Relative importance of microhabitat, plant form and photosynthetic physiology to carbon gain in two alpine herbs

Authors

  • M. J. Germino,

    1. Department of Botany, University of Wyoming, Laramie, WY 82071–3165, USA, and
    2. Department of Biology, Wake Forest University, Winston-Salem, NC 27109–7325, USA
    Search for more papers by this author
    • Present address: Thermal Biology Institute, PO Box 173120, Montana State University, Bozeman, MT 59717–3120, USA.

  • W. K. Smith

    1. Department of Biology, Wake Forest University, Winston-Salem, NC 27109–7325, USA
    Search for more papers by this author
    • Author to whom correspondence should be addressed.


Abstract

1. The effects of microhabitat and plant form on sunlight interception, leaf temperatures, frost occurrence and photosynthesis were evaluated for Caltha leptosepala DC and Erythronium grandiflorum Pursh. Both plants are perennials that commonly emerge from alpine snowbanks where the combination of cool temperatures and strong sunlight is among the most extreme for vascular plants.

2.  Caltha leptosepala occurred in microsites where colder air accumulates, and has larger, less inclined and more densely clustered leaves compared to E. grandiflorum (which has two steeply inclined leaves).

3. These differences in microsite and plant form led to leaf temperatures below 0 °C on 70% of nights during the summer growth season in C. leptosepala, compared to only 38% in E. grandiflorum. Leaves of C. leptosepala warmed more slowly on mornings following frosts compared to E. grandiflorum, due to less aerodynamic coupling between leaf and air temperature, and also a 45% smaller ratio of sunlit to total leaf area due to mutual shading among leaves.

4. As a result, night frost did not affect subsequent CO2 assimilation (A) in E. grandiflorum, while frostless nights and warmer mornings led to 35% greater A in C. leptosepala in the early morning.

5. There were no appreciable differences in the temperature and light response of photosynthesis between the two species. The apparent quantum yield of A declined only ≈8% in both species following frost and exposure to strong sunlight, indicating little adjustment of photosynthetic physiology.

6. Greater daily carbon gain probably occurs for E. grandiflorum because of its plant form and microclimate, rather than differences in photosynthetic physiology.

Introduction

In many temperate habitats, a significant portion of the growing season has cool nights followed by warm, bright days with optimum temperatures for carbon gain. Throughout the growth period at high elevations, cloudless nights commonly lead to frosts followed by clear, bright mornings (Jordan & Smith 1995). While sunlight interception is important for photosynthesis and leaf warming, it also can lead to photoinhibition, the depression of photosynthesis due to excessive sunlight (reviewed by Long, Humphries & Falkowski 1994). This combination of cool temperatures (at day or night) and strong irradiance is becoming recognized as important to the carbon gain and ecology of plants in cold regions (Ball, Hodges & Laughlin 1991; Germino & Smith 1999; Germino & Smith 2000b; Egerton et al. 2000).

Leaf size, orientation and arrangement on the plant can alter day- and night-time leaf temperatures, the occurrence of frost and the sunlight incident upon leaves (Jordan & Smith 1995; Körner 1982; Leuning & Cremer 1988), but few studies have evaluated the influence of plant form and microhabitat on carbon gain under natural conditions of night-time frost followed by strong irradiance (Germino & Smith 1999; King 1997). For example, the importance of an upright leaf orientation for minimizing photoinhibition at midday in plants experiencing water or heat stress has been recognized (Valladares & Pearcy 1997), but such an inclined orientation may also influence the occurrence of nocturnal frost in combination with strong irradiance the following morning.

Our objective was to determine whether photosynthetic responses to the combination of low temperature and strong irradiance were attributable to differences in microhabitat, plant form or photosynthetic physiology in the alpine perennial species Caltha leptosepala and Erythronium grandiflorum. These species commonly emerge directly from snowbanks, and may be surrounded by snow during their entire summer growth period (Rochow 1969; Hamerlynck & Smith 1994). Sunlight reflected from snow can increase incident sunlight on a leaf by 400–1400 µmol m−2 s−1 (photon flux density, PFD, 400–700 nm; Day, DeLucia & Smith 1989; Körner 1999), adding to the strong PFD that results from a thin atmosphere at high elevation and periods of reflected sunlight from the edges of cumulus clouds (Smith & Geller 1979).

