Phenodynamics of fi ve orchids species growing on rock outcrops in the Chapada Diamantina Mountains in northeastern Brazil

We evaluated the vegetative and reproductive phenological patterns of the orchids Cattleya elongata, Cyrtopodium aliciae, Epidendrum orchidifl orum, Epistephium lucidum, and Sobralia liliastrum (Orchidaceae) growing on sandstone outcrops in the Chapada Diamantina Mountains in northeastern Brazil. Phenological events were associated with abiotic factors, leaf longevity, life form, plant growth pattern, and pollination syndrome. Phenological observations were made for 18 months and followed the phenophases of: stem/pseudobulb emission, leaf fl ush, leaf fall, fl owering, immature fruit and mature fruit. Seasonality, synchrony, and correlations between phenophases and environmental variables were tested. Th e orchid species demonstrated aseasonal vegetative phenologies, except for pseudobulb emission. Reproductive events were seasonal, except for fl owering in E. orchidifl orum. Th ere was high fl owering overlap between the species pairs S. liliastrum and C. aliciae (dry season) and C. elongata and E. lucidum (rainy season). Dispersal occurred during both the rainy and dry seasons. Th e vegetative phenophases exhibited low synchrony, and were related to certain morphofunctional characters (stems/pseudobulbs, CAM metabolism). Th e reproductive phenophases showed high synchrony consistent with the deceit pollination strategy usually associated with the group.

