lamellate or spongiform crest basally, the petiole, postpetiole and gaster lacking spongiform appendages. First gastral sternite basally with a very conspicuous broad median flat-topped longitudinal crest which fits between the ventrolateral flanges of the petiole and the postpetiolar laterosternal lobes and restricts the degree of anterior flexion of the gaster.
shallowly concave when the blades are viewed so that t h e entire dental row is visible, basally passing through a narrow blunt curve t o their insertions. Mandibles with fine appressed short hairs, the ventral margin on each side with a single very long fine hair close t o the articulation; this is most probably a mandibular trigger hair. Anteromedian margin of labrum with a pair of elongate narrow digitate lobes which are fringed b y narrow lanceolate hairs. Apices of lobes dorsally apparently with a fused mass of very fine hairs, but this may be an artefact.  Bolton, 1983).
Mandibles longitudinally costulate-rugose, with punctures between the longitudinal components, the mandibles more densely sculptured than any other part of the body. Dentition of apical (masticatory) margins partially obscured by a mesially directed row of yellowish hairs which arise at the base of the tooth-row down the length of the margin. Counting from the base of the mandible the first (=basal), third and fifth teeth are approximately the same length (c. 0.10) and are the largest teeth on the margin; teeth 2, 4 and 7 are about the same size (c. 0.08); teeth 8 and 10 are reduced t o denticles whilst 6 , 9 and 11 (preapical)  to the head proper on each side of the median convexity. Sides of head behind level of frontal carinae with 3 pairs of short acute laterally projecting hairs, otherwise the cephalic dorsum without standing pilosity of any description. Occipital margin bounded by a low but sharply defined carina which ends in a pair of translucent cuticular flanges or lugs running down the posterolateral occipital lobes. Ventrally these lugs are confluent with the ventrolateral longitudinal carinae which run forwards to the bases of the mandibles. Eyes small, with only 7 -8 ommatidia, the maximum diameter 0.06 (c. 0 . 0 6~ HW). Sides of head between eye and mandibular base finely irregularly rugulose with roughened t o punctulate spaces between the rugulae. Above the eye the rugulac fading out and behind the eye the sides of the head mostly smooth. A secondary longitudinal rugule or weak carina runs from near the mandibular base to a point well behind the level of the eye, and is situated between the eye and the carinate ventrolateral cephalic margin. Bulging convex ventral surface of head smooth and shining. Antenna1 scapes and funiculi clothed with decumbent or appressed narrowly spatulate hairs. With the alitrunk in dorsal view the humeri rounded, the lateral marginations darker in colour than the dorsum proper, the surface lacking standing hairs except for a single pair on the transverse ridge separating the mesonotum and propodeum. A narrow arcuate ridge between the propodeal spines separates the dorsum from the declivity. Dorsum of alitrunk mostly unsculptured, with one or two faint longitudinal rugulae on each side of the mesonotum which run forward on to the posterior pronotum. Both pronotum and mesonotum with a few scattered minute appressed hairs arising from shallow pits, as on the head. Anteromedially the pronotum with a very short low carina. Propodeal declivity between the infradental lamellae exceedingly finely punctulate. Sides of alitrunk glossy and smooth, the propodeal infradental lamellae showing an internal fine reticular patterning. Petiole with sides and dorsum of node rugose, the rugae less well defined dorsally than laterally. Laminar ventral crest of petiole reticulate-punctate as is the area below the level of the spiracle. Dorsurn of petiole node with one anterior and two posterior pairs of short narrowly spatulate hairs, and with another pair posterolaterally. On the postpetiole there is a similar pair of hairs anterodorsally, two pairs posterodorsally and two pairs laterally. Petiole node in dorsal view longer than broad, broadening from front to back; postpetiole broader than long, subrectangular, margined by thick rugae anteriorly and laterally and with a more dorsally situated pair of longitudinal rugae which weakly converge posteriorly. Laterosternites of postpetiole densely punctate. First gastral tergite finely superficially reticulate everywhere, without pilosity and with a few short but sharply defined basal costulae arising from the slightly thickened basal margin. First gastral sternite smooth and convex, the broad basal carina thin-walled and translucent close to the base of the sclerite; the true basal surface of the sternite concave and thickly sclerotized below the postpetiole. Apex of gaster with a few inconspicuous projecting hairs. Femora and tibiae of middle and hind legs with appressed narrowly spatulate hairs. Colour of head, alitrunk and pedicel segments glossy medium brown t o reddish brown, the frontal carinae, antennae, legs and gaster lighter brown to yellowish brown. Mandibles darker in colour than head-shield; marginatfons of head ventrolaterally and of alitrunk laterally blackish. Holotype

