Journal of Plant Ecology Advance Access published August 31, 2016

Running Title: Diversity and functioning on Brazilian Páramos

Title: Soil and altitude drives diversity and functioning of Brazilian Páramos (Campo

de Altitude)

Andreza Viana Neri1*,Gladson Ramon Alves Borges1, João Augusto Alves Meira

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Neto1, Luiz Fernando da Silva Magnago2, Ian Michael Trotter3, Carlos Ernesto G. R.

Schaefer4, and Stefan Porembski5

1
Universidade Federal de Viçosa, Departamento de Biologia Vegetal, Laboratório e

Ecologia e Evolução de Plantas, PH Rolfs, S/N, Campus Universitário, 36570-900,

Viçosa, Minas Gerais, Brazil

2
Universidade Federal de Lavras, Lavras, Departamento de Biologia, Setor de Ecologia

e Conservação, Minas Gerais 37200-000, Brazil

3
Universidade Federal de Viçosa, Departamento de Economia Rural, PH Rolfs, S/N,

Campus Universitário, 36570-900, Viçosa, Minas Gerais, Brazil

4
Universidade Federal de Viçosa, Departamento de Solos, PH Rolfs, S/N, Campus

Universitário, 36570-900, Viçosa, Minas Gerais, Brazil

5
Universität Rostock, Institut für Biowissenschaften, Allgemeine und Spezielle Botanik,

Wismarsche Str. 8, D-18051 Rostock, Germany

*Correspondence address. Tel.:+55 31 38991954; Fax:+55 31 3899 2580; E-mail:

[Link]@[Link]

© The Author 2016. Published by Oxford University Press on behalf of the Institute of Botany, Chinese Academy of Sciences and the Botanical
Society of China. All rights reserved. For permissions, please email: [Link]@[Link]
 2

Abstract

Aims: The vegetation on Brazilian Páramos consists of assemblages that are driven

mainly by the influence of strong environmental filtering. It is very important to

understand the effect of environmental variation on taxonomic diversity and on

functional diversity. Considering the lack of information about the functional diversity

in Brazilian Páramos, we analysed for the first time the effects of altitude and edaphic

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

attributes on functional traits, as well as on taxonomic and functional diversity. We also

wanted to answer the questions: Which ecological strategies are favourable in high

altitude grassland? Does soil attributes determine distributions of traits in high altitudes

grassland? Considering the studied altitudinal gradient is altitude an important variable

in the community assembly?

Methods: the study was conducted on three mountains: Mammoth (1850 m), Elephant

(1790 m) and Totem (1690 m) in Serra do Brigadeiro State Park, Minas Gerais State,

Brazil. Those mountains represent the “Serra das Cabeças”, a smaller ridge which is

surrounded by the Atlantic Forest, one of the 25 hotspot of biodiversity. The samples

were taken using 100 plots of 1m2 per mountain that were randomly distributed. All

plants except mosses were sampled. The taxonomic diversity was evaluated using

richness, Shannon diversity, effective number of species and Pielou evenness. For the

functional diversity we considered the functional richness (FRic), functional evenness

(FEve) and functional divergence (FDiv). Generalized linear models (GLM/Poisson and

Quasi-Poisson) were used to evaluate the effect of abiotic variables (altitude, soil depth

and soil chemical attributes) on biotic variables (number of species and individuals, life

form, dispersal and fruit type) and ordinary least squares regression to evaluate the

effect of abiotic variables on the functional and taxonomic diversity.

 3

Important Findings: The soil variables presented a considerable edaphic gradient

associated with altitude. The soil in Serra das Cabeças plays an important role for the

plant diversity: richness and diversity index were positively related with fertility. With

regards to the life form, nanophanerophytes tended to increase with altitude and soil

depth, while therophytes tended to decrease with altitude. The dispersal type was also

associated with the abiotic variables: autochory decreases with altitude, while zoochory

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

increases. Functional richness increases with fertility and the functional evenness with

altitude. The studied gradient showed that altitude is working as a filter for functional

traits and indices, and is, together with soil attributes, an important determinant for the

distribution of plants on Brazilian Páramos.

Key words: Mountain conservation, vegetation islands, functional diversity, life form,

highland grassland, functional traits.

INTRODUCTION

Species in mountain habitats are the most vulnerable to climate change. An increase

in temperature can possibly cause extinction, since those species usually are endemic

and also isolated in specific areas (Gottfried et al. 2012; Leão et al. 2014; Pickering et

al. 2008; Wehn et al. 2014). For this reason, the Convention on Biological Diversity

(CBD) included a working program on mountain biological diversity. The CBD

recognises that mountains retain a high concentration of hotspots, including high

ecosystem diversity and high numbers of endemic and endangered species. The CBD

also pointed out the fragility of mountain ecosystems

([Link] Mountains are storehouses of

 4

biological diversity, and worldwide home to endangered and endemic species (Beniston

2003). Additionally, there is too little ecological information for more comprehensive

investigations of tropical highlands under a global change scenario.

The present study was carried out in Brazilian Páramos. This vegetation type

consists of assemblages that are driven mainly by the influence of strong environmental

filtering. It is very important to understand the effect of environmental variation on the

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

composition (i.e. taxonomic diversity) and on function (i.e. functional diversity),

especially when taking into account that climatic conditions may change in the coming

decades (Elkin et al. 2013; Engler et al. 2011; Ndiribe et al. 2014). The Brazilian

Páramos, as a particular type of mountain vegetation, occurs on isolated peaks and is

characterized by freezing temperatures on the highest peaks, in which case the climate is

probably the main filter for the floristic composition.

Brazilian Páramos form an outstanding ecosystem that is associated with the

Atlantic Forest domain, one of the 25 global biodiversity hotspot (Myers et al. 2000). It

is a high altitude grassland above 1700 m, although there are also similar formations on

the summits of lower mountains (Safford and Martinelli 2000; Mocochinski and Scheer

2008). Brazilian Páramos represents a small portion of the earth’s surface, as the total

area is probably less than 350 km2 (Safford 1999a).

