Floral and reproductive biology of Alcantarea nahoumii ( Bromeliaceae ) , a vulnerable endemic species of the Atlantic Forest

Alcantarea nahoumii occurs exclusively in the state of Bahia, Brazil, and is classifi ed as vulnerable due to deforestation and frequent fi res in the region. Knowledge of fl oral and reproductive biology is fundamental to understanding ecological interactions, as well as the reproductive success of plant species. Th e objective of this study was to evaluate the fl oral and reproductive biology of A. nahoumii in an Atlantic Forest fragment with regard to phenology, pollen viability, stigma receptivity, pollination ecology and reproductive systems, all of which are important parameters for of the development of conservation strategies for the species. Anthesis is diurnal and heterogeneous, starting at 6:30 a.m. and lasting until 8:00 a.m. Highest germination percentages and greatest pollen tube lengths were obtained in BK culture medium. Histochemical tests revealed high pollen viability (89.71 %). Stigma receptivity occurred during anthesis and lasted for up to 24 hours after fl oral opening. Alcantarea nahoumii exhibited preferential allogamy and self-compatibility, and required a pollinator to production of viable seeds. Sixteen species of pollinators were observed visiting A. nahoumii, among which were fi ve hummingbird species. Even though its reproductive system is effi cient, this bromeliad remains threatened mainly due to habitat fragmentation caused by deforestation, burning and predatory extractivism.


Introduction
Th e Brazilian Atlantic Forest is one of the main centers of diversity and endemism of the family Bromeliaceae, with a total of 653 species recorded in the last decade (Forzza et al. 2013).Fifty-four of these species are classifi ed as critically endangered, 89 as endangered, 182 as vulnerable, 17 as rare and three as extinct in nature (Forzza et al. 2013); however, these numbers are probably underestimates due to the lack of knowledge regarding the actual status of wild populations of species of Bromeliaceae.Th e genus Alcantarea is native to Brazil and has approximately 40 species (Versieux & Wanderley 2015).Th e plants of this genus show various adaptations that enable them to grow in habitats with water stress and high luminosity, such as on granite boulders and other outcrops of the Serra do Espinhaço in the states of Minas Gerais and Bahia (Versieux et al. 2012;Versieux & Wanderley 2015).Alcantarea nahoumii occurs exclusively in Bahia and is classifi ed as vulnerable (Forzza et al. 2013).It is estimated that in recent years the population has declined by 30 % due to deforestation and frequent fi res (Versieux & Wanderley 2007;2015;Forzza et al. 2013).
1 Universidade Federal do Recôncavo da Bahia, Rua Rui Barbosa, s/n, Campos Universitário, 44380-000, Cruz das Almas, BA, Brazil 2 Embrapa Mandioca e Fruticultura, Rua Embrapa, s/n, Chapadinha, 44380-000, Cruz das Almas, BA Brazil 3 Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, 70040-031, Brasília, DF, Brazil Maria Josirene Souza Moreira Bastos, Lucimário Pereira Bastos, Everton Hilo de Souza, Taliane Leila Soares, Daniel Vieira Morais, Fernanda Vidigal Duarte de Souza and Maria Angélica Pereira de Carvalho Costa Because of over-exploitation and ongoing destruction of the Atlantic Forest, combined with the growing interest of traders in ornamental species, bromeliads are at serious risk of disappearing, making it necessary to develop conservation policies (Zanella et al. 2012).However, the adoption of conservation strategies requires basic knowledge for the development of correct management practices, in situ or ex situ.
Knowledge of floral and reproductive biology is fundamental for understanding the interactions among pollen grains, stigma, flowers and pollinators, as well as the reproductive success of plant species through survival strategies and preservation mechanisms (Negrelle & Muraro 2006).Likewise, understanding phenology is important and has been of increasing significance over the past decade due to its relevant role in the management and conservation of native plants (Elmendorf et al. 2016).Aspects of the phenology of Bromeliaceae have not been widely studied considering the large number of species.Understanding phenology can shed light on the life cycle of plants, their phases in time, the resources within communities and plantanimal interactions (Tschapka 2004).However, phenological events of plants can be interrupted due to climatic and landscape changes, which together with other anthropic effects can pose threats to local populations (Dai et al. 2013).
The family Bromeliaceae exhibits extensive variation in reproductive systems (Benzing 2000).Allogamy and self-compatibility are present in the majority of species (Wendt et al. 2001;2008;Matallana et al. 2010;2016), but there are also records of agamospermy, autogamy, and selfincompatibility (Martinelli 1994;Scrok & Varassin 2011;Paggi et al. 2013;Souza et al. 2017).Studies of the floral and reproductive biology of A. nahoumii are important for formulating management and conservation programs, especially since this species is subjected to pressures from habitat fragmentation, including shortages of specific pollinators and reduced population sizes, which can lead to pollen limitation.
The objective of this study was to investigate the floral and reproductive biology of Alcantarea nahoumii in an Atlantic Forest fragment, and acquire novel data related to reproduction forms and pollination mechanisms of this endemic and vulnerable species.

