Growth and maintenance respiration for individual plants in hierarchically structured canopies of Medicago sativa and Helianthus annuus: the contribution of current and old assimilates


  • Markus Lötscher,

    Corresponding author
    1. Plant Science Department, Technische Universität München, Am Hochanger 1, D-85350 Freising-Weihenstephan, Germany
      Author for correspondence: Markus Lötscher Tel: +49 (0)8161 713722 Fax: +49 (0)8161 713243 Email:
    Search for more papers by this author
  • Katja Klumpp,

    1. Plant Science Department, Technische Universität München, Am Hochanger 1, D-85350 Freising-Weihenstephan, Germany
    Search for more papers by this author
  • Hans Schnyder

    1. Plant Science Department, Technische Universität München, Am Hochanger 1, D-85350 Freising-Weihenstephan, Germany
    Search for more papers by this author

Author for correspondence: Markus Lötscher Tel: +49 (0)8161 713722 Fax: +49 (0)8161 713243 Email:


  • • Respiratory costs of Medicago sativa and Helianthus annuus individuals growing in hierarchically structured stands in a controlled environment were analysed with regard to the daily rate of carbon (C) assimilation.
  • • Net assimilation of new C (An, g C d−1) and respiration rates of new (Rnew, g C d−1) and old C (Rold, g C d−1) were assessed by 13CO2 labelling and gas exchange measurements.
  • • Specific respiration rate of old C (rold, g C g−1 C d−1) decreased exponentially with increasing shoot biomass, but was not affected by the instantaneous relative growth rate (Δwi). The growth coefficient g (Rnew: An) was c. 0.32. In the most severely shaded subordinate plants, g was < 0.2, but low g stimulated rold. The contribution of Rnew to total respiraton (fR, new) and the carbon use efficiency CUE (1 – R/(An +Rnew)) were c. 0.68 and 0.62 for Δwi > 0.1, respectively. For Δwi < 0.1, fR, new and CUE decreased with decreasing Δwi in both dominant and subordinate plants.
  • • The results suggest that Rold was closely related to maintenance, whereas Rnew was primarily involved in growth.
Ag and An

gross and net assimilated C allocated to shoots


total respiration rate


respiration rate of new C


respiration rate of mobilized, old C

an, r, rnew and rold

specific rates of assimilation and respiration per unit shoot C

fR, new

fraction of new C in total respiratory CO2


instantaneous relative growth rate


relative growth rate


Height often varies considerably between individual plants within a stand. These size variations may be the result of differences in germination time and recovery from defoliation or genetic differences in growth rate and shoot architecture. Initial small differences can be increased by competition for resources. As a dense stand develops, the growth rate of the smallest individuals will decrease and they will eventually die due to suppression by larger individuals. In order to understand the dynamics of plant size structure, plant density and species composition in plant communities, growth models are needed that incorporate the growth characteristics and carbon balance of individuals. Effects of lowering irradiance on morphological adaptation of shoots and changes of the photosynthetic apparatus have been well described (Evans & Poorter, 2001; Noguchi et al., 2001a; Oguchi et al., 2003). However, there is still little information about the respiratory costs of plants growing in different hierarchical positions within the canopy.

The two-component respiration model estimates growth and maintenance respiration by calculation of a linear regression between specific respiration and specific growth rate of structural biomass (Amthor, 1989). Growth respiration is associated with growth processes such as synthesis of new structures in growth, nutrient uptake, nitrogen (N) reduction and phloem loading, whereas maintenance respiration is associated with protein and membrane turnover and maintenance of ion concentrations and gradients (Amthor, 1989; Thornley & Cannell, 2000). These processes are not constant and may change with size, ontogeny and chemical composition of plants. For example, increasing lignification with plant age (Sanderson & Wedin, 1988; Engels & Jung, 1998) may generate higher respiratory costs due to the relatively high construction costs of lignin (Penning de Vries et al., 1974). On the other hand, specific respiration (expressed per unit biomass) assigned to maintenance decreases with plant age (McCree, 1983; Stahl & McCree, 1988). This occurs in parallel with decreasing concentrations of N and nonstructural carbohydrates during growth. As these processes are related to plant size, variation in the respiratory costs between dominant and subordinate plants would simply be attributed to the developmental stage of plants rather than to growth conditions, which may be affected by competition. However, there are some indications that shading of plants reduces maintenance costs (McCree, 1982). Moreover, construction costs might be lower for plants growing in low light and with low N supply (Lafitte & Loomis, 1988; Sims & Pearcy, 1991), but such differences were assumed to be small (Poorter, 1994).

Neighbour effects can be seen as direct effects, which reduce resource availability to target plants and indirect effects in that neighbours reduce the growth rate and size of the target. Thus, by comparing the respiratory costs of subordinate (small, shaded) and dominant (tall, well lit top leaves) plants it must be considered that the respiratory costs are influenced by both the effects on current growth (direct neighbour effect) and effects of plant size (ontogenetic effect).

