Movement of Water from Old to Young Leaves in Three Species of Succulents

Abstract

A hypothetical adaptive response of succulent plants to drought‐stress is the redistribution of water from old to young leaves. We examined the effects of possible movement of water from old to young leaves in three succulent species, Carpobrotus edulis (weak CAM‐inducible), Kalanchoe tubiflora (CAM) and Sedum spectabile (possibly a CAM‐cycler or CAM‐inducible). Old leaves were removed from plants, and photosynthesis, transpiration, f. wt : d. wt ratios, diurnal acid fluctuations, stomatal conductance and internal CO₂ concentrations of the remaining young leaves were measured during drought‐stress. Comparison was made with plants retaining old leaves. There was no evidence that water moved from old to young leaves during drought‐stress as previously hypothesized. Only in drought‐stressed plants of K. tubiflora, were photosynthetic and transpiration rates of young leaves greater on shoots with old leaves removed compared with attached. There was a trend in all species for greater fluctuations in acidity in young leaves on shoots that lacked older leaves. For two of the three species studied, the f. wt : d. wt ratios of young leaves were greater under drought‐stress, on shoots with old leaves removed than with them attached. Absence of old leaves may reduce competition for water with young leaves, which consequently have higher water content and greater photosynthetic rates.

INTRODUCTION

Water is a vital resource for plant growth and survival and is often limiting, especially in arid regions. For this reason, many plants have adaptations that conserve water, such as thick cuticles, pubescence that creates a thicker boundary layer on the leaf surface, sunken stomata, and succulence (large amounts of water‐storage tissue). Succulent plants frequently assimilate CO₂ via Crassulacean acid metabolism (CAM), in addition to the C₃ mechanism. CAM photosynthesis is the assimilation of CO₂ primarily at night, when stomata are open, in contrast to C₃ plants, which assimilate CO₂ in the light when their stomata are open. Nocturnal stomatal opening, which substantially decreases transpirational losses as a result of the cooler temperatures and greater atmospheric humidity at night than during the day, conserves water in the plant (Kramer, 1983; Taiz and Zeiger, 1998; Chandler and Bartels, 1999).

Another possible adaptive response of plants to drought‐stress is the redistribution of water from old to young leaves. This is potentially of great importance in succulent plants, where the volume of water in tissues is considerable, and the slower rates of water loss may allow water from old leaves to maintain the water content of young leaves, and thereby leaf function. This water shuttling hypothesis has been frequently mentioned (Evans, 1932; Kramer, 1959; Barcikowski and Nobel, 1984; Schäfer and Lüttge, 1987; Schmidt and Kaiser, 1987; Schulte and Nobel, 1989; Tüffers et al., 1995; Herrera et al., 2000), but the few studies that address the phenomenon are problematic and of questionable ecological significance.

Evans (1932) focused attention on succulent plants, with their large amounts of water storage tissue, and considered that older leaves were capable of supplying younger leaves with water during periods of water shortage. Paul J. Kramer (1959) described this phenomenon: ‘Whenever water deficits occur in plants, competition for water develops among various organs and tissues . . .. Young leaves usually can obtain water at the expense of older leaves and the latter usually die first during drought.’

Schäfer and Lüttge (1987) suggested that younger leaves of succulents are more sensitive to drought than older leaves and that water translocation from old to young leaves compensates for transpirational water loss from the young leaves during drought‐stress. Although a difference in water content between the different‐aged leaves was not always observed in well‐watered plants of two species of succulents, it was more commonly observed in plants under drought‐stress (Tüffers et al., 1995).

Several studies have suggested that movement of water within a leaf, from tissue‐to‐tissue, occurs during drought‐stress. This preferential loss of water from hydrenchyma to the chlorenchyma maintains activity of the photosynthetic tissue during drought‐stress (Barcikowski and Nobel, 1984; Schulte and Nobel, 1989; Herrera et al., 2000). It has been hypothesized that this water movement is facilitated by an osmotic gradient in which the hydrenchyma, with a low solute concentration, donates water to the chlorenchyma tissue, which has a greater solute concentration.

