, Volume 23, Issue 3, pp 509–519

Physiological and morphological response to water deficit in seedlings of five provenances of Pinus canariensis: potential to detect variation in drought-tolerance


  • Rosana López
    • U. D. Anatomía, Fisiología y Genética vegetal. ETSI MontesUniversidad Politécnica de Madrid. Ciudad Universitaria
  • Jesús Rodríguez-Calcerrada
    • U. D. Anatomía, Fisiología y Genética vegetal. ETSI MontesUniversidad Politécnica de Madrid. Ciudad Universitaria
    • U. D. Anatomía, Fisiología y Genética vegetal. ETSI MontesUniversidad Politécnica de Madrid. Ciudad Universitaria
Original Paper

DOI: 10.1007/s00468-008-0297-5

Cite this article as:
López, R., Rodríguez-Calcerrada, J. & Gil, L. Trees (2009) 23: 509. doi:10.1007/s00468-008-0297-5


To assess the potential of short-term screenings for drought resistance at the seedling stage to detect ecotypic variation and predict field performance, we studied the responses to water deficit of seedlings of Pinus canariensis from five geographic origins under controlled conditions and compared these responses with the performance of provenances in a multi-site field trial. Leaf water potential, the osmotic component, leaf chlorophyll fluorescence and growth and biomass partitioning were measured as seedlings were subjected for 11 days to two levels of osmotic potential generated by polyethylene glycol (PEG 6000), −1 MPa (slowly imposed water deficit; S) and −1.5 MPa (fast imposed water deficit; F), and a control treatment (no PEG added to the nutrient solution; C). Leaf water potential declined to final mean values of −1.2, −2.7 and −4.7 MPa in the C, S and F treatments, respectively. The ratio of variable to maximum chlorophyll fluorescence declined to final mean values of 0.77, 0.66 and 0.40 in the C, S and F treatments, respectively, with no differences amongst provenances. All provenances showed an active osmotic adjustment (OA) in response to water deficit which varied depending on the drying rate. A slow imposition of water deficit favoured solute accumulation. Pooling all treatments, the index of OA ranged from 0.28 to 0.40, but rose considerably when only C and S treatments were considered (0.56 to 0.70). There was a positive and significant correlation between the overall index of OA (all treatments pooled) and the drought period in the site of origin, suggesting ecotypic variation in OA as a result of drought duration. Seedlings allocated more dry matter to roots than shoots when subjected to moderate and slowly imposed water deficit; only one provenance showed no increase in the root to shoot ratio at the end of the treatment period compared with control seedlings. Responses to controlled water deficits were only qualitatively related to performance (survival and growth) of provenances in several field sites, indicating the involvement of complex mechanisms to cope with drought under natural conditions. However, the provenance with the highest overall index of OA outgrew and outsurvived the other provenances in the most arid site, and the only provenance not modifying the root to shoot ratio in response to water deficit survived the least in all field sites. Acclimation of root to shoot ratio and net solute accumulation to water deficit could hence favour drought-tolerance beyond the seedling stage and be used as preliminary predictors of field performance.


DroughtDrying ratesOsmotic adjustmentChlorophyll fluorescenceDry matter allocationEarly screening


In Mediterranean-type ecosystems, seasonal water shortage is the main factor constraining survival and growth of plants (Di Castri et al. 1981). Therefore, the assessment of response mechanisms to water deficit can help to identify life history traits associated with drought-tolerance, and in turn, they can be used as predictors of species performance. Genetic improvement programs aim at reducing establishment failure in seasonally dry ecosystems by identifying and selecting genotypes more resistant to drought. Selection criteria are based on the intraspecific comparison of attributes potentially conferring drought resistance and the evaluation of performance of genotypes in field trials. Physiological (e.g., water relations) and morphological (e.g., growth and dry matter partitioning) features, and their plastic responses to water deficit, are commonly monitored in screenings for drought resistance, but their extrapolation to field performance is problematic, especially for tree species.