The aims of this study were first to quantify the contribution of currently assimilated C to respiration, and second to assess its contribution to growth and maintenance respiration in shoots of plants competing for light in dense plant stands. Root competition was excluded by growing plants individually in pots. Current assimilates of individuals were labelled with 13C for one photoperiod by moving plants to another near-identical stand growing in the presence of CO2 of altered 13C : 12C composition. The contribution of currently assimilated C to respiration was assessed by measuring the rate of respiration and the 13C composition of the respired CO2 in the following dark period. Respiratory costs of young (small) and old (tall) dominant plants were compared with those of old, subordinate (small) plants in order to differentiate between ontogenetic and direct neighbour effects.

Materials and Methods

Plant material

Medicago sativa cv. Planet and Helianthus annuus cv. Optisol were used in the experiments since both species quickly form a dense canopy with a relatively homogeneous distribution of leaves. Shoots of M. sativa were clipped a few weeks after germination to initiate growth of branches and thus to increase shoot mass per plant and leaf area index.

Experiment 1: ontogenetic effects on respiratory costs  Plants were grown in growth cabinets (E15, Conviron, Winnipeg, Canada). Seeds of M. sativa were germinated in seed trays on sand. One week after germination, seedlings were transplanted into pots (5 cm diameter, 35 cm high) filled with quartz sand (particle size 0.3–0.8 mm). Each pot contained a single plant. Pots were placed into containers (76 cm long, 56 cm wide, 32 cm high) to form stands with 400 plants m−2. There were two containers in each of the four growth cabinets. Growth conditions were adjusted to 16/8 h day/night period, 22/18°C day/night temperature, 75% rh and 350 µmol m−2 s−1 PFD (at pot height). Twice a day the containers were flooded for 1 h with nutrient solution (2.5 mmol l−1 KNO3, 2.5 mmol l−1 Ca(NO3)2, 0.5 mmol l−1 KH2PO4, 1 mmol l−1 MgSO4, 0.12 mmol l−1 Fe as EDTA and micronutrients, pH 6.5). Once a week the pots were flushed with water. Four week after sowing, plants were cut to a stubble height of 5 cm. All plants reached the top of the canopy and were thus qualified as dominant plants and assigned to the I+ treatment. In the 4 wk after defoliation, 16 plants per week were used for labelling and respiration measurements in order to assess respiratory costs during the vegetative growth period.

Experiment 2: effect of PFD gradients on respiratory costs  Plant material, establishment procedures of the plants, nutrient solution, day/night temperature and relative humidity in the four growth cabinets were the same as in exp. 1. The purpose of the treatments was to simulate growth conditions for dominant plants (high PFD at the top leaves, I+) and subordinate plants (low PFD at all leaves, I–). Sowing dates of the treatments were staggered by a week to overcome the limited number of growth cabinets. Three weeks after sowing, seedlings were cut to a stubble height of 3 cm. Afterwards, in both treatments 36 plants were placed in a 0.4 m2 stand formed by artificial plants. The artificial plants were made of a green plastic stem and green, opaque polyvinylchloride (PVC) leaves arranged in layers at intervals of 10 cm. The stands contained 200 artificial plants per m2. The leaf area index (LAI) in each layer was 0.61. A new leaf layer was added to the artificial plants when 50% of the plants overtopped the uppermost artificial leaves by 12 cm. At the beginning of the experiment lighting in the growth cabinets was adjusted to 340 µmol m−2 s−1 PFD (I+) and 85 µmol m−2 s−1 PFD (I–) at pot height, respectively. During the regrowth period, four plants per treatment were harvested at weekly intervals for growth analysis. Eight weeks after sowing, 12 plants per treatment were used for labelling and respiration measurements.

Experiment 3: effect of plant neighbours on respiratory costs  Plant stands (400 plants m−2) were established with H. annuus as described in exp. 1. Growth cabinets were adjusted to 400 µmol m−2 s−1 PFD at pot height during the 16 h light period, 20/16°C day/night temperatures and 70% rh. In order to get plants of different size, sowing occurred on two dates with half of the plants sown a week later than the first half. The first sown (200 m−2) and half of the later sown plants (100 m−2) were supplied with nutrient solution containing 5.0 mmol l−1 nitrate-N. The remaining plants were supplied with 1.0 mmol l−1 nitrate-N. In the following, only data from the high N supply will be presented. Watering with nutrient solution was automated, with a drip irrigation system supplying individual plants. Plants were arranged alternately so that each of the later sown plants was surrounded by four plants from the first sowing. The latter developed into dominant plants with high PFD at the top leaves (I+ treatment), whereas the later sown plants remained subordinate (I– treatment). Some plants of the I+ treatment showed transition from vegetative to generative growth. Four plants per treatment were harvested weekly for growth analyses. Five weeks after the first sowing, 10 plants per treatment were used for labelling and respiration measurements.

