Germination of CO2-enriched Pinus taeda L. seeds and subsequent seedling growth responses to CO2 enrichment



  • 1Pinus taeda seeds, developed under ambient or elevated (ambient + 200 µl l−1) [CO2], were collected from Duke Forest, North Carolina, USA in October 1998. Seeds were germinated in nutrient-deficient soil in either ambient or elevated [CO2] (ambient + 200 µl l−1) greenhouse chambers and allowed to grow for 120 days.
  • 2 Seeds that developed in elevated [CO2] had 91 and 265% greater weight and lipid content, respectively, and three times the germination success, compared to those developed in current ambient [CO2].
  • 3 Seedlings from the elevated [CO2] seed source had significantly greater root length and more needles regardless of greenhouse chamber, but there were no treatment effects on tissue or total biomass.
  • 4 Severely limiting nutrient conditions resulted in significant photosynthetic downregulation by seedlings grown in greenhouse chambers with elevated [CO2], regardless of seed source.
  • 5 Our hypothesis that greater seed reserves from CO2 enrichment would synergistically affect seedling growth responses to elevated [CO2] was not strongly supported. Nonetheless, seeds produced in a CO2-enriched environment may have fundamental changes in their viability, chemistry and germination that may affect reproduction.


Increasing atmospheric carbon dioxide concentration, [CO2], has profound effects on growth and development of trees. A doubling of [CO2] generally stimulates photosynthesis (Murray 1995; Saxe, Ellsworth & Heath 1998) and can lead to a substantial increase in tree growth (Poorter 1993). For example, doubling [CO2] increased the rate of photosynthesis per unit leaf area (A) of Pinus taeda by 14–146% (Ellsworth 1999; Liu & Teskey 1995; Thomas, Lewis & Strain 1994; Tissue et al. 1996; Tissue et al. 1997), and increased relative growth by 40–233% (DeLucia et al. 1999; Sionit et al. 1985; Tissue et al. 1996; Tissue et al. 1997). While it is well known that nutrient availability can greatly affect the CO2 response (Kubiske et al. 1998; Pregitzer et al. 1995; Thomas et al. 1994), we suggest that in very young seedlings, elevated [CO2] effects may interact with internal factors such as seed N or C reserves. Thus internal and external factors that regulate growth at the seedling stage may act collectively to regulate subsequent seedling growth (Corbineau & Come 1995).

Virtually all elevated [CO2] work on trees has been conducted on seedlings grown from seed that developed under current ambient CO2 conditions. A parental environment of elevated [CO2] may produce seeds with increased amounts of carbohydrates, lipids and proteins to support early seedling growth. The contents of these substances in seed have profound effects on germination, because they are utilized as both substrate and energy source for the growing seedlings (Kermode 1995). For the same reasons, the reserve content of seeds may also affect seedling responses to atmospheric [CO2]. For example, Miao (1995) grew plants from Quercus rubra acorns and found that seed size acted synergistically with atmospheric [CO2] to enhance seedling growth. Since developing seeds are usually very strong C sinks (Zamski 1995), it is very likely that enhanced photosynthetic production will improve the energy reserves of seed. Thus an important synergistic effect may be missing from virtually all elevated [CO2] studies on trees: that of seed energy reserves in combination with atmospheric [CO2] enrichment. Such studies are essential to fully understand growth and development of tree seedlings in a future, CO2-enriched ecosystem.

Until now, it has been impractical to study seed production and development of forest trees. The development of free-air CO2 enrichment (FACE) technology for forest canopies now makes such studies possible (Hendrey et al. 1999). Several FACE research facilities are presently operating in forest communities around the world, including Pinus taeda in North Carolina, USA (DeLucia et al. 1999; Ellsworth 1999; Myers, Thomas & DeLucia 1999); Populus tremuloides , Betula papyrifera and Acer saccharum in Wisconsin, USA (Karnosky et al. 1999a; Karnosky et al. 1999b); Liquidambar styraciflua in Tennessee, USA (Gunderson et al. 1999); Populus spp. at Viterbo and Siena, Italy (Tognetti et al. 1999); and tropical forest species at Sardinilla, Panama. We collected Pinus taeda seed that developed under FACE to study the subsequent response of seedlings under similar elevated [CO2] conditions. It was hypothesized that applying elevated [CO2] to the maternal parent would have a positive effect on seed energy reserves which, in turn, would result in greater early growth of seedlings. This expectation was based on the premise that the increased seed reserves would provide additional substrates for growth, and also reduce feedback inhibition of seedling photosynthesis under elevated [CO2].