According to Safford (2007), this vegetation consists of islands of temperate climate

above the tropical Atlantic Forest domain. Analyzing the macroclimatic patterns,

Safford (1999b) suggested that the Brazilian Páramos are essentially a high latitude

variant of tropical Andean páramos, with enhanced seasonality. Brazilian Páramos

consist of different habitat types, forming a mosaic of synusiae with grassland, shrub,

scrub and rock outcrop vegetation (Safford and Martinelli 2000). Differing soil fertility
 5

and soil depth conditions promote the establishment of different life forms and species

(Benites et al. 2003), contributing to the diversity and thus shaping the landscape of the

Brazilian Páramos.

Despite the importance of these habitats for biodiversity conservation, considering

the richness and the number of endemic species, these ecosystems have been neglected

with regards to inventories and ecological studies. Little research has been done, and

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

usually only related to the floristic composition (Caiafa and Silva 2005, 2007; Iganci et

al. 2011; Mocochinski and Scheer 2008; Tinti et al. 2015). Ecological studies

investigating the effect of soil characteristics on plant community structures on

Brazilian rock outcrops have been done mainly on quartzite outcrops (campos rupestres)

(Conceição et al. 2007; Le Stradic et al. 2015; Nunes et al. 2015; Silva 2013; Vincent

and Meguro 2008). For Brazilian Páramos there is an overview showing the main soil

types associated with the vegetation on high altitudinal rupestrian complexes (Benites et

al. 2003).

Some studies have shown that the distribution of species is related to topography,

altitude and soil, but it is difficult to separate the effect of each factor individually

(Ruokolainen and Tuomisto, 2002; Davidar et al. 2007), especially because plant

distribution is usually related to the spatial heterogeneity of these factors (Zuquim et al.

2007). Duivenvoorden et al. (2002) stated that environmental heterogeneity is a

determinant for species composition on a local scale, however on a regional scale the

pattern of species composition is associated with the dispersion capacity of each

individual (Hubbell 2006). Porembski (2007) pointed out that isolated mountain peaks

can be considered terrestrial islands. Those naturally fragmented and isolated habitats

are characterized by spatial and ecological isolation that act as a barrier against
 6

dispersion and migration. Certain studies have shown low degrees of floristic similarity

amongst rock outcrops (Rezende et al. 2013; Sarthou et al. 2003; Tinti et al 2015; de

Paula et al. 2016) and have also detected low genetic exchange rates between

populations from different rock outcrops (Barbará et al. 2007; Boisselier-Dubayle et al.

2010; Domingues et al. 2011; Palma-Silva et al. 2011; Hmeljevski et al. 2014).

Brazilian rock outcrops comprise a large number of locally endemic species

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

(Barthlott et al. 1993; Giulietti & Pirani 1988; Safford & Martinelli 2000; Alves &

Kolbek 2010). This isolation has been attributed to their altitude and also to edaphic

isolation (Harley 1995, Alves et al. 2007).

Granitic and gneissic outcrops are recognized as biodiversity hotspots for plants

possessing particular adaptive traits such as desiccation tolerance (Porembski 2007).

However, the lack of knowledge about their flora and ecology on Brazilian mountains is

remarkable when compared with the surrounding Atlantic Forest (Porembski 2007;

Martinelli 2007; Pessanha et al. 2014). This is extremely concerning, considering that

the rocky outcrops are the ecosystems associated with Brazilian Atlantic Forest with the

greatest probability of extinction, along with their species (Leão et al. 2014).

Studies of biodiversity have traditionally considered the numbers of species and the

evenness of abundance distribution among species. However, the indices based on

taxonomic identity provide an incomplete view of biodiversity, since they do not take

into account the differences between the species and their role in the ecosystem

processes (Petchey 2004; Villéger et al. 2010). On the other hand, functional diversity

generally shows the range of functions that organisms perform in communities and

ecosystems and try to explain and predict the impact of those organisms on ecosystems

(Petchey and Gaston 2002, 2006). Functional ecology considers the functional traits,
 7

which are defined as biological attributes that influence the species performance (Violle

et al. 2007). The diversity of a community can therefore be measured by the diversity of

traits present in the community, weighted by their abundance (Petchey and Gaston

2006). In this case, the functional diversity helps to understand the importance of

species in the ecosystem process (Loreau 2000) and the ecosystem dynamic through a

more detailed analysis of form, function and plant-environmental interaction (Gillison

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

2013).

Considering the lack of information about the functional diversity in Brazilian

Páramos, we analysed for the first time the effect of altitude and edaphic attributes on

functional traits, as well as on taxonomic and functional diversity indices. We also

wanted to answer the question: Which ecological strategies are favourable in high

altitude grassland? Does soil attributes can determine the traits distributions in high

altitude grasslands? Considering the studied altitudinal gradient is altitude an important

variable in the community assembly? We hope to advance the knowledge about

Brazilian Páramos and also to contribute to their conservation. It is extremely important

to understand which factors drive the distribution of their species in order to make

decisions about habitat conservation and also for monitoring the dynamic of highland

vegetation in response to global change.

MATERIALS AND METHODS

Study area

The study was conducted on the peak of three mountains: Mammoth (1850 m), Elephant

(1790 m) and Totem (1690 m). Those mountains form the “Serra das Cabeças” – Three

Head Range. The Serra das Cabeças is part of Serra do Brigadeiro State Park, which is
 8

located in the Zona da Mata of MG, Brazil and surrounded by Atlantic Rain Forest. The

annual average rainfall is 1300 mm and the annual average temperature is 18ºC. The

terrain is rough and consists of different landforms and habitats: plateaux, steep slopes,

escarpments, valleys and swamps (Schaefer et. al. 2007). Due to the landscape and the

altitude, the peak of the mountains experience much lower temperatures than the

surrounding region. The study area is located in the southeast of Brazil and is part of the

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Serra da Mantiqueira mountain range, with a predominance of plutonic rocks (granite)

and high grade metamorphics (migmatites, gneiss) (Benites et al. 2007).

On the peaks of the mountains there are different vegetation types: cloud forest,

scrub, shrub and rocky outcrops. The floristic composition varies between the

vegetation types and is usually associated with soil depth or exposure of the rock. We

focus our study on high altitude rock outcrops, which are part of the Brazilian Páramos.

