Stomatal density, anatomy and nutrient concentrations of Scots pine needles are affected by elevated CO2 and temperature

Authors


Elina Vapaavuori. Fax: +358 (0)10 211 4801; e-mail: Elina.Vapaavuori@metla.fi

ABSTRACT

Stomatal density, anatomy and nutrient concentrations of Scots pine (Pinus sylvestris L.) needles were studied during 3 years of growth at elevated CO2 (693 ± 30 µmol mol−1), at elevated temperature (ambient +2·8–6·2 °C depending on the time of the year) and in a combination of elevated CO2 and temperature in closed-top chambers. The treatments were started in August 1996. At elevated temperature, the needles that were grown in the first year (i.e. the 1997 cohort) were thinner, had thinner mesophyll in the abaxial side, thinner vascular cylinder and lower stomatal density than those grown at ambient temperature. The proportion of mesophyll area occupied by vascular cylinder or intercellular spaces were not changed. Lower stomatal density apparently did not lead to decreased use of water, as these needles had higher concentrations of less mobile nutrients (Ca, Mg, B, Zn and Mn), which could indicate increased total transpiration. In the 1997 and 1998 cohorts, elevation of temperature decreased concentrations of N, P, K, S and Cu. In the 1999 cohort, contradictory, higher concentrations of N and S at elevated temperature may be related to increased nutrient mineralization in the soil. Elevation of CO2 did not affect stomatal density, needle thickness, thickness of epidermis or hypodermis, vascular cylinder or intercellular spaces. Concentrations of N, P, S and Cu decreased at elevated CO2. Reductions were transient and most distinct in the 1997 cohort. The effects of CO2 and temperature were in some cases interactive, which meant that in the combined treatment stomatal density decreased less than at elevated temperature, and concentrations of nutrients decreased less than expected on the basis of separate treatments, whereas the thickness of the epidermis and hypodermis decreased more than in the separate treatments. In conclusion, alterations in the anatomy and stomatal density of Scots pine needles were more distinct at elevated temperature than at elevated CO2. Both elevated CO2 and temperature-induced changes in nutrient concentrations that partly corresponded to the biochemical and photosynthetic alterations in the same cohorts (Luomala et al. Plant, Cell and Environment26, 645–660, 2003) Reductions in nutrient concentrations and alterations in the anatomy were transient and more evident in the needle cohort that was grown in the first treatment year.

INTRODUCTION

Increasing atmospheric CO2 concentration and elevation of mean global temperature (IPCC 2001) are likely to affect plant growth and carbon balance of ecosystems by influencing directly or indirectly various metabolic processes (see review by Morison & Lawlor 1999). While biochemical acclimation at elevated CO2 has received intensive attention in recent decades, anatomical features of leaves in changing climate have been less studied. The anatomy of leaves is, however, highly flexible and is modified by environmental factors like irradiation (sun leaves/shade leaves, e.g. Lambers, Chapin & Pons 1998), nutrients (e.g. Jokela, Sarjala & Huttunen 1998), drought (e.g. Bosabalidis & Kofidis 2002) and ozone (e.g. Oksanen, Sober & Karnosky 2001; Oksanen et al. 2004). In connection with the studies of photosynthetic acclimation at elevated CO2, stomatal density is of special interest, because it sets the limit for maximal stomatal conductance for gas exchange and has thus a potential to affect carbon gain and water use efficiency of a plant (Beerling 1997). Stomatal frequency of fossil plant samples has been used to estimate atmospheric CO2 concentration in past environments (e.g. Retallack 2001; Royer et al. 2001). Correlation between stomata and CO2 concentration is, however, weaker in short-term experiments at elevated CO2, and the observations on plants grown at elevated CO2 have ranged from no changes (e.g. Reddy et al. 1998; Vanhatalo, Huttunen & Bäck 2001) to decreases (e.g. Ferris & Taylor 1994; Ferris et al. 1996, 2002; Lin, Jach & Ceulemans 2001; Tognetti et al. 2001) or increases (e.g. Ferris & Taylor 1994; Ferris et al. 1996, 2002; Visser et al. 1997) in stomatal density. In some woody shrubs, the sensitivity of stomatal response to elevated CO2 declines at concentrations exceeding about 350 µmol mol−1 (Woodward & Kelly 1995; Woodward & Bazzaz 1988). In Scots pine, however, stomatal density may be sensitive to CO2 concentrations that are expected to prevail in the near future (Beerling 1997; Lin et al. 2001). Studies exploring the interactive effects of elevated CO2 and temperature on stomatal numbers (Beerling & Chaloner 1993; Morgan et al. 1994; Ferris et al. 1996; Beerling 1997; Reddy et al. 1998; Apple et al. 2000) and particularly on the anatomy of leaves (Ferris et al. 1996) are limited. Responses of stomatal density or index to experimental or seasonal warming have been variable, showing no changes, increases or decreases in stomatal frequency (Beerling & Chaloner 1993; Morgan et al. 1994; Ferris et al. 1996; Beerling 1997; Reddy et al. 1998; Apple et al. 2000).

The anatomy of mesophyll may, in some cases, be even more responsive to changes in CO2 concentration than that of epiderm (Masle 2000). Elevated CO2 has been observed to stimulate cell division (Ferris & Taylor 1994; Kinsman et al.; 1996; Masle 2000; Ferris et al. 2001) and cell expansion (Ferris & Taylor 1994; Taylor et al. 1994; Masle 2000; Ferris et al. 2001), and to result in thicker leaves (Yin 2002) with higher numbers of cells or cell layers and/or larger cells (Radoglou & Jarvis 1992; Masle 2000). Alterations in relative volumes occupied by intercellular air spaces (Masle 2000; Oksanen et al. 2001), palisade and spongy mesophyll and vascular elements (Pritchard et al. 1997; Lin et al. 2001; Oksanen et al. 2001; Engloner et al. 2003) have also occurred at elevated CO2. Anatomical changes in the mesophyll and vascular elements are likely to affect gas exchange by altering resistance for CO2 diffusion, to influence water transport and to affect assimilate transport and thus the capacity to exploit extra carbon produced at elevated CO2.

The rate of cell division is tightly regulated by temperature and the number of cell divisions involved in the formation of a new leaf is drastically reduced in cold climates (Körner & Larcher 1988). Plants grown in cool climates have commonly thicker leaves, thicker epidermal cell walls and more numerous stomata than their counterparts grown in warmer climates (Körner & Larcher 1988; Loveys et al. 2002). Needle length in Scots pine is strongly dependent on the temperature of the current growing season (Junttila & Heide 1981; Junttila 1986). In Douglas fir, elevated temperature increased elongation rate of needles, but the net effect of temperature on needle length varied by year (Olszyk et al. 1998; Apple et al. 2000). On the basis of these observations, elevation of temperature could be expected to lead to formation of thinner and possibly longer needles with less stomata than at ambient temperature.

Altered growth rate and allocation of growth due to elevated CO2 and temperature also affects the use of mineral nutrients, and the use and retranslocation of mobile nutrients in particular (Conroy, Milham & Barlow 1992; Poorter et al. 1997; Curtis & Wang 1998; Medlyn et al. 1999; Roberntz & Linder 1999; Sigurdsson 2001; Sallas et al. 2003). Mineral nutrition is indirectly coupled with alterations in stomatal frequency and stomatal conductance since any changes in transpiration stream will also influence the availability of nutrients to the foliage. At elevated temperature, in turn, nutrient supply in the soil could be expected to increase, as soil respiration and mineralization of nutrients are strongly dependent on the temperature (Bonan & Van Cleve 1992). Thus, elevated temperature could compensate for a decrease in a nutrient concentration at elevated CO2. Accordingly, higher concentration of foliar N has previously been observed in Scots pine (Kellomäki & Wang 1997) and in Douglas-fir (Pseudotsuga menziesii) (Hobbie et al. 2001) grown at elevated temperature. Concentrations of nutrients are tightly linked with biochemical capasity for photosynthesis and growth, but may also regulate the anatomy of leaves (Jokela et al. 1998). The interaction of nutrients to anatomical features of conifer needles with a reference to biochemical acclimation at elevated CO2 and temperature has not been vastly studied.

