Leaf classes, foliar phenology and life forms of selected woody species from the tropical forests of central and southern Eastern Ghats, India

A checklist of selected woody species of Angiosperms is provided with the aim of classifying their life forms, foliar phenology and leaf classes from the tropical forests of central and southern Eastern Ghats, India. Though there are checklists available for the plants of the Eastern Ghats, a comprehensive listing of quantitative foliar measurements as done in other parts of the world, leading to valuable inputs for Plant Functional Type (PFT) classification, has not thus far been done for this key biogeographic zone of India. The list, gathered from 388 individual plants through the study area, encompasses 156 species and 3 infraspecific taxa which belong to 116 genera and 50 families. Of the total 159 taxa, 83 are evergreen and 76 are deciduous. 135 taxa are trees, 13 are shrubs 10 are climbing shrubs and one hemiparasite. Among the leaf classes of species, mesophyll dominated with 87 species, followed by notophyll (39), microphyll (24) and macrophyll (9). Hence, quantitative leaf trait measurements for selected woody species and the methodology for such studies in the tropics is the unique contribution of the present paper to the existing state-of-the-art.


Introduction
Understanding vegetation through plant functional types (PFTs) is important at a time of increased international concern on deforestation and fragmentation of forests strongly influenced by human impact. Leaves play a major role in the functional classification of plants because important functions such as photosynthesis, respiration and transpiration that greatly modify the morphology of the leaves and, in turn, the plant, occur here. Leaves closely interact with external environment and can be grouped into PFTs (Wright et al. 2004). As demonstrated clearly by Wright et al. (2004; and Barboni et al. (2004), actual trait measurements and analyses do add a distinct dimension to the field of functional ecology, especially from the point of view of conservation and management of forests and/or natural and human modified ecosystems Landsberg et al. 1999). Functional grouping of species is also relevant for vegetation dynamics models that use leaf trait measurements for modeling global climate change and subsequent shifts in vegetation boundaries due to shifts of land use and climate change Moorcroft et al. 2001;Bonan et al. 2002;Wright et al. 2004).
Though the Indian subcontinent, with its rich biodiversity, is one of the 12 mega-diversity regions in the world (McNeely et al. 1990), the high rate of degradation and change of forest cover, due to the so called developmental activities and the ever increasing human population in the forested areas (Jayakumar et al. 2002) results in increasing species loss that could lead to ecosystem collapse (Naeem 2002;Wackernagel et al. 2002). These changes have a number of effects on ecosystem processes (Naeem et al. 1999) and local populations in this region (Rao and Pullaiah 2007). Anthropogenic pressure in the form of clearing forest for areas encompasses an area of ca. 260,000 km 2 . Its area of extent is 1,130 km North-South and 1,050 km East-West. Based on geological and tectonic considerations, this mountain range is formed of four sections or provinces that are not fully or clearly established (Krishnan 1961;1974;Meher-Homji 2001;Dobmeier and Raith 2003;Mukhopadhyay and Basak 2009). The Eastern Ghats, falling under tropical monsoon climate, receives rainfall from both south-west monsoon and north-east retreating monsoon. The natural vegetation, though primarily dominated by tropical deciduous species, may be grouped into Evergreen forests, Tropical semi-evergreen forests, Tropical moist deciduous forests, Dry deciduous forests, Tropical dry evergreen forests and Dry evergreen scrubs (Champion and Seth 1968;Legris and Meher-Homji 1984;Pragasan and Parthasarathy 2009).
