| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ecology |
2Department of Biological Sciences, University of Missouri–Columbia, Tucker Hall, Columbia, Missouri 65201 USA; 3Department of Biological Sciences, The Open University, Walton Hall, Milton Keynes, MK7 6AA, UK
Received for publication August 9, 2004. Accepted for publication May 17, 2005.
ABSTRACT
Nicotiana section Alatae exhibits great diversity among species in floral morphology, mating system, and predominant pollinators. As a first step towards estimating nectar's role in floral evolution, we studied nectar traits to determine whether they vary in association with predominant pollinators and mating system. Daily phenology determines when nectar becomes available to pollinators and differed between hummingbird- and moth-pollinated species. Nectar volume and concentration varied significantly among most species and pollinator groups, but were inversely correlated, so that total energy was similar among most species. In general, nectar volume was positively correlated with corolla length. The autogamous species, N. plumbaginifolia, had a nectar volume that matched expectations based on corolla length, but with lower concentration and total energy than predicted by corolla length, while nectar volume was lower than predicted by corolla length in the autogamous population of N. longiflora. Sugar and amino acid components (determined through HPLC) were similar among species, although differences did exist. The nectar of most species was sucrose-dominant, but the autogamous N. plumbaginifolia had nectar that contained similar proportions of sucrose, glucose, and fructose. Many nectar traits varied in association with the predominant pollinators and, in some cases, with the mating system.
Key Words: amino acids • Argentina • Brazil • hummingbirds • moths • nectar • Nicotiana • Solanaceae
Nectar chemistry is an important component of floral biology. Nectar drives pollination efforts by being the primary floral reward for most pollinators. Pollinators often exhibit a preference for certain types of nectar over others (Hainsworth and Wolf, 1976
; Stiles, 1976
; Pyke and Waser, 1981
; Tamm and Gass, 1986
). If pollinator preferences for certain nectar traits are strong enough, it may be possible for pollinator-mediated selection to cause ethological isolation, which has the potential to promote species divergence or maintenance during secondary contact (Grant, 1994
). But pollinator roles in speciation have been called into question recently, mainly because the majority of plant–pollinator associations are generalized, while specialization is thought to be necessary to promote speciation (Ollerton, 1996
; Waser et al., 1996
; Waser, 1998
).
Pollinators are known to respond to floral morphology (Cresswell and Galen, 1991
; Schemske and Bradshaw, 1999
; Ippolito, 2000
; Galen and Cuba, 2001
), color (Waser and Price, 1981
; Jones and Reithel, 2001
) and even nectar-related quantitative trait loci (Schemske and Bradshaw, 1999
). Many pollinators preferentially visit particular flower types, making it possible to predict the primary pollinator of a flower by some of its distinguishing characteristics. Grant and Grant (1968)
described flowers that are adapted primarily for the feeding and pollination of hummingbirds as pendulous, solitary or loosely clustered, having a thick-walled, red (or red with yellow) corolla, yielding large quantities of nectar at the base of a long, stout floral tube. "Hummingbird flowers" also generally lack a detectable odor (Baker, 1961
; Grant, 1966
; Grant and Grant, 1968
; Raven, 1972
). On the other hand, nocturnal moth-pollinated flowers are characterized as having corollas that are white or pale in color and as emitting a strong, sweet scent when open, which is usually in the evening or at night (Baker, 1961
; Percival, 1965
; Faegri and van der Pijl, 1966
). Flowers adapted primarily for pollination by hawkmoths have a longer, more slender floral tube than typical "hummingbird flowers" (Grant and Temeles, 1992
; Ippolito, 2000
). Smaller perching moths are attracted to typical "moth flowers," but with shorter corolla tubes that fit their proboscis length (Faegri and van der Pijl, 1966
).
Nectar traits can affect pollinator behavior (Hainsworth and Wolf, 1976
; Stiles, 1976
; Pyke and Waser, 1981
; Zimmerman, 1983
; Galen and Plowright, 1985
; Tamm and Gass, 1986
; Cresswell and Galen, 1991
; Martínez del Rio et al., 1992
; Hodges, 1995
; Meléndez-Ackerman et al., 1997
; Schemske and Bradshaw, 1999
) and, presumably, pollinator behavior can affect the evolution of nectar traits. From an outcrossing plant's perspective, flowers are most likely to be effectively pollinated when nectar reward is abundant enough to attract the pollinator, but small enough to force the pollinator to make numerous plant to plant visits (Heinrich and Raven, 1972
; Heinrich, 1975
; Baker, 1975
). Large per-plant nectar rewards can increase pollinator visitation on a single plant, increasing the chances of geitonogamy (Galen and Plowright, 1985
; Real and Rathcke, 1991
; Hodges, 1995
; Ferdy and Smithson, 2002
). But pollinators are expected to prefer a nectar volume and concentration that optimizes foraging efficiency (Baker, 1975
; Hainsworth and Wolf, 1976
). Autogamous species, which are less dependent upon pollinator visitation, may evolve to produce less nectar than outcrossing species (Spira, 1980
).
Many forces could affect the evolution of nectar traits, including environmental conditions, the plants' energy budget, water relations, and coevolution with nectar robbers, florivores, and pollinators (Galen, 1999
, 2000
). Evolution due to any of these pressures may be constrained by lack of genetic variation in a population genetic sense, a phylogenetic sense, or due to antagonistic pleiotropy. Pollinators are unique in their evolutionary effects on plant traits because they not only contribute to plant fitness, but are also agents of gene flow and so could induce a phylogenetic split if different pollinators prefer different nectar traits (Grant, 1994
). However, the occurrence of this scenario has been recently questioned because of the large number of species and guilds of pollinators that visit many flowers (Ollerton, 1996
; Waser et al., 1996
; Waser, 1998
). We chose to study a group that seemed to be morphologically and phenologically adapted to hawkmoth, small moth, hummingbird, and autogamous pollination.
