Floristics and Reproductive Phenology of Trees and Bushes in Central West Brazil

Environmental conditions such as temperature, soil, photoperiodic factors and precipitation can determine the physical environment favoring the occurrence of given species and interfere with the reproductive period of plants. This work involved monthly excursions to a slope in the Serra de Maracaju between December 2009 and November 2010. Samples were collected in eight 25 x 50 m plots, transverse to the slope, totaling 1 ha sampled. Floristic richness included 79 species distributed into 58 genera and 31 families. Floristic Similarity Analysis and Principal Component Analysis showed the occurrence of two plant formations on the slope: seasonal deciduous forest and cerradão on the eutrophic soil of the slope. Flowering and fruiting occured in the rainy season; correlation with rainfall, mean temperature and photoperiod were negative for flowering and positive for fructification. Trees and bushes presented uniform flowering and a moderately seasonal fructification, which is unexpected in seasonal environments. We discuss the implications of the patterns found and make comparisons with others studies.


INTRODUCTION
Environmental conditions such as temperature, soil, photoperiodic factors and precipitation can determine the physical environment favoring the occurrence of given species and interfere with the reproductive period of plants (Towsend et al. 2006).In tropical and subtropical regions, phenological events are mainly associated with rain (Frankie et al. 1974) and soil moisture conditions (Morellato et al. 2000).In temperate areas, phenophases depend more on length and temperature of the day (Opler et al. 1976).
The Cerrado, a Savanna-like formation, is the second most important Brazilian morphoclimatic zone (Rodrigues 2005), gathering different natural phytophysiognomies (Eiten 1982), according to local edaphic, topographic and climatic factors (Oliveira and Marquis 2002).Its significant species richness may be surrounded by different vegetation types such as Amazonian and Atlantic forests, thorn forest (Chaco) and caatinga, subtropical fields, and wet and dry areas (Simon et al. 2009).
The fact that dry forests are interspersed with Cerrado in great tracts of land throughout South America (Mendonça et al. 1998) can be explained http://dx.doi.org/10.1590/0001-3765201420130042785-799 VIVIAN A. ASSUNÇÃO, JOSÉ C. CASAGRANDE and ÂNGELA L.B. SARTORI by climatological events that go back to the typically dry Quaternary period, which would have enabled the expansion of these forests and of more opened plant formations as fields, Cerrado and caatinga.Areas with dry forests differ in that they have more fertile soils, a higher pH and a concentration of certain nutrients (Mendonça et al. 1998).
Cerrado and forest remnants are often interspersed on Brazilian Serra slopes in the states of Minas Gerais (Oliveira-Filho et al. 1994), Goiás (Felfili et al. 2007) and Mato Grosso do Sul (Salis et al. 2004, G.A. Damasceno-Júnior, unpublished data, B.G. Fina, unpublished data, Ramos andSartori 2013).More recent rugged landscapes are usually richer in given soil elements, which can favor the settlement of different, albeit geographically close plants (Reatto et al. 2008).Phenological studies at a community level can help functionally classify vegetation (Vasconcelos et al. 2010) and are crucial to understand plant reproduction and ecology (Oliveira 2008).
Studies in Venezuelan savanna and semideciduous forest communities report that most species blossom in the rainy season and in the rainy and dry seasons, respectively (Monasterio and Sarmiento 1976).In Cerrado communities located in different Brazilian regions, flowering usually occurs at the beginning of the rainy season (Batalha and Mantovani 2000) or, more rarely, during the dry season (Miranda 1995).However, other studies suggest that in transition periods flowering intensity in the Cerrado could be influenced by photoperiod (Batalha and Martins 2004) or by the association of the biological features of Cerrado species (Sarmiento and Monasterio 1983) and edaphic conditions (Vasconcelos et al. 2010).
The basalt plateau of the Paraná Basin extends from South-Central Brazil to Uruguay.Within the state of Mato Grosso do Sul in West-Central Brazil, it comprises a set of escarpments called the Serra de Maracaju (Martins 2003).Although the southern part of this Serra has rugged landscapes and steep hillsides with different altitudes, phytophysiognomies, and floristic compositions, it has been scarcely studied (B.G. Fina, unpublished data, Ramos and Sartori 2013) and no work has conjointly addressed its floristic and phenological aspects.
Based on the physical proximity of the vegetation growing on the same side of this Serra de Maracaju slope and considering that they are part of the Cerrado area, we hoped to find both different formations influenced by soil fertility and reproductive phenophases regulated by abiotic factors such as rainfall, day length and temperature (monthly means).
The purpose of this study was to discover 1) if the floristic composition corroborates the fact that different phytophysiognomies grow on the slope, 2) whether there is a relation between soil fertility and floristics, 3) if flowering, fructification and dispersal syndromes are uniform throughout the year, and 4) whether flowering and fructification periods relate to abiotic factors.

