Fruiting phenology and dispersal syndromes in a sandy coastal plain in southeastern Brazil

Fruits have a wide variety of morphological and phenological characteristics that have been related to environmental conditions and seed dispersal mode. In this paper, we describe the fruit morphology, the fruiting phenology and infer dispersal patterns of 52 species from restinga of Maricá, Rio de Janeiro, in order to understand the richness and temporal variation of these resources in the community. Fleshy, indehiscent, and colored fruits, typical of zoochory, predominate in the restinga (77.8%). Anemochoric fruits represent 13.3%. In 42% of zoochoric species, fruits go through three to five colors until maturity, and different stages of ripeness can be observed on the same plant. A constant supply of zoochoric and anemochoric fruits was observed throughout the year. Unlike flowering, there were no significant correlations between fruiting activity and intensity and abiotic factors. For the community studied, the fruiting pattern observed also contrasts with flowering, due to the lower seasonality, and intensity suggesting that biotic factors, such as seed dispersers (in the case of zoochoric fruits) may have relevance in determining fruit ripening and seed dispersal periods in coastal environments.


Introduction
Fruits show a large range of sizes, shapes, number of seeds, colors, odors, ripening phenology, and chemical properties. Many of these features have been related to the environment conditions, mode of seed dispersal, and the animals involved in seed dispersion, being usually gathered to describe dispersal syndromes (van der Pijl 1982;Gautier-Hion et al. 1985;Wheelwright & Janson 1985). Color is specially one of the most important cues used by frugivores to find fruits (Lomáscolo & Schaefer 2010). Fruit color convergence in unrelated plants is independent of phylogeny and supports the hypothesis that frugivores play an important role in fruit evolution (Lomáscolo & Schaefer 2010;Valenta et al. 2018). Beyond a host of animal species (zoochoric dispersal), diaspores (seeds, fruits, infructescences and other dispersal units) can reach the ground by action of wind (anemochory), water (hydrochory) or by selfdispersal (van der Pijl 1982).
M o r p h o l o g i c a l a n d p h e n o l o g i c a l characteristics of fruits contribute to the understanding of the temporal variation of these resources and their relationship with seed dispersers (Wheelwright & Janson 1985;Schaik et al. 1993). Influence of abiotic (temperature, precipitation, wind, humidity) and biotic factors (primary and secondary dispersers of fruits and seeds, animal predators) results in a most favorable period for seed dispersal (Rathke & Lacey 1985;Opler et al. 1976;Schaik et al. 1993;Morellato et al. 2000). Seasonality in the number of plant species bearing ripe fruits decreases from temperate to tropical forests, largely as a result of the increase in the average duration of the fruiting phenophases (Jordano 2000), resulting in a continuous supply of resources for animals in tropical areas Zamith & Scarano 2004;Reys et al. 2005;Marchioretto et al. 2007). Zoochorous species prevail in tropical environments (van der Pijl 1982;Rathke & Lacey 1985). In the Atlantic Forest, high frequency of animal-dispersed species, specifically bird-dispersed ones, indicate that frugivorous animals are very important for the maintenance of communities (Martins et al. 2014). In restinga vegetation (habitat marginal to the Brazilian Atlantic Forest), where conditions are extreme (high temperatures, steady winds, high salinity and sandy soil poor in nutrients), predominance of zoochory was also recorded (Zamith & Scarano 2004;Martins et al. 2014).
