A unique cricetid experiment in the northern high-Andean Páramos deserves tribal recognition

Abstract

While hypsodonty mostly is associated with medium to large body sizes in sigmodontine rodents, high-crowned molars combined with small bodies rarely are recorded. This latter condition is present in Neomicroxus (Sigmodontinae, incertae sedis), a genus of high-Andean cricetids also characterized by a noticeable set of cranial traits, including enlarged turbinals and rostrum, slanting zygomatic plate, and a marked backward displacement of the vertical ramus of the dentary, linked with an enlargement of the basicranial region. These morphological features, combined with the isolated position of this lineage in molecular-based phylogenies, indicate that Neomicroxus should be situated in a new tribe. We name and describe this Páramo novelty monotypic clade here. As a working hypothesis, the hypsodonty displayed by this group is considered an evolutionary response to continued volcanic ash falls that characterized the region during the Neogene. A reappraisal of tribe recognition within the two cricetid largest subfamilies, arvicolines and sigmodontines, is made, coupled with a discussion about the role of morphological convergence in “long-nose” cricetids.

Mientras que la hipsodoncia normalmente es asociada con tamaños corporales medianos a grandes en roedores sigmodontinos, la combinación entre molares de coronas altas y formas con cuerpos pequeños es una condición rara en ese grupo. Esta condición está sin embargo presente en Neomicroxus (Sigmodontinae, incertae sedis), un género de cricétidos altoandinos también caracterizados por un notable conjunto de rasgos craneales, incluyendo rostro y turbinales alargados, placa cigomática baja y un marcado desplazamiento hacia atrás de la rama vertical del dentario, vinculado con un alargamiento de la región del basicráneo. Estas características morfológicas, combinadas con la posición aislada de este linaje en las filogenias moleculares, indican que el género Neomicroxus merece plaza en una nueva tribu. Nombramos y describimos aquí este novedoso clado monotípico del Páramo. Como hipótesis de trabajo, la hipsodoncia mostrada por este grupo es considerada una respuesta evolutiva a la continua lluvia de cenizas volcánicas que ha caracterizado la región durante el Neógeno. Se hace un resumen sobre el reconocimiento de tribus dentro de las dos mayores subfamilias de roedores cricétidos, los arvicolinos y sigmodontinos, a la par de una discusión sobre el papel de la convergencia morfológica en los cricétidos “hocicudos.”

cenizas volcánicas, Colombia, convergencia, Ecuador, Neomicroxus, Sigmodontinae, Venezuela

Colombia, convergence, Ecuador, Neomicroxus, Sigmodontinae, Venezuela, volcanic ash

Almost all tribes of sigmodontine rodents, the largest cricetid subfamily, have members that achieved hypsodonty (Hershkovitz 1962). What’s more, some of these suprageneric groups are entirely composed of hypsodont genera (e.g., Euneomyini, Andinomyini—Pardiñas et al. 2015; Salazar-Bravo et al. 2016). Although widespread across the sigmodontine radiation, hypsodonty neither is necessarily common nor in geographic terms equally distributed. According to available numbers, Oryzomyini, the largest tribe, composed of 40 genera (Pardiñas et al. 2017), shows an incidence of hypsodonty below 10%. However, a redundant combination is observed: hypsodonty is associated with larger body sizes; in fact, most of the sigmodontines that achieved high-crowned molars surpass a mass of 50 g and have herbivorous diets (Hershkovitz 1962; Madden 2015; Ronez et al. 2020). The opposite condition (i.e., hypsodonty in small-bodied sigmodontines) is rare, mostly restricted to few specialized granivores and carnivores (Hershkovitz 1962; Voss 1988).

NeomicroxusAlvarado-Serrano and D’Elía, 2013, was erected to distinguish small and inconspicuous cricetids that are patchily distributed in high-northern Andean ranges (Alvarado-Serrano and D’Elía 2013, 2015). This clade, early on envisioned based on morphology (see Patton et al. 1989; Voss 2003), was definitively removed from Akodon based on molecular evidence. Alvarado-Serrano and D’Elía (2013) published an initial study, using mitochondrial and nuclear gene sequences that firmly separated Neomicroxus latebricola (Anthony, 1924) from Akodon mimus (Thomas, 1901). Concomitantly, the genetic evidence retrieved Neomicroxus outside the Akodontini and any other recognized tribe (Alvarado-Serrano and D’Elía 2013).

Morphological approaches to Neomicroxus overlooked one of the most noticeable molar features displayed by these cricetids: hypsodonty. What's more, when the degree of hypsodonty is combined with small body size (< 20 g) and geographic isolation, Neomicroxus emerges as a unique experiment within the sigmodontine radiation. We hypothesize that recognition of a new tribe is supported by the integration of this evidence, extending that derived from molecular markers, and here adding data from a second species of the genus, Neomicroxus bogotensis (Thomas, 1895). The main purpose of the present contribution is to name and describe this clade and also provide an evolutionary framework for its establishment in the high-Andean Páramo.

Materials and Methods

Anatomy.

Specimens examined ( Appendix I) include young and adults composed by skulls, dry skins, and fluid-preserved bodies belonging to N. bogotensis and N. latebricola. Comparisons between Neomicroxus and members of the 11 recognized tribes of sigmodontines, as well as those genera incertae sedis, were undertaken using literature (Hershkovitz 1962; Voss 1988; Steppan 1995; Pacheco 2003; Weksler 2006), and a large list of specimens from collections detailed in Ronez et al. (2020:supplementary material table S2). We worked by direct comparison, examination of photographs, and CT-scan models obtained by one of the authors (JB). Three skulls of Neomicroxus, and several of other sigmodontines, were scanned with a micro-CT (Bruker Skyscan 1173; Zoologisches Forschungsmuseum Alexander Koenig, Bonn, Germany) using a 70-kV voltage source, with a resulting resolution of 29 μm; the raw data were reconstructed with NRecon (version 1.7.1.6, Bruker micro-CT), and 3D models built with the CTvox software, version 3.0.0 r1114, Bruker micro-CT (https://www.bruker.com). Anatomical terms and concepts used in this contribution were derived mainly from Hershkovitz (1962), Voss (1988), and Carleton and Musser (1989). Molar nomenclature follows Reig (1977). Hypsodonty definitions are from Hershkovitz (1962) and Koenigswald (2011). Stomach anatomy follows Carleton (1973) and Vorontsov (1967).

Sequence acquisition.

