CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Available online at www.sciencedirect.com

South African Journal of Botany 81 (2012) 61–70 www.elsevier.com/locate/sajb

Water use strategies of seedlings of three Malagasy species under drought ⁎ T. Randriamanana a, , F. Wang a, T. Lehto b, P.J. Aphalo a

a Department of Biosciences, P.O. Box 65 (Viikinkaari 1), 00014 University of Helsinki, Finland b School of Forest Sciences, University of Eastern Finland, P.O. Box 111, 80101, Finland

Received 21 August 2011; received in revised form 17 May 2012; accepted 25 May 2012 Available online 21 June 2012

Abstract

Adansonia species, known by the common name “baobab,” have a very low regeneration rate in Madagascar. In order to determine if Malagasy Adansonia seedlings' vulnerability to drought may account for this low rate of regeneration, we compared growth, photosynthetic behavior and water use strategy of three species of Malagasy Adansonia (A. grandidieri, A. madagascariensis, A. rubrostipa). Our results indicated that drought depressed the growth, net assimilation rate, stomatal conductance and transpiration rate of Adansonia seedlings but increased their water use efficiencies. Adansonia species are able to withstand drought by reducing water loss through stomatal closure and their ability to store water within roots. Interspecific differences were attributed to diversity in water-use strategies, relative water content and biomass allocation. A. rubrostipa and A. grandidieri appeared to be more adapted to arid environments than A. madagascariensis. Ecological implications of these results are discussed. © 2012 SAAB. Published by Elsevier B.V. All rights reserved.

Keywords: Baobabs; Growth; Madagascar; Photosynthetic responses; Water stress; Water use

1. Introduction Malagasy western forests. Species of the genus Adansonia are classified as “keystone mutualists” due to their likely role in Tropical dry forests have received less conservation and stabilizing dry forest ecosystems (Baum, 1996). Apart from their scientific attention in comparison with tropical wet forests ecological role (Baum, 1995, 1996; Metcalfe and Trevelyan, (Fajardo et al., 2005; WWF, 2010). Over 97% of Madagascar's 2007), they are multipurpose trees and have numerous economic western dry deciduous forests have been lost and are consequently functions in some Malagasy rural communities (Baum, 1996; considered as “Critical or Endangered” ecoregion according to the Marie et al., 2009; Seddon et al., 2000; Wickens and Lowe, Global 200 representation of WWF (Olson and Dinerstein, 1998; 2008). In Madagascar, Adansonia fruit pulps which are rich in WWF, 2010). vitamin C and with high antioxidant activity (Assogbadjo et al., Adansonia (BOMBACACEAE sensu stricto, 2008; Chadare et al., 2009) are consumed by local communities. sensu lato), commonly known as “Baobab”, is a xeric genus that Moreover, Malagasy Adansonia leaves possess higher nutritional symbolizes Madagascar's floral biodiversity, as six out of eight value in terms of leaf vitamin and crude protein contents when world-wide species are endemic to the island. It was chosen as a compared to African baobab A. digitata L. (Maranz et al., 2007), model species in our studies because of its predominance in and can be consequently used to improve local population diet. In Madagascar, dried leaves of Adansonia are used for food ⁎ seasoning (Marie et al., 2009). Besides, Adansonia leaves are Corresponding author at: Natural Product Research Laboratory, Department of used in Malagasy traditional pharmacopeia, its bark and fibers are Biology, University of Eastern Finland, PO Box 111, 80101 Joensuu, Finland. Tel.: +358 504423402. confectioned to make rope, mats, baskets and roofing materials, E-mail address: tendry.randriamanana@uef.fi (T. Randriamanana). their seeds produce edible oil, and medicines (Baum, 1996; Marie

