Materials and methods

The collections of D. integrifolia were made in the municipality of Tenejapa, which is in the region known as the Highlands of Chiapas (^{Villaseñor et al. 2013a}). A single collection was made in November 2019 and was keyed to species using ^{Pruski (2018)}. Additional specimens almost at the same locality (ca. 500 m away) with several duplicates were collected in January 2020. The specimens were deposited in the National Herbarium of Mexico (MEXU) and in the herbarium of El Colegio de la Frontera Sur (CH) (additional duplicates to be distributed).

Ecological niche modeling . Records of D. integrifolia in the Global Biodiversity Facility (^{GBIF 2020}) were downloaded. The records that represented collections of the species in native locations in the Old World were used to generate the model of ecological niche. A total of 312 georeferenced records from Africa and Asia were used, plus the four adventive records from the New World, i.e., the two collections from Chiapas, Mexico, and the two reports from Guatemala (^{Pruski 2011}, ²⁰¹⁸). The model was generated using the Wallace platform (^{Kass et al. 2018}), in which the 312 records were reduced to 280 through the module SpThin (^{Aiello‐Lammens et al. 2015}) by applying a distance of 5 km (^{Baldwin 2009}). The 280 filtered records were modelled on Maxent using 19 bioclimatic variables at 30 seconds of resolution obtained from WorldClim 2 (^{Fick & Hijmans 2017}). A mask related to the countries with native records of the species was implemented. Several models employing all combinations of feature classes and regularization multiplier values were made to select the most appropriate model (^{Morales et al. 2017}). The model with a feature class of Linear-Quadratic-Hinge (LQH), and a regularization multiplier value of 3.0, was chosen based on the Akaike information criterion (^{Warren & Seifert 2011}). To ascertain that the projection of the selected model was implemented correctly a partition of the records was made. This was achieved by using a special partition implementing the Checkerboard type with an aggregation factor of 2 (^{Peterson et al. 2007}, ^{Muscarella et al. 2014}). The model was projected for Mexico and Guatemala to determine whether it would predict the four recorded collections from this region. No cutoff threshold was used and the complete logistic output of the model is shown, in order to allow visualization of all zones with environmental suitability for the models.

Taxonomy. The morphology of the species was also examined in the herbaria using dissection microscope to make a morphological description of the species. Pictures of the habitat and habit of the species were taken in the field. A more detailed account of the morphology and synonymy of the species and genus can be found in ^{Pruski (2011)}.

Results

With an AUC of 0.902 the implemented model yields a high prediction for the occurrence of Dichrocephala integrifolia in their native Africa and Asia (Figure 1). Likewise, the projection for Mexico and Guatemala correctly predicts the four records currently known from this region (Figure 2). The model also shows the areas that are the most ecologically suitable for the species. In Mexico, these regions are the Mexican Transvolcanic Belt, all Sierras Madres (south, western, and eastern), the Highlands of Chiapas, and the Soconusco region, also in Chiapas. The Mexican states that have most suitable ecological conditions for the species are Mexico City, Tlaxcala, State of Mexico, Puebla, and Hidalgo (Figure 2). The only Mexican states in which the model does not project any surface are those of the Yucatan peninsula (Campeche, Quintana Roo, Yucatan) and the adjacent state of Tabasco. In Guatemala, the entities with the highest putative affected areas are Sacatepequez, Sololá, Totonicapán, Baja Verapaz, and Chimaltenango. The model also predicts considerable potentially affected areas in Costa Rica, Panama, and Honduras, and smaller regions in Nicaragua, Belize, and El Salvador. In the Caribbean, the model predicts affected regions in Dominican Republic, Haiti, Jamaica, Cuba, and Puerto Rico (Figure 2). Considering the ecoregions of the WWF (^{Olson et al. 2001}) the model indicates that the pine and pine-oak forest might be the most affected communities in Mexico and Guatemala (67 %), and in the Caribbean, both the humid forest and pine forest might be more suitable for D. integrifolia.

Figure 1 Ecological niche model of Dichrocephala integrifolia in countries of Africa and Asia, where the species is native. Blue points represent collecting sites. 

