GROWTH , LEAF AND STOMATAL TRAITS OF CRABWOOD ( Carapa guianensis Aubl . ) IN CENTRAL AMAZONIA

Crabwood (Carapa guianensis Aubl.) is a fast growing tree species with many uses among Amazonian local communities. The main objective of this study was to assess the effect of seasonal rainfall pattern on growth rates, and seasonal and diurnal changes in leaf gas exchange and leaf water potential (ΨL) in crabwood. To assess the effect of rainfall seasonality on growth and physiological leaf traits an experiment was conducted in Manaus, AM (03° 05 ́ 30" S, 59° 59 ́ 35" S). In this experiment, six 6-m tall plants were used to assess photosynthetic traits and ΨL. In a second experiment the effect of growth irradiance on stomatal density (SD), size (SS) and leaf thickness was assessed in 0.8-m tall saplings. Stomatal conductance (gs) and light-saturated photosynthesis (Amax) were higher in the wet season, and between 09:00 and 15:00 h. However, no effect of rainfall seasonality was found on ΨL and potential photosynthesis (CO2-saturated). ΨL declined from -0.3 MPa early in the morning to -0.75 MPa after midday. It increased in the afternoon but did not reach full recovery at sunset. Growth rates of crabwood were high, and similar in both seasons (2 mm month-1). Leaf thickness and SD were 19% and 47% higher in sun than in shade plants, whereas the opposite was true for SS. We conclude that ΨL greatly affects carbon assimilation of crabwood by reducing gs at noon, although this effect is not reflected on growth rates indicating that other factors offset the effect of gs on Amax.

A espessura foliar e a S D foram 19% e 47% maiores nas folhas de sol, sendo o contrário observado quanto ao S S .Conclui-se que o Ψ L afeta a assimilação de carbono da andiroba pela redução da g s ao meio-dia, entretanto esse efeito não se reflete na taxa de crescimento, indicando que outros fatores podem compensar o efeito de g s em A max .

INTRODUCTION
Crabwood (Carapa guianensis Aubl.) is a fastgrowing, evergreen or deciduous canopy tree (HUC et al., 1994) that can establish in the forest understory (MCHARGUE; HARTSHORN, 1983), and it is often classified as a climax shade-tolerant species (FETCHER et al., 1994).It is native to tropical America and its distribution ranges from the West Indies and Central America to the Guyanas and Brazilian Amazonia.The tree is abundant in the Amazon flood plains in the States of Pará and Amazonas.It is a monoecious and insect-pollinated species, and in the juvenile stage large leaves are produced in the monocaulous, which decrease in size when branching is initiated.Trees of this species often reach over 1 m in diameter and over 40 m in height at maturity.It produces a wood of high quality and intermediate wood density, about 0.6 g cm -3 (BAUCH; DUNISH, 2000).The timber of crabwood is used for furniture, shingles, millwork, interior trim, boxes, crates, flooring, masts, rafters, sheathing, and other items used in house constructions; thereby, it is one of the most preferred substitute for mahogany (Swietenia macrophylla) in areas where it grows.The bark of this tree contains carapina, an alkaloid presumed to have medical properties, and tannins.Its seeds produce cream-colored and bitter oil (carapa oil) of high acidity which may be employed to cure many ailments including coughs and colds, whooping cough, and skin problems.Young trees grow much faster (10 mm yr -1 in diameter) than adult trees, which grow at a rate of about 2.5 mm yr -1 (BAUCH; DUNISCH, 2000;DUNISCH et al., 2002).Despite the importance of its oil and timber there is a dearth of information on leaf traits of crabwood, particularly on the effect of diurnal and seasonal changes in leaf water potential (Ψ L ) on stomatal functioning and photosynthetic characteristics.
Excessive water loss by transpiration as a result of a high leaf to air vapour pressure difference (VPD) may induce stomatal closure (REIS et al., 1988;MARENCO et al., 2006).In isohydric plants, stomata exhibit a strong control of transpiration so that Ψ L of droughted and watered plants remains constant during the day, whereas in anysohydric plants larger oscillations in Ψ L can be observed during the day, as a result of low stomatal sensitivity to evaporative demand and leaf water content (TARDIEU; SIMONNEAU, 1998).
Understanding how Amazonian rainforest trees, such as crabwood, react to changes in Ψ L is important because some climate models predict an increase in temperature and a decline in rainfall in decades to come (MYERS, 1988) which may affect plant functioning in an altered environment.Thus, the main objective of this study was to assess the effect of rainfall seasonality on growth rates, leaf gas exchange and Ψ L in Carapa guianensis.Stomatal characteristics under sun and shade conditions were also examined.