0254-6299/$ -see front matter © 2012 SAAB. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.sajb.2012.05.005 62 T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70 et al., 2009; Seddon et al., 2000; Wickens and Lowe, 2008). Their water for 12 h. Seeds were then sown at 1 cm depth in pots (at a trunks are sometimes hollowed out as tanks for water storage in density of 30 seeds per pot of 50 cm×30 cm dimension) dry areas of Madagascar (Baum, 1996; Marie et al., 2009). In containing sterile sand moistened with distilled water. The pots addition, due to their impressive appearance, species of the genus were kept at 30±1 °C under a 12 h photoperiod, illuminated with − − Adansonia are associated with some Malagasy superstitions and 175 μmol m 2 s 1 photosynthetic photon flux from fluorescent folklore which consider certain trees as sacred (Baum, 1995; lamps at ambient CO2. After 2 to 3 days, radicules emerged from Marie et al., 2009). almost 95% of the seeds and the germinating seeds were Field observations have reported a noticeably low survival transplanted to 410 mL pots containing vermiculite and white of seedlings in all species of Malagasy Adansonia especially in Sphagnum peat 1:1 (v/v). All pots were filled with the same arid areas of Madagascar (Baum, 1995, 1996; Miège and Morat, quantity of dry mixture (40 g perpot)andwerewateredevery 1974; Razanameharizaka, 2009; Wickens and Lowe, 2008). other day, ensuring that excess water drained through the 75% of newly established seedlings (particularly A. rubrostipa bottom of the pots. The content of fertilizer already in the limed Jum. & H. Perr.) die before their first rainy season in their natural peat (White 420 W, Kekkilä, Vantaa, Finland) was 1 g L− 1 of environment, apparently due to combined water deficit and high N–P2O5–K2O (14:9:24) plus micronutrients (Kekkilä starter light intensity (Razanameharizaka, 2009). Studies undertaken by fertilizer 1) plus 6.2 g L− 1 of dolomite lime. At the beginning Venter and Witkowski (2010) on African A. digitata reported that of the experiment, each seedling was watered with 40 mL the predicted drop in rainfall attributed to climate change may liquid fertilizer (Taimi Superex®, Kekkilä, Vantaa, Finland) increase Adansonia trees rate of mortality in the future. containing 19% N, 4.4% P, 20.2% K, 1.2% Mg, 0.03% B, The goal of this study is to investigate the physiological 0.001% Co, 0.008% Cu, 0.17% Fe, 0.08% Mn, 0.005% Mo, responses of Malagasy Adansonia seedlings to water stress and and 0.012% Zn (w/w) at a concentration of 1 g L− 1. whether such responses may explain their low regeneration rate Versatile MLR-350 294L test chambers (Sanyo Electric and their pattern of distribution across the island. We tested the Biomedical Ltd., Japan) were used to grow the seedlings under hypothesis that species-specific differences influence the effects of controlled environmental conditions. Photosynthetic photon drought on seedling growth and photosynthetic activity in three irradiance inside the growth chambers was measured using a species of Adansonia endemic to Madagascar, and predicted that LI-190 quantum sensor plus a LI-250A meter (LI-COR®, seedlings of Adansonia madagascariensis Baill. from the northern Biosciences, Lincoln, NE, USA). Photosynthetically Active dry deciduous and subhumid forests of Madagascar would be Radiation (PAR) photon irradiance was set at 175 μmol m−2 s−1 less drought tolerant during water stress when compared to with a photoperiod of 12 h. Temperatures were held at 29±1 °C A. rubrostipa and A. grandidieri Baill. which are from the arid during the day and 24±1 °C during the night, while air humidity spiny thickets and succulent woodlands of the south. To test this ranged between 26% during the day and 31% at night. hypothesis, we assessed under controlled conditions the photo- Temperature and relative humidity were logged at 10–15 cm synthetic behavior and water use strategy of these three species of above height in all chambers (iButton, DS1922L and Adansonia. DS1923, Maxim Integrated Products, Sunnyvale, CA, USA). Pots were assigned to three different complete randomized 2. Materials and methods blocks based on seedling size so that all chambers would have equal numbers of same-height seedlings. After transplantation Seeds from three species of Adansonia (A. rubrostipa, in pots, seedlings were allowed to grow for 28 days before A. grandidieri,andA. madagascariensis) were collected from the watering treatment so that they acquire well developed leaves western part of Madagascar at Ifaty, Marofandilia and Mahajanga and roots. Three replicates of 10 seedlings per species and per (Table 1; Fig. 1)by“URP Forêts et Biodiversité” (Madagascar). watering regime were grown in five separate growth chambers; each chamber contained two species. 2.1. Plant material establishment and experimental design 2.2. Watering treatment The experiment was carried out at the Department of Biosciences, University of Helsinki (Finland). Seeds of A. Starting 28 days after germination, the seedlings were subjected rubrostipa and A. madagascariensis were scalded in boiling to two water regimes: watering to 100% (“control”) and to 50% water for 60 s while seeds of A. grandidieri were soaked in tap field capacity (“water stressed”). Field capacity was determined in

Table 1 Seed origin, altitude, annual rainfall, mean annual maximum and minimum temperatures for the three species of Adansonia collected in the western part of Madagascar in 2004. Seed origin Altitude (m) Annual rainfallª (mm) Mean annual maximum Mean annual minimum temperature a (°C) temperature a (°C) A. madagascariensis Mahajanga (15°40′S, 47°03′E) 89 2236 31.93 18.2 A. grandidieri Ifaty (20°08′S, 44°33′E) 36 1156 31.72 20.85 A. rubrostipa Marofandilia (23°08′S, 43°36′) 16 302 30.65 20.19 a Source: National meteorology service, 2004. T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70 63