13C labelling procedures

The four growth cabinets were part of a steady-state near-natural abundance 13CO2 : 12CO2 labelling system (Schnyder et al., 2003). Air flow through the growth cabinets was controlled in a way that CO2 concentration in the outlet air was constant near 350 µl l−1. In exp. 1, two cabinets were supplied with CO2 having a δ of −2.0‰; in the other two cabinets the δ was −43.7‰. For labelling, individuals were randomly selected and swapped between stands which received CO2 with different δ. In exp. 2, plants were established in the presence of CO2 with a δ of −2.4‰. At the time of labelling, an artificial stand, with the same characteristics as the established stand, was set up in a cabinet that received CO2 with a δ of −46.8‰. Eight individuals per treatment were transferred into this system for 13C labelling. In exp. 3, plant stands were established in the presence of CO2 with a δ of −1.8‰ and −43.7‰, respectively. For labelling, plants were transferred from the cabinets with δ−1.8‰ into those with δ−43.7‰. Transfer of the individuals was at the beginning of the dark period preceding the labelling photoperiod to allow for recovery from any disturbance during transfer. At the end of the labelling light period, the plants were transferred into the respiration cuvettes.

Automated system for measurement of shoot and root respiration

A respiration cuvette consisted of a tube and a top and bottom plate (all made of PVC) (Fig. 1) which could be opened and closed quickly to insert a pot. The space of the shoot chamber was adapted to the shoot size by using tubes of different lengths. A bushing that matched exactly the size of the pot was glued into the bottom plate. A similar system was used to seal the bottom of the pot. Rubber seals were used to make sure that each of the four cuvettes were gas tight.

Figure 1.

Schematic representation of the respiration cuvette. IS, inlet shoot chamber; OS, outlet shoot chamber; OR, outlet root chamber; P, bottom and top Plate 30 × 30 cm, 1.5 cm thick; S, rubber seal (Glisseal, Borer Chemie, Zuchwil, Switzerland); T, tube (Ø 20 cm, 30–90 cm height); TS, temperature sensor; V, axial fan (0.9 m3 h−1, KDE1204PKS, Sunon, Taiwan).

To assess continuously respiration and δ of the respired CO2, an open gas exchange system was constructed with an infrared gas analyser (IRGA) and a continuous-flow isotope-ratio mass spectrometer operated on-line (MS) (Fig. 2). Air from a compressor was scrubbed free of CO2 and water with a self-regenerating adsorption dryer. Mass flow controllers (MFC) were used to mix CO2-free air with CO2 of known δ to obtain a CO2 concentration of 200 µl l−1. Air flow (0.3–1.0 l min−1) into the respiration cuvettes was controlled by a MFC. Before entering the cuvettes, air was humidified to 50% rh. The cuvettes were arranged within a plant growth cabinet that held a constant temperature (18°C in exp. 1 and 2, 16°C in exp. 3). About 0.15–0.45 l min−1 of the cuvette air was drawn through the root compartment with a vacuum pump. The air was then dried and the flow to a multiway valve block controlled by a MFC. The remaining air was directly diverted from the shoot compartment and conducted to the valve block. Reference air (0.9 l min−1) from the mixing vessel was conducted through a second vessel of similar size to the humidifiers, so that any small variation of the CO2 concentration between reference and inlet air was minimized (±1 µl l−1 h−1). The reference air line was also connected to the valve block. With a 3 min interval, air from a single valve was passed to the IRGA and pumped (0.18 l min−1) to the MS. Two minutes after the opening of the valve, a 300 µl sample of air was injected via a GC–GP interface (Finnigan MAT, Bremen, Germany) into the ion source of the MS (Schnyder et al., 2003). Switching of the valves was automated and connected with the programme of the MS to synchronise the sequence of measurements. Readings of the IRGA and MS were continuously stored in files.

Figure 2.

Flow diagram of the automated system for measurement of shoot and root respiration and isotopic composition of the respired CO2. Ad, Adsorption dryer (KEN 3100; Zander, Essen, Germany; and molecular sieve: activated aluminium oxide F200; Alcoa, Houston, TX, USA); Co, screw compressor (S40; Boge, Bielefeld, Germany); CO2 cylinder with known 13CO2 : 12CO2 (Messer Griesheim, Frankfurt a. M., Germany); Ct, condensation tubes to reduce relative humidity to 10%; CC, plant growth cabinet; Hu, humidifier to regulate air humidity by flowing air over a temperature-controlled water surface; IRGA, infrared gas analyser (Li-6262, LICOR, Nebraska, USA); MFC, mass flow controller (MFC1, 0–20 l min−1; MFC2, 0–5 ml min−1; MFC3, 0–10 l min−1; MFC4, 0–5 l min−1; Tylan General, Eching, Germany); MS, mass spectrometer (Delta plus, Finnigan, Bremen, Germany); MV, mixing vessel; RC, respiration cuvette; S, multiway valve block; PMS, air pump; Proot, gas-tight Teflon-lined peristaltic membrane pump (Z.N 811.KT.18, KNF, Neuberger, Freiburg, Germany); PG, pressure regulators; Vv, flow controller. Automated data collecting system (DMP, Hegnau-Volketswil, Switzerland), all air tubes were made of Teflon.

