DROUGHT UNDER NATURAL CONDITIONS AFFECTS LEAF PROPERTIES, INDUCES CAM AND PROMOTES REPRODUCTION IN PLANTS OF Talinum triangulare

María Angélica Taisma and Ana Herrera

María Angélica Taisma. Biologist, Universidad Central de Venezuela (UCV). D.Sc. in Botany, UCV. Lecturer, Laboratory of Plant Morphology and Anatomy, Instituto de Biología Experimental (IBE-UCV). Address: Apartado 47114, Caracas 1041 A. e-mail: ataisma@cantv.net

Ana Herrera. Biologist, UCV. Ph.D. in Plant Sciences, University of London. Professor, Laboratory of Xerophyte Ecophysiology (IBE-UCV). Address: Apartado 47577, Caracas 1041 A. e-mail: aherrera50@netscape.net

Summary

Talinum triangulare is a perennial, deciduous shrub commonly found in tropical seasonally-dry environments. In this study an integrative assessment of the responses of this species to drought is provided. Water availability was manipulated under natural conditions and seasonal changes followed in a range of leaf characteristics and reproductive output. The leaf responses measured included rolling, paraheliotropic movements, increase in reflectance and induction of Crassulacean acid metabolism (CAM). Plants growing in a seasonally dry location and subjected to the natural rain regime were studied for approximately one year; 30 plants of this population were frequently watered. Nocturnal acid accumulation and leaf rolling, angle and reflectance were higher throughout the year, except during the rainy season, in plants receiving the natural water supply than in watered plants. The former had a higher production of flowers and fruits than experimentally watered plants. No significant differences between the two groups of plants were found in the frequency of pollinator visits. Leaf angle, rolling and reflectance, and the ratio reproductive mass/leaf mass showed positive correlations with night-time acid accumulation (r²=0.83, 0.64, 0.59 and 0.64, respectively). The reproductive mass/leaf mass ratio was positively correlated with leaf rolling (r²=0.51) and leaf reflectance (r²=0.79) but not with leaf angle. Results suggest that drought causes the increase in leaf angle, rolling and reflectance, as well as CAM activity, and serves as a cue for the increase in fecundity in plants of T. triangulare growing in the field.

Resumen

Talinum triangulare es un arbusto perenne y caducifolio que crece comúnmente en ambientes tropicales estacionalmente secos. Este estudio aporta un enfoque integrador de las respuestas de esta especie a la sequía. La disponibilidad de agua fue manipulada en condiciones naturales y se midieron los cambios estacionales en un grupo de características foliares y órganos reproductivos. Las respuestas foliares medidas incluyen enrollamiento, movimientos paraheliotrópicos, aumento de reflectividad e inducción del metabolismo ácido de Crasuláceas (CAM). Durante aproximadamente un año se estudiaron plantas que crecían en una localidad estacionalmente seca, sujetas al régimen natural de lluvias; 30 de estas plantas fueron regadas frecuentemente. La acumulación nocturna de ácidos y el enrollamiento, ángulo y reflectividad foliar fueron mayores durante todo el año, excepto durante la época de lluvias, en plantas con suministro natural de agua que en plantas regadas. Las primeras tuvieron mayor producción de flores y frutos que las plantas experimentalmente regadas. No se encontraron diferencias significativas en la frecuencia de visitas de polinizadores a ambos grupos de plantas. El ángulo, enrollamiento y reflectividad foliar, y el cociente masa reproductiva/masa foliar mostraron correlaciones positivas con la acumulación nocturna de ácidos (r²=0,83; 0,64; 0,59 y 0,64 respectivamente). El cociente masa reproductiva/masa foliar se correlacionó positivamente con el enrollamiento foliar (r²=0,51) y la reflectividad (r²=0,79) pero no con el ángulo foliar. Los resultados sugieren que la sequía ocasiona un aumento en el ángulo, enrollamiento y reflectividad foliar, así como en la actividad del CAM, y actúa como una señal para el aumento de la fecundidad en plantas de T. triangulare en el campo.

