Growth and ethylene evolution by shade and sun ecotypes of Stellaria longipes in response to varied light quality and irradiance

Authors


Leonid V. Kurepin. E-mail: lkurepin@ucalgary.ca

ABSTRACT

Plants growing in the shade receive both low light irradiance and light enriched in far red (FR) (i.e. light with a low red (R) to FR ratio). In an attempt to uncouple the R/FR ratio effects from light irradiance effects, we utilized Stellaria longipes because this species has two distinct natural population ecotypes, alpine (dwarf) and prairie (tall). The alpine population occupies the open, sun habitat. By contrast, the prairie population grows in the shade of other plants. Both ‘sun’ and ‘shade’ ecotypes responded with increased stem elongation responses under low irradiance, relative to growth under ‘normal’ irradiance, and this increased growth was proportionally similar. However, only the shade ecotype had increased shoot elongation in response to a low R/FR ratio. By contrast, the sun ecotype showed increased stem elongation in response to increasing R/FR ratio. Varying the R/FR ratios had no significant effect on ethylene evolution in either sun or shade ecotype. Under low irradiance, only the sun ecotype showed a significantly changed (decreased) ethylene evolution. We conclude that R/FR ratio and irradiance both regulate growth, and that irradiance can also influence ethylene evolution of the sun ecotype. By contrast, R/FR ratio and irradiance, while having profound influences on growth of the shade ecotype, do not appear to regulate these growth changes via effects on ethylene production.

INTRODUCTION

Plants that exhibit a shade-avoidance syndrome grow under light conditions characterized by a low-red (R) to far-red (FR) ratio and low light irradiance [low photosynthetically active radiation (PAR)]. A low R/FR ratio (or FR enrichment) is believed to be the more important signalling factor (Ballare, Scopel & Sanchez 1990; Smith & Whitelam 1997; Smith 2000), although a low PAR is also an important signal in shade avoidance (Ballare, Scopel & Sanchez 1991; Ballare 1999; Franklin & Whitelam 2005). Plants that exhibit the shade-avoidance syndrome normally have etiolated internodes or stems and a reduced leaf development, and often exhibit early flowering (Smith 2000). When sunlight is simulated so that FR content can be varied, ‘sun’ plants (plants that normally grow in an open-field habitat) respond to an increasing FR content (as a proportion of total PAR) with higher rates of stem extension relative to ‘shade’ plants (plants that normally grow in a shaded environment) (Smith 1982).

Ethylene is a gaseous plant hormone that can inhibit stem elongation (Abeles, Morgan & Saltveit 1992), although in a few cases a growth-stimulatory role for endogenous ethylene has been reported (Raskin & Kende 1984; Jackson 1985; Rijnders et al. 1997). It has been suggested that ethylene may play a major role in the shade-avoidance syndrome. For example, a low concentration of applied ethylene can increase stem elongation in tobacco under shade (Pierik et al. 2003). Furthermore, ethylene-insensitive transgenic tobacco plants failed to elongate to the same extent as wild-type (WT) plants in response to a low R/FR ratio (Pierik et al. 2003). Leaf expansion in sunflower (Helianthus annuus) plants is promoted by low levels of ethylene, but inhibited by higher ethylene levels (Lee & Reid 1997). In another study, Sorghum bicolor plants that contain a null mutation in the gene encoding phytochrome B (phyB-1), and thus have a constitutive phenotype similar to shade plants, showed a constitutive overproduction of ethylene (Finlayson, Lee & Morgan 1998; Finlayson et al. 1999).

Stellaria longipes (L.) Goldie (Caryophyllaceae) is a good model species to study the effect of light quality and irradiance as it has two distinct population ecotypes with varying growth and morphological responses to sun and shade habitats. One population of S. longipes in Alberta, Canada is from an alpine habitat and grows in the open under bright sun conditions. We have chosen one genotype from this population, 1D, as our sun plant. Another population grows at a lower elevation in a foothill ‘prairie’ habitat. Here, mutual shading from neighbouring shrubs and grasses modifies appreciably both light quality and irradiance (Emery, Chinnappa & Chmielewski 1994a). The genotype chosen to represent this shade habitat is 7B, which showed very increased stem elongation in response to FR enrichment (i.e. low R/FR ratio) relative to stem growth under a normal R/FR ratio (Alokam, Chinnappa & Reid 2002). By contrast, the alpine 1D genotype showed only a slight increase in stem elongation under FR enrichment (Alokam et al. 2002).

