Cyatta abscondita,

Jeffrey Sosa-Calvo, Ted R. Schultz, Carlos R. F. Brandao, Christiana Klingenberg, Rodrigo M. Feitosa, Christian Rabeling, Mauric, 2013, Cyatta abscondita: Taxonomy, Evolution, and Natural History of a New Fungus-Farming Ant Genus from Brazil, PLOS ONE 8 (11), pp. 1-20: 5-16

publication ID

http://doi.org/ doi:10.1371/journal.pone.0080498

persistent identifier

http://treatment.plazi.org/id/7C5BBA0F-A67B-12EF-258D-D2BC9D4CFCDE

treatment provided by

Christiana

scientific name

Cyatta abscondita
status

sp. n

Cyatta abscondita  , sp. n

urn:lsid:zoobank.org:act:3260572C- B7CE-429C-9440-27613BDBE 69E

Holotype, worker. labeled: " BRAZIL: DF: Brasilia; Faz. Aegua Limpa ; 1106 m; 47.90133° W 15.9524° S ± 5m; 20.ix. 2011; (J. Sosa-Calvo, T.R. Schultz, C.T. Lopes); nest series; Cerrado sensu stricto; in ground; JSC110920-01 " [ MZSP, unique specimen identifier No. USNMENT00758173]GoogleMaps  .

Paratypes. same data as holotype [3w, MZSP, USNMENT00758172, 00756921, 00758307], [1dg, 4w, USNM, USNMENT 00758174, 00758223, 00758316-18] [3w, USNM, USNMENT00521881 (EtOH vial)].GoogleMaps  Same data as holotype, but " 19.ix.2011; JSC110919-02 " [1dg, MZSP,

USNMENT00758175], [2w, USNM, USNMENT00758176-77] [1w, USNM, USNMENT00521910 (EtOH vial)]. Same data as holotype, but " 16.iv.2010; JSC100416-04 " [1w, MZC,

USNMENT00758180] [1w, USNM, USNMENT00521908 (EtOH vial)]. Same data as holotype, but " 1099 m; 47.90129° W 15.95242° S ± 3m; 23.ii.2009; (J Sosa-Calvo & TR Schultz); nest series on ground; JSC090223-26 " [1w, USNM, USNMENT00758178]GoogleMaps  . Same data as previous entry, but " 16.iv. 2010; nest series underground; JSC100416-01 " [1w, DZUP, USNMENT00758319] [2w, 1dg (currently misplaced) USNM, USNMENT00758323-24, 00758325] [3w, USNM, USNMENT00521886 (EtOH vial)]GoogleMaps  . Same data as holotype, but " Garden near dorms ; 1071 m; 47.93567° W 15.94938° S ± 3m; 12.iv.2010; (J Sosa-Calvo, TR Schultz, CT Lopes); nest series; back yard; under ground; JSC100412-01 " [1w, DZUP, USNMENT00758224], [1w, MZSP, USNMENT00758179], [3w, USNM, USNMENT00758320-22] [3w, USNM, USNMENT00521917 (EtOH vial)]GoogleMaps  .

GoogleMaps 

GoogleMaps 

Worker measurements. Holotype (Paratypes, n=9). TL 2.40 (2.29-2.56), WL 0.62 (0.58-0.65), HL 0.53 (0.50-0.55), HW 0.48 (0.48-0.51), SL 0.48 (0.43-0.50) ML 0.34 (0.32- 0.36), EL 0.13 (0.12-0.13), PL 0.17 (0.14-0.19), PPL 0.22 (0.20-0.24), GL 0.52 (0.49-0.64), CI 90 (90-95), SI 101 (90- 100), MI 64 (63-68), FLD 0.17 (0.17-0.18).

Additional material examined. Same data as holotype, but " 1099 m; 47.90129° W 15.95242° S ± 3m; 16.iv.2010; (J Sosa-Calvo & TR Schultz); stray worker; JSC100416-08 " [1w, USNM, USNMENT00521890 (EtOH vial)]GoogleMaps  . Same data as previous entry, but " 15.iv.2010; (J Sosa-Calvo & TR Schultz); nest entrance; JSC100415-03 " [4w, USNM, USNMENT00521907 (EtOH vial)]GoogleMaps  . BRAZIL: CE: Crateüs; Croatä; RPPN Serra das Almas ; [330m; 40.67978° W 05.16479° S]; 20-30.iv.2003; (Y Quinet); pitfall trap; Caatinga , Carrasco ; SP 60 [1w, MZSP]GoogleMaps  . DF: Brasilia; Reserva Ecol. IBGE ; Parcela Bienal Tardia ( Projeto Fogo ); 1100m; 48.05000° W 15.85000° S; 3012008; (J Maravalhas); pitfall trap;. Cerrado sensu stricto; B2/CT [2w, MZSP]GoogleMaps  . MG: Paineiras; Fazenda Olho D'Aegua ; [611m; 45.53527° W 18.90628° S]; 22-24.v. 1999; (AA Tavares); Winkler #2; Cerrado [1w, MZSP]GoogleMaps  . MG: Paracatu; Reserva Particular de Patrimoenio Natural do Acangau ; 671m; 47.0583056° W 17.1790833° S; 12.iv.2011; (TLM Frizzo); pitfall trap; Cerrado sensu stricto [2w, MBC- UFU]GoogleMaps  . SP: Itirapina; 9.5 km northwest of Itirapina; Broa preserve; 530m; 47.87754° W 22.18517° S; 21.vii.2011; (C Rabeling & M Bacci Jr); nest series; Cerrado sensu stricto; under ground; CR110721-04 [1w, MZSP, USNMENT00758220; 1m, MZSP, USNMENT00758204], [1w, MZC, USNMENT00758221], [2w, USNM, USNMENT00758219, USNMENT00758222], [6w, BLME & CRC, USNMENT00758190-95].GoogleMaps  Same data, but CR110721-05 [1w, MZSP, USNMENT00758217], [1w, MZC, USNMENT00758218], [1w, USNM, USNMENT00758206], [10w, BLME & CRC, USNMENT00758207, USNMENT00758253-61].GoogleMaps  Same data, but CR110721-08 [3w, BLME & CRC, USNMENT00758196-98].GoogleMaps  Same data, but 26.vii.2011, CR110726-09 [3w, BLME & CRC, USNMENT00758199-201].GoogleMaps  Same data, but 28.vii.2011, CR110728-01 [3w, BLME & CRC, USNMENT00758262-65], [1w, USNM, USNMENT00758205]GoogleMaps  . SP: Pindorama; Fazenda