As a first step toward understanding pollinators' roles in the evolution of nectar traits, we investigated the associations between presumed pollination syndromes and several nectar traits in Nicotiana section Alatae (Fig. 1). Although it may be viewed as an oversimplification to interpret data based on pollination syndromes (Waser et al., 1996
), identification of trends or differences among and between plant species with different pollinators can aid in the understanding of floral diversification mechanisms (Fenster et al., 2004
). The likelihood that pollinators were important in the evolution of these species would be increased if the nectar traits were shown to vary in association with the pollinators and their preferences. Phylogenetic, quantitative genetic, and field experiments are underway separately to further test the importance of pollinator-mediated floral evolution. The primary pollinator associations in this group are hawkmoths (2 spp.), hummingbirds (4 spp.), and small settling moths (1 sp.). Autogamy is also found in one species that is derived from a hawkmoth-pollinated ancestor and still bears some hawkmoth syndrome features (Ippolito, 2000
; Chase et al., 2003
; J. Murfett and T. Holtsford, unpublished data) and in one population of an otherwise hawkmoth-pollinated species. We also evaluated the accessibility of nectar through daily phenology measurements for these eight Nicotiana species in order to determine whether nectar presentation matched periods of pollinator activity.
|
Study system
Nicotiana section Alatae is a monophyletic group (Ippolito, 2000
; Chase et al., 2003
) comprised of seven or eight species (Fig. 1). The relationship of these species could not be completely resolved through ITS sequences, although species with the same chromosome number are more closely related (Ippolito, 2000
; Chase et al., 2003
). The self-incompatible species, N. alata Link and Otto, N. forgetiana hort. ex Hemsl., N. langsdorffii Schrank, N. bonariensis Lehm., the newly described N. mutabilis Stehmann and Semir (Stehmann et al., 2002
), and the putative species "Rastroensis" (chromosome n = 9), are restricted to southeastern Brazil, Paraguay, Uruguay and eastern Argentina. The self-compatible species, N. longiflora Cav. and N. plumbaginifolia Viv. (chromosome n = 10), have a more extensive range. Nicotiana longiflora is found in northern Argentina, southern Bolivia, Paraguay and Uruguay. Nicotiana plumbaginifolia, which evolved from N. longiflora (nrITS: Ippolito, 2000
; 256 ISSR bands: J. Murfett and T. Holtsford, unpublished data), is autogamous and weedy. It can be found from northwestern Argentina, north through Central America into Mexico. It has also been found on multiple Caribbean islands, as well as in India (Goodspeed, 1954
).
More than 200 h of field observations over 5 years suggest that there are three different pollinator groups for these Nicotiana species (Ippolito et al., 2004
; T. Holtsford, R. Kaczorowski, and A. Ippolito, unpublished data). Hawkmoths are the predominant (most common) pollinators of species with long-tubed, white flowers (N. alata and N. longiflora). Nicotiana plumbaginifolia has a similar floral morphology, suggesting hawkmoth visitation (Cocucci, 1988
), but is smaller in size and autogamous. Hummingbirds are the predominant pollinators of species with short-tubed flowers of various other colors, ranging from red to pink to greenish-yellow (Rastroensis, N. forgetiana, N. mutabilis, and N. langsdorffii), but Halictid bees have also been seen visiting, and apparently collecting pollen from, all four species (R. Kaczorowski, unpublished data). Bumblebees have also been seen visiting, and apparently collecting nectar from, N. langsdorffii in one of five populations observed (T. Holtsford, unpublished data) and N. mutabilis in one population observed (R. Kaczorowski, unpublished data). Small perching moths, which land on the long lower limbs of N. bonariensis, are thought to be the predominant pollinators of this species. These moths have probosci that match the short corolla tube length very well, but few such moths (nor any other pollinator) have been observed in the field in over 30 h of observation in four populations. Florivorous beetle larvae can be common inside N. bonariensis flowers in some populations, but their role in pollination was not apparent (R. Kaczorowski, personal observation).
Experimental design
The 137 Nicotiana plants used for the nectar volume and concentration experiments were grown beginning in late August 2000 from seed collected from various populations in Brazil and Argentina. Beginning in May 2002, a new set of plants was grown for the sugar and amino acid determinations, of which 118 plants survived. Multiple populations within each species were sampled, when available (Table 1). The plants were raised in 3.8-L pots in a common greenhouse environment (14-h days at c. 24°C and 10-h nights at c. 13°C at the University of Missouri–Columbia) as part of a randomized complete block design. The design included three blocks (different benches within the same greenhouse bay), each of which contained one progeny from three different maternal plants from up to three populations within each species. The plants were watered and fertilized with Peters Pro 20-10-20 (Scotts Co., Marysville, Ohio, USA) as necessary and regularly pruned to keep plant size manageable. Data for this study were collected between February 2001 and February 2003.
|
Nectar volume and concentration
Plants from one of three blocks were sampled on any given day for one of the three relative flower ages (approximately the time of anthesis [0 h], 12, and 24 h after opening). Day-to-day variation was therefore included in the block term. For the 0 and 24-h measurements, nectar collection started in the early afternoon (about 1400 CST) with random sampling of all N. langsdorffii plants on that block, then N. mutabilis, and then N. forgetiana. Rastroensis was considered a variant of N. forgetiana at the time of this experiment and therefore randomly sampled at the same time as N. forgetiana. Random sampling of all nocturnal species began around dusk (approximately 1730 hours). For the 12-h measurement, sampling started in the early morning (about 0200 hours) and followed the same progression.
