The Red-gartered Coot Fulica armillata (Vieillot, 1817) (Aves: Rallidae) is one of the six species of this genus occurring in limnetic wetlands of the southern Neotropical region, including Chile, Argentina, Uruguay and southeastern Brazil (^{Couve et al. 2016}). In Chile, this coot has a wide latitudinal distribution: from the Río Copiapó in the north (~27ºS) to Isla Grande de Tierra del Fuego and Isla Navarino in the south (~54ºS) (^{Medrano et al. 2018}). This species has been described as a primarily herbivorous waterbird, although its diet also includes invertebrates although in lower proportion than plants (^{Ruíz 1993}; ^{García et al. 2008}).

The abundance of F. armillata in the southern Neotropical region has been estimated in ~1,000,000 birds (Waterbird Population Estimates 2017). Thus, and due to its herbivorous feeding habits, this bird may potentially play an important role both as controller of the biomass of aquatic macrophytes and as seed disperser (^{Ruíz 1993}; ^{Bortolus et al. 1998}; ^{Charalambidou & Santamaría 2005}), as shown for other coots such as Fulica cristata (Gmelin, 1789) (^{Stewart & Bally 1985}), Fulica atra (Linnaeus, 1758) (^{Perrow et al. 1997}) and Fulica americana (Gmelin, 1789) (^{Esler 1989}). Despite the important ecological role of coots, the natural history of F. armillata in Chile is scarcely known. ^{Cody (1970}) and Riveros et al. (1981) studied the spatial distribution of this bird in wetlands of central Chile (c. 30-38°S), ^{Silva et al. (2011}) described its reproductive biology in Laguna Santa Elena (central Chile; ~36°S), while ^{Kennedy (1977}) and Ruíz (1993) studied ecological and trophic aspects at the wetland of Río Cruces (southern Chile; ~39°S). However, no studies have been carried out on populations of F. armillata along the semiarid coastal region of northern Chile, where coastal wetlands host a different suite of waterbirds compared to their southern counterparts (^{Sielfeld et al. 2012}). An important wetland at that region is Punta Teatinos (PT hereafter), located north of the conurbation of Coquimbo and La Serena in north central Chile (~29ºS; Fig. 1a, b), where field observations carried out during 2016-17 showed that F. armillata is one of the most abundant resident herbivorous waterbirds. Therefore, this study was aimed to provide a first assessment of the use of plant food resources by F. armillata at the PT wetland, based on a microhistological analysis of its feces and a comparison with published results.

The PT wetland is a coastal lagoon of ~69 ha with freshwater inflows from the ravine El Romeral and salty water influence from the nearby sand beach surf zone during extreme high tides (^{Jorge et al. 1998}). Stuckenia pectinata (L.) Börner is the most abundant submerged macrophyte at this lagoon (C. Velásquez, obs. Pers.). The western side of the wetland (see Fig. 1c) is bordered by the upper shore sandy beach level, which F. armillata used as dropping area. Thus, during May 2016, fresh feces of the Red-gartered Coot (n = 30) were randomly collected along the western sandy bank of PT, and then stored in sealed plastic bags with alcohol (70 %) to preserve them. Additionally, with the aim of maximizing the likelihood that collected feces belonged to different coot individuals, the sampling was carried out through the entire border of the lagoon, at places used by red-gartered coots as resting or passing spots.

In order to identify the species of plants in the feces, a series of microhistological techniques (after ^{Johnson et al. 1983}) to analyze food habits of herbivores were used to (i) prepare samples of epidermal cells collected from the macrophytes of the PT wetland, (ii) compare those samples with the plant epidermal cells found intact in coot feces, and (iii) quantify the examined material. The comparisons were carried out with an optic microscope (10x magnification) connected to the software Micrometrics Premium to obtain high resolution digital pictures (ACCU-SCOPE camera) (see ^{Velásquez et al. 2018}).

FIGURE 1 Geographic location of the wetland of Punta Teatinos, in north central Chile (a) and at the northern side of Bahía de Coquimbo (b). A general picture of the wetland is also included (c); the red arrows indicate the western side of the wetland where the feces of coots were collected. / Localización geográfica del humedal de Punta Teatinos, en el centro norte de Chile (a) y en el costado norte de Bahía de Coquimbo (b). Se incluye una foto general del área de estudio (c); las flechas rojas indican el lado oeste del humedal donde se recolectaron las fecas de las taguas. 

