Transpiration and leaf growth of potato clones in response to soil water defi cit

Potato (Solanum tuberosum ssp. Tuberosum) crop is particularly susceptible to water defi cit because of its small and shallow root system. The fraction of transpirable soil water (FTSW) approach has been widely used in the evaluation of plant responses to water defi cit in different crops. The FTSW 34 threshold (when stomatal closure starts) is a trait of particular interest because it is an indicator of tolerance to water defi cit. The FTSW threshold for decline in transpiration and leaf growth was evaluated in a drying soil to identify potato clones tolerant to water defi cit. Two greenhouse experiments were carried out in pots, with three advanced clones and the cultivar Asterix. The FTSW, transpiration and leaf growth were measured on a daily basis, during the period of soil drying. FTSW was an effi cient method to separate potato clones with regard to their response to water defi cit. The advanced clones SMINIA 02106-11 and SMINIA 00017-6 are more tolerant to soil water defi cit than the cultivar Asterix, and the clone SMINIA 793101-3 is more tolerant only under high solar radiation.


Introduction
The shortage of water in the soil has been, for a long time, the environmental factor that most limits crop yield worldwide (Oliveira et al., 1982;Shao et al., 2008).Potato (Solanum tuberosum ssp.Tuberosum) crop is particularly susceptible to soil water defi cit because of its small and shallow root system (Weisz et al., 1994).
The response of plant transpiration and growth parameters to FTSW has been previously studied for two potato cultivars in the Netherlands (Weisz et al., 1994) and two cultivars in Brazil (Lago et al., 2012).New potato clones are released almost annually and some of them were recently introduced in the State of Rio Grande do Sul -RS (Bisognin et al., 2008), South Brazil, but their response to water defi cit is unknown.Because there has been an increasing frequency of years with rainfall below normal in the fi rst decade of this century in southern Brazil (Streck et al., 2009), an important step is to identify, from the already developed potato clones, those with a difference in the threshold fraction of transpirable water (FTSW).
The objective of this study was to evaluate the FTSW threshold for decline in transpiration and leaf growth in a drying soil to identify potato clones tolerant to water defi cit.

Materials and Methods
Two greenhouse experiments (Experiment one -E1 and Experiment two -E2) were conducted in Santa Maria, RS, Brazil (29°43' S; 53°43' W; 95 m a.s.l.) during 2010 and 2011.The greenhouse was covered with double polycarbonate plates that have a transmissivity to incoming solar radiation of about 80 %.An automatic maximum temperature control system, was set at 32 °C, and the minimum temperature was not controlled.
The experiments were conducted in 8 L (24-cm diameter × 23-cm tall) plastic pots in the spring, planting starting on 08 Sept 2010 (E1), and in the fall, on 01 Apr 2011 (E2).The experimental design was completely randomized, with two treatments: irrigation (T1) and no irrigation (T2), and ten replicates (one pot with one plant) Four potato genotypes were selected for this study based on the results of Bisognin et al. (2008) for genotype differences adapted to subtropical environments: cultivar Asterix and the advanced clones SMINIA 793101-3, SMINIA 00017-6 and SMINIA 02106-11.Asterix is an old cultivar developed in Europe, and currently extensively grown in south Brazil.The advanced clones SMINIA 793101-3 and SMINIA 00017-6 have good processing  (Bisognin et al., 2008).In a previous study, the advanced clone SMINIA 793101-3 was identifi ed by Lago et al. (2012) as being tolerant to water defi cit as compared with the Macaca, an old cultivar also used in South Brazil, and was, therefore, used in this study as a reference clone with tolerance to water stress.The response of the other genotypes to a drying soil is unknown.
The pots were painted white to reduce absorption of solar radiation by the outer walls, fi lled with A horizon soil samples of a Hapludalf, and placed on a 70 cm high bench.Soil contained 28 g kg -1 organic matter, 0.092 mg g -1 K, and 0.0084 mg g -1 P using sodium dichromate (for organic matter) and Mehlich-1 (for K and P) extractors, respectively.The correction of soil nutrients and acidity was made according to soil analysis and local recommendations for potato.NPK rates applied were 0.56, 0.80 and 0.88 g per pot of N, P and K, respectively.