Species and study area
Plants from a natural population of Alcantarea nahoumii (Leme) J. R. Grant, located in Serra da Jibóia in the municipality of Santa Teresinha, Bahia,Brazil (12°51'08.19''S 39°28'34.32''W),were used to analyze phenology, reproductive systems and different aspects of floral and pollination biology.This mountain range encompasses approximately 59 km 2 , with a maximum altitude of 850 m and a climate varying from moist tropical to subtropical.A dried voucher specimen was deposited in the HURB herbarium (voucher 9416) (Universidade Federal do Recôncavo da Bahia, UFRB).

Phenology
Monthly observations were conducted between January 2012 and December 2015, for a total of 48 months.Climate data (minimum, maximum and mean temperature and rainfall) were obtained from the National Meteorology Institute (INMET 2016).To determine the timing of floral opening and the months of blooming, 40 randomly chosen individuals were analyzed within the population.Peak blooming was defined as when 50 % of the individuals sampled had flowers.The number of flowers opened per day and the flower availability period (interval when the flower remains open and available to visitors, from anthesis to senescence) were recorded during the entire flowering period.To describe flowering phenology, the number of flowers opened each day and the period of anthesis and of fruit shedding were quantified.Phenological study followed the recommendations of Machado & Semir (2006) and Marques & Lemes Filho (2008).
The experimental design used to study pollen germination was completely randomized in a (4 x 3) factorial scheme, with four culture media (BM, BK, MBK and SM) and three times (8 a.m., 12 noon and 6 p.m.), with 12 repetitions (each repetition being a Petri dish).
The number of germinated pollen grains was counted and the length of the pollen tubes was measured 24 hours after inoculation in the culture medium, as described by Soares et al. (2008).Photographs were obtained with a Leica EZ4 D stereomicroscope (Leica, Wetzlar, Germany).
Percentage data were transformed to arcsin (√x/100) prior to statistical analysis.To compare means, the data were submitted to analysis of variance and the Tukey test (p<0.01),with the SAS program (SAS Institute 2010).
Floral and reproductive biology of Alcantarea nahoumii (Bromeliaceae), a vulnerable endemic species of the Atlantic Forest The stained pollen grains were observed by random sampling using the slide scanning method with a Carl Zeiss 77081 optical microscope (Jena, Germany), to count 100 grains per slide, with three repetitions per treatment, for a total of 300 grains for each stain.

Reproductive systems
Reproductive systems were investigated in the field using the method described by Kearns & Inouye (1993).The treatments included: a) open pollination -floral buds were only selected and identified; b) manual selfpollination -flowers were protected during pre-anthesis and at anthesis were pollinated with grains from the same plant; c) spontaneous self-pollination -flowers during pre-anthesis were protected and remained that way until fruiting; d) manual cross-pollination -flowers were protected during pre-anthesis and anthesis and were pollinated with grains from other plants; and e) agamospermy -flowers were emasculated (anthers removed) and protected until fruiting.In treatments (b-e), flowers were protected from floral visitors with voile fabric bags (15 x 10 cm) of about 5 cm mesh, from pre-anthesis until fruiting.
Pollination was performed at 8:00 a.m.(anthesis) under the same conditions of temperature (22 ± 3 o C) and relative humidity (88 %).Different numbers of flowers were used for each reproductive treatment, according to their availability (Tab.1).Fruiting was monitored weekly until ripening, and fruits were collected before falling.The following parameters were determined for each treatment: fruiting percentage, length (cm), diameter (cm), number of seeds, and in vitro germination rate (%).
The autogamy index (AI) and self-incompatibility index (SII) were calculated according to Ramirez & Brito (1990).A species is considered autogamous or partially autogamous when the AI is greater than 0.30, and non-autogamous when it is lower than 0.30.In turn, a species is considered compatible or partially compatible with a SII greater than 0.30, while self-incompatible with a value below a 0.30.
The number of pollen grains produced per flower was estimated according to Kearns & Inouye (1993), in three flowers.The number of ovules was counted in each locule of the ovary with a Leica EZ4 D stereomicroscope (Wetzlar, Germany), using three different flowers from each individual.The pollen/ovule ratio was determined according to Cruden (1977).

