| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ecology |
Department of Botany, University of Wyoming, Laramie, Wyoming 82071-3165 USA; and Department of Biology, Wake Forest University, Winston-Salem, North Carolina 27109-7325 USA
Received for publication February 1, 2000. Accepted for publication October 24, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
10%/min prior to rainfall, in close correspondence to declines in air and corolla temperatures. Identical floral behavior was also induced experimentally in the field and laboratory by artificial changes in corolla temperature. Corolla closure did not occur during experiments that simulated natural changes in solar irradiance, wind, or absolute humidity during a thunderstorm. Furthermore, individual G. algida plants forced experimentally to remain open during rain had substantial losses of pollen after single rain events (up to 34%) and if forced to remain open for the entire flowering period (59%). Subsequent seasonal reductions in female fitness (up to 73%) also occurred, including seed size and mass, number of ovules produced, number of viable seeds produced per ovule, and seed germination. Thus, corolla closing and opening in G. algida associated with frequent summer thunderstorms may be a behavioral adaptation that improves both paternal and maternal reproductive effort.
Key Words: alpine • floral movements • Gentiana • reproduction • temperature
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Hart, 1990
). These orientational changes can occur over days due to alterations in growth patterns or much more rapidly (within minutes) as the result of changes in cell turgor pressure. Despite a large number of plant families implicated in studies of plant movement, relatively few species appear to be able to move reproductive structures on a responsive, repetitive, and reversible basis. Simons (1992)
calculated that a total of 1000 angiosperms demonstrate some form of repeatable movements in either petals, carpels, or stamens. Comparing these figures with the number of flowering species that are known to exist (235 000; Raven and Johnson, 1989
), floral movements of any kind appear to be rare in the plant kingdom.
Floral movements in response to temperature may be particularly uncommon. Goldsmith and Hafenrichter (1932)
reported that the flowers of Gentiana algida responded to changes in temperature under laboratory conditions, while flowers of Crocus vernus (Iridaceae) were also observed to open and close as temperatures changed in the laboratory (Andrews, 1929
; Crombie, 1962
). With the exception of these limited observational data, and a few other observational reports (e.g., Percival, 1965
; Stanton and Galen, 1989
; and Simons, 1992
), there appears to be no comprehensive information available regarding floral movements in response to temperature change in natural environments. No information appears to exist concerning the adaptive benefits of floral movements.
Gentiana algida Pallas (Gentianaceae) is an herbaceous perennial that inhabits arctic and alpine environments of Asia and North America. In North America it is distributed southeastward from arctic regions in Alaska and the western Yukon Territory, with notable disjunction into the alpine regions of the central and southern Rocky Mountains (Polunin, 1959
; Scoggan, 1979
; Cronquist et al., 1984
). In Wyoming, where portions of both the central and southern Rocky Mountains occur (see Hunt, 1974
), G. algida is found in alpine meadows and fellfields of the Absaroka, Beartooth, Medicine Bow, and Wind River Mountains (Scott, 1996
). Additionally, Dorn (1992)
reported that populations of G. algida may occur below timberline within these same areas, and Nelson (1984)
stated that the species can be found growing in boggy sites of the Medicine Bow Mountains of southeastern Wyoming, as well as in meadows within the alpine zone.
Gentiana algida has a caespitose growth habit and produces relatively large sympetalous flowers. The species is sometimes referred to as the "narrow-leafed gentian" for its grass-like leaves that originate each year from a persistent caudex at ground level (Polunin, 1959
). This ground-hugging tuft may produce a solitary stem as short as 2 cm, with 1–3 flowers in a simple terminal cyme. The conspicuous flowers approach 5 cm in length and have an erect, funnelform corolla comprising five connate petals (Fig. 1). While both sexes are housed within the corolla, they are temporally and spatially separated by protandry (differential maturation) and herkogamy (differences in position). The gynoecium comprises a solitary bicarpellate ovary, with a single compartment and parietal placentation, and the androecium consists of five epipetalous stamens. There are also five floral nectaries that develop as ridges on the base of the stipitate pistil (Lindsey, 1940
). Arctic gentians reproduce both sexually and vegetatively from rhizomes.