The present study focuses on the area consisting of the Nallamalai hills (Andhra Pradesh) section of central Eastern Ghats and Javadi, Kalrayan, Chitteri, Shevaroy, Kolli and Pachamalai hills (Tamil Nadu) section of southern Eastern Ghats (Figure 1). The study sites were selected based on the floristic composition and priority of conservation (Kadavul and Parthasarathy. 1999a, b;Jayakumar et al. 2002;Natarajan et al. 2004;Soosairaj et al. 2004Soosairaj et al. , 2007Reddy et al. 2008;Pragasan and Parthasarathy 2009). Out of 38 sites, 30 sites are from the disjunct hillocks of the southern Eastern Ghats and 8 sites from the more or less contiguous Nallamalai hills of central Eastern Ghats representative of the major vegetation types of the region (Reddy et al. 2008). Although the altitude of these studied sites ranges from 16 to 1362 m a.s.l., the variations in the mean annual rainfall are within the narrow range of 800 to 1100 mm only.

Data Collection
Leaf trait measurements were carried out and phenology observed on a set of plant species selected based on the following criteria: 1) woody species, mainly trees and also some important shrubs and lianas were selected; 2) simple leaves were selected preferentially not only for ease of measurements but also because the existing protocols for trait measurements recommended this (Turner and Tan, 1991;Roderick et al. 1999;Cornelissen et al. 2003); 3) some species with large compound leaves were also chosen, as species with compound leaves were dominant in the study area (Reddy et al. 2009) -further because it is important to standardize the protocol as well. Phenology observations were cross-checked, where required, with record of individuals in the herbarium of French Institute of Pondicherry (HIFP).
Standard protocols proposed by Cornelissen et al. (2003); Garnier et al. (2001) and Roderick et al. (1999) have been adopted with minor changes for leaf trait measurements. Leaf-size was determined using the geometric leaf-size classification of Raunkiaer (1934) as modified by Webb (1959) and recently by Gillison (2002). Six leaf classes were recognized by Raunkiaer (1934). This was later modified into seven by the addition of one more leaf size class 'notophyll' by Webb (1959). Recently Gillison (2002) recognized 10 leaf size classes (Table 1). Floras such as Hooker (1872Hooker ( -1879, Gamble andFischer (1915-1935), Nair and Henry (1983), Henry et al. (1987)   and Matthew (1983) were used for identification of plant species. Voucher specimens were collected and confirmed with the HIFP and deposited there. We have used APG III (2009) classification for enumeration of families. Leaf size, thickness and weight were measured on the field using hand drawn sketches, thickness gauge and portable weighing machine respectively. The hand drawn sketches were brought to the lab and the area was measured using foliar surface area calculation software developed by French Institute of Pondicherry. Leaf area alone is considered for the leaf size classification in this study. Totally, 143 species with simple leaves and 16 species with compound leaves were selected for the present study. For simple leaves, four mature "full sun" leaves (Cornelissen et al. 2003) were measured from every individual. Similarly for compound leaves, leaflet from the species was considered as the unit of study as it is analogous to the simple leaf (Raunkiaer 1934;Cain et al. 1956). The average value was taken as final value for the individual. Southern Eastern Ghats were sampled during February to April 2007 whereas Central Eastern Ghats were sampled in January 2008. All leaf samples were collected between 09:00 h and 14:00 h in order to avoid any effect of temperature, irradiance or vapour pressure deficit of the air on the leaf traits. For most of the species, photographic documentation was carried out and is presented (Figures 2-10), with an emphasis on the species not photo-documented previously. Pearson's correlation coefficient test (parametric test) was done to know the correlation between various parameters in XLStat ® . Official permissions to carry out the field works for this study were provided by the Headquarters and Divisional offices of the forest departments of Andhra Pradesh and Tamil Nadu.