In vitro germination of seeds
Seeds resulting from the different reproductive treatments were removed from ripe fruits and disinfested in a solution of sodium hypochlorite (2 % active chlorine) and distilled water (2:1) for 20 minutes, followed by washing three times in distilled water and sterilization in a laminar flow chamber.The seeds were distributed in Petri dishes containing 25 mL of MS culture medium (Murashige & Skoog 1962) supplemented with 3 % sucrose and solidified with 2 g L -1 of Phytagel ® , with the pH adjusted to 5.8.The seeds were kept in a growth room at a temperature of 25 ± 2 ºC, and with luminosity of 22 mmol m -2 s -1 and a 16hour photoperiod.The experimental design was completely randomized, with 10 repetitions of groups of eight seeds.The germination percentage was determined for each reproductive treatment.

Nectar production
Floral buds were protected on the day before collection and the nectar produced was collected hourly from 8:00 a.m. to 6:00 p.m.The volume of nectar produced was quantified  Maria Josirene Souza Moreira Bastos, Lucimário Pereira Bastos, Everton Hilo de Souza, Taliane Leila Soares, Daniel Vieira Morais, Fernanda Vidigal Duarte de Souza and Maria Angélica Pereira de Carvalho Costa by introducing a needle attached to a graduated microsyringe (Hamilton 50 µL) through the corolla and into the floral nectary of each flower in the direction toward the base of the petals.After quantification, the nectar was placed in a Reichert digital refractometer and the refraction index was converted into sugar concentration (sucrose equivalent) according to Bolten et al. (1979).To transform the data into Kcal, the total volume of nectar per flower (mg) was multiplied by the factor 4.0 (Galetto & Bernardello 2005).
The experimental design was completely randomized and composed of 20 flowers from different individuals.Pearson correlation coefficients were calculated for the traits and their significance was measured by the t-test at 1 % probability, using Genes 7.0 software (Cruz 2008).

Floral visitors
The floral visitors were observed with a digital infrared camera directly in the field during two time periods (five consecutive days in October and eight alternating days in November), from 6:00 a.m. to 6:00 p.m. and from 6:00 p.m. to 6:00 a.m., for a total of 156 daytime hours and 96 nighttime hours.The following parameters were observed: behavior, floral resource, time of day, visit duration and relative frequency of each floral visit per day.The relative frequency of visits of each species was calculated as the number of visits performed by the species in relation to the total number of visits recorded, and denoted as a percentage.Visitors were photographed with the digital camera from three angles and invertebrates were collected in an insect net and sacrificed in a bottle containing ethyl acetate.The dead invertebrate specimens were sent for identification by the Insect Group of Universidade Federal do Recôncavo da Bahia.All occurrences of agonism between hummingbirds and flowers were recorded, and the birds were identified from the photos by a specialist from the Zoology Museum of Universidade Estadual de Campinas.