Co m m en ts
The holotype is damaged, lacking its left middle and hind legs which were missing on collection.
The single known specimen of I.ascitaspis was recovered from a Winkler bag sample of previously sieved leaf litter collected at Camp 5 in Gunong Mulu National Park during the Royal Geographical Society expedition there in 1978. Camp 5 lay beside the Melinau River in lowland rainforest at the edge of the limestone outcrop which slopes up to the spectacular Limestone Pinnacles of the park.
Diagnostic characters which in combination isolate fshakidris from the workers of all other myrmicine ant genera are as follows.
7. Frontal carinae flanking lateral portions of clypeus and forming the anterolateral corners of the head (Fig. 2).

Relationship with Phalacromymex Kempf
The singie species of the Brasilian genus Phalacromyrrjzex, P.fugax Kempf, is known from a short series of workers recovered from a sample of topsoil taken at IbicarC, Santa Catarina State. This species shows a marked similarity in habitus t o Ishakidris and also shares the following suite of apomorphic characters.

.
Mandibles massive and strongly downcurved in profile. laterally.
FIGS. 5-7. PhuiUcromyrmru p g u x Kempf, paratypc worker: 5 , body in profile, antennae and legs omitted; 6 , head in full-face view; 7 , left mandible to show dentition. [The assumed plesiomorphic states for these characters at this level of analysis are: 1. Mandibles relatively slender and feebly downcurved; 2. mandibles with teeth regularly decreasing in size from apex t o base; 3. apical tooth the largest on the masticatory margin; 4. labrum simple t o weakly medially indented ; 5. palp formula > 3,2 (to a maximum at PF 6,4); 6., 7. frontal carinae absent; 8. antennae with 12 segments; 9. antennae filiform; 10. antenna1 scrobes absent; 11. eyes lateral; 12. alitrunk transversely convex, immarginate; 13. propodeal spiracle approximately at midlength; 14. propodeal spines not laminate; 15. mesonotum and propodeum subequal t o pronotum in width; 16. pectinate tibial spurs present on middle and hind legs. These assumed plesiomorphic states are abstracted from the conditions presently thought t o accord most accurately with the hypothetical ancestral condition of the myrmicine ants, and assumes that the subfamily Myrmicinae constitutes a holophyletic taxon.] The similar habitus and wealth of detailed correspondence in specialized characters between Phalacromyrmex and Ishakidris leads t o the conclusion that a true phylogenetic relationship, rather than a strong convergence, is indicated here. Particularly convincing are t h e remarkable specializations shown. by the head and mandibles, and the construction of the antennae.
Characters which genera are as follows.  (Fig. 8 ) and G.africanus Bolton (Fig. 9) have been selected specifically because of their habitus similarity to Ishakidris, other species are not nearly so like I.ascitaspis in general appearance. The mandible of G.sistrurus Bolton (Fig. 10) is in reality very strongly arched-downcurved but has been drawn here as if flat so that t h e entire dental array can be seen, and is thus semidiagrammatic; the prominent large basal lamella, characteristic of the genus, is stippled. Eighteen species of Glamyromyrmex have been described from the Neotropical and Afrotropical regions, and further undescribed species are known from the Oriental and Indo-Australian regions. It is suspected that a few species from Australia and the Neotropics which are presently placed in Codiomyrmex Wheeler may also really belong in Glamyrom yrm ex.