Stellaria longipes alpine 1D and prairie 7B genotypes differ not only in their response to varied light quality, but also in their abilities to synthesize and respond to ethylene. For example, upon exposure to wind stress, the 1D genotype (but not the 7B genotype) showed a significant increase in ethylene evolution and a reduction in stem growth (Emery, Reid & Chinnappa 1994b).

In the present study, we uncoupled the two signals (R/FR ratio and PAR) of the shade-avoidance syndrome to understand better how each of light quality and light irradiance affects the growth of sun and shade plants. We have also examined the possible role of ethylene in mediating those responses.

MATERIALS AND METHODS

Plants

Two ecotypes of S. longipes, the 1D genotype (sun plants from an alpine habitat) and the 7B genotype (shade plants from a lower-elevation foothill grassland habitat) were collected originally from Plateau Mountain (2453 m) and Chain Lakes (1310 m) areas in south-western Alberta, Canada. Clonal ramets of both genotypes are maintained in glasshouses at the University of Calgary, with sequential treatments of short-day, cold (SDC) conditions [8:16 h light:dark, at 8:5 °C day:night, relative humidity (RH) approximately 50%] and long-day, warm (LDW) conditions (16:8 h light:dark, at 22:16 °C day:night, RH approximately 50%).

Prior to each Exp, ramets of both genotypes were excised at the soil surface and transferred to 10-cm-diameter (100 mL) pots with a soil mixture consisting of crushed baked clay medium and peat (1:2). After 3 weeks of growth under LDW in the glasshouses, the shoot tips were removed, leaving only two internodes above the soil surface. The ramets were then transferred to growth chambers (Controlled Environments, Winnipeg, Manitoba, Canada) maintained under SDC for a minimum of 60 d. Lights in these SDC growth chambers came from Philips 160 W fluorescent bulbs (Philips; Markham, Ontario, Canada), and PAR was maintained at 150 µmol m−2 s−1 (plant height) measured with Li-1800/22 quantum sensor (Li-Cor, Lincoln, NE, USA).

Experimental set-up

For light-quality Exps, we also used growth chambers (Controlled Environments) equipped with Sylvania cool white 60 W fluorescent bulbs (Sylvania, Mississauga, Ontario, Canada) and Philips 60 W incandescent bulbs. The R and FR light sources were light-emitting diode (LED) units (Quantum Devices, Barneveld, WI, USA) with R and FR light emissions that peak at 670 and 725 nm, respectively. Three R/FR ratios were used: 2.7 (abnormally high R enrichment); 1.9 (near-normal sunlight, emulating the un-shaded alpine habitat) and 0.7 (FR enrichment, a low R/FR ratio which emulates the shaded, lower-elevation grassland habitat). The R/FR ratio was measured as photon irradiance between 655 and 665 nm for R, and between 725 and 735 nm for FR. Under LDW growth chamber conditions, we maintained the PAR at the relatively low level of 115 µmol m−2 s−1 (measured at the soil surface). Thus, all R/FR ratio treatments were compared under LDW at this low PAR irradiance.

We also tested light irradiance as an experimental variable using the growth chambers (Controlled Environments) with Philips 160 W fluorescent and 60 W incandescent bulbs. One PAR treatment was 611 µmol m−2 s−1 (close to the PAR of normal sunlight, thereby emulating the un-shaded alpine habitat). The other was 109 µmol m−2 s−1 (low PAR, emulating the shaded, lower-elevation grassland habitat). The R/FR ratio was kept constant at 1.22 for both PAR treatments.

Assessment of plant growth

The stem elongation of each ramet was measured from the bottom of the first internode to the shoot tip. The leaf area of the pair of leaves below the third internode from the top was measured with a ΔT area meter (Delta T Services, Cambridge, UK). The stem growth, leaf length and leaf area were assessed in three experimental trials (10 ramets each) for each of light irradiance (PAR) and light quality (R/FR ratio) treatments, and trends were similar for each trial. Thus, individual measurements from 30 ramets were used to calculate the mean and its SE. In a similar fashion, individual measurements from 30 ramets were used to calculate the mean and its SE for measurements of ethylene evolution. All data were analysed by analysis of variance (anova), and the significance of differences between mean values was tested using Tukey's test on Statistical Package for the Social Sciences (SPSS) software, version 11.5 (SPSS, Chicago, IL, USA).

