Microlicia (Melastomataceae) is a Neotropical genus nearly restricted to southeastern Brazil, and the Guiana Shield in Venezuela, with a few species in some places in the Andes of Bolivia and Peru. A new species of Microlicia endemic to the mountains of eastern Andes of Colombia is described and illustrated. Its affinities with other morphologically similar species from Venezuela are also documented. This novelty is the first record of the genus for Colombia and the northern Andes. It is argued that this disjunct distribution of the genus is attributable to the phenomenon of long-distance seed dispersal by wind.

Northeastern Andes, endemism, long-distance seed dispersal, northern Andes, sub-páramo

Microlicia D. Don is a genus of approximately 140 species, 135 of which occur in Brazil, two species grow in the Guaiana Shield of Venezuela, and two in the Andean mountains of Peru and Bolivia (Michelangeli and Cotton 2008, Romero 2013, Romero and Woodgyer 2015, Flora do Brasil 2019). The genus is extremely diverse in the campos rupestres and savannah vegetation of the states of Bahia, Minas Gerais and Goiás in Brazil (Romero 2003a, 2013).

The species of Microlicia are generally characterized by solitary flowers with five or occasionally six petals, a superior ovary with three or occasionally five locules, and capsules dehiscing longitudinally from the apex to the base (Almeda and Martins 2001, Romero 2003b). Difficulty in the delimitation of closely related species may explain the high number of names proposed for several species complexes in the genus (Romero and Woodgyer 2014).

This novelty represents the first record of the genus for Colombia and the northeastern Andes. The area from which this new taxon comes is poorly collected due to the difficult access and social conflict in the region. The current peace process in Colombia and the project Boyacá BIO developed by the Alexander von Humboldt Institute and the Government of the Department of Boyacá has allowed access to these new biologically unexplored areas.

This new species was found in the course of reviewing collections in regional herbaria of Colombia generated by the Boyacá BIO project. After several collections were located in the Federico Medem (FMB, Villa de Leyva) and Universidad Distrital (UDBC, Bogotá) herbaria, it was possible to visit the localities and collect new samples. Therefore new collections, field photographs and samples of flowers, fruits, and seeds stored in 70% alcohol were made.

Measurements of vegetative parts were made in the dry herbarium material with a digital caliper with a precision of 0.1 mm. The measurements of floral parts were based on fresh flowers preserved in alcohol from the type specimens. A Leica S8AP0 microscope was used. Photographs of leaves, flowers, fruits, and seeds were taken in the field and laboratory from fresh material using an MC190 HD camera.

The distribution of the Microlocia was associated with South America wind currents to evaluate the possible cause of the presence of the new taxon in the Andes of Colombia. For this, the map of winds patterns from the Earth Nullschool (https://earth.nullschool.net) was used, on which the distribution records of Microlicia from the electronic Tropicos database of the Missouri Botanical Garden (http://www.tropicos.org) were mapped using Arc-GIS version 10.2.1.

Microlicia colombiana is recognized by its strigose nodes with long-stalked glands curled trichomes, and early caducous glands, leaves 3-nerved or 3-plinerved, hypanthium 10-costate, and pale trichomes between the calyx lobes. Other characters, although not unique within the genus but that together help to differentiate M. colombiana, are 5-mereous flowers, petals with rounded apex, without seta, glabrous stamens, and polysporangiate anthers. The anthers of species that are polysporangiate have both of their thecae divided into numerous small locules in a way that resembles the structure of a honeycomb (Almeda and Martins 2001), character that can be observed externally without dissecting the anther. The most morphologically similar species is Microlicia benthamiana that occurs in Cerro Roraima in Venezuela and Brazil. Both have glabrescent subssesile and 3-nerved leaves, and 10–ribbed hypanthia. In M. benthamiana, however, the internodes are shorter, branches and leaves do not have spherical glands and pale trichomes, and the flowers are larger (Table 1). Microlicia colombiana also bears some similarity to M. guanayana, which occurs on Cerro Guanay in the state of Amazonas, Venezuela. Both have leaves of similar shape and size; however, Microlicia guanayana has leaves 1-nerved (vs. 3-nerved in M. colombiana), internodes and leaves with strigose indumentum with trichomes > 1 mm long (vs. without setae or only over veins and 0.3 mm long), petals >10 mm long (vs. <8.5 mm long) and setulose and unribbed hypanthia (vs. 10-ribbeb) (Table 1).