The seed tubers had high physiological and sanitary quality, with a diameter of about 30 mm.One tuber was planted per pot and the fi rst sprout that emerged on the soil surface was regarded as the main stem, while the other sprouts were eliminated as soon as they emerged to standardize one stem (plant) per pot.All pots were maintained in a well-watered condition until the start of the experiment (Sinclair and Ludlow, 1986;Lago et al., 2012).
From a total of 40 pots planted with each clone, a selection was made of 20 pots per clone (to reduce plant variability), ten of them were used as control (T1), in which water defi cit was not applied, and ten pots in which water defi cit was applied (T2).In addition to the pots with plants, there were also six pots kept without plants, three of them had the evaporated water replenished (after weighing pots every day, see details below) whereas in the other three pots water was not replenished.The daily mean values of water lost by evaporation from the pots without plants were subtracted from the pots in T1 and T2, respectively, in order to ascertain the water lost by evaporation from the pots with plants, and, therefore, quantify plant transpiration only.
The soil drying experiment followed the protocol originally described by Sinclair and Ludlow (1986) that has been widely used in other studies (Muchow and Sinclair, 1991;Ray and Sinclair, 1997;Devi et al., 2009;Gholipoor et al., 2012).Prior to the application of water defi cit, all pots were saturated with water (Sinclair and Ludlow, 1986;Ray and Sinclair, 1997) and allowed to drain for 23 h (Lago et al., 2012), when soil was assumed to be at fi eld capacity.On completing the 23 h after saturation, all pots (with plant and with no plants) were covered with a white opaque plastic fi lm, and the initial weight of each pot was determined.This day was the onset of the experiments (05 Oct 2010 in E1 and 03 May 2011 in E2).Thereafter, the water defi cit was ap-plied in the T2 pots, which were irrigated no further until the end of the experiment when transpiration of the plants in T2 was 10 % or lower as compared with plants in T1 (Sinclair and Ludlow, 1986;Muchow and Sinclair, 1991).Water defi cit was applied when plants had nine to twelve leaves on the main stem (Weisz et al., 1994;Lago et al., 2012).The dehydration of the soil was slow and there was thus no need to add water to the drying pots, as is sometimes reported in experiments that use this protocol (Devi et al., 2009;Gholipoor et al., 2012).
Every day, starting at 16h00, all pots were weighed on an electronic scale with a capacity of 50 kg and an accuracy of 5 g.After the weighing, each pot in T1 was irrigated with the amount of water that had been transpired by the plant since the previous day, as determined by the difference between the weight of the pot on that specifi c day and the initial weight (Sinclair and Ludlow, 1986;Muchow and Sinclair, 1991).
Daily minimum and maximum air temperatures were measured during the two experiments inside a shelter located in the middle of the experiment.In the second experiment, the air vapor pressure defi cit (DPV) was also measured daily at 15h00 with a psychrometer.Incoming solar radiation inside the greenhouse during the two experiments was estimated from solar radiation measured at a weather station located approximately 300 m from the greenhouse multiplied by a factor of 0.8.
The relative transpiration (TR) was calculated by the equation (Sinclair and Ludlow, 1986): where MT2 is the mass of each pot in T2 (g per pot), MT1 the mass of each pot T1 (g per pot), 'j' refers to the day, and 'n' represents the number of replicates (plants) in T1.
The experiments fi nished when all plants in T2 presented TR ≤ 0.1 (10 %), as recommended by the original protocol, assuming that below this rate of transpiration the stomata are closed and the water loss is due only to the epidermal conductance (Sinclair and Ludlow, 1986;Muchow and Sinclair, 1991).The fi nal mass was the weight of the pot when TR ≤ 0.1.