|
). These conclusions were supported by data on viable seed production from plants that were bagged and/or emasculated in the field during the flowering period. Declines in several female reproductive variables after bagging and/or emasculation indicated that G. algida is an entomophilous species dependent upon pollinators for seed set. Also, the temporal and spatial separation of male and female reproductive structures (protandry and herkogamy), plus the lack of seed set for emasculated flowers, indicated that this species is at least facultatively xenogamous (fertilization between different plants), if not predominantly xenogamous (Cruden, 1977
; Kearns and Inouye, 1993
). Finally, our field observations revealed that bumble bees were the primary, if not sole, pollinators of G. algida, matching the relatively large size of the corolla opening in this species (Fig. 1). The present study investigates environmental stimuli (e.g., temperature) that may elicit observed floral movements in G. algida, as well as their potential adaptive value to reproductive effort. Specifically, the flowers of G. algida are conically shaped, vertically oriented, and have been observed to close in response to approaching thunderstorms, and subsequently reopen as conditions become more clement (R. W. Scott, personal communication, Riverton, Wyoming). We questioned whether the flowers of G. algida close during the day in response to declines in either solar radiation, air temperature, or both, associated with the frequent afternoon thunderstorms that occur in this high-elevation zone during summer. For these upright and tubular flowers, it seemed logical that exposure to rainfall could reduce reproductive success due to the physical removal of pollen from anthers by raindrops ("rainwash") or some other form of water damage. Hence, the two primary objectives of the study were (1) to observe and document floral movements in the field while evaluating the nature of the environmental stimuli and (2) to evaluate the potential adaptive benefits of such movements according to measured impacts on reproductive effort.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Study sites
Two research sites were located in the Medicine Bow Mountains of southeastern Wyoming where the southern Rocky Mountains approach their northern limit. One site was established in a subalpine meadow (SAM site) at 3289 m elevation, while the second site was located within an alpine fellfield at 3344 m (ALF site). Both the SAM and ALF sites are situated on the leeward side of the northeastern shoulder of Browns Peak (41°23' N; 106°15' W) and have similar slopes (8° and 11°, respectively) and azimuths (89° and 94°, respectively). Both sites are fed by meltwater from snowdrifts that persist into midsummer. The SAM site is dominated by two graminoid species, Deschampsia caespitosa (L.) Beauv. and Carex scopulorum Holm, while plants such as Artemisia scopulorum Gray, Pedicularis groenlandica Retz., Polygonum bistortoides Pursh, and Geum rossii (R. Br.) Ser. occur as subdominant herbs. Senecio cymbalarioides Buek and Potentilla diversifolia Lehm. constitute a minor component of this meadow community. The ALF site is codominated by Geum rossii (R. Br.) Ser., Artemisia scopulorum, Deschampsia caespitosa, and Carex nova Bailey, with notable numbers of Polygonum viviparum L., P. bistortoides, Potentilla diversifolia, Trisetum spicatum (L.) Richt., and Salix arctica Pallas. Also, Chionophilia jamesii Benth., Erigeron simplex Greene, Sibbaldia procumbens L., and Trifolium parryi Gray occur as occasional members in this alpine community.
Field research on G. algida was also conducted at a third site on the Beartooth Plateau located in the Beartooth Mountains of north–central Wyoming (central Rocky Mountains). An alpine site was selected near Garner Lake (GL site; 44°59' N and 109°27' W) and consisted of a small meadow at 3115 m elevation with an approximate aspect of 264° and slope of 6°. This meadow was inhabited by Deschampsia caespitosa, Geum rossii, Potentilla diversifolia, Artemisia scopulorum, Polygonum bistortoides, P. viviparum, and Salix reticulata L. Gentiana prostrata Haenke is an infrequent component of this community, while Salix glauca L. formed ankle-high thickets bordering the study site.
Precipitation pattern
To evaluate the impact of rainfall on the floral movements and reproductive effort of G. algida (e.g., non-vector-induced pollen loss, or the type of damage described in Inouye et al., 1994
), the amounts and frequency of rain events, as well as the fall-line angle (from vertical) and azimuth direction were determined. Meteorological data were obtained from the USDA Forest Service at a subalpine site 1.5 km south–southeast of the ALF and SAM research sites at slightly lower elevation (Brooklyn Lake Meteorological Station, 3138 m elevation). This station is equipped to monitor hourly precipitation (tipping-bucket rain gauge at 3 m above ground) and air temperature (recording thermistor) at 2 m (B. Susemihl and A. Ellsworth, personal communication, Rocky Mountain Forest and Range Experiment Station). Precipitation data for the summers of 1991–1994 were compiled into a single data set and analyzed in terms of hourly frequency and quantity. Fall-line angles and azimuth angles for falling raindrops were visually measured for individual rain events using a compass and inclinometer on six different days in late July and mid-August (N = 11 rainfall events).
Floral orientation and structure
Floral orientation was measured in order to assess corolla susceptibility to rainfall interception and rainwash. In separate samples of 100 randomly selected flowers at each site (ALF and SAM), the angle of inclination (corolla angle) from horizontal was determined with a clinometer, while azimuth orientation of the corolla was measured with a compass. All angles were measured to the nearest ±2° at midday on clear days when corollas were fully open.
Corolla structure was also measured at the ALF (N = 36) and SAM (N = 68) sites. The distance between the base of the corolla, at the point of attachment with the pedicel, and the proximate edge of the sinus (the indentation between adjacent lobes on the corolla) was measured to estimate the depth of the corolla. A second measure was taken between the base of the corolla and the distal end of a corolla lobe to determine overall length. In addition, the diameter (d) of corolla throats, or opening at the juncture of the corolla tube and corolla lobes, was measured for 122 randomly selected flowers at both sites (N = 68, ALF; N = 54, SAM). All measurements were taken manually using either a thin, cylindrical probe with 0.2 mm graduations or a precision caliper with 0.1 mm resolution.
