Seed size variation in cold and freezing tolerance during seed germination of winterfat (Krascheninnikovia lanata) (Chenopodiaceae) Ruojing Wang, Yuguang Bai, Nicholas H. Low, and Karen Tanino
Abstract: Native plants have adaptations to their local environments and elucidation of these traits has implications in both agronomy and restoration ecology. Winterfat (Krascheninnikovia lanata (Pursh) A.D.J. Meeuse & Smit) is a native perennial shrub in North America capable of germinating at low temperatures. The effect of seed size on germination ability at low and subzero temperatures and the physiological mechanisms were investigated. Winterfat seeds achieved 50%– 72% germination at –3 8C, a temperature slightly above the base temperature estimated using thermal time models. Small seeds required a longer time to reach 50% germination at subzero temperatures than large seeds. Large seeds maintained stable water uptake rate for both the seed and the embryo when temperatures decreased from 5 to –1 8C. In contrast, faster water uptake and greater relative K+ leakage in small seeds indicated possible damage to membrane integrity at subzero temperatures. Carbohydrate conversion efficiency (Rq/RCO2) of large seeds was significantly higher than that of small seeds at 10 8C but not at 20 8C. Higher cold resistance in large seeds was also correlated with higher concentrations of glucose, raffinose, and sucrose. This study revealed the potential basis of the low-temperature germination advantage of large seeds and provided the first direct evidence of germination under freezing temperatures in winterfat. Key words: calorimetry, freezing tolerance, potassium leakage, seed size, soluble sugars, thermal time, winterfat (Krascheninnikovia lanata). Re´sume´ : Les plantes indige`nes posse`dent des adaptations a` leurs environnements locaux, et la compre´hension de ces caracte`res a des implications en agronomie aussi bien qu’en e´cologie de la restauration. L’eurotie laineuse (Krascheninnikovia lanata (Pursh) A.D.J. Meeuse & Smit) est un arbuste indige`ne pe´renne de l’Ame´rique du Nord, capable de germer a` basse tempe´rature. Les auteurs ont e´tudie´ l’effet de la dimension de la graine sur la capacite´ de germination a` des tempe´ratures basses, ou sous ze´ro, ainsi que les me´canismes physiologiques sous-jacents. Les graines de l’eurotie germent de 50 % a` 72 % a` –3 8C, une tempe´rature le´ge`rement au-dessus des tempe´ratures de base estime´es par les mode`les de dure´e thermique. Aux tempe´ratures sous ze´ro, les petites graines ne´cessitent plus de temps pour atteindre une germination de 50 %, que les grosses graines. Les grosses graines maintiennent un taux constant d’absorption de l’eau, lorsque les tempe´ratures chutent de 5 a` –1 8C. Au contraire, une absorption plus rapide de l’eau et une perte relativement plus grande de K+, chez les petites graines, indiquent la pre´sence de dommage potentiel a` l’inte´grite´ des membranes, aux tempe´ratures sous ` 10 8C, l’efficacite´ de la conversion des glucides (Rq/RCO2), chez les grosses graines, est significativement plus e´leze´ro. A ve´e que chez les petites graines, mais pas a` 20 8C. La re´sistance plus e´leve´e au froid, chez les grosses graines, est e´galement corre´le´e avec les teneurs en glucose, raffinose et sucrose. Cette e´tude re´ve`le la base potentielle de l’avantage de la germination a` basse tempe´rature des grosses graines, et constitue la premie`re preuve directe de la germination de l’eurotie laineuse a` des tempe´ratures sous le point de conge´lation. Mots cle´s : calorime´trie, tole´rance au gel, perte de potassium, grosseur des graines, sucres solubles, dure´e thermique, eurotie laineuse (Krascheninnikovia lanata). [Traduit par la Re´daction]
Introduction Received 27 June 2005. Published on the NRC Research Press Web site at http://canjbot.nrc.ca on 15 March 2006. R. Wang,1 Y. Bai,2 and K. Tanino. Department of Plant Sciences, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK S7N 5A8, Canada. N.H. Low. Department of Applied Microbiology and Food Science, University of Saskatchewan, 51 Campus Drive, Saskatoon, SK S7N 5A8, Canada. 1Present
address: Agriculture and Agri-Food Canada, Saskatoon Research Centre, 107 Science Place, Saskatoon, SK S7N 0X2, Canada. 2Corresponding author (e-mail: [email protected]
). Can. J. Bot. 84: 49–59 (2006)
The inability of many agronomic crops to germinate under low-temperature conditions in early spring limits agricultural productivity in northern environments. To improve germination of these crops, the understanding of mechanisms involved in the germination of native species that have adapted to the low-temperature conditions of their native habitats would be beneficial. Winterfat (Krascheninnikovia lanata (Pursh) A.D.J. Meeuse & Smit) is a long-lived native shrub with superior forage quality for livestock and wildlife (Coupland 1950; Smoliak and Bezeau 1967) and is also an important ecological component of the mixed prairie of North America