Resumo

Talinum triangulare é um arbusto perene e caducifólio que cresce comumente em ambientes tropicais com estações secas. Este estudo aporta um enfoque integrador das respostas desta espécie à seca. A disponibilidade de água foi manipulada em condições naturais e mediram-se as mudanças de estações num grupo de características foliares e órgãos reprodutivos. As respostas foliares medidas incluem enrolamento, movimentos para-heliotrópicos, aumento de refletividade e indução do metabolismo ácido de Crassuláceas (CAM). Durante aproximadamente um ano estudaram-se plantas que cresciam numa localidade com estação seca, sujeitas ao regime natural de chuvas; 30 destas plantas foram regadas freqüentemente. A acumulação noturna de ácidos e o enrolamento, ângulo e refletividade foliar foram maiores durante todo o ano, exceto durante a época de chuvas, em plantas com subministro natural de água que em plantas regadas. As primeiras tiveram maior produção de flores e frutos que as plantas experimentalmente regadas. Não se encontraram diferenças significativas na freqüência de visitas de polinizadores a ambos grupos de plantas. O ângulo, enrolamento e refletividade foliar, e o cociente massa reprodutiva / massa foliar mostraram correlações positivas com a acumulação noturna de ácidos (r²=0,83; 0,64; 0,59 e 0,64 respectivamente). O cociente massa reprodutiva / massa foliar se correlacionou positivamente com o enrolamento foliar (r²=0,51) e a refletividade (r²=0,79), mas não com o ângulo foliar. Os resultados sugerem que a seca ocasiona um aumento no ângulo, enrolamento e refletividade foliar, assim como na atividade do CAM, e atua como um sinal para o aumento da fecundidade em plantas de T. triangulare no campo.

KEYWORDS / Crassulacean Acid Metabolism / Fecundity / Leaf angle / Leaf Rolling / Leaf reflectance / Paraheliotropic movement /

Received: 07/04/2002. Modified: 04/28/2003. Accepted: 04/30/2003

Introduction

Deciduous plants subjected to drought in seasonally-dry environments may display a range of responses that are thought of as adaptive to water deficit. Frequent responses of plants to drought include leaf rolling, paraheliotropic movements and increase in leaf reflectance (Begg, 1980; Ehleringer, 1980). Also, Crassulacean acid metabolism (CAM) may be induced by drought in facultative plants and is generally considered an adaptive response to water deficit because its operation increases water use efficiency, i.e. the ratio between assimilated CO₂ and transpired water (Winter and Smith, 1996).

Water deficit may cause a severe reduction in reproductive output (Lee and Bazzaz, 1986; Lee, 1988); this reduction may be due to abortion of different reproductive parts: buds, ovules, fruits or seeds. Many tropical plants, including herbs and shrubs, flower and set fruit during the rainy season (Monasterio and Sarmiento, 1976; Seghieri et al., 1995), while many others do it during the dry season (Monasterio and Sarmiento, 1976), when photosynthetic activity must be lower.

In plants of the inducible-CAM species, Mesembryanthemum crystallinum, an extremely large seed production during the dry season was correlated with the extension of the life cycle due to CAM operation (Winter, 1985). In greenhouse-grown plants of Talinum triangulare with CAM induced by drought, fecundity was higher than in watered plants, survival remaining unchanged (Taisma and Herrera, 1998).

Plants of the shrub T. triangulare (Jacq.) Willd. (Portulacaceae) have a wide range of responses to water deficit which may be adaptive. Well-watered plants present a typical C3 metabolism and, within a few days of drought, switch to CAM, with a low night-time CO₂ assimilation rate (Herrera et al., 1991) to eventually enter the gas-exchange and acidity pattern known as CAM-idling (Winter and Smith, 1996). Under drought, leaves have a paraheliotropic movement whereby their angle relative to the horizontal plane increases with drought; also, they appear more reflective and the lamina rolls about its margins onto the abaxial surface. By reducing heat load and increasing water saving, these features, together with CAM induction, may help explain how leaves can remain green after two months of drought.

The above-mentioned results and observations suggested the present study. Plants of T. triangulare growing in the field were subjected to either the natural rainfall or frequent watering through the seasons. The objective of the experiment was to prevent by watering changes in leaf responses and thus establish the possible relationship between these responses, allegedly adaptive to drought, and fecundity in plants of T. triangulare.

