Cardiocondyla yamauchii Terayama 1999

Seifert, Bernhard, 2024, A taxonomic revision of the ants of the Cardiocondyla wroughtonii group (Hymenoptera: Formicidae) with a checklist of the Cardiocondyla species of the world, Soil Organisms (London, England) 96 (2), pp. 113-144 : 128-129

Cardiocondyla yamauchii Terayama 1999

[types investigated]

Okinawa / Japan 3 worker paratypes from the same sample as holotype: ‘Ada, Okinawa-jima Okinawa Pref. 12.VI.1991 K.Yamauchi leg.’ and ‘ Cardiocondyla yamauchii Terayama, 1999 , Paratype’, SMN Goerlitz. The type workers are allocated to the C. wroughtonii cluster with p = 0.9958 if run as wild-card in a LDA (for details see below).

All material examined. Numeric phenotypical data were available in 64 samples (51 nest samples and 17 single-specimen stray samples) with 140 workers. For details see supplementary information SI1, SI2. Excluding single-specimen samples with unclear separation from C. obscurior , this material originated from Australia (10 samples), Brunei (1), the Comores (1), Egypt (1), Hawaii (4), India (1), Indonesia (7), Japan (3), Kenia (1), Malaysia (13), Mauritius (3), Nepal (1), Philippines (5), Papua New Guinea (2), Singapore (1), Sri Lanka (1), Taiwan (1), Tanzania (1), Tailand (4), USA (5).

Geographic range. As a tramp species of putatively Southeast Asian origin this species is widely distributed over the tropical regions of the world. However, it has not been confirmed so far for South and Central America where C. obscurior is found. Occurrence in buildings in the temperate zone is so far not verified by voucher specimens.

Diagnosis: --Worker ( Tab. 1, Figs 28–31, key). Very small and slightly smaller than obscurior, CS 410

µm. Head short, CL/CW 1.120. Anterocentral clypeal margin straight or slightly notched; central occipital margin usually straight or with a very weak concavity. Postocular distance large, PoOc/CL 0.440. Frons broad (FRS/CS 0.272), frontal carinae caudal of FRS level parallel (FL/FR 1.008). Eye medium-sized, EYE/CS 0.233. Scape slightly shorter than in obscurior, SL / CS 0.772. Promesonotal plane in dorsal view without ‘shoulders’ due to a strongly convex frontal margin and rather straight lateral margins ( Fig. 23A ). Metanotal groove in lateral view deep (MGr/CS 3.65 %) and usually with steep anterior and posterior slopes. Propodeal spines longer than in obscurior (SP/CS 0.193) and their bases more approached (SPBA/CS 0.278). Petiole narrower and slightly lower than in obscurior (PeW/ CS 0.277, PeH/CS 0.328), Axis of petiolar peduncle in lateral aspect deviating by 30° from the longitudinal axis of the petiole node. Postpetiole narrower and lower than in obscurior (PpW/CS 0.437, PpH/CS 0.303), the sternite anterolaterally with a rounded lobiform protrusion on each side which clearly elevates above the level of median surface of sternite; the surface of the sternite thus appearing deeply concave in frontal view. Whole head and mesosoma without any notable rugosity. Paramedian and lateral vertex with deep (sometimes shallower) foveolae of 17–21 µm diameter which are frequently arranged in dense honey-comb arrangement and show an inner elevation around the hair bases of 6–9 µm diameter ( Fig. 31 ); median vertex with slightly smaller and weaker foveolae, occasionally weakly foveolate-carinulate. Whole surface of mesosoma and waist densely foveolate-reticulate, the meshes with lower diameters than on vertex ( Figs 29, 30 ). Pubescence on first gaster tergite moderately long and dilute (PLG/ CS 6.3 %, sqPDG 5.05). Color dimorphism. Light morph (76 % of samples): whole ant entirely light yellowish except for a diffuse brown band in the posterior half of 1 st gaster tergite; this band may be interrupted in the centre to form 2 separated lateral patches (as seen in Wheeler´s var. bimaculata ), PigG1 9.5 ± 8.2 [0, 45] % n= 108. Dark morph (24 % of samples): Head, mesosoma, waist, and appendages light yellowish brown; funicular club blackish brown; 1 st gaster tergite and sternite dark brown, the following segments substantially lighter, PigG1 86.8 ± 19.1 [55, 100] %, n=34.