Äguas Claras; 494m; 48.6873° W 21.4023° S; 16.viii.2011; (GA Castilho); Floresta Estacional Semidecidual ; pitfall B1 [1w, MZSP]. SP: Sales; Estacaeo Experimental ; 414m; 49.3013° W 21.5222° S; 17.viii.2011; (GA Castilho); Floresta Estacional Semidecidual ; pitfall I4 [1w, MZSP]GoogleMaps  .

GoogleMaps 

Measurements. TL 2.03-2.60, WL 0.52-0.66, HL 0.48- 0.55, HW 0.46-0.54, SL 0.42-0.51, ML 0.20-0.38, EL 0.10- 0.14, PL 0.13-0.19, PPL 0.21-0.25, GL 0.47-0.59, CI 87-95, SI 91-104, MI 40-68, FLD 0.16-0.18 (n=13).

Etymology. The specific name " abscondita  " refers to the exceedingly secretive nature of this species, which, after being recognized from a few rare specimens, proved frustratingly elusive during multiple attempts to locate it in the field.

Worker. Head. in full-face view subrectangular, slightly longer than wide ( CI 87-95); sides subparallel. Mandible subtriangular with four well-developed teeth; apical tooth twice as large as subapical tooth; diastema between subapical tooth and 3rd tooth shorter or slightly shorter than diastema between 3rd and 4th teeth (Figure 2c); dorsum of mandible reticulate and with appressed hairs (Figure 2d); masticatory margin of mandible, including apical tooth, smooth, shining, and darker in color than rest of head, with long, simple hairs. Clypeal apron (anteclypeus) convex to almost triangular, smooth, and shining; unpaired median setae (length 0.07-0.10 mm) originating slightly before (anterior to) posterior edge of clypeal apron and almost or as long as antennal pedicel, not reaching apex of insertion. Medially these carinae developed into lamellae mandible (Figures 2c,d); clypeus with pair of lateral transverse perpendicular to clypeal face, thus forming a wall that divides carinae, each extending from below frontal lobe to mandibular the clypeus laterally into anterior and posterior areas, very almost or as long as antennal pedicel, not reaching apex of mandible (Figures 2c,d); clypeus with pair of lateral transverse carinae, each extending from below frontal lobe to mandibular insertion. Medially these carinae developed into lamellae perpendicular to clypeal face, thus forming a wall that divides the clypeus laterally into anterior and posterior areas, very likely homologous with clypeal morphology of closest relative, Kalathomyrmex emeryi  (Figure 2c); medially clypeus is not so divided, face extending posterad between frontal lobes. Frontal lobes reduced, convex, barely covering antennal insertions (Figures 1a, 2c). Frontal carina fading out posteriorly at midlength of compound eye (Figures 1a, 2c). Well marked triangular area with concave anterior margin between frontal lobes reticulate, bordered anteriorly by rounded finger of clypeus, which extends broadly posterad between frontal lobes. Compound eye set slightly before middle of head, with 7-9 ommatidia at maximum length and 6 ommatidia at maximum width (33-47 ommatidia in total). Antennal scape covered with minute, simple, appressed hairs; antennal scape wider at seven-tenths of its length, and slightly surpassing posterolateral corners of head when laid back over head capsule; first funicular segment (pedicel) slightly longer than or as long as second and third funicular segments combined. In full-face view, cephalic margin deeply notched medially (i.e., at vertex) and rounded laterally (Figure 1a), shallow vertexal sulcus extending medially towards frontal lobes, fading at eye level; in lateral view, ventral face of head slightly convex. Hypostomal teeth absent. Palp formula 4,2 (Figures 2a,b).

Mesosoma. Profile of promesonotal dorsum in lateral view distinctly tuberculate, tubercles attenuate and blunt (Figures 1c,e, 2a). In dorsal view, promesonotum with raised shield-like area, broad anteriorly and narrowing posteriorly, distinctly separated from lower, lateral promesonotum; raised area formed anteriorly by triangular lateral pronotal tubercles and two median low and approximate pronotal tubercles and posteriorly by eroded remnants of promesonotal tubercles; lower, lateral area of promesonotum in dorsal view subtriangular, broader and anterolaterally angled anteriorly; in lateral view, inferior corner of pronotum rounded, lacking spines or angles. Anepisternum indistinctly separated from katepisternum. Metanotal groove relatively broad and strongly impressed, in lateral view extending to antero-ventral margin of metapleuron. Metapleura ventrally with two spiniform processes between mid and hind coxae, best seen by removing hind legs. Basal (dorsal) face of propodeum in lateral view a low, rounded, protuberance posterior to metanotal groove; in dorsal view, basal face very small, raised above remainder of propodeum, and narrowing anteriorly; declivous face of propodeum behind protuberance concave; propodeal spines triangular (Figure 1c), obliquely directed upwards and strongly diverging in dorsal view; declivity of propodeum much longer than base (dorsum); propodeal spiracle opening in an angle of 45° in relation to main body axis; in lateral view, propodeal lobes rounded without posterior projections.