At least three flowers from each plant were destructively sampled for nectar volume and concentration measurements. For most species, the calyx and corolla were separated and the corolla tube gently squeezed to bring the nectar to the base of the tube, where it was collected. Nicotiana bonariensis gave almost no nectar when sampled this way; dissection of flowers was necessary to collect the trace amounts of nectar along the corolla tube. Therefore, nectar was resampled from all N. bonariensis plants on different days than the other species (a day each for 0 and 24-h measurements, 12-h measurements were not possible because reduced turgor during daylight hours complicated dissection and collection). Other species' nectar sampled with N. bonariensis did not differ from the results (presented later), so we presume that the N. bonariensis measurements are comparable. Nectar was collected with glass micropipette tubes, and the volume was recorded. Nectar samples from individual flowers on the same plant were pooled to obtain a single concentration measurement from a temperature-compensated refractometer (Leica Microsystems, Buffalo, New York, USA, but Bellingham and Stanley, Lawrenceville, Georgia, USA, with a low volume sample window, used for N. bonariensis). Dilutions were performed as necessary to keep the concentration readings within the range of the refractometer. The refractometers measure concentration as the percentage solids in solution (sucrose equivalents, wt/wt, keeping manufacturer's usage).
Nectar volume and concentration analysis
We used population, block, and time in a randomized complete block (RCB) split-plot design (PROC MIXED in SAS 6.12 [SAS Institute Inc., Cary, North Carolina, USA]) to analyze average volume, nectar concentration, and total energy (average nectar volume multiplied by nectar concentration, which was first converted from wt/wt to wt/vol, as suggested by Bolton et al., 1979
). A second analysis, without the 12-h measurements, was also run to estimate the relationship of N. bonariensis to the other species. Two species were represented by only one population each, so a population nested within species design could not be used. We formulated contrasts that grouped populations within species to test for among species differences, as well as grouping species within pollinator groups to test for among pollinator group differences. Because the variances were unequal among populations and did not equalize with various transformations, the data were ranked. Because multiple analyses were being performed, a more rigorous alpha of 0.01 was chosen to reduce type II errors. To make pairwise comparisons of slopes and quadratic terms of nectar accumulation and concentration changes, we fit least square models and tested the differences of first and second-order terms using software from UMC Statistics Department (Critical SS for each contrast = Error MS x F from the RCB split-plot analysis described). Nicotiana bonariensis could not be included in this testing because only two time measurements were taken for this species.
Sugar and amino acid composition
Four to six plants per species, from a different set of plants, were sampled for the sugar and amino acid analysis (Table 1). All flowers were sampled during the day of anthesis to minimize effects of flower aging that have been shown to affect amino acid concentrations in nectar samples (Gottsberger et al., 1990
; Petanidou et al., 1996
). All species were destructively sampled because it would otherwise be impossible to collect nectar from some species without contaminating with pollen or floral tissue. The flower to be sampled was taken off the plant, held upside-down, and cut below the nectar pool (the nectar pool can be seen through the corolla). This allowed the stamens to be removed, leaving full access to the nectar pool and minimizing possible contamination with pollen, which can release free amino acids in solution (Linskens and Schrauwen, 1969
). The nectar was withdrawn using 10-µL glass micropipette tubes (Drummond "Microcaps" [Drummond Scientific Co., Broomall, Pennsylvania, USA]). Care was taken to avoid touching the cut edges of the corolla. Each sample was aspirated into a glass chromatography vial using a microcap bulb. The samples were preserved by addition of 50 µL of 80% ethanol shortly afterwards and kept sealed with polytetrafluoroethylene push-fit tops. This formed a tight seal and prevented any loss of sample due to evaporation. On return to the laboratory, the sample vials were cooled in the refrigerator (at 4°C) to condense any vapor before opening the vial. Ten microliters of sample were removed and used for sugar analysis. The remainder was analyzed for amino acids.
The samples were prepared and the amino acids derivatized using the AccQtag protocol (Waters Corp., Milford, Massachusetts, USA; Cohen and Micheaud, 1993
) in a 0.02 M borate buffer (pH 8.59). High-performance liquid chromatography (HPLC) was performed, with standards every four samples, using the following equipment: Waters 712 WISP autosampler, Waters 600 pump controller, Waters 600 HPLC pump with 510 pump-heads (Waters Corp.). Separation was achieved using a Waters Novapak C18 (15 x 0.46 cm) cartridge with guard column. The binary solvent system was a 6 : 4 acetonitrile : water mix and a (triethylamine)-phosphate (pH 5.04) buffer. Detection was via a Waters 474 scanning fluorescent detector (excitation at 295 nm and detection at 350 nm). The system was monitored and data collected using the Waters Millenium32 software. Chromatograms were analyzed and compared to standards for identification of individual amino acids. Standard amino acids were made up to a concentration of 100 pmol/µL. In addition to all the protein-building amino acids, standards of hydroxyproline (hyp), ornithine (orn), taurine (tau), 
-aminobutyric acid (AABA) and 
-aminobutyric acid (GABA) were used. Peak areas were compared to standards to determine the concentration of individual amino acids. From these data, the total concentration of all amino acids was determined, and the proportion that each made to the total was also calculated as a percentage.
Sugars were analyzed using an Alltech Varex Mk III evaporative light scattering laser (ELSD) system (Alltech Associates, Deerfield, Illinois, USA) and HPLC. Each sample was diluted with 30 µL of eluant. Separation was achieved using an Alltech 525 pump and pulse dampener fitted with a 5-µL injection loop. The eluant system was a 3 : 1 acetonitrile/water mix with an isocratic flow of 0.5 mL/min through a Capital NH2 Optimal narrow-bore column (5-µm particle size, 250 x 3.2 mm; Capital HPLC Ltd., West Lothian, Scotland, UK) and C18 guard cartridge (4 x 3 mm, Phenomenex, Torrence, California, USA) . Detection was via an ELSD with a gas-flow rate of 3 L/ min and drift tube temperature of 105°C. Output was monitored on a Shimadzu C-RIB integrator (Shimadzu Corp., Columbia, Maryland, USA). Chromatograms were analyzed and compared to standards for identification. Peak areas were compared to standards to determine the concentration of individual sugars (sucrose, fructose, and glucose).