Measurement of ethylene evolution

Ethylene evolution by the top portion of the shoot (three to four actively growing internodes and associated leaves) was measured by incubating c. 0.1 g fresh weight (FW) in a 5 mL syringe (1.5 mL volume) for 15 min. Then 1 mL of gas was collected and injected into a Photovac 10Splus gas chromatograph (GC) (Photovac, Markham, Ontario, Canada) with a photo-ionization detector and a 40/60 Carbopack B column (Supelco Canada, Oakville, Ontario, Canada), as described by Emery et al. (1994b).

RESULTS

Stem elongation

Figure 1b shows that the cumulative shoot elongation growth of the prairie 7B genotype was significantly greater when the plants were grown under a low R/FR ratio of 0.7, than under the near-normal R/FR ratio of 1.9, or the abnormally high R/FR ratio of 2.7. As noted earlier, the PAR was maintained at a fixed, low irradiance of 115 µmol m−2 s−1 for all three R/FR ratio treatments. When the irradiance level was made the variable, a similar result (increased elongation growth) was obtained for the prairie 7B genotype grown under low PAR, relative to normal PAR (Fig. 2b). Here, the R/FR ratio was maintained at a near normal, 1.22 (sunlight).

Figure 1.

Cumulative stem elongation of the alpine 1D (a) and prairie 7B (b) genotypes of Stellaria longipes after transfer from short-day, cold (SDC) to long-day, warm (LDW) chambers having varied low red (R)/far red (FR) ratios (●, low R/FR; ○, normal R/FR and ▾, high R/FR ratio) and a fixed (low) photosynthetically active radiation (PAR) of 115 µmol m−2 s−1. The error bars indicate one SE of the mean. Using Tukey's multiple comparison test, the stem length for both Stellaria longipes genotypes (a & b) grown under the high R/FR ratio differed significantly (P = 0.05) from both of the other R/FR ratio treatments from days 12 to 24. Differences in stem length between the normal and low R/FR ratio treatment plants were NS.

Figure 2.

Cumulative stem elongation of the alpine 1D (a) and prairie 7B (b) genotypes of Stellaria longipes after transfer from short-day, cold (SDC) to long-day, warm (LDW) chambers having varied photosynthetically active radiation (PAR) (●, low PAR; ○, normal PAR) and a fixed (near-normal) low red (R)/far red (FR) ratio of 1.22. The error bars indicate one SE of the mean. Using Tukey's multiple comparison test, the stem length for plants of the alpine 1D genotype (a) grown under low PAR differed significantly (P = 0.05) by day 10 from the stem length of plants grown under the near-normal PAR. For the prairie 7B genotype (b), significance in stem length was gained by day 8.

The alpine 1D genotype also exhibited significant stem elongation (etiolation) in response to low PAR (Fig. 2a). Surprisingly, the alpine 1D genotype showed significantly greater cumulative shoot elongation under the high R/FR ratio of 2.7, than under either of the lower R/FR ratios (1.9 or 0.7) (Fig. 1a).

The alpine 1D genotype showed a relatively constant and rapid rate of stem elongation growth until day 21 in response to all three R/FR ratio treatments (Fig. 1a). By contrast, the prairie 7B genotype maintained a rapid and constant growth up to day 21 only under a low R/FR ratio (FR enrichment). Under the near-normal and abnormally high R/FR ratios, the growth rate of 7B slowed down after day 15 (Fig. 1b).

In response to varied PAR treatments, the alpine 1D genotype maintained a rapid and constant rate of shoot elongation until day 16, after which the growth rate decreased slowly (low PAR) or quickly (normal PAR) (Fig. 2a). The prairie 7B genotype maintained its rapid and constant rate of growth until day 20 under low PAR, after which its rate of growth slowed somewhat (Fig. 2b). Under high PAR, the rate of elongation of the prairie 7B genotype was constant until day 24, after which it essentially ceased. The proportional elongation responses to low PAR irradiance (relative to normal PAR) were similar (i.e. 1.9-fold) for both alpine and prairie ecotypes at day 28 (compare Fig. 2a with Fig. 2b).