Peduncle of petiole vestigial; in lateral view, petiolar node well developed, subquadrate, with anterior face almost straight and vertical; dorsum of petiolar node short and almost flat, meeting vertical posterior face in slightly rounded angle; ventral face of petiole slightly concave or straight medially, lacking petiolar process (Figures 1c,e); in fronto-dorsal view, node of petiole shallowly V shaped, dividing node into a pair of rounded tubercles. Postpetiole robust, almost twice as long as petiole and slightly less than 0.5x gaster length; in dorsal view, postpetiole subtriangular, anterior margin rounded, lateral margins slightly diverging posteriorly; posterior margin with deep median impression, forming two distinct small lobes; in lateral view, anterior portion of postpetiole convex, postpetiole relatively compressed dorsoventrally; ventral projections absent (Figures 1c,e). In profile, gaster elliptical and dorsally finely reticulo-striate; in dorsal view, apical margin of pygidium (gastral segment IV, i.e., abdominal segment A 7) medially emarginate, bilobed; gastral sternite IV (hypopygium, i.e., A 7) covered with simple decumbent hairs; in lateral view, pygidium rounded, laterally overlapping and concealing the hypopygium; pygidium weakly reticulate and shiny; in ventral view, pygidium posteromedially emarginate (i.e., V-shaped), the triangular hypopygium fitting within the emargination of the pygidium (Figures 2e,f). Sting apparatus present, protruding through emargination on pigydium.

Color pale yellow to light brown; antennae, mandibles, and legs lighter than rest of body. Body integument areolate, with short appressed simple hairs, appearing almost hairless.

Gyne. As in worker description, but with caste-specific morphological differences as follows. All gynes studied are dealate. Head: Eyes large, with 10-11 ommatidia in maximum length and 9 ommatidia in maximum width, ~65 ommatidia total; median ocellus rounded, located in a median sulcus extending almost from the occipital carina in the back of the head to the middle of the frons; integument surrounding ocelli darker in color than elsewhere. Clypeus with unpaired median seta arising on short transverse wrinkle-like ridge that crosses clypeal apron; two to four short simple appressed hairs on clypeal apron on each side of median clypeal seta. In full-face view, cephalic border with median (vertexal) notch, not as deep as in worker. Mandibles dorsally coarsely rugose. Mesosoma. Pronotal dorsum conspicuously areolate, lacking anterior pronotal tubercles; lateral pronotal tubercles present, blunt and small; humeral pronotal tubercles vestigial. Mesoscutum, in dorsal view, rounded to slightly ovate and overall reticulo-rugose; dorsum of mesoscutum, in profile, almost flat; mesoscutal sulcus, in dorsal view, short, not extending more than 1/4 length of mesoscutum; notauli absent; parapsidal lines short, inconspicuous, and extending nearly to lateral margin of mesoscutum; transscutal suture conspicuous. Scuto-scutellar sulcus deep and with ~7 transverse carina; margin of axilla rounded, dorsally reticulo-rugose. Scutellum posteriorly weakly bidentate, dorsally rugose and with a shallow median longitudinal groove. Anapleural sulcus deep, with transverse carinae, dividing mesopleuron into anepisternum and katepisternum. Metanotal groove extended into a complete metanotal-propodeal suture (sensu Serna & Mackay [90]). In profile, metanepisternun ( sensu Serna & Mackay  [90] present, small. In profile, metanotal groove conspicuous, continuous with mesometapleural suture. Ventral metapleural processes present as a pair of spiniform tubercles between the mid and hind coxae, similar to, but longer than, those present in worker. Propodeum with pair of short, rightangled denticles; dorsum, lateral margin, and declivity of propodeum reticulate. Metasoma: petiole as in worker. Dorsum of postpetiole reticulo-rugose. Dorsum of gastral tergite IV ( A 7) rugulose; pilosity as in worker; gastral tergite and sternite IV (pygidium and hypopygium; A 7) as in worker.

Measurements. TL 3.27-3.32, WL 0.86-0.87, HL 0.65- 0.66, HW 0.60-0.63, SL 0.56, ML 0.41-0.42, EL 0.16-0.17, PL 0.25-0.28, PPL 0.29-0.30, GL 0.79-0.81, CI 91-96, SI 89-93, MI 62-65, FLD 0.20-0.21 (n=2).