Fig. 1. Map of Madagascar's primary vegetation with occurrence zones (black symbols) of Adansonia rubrostipa (represented as stars), A. madagascariensis (represented as circles) and A. grandidieri (represented as triangles). Sources: adapted from Moat and Du Puy (1997) “With the permission of the Trustees of the Royal Botanic Gardens, Kew” and MBG (2009). pots containing overwatered mixture and allowed to drain in the bulked dried and ground leaf blades at the Waikato Stable overnight. To reach field capacity, 200 mL water per pot was Isotope Unit (University of Waikato, New Zealand) using a gas needed. For water-stressed and control treatments, pots were chromatograph and magnetic sector mass spectrometer GC-MS enclosed within plastic bags sealed at the stem base to allow (Europa Scientific Ltd., England). The main objective of this measurement of water use. The control seedlings were weighed measurement was to confirm that all three species use the C3 every second day, checked and watered to reach the 100% field photosynthetic pathway. capacity by replacing the amount of water transpired. Stressed were re-watered to reach only 50% of field capacity. This 2.4. Water use measurements water content was reached in most pots. Pots and their contents (soil plus plant) were weighed, then 2.3. Stable carbon isotope analysis were re-watered every second day for 6 days. For each seedling, the difference between the weight after watering the pot in the Leaf samples from well-watered seedlings of all three species previous weighing and the weight before watering in the current (n=12) were oven-dried at 60 °C and then used for stable carbon weighing was used to calculate water use in the intervening 2 days. isotope analysis. Carbon isotope composition (δ13C) was measured Water use was measured by recording the amount of water needed 64 T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70 to restore initial soil water content (50% or 100% field capacity 2.8. Growth measurements depending on the watering treatment). Water use measurements were done three times during the watering treatment: after 2, 4, and Stem diameter was measured with digital calipers at 7 and 6 days of water stress. 14 days of drought conditions for all species and both watering regimes (n=144). Seventeen days after the beginning of watering treatments, stem, leaves, and roots of all the seedlings were 2.5. Gas exchange measurements harvested and oven dried at 105 °C for 48 h to a constant weight. Thereafter, leaf weight ratio (LWR), stem weight ratio (SWR), For the gas exchange measurements, 35- and 42-day old and root weight ratio (RWR) were calculated according to the seedlings were used (n=48). Light response curves were first equations: measured in order to determine optimal conditions of measure- ments. Light response curves were obtained by measuring net LWR ¼ dry weight of leaves=total dry weight assimilation rate on the youngest fully expanded leaves at varying SWR ¼ dry weight of stems=total dry weight μ −2 −1 increasing levels of PAR (from 0 to 1000 mol m s ). RWR ¼ dry weight of roots=total dry weight Stomatal conductance (gs), net photosynthesis rate (A), and transpiration rate (E) were measured using a GFS-3000 gas 2 Specific leaf area was calculated from the same samples as used exchange system (Heinz Walz GmbH, Germany), with a 4 cm for RWC measurements according to SLA=area/ dry weight. aperture plate in the leaf cuvette, and red plus blue LEDs light source (3040-L, Heinz Walz GmbH). Gas exchange measure- 2.9. Statistical analysis ments were performed on the youngest fully expanded leaves of the seedlings from 1000 h to 1300 h after 1 week and after Two-way analysis of variance (ANOVA), for testing signifi- 2 weeks of watering treatment. Measurements were undertaken at cance of main effects and interactions, were undertaken with SPSS μ −1 μ −1 370 mol mol CO2,flowrateof750 mol s , a saturation 16.0 software. For significant effects, differences between species μ -2 −1 photon flux density of 1000 mol m s , 30 °C, 54% relative were compared using Tukey's HSD test for every growth and humidity (RH) and 20.5 Pa/kPa leaf to air vapor pressure photosynthetic parameter. Differences were considered significant difference (VPD). Water Use Efficiency (WUE) was calculated at the Pb0.05 level. Error bars represent standard error of mean as WUE=A/E. (±SE).

2.6. Chlorophyll content measurements (Chl) 3. Results

Chlorophyll content of fresh leaves was determined using a 3.1. Carbon isotope composition SPAD-502 chlorophyll meter (Konica Minolta sensing Inc., δ13 ≤ Japan) at the beginning and at the end of the watering treatment. The C of the three species differed significantly (P 0.001). δ13 A calibration curve for the SPAD readings was established The C of well-watered seedlings belonging to the three from spectrophotometric chlorophyll measurements in N, Adansonia species were typical of C3 photosynthetic metabolism. δ 13 − ‰ N-dimethylformamide extracts from discs 6.4 mm in diameter The CofA. rubrostipa ( 27.01 )waslowerthanthoseof − ‰ − ‰ obtained from one leaf per seedling of the three species (n=144). A. madagascariensis ( 24.35 ), and A. grandidieri ( 25.65 ). The equation (Chl)=10^ (SPAD^0.267) was used and the calibration curve was fitted using the procedure of Markwell et 3.2. Water use (WU) al. (1995), yielding a calibration coefficient which is very similar to that obtained by Markwell et al. for other species (Glycine max Water use values varied from 4.94 to 32.96 mL per day for and Zea mays). A. rubrostipa, from 10.91 to 41.67 mL per day for A. madagascariensis andfrom6.74to33.15mLperdayforA. grandidieri (from water-stressed to well-watered seedlings). 2.7. Leaf water status measurements (RWC) Interaction between species and watering treatment was significant (P ≤ 0.05). WU was significantly higher with Fourteen days after the start of watering treatment, leaf relative well-watered seedlings (P ≤ 0.001). The average WU of A. water content for each species (n= 60) was measured according to grandidieri and A. rubrostipa were significantly lower than that the formula: of A. madagascariensis (Fig. 2). RWC=100(Freshweight− Dry weight)/(Full-turgor weight− Dry weight) 3.3. Gas exchange measurements Discs of 6.4 mm diameter were cut from leaves and weighed (fresh weight), then incubated in darkness and at room Light curves showed light compensation points of about temperature in Petri dishes containing distilled water for 12 h 20 μmol m−2 s−1 that were similar for the three species as shown and weighed again to obtain the full-turgor weight. Finally, disks in Fig. 3A–C. For all gas exchange measurements, interaction were oven-dried at 85 °C for 24 h and weighed to obtain dry effects between main factors were not statistically significant weight. (P=0.05) except for the drought duration×watering treatment T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70 65