Here, we described the fruit morphology (size, shape, color and number of seeds) and followed up the fruiting phenology of 52 species from an area of restinga vegetation in southeastern Brazil. These gathered data support us to infer important aspects related to seed dispersion in these areas. It is noteworthy that seed dispersal plays a crucial ecological role in the maintenance of biodiversity and natural regeneration of habitats worldwide (Morellato et al. 2016). Despite its great importance, the connection between morphological features, dispersal patterns, and phenology of fruits in restinga ecosystem are unexplored. In this context, we aim in this study: 1-to examine the occurrence and the diversity of seed dispersal syndromes in restinga vegetation; 2-to characterize the distribution throughout the year of the different seed dispersal syndromes; 3-to investigate the relationship between the fruiting periods and climatic factors; 4-to assess the temporal availability of fruit resources for the potential dispersing fauna. In particular, we explore the coloring of the fruits by monitoring of ripeness, the presentation of colors in the plant and the distribution of fruit colors throughout the year. We expect that this study can help incrementing successful restoration and conservation actions for restinga areas, which have been under intense anthropic occupation (Rocha et al. 2007;Marques et al. 2015).

Materials and Methods
The study was conducted in restinga (sandy coastal plains) within the Environmental Protection Area of Maricá (APA of Maricá -496 ha), Rio de Janeiro state, Brazil (22 o 52'-22 o 54'S and 42 o 49'-42 o 54'W). The climate of the region is Aw tropical humid (Alvares et al. 2014), with the weather being wetter between October and March (warm/rainy season). We conducted the study over a 1-year period from April 2010 to March 2011. The mean annual temperature was 23.7 °C (maximum mean in February with 28.5 °C and minimum mean in June with 19.5 °C). The total precipitation was 1,168.7 mm, with 552.6 mm between October and March and 616.1 mm between April and September. We observed a water stress in February (Instituto Nacional de Meteorologia, INMET-RJ).
We carried out fruiting phenological monitoring of 402 individuals of shrubs and trees (all individuals over 50 cm in height). Plants were marked and numbered along a transect of 500 m on the internal sandy ridge (sandy soil where the vegetation grows in thickets interspersed with almost completely clear spaces), divided into 50 plots of 10 m × 10 m. Sampled species represent approximately 80% of woody species in this area (Rodarte 2008). We organized the floristic list according to Flora do Brasil (2020) and The Plant List (2019).
In this paper, we used the botanical term "fruit" in a broad sense to describe types of diaspores irrespective of their origin and structure (i.e., "true" fruit, pseudo-fruit, etc.). We collected ripe fruits from at least five individuals for each species. For each fruit we measured the length (cm) using a caliper, and we counted the number of seeds. We also registered the dehiscence (dehiscent or indehiscent) and the consistency (dry or fleshy). Dehiscent fruits with arillate seeds were regarded as fleshy fruits because they constitute a form of food resource. We classified the colors of ripe fruits of each species by human perception as black, dark brown, brown, red, dark red, yellow, orange, green, and multicolored (color of aril and/ or seed contrasting with the outer or inner surfaces of fruit capsules). We followed the color change of the fruits through the course of maturation. We performed inferences about the dispersion syndrome, according to van der Pijl (1982). For eight species we obtained the information about fruit length, seed number, and dispersal syndrome from the literature.
The study involved only the phenophase of fruit set. Two classical methods for the phenological analysis were employed: activity and intensity percentages. Activity method (based in presence or absence of phenophase of the individuals) was used to indicate percentage of individuals in the population that were manifesting fruiting event (Bencke & Morellato 2002). We visually assessed intensity of events using a semiquantitative scale (Fournier 1974, adapted for four classes), namely: (0) absence of the event; (1) 1 to 33%; (2) 34% to 66% and (3) 67% to 100% of the canopy showing the phenophase. Each individual was evaluated weekly and scored from 0 to 3 for fruiting event during the study period. We plotted the data monthly, using the most representative score for the month. We expressed intensity of events in each month by the formula: % of intensity = (ΣFournier/3 × N) × 100), where: ΣFournier is sum of semi-quantitative data attributed to each individual; 3 is maximum value of categories adopted; and N is total number of plants evaluated (Fournier 1974).
We tested correlation (Spearman rs at a significance level of 0.05) between events of fruiting (intensity and activity of fruit set of plant community, and species grouped per syndrome) and some climatic data [precipitation (mm), day length (h), and mean, minimum and maximum temperature ( o C) in studied months] (Statsoft 2005). The National Institute of Meteorology INMET/RJ-Maricá Station provided temperature and precipitation data. We calculated day length data according to Pereira et al. (2001) and Varejão-Silva (2000).