Sugar and amino acid analysis
Statistical analyses of nectar chemistry were carried out using Statistica (StatSoft, Tulsa, Oklahoma, USA). All concentration data were log transformed (natural logarithm), and proportion data were angular transformed (arcsine) to improve the distribution and homoscedasticity of the residuals. We used species as an independent variable in two separate multiple analysis of variance tests (MANOVA), one to analyze amino acid and another to analyze sugar data, although there could be a lack of independence. Among species differences in individual model components were analyzed subsequently using post-hoc tests (Tukey's "honestly significantly different" [HSD] test). Differences among pollinator groups were determined from post-hoc tests (Tukey HSD) in nested ANOVAs (separate for sugars and amino acids) in which species was nested within pollinator groups. Because multiple analyses were being performed a more rigorous alpha of 0.01 was chosen to reduce type II errors.
RESULTS
Daily phenology
Nicotiana langsdorffii generally opened in the early afternoon, and N. mutabilis opened between early and late afternoon, exhibiting diurnal anthesis (Fig. 2). Nicotiana forgetiana and Rastroensis generally opened in the late afternoon, exhibiting a crepuscular anthesis. The four white-flowered species (N. alata, N. longiflora, N. plumbaginifolia, and N. bonariensis) opened around dusk, exhibiting a nocturnal anthesis. Nicotiana langsdorffii, the species with the smallest flowers, was the earliest to begin opening and have all of its flowers fully open. It also had the greatest ability to avoid turgor loss, and thus all flowers remained open until senescence. Most flowers of Nicotiana mutabilis and Rastroensis also remained open after anthesis, but some lost turgor during the heat of the day. Nicotiana forgetiana flowers were quite susceptible to turgor loss during the day, although some usually remained open. The flowers of the nocturnal species (N. alata, N. longiflora, N. plumbaginifolia, and N. bonariensis) remained open throughout the night and lost turgor during the day. In N. forgetiana and the nocturnal species N. alata, N. longiflora, and N. bonariensis, the third- and fourth-day flowers opened earlier and stayed open longer. The flowers of N. plumbaginifolia remained open only two nights, while the flowers of other nocturnal species continued to open night after night, until the onset of senescence (about 4–7 days after anthesis). The brief flower life in N. plumbaginifolia is likely due to it being autogamous, because we observed that pollination shortened the flowering span of the other species (data not shown).
|
|
|
Many among-species contrasts for nectar volume and concentration were significantly different (at = 0.01), although some species grouped together with similar nectar volumes or concentrations (Fig. 4; only 24-h data shown). Hawkmoth-pollinated species (N. alata and N. longiflora) tended to have more nectar at a lower concentration than the hummingbird- and small moth-pollinated species. Pollinator group contrasts found that hawkmoth-visited species were significantly different from hummingbird-pollinated species in both average nectar volume and nectar concentration based on all time measurements (F1,30 = 118.67, P < 0.0001 and F1,30 = 983.24, P < 0.0001, respectively). The second analysis (0 and 24 h only) found that the small moth-pollinated N. bonariensis had significantly lower nectar volume than the hawkmoth-visited and hummingbird-pollinated species (F1,36 = 420.38, P < 0.0001 and F1,36 = 190.58, P < 0.0001, respectively) and significantly greater nectar concentration than the hawkmoth-visited species (F1,35 = 190.94, P < 0.0001), but not from the hummingbird-pollinated species (F1,35 = 4.41, P = 0.0429, where = 0.01). Nectar volume and concentration tend to be more similar among species with the same predominant pollinator than between species with different predominant pollinators.
|
Total energy
Total energy comparisons among pollination systems showed that hummingbird-pollinated species had significantly more total energy than hawkmoth-visited species (Fig. 5; F1,30 = 16.02, P = 0.0004), and the small moth-pollinated species had significantly less total energy than hawkmoth-visited and hummingbird-pollinated species based on all time measurements (F1,35 = 166.09, P < 0.0001 and F1,35 = 278.56, P < 0.0001, respectively). However, there was no difference in total energy between hawkmoth and hummingbird-pollinated species when the autogamous N. plumbaginifolia was excluded from the hawkmoth group (df = 105, t = –0.475, P = 0.318).
|
= 0.01), although many species grouped together with similar total energy (Fig. 5; only 24-h data shown). Nicotiana mutabilis had significantly more total energy than all other species, except N. forgetiana; it tends to produce more nectar than the other hummingbird-pollinated species, and the nectar is often more concentrated (Table 2). Rastroensis had significantly less total energy than the other hummingbird-pollinated species, mainly due to its lower nectar volumes (Table 2). Rastroensis had significantly lower total energy than N. forgetiana and significantly higher total energy than N. bonariensis, the two species most morphologically similar, and possibly most closely related, to this putative species. The second analysis (0 and 24 h only) found N. bonariensis to have significantly less total energy than all species, except N. plumbaginifolia (Fig. 5), because it produced only trace amounts of nectar (Table 2). There was a significant population effect (F15,30 = 7.23, P < 0.0001) and time effect (F2,61 = 103.68, P < 0.0001; data not shown) for total energy across all species, although the population by time interaction was not significant (F30,61 = 1.13, P = 0.3383). The only population effect within a species was found within N. longiflora, where the selfing Jujuy population had significantly less total energy in its nectar than both the Calilegua and Universidad Nacional de Nordeste populations (df = 30; t = 5.09 and t = 5.70, respectively; P < 0.0001; data not shown).