During 3 years of growth at elevated CO2 and temperature, we measured stomatal density, anatomy and nutrient concentrations in needles of Scots pine trees. It was anticipated that at elevated CO2, needles would be thicker and have smaller stomatal density and nutrient concentrations than at ambient CO2, whereas elevated temperature could counteract changes in the thickness and nutrient concentrations of needles at elevated CO2. Our aim was to study whether possible alterations in the thickness of a needle are caused by alterations in a specific tissue of a needle (epidermis, hypodermis, mesophyll, vascular cylinder, intercellular spaces) or have all of these changed in parallel. It was further studied whether stomatal density and intercellular spaces exhibit changes that would have implications on gas exchange. Finally, we discuss whether possible anatomical changes at elevated CO2 and temperature are connected to alterations in nutrient concentrations of needles and whether alterations in nutrient concentrations are parallel to biochemical acclimation of the same cohorts (Luomala et al. 2003).

MATERIALS AND METHODS

Site and chambers

The experiment was conducted in a naturally regenerated Scots pine (Pinus sylvestris L.) stand close to Mekrijärvi Research Station (University of Joensuu) in eastern Finland (62°47′ N, 30°58′ E, 145 m a.s.l.). The site is characterized by cold winters with persistent snow cover and a short growing season. Mean annual precipitation (in 1961–94) is 740 mm, of which about 38% is received as snow. The monthly mean temperature is −10·4 °C in January and 15·5 °C in July, with minimum and maximum temperatures of about –40 °C and 32 °C recorded in January and July, respectively. The soil at the site is a podsolized sandy loam with low nitrogen availability and has ground vegetation of Calluna type (for closer description of the soil, see Niinistö, Silvola & Kellomäki 2004). The mean density of the Scots pine stand is 2500 trees ha−1. The trees of the stand were about 20 years old and had a mean height of 30 m at the beginning of the experiment.

In the summer of 1996, 16 Scots pine trees with about the same crown size and height were chosen and enclosed individually in closed-top chambers. To reduce shading by other trees, all nearby trees within 2 m from the chambers were cut down 1 year before the start of the experiment. Four trees were subjected to elevated CO2 concentration (+CO2), four to elevated temperature (+T), four to a combination of elevated CO2 and elevated temperature (+CO2+ T) and four trees were controls with ambient CO2 and ambient temperature in the chamber (AmbC). The treatments were started on 24 August 1996 and continued year-round.

Each chamber covered a ground area of 5·9 m2 and had an internal volume of 26·5 m3. The walls of the chambers were built in an octagonal ground structure. Four sides facing southwards were of heatable glass consisting of two elements (clear glass and antisun green; Eglas OY, Imatra, Finland) and the other four of double-wall acrylic channel sheets (standard 16 mm). During the growing season, solar radiation in the chambers was 50–60% less than outside for 82% of the time (Kellomäki, Wang & Lemettinen 2000). The walls of the chambers extended 50 cm below the mean ground surface level in order to sever any root connections and to protect the soil in the chambers from freezing. The temperature of the chambers was automatically regulated by a heat exchanger linked to a cooler (CAJ-4511YHR; L’Unité Hermetique, Barentin, France) installed in the top of each chamber. Unfiltered air was blown into the chamber at the rate of 0·04–0·14 m3 s−1 from a height of 3·5 m by a fan blower. In the growing seasons of 1997 and 1998, the hourly mean vapour pressure deficit (VPD) was 0–0·4 kPa higher  inside  the  elevated  temperature  chambers,  but  0–0·3 kPa lower in the control chambers than that outside the chambers (Kellomäki & Wang 1998b, 2001). The CO2 concentration was elevated by injecting pure CO2 and mixing it with the outside air. During the growing season each chamber was irrigated with 40 L of well water twice a week, and during winter, snow was added to match the snowfall outside. Technical details and the performance of the chambers are presented in Kellomäki et al. (2000).

In +CO2 treatment, CO2 concentration was approximately doubled (693 ± 30 µmol mol−1) compared with the ambient chamber (362 ± 43 µmol mol−1) all day and night throughout the year. The CO2 concentration in the ambient air was 363 ± 38 µmol mol−1. During the growing season (April 15 to September 15), CO2 concentration in the +CO2 chambers was 600–725 µmol mol−1 for 82% of the time. The temperature elevation was adjusted to follow the climatic scenario predicted for the site after doubling of the atmospheric CO2 concentration (Kellomäki & Väisänen 1997). The set temperature increase was 7·5 °C during December to February, 6 °C during March to May, 4·5 °C during June to August and 6 °C during September to November. The actual increase in temperature was, however, lower, on average 2·8 and 6·2 °C above the ambient temperature during the warmest summer and coldest winter months, respectively. For seasonal patterns of CO2 concentration and temperature, see Kilpeläinen et al. (2005) and Niinistöet al. (2004).

Sampling and measurements

For all analysis, needle samples of the third to fifth whorl from the stem apex were collected about 1 year after the start of the treatments in August 1997 and thereafter in August in the next two growing seasons. Results are presented for each needle cohort separately; for example, needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. These same shoots were previously studied also for gas exchange and biochemical properties on the same sampling occasions (Luomala et al. 2003).

Twenty needles per sampled shoot and needle generation of each tree were collected for analysis of fresh weight, projected needle area, dry weight (dried at 60 °C for 48 h), needle length and N concentration. In 1997, needle length was measured with a caliper and projected needle area with a leaf area meter (LI-3050A; Li-Cor Inc., Lincoln, NE, USA); and in 1998 and 1999, needle length and projected area were scanned and measured with WinNeedle 3·1 (Regent Instruments Inc, Quebec City, Canada). For comparison of the two methods, the projected needle area was measured in June 1998 with both methods, which were observed to be consistent (needle area measured by scanning = 0·9918 × needle area measured with leaf area meter + 0·1898, r2= 0·8452). Concentration of N in the dried needles was determined with a CHN Elemental Analyser (model 1106; Carlo Erba Strumentazione, Milan, Italy). In 1997, concentrations of other mineral elements were determined using atomic absorption spectroscopy (Polarized Zeeman Atomic Absorption Spectrophotometer Z-600; Hitachi, Tokyo Japan) and in 1998 and 1999, from wet-digested samples using plasma emission spectrophotometric analysis (ICP; ARL 3800; Thermo Electron Corp., Waltham, MA, USA).

Two samples consisting of 20 needles per shoot and per needle generation of each tree were analysed separately for starch concentration, and the mean of these two concentrations was used as a replicative value of a tree. In 1997, starch was measured by an enzymatic method (Steen & Larsson 1986). The method measures non-structural polysaccharides which are not extracted in a 0·05 m acetate buffer at 60 °C and are hydrolysed to glucose by α-amylase and amyloglucosidase. In 1998 and 1999, starch was measured colorimetrically (Hansen & Møller 1975). Starch concentration (data presented in Luomala et al. 2003) was used for determination of structural dry weight of needles, which was calculated by subtracting the proportion of starch from the total dry weight of a needle.

One needle per sampled shoot and per needle generation of each tree was taken for microscopic studies. Immediately after removing from the shoot, the needles were put in fixative solution containing 1·5% glutaraldehyde, 1·5% paraformaldehyde and 0·05 m cacodylate buffer (0·15 m sucrose and 2 m m CaCl2, pH 7·0). The fixative tubes were kept in a cold bag at +4 °C and transported to the laboratory. Within 2 d, samples of size 0·5–1 mm were cut from a point about 10 mm from the tip of each needle and kept in fixative solution for about 22 h. Further fixation was carried out according to Soikkeli (1980).

Cross-sections for light microscopy were cut from all the needles sampled. The cross-sections were about 2 µm thick and were double-stained first with 2% toluidine blue and then with 1%p-phenyldiamine as described by Sutinen (1987). Each cross-section was digitally photographed (Kodak DCS-460; Kodak, Rochester, NY, USA) with a light microscope (Zeiss Axioplan; Carl Zeiss, Jena, Germany) at 10 × magnification and always parallel to the abaxial side of the needle so that adaxial and abaxial mesophyll and epidermis of both surfaces could be seen. The middle point of the central cylinder was always adjusted to the middle point of the picture. In this way, the same length of each needle cross-section was taken for measurement (Fig. 1). Final measurements were done with Adobe Photoshop 6·0 (Adobe Systems Inc., San Jose, CA, USA) at final magnification of 300 ×. The thickness of epidermis, hypodermis and mesophyll tissue from adaxial and abaxial sides and the thickness of central cylinder were measured from three different locations, which were always the same points of the cross-sectional picture as shown in Fig. 1. For calculations, the mean value of the three points was used. Using the point counting method, the area of the central cylinder as well as adaxial and abaxial areas of the mesophyll tissue and intercellular area were measured. For final analysis, the percentage proportions of intercellular spaces and vascular cylinder of mesophyll tissue were calculated.