Male. A medium-sized male with head large relative to size of the mesosoma. Mandibles broadly triangular, apical and subapical teeth present large; remaining tooth minute, indistinct; texture coarsely granulate. Palp formula 4,2. Clypeus broadly trapezoidal in frontal view; anterior margin convex, with a long median seta (0.11 mm) originating at the anterior margin and projecting over the mandibles; in lateral view anterior margin of clypeus forming a lamella projecting over the mandibles. Frontal lobes triangular, only partly covering the condylar bulbs of the scape in full face view. Antennae with 13 segments; scape surpassing the posterior border of the head by 1/3 of its length. Antennal funicular segment II (0.08 mm) almost as long as funicular segment I (pedicel; 0.11 mm) (Figures 3a,b). Eyes conspicuously large, at maximum diameter approximately half as long as the entire head, counting ~15 ommatidia in maximum width and ~23 ommatidia in maximum length. Ocelli large, elevated above the remainder of the head. Surface of head coarsely granulate, finely rugulose around the ocelli. Tergum of promesonotum not distinctly enlarged, giving the mesosoma a rather slender appearance in lateral view. In dorsal view, lateral pronotal teeth pyramidal, twice as wide at the base than high, with sharp tips. Propodeal spines reduced to broad teeth with rounded tips. Anterior peduncle of the petiole about the same length as the petiolar node. Postpetiole wider than long; trapezoidal in dorsal view; posterior margin slightly concave. In lateral view, postpetiole with a broadly rounded ventral lobe. Reticulate sculpture on gaster finer in appearance than on the remainder of body, which tends to be areolate; gastric tergites moderately lustrous; rest of body with a weak silky shine. Body surface sparsely covered with short appressed setae, only ventral side of postpetiole with 10 erect setae. Body dark, blackish brown; legs and antennae slight lighter in color, yellowish to dark brown. Forewings with closed basal ( BC), costal ( CC), submarginal ( SMC 1), marginal ( MC), and subbasal ( SBC) cells; submarginal cell 2 and discal cell 2 open (Figure 3d). Hindwings with closed basal cell and open marginal subbasal and discal cells (Figure 3d). Left forewing with closed discal ( DC 1) cell, whereas the right forewing lacking this cell. The presence of a closed discal cell in the forewing is, so far as is known to us, unique in the Attini  HNS  ; this character is absent in all other attine genera, including the closest relative of C. abscondita, Kalathomyrmex emeryi. A closed discal cell is plesiomorphically present in many ant species, including those in genera closely related to the Attini  HNS  such as the Blepharidattini  HNS  , Cephalotini  HNS  , Dacetonini  HNS  , and Pheidolini  HNS  .

Measurements. EL 0.25, EW 0.27, FL 0.81, FLD 0.21, GL 0.87, HL 0.58, HW 0.73, IOD 0.42, ML 0.21, MI 37, PL 0.27, PPL 0.25, PPW 0.38, PrW 0.54, PW 0.21, SL 0.54, TL 3.21, WL 1.04, CI 126, SI 127 (n=1).

Larva. Description based on SEM study of two specimens, late- (probably fourth) instar larvae of uncertain (but probably worker) caste. Due to collapsed condition of specimens, habitus profile could not be characterized with certainty, but is consistent with the "attoid" profile category of Wheeler & Wheeler [104], i.e., with a moderately curved, ventrally shortened profile. Thoracic-abdominal articulation apparently absent, thoracic intersegmental constrictions superficial, deep lateral depressions associated with abdominal spiracles absent, all states shared with other Attini  HNS  . Remarkably, body hairs present dorsally and laterally, a condition otherwise common in the Myrmicinae  HNS  but rare in the Attini  HNS  , in which larvae usually lack dorsal and lateral hairs. In the Attini  HNS  , such hairs are known to be present only in the larvae of Mycocepurus goeldii  and M. smithii  [92], where their presence may be plesiomorphic, and in Sericomyrmex  and some Acromyrmex  species, where their presence is likely secondarily derived. Predominant hair type bifurcate with "anchor tips" (Figures 4a,d). Two rows of dorsomedian, very long anchor-tipped hairs present (Figure 4e). Labrum monolobate, narrow, bulging. Anterior setae present as papillae. Mandibles typically attine: short, fleshy, subconical. A distinct, undivided apical mandibular tooth and no subapical teeth; spinules evenly distributed on all mandibular surfaces. Mandibular gnathobases absent. Basal portions of maxillae fused with head capsule. Maxillary palp widely removed laterad from galea, a synapomorphy for the Neoattini  HNS  . Galea reduced, present as two sensilla surmounting a low protuberance, as in all Attini  HNS  except for some Myrmicocrypta  species. Maxillary palp digitiform, maxillary accessory palpal sensillum absent. A single seta present laterad of maxillary palp, a character shared with Mycocepurus  species. As in most attines, labium feebly protruding, lateral sericteral protuberances absent, labial palps digitiform. Labial spinules present on anterior surface slightly dorsal to the sericteries. Hypopharyngeal spinules multidentate and apparently densely distributed. On the head, genal lobes absent, a state in the Attini  HNS  shared with Myrmicocrypta  , Apterostigma  , Mycocepurus  , and Mycetarotes  species. Supraantennal setae present and abundant, a condition common in the subfamily but otherwise present in the lower Attini  HNS  only in M. goeldii  . Subantennal (genal) setal arrangement plesiomorphic for the tribe, consisting of around 12 setae on each gena. Supraclypeal setae present and setiform. Two clypeal setae present. Spinules absent on the head dorsad of the labrum, the state common to most attines. Due to the poor condition of specimens, most ventral thoracic/abdominal characters could not be studied, including the presence/ absence of: leg vestiges, prothoracic food anchor, ventromedian protuberances on various segments, papilliform spinules, and hairs.

Two setal sockets occur ventral of the anal opening on abdominal segment IX (Figure 4f). No other setae are associated with the anal opening. Ventral anal lip absent.