Samples

The field samples comprised 300 plots of 1 x 1 m, that is, 100 plots per peak randomly

distributed and considering only the rocky outcrop areas. Samples were taken between

June 2012 and January 2013. All type of plants except the mosses were exhaustively

inventoried and the species cover was estimated using the Braun Blanquet (1979) cover

value and abundance scale. In addition, the number of individuals and altitude per plot

were variables considered in the samples. For species that form tussocks, each clump

was considered one individual.

The fertile botanical samples were stored in the herbarium VIC at Universidade

Federal de Viçosa (Minas Gerais, Brazil). Identifications were made by comparisons

 9

with exsiccates in the herbarium VIC, by consulting specialists and the literature. The

classification of families and confirmation of accepted names followed Tropicos (2015).

Apart from the altitude, the depth and chemical features of the soil were sampled as

abiotic variables. The soil depth was measured with a graduated ruler in the centre of

each plot. Soil samples were taken from each plot. The samples were air-dried and

sieved to 2 mm. In fraction < 2 mm, chemical soil properties ( pH, P, K, Ca2+, Mg2+,

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Al3+, H+Al3+ , SB, t, T, V, m, OM, Prem) were determined in the Soil Laboratory at the

Universidade Federal de Viçosa.

Data analysis

It was necessary to eliminate 90 plots for the functional analysis. Some of the plots had

lower numbers of species than traits, and in some plots not all species could be

functionally classified. In addition, some plots were eliminated because they contained

species with mostly the same traits, occupying less than three points in the functional

space, and therefore it was not possible to calculate the volume occupied in the

functional space, which is a prerequisite for calculating some functional indices. In

total, we performed all the calculations with 210 plots.

Environmental gradient

The dimensionality of the dataset from the chemical features of the soil was reduced

using principal component analyses (PCA). The analysis was performed on the

logarithm of the variables. We used the results of the PCA to investigate a possible

edaphic gradient related to altitude and also to investigate possible correlations between

the abiotic and biotic variables.

 10

Functional traits

The species were classified according to Raunkiaer’s life form (1934), as described by

Braun Blanquet (1979). The classification is related to the perennial tissues and

meristems that allow plants and seeds survive during unfavourable season. The location

of perennial tissues and meristems is an essential feature of the plant’s adaptation to

climate. Some life forms are therefore unlikely to survive in harsh climates or will have

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

few individuals (Cornelissen et al. 2003; Pérez-Harguindeguy et al. 2013).

A regenerative trait, dispersal syndrome, was evaluated as well. This trait was

considered in the analyses since it is very important for colonising new sites. The plants

were classified in three dispersal modes: anemochory, zoochory and autochory. The

classification was done by consulting the literature about the species and also based on

the propagule characteristics (Cornelissen et al. 2003; Pérez-Harguindeguy et al. 2013).

The type of fruit was another trait considered in our analyses: dry fruit (fruit which

the pericarp is hard) and fleshy fruit (pericarp fleshy or the fruit has a fleshy appendix,

such as aril). Functional and taxonomic diversity

The taxonomic diversity was evaluated using the Shannon diversity index, the

Pielou evenness and effective number of species (Jost 2006). For the functional

diversity, we considered the three complementary indices, which measure independent

facets of functional diversity: functional richness (FRic), functional evenness (FEve),

and functional divergence (FDiv) (Villéger et al. 2008).

FRic is the functional space occupied by the species within a community and the

measurement of functional richness depends on the distribution of the abundance of

each species through niche space; low FRic indicates reduction of productivity and also

that some resources are not used (Petchey 2003, Mason et al. 2005, Villéger et al.
 11

2008). FEve is the regularity of the distribution of species abundances in the functional

space; a low FEve indicates that some parts of the niche are under-utilized, which tends

to decrease productivity and reliability (Mason et al. 2005, Villéger et al. 2008). FDiv

indicates how species abundance diverges from the centre of the functional space; high

FDiv indicates high degree of niche differentiation and low resource competition

(Mason et al. 2005, Villéger et al. 2008). Those indices are important for biological

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

assemblage studies, since they are enable to predict the rate and reliability of ecosystem

processes (Mason et al. 2005). For the calculations, we considered the Raunkiaer’s life

form, dispersion syndrome and type of fruit as functional traits.

Statistical analyses

The analysis of variance (ANOVA) was run to test if there was a difference between the

environmental variables (PC1, PC2, PC3, soil depth and altitude) considering the peak

as a factor. The PC1, PC2 and PC3 represented the soil chemical features in further

analysis. The differences in the vegetation among the plots were explored by a non-

metric multidimensional scale (NMDS) using abundance of species.

Generalized linear models using a Poisson distribution (GLM) were used to evaluate the

effect of abiotic variables (altitude, soil depth and soil chemical features) on the biotic

variables (number of species and individuals, life form, dispersal syndrome and type of

fruit), since the dependent variables represented frequencies. We also used Quasi-

Poisson distribution in the cases where the variables presented over or underdispersion.

Multiple linear regression analysis was used to evaluate the effect of the abiotic

variables on the functional and taxonomic diversity indices, since these are continuous

variables.
 12

All analyses were performed in R, version 3.1.2 (R Core team. 2015) using the

following packages: vegan, version 2.4.0 (Oksanen et al. 2016); FDind (Villéger et al.

2010); plyr, version 1.8.4 (Wickham 2011); AER, version 1.2-4 (Kleiber and Zeileis

2008) and memisc, version 0.99.6 (Elff 2016).

RESULTS

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

In the 210 plots that were used for the calculations we found 6,507 individuals and 69

species (Table S1). The life forms nanophanerophytes, hemicryptophytes and

chamephytes were the most common ones, although with slightly differing proportions

between the mountains (Figures S2, S3).

Environmental and vegetation gradient

The soil variables reflect a considerable edaphic gradient associated with altitude. In

the PCA (Fig. 1) the gradient related with altitude is obvious from the fact that the plots

from the three peaks form distinct groups in the projections along the two first principal

components: the tallest peak (whose plots are represented by crosses) generally

occupies the space in the middle of the first component and the bottom end of the

second component, the lowest peak (represented by triangles) seems to form a

downward sloping line, whereas the middle peak (represented by circles) forms an

upward sloping line.

