Botanical Journal of the Linnean Society (2001), 136: 231–238. With 8 figures doi:10.1006/bojl.2000.0430, available online at http://www.idealibrary.com on

Leaf anatomical variation in triplinervia (Spreng) Mu¨ ll. Arg. () under distinct light and soil water regimes

G. ROˆ C¸ AS and F. R. SCARANO∗ FLS

Universidade Federal do , Departamento de Ecologia, CCS, Ilha do Funda˜o, 21941-970, Rio de Janeiro, RJ,

C. F. BARROS

Jardim Botaˆnico do Rio de Janeiro, Laborato´rio de Botaˆnica Estrutural, 22460-030, Rio de Janeiro, RJ, Brazil

Received July 2000; accepted for publication October 2000

Alchornea triplinervia (Spreng.) Mu¨ ll. Arg. (Euphorbiaceae) is a tree which occurs in a broad range of habitats in Brazil. In the State of Rio de Janeiro, it occurs from montane forests to swamplands at sea level. A quantitative approach was used to examine the role of light and soil water regime on the variations found in anatomical traits of the palisade and spongy parenchyma, outer epidermal cell wall of the abaxial and adaxial surfaces, the percentage of sclerenchymatous area in relation to the total midrib area and the ratio of palisade to spongy parenchyma for five distinct ecological populations: M1 – late secondary montane forest (shaded, unflooded); M2 – early secondary montane forest (semi-exposed, unflooded); S1 – primary swamp forest (semi-exposed, flooded); S2 – secondary swamp forest (exposed, flooded); and D – deforestation area (exposed, unflooded). Tukey tests were carried out for multiple comparisons, while one-way factor variance analysis was used to test for differences among ecological populations. A principal component analysis was used to order the populations as well as to find the higher variance component. These populations developed different response levels to the environmental factors studied, namely light and soil water regime. Light accounted for the variations found in palisade and spongy parenchyma while the combination of light and soil water determined the variations found regarding the outer epidermal cell wall of the abaxial surface, the percentage of sclerenchymatous area in relation to the total midrib area and the compaction of the spongy parenchyma. The separation of S1/M2 and S2/D populations into two groups was due to similar light regimes, which suggested that light was affecting the leaf anatomical variation of A. triplinervia more than the interaction of light and soil water regime.  2001 The Linnean Society of London

ADDITIONAL KEY WORDS: – intraspecific variation – tropical tree.

INTRODUCTION and anthropogenic impact. These are often highly ad- aptable when faced with environmental vari- Widespread, common species are not as easily ation, which may be of key relevance in a global change threatened by extinction as rare and/or endemic spe- scenario (Parsons, 1990). cies, with the result that conservation initiatives often The Atlantic rainforest complex, one of the world’s focus on the latter. However, populations of widespread 25 biodiversity hotspots, is notable for its high species plants can be locally extinguished (McKinney, 1997) diversity and high level of endemisms (Myers et al., due to vulnerability to local environmental pressures 2000). However, it also has many interesting cases of generalist plants which are seldom studied from an intra-specific variation perspective (Roˆc¸as, Barros & Scarano, 1997). The poor conservation status of this ∗ Corresponding author. E-mail: [email protected] biome and high level of habitat fragmentation (only 231 0024–4074/01/060231+08 $35.00/0  2001 The Linnean Society of London 232 G. ROˆ C¸ AS ET AL.

