Author for correspondence: M. B. Lavigne Tel: +1 506 452 3384 Fax: +1 506 452 3525 Email: firstname.lastname@example.org
• To determine if stem respiration (r) varied during cambial reactivation, r was measured during March–July in untreated trees and seedlings, debudded seedlings and girdled seedlings of white ash (Fraxinus americana), red maple (Acer rubrum) and balsam fir (Abies balsamea).
• The r was measured using an infrared gas-analysis system. Cambial reactivation was monitored by light microscopy.
• After increasing modestly about the time of cambial cell swelling, r declined to a minimum for several weeks and then increased markedly as rapid xylem production (XP) began. Growth respiration (Rg) over the experimental period was positively correlated with XP over the same time span, with differences in wood anatomy and XP-measurement method accounting for differences among species. A weak, positive trend was observed between maintenance respiration (Rm) and XP. Rm varied among species.
• The marked springtime increase in r is a nondestructive marker for the onset of rapid XP. Measurements of r made after cambial cell swelling and before rapid XP are appropriate for applying the mature-tissue method to estimate Rg and Rm. Rg reflects XP, particularly the width of differentiating xylem.
Stem respiration rate (r) varies during the year in response to changes in external factors, such as temperature, and to internal factors, such as the annual cycle of activity and dormancy exhibited by the vascular cambium (Vose & Ryan, 2002). A better understanding of the environmental and phenological controls over r would address a shortcoming in process-based models of carbon cycling in forests (Cannell & Thornley, 2000), thereby improving our ability both to explain observed differences in net ecosystem exchange of carbon among sites, species and years (Lavigne et al., 1997; Goulden et al., 1998; Law et al., 2001), and to predict future carbon cycling in response to global warming and associated environmental changes (Thornley & Cannell, 2000).
Accumulating evidence indicates that the annual period of cambial dormancy in woody species native to the northern temperate zone consists of two stages: rest and quiescence (Reinders-Gouwentak, 1965; Little & Bonga, 1974; Little & Pharis, 1995). Rest occurs at the start of the dormant period in late summer–early autumn, in association with the cessation of cambial growth, and is imposed by internal factors that prevent cambial activity even under environmental conditions favorable for growth. Exposure to chilling temperatures (c. 2–10°C) during autumn and early winter gradually changes rest into quiescence, during which stage the dormancy is imposed by unfavorable environmental conditions, typically low temperature. The transition between quiescence and cambial growth (cambial reactivation) is associated with rising air temperatures during the spring.
Cambial reactivation is a prolonged, complex process that varies between species in time and space (Larson, 1994). In addition to myriad ultrastructural and biochemical changes (Catesson, 1994; Larson, 1994; Lloyd et al., 1994; Little & Pharis, 1995), it involves several developments that can be readily observed under a light microscope, beginning with the swelling of cambial zone cells. This swelling is associated with rehydration, protoplasm changes leading to vacuolation, and thinning of the radial wall. It may also be accompanied by the maturation of phloem elements that overwintered in a partially differentiated state. The eventual onset of cambial cell periclinal division signals cambial reactivation per se, which may occur first on either the xylem or the phloem side of the cambium, depending on the species. The differentiation of cambial derivatives into xylem and phloem elements follows, in the case of xylem cells involving enlargement of the surface area of the primary wall, thickening and lignification of the secondary wall, and then protoplast autolysis. Bark slippage and the appearance of first-formed xylem cells are commonly used as indirect indicators of cambial reactivation (Priestley et al., 1933; Atkinson & Denne, 1988).
We investigated r in relation to cambial reactivation, with two purposes in mind. First, to determine whether specific aspects of the cambial reactivation process are marked by changes in stem respiration. Such a marker would facilitate cambial studies because microscopical observation of this meristem is destructive and laborious whereas measuring stem respiration is nondestructive and quickly done. Our second purpose was to determine whether the results obtained would improve the application of the mature tissue method for separating r into Rm and Rg.
