Ectatomma species

Nettel-Hernanz, Alejandro, Lachaud, Jean-Paul, Fresneau, Dominique, López-Muñoz, Román A. & Poteaux, Chantal, 2015, Biogeography, cryptic diversity, and queen dimorphism evolution of the Neotropical ant genus Ectatomma Smith, 1958 (Formicidae, Ectatomminae), Organisms Diversity & Evolution (New York, N. Y.) 15 (3), pp. 543-553 : 548-550

Ectatomma species relationship and diversification

Until recently, Ectatomminae were considered as poneromorphs due to their ancestral ant morphology and behaviour ( Bolton 2003); today, they are nested within the Formicoid clade ( Moreau and Bell 2013; Ouellette et al. 2006). The taxonomy of Ectatomma has been revised several background shadings and dotted lines. Circled terminal clades represent the smallest monophyletic groups from our phylogenetic analysis that present small queens. Bold triangles indicate nodes that represent dispersal events to Central America according to our ancestral area reconstruction. Arrows represent the most recent common ancestor of Ectatomma times, mainly by Brown (1958) and Kugler and Brown (1982) who recognized 12 species of Ectatomma but synonymized some species. Three species have been described or redescribed after these revisions ( Almeida 1987; Arias-Penna 2006; Feitosa et al. 2008). However, the identity of several species remains uncertain. Our phylogenetic analysis unveils cases of unrecognized or cryptic diversity in Ectatomma with the exception of clade 4, where all species appear as relatively compact, monophyletic groups. The rest of the tree shows relatively deep lineage diversification within species going as far back as the early Miocene (the earliest of their 95 % highest posterior density divergence time estimates) in E. edentatum or 5.5 and 13 MYA, in E. ruidum and E. tuberculatum , respectively. This deep lineage diversification is consistent with the above species-level divergence as shown for other Hymenoptera ( Jansen et al. 2010; Lucky and Sarnat 2010; Lucky 2011; Blaimer 2012).

Within clade 1, E. edentatum presents paraphyletic lineages scattered within E. suzanae and E. muticum (clade 1, Fig. 1 ). Ectatomma edentatum was first considered as two geographically separated species ( Brown 1958): E. edentatum , restrict- ed to southern South America, and Ectatomma morgani , present in the Amazon and Orinoco basins. The species E. morgani was later synonymized under E. edentatum by Kugler and Brown (1982), with an extension of its range to include Panama and Costa Rica. E. suzanae , on the other hand, is a recently described species ( Almeida 1986). Most lineage splits within clade 1 represent relatively deep divergences; a new revision of all these taxa is required. Brown (1958), in an early examination of the genus, considered the morphology of E. edentatum as very similar to E. ruidum ; however, in our analyses, they prove to be unrelated.

In the case of the two most common species, E. ruidum and E. tuberculatum , our results show that both represent independent evolutionary lineages. E. ruidum presents two differentiated clades that colonized Central America independently. These two clades differentiated between 5.5 and 2.5 MYA.

E. permagnum

E. lugens C

L

A

D E. brunneum

E

4 E. opaciventre

C

E. ruidum

L

A

D

E

3

E. gibbum

E. vizottoi

C

L

A

D E. tuberculatum E

2

E. parasiticum

C E. edentatum L

A

D

E. suzanae E

E. muticum

E. edentatum 1

to biogeographic regional codes for each sample: (A) Parana, (B) Amazonian, (C) Caribbean–Mesoamerican, (D) Chacoan, and (E) Caribbean– north-west South American