Figure 1.

Typical cross-sectional area of a Scots pine needle analysed by light microscopy. The lines show the locations for measurements of thickness of the tissues. The thickest line shows the middle point of the measured area. The abaxial side is on the upper part and adaxial side on the lower part of the picture. The vascular cylinder is encircled with endodermis (En on the middle of the picture). One layer of epidermal cells and under that one layer of hypodermal cells can be seen as outermost layers on the abaxial and adaxial sides of the needle. Ic, intercellular space in the mesophyll tissue; R, resin duct. Scale bar = 40 µm.

Two needles per sampled shoot and per needle generation of each tree that had been collected in fixative solution were rinsed with distilled water and dried with tissue paper. Stomatal density was studied by making cyanoacrylate imprints of needle epidermis (Wilson, Pusey & Otto 1981). Cyanoacrylate adhesive glue (Loctite Super Attack Matic®; Henkel, Avon, OH, USA) was applied onto a mounting glass and a needle was gently pressed on the glue with Parafilm® (American National CM®, Menasha, WI, USA) and held there for 140–150 s. Imprints were viewed and photographed with a combination of a transmission light microscope (160 × magnification; Zeiss Axioplan; Carl Zeiss) and a digital camera (Kodak DCS-460). Images were transported to Adobe PhotoShop 5 (Adobe Systems Inc.). Stomata were counted on a paper copy of the image and stomatal density was determined with the help of a 2 mm scale photographed with the imprints. A section of the tip, centre and base on both surfaces of each needle was studied. For calculations of stomatal densities on the adaxial and abaxial surface of a needle, the mean value of the three sections was used.

Statistical analysis

The effects of the treatments and time as well as their interactions were analysed statistically using a mixed model program MLwiN 1·1 (Multilevel Models Project, Institute of Education, University of London, UK). CO2 and T treatments, time effects and their interactions were regarded as fixed treatment effects, and replicate samples and chambers as sources of random variation (random effects). A maximum likelihood method was used to estimate the linear functions of the fixed effects and the variances associated with the random effects. The statistical significance of the differences between treatments was determined by a likelihood ratio test with a chi-square distribution. The effects of +CO2 and +T were tested using AmbC as the basic level. Thus, the chamber effect was excluded from the effects of the treatments. In order to distinguish whether the effects of +CO2 and +T are purely additive or not, the response at +CO2+ T was tested as the interaction of +CO2 and +T. There is an interaction between +CO2 and +T, if the response at +CO2+ T differs from the additive response of +CO2 and +T. A positive interaction means that the level of the variable at +CO2+ T is greater than the summative effects of +CO2 and +T; and a negative interaction means that the level of the variable is lower. The data for different cohorts were tested separately. For example, the needles grown in 1997 (the 1997 cohort), which were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998, were analysed separately from other data. This approach tells us how the treatments affect needles developed in a specific year, at different needle ages.

RESULTS

Needle dimension

Fresh weight, dry weight and structural dry weight of needles were not significantly affected by the treatments (data not shown). In the 1997, 1998 and 1999 cohorts, average fresh weight of the different treatments varied between 20·5 and 36·6 mg needle−1, average dry weight was 8·2 to 17·2 mg needle−1 and average structural dry weight was 7·3 to 15·4 mg needle−1.

Projected area (Fig. 2) and length of a needle (data not shown) were well correlated (r2= 0·86–0·92) at different treatments and similarly affected by the treatments. At AmbC, needle length of the current-year needles was on average 44·7 mm and of the 1-year-old needles 43·2 mm in the 1997 cohort; 44·7 and 46·7 mm in the 1998 cohort, respectively; and 52·5 mm in the 1999 cohort. In the 1997 cohort as 1-year-old, +T reduced projected area (Fig. 2. P= 0·031) and needle length (data not shown, P= 0·043). Specific leaf weight (g fresh wt m−2 projected needle area), which was studied earlier (Luomala et al. 2003), was stable and, in general, not affected by the treatments.

Figure 2.

Projected area of a needle (mm2 needle−1) in Scots pine grown at ambient control chamber (AmbC), at elevated CO2 (+CO2), at elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T). (a) Needles that were grown to the full length in 1997, i.e. the 1997 cohort; (b) the 1998 cohort; and (c) the 1999 cohort. Treatments were started in August 1996. Values are means of two to four replicate trees (SEM indicated by an error bar).

Stomatal density

Stomatal density was higher on the adaxial surface (on average 75 stomata mm−2 at AmbC) than on the abaxial surface of the needles (on average 63 stomata mm−2 at AmbC) (Table 1). Stomatal density was lower on the base than on the centre and on the tip of the needles (data not shown).

Table 1.  Stomatal density on abaxial (stomata mm−2) and adaxial surface (stomata mm−2) of needles of Scots pine grown at ambient control chamber (AmbC), at elevated CO2 (+CO2), at elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T)
 Stomatal density on abaxial surface (stomata mm−2)Stomatal density on adaxial surface (stomata mm−2)
Current-year1-year-oldCurrent-year1-year-old
  1. Results are presented for each needle cohort separately, for example, needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. Treatments were started in August 1996. Values are means of two to four replicate trees (± SEM). Increases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↑), ↑, ↑↑, ↑↑↑, respectively. Decreases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↓), ↓, ↓↓, ↓↓↓. The responses at +CO2+ T were tested as an interaction of CO2 and T. Positive interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (+), +, ++, +++, respectively, and negative interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (), −, −−, −−−, respectively. Hence, the signs + and − do not tell directly how the response at +CO2+ T differs from that at AmbC, but rather in which direction the response differs from the additive response of +CO2 and +T.

1997 Cohort
 AmbC59·7 ± 4·166·4 ± 4·378·6 ± 5·776·9 ± 4·7
 +CO260·3 ± 7·362·5 ± 1·374·4 ± 4·972·2 ± 6·5
 +T49·6 ± 3·7 ↓50·5 ± 4·1 ↓58·9 ± 7·0 ↓↓58·5 ± 6·2 ↓↓
 +CO2+ T54·7 ± 3·855·8 ± 4·865·7 ± 2·464·2 ± 4·5
1998 Cohort
 AmbC66·0 ± 1·561·9 ± 2·970·9 ± 1·781·0 ± 3·0
 +CO259·5 ± 6·962·3 ± 10·270·4 ± 8·178·5 ± 8·8 ↓
 +T60·8 ± 2·553·8 ± 3·166·1 ± 4·472·8 ± 1·1
 +CO2+ T64·2 ± 2·963·0 ± 1·270·7 ± 3·179·9 ± 7·0 +
1999 Cohort
 AmbC63·3 ± 4·0 75·0 ± 0·4 
 +CO265·2 ± 7·5 79·7 ± 8·1 
 +T55·6 ± 2·0 69·3 ± 0·5 
 +CO2+ T60·1 ± 1·8 76·8 ± 4·1 

At +T in general, stomatal density was lower than at AmbC, but significant reductions were found only in the 1997 cohort (Table 1). The +CO2 treatment reduced stomatal density only on the adaxial surface of the 1-year-old needles of the 1998 cohort. In the 1999 cohort, there were no significant changes in stomatal density in any of the treatments. At +CO2+ T, there was commonly a weak positive interaction on stomatal density, and as a consequence, stomatal density did not decrease at +CO2+ T as much as at +T.