Comments. The new genus Cyatta  shares with other genera belonging to the tribe Attini  HNS  : (i) the presence of a thick unpaired median seta arising from the clypeal apron (considered a synapomorphy for the tribe by Brandao & Mayhe-Nunes [105]; but, along with the clypeal apron, presumed secondarily lost in Kalathomyrmex  [30]); (ii) 11 antennal segments in the worker and gyne, 13 in the male (the latter secondarily reduced in some Cyphomyrmex  , Mycetagroicus  , Sericomyrmex  , Trachymyrmex  , and social parasites); (iii) palpal formula 4,2 (plesiomorphic for the Attini  HNS  , secondarily reduced in Apterostigma  and some social parasites). Larvae of Cyatta  share with the larvae of other attine genera the: (iv) thoracic-abdominal articulation apparently absent; (v) thoracic intersegmental constrictions superficial; (vi) deep lateral depressions associated with abdominal spiracles absent; (vii) short, narrow labrum; and (viii) fleshy, subconical mandibles. Behaviorally, Cyatta  shares with other Attini  HNS  the cultivation of fungi for food.

The genus Cyatta  shares with other members of the neoattine clade: (i) the antennal scape of the male long, longer than the sum of the length of antennal funicular segments I-III; (ii) the first funicular segment (pedicel) of the antenna of the male longer than second funicular segment; (iii) the petiole in workers somewhat sessile; (iv) the lack of hypostomal teeth in workers and gynes; and (v) the maxillary palp of the larva widely removed laterad from the galea.

Cyatta  shares with its sister genus, Kalathomyrmex  , (i) the lack of a tubercle or spine on the inferior lateral margin of the pronotum, a symplesiomorphy shared with the paleoattine clade; (ii) the clypeus with a pair of lateral transverse carinae, each extending from the frontal lobe to the mandibular insertion and each medially developed into a lamella perpendicular to the clypeal face, thus forming a wall that divides the clypeus laterally into anterior and posterior areas; and (iii) the mandibles of the male with three teeth, of which the apical and preapical teeth are the largest and have a multidentate (sawlike) margin.

Cyatta  differs from its sister genus Kalathomyrmex  , however, by (i) having, on the forewing of the male (forewing of gyne unknown), a closed marginal cell (Figure 3d) (open in the forewings of both the male and gyne of Kalathomyrmex  [Klingenberg and Brandao ([30]), therein as radial cell]); (ii) the mesoscutum of the male with strongly impressed notauli (absent in the male of Kalathomyrmex  ); (iii) the pronotum of the male with lateral pronotal tubercles present, pyramidal (the pronotum in the male of Kalathomyrmex  lacks any tubercles); and (iv) the psammophore absent in the worker, the gyne, and the male.

In addition to the previously mentioned absence of an inferior pronotal tubercle in adult workers, shared with Kalathomyrmex  , Cyatta  differs from all or most other Neoattini  HNS  in a number of larval character states shared with the Paleoattini  HNS  , suggesting that they may be retained symplesiomorphies, including: (i) dorsal and lateral body hairs present and abundant, shared with Mycocepurus  species; (ii) a single seta present laterad of the maxillary palp, shared with Mycocepurus  species; (iii) supra-antennal setae present and abundant, shared with Mycocepurus goeldii  ; (iv) genal lobes absent, shared with the paleoattines and the neoattine genus Mycetarotes  . Larval characters are unstudied in Kalathomyrmex  . In addition, the worker and gyne of Cyatta  differ from members of the neoattine clade by (v) the node of the petiole well developed, high (Figures 1c,d, 2a). Most notably, Cyatta  differs from all other attine genera and species by the following autapomorphies: (i) mandible of the worker and gyne with four teeth (Figures 1b, 2c); (ii) in ventral view, metapleura of the worker and gyne with two spiniform processes between the mid and hind coxae, apparently absent in the male; (iii) apical margin of the pygidium medially emarginate, V-shaped (Figures 2e,f); and (iv) forewing of the male with a closed discal cell (Figure 3d).

Based on the extreme degree of morphological divergence and the results of the divergence dating analyses (see below), we have chosen to describe Cyatta  as a new genus rather than to describe it as a species within the genus Kalathomyrmex  .

Discovery history. In 2003, a single stray worker of C. abscondita  was taken in a pitfall trap as part of an ant survey conducted at the Reserva Particular do Patrimonio Natural Serra das Almas, Crateus, CE, Brazil, a relatively undisturbed area of Caatinga, a biome characterized by deciduous thorny woodland vegetation [106]. The specimen was deposited in the MZSP ant collection, where it was at first associated with the Mycetophylax  species group, but subsequently recognized as a new neoattine genus by CK and CRFB. This isolated specimen inspired the first attempt to locate C. abscondita  in the field in Serra das Almas in 2009 by CRFB and RMF. Unfortunately, it was the end of the rainy season and the soil was covered by a dense layer of grass, impairing observations of all small and inconspicuous ants. Visual searching and leaf-litter extraction failed to locate additional specimens, as did subsequent surveys at the same locality.

In 2008, two workers were taken in pitfall traps in the Instituto Brasileiro de Geografia e Estatistica ( IBGE) Cerrado preserve, near FAL in Brasilia, DF, Brazil. These specimens, deposited in the MZSP, inspired attempts by JSC, TRS, CTL, and HLV to locate the species at this locality beginning in 2009. The first such attempt yielded only the collection of a series of stray workers and an unsuccessful nest excavation; however, subsequent visits resulted in the excavations of multiple nests and collections of gynes, larvae, and cultivated fungi.

The only known male of the species was fortuitously collected in 2011 by CR and MB when they accidentally encountered two nests of C. abscondita  while excavating a nest of Mycocepurus goeldii  in the Broa Preserve, Itirapina, SP, Brazil.