Leaf length and area

For the alpine 1D genotype grown under a low PAR, the leaf length assessed at day 21 was significantly shorter as R enrichment increased (Table 1). However, this inhibitory effect of R enrichment on leaf length was not translated into a significant reduction for leaf area (Table 1) for the alpine 1D genotype. In a similar manner, the prairie 7B genotype, when grown at a low PAR, also had significantly shorter leaves with increased R enrichment (Table 1). However, the prairie 7B genotype did show an appreciable and significant reduction in leaf area as the R/FR ratio was increased from 0.7 (23.6 mm2) to 1.9 or 2.7 (c. 10 mm2) (Table 1).

Table 1.  Leaf length and area of the alpine 1D and prairie 7B genotypes of Stellaria longipes at day 21 after transfer from short-day, cold (SDC) to long-day, warm (LDW) chambers having either varied low red (R)/far red (FR) ratios with a fixed (low) photosynthetically active radiation (PAR) of 115 µmol m−2 s−1, or varied PAR with a fixed (normal) R/FR ratio of 1.22
Leaf length/area (mm)Low PARNormal R/FR
Low R/FR (0.7)Normal R/FR (1.9)High R/FR (2.7)Low PAR (109 µmol m−2 s−1)Normal PAR (611 µmol m−2 s−1)
  1. Mean values with the same superscript letter do not differ significantly (P = 0.05) based on Tukey's multiple comparison test.

1D length16.23b15.11c14d10.6c9.5c
7B length25.42a16.3b13.6d26.43a19.03b
1D area15.1b13.4b13.8b21.53b20.67b
7B area23.6a9.3c10c37.53a16.03c

The light irradiance (PAR) effects on leaf length and area were negligible for the alpine 1D genotype plants grown under a near-normal R/FR ratio (Table 1). By contrast, the prairie 7B genotype plants showed appreciable and significant reductions in both leaf length and area as the PAR was increased from 109 to 611 µmol m−2 s−1 (Table 1). Thus, only the shade ecotype of S. longipes showed leaf expansion with low light irradiance.

Ethylene evolution: effects of varied R/FR ratios under a low level of light irradiance

The ethylene evolution for the first 14 d under varying R/FR ratios (and a fixed, low PAR) for the alpine 1D genotype was relatively constant at 2260–2536 nL g FW−1 h−1, and the trend for an increased R enrichment to reduce ethylene evolution was NS (Table 2). The ethylene evolution then decreased to 1428 nL g FW−1 h−1 by day 21. At this point in time, the shoot tissue of the alpine 1D genotype taken from plants grown under the low R/FR ratio treatment showed a modest (but significant) increase in ethylene production, relative to treatments with higher levels of R enrichment (Table 2).

Table 2.  Ethylene evolution by the alpine 1D and prairie 7B genotypes of Stellaria longipes after transfer from short-day, cold (SDC) to long-day, warm (LDW) chambers having varied low red (R)/far red (FR) ratios and a fixed (low) photosynthetically active radiation (PAR) of 115 µmol m−2 s−1
Ethylene evolution (nL g FW−1 h−1)Day 7Day 14Day 21
  1. Mean values with the same superscript letter do not differ significantly (P = 0.05) based on Tukey's multiple comparison test.

  2. FW, fresh weight.

1D, low R/FR2261ab2539.84a1429.04c
7B, low R/FR1382.44b1541.12ab521.76d
1D, normal R/FR2178.56ab2137.32ab1065.96d
7B, normal R/FR1520.48ab1473ab580.96d
1D, high R/FR1957.12b2078.12ab1121.56d
7B, high R/FR1670.2ab1796.64a957.48c

By contrast, the prairie 7B genotype showed trends for ethylene evolution in response to R enrichment that tended to be just the opposite of that seen for the alpine 1D genotype (Table 2), although they did not gain significance (P = 0.05) until day 21 (at which time the tissue from the plants grown under the abnormally high R/FR ratio of 2.7 had significantly increased ethylene production) (Table 2).

At all times measured, days 7, 14 or 21, the alpine 1D genotype showed significantly increased levels of ethylene production (across all three R/FR treatments and also across both levels of PAR irradiance), relative to the low-elevation prairie 7B genotype (data not shown).