significantly higher in well-watered seedlings compared to water- stressed ones when subjected to drought during 7 or 14 days (Fig. 4A–B–C). The same effect was found with transpiration rate (Pb0.001; Fig. 4D–E–F). Additionally, transpiration rate of all species slightly increased from the first to the second week of drought but such changes were not statistically significant. In a species comparison, the lowest transpiration rate was found in A. rubrostipa regardless of watering treatment. Stomatal conduc- tance decreased markedly (Pb0.001) with drought for all species (Fig. 4G–H–I). Decrease in all those three parameters led to an Fig. 2. Water use per seedling (WU) of three Adansonia species. Each measurement increase in Water Use Efficiency which was then significantly is the mean of three measurements after 2, 4, and days of drought. Mean±SE b (n=144 observations on 24 seedlings per species and treatment). higher (P 0.001) in all water-stressed conditions of all three species (Fig. 4J–K–L). interaction of the assimilation rate parameter (Table 2) which was 3.4. Chlorophyll content (Chl) significant (P≤0.01). Moreover, no statistical effects of either drought duration (7 vs. 14 days) or interspecific variations were Species×watering treatment×drought duration interaction as observed. In addition to that, net assimilation rate (P≤0.001) was well as species×watering treatment interaction were not statisti- cally significant. Species×drought duration interaction was significant at P=0.03.A. grandidieri displayed the highest amount of chlorophyll among the three species (Fig. 5A–B–C, Pb0.001). Moreover, chlorophyll content increased after 14 days for both watering regimes (Pb0.001).

3.5. Growth and leaf water status measurements

For all growth and leaf water status parameters, drought duration×species×watering treatment interaction, species×water- ing treatment interaction, and species×drought duration interac- tion were not statistically significant. Watering treatment×drought duration interaction was statistically significant (P≤ 0.001). The stem diameter increased significantly with chronological age among the well-watered seedlings (P ≤0.001), while it remained invariable in water-stressed seedlings whatever their ages (Fig. 6A–B–C). A. grandidieri and A. madagascariensis displayed significantly higher stem diameters than A. rubrostipa (P≤0.001). As interaction between species and watering treatment factors was not significant for RWC and all biomass related parameters, data were pooled and the variation attributed to species (Table 3) and drought (Table 4)wereseparately assessed. The highest RWC was found with A. rubrostipa (62.62%). A. grandidieri and A. rubrostipa invested significantly more biomass in roots and in stems than did A. madagascariensis (P≤0.001). This latter species allocated more biomass to leaves compared to the two other species (P≤0.001). Relative water content was higher for well-watered seedlings (P≤0.01). Com- pared to water-stressed seedlings, well-watered ones allocated significantly higher amounts of biomass to stems than they did to the other parts of the plants (P≤0.01). Drought did not affect SLA but the three species differed significantly from each other. A. madagascariensis had the highest SLA, A. grandidieri had intermediate SLA and A. rubrostipa the smallest SLA (Table 3). Fig. 3. Typical response of net assimilation rate to PAR irradiance in Malagasy SLA was 50% larger in A. madagascariensis than in A. Adansonia grandidieri (A), A. madagascariensis (B) and A. rubrostipa (C) seedlings rubrostipa, while SLA in A. grandidieri differed from that in under well-watered conditions. A. rubrostipa by only 9%. 66 T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70

Table 2

Comparison of P-values from factorial ANOVA of Net assimilation rate (A), Transpiration rate (E), Stomatal conductance (gs), Water Use Efficiency (WUE) in terms of drought duration, species, watering treatments, and drought duration×species, drought duration×watering treatment, species×watering treatment, drought duration×species×watering treatment interactions in Adansonia. − 2 −1 − 2 − 1 −2 −1 − 1 A (μmol m s ) E (mmol m s ) gs (mol m s ) WUE (mol mol ) Drought duration 0.57 ns 0.89 ns 0.09 ns 0.02* Species 0.16 ns 0.07 ns 0.09 ns 0.32 ns Watering treatment 0.000*** 0.000*** 0.000*** 0.000*** Drought duration×Species 0.17 ns 0.74 ns 0.65 ns 0.91 ns Drought duration×Watering treatment 0.008** 0.72 ns 0.84 ns 0.94 ns Species×Watering treatment 0.58 ns 0.26 ns 0.19 ns 0.50 ns Drought duration×Species×Watering treatment 0.91 ns 0.86 ns 0.94 ns 0.59 ns * 0.01bPN0.05; **, 0.001bPN0.01 ; ***, P ≤0.001 ; ns, non-significant.