Pines show several desiccation-avoiding adaptations such as deep roots, sclerophyllous leaves, deeply sunken stomata, thick leaf cuticles and abundant leaf waxes. Furthermore, pines, like most species, exhibit morphological and physiological modifications when exposed to environmental changes (phenotypic plasticity), which in heterogeneous and unpredictable ecosystems, such as Mediterranean, can be considered advantageous (Bradshaw 1965). Osmotic adjustment (OA), i.e., the lowering of osmotic potential resulting from the net accumulation of solutes, is considered a major component of drought resistance (Kozlowski and Pallardy 2002), since it allows plants to enhance their ability to extract water from dry soil while maintaining turgor pressure, cell expansion and stomatal conductance (Hsiao et al. 1976; Jones and Turner 1980). Previous studies have shed some light on OA in populations of Mediterranean pines, and it seems more intense in plants from dry sites (Calamassi et al. 2001; Nguyen-Queyrens and Bouchet-Lannat 2003). Pines are characterised by high stomatal sensitivity to dehydration. Preventing desiccation in this way reduces carbon assimilation owing to CO2 shortage (Martínez-Ferri et al. 2000), associated decrease in Rubisco activity (Flexas et al. 2006), and to a lesser extent, reduced photochemical efficiency of photosystem II. However, leaf photo-damage is usually prevented by concomitant reductions in light harvesting reducing excess energy (Baquedano and Castillo 2007). Assessing the integrity of the thylakoid membrane using chlorophyll fluorescence provides a rapid and accurate technique of detecting and quantifying plant tolerance to environmental factors such as drought (Epron et al. 1992).

The wide range of ecological conditions found in the native range of the Canary Island pine (Pinus canariensis Chr. Sm. Ex DC) is due to a high neutral genetic differentiation (Gómez et al. 2003), and a high diversity of quantitative traits related to colonisation capacity (Climent et al. 2006; López et al. 2007). This species covers large areas of the Canary Archipelago where the existence of other tree species is extremely unlikely. In the most xeric locations, drought in the first stages of development is the main factor limiting survival (Gieger and Leuschner 2004). Field experiments located in the Canary Islands and Israel showed that plants from dry and xeric populations survived better than those from wet and more productive origins (López et al. 2007). Changes in physiological traits have been observed in stands covering an altitudinal gradient from 200 to 2,000 m.a.s.l. (Jiménez et al. 2005). Flexibility to thrive in distinct climatic areas will result from either or both phenotypic plasticity and ecotypic differentiation.

In the present study, we evaluated intraspecific variation of morphological and physiological responses to water deficit in juvenile seedlings of P. canariensis from five provenances from contrasting environments.

Hydroponics have proved useful for evaluating mechanisms related to drought resistance, since water availability can be controlled precisely, water deficit is applied in a homogeneous manner, and problems related to differential rates of water consumption by different genotypes are avoided. We hypothesised that plants from drier provenances would experience either a greater OA, a lower reduction in photochemical efficiency or an increased relative allocation to roots in response to water deficit than plants from wet provenances. We contrasted our results with the performance of all five provenances in a field experiment, to assess the convenience of some attributes for selecting drought-tolerant genotypes.

Materials and methods

Plant material and drought treatments

Seeds of five provenances of P. canariensis were collected in natural populations covering a wide range of environments (Table 1). After seed germination, 375 seedlings per provenance were transferred to an aerated Hoagland nutrient solution oxygenated with air diffusers to prevent root asphyxia. The growing media was renewed once a week. Nutrient concentrations were gradually increased: 1/16, 1/8, 1/4, 1/2 every week until the complete formula. The whole experiment was conducted in a growth chamber with controlled photoperiod (16 h), temperature (25/20°C day/night) and relative humidity (60/80%). After 2 months in the growing media, Polyethyleneglycol (PEG 6000) was added every other day to gradually lower the solution water potential down to −1 MPa in the slowly imposed water deficit treatment (S; 0.1 MPa day−1) and −1.5 MPa in the fast imposed water deficit treatment (F; 0.15 MPa, day−1). In addition, 125 plants were maintained in the growing media without addition of PEG (C). Solution water potential was determined with a psychrometer–hygrometer, based on the voltage versus solution water potential relationship previously derived from the relationship between different PEG 6000 concentration and voltage readings (HR-33T, Wescor, Logan, UT). During the treatment period, which lasted for 11 days, culture solutions were replaced three times, and were always kept aerated.
Table 1