Regarding the Andean species that grows in Peru and Bolivia, the most affine is M. sphagnicola Gleason, nevertheless the latter has flowers 4-merous and glandular marginate connective in outter (antesepalous) stamens.

Microlicia is a highly diversified genus in south-central Brazil, with a few disjunct species in the Andes of Bolivia and Peru and the Guiana Shield of Venezuela (Romero 2003a). The most morphologically similar species to M. colombiana grow in high areas (1000–2700 m elevation) of the Guiana Shield around 650 km (M. guanayana) and 750–1300 km (M. benthamiana) away from the northeastern Andes of Colombia (Figure 4A).

This distributional pattern could be explained by long-distance seed dispersal events from the East (southeastern Brazil or Guiana Shield) towards the West (central and northern Andes). In Melastomataceae, there are several different genera with disjunct distribution between the northern Andes and the Guiana Shield, such as in Boyania Wurdack (Mendoza 2010) and Phainantha Gleason (Ulloa and Neill 2009). Also, there are species with known disjunct distributions between the Guiana Shield and the Andes without records in the lowland of Amazonia and Orinoquia, such as Marcetia taxifolia (A. St.-Hil.) DC. (Martins 1989), Graffenrieda intermedia Triana and G. weddelli Naudin (Almeda et al. 2016), and Monochaetum bonplandii (Kunth) Naudin (Alvear and Almeda in press).

The majority of these groups with disjunct distributions have capsular fruits with seeds dispersed by wind. Renner et al. (2001) and Renner (2004) discussed the preponderant role of long-distance dispersal in the evolution of the distribution of Paleotropical Melastomataceae. The small seeds of some of the genera in the Melastomataceae are likely to travel great distances in air currents. The disjunctive distribution of Microlicia could possibly be associated with the wind currents in the north of South America that run from East to West (Figure 4B). However, it is possible that the distribution in the center of South America is a combination of stepping-stone route with consecutive short dispersal events.

Another condition is the type of ecosystem, like sandy enclaves in highlands Andes. Successful colonization and establishment are more likely in environments that approximately match the source environment (Cavendar-Bares et al. 2009). The most affine species to M. colombiana, grow in what is known as Pantepui (Wurdack 1973), which includes the characteristically flat topped mountain summits of the Guiana highlands, between 1500 and 3000 m a.s.l., encompassing a range of meso‐ to submicrothermic temperature regimes (MAT 18–8 °C), the extensive herbaceous ecosystems are developed on open sandstone surfaces (Huber 2006). These temperature conditions, herbaceous or shrub vegetation and sandy soils occur in some areas along the eastern slope of the eastern Andes in Colombia as the type locality of M. colombiana.

Dispersal by birds that transport seeds adhered in mud may also be a possibility to explain the disjunct distribution of Microlicia in the northern Andes. However, there is no greater evidence to document this. Bird migrations occur mostly from north-south and not from East to West (Gillespie et al. 2011), and the possibilities of transporting seeds in mud from areas where soils are sandy are unlikely.

This study was financed by the Boyacá Bio project, with funds from the General System of Royalties of Colombia that were administered by the National Planning Department (BPIN 20170000100009) executed by the Government of Boyacá and operated by the Institute of Research of Biological Resources Alexander von Humboldt (cooperation agreement No. 17170). We thank Hernando García (Vice Principal of Institute Humboldt) for supporting the elaboration of the article, to the herbaria FMB and UDBC and Francisco Castro for acces to collections, Cristina Rueda Uribe for reviewing the text, and to Andres Rymel Acosta for preparingthe distribution maps.