Microenvironmental temperatures and floral movements
At the ALF and GL sites, air and plant temperatures were monitored at six different locations on individual G. algida plants (N = 6), while corolla movements (widths) of terminal flowers (N = 2–3) were also measured. Every half-hour, from pre-dawn to post-dusk, temperatures were measured inside corollas using a fine-wire thermocouple (36 ASG, 0.014 mm diameter); leaf temperatures of mid-cauline leaves and air temperatures immediately adjacent (<2 cm) to individual flowers were also measured. Corolla temperatures near the base of the ovary were monitored by inserting a beaded thermocouple into the corolla tube to a depth approaching that of the floral nectaries. Leaf temperatures were taken by gently pressing the thermocouple to the shaded, adaxial surface. The width of the corolla opening was measured with precision calipers as the maximum distance between lobe tips (Fig. 1). Measurements at the ALF site were taken on 26 July, 13 and 14 August, and 20, 21, and 22 August at the GL site.
Floral temperatures were manipulated experimentally for plants of G. algida in the field to evaluate the influence of temperature on the observed closing and opening responses of flowers. Open flowers (N = 7) were given a chilling treatment under clear skies (full sunlight) during the day by encircling plants with a doughnut-shaped wall of snow, avoiding contact with the plant and any shading of flowers. In some cases, corolla temperatures were reduced to as low as 3°C, which was also >10°C below air temperature. Flowers were also cooled to below ambient air temperature by blowing cold air (passed through a metal pipe wrapped with dry ice) onto flowers. In addition, closed flowers (N = 7) were warmed at night by placing chemical handwarmers, or heated waterpacks, in close proximity to the flowers. Potted plants (N = 6) in the laboratory were moved from a refrigerator (7°–8°C) to room conditions (22°–24°C) for qualitative observations of floral movements. Finally, identical experiments were also conducted using dry vs. moist air (22% vs. 78% relative humidity at the same approximate air temperature using air streams directed through desiccant or moistened tubing) and high vs. low wind speed (0.5 m/sec vs. 3.4 m/sec) in the field (N = 6) and laboratory (N = 5) to test for possible effects of humidity changes during thunderstorms (pulvinar responses) or increased wind (thigmotactic responses). In all of the above experiments, corolla widths were measured before and after exposure to a new treatment level for 5 min.
Experimental coning and pollen rainwash
To investigate the effects of rainfall on pollen contents of individual corollas, natural flowers of G. algida were compared with flowers that were prevented from closing during afternoon thunderstorms by inserting wax-paper cones inside corollas. These cones were inserted into the throat of each corolla just prior to and during the duration of individual rain events. To adjust for differences in corolla throat diameters, the length (height) of the cone from one or both ends was altered. Corollas were selected as experimental units based on the stage of floral development observed. Male-phase flowers exhibiting dehiscent, pollen-producing anthers and a nonreceptive pistil (i.e., style may have been exserted above the anthers, but the stigma was not bifurcate) were considered qualitatively similar in terms of maturation (Devlin and Stephenson, 1985
). Field experiments were conducted on 9 August in the ALF population (N = 32) and on 20 August in the SAM population (N = 68). Precipitation was recorded with a portable rain gauge located near the center of each sampled population. Immediately following each storm sampled, flowers were removed and temporarily stored upright in open containers for drying. After returning to the laboratory, the five stamens of each flower were excised, rinsed with 95% ethanol, and stored in cuvettes. After the rinse had evaporated, each sample cuvette was sealed with a parafilm lid. Prior to counting, pollen was removed from the anther and placed on a microscope slide with cover slip.
Pollen counts
Pollen grains were counted using a photometric approach and regression analysis. The method was tested using a standard curve for predicting floral pollen counts in control and rainwashed populations using data from a subsample of G. algida flowers harvested during the experiments. Individual pollen samples were suspended in 3 mL of anhydrous glycerol and the optical density of each sample was measured with a spectrophotometer (Beckman Model 60). Optical density (absorptance) measurements were made at 385 nm after an initial analysis (i.e., an absorption spectrum analysis using a range of wavelengths between 330 and 1000 nm) indicated that a relatively strong absorptance maximum occurred at this wavelength. Preparatory to absorptance measurements, samples were thoroughly mixed and the spectrophotometer was calibrated to zero using a blank cuvette containing glycerol and no pollen. After absorptance readings were recorded for each sample, actual pollen counts were made in subsamples using a microscope and a Sedgwick-Rafter (S-R) cell. The S-R cell is a 1 mL (50 x 20 x 1 mm) counting chamber used for quantifying aquatic microorganisms following a technique prescribed by the American Public Health Association (APHA, 1989
). Pollen samples (3 mL) were stained with five drops of cotton blue in lactic acid and a 1-mL pollen sample was drawn for microscopic enumeration in the S-R cell. Prior to enumeration, the microscopic field of view at 100x was calibrated using a stage micrometer. Transects were selected randomly for sample slides, and pollen was counted in the 1-mL subsample.