Despite their spatial proximity and similar life histories, these two species occur in different microsites and have contrasting leaf, crown and canopy structures. Caltha leptosepala occurs along downhill margins of snowbanks, where colder air and snowmelt accumulate, and has larger, broader, less inclined, and more densely packed leaves than E. grandiflorum. We hypothesized that these differences in microhabitat and plant form between the two species would have a strong influence on the occurrence of night-time frost, incident sunlight the following morning, and photosynthetic carbon gain throughout the summer growth period.

Materials and methods

The separate effects of microhabitat, plant form and photosynthetic physiology were evaluated in situ for Caltha leptosepala and Erythronium grandiflorum. Effects of leaf and crown architecture on night-time temperature, and incident sunlight and temperature in the morning, were evaluated in combination with photosynthetic responses to PFD, leaf temperature, and previous night frost. Requirements for parametric tests were satisfied for all statistics in this study, and all values reported here are means ± 1 SE (Zar 1999; jmp version 3·2·2, SAS Institute, Cary, NC, USA).

Site and species description

All study sites were located in the Snowy Range of the Medicine Bow National Forest in south-east Wyoming, USA (N 41°22′25″, E 106°14′00″) at ≈3250 m above sea level. Plants were examined at five locations within a 2 km radius that had snowbanks persisting for most of the growing season. These sites were characterized by east-facing, leeward slopes that collect drifting snow from the prevailing westerly winds (Billings & Bliss 1959).

Caltha leptosepala DC (marsh marigold) and Erythronium grandiflorum Pursh (glacier lily) are herbaceous perennials that frequently begin flowering even before the snow has melted from the base of the plant (Hamerlynck & Smith 1994). The boundaries of all populations of C. leptosepala and E. grandiflorum within the study area were identified and mapped. The centres of populations of E. grandiflorum were a mean 2·6 ± 0·6 m higher in elevation than adjacent populations of C. leptosepala at the snowbank sites examined (n = 12 snowbank sites having both species, Table 1, paired t-test, P < 0·0001). Erythronium grandiflorum occurred along the upper edges of snowbanks, adjacent to, but not overlapping with, C. leptosepala, which generally occurred downslope of snowbanks in topographic depressions.

Table 1.  Leaf, crown and microsite differences in Erythronium grandiflorum and Caltha leptosepala
SpeciesWidth (cm)Length/widthFolding ratio% angles >60°STAR (morning/midday)Elevation (m)
  1. ‘Width’ is for unfolded leaves; ‘Folding ratio’ is the ratio of the width of the leaf in its natural, folded state, and the width of the leaf when unfolded. ‘% angles >60°’ is the proportion of leaf angles greater than 60° from horizontal. STAR is the ratio of silhouette to total leaf area measured at morning or midday (≈09·00 or 12·00 h, respectively). Values are means ± SE. n = 77 for E. grandiflorum and 56 for C. leptosepala, except for STAR (n = 7 for E. grandiflorum;n = 4 for C. leptosepala) and elevation (n = 12). All values were statistically different between the two species (P < 0·001, two-tailed t-tests, paired t-test for elevation).

E. grandiflorum2·5 ± 0·144·9 ± 0·20·5 ± 0·060·570·18 ± 0·02/0·15 ± 0·023215 ± 8
C. leptosepala6·6 ± 0·241·2 ± 0·030·8 ± 0·020·380·10 ± 0·02/0·13 ± 0·023218 ± 8

Erythronium grandiflorum is acaulescent, with two leaves arranged oppositely, while individuals of C. leptosepala typically have three to six leaves and occur most often in dense clusters of about five or more individuals, with >20 leaves per cluster. In addition to being more linear and folded (adaxially), leaves of E. grandiflorum had much steeper orientations than those of C. leptosepala (Table 1). Moreover, leaves of E. grandiflorum that did have shallow angles of inclination from horizontal (<45°) were 35% more folded than leaves with steeper orientations (two-tailed t-test, P < 0·05; folded/unfolded width = 0·34 ± 0·023 for leaves with a mean inclination of 30 ± 1·8°, compared to 0·52 ± 0·02 for leaves inclined 64 ± 1·8°, n = 30). Leaf lamina were oriented closer to perpendicular and were more widely spaced in E. grandiflorum than the more horizontal and clustered leaves of C. leptosepala.