The three first axes explained 75.7% of the variance. The first principal component

(PC1) was positively correlated with the sum of bases (SB), K, Ca2+, Mg, and bases

saturation (V), and negatively correlated with Al saturation (m). The second component

(PC2) was mostly negatively correlated with aluminum and cation exchange capacity.
 13

The third component (PC3) was negatively correlated with P remaining, P and Al3+, and

positively correlated with pH. Therefore, we consider the first component, PC1, to

represent fertility, and the second component, PC2, shows a negative correlation with

aluminium. The ordination had an arch effect, which indicates that the PC1 and PC2

pick up the same gradient what is nonlinear. It did not mean any problem for the other

analysis, since we used both PC1 and PC2.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Considering the peak as a factor the variance analysis (ANOVA) of abiotic features

evaluated (soil depth, altitude, PC1, PC2 and PC3) showed that there are significant

differences between the peaks with p> 0.001(Tab. 1 and S4).

The vegetation also showed a small gradient as can be seen on the NMDS. Despite

the clear gradient the stress with two dimensions was great, stress equal 0.22. However,

with three dimensions we got a better result, stress equal 0,16 and the ordination of plots

separated the three mountains (Fig. 2).

Taxonomic diversity and environmental features

The soil in the Serra das Cabeças plays an important role for the spatial distribution

of plant diversity; together with altitude it is an important factor for the gradient. The

regression analysis in table 2 showed that richness decreases with altitude and increases

with fertility (PC1). The diversity index (H’) was also related with soil features

represented by PC1. Considering the taxonomic diversity, altitude was also an important

factor for J’, showing that this index increases with altitude.
 14

Functional groups, traits and environmental features

Considering the relationship between life forms and abiotic variables (soil and

altitude), the nanophanerophytes were positively associated with altitude, as opposed to

chamaephytes and therophytes, which were negatively associated (Tab. 3). Table 3

shows that soil was an important variable that controls life form distribution. In general,

phanerophytes were related with soil depth and fertility.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

The soil depth was the abiotic variable that was correlated with all kind of life forms

analysed, except for geophytes. The soil depth is important for species that apparently

have deeper root systems, such as phanerophytes, or need deeper soil for establishing

seed banks, such as therophytes.

When analysing the number of species per life form (Table S5), the correlations

were very similar to when we considered the number of individuals. However,

hemicryptophytes tend to increase with altitude and geophytes showed a negative

relation with soil depth.

The dispersal types were also associated with abiotic variables and can be

emphasized that zoochory tends to increase with altitude, soil depth and fertility.

Anemochory decreases with soil depth, but increases with soil fertility. In contrast,

autochory is negatively correlated with soil fertility (Tab. 4 and Table S6). Since

zoochory is usually associated with fleshy fruits, the same patterns were found when

considering the relationship between fruit types and abiotic variables (Tab. 5), though

dry fruits were negatively correlated with soil depth following the same pattern as we

found for anemochory.

 15

Functional diversity

The regressions between the functional diversity indices and environmental

variables showed that the functional evenness increased with altitude (Tab. 6). The

functional richness was positively correlated to PC1, which means that, in the studied

gradient, greater FRic is associated with fertility. The functional divergence was

negatively associated with soil depth and positively with PC2 (axis that is negatively

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

related with aluminum).

DISCUSSION

In this study we investigated the effect of environmental filters (soil and altitude) on

taxonomic and functional diversity of Brazilian Páramos. Our results showed that not

only altitude, but also the soil attributes are responsible for the species and traits

distribution. The PCA a clear edaphic gradient and together with altitude, these

variables play an important role for plant functional diversity. In the taxonomic

diversity the evenness shows that the abundance of species tends to be more

homogenous with altitude whereas richness and diversity tend to increase with fertility.

The life form nanophanerophyte is positively related with altitude while therophyte

shows an opposite trend. The functional indices also showed a relation with the

environmental filters, what is particular true for FEve that tended to increase with

altitude and FRic with fertility.

As suggested by our results there is a strong variation in soil properties on rocky

outcrops and this variation concerns particularly the Al3+ content and soil fertility that

play an important role in the distribution of species (Benites 2007; Le Stradic et al.

2015; Silva 2013). These authors pointed out the strong relationship between soil and
 16

vegetation, showing that the soil changes together with the floristic composition and

physiognomy of the vegetation. However, in the present study we could also

demonstrate that soil plays an important role concerning the spatial distribution of life

forms, dispersal types and functional diversity indices.

Life forms were very sensitive to altitude with nanophanerophytes and

hemicriptophytes having the greatest abundance and species richness in the Serra das

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Cabeças (S2, S3). Our results are similar to those of Safford (2007) who found the

same trend in the spatial distribution of life forms considering only genera

identification. The proportion of hemicryptophytes tends to be higher than that of

phanerophytes in highland grassland.

Safford (2007) evaluated the dispersal syndrome on the genus level for tropical

mountains, and concluded that in general wind and gravity dispersal is predominant. We

considered species rather than genera, and in this case anemochory was also the richest

and most abundant dispersal syndrome (S7). However, we could not find any

correlation between altitude and anemochory. In this case the great proportion of

anemochory and autochory indicates a change in the abundance of dispersal syndrome,

which is more similar to the temperate zones (Safford 2007).

In our study perennials life cycle such as nanophanerophytes and hemicryptophytes

dominated whereas therophytes were rare (S2, S3) which tend to decrease with altitude

(Porembski et al 1998). Species of plants that occur in high elevations are naturally

affected by low temperatures, radiation, wind and storm or insufficient water

availability. As a response, they present morphological and physiological adjustments

such as growth form, small leaves, reproductive strategies and low thermal requirements

(Beniston 2003; Körner and Larcher 1988). In addition, the plants in high altitudes
 17

require adaptation and frost tolerance (Safford 2007), and also the plants should present

characteristics that are common in alpine vegetation – low stature, cushion shape,

perennial life cycle and ecophysiological adaptations – to cope with high light intensity

(Körner 2003).

