Camponotus candiotes

Nothing is known concerning the biology of this species (Seifert, 2019).

Identification
Seifert, 2019: The character combinations to identify this species can be derived from the key, Tab. 3 and the following pictures in AntWeb.org: CASENT0281578 (minor worker), CASENT0281579 (minor), CASENT0905389 (minor, syntype of C. candiotes).

Distribution
Seifert, 2019: Crete, Rhodes, Asia Minor, Georgia. Eastern parapatric sibling species of Camponotus piceus without safely confirmed range overlap. The single-specimen sample from Holomontas: Stagira, which is outside the known range of C. candiotes has a low posterior probability of p = 0.820 when run as wild-card in an LDA against C. piceus and may be misidentified. There is sympatric occurrence with Camponotus atricolor in Georgia.

Distribution based on Regional Taxon Lists
Palaearctic Region: Greece, Turkey.

Nomenclature

 *  candiotes. Camponotus lateralis var. candiotes Emery, 1894j: 10 (w.) GREECE. Combination in C. (Myrmentoma): Emery, 1925a: 68; Emery, 1925b: 120. Subspecies of piceus: Emery, 1925a: 68; Menozzi, 1936d: 304. Raised to species: Agosti & Collingwood, 1987a: 58.

Type Material
Seifert, 2019: Investigated were 4 syntype workers labeled ‘Creta (Cecconi) Omalo s Cata....’ [last word of label illegible], ‘SYNTYPUS Camponotus lateralis candiotes Emery, 1894’, ‘C. lateralis var. candiotes Eme’ and 3 syntype workers labeled ‘Creta (Cecconi) La C....’ [last word of label illegible], ‘SYNTYPUS Camponotus lateralis candiotes Emery, 1894’, ‘ANTWEB CASENT0905389’; all material.

Taxonomic Notes
Seifert, 2019: Both syntype series are clearly allocated to the cluster of 17 samples given in the next paragraph if run as wild-card in a 5-class LDA considering the five related species shown in Tab. 3 – the series with four syntypes is assigned with p = 0.9998 and that with three syntypes with p = 1.0000.

The exploratory data analyses NC-part.hclust, NC-part.kmeans, NC-NMDS-k-means and NCWard provided partially contradictory results regarding the heterospecificity of C. candiotes and C. piceus when CS and all 12 RAV-corrected shape and seta characters are considered. NC-Ward suggested C. candiotes to form a separate cluster with only one sample being misplaced (error 1.1 % in 87 samples). However, both NC-part. hclust and NC-part.kmeans did not confirm the presence of more than one cluster. Accepting the hypothesis formed by NC-Ward, a stepwise LDA was run which reduced the considered data set to the characters CL/CW1.25, SL/CS1.25, ScI1.25, MGr/CS1.25, nSc1.25 and PrL/CS1.25. Under this setting, NC-part.hclust fully confirmed the hypothesis formed by NC-Ward with three samples of C. piceus remaining unclassified – i.e., being placed as outliers. NC-NMDS-k-means clustering fully confirmed the hypothesis of NC-Ward but NC-part.kmeans, however, failed again to confirm the presence of two clusters (Fig. 13). With three exploratory data analyses confirming the final species hypothesis and one failing, I hypothesize C. candiotes to represent an eastern parapatric sibling species of C. piceus. The classification error in 180 individual workers is 1.1 % by an LDA and 1.7 % by a leave-one-out cross-validation LDA.