DNA sequences for the first portion of the mitochondrial protein-coding cytochrome b gene (Cytb) and the first exon of the interphotoreceptor retinoid-binding protein (IRBP) for sigmodontines and some outgroup taxa mostly were obtained from GenBank. For Neomicroxus we extracted the genomic DNA from five specimens (N. bogotensis = 3, N. latebricola = 2) and sequenced both genes. DNA extraction was carried out using the protocol of the Wizard Genomic DNA Purification kit, with fresh liver tissues as starting material. We used the MVZ05 and MVZ16 (Smith and Patton 1993; Cytb), and the A1 and F1 (Jansa and Voss 2000; IRBP) primers for the amplification and sequencing processes. Amplification conditions followed Da Silva and Patton (1993) for Cytb and Jansa and Voss (2000) for IRBP. All taxa analyzed as well as the vouchers that were the sources for Cytb and IRBP sequences are listed in Supplementary Data SD1.

Phylogeny.

We calculated the observed values on Cytb sequence divergence (p-distance) within and between tribes in Sigmodontinae with MEGA7 software (Kumar et al. 2016), ignoring those sites with missing data (Table 1). Phylogenetic analyses were carried out on the concatenated matrix using maximum likelihood (ML—Felsenstein 1981), and Bayesian inference (BI—Huelsenbeck et al. 2001) approaches. In the cases that miss information for a locus, we completed the matrix with missing data or ambiguous state characters (i.e., N). The partition scheme and substitution models were identified previously by PartitionFinder2 (Lanfear et al. 2017), being SYM+I+G for the first and second cytb codon position, and first, second, and third IRBP codon position, and GTR+I+G for third cytb codon position. ML analysis was carried out on the IQ-TREE version 1.6.0 software (Nguyen et al. 2015) implemented in the IQ-TREE webserver (http://iqtree.cibiv.univie.ac.at/—Trifinopoulos et al. 2016). Support for each individual node of the ML tree was estimated using 1,000 iterations of the ultrafast bootstrap value (BT). BI analysis was undertaken using MrBayes 3.2 (Ronquist et al. 2012). Two independent runs, each with three heated and one cold Markov chains, were allowed to proceed for 10⁷ generations and were sampled every 1,000 generations. Log-likelihood values against generation time for each run were plotted in Tracer v1.7.1 (Rambaut et al. 2018). The first 25% of trees were discarded as “burn-in” and the remaining trees were used to compute a 50% majority rule consensus tree and obtain posterior probability (PP) estimates for each clade. Nodes with bootstrap ≥ 80 and posterior probability ≥ 0.8 were considered well-supported.

Table 1.

Observed values of Cytb sequence divergence (p-distance) within and between recognized tribes in Sigmodontinae; nc (= not comparable) highlights those tribes where the mean divergence within the tribe cannot be calculated because are composed of a single extant genus.

		Between tribes
	Within tribes	1	2	3	4	5	6	7	8	9	10	11
(1) Abrotrichini	13.05
(2) Oryzomyini	14.67	17.46
(3) Phyllotini	16.31	17.26	17.58
(4) Andinomyini	12.36	16.54	16.97	17.28
(5) Thomasomyini	15.33	17.82	17.14	18.15	16.35
(6) Euneomyini	18.19	18.91	19.21	20.14	19.16	18.53
(7) Neomicroxini	6.35	15.60	16.63	16.42	15.60	15.91	17.84
(8) Ichthyomyini	14.36	19.13	18.34	19.42	19.79	18.67	21.14	18.69
(9) Wiedomyini	15.45	16.52	17.72	18.01	18.30	18.39	19.99	16.72	19.30
(10) Reithrodontini	nc	20.82	20.96	21.95	20.66	20.76	20.89	20.37	23.28	22.22
(11) Sigmodontini	nc	19.29	18.55	19.16	16.54	17.91	20.22	16.37	20.72	20.27	21.60
(12) Akodontini	15.07	17.59	17.47	18.41	16.37	16.81	18.82	16.37	19.51	17.76	21.12	18.51

Results and Discussion

Selected anatomical traits.

Voss (2003:21) proposed the first clarified set of craniodental traits amalgamating Microxus latebricola with bogotensis:

among other shared similarities, both species differ from typical Akodon by their very small size; possession of a slender, tapering rostrum flanked by very shallow zygomatic notches (versus a shorter, stouter rostrum flanked by deeper zygomatic notches); origin of the superficial masseter from an indistinct tubercle or scar on the anterior margin of the zygomatic plate (versus from a scar posteroventral to the anterior edge of the zygomatic plate); confluence of the buccinator–masticatory foramen and foramen ovale (versus buccinator masticatory foramen and foramen ovale accessorius separated by a vertical strut of the alisphenoid); proportionately shorter incisive foramina, wider parapterygoid fossae, and more inflated bullae; and highly distinctive molars with opposite (versus alternating) cusps.

Because this constitutes the morphological foundation of Neomicroxus (see Alvarado-Serrano and D’Elía 2013) and, by monotypy, of the tribe herein proposed, the listed traits deserve closer inspection.

Voss (2003) emphasized the small size of the members of Neomicroxus. Because adult masses in N. latebricola (slightly larger than N. bogotensis) average about 16 g (Curay 2019), they are close to the lower limit of body mass recorded in sigmodontines, determined by Salinomys delicatus (9–14 g; see Pardiñas et al. 2017). In the Páramo ranges inhabited by Neomicroxus, a few other comparable small-bodied cricetids occur, including Microryzomys altissimus (Osgood, 1933), Reithrodontomys soederstroemiThomas, 1898, and Thomasomys ucuchaVoss, 2003, all under 20 g (Tirira 2017).

A cranial feature that early on cemented the concept of a polyphyletic Microxus was the possession of a pointed, tapering rostrum, externally expressed as a long muzzle associated with small eyes. Typically, stuffed skins fail to retain these traits. In fact, Thomas (1920:240–242) was adamant when he obtained the first fluid-preserved animals noting that “A study of these [spirit specimens] shows very strongly the essential difference between the two animals under discussion [i.e., Microxus torques and Akodon mollis]. The long head, especially the long muzzle, and the small eyes, give the Microxus quite a different aspect to that of the Akodon with its blunt snout and normal eyes, and I now feel no hesitation in considering them as belonging to different genera.” Taken as a whole, the long rostrum in Neomicroxus (Fig. 1A) is formed by a combination of anteriorly projected nasal bones coupled with the anterior extension of the premaxillary bones, including a medium-sized gnathic process, to produce a moderately expressed “nasal-tube” or “trumpet” (Hershkovitz 1994). In frontal view, the rostrum opening is broad, and two pairs of turbinals largely are exposed (Fig. 1B). However, viewed ventrally, the diastemal region is not completely occupied by the incisive foramina, leaving a noticeable anterior portion with a tiny Hill foramen, suggesting that the enlarging of the rostrum mostly operated affecting this distal region (Fig. 1C).

Fig. 1.