7.5% of its original area remains; Myers et al., 2000) In the Poc¸o das Antas Biological Reserve, mu- demands a deeper scientific knowledge regarding spe- nicipality of Silva Jardim, the plants were collected in cies acclimation to environmental heterogeneity. a primary swamp forest (S1), a vegetation type which The euphorb tree Alchornea triplinervia (Spreng.) used to be typical of the Atlantic forest lowlands, Mu¨ ll. Arg. is a widespread neotropical species which and is now reduced to a few remnants. The sunlight occurs in various habitats within the Atlantic rain- penetrates through the canopy semi-exposing the forest complex. In a previous study we (Roˆc¸as et al., understorey. The soil is waterlogged for c. 10 months 1997) compared leaf anatomy of specimens of this of the year (Scarano et al., 1997). Tree mean DBH was species in relation to light intensity in two remnants c. 15 cm and the mean height c. 15 m. of secondary forest at distinct successional stages. The Jacarepia´ State Ecological Reserve, mu- Subsequently, we found three further populations, two nicipality of Saquarema, is a sandy coastal plain loc- of them subject to flooding. In this paper, we examine ated between the Atlantic Ocean and the Jacarepia´ the role of both periodic flooding and light on leaf lagoon. Specimens were collected in a secondary swamp anatomical variation for this species, aiming to discern forest (S2). Adult plants had an average height of the role played by each of these factors on the variations 5 m and DBH was c. 10–12 cm. Sunlight penetrated found. Although acclimation cannot be predicted from through canopy gaps formed by previous anthropogenic leaf anatomy alone (Strauss-Debenedetti & Berlyn, disturbances. Thus, the understorey was classified as 1994), the variation spectrum provides an indication exposed to sunlight. Here, the soil is waterlogged for of the species’ ecological plasticity. 10–11 months per year. In a neighbouring impacted Studies of structural modifications in plants oc- site which suffered deforestation (D), another sample cupying distinct habitats have often lacked a sound was collected, which was totally exposed to sunlight quantitative basis (e.g. Zagdanska & Kozdo´j, 1994) and growing on unflooded sandy soil. These plants and more commonly examine the effect of a single were approximately 2–3 m in height (Sa´, 1992). environmental factor on the character(s) under Thus, A. triplinervia plants in M1 were shaded and scrutiny (Smirnoff & Crawford, 1987; Sojka, 1992). unflooded, in M2 were semi-exposed and unflooded, in Currently, a more quantitative approach is being em- D were exposed and unflooded, in S1 were semi-exposed ployed (Loreti & Oesterheld, 1996; Schreiber & Rie- and flooded and in S2 were exposed and flooded. derer, 1996; Rozema et al., 1997), although studies attempting to assess which environmental factor ac- counts for which variation pattern are still lacking. LEAF ANATOMY Palisade and spongy parenchyma, and outer epidermal MATERIAL AND METHODS cell wall of the abaxial and adaxial surfaces were measured by using sections embedded in methacrylate STUDY SITES resin (hydroxyethyl methacrylate, Reichert-Jung, his- Five leaves were taken at the middle-third of the toresin embedding kit), and stained by toluidine blue canopy of each of 3–5 mature individuals of Alchornea 0.1% (Feder & O’Brien, 1968). These tissues are known triplinervia (Spreng.) Mu¨ ll. Arg. in each of five sites to reveal responses related to light and soil water distributed along three distinct vegetation types in Rio variations (Esau, 1972; Levitt, 1980; Rozema et al., de Janeiro State, south-eastern Brazil (Fig. 1). These 1997). Histochemical tests for lignin, tannins, and sites were chosen because they represent distinct light calcium oxalate crystals were carried out according to and water regimes within a single vegetational domain, Johansen (1940) and Jensen (1962). The percentage of the Atlantic rainforest complex. Geographic and cli- sclerenchymatous area in relation to the total midrib matic features of these sites are depicted in Table 1. area was calculated by using images obtained by the Leaves were harvested at all sites at the same seasonal camera lucida of a light microscope which were sub- time and from individuals with similar physiological sequently digitized (Genitizer GT-1212B). The software age. used was Autocad R13 for Windows. The sample size Two sites were located at Macae´ de Cima Ecological established was 25 fields for each measurement (N= Reserve, municipality of Nova Friburgo: a late-sec- 25), from which were calculated mean and standard ondary montane forest (M1) and a neighbouring early deviation. Tukey tests were carried out for multiple secondary montane forest (M2). M1 was characterized comparisons, while one-way analysis of variance was by a dense canopy cover, a shaded understorey and a used to verify the existence of differences among eco- humid clay-sandy soil, while M2 had large canopy logical populations. Principal components analysis gaps, higher light penetration and a dry clay-sandy (PCA) was used to order the populations as well as soil. The mean trunk diameter at breast height (DBH) to find the higher variance components (Schlichting, of A. triplinervia was c. 15 cm and the mean height 1986; Zar, 1996). The statistical analyses were done c. 20 m for both sites (Roˆc¸as et al., 1997). with Statistica v. 4.2. LEAF ANATOMICAL VARIATION IN ALCHORNEA TRIPLINERVIA 233

Espírito Santo 410 210 450 210

M1 M2

Minas Gerais

Macaé de Cima Ecological Reserve Poço das Antas S1 Biological Reserve Jacarepía State São Paulo Ecological Reserve Rio de Janeiro

S2 D Atlantic Ocean 60 km State limit

Figure 1. Rio de Janeiro State, south-eastern Brazil, study sites and the leaf size of each sample. Scale bar (inside the leaves)=3 cm. M1 – late secondary montane forest; M2 – early secondary montane forest; S1 – primary swamp forest; S2 – secondary swamp forest; and D – deforestation area.