In the first of two experimental approaches, r was monitored during the cambial reactivation period at the base of the main stem of large trees of three coexisting species: white ash (Fraxinus americana L.), red maple (Acer rubrum L.) and balsam fir (Abies balsamea (L.) Mill.), chosen to represent a ring-porous deciduous hardwood, a diffuse-porous deciduous hardwood and an evergreen conifer, respectively. Based on reported differences in phenology (Priestley, 1935; Wareing, 1951; Lechowicz, 1984; Atkinson & Denne, 1988; Suzuki et al., 1996), it was anticipated that xylem production (XP) at the stem base would begin later in red maple than in white ash or balsam fir, whereas budbreak in the crown would occur earlier. In the second experimental approach, r was measured in white ash, red maple and balsam fir seedlings subjected to debudding or bark girdling in order to inhibit cambial growth below the treatment point (Larson, 1994; Little & Pharis, 1995). Both debudding and bark girdling were imposed because these treatments could affect cambial reactivation in each species differently, possibly leading to greater understanding of the relationship between stem respiration and cambial reactivation.
Materials and Methods
In the first experiment (denoted woodlot experiment), an area of c. 100 × 100 m was selected in the University of New Brunswick woodlot at Fredericton, NB, where pole-stage trees of white ash, red maple and balsam fir occurred together. Twelve trees of similar size per species were selected, and if necessary their bottom branches were removed to a height of 6–7 m on 23 July 1999 to obtain branch-free stems of similar length. In early March 2000 three of these pruned trees per species were selected, and their main stem was sampled 1.3 m above ground on 22 March, 4 May and 6 June to monitor the progress of cambial reactivation. A second subsample of the pruned trees was selected (Table 1) to measure stem respiration 1.3 m above ground two to three times a week between 23 March and 20 July. Cambial growth of these trees was determined on 21 July, at the same position where stem respiration was measured.
Table 1. Size of white ash (Fraxinus americana), red maple (Acer rubrum) and balsam fir (Abies balsamea) trees used to measure stem respiration in the woodlot experiment
Diameter (cm) at 1.5 m
Total height (m)
Height to live crown (m)
Mean SE measured at beginning of experimental period in March.
15.9 ± 0.6a
16.0 ± 0.3b
6.9 ± 0.2a
16.3 ± 0.5a
15.3 ± 0.7b
6.4 ± 0.3a
16.5 ± 0.4a
12.1 ± 0.2a
6.8 ± 0.1a
The second experiment (denoted nursery experiment) involved the use of white ash, red maple and balsam fir seedlings cultivated in 3 dm3 (white ash and red maple) or 14 dm3 (balsam fir) pots containing peat moss/perlite/crushed rock/loam (2/1/1/1; v/v/v/v) at the Canadian Forest Service tree nursery located 4 km north of the woodlot site. In the case of white ash and red maple, vigorous seedlings with a straight stem were selected in July 1999 from a large group of 3-yr-old, 1–2-m-tall material. In March 2000 a group of 12 seedlings per species was selected to monitor cambial reactivation on 31 March, 1 May, 15 May and 13 June (three trees per species per date). In addition, three groups of seven seedlings per species were selected such that the average length of the main stem was the same for each group per species. A 2-cm-long segment of the main stem, located 23–25 cm above the root system, was ringed in all 21 seedlings per species. In one group of seven seedlings per species the ringing was classic bark girdling whereby all tissues external to the latewood in the xylem of the previous year's annual ring were removed using a scalpel (denoted girdled seedlings). In the remaining two groups of seven seedlings per species only the periderm and subjacent chloroplast-containing cells were removed, leaving the underlying cambium and functional phloem intact. The seedlings in one of these two groups received no additional treatment (denoted control seedlings), whereas those in the second group were completely debudded (denoted debudded seedlings). Debudding was accomplished by decapitating the main stem below the apical whorl of branches and pinching off all buds present elsewhere on the seedling. In the case of balsam fir, 7-yr-old seedlings that had been pruned and defoliated below the 1996 whorl of lateral branches at the start of the 1998 growing season were used. Only 12 seedlings were available; in late March 2000 these were divided into three groups of four seedlings such that the average length of the main stem was the same for each group. One group of seedlings was left untreated as control; the main stem of seedlings in the second group was girdled 5 cm below the 1996 whorl of branches; and the third group of seedlings was completely debudded. This debudding was accomplished by severing the apical whorl of buds from each shoot with clippers and removing all visible axillary buds using a scalpel. For all species the ringing site was coated with lanolin and wrapped with aluminum foil. Subsequently, epicormic shoots appearing anywhere on debudded seedlings or below the ringing site of girdled seedlings were removed at weekly intervals. In addition, callus that developed within the ringing site of girdled seedlings was removed at monthly intervals. Stem respiration rate was measured weekly in all control, girdled or debudded seedlings between 30 March and 13 July below the ringing site in white ash and red maple, and between 26 April and 13 July at the midpoint of the 1995 internode of the main stem in balsam fir. Stem respiration was also measured above the ringing site of white ash and red maple seedlings on the final measurement date. The seedlings were harvested on 14 July (white ash) or 17 July (red maple and balsam fir), and cambial growth was determined at the midpoint of the length of the chamber used to measure stem respiration.