Needle anatomy

The +CO2 and +T increased or decreased occasionally the thickness of epidermis and hypodermis on the abaxial (Table 2) and on the adaxial side of the needles (data not shown). In the 1997 cohort, interaction of CO2 and temperature was negative indicating that epidermis and hypodermis in needles grown at +CO2+ T were thinner than in those grown at AmbC. Mesophyll tissue in both adaxial and abaxial sides tended to be thinner at +CO2 and at +T, but significant differences compared to AmbC were found only on the abaxial side (Table 2). In the current-year needles of the 1997 and 1999 cohorts, interaction of CO2 and temperature on the abaxial side was positive, and the thickness of the mesophyll at +CO2+ T was similar in comparison with that at AmbC.

Table 2.  Thickness of abaxial epidermis (µm), abaxial hypodermis (µm) and abaxial mesophyll (µm) in needles of Scots pine grown at ambient control chamber (AmbC), at elevated CO2 (+CO2), at elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T)
 Abaxial epidermis (µm)Abaxial hypodermis (µm)Abaxial mesophyll (µm)
Current-year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
  1. Results are presented for each needle cohort separately, for example, needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. Treatments were started in August 1996. Values are means of two to four replicate trees (± SEM). Increases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↑), ↑, ↑↑, ↑↑↑, respectively. Decreases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↓), ↓, ↓↓, ↓↓↓. The responses at +CO2+ T were tested as an interaction of CO2 and T. Positive interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (+), +, ++, +++, respectively, and negative interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (−), −, −−, −−−, respectively. Hence, the signs + and − do not tell directly how the response at +CO2+ T differs from that at AmbC, but rather in which direction the response differs from the additive response of +CO2 and +T.

1997 Cohort
 AmbC18·89 ± 1·3620·00 ± 1·4312·78 ± 0·7211·67 ± 0·72161·7 ± 11·2167·8 ± 4·6
 +CO220·28 ± 0·9519·44 ± 0·9612·50 ± 0·5313·61 ± 0·83 ↑↑141·9 ± 7·0 ↓171·7 ± 7·9
 +T19·17 ± 0·2819·26 ± 0·9812·50 ± 0·2811·48 ± 0·37145·6 ± 8·5 (↓)141·9 ± 7·6 (↓)
 +CO2+ T17·04 ± 0·37 −17·50 ± 0·83 −11·48 ± 0·9810·83 ± 0·28 −154·4 ± 12·0 +143·3 ± 9·9
1998 Cohort
 AmbC18·61 ± 0·8317·41 ± 0·3713·06 ± 0·7012·22 ± 0·00154·2 ± 11·2148·1 ± 12·7
 +CO218·89 ± 0·4518·89 ± 1·1111·39 ± 0·95 ↓12·96 ± 0·74132·5 ± 3·3 ↓145·2 ± 1·3
 +T17·78 ± 1·1119·26 ± 0·98 ↑12·59 ± 1·3411·11 ± 0·00119·3 ± 12·4 ↓↓143·7 ± 10·2
 +CO2+T17·22 ± 0·9618·33 ± 0·5612·22 ± 1·5013·89 ± 0·56115·0 ± 6·4146·7 ± 4·4
1999 Cohort
 AmbC18·52 ± 0·98 11·11 ± 0·00 134·1 ± 5·5 
 +CO218·89 ± 1·28 12·59 ± 0·74 (↑) 128·9 ± 2·2 
 +T17·41 ± 0·37 12·22 ± 1·11 123·0 ± 7·1 (↓) 
 +CO2+T18·15 ± 0·74 12·96 ± 0·37 147·4 ± 7·9 ++ 

In the current-year and 1-year-old needles of the 1997 cohort and in the current-year needles of the 1998 cohort, vascular cylinder and whole needle were significantly thinner at +T than at AmbC (Table 3). The +CO2 treatment reduced the thickness of current year needles in the 1997 cohort, in which CO2 and temperature also showed a positive interaction. As a consequence, the thickness of needles at CO2+ T did not decrease as much as if the effects of +CO2 and +T were additive. The proportion of the vascular cylinder of the measured cross-sectional area varied between 47 and 54% without any differences between the treatments (data not shown). In the 1998 cohort, the proportion of abaxial mesophyll occupied by intercellular spaces was greater at +CO2 and at +T than at AmbC (Table 4).

Table 3.  Thickness of vascular cylinder (µm) and thickness of a needle (µm) in Scots pine grown at ambient control chamber (AmbC), at elevated CO2 (+CO2), at elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T)
 Vascular cylinder (µm)Thickness of a needle (µm)
Current- year1-year-oldCurrent-year1-year-old
  1. Results are presented for each needle cohort separately, for example, needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. Treatments were started in August 1996. Values are means of two to four replicate trees (± SEM). Increases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↑), ↑, ↑↑, ↑↑↑, respectively. Decreases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↓), ↓, ↓↓, ↓↓↓. The responses at +CO2+ T were tested as an interaction of CO2 and T. Positive interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (+), +, ++, +++, respectively, and negative interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (−), −, −−, −−−, respectively. Hence, the signs + and − do not tell directly how the response at +CO2+ T differs from that at AmbC, but rather in which direction the response differs from the additive response of +CO2 and +T.

1997 Cohort
 AmbC300·8 ± 12·1277·2 ± 11·1660·6 ± 27·7650·6 ± 13·7
 +CO2290·6 ± 17·5302·2 ± 14·1614·7 ± 28·5 (↓)684·7 ± 14·9
 + T253·6 ± 3·7 ↓↓245·9 ± 12·9 ↓↓592·8 ± 19·0 ↓↓↓582·2 ± 25·7 ↓↓↓
 +CO2+T270·7 ± 11·3231·1 ± 19·8 (−)612·2 ± 27·9 (+)593·6 ± 31·8
1998 Cohort
 AmbC279·4 ± 17·3274·8 ± 6·0628·6 ± 35·5611·5 ± 35·4
 +CO2264·2 ± 13·4291·5 ± 7·7577·8 ± 20·4647·4 ± 21·8
 + T227·0 ± 11·5 ↓↓260·0 ± 17·7524·1 ± 34·5 ↓↓597·0 ± 34·1
 + CO2+T222·8 ± 17·8286·1 ± 33·9523·6 ± 18·5630·0 ± 35·6
1999 Cohort
 AmbC251·9 ± 13·8 568·5 ± 20·8 
 +CO2284·8 ± 20·5 598·5 ± 24·0 
 +T241·9 ± 21·5 552·2 ± 25·2 
 +CO2+T282·2 ± 21·7 625·6 ± 31·2 
Table 4.  The proportion of mesophyll area occupied by intercellular air spaces in abaxial (%) and in adaxial side (%) in needles of Scots pine grown at ambient control chamber (AmbC), at elevated CO2 (+CO2), at elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T)
 Intercellular spaces in abaxial side (% of mesophyll area)Intercellular spaces in adaxial side (% of mesophyll area)
Current- year1-year-oldCurrent-year1-year-old
  1. Results are presented for each needle cohort separately, for example, needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. Treatments were started in August 1996. Values are means of two to four replicate trees (± SEM). Increases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↑), ↑, ↑↑, ↑↑↑, respectively. Decreases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↓), ↓, ↓↓, ↓↓↓. The responses at +CO2+ T were tested as an interaction of CO2 and T. Positive interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (+), +, ++, +++, respectively, and negative interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (−), −, −−, −−−, respectively. Hence, the signs + and − do not tell directly how the response at +CO2+ T differs from that at AmbC, but rather in which direction the response differs from the additive response of +CO2 and +T.