Voucher specimen were deposited in Herbarium of Museu Nacional (R), Universidade Federal do Rio de Janeiro and Faculdade de Formação de Professores (RFFP) from Universidade do Estado do Rio de Janeiro.
Fleshy fruits were recorded in 82% of the species, and indehiscent pattern (71%) was predominant (Tab. 3). Andira legalis (Fabaceae) was the only zoochoric species with dry fruit. We observed that 58% of the species studied have fruits with one or two seeds (Tab. 2).
In relation to fruits maturation, we observed that 42% (19) of species, especially the ones whose fruits seems to be dispersed by animals ( Fig. 1; Tab. 3), the fruits pass through three or five hues until complete maturation, and different maturation stages (different colors) are present in the same plant (Calyptranthes brasiliensis, Eugenia astringens, Myrcia ilheosensis and Neomitranthes obscura). Some species did not change the fruit color during the maturation (e.g., Andira legalis) (brown fruits), Couepia ovalifolia and Byrsonima sericea (both with green fruits). Considering zoochoric and zoochoric/ self-dispersion fruits, there is a predominance of black in ripe fruits, taking up 39.5% of all species, followed by red and multicolored (13.2% each),  The set of plants studied showed fruit events throughout the study period (Fig. 2). Lower levels of activity and intensity were recorded in August (2.2% and 1.2%, respectively) and higher levels were recorded in January (12.7% and 6.6%, idem). In relation to zoochoric species, we observed availability and dispersing of fruits throughout the year, with higher rates of intensity and activity in the months of May and September and peaking in January. For anemochoric plants, the highest percentage of dispersal was in February (Fig. 2).  Spearman correlation coefficients obtained between phenological phases (intensity and activity of fruit set of all species, of anemochoric and zoochoric species) and abiotic factors showed no significant correlation (Tab. 4). Considering the availability of fruits based in fruit color of zoochoric and self-dispersion species, we observed black and multicolored fruits throughout the year. Black fruits had higher percentages of intensity and activity in May (conspicuously the fruit set of Calyptranthes brasiliensis, Myrcia ilheoensis and Vitex polygama) and September (Eugenia umbelliflora, Guapira opposita and Neomitranthes obscura were the most expressive species). Multicolored fruits showed higher fruiting from October to March. Red and dark red fruits showed higher percentages in January, especially the species Erythroxylum ovalifolium and Alchornea triplinervea. Yellow fruits were restricted from August to January and April (Garcinia brasiliensis, Cathedra rubricaulis and Cupania emarginata). Green fruits were restricted from January to March (only Byrsonima sericea). Orange fruits had higher percentages in December with more expressive fruit set of Allagoptera arenaria and Myrciaria floribunda. Brown and dark brown fruits were present from June to March, with low percentages (Fig. 3).
In restinga areas, animals that feed on fruit comprise resident and migratory birds, small mammals, lizards, and anuras (Cerqueira et al. 1993;Fialho et al. 2000;Gomes 2006;Silva & Britto-Pereira 2006). In restinga of Jurubatiba (Rio de Janeiro state), Gomes (2006) (Fournier 1974) and activity for the fruit set of all species studied; c. percentage of intensity (Fournier 1974) and activity for the fruit set of zoochoric and zoochoric/self-dispersal plants; d. percentage of intensity (Fournier 1974) and activity for fruit set of anemochoric plants.