Some studies of leaf‐to‐leaf water movement are difficult to relate to an ecological context. For example, shoots were detached from the roots in studies of several non‐succulents (Čatský, 1962; Milburn, 1979) or, measurements were made on young succulent shoots detached from the roots (von Willert et al., 1992; Tüffers et al., 1995)—an artificial situation. In succulent CAM plants, gas exchange (CO₂ and water vapour uptake and loss) needs to be measured throughout the whole day–night cycle in order to analyse the effects of leaf removal on leaf water status; it is also important to evaluate 24‐h gas exchange to account for the variety and timing of photosynthetic pathways in succulents.

The general goal of this investigation was to determine if water is translocated from older leaves to the younger leaves during drought‐stress in three species of succulents, Kalanchoe tubiflora, Carpobrotus edulis and Sedum spectabile. The species were selected for three reasons. First, all have a large water‐storage capacity in thick leaves. Secondly, the old and young leaves are arranged on a single shoot, aiding measurement of gas exchange of the young leaves. Thirdly, they have different photosynthetic pathways. Kalanchoe tubiflora has been shown to have a typical CAM photosynthetic pathway (Stidham et al., 1983). Carpobrotus edulis has exhibited inducible CAM with drought‐stress (Earnshaw et al., 1987) and when subjected to high salt concentrations (Winter, 1973). The photosynthetic pathway of S. spectabile is unknown; however, some droughted Sedum spp. are known CAM‐cyclers, with diel acid fluctuations but with most of the CO₂ uptake occurring during the day (Borland, 1996), and some exhibit inducible CAM (Winter and Lüttge, 1976; Kluge, 1977).

The specific goal was to test the hypothesis that water is transferred from old to young leaves of succulents. To do so, the physiological consequences of water‐shuttling were investigated by measuring the photosynthetic rate (CO₂ uptake) and transpirational water loss of the young leaves, and water content and diel acid fluctuations of old and young leaves, in order to determine how the presence or absence of old leaves affects the young leaves. If water was translocated from the old leaves to the young leaves, particularly in a drought situation, one would expect greater photosynthetic rates and f. wt : d. wt ratios for young leaves on shoots with old leaves attached compared with young leaves on shoots with old leaves detached.

In order to answer the questions posed, gas exchange was measured for at least 36 h consecutively so that the physiological responses of all plants could be accurately assessed at night and during the day; also, the effects of the presence versus the absence of old leaves on the physiology of young, fully developed leaves were examined both under well‐watered and drought‐stressed conditions.

MATERIALS AND METHODS

Plant species

Kalanchoe tubiflora (Harvey) Hamet plants were obtained from University of Kansas glasshouse stock and from three other sources (Glasshouse Works, Stewart, OH, USA; Huntington Library, San Marino, CA, USA; and Cactus King, Encinitas, CA, USA) in order to ensure genetic diversity in these vegetatively reproducing plants. Carpobrotus edulis (L.) Bol. plants were collected from a roadside in San Diego, California. Sedum spectabile Boreau plants were obtained from diverse stock in the University of Kansas glasshouse. All plants were grown in Premier Horticulture Pro‐Mix (Red Hill, PA, USA) composed of sphagnum peat moss (75–85 %), perlite and vermiculite in 1·5‐L pots.

All plants were mature or grown from cuttings in the glasshouse under a natural photoperiod with a photosynthetic photon flux (PPF) ranging from 200 µmol m^–2 s^–1 during cloudy conditions to 1000 µmol m^–2 s^–1 in full sunlight, relative humidity 50/80 % day/night and averaged air temperature ranges of 25–30 °C/15–20 °C day/night. Plants were watered at least once every 2 d and fertilized weekly with 20 : 10 : 20 N : P : K (with traces of Mg, B, Cu, Fe, Mn, Mo and Zn) (Scott’s Peters Professional Water Soluble Fertilizer, Marysville, OH. USA).

There were four treatments per species: (1) well‐watered with old leaves attached (hereafter called ‘watered attached’ and ‘Watt’) or (2) detached (‘watered detached’ and ‘Wdet’), (3) drought‐stressed with old leaves attached (‘drought attached’ and ‘DRTatt’) or (4) detached (‘drought detached’ and ‘DRTdet’).

