By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
N. J. Livingston Fax: +250 721 7121; e-mail: firstname.lastname@example.org
A, assimilation rate a, fractionation against 13C for CO2 diffusion through air b, net fractionation against 13C during CO2 fixation Ca, ambient CO2 concentration Cc, CO2 concentration at the chloroplast Ci, intercellular CO2 concentration D, vapour pressure deficit En, needle transpiration rate Ep, whole plant water use gw, leaf internal transfer conductance to CO2 gs, stomatal conductance to water vapour L, projected leaf area NUE, nitrogen use efficiency PEP, phosphoenolpyruvate Rubisco, ribulose-1,5-biphosphate carboxylase TDR, time domain reflectometry WUE, water use efficiency Δ, carbon isotope discrimination δ13C, carbon isotope abundance parameter δ13Ca, carbon isotopic composition of atmospheric CO2 θ, volumetric soil water content
The effect of nitrogen stress on needle δ13C, water-use efficiency (WUE) and biomass production in irrigated and dry land white spruce (Picea glauca (Moench) Voss) seedlings was investigated. Sixteen hundred seedlings, representing 10 controlled crosses, were planted in the field in individual buried sand-filled cylinders. Two nitrogen treatments were imposed, nitrogen stressed and fertilized. The ranking of δ13C of the crosses was maintained across all combinations of water and nitrogen treatments and there was not a significant genetic versus environmental interaction. The positive relationships between needle δ13C, WUE and dry matter production demonstrate that it should be possible to use δ13C as a surrogate for WUE, and to select for increased WUE without compromising yield, even in nitrogen deficient environments. Nitrogen stressed seedlings had the lowest needle δ13C in both irrigated and dry land conditions. There was a positive correlation between needle nitrogen content and δ13C that was likely associated with increased photosynthetic capacity. There was some indication that decreased nitrogen supply led to increased stomatal conductance and hence lower WUE. There was a negative correlation between intrinsic water use efficiency and photosynthetic nitrogen use efficiency (NUE). This suggests that white spruce seedlings have the ability to maximize NUE when water becomes limited. There was significant genetic variation in NUE that was maintained across treatments. Our results suggest that in white spruce, there is no detectable effect of anaplerotic carbon fixation and that it is more appropriate to use a value of 29‰ (‘Rubisco only’) for the net discrimination against 13C during CO2 fixation. This leads to excellent correspondence between values of Ci/Ca derived from gas exchange measurements or from δ13C.
Measurements of the relative abundance of the stable isotopes 13C and 12C (δ13C) in plant organic tissue can provide useful long- or short-term indicators of water use efficiency (WUE) and metabolism, and have the potential to be used as a powerful tool for the selection of genotypes with improved WUE and productivity. Leaf carbon isotopic composition is a function of both the supply of CO2 to the sites of carbon fixation and photosynthetic capacity (chloroplast demand for CO2). These relations are formalized for C3 plants in the model developed by Farquhar et al. (1982) that links photosynthesis to δ13C through Ci/Ca, the ratio of intercellular to atmospheric CO2 partial pressure.
In a previous paper (Sun et al. 1996), we reported that in white spruce (Picea glauca (Moench) Voss) there was highly significant genetic variation (1·6–2·0‰) in needle δ13C and that there was not a significant genotype by environment interaction (G×E) for well fertilized seedlings grown under irrigated or dry land conditions. Further, we established that there was a positive correlation between δ13C and the ratio of the maximum assimilation rate to stomatal conductance (gs), and long-term WUE (calculated as the ratio of total plant dry mass production to water used, and determined over the growing season), and between δ13C and productivity, suggesting a correlation resulting from variation in photosynthetic capacity. We concluded that it should be possible to use δ13C as a surrogate for WUE and to select white spruce genotypes for increased WUE without compromising yield.