The functional index FEve increases with altitude, which suggests that the highest

parts of the “Serra das Cabeças” are more even, considering the distribution of species

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

in the functional space. This might indicate that the niches available have been used

efficiently by species and for now, invaders have low chance for getting established

(Mason et al. 2005). The functional redundancy can be very important for the

maintenance of the ecosystem functioning (Flynn et al., 2009). The FRic increased with

fertility and it indicates that the potential resources (niches) have been used and our

results also showed that increasing resources (fertility) also increases FRic.

It is a general rule that high species richness occurs on mid-altitude peak areas

(Colwell 2000), since low-altitude areas usually suffer high level of human impact and

high-altitude areas also support less species, but in this case because of the harsh

climate conditions (Jiménez-Valverde and Lobo 2007; Wilson et al. 2007; Zhou et al.

2015). Our results showed that richness decreases with altitude and in this case we can

consider that we were working from mid-altitude to high altitude, since we did not

sample low-altitude rocky outcrops. As we did not test the mid-domain effect of

Colwell (2000), we cannot reject or confirm, but some of our result indicated a

confirmation, like the number of species per peak. The richest peak was the one at

intermediate elevation. However, further statistical tests are necessary to support such a

statement (Colwell 2000).

 18

The studied gradient showed that altitude is acting as an environmental filter that

controls the abundance of certain functional traits. Our study showed that there is a

strong association between biotic and abiotic variables and this relationship should be

further investigated in the future. In our study, we found that altitude, soil depth and

other soil attributes are important determinants for the distribution of plants and also

have effects on functional traits in Brazilian Páramos.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

SUPPLEMENTARY MATERIAL

ACKNOWLEDGEMENTS

We thank the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (Fapemig)

and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq -

206814/2014-3, Postdoctorate scholarship of A.V.N.). We also would like to thank the

staff of Serra do Brigadeiro State Park for logistical support and the specialists that

helped with the identification.

REFERENCES

Alves RJV, Cardin L, Kropf MS (2007) Angiosperm disjunction ‘‘Campos Rupestres-

Restingas’’- a re-evaluation. Acta Botanica Brasilica 2: 675–685.

Alves RJV, Kolbek J (2010) Can campo rupestre vegetation be floristically delimited

based on vascular plant genera? Plant Ecology 207: 67–79.

Barbará T, Martinelli G, Fay MF, Mayo SJ, Lexer C (2007) Population differentiation

and species cohesion in two closely related plants adapted to neotropical high-

altitude “inselbergs”, Alcantarea imperialis and Alcantarea geniculata

 19

(Bromeliaceae) Molecular Ecology 16: 1981–[Link] W, Gröger A,

Porembski, S (1993) Some remaks on the vegetation of tropical Inselbergs: diversity

and ecological differentiation. Biogéographica 69: 105-124.

Beniston M (2003) Climatic Change in Mountain Regions: A Review of Possible

Impacts. Climatic Change 59: 5–31.

Benites VM, Caiafa AN, Mendonça ES, Schaefer CEGR, Ker JC (2003) Solos e

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

vegetação nos complexos rupestres de altitude da Mantiqueira. Floresta e Ambiente

30: 76–85.

Benites VM, Schaefer CEGR, Simas FNB, Santos HG (2007) Soils associated to rock

outcrops in the Brazilian mountain ranges Mantiqueira and Espinhaço. Revista

Brasileira de Botânica 30: 569-577.

Boisselier-Dubayle MC, Leblois R, Samadi S, Lambourdière J, Sarthou C (2010)

Genetic structure of the xerophilous bromeliad Pitcairnia geyskesii on inselbergs in

French Guiana – a test of the forest refuge hypothesis. Ecography 33: 175–184.

Braun-Blanquet J (1979) Fitosociologia Bases para el estudio de las comunidades

vegetales. H Blume Ediciones.

Caiafa AN, Silva AF (2005) Composição florística e espectro biológico de um campo

de altitude no Parque Estadual da Serra do Brigadeiro, Minas Gerais—Brasil.

Rodriguésia 56: 163–173.

Caiafa AN, Silva AF (2007) Structural analysis of the vegetation on a highland granitic

rock outcrop in Southeast Brazil. Brazilian Journal of Botany 30: 657–664.

Conceição AA, Pirani JR, Meirelles ST (2007) Floristics, structure and soil of insular

vegetation in four quartzite-sandstone outcrops of “Chapada Diamantina”, Northeast

Brazil Brazilian. Journal of Botany 30: 641–656.

 20

Cornelissen JHC, Lavorel S, Garnier E, Díaz S, Buchmann N, Gurvich DE, Reich PB,

Steege H ter, Morgan HD, Heijden MGA van der, Pausas JG, Poorter H (2003) A

handbook of protocols for standardised and easy measurement of plant functional

traits worldwide. Australian Journal of Botany 51: 335–380.

Davidar P, Rajagopal B, Mohandass D, Puyravaud JP, Condit R, Wright SJ, Leigh EG

(2007) The effect of climatic gradients, topographic variation and species traits on

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

the beta diversity of rain forest trees. Global Ecology and Biogeography 16: 510–

518.

Domingues R, Machado MA, Forzza RC, Melo TD, Wohlres-Viana S, Viccini LF

(2011) Genetic variability of an endangered Bromeliaceae species (Pitcairnia

albiflos) from the Brazilian Atlantic rainforest. Genetics and Molecular Research

10: 2482–2491.

Duivenvoorden JF, Svenning JC, Wright SJ (2002) Beta Diversity in Tropical Forests.

Science 295: 636–637.

Elff M (2016). memisc: Tools for Management of Survey Data and the Presentation of

Analysis Results. R package version 0.99.6. [Link]

[Link]/package=memisc

Elkin C, Gutiérrez AG, Leuzinger S, Manusch C, Temperli C, Rasche L, Bugmann H

(2013) A 2 °C warmer world is not safe for ecosystem services in the European

Alps. Global Change Biology 19: 1827–1840.