Gas exchange

Ten days prior to gas‐exchange measurements, plants were placed in a controlled environmental growth chamber (EGC) (Environmental Growth Chambers, Chagrin Falls, OH, USA), with a 12‐h photoperiod with 500–600 µmol m^–2 s^–1 PPF, 30 °/20 °C day/night air temperatures, 50–60 % daytime relative humidity and 65–80 % /night‐time relative humidity. Approx. 15–25 old leaves were detached from the plants 5 d prior to gas‐exchange measurements, leaving approx. four to ten (dependent upon the species and leaf spacing and size: four to six for C. edulis, eight to ten for K. tubiflora, five to six for S. spectabile) young, fully developed leaves on the stem. In addition, approx. two to three intermediate‐aged leaves were removed from the plants from which leaves were otherwise not removed to ensure a proper seal of the shoot in the gas‐exchange cuvettes. Two old leaves were left on the plants from which the old leaves were removed; these were sampled for acidity and f. wt : d. wt ratios. Acidity and f. wt : d. wt ratios were measured on both young and old leaves. Plants were watered at least every other day in the EGC. The drought‐stress treatment began 5 d prior to, and continued throughout, the gas‐exchange measurements.

Gas exchange was measured under laboratory conditions in a custom‐built gas‐exchange system with open‐flow, using three polycarbonate cuvettes with fans for air circulation and an infrared gas analyser (IRGA) (LI‐COR LI‐6262; LI‐COR, Lincoln, NB, USA). Plants were sealed into the gas‐exchange cuvettes with a mastic (adhesive mounting putty; Manco, Inc., Avon, OH, USA). The gas‐exchange system was calibrated two or three times daily. Measurements were recorded in a cycle every 5 min for 15 min per cuvette for 36 h consecutively or more. Plants in the cuvettes were exposed to the same environmental conditions as in the EGC.

After sealing in the cuvettes, plants were allowed the first afternoon and evening to acclimate to the measurement conditions in the gas‐exchange system. Measurements were then taken on the young leaves on the second day and night. The system did not allow measurement of gas exchange of the old leaves. Gas‐exchange rates were calculated according to the equations of Šesták et al. (1971) and Farquhar and Sharkey (1982). The area under the curves for net CO₂ uptake and net H₂O vapour loss were calculated using SigmaPlot (SPSS Inc., San Rafael, CA, USA).

Fresh and dry mass measurements and acidity

On the third morning of the gas‐exchange measurements, one old leaf and one young leaf were removed from each plant and weighed before the lights were turned on (0830 h). Leaves were stored frozen (–10 °C) for 1 week prior to titrations for measurement of total acidity. At the end of the third day before darkness (2030 h), another old leaf and young leaf were removed, weighed and frozen. Before the lights were turned off, the remaining tissue in the gas‐exchange cuvette was removed, weighed and frozen for determination of shoot fresh and dry mass. The latter were used for calculations of gas‐exchange measurements.

Acidity was measured by grinding the frozen leaves in deionized water using a mortar and pestle. The resultant slurry was titrated to pH 7·00 using 0·01 m NaOH. After titration, the slurry was dried at 60 °C for at least 3 d and weighed for fresh and dry mass calculations and gas‐exchange measurements on a dry mass basis.

Statistical analysis

Variances of means for attached and detached treatments for each species were compared using an F‐test to determine their degree of heteroscedasticity (Sokal and Rohlf, 1995). A pooled variance two‐sample t‐test was employed on non‐heteroscedastic data; a two‐sample t‐test using non‐pooled variances was used for means with heteroscedastic variances (Minitab v. 12.0). Significant differences between attached and detached treatment means were inferred when P < 0·05.

RESULTS

For well‐watered and drought‐stressed plants of C. edulis, there were no significant differences between gas‐exchange parameters measured for plants with leaves attached or detached (Table 1). Net CO₂ uptake (Fig. 1A) and H₂O vapour loss (Fig. 1B) occurred only in the light, although CO₂ efflux (under well‐watered and droughted conditions) and water vapour uptake (drought‐stressed conditions only) occurred in darkness (Tables 2 and 3). There were no significant effects of detachment of older leaves on stomatal conductances, C_i (CO₂ concentrations inside the leaf), f. wt : d. wt ratios, or fluctuations in acidity in this species either under well‐watered or droughted conditions (Table 1).

Table 1.