Ishakidrik
Apart from habitus similarity the following characters also correspond in Ishakidris and Glamyromyrmex workers.  The supposedly glandular area of the mesopleuron, well defined and subtended by a posteroventrally directed open groove in Ishakidris, is variously developed in Glamyro-myrmex. The anterodorsal excavation of the mesopleuron, just behind the posteroventral corner of t h e pronotal laterotergite, is always present and sometimes has a few short hairs traversing it (thus in G.sistrurus Bolton, crypturus Bolton, tetragnathus (Taylor), dagon Bolton, sahurus Bolton, tukultus Bolton, thuvidus Bolton, and a n undescribed species from Paraguay). In other species the excavation is present and traversed by numerous hairs, but is not subtended by a groove (G. semicomptus (Brown), aztecus Kempf, excisus (Weber), and an undescribed species from Sarawak). In an undescribed species from Hong Kong the mesopleural impression is well developed and fringed with fine flocculent material. Given this variation of development in Glamyromyrmex (and in some other smithistrumiform genera), and the corresponding amount of variation in this supposed glandular structure in other genera of similar habitus, where it is small in Phalacromyrmex, moderately developed in Ishakidris and very large and subcircular in Pilotrochus, it seems most likely that it has developed independently several times and cannot be utilized in assessing the relationships of these genera.
In spite of t h e characters listed above I consider the apparent similarities between Ishakidris and Glamyromyrmex t o have been acquired convergently, and view the resemblances as superficial. This conclusion is justified by a comparative analysis of the diagnostic characters of Ishakidris with all the eighteen presently recognized species of Glamyromyrmex, tabulated below.

Teeth alternating in
size down the length of the mandibular masticatory margin 2. Mandibles without a basal lamella (Fig. 3) 3. Clypeus in full-face view not forming a part of the side of the head in front of the frontal carinae (Fig. 2) flanked by the expanded frontal carinae, which form the anterolateral angles of the head

Clam yrom yrmex
Teeth not alternathg in size down the length of the mandibular masticatory margin Mandibles with a basal lamella (Fig. 10) Clypeus in full-face view forming a part of the side of the head in front of the frontal carinae ( Fig. 9) Lateral clypeal margins not flanked by the expanded frontal carinae, the clypeus forming the anterolateral angles of the head part of the diagnosis of the smithistrumiform genus-group or of the Dacetini as a whole. Character 1 is the only remaining correspondence, but the reduced apical tooth appears almost certainly acquired convergently as the mandibles have no other correlating detailed structures. These considerations imply most strongly that the apparent similarity of Ishakidris and Glumyromyrmex is superficial and the result of convergence, and that the two genera d o not share a close pliylogenetic relationship. morphic at this level of analysis are 1, 3 , 4 and 10; those considered apomorphic in Glamyrornyrmex are 2 , 5 , 6, 7, 9 and 1 1 . Character 8 remains as one whose plesiomorphic state is indeterminate at this level. Chmyroinyrmex belongs t o the smithistrumiform dacetines (Brown, 1953;Bolton, 1983), a group of sixteen genera diagnosed together as those members of the tribe Dacetini which have the eyes situated ventrolaterally on the head within or beneath the scrobes, have the antennae 6-or 4-segmented with the apical antennomere much the longest of the funicular segments, have the palp formula 1 , l and lack an apical fork of spiniform teeth on the mandibles. Throughout all the smithistrumiform genera characters 1 , 3, 4 and 10 of Ishakidris are universally absent. Of those given above as apoinorphies for Glamyromyrrnex numbers 2, 5 , 9 and 11 are also present throughout the smithistrumiform genera. (Character 6 is generally so but some have the antennaereduced t o only 4 segments; character 7 is variably developed.) Away from Glamyromyrmex also the habitus of the smithistrumiform genera is markedly different from that of Ishakidris. In particular, o f the characters listed at the beginning of this section as corresponding between Ishakidris and Glamyromyrmex, numbers 2, 5 and 6 are generally absent in other smit histrumiform genera, and numbers 3, 4 , 7 and I) constitute quickly analysed a s superficial, as in Tutuidris the heavy mandibles are bidentate apically but otherwise lack teeth, the palp formula is 1,2, the eyes are situated at the posteriormost apices of the scrobes, the propodeum is unarmed and the middle and hind tibiae have pectinate spurs present. Apart from these telling characters the construction of the petiole and postpetiole is radically different in Tutuidris, where the short thick sessile petiole which has a very large deep ventral process is followed by a postpetiole which is short, deep and very broad, and broadly attached to the gaster. Brown & Kempf ( 1 967) included Tatuidris in the tribe Agroecomynnecini, along with the fossil genera Agroecomyrmex Wheeler and Eulithomyrmex Carpenter. They noted the superficial similarity of Tutuidris with Phulacromyrmex and Glumyromyrmex but dismissed any suspicion of true phylogenetic relationship, concluding that Tatuidris 'is very strongly isolated among living ant genera, though clearly a myrmicine'.