After the end of each experiment, the FTSW for each T2 pot was calculated, each day, by the equation (Sinclair and Ludlow, 1986): where MT2 is the mass of each pot in T2 (g per pot), 'j' refers to the day, 'initial' indicates MT2 on the starting day of the application of water defi cit (beginning of the experiment) and 'end' indicates MT2 on the last day of the experiment.Leaf area (LA) was determined daily throughout the experiment by measuring the longest length (L) and the largest width (W) of the apical leafl et of each leaf on the plant.The LA of each leafl et was calculated multiplying the length and width by a shape factor: LA = 0.7067*L*W (r 2 =0.994),LA = 0.7189*L*W (r 2 = 0.996), LA = 0.6978*L*W (r 2 = 0.994) and LA = 0.7026*L*W (r 2 = 0.993), respectively, for clones SMINIA 00017-6, SMIN-IA 02106-11, SMINIA 793101-3 and Asterix.The shape factor for each clone was estimated using simple linear regression between L*W and the measured area (scanned and digitally measured) of 60 individual leafl ets.
The LA of all leafl ets was summed for each plant on each day and the daily relative leaf growth (LGR) for each clone was calculated by: where LAT2 is the leaf area of all leafl ets measured on each plant in T2 (cm² per plant), LAT1 is the leaf area of all leafl ets measured on each plant in T1 (cm² per plant), 'j' refers to the day and 'n' represents the number of replications (plants) in T1.
The amount of water, in liters, necessary for the production of 1 kg of biomass (L kg -1 ) of dry mass, is the coeffi cient of transpiration (CT), which is an indicator of the effi ciency of the use of water by plants (Hsiao and Acevedo, 1974).Three plants of each clone for each experiment prior to application of water defi cit and all plants at the end of the experiments were sampled and oven dried at 60 °C to determine the initial and fi nal total dry biomass and root dry biomass.CT was determined by: where TT is the total transpiration (L per plant), BF is the fi nal total dry biomass (kg per plant) and BI is the initial total dry biomass (kg per plant).
The variables transpiration and leaf growth were subjected to two normalizations (Sinclair and Ludlow, 1986;Ray and Sinclair, 1997).The fi rst normalization was the application of equations 1 and 3 to TR and LGR, respectively, so that TR and LGR varied from zero to one and enabled large daily environmental variations during the experimental periods to be minimized.The second normalization aimed to reduce variations between plants caused mainly by differences in plant size.For the second normalization, TR and LGR of individual plants were divided by the mean of TR and LGR before water stress developed in the soil for each plant, i.e. while FTSW was equal to or greater than 0.5 (Lago et al., 2012), as recommended by the protocol (Devi et al., 2009;Gholipoor et al., 2012), resulting in the NTR and NLGR variables.
The NTR and NLGR data (after the second normalization) were fi tted to the logistic-type equation Y = 1/{1+ exp[-a(X -b)]} (Sinclair and Ludlow, 1986;Ray and Sinclair, 1997), where Y is the dependent variable (NTR and NLGR), X is FTSW and "a" and "b" are empirical coeffi cients estimated by nonlinear regression analysis.As an indicator of tolerance to soil water deficit, the value of the FTSW threshold for NTR and NLGR was estimated as the value of FTSW when the NTR or NLGR from the logistic equation equaled 0.95 (Lago et al., 2012).This approach is similar to the two-segment linear regression approach used by Devi et al. (2009) and Gholipoor et al. (2012).
The variables total transpiration (total water extracted by plants during the experiment, in g per plant), CT (L kg -1 dry mass) and FTSW threshold for NTR and NLGR were analyzed according to a 4 × 2 factorial scheme with each plant being a replicate, A factors were the clones (four levels) and B factors were the soil water regime treatments (two levels = irrigated and non-irrigated).These data were subjected to analysis of variance (ANOVA) and means were compared by the Tukey test at 5 % probability of type I error.Normality of the data was tested with the Shapiro-Wilk test before performing ANOVA.

Results and Discussion
The maximum daily air temperature varied little during the experimental period, mainly in E1, due to the control system of the greenhouse that was set to 32 ºC, while the daily minimum air temperature varied considerably (Figure 1A).The minimum temperatures varied from 9.2 to 20.0 °C in E1 and from 4.0 to 19.6 °C in E2, and the maximum temperatures varied from 26.4 to 32.0 °C in E1 and from 21.7 to 31.6 °C in E2.