Materials and Methods

Plant material and field site. Talinum triangulare (Jacq.) Willd. (Portulacaceae) is a perennial, deciduous shrub with lignified, non-photosynthetic stems; the leaves can last as long as two months after the onset of water deficit. The plant is iteroparous, i.e. it is capable of reproduction several times during its life span. The field work was carried out at the Henri Pittier National Park, Estado Aragua, Venezuela, 10º13’N - 67º14’W, at approximately 20masl, in a site near the village of Cuyagua by a road outside the forest, under full sun exposure. Plants were studied on five dates during 1990-91, covering the dry and rainy seasons and the transitions thereof. Rainfall data (Figure 1a) were collected at the nearby Ocumare de la Costa weather station; mean annual precipitation is 800mm. In an area of approximately 100m² with a dense population of T. triangulare, the dominant species, 30 individuals were watered immediately after the first measurements were taken, in December 1990, and weekly thereafter, until September 1991. Another group of 30 individuals which remained under natural watering conditions was labeled with plastic tags. The plants were 50cm tall on average and had profuse lignified branching. Those individuals that were tagged at the beginning of the experiment remained alive on each sampling date.

Microclimate. Photosynthetic photon flux density (PPFD) was measured with an LI-170 quantum sensor connected to an LI-185 meter (LI-COR Inc., Lincoln, Nebraska, USA). Air temperature was measured with thermistors connected to a telethermometer (Yellow Springs Instruments Co., Yellow Springs, Ohio, USA).

Gas exchange. Measurements of CO₂ and H₂O exchange were made on three randomly selected leaves from different individuals per treatment on each date, with an LCA2 IRGA connected to a PLC(B) leaf chamber and an ASU(MF) air supply unit (Analytical Development Co. Ltd., Hoddesdon, U.K.). Leaf temperature during measurements was approximately 2ºC higher than air temperature. Illumination was natural and measurements were taken under clear skies. Between 08 and 16h, PPFD ranged from 500 to 2000µmol·m^-2·s^-1, sufficient to saturate light curves in T. triangulare (Herrera et al., 1991). In December 1990 a 24h gas exchange course was done; for the rest of the sampling period only daytime gas exchange was measured since nocturnal sampling was dangerous due to the presence of snakes. Instantaneous water use efficiency (IWUE) was calculated as the average of the photosynthetic rate/transpiration rate, measured every hour during the light period.

Natural carbon isotopic discrimination (d¹³C). Dried leaf and fruit samples were ground in a mill and analyzed for d¹³C at the Laboratory of Isotopic Ecology, Centro de Energia Nuclear na Agricultura, Universidade de Sâo Paulo, (Piracicaba, Brazil).

Fecundity schedules. On each study date, the number of inflorescences per branch was counted on five labeled plants of each treatment. In order to determine the mass of buds, flowers and fruits per inflorescence, fifty inflorescences per treatment were randomly collected among all plants of each treatment. Fifty flowers and fruits per treatment were randomly collected and preserved in 70% ethanol for determination of the number of ovules per flower, and viable and abortive seeds per fruit. As a measure of the influence of changes in leaf properties on fecundity, the relative reproductive mass was calculated in plants, collected as indicated below, as the ratio of reproductive mass to leaf mass (RRM, as %), where reproductive mass included peduncles, buds, flowers and fruits. A similar calculation was done for individual organs. Total seed number/leaf mass was calculated multiplying the numbers of seeds/fruit by fruits/inflorescence by inflorescences/plant and dividing by leaf mass.

Biomass allocation. Unlabeled naturally watered plants (n=5) as well as experimentally watered plants (n=5) were harvested on each date and their organs separated. Dry mass (48h at 60ºC) of organs was weighed on an analytical balance.

Nocturnal acid accumulation. Samples (n=6) were collected at dawn and dusk and kept in liquid N₂. Boiled aqueous leaf extracts were titrated to pH 7.0 with 10mM KOH (Nobel, 1988). Nocturnal acid accumulation (DH⁺) was calculated as the difference between dawn and dusk values.

Leaf angle, area reduction and reflectance. Between 20 and 50 randomly selected leaves per treatment were measured for each variable. Leaf angle relative to the horizontal plane was measured with a protractor; values higher than 90º indicate that leaves became pendulous. Reflectance was estimated as the fraction of incident PPFD reflected by the adaxial surface of leaves; a measurement was taken of the incident PPFD and immediately of the PPFD reflected by the leaf at times of peak solar radiation, placing the sensor perpendicular to the lamina without modifying leaf angle. The reduction in leaf area by rolling was calculated as the ratio of rolled to unrolled lamina area (as %).