Relationship with Pilotrochus Brown
Pilotrochus Brown (1977) is a very strange myrmicine ant known only from il single worker specimen collected in a Berlese sample of forest litter and humus along the road t o AnosibC, 3 3 km south of Moramanga in east central Madagascar. The single species, P.
besmerus Brown, has I large teeth alternating with denticles on the mandibular masticatory margin, of which the apical tooth is smaller than the basals. The antennae are 8-segmented and the funiculus ends in a strong 2-segmented club. The frontal carinae are far apart and expanded, and large deep scrobes are present which have the eyes situated ventrolaterally. The propodeal spiracle is close t o the margin of the declivity and tibia1 spurs are absent. The petiole is curved-clavate and it and the postpetiole in general are similar to Ishakidris. The apparently glandular mesopleural organ seen in Ishakidris is much more strikingly developed in Pilotrochus and is roughly circular, though apparently not subtended by the open groove seen in Ishakidris.
Despite these similarities there are a number of important characters in which Pilotrochus differs from both Ishakidris and Phalacromyrmex. For instance, the mandibles are not strongly arched-downcurved as in the last two and the labrum is thickly linguiform rather than bilobate. The frontal carinae are excavated immediately behind the narrow lateral portions of the clypeusand do not flank it laterally, so that in full-face view the anterolateral corners of the head are formed by the lateral portions of the clypeus as is usual in myrmicines. The propodeum is unarmed, lacking the lamellate spines characteristic of Phalacromyrmex and Ishakidris.
Having considered the diagnostic characters of Pilo troc hus, Phalacromy rm ex and Is ha kidris, and taking into account the shortage of material in all three genera (including a complete lack of sexual and immature forms), I presently suspect that the similarities outweigh the differences sufficiently t o imply a distant common origin for all three. In particular I am impressed by the construction of the mandibles, antennae and pedicel segments, which are suspected of being synapomorphies rather than convergently acquired characters. If Ishakidris and Phalacromyrmex are indeed related t o Pilotrochus then the obvious corollary is that Pilotrochus split from the parent stock much earlier than Ishakidris and Phalacromyrmex, which resemble each other much more closely than either one resembles Pilotrochus. In summary then, whilst Pilotrochus is not immediately related t o Phalacromyrmex and Ishakidris, it may well be associated at a higher level of classification and the three together may represent relicts of a once much more common and widespread fauna.
It would probably be safe to designate a tribal-group name to hold these three genera but, as the existing tribal-level classification within the Myrmicinae is in such a decrepit state, I can see no advantage to adding yet another formal name t o the confusion. Consequently I propose an informal genus-group, the Phalacromyrmex-group, t o hold these three genera until such time as a functional tribal classification of the Myrmicinae can be achieved.
Myrmicine ants belonging t o the Phalacromyrmex-group are diagnosed by possessing the following suite of characters in comhination in the worker.