Location and ecological characteristics of the five provenances of Pinus canariensis used in this study and the four sites included in the field trial

Provenance/trial site

Latitude (N)

Longitude (W)

Elevation (m)

Pa (mm)

T (°C)

Tr (°C)

Dp (months)









































































Pa annual precipitation, T mean annual temperature, Tr annual temperature range, Dp drought period

Measurements under controlled conditions

Relative water content (RWC), water potential (Ψwp) and osmotic potential (ΨΠ) were measured five times during the experiment (days 3, 5, 7, 9 and 11) in six plants per provenance and treatment. On each day, seven needles per tree were collected at dawn, three for RWC, three for ΨΠ and one for Ψwp determination. RWC was estimated as RWC = (FW − DW)/(HW – DW), where FW (fresh weight) was obtained by weighing the needles immediately after harvest; HW (weight at full hydration) was estimated by placing the needles in distilled water at 4°C during 24 h, and DW (dry weight) was that of the oven-dried (48 h at 70°C) needles. Ψwp was measured using a pressure chamber (PMS Instrument Co., 7000, Albany, USA). To determine ΨΠ, the needles were cut in small pieces and frozen in liquid nitrogen; after thawing, ΨΠ was measured with a psychrometer-hygrometer (C-52 sample chamber Wescor) connected to a dew-point microvoltmeter (HR-33T, Wescor, Logan, UT).

Logarithmic plots of RWC against osmotic potential: lnRWC = f(–ln(–ΨΠ)) were made to determine OA (Morgan 1992). In addition, a RWC of 80%, close to the turgor loss point in other Mediterranean pines (Fernández et al. 1999), was chosen to compare OA between provenances. The index of OA proposed by Turner (2006) was also estimated. The effect of the imposition rate of water deficit in the OA was tested by removing either plants from the F treatment or the S treatment in the logarithmic plots, and fitting new regression lines.

Chlorophyll fluorescence emission of needles in six seedlings per provenance and treatment was measured twice, on days 7 and 11, with a portable pulse-modulated fluorometer (FMS 2, Hansatech Instruments, Norfolk, UK). Each sample consisted of about eight needles from each seedling. The ratio of variable to maximum chlorophyll fluorescence (Fv/Fm) was calculated as (FmFo)/Fm, where Fm and Fo were, respectively, the maximum and basal fluorescence yields of leaves darkened for 30 min before measurement; this parameter provides an estimate of the maximum photochemical efficiency of the photosystem II (PSII). Under ambient light, once the fluorescence signal was measured (Fs′), an 0.8 s saturating pulse (6,600 μmol m−2 s−1) was applied to obtain the maximum fluorescence (Fm′); the minimal fluorescence yield in the light-adapted state (Fo′) was then immediately assessed in darkened leaves exposed to far-red radiation for 5 s. The efficiency of excitation energy capture by open PSII reaction centres (Fv′/Fm′, where Fv′ = Fm′ − Fo′) and the operating quantum efficiency of PSII electron transport (ΦPSII = 1 − Fs′/Fm′) were calculated according to Maxwell and Johnson (2000).

Dry matter partitioning was measured on days 5, 9 and 11. Three plants per provenance and treatment on days 5 and 9, and seven on day 11, were harvested, divided into roots, stems and needles, dried at 70°C for 1 week, and weighed.