Materials and methods

Seed production, collection and germination

The Duke Forest (Orange County, North Carolina, USA) FACE experiment was constructed in 1996 in a 14-year-old P. taeda plantation. It consisted of a randomized block design of six 30-m-diameter open-air gas fumigation arrays in three replicate blocks. Three of the arrays fumigated the forest canopy with ambient air enriched by 200 µl l−1 CO2, and three with ambient air. The elevated [CO2] plots were fumigated for 81 and 79% of 1997 and 1998, respectively; the annual average [CO2] at the centre of each FACE plot was 199–203 ± 84 µl l−1 above the ambient concentration.

Pinus taeda is an important timber tree species in the south-east USA, and its reproductive cycle is typical of North American pines. It normally reaches reproductive maturity at age 5–10 years (Schopmeyer 1974). Pollination occurs between February and April. Fertilization takes place the year following pollination, and cones mature by October of that same year (Burns & Honkala 1990). In 1997, at age 15, a few of the canopy trees had begun to produce cones. The seeds that were collected in the present study underwent their entire developmental cycle, from pollination through maturity, under experimental treatments.

A total of 39 cones that matured in 1998 were collected from the Duke Forest experiment between October 29 and November 1, 1998. Most of the cones were open at the time of collection, and only the seeds of open cones were used in this experiment. Seeds were de-winged, floated in water for 2 days, and stratified for 120 days at 10 °C; seeds that remained floating were discarded (Schopmeyer 1974).

Following stratification, seeds were sown in 0·16 l (4 cm diameter × 21 cm long) tubes in a greenhouse at Mississippi State University, USA, on March 19, 1999. The tubes were filled with equal amounts of coarse, sterile silica sand. Seeds were placed on the sand surface ≈2 cm below the rim of the tubes. All seeds were watered for 2 min every 2 h by an automatic misting system. No additional fertilizers were used, so that early seedling development was largely a function of seed storage.

The tubes were then placed into six chambers (1·2 × 1·2 × 0·9 m), constructed in the greenhouse, that were enclosed on the top and four sides by transparent plastic and by plywood on the bottom. Each of the six chambers contained 28 seeds: 14 from FACE rings (referred to as the CO2-enriched seed source); and 14 from ambient rings (referred to as the ambient seed source), with rings being equally represented by block and treatment in each chamber, for a total of 168 seeds in the experiment. The seedling tubes were arranged randomly within each chamber. Three blowers, each connected to two chambers (one ambient chamber and one CO2-enriched chamber), circulated air through the chambers at a rate of two chamber volumes per min. Volume-flow regulators were used to dispense pure CO2 gas into the blower air stream of one chamber in each pair. The concentration of CO2 in each chamber was continually monitored with an infrared gas analyser (LI-Cor model 6252, Lincoln, NB, USA) and logged to a computer. Thus three of the chambers were continuously supplied with ambient + 200 µl l−1 CO2 (517–759 µl l−1 at 95% confidence interval, referred to as elevated), similar to the fumigation protocol used at Duke Forest, while the other three chambers (referred to as ambient) were supplied equal volumes of ambient air (365–481 µl l−1, 24 h, 95% confidence interval). Ambient chambers were vented into the greenhouse whereas elevated [CO2] chambers were vented to the outside air. Chamber and soil temperatures (measured in two tubes per chamber with copper-constantan thermocouples) averaged 34·3 and 33·9 °C, respectively, throughout the experiment.