The estimated accuracy of the mean value for the number of pollen grains was calculated using a formula adapted from Eckblad (1991)
where count accuracy is equal approximately to the t value. Absorptance values were correlated with actual pollen counts using linear regression (PROC RE, SAS Institute 1989
). Regression models with an intercept of zero were chosen given the a priori assumption that zero pollen in the glycerol blank yields a reading of zero absorptance. When G. algida pollen counts (independent variable) were regressed with absorptance readings (dependent variable), a significant and positive linear relationship occurred between the two variables. The slope was significantly different than zero (i.e., b = 0.002596, t = 16.494, P < 0.0001). Moreover, 97% of the absorptance response measured in the samples was attributable to corresponding changes in pollen concentration (r2 = 0.97). Hence, floral pollen counts were quantified using the regression equation y = 374 273x, where y = pollen number and x = optical absorptance.
Influence of floral movements on seasonal reproductive effort
The loss of pollen after individual rainfall events, from flowers forced experimentally to stay open (coned), is not conclusive evidence that reproductive effort was diminished. Thus, cones were inserted into corollas to prevent closure of corollas over the entire flowering period at both the ALF and SAM sites. All treatments were initiated on 11 August and terminated with the harvest of the flowers on 12 September. Following harvest, sampled flowers were stored in open-air containers until fruits were seasoned. The frequency and duration of individual visits by flying pollinators were also compared for coned and control plants during the peak of the flowering period at both sites (16, 17, 20 August).
To assess the effects of experimental coning (and possible rainwash of pollen) on reproductive effort in G. algida, seed set, percentage seed set, total fruit seed mass, mean individual seed mass, number of viable seeds produced, and percentage germination were measured and compared for coned vs. control plants (N = 36). Seed set, or total number of seeds produced, was determined by opening the seasoned capsules and counting all the seeds. Because percentage seed set is defined as the proportion of ovules that mature into seeds, it was also necessary to count the total number of infertile ovules within each capsule. Air-dried seeds from each capsule (60 d post-harvest) were weighed together on a 0.01-mg balance. This value was divided by the total number of seeds to estimate mean mass per seed. Additionally, seed viability and percentage germination were determined via the germination tests described below.
Prior to the germination trials, seeds were placed in manila envelopes and held in dry storage for 138 d at 15°–18°C. Seeds were then chilled in a refrigerator for 104 d at 4°–7°C, followed by a stratification treatment of 20 d. For the stratification treatment, seeds were placed in petri dishes on filter paper saturated with distilled water and allowed to imbibe under the same temperature regime as the chilling treatment. After removal from the refrigerator, seeds were irradiated in a 16-h day growth chamber under high-light conditions (940–1127 µmol·m–2·s–1 photosynthetically active radiation [PAR]). Temperatures (T) within the chamber were maintained at 20°C during the day and 5°C at night. After two full days in the growth chamber (1–2 June), seeds were removed to avoid excessive dehydration and placed in a north-facing window (Tair 
18°–20°C) during the day and outdoors under natural conditions at night (Tmin 6°C). The seeds were monitored for 25 d, and seed germination was scored according to the appearance of an emergent hypocotyl. The number of seeds that germinated was compared to the number of seeds tested and expressed as percentage germination.
Statistics
Experimental results from field and laboratory experiments were evaluated using standard statistical tests as described in Zar (1999)
. For example, two-sample t tests (one-sided) were used for effects of experimental coning, and a V test for uniform distribution (circular) was used to test for significant corolla orientation preferences. All statistical tests employed included any required evaluations of assumptions about normality in sampled populations (Wilk-Shapiro/Rankit Plots) and the equality of variances (F test).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Precipitation patterns
During August, rain events and measurable precipitation occurred with the highest frequency during the daylight period from late morning to mid-evening (Fig. 2). Of the 52 rain events recorded, 22, or 42% of all rain events, occurred within a 5-h period between 1400 and 1900. During this same period, 49.5% (51 mm) of the total precipitation (103 mm) was received, while 58% (60 mm) occurred during a 9-h period between 1100 and 2000, also the time period when 50% (26) of all rain events were logged. Consequently, rain events and rainfall were not uniformly distributed throughout an August day (Fig. 2). Chi-square tests indicate that afternoon showers, and the precipitation associated with such events, represent a significant proportion of diurnal rainfall (event 
2 = 42.5, df = 23, P < 0.01; precipitation 2 = 104.6, df = 23, P < 0.0005). In addition, the median hour for a rain event was 1530, while the average amount of rain per event was 2 mm. For a total of 11 rainfall events observed, there was no significant difference in droplet angles from vertical or azimuth angles.
|
|
14 mm (13.8 ± 1.8), while extreme values were near 11 mm and 16 mm.