Measurement of dark respiration

Respiration was measured for both labelled and nonlabelled control plants. The latter were always measured a day before or a day after the measurement of labelled plants. At the beginning of the dark period pots were removed from the stands, flushed with 0.5 l water and then filled up with nutrient solution that had been previously aerated with CO2-free air for 3 d. This procedure aimed at removing all CO2 from the root compartment. Afterwards, the pots were enclosed into the respiration cuvettes. The shoot compartments were flushed with 3 l air min−1 to drop the CO2 concentration quickly to 200 µl l−1. When this was reached the drainage of the root compartments was opened and the air fittings installed. Air (1 l min−1) was conducted through the root compartments until the CO2 concentration was near 200 µl l−1. Afterwards inlet air flow and outlet air flow from the root compartment were set according to the plant size. The automatic switching of the valves was started 10 min later. Thus, first measurements of shoot and root respiration and δ were made c. 1.5 h after the plants had been removed from the stands. Respiration rates were measured during 5–6 h. Afterwards plants were harvested.

Measurement of irradiance and harvest

A day before labelling, the vertical light gradient was measured in the stands used for labelling the individuals (Lötscher et al., 2003). PFD was measured with a photon flux meter (sensor head 12 mm wide, 288 mm long; Solems, Palaiseau, France). PFD at the top of the individual (I0) was calculated using the vertical light gradient of the stand and the height of the individuals. At harvest, shoots were separated from roots. The length of the stems was assessed with a ruler and leaf area was measured with an area meter (LI-3100, Li-Cor, NB, USA). Dry mass of shoots was determined after oven drying at 70°C for 72 h. Dried material was ground and analysed (1 mg) for total N and C and C isotope composition using an elemental analyser (Carlo Erba NA1108, Milan, Italy) interfaced to the MS.


The fraction of C assimilated during the labelling photoperiod and accumulated in shoots by the end of the dark period (fA, new) was calculated according to Schnyder & de Visser (1999): fA, new = (δP–δPC)/(δPL–δPC) where δP is the δ of a shoot harvested from the labelling cabinet, δPL is the δ of the C assimilated from the labelling CO2 and δPC is the δ of an ‘unlabelled’ (control) shoot. (Note that ‘unlabelled’ means that the plant was continuously kept in the presence of CO2 with the same δ). δPL was calculated as: δPL = [(δL–Δ)/(1000 + Δ)] × 1000, where δL is the δ of the labelling CO2 and Δ is the C isotope discrimination of shoots, as determined in shoots of unlabelled plants. The δ (in per mil, ‰) was expressed in the conventional form: δ = (13C : 12C in sample/13C : 12C in VPDB standard − 1) × 1000 (Farquhar et al., 1989).

The net assimilation rate of new C (An), that is the mass of C assimilated during the labelling photoperiod and accumulated in shoots by the end of the dark period, was calculated as: An[mg C d−1 plant−1] = fA, newM Cconc where M is the dry mass of the shoots [g] and Cconc the C concentration in the shoots [mg C g−1 dry mass]. Gross assimilation rate [mg C d−1 plant−1] was calculated as Ag = (An+Rnew). Specific net assimilation rate (an) was calculated as An per unit shoot C [mg C g−1 C d−1].

The fraction of C assimilated during the labelling photoperiod and detected in the respiratory CO2, that is the fraction of new C in the respiratory CO2 (fR, new) was calculated as: fR, new = (δR–δRC)/(δRL–δRC) where δR, δRC and δRL are the δ of respired CO2 of a labelled plant, an unlabelled control plant and a plant grown continuously with labelling CO2, respectively. δR, δRC and δRL were calculated as: δ = (δinuin–δoutuout)/(uinuout) where δin, δout, uin and uout are the δ and flow rates of the CO2 entering and leaving the shoot cuvette, respectively.

δRL was measured in exp. 1 in which plants were exchanged between cabinets receiving CO2 of different δ, and where unlabelled plants were harvested from both cabinets. For exp. 2 and 3, δRL was not determined experimentally, but was calculated as: δRL = [(δL–Δ)/(1000 + Δ)] × 1000, with Δ as: Δ[‰] = [(δNL–δRC)/(1000 + δRC)] × 1000, where δNL is the δ of the CO2 supplied to the plants before labelling.

The respiration rate of new C (Rnew) during the dark period was calculated as: Rd, new[mg C h−1 plant−1] = fR, newRd where Rd is the total respiration and Rd = Rd, new + Rd, old. Respiration rate for the whole day (8 h dark and 16 h photoperiod) was calculated as R[mg C d−1 plant−1] = (8Rd + 16Rdinline image), where Tp and Td are the temperatures of the photoperiod and the dark period, respectively. For M. sativa, Q10 was set to 2 as found in a preliminary study (M. Lötscher, unpublished) and calculated from the data of Ziska & Bunce (1994). For H. annuus, Q10 was set to 2.5 according to Tjoelker et al., 2001 (and references therein). Specific respiration rates (r, rnew, rold) indicate respiration rate per unit shoot C [mg C g−1 C d−1].

