Crop Science 40:126-130 (2000)
© 2000 Crop Science Society of America
CROP PHYSIOLOGY & METABOLISM
Effect of Flower Color and Sampling Time on Volatile Emanation in Alfalfa Flowers
Luciano Pecettia and
Aldo Tavaa
a Istituto Sperimentale per le Colture Foraggere, viale Piacenza 29, 26900 Lodi, Italy
iscfbred{at}telware.it
 |
ABSTRACT
|
|---|
Flower volatile compounds of alfalfa (Medicago sativa L.) are involved in attractiveness to pollinating insects, and thus influence seed set. This study investigated differences in volatile emanation in this species by varying flower color of the sampled material or time of flower collection. Chemical determinations were made by gas chromatography and gas chromatography–mass spectrometry analyses. The first investigation examined volatile concentration in clones characterized by seven different color ratings. Significant variation among colors was found with a trend for greater emanation from dark- than from light-colored flowers. This trend could bear an evolutionary significance to compensate for the possibly lower visibility of dark flowers in a stand. The second investigation analyzed volatiles from intact inflorescences and from florets and rachises taken separately. A substantial similarity of volatile amount and composition among the three kinds of samples was observed. The final investigation assessed the volatiles produced at different times of the day in each of the first three flowerings of the growing season, during which alfalfa pollination was expected to take place. Confirming previous evidence, temporal patterns of emanation were observed, with a sharp increase from the flowering in May to those in July and August and a strong time-of-day effect during the two summer flowerings (burst of volatiles around midday). The marked seasonal effect appeared somewhat related to temperature increase while the daily variation of emanation in summer paralleled that of solar radiation. There seems to be a coincidence between conditions for maximum volatile emanation and reported optima for insect foraging activity.
Abbreviations: GC, gas chromatography • MS, mass spectrometry
|
INTRODUCTION
|
|---|
IN THE ANGIOSPERMAE that rely on insect pollination, reproductive fitness depends, in part, on the ability of plants to produce very attractive flowers (easily visible or olfactory hints) for insects. Vision and olfaction are indeed the main stimuli that attract insects to flowers, although gustatory and tactile hints may also be important (Dobson, 1994). Odor is learned faster than and used in preference to color and this, in turn, to shape (Dafni and Neal, 1997). From afar insects probably select their target by flower color while at close distances aroma of individual plants becomes the main guide. After alighting the insect may be guided to food sources within the flower by close-range olfactory and gustatory cues, such as nectar volume and sugar concentration (Kauffeld and Sorensen, 1971; Dobson, 1994).
Alfalfa is almost certainly the most important forage species worldwide. A significant trade of seed, which represents therefore an important commodity, parallels its use as a hay or pasture crop (Smith, 1988). Pollination and subsequent seed set require "tripping" of alfalfa flowers by insects to produce adequate amounts of good quality seed (Viands et al., 1988). Emanation of flower volatiles in this species has been investigated, with particular regard to their attractiveness for honey bees, Apis mellifera, (Kauffeld et al., 1969; Loper and Waller, 1970; Loper and Lapioli, 1971; Kauffeld and Sorensen, 1971; Loper et al., 1974; Loper, 1976; Loper and Berdel, 1978; Buttery et al., 1982; Henning et al., 1992).
In a recent study, Tava and Pecetti (1997) analyzed the chemical composition of flower odor in clones pertaining to different taxonomic groups within the M. sativa complex. While ranking of individual constituents was similar in all genotypes, the total amount of volatiles varied widely, being possibly associated with the genotypic flower color. The first aim of the present investigation was to assess on a more representative sample the variation within the species to determine whether volatile emanation of flowers could be affected by their color.
As pointed out by Dafni and Neal (1997), the flower is a morphological unit not necessarily corresponding to a functional unit of the pollination process (a "pollination unit"). If flowers in inflorescences are small and the inflorescence is compact, the insects may perceive the whole as just one unit of attraction. Tava and Pecetti (1997) analyzed volatiles emitted from whole inflorescences, whereas Loper and Lapioli (1971) analyzed the aroma of florets cut from the racemes. In general, fragrances are believed to diffuse from petal tissues (Loper and Lapioli, 1971; Dobson, 1994). The second aim of this study was to examine whether intact inflorescences differ in volatile quantity and quality from florets and green rachises taken separately.
