• The effects of enhanced UV-B radiation on reproductive and pollination success were investigated in the Mediterranean annual Malcolmia maritima.
• Plants were exposed in the field to ambient or ambient plus supplemental UV-B radiation (biologically equivalent to a 15% ozone depletion over Patras, Greece, 38°14′ N, 21°44′ E) up to leaf senescence and fruit maturation.
• UV-B radiation had no effect on stem and fruit biomass, anthesis time and duration and flower number. However, flower diameter, nectary volume and nectar amount per flower (but not nectar concentration) were significantly increased by supplemental UV-B radiation. In addition UV-B treated plants showed higher reproductive success (i.e. lower abortion rates and higher fruit to flower ratio) and a trend to higher pollination success (i.e. increased number of seeds per fruit). As a result, the seed yield was increased. Seed mass, seed germination and early seedling growth were not affected by UV-B treatment of mother plants.
• It is suggested that the UV-B induced changes in flower attributes might have affected pollinators’ behaviour in a way that improved the fitness of M. maritima.
In spite of the international efforts to limit emission of ozone depleting substances, its stratospheric concentration may continue to decrease within the next decades (Madronich et al., 1995). The on-going global warming may further contribute towards this direction (Shindell et al., 1998). The concomitant increase of solar UV-B radiation reaching the surface of the earth has prompted much research on the effects that this surplus radiation may have on plants. Yet, the initial predictions for considerable negative effects on plant growth and photosynthesis have not been fully justified, as shown by field experiments using realistic doses and spectral distributions of radiation (Fiscus & Booker, 1995; Björn et al., 1997; Rozema et al., 1997). Mediterranean plants in particular, seem to be very tolerant to UV-B radiation (Manetas, 1999). However, even in the absence of any UV-B radiation effects on CO2 assimilation or biomass accumulation, the plants may respond with morphogenetic changes of various kinds (Barnes et al., 1990; Cen & Bornman, 1993; Beggs & Wellman, 1994; Ballaréet al., 1995). The responses in some cases are subtle and neutral for plant growth under optimal conditions, yet they may become critical under adversity or at specific developmental stages (Rozema et al., 1997; Manetas, 1999).
The above general conclusions are based on field studies investigating the effects of enhanced UV-B radiation on aspects of vegetative growth, morphology and physiology. However, the corresponding effects on reproductive biology are comparatively under-investigated, although the fitness of a plant depends mainly on its successful reproduction. One may argue at this point that if solar UV-B radiation is a morphogenetic factor, its effects may be better manifested as changes in flower phenology and/or flower attributes, further influencing pollination success and plant fecundity. Indeed, in-door experiments revealed considerable (yet variable) influences of UV-B radiation on various parameters, potentially related to the reproductive success (Musil & Wand, 1994; Deckmyn & Impens, 1995; Musil, 1995; Day & Demchik, 1996; Demchik & Day, 1996; Fendheim & Conner, 1996; Mark et al., 1996; van de Staaij et al., 1997; Deckmyn & Impens, 1998). Although these studies do indicate that UV-B radiation may be a critical environmental factor influencing plant reproductive development, extrapolation of their results to the field situation is risky, since the spectral balance between UV-B/UV-A and visible radiation is not realistic. This spectral balance is known to modify the responses of plants to UV-B radiation (Caldwell et al., 1994). Concerning field investigations, these have focused on seed yields of economically important cultivated species with characteristically variable results. Thus, seed yield in soybean was suppressed, stimulated or remained unchanged depending on variety, year of study or additional watering (Murali & Teramura, 1986; Sinclair et al., 1990; Teramura et al., 1990; Miller et al., 1994) while pea yield was reduced by UV-B radiation (Mepsted et al., 1996).
In the few wild plants studied so far under field conditions and with realistic ozone depletion scenarios, a consistent trend towards an increase in flower number was observed in the Mediterranean Mentha spicata (Grammatikopoulos et al., 1998) and in Colobanthus quitensis and Deschampsia antarctica in the Antarctic Peninsula (Day et al., 1999). In addition, Gwynn-Jones et al. (1997) found an increase in berry production in the subarctic Vaccinium myrtilus. However, the seed yield was not measured in these investigations. In another study with the Mediterranean shrub Cistus creticus, the number of flowers was not affected, yet pollination success was improved and seed yield increased (Stephanou & Manetas, 1998). This was ascribed to an increase in pollinators’ visit duration (Stephanou et al., 2000).
Based on the above, we grew the Mediterranean winter annual Malcolmia maritima under ambient or ambient plus supplemental UV-B radiation biologically equivalent to a 15% ozone depletion over Patras, Greece. We investigated in detail the demography of flowering and some selected flower properties, with the aim of correlating the above parameters with possible changes in the final reproductive success.