The instantaneous relative growth rate Δwi[g C g−1 C d−1] of the shoot was estimated as: Δwi = (anrold). In order to test the validity of Δwi, the relative growth rate (Δw = ΔW W−1Δt−1) of slow and fast growing plants in each treatment was estimated by classical growth analysis. To this end, for each harvest plants were grouped into two size classes of shoot dry mass (W). Growth curves were fitted to the W vs time (t) relation and used for the calculation of shoot Δw.

Maintenance and growth respiration were calculated based on the two-component respiration model (Amthor, 1989). In this model r is the sum of specific growth (rg) and maintenance (rm) respiration (Fig. 3):

Figure 3.

The two-component respiration models which estimate growth (rg) and maintenance respiration (rm) based on specific growth rate (a) and relative growth rate (b). r = gΔwg + rm and inline image. Δwg is the specific accumulation rate of new biomass and Δw = Δwgrm. Note that the y-intercept of r–Δw regression in (b) would overestimate maintenance respiration inline image and the inline image : Δwj ratio would overestimate the growth coefficient g′.

r = rg + rm = gΔwg + rm,(Eqn 1)

where inline image is the growth coefficient (C respired per unit new biomass) and Δwg is the new biomass accumulated in the plant per unit plant biomass and time. In the present paper, the net assimilation rate equals the accumulation rate of new biomass in the shoots: an = Δwg. Alternatively, when Δwg is not known, rg and rm can be estimated using the r – Δw regression:

image(Eqn 2)

where inline image. Note that inline image.


Light interception and plant morphology

At the final harvest, PFD at the base of the canopy was less than 10% of I0 in all experiments and treatments. This meant that all plants were exposed to a steep vertical light gradient. Consequently, in each treatment taller plants and – as shoot mass correlated positively with shoot height (data not shown) – heavier plants intercepted more PFD. I0 varied with time and between experiments due to the increasing heights of the plants and the light gradients in the growth cabinets. Hence, I0 increased in exp. 1 during the regrowth period due to the increasing heights of the plants, but was lower than I0 in the I+ treatments of the other two experiments where the plants were even taller (Table 1).

Table 1.  Number of labelled plants (No.), irradiance at the top of the tallest plant (I0), plant height, shoot biomass, N:C ratio of shoots (N:C), specific stem length (SSL) and specific leaf area (SLA) of Medicago sativa and Helianthus annuus
ExperimentTreatmentNo.I0 (µmol m−2 s−1)Height (cm)Shoot mass (g DM per plant)N:C mg N g−1 CSSL cm g−1 DMSLA cm2 g−1 DM
  1. Plants grown in stands harvested 1 wk (Harvest 1) and 4 wk (Harvest 4) after defoliation (exp. 1). Plants grown in stands with high/low PFD at the top of the plants (I+/I–). Values are means (SE).

 M. sativa
1I+ , Harvest 14350 5 (2)0.06 (0.02) 63.4 (3.8) 84 (5) 279 (39)
1I+ , Harvest 4750048 (1)0.68 (0.08) 67.4 (1.9) 70 (5) 409 (22)
2I+861071 (3)3.01 (0.29) 64.9 (3.6) 93 (5) 336 (27)
2I–7 9137 (3)0.34 (0.05)106.5 (2.5)349 (38) 677 (35)
 H. annuus
3I+670072 (2)4.93 (0.69) 66.3 (2.9) 27 (2) 762 (57)
3I–4 6042 (2)0.24 (0.02)179.4 (8.9)296 (24)1685 (104)

Shoot mass at the final harvest of exp. 1 was considerably lower compared with plants of the I+ treatment in the other experiments. In the latter, low PFD compared with high PFD (I– vs I+ treatments) reduced shoot mass by at least 90%, but increased the N : C ratio of the shoots, the specific stem length (SSL) and specific leaf area (SLA). In exp. 2 and 3, N : C ratio, SSL, and SLA decreased with increasing plant height, that is with increasing light interception. These relationships were found between treatments and also within treatments (data not shown).

Measurement of shoot dark respiration

For very small plants, respiration was below the limit of detection. This reduced the number of replicates for analysis of respiration especially at the first harvest of exp. 1 and in the I–treatment with H. annuus (Table 1). Data collection of the dark respiration started c. 1.5 h after the beginning of the dark period and was maintained during the following 5–6 h. Within this period, respiration rate and the fraction of new C in the respiratory CO2 (fR, new) remained constant (Fig. 4). Even in the most severely shaded subordinate plants respiration rate and fR,new were surprisingly constant, indicating that gas exchange was steady within a 24-h period. Thus, average rates of dark respiration and fR,new were used in the calculation of the daily respiration rate.

Figure 4.

(a) Rate of dark respiration (R) and (b) fraction of new C in the respiratory CO2 (fR, new) during the 8 h dark period for Medicago sativa (mean ± SE, n = 8). Values are standardised to the mean respiration rate of an individual. Plants were grown in high PFD (square, exp. 1, final harvest; open circle, exp. 2) or low PFD (closed circle, exp. 2).