Statistical significance (P‐values) for the differences in physiological parameters measured for plants of Carpobrotus edulis, Kalanchoe tubiflora and Sedum spectabile with old leaves detached vs. attached: photosynthesis (Ps), transpiration (Tr), leaf f. wt : d. wt, diurnal acid fluctuation (Δ acid), stomatal conductance (G_s) and internal CO₂ concentration (C_i)

Treatment	Ps all	Ps night	Ps day	Tr all	Tr night	Tr day	f. wt: d. wt new	f. wt: d. wt old	Δ acid (mmol g–1)	Gs (mmol g–1)	Ci (ppm)
C. edulis watered	0.56	0.77	0.87	0.27	0.24	0.12	0.29	0.68	0.36	0.94	0.95
C. edulis droughted	0.96	0.65	0.81	0.33	0.36	0.37	0.55	0.77	0.19	0.49	0.66
K. tubiflora watered	0.14	0.13	0.44	0.35	0.015	0.83	0.11	0.058	0.067	0.58	0.041
K. tubiflora droughted	0.024	0.24	0.009	0.049	0.26	0.025	0.006	0.89	0.088	0.66	0.76
S. spectabile watered	0.14	0.83	0.8	0.43	0.56	0.86	0.16	0.48	0.15	0.48	0.37
S. spectabile droughted	0.88	0.57	0.92	0.83	0.47	0.91	0.027	0.013	0.33	0.98	0.14

Significance inferred at P < 0·05; significant values in bold type.

Fig. 1. Two representative curves of net CO₂ exchange (A) and net H₂O exchange (B) for drought‐stressed Carpobrotus edulis shoots with old leaves detached (filled symbols) or attached (open symbols). Positive values indicate CO₂ uptake (A), H₂O loss (B). Dark bars indicate darkness. Time is from the start of the measurements.

Table 2.

Integrated CO₂ uptake (Ps) for three succulent species, Carpobrotus edulis, Kalanchoe tubiflora and Sedum spectabile over 24 h of measurements (Ps 24 h), second night of measurements (Ps night) and second day of measurements (Ps day) for plants with old leaves detached and attached

	Ps all		Ps night		Ps day
	Detached	Attached	Detached	Attached	Detached	Attached
C. edulis watered	4136 ± 768	4910 ± 634	–1068 ± 141	–1120 ± 103	6507 ± 928	6685 ± 579
C. edulis droughted	256 ± 248	242 ± 164	–327 ± 54	–357 ± 31	1063 ± 278	986 ± 149
K. tubiflora watered	11164 ± 959	8915 ± 1013	1445 ± 311	863 ± 115	8054 ± 1055	6882 ± 1000
K. tubiflora droughted	6800 ± 902	3840 ± 184	2296 ± 380	1763 ± 143	2162 ± 397	466 ± 91
S. spectabile watered	6175 ± 1694	5912 ± 901	328 ± 350	416 ± 199	5564 ± 2317	4885 ± 1081
S. spectabile droughted	5333 ± 2077	4856 ± 2185	495 ± 209	346 ± 150	4353 ± 2406	4026 ± 1938

Significant differences between rates of leaves on plants with attached or detached old leaves inferred at P < 0·05 as shown in bold type.

n = 6 for each mean ± s.e.

Table 3.

Integrated H₂O loss (Tr) for three succulent species, Carpobrotus edulis, Kalanchoe tubiflora and Sedum spectabile, over 24 h of measurements (Tr 24 h), second night of measurements (Tr night) and second day of measurements (Tr day) for plants with old leaves detached and attached

	Tr all		Tr night		Tr day
	Detached	Attached	Detached	Attached	Detached	Attached
C. edulis watered	2220 ± 214	2588 ± 228	387 ± 67	121 ± 191	1450 ± 214	2098 ± 313
C. edulis droughted	148 ± 95	–7 ± 117	–48 ± 21	–88 ± 36	248 ± 61	179 ± 40
K. tubiflora watered	2529 ± 300	2185 ± 188	298 ± 33	179 ± 13	1812 ± 231	1746 ± 181
K. tubiflora droughted	713 ± 137	353 ± 65	104 ± 24	72 ± 9	470 ± 88	193 ± 42
S. spectabile watered	1387 ± 777	669 ± 419	49 ± 119	–64 ± 142	1362 ± 456	1262 ± 282
S. spectabile droughted	809 ± 430	694 ± 316	–18 ± 46	–68 ± 48	875 ± 448	810 ± 306

Significant differences between rates of leaves on plants with attached or detached old leaves inferred at P < 0·05 as shown in bold type.

n = 6 for each mean ± s.e.