Materials and Methods
Plant material and radiation conditions
Malcolmia maritima (L.) R. Br. (Brassicaceae) is a winter annual species, self-sown in Greece and Albania (possibly further north along the Adriatic coast), growing on open rocky, sandy and gravelly habitats as well as on roadsides (Stork, 1972; Georgiou, 2001). Based on indirect evidence (showy flower, presence of nectar and nectar guides, sticky pollen) Stork (1972) considers M. maritima to be an allogamous species. In January 2000, young seedlings of M. maritima were carefully excavated from their natural environment (NW Peloponnisos, Nomos Ilias, Greece: locality Kounoupeli, 38°05′ N, 21°20′ E, alt. 25 m) put in 3 l clay pots (one plant per pot) with local soil and transferred to a small open nursery in the vicinity of the experimental site (Patras University Campus, 38°14′ N, 21°44′ E, alt. 125 m). Their growth was monitored for 25 d and on February 15, 2000, 56 similar seedlings (based on the rosette diameter) were selected for further experimentation. The plants were randomly assigned into eight plots (four controls, four UV-B, seven plants per plot), located on a horizontal, shade free area with no reflecting objects around. On each plot, a metal frame provided support for six Q-Panel UV-B 313 fluorescent tubes (The Q-Panel Company, Cleveland, OH, USA) suspended approx. 170 cm above plant apex and 20 cm apart from each other. In control plots, the tubes were wrapped in UV-B absorbing 0.1 mm Mylar (DuPont Co., Wilmington, DE, USA) and in UV-B plots in UV-C absorbing 0.1 mm cellulose acetate film (Courtaulds Chemicals, Derby, UK) which was preburnt in the laboratory for 12 h with the same tubes to stabilize its transmission properties. Both films were replaced after approx. 20 h of lamp operation. Attenuation of solar photosynthetically active radiation (PAR) by tubes and tube supports was < 10% on a daily basis, as shown by preliminary measurements with a linear (80 cm) quantum sensor (Decagon Sunfleck Ceptometer, Pullman, WA, USA). The additional PAR and UV-A (320–400 nm, measured with an Optronic OL 752 spectroradiometer, see below) radiation from the tubes at plant level was < 1‰ and 1%, respectively, of the corresponding solar radiation on a daily basis (Manetas et al., 1997). In this way, all plants were receiving the same levels of slightly subambient solar radiation and UV-B plants were receiving supplemental UV-B radiation simulating 15% ozone depletion over Patras (38°14′ N, 21°44′ E). This was accomplished as follows: the absolute UV-B (280–320 nm) spectral irradiance of the tubes at plant height was measured (Optronic OL 752 spectroradiometer, Orlando, FL, USA) at night (before placing the plants under the frames) and weighted with the generalized plant action spectrum normalized at 300 nm (Caldwell, 1971). The result was used in conjunction with the computer program of Björn & Murphy (1985) in order to calculate the time needed for the tubes to be on each day to obtain the required UV-B supplementation. The tubes were on and off in two steps centred at solar noon and daily tube duration was modified every 15 d to follow the natural march of ambient UV-B radiation change. Thus, at the beginning of the experiment, half of the tubes were on at 11.00–12.30 and the rest at 11.30–13.00 every day. Corresponding time of tube function at the end of the experiment was 8.30–15.00 and 9.00–15.30 every day, respectively. The experiment was started on 15 February and was terminated after leaf fall and fruit maturation on 5 June, 2000. Termination date was decided on the basis of previous experience with the same plant (O. Georgiou, unpublished); since one of the main objectives was to measure seed yield and seed germination rates, enough time was allowed after termination of anthesis for complete seed maturation in planta and under the above experimental conditions. No seed loss occurred during this period, since fruits remained closed. During the experimental period, the plants were receiving 100 ml of tap water per pot every second day in addition to natural precipitation, which was 210 mm for the whole experimental duration.
The field site was visited every morning for the whole period of anthesis. Each new flower was tagged by a small white thread and the dates of its commencement and termination were registered. Corresponding criteria of flower commencement and termination were the spread out of petals and the growth of the pistil through the petal claw ‘pipe’, respectively. The latter usually coincides with flower senescence (Stork, 1972). When needed (see measurements), an equal number of flowers from each plant/plot were removed for laboratory measurements. The number of removed flowers was c. 3% of the total. Removed flowers were not used in the calculation of reproductive success (i.e. the ratio of fruits to flowers). Since flower life span decreased with time (with the same pattern under ambient and enhanced UV-B radiation) care was taken to collect flowers which had covered about the same proportion of their life duration on each sampling date.