GENERAL CHARACTERIZATION OF THE STUDY AREA
Study area is located 47 km from Campo Grande, Mato Grosso do Sul, Brazil, between 20°05' to 20°30' S and 54°45' to 54°15' W, at an approximate altitude of 500 m above sea level (see Figure 1).Vegetation is essentially Cerrado interspersed with seasonal forests (RADAMBRASIL 1982), which, according to UNESCO (1973), are classified as tropical and subtropical drought-deciduous forests.The study area is located in the Serra de Maracaju, which practically crosses the state longitudinally FLORISTICS AND PHENOLOGY OF TREES AND BUSHES and is characterized by significant topographic variations, a great quantity of fountainheads, springs and two main watercourses: the Paraguay and the Paraná Rivers.
Local soils are classified as dystrophic purple latosol and dark red alic and litholic eutrophic latosol (Oliveira et al. 2006).The slope studied here is not deforested and evidence shows it has not burnt for at least 50 years, but cattle grazes in some of its parts.Climate is hot and dry, classified as Aw (Köppen 1948), with rain throughout the year and a mean rainfall.

DATA COLLECTION AND ANALYSIS
In 34 ha covered with native vegetation, eight 25 x 50 m plots transverse to slope were delimited (see Figure 1), totaling 1 ha sampled.All trees and bushes were sampled in each plot.To complete the floristic survey, fertile materials were collected during random walks near each plot.
Flowering and fructification periods of bush and tree individuals were followed up in those plots at community level and recorded monthly on a spreadsheet (presence/absence) between December 2009 and November 2010; the maturity stages of flowers (bud and pre-anthesis) and fruits were disregarded.
Species were assorted by habit according to Cain and Castro (1959), modified by Whittaker To compare the floristic data obtained in our plots, we used the Sørensen similarity index (Müller-Dombois and Ellenmberg 1974).A dendrogram was constructed based on the group mean values (UPGMA).Spearman correlation analysis (Zar 1984) was used to ascertain the relations between monthly rainfall, day lenght and temperature (monthly means) and flowering and fructification, during the year of study and between 2000 and 2010.The relation of VIVIAN A. ASSUNÇÃO, JOSÉ C. CASAGRANDE and ÂNGELA L.B. SARTORI the day length with flowering and fructification, was estimated only for the year study.The abiotic data used in the correlation analyses of the year of study were those measured 2 months before collection, since the response of plants to the effect of climate is usually delayed (Marques et al. 2004) We also checked if flowering and fruiting were different between floristic groups, through the Mann-Whitney test (Zar 1984), if the proportions of anemo-, zoo-and autochory were different by performing a chi-square test (Zar 1984), and, using species number, if flowering, fructification and dispersal syndromes were uniform throughout the year, using circular statistics (Zar 1984); circular statistics is highly recommended to investigate data generated from circular scales (Morellato et al. 2010).
Monthly mean day lengths were obtained from the Brazilian Observatory (Moreira 2011).A climatogram was created according to Walter (1986).Monthly mean rainfall and temperature from 2000 to 2010 was provided by the Center for Meteorological Analyses of the Universidade Anhanguera-UNIDERP (University for the Development of the State and of the Pantanal Region), which is located 32 km away from our study site (20°26'19.4"S; 54°32'18.3"W; 669 m); monthly rainfall below 100 mm was considered the least wet season.The cumulative rainfall of the year of study was 1,673.5 mm, which is higher than the 1,468.8mm mean of the previous 10 years.Mean temperature of the year was 23.6°C and that of the previous 10 years was 23.4°C.
Our climatogram (see Figure 2) showed that the wet season runs from September to March, whereas the period from April to August is less humid, although there is no water deficit.The monthly mean average rainfall during the 10 previous years was 122.4 mm, while that of the year of study was 139.5 mm.
Fifteen soil samples were collected in each plot, 0-20 cm deep.Sampled points were randomly chosen along each plot to constitute a one and only sample.Contents of macro-and micronutrients, and Al, Cation Exchange Capacity (CEC), pH, organic matter, V% and BS (base sum) were assessed.The analyses were performed at the laboratory of the Federal University of São Carlos at Araras, SP, Brazil.When comparing soil features, we performed a Principal Component Analysis (PCA) with 13 variables, excluding the graphic representations of H+Al, BS, CEC and V%.