The results were first expressed as the percent occurrence frequency of each plant species in the sample of 30 feces (replicates). However, this measure cannot clearly reflect the intensity of consumption of food items, considering that the presence of different plants in each feces is recorded only once, regardless of the amount of biomass ingested of each plant. Therefore, an exhaustive subsampling was conducted in order to obtain an estimate of the incidence of each plant species within the feces. To that purpose, two portions from each feces (herein referred to as subsamples) were randomly selected, and each subsample was uniformly spread out in a Neubauer counting chamber, where 10 optic fields of 1 mm² were also randomly selected for observation, thus examining a total of 600 fields (i.e. 10 fields x 2 chambers x 30 feces). The presence of each plant species in each set of 10 fields per chamber was recorded as a percent frequency, and the two frequency values recorded for each feces were then averaged. Finally, the incidence of each plant species was expressed as the grand average (± SE) of the 30 average values obtained from the feces collected in the PT wetland. The analysis of feces showed that the only conspicuous macrophytes consumed by F. armillata were Cotula coronopifolia L., Sarcocornia fruticosa (L.) Scott, and S. pectinata which was by far the dominant food item (Fig. 2). Remarkably, S. pectinata was present not only at all feces, with an occurrence frequency of 100.0 %, but also at each of the 600 optic fields examined, with an incidence of 100.0 ± 0.0 % (Table 1) suggesting a heavy consumption by this species. In contrast, the occurrence frequency of C. coronopifolia and S. fruticosa in the 30 feces was 66.7 % and 36.7 %, respectively, but their incidence within the feces was much lower, reaching only 13.9 ± 0.6 % and 15.3 ± 1.0 %, respectively (Table 1). In addition, these results were consistent with the different life habits of the three macrophytes (see ^{Ramírez & San Martín 2006}), indicating that coots fed predominantly on submerged (S. pectinata) rather than marshland (C. coronopifolia and S. fruticosa) plants (Table 1).

Albeit the above results do not allow a proper evaluation of feeding preferences (e.g. ^{Jacksic et al. 1979}), prior antecedents suggest that the high dietary importance of S. pectinata would not be due to its selective consumption by coots, but rather to its submerged habit and its high abundance at the PT wetland. Unlike other Chilean coots such as Fulica leucoptera (Vieillot, 1817) and Fulica rufifrons (Philippi & Landbeck, 1861), F. armillata is known to display its foraging activities mainly at the central, deeper portion of lagoons (^{Cody 1970}; ^{Riveros et al. 1981}). On the other hand, in shallower, non-lagoon wetlands such as the large wetland of Río Cruces (southern Chile), both F. armillata and F. leucoptera (see Table 1) feed primarily (~85 %) on the submerged and most abundant macrophytes (particularly Egeria densa Planch; ^{Ramírez et al. 1991}; ^{Ruíz 1993}), a pattern shown also by other common water birds such as the Black-necked Swan Cygnus melancoryphus (Molina, 1782) (Table 1; ^{Corti & Schlatter 2002}; ^{Norambuena & Bozinovic 2009}). Thus, F. armillata would behave as an opportunistic consumer, although future studies should clarify whether the submerged habit itself is a factor influencing the feeding choice of this species.

On a general level, the similar foraging patterns of the commonest herbivorous birds across wetlands of Chile and the southern Neotropical region (^{Couve et al. 2016}) might play an important role in the succession of macrophyte assemblages. Submerged macrophytes such as S. pectinata, E. densa and Ruppia maritima L. are dominant in coastal wetlands of the southern Neotropic, and the large amount of plant necromass derived from bird foraging may determine the zonation and successional patterns of the vegetation, or even retard the successional process (^{Ruíz 1993}; ^{Bortolus et al. 1998}; ^{Corti & Schlatter 2002}).

FIGURE 2 Pictures of the aquatic macrophytes consumed by Fulica armillata in the study area: Stuckenia pectinata (a), Cotula coronopifolia (b) and Sarcocornia fruticosa (c). / Fotografías de las macrófitas acuáticas consumidas por Fulica armillata en el área de estudio: Stuckenia pectinata (a), Cotula coronopifolia (b) and Sarcocornia fruticosa (c). 

TABLE 1 Frequency of occurrence of the food items of herbivorous waterbirds in coastal wetlands of Chile (Fulica armillata, Fulica leucoptera and Cygnus melancoryphus). The macrophytes are classified by their life habits (see ^{Ramírez & San Martín 2006}). Food items were determined throughout analyses of feces (*) and analyses of stomach contents (**). / Frecuencia de ocurrencia del ítem alimentario de aves acuáticas herbívoras en humedales costeros de Chile (Fulica armillata, Fulica leucoptera y Cygnus melancoryphus). Las macrófitas se clasifican según sus hábitos de vida (ver Ramírez & San Martín 2006). Los ítems alimentarios fueron determinados a través del análisis de fecas (*) y análisis de contenido estomacal (**). 

Thus, the close connection between submerged macrophytes and herbivorous waterbirds such as F. armillata would make wetlands highly sensitive to bottom- up forcing. For instance, industrial activities triggered sudden environmental changes at the Río Cruces wetland in 2004, leading to a radical decline in the spatial cover of E. densa and the population abundance of F. armillata and others birds (e.g. ^{Jaramillo et al. 2007}; ^{Lagos et al. 2008}). Only eight years later, macrophytes begun to recover and a concomitant increase in the abundance of coots was observed (Jaramillo et al. 2018a, 2018b), although the mechanisms underlying their decline and subsequent recovery remain not fully understood.

Even though F. armillata is not considered an endangered species (^{Bird Life International 2017}), a better knowledge of its trophic ecology may provide crucial information for the management and restoration of coastal wetlands over the southern Neotropical region.