Leaf and crown architecture

The ratio of silhouette leaf area (the shadow cast on a surface perpendicular to the sun’s rays) to total leaf area (STAR, Carter & Smith 1985) was used to evaluate sunlight interception efficiency for whole-plant crowns. STAR was measured twice during the day, at ≈09·00 and 12·00 h, to represent morning and midday conditions, respectively. The maximum value possible for STAR is 0·5 for a thin, flat leaf oriented perpendicular to the sun. To quantify STAR values for entire crowns, a digital camera (model Mavica FD91, Sony, Tokyo, Japan) was positioned above a plant at the position of the sun. A photograph was taken of the plant with calibration objects of known size (e.g. 1 cm2) positioned throughout the depth of view to standardize all measurements. In addition, sampled leaves were detached, laid flat and photographed to measure the total projected leaf area (approximately half of the total leaf area for these thin leaves). All digital photographs were quantified using image-processing software (scion image for Windows, Scion Corp., Frederick, Maryland, USA). Individual plants of E. grandiflorum, and clusters of C. leptosepala individuals (five to 12 plants) were measured for each replicate.

Leaf temperatures and incident sunlight

To evaluate the effects of night-time sky exposure and incident irradiance on leaf temperatures, diurnal measurements of leaf temperature and PFD were recorded at 15 s intervals and averaged over 10, 15 or 30 min throughout the 1997 growing season (9 June to 16 October) with dataloggers (model 21x, Campbell Scientific, Logan, Utah, USA) at three sites. Additionally, two plastic-coated, paperboard models of typical leaves of each species were used to evaluate the frequency distribution of the minimum night-time leaf temperatures (Tmin) for each diurnal period throughout the growth season. The leaf models were positioned within populations of each species to represent natural orientations, and were necessary for long-term microclimate measurements due to leaf deterioration and the likelihood that thermocouples on real leaves would eventually become detached. Correction factors were applied to the minimum temperatures for each leaf model for each 24 h period throughout the growth season in order to estimate the minimum temperatures of actual leaves. These correction factors were determined by field-calibrating the leaf models with actual leaves using means recorded every 15 min for three leaves over five nights (r2 = 0·99 for correlation of leaf and model temperatures).

Leaf temperatures were measured with fine-wire thermocouples (0·05 mm, Omega Engineering Inc., Stamford, CT, USA) threaded into leaf mid-veins (single stitch). Air temperatures at leaf level were measured with identical thermocouples shielded from the sky to avoid radiation error. The difference between leaf and air temperature (ΔT) was determined by subtracting the temperature of the air (Tair) from leaf temperature (Tleaf). Tair was measured either adjacent to the plant at leaf height (outside the boundary layer of the leaf) or at 1 m height (outside the leaf and plant boundary layer). With rare exceptions, leaf temperatures at night were below air temperature and ΔT was negative regardless of the sign of Tleaf or Tair. ΔTmin indicates the greatest depression of leaf temperature below air temperature and the most negative value of ΔT for a diurnal period. Incident sunlight at each site was measured with PFD sensors that were positioned horizontally within 1 m of the measured plants (model Q-190, LiCOR, Lincoln, NE, USA).

The variation in incident sunlight and leaf temperature within plants of E. grandiflorum and clusters of C. leptosepala was measured, beginning before sunrise and continuing through morning. The amount of leaf area that was oriented within ≈30° of perpendicular to the sun’s rays (Perpendicular); 30–90° of perpendicular (Oblique); or shaded from sunlight by other leaves (Shaded), was traced and quantified using the image analysis program described above. Incident sunlight and leaf temperatures were then measured for each category of leaf orientation to the sun using a PFD sensor and an infrared thermometer (model PM Plus, Raytek, Santa Cruz, CA, USA) with a field-calibrated accuracy of ±0·3 °C based on direct comparisons with thermocouple measurements (r2 = 0·98, n = 20).

Effect of leaf orientation and arrangement on incident sunlight and temperature

The effect of leaf orientation on night-time temperatures was calculated by subtracting the mean temperature of selected leaves placed horizontally from the mean temperature of adjacent, naturally oriented leaves. Leaves were oriented using one or two paperclips on individual leaf blades. Two to four horizontal and natural leaves of each species were sampled on seven different nights. Leaf temperatures throughout each of the seven nights were measured with dataloggers and thermocouples as described above.

The separate effects of natural leaf folding and inclination on leaf temperatures were also evaluated on one clear, still night for 28 individuals of E. grandiflorum. One leaf per individual was forced into either a horizontal orientation with natural folding, or a horizontal and unfolded condition, and the orientation of the other leaf was left natural to provide a control. Instantaneous temperatures were measured for ≈3 cm2 of each leaf through the night, using an infrared thermometer.