1997 Cohort
 AmbC34·28 ± 4·2220·15 ± 3·2037·78 ± 4·2129·56 ± 3·73
 +CO229·47 ± 3·3823·86 ± 1·8939·02 ± 5·1322·15 ± 3·72
 +T26·30 ± 6·2028·53 ± 5·4232·05 ± 6·0527·88 ± 6·46
 +CO2+T25·50 ± 4·0720·08 ± 3·23 (−)29·47 ± 1·9820·76 ± 3·10
1998 Cohort
 AmbC19·10 ± 3·7120·53 ± 3·4226·18 ± 3·0619·62 ± 1·97
 +CO227·61 ± 6·93 (↑)32·89 ± 4·24 (↑)25·08 ± 2·9422·65 ± 4·43
 +T27·35 ± 5·93 (↑)28·85 ± 5·31 (↑)35·08 ± 2·41 ↑19·43 ± 2·33
 +CO2+ T27·54 ± 2·8520·57 ± 4·8529·40 ± 6·7532·57 ± 6·71
1999 Cohort
 AmbC25·69 ± 3·61 26·39 ± 3·79 
 +CO226·05 ± 7·32 22·66 ± 1·68 
 +T26·49 ± 3·32 27·89 ± 0·31 
 +CO2+ T21·93 ± 7·02 25·59 ± 3·60 

Chemical composition of needles

Concentrations of nutrients in needles were close to those reported earlier for Scots pine grown at this or at a similar nutrient poor site (Helmisaari 1990; Kellomäki & Wang 1997, 1998a; Laitinen et al. 2000). Concentrations of mobile macronutrients N, P and K were lower in 1-year-old needles than in current-year needles, whereas concentrations of poorly mobile nutrients Ca, Mn and Fe tended to be higher in 1-year-old needles (Table 5). Concentrations of S, Cu and Mg, which are considered to be rather immobile from actively metabolizing leaves but moving out of senescing leaves, were commonly lower in 1-year-old needles than in current-year needles. Concentration of B was slightly higher in current-year than in 1-year-old needles. This implies that in Scots pine, B shows some mobility, as earlier suggested by, for example, Helmisaari (1990) and shown in a B-isotope study by Lehto, Kallio & Aphalo (2000). In general, mobility of B varies among plant species (Brown & Shelp 1997).

Table 5.  Concentration of N (mg g−1 DW), P (mg g−1 DW), K (mg g−1 DW), Mg (mg g−1 DW), Ca (mg g−1 DW), S (µg g−1 DW), Cu (µg g−1 DW), B (µg g−1 DW), Zn (mg g−1 DW), Fe (µg g−1 DW) and Mn (µg g−1 DW) in needles of Scots pine grown at ambient control chamber (AmbC), elevated CO2 (+CO2), elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T)
Element Optimal rangeN (mg g−1 DW) >14P (mg g−1 DW) 1·2–1·8K (mg g−1 DW) 4·5–6·0Mg (mg g−1 DW) 0·8–2·2
Current- year1-year-oldCurrent-year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
1997 Cohort
 AmbC11·93 ± 0·6910·70 ± 0·631·37 ± 0·041·09 ± 0·086·32 ± 0·355·10 ± 0·501·09 ± 0·060·66 ± 0·08
 +CO29·71 ± 0·75 ↓↓↓7·46 ± 0·52 ↓↓↓1·15 ± 0·15 ↓0·93 ± 0·06 ↓5·28 ± 0·53 (↓)4·60 ± 0·791·03 ± 0·080·75 ± 0·04
 +T12·40 ± 0·918·49 ± 0·53 ↓↓↓1·19 ± 0·03 (↓)0·89 ± 0·04 (↓)4·50 ± 0·56 ↓↓4·00 ± 0·391·30 ± 0·15 (↑)0·90 ± 0·14 (↑)
 +CO2+ T12·76 ± 0·56 +++7·78 ± 0·22 +++1·34 ± 0·11 ++0·93 ± 0·046·59 ± 0·09 +++4·20 ± 0·751·18 ± 0·300·66 ± 0·14 −
1998 Cohort
 AmbC10·84 ± 0·659·04 ± 0·401·61 ± 0·060·85 ± 0·137·90 ± 0·254·10 ± 0·350·85 ± 0·070·58 ± 0·02
 +CO29·27 ± 0·35 ↓7·95 ± 0·46 ↓1·44 ± 0·06 (↓)1·00 ± 0·058·30 ± 0·424·78 ± 0·780·96 ± 0·050·64 ± 0·05
 +T9·69 ± 0·86 (↓)9·72 ± 0·361·31 ± 0·08 ↓↓0·97 ± 0·067·15 ± 0·634·48 ± 0·380·99 ± 0·04 ↑0·91 ± 0·07 ↑↑
 +CO2+ T10·34 ± 0·63 +8·77 ± 0·931·42 ± 0·09 +0·71 ± 0·07 −7·74 ± 0·523·81 ± 0·950·91 ± 0·11 (−)0·71 ± 0·16 (−)
1999 Cohort
 AmbC11·16 ± 0·89 1·75 ± 0·17 8·89 ± 0·61 0·74 ± 0·05 
 +CO210·13 ± 0·42 1·53 ± 0·08 (↓) 8·31 ± 0·69 0·70 ± 0·09 
 +T12·24 ± 0·87 ↑↑ 1·82 ± 0·03 9·51 ± 0·24 0·95 ± 0·03 (↑) 
 +CO2+ T11·31 ± 0·11 1·56 ± 0·06 7·63 ± 0·39 0·91 ± 0·15 
Element Optimal rangeCa (mg g−1 DW)S (µg g−1 DW)Cu (µg g−1 DW) 3–6B (µg g−1 DW) 8–45
Current- year1-year-oldCurrent-year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
1997 Cohort
 AmbC1·67 ± 0·062·62 ± 0·32 788·8 ± 42·4 3·32 ± 0·14 9·90 ± 1·54
 +CO21·38 ± 0·202·57 ± 0·46 650·0 ± 27·2 ↓↓ 2·87 ± 0·25 (↓) 14·93 ± 3·04 (↑)
 +T1·50 ± 0·282·51 ± 0·37 656·3 ± 46·3 ↓↓ 2·21 ± 0·15? ↓↓↓ 15·19 ± 2·80 (↑)
 +CO2+ T1·80 ± 0·452·77 ± 0·93 700·0 ± 30·0 + 2·35 ± 0·39 9·05 ± 0·56 −
1998 Cohort
 AmbC1·48 ± 0·172·48 ± 0·33793·5 ± 35·0600·5 ± 50·14·08 ± 0·133·24 ± 0·1014·15 ± 1·1011·00 ± 2·53
 +CO21·65 ± 0·262·62 ± 0·37747·3 ± 20·2632·3 ± 48·73·72 ± 0·284·67 ± 1·02 ↑↑↑14·83 ± 1·6510·95 ± 1·65
 +T1·79 ± 0·123·15 ± 0·29 (↑)686·8 ± 42·4 ↓666·5 ± 37·12·83 ± 0·10 ↓↓2·39 ± 0·13? ↓↓12·94 ± 2·0110·89 ± 1·21
 +CO2+ T1·83 ± 0·183·05 ± 0·83749·0 ± 40·6565·0 ± 34·03·08 ± 0·062·78 ± 0·53 (−)9·64 ± 0·7310·39 ± 0·41
1999 Cohort
 AmbC1·69 ± 0·09 828·5 ± 77·5 4·27 ± 0·44 ↓ 14·08 ± 1·24 
 +CO21·59 ± 0·36 761·5 ± 39·6 3·33 ± 0·19 14·13 ± 1·82 
 +T2·52 ± 0·20 ↑ 935·8 ± 33·3 (↑) 3·65 ± 0·16 14·88 ± 1·57 
 +CO2+ T2·32 ± 0·36 827·0 ± 41·0 3·62 ± 0·47 12·34 ± 1·68 
Element Optimal rangeZn (mg g−1 DW) 25–90Fe (µg g−1 DW) 40–100Mn (µg g−1 DW) 70–400
Current- year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
  1. Optimal range of nutrient concentration is from Oleksyn et al. (2002). Results are presented for each needle cohort separately, for example, needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. Treatments were started in August 1996. Values are means of two to four replicate trees (± SEM). Increases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↑), ↑, ↑↑, ↑↑↑, respectively. Decreases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↓), ↓, ↓↓, ↓↓↓. The responses at +CO2+ T were tested as an interaction of CO2 and T. Positive interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (+), +, ++, +++, respectively, and negative interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (−), −, −−, −−−, respectively. Hence, the signs + and − do not tell directly how the response at +CO2+ T differs from that at AmbC, but rather in which direction the response differs from the additive response of +CO2 and +T.