of 35 species which are consumed by birds and concluded that most of the resident and migratory birds of restinga are frugivorous. Among all plant species registered by Gomes (2006), 13 also occur in Maricá (including O. notata and P. arrabidae), likely also used as food by birds in this area. Cerqueira et al. (1994) point out that of the ten species of small mammals studied in restinga of Maricá, eight feed on fruits, particularly the marsupial species Philander frenatus (Olfers 1818), which feed on fruits of Aechmea, Erythroxylum, Passiflora, Paullinia and Pilosocereus (Santori et al. 1997). Erythroxylum ovalifolium represent 50% of diet of the lizard species Tropidurus torquatus (Wied 1820) in restinga of Maricá (Fialho et al. 2000). Fruits of Erythroxylum ovalifolium and Monteverdia obtusifolia are important components of the diet of the threatened Anura species Xenohyla truncata (Izecksohn 1959), which inhabits restingas of Rio de Janeiro, including Maricá (Silva & Britto-Pereira 2006). The same authors report that X. truncata lives in bromeliads Neoregelia cruenta and suggest this species pass seeds through its feces to bromeliads, where M. obtusifolia seedlings are easily observed (personal observation).
A variety of colors in ripe fruits was recorded in restinga of Maricá, with a predominance of black, within 39.5% of the species, followed by red and multicolored (13.2% each). Red and black fruits are dominant in many tropical and temperate ecosystems (Wheelwright & Janson 1985;Willson & Whelan 1990;Galetti et al. 2011) and, together with multicolored fruits, have been linked to consumption mainly by birds Galetti 2002;Galetti et al. 2011;Valenta et al. 2018). According to Schmidt et al. Rodriguésia 73: e02012019. 2022 (2004) the prevalence of black and red colors in plant communities seems to be explained by their greater conspicuity against a natural background (immature fruits of different colors, leaves, bark, petiole, etc.) when compared to other colors, increasing chances of being detected by frugivores, mainly birds.
When we investigated, more specifically, color variation during fruit maturation, we observed that, in 42% of zoochoric species, individual fruits go through three to five hues until complete maturation, and different maturation stages (different colors) can be observed in the same plant. This characteristic was common in some species of Myrtaceae that pass through several colors until complete maturation. In N. obscura, for example, plants display fruits of different colors (orange, red, dark red and black), representing the different stages of maturation. Fruits often change color as they mature, and slow and gradual fruit maturation resulting in plants with fruits of different colors may be an important strategy to increase fruit display contrast and influence the detection of these fruits by birds (Willson & Melampy 1983;Wheelwright & Janson 1985). Stiles (1982) suggested plants with immature and ripe colored fruits may function as a pre-ripening fruit flag, a sign that mature fruits will become available over time.
A constant supply of mature fruits was observed in the studied community. The highest percentages were registered in January/February, but they were not much different from the other months. Unlike flowering (Rodarte 2008), there were no significant correlations between fruiting phenophase activity and intensity percentages, and abiotic factors tested (precipitation, temperature and length of day). Morellato et al. (2000) also suggested that climatic factors do not limit fruit production in Atlantic forest trees. For the restinga community studied, the fruiting pattern observed also contrasts with flowering due to the lower seasonality and intensity of events (Rodarte 2008). Phenological studies in Atlantic forest habitats have also demonstrated non-seasonal patterns in fruit production Morellato et al. 2000;Medeiros et al. 2007). In both the Atlantic rain forest and restinga, flowering is more seasonal, concentrated in the warm/rainy season (October to March) Rodarte 2008). These data reinforces the idea that the biotic factors, such as seed dispersers (in the case of zoocoric fruits) may have relevance in determining fruit ripening and seed dispersal periods in coastal environments like Atlantic forest and restinga vegetation (Morellato & Leitão-Filho 1991;Morellato et al. 2000).