Pollinator visit frequency. The number of insect visits to a single flower was counted during 3min between 07 and 08h, when anthesis takes place in this species. Ten flowers per treatment were observed on each date. No attempt was made to distinguish among types of pollinators.

Statistics. Results presented are mean ±SE. Data were analyzed through linear and non-linear regressions and pair-wise ANOVA by date (p<0.05) using the statistical package Statistica for Windows 93.

Results

Nocturnal acid accumulation was higher in naturally than in experimentally watered plants in February and June but not in August, at the peak of the rainy season, or in October, when watering was stopped (Figure 1). Microclimatic conditions in December and August were, respectively: maximum PPFD, 2000 and 1573µmol·m^-2·s^-1, and air temperature, 40.0 ±1.5 and 33.1 ±1.5ºC.

In naturally watered plants, a marked influence of water deficit on the magnitude of daytime CO₂ assimilation rate (A) was observed during the period of study, and lower values were found in naturally than experimentally watered plants in February and June (Figure 2). Reductions in A due to water deficit resulted in diminished daytime C balance; the average IWUE was higher in naturally than experimentally watered plants in February (Table I). The relative recycling of CO₂ by CAM in naturally watered plants, calculated with values of December as % recycling = 100x[(0.5xDH⁺)-integrated night-time A)]/(0.5xDH⁺), taking DH⁺ on an area basis (after Griffiths, 1988) was 83%.

Figure 3 shows the seasonal changes in the pattern of biomass allocation of plants in this experiment. Whole plant mass of naturally watered plants was not significantly different from that of experimentally watered plants, except in October, when it was higher. The proportion of whole plant mass allocated to leaves was higher in experimentally than in naturally watered plants in February as well as October, while the shoot/root ratio was always higher, except in June. The RRM, a measure of how much leaves contribute to reproduction, was higher in naturally than experimentally watered plants in February and October.

Nocturnal acid accumulation and RRM decreased with an increase in daytime carbon balance (Figure 4). Values of d¹³C of fruits from naturally watered plants were not statistically different from those of leaves of either treatment, giving a mean of -24.6 ±0.6‰.

 

The relative mass of buds, flowers and fruits per plant was equal or larger in naturally than in experimentally watered plants (Figure 5). The number of ovules per flower was higher and the seed abortion ratio lower in naturally watered plants (Figure 6). Seed abortion ratio of naturally watered plants was higher in February than in December (the corresponding values for experimentally watered plants are not available as these had no fruits). The highest number of seeds per leaf mass occurred in December (Figure 6).

Taking DH⁺ as an indicator of drought, it can be seen that the angle relative to the horizontal plane, leaf rolling, leaf reflectance and RRM increased significantly with drought in both treatments, although low values can be found in experimentally watered plants, more so at intermediate values of DH⁺  (Figure 7). Linear significant relationships were found between RRM and rolling (r²=0.51; p=0.03) and reflectance (r²=0.79; p=00007) but not leaf angle (p=0.12).

As suggested by Figures 1 and 5, significant correlations were found between DH⁺ of naturally watered plants and mass of flowers per leaf mass (r²=0.62; p=0.04) and mass of buds per leaf mass (r²=0.86; p=0.03). The correlation of DH⁺ with the number of ovules per flower and of viable seeds per fruit was very low (r²=0.07).

Differences in fecundity due to treatment were apparently not influenced by frequency of pollinator visits, which showed no significant difference between naturally and experimentally watered plants on any sampling date; rains caused a reduction in pollinator visits of approximately three times relative to moderate drought (Table II).

Discussion

The results of the present study confirm previous observations on the induction of CAM by water deficit done under laboratory conditions (Herrera et al., 1991). CAM induction was fully reversed by irrigation or by natural rainfall events. Daytime carbon balance decreased, and leaf angle, rolling, reflectance and DH⁺ increased with drought, significant relationships occurring between DH⁺ and the increase in leaf angle and rolling.

Relative reproductive mass was highly correlated with DH⁺, as previously reported in plants of T. triangulare subjected to drought in the greenhouse (Taisma and Herrera, 1998) and with leaf rolling and reflectance. The correlation between RRM and DH⁺ was much higher than that reported between the ratio droughted/watered reproductive biomass and the overnight malic acid accumulation in plants of five species of Talinum (Harris and Martin, 1991).