1997 Cohort
 AmbC67·66 ± 6·5374·35 ± 12·7648·19 ± 3·3542·15 ± 4·00369·1 ± 58·2592·3 ± 140·4
 +CO259·42 ± 9·4185·15 ± 11·08 ↑42·53 ± 11·5442·73 ± 7·33448·3 ± 87·2840·0 ± 213·5 (↑)
 +T53·05 ± 5·1856·48 ± 3·0159·80 ± 10·7029·95 ± 2·73437·6 ± 81·9734·8 ± 119·7
 +CO2+ T78·63 ± 21·73 +90·80 ± 23·2061·40 ± 14·9633·55 ± 4·45523·9 ± 177·2908·5 ± 351·5
1998 Cohort
 AmbC49·30 ± 5·7770·15 ± 8·7538·43 ± 1·9958·28 ± 7·59292·0 ± 32·1502·8 ± 82·8
 +CO258·95 ± 5·5072·18 ± 4·1945·18 ± 10·4356·05 ± 9·41474·0 ± 82·9 ↑579·5 ± 91·8
 +T49·75 ± 4·3566·85 ± 6·8436·13 ± 2·3452·90 ± 5·44415·5 ± 46·5664·3 ± 54·9
 +CO2+ T62·00 ± 10·4993·65 ± 33·35 +38·67 ± 7·3276·40 ± 34·60482·7 ± 101·1681·5 ± 218·5
1999 Cohort
 AmbC50·98 ± 2·58 76·23 ± 8·77 279·3 ± 27·6 
 +CO250·75 ± 5·06 60·28 ± 3·91 360·8 ± 97·7 
 +T63·05 ± 2·45 96·93 ± 18·37 455·3 ± 16·7 ↑ 
 +CO2+ T59·35 ± 12·75 90·15 ± 20·62 406·8 ± 84·8 

In the 1997 cohort, mobile nutrients N, P and K and intermediately mobile nutrients S and Cu showed a similar response at +CO2 or +T, as concentrations of these nutrients in general were lower than at AmbC (Table 5). In the 1998 cohort, concentrations of N and P had a decreasing trend at +CO2 and at +T, and concentrations of S and Cu were also lower at +T than at AmbC. In the 1999 cohort, there was an opposite response at +T, as concentrations of N and S were higher than at AmbC. At +CO2+ T, elevated CO2 and temperature generally had a positive interaction on concentrations of N, P, K, S and Cu, and consequently, concentrations of these nutrients were not as low as could be expected on the basis of separate treatments.

At +T, concentrations of intermediately mobile nutrients Mg and B and occasionally also concentrations of immobile nutrients Ca and Mn were higher than at AmbC (Table 5). At + CO2, concentrations of Mg, B, Zn and Mn showed occasional significant or non-significant increases.

In 1-year-old needles at AmbC, mass-based concentrations of K, B, Zn, Cu, Fe and Mn were in optimal range as considered by Oleksyn et al. (2002), whereas concentrations of N, P and Mg were below optimal (Table 5). In current-year needles, however, concentrations of Mg and P were also in optimal range. At +CO2, concentrations of N, P, K and Mg were below or at the limits of optimal. At +T, concentrations of N, P, K and Cu, and at +CO2+ T, concentrations of N, P, K, Mg and Cu were below optimal.

The N-ratios of nutrients [(concentration of a nutrient/concentration of N) × 100%] are used to study the nutritional balance of plants without confounding effects of changes in carbohydrate content and structural dry weight of the needles (Linder 1995). At AmbC, N-ratios of 1-year-old needles were above the target values given by Linder (1995) for 1-year-old needles (Table 6). When compared with the target values, the N-ratios of P and Cu were lowest and mostly at the lower limits of optimal. The N-ratios of Ca, Mn and Zn were constantly much higher than the target values, as has also been found earlier in Norway spruce (Linder 1995) and in black cottonwood (Populus trichocarpa) (Sigurdsson 2001). At +CO2, N-ratios of all nutrients in 1-year-old needles of the 1997 cohort were higher than at AmbC. In the 1998 cohort, the N-ratios of most of these nutrients were higher either in current-year or in 1-year-old needles, whereas in the 1999 cohort, there were no changes in the N-ratios at +CO2. In the 1997 and 1998 cohorts, higher N-ratios at +CO2 probably reflect merely lower concentration of N than changes in the concentrations of other nutrients. At +CO2 in general, all N-ratios were over optimal. At +T, the N-ratios of P, K, Cu and Zn were occasionally lower than at AmbC, whereas the N-ratios of Mg, Ca, S, B and Mn were higher. At +T, the N-ratios of P and Cu were at the limit or below optimal, and at +CO2+ T, the N-ratio of P was at the limit of optimal.

Table 6.  Nitrogen ratio of P, K, Mg, Ca, S, Cu, B, Zn, Fe and Mn on a dry weight basis (%) in needles of Scots pine grown at ambient control chamber (AmbC), at elevated CO2 (+CO2), at elevated temperature (+T), or at elevated CO2 and elevated temperature (+CO2+ T)
N-ratio OptimalP : N (%) 10K : N (%) 35Mg : N (%) 4Ca : N (%) 2·5
Current- year1-year-oldCurrent-year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
1997 Cohort
 AmbC11·5 ± 0·310·2 ± 0·553·7 ± 5·047·5 ± 3·3 9·2 ± 0·5 6·3 ± 0·814·1 ± 0·824·9 ± 3·3
 +CO211·7 ± 0·712·5 ± 0·4 ↑↑54·9 ± 5·561·5 ± 8·4 ↑10·7 ± 0·310·3 ± 1·0 ↑14·0 ± 1·035·2 ± 6·5 ↑↑
 +T 9·8 ± 0·8 ↓↓10·6 ± 1·036·5 ± 4·5 ↓↓↓47·6 ± 4·910·8 ± 1·710·8 ± 1·7 ↑↑12·3 ± 2·729·3 ± 3·5
 +CO2+ T10·4 ± 0·612·5 ± 1·051·6 ± 2·9 +56·7 ± 11·9 9·0 ± 1·8 (−) 8·8 ± 1·6 (−)13·8 ± 2·836·7 ± 11·2
1998 Cohort
 AmbC15·0 ± 1·110·9 ± 0·773·4 ± 3·247·9 ± 3·1 8·0 ± 1·1 6·5 ± 0·613·9 ± 1·928·7 ± 6·3
 +CO215·5 ± 0·4 (↓)12·4 ± 0·689·7 ± 6·5 ↑56·1 ± 12·310·3 ± 0·5 ↑ 8·0 ± 0·6 ↑17·7 ± 2·429·2 ± 2·9
 +T13·6 ± 0·810·0 ± 0·875·2 ± 8·349·5 ± 0·910·5 ± 1·1 ↑ 8·8 ± 0·4 ↑18·8 ± 1·631·9 ± 3·0
 +CO2+ T13·3 ± 0·3 9·872·1 ± 0·6 (−)60·7 8·7 ± 1·6 – –11·217·4 ± 2·949·5
1999 Cohort
 AmbC14·6 ± 0·5 75·7 ± 4·7  6·7 ± 0·4 15·3 ± 1·8 
 +CO214·8 ± 0·7 79·5 ± 11·1  6·3 ± 0·7 14·1 ± 4·0 
 +T15·1 ± 0·7 78·8 ± 4·3  8·0 ± 0·7 21·7 ± 3·4 (↑) 
 +CO2+ T14·2 ± 0·4 70·7 ± 1·2  9·2 ± 1·1 23·4 ± 2·2 
N-ratio OptimalS : N (%) 5Cu : N (%) 0·03B : N (%) 0·05
Current- year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
1997 Cohort
 AmbC 7·4 ± 0·4 0·031 ± 0·002 0·094 ± 0·015
 +CO2 8·8 ± 0·6 ↑ 0·039 ± 0·003 ↑ 0·199 ± 0·036 ↑↑
 +T 7·8 ± 0·7 0·026 ± 0·002 0·175 ± 0·025 ↑↑
 +CO2+ T 9·4 ± 0·1 0·031 ± 0·004 0·121 ± 0·003 −−
1998 Cohort
 AmbC7·4 ± 0·77·2 ± 0·40·038 ± 0·0030·035 ± 0·0020·132 ± 0·0120·115 ± 0·034
 +CO28·1 ± 0·37·8 ± 0·80·040 ± 0·0030·068 ± 0·010 ↑↑↑0·160 ± 0·018 ↑0·122 ± 0·015
 +T7·2 ± 0·56·7 ± 0·30·030 ± 0·002 ↓↓0·024 ± 0·001 ↓↓0·132 ± 0·0110·125 ± 0·009
 +CO2+ T7·0 ± 0·67·60·029 ± 0·0020·042 (−)0·090 ± 0·007−−−0·138
1999 Cohort
 AmbC6·8 ± 0·2 0·035 ± 0·001 0·116 ± 0·007 
 +CO27·2 ± 0·5 0·035 ± 0·001 0·126 ± 0·018 
 +T7·7 ± 0·2 ↑ 0·030 ± 0·002 (↓) 0·125 ± 0·013 
 +CO2+ T7·6 ± 0·3 0·035 ± 0·005 0·123 ± 0·008 
N-ratio OptimalZn : N (%) 0·05Fe : N (%) 0·2Mn : N (%) 0·05
Current- year1-year-oldCurrent-year1-year-oldCurrent-year1-year-old
  1. Optimal N-ratio of a nutrient is from Linder (1995). Results are presented for each needle cohort separately, e.g. needles that were grown to the full length in 1997 (the 1997 cohort) were studied as current-year needles in August 1997 and as 1-year-old needles in August 1998. Treatments were started in August 1996. Values are means of one (no SEM indicated) to four replicate trees (± SEM). Increases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↑), ↑, ↑↑, ↑↑↑, respectively. Decreases at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (↓), ↓, ↓↓, ↓↓↓. The responses at +CO2+ T were tested as an interaction of CO2 and T. Positive interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (+), +, ++, +++, respectively, and negative interactions at P≤ 0·10, P≤ 0·05, P≤ 0·01, P≤ 0·001 are shown as (−), −, −−, −−−, respectively. Hence, the signs + and − do not tell directly how the response at +CO2+ T differs from that at AmbC, but rather in which direction the response differs from the additive response of +CO2 and +T.