2006 NRC Canada
(Call and Roundy 1991; Romo et al. 1995). Winterfat is adapted to germinate under the subzero temperature conditions of early spring in these regions. Larger seeds of winterfat germinate faster and at higher percentage than smaller seeds, especially at low temperatures (Springfield 1973; Hou and Romo 1998; Wang et al. 2004). Early-emerging winterfat seedlings have greater seedling growth and a better survival rate than late-emerging plants (Hou and Romo 1998), and relatively heavier seeds are desirable for restoration. The selection for increasing seed size is positively associated with faster germination (Charlton 1989) and greater seedling vigor at low temperatures (Kelman and Forrester 1999). However, the mechanisms for the different performance among seed sizes, especially under low temperatures, are rarely studied in spite of its importance for understanding the adaptation of plants to local environments. Seed mass can vary widely within a population of a species (Zhang and Maun 1990; Zhang 1998) depending on biological and environmental constraints (Vaughton and Ramsey 1997, 1998). Populations with different mean seed weights are thought to have evolved under different selection pressures, ecologically increasing their potential fitness (Westoby et al. 1992). Variance in seed size is also pronounced within an individual plant, attributable to both parental identity and fruiting positions (Mendez 1997; Vaughton and Ramsey 1998; Simons and Johnston 2000). Seed weight is correlated with seed vigor (Lafond and Baker 1986; Berdahl and Frank 1998), seedling recruitment (Mendez 1997; Susko and Lovett-Doust 2000; Dalling and Hubbell 2002; Debain et al. 2003) and plant size and the probability of survival (Hou and Romo 1998; Simons and Johnston 2000). Physiological mechanisms related to plant cold tolerance in adult plants include the accumulation of compatible solutes such as sugars and proline, which protect macromolecules during stress (Hoekstra et al. 2001), the increase in cell membrane fluidity and integrity (Sharom et al. 1994; Saltveit 2002; Wisniewski et al. 2003), and the maintenance of higher metabolic rates under stress temperatures (Massardo et al. 2000). Dramatic changes in physiological and biochemical activities take place during seed germination and can be correlated with phases in water uptake (Bewley and Black 1994; Bewley 1997). Generally, leakage of cell solutes peaks at the initial imbibition stage during cell membrane repair (Bewley and Black 1994). Ionic leakage may be a more sensitive measurement of membrane integrity than bulk conductivity (Ouyang et al. 2002), and potassium has been shown to be a better indicator of cell membrane integrity (Dias et al. 1996). Therefore, the ionic leakage of seeds during imbibition not only affects membrane repair and functional recovery but also reflects freezing tolerance. Besides cellular function recovery, respiration is the first detectable metabolic activity providing energy for germination (Bewley and Black 1994). Plant metabolic activities can be studied using calorimetry, which measures the metabolic heat production rate (Rq) from catabolism and anabolism (Thygerson et al. 2002). Plant growth rates are correlated with heat production rates and respiration rates, and the latter can be measured by CO2 production rate (RCO2 ) from respiration (Criddle et al. 1991). Metabolic heat and CO2 production rates have been used to measure meta-
Can. J. Bot. Vol. 84, 2006
bolic responses to environmental stress (Smith et al. 2000; Thygerson et al. 2002), respiratory substrates (Criddle et al. 1997; Edelstein et al. 2001), and growth rates (Hansen et al. 1994). The ratio Rq =RCO2 provides a reference point for respiratory substrates used by the tissue (Hansen et al. 1997). Seed reserves and their composition affect respiration, metabolism, and stress tolerance. Soluble sugars such as sucrose and raffinose family are compatible solutes for maintaining enzyme structure and function and membrane integrity under stress (Bailly et al. 2001; Hoekstra et al. 2001; Uemura and Steponkus 2003). The objectives of this study were to determine physiological differences between large and small seeds of native winterfat at low and subzero temperatures and to study the mechanism associated with these differences. Specifically, we investigated the impact of seed size on (i) germination at subzero temperatures and the thermal time model, (ii) membrane integrity, soluble sugars, and water uptake characteristics at low and subzero temperatures, and (iii) metabolic activities at low temperatures. We hypothesize that large seeds of winterfat are more resistant to low and freezing temperature stress than small seeds, which will enable large seeds to germinate better under subzero temperatures.