Study area and plant material
The study was conducted at the National Institute for Research in the Amazon (Campus V8) in Manaus, AM-Brazil (03° 05´ 30" S; 59° 59´ 35" W).The city of Manaus has average temperature of 26.7 °C, average rainfall of 2260 mm (mean of 1961(mean of to 1990(mean of , Inmet, 2010)), and a relative humidity of 83%.The region has characteristics of humid equatorial climate, with a short mild dry season (July-September, with a rainfall of 50-100 mm per month).The wet season extends from November to May (200-300 mm month -1 ), whereas October is a dry-wet transition period.Thus, during most of the year (October to June) precipitation exceeds evaporation.Mean temperature oscillates from 26.1 °C in January-March to 27.5 °C in September-November.Over the year, maximum mean temperatures vary from 30.6 °C (January-March) to 32.6 °C (August-November).Minimum mean temperature is 23.3 °C with little variation over the year.Monthly relative humidity ranges between 75-80% in the dry season to 81-87% in the wet season (Inmet, 2010).
We conducted two experiments.In the first experiment, we assessed the effect of diurnal and rainfall seasonality on leaf gas exchange and Ψ L .In the second experiment, we determined the effect of leaf growth irradiance on anatomical leaf traits.
In the first experiment, six open-growth trees of Carapa guianensis Aubl.(Meliaceae) about 6 m in height and 50-80 mm in diameter were used.Diameter growth (at 1.3 m, breast height) was recorded at monthly intervals between 2006 and 2008.Gas exchange and Ψ L data were collected during four consecutive days in both the dry (August) and wet (January) season of 2008.A double sided step ladder was used to reach canopy leaves for gas exchange measurements.In the Growth, leaf and stomatal traits of... Revista Árvore, Viçosa-MG, v.36, n.1, p.07-16, 2012 second experiment, carried out in 2009, a group of 10 saplings was grown from seeds in 10 L pots under shade conditions (0.5 mol m -2 day -1 and about 10 µmol m -2 s -1 at noon).The substrate used for plant growth was soil from a natural forest collected from the first 0.2 m from the soil surface.It was amended with 5 g kg -1 of a slow release fertilizer (Basacote®).When the plants were 0.8-m tall, a group of five plants was transferred to a full sunlight environment (26 mol m -2 day -1 and about 1200 µmol m -2 s -1 at midday) for 60 days.At the end of this period, leaf trait data were collected.

Gas exchange and water potential measurements (experiment 1)
Net photosynthetic rate (A) and stomatal conductance (g s ) were measured with a portable open gas exchange system fitted with infrared gas analyzer (Li-6400, Li-Cor, Lincoln, NE, USA) and an independent control of CO 2 and photosynthetic photon flux density (Q).Data were collected on fully expanded leaves of similar age and appearance at ambient temperature.The effect time of day on A and g s was assessed by collecting data at 1-h interval from 06:00 to 18:00 h.When g s was higher (09:00-15:00 h) and the system stable (coefficient of variation less than 0.7% at a Q of 500 µmol m -2 s -1 in the leaf chamber), A-C i and light-response curves [A(Q)] were constructed.The A-C i relationship was examined by measuring, A at 380 µmol mol -1 , then the [CO 2 ] was reduced stepwise to 250, 200, 150, 100, and 50 µmol mol -1 , and after that, it was increased to 380, 500, 1000 and 2000 µmol mol -1 while illumination in the leaf chamber was kept constant at 1000 µmol m -2 s -1 .To obtain A(Q) data, photon flux density (Q) was increased to 2000 µmol m -2 s -1 and A values recorded, then light intensity was decreased stepwise.The apparent quantum yield (Φ) was calculated by using data from the A(Q) response curve at low PFD values (< 80 µmol m -2 s -1 ).Stomatal limitation (L s ) to photosynthesis was calculated as follows (FARQUHAR; SHARKEY, 1982): where, A 0 is the photosynthetic rate at 380 µmol mol -1 when stomatal resistance is zero (i.e.C i = C a ) and A is the actual photosynthetic rate at a C a value of 380 µmol mol -1 (ambient air).After Jacob and Lawlor (1991), we calculated the mesophyll limitation as: L m = 100 (A C -A S )/ A C , where A C and A S are CO 2 -saturated A (C i > 800 mol mol -1 ) of control (wet season) and stressed leaves (dry season), respectively.
Leaf water potential (Ψ L ) was determined from 07:00 to 18:00 h at 1-h intervals using psycrometric sensors (L-51, Wescor, Utah, USA) connected to a datalogger (Psypro, Wescor, Utah, USA),as previously described (COSTA; MARENCO, 2007).Ψ L was measured in four leaves per plant and in the same plants used to measure photosynthetic rates.Leaf chlorophyll content was calculated after pigment extraction in 80%acetone (v/v) and the absorbance of chloroplast solution measured at 645 and 663 nm, as described elsewhere (MARENCO et al., 2009).