Field trial measurements

Survival and height of the provenances were measured for 6 years in a multi-site provenance experiment in the Canary Islands (for more details see López et al. 2007). Annual precipitation of two of the four sites was lower than 600 mm and in the most xeric site the length of the drought period (sensu Walter and Lieth 1960) was greater than 7 months (Table 1). Mean survival and growth values per provenance in the last year of measurement were calculated (Table 2).
Table 2

Mean values ± standard errors of survival (s) and height in centimetres (h) 6 years after planting of the five provenances across the four field trial sites















0.79 ± 0.04

138 ± 7

0.64 ± 0.05

86 ± 7

0.50 ± 0.05

64 ± 7

0.61 ± 0.02

62 ± 8


0.64 ± 0.05

168 ± 10

0.61 ± 0.05

84 ± 7

0.43 ± 0.05

97 ± 9

0.32 ± 0.02

59 ± 10


0.82 ± 0.06

134 ± 9

0.71 ± 0.06

97 ± 9

0.71 ± 0.06

88 ± 7

0.57 ± 0.03

51 ± 9


0.74 ± 0.06

144 ± 9

0.64 ± 0.06

91 ± 7

0.68 ± 0.06

61 ± 7

0.54 ± 0.03

54 ± 9


0.83 ± 0.04

127 ± 7

0.69 ± 0.04

69 ± 7

0.58 ± 0.05

72 ± 7

0.75 ± 0.02

70 ± 8

Data analyses

The effect of provenances and treatments on RWC, Ψwp and ΨΠ was checked using repeated measures analyses of variance (ANOVA) and on chlorophyll fluorescence parameters and dry mass partitioning values using a GLM approach to ANOVA. Both factors were considered fixed for all the analyses. Differences between levels of significant predictors were tested by Duncan’s multiple-range tests.

Allometric relationships for the study of dry matter allocation were determined through the regression of the natural logarithms of root dry weight (RDW) and shoot (stem + needles) dry weight (SDW); changes in allocation pattern between treatments were assessed by comparison of the slopes and intercepts of regression lines.

Correlation coefficients were calculated to relate water status, chlorophyll fluorescence and dry matter traits to climatic variables of the seed sources. Spearman correlation coefficients were determined between field traits (survival and height) and the measurements obtained on the last day of the controlled experiment (RWC, OA, chlorophyll fluorescence parameters and dry matter allocation).


Time courses of RWC, water potential and osmotic potential

RWC ranged between 82% in control plants and 50% in plants exposed to rapid drying on the last day of measurement (Fig. 1a). In the control treatment, RWC decreased by 10% throughout the experiment and needle Ψwp dropped from −0.50 MPa on day 3 to −1.36 MPa on day 11. The effect of the drought treatment was significant from day 5, when solutions reached Ψ = −0.4 MPa in S and Ψ = −0.6 MPa in F.
Fig. 1

Evolution of relative water content, RWC (a), water potential (dotted lines) and osmotic potential (continuous lines) (b), during the water deficit period. Vertical bars correspond to the standard errors of the mean values of seedlings of each treatment; open circle control treatment; filled triangle slowly imposed water deficit treatment; filled square fast imposed water deficit treatment. RWC (c) and ΨΠ (d) of the five provenances of P. canariensis at day 11 of the water deficit period under control treatment (open bars), slowly imposed treatment (dotted bars) and fast imposed treatment (hatched bars)

The decline in RWC was paralleled by a substantial decrease in Ψwp and ΨΠ (Fig. 1b). The fast imposed water deficit treatment differed from the other two treatments from day 5, and the slowly imposed treatment differed from the control from day 9. Needle Ψwp dropped to final mean values of −1.36 MPa in C, −2.70 MPa in S and −4.75 MPa in F. In contrast to RWC (Fig. 1c), there was a significant provenance effect on Ψwp and ΨΠ (Table 3), with the greatest decline observed in Pajonales (−5.5 MPa in F). Amongst the other provenances, only Vilaflor exhibited lower values than Arico in F on day 11 (Fig. 1d).
Table 3

F values for repeated measures analysis of variance of relative water content (RWC), water potential (Ψ) and osmotic potential (ΨΠ)