Selected anatomical regions in the cranium of Neomicroxus: (A) rostrum in lateral right view (N. latebricola; MEPN 12715); (B) rostrum anterior opening (N. latebricola; MEPN 12715); (C) diastemal palate (N. latebricola; MECN 9803); (D) right zygomatic plate (N. latebricola; MEPN 12715).

Additional traits of the rostrum are the infraorbital foramen, the zygomatic plate, the zygomatic notch, and the nasolacrimal capsule. These elements usually are examined separately, but this operational procedure partially clouds the fact that they are intimately associated to produce a definite bauplan (Voss 1988). In Neomicroxus, when the zygomatic notch is described as shallow or directly non-existent, it is because the zygomatic plate almost lacks an upper free border (i.e., the posterior limit of the zygomatic notch), and the nasolacrimal capsule also is poorly developed (i.e., the anterior limit of the zygomatic notch). However, in dorsal view, in both species of Neomicroxus, the nasolacrimal capsule is expressed as a hook; in addition, because the zygomatic plate is laterally everted, its internal side remains dorsally exposed, contributing to define a space known as the zygomatic notch (Fig. 2A). The extremely low (more so in bogotensis than in latebricola) zygomatic plate of Neomicroxus (called the “microxine type” by Thomas 1920:240) is attached dorsally to the vault by a strong, descending, and broad maxillary zygomatic root (Figs. 1D and 2B). The shaft of the zygomatic plate is everted markedly from the sagittal plane of the cranium, leading to an ovoid infraorbital foramen (Fig. 2B). Basally, the zygomatic plate makes a twist and, at this point, a real masseteric tubercle is expressed, not just the typical scar showed by most other sigmodontines. This tubercle looks like a “bony pinch” (more marked in bogotensis than in latebricola) located where the zygomatic plate bends to run to the cranium (Fig. 1D).

Fig. 2.

Selected anatomical regions in the cranium of Neomicroxus: (A) zygomatic notch complex in dorsal right view (N. latebricola; MEPN 9803); (B) left infraorbital foramen in frontal view (N. latebricola; MECN 6136); (C) cross section of the cranium at the frontal sinuses plane (N. latebricola; MEPN 12715); (D) cross section of the cranium at the rostral anterior level (N. latebricola; MEPN 12715). Acronyms: etI = ethmoturbinal I; etII = ethmoturbinal II; etIII = ethmoturbinal III; ft1 = frontoturbinal 1; ft2 = frontoturbinal 2; it = interturbinal; ls = lamina semicircularis; mt = maxilloturbinal; nt = nasoturbinal.

A broad interorbital region is another feature noticeable in Neomicroxus. When the interior of the cranium is examined across the plane defined by the lacrimals, frontal sinuses are obvious, mostly filled by bony trabeculae belonging to the turbinal complex (Fig. 2C). Also, the entire braincase is enlarged and rounded in Neomicroxus, and the temporal region is extended strongly backwards (Fig. 3A). In lateral view, the posterior protrusion of the mastoid capsule and the occipital is noteworthy, associated with a long hamular process of the squamosal that broadens distally. Two subequally sized foramina are produced by this structure. In addition, a globular auditory bulla is present, having a large auditory meatus that exposes the entire malleus with a well-developed orbicular apophysis. The otic capsule is attached to the cranium by the petrosal, solidly overlapping the posterior suspensory process of the squamosal (Fig. 3A). The squamosal–alisphenoid groove (sag) almost is imperceptible on the external surface of the bone, but clearly marked inside the braincase (Figs. 3B and 3C). When the sag crosses the broad trough for the masticatory–buccinator nerve, a fontanelle, equal in size to the sphenofrontal foramen, is created. A large foramen ovale is not interrupted by an alisphenoid strut (Fig. 3C).

Fig. 3.

Selected anatomical regions in the cranium of Neomicroxus: (A) basicranial and temporal regions in internal right view (N. latebricola; MEPN 12715); (B) left temporal region in lateral view (N. latebricola; MEPN 12715); (C) detail of the left temporal region showing foramina associated to carotid circulation (N. latebricola; MEPN 12715). Acronyms: aalc = anterior opening of alisphenoid canal; ect = ectotympanic; fo = foramen ovale; hp = hamular process of squamosal; mbt = trough for masticatory–buccinator nerve; palc = posterior opening of alisphenoid canal; pet = petrosal; pgf = postglenoid foramen; sag = squamosal–alisphenoid groove; sfr = sphenofrontal foramen; ssf = subsquamosal fenestra; stc = stapedial artery canal; stg = groove for stapedial artery.

To accommodate these osteological features in the temporal region, the ascending portion of the dentary is notably displaced backwards (Fig. 4A). In fact, the geometry of the jaw is completely reworked, exposing the entire toothrow when viewed from the side. Conversely, in most sigmodontines the ascending ramus hides the m3. Both the condyle and the coronoid process of the dentary are well-developed structures extended posteriorly while the angular process remains shortened. The ascending ramus also is sharply angled with respect to the main axis of the jaw (Fig. 4B) and, probably as an effect associated with the enlargement of the entire region, the capsular process of the incisor is absent.

Fig. 4.

Morphology of the dentary in Neomicroxus (based on the right hemimandible of N. latebricola; MEPN 12715): (A) labial view; (B) anterolateral view; (C) lingual view.

Finally, Voss (2003:21) highlighted that Neomicroxus has “… highly distinctive molars with opposite (versus alternating) cusps.” The noticeable feature of these molars is their hypsodonty, typified here as tubercular hypsodonty because the cusps tend to produce a bowed profile (Fig. 5A). Hypsodonty is more profound in N. latebricola than in N. bogotensis, leading to a simplification of the occlusal structure in the former species, including the loss of anterolophs, mesolophs, and several additional secondary crests.

Fig. 5.

Molars and teeth roots in Neomicroxus: (A) left upper toothrow in lingual view (N. latebricola; MEPN 9803); (B) right lower toothrow in lingual view (N. latebricola; MEPN 9803); (C) roots of left upper molars (N. latebricola; MEPN 12715); (D) roots of right lower molars (N. latebricola; MEPN 12715).

Neomicroxus as a unique experiment.