Measurement of stem respiration
Carbon dioxide efflux from stems was measured with an LI-6200 portable photosynthesis system (Li-Cor, Lincoln, NE, USA) and custom-built chambers in a manner similar to that used in soil respiration studies (Norman et al., 1992). For the woodlot experiment, a baseplate consisting of a closed-cell neoprene gasket 2.5 cm thick glued to an aluminum plate was attached to the north side of the stem near breast height. The area where the base plate was attached was prepared by removing loose bark and making the contact area as smooth as possible by chiseling bark. An airtight seal between neoprene and bark was created with putty and stopcock grease, and the baseplate was held in place with hose fittings. A respiration chamber constructed from an acrylic tube cut in half lengthwise was attached to the baseplate to make the measurement. The stem area beneath the respiration chamber was 105 cm2. A thermocouple was inserted under the bark beside each baseplate and temperature was measured at the same time as respiration. For the nursery experiment, a respiration chamber that completely enclosed a section of stem was constructed by using both halves of a 10-cm-long acrylic tube cut in half lengthwise. Half-circle end walls, with notches to provide space for seedling stems when the chamber was in place, were cut from an acrylic sheet and glued to both ends of each half tube. Neoprene foam was used to make a gasket creating an airtight seal between the stem and chamber endwalls. In both experiments the same chamber was used for every measurement, and each tree or seedling was measured once per sampling date.
Stem respiration rate (r) measured at ambient temperature (T) was adjusted to stem respiration rate expected at 10°C (r10) using the following equation:
(r and r10 have units of mol m−2 stem surface s−1 and 2 is our estimate of Q10). Stem diameter was measured with microcalipers at the midpoint of the section enclosed in the respiration chamber on each measurement day.
Values of r10 were estimated for each day of the experimental period by linearly interpolating between values obtained on measurement days. The air temperature recorded hourly at a shaded location of the nursery (T) was used to estimate hourly stem respiration, which was then summed to estimate total respiration for the experimental period (R), again assuming that Q10 = 2, using the following equation:
(R is calculated in g m−2 stem surface). Stem maintenance respiration (Rm, mol m−2 stem surface) after the onset of rapid XP was estimated by using the r̄10 determined using the three to 10 measurements obtained during the 2–3 wk period immediately before rapid XP began. Stem growth respiration (Rg) was estimated as the difference between R and Rm (Rg = R − Rm).
Measurement of growth
The development of vegetative buds was monitored on every date that stem respiration was measured until the occurrence of budbreak, defined as when expanding leaves broke through the bud scales. This was assessed (using binoculars) for buds located at the base of the crown in the woodlot experiment, and for buds located at the crown apex in the nursery experiment.
Cambial reactivation and growth were deduced from counts of the number of fusiform cambial cells and phloem and xylem elements per radial file in transverse sections obtained from small stem blocks. Stem blocks were collected using a scalpel, chisel and hammer in the woodlot experiment, maximizing the distance around the stem circumference between sampling sites in trees sampled repeatedly, and using a saw in the nursery experiment. Each stem block was reduced in size using a bandsaw, and the periderm/rhytidome and excess secondary phloem and xylem were removed with a scalpel. These blocks were subdivided longitudinally, if necessary, fixed in formalin/glacial acetic acid/water/95% ethanol (1/1/4/14; v/v/v/v), and stored in 70% ethanol. Samples for sectioning were dehydrated in a tertiary-butyl alcohol series, embedded in Paraplast, softened by soaking in Gifford's #1 solution (Gifford, 1950) or glycine/70% ethanol (3/7; v/v), sectioned transversely at 10 mm on a rotary microtome, placed on glass slides, stained in safranin-fast green (Johansen, 1940) or toluidine blue (Berlyn & Miksche, 1976), and mounted in Permount. Cells were counted along rays between the last-formed cell in the latewood of the previous year's annual ring of secondary xylem and a conspicuous marker in the secondary phloem. This marker was the innermost band of sclerenchyma cells in white ash and red maple, and the innermost band of axial parenchyma cells in balsam fir. For white ash and red maple, only axial xylem cells other than vessel elements were included in the count, and cell counts were not made where vessel elements abutted rays. Cambial cells were assumed to be those fusiform cells lacking radial enlargement. Separate cell counts were made for the xylem, cambium and phloem starting from the latewood, and then the total number of cells was counted from the phloem marker back to the latewood. Counts were repeated until the sum of tissue cell counts equaled the total cell count. Two to four radial files of cells were counted per section. On occasion, separate counts were made of xylem cells with cytoplasmic staining, to determine the number of cells undergoing differentiation at the time of sampling. In addition, transverse hand-cut sections were obtained from the stem blocks harvested from nursery seedlings at the end of the experimental period, and the radial width of xylem was measured after staining the sections with an aqueous solution of phloroglucinol in 20% hydrochloric acid, as described previously (Eklund & Little, 1996). Xylem radial width is positively correlated with xylem cell number (e.g. for white ash, red maple and balsam fir on 21 July in the woodlot experiment, = 0.96, P < 0.001; see also Gregory, 1971; Little & Sundberg, 1991).