The earliest known collection of C. abscondita  was that of a stray worker taken in a leaf-litter sample in Paineiras, MG, in 1999, only recently discovered in the entomological collection at MZSP and recognized as belonging to this species. Most recently, in 2011, two workers of C. abscondita  were recovered from pitfall traps in fragments of semideciduous forests in the Sales and Pindorama municipalities in northwestern Sao Paulo state. This history of discovery indicates that C. abscondita  is rarely collected by traditional methods. The cryptic nature of foragers and of nest entrances makes it almost invisible to traditional hand collecting. The rarity of individuals in pitfall and leaf-litter samples remains puzzling, since the concentrations of nests encountered at FAL and Broa Preserve suggest that it is locally abundant. Now that the genus and species are recognized and described, we hope that additional specimens will be identified in unsorted material in collections as well as in newly collected material from ant surveys in Brazil and perhaps even elsewhere in South America.

Natural history Macrohabitat

Most collections of C. abscondita  are from Cerrado localities (Figure 5). These include Fazenda Agua Limpa ( FAL) near Brasilia, the Broa preserve in Sao Paulo, the IBGE Cerrado preserve in Brasilia, DF, the Fazenda Olho D'Agua in Paineiras, MG, and the Reserva Particular do Patrimonio Natural ( RPPN) do Acangau in Paracatu, MG, all of which are characterized by diverse Cerrado phytophysiognomies, ranging from campo limpo to Cerrado sensu stricto [107]. The predominant habitat, Cerrado sensu stricto, is a low canopy arboreal woodland that is characterized by the presence of small trees with a canopy height of less than 7 meters, shrubs, and abundant ground vegetation [108-110]. Cerrado soil is typically a red-yellow latosol, largely composed of well-drained and nutrient-poor quartz sand with moderate clay content below 15% [107,111]. Both FAL and Broa have typical Cerrado climates with a marked dry season from May to September and with a mean annual temperature and precipitation of 23°C and ~ 1420 mm, respectively [110]. For a complete account of vegetation and soil compositions at FAL and Broa, see [112] and [113]. The label data associated with the C. abscondita  worker from Fazenda Olho D'Agua indicates that it was obtained from a leaf-litter sample from a Winkler extractor. This suggests that this worker was likely taken in either a riparian forest or in "Cerradao," because it is in such areas that trees are dense enough to produce conspicuous accumulations of leaf litter.

Some specimens of C. abscondita  have been taken outside the Cerrado biome (Figure 5). Most notably, a stray worker was recovered from a pitfall sample taken in 2003 in a relatively undisturbed area of Caatinga, a biome characterized by deciduous thorny woodland vegetation [106], in the RPPN Serra das Almas, Crateus, CE, Brazil. More recently, in 2011, two workers of Cyatta abscondita  were taken in pitfall traps in fragments of semideciduous forests in northwestern Sao Paulo state. This region is considered a transition zone between Cerrado and endangered coastal Atlantic Forest [114].

Microhabitat. Four nests of Cyatta abscondita  were excavated at FAL and two nests at Broa (summarized in Table 1). Nest entrances of four additional nests were located at FAL; however, attempted excavations of these nests failed. At FAL, seven of the excavated nests occurred on the side of a littleused dirt service road in Cerrado sensu stricto (Figure 6b) and the eighth (nest 1, Table 1) on the lawn of the FAL dormitories (Figure 6a). The roadside nests were directly exposed to sunlight for most of the day, whereas the lawn nest was shaded by the adjacent building in the morning and afternoon. At the Broa preserve, both nests occurred in the shade of trees in Cerrado sensu stricto. Both Broa colonies were excavated serendipitously during excavations of Mycocepurus goeldii  nests and the C. abscondita  nest entrances were not observed.

Foraging behavior. Foraging workers of Cyatta abscondita  are difficult to locate because colony sizes are small, workers forage individually, and individuals are very small and cryptic (Figure 6c). In mid-September, which coincides with the beginning of the rainy season at FAL, three to four individuals from three different nests, including nests 3 and 4 (Table 1), were observed foraging at night between 22h and 23h. In February and April, at the end of the rainy season, ants from five FAL nests, including nests 1 and 2 (Table 1), were likewise observed foraging individually only after sunset. Only in the case of FAL nest 1 were workers observed to forage in the early afternoon on two consecutive days in April between 13h and 15 h during a time when the nest entrance was shaded from direct sunlight. Unlike the other nests, nest 1 occurred in a well-watered, human-managed grassy lawn. Nests 5 and 6 (Table 1), which were collected in July at the height of the dry season at the Broa preserve in Sao Paulo State, were located in the shade and, unfortunately, neither foraging nor nestbuilding activity was observed.

The entrance of one FAL nest ( JSC 090223 -26) was located ~4.5 centimeters from the entrance of an adjacent Mycocepurus goeldii  nest. At around 23h a Cyatta abscondita  worker was observed lurking slightly inside the nest entrance while workers of M. goeldii  foraged on bait (granules of Cream of Rice cereal) placed near the nest entrances. When M. goeldii  workers were absent, the C. abscondita  worker darted out to retrieve a piece of bait and quickly returned to its nest. This lurking and rapid foraging behavior was repeated until the supply of bait was depleted. In rare cases of contact between C. abscondita  and M. goeldii  workers, C. abscondita  workers were observed to remain motionless. Aggressive interactions were not observed.

Nest architecture. At FAL, nest entrances of Cyatta abscondita  consisted of a single, inconspicuous, hole in the ground of approximately 1 mm in diameter without any accompanying mound or turret (Figures 6c,d). As mentioned above, the entrance of one nest was located in the mound of a Mycocepurus goeldii  colony ~ 4.5 cm from the M. goeldii  nest entrance. At the Broa preserve, nest chambers of C. abscondita  were encountered serendipitously while excavating M. goeldii  nests and the nest entrances were not observed.