Between-subjects effects


0.293 ns


3.374 *


5.403 ***




80.732 ***


96.137 ***

Prov × treat

0.54 ns

Prov × treat

0.603 ns

Prov × treat

0.374 ns

Within-subjects effects


561.29 ***


159.163 ***


190.626 ***

RWC × prov

1.12 ns

Ψwp × prov

1.353 ns

ΨΠ × prov

3.052 ***

RWC × treat

82.37 ***

Ψwp× treat

24.153 ***

ΨΠ × treat

23.884 ***

RWC × prov × treat

0.74 ns

Ψwp× prov × treat

0.870 ns

ΨΠ × prov × treat

1.007 ns

ns not significant

*P < 0.05, **P < 0.01, ***P < 0.001

Osmotic adjustment

All provenances showed an initial phase of decreasing ΨΠ with no changes in RWC, pointing out that absolute OA had occurred, and a second falling phase when both RWC and ΨΠ declined (Fig. 2). The estimated constant value of RWC in the first phase was lower (90.5%), and the length of this phase slightly longer (0.36 MPa) in Pajonales than the other provenances (mean values of 94.2% of RWC and 0.32 MPa). The breakpoint osmotic potential (ΨΠ at the intersection of the two phases) ranged from −1.91 MPa in Arico to −2.32 MPa in Pajonales (Fig. 2). For all provenances, the slope of the second phase showed a marked deviation from 1, which corresponds to a mere concentration of solutes with water loss. The slopes varied amongst provenances and the index of OA ranged between 0.28 and 0.40 (Table 4) indicating different capacities for OA. The highest adjustment was found in Pajonales and Arico and the lowest in Orotava. A strong relationship was found between the length of drought period of the seed sources and the OA index (r2 = 0.93) (Fig. 3a). The index of OA increased considerably for all provenances when plants under the F treatment were removed from the plots. The highest value was reached by Arico: 0.70; Guancha doubled its index (0.60), and in the other three provenances the index of OA was equal to 0.56. When only C and F plants were considered, the ranking of provenances did match again the length of the drought period: Orotava (0.20), Guancha (0.26), Vilaflor (0.32), Arico (0.34) and Pajonales (0.36) (r2 = 0.95). When net solute accumulation was determined at RWC 80%, the ranking of the provenances changed and Arico showed a lower value of OA than Guancha and Vilaflor (Table 4).
Fig. 2

Natural logarithm plots of relative water content (RWC) against osmotic potential (ΨΠ) for five provenances of P. canariensis. The responses of each provenance have been described by three piece-wise regressions: including all treatments (continuos line), including plants under C and S (dotted line) and including plants under C and F (dashed line). Open circle control treatment, filled triangle slowly imposed water deficit treatment, filled square fast imposed water deficit treatment (ae). Comparison of the overall adjustment of the five provenances (f). Regression parameters for this adjustment are presented in Table 4

Table 4

Regression parameters of lnRWC against −ln(−ΨΠ). Osmotic adjustment at 80% RWC (OA80) and Index of osmotic adjustment (OA Index, Turner 2006)



RWC (%) (phase 1)

Πbp (MPa)

Slope (phase 2)

OA80 (MPa)

OA index



87.40 a


0.72 a





90.79 b


0.69 ab





93.94 c


0.64 b





94.17 c


0.60 b





88.09 a


0.60 b



Within columns, parameters with the same letter are not significantly different at P = 0.05

Chlorophyll fluorescence

Water deficit induced a decrease of Fv/Fm, Fv′/Fm′ and ΦPSII in all provenances. A significant treatment effect was found for the three parameters, but there was no effect of provenance on any parameter (Table 5). Divergence amongst treatments increased with time. On day 9, all parameters were similar in the C and S treatments, and different from F. All treatments were significantly different on day 11: Fv/Fm was 0.77 ± 0.02 in C, 0.66 ± 0.02 in S and 0.40 ± 0.03 in F; Fv′/Fm′ reached 0.77 ± 0.04 in C, 0.66 ± 0.04 in S and 0.37 ± 0.05 in F; and ΦPSII was 0.56 ± 0.02 (C), ± 0.41 (S) and 0.16 ± 0.03 (F). There was a positive relationship between the length of drought period of the seed sources and Fv′/Fm′ of F plants on day 11 (r2 = 0.79) (Fig. 3b).
Table 5