In this paper, we extend our discussion to the effects of nitrogen treatments on white spruce δ13C. An understanding of such interactions would be critical to any selection programme. Whilst the influence of nitrogen metabolism on the stable carbon isotopic composition of plant tissue has been discussed in detail by Raven & Farquhar (1990), very few field studies have been undertaken, particularly in conifers, to determine the effects and interactions of multiple stresses (water and nitrogen). Recently, however, Patterson et al. (1997), reported that in black and white spruce, there were not discernible trade-offs between nitrogen-use and water-use efficiency between species. Within a given species, however, there was considerable plasticity in the response to simultaneous limitations of nitrogen and water, and both species were forced to utilize each resource with less than optimal efficiency.
The objectives of this study were to determine: (i) whether the relations between δ13C, WUE and productivity for irrigated and dry land seedlings would hold even under varying nitrogen regimes; and (ii) whether there is genetic variation in nitrogen use efficiency that would be maintained across treatments. Additionally, we tested the hypotheses that: (i) if nitrogen stress has a profound influence on photosynthetic capacity by affecting the amount and activity of enzymes critical to photosynthesis, this should lead to reduced WUE and increased Ci and should be reflected in decreased (more negative) δ13C values; (ii) decreased δ13C might reflect increased Ci brought about by increased gs or a combination of increased (or unchanged) gs and reduced photosynthetic capacity. In any event, the effects of drought on δ13C, brought about by reduced gs, should be offset by reduced nitrogen supply.
MATERIALS AND METHODS
Plant material and growing conditions
The field site and plant material are described in detail in Sun et al. (1996). Briefly, 1600 one-year-old white spruce seedlings, representing 10 different crosses, were planted at the University of Victoria in April 1992. The crosses were chosen to provide a sufficiently wide range of material to establish genetic differences (Table 1). The females for six of these crosses were grafted scions of plus trees found within the Prince George selection unit, collectively covering an area from Aleza Lake in the north to Gravelle Ferry in the south (a distance of 140 km). With the exception of two individuals from the East Kootenays, pollen trees used in each of these crosses were also from the Prince George selection unit. The females for two additional crosses were from the Prince George selection unit, but one has been crossed with a specimen from eastern North America. The final two crosses were those where both parents were from outside the Prince George selection unit.
Table 1. . List of accessions and their origin, where PG = Prince George selection unit, EK = East Kootenays selection unit and ENA = Eastern North America selection unit
Seedlings were planted, ≈ 1 m apart, in buried sand-filled open-ended galvanized steel cylinders (0·15 m in diameter and 0·9 m long). The cylinder top was covered with a thin layer (0·02 m) of gravel to reduce surface evaporation, and the walls were lined with plastic. Over the two years of the study, roots did not emerge through the bottom of the cylinder (Sun et al. 1996).
Treatments were laid out in a modified split-split plot design with two blocks. Block 1 (Fertilized) had four main plots and Block 2 (Nitrogen stressed) two main plots. Liquid fertiliser (NH4NO3 at 400 g m–3 for Block 1 and at 100 g m–3 for Block 2; KH2PO4 at 113 g m–3; K2SO4 at 67 g m–3; MgSO4·7H2O at 116 g m–3; CaCl2·2H2O at 100 g m–3 and micronutrients) was applied at regular intervals throughout each growing season to both blocks. Seedlings from the fertilized treatment were also used in a larger study (Sun et al. 1996) to investigate the relationship between WUE and stable carbon isotope composition. The main plots in both blocks were irrigation and dryland, and subplot treatments were the crosses.
Details of the water treatments are described in Sun et al. (1996). These treatments were imposed from 15 July and 19 May until the end of September in the first and second year, respectively. Irrigation was carried out with reference to time domain reflectometry (TDR) measurements of soil water (see below). Irrigated seedlings each received ≈ 0·4 dm3 of water per week (applied by hand) in the first year of the study and 0·8 dm3 of water the next year. Dry land seedlings received approximately half that amount. In the first year the average soil volumetric water content (θ) of cylinders was maintained at 0·060 and 0·083 m3 m–3 for the dry land and irrigated treatments, respectively. In the second year θ was maintained at 0·028 and 0·043 m3 m–3 for the same treatments. Volumetric water contents were lower in the second year, despite the higher rates of water application, because of the higher water use by the much larger plants.