At FAL, excavations of eight nests were attempted. Four excavations ( JSC 090223 -26, JSC 110914 -02, JSC 100415 -03, JSC 100416 -04) failed (i.e., neither chambers containing fungus gardens nor gynes were found, but workers were collected at their respective nest entrances), but chambers containing fungus gardens were located in four nests (Table 1). Nests contained three to eight chambers. In FAL nests 1 and 3, which contained 4 and 3 chambers respectively, chambers were roughly arranged vertically below the nest entrance (Table 1), although it is possible that additional, laterally dispersed chambers were missed during the excavations. At FAL, nests 2 and 4 contained 7 and 8 chambers, respectively, some of which occurred at the same depth but were laterally separated from each other (Figure 6f). The shallowest chamber encountered (nest 3, FAL) was 29 cm deep and the deepest chamber (nest 6, Broa) was 195 cm deep. Because no gynes were found during the nest excavation at Broa, it is entirely possible that additional chambers occurred below a depth of 2 meters. Chambers were elliptically shaped, 1-2.5 cm wide and 2-5 cm high (Figure 6e). The largest garden chamber encountered (nest 2, FAL) was 2.5 x 5.5 cm; at Broa, a similarly sized chamber (nest 5) contained ~50 hanging garden filaments (Figure 6e). Some chambers were empty; in one case, an empty chamber contained three polydesmid millipedes.

Demography. Dealate gynes were collected in three of the eight excavated nests, suggesting that in five nests additional chambers remained undiscovered in the soil, or that gynes escaped into adjacent tunnels. In each of the three queenright nests (nests 2, 3, and 4), a single gyne was consistently encountered in the deepest chamber (see Table 1); however, additional chambers may have been present at greater depths, because excavations were generally terminated upon encountering the gyne. Brood was found only in FAL nest 4, which was collected in September, the beginning of the rainy season, suggesting that colonies of Cyatta abscondita  reduce their reproductive activities during the dry season. The maximum number of workers encountered in colonies ranged from ~20 ( FAL nest 4) to 26 (Broa nest 5). One male was collected in nest 5 at Broa on 21 July.

Garden morphology. Gardens were pendant and arranged in filamentous curtains suspended from the chamber ceiling (Figure 5e), similar to the fungus gardens of Mycocepurus  species [70,93,115,116] and of Kalathomyrmex emeryi  ( TRS, JSC, pers. obs.). Single fungal curtains were 5-6 mm long and 1-2 mm wide and a maximum number of 50 curtains were found in a single chamber. Curtains were directly attached to the soil of the chamber ceiling rather than to rootlets. In nest 4, which was maintained in laboratory culture for three months, workers attached garden filaments to the plastic ceiling of the nest box and cultivated suspended gardens. The filaments were firmly attached to the plastic ceiling by an unknown mechanism.

Phylogeny. Results of molecular phylogenetic analyses incorporating four nuclear gene sequences from Cyatta abscondita  confirm the previous finding [95] that the tribe Attini  HNS  is divided by an ancient divergence into two major clades, the Paleoattini  HNS  and the Neoattini  HNS  (Figure 7). Cyatta abscondita  occupies a relatively isolated position in the latter clade, distantly related to the monotypic genus Kalathomyrmex  Klingenberg & Brandao, the result of an early divergence in the Neoattini  HNS  . Its phylogenetic position, nested well within the paraphyletic group of "lower attine ants," strongly supports the hypothesis that C. abscondita  practices "lower attine agriculture" [58].

A relaxed-clock-divergence dating analysis conducted in BEAST using the Bayesian uncorrelated lognormal approach with a normal prior on the root node (as described in Schultz & Brady [95]), resulted in a chronogram in which Cyatta  occupies a position identical to that in the MrBayes results shown in Figure 7. The BEAST chronogram indicates that Cyatta  and Kalathomyrmex  diverged 26 Ma (95% CI = 18-34) and that the earliest possible divergence of the clade ( Cyatta  + Kalathomyrmex  ) from the rest of the Neoattini  HNS  occurred 44 Ma (95% CI = 37-51).

Discussion

The pursuit and discovery of phylogenetically informative new species are arguably among the most important enterprises in systematic biology. Numerous studies have demonstrated the significant effects of taxon representation on phylogenetic inference, including, in addition to tree topology, ancestral character state reconstruction, divergence time estimation, and inferences of evolutionary rates [117-119]. Ward et al. [118] showed that the exclusion of a single relict species resulted in the incorrect reconstruction of the phylogeny of the ant subfamily Dolichoderinae  HNS  ( Formicidae  HNS  ). In addition, the recently discovered ant genus Martialis  , a relict subterranean species known from the Amazon forest in Brazil, has been shown to be the only hitherto known representative of an early diverging branch of the ant tree of life [27]. Within the Attini  HNS  , the recently described ant genus Mycetagroicus  [105,120] was found to be the sister group to the higher attine ants ( Sericomyrmex  , Trachymyrmex  , Acromyrmex  , and Atta  ) and thus to occupy a phylogenetic position transitional between lower and higher agriculture [95]; however, until very recently its fungal cultivar association remained unknown. Subsequent field work documented that Mycetagroicus cerradensis  cultivates a lower attine fungus, suggesting that biological investigations of the genus are critical for understanding the evolutionary transition from lower to higher agriculture [94]. This strategy of reciprocal illumination (i.e., information gathered from the field informing phylogenies, and phylogenies guiding field work) plays a key role for reconstructing and understanding the macro- and micro-evolutionary processes driving the attine agricultural symbiosis.