Seedling measurements

The seeds in all chambers were monitored twice daily for emergence of radicles. Upon radicle emergence, the date was recorded and seeds were immediately covered with 1 cm of sand. Hypocotyl height and cotyledon length were measured when the seedlings reached their first resting stage, which was identified by the appearance of a terminal bud. Subsequent height-growth measurements from the sand surface to the top of seedlings were conducted on May 17, June 2 and 27, and July 11 and 16, for all seedlings. At the end of the experiment seedlings had not yet begun to produce needle fascicles. Mean relative height-growth increment for each measurement interval was calculated according to Hunt (1982):

image(eqn 1)

where lnH2 and lnH1 are the natural logs of heights at times T2 and T1, respectively.

Seedlings were harvested on July 17, 1999 and separated into leaves (all leaves were juvenile leaves), stems and roots for additional analysis. The number of needles per seedling were counted and total projected leaf areas were measured with a leaf-area meter (Li-Cor model LI-3100, Lincoln, NE, USA). The length of the first-order root (tap root) from the root collar was measured by carefully extending it along a cm scale. Nitrogen concentrations were determined using a CNS analyser (Fisons Model NA-1500, Milan, Italy) on oven-dry, ground tissues. Biomass was determined following oven drying at 70 °C for 48 h.

Shoot net assimilation efficiency was measured as the initial slope of a photosynthesis − [CO2] response function on two plants from each chamber at the time of harvest. Seedlings were randomly selected from each seed source (ambient and CO2-enriched seed) in each greenhouse chamber for a total of 12 seedlings. A portable infrared photosynthesis system (ADC, Model-LCA2, Hoddesdon, UK) was used. A gas cylinder supplied CO2-enriched air to the system and a soda-lime scrubber was used to modify the cuvette [CO2]. The entire needle-bearing epicotyl was enclosed in a conifer cuvette and exposed to 1000 µmol m−2 s−1 photosynthetically active radiation, which is saturating for Pinus taeda (Liu & Teskey 1995). CO2 and H2O vapour exchange were recorded at cuvette [CO2] ranging from 50 to 500 µl l−1. A period of 10–15 min was allowed for leaf equilibration at each cuvette [CO2]. Seedlings were considered equilibrated when a steady-state CO2 differential was observed for 3 min. Photosynthesis (A) and leaf internal CO2 concentration (Ci) were calculated based on total projected leaf area (measured with a leaf-area meter, Li-Cor model LI-3100) in the cuvette according to Von Caemmerer & Farquhar (1981).

Lipid analysis

A subset of seed was used for total lipid determination using a petroleum ether Soxhlex extraction (Horwitz 1965). Seeds were lyophilized and stored in a vacuum desiccator until being ground with a mortar and pestle. Ground seed material (98–215 mg) from each Duke Forest treatment ring was placed in a separate cellulose thimble (total of six), and lipid was extracted with boiling petroleum ether at 100 °C in a recirculating boiler–condenser. Additional petroleum ether was added periodically to maintain volume at 60 ml. After 8 h, the hot petroleum ether-lipid extract was collected into preweighed flasks and evaporated overnight. Total lipid mass was then determined by weighing the flasks.

Statistical analysis

The relationships between individual seedling height and age were compared using least-squares regression. All other data were analysed as a randomized complete block design with a split-plot treatment arrangement with subsampling. The whole-plot effect was greenhouse chamber, and the split-plot effect was seed source within chambers. For each measurement the mean values within each split-plot unit were used in analysis of variance. Significant differences between treatment means were determined using Fisher’s protected LSD (least significance differences) at P < 0·05.


The CO2-enriched seed source had 91% greater weight (F1,4 = 10·6, P < 0·05) than the ambient seed source (Table 1). In addition, the CO2-enriched seed source had 128% greater lipid concentration (F1,4 = 54·0, P < 0·05) and 265% greater lipid content (F1,4 = 26·78, P < 0·05). Seed source also significantly affected the rate and success of germination (F3,8 = 46·37, P < 0·05) regardless of chamber CO2 treatment (Fig. 1). The total germination success of the CO2-enriched seed source was more than three times that of the ambient source (mean of 95 versus 28%, respectively). In addition, the CO2-enriched seed source began germinating up to 5 days earlier than ambient seed. Greenhouse chamber treatments had no effect on germination.