An automated climate station installed at the site provided hourly averages or totals of standard meteorological variables. Victoria, located on the south-east coast of Vancouver Island (48 °28′N, 123 °18′W), is in a rain shadow and has warm dry summers. Total rainfall between the beginning of May and end of September was 117·2 mm in 1992 and 108·8 mm in 1993. In the latter year, only 32·3 mm fell between June and August. The highest mean monthly temperatures were recorded in July in both years (16·9 °C). Day time vapour pressure deficits never exceeded 3·0 kPa and typically were less than 2 kPa. The average growing season vapour pressure deficit was 0·91 kPa in 1992 and 0·93 kPa in 1993.
Soil water was measured every 10–15 d throughout each growing season using TDR. One single diode probe (Hook et al. 1992; Sun et al. 1996) was placed in each cylinder. Measurements of time delay were made with a cable tester (1502C, Tektronix Corp. Corvalis, OR) and converted to estimates of θ using the linear model of Hook & Livingston (1996). Measurement resolution and accuracy were 0·002 and 0·006 m3 m–3, respectively (Hook & Livingston 1995, 1996) so that the smallest detectable change in cylinder water was 32 cm3 (less than one day of water use by an irrigated seedling).
Seedling water use was calculated from the water balance of the cylinder with the appropriate corrections for rainfall, drainage and soil evaporation (Sun et al. 1996). Water use was calculated for 22–25 seedlings of every cross for each water and nitrogen treatment.
A portable (open) gas exchange system (LCA-3, Analytical Development Co. Ltd, Hoddesdon, UK) was used to measure needle transpiration rate (En), gs and assimilation rate (A). In the first year, measurements were made on July 22 and 23, August 27 and 28, and October 5 and 6 on five replicate individuals of each cross and treatment. The following year measurements were made on June 25 and 26, August 19 and 20, and on October 8 and 9 on 12 replicates of fertilized seedlings (for both irrigated and dryland treatments) and five to six replicates of nitrogen stressed seedlings. In both years, measurements were taken on clear days between 10 : 00 and 15 : 30 h. Within a given year, measurements were always taken on the same shoots with the exception of those made in June of the second year when shoots were removed for leaf area determination. Because of the marked sensitivity of gs to vapour pressure deficit (D), and Ci to the CO2 concentration, considerable care was taken to ensure that conditions in the gas exchange chamber were not changed between measurements. At the end of each growing season, shoots were harvested and their projected needle area determined with a leaf area meter (LI-3100, Li-Cor Inc., Lincoln, NE). As shoot growth was almost complete by mid-July (Sun 1995), errors introduced by using a single value of needle area for a given shoot were negligible.
Dry mass production and water use efficiency
Dry mass production was estimated as the difference in total dry mass (roots and shoots) between the beginning and end of the growing season. In the first year, dry mass was determined for five replicate individuals for each cross and treatment before planting, and 10 replicates in early October. The following year, eight replicates were harvested in early May and a further 25 in October. After harvesting, the stem diameter, stem height and root length were measured. Roots and shoots were then separated and oven dried at 70 °C for 48 h for the determination of dry mass and root to shoot dry mass ratio. Water use efficiency was calculated as the ratio of dry mass production to total water use for the growing season.
Carbon isotopic composition and leaf nitrogen
Current year needles were used for all determinations of δ13C (Sun et al. 1996). After oven drying, samples were ground in a Wiley Mill (Fisher Scientific, Napeon, Ontario, Canada) to pass a size 40 mesh and pulverized to submicron particle size using a stainless-steel planetary ball mill (Pulverisette, Fritsch GMBH, Germany). Sub-samples (1–2 mg) were analysed for δ13C on an isotope ratio mass spectrometer (VG-SIRA 12, Isotech, Middlewich, UK) with an elemental analyzer (Model 1106, Carlo Erba, Valencia, CA). Ten replicate individuals per cross for each treatment were analysed in the first year. In the second year, the number of replicate individuals was increased to 25.