Fig. 4. Net assimilation rate (A), Transpiration rate (E), Stomatal conductance (gs), Water Use Efficiency (WUE) in four well-watered and four drought-stressed seedlings/treatment/species of three Adansonia species: A. grandidieri (A–D–G–J), A. madagascariensis (B–E–H–K), A. rubrostipa (C–F–I–L) after 7 days and 14 days of drought treatment. Mean±SE (n=48). P-values shown in Table 2. T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70 67

Fig. 5. Leaf chlorophyll content (Chl) in well-watered and water-stressed Fig. 6. Seedling stem-base diameter in well-watered and drought-stressed seedlings of three Adansonia species: A. grandidieri (A), A. madagascariensis seedlings of three Adansonia species: A. grandidieri (A), A. madagascariensis (B), A. rubrostipa (C) after 7 days and 14 days of drought. Mean±SE (n=144). (B), A. rubrostipa (C) after 7 days and 14 days of drought. Mean±SE (n=144). P-values for watering regime (W) and for species (S) are shown in the upper P-values for watering regime (W) and for species (S) are shown in the upper right of the figure. right of the figure.

4. Discussion seedlings of all three species showed a light compensation point −2 −1 (20 μmol m s ) typical of C3 plants at 30 °C (Sharp et al., Previous studies reported that Malagasy adult Adansonia are 1984). structurally and morphologically adapted to drought (Chapotin et Like in most C3 species for which there are tradeoffs be- al., 2005, 2006a, b); however, despite several studies that focused tween maximizing growth and acclimation to survive (see for on responses of Adansonia digitata to drought (Cuni Sanchez et example Bacelar et al., 2007; França et al., 2000; Hessini et al., al., 2011; De Smedt et al., 2012; Sanchez et al., 2010), little is 2009; Wu et al., 2008; Yin et al., 2005) drought depressed known about Malagasy Adansonia seedling vulnerability and Adansonia seedling growth. As reflected by biomass measure- adaptation to water stress. To our knowledge, this is the first ments, seedlings invested less carbon in building aboveground paper dealing experimentally with such issue. organs particularly stems that are not actively involved in The use of isotope discrimination analysis ascertained that the photosynthesis to optimize their water balance during water three Adansonia species are C3 plants. Light response curves in stress. In our study, stem diameters of well-watered seedlings 68 T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70

Table 3 Comparison of relative water content (RWC), biomass, leaf weight ratio (LWR), stem weight ratio (SWR), root weight ratio (RWR), specific leaf area (SLA) and leaf area per seedling (L) between the three Adansonia species (using marginal means, because species×watering treatment interaction was not significant). Leaf area was not measured directly. Species RWC (%) Biomass (g) LWR SWR RWR SLA (cm2 g−1) L (cm2) A. grandidieri 42.91 2.34 0.26 0.35 0.39 78 47 A. madagascariensis 37.05 2.10 0.49 0.30 0.21 113 116 A. rubrostipa 58.39 1.59 0.30 0.34 0.35 85 41 PNFs *** *** *** *** *** *** – *0.01bPN0.05; **0.001bPN0.01; ***P≤0.001; ns, non significant ; –: not tested.

FS: species effect; PNFS indicates comparisons between three species. were significantly bigger than that of water-stressed seedlings for their survival. Adjustment of stomatal conductance allows for after 14 days of drought. In African A. digitata,stemswere optimization of carbon assimilation in relation to water supply reported to contain up to 20% of Adansonia seedlings total (WUE). Slight increase in assimilation rate, stomatal conduc- water content (De Smedt et al., 2012). Thus we might suggest tance, and transpiration rate in 14 days water-stressed seedlings that such differences are the result of a mere variation in stem when compared with 7 days water-stressed seedlings argues for water status and cell turgor. More precisely, stem water content of seedlings' ability to acclimatize through time. droughted seedlings might be lower than well-watered ones, We observed that well-watered