The process to collate several pieces of evidence from a number of species (e.g., Akodon bogotensis, Oxymycterus lanosus, A. mimus, Oxymycterus iheringi, the nominal components of the genus Microxus as envisioned by Thomas 1909) to a single morphospace within the Sigmodontinae radiation has been difficult, as successive contributions by Thomas (1909, 1916, 1920, 1927) denote. Clearly, one of the most important obstacles was the imperfect knowledge about the anatomy of the various species. Some craniodental traits initially were selected as indicators of the generic distinction of Akodon and Microxus. One of these features was the morphology of the zygomatic plate. As early as the end of the 19th century, Thomas (1895:369) highlighted an “outer wall of anteorbital foramina unusually short” in A. bogotensis; and also remarked that “in the unusually slender lower portion of its anterior zygoma-root it [A. bogotensis] also seems to differ from all its allies” (Thomas, 1895:370). He did not abandon this feature in his deconstruction of Oxymycterus, when Microxus was erected and Thomas (1909:237) noted “skull like that of a small Akodon, except for the characteristic narrow zygomatic plate.” Thomas (1920:240) returned to this feature when he described in full Microxus torques, and remarked that “this species is distinguishable from M. mimus by its broader and less characteristically Microxine zygomatic plate…” and concluded that “… I consider the genus Microxus, as based on mimus, bogotensis, and lanosus, as possibly valid, the characters of the very narrow bar-like zygomatic plate much more marked and Oxymycterus-like than in A. orophilus- and the whole skull smooth, papery, with large brain-case and slender upturned muzzle” (Thomas 1927:370–371).

It may appear contradictory to state that Neomicroxus constitutes a unique experiment because the recognized species in this genus have been associated with other several cricetids as the “oxymycterine” group (sensu Hershkovitz 1966:86–87). This clade, envisioned by Thomas (1909), is anchored strongly in a suite of shared morphological characters, particularly craniodental. Among them, the pointed and enlarged rostrum, the well-developed turbinals and inflated frontal sinuses, the slanting zygomatic plate, the enlarged and rounded braincase, the low and slender dentary, and the tendency to hypsodonty constitute a common signature. It is the so-called “long-nose” or “hocicudo” bauplan (see Hershkovitz 1994), examined in detail by Hinojosa et al. (1987:5–9). However, this illuminating study was partially overlooked in the rapid dynamics that characterized the decade of the 1990s, with many discoveries based on molecular markers (Patton et al. 1989, 1990). A large body of evidence now refutes the monophyly of the “oxymycterine” group, challenging the established vision of an Akodontini including both Akodon and Abrothrix (Spotorno 1986; Apfelbaum and Reig 1989; Barrantes et al. 1993). A few years later, the emergence of a new tribe (i.e., Abrotrichini) to contain the latter genus plus several associated taxa (e.g., Chelemys, Geoxus) was the “knockout” for the “oxymycterine” group (Smith and Patton 1999; D’Elía et al. 2007). However, a crucial question remains to be addressed: why do two tribes (or more, see below), Akodontini and Abrotrichini, share the “hocicudo” bauplan?

We argue here that the “hocicudo” bauplan also is present in Neomicroxus, adding a third presumptive tribe that can be distinguished by this unique set of morphological characters (Fig. 6). In addition, we propose that this bauplan almost is unrepresented in other sigmodontine lineages, such as the Andinomyini, Euneomyini, Ichthyomyini, Oryzomyini, Phyllotini, Reithrodontini, Sigmodontini, Thomasomyini, Wiedomyini, and several Sigmodontinae incertae sedis (see Pardiñas et al. 2017). Probably, a few exceptions should be noted; for instance, Aepeomys (Thomasomyini—Ochoa et al. 2001), Delomys (incertae sedis—Voss 1993), Neusticomys (Ichthyomyini—Voss 1988), although all of these genera depart from the typical bauplan in the morphology of their zygomatic plates. But the concentration of the “hocicudo” bauplan in Akodontini, Abrotrichini, and Neomicroxus invites a search for a common causation. We return to this issue in the concluding section of this contribution.

Fig. 6.

Selected anatomical traits characterizing the “long-nose” (or “hocicudo”) bauplan in three tribes of sigmodontines, including rostrum in left lateral view (left column; arrows highlight the position and development of the upper root of the zygomatic plate) and left hemimandible in labial view (right column; arrows point to the visible space behind the m3): (A, D) Neomicroxus latebricola (MEPN 12715; Neomicroxini); (B, E) Abrothrix longipilis (Zoologisches Forschungsmuseum Alexander Koenig, Germany, ZFMK 1992-0387, Isla Mocha, Chile; Abrotrichini); (C, F) Oxymycterus hiska (ZFMK 1992-0403, Yungas del Pojo, Bolivia; Akodontini).

The molecular evidence.

The overall mean divergence at the Cytb gene for Sigmodontinae is close to 17%. Values of genetic distances within each tribe or isolated genus fluctuate between 6.3% for Neomicroxus and 18.1% for Euneomyini. The tribal comparisons reveal highest divergences between Neomicroxus and Ichthyomyini (18.6%) and Reithrodontini (20.3%); these values vary between 15.6% and 16.7% (Table 1) when Neomicroxus is compared with remaining tribes. The phylogenetic analyses recovered well-supported relationships within Sigmodontinae, mostly concordant with previous studies (e.g., Parada et al. 2013; Salazar-Bravo et al. 2016; Gonçalves et al. 2020). Sigmodontinae was monophyletic (BT/PP: 100/1.0) with two major clades: Sigmodontalia (sensu Leite et al. 2014; 99/0.9) composed by Ichthyomyini and Sigmodontini; and Oryzomyalia (sensu Steppan et al. 2004; 100/1.0) composed by the remaining taxa. Most of the relationships among the tribes currently recognized in the Sigmodontinae are resolved with low or moderate strong support in the ML phylogeny, while in the BI topology all Oryzomyalia tribal relationships are unresolvedly displayed as a polytomy (Fig. 7).

Fig. 7.

Phylogenetic trees based on the combined matrix of Cytb and IRBP loci for Sigmodontinae with emphasis in the position of Neomicroxini, new tribe, obtained with (A) maximum likelihood (ML), and (B) Bayesian inference (BI). Numbers above each node indicate bootstrap and posterior probabilities support values obtained in the ML and BI analyses; nodes with bootstrap ≥ 80 and posterior probability ≥ 0.9 are considered low support and are not shown.

Neomicroxus was found to be monophyletic with high support values (99/1.0), sister to the Abrotrichini + Wiedomyini in the ML analysis (BT: 34), but without affiliation to any recognized tribe in the BI topology (Fig. 7). The phylogenetic position of Neomicroxus in previous studies has been variable, with no affiliation with any tribe (i.e., Alvarado-Serrano and D’Elía 2013; Gonçalves et al. 2020) or as the sister to Juliomys with very low support (Salazar-Bravo et al. 2016).