Stomatal characteristics in sun and shade plants (experiment 2)
To determine the effect of leaf growth irradiance on stomatal characteristics, particularly stomatal density (S D ), stomatal size (S S or guard cell length) and leaf thickness, a group of five shade-grown saplings was transferred to full sunlight for 60 days.The control was another group of five plants kept under low irradiance (0.5 mol m -2 day -1 ).Shade conditions were obtained by growing the plants under a nylon screen and beneath the canopy of adult trees.During daytime hours, temperatures oscillated from 26 to 32 °C outdoors and from 26 to 30 °C under shade conditions.Whereas relative humidity ranged between 50 to 91% in the sunny environment and 72 to 96% under shading.Stomatal characteristics were determined on fully expanded leaves and developed during the 60-dayirradiance treatment.S D and S S were determined on clear nail polish imprints (HILU; RANDALL, 1984) taken from both sides of the leaf surface, at the widest part of the leaf lamina.Stomata were counted in three leaves per plant under a light microscope (Leica DM500, Leica Microsystems, Wetzlar, Germany).S D was recorded in two 0.45-mm 2 fields of view per polish imprints of the leaf surface.S S was measured in a sample of 20 stomata per leaf at 1000x.To determine S S , an optical reticle (accuracy of 1 µm) was previously mounted on the eyepiece and calibrated against a micrometer slide (Nachet, Paris, France).The stomatal type was determined after dissociation of the leaf epidermis (a fragment of 100 mm 2 ) with a solution of water-free acetic acid (glacial acetic acid) and hydrogen peroxide (1:1 v/v) at 60 °C for 48 h (KRAUS; ARDUIN, 1997).Images were taken with a digital camera (Canon, Sx100-IS, China) coupled to the microscope.The stomatal type was defined following the criteria of Baranova (1992), by observing the arrangement of the subsidiary or neighbouring cells Revista Árvore, Viçosa-MG, v.36, n.1, p.07-16, 2012 immediately adjacent to the stomatal pore.Specific leaf area (SLA) was determined as the area to leaf mass ratio.Leaf area was measured with a portable leaf area meter (Li-3050, Li-Cor,NE, USA), whereas leaf mass was determined by using a standard laboratory balance (0.1 mg precision).Leaf thickness was measured as the transversal leaf section under microscope at 1000x magnification.Besides SLA, the sclerophyll index (S I , half dry mass to leaf area ratio) was also used to assess the effect of irradiance on leaf morphology (RIZZINI, 1976).To assess statistical significance of seasonal rainfall pattern and growth irradiance on variables examined, the Student´s test was used.Probability values of less than 0.05 were considered significant.The standard error (SE) was used to compare the effect of time of day on A max , g s and Ψ L .