The vapor pressure defi cit (DPV) at 15h00 in E2 ranged from 1.0 to 4.6 hPa (Figure 1B), which is in the very low VPD range (below 15 hPa, Kiniry et al., 1998), indicating that the plants grew and developed in an atmosphere of low evaporative demand.The low levels of VPD in the afternoon are due to the maximum temperature control system in the greenhouse which, upon reaching the maximum temperature of 32 °C, fl ipped on a ventilation system that forced air to pass through a wall of soaked expanded clay, forcing moist air into the greenhouse.Even though not measured, similar afternoon VPDs are expected to occur during E1 because during this experiment the fan system turned on even more often than during E2 due to higher solar radiation and longer days during E1 as compared with E2 (Figure 1C).The meteorological conditions (temperature, photoperiod and solar radiation) during E1 and during E2 (Figure 1) are typical of fi eld conditions when potato is grown during spring and fall, respectively, in this region (Bisognin et al., 2008).
Potato plant development is mainly driven by temperature (Streck et al., 2007).Due to higher minimum temperatures in E1 than E2 (Figure 1A), the length of the fi rst experiment was shorter (24,23,23  For the transpiration and coeffi cient of transpiration (CT) variables there were no signifi cant clone x soil water regime interactions (Table 1).There was no difference in total transpiration among clones, and plants without water defi cit had higher transpiration than plants with water defi cit, as expected.The coeffi cient of transpiration was lower in plants with water defi cit in both experiments, indicating higher water use effi ciency when plants were subjected to a drying soil.Among the clones, no differences were found in coeffi cient of transpiration when T1 and T2 values were plotted, as indicated by ANOVA (interaction clones x soil water regimes was not signifi cant).Looking at the CT values only in T2 (Table 1) clones SMINIA 00017-6 and SMINIA 793101-3 had lower CT than cultivar Asterix, indicating that these two clones are more water use effi cient.
The small differences in CT among clones (Table 1) are probably due to the short period of water deficit (10 to 13 days) beyond the FTSW threshold.Defense mechanisms against water defi cit in the long term such as increased root growth and leaf senescence need more time to act, the stomatal closure being the most effective mechanism to defend the plant from water defi cit in the short term (Sinclair and Ludlow, 1986;Lago et al., 2011).In both experiments E1 and E2, dry root matter did not differ between irrigated and non-irrigated treatments (Table 1) and no leaf senescence was observed in non-irrigated plants, indicating that long-term defense mechanisms against soil water shortage were not activated in the plants.
The response of NTR to FTSW for the four potato clones in both experiments is shown in Figure 2. The variability of the data is high, especially in the FTSW range between 0.5 and 1.0, a common feature in experiments using the FTSW approach with other crops (Sinclair and Ludlow 1986;Ray and Sinclair, 1997;Sinclair et al., 2005;Wahbi and Sinclair, 2007;Martins et al., 2008).The lower the air temperature, solar radiation and VPD, the greater the variability of the NTR data (Lago et al., 2011).
Besides the variability in the data, the logistic equation described well the NTR response to the FTSW for the different clones and experiments (Figure 2).As FTSW decreases, the NTR varied around the maximum value, and after reaching an FTSW threshold, the NTR started to decrease due to stomatal closure (Ludlow and Sinclair, 1986), one of the key defense mechanisms that the plant uses to protect from soil water defi cit in the short term (Taiz and Zeiger, 2006).Variability of the data was lower in the range of FTSW values between zero and 0.4, indicating that the response of NTR to FTSW in the range that causes stomata closure is well defi ned, even in a low VPD atmosphere.Therefore, the lower variability of the data at low FTSW indicates that the approach of using the FTSW threshold as an indicator for tolerance to water shortage is appropriate.
The FTSW threshold for NTR was higher in the spring experiment (E1) than in the fall experiment (E2) for all clones (Figure 2).The main hypothesis for the lower FTSW threshold in E2 was the lower availability of solar radiation this time of year, resulting in lower potential transpiration, as compared to E1, as plants in E2 took about 10 days longer to extract the water from the pots (duration of the experiment was, on the average of the four genotypes, 23 days and 33 days in E1 and E2, respectively).