Percentage of the explained variation due to provenance, treatment and the interaction provenance by treatment and significance values from the general linear models for biomass components in d11














11.57 **



13.51 **

3.29 **


69.43 ***

32.13 ***

74.62 ***

19.43 ***



24.25 ***

4.57 **

Prov × treat








22.76 **

ns not significant, RDW root dry weight, SDW stem dry weight, NDW needle dry weight, TDW total dry weight, RDW/SDW root dry weight/shoot dry weight ratio

*P < 0.05, **P < 0.01, ***P < 0.001
Fig. 3

Relationships between the length of the drought period at the site of origin of the provenances of P. canariensis and the index of osmotic adjustment (a) and Fv′/Fm′ at day 11 (b)

Dry matter partitioning

Drought treatment accounted for the highest proportion of the variability encountered in total dry matter, mostly because of differences in root dry mass (Table 5). Total dry mass was similar in C and F (0.49 g on the last day of the experiment), while plants under S produced almost 30% more dry matter (0.63 g). Vilaflor stood out in all treatments because of its higher productivity (Fig. 4). Provenance by treatment interaction was highly significant for root/shoot (stem + needles) ratio (Table 5). Reaction norms by provenance across treatments followed three different patterns: root/shoot ratio did not change with water deficit level in Guancha, increased as drought was more intense in Vilaflor, and only the slowly imposed water deficit treatment resulted in an increase of the root/shoot ratio in the other three provenances (Fig. 5a). The allometric analysis showed changes in allocation patterns in response to water deficit, with higher allocation toward roots in the S treatment than the other two treatments. Neither the slope nor the intercept of the regression lines fitted for F and C were significantly different (Fig. 5b). Therefore, when comparisons are made at the same sample size, there are no differences between plants in the control and rapidly drying treatments.
Fig. 4

Dry mass of roots (white), stems (light grey) and needles (dark grey) at day 11 of the water deficit period of the five provenances of P. canariensis under control treatment (open bars), slowly imposed treatment (dotted bars) and fast imposed treatment (hatched bars). Vertical bars correspond to the standard error of the mean
Fig. 5

Reaction norms of the five provenances of P. canariensis on root/shoot ratio at day 11 of the water deficit period. C control treatment, S slowly imposed water deficit treatment, F fast imposed water deficit treatment (a). Allometric trajectories for roots and shoots under open circleC, filled triangleS and filled squareF (b)

Relationships between growth chamber measurements and field performance

No clear relationships were found between traits measured under controlled conditions and either survival or growth of provenances in the field. Spearman correlation coefficients were only significant for survival in Fasnia, the dry site in Tenerife, with traits related to dry matter production: root dry mass (r = 0.84) and total dry mass (r = 0.84), and dry matter allocation: root/shoot ratio (r = 0.97) in treatment F. In treatment S, survival in Llanos was correlated with total dry mass (r = 0.90). Nevertheless, a qualitative assessment showed that Pajonales, with the highest index of OA, ranked as the best provenance for growth and survival in the most arid site and the worst for growth in the most humid (Table 2). In contrast, Guancha survived the least throughout the experiment but it grew taller in the humid sites. Vilaflor, the only provenance that increased root/shoot ratio under the F treatment, seemed to be well adapted to Fasnia, where it outlived and outgrew the other provenances (Table 2).