Phenology
Flowering occurred from 2013 to 2015, with peak blooming in October through December, and the number of flowered plants in 2013 was lower than in 2014 (Fig. 1).Only two plants bloomed in 2012, in December.This low flowering rate is related to the low rainfall that year (714 mm) and a fire in March that affected most of the experimental area (Figs. 1, 2A).
No flowering was observed for plants in any of the four years during the period from February to July (Fig. 2B).The first bracts from the floral scape were observed in August through mid-September, at the end of the rainy season, with a change in coloration from green to yellow-green (Fig. 2C).The time between emission of the floral scape and the first flower was on average 61 ± 9 days, with a further 87 ± 11 days before opening of the last flower at the apex of the inflorescence, a period covering spring and summer, from October through March (Fig. 1).
Each inflorescence contained an average of 150.76 ± 64.98 buds, with a range of 34 to 312, of which only 37.50 % produced flowers.Blooming occurred heterogeneously, with an average of 0.64 flowers per plant/day and varying from zero to six flowers from the start to end of blooming.
Anthesis was observed to be diurnal and not homogeneous, starting at 6:30 a.m. and lasting until 8:00 a.m., when the thin yellow petals were fully open and curled downward (Fig. 2C -detail).The flowers remained open 24 hours after anthesis, although exhibiting clear signs of necrosis of the petals, pistil and style, and remained adhered to the inflorescence regardless of the occurrence of pollination.
Fruiting took place from February to April and the fruits fell 178 ± 18 days after pollination, in the period from May to June, with opening of the capsules and dispersion of some of the seeds by wind (Fig. 2D).Some seeds remained adhered to the fruits and germinated in the capsules and inflorescences (Fig. 2D -detail).The seeds dispersed by the wind to form new individuals stood apart from the others by having a feathery appendage, a typical characteristic of the subfamily Tillandsioideae.
After flowering and senescence (February to March), plants formed five sprouts (lateral shoots), which started a new reproductive cycle two years later.Therefore, the reproductive, flowering, fruiting and dispersion period of this species was relatively long (approximately seven months), with reproductive structures forming in the second half of the year and the dispersion period ending in the middle of the next year.

Pollen viability
Alcantarea nahoumii exhibited high pollen viability regardless of the type of test (in vitro germination or histochemistry).We observed a significant effect of the interaction between the factors of culture media and collection time for in vitro germination percentage of pollen grains and pollen tube length.The BK culture medium was associated with the highest germination percentages during anthesis, with 95.41 % of the grains germinating, while the MBK medium was the least favorable for germination (82.00 %) (Tab.2, Fig. 2E).However, in all media the highest pollen viability occurred during the 8 hours of anthesis.
With respect to the histochemical tests, the highest pollen viability percentage was obtained with 1 % Alexander's stain, with 94 % viability, followed by 1 % TTC with 91 % (Fig. 2G-H), 1 % Lugol's iodine with 90.40 % and 1 % acetic carmine with 83.46 % (Fig. 2I-J).Also, Alexander's stain and TTC produced the best distinction, by the difference in Floral and reproductive biology of Alcantarea nahoumii (Bromeliaceae), a vulnerable endemic species of the Atlantic Forest coloration, between the viable and nonviable pollen grains, in comparison with the other two stains (Fig. 2).

Stigma receptivity
Stigma receptivity started in the beginning of anthesis and continued until 24 hours afterward (Tab.3).No preanthesis receptivity was observed by either of the two methods applied.Both methods were efficient at evaluating stigma receptivity of A. nahoumii, producing similar results (Tab.3).The use of 3 % hydrogen peroxide caused the formation of air bubbles in the stigma area, indicating very strong peroxidase enzyme activity.The solution of α-naphthyl acetate, phosphate buffer, acetone and fast blue B salt conferred a dark brown color when the stigma was receptive.
versus 82 % germination for open pollination and 75 % for manual cross-pollination (Tab.1).Alcantarea nahoumii presented an autogamy index of zero and a self-incompatibility index of 0.44.These results indicate this species is preferably allogamous, selfcompatible and requires a pollinator to maintain production of viable seeds.
For A. nahoumii, the number of pollen grains per flower in this experiment was 396,300 on six anthers, and the number of ovules per flower was 551 ± 39, or a pollen/ ovule ratio of 719.24/1.Means followed by the same lower-case letters in the column and upper-case in the rows, within the same factor, do not differ from each other by the Tukey test at 1 % probability.* Culture media: BM (Parton et al. 2002), BK (Brewbaker & Kwack 1963), MBK (Brewbaker & Kwack 1963, modified), SM (Soares et al. 2008).