Table 1.  Dry mass, total lipid concentration and content (mean ± SE) of Pinus taeda seeds collected from a free atmospheric CO2 enrichment experiment, North Carolina, USA
  1. Seeds were produced in either ambient or elevated (ambient + 200 µl l−1) atmospheric CO2 concentration (n = 3, and five to ten seeds per replicate). The CO2 effect was significant for all three parameters (P < 0·05).

Seed mass (mg) 11·6 ± 2·25 22·0 ± 2·31
Lipid concentration (%) 5·30 ± 0·5012·10 ± 0·80
Lipid content (mg) 0·43 ± 0·06 1·57 ± 0·34
Figure 1.

Cumulative germination percent versus days since sowing (DSS; based on date of radicle emergence) of Pinus taeda seed sources that developed under ambient (solid symbols) or elevated (ambient + 200 µl l−1, open symbols and crossed circles) atmospheric CO2 concentration. Seeds were germinated in either ambient (triangles) or elevated (circles) greenhouse chambers. Points represent three replicates of 14 seeds in each seed source by greenhouse chamber combination. Regression equations are significantly different (P < 0·05) (ambient seed source: % germination = −52·15 + 3·71 DSS − 0·02 DSS2, r2 = 0·59; elevated seed source: % germination = −150·22 + 13·80 DSS − 0·19 DSS2, r2 = 0·84. One elevated seed source by elevated greenhouse chamber replicate (crossed circles) was not included in the regression).

Seed source had no significant effect on seedling relative height growth rates (data not shown). However, at 98 days the relative height growth increment for the final 5 days of the study was significantly higher for the CO2-enriched seed source grown in the elevated [CO2] chambers (0·01 cm cm−1 day−1 compared to 0·005 cm cm−1 day−1 for all other treatment combinations; F1,8 = 12·55, P < 0·05). Consequently, the CO2-enriched seed source produced significantly greater (F1,4 = 0·68, P < 0·05) seedling biomass in the elevated [CO2] greenhouse chambers, but those seedlings were not significantly larger than any seedlings grown in the ambient [CO2] chambers (Table 2). The CO2-enriched seed source produced seedlings that had more needles (F1,4 = 14·77, P < 0·05), and longer first-order roots (F1,4 = 13·98, P < 0·05) than the ambient seed source regardless of chamber growth conditions. However, treatment effects on the biomass of those organs were not significant. There were no significant treatment effects on seedling allometry in terms of leaf area ratio (overall mean of 0·20 ± 0·02 cm2 projected leaf area mg−1 stem mass) or root : shoot dry mass ratio (overall mean of 1·15 ± 0·23), or on tissue N concentrations (leaves = 5·89 ± 0·02 mg g−1; stems = 4·46 ± 0·03 mg g−1; roots = 4·86 ± 0·02 mg g−1).

Table 2.  Growth characteristics (mean ± SE) of Pinus taeda seedlings grown in ambient and elevated CO2 (ambient + 200 µl l−1) greenhouse chambers from seed produced under ambient and elevated (ambient + 200 µl l−1) CO2
 Ambient greenhouse chambersElevated greenhouse chambers
ParameterAmbient seedElevated seedAmbient seedElevated seed
  1. Means are of three replicates (n = 3) consisting of three to 14 seedlings each depending on seed germination success. Means within a row not followed by the same letter are significantly different (P < 0·05).

Total seedling height (cm) 7·09 ± 0·46a 6·21 ± 0·46a 6·86 ± 0·46a 7·25 ± 0·46a
Total seedling biomass (mg) 144 ± 7a136·0 ± 7a 118 ± 7b 138 ± 7a
Leaf mass (mg) 54·8 ± 4a 50·4 ± 4a 41·8 ± 4a 50·6 ± 4a
Stem mass (mg) 19·9 ± 1 a 20·4 ± 1a 17·0 ± 1a 19·7 ± 1a
Root mass (mg) 68·8 ± 4a 65·2 ± 4 a 59·0 ± 4b 66·0 ± 4b
Leaf area per seedling (cm2)  3·9 ± 0·2a  3·8 ± 0·2a  3·6 ± 0·2a  3·9 ± 0·2a
Number of leaves   64 ± 2b   73 ± 2a   69 ± 2b   73 ± 2a
Taproot length (cm) 24·0 ± 3·0b 31·9 ± 3·0a 27·3 ± 3·0b 30·8 ± 3·0a

Seedlings grown in greenhouse chambers with elevated [CO2] had significantly lower shoot net assimilation efficiency (shallower initial slope of A–CO2 response curve) compared to those in the ambient chambers for both the ambient seed source (t = 0·019, P < 0·05) and the elevated seed source (t = 0·022, P < 0·05; Fig. 2). Seed source had no effect on whole-shoot net assimilation efficiency.