The increase caused by water supply in daytime CO₂ assimilation rates resulted in higher total daytime carbon balances but lower RRM. In contrast, in plants of Cakile edentula var. lacustris A was positively correlated with reproductive output (Dudley, 1996). By taking place during periods of water deficit, reproduction in species such as T. triangulare would depend on carbon acquired by the photosynthesis of the previous, rather than the current season.

CAM induction by drought may provide an important source of energy for reproduction. Fruits of M. crystallinum exhibited higher d¹³C values than leaves, indicating that carbon used for inflorescence development was primarily fixed through CAM late in the growing season (Winter, 1985). Plants of M. crystallinum with CAM induced exposed to normal air during the night produced 10 times more seeds than those with CAM induced but exposed to CO₂-free air during the night (Winter and Ziegler, 1992). In the present study, no difference was found in d¹³C of fruits and leaves, suggesting that, although RRM was highly correlated with DH⁺, the source of carbon for fruit set was not different from that for leaf production.

In T. triangulare, dark CO₂ assimilation amounts to only 7% of the net carbon gain of plants in the C3 mode (calculated from Herrera et al., 1991). Also, values of D¹³C in leaves of this species with or without CAM induced are not different (Herrera, 1999). These results suggest that in this species the contribution of dark CO₂ uptake to carbon balance of leaves and to reproduction is very small. Probably, a high internal CO₂ recycling is the feature of CAM best related to reproductive output.

Experimentally watered plants of T. triangulare with a DH⁺ similar to that of naturally watered plants had a significantly lower RRM in October (when watering was already stopped) than naturally watered plants, suggesting that experimentally watered plants were more prone to water stress.

Alternatively, the previous story of water deficit apparently contributed to the acclimation of plants of T. triangulare, since naturally watered plants had higher daytime CO₂ assimilation rate (A) and were larger in October than experimentally watered plants. This hypothesis may be supported by the observation that naturally watered plants apparently allocated more biomass to roots than to the shoot. Similarly, in sunflower plants subjected to an acclimation water deficit pre-treatment A was higher than in watered controls (Matthews and Boyer, 1984). Flower and fruit mass were higher in naturally than in experimentally watered plants of T. triangulare after the peak of the rainy season, suggesting that reproduction was also influenced by the previous history of water deficit.

The poor correlation between DH⁺ and number of ovules per flower and of viable seeds per fruit, rather than contradicting the positive correlation between RRM and DH⁺, may reflect the dynamics of reproduction. Based on the markedly seasonal rain regime of the study site, it should be possible to predict that the number of seeds per leaf mass after the experimental year is similar to that found when this study began, at the peak of the dry season.

In February, when IWUE was highest in naturally watered plants, the number of flowers was higher and ovule abortion ratio lower. An increase in IWUE, possibly due to the joint effect of the operation of CAM, leaf rolling and increased reflectance and angle, may help explain the observed increase in these reproductive characteristics. Water use efficiency contributed in an important fraction to the reproductive success of individuals of Xanthium strumarium grown in an experimental garden (Farris and Lechowicz, 1990).

In this study no evidence was found that the increase in reproduction associated to periods of drought reflected a decrease in survival, since labeled individuals were alive at the end of the study. In contrast, in plants of M. nodiflorum the relationship between the increase in DH⁺ and reproductive output during summer was interpreted as a strategy to delay resource allocation to reproductive biomass and to maximize reproductive yield before the end of the life-span in this annual species (Sayed and Hegazy, 1994).

Drought increased reproductive output mainly through a decrease in bud and ovule abortion. An important factor for abortion may be a limitation on the frequency and abundance of pollinators (Bierzychudek, 1981). In the present work naturally and experimentally watered plants were visited by pollinators with equal frequency, probably because of the spatial proximity between plots. The decrease in abortion observed in naturally watered plants may be due to the leaf and whole-plant tolerance responses to drought.

The results suggest that the changes brought about by drought in the leaf properties studied increased fecundity without changes in survival of plants of T. triangulare under field conditions. Drought serves as a cue to these changes, the causative agent linking leaf responses to reproduction remaining to be investigated.

Acknowledgements

This research was financed by grants CDCH (Venezuela) 03.10.2252.90 and 03.10.2252.92, and CONICIT (Venezuela) S1-1754. M.A. Taisma received a scholarship from Fundación Gran Mariscal de Ayacucho. M. Cuberos, M.D. Fernández and W. Tezara helped with field work.