Ethylene evolution: effects of varied PAR under a near-normal R/FR ratio of 1.22

The ethylene evolution of shoot tissue from plants of the alpine 1D genotype grown under a normal PAR (611 µmol m−2 s−1) and a normal R/FR ratio (1.22) increased significantly between days 7, 14 and 21 (Table 3). Under a low PAR, however, the situation was more ‘confusing’ (Table 3). An increased PAR was a significant variable (it increased ethylene evolution) at both days 14 and 21 for the alpine 1D genotype.

Table 3.  Ethylene evolution by the alpine 1D and prairie 7B genotypes of Stellaria longipes after transfer from short-day, cold (SDC) to long-day, warm (LDW) chambers having varied photosynthetically active radiation (PAR) and a fixed (normal) low red (R)/far red (FR) ratio of 1.22
Ethylene evolution (nL g FW−1 h−1)Day 7Day 14Day 21
  1. Mean values with the same superscript letter do not differ significantly (P = 0.05) based on Tukey's multiple comparison test.

  2. FW, fresh weight.

1D, low PAR2507.96c1942.76d2500.4c
7B, low PAR1276.64ab1291.88ab1442.52a
1D, normal PAR2227.84cd3379.88b3794.08a
7B, normal PAR1169.96ab1375.28a925.2b

By contrast, for the prairie 7B genotype grown under a low or a normal PAR, there was no effect of time (ramet age) or light irradiance level (Table 3). Only at day 21 was the irradiance level a significant variable (here, growing the plants under a normal PAR reduced the ethylene evolution for the prairie 7B genotype) (Table 3).

DISCUSSION

As expected of a shade plant that exhibits the shade-avoidance syndrome (Smith 1982, 2000; Alokam et al. 2002; Franklin & Whitelam 2005), the prairie 7B genotype of S. longipes showed an increased stem length growth (etiolation) in response to both a reduction in the R/FR ratio (Fig. 1b) and a reduced PAR irradiance (Fig. 2b). By contrast, under these same R/FR and PAR conditions, the alpine 1D genotype (a sun plant) showed, quite surprisingly, just the opposite stem-elongation response to a reduction in the R/FR ratio (Fig. 1a), although it still responded by increased stem growth with a reduced PAR irradiation (Fig. 2a). Our results with the alpine 1D genotype (increased growth as R irradiation is increased to yield higher R/FR ratios) are thus very different from those obtained by Alokam et al. (2002) where the alpine 1D responded in the same manner as the prairie 7B genotype (e.g. both genotypes de-etiolated as the R/FR ratio was increased).

While we cannot cogently explain this rather anomalous behaviour for the alpine 1D genotype, we do know that its response is real (replicated in three different growth trials), and that its response of increased growth was obtained under the same light irradiation regimes that gave de-etiolation (reduced stem elongation) of the prairie 7B genotype as the R/FR ratio was increased.

Here, we should note that Alokam et al. (2002) varied their R/FR ratio by increasing FR (via incandescent irradiation filtered through an FRF 700 filter [West Lake Plastics, Lenni Mills, PA, USA) as the R/FR ratio was reduced to 0.7. By contrast, we utilized the LED sources, and we reduced the R irradiance (while maintaining the FR irradiance at a constant level), while keeping the PAR constant with a supplemental white light.

Hence, it is possible that differences in methodology (with regard to establishing the differing R/FR ratios) have yielded (for the alpine 1D genotype) what would seem to be an ‘anomalous’ behaviour (i.e. increased elongation as the R/FR ratio is increased).

Furthermore, when one moves to the growth parameters of leaf length and total leaf area (Table 1) in response to reduced R/FR ratio and to low PAR treatments, the increases seen for the prairie 7B genotype were not observed for the alpine 1D genotype in response to any light treatment. This is yet another example of the reduced ‘plasticity’ in the ability of the alpine ecotype plants (as evidenced by the 1D genotype) to respond in a photo-morphogenic manner to changes in intensity of PAR irradiance or to changes (reduced R/FR ratio) in light quality.

How can these unexpected effects for the alpine 1D genotype in response to modifying R/FR ratio be explained? Firstly, the alpine ecotype (population) of S. longipes may have adapted to the ‘stresses’ of the higher PAR environment (which would include a higher R irradiance). That is, this population of genotypes may no longer be able to respond to ‘shade’ light, where the R irradiance is reduced, relative to a near-constant level of FR irradiance (i.e. the reduced R/FR ratio of shade light). Rather, they are responding only to increasing doses of R (or total PAR) in a photo-morphogenic manner.