Discrimination was calculated as:
where δ13Ca is the isotopic composition of the ambient air (assumed to be – 8‰).
where a is the discrimination during diffusion in air (4·4‰) and b is the net discrimination during carboxylation (27‰).
The leaf nitrogen concentration of current and one-year old needles was determined for a minimum of six replicate individuals per cross for each treatment. Needles were oven dried to 70 °C and ground to a fine powder. Nitrogen concentration was measured using a modification of the method of Parkinson & Allen (1975), followed by colorimetric (autoanalyser) analysis.
The ranking of needle δ13C for the 10 white spruce crosses was maintained across all combinations of water and nitrogen treatments in both years (Fig. 1, only second year data shown). All correlations are highly significant (P < 0·001). For example, the coefficient of variation (r2) for the regression (year 2) of δ13C of irrigated and fertilized seedlings against those of dry land, nonfertilized seedlings is 0·79. Crosses 1 and 4 (both parents from the Prince George selection unit) had the highest (most positive) δ13C, while crosses 9 and 6 (with at least one parent outside the Prince George selection unit) consistently had the lowest δ13C. Differences in δ13C among the remaining crosses were generally very small and were not statistically significant.
Differences in mean values of δ13C between treatments were significantly more pronounced in the second year than in the first year of the study. Because isotopic values of samples harvested in the first year might have partially reflected nursery and cold room conditions in the previous year (Sun et al. 1996), and because of the larger sample sizes and earlier imposition of the water treatments in the second year, it is likely that data for the second year are more reliable than those of the first.
Two-way analysis of variance (Table 2) revealed that there was not a significant genotype by environment interaction but that there was significant genetic variation in δ13C. In the irrigated treatment, in all but one cross, δ13C of nitrogen stressed seedlings were lower than those of the fertilized seedlings. However, there was not a significant difference in mean δ13C for all crosses between fertiliser treatments (Table 3). In the dry land treatment, six of the 10 unfertilized crosses had lower δ13C than those that were fertilized but, again, mean δ13C for all crosses did not significantly differ. Values of δ13C for irrigated seedlings were always lower than that of dry land seedlings regardless of the nitrogen treatment.
Table 2. . Three way analysis of variance for the second year of the study. The water treatments were irrigated and dry land and the nitrogen treatments were fertilized and nitrogen stressed
Table 3. . Carbon isotopic composition (δ13C,‰), water use efficiency (WUE, mg g–1), nitrogen use efficiency (NUE, g g–1), stomatal conductance (gs, mmol m–2 s–1) and the ratio of whole plant seasonal water use (Ep) to leaf area at harvest (L) (mmol m–2 s–1) of 2 year old white spruce seedlings. Seedlings were irrigated and fertilized (+ N), irrigated and nitrogen stressed (– N), droughted and fertilized and droughted and nitrogen stressed. Values of gs were determined from gas exchange measurements. Values of Ep were determined by water balance. Each number represents the mean value of 22–25 (δ13C, WUE, Ep/L) or 12 (gs) replicates of each of the 10 crosses given in Table 1. Standard deviations are given in parentheses. Mean values within each row followed by a different letter are significantly different at P = 0·05
Under irrigated conditions, nitrogen stressed seedlings generally had lower assimilation rates than fertilized seedlings (Fig. 2). Again, while gs (determined from gas exchange measurements) of nitrogen stressed seedlings was higher than that of seedlings that were fertilized, means calculated for all crosses were not significantly different in either irrigated or dry land treatments (Table 3). However, for irrigated seedlings, the ratio of seasonal water use (as determined from water balance measurements) to leaf area (at final harvest) was significantly higher for unfertilized seedlings than for those that were nitrogen stressed (Table 3).