Assimilation and respiration relative to shoot mass

The specific respiration of old C (rold) decreased exponentially with increasing shoot mass in the smaller plants, whereas it was rather constant in the heavier plants (Fig. 5a). Values were 48 and 18 mg C g−1 shoot C d−1 for the plants with the lowest and highest shoot mass, respectively. Plants with high shoot mass (> 100 mg C per plant) showed comparable rold in the two PFD treatments, whereas for smaller plants rold tended to be higher in the I– treatment than in the I+ treatment.

Figure 5.

(a) Specific respiration of old C (rold), (b) specific net assimilation rate an, and (c) specific respiration of new C (rnew) in relation to shoot mass for Medicago sativa (square, increasing size of symbols denotes harvests 1–4, exp. 1; circle, exp. (2) and Helianthus annuus (triangle exp. 3). Plants were grown in high (open symbols) (ρ) or low PFD (closed symbols). Regression line in (a) is given for plants of the I+ treatments: y = 18.1 + 38.7e−20.8x, r2 = 0.75.

In exp. 1, specific net assimilation rate (an) decreased with shoot mass although I0 increased during the experiment (Fig. 5b). In the I+ treatments of the exp. 2 and 3, an was higher than expected from the an–mass relationship of exp. 1, reflecting the 20–40% higher I0 in these experiments at the time of labelling. Low PFD in exp. 2 and 3 reduced an significantly. In these treatments, higher shoot mass corresponded with higher light interception (data not shown) and higher an. The two species showed comparable an when shoot mass and PFD treatments were considered.

Specific respiration of new C (rnew) and an showed similar relationships with shoot mass (Fig. 5b,c). In exp. 1, Rnew increased proportionally with An, and the Rnew : An ratio was c. 0.32 (Fig. 6). In the I+ treatments of exp. 2 and 3, the Rnew : An ratio tended to decreased with increasing An, whereas in the I– treatments this ratio was lowest in the most severely shaded plants which had low An.

Figure 6.

(a) Respiration rate of new C (Rnew) in relation to net assimilation rate (An). (b) Rnew : An ratio in relation to An. Definition of symbols see Fig. 5. Dashed line represents Rnew = 0.32An.

Relative growth rate

The instantaneous relative growth rate (Δwi), estimated from the gas-exchange measurement, agreed well with the relative growth rate (Δw) obtained from classical growth analysis (Fig. 7). Δwi estimated the specific net assimilation rate of (new) C in shoots minus the specific respiration rate of old C (anrold), whereas Δw estimated the daily increase of shoot C per unit shoot C. In the I+ treatment of exp. 2 and 3, Δwi was slightly higher than Δw (Fig. 7). This may reflect some loss of dead leaves which was incorporated in the calculation of the Δw, but not in Δwi. Export of old C to roots, which would affect Δw, but not Δwi, may also have contributed to the variation of the relationship. Δwi was highest in young M. sativa plants 1 or 2 wk after defoliation (exp. 1) and then decreased with shoot size. The lowest, even negative Δwi were observed in H. annuus plants exposed to low PFD.

Figure 7.

Relationship between instantaneous relative growth rate (Δwi) and relative growth rate (Δw). Δwi was derived from the gas exchange measurements with Δwi = anrold; Δw was derived from classical growth analysis with weekly harvests. Means ± SE. Definition of symbols see Fig. 5.

Respiration vs instantaneous growth rate

The fraction of new C in the respiratory CO2 (fR, new) increased with increasing Δwi up to a threshold value of c. 0.68 which was reached when Δwi was c. 0.1 (Fig. 8). This relationship was similar for both PFD treatments and species. The values of rnew increased proportionally with Δwi, whereas rold did not correlate with Δwi (Fig. 9b,c). Total specific respiration (r) increased exponentially with Δwi in the I+ treatments (Fig. 9a). In the I– treatments, rnew increased and rold decreased with Δwi so that r was rather similar for all plants.

Figure 8.

Fraction of new C in the respiratory CO2 (fR, new) in relation to instantaneous growth rate (Δwi). Definition of symbols see Fig. 5.

Figure 9.

(a) Total specific respiration rate (r), (b) specific respiration of new C (rnew), (c) specific respiration rate of old C (rold), (d) total respiration (R) per unit gross assimilation (Ag), (e) respiration of new C (Rnew) per unit Ag and (f) respiration of old C per unit Ag, all in relation to instantaneous relative growth rate (Δwi). Definition of symbols see Fig. 5.

Total respiratory costs were c. 40% of the gross assimilated new C (Ag) in shoots for Δwi > 0.05 and rold contributed c. 33% to the total respiratory costs (Fig. 9d–f). For plants with lower Δwi, the contribution of Rold to R increased with decreasing Δwi so that it accounted for most of respiratory costs in the plants with lowest Δwi.