Reward scaled by flower size
To determine how total energy and its components (nectar volume and concentration) related to floral size, we examined the functional relationship between these three nectar traits and corolla length (Fig. 6). Nectar volume was positively correlated to corolla length (r2 = 0.8202, P < 0.0001; Fig. 6A), while nectar concentration was negatively correlated to corolla length (r2 = 0.6834, P < 0.0001; Fig. 6B). As a result, total energy was not significantly correlated with corolla length (r2 = 0.0238, P = 0.5544; Fig. 6C). Total energy in autogamous groups was lower than in other species, except the small moth-pollinated species, which produced only trace quantities of nectar. Bud-selfing N. plumbaginifolia has a low concentration for its size, while the autogamous Jujuy population of N. longiflora has a low nectar volume for its size (Fig. 6).
|
= 0.01), but some significant differences do exist (Fig. 7).
|
= 0.01). Sugar ratios (sucrose to [glucose + fructose]) ranged from 0.6 in N. plumbaginifolia to 2.7 in N. forgetiana (Fig. 7). Some pollinator groups were found to be significantly different from each other in their proportions of each sugar type. Sucrose makes up the greatest proportion of the sugar concentration in all pollinator groups, but hummingbird-pollinated species had significantly greater proportions of sucrose (at = 0.01) than the hawkmoth-visited species, which had significantly greater proportions of glucose than the hummingbird-pollinated species, while the small moth-pollinated and hawkmoth-visited species had significantly greater proportions of fructose than the hummingbird-pollinated species.
There was a significant species effect for total amino acid concentration (F7,40 = 8.806, P < 0.0001), which was not significantly different among most species (by Tukey HSD at = 0.01), though a few significant differences were found (Fig. 8). The only significantly different pollinator group in amino acid concentration was the small moth species, N. bonariensis, which had significantly more amino acids than the other pollinator groups.
|
= 0.01), while others did not (Table 3). The most common amino acid found in the species with higher nectar concentrations (N. mutabilis, N. langsdorffii, N. forgetiana, Rastroensis, and N. bonariensis) was proline (concentration means ± SE, in nMol/mL: 1.66 ± 0.70, 4.49 0.98, 0.74 ± 0.13, 0.77 ± 0.12, and 4.53 ± 0.79, respectively). Nicotiana longiflora also had a substantial proportion of proline (0.99 ± 0.32 nMol/ mL), but glutamine was its most abundant amino acid (1.54 ± 0.22 nMol/mL). Glutamine was also relatively abundant in N. mutabilis and N. langsdorffii (1.97 ± 1.81 nMol/mL and 2.22 ± 0.49 nMol/mL, respectively). The most predominant amino acid found in N. alata and N. plumbaginifolia was GABA (0.57 ± 0.09 nMol/mL and 0.89 ± 0.19 nMol/mL, respectively), which also contributed a substantial proportion of amino acids to N. forgetiana and Rastroensis (0.28 ± 0.07 nMol/ml and 0.38 ± 0.05 nMol/ml, respectively; see Table 3 for amino acid proportions).
|
The species of Nicotiana section Alatae showed variation in daily phenology and several nectar traits that corresponded to their presumed pollination syndromes. Species within pollinator groups are not necessarily independent of each other because they share phylogenetic history. The phylogeny for Nicotiana section Alatae has not yet been fully resolved. It is possible that differences are due to shared history, rather than the pollination syndrome. This should be kept in mind while interpreting the results of all post-hoc tests performed on presumed pollination system.
Daily phenology
Daily phenology dictates the availability of nectar to effective pollinators. The timing of anthesis usually coincides with the time of day that the pollinators are actively foraging (Percival, 1974
). Thus, one can expect diurnal plant species to have diurnal pollinators (hummingbirds), while nocturnal plant species have nocturnal pollinators (moths). Nicotiana forgetiana had nectar traits similar to the other hummingbird-pollinated species, but its phenology differed only slightly from the moth-pollinated species. Hummingbird activity is at a peak in the evening when most N. forgetiana flowers are open, but another possible reason for N. forgetiana's phenology might be linked to its phylogenetic history. Although the ancestral pollination syndrome is unclear in section Alatae, the next most closely related group is section Suaveolens (Ippolito, 2000
; Chase et al., 2003
), a clade whose white, scented flowers suggest moth pollination. That, coupled with biogeography (N. forgetiana's range seems to be encompassed by that of N. alata), suggests that N. forgetiana has evolved from a hawkmoth-pollinated ancestor, which was likely N. alata itself or its immediate progenitor, and has retained a similar daily phenology. Although the daily phenology for each species was recorded for a relatively small number of flowers during one time of the year, personal observations suggest that the patterns hold throughout the year in the greenhouse and during the flowering season in natural populations (R. Kaczorowski, A. Ippolito, and T. Holtsford, unpublished data).
Nectar volume and concentration
Nectar volume, concentration, and the rates of their increase over 24 h varied according to the presumed pollination system. The hawkmoth-pollinated species (N. alata and N. longiflora) had a lower nectar concentration than the other species, and their nectar concentration did not increase significantly over the first 24 h after anthesis. Hummingbird-pollinated species exhibited a significant increase in nectar concentration over the first 24 h after anthesis (Table 2, Fig. 3B). Evaporation may play a small part in increasing nectar concentrations (Plowright, 1987
), but the increases were more likely due to continued sugar production throughout the day because nectar volumes were also increasing. Larger flowers tend to have more nectar (Fig. 6A), perhaps because of pleiotropic size effects, e.g., larger nectar glands and more space to hold more nectar. The differences in nectar concentration between hawkmoth-, hummingbird-, and the small-moth-pollinated species could be associated with differences in corolla size, which may be an artifact of phylogenetic history and constraints, not necessarily pollinator preferences. However, the lower total energy in autogamous lineages, scaled by flower size, is very suggestive of evolution of nectar traits in association with the evolution of pollination system (Fig. 6, see below).