sigmastat ver. 2.03 (Jandel Scientific, San Rafael, CA, USA) was used for statistical analyses. Simple linear regression was used to examine relationships between stem respiration, xylem production and stem properties for each species, and regression coefficients were compared among species by one-way ANOVA. The change in stem respiration over time was analyzed by repeated-measures ANOVA. One-way ANOVA was used to compare treatments in the nursery experiment, and treatment means were compared using Duncan's new multiple range test when ANOVA indicated significant differences (P < 0.05). In the figures, the length of the vertical line from the mean represents one standard error (SE).
Budbreak occurred first in red maple, then in white ash and finally in balsam fir, on about 1 May, 8 May and 15 May, respectively. By 6 June shoot elongation was well advanced in every species, with leaf expansion being more rapid in red maple than in white ash.
On 22 March the cambium appeared quiescent in all three species, as the bark did not peel, the cambial zone cells were radially narrow (although some swelling may have occurred in white ash), and radially enlarging cambial derivatives were absent on the phloem and xylem sides of the cambium (Fig. 1a,e,i). By 4 May, however, the cambium had reactivated in every species, as indicated by the swelling of cambial zone cells and the presence of cambial cell divisions and differentiating phloem cells. The bark peeled readily and radially enlarging xylem elements were present in white ash (Fig. 1b) and balsam fir (Fig. 1j), but not in red maple (Fig. 1f). Whether the first-formed phloem cells in any species were derived from overwintering immature cells on the periphery of the cambium could not be determined with certainty. By 6 June XP was rapid, and wide bands of xylem elements undergoing radial enlargement and secondary wall thickening and lignification were present in white ash (Figs 1c, 2a) and balsam fir (Figs 1k, 2e). In red maple XP had just begun, as the bark peeled but only one or two (if any) radially enlarging xylem elements were present (Figs 1g, 2c). On 21 July XP was rapid and many xylem elements in all stages of differentiation were present in red maple (Figs 1h, 2c), as well as in white ash (Figs 1d, 2a) and balsam fir (Figs 1l, 2e). Throughout June and July phloem cell number increased slightly in all three species (Fig. 2a,c,e), indicating that phloem production continued during this period. The decrease in the size of the last-formed vessels in white ash (Fig. 1d) and the increase in the wall thickness of the last-formed tracheids in balsam fir (Fig. 1l) indicate that latewood formation had begun in these species on 21 July.
The r10 of all three species varied significantly during the experimental period (P < 0.001). It was initially high for a short period, then declined to a minimum that continued for several weeks (Fig. 2b,d,f). Eventually r10 increased markedly, the increase occurring first in balsam fir (late May), then in white ash (early June), and lastly in red maple (mid-June). The r10 declined between mid-June (r̄10 = 2.02) and late July (r̄10 = 1.52) in balsam fir (Fig. 2f), and this decrease was associated with a reduction in the number of differentiating xylem cells (Fig. 1k,l; 13.4 0.8 and 9.0 0.5 for 6 June and 21 July, respectively, P < 0.001).
For all three species R and Rg were positively correlated with total XP, measured as xylem cell number (Fig. 3a,b; Table 2). The slope of the relationship between R and Rg vs total XP (coefficient b in Table 2) was greater for balsam fir than for white ash and red maple (P < 0.05). The relationship between Rm and XP was not significant for any species (Fig. 3c; Table 2), although in all cases Rm increased with total XP.