A. rubrostipa seedlings had a resulting in a smaller stem diameter. Incidentally the same slightly less negative δ13C than the other two species considered; behavior is also found in adult Adansonia trees whose stem A. grandidieri displayed an intermediate behavior, while A. diameter decreased during the leaf flushing period in dry seasons madagascariensis had the lowest δ13C, which would suggest that when stored water in stems is used to maintain leaf turgor A. rubrostipa had lower WUE than A. grandidieri and A. (Chapotin et al., 2006a, 2006b). madagascariensis. However, we did not find any differences in Chlorophyll content was unaffected by drought in our WUE between the species. Because δ13C increases with growth experiment. This result is consistent with results from a related irradiances (Farquhar et al., 1989) we suggest that the reason why species, Ceiba pentandra (according to Baum et al., 2004) within a δ13C failed to predict accurately WUE is the relatively low-light rainforest mesocosm; C. pentandra showed changes neither in conditions of the growth chambers contrasting with the saturating pigment (e.g., chlorophyll a, chlorophyll b, violaxanthin) concen- irradiance that we used for gas exchange measurements. trations nor in chlorophyll a/b ratio during watering treatment A. rubrostipa used the smallest amount of water, A. (Rascher et al., 2004). Drought had markedly reducing effects on madagascariensis the greatest, with A. grandidieri in between. all photosynthetic parameters and increasing effect on WUE in the These are whole-plant WU values which depend both on three Adansonia species. Same trends have also been recorded in stomatal conductance and the amount of foliage per plant. Due A. digitata (DeSmedtetal.,2012), Pseudobombax septenatum to technical issues, we did not scale WU with leaf area (Hogan et al., 1995), Ceiba pentandra (Rascher et al., 2004), Ceiba (L), but from leaf dry weights and SLA, it can be inferred samauna (Slot and Poorter, 2007) that are all included in that A. madagascariensis had approximately twice as much MALVACEAE family. As in our studies, DeSmedtetal.(2012) leaf area than the other two species (Table 3). The leaf dry did not find photosynthetic differences in seedlings from different weights were 0.83–1.19 g for A. madagascariensis,0.36–0.58 g geographic provenances. for A. rubrostipa and 0.51–0.71 g for A. grandidieri. As SLA are Our results suggest that the main cause behind the observed supposed to be the same across foliage in seedlings of the same depression in photosynthetic rate was the reduced stomatal species, multiplying leaf dry weights by SLA values enabled us to conductance resulting from partial closure of stomata. One way approximate the value of L. Moreover, specific leaf area (SLA) for Adansonia seedlings to avoid water loss is then to close their calculated from leaf discs that we used from water content data stomata to limit water loss and to maintain a RWC high enough showed that SLA significantly differed between species. The larger leaf area (L) of A. madagascariensis than of A. rubrostipa and A. grandidieri together with small and non-significant differences Table 4 among species in light-saturated stomatal conductance, both in Comparison of relative water content (RWC), biomass, leaf weight ratio well-watered and drought conditions, allow us to infer that the (LWR), stem weight ratio (SWR), root weight ratio (RWR) between well-watered and water-stressed seedlings of the three Adansonia species (using marginal means differences in WU were mainly a consequence of differences in because species×watering treatment interaction was not significant). the amount of foliage. In relative terms, the differences between Watering treatment RWC (%) Biomass (g) LWR SWR RWR species were largest under drought when WU by A. rubrostipa was approximately one half that by A. madagascariensis. Well-watered 49.50 2.43 0.34 0.34 0.30 Water-stressed 42.74 1.59 0.34 0.32 0.32 The effect of drought on growth was similar for all species.