Much of the literature claims hummingbird-pollinated species have relatively dilute nectar (20–25% sucrose equivalents; Baker, 1975
; Baker and Baker, 1982
; Cruden et al., 1983
). However, the hummingbird-pollinated species in this study had nectar concentration means ranging from 47–60% solids (individual plant concentrations ranged from 20–70% solids), approximately 24 h after anthesis. Even shortly after anthesis (0 h), means ranged from 32–49% solids. The high values may be a result of more accommodating greenhouse conditions, but preliminary field experiments found nectar concentrations in Rastroensis to be as high as 57% solids, which is still much greater than the nectar concentrations typically found in hummingbird-pollinated flowers. The higher nectar concentrations could be the result of pollinator selection, since hummingbirds have been shown to prefer more concentrated nectar (Hainsworth and Wolf, 1976
; Stiles, 1976
; Pyke and Waser, 1981
), although they were found to discriminate against nectar over 55% solids (Tamm and Gass, 1986
). There is also the possibility that bees are playing a larger role in pollination and selection than expected. Bees prefer very concentrated nectar and probably require it to ensure that foraging will be energetically profitable (Bolton and Feinsinger, 1978
). Bees are not expected to have a large role in the pollination of these species due to the low frequency of observed visits. However, more observations and pollinator efficiency experiments are necessary to examine this possibility.
The hawkmoth-pollinated species had nectar concentration means ranging from 19 to 24% solids (individual plant concentrations ranged from 18 to 28% solids), approximately 24 h after anthesis. These estimates fit well within the range of nectar concentrations typical of hawkmoth-pollinated species (15–30%, Baker and Baker, 1982
; Cruden et al., 1983
; Haber and Frankie, 1989
). However, this range is lower than the estimated concentrations (30–40%) that should maximize sucrose intake rates for hawkmoths (Josens and Farina, 2001
). Lower nectar concentrations will be less viscous and therefore more easily extracted, which is assumed to be necessary for Lepidoptera feeding through probosci (Baker and Baker, 1982
). Nevertheless, diurnal hawkmoths showed no difference in preference between solutions ranging from 20 to 50% sucrose (Josens and Farina, 1997
), and changes in viscosity per se do not significantly alter hawkmoth behavior (Josens and Farina, 2001
).
The relatively high nectar concentration for the small moth-pollinated N. bonariensis (Table 2) is surprising because these moths also feed with a proboscis and therefore should favor less concentrated, less viscous nectar. Bees are not likely to have a role in the pollination of N. bonariensis because its flowers are white and nocturnal, although residual nectar in nocturnal flowers of Silene alba has been found to encourage bee visitation early the next day (Young, 2002
). Beetle larvae can often be found inside N. bonariensis flowers, though it is not clear how they affect pollination. Field results suggest that N. bonariensis nectar concentrations may be lower in a natural habitat setting, but population differences are significant (mean at anthesis ± SE; Santa Tereza: 33.55% ± 1.77, N = 53; Morro da Igreja: 12.14% ± 1.14, N = 18).
Total energy
Total energy was equivalent between the hawkmoth- and hummingbird-pollinated species, but was much lower in the autogamous groups and the small moth-pollinated species (Fig. 5). The nectar of hawkmoth- and hummingbird-pollinated species was equivalent in total energy per flower; perhaps there is a trade-off constraining the amount of total energy per flower, but autogamy may release that constraint (Fig. 6).
Autogamy may facilitate the evolution of reduced nectar volumes, and thus total energy, because of the decreased need for pollinator attraction. The two autogamous groups in this analysis produce less total reward than expected based on floral size. The average nectar volume found in N. plumbaginifolia was close to the expected value for its corolla length (Fig. 6A), but its nectar concentration was more similar to the other hawkmoth-visited species, which was well below the expected value for its corolla length (Fig. 6B). However, the Jujuy population of N. longiflora (represented by three progeny from a single maternal plant, all exhibiting autogamy) produced less nectar than expected for its corolla length.
Energy values were estimated for individual flowers, rather than for the whole plant. It is possible that differences in energy per flower could be balanced across species through differences in the number of available flowers, though this was not measured. Total energy represents the total amount of sugar equivalents found in a flower (the product of nectar volume and concentration), while the total sugar concentration indicates the amount of sugar found per unit volume of nectar. Different plants (see Table 1) and different techniques were used to obtain the total sugar concentration (wt/vol) data (HPLC) than those used to obtain the average nectar volume and nectar concentration (wt/wt) results (refractometer). When the weight/weight values were converted to weight/volume, it was apparent that the two measurements differed significantly in each species (t tests at = 0.05). Refractometers are calibrated for percentage sucrose concentrations, but there are other dissolved solids in nectar (including fructose and glucose) that can alter the refractive index from that expected with a pure sucrose solution, producing a concentration value that often does not represent the most accurate sugar concentration (Inouye et al., 1980
). The HPLC method was selective for individual sugars and therefore is more accurate in determining the total sugar concentration than the refractometer. A certain degree of error can be expected for the total energy measurements as well, because they were determined from refractometer readings, but using the refractometer values allowed us to determine total energy for individual flowers measured within each population of each species. Only six samples per species were analyzed with HPLC, restricting total energy to only one data point per species or population. Therefore our statistical power would be greatly compromised if we used the HPLC data to determine total energy.
Sugar and amino acid composition
Sugar concentration
The small moth-pollinated species, N. bonariensis, had a significantly higher sugar concentration than the hummingbird-pollinated species, which had a significantly higher sugar concentration than the hawkmoth-visited species (post-hoc Tukey HSD on HPLC data; see also Fig. 7). The small moth-pollinated N. bonariensis was not significantly different from the hummingbird-pollinated species in nectar concentration as determined by refractometer readings (P = 0.0429, = 0.01).