Engler R, Randin CF, Thuiller W, Dullinger S, Zimmermann NE, Araújo MB, Pearman

PB, Le Lay G, Piedallu C, Albert CH, Choler P, Coldea G, De Lamo X, Dirnböck T,

Gégout J-C, Gómez-García D, Grytnes J-A, Heegaard E, Høistad F, Nogués-Bravo

D, Normand S, Puşcaş M, Sebastià M-T, Stanisci A, Theurillat J-P, Trivedi MR,

 21

Vittoz P, Guisan A (2011) 21st century climate change threatens mountain flora

unequally across Europe. Global Change Biology 17: 2330–2341.

Flynn DFB, Gogol-Prokurat M, Nogeire T, Molinari N, Richers BT, Lin BB, Simpson

N, Mayfield MM, DeClerck F (2009) Loss of functional diversity under land use

intensification across multiple taxa. Ecology Letters 12: 22–33.

Gillison AN (2013) Plant Functional Types and Traits at the Community, Ecosystem

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

and World Level. In: van der Maarel E. and Franklin J. Vegetation Ecology (eds)

John Wiley & Sons, Ltd, Oxford, UK, pp.347-386.

Giulietti AM, Pirani JR (1988) Patterns of geographic distribution of some plant species

from the Espinhac¸o Range, Minas Gerais and Bahia, Brazil. In: Vanzolini PE,

Heyer WR (eds) Proceedings of a workshop on neotropical distribution patterns.

Academia Brasileira de Ciências, pp 39–69

Gottfried M, Pauli H, Futschik A, Akhalkatsi M, Barančok P, Benito Alonso JL, Coldea

G, Dick J, Erschbamer B, Fernández Calzado MR, Kazakis G, Krajči J, Larsson P,

Mallaun M, Michelsen O, Moiseev D, Moiseev P, Molau U, Merzouki A, Nagy L,

Nakhutsrishvili G, Pedersen B, Pelino G, Puscas M, Rossi G, Stanisci A, Theurillat

J-P, Tomaselli M, Villar L, Vittoz P, Vogiatzakis I, Grabherr G (2012) Continent-

wide response of mountain vegetation to climate change. Nature Climate Change 2:

111-115.

Harley RM (1995) Introduction. In: Stannard BL, Harvey YB, Harley RM (eds) Flora of

the Pico das Almas, Chapada Diamantina, Bahia, Brazil. Royal Botanic Gardens,

Kew.

Hmeljevski KV, Freitas L, Domingues R, Pereira AR, Cancio AS, Andrade ACS,

Machado MA, Viccini LF, Forzza RC (2014) Conservation assessment of an

 22

extremely restricted bromeliad highlights the need for population-based

conservation on granitic inselbergs of the Brazilian Atlantic Forest. Flora -

Morphology, Distribution, Functional Ecology of Plants 209: 250-259.

Hubbell SP (2006) Neutral theory in ecology and the evolution of ecological

equivalence. Ecology 87: 1397-1308.

Iganci JR, Heiden G, Miotto STS, Pennington R (2011) Campos de Cima da Serra: the

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Brazilian Subtropical Highland Grasslands show an unexpected level of plant

endemism. Botanical Journal of the Linnean Society 167: 378–393.

Jiménez-Valverde A, Lobo JM (2007) Determinants of local spider (Araneidae and

Thomisidae) species richness on a regional scale: climate and altitude vs habitat

structure. Ecological Entomology 32: 113–122.

Jost L (2006) Entropy and diversity. Oikos 113: 363–375.

Kleiber C and Zeileis A (2008). Applied Econometrics with R. New York: Springer-

Verlag. [Link]

Körner Ch, Larcher W (1988) Plant life in cold climates Symposia of the Society for

Experimental Biology 42: 25–57.

Körner Ch (2003) Alpine plant life. 2nd edition, Springer, Heidelberg.

Leão TCC, Fonseca CR, Peres CA, Tabarelli M (2014) Predicting Extinction Risk of

Brazilian Atlantic Forest Angiosperms: Neotropical Plant Extinction Risk.

Conservation Biology 28: 1349–1359.

Le Stradic S, Buisson E, Wilson FG (2015) Vegetation composition and structure of

some Neotropical Mountain Grasslands in Brazil. Journal of Mountain Science 12:

864-877.
 23

Loreau M (2000) Biodiversity and Ecosystem Functioning: Recent Theoretical

Advances. Oikos 91: 3–17.

Martinelli G (2007) Mountain biodiversity in Brazil. Brazilian Journal of Botany 30:

587–597.

Mason NW, Mouillot D, Lee WG, Wilson JB (2005) Functional richness, functional

evenness and functional divergence: the primary components of functional diversity.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Oikos 111: 112–118.

Mocochinski AY, Scheer MB (2008) Campos de altitude na serra do mar paranaense:

aspectos florísticos. Floresta 38: 625-640.

Naqinezhad A, Zare-Maivan H, Gholizadeh H (2015) A floristic survey of the

Hyrcanian forests in Northern Iran, using two lowland-mountain transects. Journal

of Forest Research 26: 187–199.

Ndiribe C, Pellissier L, Dubuis A, Vittoz P, Salamin N, Guisan A (2014) Plant

functional and phylogenetic turnover correlate with climate and land use in the

Western Swiss Alps. Journal of Plant Ecology 7: 439–450.

Nunes J A, Schaefer CEGR, Ferreira-Junior WG, Neri AV, Correa GR, Enright NJ

(2015) Soil -vegetation relationships on a banded ironstone ‘island’, Carajás Plateau,

Brazilian Eastern Amazonia. Anais da Academia Brasileira de Ciências 4: 2097-

2110.

Oksanen J, Blanchet F G, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin P R,

O'Hara R B, Simpson G L, Solymos P, Stevens M H H, Szoecs E and Wagner H

(2016). vegan: Community Ecology Package. R package version 2.4-

0. [Link]
 24

Palma-Silva C, Wendt T, Pinheiro F, Barbará T, Fay MF, Cozzolino S, Lexer C (2011)

Sympatric bromeliad species (Pitcairnia spp) facilitate tests of mechanisms involved

in species cohesion and reproductive isolation in Neotropical inselbergs. Molecular

Ecology 20: 3185–3201.