In K. tubiflora, integrated CO₂ uptake over 36 h was significantly greater in young leaves when the older leaves had been removed than when they were attached, but only in drought‐stressed plants (Table 1). The effect was observed during the day (Table 2). In this species, integrated net H₂O loss of young leaves was lower at night only in well‐watered K. tubiflora plants with old leaves attached than with them detached (Table 3), whereas in the droughted plants water loss was significantly greater when leaves were detached only in the day. Net H₂O vapour exchange of drought‐stressed plants differed little over most of the 36‐h period of measurement, except for the first 60 min after illumination, when the water vapour loss was substantially greater for young leaves on plants from which the older leaves had been removed (Fig. 2B).

Fig. 2. Two representative curves of net CO₂ exchange (A) and net H₂O exchange (B) for drought‐stressed Kalanchoe tubiflora shoots with old leaves detached (filled symbols) and attached (open symbols). Positive values indicate CO₂ uptake (A), H₂O loss (B). Dark bars indicate darkness. Time is from the start of the measurements.

Sedum spectabile showed no significant differences in CO₂ exchange as a consequence of removal of older leaves (Table 1) and had substantial uptake during the day (approx. ten‐fold greater than at night) both when well‐watered and droughted (Table 2). Net H₂O losses during the day were greater than at night for both watered and droughted plants and with and without older leaves (Table 3).

In the three species, removal of the old leaves had no statistically significant effect on mean maximum CO₂ uptake values (Fig. 3). Neither did removal significantly affect changes in acidity in either well‐watered or droughted plants (Table 1). There was a trend, however, toward greater diel acid fluctuations in plants without old leaves compared with those having old leaves (Fig. 4).

Fig. 3. Means of maximum CO₂ uptake for three succulent species, Carpobrotus edulis, Kalanchoe tubiflora and Sedum spectabile, watered or droughted, and with leaves removed or attached. There were no significant differences between shoots with old leaves attached or detached for either of the water treatments in any species. n = 6 ± s.e.

Fig. 4. Acidity per unit dry mass under well‐watered and drought‐stress conditions at the end of the dark period (AM) and the end of the light period (PM) in young leaves of Carpobrus edulis (A), Kalanchoe tubiflora (B), and Sedum spectabile (C). Differences between AM and PM values (delta acidity) are shown on the right side of the figure. There were no significant differences between shoots with old leaves attached or detached for any of the treatments or species. n = 10–12 ± s.e.

The f. wt : d. wt ratios of both young and old leaves of C. edulis were much larger in well‐watered compared with droughted plants, but there was no effect of leaf removal (Fig. 5A and B). Kalanchoe tubiflora responded similarly, except that the detachment of old leaves slightly increased the f. wt : d. wt ratio of young leaves. Sedum spectabile had similar ratios in watered and droughted plants with higher ratios in shoots without older leaves, significantly so under drought (Fig. 5).

Fig. 5. F. wt : d. wt ratios for young leaves (A) and old leaves (B) for three species of succulents, Carpobrotus edulis, Kalanchoe tubiflora and Sedum spectabile, under well‐watered or droughted conditions. An asterisk indicates a significant difference between shoots with old leaves attached or detached at P < 0·05. n = 10–12 ± s.e.

DISCUSSION

Carbon uptake by Carpobrotus edulis was characteristic of C₃ plants, with CO₂ uptake limited to the day when well‐watered and droughted. Previous studies found an induction of diel acid fluctuations in this succulent caused by drought‐stress (Earnshaw et al., 1987) or salt‐stress (Winter, 1973); however, in our study there were no significant day/night differences in acidity, i.e. they did not exhibit CAM photosynthesis. Perhaps the duration and extent of the drought‐stress imposed was not sufficient to cause diel acid fluctuations; however, longer periods of drought‐stress eliminated gas exchange in preliminary studies on this species (data not shown), suggesting that it does not have the capacity to acquire CAM CO₂ assimilation.

The uptake of water by C. edulis at night is curious. Given the standard error values for these measurements (Table 3), the uptake is minimal, and does not appear to be a result of the presence or absence of old leaves.

The young leaves of C. edulis had higher f. wt : d. wt ratios under drought conditions, in other words contained more water, with the old leaves detached than attached. This result is the opposite of earlier findings (Evans, 1932; Kramer, 1959; Čatský, 1962; Milburn, 1979; Schäfer and Lüttge, 1987; Schmidt and Kaiser, 1987; Schulte and Nobel, 1989; Tüffers et al., 1995) and suggestions concerning leaf‐to‐leaf water movement (Barcikowski and Nobel, 1984; Schulte and Nobel, 1989; Herrera et al., 2000). Thus, the results presented here provide no evidence that water moves from old to young leaves during drought‐stress in this species.