Two strongly supported major clades were retrieved within Neomicroxus. One clade grouping all individuals from Colombia can be referred to what currently is understood as Necromys bogotensis; another clade comprising all samples from the Ecuadorian Cordillera Oriental and Occidental can be confidently associated to N. latebricola. The overall mean divergence at the Cytb gene for Neomicroxus exceeds 6%; meanwhile, the genetic distance between the two major clades reaches 11%. The P-values within each species are contrasting. The divergence within the N. bogotensis samples ranges from 0.1% to 6.5%, which suggests greater specific diversity in Colombia. Within the N. latebricola clade, the genetic divergence values vary between 0.3% and 1.6%, which added to the morphological evidence, support the existence of a new subspecies for Ecuador (Cañón et al. 2020).

A new tribe for Neomicroxus.

The sum of morphological traits exhibited by Neomicroxus, in addition to its isolated phylogenetic position as revealed by molecular markers, supports its recognition in its own tribe. Therefore, a new tribal group to contain the genus is established here, as follows:

Family Cricetidae

Subfamily Sigmodontinae

Neomicroxini, new tribe

(Figs. 1–5 and 8)

Fig. 8.

Selected external and soft anatomy traits in Neomicroxus: (A, B) dorsal and plantar view of right forefoot (N. latebricola; MECN 3727); (C, D) dorsal and plantar view of right hindfoot (N. latebricola; MECN 3727); (E) soft palate showing diastemal (d) and interdental (i) rugae (N. latebricola; MECN 5605); (F) gross internal morphology of the dorsal half of the stomach (N. latebricola; MECN 3789). Acronyms: b = bordering fold; co = cornified epithelium; d = duodenum; e = esophagus; ge = glandular epithelium; i = incisura angularis.

Type genus, by present designation.

NeomicroxusAlvarado-Serrano and D’Elía, 2013.

Diagnosis.

A tribe of the subfamily Sigmodontinae, “clade Oryzomyalia” (sensu Steppan et al. 2004) grouping small-sized cricetids (head and body length ~85 mm; body mass ~17 g) characterized by the following combination of morphological traits: small head and minute eyes, ears small (~14 mm), rounded, semihidden by head fur; dorsal fur long (~10 mm), dense, soft but not woolly, from grizzled to dark chestnut or rufous, blackish brown, belly slightly paler, scarcely countershaded; tail long (~80 mm) slightly shorter than head and body length, covered by short hairs; manus and pes with moderately pointed claws at the ends of the toes; palmar pad composed of five pads with digits stocky and subequal in length; plantar pad composed of six pads with a large hypothenar pad, squamate surface, and with the ungual tufts surpassing the end of the claws; three pairs of mammary glands arranged in pectoral, abdominal, and inguinal, pairs; skull delicate, with domed profile, rostrum pointed with short “nasal-tube” and globular noteworthy braincase; conspicuous nasolacrimal capsules and foramina; shallow zygomatic notches; interorbital region broad and smooth; zygomatic plate low, poorly developed, almost without free upper border, but with a noticeable masseteric tubercle; infraorbital foramen ovoid, not-constricted basally; incisive foramina broad, reaching posteriorly the first upper molar protocones; palate short (N. latebricola) or long (N. bogotensis) and wide; toothrows parallel; broad mesopterygoid fossa and parapterygoid plates; temporal region enlarged with long and thin hamular process of the squamosal applied to a robust mastoid promontorium; tegmen tympani overlapping suspensory process of squamosal; alisphenoid strut absent; carotid circulatory pattern type 1 (stapedial foramen, sphenofrontal foramen, and squamosal–alisphenoid groove, all present); auditory capsules inflated with very short Eustachian tubes; large mastoid capsules; large and rounded foramen magnum directed ventrally; dentary slender and low, with the ascending ramus strongly displaced backwards leaving all the molars laterally well-exposed; capsular process of the incisor root absent; short and pointed angular process; orthodont (in N. latebricola) to opisthodont (in N. bogotensis) gracile and ungrooved upper incisors; molars hypsodont, with crested coronal surfaces, main cusps opposite, and noticeable basal cingulum closing the main flexus/ids; anterocone of M1 with two well-developed and subequal size conulids; anterolophs and mesolophs barely present (in N. bogotensis) or fused (in N. latebricola); M3 small but bilobed (in N. bogotensis) or tending to peg-like form (in N. latebricola); mesolophids typically reduced or absent; m3 anteriorly–posteriorly compressed; upper molars three-rooted; lower molars two-rooted; axial skeleton typically composed of 13 thoracic, 7 lumbar, 4 sacral, and 23–25 caudal vertebral elements; axis with squared spinous process; second thoracic with markedly enlarged dorsal process; stomach unilocular–hemiglandular, with both types of epithelium subequally distributed (in N. latebricola); gall bladder present (in N. bogotensis) or absent (in N. latebricola); soft palate with three complete diastemal and five incomplete interdental rugae (in N. latebricola); two pairs of basal prostate glands (in N. bogotensis) (after Thomas 1895; Anthony 1924; Gyldenstolpe 1932; Ellerman 1941; Voss and Linzey 1981; Reig 1987; Voss 1991, 2003; Alvarado-Serrano and D’Elía 2013; Curay 2019; this paper).

Content.

A single genus, NeomicroxusAlvarado-Serrano and D’Elía, 2013:1008.

Geographic distribution.

Neomicroxini rodents are distributed along high-Andean (typically above 3,000 m a.s.l.) Polylepis forests and shrubland–grassland Páramo environments from northern Ecuador, through Colombia to southwestern Venezuela (Linares 1998; Alvarado-Serrano and D’Elía 2015; Fig. 9).

Fig. 9.

Neomicroxini: current schematic geographic range in South America.

Biochron.

Recent in Ecuador, Colombia, and Venezuela.

Remarks.

Because Neomicroxini contains the single genus Neomicroxus, the diagnosis constructed for it (Alvarado-Serrano and D’Elía 2013:1008) should be sufficient to diagnosis the tribe, by monotypy. However, we prefer to provide a more comprehensive diagnosis to also incorporate soft anatomy and postcranial skeleton and to add a more accurate description of molars. Microxus is not a classic Latin or Greek word; etymologically, it probably highlights the small (micr-) and pointed (oxys) rostrum of mimus, their genotype. Because Neomicroxus is a neologism composed by the prefix “Neo” (new) and the generic epithet Microxus, the name of the tribe is derived from the addition of the tribal ending -ini, hence, Neomicroxini.

Hypsodonty and volcanic ash fall in the northern Andes.