Zoochoric fruits of different colors are found throughout the year in restinga of Maricá. Black, red and multicolor fruits, commonly associated with dispersion by birds (Willson & Whelan 1990;Valenta et al. 2018), were released in fruiting phenophases with higher levels of intensity and activity throughout the year, compared to fruits of the other colors. Gonzaga et al. (2000) found significant seasonal variation in composition of the avifauna in restinga of Maricá. Gomes (2006) concluded that constant supply of fruits is important for maintenance of the community of resident and visiting birds in restinga of Jurubatiba. The other colors (mainly yellow and green) are commonly associated with the dispersion of mammals (Willson & Whelan 1990;Valenta et al. 2018). Species with fruits of these colors, although they showed less intense fruiting phenophases and a little more restricted availability throughout the year, also constitute an ample supply of edible fruits. Our results points that vegetation of restinga of Maricá presents fruits very different in colors and these colors are well distributed throughout the seasons, meaning a constant and reliable supply of fruits to the dispersing fauna, especially birds and mammals.
Myrtaceae stands out in this study, since there is always a species of Myrtaceae producing fruits throughout the year. Likewise, in Maricá, Staggemeier et al. (2017) also highlighted the family for offering a wide morphological diversity of fruits to different frugivore guilds throughout the year, in an area of restinga vegetation in São Paulo state. According to Rodarte (2008), Myrtaceae is the family with highest species richness and highest value of phytosociological importance (37.2%) in restinga of Maricá. The fruit traits of Myrtaceae species (fleshy, indehiscent, different colors and sizes) highlights the importance of this family as a year round resource for local fauna, especially frugivorous fauna.
The main fruiting peak of anemochoric species occurred in the hottest and rainy season of the year, although no significant correlation with abiotic factors evaluated (precipitation, temperature and day length). The peak in the hot and rainy season displayed by species with wind dispersion owes mainly to the activity and intensity of Agarista revoluta, Pleroma gaudichaudianum and Marcetia taxifolia fruiting. It should be noted that the peak of fruiting of the anemochoric species in the warmest and rainiest season recorded in this study coincides with the blooming peak of species pollinated by the wind, in the same community. Myrsine parvifolia, Alchornea triplinervia and species of Cyperaceae and Poaceae also bloom in the period from October to March (Rodarte 2008;Albuquerque et al. 2013). Winds, low humidity and higher temperatures are constant in the restinga ecosystem, when compared to other habitats of Atlantic Forest (Scarano 2002), mainly in areas closest to the sea (Staggemeier & Morellato 2011). The species analyzed in the present study are frequent in more open physiognomies (open shrub), thus allowing the dispersal of anemochoric seeds throughout the seasons.
We highlight that restinga environments present many key species for the germination of other plants, such as bromeliads (nurse plants). Restingas have relatively low rate of free water available in the soil (Zaluar & Scarano 2000;Scarano 2002) and success in germination and development of seedlings in restinga has been demonstrated to be higher in the microhabitat inside of the bromeliad rather than in the sand around the plant (Fialho 1990;Zaluar 1997). Bromeliads can store a considerable amount of free water (Cogliatti-Carvalho et al. 2010), and thus represent a relatively stable microhabitat since tanks tend to remain with water even during periods of drought (Krügel & Richter 1995). In fact, seedlings of various species such as Monteverdia obtusifolia, Guapira opposita, Myrtaceae, Clusiaceae, among others, are easily found within bromeliads (personal observation). Studies involving recruitment throughout seasons and its relationship with nurse plants and seed dispersers are required in this ecosystem, with flora found in sandy, permeable and nutrient-poor soils.  Table 4 -Spearman correlation coefficients (p < 0.05), obtained between phenological phases (levels of intensity and activity of fruit set) and abiotic factors [maximum, mean, and minimum temperature ( o C), precipitation (mm), and day length measure (h)] in plant community distributed over the internal sand stream of the restinga of Maricá, Rio de Janeiro, Brazil.
Our study on the fruiting phenology and the characterization of the fruits of 52 species from an area of restinga vegetation shows a predominance of fruits with zoochoric characteristics, diverse in colors and sizes and available throughout the year for a potentially rich fauna of seed dispersers. In restinga, fruiting events are, however, less seasonal and intense than flowering events and seems to be not limited by climatic factors. Studies on fruiting phenology, dispersion and seeds germination must be incentivated as they can help incrementing successful restoration and conservation actions for restinga areas.