Flower diameters were measured in situ using a ruler without removing the flowers. For nectar amount, cut flowers were put into airtight small plastic bags and immediately transferred to the laboratory, where nectar was extracted and its volume measured with graduated microcapillaries. Its sugar concentration was assessed with a microrefractometer (Bellingham & Stanley, Kent, UK). The cushion-like nectaries are present on the upper side of the bases of the two lateral (shorter) stamens (Stork, 1972). For nectary size, the sepals, petals and stamens were removed and the exposed area was examined under the stereoscope and photographed with a scale. The photographs were scanned and nectary projected area was calculated by using the Image Tool (version 1.25, UTHSCSA) computer program. The number of ovules was estimated as follows: on two sampling dates, equal number of fruits/treatment (deriving from flowers of the same age) were collected and dissected, the seeds and the nonpollinated ovules were counted and the total was used as the initial number of ovules.
After harvesting, plants were divided into stems and fruits and their dry mass was measured after air-drying to constant weight. In addition, the seeds from each fruit were taken out, counted and weighted.
For seed germination, 10 randomly selected seeds from each individual (i.e. 280 seeds per treatment) were pooled, surface sterilized in running tap water and evenly layered on filter paper in 9 cm Petri dishes (5 dishes per treatment, 56 seeds per dish) and watered with 10 ml de-ionized water. Germination tests were performed in a growth chamber under 45 µmol m−2 s−1 PAR, 17°C night temperature/27°C day temperature and a photoperiod of 12 h. After 7 d, germinated seeds were counted, the seedlings divided into shoot and radicle, oven dried at 80°C for 24 h and their dry mass was measured.
Sampling and statistics
The dry mass of various plant parts and the number of fruits and seeds were measured for all plants at harvest. Also, all plants were used for the measurement of flower number. For flower diameter, all mature flowers having the same age at the indicated sampling dates were used. For nectary projected area, 16 mature, same aged flowers per treatment (four flowers per plot, sampled from four randomly selected individuals) were used at the indicated dates. The same procedure was followed for nectar volume, with an equal number of flowers within each sampling date but different total number of flowers between different dates (Fig. 1e).
All statistical tests were performed with SPSS 9.0 statistical package. Normality (Kolmogorov-Smirnov test) and homogeneity (Levene's test) of variables were tested but data transformations were not required. For measurements taken after plant harvest, all test plants (i.e. seven plants per plot × four plots per treatment) were used. Since pots were not rotated between frames (plots), a nested ANOVA, using Type III square sums, was adopted to evaluate both within and between treatment variation and minimize the potential influence of site effects. Becuase within treatment differences were not found (P > 0.1), the results of between treatments comparisons only are given in the figures.
For parameters measured after sampling at different dates, the sample size at some of the dates was unavoidably small, thus a t-test with the pot as the experimental unit was applied. When all sampling dates were considered (Fig. 1e, nectar volume), a paired t-test was employed because nectar volume depends strongly on temperature and atmospheric humidity (Corbet et al., 1979).
Under UV-B supplementation, there was a tendency for increased stem and fruit dry mass (by 15% and 20%, respectively) measured at plant harvest (Fig. 1a). Yet, this trend was not statistically significant. Dry mass of leaves was not measured since leaves had already fallen at that time. Flower demography was also not affected by enhanced UV-B radiation. Thus, the dates of commencement and termination of anthesis and the individual flower duration (not shown) as well as the total number of produced flowers, were the same (Fig. 1b). Also, a 15% increasing trend in the number of fruits (Fig. 1b) was not significant. However, the reproductive success under UV-B supplementation (as judged by the ratio of fruits to flowers) was significantly increased (Fig. 1b). Accordingly, the calculated flower abortion rates were considerably lower in UV-B treated plants (16 vs 25%).
Supplemental UV-B radiation had a small (c. 9%) yet consistent and statistically significant positive effect on flower diameter (Fig. 1c). In addition, examination of the flower morphology under the stereoscope revealed a considerable swelling of the nectaries with 85% and 52% increases in projected nectary area for the two sampling dates, respectively (Fig. 1d). The increase in nectary volume prompted us to measure the nectar amount. As shown in Fig. 1(e), considerable variation between individual flowers and sampling dates was observed. However a c. 50% statistically significant increase in nectar volume of UV-B treated plants was found on 12/05 and considerable increasing trends for the two other sampling dates. Applying a paired t-test for all sampling dates indicated a significant increase in nectar volume under UV-B supplementation (Fig. 1e). Sugar concentration was not affected (46.70 ± 6.06 and 46.87 ± 6.06 w/w for control and UV-B plants, respectively).