Sigmodontines display varied morphological contrivances to increase the functional longevity of their molars. According to the traditional view of the subfamily’s evolution, low-crowned (= brachydont) founding population(s) radiated into a remarkable diversity of medium- (“mesodont”) to high-crowned (hypsodont) taxa to face the abrasiveness of South American environments, particularly during the transition from forests to grasslands (Hershkovitz 1962). However, the topic remains problematic, despite hypsodonty perhaps being the key innovation droving diversification. Madden’s (2015) integrative approach to understanding hypsodonty and its causes over the entire sigmodontine radiation demonstrated a significant increase in the proportion of hypsodont taxa among folivores and herbivores in faunas occurring on andisols (volcanic ash soils), particularly in montane environments. This fact presumably reflects the increased mobility of abrasive minerals (in particular, tephras) as contaminants in mammalian foodstuffs. Madden (2015) explored the possible contribution of ash-rich vulcanism and mountain uplift in the Cenozoic by analyzing the association among hypsodont sigmodontines, active volcanoes, duststorms, and glaciers, as potential sources. He concluded (Madden 2015:79–80) that “hypsodonty is a complex phenomenon involving correlations with herbivorous diets, large-scale climate variation in temperature and rainfall, and both the sources of sediment and the surface processes that mobilize environmental abrasives through the animal’s environment.” He found a clear association between a core of sigmodontine hypsodonty and the “Northern Volcanic Zone–NVZ in southern Colombia and Ecuador (between 7ºN and 0.5ºS latitude)” (Madden 2015:figure 3.8).

The Northern Volcanic Zone concentrates several of the most important examples of active and extinct volcanoes in South America (Hall 1977). According to Hall and Calle (1982:235)

the most evident indication of magmatic activity during the Pleistocene and Holocene are the numerous large stratovolcanoes, both extinct and active, that cap the northern Andes. These volcanoes form two principal parallel rows that extend from the Colombian border southward to about latitude 2°30’S, south of which they are absent... The western row, that forms the Western Cordillera, includes the volcanoes Chimborazo, Carihuairazu, Quilotoa, Iliniza, Corazón, Atacazo, Pichincha, Pululahua, Cotacachi, Cuicocha, Yana–urcu, and Chiles. The eastern row, that crests the Cordillera Real, includes the volcanoes Sangay, Altar, Tungurahua, Cotopaxi, Sincholagua, Antisana, Las Puntas, and Cayambe. Between the two volcanic rows, the Interandean Valley contains somewhat more eroded volcanoes...

Plio-Pleistocene ash fall deposits cover much of the Andes, including the Interandean Valley, and have been largely recognized in classic geological formations as the Cangahua Formation, a thick sequence of ash material (Hall and Calle 1982; Mothes et al. 1998). An explosive period of widespread vulcanism, beginning in Late Pliocene and increased during the Plio-Pleistocene interval, accompanied the rise of both cordilleras in the Ecuadorian Andes (Hungerbühler et al. 2002). It is tempting to connect hypsodonty in Neomicroxus with the continuous ash fall in northern Andes during the Neogene. The emergence of a small-bodied, markedly high-crowned lineage of sigmodontines as an evolutionary response to an increasing tephra input in the context of Andean orogeny is rather speculative. However, additional facts provide some support for this hypothesis. For example, hypsodonty is less developed in N. bogotensis than N. latebricola, and this species occurs in areas where Neogene ash fall has been more limited (Soriano et al. 1999). The diet of N. latebricola is unknown, but unsystematic observations suggest that the species is a surface-dweller, feeding on plant material, roots, and few invertebrates, and, therefore, ingesting abrasive soil particles (Brito 2013). Finally, it is striking to note that some tendency to high-crowned molars is present in other taxa that share the northern Ecuador high-Andean ranges with Neomicroxus, in particular, several species of Thomasomys. Although these sigmodontines are mainly treated as brachydont (Pacheco 2003), a moderate degree of hypsodonty is found in several other Ecuadorian species. When diagnosing T. ucucha, a high-Andean (crest of the Cordillera Oriental between ca. 3,400 and 3,700 m) species, Voss (2003:10) noted “small, hypsodont molars lacking well-developed cingula and stylar cusps” and then “hypsodont [molars] when unworn (by comparison with more brachydont congeners)” (Voss 2003:11). In resurrecting Thomasomys cinnameusAnthony, 1924, another high-Andean Ecuadorian form and one of the smallest known species of the genus (known masses range from 14 to 19 g—Voss 2003:table 5), Voss (2003:29) indicated “larger and more hypsodont molars with weakly developed cingula and stylar cusps [than in Thomasomys gracilis].” The group of species associated with Thomasomys aureus (Tomes, 1860), widespread along the Andes from Bolivia to Venezuela (Pacheco 2015; Brito et al. 2019) also is characterized by having high-crowned molars (see Supplementary Data SD2). Evidently, hypsodonty deserves further attention as a plausible evolutionary response in northern Andean sigmodontines.

Remarks

Taxonomic inflation in cricetids: numbers and categories.

Does it make any sense to name and describe monotypic tribes or does doing so simply produce an unnecessary increment in taxonomy (i.e., “taxonomic inflation”)? We are persuaded that it is a crucial task of current taxonomic practice to recognize and, therefore, to name clades, even those composed of a single genus. Monotypy could be just a temporal accident, because nothing contradicts the concept that a tribe today composed of a single genus would have flourished in the past (or would radiate in the future). In any case, it seems a more productive exercise to examine cricetid diversity to discuss the issue. We do this by comparing two large cricetid subfamilies, living Arvicolinae and Sigmodontinae (Table 2). Although this kind of analysis is not free from bias, because the arvicoline fossil record is much richer than that of sigmodontines, these subfamilies display 10 and 11 tribes, with 29 and 87 genera, respectively (Pardiñas et al. 2017). The generic ratio of 1:3 (29 versus 87 genera) suggests that tribal diversity (ca. 1:1) is underestimated in sigmodontines or overestimated in arvicolines. Conversely, the species:genus ratio expresses a similar value in both subfamilies (Table 2).

Table 2.

Diversity in taxonomic categories for the living cricetid subfamilies, arranged from top to bottom by decreasing number of tribes (counts based on data compiled by Pardiñas et al. 2017).

	Tribes	Genera	Genus/tribe	Species	Species/genus
Sigmodontinae	11	87	8	443	5
Arvicolinae	10	29	3	162	6
Neotominae	4	16	4	140	9
Tylomyinae	2	4	2	11	3
Cricetinae	1	7	7	18	3

However, there is no easy way to decide between these alternatives. According to Anderson’s (1975:1) simulations, “the number of categories required is much nearer the minimum possible number than the maximum possible; usually 11 to 16 categories will be needed for a group of 100…” This probably favors the tribal arrangement of the Arvicolinae and, consequently, suggests that tribal classification in Sigmodontinae is underestimated. Adding Neomicroxini increases sigmodontine tribal diversity to 12, and we suspect that several other genera need to be accommodated in still unnamed tribes, including Abrawayaomys, Delomys, Juliomys, Scolomys, and Zygodontomys, but not Chinchillula, given the latter could be integrated into the Euneomyini (or, more extreme, Andinomyini and Euneomyini could be united into a single tribe). Each of these genera is a suitable candidate to create monotypic tribes, because each one represents a unique experiment within the sigmodontine radiation. If 17 tribes are recognized in the subfamily, a striking effect results on the calculations in Table 2, turning the genus:tribe ratio more similar between Arvicolinae and Sigmodontinae. Perhaps it is time to face what Gippoliti and Groves (2013:10) summarized, “… the current age of so-called taxonomic inflation may be the historic answer to a long period of taxonomic deflation, and of undervaluing of the real extent of biodiversity.”