Results and Discussion
388 individuals from 38 sites of southern and central Eastern Ghats were measured for leaf traits (Figure 1; Supporting Information). This represented 156 species and 3 infraspecific taxa which fall under 116 genera, 50 families and four leaf classes (Table 2). Of the 159 woody species that were studied, 135 were trees, 13 were shrubs 10 were climbing shrubs and one hemiparasite (Table 2). Among the different life forms, 54% of the taxa measured were mesophyllous, while 25% were notophyllous, 15% microphyllous, and only 6% were macrophyllous ( Table 2). The higher proportion of mesophylls and the moderate proportions of microphylls and notophylls are characteristic features of tropical regions (Cain et al. 1956;Cain and de Oliveira Castro 1959;Richards 1957;Webb 1959;Malhado et al. 2009) and this is true for our study area. In the list, leaves of two species of Bauhinia (B. racemosa and B. vahlii) are considered as simple as in Saldanha (1984) though most of the taxonomists term them compound. Reddy et al. (2009) studied the tree wealth of Eastern Ghats of Andhra Pradesh and came up with an updated list of 510 tree species, of which 135 are with compound leaves. The present study includes 68 species from central Eastern Ghats of which 62 species are listed by Reddy et al. (2009) and of which 9 are with compound leaves; the remaining species are climbing shrubs (5) and shrub (1) ( Table 2). Foliar phenology and leaf classes of 68 species from central Eastern Ghats are new additions to the existing list of species by Reddy et al. (2009). In our study 10 species of climbing shrubs are included. Out of these 10 species, 5 are from southern Eastern Ghats, 4 are from central Eastern Ghats, 1 species (Hiptage benghalensis) occurs both in southern as well as central Eastern Ghats. Pragasan and Parthasarathy (2009) reported 272 species from tropical forests of southern Eastern Ghats of Tamil Nadu. The present checklist covers only 134 species from southern Eastern Ghats; however 40 species were not listed in that earlier study (Table 2).     In the present study, different leaf sizes were observed within the same species (Table 3). We have assigned leaf classes to these heterophyllous leaf species based on the dominant leaf class exhibited by maximum number of individuals. It is well known that environmental conditions can lead to heterophylly i.e., two or more distinct leaf forms in the same species (Vaughan and Wiehe 1939). Leaf size variation among ecologically similar species and within species may be associated with their function (Ackerly and Reich 1999). Of the total 159 taxa, 83 are evergreen and 76 are deciduous. Here, the evergreen or deciduous nature is based on the foliar phenology of the plant and irrespective of its occurrence in different vegetation types. Further we do not differentiate a brevi-deciduous category. Phenological patterns of tropical forest trees are diverse and complex and dependent on equally complex factors (Bawa and Hedley 1990; Williams-Linera 1997) such as precipitation, tree water status (Reich and Borchert 1982;Borchert 1994), irradiance (Wright and van Schaik 1994) and temperature (Ashton et al. 1988;Tutin and Fernandez 1993). The robust pattern appears to be that deciduous trees lose their leaves in a particular season but evergreen trees drop leaves throughout the year (Williams-Linera 1997). The distinction between deciduous and evergreen habit is not always straight forward (Thomas 2000). For instance, Shorea robusta is considered as an evergreen species (Krishnaswamy and Mathauda 1954) but this species was referred as tardily deciduous (Li et al. 1994), deciduous (Tiwari 1995) and observed leafless for about 8 days in Hathinala forest area of Vindhyan region . The degree of gradient of deciduousness is ascertained only by finer field observations . Not surprisingly, we found that our study assigned a different phenology to 15 species (Table 4) in comparison with Pragasan and Parthasarathy (2009). Of these, 6 that were assigned evergreen by us were characterized brevideciduous in the earlier study. However, Haldina cordifolia, observed to be deciduous in the present study was marked as evergreen by Pragasan and Parthasarathy (2009);Newton (1988) in central Indian highlands, Mishra et al. (2006) in Similipal biosphere reserve, Orissa and Desai and Patel (2010) in Satlasana range forest, north Gujarat observed H. cordifolia as deciduous. Likewise, in Ficus drupacea, leaf flushing was observed from January to mid-May in one season whereas the same was observed during September to December in the next season in the same site in Coorg of Western Ghats (Patel, 1997) and the same species has been marked as brevi-deciduous by Pragasan Table 4. Foliar phenology assignments: Pragasan and Parthasarathy (2009) and present study.