Table 1. General geographic and climatic information related to the sites where Alchornea triplinervia was sampled for leaf anatomy analyses in Rio de Janeiro State, Brazil

Sites Macae´ de Cima Ecological Poc¸o das Antas Biological Jacarepia´ State Ecological Reserve Reserve Reserve

Latitude 22°30′ and 22°33′S22°30′ and 22°33′S22°47′ and 22°57′S Longitude 42°19′ and 42°14′W42°15′ and 42°19′W42°20′ and 42°43′W Mean altitude 880–1720 m a.s.l. Sea level Sea level Climate Mesothermic, always Hot and humid Hot and humid humid Mean annual rainfall 1500–2000 mm 2260 mm 1000 mm Mean minimal temperature 13.8°C 18.6°C8°C Mean maximal temperature 21°C 30.6°C30°–32°C

RESULTS paracytic stomata and the epidermis was uniseriate. Stellate multicellular trichomes occurred on both faces Alchornea triplinervia showed phenotypic variations of all specimens studied. in regard to leaf morpho-anatomy among the ecological M1, M2 and S1 had a dorsiventral mesophyll. M1 populations studied. M1 (leaves of shaded, unflooded showed a one-layered palisade parenchyma and 3–5 plants) had mesomorphic characteristics, while the layers of spongy parenchyma. The intercellular spaces others showed xeromorphic trends. were wide, often occupying all the extension between M1 leaf length was 20–25 cm, while M2 (semi-ex- the palisade and the abaxial epidermis (Fig. 2). M2 and posed and unflooded) and S1 (semi-exposed and S1 were structurally similar; the palisade parenchyma flooded) was 10–15 cm, and D (exposed and unflooded) contained 1–2 layers of compactly arranged elongate and S2 (exposed and flooded), 8–10 cm (Fig. 1). In all cells and the spongy parenchyma a few small in- populations the leaves were amphistomatous with tercellular spaces and c. 4–5 layers. In the latter, which 234 G. ROˆ C¸ AS ET AL.

Figures 2–7. Transverse leaf section. ∗=palisade parenchyma; _=spongy parenchyma. 2. M1; 3. M2; 4. S1; 5. S2; 6. d; 7. transverse midrib section of an individual collected in D. → gelatinous fibres. Scale bars=50 m.

was closer to the abaxial epidermis, the diameter in- parenchyma layers, in both cases with a compact ar- creased towards the anticlinal wall, similar to a pal- rangement as compared with the others. They showed a isade cell (Figs 3, 4). S2 and D frequently showed completely developed second palisade near the abaxial two palisade parenchyma layers and c. 2–4 spongy epidermis (Figs 5, 6), characterizing an isobilateral LEAF ANATOMICAL VARIATION IN ALCHORNEA TRIPLINERVIA 235

Table 2. Leaf anatomical differences of ecological populations of Alchornea triplinervia. Mean±SD and F values (ANOVA, ∗ P<0.05). Means followed by different letters differ at P<0.05 (HSD Tukey test). (+) light exposure or flooded soil; (−) shade or unflooded soil; (±) semi-exposure to light. M1 was used as a control and not included in the comparative analysis. M1 – late secondary montane forest; M2 – early secondary montane forest; S1 – primary swamp forest; S2 – secondary swamp forest; and D – deforestation area

M1 M2 S1 S2 D F Light −±±++ Flooding −−++−

Palisade parenchyma 55.65±4.96 73.87±10.58a 75.11±5.47a 49.97±8.54c 55.94±5.69b 65.02∗ (PP) Spongy parenchyma 48.68±6.72 69.37±2.54a 68.23±11.11a 48.21±5.46b 51.68±46.2b 32.04∗ (SP) PP/SP 1.17±0.21 1.09±0.18a 1.12±0.17a 1.05±0.20a 1.11±0.24a 0.77 Adaxial epidermis 1.84±0.49 2.43±0.46a 2.72±1.16a 2.70±0.68a 2.52±0.57a 0.83 Abaxial epidermis 2.15±0.4 2.18±0.4a 1.71±0.62b 2.21±0.50a 1.99±0.42a 5.51∗ Sclerenchyma 28.91±3.12 29.8±5.16b 25.03±7.79c 34.47±7.47a 35.44±3.56a 14.70∗ percentage