Floral movements and temperature
Floral movements (percentage openness of corolla) vs. time and air temperatures at 10 cm above ground surface (approximate inflorescence height) are plotted for a representative day at each of the two research sites (Fig. 4). Midday fluctuations in temperatures coincided with transient cloud events and reductions in insolation. Contraction of the corolla occurred following a decrease in corolla temperature (prior to rain or hail), in response to the decreased incident sunlight and air temperatures associated with increased cloudiness. Corollas also reopened in response to warming air temperatures and the return of direct sunlight. For instance, on 6 August at the ALF site, falling temperatures were recorded in conjunction with advancing clouds and afternoon thunderstorms. The first storm cell developed and dissipated during the period between 1343 and 1450. During that time, the temperature measured at plant height (10 cm) declined rapidly from near 15°C to 7.5°C in 30 min. The same phenomenon was recorded later that afternoon (1609) when air temperature decreased from 12.6°C to 7.1°C within 14 min (0.3°C/min). Each event was accompanied by lightning, rain, and occasional light hail. Also, closed flowers subsequently reopened in the interims between squalls and flowers were never observed in the open condition when air temperatures were below 5°C during the night or day.
|
Experimental manipulation of corolla temperatures in the field was achieved by simulating air temperature changes (fan and dry ice) that occurred during natural cloud-cover events. These experimental temperature treatments also generated corolla closure and enabled a quantitative estimation of the dynamics of this floral behavior (Fig. 5). A slower rate of corolla closure appeared to occur until a temperature of near 10°C below pre-thunderstorm conditions was attained. At this point, corolla width declined rapidly to near full closure. For a total of 3 d of measurement on three different plants, maximum closure rates of almost 10% of maximum width per minute occurred after corolla temperatures fell to 10°C below maximum values under full sun exposure (28°C to 18°C, Fig. 5). Very similar results were obtained when corolla cooling was accomplished using snow "doughnuts" constructed around the base of individual plants (data not shown).
|
150–350 µmol·m–2·s–1 PAR), if corolla temperatures remained within 2.5°C of pretreatment values (24.1°–28.7°C). Also, experimental applications of drier or more humid air streams, or wind alone, did not elicit detectable corolla closure in the field (data not shown).
Rainwash effects on pollen
Pollen numbers of flowers experimentally coned to promote rainwash were significantly reduced at both the ALF and SAM sites following a single, natural rain event (2.1 mm). Coned flowers at the ALF site (N = 16) had a floral pollen count of 17 217 ± 1495 (mean ± 1 SE), while flowers that were allowed to close naturally (N = 16) had a pollen count of 22 456 ± 2053. Also, this reduction in pollen grain numbers for coned flowers (23%) was statistically significant for flowers allowed to close naturally (one-tailed t test; t = 2.045; df = 30; P = 0.0249). Similarly, pollen numbers for coned plants were reduced by 34% at the SAM site following a 4.8-mm natural rain. The pollen count in unconed, natural flowers (N = 34) was 25 076 ± 1915 grains, while the pollen count in coned flowers (N = 34) was 16 468 ± 1340 grains. Here again, flowers that were prevented from closing (coned) during this afternoon thunderstorm had a statistically significant loss of pollen (one-tailed t test; t = 3.651; df = 59; P < 0.001).
Pollen counts for coned vs. natural, control plants (N = 25 each) were also reduced substantially when measured at the beginning of the flowering period (4 August) and at the end (16 September). Mean values for control (nonconed) plants were 29 006 ± 2464 vs. 27 814 ± 2346, respectively, compared to 28 602 ± 2022 and 11 769 ± 1704 for experimentally coned plants. Thus, a substantial reduction in pollen occurred for coned plants (59%) vs. natural, control plants (4%), which was also significant at P < 0.001 (as above).
Maternal effects of rainwash
The preclusion of floral movements in G. algida by experimental coning had substantial and statistically significant effects on five out of the seven reproductive variables measured at the ALF site. For all dependent variables other than percentage germination, coning significantly (P < 0.0001 in four of five cases) decreased mean responses within treated populations (Table 1). For example, flowers receiving the coning treatment had, on average, a 71% reduction in seed set and a >60% decrease in percentage seed set and viable seed production. The effects of coning in the subalpine meadow population were similar to those observed in the alpine fellfield experiment. In five of the six response categories, declines due to coning were statistically significant (P < 0.0001 in all cases; Table 1), with percentage germination being the only exception. Finally, the number and duration of individual visitations by flying pollinators were not statistically different between coned and control plants at either the ALF or SAM sites according to 3 d of observation on 16, 17, and 20 August (N = 10, t test for paired means, P < 0.001).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, July and August are the months with the highest frequency of thunderstorms in the central and southern Rocky Mountains, and, in the case of the southern Rockies, the amount of precipitation received from showers and thunderstorms in July and August can exceed that of any other time during the year. Brown (1980)
has noted that the alpine tundra regions of Wyoming (defined by the author as those elevations in excess of 2896 m) generally experience a higher frequency of thunderstorms than other portions of the state. Meanwhile, Martner (1986
, p. 129) has mapped thunderstorm activity in Wyoming, indicating that, on average, there are between 40 and 60 d each year when thunderstorms occur within these areas. An analysis of the 4-yr data set from the Medicine Bow Mountains suggests that afternoon showers and thunderstorms are particularly common during the month of August. On average, at least one rain event was recorded every other day in August. As might be predicted from the dynamics of thunderstorm genesis, many of these rain events occur during the potentially warmest period of the day when pollinator activity should be greatest. Interestingly, we found no significant patterns in droplet fall angle, or compass direction, as well as no corresponding patterns in the positioning of the corolla opening in G. algida at either the ALF or SAM sites.