Materials and methods Seed sources and seed size classification Winterfat seeds (diaspores), collections Cela and No. 63, were purchased from Wind River Seed (Manderson, Wyoming). Both collections originated from Utah (Cela from Tooele (40831’N, 112818’W) and No. 63 from Delta (39821’N, 112834’W)) with estimated minimum germination temperatures of –3.5 and –4.5 8C (Wang et al. 2004). The diaspores were cleaned by rubbing, fanning, and passing seeds through serial sieves and blowers. Cleaned seeds were separated using a seed blower into two classes based on seed density and hereafter were referred to seed size classes ‘‘large’’ and ’’small’’. Seed weight was significantly different and seed moisture content was similar between large and small classes within each collection. Cleaned and classified seeds were then sealed in plastic bags and stored at –18 8C until use. Details of climate characteristics and seed sizes and collections have been provided previously (Wang et al. 2004). The ratio of embryo to total seed weight for two seed size classes was calculated gravimetrically. Ten seeds of each seed size class and collection were hydrated for 30 min on top of two moist filter papers in Petri dishes at room temperature (22.5 ± 1 8C). The hydrated seeds were excised under a microscope into embryo and perisperm with the seed coat. The dry weights of seeds and embryos were determined using a microbalance (±1 mg) after oven drying at 80 8C for 48 h. Germination tests at subzero temperatures Large and small seeds of collections Cela and No. 63 germinated at –1 or –3 8C, respectively, using a circulating liquid bath (ethylene glycol, temperature stability ±0.01 8C; Cole Parmer, Vernon Hills, Illinois). Three replicates of 50 seeds each were used. Seeds were placed in 5 cm Petri dishes on top of two layers of moist Whatman No. 1 filter #
2006 NRC Canada
Wang et al.
Fig. 1. Seed mass and its allocation to the embryo in large (L) and small (S) seeds of collections Cela and No. 63 in winterfat (Krascheninnikovia lanata). Means with the same letters are not significantly different within a collection at P £ 0.05.
Table 1. Final germination percentage of viable seeds (maximum germination at optimum temperature) and days to 50% germination (D50) at subzero temperatures in winterfat (Krascheninnikovia lanata). –1 8C Collection and seed size No. 63, large No. 63, small Cela, large Cela, small
Final germination % 79.9±2.5a 72.9±13.1a 80.4±9.8a 68.9±11.7a
–3 8C D50 (d) 8.1±1.4b 15.2±4.3a 8.8±2.1b 16.9±5.0a
Final germination % 69.8±12.7a 60.1±13.5a 71.9±14.6a 50.2±12.2b
D50 (d) 15.1±4.1b 25.3±6.6a 19.4±4.6b 31.1±5.6a
Note: Means (±SD) with the same letters are not significantly different within a column and a collection at P £ 0.05; D50 was estimated from a polynomial function derived from the real germination time courses (days) of both seed sizes and collections.