1997 Cohort
 AmbC0·57 ± 0·040·71 ± 0·130·41 ± 0·050·40 ± 0·033·1 ± 0·4 5·5 ± 1·2
 +CO20·60 ± 0·051·16 ± 0·16 ↑↑↑0·45 ± 0·120·59 ± 0·12 (↑)4·6 ± 0·711·2 ± 2·7 ↑↑
 +T0·44 ± 0·060·67 ± 0·050·48 ± 0·070·35 ± 0·033·6 ± 0·8 8·6 ± 1·2 ↑
 +CO2+ T0·60 ± 0·141·21 ± 0·270·47 ± 0·090·45 ± 0·084·0 ± 1·112·0 ± 4·3
1998 Cohort
 AmbC0·47 ± 0·080·83 ± 0·140·36 ± 0·040·72 ± 0·082·8 ± 0·5 5·4 ± 1·1
 +CO20·63 ± 0·04 ↑0·94 ± 0·040·50 ± 0·13 (↑)0·72 ± 0·175·0 ± 0·7 ↑↑ 7·6 ± 1·2 ↑↑
 +T0·51 ± 0·010·66 ± 0·08 (↓)0·38 ± 0·020·53 ± 0·084·4 ± 0·6 (↑) 7·1 ± 0·5 (↑)
 +CO2+ T0·60 ± 0·141·62 +++0·36 ± 0·061·42 +++4·7 ± 1·311·5
1999 Cohort
 AmbC0·47 ± 0·02 0·71 ± 0·04 2·4 ± 0·2 
 +CO20·52 ± 0·04 0·58 ± 0·04 3·8 ± 1·1 
 +T0·54 ± 0·04 0·84 ± 0·26 3·9 ± 0·4 (↑) 
 +CO2+ T0·61 ± 0·12 0·91 ± 0·21 4·1 ± 0·8 

DISCUSSION

Stomatal density and needle anatomy were affected more by temperature than CO2

During 3 years of growth at elevated CO2 and temperature, stomatal density of Scots pine needles appeared to be more responsive to temperature than to CO2. At elevated CO2, stomatal density decreased on only one occasion, whereas at elevated temperature and in the combined treatment of elevated CO2 and temperature, stomatal density commonly was lower than in ambient conditions. Significant reductions were, however, found only in the 1997 cohort that was grown in the first year of the experiment and exhibited greatest alterations also in the anatomy, nutrient concentrations and biochemical parameters connected to photosynthesis (Luomala et al. 2003) of all cohorts. In accordance to our results, earlier studies have shown that stomatal density was lower in leaves that were formed in warmer summer temperatures than in spring (Beerling & Chaloner 1993; Ferris et al. 1996). In contrast, opposite results showing higher stomatal density at elevated temperature have been reported (Ferris et al. 1996; Reddy et al. 1998), but with varying responses in stomatal index (proportion of stomatal cells of all epidermal and stomatal cells) (Ferris et al. 1996; Reddy et al. 1998; Apple et al. 2000). Higher stomatal density at elevated temperature might be related to severe water deficit experienced during needle development, as water stress appears to reduce the expansion of epidermal cells and thus increase stomatal density (Bosabalidis & Kofidis 2002), but not to affect stomatal initiation rates and stomatal index (see Royer 2001). Lower stomatal density, in contrast, might be induced by elevated temperature, when water supply is limited but there are no abrupt periods of drought. Slightly higher VPD inside the elevated temperature chambers during the growing season has probably also affected stomatal density.

In general, stomatal density is more strongly affected by environmental factors than is stomatal index, and has been found to be affected for example by ozone (Pääkkönen et al. 1998), UV-B radiation (Visser et al. 1997), photon flux density (Poole et al. 1996), drought stress (Pääkkönen et al. 1998; Bosabalidis & Kofidis 2002), growth altitude (Hultine & Marshall 2000) and with leaf position in the plant (Poole et al. 1996; Lin et al. 2001). Stomatal density may also differ between abaxial and adaxial surfaces (Ferris et al. 1996, 2002) and within the tip, centre and base of a leaf (Ferris et al. 1996), as was also found in this study. Stomatal index is a direct measure of a proportion of epidermal cells differentiated into stomata, whereas stomatal density is influenced also by epidermal cell size, which may be modified, e.g. by elevated CO2 (Ferris et al. 2001, 2002). In this study, we cannot conclude whether the lower stomatal density at elevated temperature and in the combined treatment is caused by greater size of epidermal cells or by a real reduction in the proportion of stomata of all epidermal cells. The thickness of epidermis and hypodermis were significantly changed by CO2 or temperature on only a few occasions, which could imply that in the separate treatments, the expansion growth of epidermal cells in general was not stimulated. In the combined treatment, however, CO2 and temperature had a negative interaction on epidermis and hypodermis in the 1997 cohort, as the epidermis and hypodermis were thinner than in the separate treatments and in ambient conditions.

A decrease in stomatal density gives a potential to reduced use of water, when compared with similar leaf area with higher stomatal density. In this study, however, lower stomatal density observed at elevated temperature in the 1997 cohort did not lead to lower stomatal conductance or transpiration (Luomala et al. 2003). Stomata have apparently acclimated to elevated temperature both functionally by adjustment of stomatal aperture and anatomically by alterations in stomatal density. Internal resistance to gas exchange is partly regulated by the extent of which mesophyll cells face intercellular air spaces. This was studied by measuring the proportion of intercellular spaces in the mesophyll. At elevated temperature, the proportion of intercellular spaces was significantly higher only in the 1998 cohort, whereas in general, treatments had a little impact on intercellular spaces, which implies that internal resistance to gas exchange at this level was not changed. At elevated temperature, however, water use at the canopy level increased, as daily total water flux in shoots (Wang et al. 2003) and daily total sap flow (Kellomäki & Wang 1998b, 2000) were greater than in ambient conditions. Greater sap flow was concluded to be caused by larger total needle area, by reduced stomatal sensitivity to high levels of vapour pressure deficit and by changes in stomatal conductance (Kellomäki & Wang 1998b, 2000).