Table 2. Results of simple linear regression between stem total respiration (R), growth respiration (Rg) or maintenance respiration (Rm), and total xylem production (XP) measured as xylem cell number (e.g. R = a + b × XP) for trees in the woodlot experiment
SEE denotes standard error of estimate. n = 7 for white ash (Fraxinus americana); n = 6 for red maple (Acer rubrum) and balsam fir (Abies balsamea).
Total respiration (g m−2)
Growth respiration (g m−2)
Maintenance respiration (g m−2)
The resumption of growth in seedlings was comparable to that observed in large trees of the woodlot experiment. Budbreak was earlier in red maple (about 5 May) than in white ash and balsam fir (about 12 May). However, cambial reactivation began sooner in white ash than in red maple, as indicated by the relative timing of bark peeling, appearance of first-formed xylem cells and rate of increase in XP. The bark peeled on 1 May in white ash, but not until 15 May in red maple. Radially enlarging vessels were present in white ash, but not in red maple, on 15 May (Fig. 4c,g). On 13 June wide bands of xylem elements undergoing radial enlargement and secondary wall thickening and lignification were present in white ash, whereas only a few radially enlarging xylem elements were found in red maple (Figs 4d,h, 5a,c).
For all species, girdling did not affect the timing of budbreak; however it not only resulted in the development of relatively small, chlorotic leaves, but also decreased total XP and narrowed the zone of differentiating xylem elements. The inhibitory effect of girdling on XP was greater below than above the ringing site (Figs 6a–c, 7). Debudding also markedly inhibited XP, but to a much smaller extent in balsam fir than in white ash or red maple (Figs 6a–c, 7). In debudded seedlings of both deciduous species, XP was equally inhibited above and below the ringing site (Fig. 6a,b).
The r10 in seedlings of all species varied significantly over time (P < 0.001). For balsam fir r10 was low in late April and early May, began to increase in mid-May, continued to increase until mid-June, and subsequently remained the same or decreased slightly (Fig. 5e). However, in white ash and red maple seedlings the increase in r10 occurred later (late May and mid-June, respectively) than in balsam fir, and thereafter it continued to increase until the end of the experimental period in mid-July (Fig. 5b,d). At this time of year the zone of differentiating xylem elements was wider in seedlings than in large trees, particularly of white ash (cf Fig. 7a with Fig. 1d; Fig. 7e with Fig. 1h; Fig. 7i with Fig. 1l). Both girdling and debudding prevented an increase in r10 below the ringing site in white ash and red maple (data not shown). Measured in mid-July, girdling decreased the r10 below the ringing site in all species, and to a small extent above the ringing site in white ash (Fig. 6d–f). Debudding did not significantly decrease the r10 in balsam fir, but did so above and below the ringing site in both white ash and red maple (Fig. 6d–f).
R and Rg were positively correlated with total XP (Fig. 8a,b; Table 3). The slopes of these relationships were greater for white ash than for the other two species (P < 0.01). Rm was also greatest for white ash (P < 0.001) (Fig. 8c; Table 3). Rm was positively related to total XP in balsam fir: in white ash and red maple the relationships were also positive, but not significantly.
Table 3. Results of simple linear regression between stem total respiration (R), growth respiration (Rg) or maintenance respiration (Rm) and total xylem production (XP) measured as xylem radial width (e.g. R = a + b × XP) for seedlings in the nursery experiment
SEE denotes standard error of estimate. n = 12 for white ash (Fraxinus americana) and red maple (Acer rubrum); n = 10 for balsam fir (Abies balsamea).
Total respiration (g m−2)
Growth respiration (g m−2)
Maintenance respiration (g m−2)
Most studies that measured r over extended periods (Linder & Troeng, 1981; Sprugel, 1990; Edwards & Hanson, 1996; Lavigne & Ryan, 1997; Stockfors & Linder, 1998; Maier, 2001; Damesin et al., 2002; Edwards et al., 2002) can be interpreted to show that the temperature-corrected r determined following the end of stem diameter growth in autumn was lower than that observed during the several weeks in spring preceding the onset of XP. Maier (2001) found that the use of autumn measurements of r to estimate Rm produced substantially lower estimates than using spring measurements. We believe this seasonal difference in r reflects variation in metabolic activity in the cambial region (Little & Bonga, 1974; Riding & Little, 1984; Little & Pharis, 1995). The cessation of cambial growth in late summer/early autumn is associated with the entry of the cambium into rest. This stage of dormancy is characterized by the presence of nondividing cambial zone cells and metabolic activity associated with the differentiation of a decreasing number of xylem elements and the development of frost tolerance in cambial region cells. Metabolic activity declines to a minimum on completion of xylem differentiation and frost hardening during autumn, at which time the cambium is quiescent and little maintenance respiration is required. The changeover from quiescence to activity during the following spring (cambial reactivation) involves developments that result in increased metabolic activity, requiring substantially more maintenance respiration to support (see Introduction). We propose that the physiological condition of cambial region cells in spring, during the interval between cambial cell swelling and the onset of rapid XP, represents the metabolic state requiring maintenance during the growing season better than does the autumnal physiological condition associated with quiescence.