The first clade included samples from a Mexican population of E. ruidum without microgynes together with samples from Venezuela, while samples from a Mexican population with microgynes grouped with samples from Ecuador and Panama. However, to date, no microgynes have been reported in E. ruidum from Ecuador or Panama ( Breed et al. 1990, 1999; Guénard and McGlynn 2013). E. ruidum microgynes have only been found in samples from Mexico ( Lachaud et al. 1999a, b). The two lineages of Mexican E. ruidum live in sympatry on a coastal plain in southern Mexico, with less than 15 km and no noticeable geographical barriers separating the sampled populations. Although some life traits are common in both E. ruidum lineages (e.g. feign of death when workers are disturbed, Cupul-Magaña 2009), in addition to the presence/ absence of microgynes, they are differentiated by other biological characters: date of sexuals production (Lachaud, unpublished data) and chemical compounds in their hydrocarbon profiles (Poteaux, unpublished data). At the beginning of the 20th century, another species was described in central–southwestern Mexico (Michoacan State) by Emery (1901): Ectatomma aztecum . It was revised by Kugler and Brown (1982) and is currently considered to be a junior synonym of E. ruidum . However, the morphology of both E. ruidum forms analyzed here is distinct from the described E. aztecum ’s morphology (F. Fernandez, personal communication). Therefore, the two E. ruidum lineages may represent two different species: one corresponding to the ‘true’ E. ruidum and the other to a new, undescribed species.

Furthermore, Kugler and Brown (1982) considered in their revision that Ectatomma confine is a very close relative of E. ruidum (only differing in the processes of the mesosoma) and could eventually represent an extreme variant of this species (locality series). We can thus hypothesize that the diversity in E. ruidum —as in other Ectatomma lineages—is underestimated over its distribution range and that a new taxonomic revision based on a holistic approach, combining different fields such as morphology, ecology, and molecular biology, is required.

E. gibbum was first described as a new species by Kugler and Brown (1982), with morphology very similar to E. ruidum . Our results show that E. gibbum represents a well-differentiated lineage that might share a common ancestor with the E. ruidum lineage. Our divergence time estimation and ancestral area reconstruction results place the origin of clade 3 ( E. ruidum + E. gibbum ) in Central America during the Middle or Late Miocene. The closure of the Central American Seaway and consequent formation of the Central American Isthmus occurred during the end of the Pliocene, implying that the ancestor of these species was able to migrate northwards either to an island or to the southernmost part of North America before the isthmus closed. This hypothesis would have to be confirmed with a finer evolutionary reconstruction that includes samples of E. gibbum from northern South America.

Within E. tuberculatum , most samples formed two clearly differentiated clades that diverged between 8 and 5 MYA. One of those clades includes samples from Brazil and French Guiana; the other—where E. parasiticum is nested—includes samples from Mexico, Ecuador, and French Guiana. E. tuberculatum is reported to be a facultative polygynous ant in Brazil with a functional polygyny and a mean relatedness of 0.3 among nestmate workers ( Hora et al. 2005; Zinck et al. 2007). However, in Mexico, E. tuberculatum is known to be monogynous in different places (e.g. only 1 % of 253 colonies collected in the south of the country were polygynous, including only two queens, and the parasite was never found, see Pérez-Lachaud et al. 2011) except in Apazapan (Veracruz) where E. parasiticum was discovered ( Fénéron et al. 2013). Furthermore, in this population, polygyny was functional but with oophagy events and unequal sharing of reproduction between queens ( Hora et al. 2007).

A single E. tuberculatum sample forms a clade along with the recently described E. vizottoi ( Almeida, 1987) both from the Parana Region in Brazil; the level of divergence within these two specimens is greater than the differentiation found within E. ruidum , and therefore, a thorough taxonomic revision is also recommended.

We acknowledge that our divergence estimates would be greatly improved with the addition of more nuclear markers that are inherited independently and have different coalescence times and historical migration patterns to our predominantly mitochondrial data set. Divergence time estimates from fast-evolving mitochondrial genes could result in overestimation of species divergence times because of their high substitution rate ( Dornburg et al. 2014; Zheng et al. 2011). Furthermore, mitochondrial genomes in some groups tend to introgress between related species at a high rate, which would yield inaccurate phylogenetic reconstructions and would impact species divergence time estimations under total evidence analysis ( Near and Keck 2013). As stated above, we propose a thorough revision of some specific relationships between Ectatomma species and within E. ruidum and E. tuberculatum , which should include a higher number of markers, as well as other biosystematic studies.

Figures

Fig. 1 Ectatomma ant species Bayesian phylogram obtained by Mr. Bayes analyses of two mitochondrial and one nuclear gene sequences. Node support is shown by two numbers, the first one corresponding to Bayesian posterior probabilities and the second one to bootstrap support