Photosynthesis in the field

Photosynthesis was measured periodically throughout the summer growth period, on the adaxial face of individual, intact leaves that were oriented perpendicular to the sun. The response of photosynthetic CO2 exchange to nights with and without frost (Tmin of leaves <0 °C or >0 °C, respectively) was measured under natural sunlight and temperatures. Before measurements, any frost or dewfall was carefully removed from the leaf surface with tissue paper, and the temperature of the leaf was measured with the infrared thermometer. The responses of photosynthetic CO2 assimilation (A) to leaf temperature and incident sunlight were used to compare photosynthesis between the species.

A was measured with a portable, closed-flow photosynthesis system (model 6200, LiCOR) at ambient CO2 concentrations of ≈365 µl l−1. All measurements of A were for light-saturated photosynthesis, except when indicated (see Results), and were expressed on a projected leaf area basis, which was effectively silhouette leaf area because sample leaves were oriented perpendicular to the sun (Smith, Schoettle & Cui 1991). Tracings of leaf perimeters were digitized either manually or with a scanner (ScanJet ADF, Hewlett Packard, Palo Alto, CA, USA), and leaf area was obtained using the image analysis program. The temperature response of A was determined using numerous measurements collected with the closed-flow system on 11 days for C. leptosepala and nine for E. grandiflorum from ≈2 h after sunrise to solar noon. The temperature–response curves were then fitted to second-order polynomial equations using least-squares minimization, and the statistical significance of the fit was determined using regression.

An open-flow photosynthesis system (model 6400, LiCOR) equipped with a CO2 controller (CO2 concentrations always 365 µl l−1) was used to measure light–response curves under constant temperature and humidity. Light–response curves were obtained using a red–blue LED lamp, and the apparent quantum yield (φ; mol CO2 mol−1 incident photons) was estimated by regression analysis of A at PFD <100 µmol m−2 s−1.

The effect of differences in STAR on whole-plant photosynthesis was estimated for one point in time, using the mean air temperature and PFD of the growth season at 09·00 h. The incident PFD and corresponding Tleaf for each leaf surface in the crown was determined as described above. A was then predicted for each leaf surface from the relationships between A, PFD and Tleaf, then summed for all leaves to give an estimate of A for the whole plant. The contribution of interspecific differences in A under similar conditions, incident PFD, and Tleaf to whole-plant A were determined separately by comparing the percentage change in interspecific difference in whole-plant A if each variable was held constant.

Results

Leaf temperatures

Minimum night-time leaf temperatures were below 0 °C on 70% of 106 nights examined for C. leptosepala, compared to 38% of nights for E. grandiflorum (Fig. 1). In both species, minimum temperatures were within 2 °C of 0 °C on ≈40% of all nights sampled. Leaves of C. leptosepala were up to 15 °C warmer than air temperature (measured at 1 m above ground) during the day, and up to 8 °C cooler than air at night, while Tleaf in E. grandiflorum was typically within ≈2–4 °C of Tair during the day and night (Fig. 2). At time = 04·00 h in Fig. 3, Tair was ≈−1·5 °C at leaf height and adjacent to the crown of C. leptosepala, compared to 3 °C near leaves of E. grandiflorum. The Tmin of each day occurred most often before dawn in both species, although Tmin occurred after sunrise on 32% of days from 26 August to 7 October 1997 in C. leptosepala (not shown). During this period, Tleaf increased more rapidly above air temperature in E. grandiflorum than in C. leptosepala (Fig. 3). In C. leptosepala, the greatest depression in Tleaf below TairTmin; Tair measured at leaf height) occurred after sunrise on 36% of days (not shown), when dew was evaporating from the leaf surface and Tair near leaves rose faster than Tleaf (Fig. 3).

Figure 1.

Cumulative frequency distribution of minimum temperatures during each diurnal period throughout the growing season (106 nights) for leaves of Caltha leptosepala and Erythronium grandiflorum. Values are based on daily minima of means recorded every 15 min from 9 June to 16 October 1997. Measurements were made on leaf models in the field and corrected to estimate actual leaf temperatures; n = 2 leaves per night.

Figure 2.

Difference between leaf and air temperature (Tleaf −Tair; Tair measured at 1 m height) for leaves of Caltha leptosepala (dashed) and Erythronium grandiflorum (solid). Data are means of 15 s samples recorded every 30 min from 6 to 16 September 1997. n = 2.

Figure 5.