Taxonomic comments and clustering results. C. wroughtonii is extremely similar to C. obscurior and the occurrence of a dark morph in wroughtonii restricts the use of pigmentation for species separation. Species hypotheses were formed in 84 samples of both species with 205 worker individuals under exclusion of single-specimen samples. In the first step of analysis, the 15 characters CS, CL/CW, PoOc/CL, SL/CS, EYE/CS, FRS/ CS, SPBA/CS, SP/CS, PeW/CS, PeH/CS, PPW/CS, PpH/CS, MGr/CS, sqPDG and PLG/CS were run in the exploratory data analyses NC-Ward, NC-NMDS.kmeans and PCA. Hypotheses were fixed in the controlling LDA when the classifications in the three exploratory data analyses coincided whereas samples with controversial classification were run as wild-cards. In the next run of the LDA the corrected classifications were accepted but now the type samples were run as wild-cards. The posterior probabilities for allocation to the C. wroughtonii cluster were 1.0000 in the types of wroughtonii , 0.9999 in the type of hawaiensis, 0.9997 in the types of longispina , 0.9995 in the types of quadraticeps, 0.9984 in the types of bimaculata and 0.9958 in the types of yamauchii . The posterior probabilities for allocation to the C. obscurior cluster were 0.9993 in the neotype series of obscurior and 0.9992 in the types of bicolor . As a third step, a stepwise character reduction to CS, CL/CW, SL/CS, EYE/CS, FRS/ CS, SP/CS and PPW/CS was performed.This increased the agreement of the exploratory data analyses with the final species hypothesis determined by the controlling LDA to 96.4 % in NC-Ward, 98.8 % in NC-NMDS.kmeans, 98.8 % in NC-part.kmeans ( Fig. 52) and to 98.8 % and in the PCA ( Fig.53 ). This is a mean error in four forms of analysis of only 1.8 %. In laboratory experiments, males of C. obscurior and gynes of C. wroughtonii produced hybrids but the opposite mating combination so far not (K. Yamauchi pers. comm. 2000). Hybridisation is supposed to be rare or absent under natural conditions as concluded from the strong separation of the clusters. However, the available sample size, in particular the low number of 2.4 workers per sample, makes assessment of hybridization frequency difficult. There is so far only one suggestion on a possible hybridization: the only sample of obscurior with a gaster pigmentation corresponding to light morph of wroughtonii (SaNo 1072 Singapore) was allocated to the obscurior cluster with p=0.9456 if run as a wild-card in a LDA of the reduced 7-characters data set and was placed by the PCA in a marginal position ( Fig. 53 ). It was found under the bark of a tree in 1.50 m height which supports the determination by the LDA.

Biology. In contrast to its sibling species C. obscurior , it was reported to nest near to or on the ground; it was found in hollow stems of dead Eulalia grasses (Okinawa), in a dead twig on the ground (New Orleans/ USA), between layers of Eugenia jambolana leaves ( India), in litter (Sulawesi), and ‘under leaves in a silk patch’ ( Tanzania). Nest populations are polygynous and adopt alien queens. There is polyphenism showing winged and wingless (ergatoid) males. The wingless males have sickle-shaped mandibles used to kill male callows or pupae whereas they besmear adult rivals with a secretion that elicits worker aggression.

Figures

Figure 23. Mesosoma in dorsal view.(A) Cardiocondyla obscurior, neotype, Nago / Okinawa, 3 June 2003; (B) Cardiocondyla shagrinata, paratype, South Konkan / India. Figure 24–27. Cardiocondyla shagrinata, paratype, South Konkan / India. (24) head in dorsal view; (25) lateral view; (26) dorsal view; (27) head surface between inner eye margin and paramedian vertex. Figure 28. Cardiocondyla wroughtonii, Lundu / Malaysia, 27 May 2007. (28) head in dorsal view.

Figure 53. Principal component analysis of 84 nest samples of Cardiocondyla wroughtonii and C. obscurior based on 7 morphometric characters taken in 205 worker individuals.The arrow points to the sample determined as C. obscurior by a wild-card LDA with p = 0.9456 which, however, showed a pigmentation typical for C. wroughtonii. The mean error of the four exploratory data analyses relative to the controlling LDA (final.sp.hyp) is 1.8%.

Figure 29–31. Cardiocondyla wroughtonii, Lundu / Malaysia, 27 May 2007.(29) lateral view; (30) dorsal view; (31) head surface between inner eye margin and paramedian vertex. Figure 32–34. Cardiocondyla obscurior, neotype, Nago / Okinawa, 3 June 2003. (32) head in dorsal view; (33) lateral view; (34) dorsal view.