paper. To prevent ice formation at subzero temperatures, 3 mL of polyethylene glycol-6000 solution of –0.20 MPa at each temperature was used for germination. Preliminary tests indicated that ice did not form in this solution and seed germination rate and percentage at –0.20 MPa were similar to those at 0 MPa at temperatures >2 8C. Petri dishes were sealed with several layers of plastic bags to ensure that ethylene glycol solution did not enter the Petri dishes. They were placed in the middle of the circulating liquid bath. Germination was monitored at 7 d intervals for –1 8C and 15 d intervals for –3 8C up to 30 d. Germination was checked on an ice bed at each time interval. Seed water uptake at low temperatures as affected by seed size Five replicates of 20 seeds each from two seed size classes and collections were imbibed in 5 cm Petri dish with 3 mL of distilled water on top of the saturated filter paper at 5 or –1 8C, respectively. The weight of an individual seed from each replicate was repeatedly measured using a microbalance (±1 mg) at 2, 4, 12, 24, and 48 h intervals during imbibition. Embryos were excised and weighed from
different seeds of each replicate at each time interval. Seeds or embryos were blotted gently with tissue paper to remove surface water before weighing. Dry weights of seeds or embryos were determined gravimetrically after oven drying at 80 8C for 48 h and changes in water content were calculated on a dry weight basis. Cell membrane integrity at low temperatures as measured by K+ leakage Ten seeds, free of morphological damage, for each treatment were selected from each seed size class and collection and weighed individually to determine their air dry weight. Each seed was imbibed with 1 mL of deionized water in 1.5 mL microcentrifuge tubes at 30 8C for either 1.5, 3.5, 5.0, or 6.5 h, at 10 8C for either 4, 8, 12, or 16 h, or at –1 8C for either 24, 48, 72, or 96 h, respectively. Imbibition intervals for different temperatures were determined based on time to reach similar thermal time accumulation (Wang et al. 2004) and comparisons of different temperatures were expressed on a thermal time course. The leaching solutions were kept at 5 8C until use after removing seeds at each time interval. Two drops of polyethylene glycol-6000 #
2006 NRC Canada
Can. J. Bot. Vol. 84, 2006
Fig. 2. Predicted and observed thermal time courses (degree-day) of germination at –3 and –1 8C for large (L) and small (S) seeds of collections Cela and No. 63 in winterfat (Krascheninnikovia lanata). Predicted values were calculated from thermal time models with estimated parameters in Wang et al. (2004). Fitted lines were based on the least squares fit sigmoid function.
solution at –0.20 MPa were added to deionized water to prevent ice formation at subzero temperatures. An atomic absorption spectrometer (Spectr AA-220; Varian Scientific Instrument, Palo Alto, California) was used to measure K+ concentration in each solution. Deionized water and deionized water containing the polyethylene glycol-6000 solution were used as the control for temperatures above and below 0 8C, respectively. Relative K+ leakage was calculated based on maximum leakage, which was the K+ concentration after boiling seeds for 30 min. Air dry weight of seeds was used to calculate relative K+ leakage as recommended by the Seed Vigor Test Committee of the Association of Official Seed Analysts (1983). Heat production rate and respirationary CO2 production rate Microcalorimetry (multicell differential scanning calorimeter MDSC 4100DSC; Calorimetry Science Corporation, American Fork, Utah) was used for studying the rate of heat production from seeds. Five large seeds (total weights of 16–20 mg) and eight small seeds (total weights of 12– 18 mg) from collection No. 63 were weighed, placed in plastic cassettes, and soaked in 1% sodium hypochlorite for 10 min for surface sterilization. Sterilized seeds germinated in a sterilized Petri dish with filter paper were free of microorganism contamination. Rinsed seeds were placed in a 5 8C incubator for 16 and 24 h, respectively. Metabolic heat production rates, Rq, were measured up to 3000 s at 10 8C and 2000 s at 20 8C after 10–20 min of equilibrium, respectively
(Smith et al. 2000). The Rq was measured both before and after measuring the CO2 respiration rate and the average value was used. CO2 respiration rate, RCO2 , was measured as the rate of heat production from respired CO2 forming carbonate in a NaOH trap and calculated as the difference in the rates of heat production with and without the carbonation. The reaction of CO2 with NaOH solution to form carbonates produces –108.5 kJ/mol and there was 25 mL of 0.4 mol NaOH/L placed in the trap (Criddle et al. 1997). The Rq and the RCO2 were converted to mW/mg air dry weight and the ratio Rq =RCO2 was multiplied by 108.5 kJ/mol (Criddle et al. 1997). Changes in soluble sugar concentration during imbibition at low temperatures As a side test, four replicates of 10 large seeds and 10 small seeds of collection No. 63 were imbibed at 5 8C in distilled water for each of 0, 12, 24, 48, and 72 h. Soluble sugars were extracted from these seeds with seed air dry weight of 15–30 mg. Two millilitres of distilled deionized water (Milli QTM water system; Millipore, Milford, Massachusetts) at 98 ± 2 8C was added to seed samples and the samples were grounded to a fine mixture using a pestle and mortar. The mixture was incubated at 90 8C for 10 min to inactivate enzymes using a microtube incubator (incuBlock; Denville Scientific Inc., Metuchen, New Jersey) and then centrifuged at 10 400g for 10 min. The supernatants were diluted to an appropriate concentration (within 40 mg/L) and passed through a 13 mm diameter nylon syringe filter #
2006 NRC Canada
Wang et al.