The species reported here correspond to those with different attributions in the two studies and Parthasarathy (2009). The added value of the present study is that it draws attention to the fact that long-term observations at a finer spatial scale are particularly important in deciduous forests.
In the present study, leaf area is significantly correlated with the altitude (two-tailed p-value is 0.002 where alpha is 0.05). Leaf area increased with increasing altitude. Leaf area is not significantly correlated with mean annual temperature (MAT) and mean annual precipitation (MAP). The narrow difference in temperature and precipitation has no effect on the leaf size. The dominant mesophyllous species are highly adaptive within this short range of MAT and MAP as seen in Amazonian rainforest trees where only megaphyllous trees correlated significantly with environmental factors (Malhado et al. 2009). Specific leaf area, stomatal length and index increased with increasing altitude below 2800 m, but decreased with increasing altitude above 2800 m (Li et al. 2006). Royer et al. (2008) suggests that leaf size is not sensitive to MAT and MAP and this is due to wind speed and water availability which underpin these elevation patterns. But still leaf size may be used to know major fluctuations in, for example, mean annual biotemperature in regions such as Costa Rica, Barro Colorado Island, Panama (tropical) and 17 eastern North American sites (temperate) (Dolph and Dilcher 1980a;Royer et al. 2005).
Earlier it was thought that leaf size cannot be used to identify specific life zone or climates in either extant or fossil floras (Dolph and Dilcher 1980b) but recent studies suggest that leaf sizes in modern vegetation accurately predict MAP and can be used in the prediction of climate of fossil leaf assemblages (Wilf et al. 1998). Wolfe (1971Wolfe ( , 1990Wolfe ( , 1995 and Wolfe and Upchurch (1987a, b) have undertaken a comprehensive analysis of palaeofloristic data bearing on the interpretation of Late Cretaceous climates using the extant leaf physiognomy study. Correspondence analysis of Dicotyledones leaf physiognomy of modern vegetational samples from a wide range of environments indicates that more than 70% physiognomic variation corresponds to water or temperature factors, or both. Wolfe (1990) applied this climate-leaf analysis multivariate program (CLAMP) to leaf assemblages from the Cretaceous/Tertiary boundary for the inference of climate characteristics from physiognomic analyses.
South American models for climate reconstruction indicate that MAT -leaf morphology correlation equations based on Northern Hemisphere or Australian data do not accurately predict the MAT of South American sites but the use of tropical south America data itself may help in predicting the MAT of low temperature and high elevation modern sites as well as fossil sites (Kowalski 2002). In this sense, the data presented in this paper can be used to predict the climate variables using leaf size with additional inputs from various other regions of India. The prediction of MAT by analyzing leaf size is largely dependent on the habitat of the plant species (Burnham et al. 2001). It was observed that leaf size tends to increase from colder to warmer environments in humid sites (Grubb 1977;Tanner 1980;Niinemets et al. 2007); however, in the present study the narrow range of temperatures does not allow us to discern any predictable relationship between leaf size and MAT.
The present study standardized the methodology for the leaf trait measurements in the tropical region and also initiated a baseline data on leaf area measurements (Supporting Information). The quantitative trait measurements from the Eastern Ghats do add a distinct dimension to the field of functional ecology: even within the narrow range of temperature and precipitation in the study area, heterophylly is conclusively recorded in 10 species; further, leaf area is significantly correlated with altitude, showing that by extending such studies wider, a definite relationship with climatic parameters, such as precipitation or water availability may also be established. Using information from and comparing with existing checklists from this region has also yielded interesting results and pointers for further research: the need for spatially finer long term monitoring studies to reach definitive conclusions about phenology in deciduous forests and the importance of establishing quantitative measurements for species with compound leaves that are more prevalent in this study area characterized by higher temperatures and lower mean annual precipitation, among them. Leaf area measurements and corresponding climatic data derived from the WorldClim data set (Hijmans et al. 2005) are provided in Supporting Informationcontributing to the CLAMP data base.