The results reported here indicate that Cyatta abscondita  possesses an intriguing mosaic of characters, some that are shared with paleoattines, others that are shared with neoattines, and at least one that is shared only with non-attine ants. Because the Neoattini  HNS  , Paleoattini  HNS  , and the non-attine Myrmicinae  HNS  span the ancestral node of the tribe Attini  HNS  , this combination of character states suggests that the morphology, behavior, fungal associations, nest architecture, and other biological characters of C. abscondita  are potentially informative about plesiomorphic character states within the tribe and, consequently, about the early evolution of ant agriculture. At least one character of C. abscondita  , the presence of a closed discal cell in the forewing of the male, is unknown in all other Attini  HNS  . If, as we suspect, this is a retained plesiomorphy rather than an autoapomorphy, then the absence of the discal cell in other Attini  HNS  must be the result of at least three parallel losses, one in the Paleoattini  HNS  , one in the Neoattini  HNS  , and one in Kalathomyrmex  . Another character previously unknown in the Attini  HNS  is the presence of at least two rows of elongate anchor-tipped hairs (Figures 4a,e) on the mid-dorsum of the larva. The function of such anchor-tipped hairs has recently been studied in the nonfungus-farming ant Pheidole rhea  , which utilizes these specialized setae to hang fourth-instar larvae from the nest walls [121]. The widespread presence of this character state in non-fungus-farming ants strongly suggests that it is plesiomorphic for the Attini  HNS  and that its presumably derived absence in most Attini  HNS  may be connected to the fact that larvae are usually more or less enveloped in mycelium deep within the fungus garden rather than hung from the chamber wall. Presumed plesiomorphic adult character states shared by C. abscondita  and the Paleoattini  HNS  include the presence of a rounded inferior lateral margin of the pronotum, also retained in Kalathomyrmex emeryi  . All other neoattine genera, from Mycetarotes  to the leaf-cutter ants, have a denticle or tooth in this position. A number of larval character states are shared with species of the paleoattine genus Mycocepurus  , including the presence of dorsal and lateral hairs, the presence of a single seta laterad of the maxillary palp, and abundant supraantennal setae. The absence of genal lobes is shared with all Paleoattini  HNS  as well as, in the Neoattini  HNS  , with Mycetarotes  species. (Larval characters of Kalathomyrmex  have not yet been documented.) The striking pendant, curtain-like morphology of the fungus garden is a character state shared with Mycocepurus  species as well as with Kalathomyrmex  . Presumed neoattine synapomorphies shared by C. abscondita  and other Neoattini  HNS  include, in adults, (i) the lack of hypostomal teeth, also secondarily lost in some species of Apterostigma  ; (ii) the antennal funicular segment II of males short, as long as or slightly longer than funicular segment I (pedicel), whereas in the males of the paleoattine genera the funicular segment II is long, almost twice as long than the pedicel; and (iii) the wide separation of the maxillary palp from the galea in the larvae. Although both molecular and morphological data indicate that Cyatta abscondita  is the sister species of Kalathomyrmex emeryi  , it is a very distant sister, having diverged from their most recent common ancestor approximately 26 mya. As far as is currently known, C. abscondita  shares with Kalathomyrmex  two unique morphological characters, the form of the clypeus and the morphology of the mandibles in the male. In common with the paleoattines, but differing from all other neoattines, C. abscondita  also shares with Kalathomyrmex  a rounded inferior pronotal corner. In contrast to these shared character states, two synapomorphic and one symplesiomorphic, C. abscondita  notably lacks the defining feature of Kalathomyrmex  , the basket-like psammophore (Gr. kalathos = "basket") and differs not only from Kalathomyrmex  but from all other Attini  HNS  in a number of striking characters, including the 4-toothed mandible, the presence of paired ventral pleural spiniform processes, and the presence of a discal cell in the wing of the male. In fact, the morphology of C. abscondita  is a mosaic of characters of the paleoattine and neoattine clades as well as of closely related non-attine myrmicines. For these reasons we choose to recognize Cyatta abscondita  as a distinct genus within the fungus-farming ants. The discovery, description, and mapping of biological diversity is essential for devising strategies for protecting biodiversity hotspots, i.e., areas that support high concentrations of endemic species and that are threatened due to the rapid loss of habitat as a result of years of unsustainable human exploitation [86], [122]. The Brazilian Cerrado is one such biodiversity hotspot. With only 20% of the original primary habitat remaining, and with only 6.2% of that habitat protected, the Brazilian Cerrado sustains more than four thousand species of endemic plants and more than a hundred species of endemic vertebrates [86]. The Caatinga is even less protected than the Cerrado [123]. The discovery of Cyatta abscondita  in the Cerrado and in the poorly explored Caatinga habitats of Brazil suggests that increasing the search for cryptic and inconspicuous species will lead to discoveries that will fundamentally alter our understanding of insect evolutionary history.

MZSP

Brazil, Sao Paulo, Sao Paulo, Museu de Zoologia da Universidade de Sao Paulo

USNM

USA, Washington D.C., National Museum of Natural History, [formerly, United States National Museum]

MZC

Portugal, Coimbra, Universidade de Coimbra

DZUP

Brazil, Parana, Curitiba, Universidade Federal do Parana, Museu de Entomologia Pe. Jesus Santiago Moure

Kingdom

Animalia

Phylum

Arthropoda

Class

Hexapoda

Order

Hymenoptera

Family

Formicidae

Genus

Cyatta