RESULTS
Floristic richness included 75 species distributed into 58 genera and 31 families (see Table I).In the plots, 54 species belonging to 40 genera and 23 families were sampled (see Table II).The families with highest species richness were: Fabaceae, 10 spp or 13.3%, Myrtaceae and Malvaceae, seven spp or 9.3% each (see Table I).The remaining families were represented by 51 species (see Table I).Based on the floristic plot data, group mean (UPGMA) revealed the presence of two plant groupings (see Figure 3).Grouping 1 (A1) was formed by plots 1, 2, 3 and 4 and grouping 2 (A2) by plots 5, 6, 7 and 8.Both A1 and A2 comprised 38 species, 16 of which were exclusive to each (see Table I); the Sørensen index yielded a 0.71 cophenetic correlation coefficient.
Ordination through Principal Component Analysis (PCA) of the soil of the eight plots analyzed (see Figure 4) presented a cumulative variance of 59.0%, and eigenvalues of axes 1 and 2 were 2.11 and 2.56, respectively.These results were similar to those of the floristic groupings (see Figure 3), the only difference being the distance between plots 4 and 5. Considering the groupings found through floristic similarity, A1 and A2 were practically equal for fertility (p=0.74).Among the floristic groupings, phosphorus (P) (p<0.01),copper (Cu) (p<0.01) and zinc (Zn) (p=0.01)differ significantly (see Table III).
During the period of the study, 54 species were monitored for reproductive phenology at community level: 25 flowered and 52 fruiting (see Table II; see Figure 2).As for the 38 species found in each area, 19 and 22 species were observed in flower in A1 and A2, respectively (see Table II).The difference between both groupings was not significant (U=53.5;p=0.28).Flowering was uniform during the analyzed year for both groupings A1 and A2; mean angles were 268.6° (A1+A2), 249.0° (A1), corresponding to the month of September, and 236.8° (A2), corresponding to August (see Table V).
The number of flowering species was negatively related to rainfall in A2 and presented no correlations to temperature or day length in either grouping (see Table IV).
Fructification occurred in 35 species in A1 and in 36 species in A2 without any significant difference (U=59.5;p=0.47) and was moderately seasonal for both groupings A1 and A2; mean angles were 10.1°, 11.4° and 11.9°, respectively, corresponding to January for all (see Table V).Fructification was explained by the abiotic factors analyzed (see Table IV).
With regards to dispersal, zoochory predominated in 32 species, followed by anemo-and autochory with 13 and 11 species, respectively.Zoochory also prevailed in both groupings (see Table II), and there were no differences between the distribution of dispersal syndromes (x 2 =1.173; p=0.56).All syndromes were seasonal throughout the year in both groupings.Mean Vochysiaceae Callisthene fasciculata Mart.Tree X X X X X X X X 31453 Qualea parviflora Mart.Tree X X X X X X X X 31454

TABLE I (continuation)
Figure 3 -Dendrogram produced through group mean (UPGMA) for the plots analyzed through the Sørensen similarity index in the floristic study on an escarpment slope in the Serra de Maracaju, Mato Grosso do Sul, Brazil (cophenetic correlation = 0.71).32.6° and A2=45.0°corresponding to February, and A1=29.6°,corresponding to January, for autochory (see Table V).These months represent the largest production for each syndrome and suggest that they prevail in the rainy period.FLORISTICS AND PHENOLOGY OF TREES AND BUSHES