Whilst there was good correspondence (r2 = 0·63) between values of Ci/Ca derived from either gas exchange measurements or from measurements of δ13C and Eqns 1 and 2, values of Ci/Ca derived from gas exchange measurements were consistently lower than those obtained from δ13C. The slope and intercept of the regression line are significantly different from the 1 : 1 line (Fig. 3).
Under irrigated conditions, root to shoot dry mass ratios were higher in fertilized seedlings than those that were nitrogen stressed (Fig. 4). However, in contrast, under dry land conditions, root to shoot dry mass ratios were generally lowest in the fertilized seedlings.
There was a positive relation between needle nitrogen concentration (on a unit leaf dry mass basis) and δ13C (y = – 29·7 + 6·08x, r2 = 0·42, data not shown). The coefficient of variation was higher for the dry land than the irrigated seedlings but both were highly significant (P < 0·002). The average leaf N concentration was higher in dry land than in irrigated seedlings. This might have been the result of fundamental alterations in N metabolism (Belesky et al. 1982) but more likely, was simply a dilution effect as dry land seedlings were much smaller than irrigated seedlings (Sun et al. 1996).
There were positive relationships between WUE and δ13C (Fig. 5), between WUE and dry mass production (Fig. 6a) and between δ13C and dry mass production (Fig. 6b). The relation between water use efficiency and dry mass production, while sensitive to water stress, was not sensitive to nitrogen stress. Conversely, the genetic relationship between δ13C and WUE (Fig. 5) and between δ13C and dry mass production (Fig. 6b) appeared to be sensitive to both stresses.
Generally, the ranking of crosses (in terms of WUE and dry mass gains) held across nitrogen treatments with crosses 9 and 6 having the lowest, and crosses 1 and 4 having the highest WUE and dry mass gains (consistent with their respective δ13C values). Crosses 7 and 10 performed well under nitrogen stressed conditions as shown by their high WUE and dry mass production. This might have been because of their relatively high leaf nitrogen content under these conditions (data not shown).
There were significant (P < 0·01) differences in photosynthetic nitrogen use efficiency determined either as the ratio of photosynthesis rate to leaf nitrogen content or, following Patterson et al. (1997), the ratio of total plant carbon to nitrogen (Fig. 7), between crosses. However, in both cases there was not a significant G versus E interaction. Mean values of NUE (derived from C : N ratios) for all crosses were generally highest in nonfertilized treatments in both irrigated and dry land seedlings, but differences were not statistically significant (Table 3). However, in the irrigated treatments the mean WUE of fertilized seedlings was significantly higher than that of seedlings that were nitrogen stressed (Table 3).
Our results show that even in nitrogen deficient environments under well watered or dry land conditions, the genetic variation in white spruce needle δ13C is maintained. Variation in δ13C between individuals within a given genotype and exposed to the same environment was relatively small (Fig. 1), so that the superior or poorest genotypes could be identified. Our results are consistent with those of other studies on a wide range of species, which have shown little genotype (or accession) by environment interaction (G×E) for δ13C (Condon et al. 1987; Hubick et al. 1988; Guy & Holowachuk 1993; Zhang & Marshall 1994). Condon & Richards (1992) reported that G×E accounted for only a small fraction (5%) of the total genetic variation in wheat. However, they predicted that in more contrasting environments, G×E might increase. Our results and those of our previous study (Sun et al. 1996) do not support that prediction.
The positive relation between dry mass production and δ13C and between dry mass production and long-term WUE (as determined by water balance) suggest that the best performing crosses (1 and 4) had increased WUE because of higher inherent photosynthetic capacity (and that high WUE must have been primarily due to high dry mass production). The gas exchange data provide further support in that the larger crosses typically had the highest photosynthetic rates (Fig. 2). Also, there was a positive relation between intrinsic water use efficiency (defined as the ratio of assimilation rate to stomatal conductance, Osmond, Bjorkman & Anderson 1980; Ehleringer, Hall & Farquhar 1993) and both WUE (r2 = 0·41, P < 0·01, data not shown) and δ13C (r2 = 0·39, P < 0·01). Intrinsic water use efficiency reflects the compromise between photosynthesis and transpiration at the time of measurement while WUE reflects the compromise between biomass gain and water loss over the growing period. The correspondence between intrinsic and long-term WUE suggests that the response to drought or changed nitrogen supply was primarily physiological rather than morphological.