Floral and reproductive biology of Alcantarea nahoumii (Bromeliaceae), a vulnerable endemic species of the Atlantic Forest
Table 3. Stigma receptivity of Alcantarea nahoumii (Leme) J. R. Grant (Bromeliaceae) evaluated between pre-anthesis and postanthesis, by two methods.
High correlations were observed between the presence of floral visitors (hummingbirds, bees, other birds, butterflies, wasps, beetles and ants) and sugar concentration (r= 0.71**).With respect to time of collection, there was a significant positive correlation with sugar concentration (r=0.77**).No significant correlations were observed between collection hour and nectar volume (r= -0.048 ns ).
Visits by hummingbirds started at 9:00 a.m., extended until 4:00 p.m., and were more frequent in the morning hours, coinciding with the greatest nectar availability, with intervals of 10 to 20 minutes between visits and an average of 15 visits per plant.
Visits were fewer and spaced at longer intervals in the afternoon, (from 30 to 40 minutes, with an average of three visits per inflorescence).The hummingbirds Chrysolampis mosquitus (Fig. 4A), Thalurania glaucopis (Fig. 4B) and Florisuga fusca (Fig. 4C) remained for three to six seconds at each flower and then moved to other inflorescences or rested on nearby stems, without leaving the area, waiting to make new visits.The species Chrysolampis mosquitus and Thalurania glaucopis were the only visitors that exhibited agonistic interactions, and were the only species to visit flowers in the afternoon.Females of Anthracothorax nigricollis (Fig. 4D) were observed with less frequency (11.13 %) than the other hummingbird species, with the exception of Eupetomena macroura (Fig. 4E), which exhibited isolated behavior, with frequency of 3.17 %, only landing on plants near wooded areas at the edges of the study areas.Of the five hummingbird species observed in this study, all collected nectar by introducing their long beaks between the petals toward the base of the corolla, where the nectar chamber is located.
Coereba flaveola (bird) (Fig. 4F) paid only one illegitimate visit, collecting exposed pollen and nectar, so this species can be classified as a robber.The bees Trigona spinipes and Apis melifera (Fig. 4G) visited the flowers individually and remained on the flowers for long periods.In contrast, Oxytrigona tataira (Fig. 4H) visited the flowers in groups of 8 to 10 individuals, and only in the mornings, most intensely between 6:30 to 11:30 a.m.All these bee species were present throughout the blooming period and collected nectar.In turn, the bee Augochlorella sp. was found only on the anthers collecting pollen, and remained for an average of 10 seconds.The butterfly Ascia manuste (Fig. 4I) landed on flowers only to collect nectar and remained for only 8 seconds on average, with a relative frequency of 2.65 %.
In addition to the species mentioned sthuso far, we also observed one species of the order Coleoptera (Diabrotica septenliturata) and four of Hymenoptera: Polybia sp, Polistes sp, Polybia platycephata and Formicidae (not identified to species level).None of these visitors showed any interest in floral resources and mainly visited in the morning hours, except for the ants, which were present throughout the day.

Phenology
Alcantarea nahoumii displays a two-year cycle after sprouting of lateral shoots and an additional seven months from flowering to seed generation.The long reproductive period appears to be a characteristic of the subfamily Tillandsioideae, as previously reported by Marques & Lemos Filho (2008).According to these authors, in general,  bromeliads that flower in the rainy season inhabit rocky regions at high elevations (1,400 m), because under these conditions their entire vegetative and reproductive cycle is optimized by the morphological traits of the bromeliads, such as tanks (phytotelmata), absorptive roots and scaly/ peltate trichomes.In contrast, in species that develop at elevations lower than 1,300 m, as is the case of A. nahoumii (at 850 m), flowering happens in the dry season, which coincides with a greater availability of pollinators, thus avoiding the fog and lower temperatures of higher elevations.The diurnal floral anthesis and prolonged open period of the flowers (24 hours) increased visitation by pollinators and the possibility of fertilization.Knowledge of floral phenology allows the establishment of the moment, duration and intensity of a species' flowering, as well as the influence of these factors on the behavior of pollinators (Aker 1982;Tschapka 2004).However, the timing of these events is subject to the influence of environmental perturbations and landscape alterations.