Response of provenances to contrasted levels of water deficit

RWC, water potential and osmotic potential of needles decreased progressively with time in all three treatments. The decrease of these parameters in the control treatment, with no PEG added, was probably caused by waterlogging due to insufficient aeration of the growing medium (Gibeaut et al. 1997). Values of RWC and Ψwp at the end of the drought cycle were lower than those encountered in other drought-resistant pine species subjected to similar levels of water deficit (Villar-Salvador et al. 1999; Calamassi et al. 2001) or under desert-like natural conditions (Atzmon et al. 2004). Solutions with high molecular weight PEG have high viscosity, which inhibits root water transport (Chazen et al. 1995) and contributes to lowering RWC in plants more than in soil-dried seedlings at given xylem water potentials (Fan and Blake 1997). Therefore, P. canariensis seedlings in the present work were exposed to lower water potentials than those merely caused by the imposed water deficit, but still within the capacity of the plants to recover (Cregg and Zhang 2001). RWC at the end of the drought treatment was close to the permanent wilting point found in other pine species adapted to long periods of water deficit (Parker 1952; Fan et al. 1994; Lee et al. 2004) or in olive trees, whose RWC reached 40% after 13 days of high water deficit (−5.35 MPa leaf water potential) and photosynthetic activity still continued (Dichio et al. 2006).

Osmotic potential decreased in all provenances during the treatment period. All provenances showed a biphasic response to leaf dehydration. This trend is not universal; the initial phase of absolute OA without changes in RWC was absent in some species (Rodríguez-Maribona et al. 1992; Ma et al. 2004) and present in other crops and trees (Morgan et al. 1991; Lilley and Ludlow 1996; Nguyen-Queyrens and Bouchet-Lannat 2003; Warwick and Thukten 2006). The degree of OA varied amongst provenances, according to how it was quantified and the rate of water stress imposition. Considering the index of OA (calculated from the slope of the second phase; Turner 2006) Arico and Pajonales showed the largest adjustments (0.40) and Orotava the least (0.28). These values are similar to values found in different provenances of Pinus pinaster (Nguyen-Queyrens and Bouchet-Lannat 2003) and sections of Acacia (Warwick and Thukten 2006), but lower than values reported by Turner (2006) for sorghum, some lines of wheat and grape vine leaves. OA calculated at 80% RWC (OA80) showed a different ranking of provenances. While Pajonales, Vilaflor and Orotava kept their positions, those of Arico and Guancha changed due to differences in the ΨΠ at the intersection of the two linear phases amongst provenances (Fig. 2). OA80 was higher than that achieved by the drier provenances of P. pinaster (Nguyen-Queyrens and Bouchet-Lannat 2003) or four sections of Acacia (Warwick and Thukten 2006). OA80 is extensively used in literature but the index proposed by Turner has the advantage of taking into account the shape of the response to dehydration. Furthermore, the rate of drying significantly influenced OA, since a slow development of water deficit favours solute synthesis and the maintenance of turgor (Jones and Rawson 1979). In the present study, a rapid rate of water deficit imposition seemed to exceed the capacity for acclimation of mesic provenances via OA, which accumulated solutes as fast as the drier provenances when the water deficit was imposed more slowly. Comparisons amongst experiments with different drying rates or indexes could thus be controversial. The index of OA was positively correlated with the length of the drought period at the site of seed origin (Fig. 3a), particularly when excluding the S treatment, suggesting that OA could have an adaptive role in drought resistance. This relationship could reflect an ecotypic variation caused by xericity, as populations in different islands or separated by high elevation volcanoes are exposed to different climates. As suggested by Nguyen-Queyrens and Bouchet-Lannat (2003) leaf OA could form part of a drought-tolerance syndrome conferring greater growth in dry sites based on an enhanced soil-to-leaf water circulation.

In contrast to OA, photochemistry did not help in discriminating amongst provenances, which may reflect either a short time of treatment or a conserved feature in P. canariensis. A relationship between OA and PSII photochemistry was expected, since OA has been correlated with a smaller decline in photosynthetic capacity under water deficit (Comstock and Ehleringer 1984). OA plays a role in maintaining water uptake, and cell turgor and volume at low water potential, and consequently, it helps to maintain CO2 supply and photosynthetic enzyme activity (Clifford et al. 1998). Higher catabolisation of ATP and NADPH in CO2 fixation process in leaves with high net solute accumulation might thus lead to a lower down-regulation of PSII photochemistry. The independence of PSII photochemistry and OA across provenances may also reflect a small difference in retention of leaf turgor amongst provenances at the end of the experiment in all three treatments. However, it was notable that provenances from sites with longer droughts exhibited a lower reduction in the efficiency of open reaction centres in the last day of the F treatment than those from sites with shorter drought periods (Fig. 3b). At such low water potentials (≈−5 MPa), it is unlikely that differences in activation of reaction centres were related to variation in photosynthetic processing of energy, but rather that provenances from sites with longer drought periods absorb less energy (e.g., through a decrease in chlorophyll content) and reduce excess energy. Keeping PSII efficiency by reducing energy absorption can prevent oxidative damage (Baquedano and Castillo 2006) and hence being adaptive in sites with prolonged periods of drought and intense radiation.