There was no evidence that the more productive crosses allocated a greater fraction of dry mass production to the shoots. Indeed, cross 9, one of the least productive crosses, consistently had the lowest root to shoot ratio. Overall there was a positive correspondence between dry matter production and root to shoot dry mass ratio (r2 = 0·57, P = 0·001, data not shown) and no correspondence, across all treatments, between the root to shoot dry mass ratio and WUE. There was a slight tendency for δ13C to increase with decreasing root to shoot ratio (r2 = 0·09 P = 0·01, data not shown), consistent with the data of Poorter & Farquhar (1994) who reported a negative correlation between fractional biomass allocation to roots and δ13C over a range of herbaceous species differing in relative growth rate.
Our data suggest that decreased nitrogen supply, at least under irrigated conditions, led to increased stomatal conductance and that this contributed, in part, to the significantly lower Ep/L and WUE (Table 3) in these treatments. While not statistically significant, the general trend of lower δ13C (and higher gs as measured by gas exchange) in nitrogen stressed than in fertilized seedlings (Fig. 1, Table 3) suggest that there was increased discrimination against 13C in nitrogen stressed seedlings because of their relatively high average Ci.
There are alternative, but less probable, explanations for the lower δ13C in the nitrogen stressed seedlings. For example, phosphoenolpyruvate (PEP) carboxylase operates in parallel with Rubisco in the provision of carbon skeletons for N assimilation (Turpin et al. 1991). Discrimination by PEP carboxylase is only 5·7‰ with respect to air CO2 (Farquhar et al. 1989). Where N assimilation is restricted to the roots, as in gymnosperms (Smirnoff & Stewart 1985), the carbon fixed by PEP carboxylase is largely of respiratory origin and has an isotopic composition approximating that of the plant. Taking Raven & Farquhar's (1990) estimate of 0·346 mol C fixed by non-Rubisco carboxylases per mol N assimilated, we calculate that anaplerotic C fixation would lower δ13C values by a mere – 0·03 to – 0·09‰ across the full range of leaf C : N ratios observed in our seedlings. This contribution is clearly inconsequential.
Roots and stems tend to be slightly enriched in 13C relative to leaves (O’Leary 1981; Brown et al. 1995). If nitrogen stress results in the allocation of more dry matter to roots than shoots, then mass balance considerations may cause a lower δ13C in leaves. Our results do show an increase in the root to shoot ratio for nitrogen stressed seedlings under dryland conditions but a decrease for seedlings that were irrigated (Fig. 4).
It is possible that the trend of lower values of δ13C in nitrogen stressed plants resulted from changes in the internal conductance to CO2 diffusion from intercellular air spaces to the site of carbon fixation (internal transfer conductance, gw). The development of ‘on-line’ systems (Evans et al. 1986) in which carbon isotope discrimination is measured concurrently with gas exchange, has revealed that discrimination is primarily determined by the CO2 concentration within the chloroplast (Cc) rather than Ca (Gillon & Griffiths 1997; Flanagan et al. 1994). Thus, if nitrogen stress leads to an increase in gw (over and above any increase in gs), then there would be a corresponding increase in Cc and decrease in δ13C (Farquhar & Lloyd 1993; Harwood et al. 1998). Changes in gw with nitrogen supply might also account for the shift in the relationship between δ13C and dry mass production, and between δ13C and WUE with nitrogen stress.