Photosynthetic temperature (a) and light response (b) of Caltha leptosepala (solid symbols and lines) and Erythronium grandiflorum (open symbols, dashed lines). (a) Leaf temperatures were the average of temperatures prior to and during the measurement, and all measurements of photosynthesis were at PFD > 1500 µmol m−2 s−1. Curves are second-order polynomials, A = −0·023Tleaf2 + 1·052Tleaf + 0·067 for C. leptosepala;A = −0·029Tleaf2 + 1·247Tleaf + 0·251 for E. grandiflorum;r2 < 0·25 and P < 0·0001 for both (n = 127 for C. leptosepala and n = 69 for E. grandiflorum). (b) Data were collected for both sun and shade leaves, values are means ± SE; n = 6.

Figure 3.

(a) Photon flux density (PFD); (b) temperatures of leaves (solid lines) and air at 1 m height (dashed line); (c) difference between leaf and air temperature adjacent to the plant (ΔT = Tleaf − Tair) for representative leaves of Caltha leptosepala (solid) and Erythronium grandiflorum (open). In (a), PFD was measured on a horizontal plane (solid), as well as perpendicular to the solar beam to represent the maximum incident values (dashed). The approximate times of frost and dew occurrence are indicated in (c). Data are means of 15 s samples recorded every 10 min from a typical clear, still day (25 September 1997); n = 2.

Natural leaf inclination from horizontal resulted in a 54% greater increase in minimum leaf temperatures in E. grandiflorum than in C. leptosepala on nights with slow wind speeds.Mean Tmin for natural leaves of E. grandiflorum was 1·6 °C higher than horizontally positioned leaves (0·9 ± 0·8 °C for natural leaves; −0·7 ± 0·7 °C for horizontal leaves, two-tailed t-test, P < 0·001, n = 15). In C. leptosepala mean Tmin for natural leaves was ≈1·0 °C higher compared to horizontal leaves (1·7 ± 1·2 °C for natural leaves; 0·6 ± 1·1 °C for horizontal leaves, t-test, P < 0·01, n = 9).

In a separate experiment using E. grandiflorum only, natural leaf orientation led to 2·1 ± 0·1 °C higher Tleaf compared to unfolded and horizontal leaves (P < 0·001). The separate effects of inclination and folding each contributed ≈1 °C to higher Tleaf (instantaneous values measured on one night; paired t-tests, P < 0·001, n = 28). The maximum difference in Tleaf between leaves oriented naturally and experimental leaves that were positioned horizontally and unfolded was 5 °C.

The ratio of silhouette to total leaf area (STAR, Table 1) after sunrise (≈09·00 h) was 80% greater in E. grandiflorum (0·18 ± 0·02) compared to C. leptosepala (0·10 ± 0·02), indicating greater sunlight interception per unit leaf area in E. grandiflorum in the morning (two-tailed t-test, P < 0·05). The proportion of leaf area oriented within 30° of perpendicular to the sun after sunrise was ≈50% greater, and the proportion of leaf area in the shade of other leaves was sixfold smaller in E. grandiflorum than in C. leptosepala (F1,10; P = 0·06 and P < 0·001, respectively; Fig. 4). As a result, leaf areas of E. grandiflorum oriented within 30° of perpendicular to the sun were ≈2–4 °C warmer than leaf areas with an oblique or shaded orientation to the sun following sunrise (Tleaf − Tair < 1 °C for shaded leaves; Fig. 4). Leaf areas of C. leptosepala with a perpendicular orientation to the sun were >6 °C warmer than leaf areas with an oblique orientation, and 9 °C warmer than shaded leaf areas, which were >2 °C below Tair after sunrise. Based on Fig. 4, the computed mean Tleaf of all leaf areas in E. grandiflorum was 2·8 °C warmer than in C. leptosepala after sunrise (14·2 versus 11·6 °C, respectively) although the range in Tleaf at this time was considerably broader in C. leptosepala.

Figure 4.

Mean proportions and temperatures of leaf area oriented perpendicular or oblique to the sun, or in the shade of other leaves after sunrise (09·00–09·30 h), for Caltha leptosepala (hatched) and Erythronium grandiflorum (open). ‘Perpendicular’ leaf areas were oriented within ≈30° of perpendicular to the sun (PFD ≈ 1600–1900 µmol m−2 s−1); ‘oblique’ leaf areas were oriented ≈ 30–90° from the sun (PFD ≈ 200–1600 µmol m−2 s−1), and ‘shaded’ leaf areas were shaded by other leaves on the plant (or neighbouring plants in the same cluster of C. leptosepala, PFD < 150 µmol m−2 s−1). The shaded portions of bars indicate abaxial leaf area, for leaf area with greater PFD incident on the abaxial compared to the adaxial leaf surface. anova (F1,10) results for significant differences between the species for each category of incident sunlight are shown as *** for P < 0·001, ** for P < 0·01, and * for P < 0·1. Values are means ± SE (SE for the combined abaxial and adaxial values); n = 4 for C. leptosepala and n = 7 for E. grandiflorum.