Paula LFA, Forzza RC, Neri AV, Bueno ML, Porembski S (2016) Sugar Loaf Land in

south-eastern Brazil: a centre of diversity for mat-forming bromeliads on inselbergs.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Botanical Journal of the Linnean Society 181: 459-476.

Pérez-Harguindeguy N, Díaz S, Garnier E, Lavorel S, Poorter H, Jaureguiberry P, Bret-

Harte MS, Cornwell WK, Craine JM, Gurvich DE, Urcelay C, Veneklaas EJ, Reich

PB, Poorter L, Wright IJ, Ray P, Enrico L, Pausas JG, de Vos AC, Buchmann N,

Funes G, Quétier F, Hodgson JG, Thompson K, Morgan HD, ter Steege H, Sack L,

Blonder B, Poschlod P, Vaieretti MV, Conti G, Staver AC, Aquino S, Cornelissen

JHC (2013) New handbook for standardised measurement of plant functional traits

worldwide. Australian Journal of Botany 61: 167-234.

Petchey OL (2003) Integrating methods that investigate how complementarity

influences ecosystem functioning. Oikos 101: 323–330.

Petchey OL (2004) On the statistical significance of functional diversity effects.

Functional Ecology 18: 297–303.

Petchey OL, Gaston KJ (2002) Functional diversity (FD), species richness and

community composition. Ecology Letters 5: 402–411.

Petchey OL, Gaston KJ (2006) Functional diversity: back to basics and looking forward.

Ecology Letters 9: 741–758.

 25

Pickering C, Hill W, Green K (2008) Vascular plant diversity and climate change in the

alpine zone of the Snowy Mountains, Australia. Biodiversity and Conservation 17:

1627–1644.

Porembski S, Martinelli G, Ohlemüller R, Barthlott W (1998) Diversity and ecology of

saxicolous vegetation mats on inselbergs in the Brazilian Atlantic rainforest.

Diversity and Distributions 4: 107–119.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Porembski S (2007) Tropical inselbergs: habitat types, adaptive strategies and diversity

patterns. Revista Brasileira de Botânica 30: 579-586.

R Development Core Team (2014) R: A language and environment for statistical

computing R Foundation for Statistical Computing, Vienna, Austria Disponível em:

<[Link]

Rezende MG, Elias RCL, Salimena FRG, Menini Neto L (2013) Flora vascular da Serra

da Pedra Branca, Caldas, Minas Gerais e relações florísticas com áreas de altitude

da Região Sudeste do Brasil. Biota Neotropica 13:201-224

Ruokolainen K, Tuomisto H (2002) Beta-Diversity in Tropical Forests. Science 297:

1439–1439.

Safford HD (1999a) Brazilian Páramos I: An introduction to the physical environment

and vegetation of the campos de altitude. Journal of Biogeography 26: 693–712.

Safford HD (1999b) Brazilian Páramos II: Macro- and mesoclimate of the campos de

altitude and affinities with high mountain climates of the tropical Andes and Costa

Rica. Journal of Biogeography 26: 713–737.

Safford HD, G Martinelli (2000) Southeast Brazil. In: Porembski S, Barthlott W (eds)

Inselbergs: biotic diversity of isolated rock outcrops in tropical and temperate

regions. Heidelberg, Ecological Studies, Springer-Verlag, 339-389.

 26

Safford HD (2007) Brazilian Páramos IV: Phytogeography of the campos de altitude.

Journal of Biogeography 34: 1701–1722.

Sarthou C, Villiers J F, Ponge JF (2003) Shrub Vegetation on Tropical Granitic

Inselbergs in French Guiana. Journal of Vegetation Science 14: 645-652

Schaefer C E G R, Simas F N B (2007) Elementos do meio físico e a gestão de

Unidades de Conservação. Ação Ambiental 37: 12-15.

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

Silva, W A (2013) Gradiente vegetacional e pedológico em complexo rupestre de

quartzito no Quadrilátero Ferrífero, Minas Gerais, Brasil. MSc dissertation,

Universidade Federal de Viçosa, Brasil.

Tinti BV, Schaefer CERG, Nunes JA, Rodrigues AC, Fialho IF, Neri AV (October 30,

2015) Plant diversity on granite/gneiss rock outcrop at Pedra do Pato, Serra do

Brigadeiro State Park, Brazil. Check List 10.15560/11.5.1780

Tropicos. [Link]

Villéger S, Mason NW, Mouillot D (2008) New multidimensional functional diversity

indices for a multifaceted framework in functional ecology. Ecology 89: 2290–2301.

Villéger S, Miranda JR, Hernández DF, Mouillot D (2010) Contrasting changes in

taxonomic vs functional diversity of tropical fish communities after habitat

degradation. Ecological Applications 20: 1512–1522.

Vincent RC, Meguro M (2008) Influence of soil properties on the abundance of plant

species in ferruginous rocky soils vegetation, southeastern Brazil. Brazilian Journal

of Botany 31: 377–388.

Violle C, Navas M-L, Vile D, Kazakou E, Fortunel C, Hummel I, Garnier E (2007) Let

the concept of trait be functional! Oikos 116: 882–892.

 27

Wehn S, Lundemo S, Holten JI (2014) Alpine vegetation along multiple environmental

gradients and possible consequences of climate change. Alpine Botany 124: 155–

164.

Wickham H (2011). The Split-Apply-Combine Strategy for Data Analysis. Journal

of Statistical Software 40: 1-29. URL [Link]

Wilson RJ, Gutiérrez D, Gutiérrez J, Monserrat VJ (2007) An elevational shift in

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

butterfly species richness and composition accompanying recent climate change.

Global Change Biology 13: 1873–1887.