Corolla response to temperature
When microenvironmental temperatures were monitored on and around the plants of G. algida, it was evident that natural, daily changes in temperature generated corresponding changes in corolla width. Above a response threshold of 4°–5°C, corollas opened as temperatures increased or closed as temperatures fell. In turn, opening and closing occurred at the same temperature depending upon thermal trends. For example, corollas may have been opening at 15°C when the temperatures were rising, but closing at 15°C if temperatures were falling. In addition to being reversible, floral movements in G. algida could be rapid. Flowers were observed to close fully within 6–10 min on several occasions when rapid temperature changes (>0.5°C/min) occurred with the approach of afternoon thunderstorms. When conditions moderated, the corollas reopened within 25–40 min. Similarly, when flowers were given a daytime chilling treatment, the time required for closure was 7.6 ± 2.8 min (mean ± 1 SD, N = 7). When flowers were experimentally warmed at night, they opened within 44.8 ± 10.2 min (N = 7). Thus, the flowers of G. algida can cycle from completely closed to fully open and to completely closed again, within the span of much less than an hour. This cycle could be repeated several times during a day, especially in the afternoons when typical advective thunderstorms develop.
The above field observations are consistent with the findings of Goldsmith and Hafenrichter (1932)
, who studied the floral movements of G. algida in a controlled laboratory setting >60 yr ago. These authors report working with Gentiana frigida Haenke, but G. frigida is a gentian of European origin that does not occur on this continent. However, G. frigida is a recognized synonym for G. algida Pallas in North America (Hitchcock et al., 1959
). Goldsmith and Hafenrichter (1932)
removed entire plants of G. algida from natural habitats at 4167 m near Colorado Springs, Colorado and placed detached flowers in a growth chamber where temperature, light, and humidity were regulated at various levels. They reported that G. algida flowers did not demonstrate diurnal rhythms in opening times and would remain open under constant temperature conditions. Likewise, flowers remained closed under a constant temperature of <5°C. Hitchcock et al. (1959)
also noted that movement in either direction was reversed by a change in the temperature stimulus, regardless of the time of day, and that the rate of movement was dependent upon the magnitude of the temperature change. Hence, Goldsmith and Hafenrichter (1932)
concluded that "It may be that in the flowers of G. frigida [G. algida] there is an irritability phenomenon for which the sudden temperature change is the stimulus."
Effects of rainwash on reproductive effort
It may be adaptive to reduce wettability, or otherwise avoid wetting in flowers, in order to promote the processes associated with sexual reproduction. For instance, wetting of anthers in terrestrial plants can temporarily prevent dehiscence (Percival, 1955, 1965
), while wetting of pollen is known to affect viability and germinability, in some species, by either increasing mortality or inducing germination prior to pollination (see Fægri and van der Pijl, 1979
; Corbet and Plumridge, 1985
; Corbet, 1990
; Galil and Meiri, 1992
). In addition, wet flowers may discourage pollinators, with the effect that pollen collection and dispersal are deferred (Percival, 1947
).
Another important implication associated with pollen wetting in G. algida is that the pollen grains, which were not dislodged during the rain event, tended to become clumped and cemented to the pollen-bearing surfaces of the anther during the drying process. This became evident when pollen was removed from rainwashed anthers. Considerable manipulation was required to free the pollen, compared to pollen removal on non-rainwashed anthers. Consequently, if pollen placement (i.e., pollen transfer from anther to vector; see Inouye et al., 1994
) is encumbered by the effects of wetting, pollen dispersal could be restricted.
Not only might precipitation affect pollen dispersal in G. algida through a reduction in pollen numbers or a change in pollen consistency, rain could potentially affect pollen dispersal by altering flower–pollinator interactions. Manipulated (coned) flowers that were forced open during a rain event were observed to gather and retain rainwater for several hours. Presumably that nectar was diluted by the presence of the rainwater. Percival (1965)
has suggested that pollen donation and deposition may decline if nectar concentration is affected by the external environment. This statement is reasonable in light of the fact that nectar is the preeminent reward for floral visitors (Simpson and Neff, 1983
), and it is well known that bumble bees, which visit arctic gentians regularly (M. R. Bynum, personal observation), are sensitive to differences in nectar quality (Heinrich, 1979a
; Marden, 1984
). As a result, bumble bees tend to exploit nectar resources with high-energy rewards, while minimizing visits to flowers with low-energy rewards (Heinrich, 1979b, 1983
; Waddington, 1983
; Marden, 1984
; Harder, 1986
). Therefore, it is likely that rainfall could reduce nectar quality and result in fewer pollinator visits to G. algida flowers.