Photosynthesis

The photosynthetic light–response curves measured from 10·00 to 12·00 h for each species were similar (14·4 compared to 12·5 µmol CO2 m−2 s−1 for C. leptosepala; two-tailed t-test, P = 0·089; Fig. 5). The PFD required for 85% saturation of photosynthesis (895 ± 77 µmol m−2 s−1 in C. leptosepala; 1000 ± 78 µmol m−2 s−1 in E. grandiflorum), the initial slope (φ), sunlight compensation points (all two-tailed t-tests, P > 0·3), and temperature response of A (Fig. 5) were similar between the species. In C. leptosepala, the maximum A (12·1 µmol CO2 m−2 s−1) was observed at Tleaf = 23 °C, and the Tleaf required for 85% of maximum photosynthesis was between 14 and 31·7 °C. In E. grandiflorum, the maximum A (13·7 µmol CO2 m−2 s−1) occurred at Tleaf = 21·5 °C, and the Tleaf required for 85% of maximum photosynthesis was between 13·2 and 29·5 °C.

Frostless nights (Tmin of leaves >0 °C) led to 35% greater CO2 assimilation (A) for ≈2 h following sunrise (Fig. 6) in C. leptosepala. In frosted plants of C. leptosepala (Tmin ≈ −3·5 °C), the recovery of A was preceded by the warming of their leaves towards temperatures measured for non-frosted leaves (for difference in Tleaf in early morning, two-tailed t-test, P < 0·05; Fig. 6). Photosynthesis in E. grandiflorum did not appear to be significantly influenced by night-time frost (Fig. 6). These findings were supported for both species by data collected on 13 other days with and without frost (n = 20, t-tests, P < 0·0001 for C. leptosepala; not shown).

Figure 6.

CO2 assimilation, photon flux density (PFD), square symbols, same for both species), and leaf temperature (Tleaf) for Caltha leptosepala and Erythronium grandiflorum following frost (solid) or non-frost nights (open). Leaf temperatures were measured before enclosing the leaf in the gas-exchange cuvette. Photosynthetic measurements were made following a frost night with minimum temperatures of ≈−3·5 °C for C. leptosepala and −1·5 °C for E. grandiflorum, and following frostless nights with minimum temperatures above 2 °C. Microsite PFD values for C. leptosepala were the same as shown for E. grandiflorum;n = 3–6.

By solar noon following a light frost, leaves in both species that had been oriented perpendicular to the beam of the sun had ≈8% less apparent quantum yield (φ) than leaves that were shaded from full sun by other vegetation or oriented away from the sun (n = 3; two-tailed t-tests, P < 0·05; for C. leptosepala, 0·054 ± 0·002 in strong PFD, 0·058 ± 0·006 in weak PFD; for E. grandiflorum, 0·056 ± 0·003 in strong PFD, 0·061 ± 0·001 in weak PFD; see Fig. 4 for approximate PFD data for different orientations).

Discussion

Plant form has previously been shown to influence daytime microclimate, photosynthesis and transpiration at high elevation (Körner 1982; Smith & Carter 1988; Smith & Geller 1979). In the current study, interspecific variation in microhabitat and plant form led to differences in daytime and night-time microclimate, and differences in the response of CO2 assimilation to nocturnal frost followed by strong sunlight. In contrast, differences in photosynthetic physiology were minimal between the species.

Cool leaves and strong sunlight

Although these two species typically grow within metres of one another, C. leptosepala experienced nearly twice as many frost nights (Fig. 1) and had colder leaves in the morning than E. grandiflorum (Fig. 3), due to differences in Tair between their microsites (cold-air drainage; Table 1) and differences in Tleaf due to plant form. In Fig. 3, the 7·5 °C difference in Tleaf between the species at night was due to a ≈4·5 °C difference in Tair between their microsites, and a ≈3 °C difference in Tleaf due to plant form.