Seasonality and synchrony in plant phenology are traditionally associated with environmental variables, although variations in leaf and reproductive rhythms among species occupying similar environments indicate that phenological responses may also be mediated by intrinsic species attributes such as life form, plant growth pattern, leaf longevity, and dispersal and pollination modes (Williams-Linera & Meave 2002).Th e Chapada Diamantina Mountains, the northern extension of the Espinhaço range in Brazil, exhibits a vegetation mosaic with campo rupestre (open, rocky field vegetation) predominating above 900 m a.s.l.(Funch et al. 2009).Orchidaceae is one of the principal groups found in that habitat (Conceição et al. 2007a), with approximately 300 species (Brito & Cribb 2005).This diversity is associated with morphofunctional characters such as CAM metabolism and the presence of pseudobulbs that retain water resources for metabolic activities and facilitate survival in otherwise unfavorable environments (Silveira et al. 2009;Arévalo et al. 2011).There has been a notable lack of studies addressing the phenologies of the Orchidaceae family in Brazil, and the available information is generally restricted to records of their flowering periods and floral biology (Pansarin et al. 2012;Pansarin & Pansarin 2014).Here, we evaluated the vegetative and reproductive phenological patterns of five Orchidaceae species (Cattleya elongata Barb.Rodr., Cyrtopodium aliciae L. Linden & Rolfe, Epidendrum orchidiflorum Salzm.Ex Linndl., Epistephium lucidum Cogn., and Sobralia liliastrum Salzm.Ex Lindl.) in an area of campo rupestre vegetation (12°34´28´´S x 41°24´31´´W; at 614 m a.s.l.) in the Chapada Diamantina Mountains, where those species grow abundantly on rocky outcrops.Considering the regional seasonality, with marked rainy and dry periods (Funch et al. 2002), the relatively low availability of water in rocky mountain top microhabitats (Conceição et al. 2007b), and well-known aspects of the structure and function of orchids that allow them to exploit seasonally dry habitats (Benzing et al. 1982), we hypothesized that their vegetative phenology would exhibit pseudobulb/stem emissions and leaf flush during the rainy season, associated with life form, plant growth pattern, and leaf longevity; as their reproductive dynamics would be aggregate adjusted to the seasonality, and mode of pollination and wind dispersal of their seeds (Borba et al. 2003;Smidt et al. 2006;Spaethe et al. 2007;Vale et al. 2011).
The regional climate is type Aw by the Köppen's climate classification (Alvares et al. 2014), marked by maximum rainfall in the Austral summer (November to April), and a dry winter (July to October).The temperature, rainfall and day lengths for the region were published by Souza et al. (2012).Campo rupestre is a complex vegetation type comprising numerous microhabitats on rocky mountain tops with water and soil nutrients restrictions and soil depths that severely limit plant establishment and growth (Conceição et al. 2007).The specimens of Orchidaceae examined here (n=89) were growing directly on rock surfaces (C.elongata n=24, C. aliciae n=17, E. orchidiflorum n=18, and S. liliastrum n=21) and in narrow cleft microhabitats with shallow soils contained by rock walls (E.lucidum n=9).
Phenological observations were made on a monthly basis from June/2009 to November/2010 on marked individuals.Individuals were considered to be distinct plants when separated from one another by approximately 1 m.The phenophases considered were: pseudobulb/stem emission, leaf flush, leaf fall, flowering, immature fruit and mature fruit.The phenophases were recorded on a presence-absence basis.Four individuals of each species were closely accompanied during the phenological observations to determine leaf longevity by sketching them on paper and thus recording and numbering existing leaves and all new leaves.We also analyzed life forms (Raunkiaer 1934), plant growth patterns, pollination syndromes (Faegri & Pijl 1979), and floral and inflorescence characteristics (such as color, size, the presence of nectar or oil) by observations in the field or by consulting the literature (Chingel 2001).The Shapiro and Wilk test was used to determine the normality of the distributions of the phenological and environmental data (Zar 1996).The Spearman correlation index was calculated for data that did not have normal distributions, and the Pearson correlation index was calculated for the data that showed normal distributions, using Past 2.1.7software.The seasonality and synchrony of the phenophases in each species and the overlap in flowering species pairs were evaluated using circular statistics (Morellato et al. 2010), run on Oriana software.
Four Orchidaceae species were classified as hemicryptophytes, having rhizomes flush to the ground/ rock surface; E. lucidum was classified as a cryptophyte, as its rhizomes grow below the soil surface in narrow rock-cleft microhabitats.The orchid species all showed sympodial growth pattern, with pseudobulbs (C.Reproductive events were markedly seasonal and synchronous (Tab.1).There was high flowering overlap in certain species pairs, with S. liliastrum and C. aliciae flowering during the dry season (r = 0.83), and C. elongata and E. lucidum flowering during the rainy season (r = 0.87); the only species that demonstrated aseasonal flowering was E. orchidiflorum (Fig. 1).S. liliastrum and E. orchidiflorum produced fruits during the dry season that matured during the rainy season; C. elongata, C. aliciae, and E. lucidum initiated fruit production during the rainy season, with maturation occurring during the dry season (Fig. 1).In general, the species demonstrated correlation of their flowering and fruiting with photoperiod and temperature (Tab.2).Melittophily predominated, except in E. orchidiflorum, which was identified as sphingophilous.The species produced apical racemes, with up to 15 flowers,  except C. aliciae which produces panicles with up to 300 flowers; C. elongata, E. lucidum, and S. liliastrum produce large flowers (5-8 cm diam.) and C. aliciae and E. orchidiflorum produce small flowers (1.5-3.5 cm diam.).C. elongata and E. lucidum produced white flowers while C. aliciae and S. liliastrum produce rose-colored blooms, and E. orchidiflorum greenish flowers (Fig. 1).Only E. orchidiflorum and E. lucidum offered nectar rewards to pollinators.Precipitation has been recognized as the principal factor influencing plant phenological patterns in tropical environments (Brearley et al. 2007).The absence of any correlations between the phenophases and precipitation among the orchid species evaluated in the present study was probably related to the fact that they all use water resources very efficiently and demonstrate CAM metabolism (Silveira et al. 2009).The pseudobulbs/stems store water for metabolic activities and facilitate their survival in otherwise unfavorable environments (Silva & Milaneze-Gutierre 2004); the seasonality and synchrony observed for pseudobulb emission with few apical leaves can be understood in terms Karoline Coutinho de Santana, Isys Mascarenhas Souza, Lia d'Afonsêca Pedreira de Miranda and Ligia Silveira Funch of their construction costs, in contrast to aseasonal stem emissions with continuous leaf production and leaf fall.