The value of convergence.

If the “hocicudo” bauplan appears in three tribes and these tribes are not recognized as phylogenetically close, convergence is the most parsimonious explanation. However, to firmly establish that, we need a resolved phylogeny, which is not available for the sigmodontines, where most tribes emerge from a polytomy or, alternatively, the few retrieved basal relationships have poor statistical support (Parada et al. 2013; Salazar-Bravo et al. 2016; Steppan and Schenk 2017; Gonçalves et al. 2020; this paper).

Rostral specializations, and their mirrored osteological counterparts, are related to three main components of life in small mammals: olfaction, heat and moisture conservation, and feeding (Voss 1988; Martinez et al. 2018, 2020). Martinez et al. (2018) highlighted an important degree of convergence among murine “shrew-rats” that colonized the Indo-Australian Archipelago, suggesting that the process mainly was driven by dietary constraints in altitudinal gradients. According to their study, “compared to carnivores and omnivores, vermivores [vermivorous murines] should have significantly better olfactory capacities, based on both the larger surface area and higher complexity of their olfactory turbinals… these bony specialisations are related to an improvement of their olfactory adaptations allowing them to detect prey that are underground or invisible within wet leaf litter… Such prey may be especially elusive and difficult to detect for more generalist, opportunistic rats.” The variety of rostral morphologies (extreme facial elongation, reduction of muscle attachment points) displayed by these murines (e.g., Echiothrix, Melasmothrix, Rhynchomys, Sommeromys; see Musser and Durden 2002; Balete et al. 2007) certainly is unique and partially unparalleled among living sigmodontines. Interesting to note, extreme rostral elongations do not always result from equally comparable osteological features, as can be easily appreciated from the contrast between Sommeromys and murid “shrew-rats” (see discussion in Musser and Durden 2002:15).

Is the “hocicudo” bauplan one of the morphological responses to a common problem faced by cursorial, forest-dwelling, sigmodontines: the searching for food in deep leaf litters under humid-saturated environmental conditions? For the second time in the present contribution, we are tempted to cement incompletely known pieces into a working hypothesis. Rostral specializations have been highlighted in sigmodontine rodents for more than a century, leading to the construction of polyphyletic groups (Hinojosa et al. 1987). In addition, a complex interplay and possibly trade-offs among the several principal skull components involved (e.g., zygomatic plate, ascending ramus, nasal, gnathic process), retrieved a variety of similar morphologies but also obvious or subtle differences. The case of Blarinomys, recently the object of a detailed skull exploration (Missagia and Perini 2018), illustrates this issue. This fossorial sigmodontine was added to the “oxymycterine” group by Hinojosa et al. (1987:3). According to Missagia and Perini (2018:151), the zygomatic plate in Blarinomys is “… slanted posteriorly… and not vertically inserted on the rostrum… the angle of the inferior root of the zygomatic plate relative to the axis of the skull wall, which approximates a right angle in Blarinomys… This configuration gives a more or less squared-off appearance to the zygomatic plate in dorsal view… and may contribute to the oval shaped and wide lumen of the infraorbital foramen...” The described zygomatic plate as well as the infraorbital foramen configuration approach those observed in Neomicroxus but differs markedly with the conditions in Abrothrix lanosa (Fig. 6). Missagia and Perini (2018:159) associated several of the skull traits detected in Blarinomys with fossoriality but also with diet, highlighting that

… some of the similarities [of Blarinomys] with Oxymycterus may be due to a shared insectivorous diet, which can be related to some morphological features of the skull, such as the narrow zygomatic plate, inconspicuous zygomatic notch, and a narrow and elongated mandible… Skulls of species of Oxymycterus also closely resemble that of Brucepattersonius in general morphology… sharing similar features as the elongated rostra and narrow zygomatic plates… Brucepattersonius and Oxymycterus are not closely related… indicating that features indicative of an insectivorous diet appeared convergently different times within Akodontini.

The history of construction and deconstruction of the “oxymycterine” group and the widespread “hocicudo” bauplan across living sigmodontines raises the complex issue of convergence and adaptation in a subfamilial radiation (Losos 2011). How many “signatures” (i.e., diet, substrate, thermal tolerance), if any, are superimposed in this bauplan? Fortunately, several recent studies tackled these or related questions using a large quantity of molecular and morphological evidence (e.g., Parada et al. 2015; Alhajeri et al. 2016; Maestri et al. 2016, 2017; Steppan and Schenk 2017). According to Maestri et al. (2017:626) “… ecological variables (diet and life-mode) explain little of the shape and size variation of sigmodontine skulls [sic] and mandibles” and also

… insectivorous mice evolve at a faster rate than other groups. This suggests that diet can affect the rate of evolution, although not necessarily leading to convergent evolution (e.g., Alfaro et al. 2004). Insectivorous sigmodontines typically share general resemblance, such as an elongated, gracile rostrum… but insectivorous genera such as Rhagomys and Blarinomys lack these features” (Maestri et al. 2017:628), to conclude “if sigmodontines retain a phenotype that permits them to exploit a wide range of dietary resources in a given area (i.e., functional versatility), then convergence is unlikely… a generalist morphology can be coupled with functional specialization… favoring a probable one-to-many mapping between form and function (i.e., one form, many functions).

Some of the obtained results are as challenging as intriguing. Among the former is the statement that “insectivorous mice evolve at a faster rate than other groups” being the same demonstrated for another group of muroids (see Martinez et al. 2018). Although our knowledge about diets in sigmodontines is poor (Pardiñas et al. 2017), stomach diversity (Carleton 1973) and isotopic data (Missagia et al. 2019) favor animalivory as the widespread dietary strategy in Akodontini (review by Pardiñas et al. 2020) and some Abrotrichini (Pearson 1984). If evolutionary rates are faster in invertebrate-eaters, perhaps the chance of independent acquisition of skull traits leading to a “hocicudo” bauplan also is magnified. An intriguing aspect is if the data collection methods employed in these studies are capable of accurately resolving the problems. Maestri et al. (2016:figure 1, 2017:figure 1) assessed cranial variation through landmarks digitized on ventral and lateral views, but none of these points “describes” the zygomatic plate, a crucial cranial structure related to masseteric muscle attachments and, transitively, to diet (Carleton 1980; Voss 1988). A more delicate point emerges from recently published criticisms that undermine the entire architecture of these evolutionary reconstructions (Louca and Pennell 2020), producing a mathematical quagmire reminiscent of the proposals by the theoretical physicist Richard Feynman.