DISCUSSION
The high specific richness of Fabaceae and Myrtaceae in western and southeastern Brazil (B.G. Fina, unpublished data, Ruggiero et al. 2002) consolidates the link between the Cerrado and the Atlantic Forest (Pinto and Oliveira-Filho 1999).The present study showed that Fabaceae      (2007), this family is common in savanna, "scrub" and forest edges.
Grouping 1 (A1) corresponds to a remnant of seasonal deciduous forest characterized by the occurrence of the same species, significant levels of leaf deciduousness in the dry period, according to field observations, and eutrophic soil.The presence of characteristic species of dry forests as Sebastiana brasiliensis, Bauhinia ungulata, Rollinia emarginata, Randia armata and Casearia rupestris is reported in other studies on similar formations (Nascimento et al. 2004, Salis et al. 2004, G.A. Damasceno-Junior, unpublished data, Pott et al. 2011).Grouping 2 (A2) here considered as cerradão shows characteristic species such as Curatella americana, Jacaranda cuspidifolia Annona crassiflora, Byrsonima intermedia, Davilla elliptica, Roupala montana.Our results are in agreement with previous studies such as Ribeiro andWalter (2008), andRatter et al. (2006).Curatella americana is common in savannas from Central America to southern Brazil, and the presence of Roupala montana indicates soil fertility (Ratter et al. 2006).
Soil fertility analysis with a pH between 5.5 and 6.0, base saturation higher than 50% and rugged landscapes, suggest eutrophic soil, according to the parameters listed by Ratter et al. (2003) and Reatto et al. (2008).The high content of phosphorus, zinc and copper available in A1 together with deeper soils reinforce the fact that it is a deciduous forest, according to Eiten (1982).A2 is a eutrophic cerradão since its pH is higher than expected dystrophic Cerrado in central Brazil, where it usually varies between 4.0 and 4.8 (Ratter et al. 1978).In addition, such plant formations prevail in basalt formations (Reatto et al. 2008), which is the case of the Serra de Maracaju (Martins 2003).Deciduous forest with fertile soil interspersed with Cerrado (cerradão) could have resulted from dry climate episodes during the Pleistocene, as described by Prado (2000) for South American areas.These formations in the Serra de Maracaju would have acted as corridors for species since they can shelter those common to other plant formations, such as the thorn forest (Chaco) and dry forests (Salis et al. 2004).
The uniform flowering and the negative correlation between flowering and rain, day length and mean temperature in the area of study can be explained by the fact that the soil contains water all year round, since it is interspersed with streams and springs in addition to rocky outcrops that possibly favor water retention.This pattern suggests that the tree and shrub species do not depend directly on rain to flower and probably rely on their deep roots to maintain their water supply (Sarmiento andMonasterio 1983, Batalha andMartins 2004 The moderately seasonal fructification and positive correlation between fructification and average rainfalls stresses the importance of favorable water conditions to yield fleshy fruits since, during the rainy season, easily acess to water contributes to fruit development (Marco and Paéz 2002); the positive correlation between fructification and both day length and mean temperature was expected in this seasonal formation (Morellato et al. 1990, Morellato et al. 2000).
The preponderance of zoochory followed by anemochory and autochory has already been reported in semi-deciduous forest and semideciduous altitude forests (Morellato and Leitão-Filho 1992, Kinoshita et al. 2006), in Cerrado (Batalha and Martins 2004), and it also prevails in tropical forests (Fleming 1979).The period when most species blossomed and fruited, independently of their respective dispersal modes, coincided with that of studies in Cerrado formations of west-central Brazil (Batalha and Martins 2004) and in semi-deciduous altitude forests and semideciduous forests of the southeastern region of Brazil (Morellato and Leitão-Filho 1992).
The intense deciduousness of some species observed during a given period of the year apparently suggests that the area of the Serra de Maracaju studied here is constituted by a one and only plant formation.Nevertheless, two adjacent phytophysiognomies, seasonal forest and cerradão, could be distinguished through floristic composition and soil fertility.These must be further analyzed both to confirm this classification and to conduct other studies on the slopes of this serra.The set formed by given tree and shrub species suggests that this portion of the Serra de Maracaju is floristically linked to the Atlantic Domain (dry forests) and the Cerrado (cerradão), a biogeographical aspect that should be further explored in future studies.
The correlations between reproductive phenophases and the abiotic variables did not differ between deciduous forest and cerradão.The phenological pattern obtained for flowering differed from what was expected in seasonal environments.This is possibly related to the specificities of the plant communities assessed and their evolutionary history.The preponderance of zoochory highlights the importance of animal-plant interaction to maintain the studied community.Nevertheless, abiotic syndromes do play a crucial role in the Serra de Maracaju formed by many slopes that create physical barriers to biotic agents.

ACKNOWLEDGMENTS
The first author wishes to thank the Fundação de Apoio ao Desenvolvimento do Ensino, Ciência e Tecnologia do Estado de Mato Grosso do Sul (FUNDECT) for the fellowship granted, as well as her colleagues and friends who helped her during field collection, and the Pró-Reitoria de Pesquisa e Pós-Graduação (PROPP -Dean for research and post-graduation) and the Program in Vegetal Biology for their logistical support.

Figure 2 -
Figure 2 -Climatic diagram (2000 to 2010) built from the data obtained from the Center for Meteorological Analyses of the University for the Development of the State and of Pantanal Region (Universidade Anhanguera-UNIDERP), Mato Grosso do Sul, Brazil.
III Mean and standard deviation from mean (SDM) of soil fertility at 0-20 cm depth in the eight plots sampled on an escarpment slope in the Serra de Maracaju, Mato Grosso do Sul, Brazil.Elements S, B, Cu, Fe, Mn and Zn in mg/dm 3 , organic matter (M.O.) in g/dm 3 .Except for pH, the other elements and parameters are in mmolc/ dm 3 .V% = percentage of base saturation; BS = base sum (

TABLE I (
continuation) FLORISTICS AND PHENOLOGY OF TREES AND BUSHES

TABLE V Circular analysis of the reproductive phenology of number of trees and bush species on an escarpment slope in the Serra de Maracaju, Mato Grosso do Sul, Brazil. (A1= plots 1,2,3 and 4; A2= plots 5,6,7 and 8; z= result of circular analysis; p= significance
). ).FLORISTICS AND PHENOLOGY OF TREES AND BUSHES