While modelled Ci/Ca (from δ13C) and measured Ci/Ca (from gas exchange) were clearly related (Fig. 3), there was a significant offset between the respective values with the modelled values greater than expected. This is a similar result to that obtained by Zhang et al. (1993) for a 15-year-old Douglas-fir plantation. As outlined above and discussed by Raven & Farquhar (1990), the computed decrease in δ13C in a terrestrial C3 plant relative to a hypothetical ‘Rubisco only’ value of δ13C is a function of the nitrogen supply and the site of N assimilation. The smallest effect occurs when N is assimilated in the roots as NH4+. Recently, Kronzucker et al. (1997) demonstrated that white spruce assimilates all nitrogen in the roots, predominantly in the form of NH4+, regardless of the nitrogen supply. Thus it would seem appropriate to use a discrimination factor (b in Eqn 2) of 29‰ (i.e. the net discrimination of Rubisco) rather than 27‰ which accounts for anaplerotic reactions. As is shown in Fig. 3, when this higher value is used, the offset between the modelled and measured Ci/Ca is no longer statistically significant (P = 0·001).
Periodic measurements of δ13Ca during one month in the second year of the study, which showed that the isotopic composition of atmospheric CO2 was not significantly different from –8‰, confirmed that any anthropogenic effects on values of δ13C could be discounted. Further, any effects associated with changes in gw could not have been a factor, because Cc would have had to be greater than Ci to bring about increased discrimination, clearly an impossibility.
The negative correlation between WUE and nitrogen use efficiency (and between intrinsic water use efficiency and photosynthetic nitrogen use efficiency) is a similar result to that reported for California evergreen and American elm (Field et al. 1983 and Reich et al. 1989, respectively). Typically, under well watered conditions, A in C3 plants is not saturated with respect to CO2, and A increases with increasing gs (Wong et al. 1979). This does not require the investment of additional nitrogen in photosynthetic enzymes and therefore should give rise to a positive relationship between gs and photosynthetic nitrogen-use efficiency. However, any increase in gs will also lead to a decrease in intrinsic water-use efficiency because of a proportionately greater increase in E than A. Under well watered conditions, but limited nitrogen supply, there might be an increase in gs in order to raise Ci and nitrogen use efficiency (Patterson et al. 1997). Our data is generally consistent with such an argument in that under irrigated conditions there was a significant decrease in Ep/L (Table 3) with nitrogen stress. The decrease in gs, whilst apparent, was not statistically significant.
Under dry land conditions, plants need to expend more nitrogen to maintain carbon assimilation and growth. Additional nitrogen supply might also limit the severity of water stress (Bennett et al. 1986). In such a case, NUE will be lower but WUE higher. Thus the cost of a high WUE is low NUE and vice versa. Plants in wet environments might fully use nitrogen resources to maximize their assimilation gain per unit of leaf nitrogen, despite decreasing their WUE. Conversely, plants growing in dry land conditions might well under-utilize their nitrogen source in order to maximize their WUE.
The ranking of 10 white spruce crosses in terms of δ13C was maintained across a range of multiple stress (water plus nitrogen) treatments. There was a positive correlation between needle δ13C and WUE and dry matter production in both fertilized and nitrogen stressed treatments. Increased nitrogen supply generally led to increased δ13C, likely a result of increased A and reduced Ci. There was some indication that decreased nitrogen supply led to increased stomatal conductance and hence lower WUE. When water is limited, the trade off between WUE and NUE reflects the ability of these white spruce crosses to maximize the resource use efficiency.
Our results suggest that in white spruce, there is no discernable effect of anaplerotic carbon fixation and that it is more appropriate to use a value of 29‰ (‘Rubisco only’) for the net discrimination against 13C during CO2 fixation. This leads to excellent correspondence between values of Ci/Ca derived from gas exchange measurements or from δ13C.
Funding for this research was provided by the Science Council of British Columbia and by the Natural Science and Engineering Research Council of Canada. We thank Mr K. Kiss, B.C. Ministry of Forests, Kalamalka, for supplying the seeds used in this study. We also thank B.M. Binges and S. Lotz for their technical support and Dr M. Whiticar, T. Cederberg and F. Harvey-Kelly who operate the isotope mass spectrometry facilities at the University of Victoria. Finally, we thank Professor P. G. Jarvis and two anonymous reviewers for their constructive comments on the paper.