Assessment of the two-component model

The two-component model presupposes that specific maintenance respiration (rm) and growth coefficient g are similar for all plants used in the calculation. Thus, only plants with similar shoot mass (100–400 mg C shoot−1) were selected for the analysis shown in Fig. 10, because rm depends on plant size. This selection considered M. sativa plants from exp. 1 and 2 including both PFD treatments. The use of eqns 1 and 2 resulted in almost identical estimates for the growth coefficient (g = 0.289, g′ = 0.291) and maintenance respiration (rm = r’m = 21.2 mg C g−1 C d−1), respectively (Fig. 10a,b). In these plants the Rnew : An ratio (0.29), that is the slope in the rnew – an regression and the average rold (20.2 mg C g−1 C d−1) were comparable with g and rm (Fig. 10b).

Figure 10.

Specific rates of respiration of total C (r; circle), new C (rnew; triangle) and old C (rold; square) in relation to the instantaneous relative growth rate (a) and the specific net assimilation rate of new C (b) of plants grown with high PFD (open symbols) or low PFD (closed symbols). The regression line in (a) represents inline image, with g′ = 0.29 (0.04 SE), inline image = 0.021 (0.003) g C g−1 C d−1 and r2 = 0.76. Regression lines in (b) represent r = 0.29 (0.04) an + 0.021 (0.004), r2 = 0.77 (solid line); rnew = 0.29 (0.04) an + 0.001 (ns), r2 = 0.73 (dashed line) and rold = 0.004 (ns) an + 0.020 (0.002), r2 = 0.002 (dash-dot line). ns, not significant.



The experiments aimed to analyse respiratory costs in shoots of individuals from different hierarchical positions of dense canopies in relation to the daily C assimilation rate. Dominant plants are exposed to steep vertical light gradients with high PFD at the uppermost leaves. Thus, all plants in exp. 1 and those from the I+ treatments in exp. 2 and 3 experienced light environments typical for dominant plants. By contrast, low PFD on all leaves, as induced in the I– treatments, is typical for subordinate plants. An important characteristic of the study was the analysis of plants sampled from dominant positions in the early growth period (exp. 1) and from dominant and subordinate positions after several weeks of growth (exp. 2 and 3). Thus, ontogenetic effects were distinguished from position effects. The use of two species and both approaches, growing plants in dense ‘natural’ canopies (exp. 1 and 3) and in artificial canopies with altered irradiance (exp. 2), led to the same conclusions.

In general, rnew increased with the specific assimilation rate and relative growth rate, whereas rold was closely correlated with plant size. These relationships were in accordance with the two-component respiration model in which respiration is related to growth and maintenance processes and proportional to the growth rate and plant size, respectively (Amthor, 1989 and references therein).

Maintenance respiration

There are several indications that respiration rate of mobilized C (Rold) can be attributed to processes referred to as maintenance respiration (Rm). First, for plants of similar size rold corresponded to the estimated rm, which was derived from the two-component respiration model (Fig. 10). Second, in studies of others, the respiration rate in prolonged darkness was assumed to reflect maintenance respiration and was on average c. 37% of the respiration rate measured at the end of a regular night in whole plants (McCree, 1974, 1982; McCree & Kresovich, 1978) and 50% in expanding leaves (Hendershot & Volenec, 1989). In accordance with those results, Rold was c. 32% of the total respiration in shoots of plants with Δwi > 0.1 (Fig. 8). In Lactuca sativa, van Iersel (2003) found that the fraction of maintenance in total respiration was c. 25% in plants with Δwi > 0.12. Below this threshold, the maintenance fraction increased comparably with the fraction of Rold in the present study (Fig. 8). These results indicate that maintenance dominates the respiratory costs when Δwi is low, where low Δwi can be due to high plant mass or low radiation interception.

Thirdly, the specific respiration rate of old C (rold, 1.0–3.2 mg CO2 g−1 dry matter h−1) was comparable with earlier reported results for shoot maintenance respiration (Amthor, 1989). Further, the decreasing values of rold (48–14 mg C g−1 C d−1) with increasing plant size for H. annuus (an annual species) and M. sativa (a perennial species) were comparable with values reported for maintenance respiration of whole plants of white clover (McCree, 1982) and sorghum (McCree, 1988). It is concluded that, apart from the most severely shaded, subordinate plants (see Effects of neighbourhood and ontogenesis on respiratory costs section), Rold represented the respiration which is generally attributed to maintenance. This suggests that old C is the main source for maintenance processes such a protein turnover and maintenance of ion concentrations.

Growth respiration

Increasing respiration of new C (Rnew) with net assimilation rate (An) (Fig. 6a) indicates that respiration of currently assimilated C was mainly associated with growth. Furthermore, the Rnew : An ratio was comparable with the growth coefficient g suggesting that growth respiration used mainly new C. The Rnew : An ratio was in the range of 0.25–0.4, which is similar to g-values found for many species and growth conditions (Cannell & Thornley, 2000; Thornley & Cannell, 2000). These values include respiratory costs for construction of new biomass and costs associated with phloem loading and nitrate reduction. For comparison g-values of 0.17–0.33 are thought to represent the direct biochemical costs of synthesis of new biomass (Cannell & Thornley, 2000). Some H. annuus of the I+ treatment showed rather low Rnew : An ratios (Fig. 6b). It is not clear if this result was due to the transition from vegetative to generative growth or due to higher demand of new C for maintenance.