Figure 2 Ecological niche model of Dichrocephala integrifolia in Mexico, Central America, and the Caribbean. The map shows the records from Chiapas, Mexico, and Guatemala. The two records from Mexico appear as a single point due to its proximity. The record from the Sierra Los Cuchumatanes reported in ^{Pruski (2011)} has no coordinates and due to this the dot in the map is the centroid for this area with coordinates 91.443951° W and 15.583357° N. 

Taxonomy. Dichrocephala integrifolia (L. f.) Kuntze, Revis. Gen. Pl. 1: 333. 1891. Hippia integrifolia L. f., Suppl. Pl. 389. (1782). = Dichrocephala latifolia DC.

Erect annual herbs, 30-70 cm high, with tap roots, moderately to scarcely pilose to hirsutulous, the hairs often adpressed. Leaves alternate, lyrate-pinnatifid to simple, petiolate, blades 3-10 × 2-6.5 cm, with 3-5 lobes, the apical one larger, ovate to subcordate, 2.5-8 × 1.5-5 cm, palmately veined, the lower lobes often reduced to absent, lobe margins serrate, petioles 0.9-4 cm long. Capitulescence cymose paniculiform. Heads small, ovoid to subglobose, peduncles 1.5-3 mm long, heterogamous, disciform; involucre saucer-shaped, 1 × 3 mm, phyllaries biseriate, equal or subequal in length, lanceolate, green. Peripherical flowers numerous (80+), in 6-8 series, pistillate and fertile, corollas white, tubular, 0.4-0.5 mm, with 4 apical lobes. Disk flowers 12-16, perfect, fertile, corollas green to yellowish, campanulate, with a short tube and four apical lobes, 0.8-1.0 mm, with few resinous dots, anthers brown, style with two extremely short deltoid branches. Cypselae obovate to oblanceolate, 1.0-1.2 × 0.4 mm, compressed, greenish to pale brown, with glands near apex, somewhat glutinous, pappus absent or of 1-2 readily deciduous, white bristles, 0.4 mm long. Receptacle convex-columnar. (Figures 3-6).

Figure 3 A. Young blooming plants of Dichrocephala integrifolia. B. Note variation from simple to lyrate-pinnatifid leaves. 

Figure 4 Disciform heads of Dichrocephala integrifolia. Note numerous outer pistillate flowers with white tubular corollas disposed in several series and less numerous central hermaphrodite disk flowers with yellowish-green corollas. 

Figure 5 Peripherical flower (left), disk floret (center), and disk cypsela (left) of Dichrocephala integrifolia. Note more campanulate corolla of the disk flower and one slender pappus bristle in the cypsela at the left.  

Figure 6 A. Dichrocephala integrifolia, young blooming plant. B. Capitulescence. C. Disciform head. D. disk flower. E. Peripheral pistillate flower. F. Cypsela. 

An Old World species native to tropical Africa and Asia, introduced into Europe, SW Asia (Azerbaijan, Georgia, Turkey), Australia, Pacific Islands, and in the New World (^{Chen & Brouillet 2011}, ^{Pruski 2011}, ²⁰¹⁸). In the Americas it was previously known from Guatemala (^{Pruski 2011}, ²⁰¹⁸), and is here reported from the adjacent state of Chiapas, in southern Mexico. Medicinal uses (indigestion, anti-inflammatory, and antiemetic) are reported from China (^{Chen & Brouillet 2011}).

Habitat in Mexico. Wet and shady trail sides, near small streams and crops in shady hillsides of disturbed humid mountain or pine-oak forests, with Pinus tecunumanii F. Schwerdtf. ex Eguiluz & J.P. Perry, Quercus sp., Alnus acuminata Kunth, Montanoa hexagona B.L. Rob. & Greenm., Sambucus nigra L., Cornus excelsa Kunth, Crataegus pubescens (C. Presl) C. Presl, Equisetum hyemale L., and Plantago australis Lam. (Figure 7).