Among the potato clones, in E1 the lowest FTSW threshold for NTR was for Asterix (0.28) and the highest for SMINIA 02106-11, SMINIA 00017-6 and SMINIA  Means followed by the same lowercase letters in rows and uppercase letters in columns are not differ (Tukey test, p < 0.05).CV = coeffi cient of variation.
793101-3 (0.34, 0.37 and 0.41) respectively, and in E2 the lowest FTSW threshold was for SMINIA 793101-3 (0.19), followed by Asterix (0.24), SMINIA 00017-6 (0.26) and SMINIA 02106-11 (0.28).The range of FTSW threshold values in this study (0.19 to 0.41) is slightly greater than the range of values found in other studies on potato, such as found in Weisz et al. (1994) (from 0.20 to 0.35) for cultivars BelRus and Katahdin, respectively, and overlaps partially the range reported by Lago et al. (2012) of 0.28 to 0.49 for the Macaca cultivar and for advanced clone SMINIA 793101-3, respectively.The advanced clone SMINIA 793101-3 had the highest FTSW threshold in E1 (Figure 2E) and the lowest FTSW threshold in E2 (Figure 2F).The response of NLGR to FTSW for the four potato clones in both experiments is shown in Figure 3.The variability of the data is greater than in Figure 2. In previous studies with other crops, the leaf growth response to FTSW also had greater variability than the relative transpiration (Muchow and Sinclair, 1991;Lago et al., 2011).The FTSW threshold was higher in E1 than in E2, except for the cultivar Asterix.In E1, FTSW threshold for NLGR was the greatest for clone SMINIA 00017-6 (0.80) (Figure 3C) and the lowest for cultivar Asterix (0.31) (Figure 3G) whereas in E2 the lowest FTSW threshold was for clones SMINIA 02106-11 and SMINIA 793101-3 (0.29) (Figure 3B, 3F), followed by clone SMI-NIA 00017-6 (0.37) (Figure 3) and cultivar Asterix (0.88) (Figure 3H).The range of FTSW threshold values ob-tained here (0.29-0.88) is higher than the range reported by Lago et al. (2012), from 0.29 to 0.49, and includes the 0.6 value obtained by Weisz et al. (1994), with clones adapted to temperate environments.
The NLGR began to decrease before (at a higher FTSW, Figure 3) NTR (at a lower FTSW, Figure 2), for all clones in both experiments except for the clone SMI-NIA 02106-11 in E1 (Figure 3A).This anticipation of the reduction in leaf growth compared to the reduction in transpiration in a drying soil indicates a passive hydraulic response of potato leaves to water defi cit.A decrease in leaf cell turgor led to the activation of the stomatal closure mechanism, thus causing a reduction in leaf cell division and expansion and, consequently, reducing leaf growth (Marenco and Lopes, 2005;Taiz and Zeiger, Tolerance to water defi cit is a genotype-dependent trait.In a previous study, Lago et al. (2012) found a higher tolerance to soil water defi cit in the clone SMINIA 793101-3, when compared with Macaca, an old cultivar but still widely used by potato producers in Rio Grande do Sul.In this study, the clone SMINIA 793101-3 had the highest FTSW threshold in E1 (Figure 3E), confi rming results of Lago et al. (2012), but it had the lowest FTSW threshold in E2 (Figure 3F), which does not agree with results reported by Lago et al. (2012).In E2 solar radiation was lower than in E1 (fi gure 1C), typical of the fall weather in this location and it took more time for plants for depleting soil water than in E1.Lago et al. (2012) also found a lower FTSW threshold (0.33) for the clone SMINIA 793101-3 when the experiment was conducted in the fall (average solar radiation was 8.6 MJ m -2 day -1 ) as compared with the FTSW threshold (0.49) when their experiment was conducted in the spring (average solar radiation was 16.1 MJ m -2 day -1 ).Therefore, results from our experiments and those from Lago et al. (2012) consistently indicate that when clone SMINIA 793101-3 grows under high solar radiation, stomata close at a higher FTSW whereas when it grows under low solar radiation, stomata close at a much lower FTSW than other clones.Gholipoor et al. (2012) evaluated sixteen sorghum genotypes and identifi ed two groups of hypothesized traits: one group exhibited high FTSW thresholds for conserving water in a soil drying cycle, while another group allowed transpiration at a maximum rate until the soil dried to a lower FTSW threshold.In order for the second group to conserve water and have some tolerance to low soil water contents, these genotypes must have low transpiration rates while soil is still in the well-watered phase, mainly early in the season (Gholipoor et al., 2012).In our study, the cultivar Asterix (which had the lowest FTSW threshold) had one of the highest transpi-ration rates in well-watered plants during experiment E1 (Table 1), indicating that this cultivar has lower tolerance in a drying soil as compared with the other three clones.This cultivar is currently widely used by potato growers in the State of Rio Grande do Sul.The clones SMINIA 02106-11 and SMINIA 00017-6 consistently had a higher FTSW threshold (Figure 3A, 3B, 3C, 3D) as compared with cultivar Asterix (Figure 3G, 3H) in both experiments.These results indicate that these two clones that are adapted to subtropical environments have good quality chips are resistant to major diseases (Bisognin et al., 2008), and also have tolerance to water defi cit in the soil.