By contrast, the prairie ecotype (population) is fully adapted and highly responsive to reduced R irradiance (a lowered R/FR ratio) as a photo-morphogenic factor. Speculatively, it is also possible that the alpine ecotype plants are able to utilize the increased R irradiance in our high R/FR ratio treatment more effectively for photosynthesis, relative to the prairie ecotype plants.

The ethylene evolution of the alpine 1D genotype is almost double that of the prairie 7B genotype irrespective of light treatment (compare Table 2 with Table 3). However, the increase in ethylene evolution in response to light enriched with FR that was shown for tobacco plants (Pierik et al. 2004) was not detected in our study. That said, there is a trend (although NS) of increased ethylene evolution in response to FR enrichment for the prairie 7B genotype. By contrast, for the alpine 1D genotype, there is a reversed trend (NS) of increased ethylene evolution in response to enrichment with R.

For entire Arabidopsis (rosette leaf stage) plants, a low PAR irradiance, the other signal that induces the shade-avoidance syndrome, gave an increased ethylene evolution (Vandenbussche et al. 2003). However, in S. longipes, we found no change in ethylene evolution during the most rapid period of growth by the prairie 7B genotype in response to low PAR. Furthermore, there was a decrease in ethylene evolution by the alpine 1D genotype under low PAR, relative to normal PAR (Table 3, days 14 and 21), and this decreased ethylene evolution accompanied the increased shoot elongation seen under the reduced PAR.

Thus, while the shade prairie 7B genotype showed a good response to both FR enrichment and low irradiance in terms of increased stem elongation and leaf growth, this was not accompanied by an increased ethylene evolution, at least not during the most rapid growth period (days 7 and 14), as shown previously for sorghum (Finlayson et al. 1998, 1999), tobacco (Pierik et al. 2004) and Arabidopsis (Vandenbussche et al. 2003). Hence, the increase in stem elongation of the shade 7B genotype in response to both FR enrichment and low PAR appears not to be mediated by an increased ethylene production. Rather, an increased growth of S. longipes under these two treatments may be mediated by other hormones (i.e. gibberellins and auxin).

By contrast, the sun 1D genotype showed responses (growth and ethylene evolution) that were opposite to the shade 7B genotype under both FR enrichment and low irradiance treatments. Thus, for 1D genotype, FR enrichment tended (NS) to decrease ethylene evolution, while low PAR significantly reduced ethylene evolution. This light–ethylene interaction is not explained easily, as the increase in stem elongation observed in response to R enrichment appears not to be related directly to ethylene. By contrast, the increase in stem elongation induced by low irradiance for the alpine 1D genotype was associated with an appreciable decrease in ethylene evolution. This reduced ethylene evolution may thus act, along with changes in gibberellins and auxin, to mediate the increased stem elongation seen under low irradiance PAR.

To conclude, the sun (1D) and shade (7B) ecotypes of S. longipes respond quite differently (in terms of growth) to the key signals of the shade-avoidance syndrome, FR enrichment and low PAR. Significant reductions in ethylene evolution were correlated with increased stem elongation, but only for the alpine 1D genotype, and only in response to the low irradiance signal. Unlike sorghum (Finlayson et al. 1998, 1999), tobacco (Pierik et al. 2004) and Arabidopsis (Vandenbussche et al. 2003), it appears that the increased growth seen for both S. longipes genotypes under low PAR (Fig. 2a & b) is not accompanied by an increased ethylene evolution. However, the reduced shoot growth seen for the alpine 1D genotype under normal PAR (Fig. 2a) is accompanied by significant increases in ethylene evolution. Hence, the dwarf (shoot) phenotype of the alpine ecotype may be mediated by increased levels of ethylene production, even when the plants are grown in ‘temperate’-controlled environment conditions. A reasonable conclusion, then, is that the low-stature phenotype for plants in this alpine population is a product of selection for increased ethylene production under an environment that would have included very high winds (Emery et al. 1994b) and high levels of light irradiance.

ACKNOWLEDGMENTS

We would like to thank Dr Gillian M. Donald for her expertise in ethylene analysis, and Ms Bonnie Smith and Mr Ken Girard for their excellent glasshouse assistance. This work was funded by the Natural Sciences and Engineering Research Council (NSERC) (Canada) grants to D.M.R., R.P.P. and C.C.C.

Ancillary