Results for K. tubiflora were typical of a species with intermediate C₃‐CAM photosynthesis (both daytime C₃ and night‐time CAM CO₂ uptake). Well‐watered plants showed a C₃‐CAM intermediate pattern of carbon uptake and acid fluctuations. Drought‐stressed plants exhibited a classic CAM pattern of gas exchange and diel acid fluctuations, in agreement with previous studies (Stidham et al., 1983). Diel acid fluctuations were slightly larger under well‐watered than droughted conditions, in agreement with previous studies on obligate CAM plants (Szarek et al., 1973; Szarek and Ting, 1974, 1975; Lüttge and Ball, 1977). Generally, acid fluctuations were greater in shoots without old leaves than in shoots having old leaves. This applies to all three of the succulent species examined. This greater fluctuation in acidity is a direct result of greater CO₂ uptake in the dark when the old leaves had been removed from K. tubiflora. F. wt : d. wt ratios were greater in young leaves of drought‐stressed K. tubiflora plants without old leaves than in shoots with them; however, the f. wt : d. wt ratios in old leaves of these plants were nearly identical. This indicates that when drought‐stressed, the young leaves may gain more water from another source (i.e. the roots, stems, or other parts of the plant or directly from the soil) rather than from the old leaves.

Sedum spectabile also has an intermediate C₃‐CAM type of gas exchange when well‐watered, with some fluctuations in diurnal acidity. Dark CO₂ uptake increased, while CO₂ uptake in the light decreased, with drought. Winter and Lüttge (1976) and Kluge (1977) observed induction of CAM in droughted Sedum spp. The diel fluctuations in acidity in S. spectabile increased with drought‐stress. There was no significant difference between the fluctuations in acidity of young leaves on shoots with old leaves detached or attached. The f. wt : d. wt ratios of young leaves were significantly higher when old leaves were removed, compared with plants with old leaves attached.

For the three species, f. wt : d. wt ratios of the young leaves were greater (although significant only in a few cases) when the old leaves were removed for both well‐watered and droughted plants, and also the ratio for young leaves was greater than for the old leaves on the plant (except in drought‐stressed C. edulis), suggesting that the young leaves acquired more water when the old leaves had been removed. Results of CO₂ uptake and H₂O loss for drought‐stressed plants of C. edulis and S. spectabile did not indicate that removal of old leaves significantly improved the physiology of younger leaves. In K. tubiflora, however, rates of CO₂ uptake and H₂O loss of the young leaves were greater when the old leaves had been removed, particularly in drought‐stressed plants.

In accordance with previous studies, (Barcikowski and Nobel, 1984; Schmidt and Kaiser, 1987; Schäfer and Lüttge, 1987; Schulte and Nobel, 1989; Tüffers et al., 1995; Herrera et al., 2000), detachment of old leaves from a drought‐stressed succulent plant, such as K. tubiflora, can influence physiological traits of plants under similar conditions. The nature of this impact, however, is opposite of what one would expect if the young leaves were acquiring water at the expense of the older leaves. An explanation for why greater photosynthetic rates and greater f. wt : d. wt ratios were found in young leaves after the removal of old leaves relates to within‐plant competition for water. The presence of old leaves may limit the photosynthetic rates and water content of the young leaves by diverting some of the water taken up by the plant, reducing the amount available to the young leaves. This study refutes previous suggestions (Evans, 1932; Kramer, 1959; Schäfer and Lüttge, 1987; Tüffers et al., 1995; Herrera et al., 2000) that water is moved from old to young leaves in succulents during drought‐stress. Moreover, these results indicate that removal of the old leaves leads to greater water content and greater photosynthetic rates in the remaining young leaves, at least in one succulent species examined.

ACKNOWLEDGEMENTS

We thank John Trager at the Huntington Botanical Gardens in San Marino, CA and Katie Nus and Chip Taylor at the University of Kansas for providing plant specimens.

Received: 15 February 2003; Returned for revision: 17 March 2003; Accepted: 11 June 2003    Published electronically: 7 August 2003