The trait tolerance to water shortage of the three clones is particularly important if the trend of years with rainfall below normal observed in Rio Grande do Sul during the fi rst decade of this century continues in the coming years (Streck et al., 2009).Thus, the fi rst step which is to identify from the already developed potato clones those with a difference in FTSW threshold in drying soil, has been addressed.
We used the FTSW approach (Sinclair and Ludlow, 1986) to identify differences in tolerance to water defi cit in potato genotypes.This approach assumes that the water content in the soil used by the plant for transpiration varies between fi eld capacity (when transpiration is at the maximum rate) and the water content in the soil when transpiration of plants under stress equals 10 % of maximum transpiration (Sinclair and Ludlow, 1986).Compared to other approaches such as the total extractable soil water (Ritchie, 1973(Ritchie, , 1981)), the FTSW is a better indicator of the actual amount of soil water that can be extracted by plants for transpiration (Ludlow and Sinclair, 1986).
The response of plant transpiration to FTSW has two phases (Ludlow and Sinclair, 1986): (i) phase I, in which transpiration takes place at the maximum rate in the range of FTSW from one (fi eld capacity) to the onset of transpiration reduction due to stomatal closure (threshold FTSW), and (ii) phase II, in which transpiration decreases almost linearly to a minimum value as soil water decreases from an FTSW threshold down to zero.The greater the FTSW threshold the earlier stomata close in a drying soil, which allows for conservation of water in the soil and, therefore, more time is available for the plant to activate long-term strategies for accessing more or conserving the available soil water so as to increase the growth of roots and to reduce the transpiring leaf area.
In addition to helping in identifying water-defi cit tolerant genotypes, the response curves of transpiration and leaf growth versus FTSW (Figures 2 and 3) also have a direct application in potato simulation models.Fagundes et al. (2010) calibrated such models for the Rio Grande do Sul conditions, but the effect of water defi cit was not taken into account, that is, it was considered that the crop was irrigated (near optimum conditions).The next step, now possible, is to take into account the effect of water defi cit on transpiration and incorporate this effect into models of the potato cultivars Asterix and Macaca and the new clones.

Conclusions
The fraction of transpirable soil water (FTSW) approach is an effi cient method to separate potato clones differing in their responses to water defi cit.The ad-vanced clones SMINIA 02106-11 and SMINIA 00017-6 are more tolerant to soil water defi cit than the cultivar Asterix under both high and low solar radiation.The clone SMINIA 793101-3 is more tolerant to soil water defi cit than the cultivar Asterix under high solar radiation.

Figure 1 -
Figure 1 -(A) Daily minimum (Tmin) and maximum (Tmax) air temperatures during the two experiments (E1 and E2), (B) daily air vapor pressure defi cit (VPD) at 3 p.m. during the second experiment (E2) and (C) photoperiod (P) and solar radiation (SR) during the two experiments (E1 and E2).Santa Maria, RS, Brazil, 2010-2011.Each experiment began the day after soil in the pots was water saturated.