Pollen viability
In this study, A. nahoumii exhibited high pollen viability regardless of the type of test (in vitro germination or histochemistry).In vitro germination does not completely reproduce the same pollen tube growth as in vivo fertilization, but the results obtained are near those that occur in vivo, making it important to perform this technique carefully for each species (Soares et al. 2008;Souza et al. 2015).In vitro germination is the result of interactions between the composition of the culture medium and the peculiarities of each species or genotype.For this reason, it is necessary to adjust conditions to find the best combination for each type of material.
According to Souza et al. (2002), pollen viability values above 70 % are considered high and 30 % or below as low.They also stated that pollen viability is considered to be a measure of masculine fertility, and can be assessed by  (Souza et al. 2015).Based on this comment, and according to the results obtained in this study, the pollen grains of A. nahoumii exhibited high viability and can be used in controlled pollination.Among the histochemical methods studied, staining with Alexander's solution produced results closest to those obtained by pollen viability estimation through in vitro germination.Similar behavior has been observed for other plant species, such as Arabidopsis thaliana, which exhibited pollen viability greater than 90 % when using Alexander's stain, results very close to those obtained with in vitro pollen germination (Swanson et al. 2016).

Stigma receptivity
Another important aspect for understanding reproductive biology is stigma receptivity.Normally stigma receptivity is associated with floral opening and directly influences the reproductive system, because it is at this moment that the pollen grains should be deposited on the stigma to assure fecundation (Souza et al. 2016).Thus, for A. nahoumii under the conditions studied, the period from 8:00 a.m. to 6:00 p.m. is the best for performing fertilization, since the pollen grains exhibited the highest viability rates and the stigma was receptive.
These results corroborate those reported by Souza et al. (2016) who, when evaluating stigma receptivity of A. nahoumii in a controlled environment (greenhouse), also observed that receptivity occurred during anthesis and continued for 24 hours post-anthesis.They also suggested that controlled crosses should be carried out during this interval to attain the highest probability of fecundation and seed production.

Reproductive systems
The in vivo pollination tests of A. nahoumii revealed greater fruit formation from controlled pollination.This can be partly attributed to variation in the number of pollen grains deposited on the stigma.In manual pollination, the number of grains deposited on stigmas is not only more uniform among flowers, but also likely greater than those deposited by insect visitors in the case of open pollination.This assumption is supported by other studies showing that the number of seeds and quality of fruits are closely related to the number of pollen grains deposited on stigmas (Park et al. 2009;Wetzstein et al. 2013).
Open pollination led to the formation of fruits, but at only 35% the percentage was low.This result can be attributed to the observed low frequency of floral visits.The fruiting rate was low (10%) for flowers submitted to manual self-pollination, and no fruits were formed in the spontaneous self-pollination and agamospermy treatments.This absence, or low percentage, of fruiting might have been the result of some degree of incompatibility or inbreeding depression of A. nahoumii.
Inbreeding depression can occur with manual selfpollination, as reported by Vaughton et al. (2010) for the species Cyrtanthus breviflorus and by Paggi et al. (2013) for Vriesea friburgensis, although no studies have been carried out to evaluate the endogamic effects on the progeny of A. nahoumii.On the other hand, cross-pollination and/or open pollination promote gene flow between individuals, which can increase the genetic variability within and between populations, both of which are important features for diversification and maintenance of species (Vosgueritchian & Buzato 2006).
With respect to the number of seeds produced, plants submitted to cross-pollination produced the greatest number of seeds, followed by open pollination.The success of open pollination can probably be attributed to greater gene flow between plants due to the action of pollinators, which can travel long distances carrying pollen grains from one plant to another (Barbará et al. 2009;Matallana et al. 2010).
Another factor that can affect the number of seeds is pollination efficiency, determined by the number of pollen grains deposited on the stigma and the number of ovules per flower (Mione & Anderson 1992).The pollen/ovule ratio observed for A. nahoumii was in line with the classification of allogamous species described by Cruden (1977).