mesophyll. The ratio of palisade to spongy parenchyma significantly from the others. These results indicated of all populations did not show statistical differences, that the magnitude of the tissue responses was dif- probably because the ontogeny of these tissues was ferent when they were submitted to distinct en- not under ecological control (Table 2). Druses of calcium vironmental factors. oxalate and phenolic substances were found in both The PCA indicated that the anatomical char- tissues. acteristics varied according to the first two components, The structural arrangement of the midrib was sim- which together explained 62.96% of the total variance. ilar in all sites. The uniseriate epidermis is followed PC1, which corresponded to the light factor, was re- by an angular collenchyma in both faces. The arch- sponsible for 42.46% of the total variance, and was shaped vascular system showed five bundles disposed strongly influenced by palisade and spongy par- near the adaxial epidermis (Fig. 7). Different levels of enchyma and sclerenchyma percentage value. PC2, sclerification and gelatinous fibres were observed. A responsible for 20.46% of the total variance, was mainly lower proportion of sclerenchyma occurred at S1, fol- influenced by the outer epidermal cell wall of the lowed by M1-M2 and then S2 and D (Table 2). abaxial surface, and secondarily by the sclerenchyma The multiple comparison tests and the PCA (Fig. 8A) percentage value (Fig. 8B), and thus corresponded to showed that M1 formed an isolated and intermediate the interaction of light and water. group in relation to the other groups (M2/S1 and D/ There was a high positive correlation between pal- S2). In the field, these individuals were taller (c. 20 m isade and spongy parenchyma, which were negatively tall), had bigger leaves and a broader canopy, sug- correlated to the sclerenchyma percentage value. The gesting that these were under less stressful conditions populations were distributed more widely in relation than the other populations. For this reason, M1 was to PC1 (light) and more narrowly in relation to PC2 used as a natural control and was excluded from the (light/flooding). PC1 was primarily responsible for the comparative statistical tests. population arrangement and the anatomical character The phenotypic differences regarding leaf anatomy variation (Fig. 8B). The proximity between the sample- among the populations of A. triplinervia were tested points M2/S1 and S2/D indicated an intrinsic similarity by a multiple comparison test (Table 2). M2 and S1 between these groups. Palisade and spongy par- showed the highest palisade and spongy parenchyma enchyma seemed to play a key role in the ordering of thickness and S2 the smallest. The outer epidermal PC1, while the outer epidermal cell wall of the abaxial cell wall of the adaxial surface did not vary significantly surface and the sclerenchyma percentage value best among the distinct populations. Although M2, S2 and explain the arrangement of PC2. Again, light is the D differed significantly among them for the outer epi- main factor for the M2/S1 and S2/D grouping. dermal cell wall of the abaxial surface, they formed an isolated group in relation to S1. The highest thick- DISCUSSION ness of the outer epidermal cell wall of the abaxial surface was found for S2 and M2. D and S2 showed Leaf anatomy of the ecological populations of A. tri- the highest sclerenchyma percentage values, differing plinervia revealed different magnitudes of response to 236 G. ROˆ C¸ AS ET AL.

6.0 A 5.0

4.0

3.0

2.0

PC2 1.0

0.0

–1.0

–2.0

–3.0 2.5 B

1.5

0.5 2 1 4 PC2 –0.5 3

–1.5

–2.5 –3 –2 –10123 PC1

Figure 8. Principal components analysis (PCA): variable ordering (loadings) and population ordering (scores). M1 (Α); M2 (—); S1 (Φ); S2 (+); and D (Β). (A) M1 included in the analysis; (B) M1 excluded from the analysis. 1=palisade and spongy parenchyma; 2=sclerenchyma percentage; 3=abaxial epidermal cells; 4=adaxial epidermal cells.