Finally, supplemental UV-B radiation caused an 11% (yet nonsignificant) increase in the number of seeds per fruit (Fig. 1f). Since the number of ovules was not affected by UV-B radiation (43.51 ± 6.96 and 42.91 ± 7.78 for control and UV-B plants, respectively; n = 35), we may consider a tendency for improvement of pollination success by supplemental UV-B radiation. The combination of higher reproductive and pollination success resulted in a significant 30% increase in seed yield (Fig. 1f). Seed mass, seed germination rates and early seedling growth (assessed from radicle length and mass measurements) were independent of UV-B exposure level of mother plants (not shown).
The results of the present investigation do not confirm the hypothesis that supplemental UV-B radiation could affect flower phenology and demography in Malcolmia maritima. Theoretically, this hypothesis could be based on the fact that in UV-B supplementation experiments performed during the period of ascending ambient UV-B fluencies and applying a constant percent of above ambient UV-B radiation supplement, the plants receive mild UV-B radiation doses a few days or weeks in advance of their normal occurrence, depending on the extent of the applied ozone depletion scheme (Manetas, 1999). This surplus radiation, acting as a premature environmental signal, could alter the timing (Ziska et al., 1992; Saile-Mark & Tevini, 1997) or the number of flowers (Musil, 1995; Grammatikopoulos et al., 1998; Day et al., 1999), having important consequences for synchronization and availability to pollinators and therefore for plant fitness. The insensibility of flower demography and phenology in Malcolmia maritima, compared with the previously cited examples with other plant species, indicates that the UV-B radiation effects on reproduction are species-specific. The same is true for growth responses emphasizing the possibility of altered inter–species competitive interactions at the ecosystem level (Rozema et al., 1997).
The results, however, showed that UV-B radiation did affect some flower attributes potentially related to pollination success. Thus, the increased flower diameter can be considered as an improved optical signal affecting the frequency of pollinators’ visits (Proctor et al., 1996). In addition, the enlarged nectary (also observed in Cistus creticus, Stephanou et al., 2000) and the concomitant increase in nectar volume could work towards the same direction, further increasing not only the frequency of visits (Heinrich, 1975; Real & Rathcke, 1991) but their duration as well (Zimmerman, 1988; Manetas & Petropoulou, 2000). Both the frequency and the duration of pollinator visits are positively correlated with the number of pollen grains deposited on the stigma (Thomson & Plowright, 1980) and, accordingly, to fruit set and seed yield (Real & Rathcke, 1991; Pyke et al., 1988; Manetas & Petropoulou, 2000). In our case, pollinator behaviour was not monitored, since the visiting insects had a very small size. We can reasonably assume, however, that their behaviour might have been changed in accordance with the observed improvements in both the optical signals and the offered rewards by the UV-B treated plants. This assumption is strengthened by the decreased abortion rates and the increased number of seeds per fruit and total seed yields under UV-B supplementation, all implying an improved pollination success. Alternatively, one could speculate that although the optical signals and the increased rewards were encouraging for pollinators, their behaviour could also be directly modified by the UV-B supplementation field. Indeed, many arthropods can perceive UV-B radiation (Tovée, 1995) and Mazza et al. (1999) showed that phytophagous thrips avoid solar UV-B. In this case, the tubes used for field UV-B enhancements may act antagonistically to flowers for pollinators attraction, tending to reduce rather than increase visit frequencies and duration. It is worth noting at this point that field video-recording of insect behaviour on nectar-rich and nectar-poor flowers of Cistus creticus with the UV-B tubes on or off, showed that both pollinating bees and nectar thieves prolonged the time spent on nectar-rich flowers only when the UV-B tubes were off (Stephanou et al., 2000). This indicates that the insects were probably annoyed by supplemental UV-B radiation. This may be true for herbivorous insects as well, since filtering out solar UV-B radiation increased the leaf area consumed by moth larvae in Gunnera magellanica (Rousseaux et al., 1998). This change could not be ascribed to an alteration of leaf nutritional value. Based on these grounds, we consider that the UV-B induced increase in pollination success of Malcolmia maritima was possibly mediated by corresponding positive reaction of pollinators to the improved optical signals and quantity of rewards.
An increase of seed yield does not necessarily imply an improvement of plant fitness. For example, in glasshouse experiments it was shown that although seed yield in some plants was increased by UV-B radiation, their germination ability was dramatically reduced (Musil & Wand, 1994; Musil, 1995). In our experiment (and also in the case of field grown Cistus creticus, Stephanou & Manetas, 1998), the germination and early seedling growth were independent of UV-B treatment of mother plants.
In conclusion, this study showed that enhanced UV-B radiation can improve the fitness of Malcolmia maritima by increasing its fecundity and contribution to the seed bank.