Figure 2.

Shoot net CO2 assimilation rate of Pinus taeda seedlings grown in either ambient (triangles) or elevated CO2 (ambient + 200 µl l−1, circles) greenhouse chambers, and from seed that developed under ambient (solid symbols, a) or elevated CO2 (open symbols, b). Regression lines were fitted to data for three seedlings from each treatment combination. Regression equations are: ambient CO2-grown seed in ambient CO2 greenhouse chambers (▴), A = 0·416 + 0·011Ci, r2 = 0·97; ambient CO2-grown seed in elevated CO2 greenhouse chambers (●), A = −0·237 + 0·007Ci, r2 = 0·94; CO2-enriched seed in ambient CO2 greenhouse chambers (▵), A = −0·414 + 0·012Ci, r2 = 0·94; CO2-enriched seed in elevated CO2 greenhouse chambers (○), A = 0·787 + 0·006Ci, r2 = 0·98.


Elevated [CO2] dramatically increased the size and lipid content of Pinus taeda seeds. Herbaceous and crop species often exhibit increases in seed size (Baker et al. 1989; Kimball & Mauney 1993) or lipid content (Rogers, Thomas & Bingham 1983) under elevated [CO2]. We know of only one other study that examined such effects in trees: Connor et al. (1998) found a significant increase in Cornus florida fruit mass under elevated CO2 and temperature using branch chambers.

Lipids comprise the largest portion of C storage in Pinus seed, four times that of carbohydrates (Bewley & Black 1994). Various lots of P. taeda seed contained 7·6–9·7% total fatty acids, in agreement with lipid concentrations reported here (Marquez-Millano 1989). Storage lipids are synthesized from photoassimilated sucrose translocated to seeds via phloem (Slack & Browse 1984). Developing fruits and seeds are high priority C sinks, and carbohydrates are usually allocated to favour reproductive versus vegetative structures (Wardlaw 1990). The capacity of vegetative structures to utilize assimilates depends in part on the efficiency of symplastic transport at the site of phloem unloading (Thorne 1985). In contrast, high sink strength of developing fruit and seed results largely from a low-resistance, apoplastic phloem-unloading pathway; the ability of developing seed to utilize photosynthate is more strongly dependent on carbohydrate supply compared to vegetative sinks (Cannell & Dewar 1994; Thorne 1985). Increased assimilation under elevated [CO2], such as that reported for P. taeda at the Duke FACE experiment (DeLucia et al. 1999; Ellsworth 1999; Myers et al. 1999), should generally increase C supply to developing fruit and seed. Moreover, this response should occur even where no vegetative growth response of the maternal parent occurs.

The elevated [CO2] seed source in our study had three times the germination success (90% germination) and earlier germination times compared to the ambient seed source. Both results were probably related to the 265% difference in seed lipid content. During seed development lipids in the nucellar layer, which lies between the seed coat and the megagametophyte, appear to provide metabolic energy for maturation of the embryo (Tillman-Sutela et al. 1996). Also, the lipid content of that layer, composed primarily of polar glyco- and phospholipids that do not restrict the passage of water, was related to imbibition and germination of P. sylvestris (Tillman-Sutela et al. 1996). However, the bulk (85%) of P. taeda seed energy reserves lie in the megagametophyte lipid pool (Stone & Gifford 1999). Germination is an energy-demanding process that is fuelled by respiration of fats. The germination success of P. taeda declined in proportion to the loss of fatty acid content with seed ageing (Marquez-Millano 1989). During germination, most lipid reserves are broken down into their constituent fatty acids, converted to sugars, and transported to the cotyledons and axis for dry weight gain and maintenance (Doman et al. 1982; Gori 1979; Shewry & Stobart 1973; Stone & Gifford 1999; Vanni, Vincenzini & Vincieri 1975). Doman et al. (1982) pointed out that axis dry weight gain in the first 48 h of germination was almost exclusively root growth, which would agree with our findings of earlier radicle emergence and generally longer roots of the CO2-enriched seed source (Fig. 1; Table 2).