The needles of the 1997 cohort appeared to be smaller at elevated temperature than at ambient temperature, as they were thinner and shorter and had a smaller projected area. The needles of the 1998 cohort were thinner as well. Reduced needle thickness was associated with thinner mesophyll on the abaxial side and with thinner vascular cylinder. The relative area occupied by vascular cylinder did not change, which suggests that at anatomical level, capacities for water transport and for translocation of carbohydrates were not affected, and that proportion of photosynthetic mesophyll cells was not altered. In Scots pine, the thickness of phloem has been observed to be greater in the needles of mature trees than in those of young trees and to increase during the ageing of needles (Kivimäenpää 2003) and at elevated CO2 (Lin et al. 2001). In older trees, the increase in the volume occupied by vascular cylinder may be related to higher demand for water transport (Apple et al. 2002), and at elevated CO2, thicker phloem has been linked to the increased production and transport of photo-assimilates (Lin et al. 2001).

Although there were reductions in the thickness of needles at elevated temperature, no changes were found in the dry weight of needles (data not shown) or in specific leaf weight (Luomala et al. 2003), which implies that the density of needles may have been greater at elevated temperature than at ambient temperature. A higher density of needles may be brought about by accumulation of starch, by thicker cell walls, by smaller intercellular air spaces or by increased secondary lignification and secondary phloem formation (Ewers 1982). Accumulation of starch (Luomala et al. 2003) or a reduction in intercellular air spaces were not the mechanisms leading to higher density here. Rather, thickening of cell walls and/or formation of secondary phloem may have occurred. Contradictory to numerous studies at elevated CO2 (Yin 2002), needle thickness or specific leaf weight (Luomala et al. 2003) were not increased by CO2 in this study. In agreement, accumulation of starch was seldom observed here at elevated CO2 (Luomala et al. 2003).

Nutrient concentrations were correlated with alterations in the anatomy and in the photosynthetic capacity

At this site, foliar concentrations of nutrients were low, and at ambient chamber, concentrations of nitrogen, phosphorus and magnesium were lower than considered optimal for 1-year-old needles (Oleksyn et al. 2002). Nutritional balance in needles of Scots pine was studied also in terms of N-ratio, which suggested that at this site, P and Cu may have been even more limiting to growth than N. In northern coniferous stands, low P availability may commonly limit growth. Elevation of CO2 may improve the uptake of P because of increased mycorrhizal infection and exudation of oxalate from roots (DeLucia et al. 1997). In this study, however, growth at elevated CO2 led to lower foliar concentration of P, as well as to reductions in concentrations of mobile nutrients N and K and of intermediately mobile nutrients S and Cu. These reductions were most distinct in the 1997 cohort and were associated with down-regulation of photosynthetic capacity, as concentrations of Rubisco and chlorophyll were lower at elevated CO2 than at ambient CO2 (Luomala et al. 2003). In general, changes in concentration of N on a structural dry mass basis were as large as changes on a dry mass basis (Luomala et al. 2003) and since starch concentration at elevated CO2 increased on only a few occasions (Luomala et al. 2003), the reductions in nutrient concentrations are not likely to be caused by a dilution effect of carbohydrates. Neither were there indications of earlier ageing or an ontogenetic drift, as the decreases in nutrient concentrations were generally as large in current-year needles as in 1-year-old needles. Then, lower concentrations of foliar nutrients may be related to a smaller uptake of nutrients and/or to a dilution effect caused by increased growth that was observed in this experiment at elevated CO2 (Kilpeläinen et al. 2005).

Elevation of temperature decreased foliar nutrient concentrations in the 1997 and 1998 cohorts, as concentrations of mobile nutrients N, P and K and of intermediately mobile nutrients S and Cu were mostly lower than at ambient temperature. Higher concentrations of less mobile nutrients (Ca, Mg, B, Zn and Mn) observed here at elevated temperature and occasionally also at elevated CO2 could indicate increased total transpiration in these treatments. In contrast, higher concentrations of N and S at elevated temperature in the third year may be related to increased soil respiration and nutrient mineralization in the soil, as these processes are dependent on temperature (Bonan & Van Cleve 1992) and are stimulated in long-term studies with elevated temperature (see Rustad et al. 2001). This was supported by increased soil CO2 efflux at elevated temperature, although the response of soil CO2 efflux to warming was not greater in this third year than in the two previous years (Niinistöet al. 2004). Nevertheless, higher concentrations of foliar nutrients, especially that of nitrogen, were reflected in the photosynthetic capacity, as the amount and activity of Rubisco enzyme increased at elevated temperature in the 1999 cohort (Luomala et al. 2003). In the combined treatment, elevated temperature compensated for decreases in nutrient concentrations caused by elevated CO2, as at elevated CO2 and temperature nutrient concentrations decreased less than expected on the basis of separate treatments.

At elevated temperature, the reduction in the thickness of needles and the lowering of stomatal density were most distinct in the 1997 cohort. In this cohort, foliar concentration of phosphorus and potassium were also very low. Since K is a principal cation in vacuoles creating osmotic potential and turgor pressure during cell extension (Marschner 1995), low K concentration may have reduced expansion of cells via low cell turgor pressure. Phosphorus deficiency, as well, impairs extension of epidermal cells and reduces expansion of leaves (Marschner 1995). Leaf expansion is more severely affected than chloroplast and chlorophyll formation, and phosphorus deficient plants may even have higher chlorophyll content (Marschner 1995). Current-year needles of the 1997 cohort had, indeed, slightly but not significantly higher chlorophyll concentration at elevated temperature than at ambient temperature (Luomala et al. 2003). In contrast, in the 1999 cohort of all treatments concentrations of P and K were higher, and concurrently, current-year needles were bigger than in the other cohorts. In addition, K is involved in photosynthesis at various levels, as in functioning of stomata, in ATP synthesis as a counterion to the light-induced proton flux across the thylakoid membranes of chloroplasts, in phloem loading and in phloem transport by creating osmotic pressure to drive mass flow in sieve cells (Marschner 1995). On this basis, the reduction of photosynthetic components observed at elevated CO2 in the 1997 cohort (Luomala et al. 2003) may be linked to lower concentration of K in this cohort.

Based on the data of N-ratios, P and Cu could be considered also as limiting nutrients for growth at elevated temperature, and P at elevated CO2 and temperature. Nutritional balance in terms of N-ratio has to be, however, considered carefully, as it is likely that changes in growth conditions could alter the requirement of N for maximal biomass production. At elevated CO2, critical concentration of N required for maximal growth is likely to change because of the importance and magnitude of nitrogen in photosynthetic machinery (Evans 1989). Requirements for other nutrients do not necessarily change in parallel to the requirements for N (Conroy et al. 1990; Johnson, Ball & Walker 1995).

In conclusion, during 3 years of growth at elevated CO2 and temperature, alterations in the anatomy and stomatal density of Scots pine needles were more distinct at elevated temperature than at elevated CO2. Both elevated CO2 and temperature induced changes in nutrient concentrations that were partly corresponding to the biochemical and photosynthetic alterations in the same cohorts (Luomala et al. 2003). Reductions in nutrient concentrations and in the components of photosynthesis (Luomala et al. 2003) were transient and more marked in the needle cohort that was grown in the first treatment year. Thus, acclimation is not similar from year to year, which warrants long-term studies at elevated CO2 and temperature in order to be able to estimate average responses over years. The differential responses of the cohorts may be linked to the general physiological state of the tree and the cohorts, which was observable here as differences in the concentrations of nutrients in ambient conditions. In the combined treatment of elevated CO2 and temperature, stomatal density and concentrations of nutrients were not changed as much as could be expected on the basis of separate treatments. Interactive effects on some needle anatomy parameters were, however, opposite, as they changed more in the combined treatment than in the other treatments.

ACKNOWLEDGMENTS

We thank Ms Mervi Ahonpää, Ms Marja-Leena Jalkanen, Ms Anneli Reentie, Mr Pekka Voipio and Ms Anna-Maija Väänänen for invaluable technical assistance and Mr Alpo Hassinen, Mr Matti Lemettinen and the staff at the Mekrijärvi Research Station for setting up and maintaining the experiment. We are grateful to Dr Juha Lappi for statistical guidance. This work was supported by funding provided by the Graduate School in Forest Sciences, Maj and Tor Nessling Foundation, The Finnish Cultural Foundation and University of Kuopio to E.-M.L. Funding by the Academy of Finland to the Mekrijärvi site (Project no. 64308), co-ordinated by Professor Seppo Kellomäki, is gratefully acknowledged.

Ancillary