The reproductive potential in G. algida could be negatively impacted via several mechanisms if flowers were not able to close in response to the falling temperatures that precede a typical summer rainfall event. Precipitation can reduce pollen numbers, while rainfall effects may also limit pollen placement and reduce nectar quality. In turn, pollen donation and deposition, and therefore paternal contribution and female fecundity, could be adversely affected. In the present study, female reproductive effort was also significantly impacted, including reductions in the number and size of seeds and fruits, as well as viable seed production, when compared to control plants.
Summary and conclusions
According to the results presented here, the frequent thunderstorms characteristic of these high-elevation habitats during summer would reduce reproductive effort in G. algida, a species with upright, tubular corollas that appear susceptible to rain damage without prior closure. Both male and female fitness would be at risk. This threat was evidenced by the measured pollen loss and reduction in maternal fitness of flowers experimentally forced (by coning) to remain open during rain events. Thus, the corolla-closing response described here appears to be adaptive for the simple reason, at least, that reproductive structures are protected from damage or removal by rainwash in individual flowers. More research is needed to establish a more comprehensive understanding of the ecological importance of floral movements in species with similar floral forms and susceptibility to rainwash effects. There may also exist interesting evolutionary trade-offs associated with the effectiveness of attracting flying pollinators, as well as the pollination process itself, for species with upright vs. pendant corollas (e.g., it may be easier to pollinate an upright, tubular corolla). It is also noteworthy that most climate change scenarios now proposed for global warming trends include increases in afternoon, advective thunderstorms in mountainous regions of North America (Kattenberg et al., 1996
).
|
FOOTNOTES |
|---|
2 Author for reprint requests (tel: 336-758-5779; FAX: 337-758-6008;e-mail: smithwk{at}wfu.edu
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Andrews F. M. 1929 The effect of temperature on flowers. Plant Physiology 4: 281-284
Bliss L. C. 1971 Arctic and alpine plant life cycles. Annual Review of Ecology and Systematics 2: 405-438
Brown R. H. 1980 Wyoming: a geography. Westview Press, Boulder, Colorado, USA
Bynum M. R. 1995 Floral movements in Gentiana algida: benefits for male and female fitness in an alpine environment. Master's thesis, University of Wyoming, Laramie, Wyoming, USA
Corbet S. A. 1990 Pollination and the weather. Israeli Journal of Botany 39: 31-42
———, and J. R. Plumridge 1985 Hydrodynamics and the germination of oil-seed rape pollen. Journal of Agricultural Science 104: 445-451
Costelloe B. H. 1988 Pollination ecology of Gentiana andrewsii.. Ohio Journal of Science 88: 132-138[ISI]
Crombie W. M. L. 1962 Thermonasty. In W. Ruthland [ed.], Handbuch der Pflansenphysiologie, vol. 17, part 2, 15–28. Springer-Verlag, Berlin, Germany
Cronquist A. A. H. Holmgren N. H. Holmgren J. L. Reveal andP. K. Holmgren 1984 Intermountain flora: vascular plants of the intermountain west, U.S.A., vol. 4. New York Botanical Garden, Bronx, New York, USA
Cruden R. W. 1977 Pollen–ovule ratios: a conservative indicator of breeding systems in flowering plants. Evolution 31: 32-46
Darwin C. andF. Darwin 1881 The power of movement in plants. D. Appleton and Company, New York, New York, USA (1966 reprint, Da Capo Press, New York, New York, USA)
Devlin B. andA. G. Stephenson 1985 Sex differential floral duration, nectar secretion and pollinator foraging in a protandrous species. American Journal of Botany 72: 303-310[CrossRef][ISI]
Dorn R. D. 1992 Vascular plants of Wyoming. Mountain West Publishing, Cheyenne, Wyoming, USA
Eckblad J. W. 1991 How many samples should be taken?. BioScience 41: 346-348[CrossRef][ISI]
FÆgri K. andL. van der Pijl 1979 The principles of pollination ecology, 3rd ed. Pergamon Press, Oxford, UK
Galil J. andL. Meiri 1992 Effect of wetting on the germinability of the pollen of Arisarum vulgare Targ.-Tozz. Flora 186: 145-152[ISI]