Fig. 3. Water uptake with increment of imbibition thermal time at 5 and –1 8C for large and small seeds of collections Cela and No. 63 in winterfat (Krascheninnikovia lanata). Thermal time was based on estimated base temperatures (Tb) in Wang et al. (2004). The estimated Tb values for large and small seeds were –4.8 and –4.2 8C for collection No. 63 and –3.8 and –3.3 8C for collection Cela. Fitted lines were based on the least squares fit sigmoid function.
(0.2 mm pore size; Chromatographic Specialties Inc., Brockville, Ontario) prior to analysis. The concentrations of fructose, glucose, raffinose, and sucrose were determined with a Dionex Bio LC 4000 gradient liquid chromatography unit (Dionex, Sunnyvale, California) equipped with a pulsed amperometric detector (Dionex). The working gold electrode, during operation, was maintained at the following potentials and durations: E1 = 0.05 V (t1 = 0.299), E2 = 0.60 V (t2 = 0.299), and E3 = –0.80 V (t3 = 0.499) at a sensitivity of 10 000 nA. Isocratic (80 mmol NaOH/L) separation was achieved using a Dionex CarboPac PA1 anion-exchange column (250 mm 4 mm) equipped with a Dionex CarboPac PA1 guard column (50 mm 4 mm). The mobile phase flow rate was 1 mL/min and the sample volume was 50 mL. Standard curves for glucose, fructose, raffinose, and sucrose were constructed at concentrations ranging from 10 to 188 mg/L. Standard curves of each sugar had correlation coefficients of 0.99 or better. Approximate retention times for fructose, glucose, sucrose, and raffinose were 5.5, 6.3, 9.8, and 25.8 min, respectively. The presence of raffinose in seeds was confirmed by the analysis of two samples with raffinose standard (100 mg/L) and spiking experiments. Lyophilized samples were derivatized by adding 0.5 mL of Tri-Sil Z (Chromatographic Specialties Inc., Brockville, Ontario). The vials were capped and the solutions heated at 80 8C for 1 h. The trimethylsilylated carbohydrates
were analyzed with a gas chromatography (HP6890; Hewlett-Packard, Mississauga, Ontario) equipped with an autosampler. Data analysis Thermal time (degree-hour) is a measure of biological time rather than actual time and was used in germination time expressions to make germination or physiological responses at different temperatures comparable. According to the thermal time model (Garcia-Huidobro et al. 1982), thermal time to germination percentage g, yT(g), is yTðgÞ ¼ ðT Tb ÞtðgÞ where T is actual temperature, Tb is the base temperature for germination, and t(g) is the time to germination percentage g. The values of Tb were estimated from the intercepts of linear response of germination rates (1/t(g)) to temperatures at suboptimal temperature range. The estimated Tb values for large and small seeds were –4.8 and –4.2 8C for collection No. 63 and –3.8 and –3.3 8C for collection Cela (Wang et al. 2004). Germination percentage was scaled by the mean maximum germination percentage for each seed collection and each seed size class, which ranged from 83% to 95%. Percentage data were arcsine transformed for mean separation. The general linear model procedure and least significant difference in SAS1 Windows version 8.2 at a significance #
2006 NRC Canada
Can. J. Bot. Vol. 84, 2006
Fig. 4. Relative K+ leakage at different imbibition time intervals at 30, 10, and –1 8C for large and small seeds of collections Cela and No. 63 in winterfat (Krascheninnikovia lanata). The relative leakage was calculated as the percentage of maximum leakage, which was the leakage after seeds were boiled for 30 min.
level of 0.05 were used to analyze all data. SigmaPlot1 was used for curve fitting and estimation of days to 50% germination.