The close correspondence of the Rnew : An ratio and g over a large range of experimental conditions indicates that the Rnew : An ratio can be used as a surrogate of g. In fact, the use of Rnew : An offers significant advantages over g estimated by the two-component respiration model, since the Rnew : An ratio can be estimated for every individual whereas the two-component respiration model gives only an accurate estimate of g when all analysed individuals have the same g and rm.

The estimate of g depends on the assumption made for temperature dependence and light inhibition of the dark respiration. It has been shown that light inhibits dark respiration between 16 and 77% in leaves (Villar et al., 1994; Atkin et al., 2000; Wang et al., 2001) and c. 50% in a ryegrass stand (Schnyder et al., 2003). Therefore, g was calculated with different assumptions of the Q10 value and light inhibition. For example, increasing the Q10 value from 2.0 to 2.5 increased g of M. sativa by 7%, whereas an estimated 50% reduction of dark respiration due to light inhibition decreased g by 27%. Light inhibition of dark respiration was found to be similar within a range of 12–800 µmol m−2 s−1 PFD (Atkin et al., 2000). Thus, variations in g between plants of the I+ and I– treatments were likely not due to differences in light inhibition.

However, first respiration readings were obtained only c. 1.5 h after the beginning of the dark period. Thus, we failed to detect a sharp decrease of respiration rate within the first hours of darkness (McCree, 1982; Noguchi et al., 2001b; Schnyder et al., 2003) and respiration rate might have been underestimated. Nevertheless, the strong correlation between Δwi and Δw (Fig. 7) indicates that the calculated respiration rates were a good approximation of the daily respiratory costs.

Effects of neighbourhood and ontogenesis on respiratory costs

This is the first comprehensive study that quantifies respiratory costs of plants growing in different hierarchical positions. Respiratory costs were mainly defined by the actual size of the plant. Plant size was characterised by shoot mass which had to be maintained and shoot height which affected the light interception and as such the plant morphology, the C assimilation rate and relative growth rate.

Reduced PFD due to plant neighbours (I– treatments) resulted in relatively high specific leaf area and stem length (Table 1). This indicates that these plants invested little in costly secondary growth such as lignification. Thus, construction costs of these plants might be lower than for dominant plants. Indeed, the most severely shaded plants showed a low Rnew : An ratio (Fig. 6). However, as indicated by the higher rold in these plants (Fig. 5a), shortage of new C might have stimulated rold. For these plants, the additional specific respiratory demand for new C, which would have maintained a Rnew : An ratio close to that of dominant plants, was c. 8 mg C g−1 C d−1. Although statistically not significant, this value corresponded with the difference of rold between dominant and subordinate plants when calculated for similar shoot mass. Thus, the data indicate that the construction costs of dominant and subordinate plants were rather similar, although they might have varied in chemical composition. Indeed, Poorter & de Jong (1999) and Martínez et al. (2002) found similar construction costs for species with considerable difference in chemical compositions.

Reduction of construction costs would have been of little importance as maintenance (rold) dominated the total respiratory costs in the most severely shaded subordinate plants (Fig. 9). Thus, modification of g in subordinate positions would have little effect on reducing costs and hence increasing the probability of survival when biomass and, as a consequence, maintenance costs cannot be reduced.

The respiratory costs per unit gross accumulated new C (R : Ag ratio) depended on Δwi and hence, it was similar for dominant and subordinate plants (Fig. 9). For Δwi > 0.05, the R : Ag ratio was rather constant with values of 0.3–0.45 which are in the range found for many environmental conditions and plant sizes (Amthor, 2000; Cannell & Thornley, 2000). The respiration of old C increasingly dominated the respiratory costs in plants with decreasing Δwi. In Lactuca sativa plants a decline of Δwi from 0.17 to 0.01 reduced the carbon use efficiency (CUE = 1 – R/Ag) from c. 0.6–0.2 (van Iersel, 2003), following a similar relationship as in the present study (Fig. 9d). The R : Ag ratio, which is typically higher in ‘natural’ vegetation than in a crop (Amthor, 2000), is partly explained by the close relationship between the R : Ag ratio and Δwi. In a ‘natural’ self-thinning stand Δw differs among individuals where some plants have very low Δw, so that the average R : Ag ratio is higher than in a crop stand with homogeneous Δw.


Labelling individuals in a dense canopy with 13C and subsequent on-line measurement of respiration proved to be a promising method to quantify C allocation and its use in respiratory processes in dominant and subordinate plants. The correlation between respiration of new and old C with the estimated growth and maintenance respiration, respectively, indicated that old, mobilized C seemed to be of little importance for processes related to growth in dominant and moderately shaded plants, but it may have contributed to growth in severely shaded plants.


We thank Angela Ernst-Schwärzli, Wolfgang Feneis and Dr Rudi Schäufele (Technische Universität München, Germany) for technical assistance. The work was supported by the Deutsche Forschungsgemeinschaft (SFB 607).