RESULTS
Photosynthetic photon flux density (Q) was higher in the dry season than in the rainy season (Figure 1A) because of higher cloudiness in rainy months.Irrespective of season of the year, A and g s were higher between 09:00 and 15:00 h, indicating that stomata remained closed or partly closed early in the morning and at the end of the afternoon.Regardless of light intensity in the leaf chamber, g s followed the same trend as A (Figures 1B-C).A max , g s and the C i /C a ratio were higher in the wet season than in the dry months of the year (Table 1), whereas Φ, A pot , and the leaf chlorophyll content were not affected by rainfall seasonality.There was a curvilinear and highly significant relationship between A and g s (R 2 = 0.93, p < 0.01, Figure 2A) confirming that photosynthetic rates (A) are strongly dependent on g s .Irrespective of the C i /C a relationship, at low or very low g s the C i /C a ratio tended to increase (Figure 2B), particularly in the dry months suggesting the occurrence of stomatal patchiness.L m accounted for about 5% of carbon uptake limitation, whereas the photosynthetic constraint imposed by L s ranged from 43% in the wet to 64% in the dry season indicating that stomatal functioning was the main factor limiting carbon uptake in crabwood trees (Table 1).
Despite the effect of season of the year on A max , growth in stem diameter was similar in both seasons (around 2mm month -1 ) indicating that the apparent reduction in A max was offset by the high irradiance observed during the dry season, so that carbon gain at the plant level remained almost unchanged.Ψ L declined from -0.3 MPa early in the morning to about -0.75 MPa at noon without any significant effect of the season of year (Figure 1D).After midday, Ψ L increased but it did not reach full recovery at sunset.Stomata were found only on the abaxial surface (hypostomatic leaves).Sun leaves had a higher (90%) stomatal density (S D ) than shade leaves (Table 2); on the contrary, stomatal size (S S ) was larger in shade leaves (Table 2).Irrespective of growth irradiance stomata were located in epidermal depressions or crypts (Figure 3).The stomata were flanked by three or more subsidiary cells encircling the guard cell pair; anticlinal walls to separate adjacent subsidiary cells radiating from guard cells (laterocytic