The drying rate also affected dry matter production. Compared with control and fast imposed water deficit treatments, the slowly imposed water deficit treatment resulted in a significant increase in total dry matter, attributable to the increment of RDW (Fig. 4), and a shift in carbon allocation to roots (Fig. 5b). Previous studies have reported a larger absolute growth of roots in plants exposed to water deficits than in well watered plants (Nguyen and Lamant 1989), consistently with optimal partitioning theory (Bloom et al. 1985) that plants shift carbon allocation to the organs collecting the most limited resource. However, this was not observed in the fast imposed water deficit treatment, probably because the intensity and imposition rate of water deficit exceeded a threshold for morphological acclimation in the seedlings. Provenances displayed different reaction norms of root/shoot ratio (Fig. 5a). However, the rank of provenances by OA did not tally with the rank by dry matter production or allocation. Although the capacity to delay turgor loss during water deficit via OA ensures prolongation of photosynthesis (Jones and Turner 1980; Cochard et al. 2002), accumulation of soluble sugars as solutes (e.g., Epron and Dreyer 1996) might counteract any positive effect of turgor retention on growth (Munns 1988).

Scaling of screening for drought-tolerance at early stages to field performance

Based on the greater capacity for overall OA and acclimation of the root to shoot ratio in response to moderate and slowly imposed water deficit observed in plants from Pajonales and Arico, the two driest sites, we would expect a better performance of plants from these two provenances in sites under severe drought conditions with respect to plants from Guancha, which showed less overall OA and no change in dry matter partitioning in response to water deficit. Our results showed, though, qualitative more than quantitative relationships between descriptors of field performance in arid locations and physiological traits related to drought resistance at early stages of development. Pajonales survived well in sites with harsh conditions and exhibited a conservative behaviour for growth, outgrowing the rest of provenances in the most arid site but growing less in the most fertile location. Different values in dry matter allocation of Vilaflor under the slowly imposed water deficit treatment could be related to its high survival and growth in Fasnia, a moderately dry site. Guancha displayed an opposite behaviour to Pajonales, growing taller than the other four provenances in the humid sites but having the lower rates of survival across sites. These findings support evidence that there is no single trait or strategy to effectively cope with drought. For example, provenances showing lower solute accumulation in response to water deficit could reduce turgor decline with water loss by increasing the elasticity of cells and lowering the RWC at which turgor loss occurs (Saito and Terashima 2004). Changes in symplast/apoplast water partitioning may also play a role in turgor maintenance by diminishing the osmotic potential at turgor loss (Serrano et al. 2005). Furthermore, preliminary assessments under controlled conditions neglect both the influence of multiple environmental interactions (Tschaplinski et al. 1998) and the scaling of drought-tolerance mechanisms across ontogenetic stages (Mediavilla and Escudero 2004). Selection of tree genotypes suited to specific environments can hardly be made based on single variables. However, acclimation of the root to shoot ratio and net solute accumulation in response to water deficits in young seedlings seems to favour drought-tolerance beyond the seedling stage, and could be used as preliminary predictors of trends of survival in arid sites.


We thank Martin Venturas for his technical assistance and Ismael Aranda for his useful comments. Anat Madmony and Ami Zehavi encouraged us to try hydroponics with Pinus canariensis. Seeds were provided by the Cabildos of Tenerife and Gran Canaria.

Copyright information

© Springer-Verlag 2008