Results Seed mass and relative embryo mass allocation in large and small seeds Seed mass was significantly different between the two seed size classes in both collections (P < 0.0001) (Fig. 1). However, the ratios of the embryo to seed were not significantly different (P = 0.119) between large and small seeds in the two collections. More than 70% of the seed mass was allocated to the embryo, indicating that the embryo to seed ratio is a conserved trait in winterfat. Germination at subzero temperatures Winterfat seeds germinated from 69% to 80% at –1 8C and from 50% to 72% at –3 8C after 30 d (Table 1, Fig. 2). The germination percentage was 20%–40% lower than that predicted by the thermal time model (Wang et al. 2004). Parameters of the thermal time model were estimated using germination data at above-zero temperatures. The observed germination time courses diverged from that predicted by
the thermal time model, which under- or overestimated germination at –1 and –3 8C, respectively. Although large seeds in both collections tended to have a higher final germination percentage than small seeds at –1 and – 3 8C, the data were only statistically significant at –3 8C for collection Cela (Table 1). The average days to 50% germination was much longer in small seeds than in large seeds according to real germination time courses (Table 1). The days to 50% germination for small seeds was nearly double that for large seeds in both collections. Water uptake is affected by seed size The water contents of small seeds and their embryos, but not those of large seeds and their embryos, were higher when the temperature was decreased from 5 to –1 8C (Fig. 3). Small seeds and their embryos generally accumulated water faster at –1 8C than at 5 8C, especially for embryos. Average water uptake rates (mgmg–1h–1 100, the first 7 d) at –1 8C were 2.4 and 1.3 for small seeds and 1.6 and 0.9 for large seeds in collections Cela and No. 63, respectively. Water uptake rates at –1 8C were significantly different between large and small seeds (P = 0.015 and 0.039 for Cela and No. 63, respectively). However, there were no significant differences in water uptake rates be#
2006 NRC Canada
Wang et al.
Fig. 5. Relationships between heat production rates (Rq) and seed mass at 10 and 20 8C following 16 and 24 h of imbibition at 5 8C in winterfat (Krascheninnikovia lanata) (collection No. 63). The P value indicates the significance level of the regression equation.
Fig. 6. Relationships between respiration CO2 production rates (RCO2 ) and seed mass at 10 and 20 8C after 16 and 24 h of imbibition at 5 8C in winterfat (Krascheninnikovia lanata) (collection No. 63). The P value indicates the significance level of the regression equation.
2006 NRC Canada
Can. J. Bot. Vol. 84, 2006
Fig. 7. Changes in soluble sugar concentrations during imbibition at 5 8C for large and small seeds of winterfat (Krascheninnikovia lanata) (collection No. 63).
tween large and small seeds at 5 8C, ranging from 1.4 to 1.7. The relative increase in water content from –1 to 5 8C was greater in small seeds and their embryos than in large seeds and their embryos. Relative K+ leakage as affected by seed size The relative K+ leakage during imbibition at 10 or 30 8C was similar in large and small seeds in three of the four comparisons; the only significant difference occurred for No. 63 at 30 8C. Significant differences between large and small seeds in both collections were evident at –1 8C for both collections (P = 0.02 and 0.03 for No. 63 and Cela, respectively) (Fig. 4). Large seeds had lower relative K+ leakage than small seeds at the early stage of imbibition, suggesting differential membrane integrity between large and small seeds at –1 8C. K+ leakage of small seeds was also increased at 30 8C in collection No. 63, which might be an indication of high temperature stress. Metabolic rates of large and small seeds during imbibition Rates of heat production per milligram air dry weight were negatively correlated with seed mass at 10 and 20 8C after 16 and 24 h following imbibition at 5 8C (Fig. 5). Rates of respiratory CO2 production were also negatively correlated with seed mass, except for 10 8C after 16 h of imbibition at 5 8C (Fig. 6). The Rq and RCO2 were temperature dependent, increasing with increasing temperature. The lower the values of Rq =RCO2 , the higher the efficiency. The ratios Rq =RCO2 varied greatly among seeds with no significant differences between large and small seeds, except at
10 8C after 24 h of imbibition at 5 8C (P = 0.01); the ratio Rq =RCO2 was 382.2 ± 18.6 (SE) and 458.1 ± 17.2 (SE) kJ/mol for large and small seeds, respectively. Dynamics in soluble sugar concentration during imbibition as affected by seed size Initial concentrations of glucose and raffinose per unit seed mass were significantly (P