Zhou T, Chen B-M, Liu G, Huang F-F, Liu J-G, Liao W-B, Wang Y-Y, Ren S-J, Chen

C-Q, Peng S-L (March 12, 2015) Biodiversity of Jinggangshan Mountain: The

Importance of Topography and Geographical Location in Supporting Higher

Biodiversity. PLoS ONE 10.1371/[Link].0120208

Zuquim G, Costa F R C, Prado J (2007) Fatores que determinam a distribuição de

espécies de pteridófitas da Amazônia Central. Revista Brasileira de Biociências 5:

360-
 28

Table 1 Summary of variance analysis (ANOVA) of the abiotic variables evaluated in

three Mountains of Serra do Brigadeiro State Park. PC represents the three first
principal components of the PCA (PC1 – axis 1, PC2 – axis 2, PC3 – axis 3).

Df Sum Sq Mean Sq F value P (>F)

Soil Depth 2 1913 956.7 14.32 1.5e-06***
Altitude 2 636550 318275 1773 <2e-16***
PC1 2 368.1 184.06 45.9 <2e-16***
PC2 2 195.2 97.6 41.7 6.08e-16***
PC3 2 73.6 36.81 20.97 5.08e-09***
Df: degree of freedom, Sum Sq: sum of square, Mean sq: mean of square

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

 29

Table 2. Regression of taxonomic diversity and the environmental variables. The Quasi-
Poisson distribution was used for S, due to underdispersion. S: number of species, H’:
Shannon diversity and J’: Pielou evenness, PC: Principal component of the PCA (PC1 –
axis 1, PC2 – axis 2, PC3 – axis 3)
S H’ J’
(Intercept) 3.456*** 1.438 0.072
(0.806) (0.904) (0.298)
Altitude -0.001* 0.000 0.000*
(0.000) (0.001) (0.000)
Soil Depth -0.002 -0.001 0.000
(0.003) (0.003) (0.001)
PC1 0.034*** 0.038*** 0.006

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

(0.008) (0.009) (0.003)
PC2 -0.008 -0.007 -0.001
(0.013) (0.015) (0.005)
PC3 0.031* 0.036* 0.010
(0.014) (0.016) (0.005)
Sigma 0.311 0.103
R-squared 0.102 0.119
F 4.613 5.433
P 0.064 0.001 0.000
N 210 210 210
 30

Table 3. Regressions (GLM/Quasi-Poisson) of number of individuals of each life form

and the environmental variables. NP: nanophanerophyte, H: hemicryptophyte, P:
phanerophyte, G: geophyte, C: chamephyte, T: terophyte, PC: Principal component of
the PCA (PC1 – axis 1, PC2 – axis 2, PC3 – axis 3)

N H F G C T
(Intercept) -9.332** 2.532 -9.132 3.604 10.586*** 49.462***
(3.341) (2.788) (9.868) (12.208) (2.710) (14.207)
Altitude 0.006** 0.000 0.004 -0.002 -0.005** -0.030***
(0.002) (0.002) (0.006) (0.007) (0.002) (0.008)
Soil Depth 0.033*** -0.050*** 0.039* -0.067 -0.034** 0.084**
(0.008) (0.010) (0.018) (0.052) (0.011) (0.029)

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

PC1 0.094** -0.006 0.241** 0.076 0.009 -0.105
(0.032) (0.029) (0.092) (0.114) (0.029) (0.166)
PC2 -0.079 0.004 0.030 0.029 -0.073 -0.145
(0.054) (0.046) (0.143) (0.177) (0.044) (0.174)
PC3 -0.063 0.139** 0.467** -0141 -0.071 -0.333
(0.053) (0.051) (0.161) (0.208) (0.051) (0.257)
P 0.000 0.000 0.000 0.001 0.000 0.000
N 210 210 210 210 210 210
 31

Table 4. Regressions (GLM/Quasi-Poisson) of number of individuals of each dispersion

type and the environmental variables. Ane: anemochory, Auto: autochory, Zoo:
zoochory, PC: Principal component of the PCA (PC1 – axis 1, PC2 – axis 2, PC3 – axis
3)

Ane Auto Zoo

(Inercept) 5.392* 0.282 -37.604***
(2.098) (2.748) (7.961)
Altitude -0.001 0.001 0.021***
(0.001) (0.002) (0.004)
Soil depth -0.025*** -0.016 0.047***

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

(0.007) (0.009) (0.011)
PC1 0.005 0.031 0.223***
(0.022) (0.027) (0.062)
PC2 -0.010 -0.023 -0.295**
(0.034) (0.044) (0.099)
PC3 0.108** -0.190*** 0.200*
(0.039) (0.048) (0.091)
P 0.000 0.000 0.000
N 210 210 210
 32

Table 5. Regression (GLM/Quasi-Poisson) of number of individuals of each fruit type

and the environmental variables. PC: Principal component of the PCA (PC1 – axis 1,
PC2 – axis 2, PC3 – axis 3)

Dry Fleshy
(Intercept) 3.747* -30.150***
(1.752) (7.875)
Altitude -0.000 0.016***
(0.001) (0.004)
Soil depth -0.022*** 0.057***
(0.006) (0.011)

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

PC1 0.011 0.259***
(0.018) (0.065)
PC2 -0.015 -0.315**
(0.029) (0.105)
PC3 0.025 0.210*
(0.032) (0.098)
P 0.000 0.000
N 210 210
 33

Table 6. Regressions of functional indices and the environmental variables. FRic:

functional richness, FEve: functional evenness, FDiv: functional divergence, PC:
Principal component of the PCA (PC1 – axis 1, PC2 – axis 2, PC3 – axis 3)

FRic FEve FDiv

(Intercept) -3.829 -1.665*** 1.038
(4.333) (0.476) (0.272)
Altitude 0.004 0.001*** -0.000
(0.003) (0.000) (0.000)
Soil Depth -0.007 0.002 -0.004***
(0.014) (0.001) (0.001)
PC1 0.195*** 0.004 -0.002
(0.045) (0.005) (0.003)

Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016

PC2 -0.001 -0.004 0.013**
(0.072) (0.008) (0.004)
PC3 0.031 0.007 0.002
0.079 (0.008) (0.005)
Sigma 1.526 0.164 0.094
R-squared 0.119 0.217 0.173
F 5.501 11.334 8.542
P 0.000 0.000 0.000
N 210 210 210
 Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016
34

Figure 1
 Downloaded from [Link] at Universidade Federal de Viçosa on September 1, 2016
35

Figure 2