Figure 7 Local environment where Dichrocephala integrifolia was collected in the municipality of Tenejapa, Chiapas, southern Mexico. 

Elevation in Mexico. 2,057-2,200 m.

Specimens examined. Mexico, Chiapas: 1.9 km NE of Tenejapa, on a wet trail side, 16° 49ʹ 24ʺ N, 92° 31ʹ 17ʺ W, 2,083 m, Nov-13-2019, Oscar Hinojosa-Espinosa # 734, Mario Ishiki, & Alfonso Luna (MEXU); 1.7 km NE de Tenejapa, shady and wet sites on hillside near crops of corn, bean, and squashes, 16° 49ʹ 21ʺ N, 92° 31ʹ 01ʺ W, 2,057 m, Jan-20-2020, Mario Ishiki # 3675, Henry Castañeda, Alfonso Luna, Mario Sánchez-Gutiérrez, & Erick A. Luna (CH, MEXU, and several duplicates to be distributed).

Discussion

The fact that the selected model of ecological niche correctly predicts actual collections of Dichrocephala integrifolia in Africa and Asia, and the currently known occurrences in the Americas, suggests that the model is accurate and reliable. According to the model, the Mexican states with more suitable set of conditions for the species are in the mountainous regions with pine, pine-oak, and humid mountain forests. They are among the states with a higher density of alien plant species (^{Villaseñor & Espinosa-García 2004}), except for the state of Tlaxcala. Indeed, the state of Chiapas, in which the species was collected, is the state with the second highest density of alien species in Mexico (^{Villaseñor & Espinosa-García 2004}). Most of the exotic plant species in Mexico arrived from the Old World (^{Villaseñor & Espinosa-García 2004}) as is the case for Dichrocephala integrifolia.

It is unclear from where and how this species was introduced into the New World. ^{Pruski (2011)} suggested that the species could have been introduced from Africa, based on its apparent restriction to the Atlantic watershed in Guatemala, although he did not discard the possibility of a long dispersal from Asia, or as a result of anthropogenic activities. It seems somewhat difficult that the small cypselae can survive such long dispersals through sea water; however, some physiological related investigations may shed light on it. Therefore, we find more likely that the species was accidentally carried into the New World by people, as the rise of human movement has increased the opportunities for species transportation to novel sites (^{Peterson 2003}). Such diminutive cypselae (1-1.2 mm, Figure 5) might be inadvertently carried (e.g., in a goods shipped, in the groove of a sole shoe, a bag pack, etc.). Moreover, as discussed above, D. integrifolia has affinities for human disturbed places (e.g., trails, crops, roadsides) both in its native and non-native distribution range (Figure 7).

^{Pruski (2018)} expected that D. integrifolia would become naturalized in the Antilles and Central America, following the pattern of other exotic asters in Guatemala, which were presumably introduced into the New World from Africa. Although the species has not been found in these regions yet, our model strongly predicts areas in both the Caribbean and Central America that are environmentally appropriate for D. integrifolia supporting the hypothesis of ^{Pruski (2018)}. We also expect the species to become naturalized and more widespread in Mexico based on the results of the ecological niche model and on its weedy behavior. In west tropical Africa, where the species is native, it is regarded as a high ground weed (^{Hutchinson & Dalziel 1963}). In China, D. integrifolia is reported from a variety of habitats including meadows, slopes, shaded valleys, but also from altered habitats, such as trails and roadsides (^{Chen & Brouillet 2011}).

It is unknown how long the species has been dwelling in Chiapas. This state is one of the most biodiverse of Mexico: it harbors some 1,912 and 8,790 native genera and species of plants, respectively (^{Villaseñor 2016}). Even though D. integrifolia has not been considered a noxious weed, an effort to avoid its continuing expansion and further introductions would be desirable, because weeds might have a negative impact on the native flora, and/or local natural ecosystems and agroecosystems (^{Villaseñor et al. 2013b}). We hope this contribution will provide useful information to help identify D. integrifolia as well as the regions in which the species might be present, and therefore, help to monitor its spread and develop strategies to prevent further introductions.