As reported for species of Eucalyptus in a previous study (Leuning & Cremer 1988), natural leaf inclination resulted in 1–2 °C warmer night-time temperatures in C. leptosepala and E. grandiflorum. Warmer nocturnal temperatures for upright leaves may be due to decreased sky exposure and less thermal re-radiation to the cold night sky, plus an approximately twofold greater coefficient of free convective heat exchange from upright versus horizontal surfaces (Campbell & Norman 1997). In addition, a more upright, vertical orientation places the leaf into warmer portions of the air-temperature profile above the ground at night (for night-time temperature profile above ground see Jordan & Smith 1995).

In E. grandiflorum, the 1·6 °C warming that resulted from natural leaf inclination (compared to experimental leaves held horizontally) could alone decrease the frequency of nights when leaves have frost by 20% (based on Fig. 1). Similarly, C. leptosepala leaves were also ≈1 °C cooler at night when held horizontally, even though the average leaf orientation is already more horizontal in C. leptosepala (Table 1). Leaves of C. leptosepala have relatively less convective heat exchange, and a greater sensitivity of leaf temperatures to radiation exchange (Fig. 2). Moreover, denser leaf clustering in the crown and canopy of C. leptosepala decreases convection for individual leaves (see Smith & Carter 1988 for conifer branches), leading to cooler leaves at night. The clustering of large and horizontal leaves in C. leptosepala led to greater frosting of its leaves, more mutual shading, and cooler temperatures during mornings following frosts compared to the upright and widely spaced leaves of E. grandiflorum.

Frost accumulation may cause 1–2 °C higher leaf temperatures at night (Leuning 1988). However, the evaporation of dewfall (accumulated during nights as frost) appeared to cause substantial latent heat loss from the leaf during the mornings, especially for C. leptosepala (Fig. 3). In addition to radiative cooling, this latent heat loss appeared to prevent leaf warming in C. leptosepala, even while the sun was rising and air temperatures near leaves were increasing. A large fraction of the total leaf area in C. leptosepala remained shaded from the sun, was colder than surrounding air and often remained frosted until after sunrise (Figs 3 and 4). In contrast, 80% greater sunlight interception per unit leaf area and fivefold less mutual shading contributed to warmer leaves in the morning for E. grandiflorum, although the fewer sunlit leaves of C. leptosepala became warmer (after drying) than sunlit leaves of E. grandiflorum (Fig. 4). Significantly, air temperatures at leaf height were still cooler than those required for 85% of maximum photosynthesis 3 h after sunrise for both species on 50% of the days represented in Fig. 1. Photosynthetic responses to frost in C. leptosepala may be influenced not only by the history of leaf temperatures below 0 °C at night, but also by cool temperatures during sunrise. While direct sunlight interception is important for leaf warming in these plants, especially on mornings following frost, it may lead to photoinhibition, as shown for other species (Ball et al. 1991; Germino & Smith 1999). Moreover, dewfall and frost may also limit CO2 diffusion into leaves (Smith & McClean 1989), reducing A and possibly exacerbating photoinhibition as well as significantly influencing leaf temperature.

Photosynthetic responses to sunlight and temperature

Only subtle differences were observed in temperature and sunlight-response curves between the species (Fig. 5), despite substantially more mutual shading (smaller STAR values, Table 1) and a greater range of leaf temperature extremes experienced regularly by C. leptosepala (Figs 2 and 4). The temperature required for maximum A was only ≈1·5 °C higher, and the estimated range of temperatures required for 85% of maximum A was only 1·5 °C (10%) broader for C. leptosepala than for E. grandiflorum. The maximum A was only ≈15% smaller for C. leptosepala than E. grandiflorum, and the PFD required for light saturation, φ and light-compensation point were all similar between the species. Photosynthetic gas-exchange char-acteristics were more similar between the species than were their respective amounts of incident sunlight and range of Tleaf experienced daily.

Photosynthesis following frost

A appeared at least as resistant as other alpine plants to frost (Körner 1999), though a small effect of frost on subsequent A was detected in C. leptosepala (Fig. 6). Differences in frost occurrence and subsequent Tleaf between C. leptosepala and E. grandiflorum were important for daytime carbon gain. Greater physiological acclimation to frost was expected for C. leptosepala because of fewer structural or microsite advantages for frost avoidance, but A increased more in C. leptosepala than in E. grandiflorum following frostless (versus frost) nights. Warmer temperatures after sunrise in C. leptosepala (as in Fig. 3) may have been the more direct cause of greater A following frostless nights, as opposed to warmer night-time temperatures per se (Fig. 6). Warmer leaves in the morning (resulting from an upright orientation and greater sunlight interception efficiency; Fig. 4) may have also contributed to the relatively smaller impact of frost on subsequent A in E. grandiflorum (Fig. 6). If C. leptosepala had the same plant form and microsite features that lead to warmer nights for E. grandiflorum, an estimated 7% more daily CO2 uptake per unit leaf area might occur on 32% more days that would be preceded by a frostless night. This estimate is based on 35% greater A for ≈2 h of a 10 h solar period (Fig. 6), and considering only PFD greater than saturating, and not differing degrees of frost severity.