Because of its relatively small seed size, Pinus taeda seedlings begin to depend upon current photoassimilates during cotyledon expansion (D. J. Gifford, personal communication). Both seed sources showed significant down-regulation of shoot assimilation efficiency when grown under CO2 enrichment. (It is not likely that stem respiration differentially affected our estimates of shoot assimilation, because leaf-area ratios did not differ among treatments.) Photosynthetic down-regulation under elevated [CO2] is often reported where limited rooting environment restricts the belowground C sink, or where severe nutrient limitations restrict overall plant growth (Arp 1991; Eamus & Jarvis 1989; Sage 1994; Stitt 1991; Thomas & Strain 1991; Tissue et al. 1996). Our seedlings were not constrained by rooting volume at the time of harvest, but they were severely nutrient-limited, which probably contributed to the overall decrease in assimilation efficiency (cf. Kubiske et al. 1998; Medlyn et al. 1999; Thomas et al. 1994) and the overall lack of a CO2 effect on growth.

Growth increases of up to 233% with CO2 enrichment have been reported for P. taeda (Gebauer et al. 1996; Tissue et al. 1996, Tissue et al. 1997). DeLucia et al. (1999) reported a sustained growth increase for 2 years for P. taeda at the Duke Forest FACE experiment. Strong growth responses to elevated [CO2] were not observed in our experiment, and our hypothesis that the elevated CO2 seed source would enjoy a synergistic growth increase under [CO2] enrichment was not strongly supported. The small sample size, severe nutrient limitations and photosynthetic down-regulation probably contributed to unimpressive growth responses.

Our objective in this study was to understand the potential interactive responses of elevated [CO2] on seed development and subsequent seedling performance of an important forest tree species. Seedling responses play an important role in tree regeneration and succession, because germination and initial seedling growth set the pattern for future growth (Miao 1995). For example, in P. taeda, larger seed had significantly shorter germination times than smaller seed, which influenced seedling height after 28 days (Dunlap & Barnett 1983). The growth advantage gained from larger seed mass was observed in standing tree volumes 15 years after sowing (Robinson & van Buijtenen 1979). Thus a small growth advantage due to earlier germination time or seedling vigour often compounds over time to affect tree-growth characteristics years later.

Our results have profound implications for reproductive strategies of trees if elevated [CO2] has dissimilar effects on seed development among species. For example, P. taeda seed reserves directly influence seedling growth for only a very short period of time, after which new photoassimilates support growth. It is entirely possible that larger-seeded competitors, such as Quercus or Carya, would respond in a different way. Although the seedling growth phase of this study did not conclusively support our main hypothesis, there were striking effects of elevated [CO2] on the size, chemistry and germination of the seeds themselves. As our seed were extracted from open cones, it is not clear if our findings are representative of all P. taeda seed produced at the Duke Forest experiment in 1998. One possibility is that individual cones under elevated CO2 contained a larger proportion of lipid-rich, viable seed, and the population of viable seed from the ambient seed source had shed prior to our collection. Another possibility is that the manner of seed shedding varied between CO2 treatments. Additional work is needed to resolve these questions, so that the ecological ramifications of elevated [CO2] effects on Pinus seed production can be assessed.


The Duke FACE project is supported by the US Department of Energy. The greenhouse work was funded in part by competitive grant no. 97-35106-5056, US Department of Agriculture, National Research Initiative. Jeffery Hutchinson aided in construction and maintenance of the greenhouse chambers, and Juanita Mobley helped with the N analysis. Sam Land, Emily Schultz and Jack Vozzo provided critical reviews of the manuscript. Approved for submission to peer review as manuscript number FO145 of the Forest and Wildlife Research Center, Mississippi State University.

Received 6 July 2000; accepted 27 November 2000