Sugar components
All of the Nicotiana species studied here had sucrose-dominant nectar [sucrose to (glucose + fructose) > 1.0; (Baker and Baker, 1982
, 1983
)] except N. plumbaginifolia, which had a sucrose-rich nectar [sucrose to (glucose + fructose) > 0.5], but was relatively close to having a more balanced nectar (Fig. 7). Nevertheless, there were significant differences among sugar components that varied according to presumed pollination syndrome. Hummingbird-pollinated species had significantly greater proportions of sucrose than the hawkmoth-visited species, which had significantly greater proportions of glucose than the hummingbird-pollinated species, while the small moth-pollinated and hawkmoth-visited species had significantly greater proportions of fructose than the hummingbird-pollinated species. Hummingbirds prefer sucrose to other sugar types (Hainsworth and Wolf, 1976
; Stiles, 1976
), and at least one hawkmoth species has been shown to prefer sucrose as well (Kelber, 2003
). Sugar type may also be a highly conserved character, as signs of phylogenetic constraint can be found in some families (Baker and Baker, 1982
, 1983
). Therefore it is possible that the similarity in sugar types among these Nicotiana species is due to some degree of phylogenetic constraint. However, the abundance of sucrose in these species may also be due to their morphology, as concealed nectaries tend to be associated with sucrose-dominated nectar (Percival, 1961
; Gottsberger et al., 1984
).
Amino acid concentration
There was significant variation in total amino acid concentration among some species (Fig. 8). Pollinator groups did not have significantly different amino acid concentrations, with the exception that N. bonariensis, the only small-moth-pollinated species, had a significantly greater amino acid concentration than the other pollinator groups. This finding coincides with data compiled by Baker and Baker (1982)
, showing that perching moth-pollinated flowers have significantly more amino acids than hummingbird- or hawkmoth-pollinated flowers. The reason for this pattern is presumably due to the lack of alternative protein sources for moths. Because hummingbirds augment their nectar diet with insects (Wagner, 1946
; Baker and Baker, 1982
; Brice and Grau, 1991
), a lower amino acid concentration might be expected in the species they typically pollinate. Hummingbirds have even been shown to prefer nectar with lower amino acid concentrations (Hainsworth and Wolf, 1976
). Hawkmoths, on the other hand, also lack an alternative protein source, and therefore the lower amino acid concentration found in flowers they typically pollinate is somewhat surprising. It is likely, though, that they may accumulate sufficient amino acids because they collect relatively large quantities of nectar each night (Baker and Baker, 1982
). Also, hawkmoths have a relatively short lifespan (up to 3 weeks in Manduca sexta) and use reserves stored during larval growth and may not need to build up protein as an adult (Ziegler, 1991
). Amino acid concentration could also be affected by fertilizer treatments in the greenhouse (Gardener and Gillman, 2001b
). Therefore, it is possible that the amino acid concentrations found in this study would be different from those of plants in situ. However, barring strong species by fertilizer interaction, we expect the deviations would be consistent across species. It is also possible that the pollinators would dislodge pollen into the nectar while collecting it, which would significantly increase the amino acid concentration (Baker and Baker, 1982
).
Amino acid components
Amino acid complements have been shown to vary little (on a presence or absence basis) within a species (Baker and Baker, 1977
, 1982
; Gardener and Gillman, 2001a
). Nevertheless, Gottsberger et al. (1984)
found that many species vary in amino acid complements across different samples of the same species. Lanza et al. (1995)
found variation between different populations of Impatiens campensis, as well as within a single population. Gardener and Gillman (2001a)
found that amino acid concentrations were much more variable than the relative proportions of amino acids within a species. In multiple cases, at least one sample per species differed in whether a particular amino acid was present or absent (data not shown). The possibility of contamination by pollen or damaged cell contents cannot be ruled out, although much care was taken to avoid any form of contamination.
Amino acid complements are thought to be important in determining the taste of the nectar to the pollinator (Baker and Baker, 1977
, 1982
; Gardener and Gillman, 2002
), although sugars are much more abundant than amino acids and most likely dominate the taste of nectar. Shiraishi and Kuwabra (1970)
found that amino acids could be categorized into four classes based on their effects on chemosensory cells of two fly species: no effect (asn, gln, ala, cys, gly, ser, thr, tyr), general inhibitory (arg, asp, glu, his, lys), salt cell stimulatory (pro and hyp), and sugar cell stimulatory (ile, leu, met, phe, trp, val). The nectar of each species contains at least some amino acid from each of these categories. The effects of amino acids have not been investigated in other insects or hummingbirds and may differ from the effects found on flies. Abundant amino acids were not restricted to any one biochemical category. There were no species that contained all 10 of the essential amino acids (see Table 3) for insects (Haydak, 1970
; Dadd, 1973
) or birds (Brue, 1994
).
Baker and Baker (1982)
found alanine and arginine to be the most common amino acid found in the nectar of species they examined. Arginine was found in all species in this study and contributed a substantial proportion of amino acids to N. forgetiana nectar. Although alanine was found in relatively large proportions in a few species, it was absent from N. forgetiana nectar. Glutamine was predominant in N. longiflora nectar, but was also abundant in N. mutabilis and N. langsdorffii nectar. GABA, a neurotransmitter derived from glutamate, was predominant in N. alata and N. plumbaginifolia, but was also abundant in the nectar of N. forgetiana and Rastroensis. Proline was the predominant amino acid found in the hummingbird- and small-moth-pollinated species, but also contributed a substantial proportion to N. longiflora nectar. Baker (1978)
found proline to be a very common amino acid in nectar, present in 87% of the species in his study, but some authors argue that the abundance of proline is only an indication of pollen contamination (Linskens and Schrauwen, 1969
; Gottsberger et al., 1984
). It is interesting that the species in our study that have a predominance of proline are those with shorter and wider corolla tubes, offering a slightly better chance for pollen to fall into the corolla tube as the flowers move. However, contaminated samples would probably be detectable because specific profiles are produced by pollen, which were not detected in these samples (M. Gardener, unpublished data). It is also possible that the destructive sampling of flowers altered the amino acid content, even if contact with damaged cells was avoided (Gottsberger et al., 1984
). Tryptophan was not present in any of the species in this study, which is surprising because it is considered an essential amino acid for both insects and birds (Haydak, 1970
; Dadd, 1973
; Brue, 1994
). AABA was also not detected in any sample of any of the species, and there were several other amino acids that were found in very small amounts (see Table 3).