PNFT * *** ns * ns Plant dry mass at the end of the experiment was largest for * 0.01bPN0.05; **, 0.001bPN0.01 ; ***, P ≤0.001 ; ns, non‐significant. A. madagascariensis, indicating a faster growth rate, which can

FT: Watering treatment effect; PNFT indicates comparisons between watering be explained by a larger allocation of growth to foliage than in the regimes of all three species. other two species. This was accompanied by a reduced allocation T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70 69 to roots, results that are in accordance with A. digitata's responses cannot be ruled out. Dispersal elements may also be another factor to drought (Cuni Sanchez et al., 2011; De Smedt et al., 2012). that may account for this low rate of regeneration. For instance, it Apart from developing deep fine roots to access underground might be linked to limited seed dispersion (probably due to the lack water (like most drought-tolerant species), Adansonia seedlings of suitable agents of dispersion) or high seed predation rate. develop thick taproots enabling them to store water (De Another possibility also is that regeneration of A. rubrostipa Smedt et al., 2012; Wickens and Lowe, 2008). A. grandidieri depends on light brought by occasional canopy gaps as suggested and A. rubrostipa which both originate from drier environments by Metcalfe and Trevelyan (2007). allocated more resources to such root development in comparison In conclusion, Malagasy Adansonia seedlings of different with A. madagascariensis. A. madagascariensis also had the species use the same mechanism, mainly partial stomatal closure, lowest RWC in leaves, suggesting that the other two species had to withstand drought events; however, seedlings of the more some degree of leaf water retention. Our results are in accordance drought-tolerant species are structurally different with larger root with studies on Adansonia digitata which reported that Adansonia systems and proportionally less foliage. The short-term response morphology and biomass allocation depend on seedlings to drought leads to reduced stem diameter, biomass and growth provenances, and that seedlings from drier provenances have when seedlings face a need to maximize water use efficiency. higher water content and invest more biomass in their root With their higher root biomass, compound leaves and lower WU, system but less in foliage (Cuni Sanchez et al., 2011; De however, A. rubrostipa and A. grandidieri seedlings are better Smedt et al., 2012). adapted to the arid region in the southwestern part of the Island A. rubrostipa stands out with a significantly higher RWC and where they naturally co-exist; A. madagascariensis which does A. madagascariensis with the highest WU. A. rubrostipa and not present such characteristics is present in less arid areas of the A. grandidieri probably used water slightly more efficiently northwestern region. than A. madagascariensis. A. grandidieri had water use similar to A. rubrostipa but RWC and transpiration pattern similar to Acknowledgments A. madagascariensis. The differences between species seem to be due mainly to differences in biomass allocation; higher allocation We are grateful to the “University of Helsinki's exchange to roots should be even more effective for drought tolerance in program” which provided T. Randriamanana with a grant for her plants growing in the open ground than in pots. All these results stay in Helsinki. The authors also thank “URP Forêts et suggest that A. madagascariensis is the least drought tolerant of Biodiversité” Madagascar, especially Dr. Pascal DANTHU for the three species and A. rubrostipa the most tolerant one, with A. having generously provided the seeds and helped with adminis- grandidieri having intermediate tolerance. Even though our trative processes needed for seeds exportation. We also thank results are limited by the fact that drought duration during our Dr. Aro Vonjy RAMAROSANDRATANA, (Laboratory of experiment was very short compared to the three- to five-month Plant Physiology, University of Antananarivo, Madagascar) for long drought under natural conditions these relative degrees of his valuable comments and Sharon Bowen, E.L.S. for her edits on drought tolerance correlate with the water availability in the the earlier versions of the manuscript. natural ranges of the species: A. madagascariensis is mostly found in subhumid regions (dry deciduous forests between Mahajanga and Antsiranana, Sambirano) and the two other species are found References in drier areas, with A. grandidieri subjected to higher rainfall than A. rubrostipa (Baum, 1995, 1996). The fact that A. rubrostipa is Assogbadjo, A.E., Glèglè Kakaï, R., Chadare, F.J., Thomson, L., Kyndt, T., the most tolerant species among the three might explain why it has Sinsin, B., Van Damme, P., 2008. Folk classification, perception, perception, and preferences of baobab products in West Africa: conse- greater ecological plasticity and a larger area of distribution than quences for species conservation and improvement. Economic Botany 62, the other two species. Therefore, drought might be a factor 74–84. influencing the distribution patterns of Adansonia in Madagascar Bacelar, E.A., Moutinho-Pereira, J.M., Gonçalves, B.C., Ferreira, H.F., Correia, Island. C.M., 2007. Changes in growth, gas exchange, xylem hydraulic properties From our results, we suggest that even though water stress and water use efficiency of three olive cultivars under contrasting water availability regimes. Journal of Experimental Botany 60, 183–192. limits growth and photosynthesis in Adansonia seedlings, their Baum, D.A., 1995. A systematic revision of Adansonia (Bombacaceae). Annals efficient water use strategy and biomass allocations enable of the Missouri Botanical Garden 82, 440–470. them to withstand drought. Moreover, in their natural environ- Baum, D.A., 1996. The ecology and conservation of the baobabs of Madagascar. ments A. rubrostipa has the lowest rate of regeneration while Primate Report 46, 311–326. A. madagascariensis and A. grandidieri have medium rates of Baum, D.A., Dewitt, S., Yen, A., Alverson, W.S., Nyffeler, R., Whitlock, B.A., Oldham, R.L., 2004. Phylogenetic relationships of Malvatheca (Bombacoideae regeneration (Razanameharizaka, 2009),whichisincontrastwith and Malvaoideae; Malvaceae sensu lato) as inferred from plastid DNA our results where A. rubrostipa is the most adapted to drought. sequences. American Journal of Botany 91, 1863–1871. Therefore other environmental factors might explain the natural Chadare, F.J., Linnemann, A.R., Hounhouigan, J.D., Nout, M.J.R., Van Boekel, regeneration of Adansonia. For example, the environments where M.A.J.S., 2009. Baobab food products: a review on their composition and – Adansonia occur are subjected to frequent slash and burn practices. nutritional value. Critical Reviews in Food Science and Nutrition 49, 254 274. Chapotin, S.M., Razanameharizaka, J., Holbrook, M., 2005. Baobab trees Even though surveys undertaken by Marie et al. (2009) found (Adansonia) in Madagascar use stored water to flush new leaves but not to evidence that people are protecting adult and juvenile Adansonia support stomatal opening before the rainy season. The New Phytologist 169, trees, the probability of increased seedling mortality due to fire 549–559. 70 T. Randriamanana et al. / South African Journal of Botany 81 (2012) 61–70