the light and soil water regime. Light accounted for Sprugel & Hinckley, 1996; Rozema et al., 1997; Roˆc¸as variations in the thickness of the outer epidermal et al., 1997), the mesophyll of leaves exposed to high cell wall (see Schreiber & Riederer, 1996) and in the levels of light was thicker and had a well-developed thickness and number of cell layers of the palisade palisade layer with columnar-shaped cells, while that and the spongy parenchyma. Field observations re- of those developed under shade or low irradiance were vealed that the young leaves remained rolled up to the thin. The columnar shape of palisade cells facilitates moment of their complete differentiation. During this light penetration to the spongy tissue (Vogelmann & period, the abaxial epidermis is exposed while the Martin, 1993). The distribution of light through the adaxial one is protected. Thus, the former is constantly leaf and the acclimation of chloroplasts in habitats exposed to environmental factors during its de- with high levels of light may maximize the rate of velopment while the latter is subjected to physical photosynthesis and reduce susceptibility to photo- factors only when the leaf is already fully grown. The inhibition (De Lucia et al., 1996). As exposure to light thickness of the outer epidermal cell wall of the former increased, we observed a tendency for the development varied significantly among the collecting areas, while of another palisade layer near the abaxial epidermis; that of the latter did not. at its most extreme this resulted in a total development As in other studies (Chabot & Chabot, 1977; Brooks, of an isobilateral structure at S2 and D. This kind of LEAF ANATOMICAL VARIATION IN ALCHORNEA TRIPLINERVIA 237 structure is common in xeromorphic plants (Esau, ACKNOWLEDGEMENTS 1972). In all areas the ratio of palisade to spongy parenchyma was similar, suggesting that the ontogeny We thank C. G. Costa, D. S. D. Araujo and G. W. of these tissues under genetic control, while number Fernandes for critical readings of the manuscript; the of cells, shape, size and compaction of the mesophyll staff at the Anatomy Laboratory of the Rio de were environmentally controlled. Janeiro Botanical Garden; CNPq (Brazilian Research According to Esau (1972) and Levitt (1980), light Council) for research grants nos. 520494/93-8 and enhances the development of palisade parenchyma and 301056/92-7; and the Atlantic Forest Program (Shell, the compaction of spongy parenchyma, while water Margaret Mee Foundation Amazon Trust, MacArthur deficit hinders normal cell growth, inducing the ap- Foundation) for funding. pearance of smaller cells and a higher thickness of parenchyma due to an accumulation of photosynthetic products. Kozlowski, Kramer & Pallardy (1991) argued REFERENCES that turgor reduction impairs cellular growth and en- largement, while light directly affects the extensible Brooks JR, Sprugel DG, Hinckley TM. 1996. The effects properties of the cell wall. Turner (1986) concluded of light acclimation during and after foliage expansion on that leaf growth may be affected by hydric root status photosynthesis of Abies amabilis foliage within the canopy. as well as leaf turgor. It is likely that the light and Oecologia 107: 21–32. soil water regimes of all the areas studied contributed Chabot BF, Chabot JF. 1977. Effects of light and tem- to the anatomical arrangement observed and these perature on leaf anatomy and photosynthesis in Fragaria vesca. Oecologia 81: 124–131. tissues may thus be used as climate indicators. We DeLucia EH, Nelson K, Vogelmann TC, Smith WK. 1996. used the sclerophyll index as another climatic indicator Contribution of intercellular reflectance to photosynthesis of light and water deficit (Nu´ n˜ ez-Olivera, Martı´nez- in shade leaves. Plant Cell Environment 19: 159–170. Abaigar & Escudero, 1996). The increase of the scler- Esau K. 1972. Anatomı´a vegetal. Barcelona: Omega. enchyma at the midrib suggests a tendency to xero- Feder N, O’Brien TP. 1968. Plant microtechnique: some morphism (Witkowski & Lamont, 1991), since this principles and new methods. American Journal of Botany tissue enhances leaf hardening, contributing to a re- 55(1): 123–142. duction of structural damage during dry and windy Jensen WA. 1962. Botanical histochemistry (principles and spells (Turner, 1994). Probably these factors played a practice). San Francisco: Freeman & Co. role in the development of sclerenchyma in all popu- Johansen D. 1940. Plant microtechnique. New York: lations except M1 (shaded and unflooded). McGraw-Hill. The separation of S1/M2 and S2/D populations Kozlowski TT, Kramer PJ, Pallardy SG. 1991. The physio- into two groups was intrinsically related to light logical ecology of woody plants. San Diego: Academic Press. exposure which acted upon thickness and number of Levitt J. 1980. Responses of plants to environmental stresses. cell layers of palisade and spongy parenchyma, while Vol. 2. New York: Academic Press. the interaction of light and soil water regime affected Loreti J, Oesterheld M. 1996. Intraspecific variation in the thickness of the outer epidermal cell wall of the the resistance to flooding and drought in populations of abaxial surface, the percentage of sclerenchyma and Paspalum dilatum from different topographic positions. the compaction of the mesophyll. A. triplinervia Oecologia 108: 279–284. showed both mesomorphic and xeromorphic char- McKinney ML. 1997. How do rare species avoid extinction? acteristics depending on the population. Tilman A palaeontological view. In: Kunin WE, Gaston KJ, eds. The (1988) suggested that resistance to particular en- biology of rarity: causes and consequences of rare–common vironmental factors, which results in particular re- differences. London: Chapman & Hall, 110–129. Myers N, Mittermeier RA, Mittermeier CG, Fonseca source allocation patterns, should represent a GAB, Kent J. 2000. Biodiversity hotspots for conservation structural conflict since it may trigger the de- priorities. Nature 403: 853–858. velopment of adaptive strategies to both factors. This Nun˜ ez-Olivera E, Martı´nez-Abaigar J, Escudero JC. may account for the broad ecological plasticity of A. 1996. Adaptability of leaves of Cistus ladanifer to widely triplinervia. varying environmental conditions. Functional Ecology 10: Ecological plasticity of the extent shown by A. tri- 636–646. plinervia is possibly linked to its ability to adjust leaf Parsons PA. 1990. The metabolic cost of multiple en- anatomy according to environmental variation, and vironmental stresses: implications for climatic change and partly explains its commonness in a broad range of conservation. Trends in Ecology and Evolution 5(9): 315– neotropical habitats. This intrinsic ability of common 317. species to adjust to distinct ecological extremes is Roˆ c¸as G, Barros CF, Scarano FR. 1997. Leaf anatomy an often neglected component of biodiversity in the plasticity of Alchornea triplinervia (Euphorbiaceae) tropics. under distinct light regimes in a Brazilian montane 238 G. ROˆ C¸ AS ET AL.