Evolution of nectar rewards
This study shows that there are significant differences in nectar traits between different species associated with different presumed pollination systems (see Table 4). Hummingbird-pollinated species had relatively low nectar volumes with high sugar and low amino acid concentrations. Hummingbirds prefer relatively high sugar (Hainsworth and Wolf, 1976
; Stiles, 1976
; Pyke and Waser, 1981
) and low amino acid concentrations (Hainsworth and Wolf, 1976
), which is supported by our results. Hawkmoth-pollinated species had relatively high nectar volumes with low sugar and amino acid concentrations. Although there is a paucity of hawkmoth preference tests, our results support previous findings that hawkmoth-pollinated species tend to have relatively low sugar and amino acid concentrations (Baker and Baker, 1982
). The small-moth-pollinated species had an extremely low nectar volume, but very high sugar and amino acid concentrations. Although the high sugar concentration contradicts assumptions that Lepidoptera should prefer less-concentrated nectar for ease of extraction, the high amino acid concentration supports previous findings (Baker and Baker, 1982
). The autogamous groups had lower total energy than expected, through lower concentration (N. plumbaginifolia) or less nectar (Jujuy population of N. longiflora), than their flower size would predict. Autogamous species often exhibit smaller flower sizes and lower nectar volumes (Spira, 1980
), also suggested by our results.
|
; Pacini, 2003
), but this was accounted for in the experimental design of this study. Floral size probably has a large role in determining nectar volume and concentration. Nectar volume tends to increase, while concentration decreases, with increasing floral size (Fig. 6). Total energy in hawkmoth vs. hummingbird species could be constrained, but if so, that constraint appears to have been released in the autogamous lineages. Nectar traits are sometimes associated with taxonomic families (Baker and Baker, 1982
, 1983
), suggesting phylogenetic constraints had a role in nectar evolution. A more-resolved phylogeny is needed to determine the likelihood that Nicotiana section Alatae exhibits phylogenetic constraints in nectar traits, although the group may be too small to provide enough independent comparisons among pollination systems.
Many of the nectar traits examined in this study vary in association with the presumed predominant pollinator. The variation in these traits makes it possible for pollinators to select for or discriminate against certain types of nectar. Some studies have demonstrated pollinator preference or discrimination for certain nectar (Hainsworth and Wolf, 1976
; Stiles, 1976
; Pyke and Waser, 1981
; Zimmerman, 1983
; Galen and Plowright, 1985
; Tamm and Gass, 1986
; Cresswell and Galen, 1991
; Martínez del Rio et al., 1992
; Hodges, 1995
; Meléndez-Ackerman et al., 1997
; Schemske and Bradshaw, 1999
), making pollinator-mediated selection a real possibility. The nectar trait variation in Nicotiana section Alatae could be a result of past selection pressures from pollinators, but nectar traits could have evolved in association with other floral traits with or without the aid of pollinators. More pollinator preference tests, especially with moths, are necessary to determine whether preferences truly match the nectar traits offered in the plant species they pollinate. Eco-genetic experiments are also needed to determine whether those preferences are strong enough to affect evolution.
FOOTNOTES
1 The authors thank Mark Ellersieck for repeated assistance with statistical analyses, Allena Volskay for use of her corolla length data, Candace Galen for helpful comments on an earlier draft, and Jane Murfett for help in revising one of the final drafts. This study was supported in part by a GAANN fellowship through the U.S. Department of Education, NSF DEB 9727037, and UM Research Board 97-033. 
4 Author for correspondence (e-mail: HoltsfordT{at}missouri.edu
)
LITERATURE CITED
Baker H. G. 1961 The adaptation of flowering plants to nocturnal and crepuscular pollinators. The Quarterly Review of Biology 36: 64-73[CrossRef]
Baker H. G. 1975 Sugar concentrations in nectars from hummingbird flowers. Biotropica 7: 37-41[CrossRef]
Baker H. G. 1977 Non-sugar chemical constituents of nectar. Apidologie 8: 349-356[CrossRef][ISI]
Baker H. G. 1978 Chemical aspects of the pollination biology of woody plants in the tropics. In P. B. Tomlinson and M. H. Zimmerman [eds.], Tropical trees as living systems, 57–82. Cambridge University Press, New York, New York, USA
Baker H. G. I. Baker 1977 Interspecific constancy of floral nectar amino acid complements. Botanical Gazette 138: 183-191[CrossRef]
Baker H. G. I. Baker 1982 Chemical constituents of nectar in relation to pollination mechanisms and phylogeny. In H. M. Nitecki [ed.], Biochemical aspects of evolutionary biology, 131–171. University of Chicago Press, Chicago, Illinois, USA
Baker H. G. I. Baker 1983 Floral nectar sugar constituents in relation to pollinator type. In C. E. Jones and R. J. Little [eds.], Handbook of experimental pollination biology, 117–140. Scientific and Academic Editions, New York, New York, USA
Bolton A. B. P. Feinsinger 1978 Why do hummingbird flowers secrete dilute nectar?. Biotropica 10: 307-308[CrossRef][ISI]
Bolton A. B. P. Feinsinger H. G. Baker I. Baker 1979 On the calculation of sugar concentration in flower nectar. Oecologia 41: 301-304[CrossRef][ISI]
Brice A. T. C. R. Grau 1991 Protein requirements of Costa's hummingbirds Calypte costae. Physiological Zoology 64: 611-626
Brue R. N. 1994 Nutrition. In B. Ritchie, G. Harrison, and L. Harrison [eds.], Avian medicine: principles and application, 63–95. Wingers Publishing, Lake Worth, Florida, USA
Chase M. W. S. Knapp A. V. Cox J. J. Clarkson Y. Butsko J. Joseph V. Savolainen A. S. Parokonny 2003 Molecular systematics, GISH and the origin of hybrid taxa in Nicotiana (Solanaceae). Annals of Botany 92: 107-127