Nectar production and floral visitors
The lack of a correlation between nectar volume and sugar concentration can be associated with differences in evaporation and nectar secretion rates, since the concentration of sugars in nectar depends on the rate of pollinator visitation and environmental conditions (solar radiation, rain, wind and temperature) (Kajobe 2007).These results corroborate those reported by Aguilar-Rodríguez et al. (2014), who observed a non-significant correlation between volume and sugar concentration for the nectar of the bromeliad Tillandsia macropetala.On the other hand, Aguilar-Rodríguez et al. (2016) observed a negative and significant correlation between collection hour and volume produced, as well as between collection hour and sugar concentration.
Variation in nectar volume and sugar concentration is common among some species, such as Aechmea pectinata, A. nudicaulis, A. organensis, A. ornata, Billbergia amoena, Nidularium innocentii, N. Rubens, Tillandsia geminiflora, T. stricta, T. tenuifolia, Vriesea altodaserrae, V. carinata, V. flammea, V. incurvata, and V. philippocoburgii.Canela & Sazima (2003) and Machado & Semir (2006) stated that for these species this variation may be related to the frequency of floral visits: flowers with greater nectar volume and sugar concentration attract more pollinators.On the other hand, Lenzi et al. (2006) and Scrok & Varassin (2011), studying Floral and reproductive biology of Alcantarea nahoumii (Bromeliaceae), a vulnerable endemic species of the Atlantic Forest A. lindenii and A. distichantha, respectively, observed that nectar volume and sugar concentration remained constant during the day.Krömer et al. (2008) studied the sugar composition and concentration of nectar in Bromeliaceae, covering 111 species belonging to all three subfamilies, and concluded that the characteristics of the nectar of Bromeliaceae are predominantly determined by putative adaptations of nectar sugars to preferences of the pollinators rather than by phylogeny.
In the case of bromeliads, nectar is the floral resource that most attracts hummingbirds.It is produced in septal nectaries and stored at the base of the corolla (Sajo et al. 2004).This induces hummingbirds to penetrate the flower with their beaks (generally long) while collecting nectar, during which they perform pollination by causing contact of the anthers with the receptive stigma.
Due to the lack of other resources, some bromeliads that grow in rocky fields, such as Hohenbergia ramageana, A. bromeliifolia, N. bahiana, O. albopictum, and Dyckia dissitiflora, can provide a significant portion of the diet of hummingbirds, with percentages varying from about 30 to 70 %, being influenced by various factors such as competition, resource availability and seasonality (Machado et al. 2007).
The nectar robbing behavior observed for Coereba flaveola has been reported in relation to a variety of plant species (Navarro 2000;2001), including some Bromeliaceae (Fumero-Cabán & Meléndez-Ackerman 2007;Santana & Machado 2010;Queiroz et al. 2016).Studying floral visitors of Pitcairnia angustifolia, Fumero-Cabán & Meléndez-Ackerman (2007) reported that C. flaveola was the most frequent species and acted as a primary nectar robber, including piercing the base of the corolla of some flowers, yet they performed pollination incidentally by depositing pollen grains on stigmas.
The results of this study increase knowledge regarding the forms of reproduction and pollination mechanisms of A. nahoumii, which is considered endemic and vulnerable in the regions studied.This information can be used in future studies related to programs to preserve genetic resources, since the species is preferentially allogamous and self-compatible and floral visitors are available in the region.Urgent measures are necessary for the conservation and restoration of the entire area to assure that this species habitat is maintained.
number of fruits /number of flowers produced are in parentheses.(AI) Autogamy Index; (SII) Self-Incompatibility Index.The means with different letters in the column differ significantly by the Tukey test at 1 % probability.

Figure 1 .
Figure 1.Climate data in the region of Santa Teresinha, Bahia (A) and flowering frequency of Alcantarea nahoumii plants (B) from Serra da Jibóia 2012 to 2015.

Table 1 .
Number, length and diameter of fruits, seeds produced and seed germination percentage of Alcantarea nahoumii (Leme) J. R. Grant (Bromeliaceae) in function of different reproductive treatments.

Table 2 .
In vitro germination percentage of pollen grains and pollen tube length (mm) of Alcantarea nahoumii (Leme) J. R. Grant (Bromeliaceae) in different culture media and collection hours.

Maria Josirene Souza Moreira Bastos, Lucimário Pereira Bastos, Everton Hilo de Souza, Taliane Leila Soares, Daniel Vieira Morais, Fernanda Vidigal Duarte de Souza and Maria Angélica Pereira de Carvalho Costa different
techniques, such as in vitro or in vivo germination, or by indirect methods based on cytological parameters such as coloration, as discussed above