Atlantic rain forest. Trees Structure and Function 11(8): Sojka RE. 1992. Stomatal closure in oxygen-stressed plants. 469–473. Soil Science 154(10): 269–280. Rozema J, Chardonnens A, Tosserams M, Hafkenscheid Strauss-Debenedetti S, Berlyn GP. 1994. Leaf anatomical R, Bruijnzeel S. 1997. Leaf thickness and UV-B absorbing responses to light in five tropical Moraceae of different pigments of plants in relation to an elevational gradient successional status. American Journal of Botany 81: 1582– along the Blue Mountains, Jamaica. Plant Ecology 128: 1591. 150–159. Tilman D. 1988. Plant strategies and the dynamics and Sa´ CFC. 1992. A vegetac¸a˜o da Restinga de Ipitangas, Reserva structure of plant communities. New Jersey: Princeton Uni- Ecolo´gica Estadual de Jacarepia´, Saquarema (RJ); fisio- versity Press. nomia e listagem de angiospermas. Arquivos do Jardim Turner NC. 1986. Crop water deficits: a decade of progress. Botaˆnico do Rio de Janeiro 31: 87–102. Advances in Agronomy 39: 1–51. Scarano RF, Ribeiro KT, Moraes LFD, Lima HC. 1997. Turner IM. 1994. Sclerophylly: primarily protective? Func- Plant establishment on flooded and unflooded patches of tional Ecology 8: 669–675. freshwater swamp forest in south-eastern Brazil. Journal Vogelmann TC, Martin G. 1993. The functional significance of Tropical Ecology 14: 793–803. of palisade tissue: penetration of directional versus diffuse Schlichting CD. 1986. The evolution of phenotypic plasticity light. Plant Cell and Environment 16: 65–72. in plants. Annual Review Ecological Systematic 17: 667– Witkowski ETF, Lamont BB. 1991. Leaf specific mass 693. confounds leaf density and thickness. Oecologia 88: 486– Schreiber L, Riederer M. 1996. Ecophysiology of cuticular 493. transpiration: comparative investigation of cuticular water Zagdanska B, Kozdo´ j J. 1994. Water stress-induced permeability of plant species from different habitats. Oe- changes in morphology and anatomy of flag leaf of spring cologia 107: 426–432. wheat. Acta Societatis Botanicorum Polonie 63(1): Smirnoff N, Crawford RMM. 1987. Variation in the 61–66. structure and response to flooding of root aerenchyma in Zar JH. 1996. Biostatistical analysis. 3rd ed. New Jersey: some wetland plants. Annals of Botany 51: 237–249. Prentice Hall.