Goldsmith G. W. andA. L. Hafenrichter 1932 Anthokinetics: the physiology and ecology of floral movements. Publication Number 420. Carnegie Institution of Washington, W. F. Roberts Company, Washington, D.C., USA
Harder L. D. 1986 Effects of nectar concentration and flower depth on flower handling efficiency of bumble bees. Oecologia 69: 309-315[CrossRef][ISI]
Hart J. W. 1990 Plant tropisms and other growth movements. Unwin Hyman, London, UK
Heinrich B. 1979a Bumblebee economics. Harvard University Press, Cambridge, Massachusetts, USA
———. 1979b Resource heterogeneity and patterns of movement in foraging bumblebees. Oecologia 40: 235-246[CrossRef][ISI]
———. 1983 Insect foraging energetics. In C. E. Jones and R. J. Little [eds.], Handbook of experimental pollination biology, 187–214. Scientific and Academic Editions, New York, New York, USA
Hitchcock C. L. A. Cronquist M. Ownbey andJ. W. Thompson 1959 Vascular plants of the Pacific Northwest, part 4. University of Washington Press, Seattle, Washington, USA
Hunt C. B. 1974 Natural regions of the United States and Canada. W. H. Freeman, New York, New York, USA
Inouye D. W. D. E. Gill M. R. Dudash andC. B. Fenster 1994 A model and lexicon for pollen fate. American Journal of Botany 81: 1517-1530[CrossRef][ISI]
Kattenberg A. F. Giorgi H. Grassi G. A. Meehl J. F. B. Mitchell R. J. Stouffer T. Tokioka A. J. Weaver andT. M. L. Wigley 1996 Climate models: projections of future climate. In J. T. Houghton, L. G. Meira Filho, B. A. Callender, H. Harris, A. Kattenberg, and K. Maskell [eds.], Climate change 1995. The science of climate change, 285–358. Cambridge University Press, Cambridge, UK
Kearns C. A. andD. W. Inouye 1993 Techniques for pollination biologists. University Press of Colorado, Niwot, Colorado, USA
Lindsey A. A. 1940 Floral anatomy in the Gentianaceae. American Journal of Botany 27: 640-652[CrossRef]
Marden J. H. 1984 Remote perception of floral nectar by bumblebees. Oecologia 64: 232-240[CrossRef][ISI]
Martner B. E. 1986 Wyoming climate atlas. University of Nebraska Press, Lincoln, Nebraska, USA
Nelson B. E. 1984 Vascular plants of the Medicine Bow Range. Jelm Mt. Press, Laramie, Wyoming, USA
Percival M. S. 1947 Pollen collection by Apis mellifera.. New Phytologist 46: 142-165[CrossRef][ISI]
———. 1955 The presentation of pollen in certain angiosperms and its collection by Apis mellifera.. New Phytologist 54: 353-368[CrossRef]
———. 1965 Floral biology. Pergamon Press, Oxford, UK
Polunin N. 1959 Circumpolar arctic flora. Oxford University Press, London, UK
Raven P. H. andG. B. Johnson 1989 Biology. Times Mirror/Mosby College Publishing, St. Louis, Missouri, USA
Reifsnyder W. E. 1980 Weathering the wilderness: the Sierra Club guide to practical meteorology. Sierra Club Books, San Francisco, California, USA
SAS. 1989 SAS proprietary software release, Version 6.06.01. SAS Institute, Cary, North Carolina, USA
Scoggan H. J. 1979 The flora of Canada, part 4, 1117–1711. National Museum of Natural Sciences, Ottawa, Ontario, Canada
Scott R. W. 1996 The alpine flora of the Rocky Mountains,. vol. 1, The middle Rockies. University of Utah Press, Salt Lake City, Utah, USA
Simons P. 1992 The action plant: movement and nervous behavior in plants. Blackwell, Oxford, UK
Simpson B. B. andJ. L. Neff 1983 Evolution and diversity of floral rewards. In C. E. Jones and R. J. Little [eds.], Handbook of experimental pollination biology, 142–159. Scientific and Academic Editions, New York, New York, USA
Stanton M. L. andC. Galen 1989 Consequences of flower heliotropism for reproduction in an alpine buttercup (Ranunculus adoneus). Oecologia 78: 477-485[CrossRef][ISI]
Waddington K. D. 1983 Foraging behavior of pollinators. In L. Real [ed.], Pollination biology, 213–239. Academic Press, Orlando, Florida, USA
Zar J. H. 1999 Biostatistical analysis, 4th ed. Prentice Hall, Upper Saddle River, New Jersey, USA
This article has been cited by other articles:
|
|
 |
|
  C. Galen and M. L. Stanton Sunny-side up: flower heliotropism as a source of parental environmental effects on pollen quality and performance in the snow buttercup, Ranunculus adoneus (Ranunculaceae) Am. J. Botany, May 1, 2003; 90(5): 724 - 729. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-Q. Huang, Y. Takahashi, and A. Dafni Why does the flower stalk of Pulsatilla cernua (Ranunculaceae) bend during anthesis? Am. J. Botany, October 1, 2002; 89(10): 1599 - 1603. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