Chapotin, S.M., Razanameharizaka, J., Holbrook, M., 2006a. A biomechanical Missouri Botanical Garden (MBG), 2009. Retrieved May 10, 2012; Accessed perspective on the role of large stem volume and high water content in through Tropicos at bhttp://www.tropicos.orgN. baobab trees (Adansonia spp; Bombacaceae). American Journal of Botany Moat, J., Du Puy, D.J., 1997. “Madagascar, Remaining Primary Vegetation”, 93, 1251–1264. Royal Botanic Gardens, Kew (Eds.), London; Retrieved May 24, 2012; Chapotin, S.M., Razanameharizaka, J., Holbrook, M., 2006b. Water relations of Accessed at bwww.kew.org/gis/projects/madagascar/veg_mapping.htmlN. baobab trees (Adansonia spp) during the rainy season: does stem water Olson, D.M., Dinerstein, E., 1998. The Global 200: a representation approach to buffer daily water deficits? Plant, Cell & Environment 29, 1021–1032. conserving the Earth's most biologically valuable ecoregions. Conservation Cuni Sanchez, A., De Smedt, S., Haq, N., Samson, R., 2011. Variation in Biology 12, 502–515. baobab seedling morphology and its implications for selecting superior Rascher, U., Bobich, E.G., Lin, G.H., Walter, A., Morris, T., Naumann, M., planting material. Scientia Horticulturae 130, 109–117. Nichol, C.J., Pierce, D., Bil, K., Kudeyarov, V., Berry, J.A., 2004. De Smedt, S., Cuni Sanchez, A., Van den Bilcke, N., Simbo, D., Potters, G., Functional diversity of photosynthesis during drought in a model tropical Samson, R., 2012. Functional responses of baobab (Adansonia digitata L.) rainforest—the contributions of leaf area, photosynthetic electron transport seedlings to drought conditions: differences between western and south- and stomatal conductance to reduction in net ecosystem carbon exchange. eastern Africa. Environmental and Experimental Botany 75, 181–187. Plant, Cell & Environment 27, 1239–1256. Fajardo, L., Gonzalez, V., Nassar, J.M., Lacabana, P., Portillo, Q., Carlos, A., Razanameharizaka, J., 2009. Régénération, Démographie, Physiologie de la graine Carrasquel, F., Rodriguez, J.P., 2005. Tropical dry forests of Venezuela: et des plantules du genre Adansonia à Madagascar, PhD dissertation, pp. 170. characterization and current conservation status. Biotropica 37, 531–546. Sanchez, A.C., Haq, N., Assogbadjo, A.E., 2010. Variation in baobab Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989. Carbon isotope discrimi- (Adansonia digitata L.) leaf morphology and its relation to drought nation and photosynthesis. Annual Review of Plant Physiology and Plant tolerance. Genetic Resources and Crop Evolution 57, 17–25. Molecular Biology 40, 503–537. Seddon, N., Tobias, J., Yount, J.W., Ramanampamonjy, J.R., Butchart, S., França, M.G.C., Thi, A.T.P., Pimentel, C., Rossiello, R.O.P., Zuily-Fodil, Y., Randrianizahana, H., 2000. Conservation issues and priorities in the Mikea Laffray, D., 2000. Differences in growth and water relations among Phaseolus Forest of south-west Madagascar. Oryx 34, 287–304. vulgaris cultivars in response to induced drought stress. Journal of Experimental Sharp, R.E., Matthews, M.A., Boyer, J.S., 1984. Kok effect and the quantum Botany 43, 227–237. yield of photosynthesis. Plant Physiology 75, 95–101. Hessini, K., Martínez, J.P., Gandour, M., Albouchi, A., Soltani, A., Abdelly, C., Slot, M., Poorter, L., 2007. Diversity of tropical tree seedling responses to 2009. Effect of water stress on growth, osmotic adjustment, cell wall elasticity drought. Biotropica 39, 683–690. and water-use efficiency in Spartina alterniflora. Journal of Experimental Venter, S.M., Witkowski, E.T.F., 2010. Baobab (Adansonia digitata L.) density, Botany 67, 312–319. size-class distribution and population trends between four land-use types in Hogan, K.P., Smith, A.P., Samaniego, M., 1995. Gas exchange in six tropical northern Venda, South Africa. Forest Ecology and Management 259, 294–300. semi-deciduous forest canopy tree species during the wet and dry seasons. Wickens, G.E., Lowe, P., 2008. The Baobabs: Pachycauls of Africa, Madagascar, Biotropica 27, 324–333. and Australia. Springer, London, p. 498. Maranz, S., Niang, A., Kalinganire, A., Konaté, D., Kaya, B., 2007. Potential to World Wildlife Fund (Lead Author), Michael Hogan, C., PhD. (Contributing harness superior nutritional qualities of exotic baobabs if local adaptation Author), Sahotra Sarkar, Mark McGinley (Topic Editor) "Madagascar dry can be conferred through grafting. Agroforestry Systems 72, 231–239. deciduous forests", In: Cleveland, C.J. (Eds.), Washington, D.C.: Environ- Marie, C.N., Sibelet, N., Dulcire, M., Rafalimaro, M., Danthu, P., Carrière, mental Information Coalition, National Council for Science and the S.M., 2009. Taking into account local practices and indigenous knowledge Environment. Encyclopedia of Earth. [First published in the Encyclopedia of in an emergency conservation context in Madagascar. Biodiversity and Earth April 19, 2010; Last revised Date November 5, 2010; Retrieved February Conservation 18, 279–2777. 25, 2011 bhttp://www.eoearth.org/article/Madagascar_dry_deciduous_forests? Markwell, J., Osterman, J.C., Mitchell, J.L., 1995. Calibration of the Minolta topic=49597N. SPAD-502 leaf chlorophyll meter. Photosynthesis Research 46, 467–472. Wu, F.Z., Bao, W.K., Li, L.F., Wu, N., 2008. Effects of drought stress and N Metcalfe, D.J., Trevelyan, R., 2007. Distribution and population structure of supply on the growth, biomass partitioning and water-use efficiency of Sophora Adansonia rubrostipa in dry deciduous forest in western Madagascar. davidii seedlings. Environmental and Experimental Botany 63, 248–255. African Journal of Ecology 45, 464–468. Yin, C., Wang, X., Duan, B., Luo, J., Li, C., 2005. Early growth, dry matter Miège, J., Morat, P., 1974. Etude du genre Adansonia L. II. Caryologie et allocation and water use efficiency of two sympatric Populus species as blastogenèse. Candollea 29, 457–475. affected by water stress. Journal of Experimental Botany 53, 315–322.

Edited by OM Grace