Photosynthesis in high sunlight following frost

Previous work suggested that reduced sunlight interception – due to mutual shading of leaves, leaf inclination and cloud cover – was most limiting to carbon gain during the snow-free season for several herbaceous, alpine plants (Körner 1982; Körner 1999). Also, a greater resistance to strong sunlight and photoinhibition has been inferred for alpine (as opposed to lowland) plants that have biochemical adaptations for tolerating excess sunlight at low temperatures (Fetene et al. 1997; Lütz 1996; Manuel et al. 1999; Streb et al. 1997; Streb et al. 1998; Wildi & Lütz 1996). In the current study, A declined in C. leptosepala and E. grandiflorum as PFD reached ≈85% of maximum daily flux densities, regardless of previous night temperature. However, the maximum reductions in φ for both species following exposure to strong sunlight were small (<8%), especially compared to the two- to tenfold greater reductions in φ reported for a variety of chilled C3 crops (Bongi & Long 1987). Maximum declines in Fv/Fm (the ratio of variable to maximum fluorescence for dark-adapted leaves) for C. leptosepala and E. grandiflorum were <8% due to frost and exposure to strong sunlight, and were reversible within the diurnal period (Germino & Smith 2000a). Similarly to φ, Fv/Fm is an indicator of photochemical efficiency, and lower values indicate photoinhibition (reviewed by Krause 1994). Caltha leptosepala and E. grandiflorum may be particularly resistant to frost-induced photoinhibition, and the warming and photosynthesis resulting from greater sunlight interception provided greater benefit to photosynthesis than the inhibition caused by high PFD.

Greater sunlight interception per unit leaf area might lead to an estimated twofold greater CO2 assimilation per unit leaf area at the whole-plant level in E. grandiflorum compared to C. leptosepala following sunrise (based on Figs 4–6, clear sky conditions, and Tair = 11 °C). Of this twofold difference in A, ≈15% is attributable to greater A at similar PFD in E. grandiflorum; 75% to insufficient sunlight for photosynthesis as a result of mutual shading in C. leptosepala; and 10% is attributable to secondary effects of sunlight on leaf temperature (the latter fraction would become substantially larger at early morning hours and in cooler air). At midday, photoinhibition might lead to a maximum of 8% less A in leaf surfaces that intercept full sunlight during the morning (i.e. upright and east-facing orientation, low PFD at midday). While this slight photoinhibitory effect would be expected to influence carbon gain more in E. grandiflorum, C. leptosepala would still have a substantially smaller carbon gain per unit leaf area as a result of mutual shading (Fig. 4). Future research should evaluate the importance of contrasting night-time and daytime microclimate on respiration and carbon allocation. Combined with leaf area to leaf mass relationships, these data would enable an extension of the effects of plant form and microhabitat on CO2 assimilation to whole-plant carbon balance.

Conclusions

Measured leaf temperatures and sunlight regimes varied considerably according to differences in microhabitat and plant form for C. leptosepala and E. grandiflorum, two species that can experience an extraordinary combination of low temperatures and exposure to strong sunlight. Their photosynthetic responses to light and temperature were similar. Differences in daily CO2 uptake between these two species were due largely to differences in plant form and microhabitat, not to photo-synthetic physiology. The clustering of large leaves contributed to cooler leaf temperatures and less carbon gain in C. leptosepala. In contrast, the clustering of slender needles in conifers (Picea and Abies spp.) at high elevation functions to avoid strong sunlight while still warming leaves to more optimal temperatures for photosynthesis (Germino & Smith 1999; Germino & Smith 2000b; Smith & Carter 1988). These differences in conifer needles and herbaceous broadleaves partly explain the functional, adaptive benefits of such dramatic variation in plant form.

Acknowledgements

Support for this study was provided by an NSF Ecological and Evolutionary Physiology grant awarded to W.K.S. and a NASA fellowship administered by the Wyoming Space Grant Consortium to M.J.G. We thank Matt Lokken for his valuable assistance.

Received 30 June 2000; accepted 10 November 2000

Ancillary