Phenodynamics of five orchids species growing on rock outcrops in the Chapada Diamantina Mountains in northeastern Brazil
The species studied demonstrated reproductive phenophases correlated with seasonal changes in photoperiod and temperature -the most predictable environmental parameters in this region.Photoperiodicity has been found to be associated with breaking and/or induction of bud dormancy, making it the trigger for growth and flowering initiation in many plant species (Stevenson et al. 2008).The high observed overlap of flowering between pairs of species with similar flower colors during the dry and rainy seasons (including E. lucidum which produces nectar) appears to be a viable reproductive strategy as most orchids mimic melittophilous flowers and, as they offer no floral rewards, are pollinated using deceit strategies (Pansarin & Pansarin 2014).The flowers of species utilizing deceit strategies will be visited less frequently than the flowers of species that offer some type of reward (as insects have the capacity to learn) -and will consequently show less fruiting (Borba & Semir 2001); as such, the high synchrony and overlapping between them presumably favors pollination processes by accentuating floral displays.E. orchidiflorum was the only species showing aseasonal flowering, which appears to reflect nectar production by their flowers (with characteristics of sphingophily).
A low fruiting rate is considered an adaptation among orchids, especially among epiphytic and rupiculous species, with limited access to resources; their fruits generally result from cross-pollination events and their seeds show high genetic viability -so that the production of even small numbers of fruits will allow the maintenance of viable populations (Ackerman & Zimmerman 1994).The maturation of dry fruits is favored by dry periods with severe water restrictions, and the low relative humidity of the air contributes to their desiccation and to the dispersal of anemochoric seeds (Pezzini et al. 2008).As Orchidaceae seeds are very small and essentially without nutritional reserves their dispersal during dry periods might be expected to hinder germination and establishment, but they demonstrate dormancy due to the presence of lipidic reserves on their surfaces (which will also provide for their initial sustenance) (Baskin & Baskin 2001), allowing dispersal in both the dry and rainy seasons.
The low intraspecific synchrony and aseasonality demonstrated by the vegetative phenophases of the orchid species examined in the present study appeared to be a reflection of the influences of the morphofunctional characters (pseudobulbs/stems and CAM metabolism), and the high synchrony and seasonality of their reproductive events is likewise related to the pollination mechanisms of the group, with overlapping of flowering among species with similar flower colors and the production of nectar by at least one of them.
elongata and C. aliciae) or stems (E.orchidiflorum, S. liliastrum, and E. lucidum) united by rhizomes; stems were numerous in E. orchidiflorum, S. liliastrum, while E. lucidum produced only a single stem.Plants with stems showed numerous leaves along the stem, with continuous leaf flush and leaf fall.Plants with pseudobulbs produced just a few apical leaves; the pseudobulbs and leaves of C. aliciae persisted for up 12 months; the leaves of C. elongata remained attached to the older pseudobulbs for periods exceeding 18 months.The orchids examined demonstrated aseasonal and asynchronous vegetative phenology during the study period, except for pseudobulb emission in C. elongata and C. aliciae and leaf fall in C. aliciae, which were seasonal and synchronous (Fig. 1, Tab. 1).

Figure 1 .
Figure 1.Habit, flower and phenology of orchids species from Chapada Diamantina, Northeast Brazil.The shaded area indicates the rainy season.Continuous lines with black circles are pseudobulbo; continuous lines with green square are leaf flushing; continuous lines with red triangles are leaf fall; dotted lines with stars are flower continuous lines with yellow square are immature fruit; continuous lines with blue rhombus are mature fruit.

Table 2 .
Correlation between phenophases and environmental variables of five species of Orchidaceae, Chapada Diamantina, Bahia, Northeast, Brazil.