Considered together, the results of our respiration and anatomical measurements indicate that changes in r monitor the progress of cambial reactivation. A modest increase in r10 occurred early in spring, most obviously in the woodlot experiment in which frequent measurements were made (Fig. 2b,d,f). This transient increase is attributed to growth respiration being expended in association with the onset of cambial reactivation, particularly cambial cell swelling. The marked increase in r10 observed in late spring (Figs 2b,d,f, 5b,d,e) coincided closely with the onset of rapid XP in each species (Figs 1, 2a,c,e, 4, 5a,c). The concomitant delay of the increase in r10 and the initiation of rapid XP in red maple, compared with white ash and balsam fir, was particularly striking. Additional evidence for a causal relationship between the late spring increase in r10 and the onset of rapid XP is the finding that r10 increased only where rapid XP occurred, i.e. above but not below the girdle in all species of girdled seedlings, and in debudded seedlings of balsam fir but not of white ash or red maple (Fig. 6). In both trees and seedlings the r10 appeared maximal in June for balsam fir and in July or later for red maple and white ash (Figs 2b,d,f, 5b,d,e), suggesting that XP peaked earlier in balsam fir than in the deciduous species. This view is supported by the observations that the zone of differentiating xylem decreased in width between the last two sampling dates in the woodlot balsam fir (Fig. 1k,l).
Four observations provide evidence that the positive relationship between Rg and XP is mediated through the width of the zone of differentiating xylem cells (Sprugel & Benecke, 1991). First, the marked increase in r10 at the onset of rapid XP was associated with the appearance of a wide band of differentiating xylem (Figs 1, 2b,d,f, 4, 5b,d,e). Second, treatments (debudding, girdling) that inhibited XP and decreased r10 also reduced the width of differentiating xylem (Figs 6, 7). Third, the decline in r10 between mid-June and late July in balsam fir trees in the woodlot experiment (Fig. 2f) was accompanied by a decrease in the number of differentiating xylem cells per radial file (from 13.4 to 9.0, P < 0.001; Fig. 1k,l). Fourth, both r10 and the width of the zone of differentiating xylem were markedly greater in seedlings than in trees of white ash (cf Fig. 1d with Fig. 7a; Fig. 2b with Fig. 5b).
Differences among species in stem growth respiration per unit of xylem produced (rg), estimated by the slope of the relationship between Rg and total XP (b in Tables 2, 3), can be attributed partly to differences in wood anatomy and the methods used to measure total XP, and partly to differences in the chemical composition of the wood. The value of rg was higher for balsam fir than for both deciduous species in the woodlot experiment because the xylem cells counted in balsam fir were larger in diameter and had thicker walls than the fibers counted in the deciduous species and thus were more expensive to produce (Chung & Barnes, 1977). Similarly, white ash had a higher rg than the other species in the nursery experiment, in which XP was measured as xylem width, because wood density was higher in white ash than in the other species (Panshin & de Zeeuw, 1970). Accordingly, we conclude that stem growth respiration per unit of biomass or carbon gain was much more similar for the three species than was rg based on number of cells produced or per millimeter of radial enlargement, provided that r measurements made in late spring during the 2–3 wk before rapid XP were used to apply the mature tissue method for separating Rm and Rg. If this conclusion is generally true, then the variability in rg among species and sites reported in previous studies (Sprugel et al., 1995; Edwards & Hanson, 1996; Lavigne, 1996; Lavigne & Ryan, 1997; Maier et al., 1998; Maier, 2001; Ceschia et al., 2002; Damesin et al., 2002; Edwards et al., 2002; Vose & Ryan, 2002) could be partly caused by making the r measurements used to estimate Rm at an inappropriate time.