References

1. Begg JE (1980) Morphological adaptations to water stress. In Turner NC, Kramer PJ (Eds.) Adaptations of Plants to Water and Temperature Stress. John Wiley and Sons. New York, USA. pp. 33-42.        [ Links ]

2. Bierzychudek P (1981) Pollinator limitation in plant reproductive effort. American Naturalist 117: 838-840.        [ Links ]

3. Dudley SA (1996) Differing selection on plant physiological traits in response to environmental water availability: a test of adaptive hypotheses. Evolution 50: 92-102.        [ Links ]

4. Ehleringer J (1980) Leaf morphology and reflectance in relation to water and temperature stress. In Turner NC, Kramer PJ (Eds.) Adaptations of Plants to Water and Temperature Stress. John Wiley and Sons. New York, USA. pp. 295-308        [ Links ]

5. Farris MA, Lechowicz MJ (1990) Functional interactions among traits that determine reproductive success in a native annual plant. Ecology 71: 548-557.        [ Links ]

6. Griffiths H (1988) Crassulacean acid metabolism: a reappraisal of physiological plasticity in form and function. Adv. Botanical Res. 15: 43-92.        [ Links ]

7. Harris FS, Martin CE (1991) Plasticity in the degree of CAM-cycling and its relationship to drought stress in five species of Talinum (Portulacaceae). Oecologia 86: 575-584.        [ Links ]

8. Herrera A (1999) Effects of photoperiod and drought on the induction of CAM and the reproduction of plants of Talinum triangulare. Can. J. Botany 77: 1-6.        [ Links ]

9. Herrera A, Delgado J, Paraguatey I (1991) Occurrence of inducible Crassulacean acid metabolism in leaves of Talinum triangulare (Portulacaceae). J. Exp. Botany 42: 493-499.        [ Links ]

10. Lee TD (1988) Patterns of fruit and seed production. In Lovett-Doust J, Lovett-Doust L (Eds.) Plant Reproductive Ecology. Patterns and Strategies. Oxford University Press. Oxford, UK. pp. 179-202.        [ Links ]

11. Lee TD, Bazzaz FA (1986) Maternal regulation of fecundity: non-random ovule abortion in Cassia fasciculata Michx. Oecologia 68: 459-465.        [ Links ]

12. Matthews MA, Boyer JS (1984) Acclimation of photosynthesis to low leaf water potentials. Plant Physiol. 74: 161-166.        [ Links ]

13. Monasterio M, Sarmiento G (1976) Phenological strategies of plant species in the tropical savanna and the semi-deciduous forest of the Venezuelan Llanos. J. Biogeogr. 3: 325-356.        [ Links ]

14. Nobel PS (1988) Environmental Biology of Agaves and Cacti. Cambridge University Press. Cambridge, UK. pp. 54-55.        [ Links ]

15. Sayed OH, Hegazy AK (1994) Growth-specific phytomass allocation in Mesembryanthemum nodiflorum as influenced by CAM induction in the field. J. Arid Environ. 27: 325-329.        [ Links ]

16. Seghieri J, Floret Ch, Pontanier R (1995) Plant phenology in relation to water availability: herbaceous and woody species in the savannas of northern Cameroon. J. Tropical Ecol. 11: 237-254.        [ Links ]

17. Taisma MA, Herrera A (1998) A relationship between fecundity, survival and the operation of CAM in T. triangulare. Can. J. Botany 7: 1-8.        [ Links ]

18. Winter K (1985) Crassulacean acid metabolism. In Barber J, Baker NR (Eds.) Photosynthetic Mechanisms and the Environment. Elsevier. Amsterdam, Netherlands. pp. 329-387.        [ Links ]

19. Winter K, Smith JAC (1996) Crassulacean acid metabolism: current status and perspectives. In Winter K, Smith JAC (Eds.) Crassulacean acid metabolism. Biochemistry, Ecophysiology and Evolution. Springer-Verlag. Berlin, Germany. pp. 389-436.        [ Links ]

20. Winter K, Ziegler H (1992) Induction of crassulacean acid metabolism in Mesembryanthemum crystallinum increases reproductive success under conditions of drought and salinity stress. Oecologia 92: 475-479.        [ Links ]