The “hocicudo” bauplan illuminates a key aspect of sigmodontine contemporary studies, the need for refined anatomical research. Either using traditional approaches, high technology (e.g., CT scan), or a combination of both, the amount of basic morphological data needed to support the growing set of evolutionary hypotheses must be expanded. Some progress has been made (e.g., Tulli et al. 2016; Missagia and Perini 2018; Barbière et al. 2019; Pardiñas et al. 2020) and near future prospects are exciting, because several anatomical systems have been explored in a preliminary manner (e.g., petrosal—Wible and Shelley 2020; turbinals—Martinez et al. 2020). In a circumstance like this, the concluding statement of Patton (2005:274) is perceptive, “our abilities in the laboratory, and especially with increasingly sophisticated analytical methods, are almost boundless now, but we all need to take the time, and care, to think about the questions we ask and to use methods that are appropriate and for which there is sound theoretical understanding.”

Supplementary Data

Supplementary data are available at Journal of Mammalogy online.

Supplementary Data SD1.—Access numbers and vouchers for taxa used in the phylogenetic analyses.

Supplementary Data SD2.—Lingual views of upper (A, C, E) and lower (B, D, F) molars showing contrasting degrees of hypsodonty in Neomicroxus latebricola (A, B; MEPN 9803) and two species of Thomasomys (C, D: T. aureus, MEPN 6144; E, F: T. cinereus, QCAZ 16173) recorded in northern Ecuadorian Andes.

Version of Record, first published online 17 March 2021, with fixed content and layout in compliance with Art. 8.1.3.2 ICZN.

Nomenclatural statement: A Life Science Identifier (LSID) number was obtained for this publication: urn:lsid:zoobank.org:pub:07773773-CBD0-448D-A2BA-958AB62EE452

Acknowledgments

Study travels of the senior author to Quito were funded by Fundación Ecominga and the generosity of Javier Robayo. The work of JB scanning specimens in Germany was possible thanks to the “Germany–Brazil–Ecuador Trilateral Cooperation Program,” financed by GIZ international cooperation, and the invaluable help of Claudia Koch and Rainer Hutterer; Claudia Koch deserves also the recognition for obtaining models employed in Fig. 6. Richard Madden generously shared his huge knowledge about hypsodonty. Jim Patton improved the manuscript through an enriching critical reading; the same was done by Robert Martin acting as reviewer. Néstor Cazzaniga enlarged our understanding of tribal naming. Quentin Martinez corrected our primary and erroneous assessment about turbinal nomenclature. Reed Ojala-Barbour and Glenda Pozo kindly assisted us during field works at Bosque de Polylepis. This contribution was made under the economic support and laboratory equipment provided by grant Agencia Nacional de Promoción Científica y Tecnológica 2014-1039 (to UFJP). We are deeply indebted to the abovementioned persons and institutions. This is GEMA’s (Grupo de Estudios de Mamíferos Australes) contribution #31.

Appendix I

Specimens examined belong to the following mammal collections: Colección de Mamíferos del Centro Nacional Patagónico (CNP; Puerto Madryn, Chubut, Argentina); Colección de Mamíferos del Instituto de Investigación de Recursos Biológicos “Alexander von Humboldt” (IavH; Bogotá, Colombia); Colección de Mamíferos “Alberto Cadena García” del Instituto de Ciencias Naturales de la Universidad Nacional de Colombia (ICN; Bogotá, Colombia); Museo de Zoología de la Pontificia Universidad Católica del Ecuador (QCAZ; Quito, Ecuador); Instituto Nacional de Biodiversidad (MECN; Quito, Ecuador); Instituto de Ciencias Biológicas de la Escuela Politécnica Nacional (MEPN; Quito, Ecuador); and National Museum of Natural History, Smithsonian Institute (USNM; Washington DC, United States). GenBank access numbers of the five sequenced specimens in this study are indicated (* Cytb, ** IRBP).

Neomicroxus bogotensis (n = 11): Colombia, Boyacá, Municipio Guacamayas, vereda Alfaro, sitio Piedras Blancas 6.416°, −72.505° (ICN 14722); Cundinamarca, Junín, Reserva Biológica Carpanta 4.563°, −73.683° (ICN 11027, ICN 11028, ICN 11029); Cundinamarca, PNN Chingaza (IAvH 5777 – MT240521*); Santander, Santa Bárbara, Páramo del Almorzadero, Vereda Volcanes 7.076°, −72.848° (UIS-MZ 1596 – MT240520* MT249798**); Santander, Santa Bárbara, Vereda Esparta 7.019°, −72.892° (UIS-MZ 1299 – MT240522* MT249797**); Norte de Santander, Cucutilla, Sisavita, Romeral, Predio Greystar [no coordinates available] (UIS-MZ 907). Venezuela: Mérida, Tabay, 7 Km SE Tabay, La Coromoto 8.6°, −71.02° (USNM 374611, USNM 374612, USNM 374613).

Neomicroxus latebricola (n = 45): Ecuador, Carchi, Espejo, La Libertad, Sector Bosque de Polylepis 0.712202°, −77.981639°, 3,650 m (MECN 3717, 3718, 3719, CNP 6397 – MT240523* MT249799**, CNP 6533 – MT240524* MT249800**, 3735, 3736, 3739, 3740, 3748, 3776, 3777; QCAZ 11142, 11158, 11145, 12504, 12503, 9814; MEPN 10869, 10870, 10887, 10886, 12716, 10644, 12718, 12715B, 10485, 12712, 12715); Carchi, Tulfán, Tufiño, Páramo del Artesón, Comuna La Esperanza (QCAZ 9801); Imbabura, Pimampiro, Mariano Acosta, Laguna Blanca 0.22367°, −77.97867°, 3,400 m (MECN 4763); Imbabura, Zuleta, Faldas del Imbabura 0.248372°, −78.15425°, 3,610 m (MECN 6134, 6135, 6136); Imbabura, Cotacachi, Bosque Protector Neblina 0.342024°, −78.412935°, 2,990 m (MECN 5605, 5606); Napo, Quijo, bosque administrado por la fundación TERRA −0.33422°, −78.1433°, 3,400 m (QCAZ 4090, 4121, 4160, 4167, 5230, 5236, 5239); Tunguragua, Pisayambo, km. Parque Nacional Llanganates −1.044686°, −78.345828°, 3,102 m (CNP 6396).