DISCUSSION
The lower irradiance observed in this study in the wet season was not unexpected as a spell of cloudy days is not uncommon in central Amazonia during the rainy season.A max and g s showed a diurnal pattern with lower values early in the morning and at the end of the day.Several environmental factors may lead to variations in g s during the day, such as light intensity, [CO 2 ] and leaf water content (LAWSON et al., 2008).As CO 2 concentration in the leaf chamber, boundary layer conductance and light intensity remained constant during photosynthetic measurements we cannot attribute the stomatal closure observed early in the morning and late in the afternoon to variations in these environmental factors.On the other hand, diurnal oscillations in stomatal opening can be explained by changes in leaf water content (Figure 1D).We cannot rule out, however, the effects of circadian rhythms on stomatal functioning as they have been observed in other tree species in the Amazon (DOUGHTY et al., 2006;MAGALHÃES, 2010;MENDES;MARENCO, 2010).A max differences among seasons most likely reflect the effect of rainfall seasonality on g s .A max was about 41% of A pot during the rainy season, but it declined to only 27% in dry months.The drop in A max in the dry season can be attributed, at least in part, to a higher stomatal limitation (L s ) during the dry season (Table 1) when lower C i /C a values were observed.This hypothesis is supported by the fact that A pot remained almost unaltered between rainfall periods (p > 0.05, Table 1).Thus, we conclude that in crabwood, the response of A to leaf water content can be described as Type 1(LAWLOR; CORNIC, 2002), whereby a wild water deficit does not lead to a reduction in A pot .
The low A max in the dry season could reflect either low water availability in the soil or the effect of a high vapor pressure difference (VPD) on dry months, as stomata often tend to close when VPD is high (COSTA and MARENCO, 2007;EAMUS et al., 2008).The close correlation between A and g s (Figure 2A, r 2 = 0.93, p < 0.01) confirms that photosynthetic rates are strongly dependent on g s (p < 0.05), as reported by others (CHAVES et al., 2004;MENDES;MARENCO, 2010).Also, it is worth noting that at low or very low g s the C i /C a ratio was higher than it was expected (Figure 2B).This was attributed to non-uniform stomatal opening, which may cause actual C i values deviate from calculated ones, as reported by Marenco et al. (2006).
Although A max was lower in the dry season, growth rates remained rather constant along the year (Table 1) which suggests that the detrimental effect of the dry season on A max is offset by the high irradiance that prevails during the dry season (Figure 1A).Differences in A max among seasons are not probably related to variation in Φ or chlorophyll concentrations, as Φ and leaf chlorophyll content was similar in both seasons.The high growth rate found in this study is a little higher than the crabwood growth rate reported by Dunisch et al. (2002) for a monoculture plantation, a mean of 18.7 mm yr -1 during the first seven years of development.High growth rate observed in crabwood led us to classify crabwood as an intermediate species in the successional status, rather than shade tolerant as suggested by Fetcher et al. (1994).
In the afternoon, Ψ L increased but it did not reach full recovery at the end of the day.Diurnal variations in Ψ L found in this experiment are similar to those observed in other studies (BARKER; PEREZ-SALICRUP, 2000;FISHER et al., 2006).On average, Ψ L did not decrease below -0.8 MPa, which appears to be the minimum critical Ψ L for this species, at least under our experimental conditions.Similar diurnal patterns of Ψ L between rainfall seasons suggest that g s had an important role in controlling water loss via transpiration.This is in agreement with results reported by Fisher et al. (2006) for other Amazonian tree species In this experiment, the drop in Ψ L had an effect on both A max and g s , which was consistently lower in the dry season.The effect of a decline of Ψ L on A max was more pronounced in the afternoon, particularly in the dry season, as restrictions in stomatal aperture reduce carbon uptake.
Stomatal density and S S found in crabwood leaves are within the range of values reported for other hypostomatous leaves in the Amazon (CAMARGO; MARENCO, 2011).In comparison with shade leaves, under high irradiance (full sunlight) S D was higher and individual stomata were smaller.
Revista Árvore, Viçosa-MG, v.36, n.1, p. [7][8][9][10][11][12][13][14][15][16]2012 This probably occurs because under high irradiance, smaller stomata confer some adaptive advantage to the plant, as they can open and close more rapidly and thereby enhancing water use efficiency, the ratio of photosynthesis to transpiration (HETHERINGTON; WOODWARD, 2003).Indeed, the increase of S D and leaf thickness with irradiance seems to be a quite common response in several plant species (FAHN;CUTLER, 1992;VOLENIKOVÁ;TICHÁ, 2001;LIMA JUNIOR et al., 2006).As S D is somewhat related to g s and carbon uptake, it has been suggested that high leaf S D may contribute to improve biomass accumulation (AL AFAS et al., 2006).Stomatal encryptation has often been considered a xerophytic trait, as the location of stomata in depression reduces water loss by transpiration.However, it should be noted that sunken stomata also occurs in plants growing in wet and even flooded environments (JORDAN et al., 2008).This makes difficult to indubitably ascribe the presence of crypts to variations in water regime during plant growth.Indeed, it has been suggested that crypts have a role in facilitating CO 2 diffusion from abaxial to adaxial palisade cells (HASSIOTOU et al., 2009).Although the correlation between S D and S S was significant (r = 0.65, p < 0.05), it seems that encrypted stomata tend to be less responsive to variation in size, due to reduced peristomatal transpiration, perhaps because in comparison with non-sunked stomata, guard cells of encrypted stomata are less prone to suffer dehydration.
Under sunlight, SLA values were 90% higher than those observed under shading.On the contrary the S I , an indication of the amount of vascular and sclerenchyma tissue in the leaf, increased by 140% in sun leaves.Both low SLA and high S I values indicate a tendency to xeromorphism (WITKOWSKI; LAMONT, 1991), thereby to improved water use efficiency.In addition, as an inverse relationship between wind speed and SLA has been found (WOODWARD, 1983), it seems that a high proportion of hard sclerenchymatous tissue in a leaf may reduce structural injuries under dry and windy conditions.We concluded that although A max is higher in the wet than in dry season, monthly growth rate remains unaltered along the year, most likely due an increase in irradiance in the dry season.The tight control of g s during the day seems to be responsible for the relative constancy of Ψ L values between seasons, so we classified C. guanensis as an isohydric species.
The high growth rate of crabwood and its capability to adjust leaf traits in response to variation in light environment makes of crabwood a good choice in reforestation programs, where fast growing species are recommended for accelerating forest recovery and successional processes of deforested lands.Besides, such species may also contribute to effectively mitigate the effect of CO 2 deposition in the atmosphere.Finally, as a remarkable change in S I and other leaf traits were observed in crabwood in response to variations in environment during growth, it can be suggested that this species has a promising potential to acclimate to the altered environment predicted for the near future for most part of the Amazon region.

Figure 2 -Figure 3 -
Figure 2 -Photosynthetic rates, A (A) and the C i /C a ratio (B) as a function of g s in the dry ( , open circle) and rainy season ( , closed circle).High values of C i /C a at low g s may indicate non-uniform stomatal aperture.** Significant at p < 0.01.Figura 2 -Taxas fotossintéticas, A (A) e razão C i /C a (B) em função de g s nas estações seca ( , círculos abertos) e chuvosa ( , círculos fechados).Altos valores de C i /C a em baixa g s indicam uma não uniforme abertura estomática.** Significativo com p < 0,01.

Table 1 -
Photosynthetic traits, chlorophyll content and diameter growth rates in crabwood in the dry and wet season (mean ± standard deviation).