Flower volatile emissions follow temporal patterns, conditioned by temperature, photoperiod, and physiological factors (Dobson, 1994). Daily rhythms of emanation were also observed in alfalfa flowers in controlled growth chambers (Loper and Lapioli, 1971). Seasonal changes in the amount of volatiles in this species under different irrigation treatments were reported by Loper and Berdel (1978). The final aim of the present study was to quantify the flower volatiles produced under field conditions at different times of day during the first three flowerings in the growing season. Late spring and early summer is the period during which alfalfa pollination is expected to take place in northern Italy (Marletto et al., 1985). An attempt was made to associate climatic conditions at the time of sampling with the amount of volatiles.
|
Materials and methods
|
|---|
Plant Materials
Twenty-one genotypes, from a breeding program for the development of grazing-tolerant alfalfa (Piano et al., 1996), were chosen to assess volatile emanation of different colored flowers. These genotypes encompassed different taxa within the M. sativa complex. Seven color levels of those proposed by Barnes (1972) for visually scoring alfalfa flower color were tested, with each color level represented by three genotypes. Color rating 1.1 and 1.3 corresponded to dark and light purple flowers, respectively; 2.1 to dark purple variegated flowers; 2.4 to very light purple variegated flowers; 2.7 to dark green flowers; 2.8 to dark yellow-green flowers; and 4.3 to bright yellow flowers (Table 1)
.
The genotypes were field-grown under rainfed conditions in completely randomized plots of nine propagules, spaced 30 cm apart, obtained by rooting stem cuttings of each mother plant. Flowers were collected from 4-yr-old propagules at the full flowering stage during the first flowering of the season. Date of sampling occurred between 15 May and 19 May and time of the day was between 0830 and 0930 h local solar time; the weather was sunny. For each genotype, about 20 to 30 inflorescences with freshly open florets were collected across different plants by clipping the raceme just below the insertion of the bottom-most floret. In the data analysis of variance (ANOVA), individual genotypes (three) were treated as replicates of each flower color level.
Flowers for the assessment of volatile emanation from intact inflorescences, florets and rachises were collected from the cultivar Algonquin. Sixteen spaced plants, 30 cm apart, were grown from seedlings in each plot, in a randomized complete block (RCB) design with two replicates. All plants had purple flowers. During the first flowering of the third year of growth, 40 inflorescences with freshly open florets were collected across plants at the full flowering stage in each plot. The date of sampling was 19 May, at 0930 h local solar time with sunny weather. Half of the inflorescences were used for chemical analyses while the other half were subdivided into florets and green rachises. Mean values across replicates of volatile amount in intact inflorescences, florets and rachises were compared by ANOVA.
The material for the assessment of volatile emanation at different dates and different times of the day was obtained from the cultivar Equipe grown under the same RCB design as cultivar Algonquin. All plants of Equipe also showed purple flowers. About 20 intact inflorescences with freshly open florets were collected across different plants in each replicate at 0830, 1230, and 1700 h local solar time at the full flowering stage of plants during the first three flowerings of the season which terminated, respectively, the regrowth after winter dormancy, the regrowth after the first clipping, and the regrowth after the second clipping. Flower sampling dates were 28 May, 7 July, and 12 August, respectively. The weather was sunny at all samplings. Air temperature, relative humidity and solar radiation were always recorded at a weather station adjacent to the plots. An ANOVA was carried out on the total volatile amount, including the main factors flowering date, time of the day of flower collections, and block. The variance of the first two factors and of their interaction were tested against the residual experiment error variance. All statistical analyses were performed with the SAS software.
Volatile Extraction and Chemical Analyses
After clipping, the inflorescences of all samples were placed at once into a stoppered flask kept in an insulated box, and brought quickly to the laboratory, less than 5 min walking from the plots. As soon as they got to the laboratory, chemical determinations were made on samples from the first and third experiments. For the samples from the second experiment, half of the inflorescences were also immediately routed to the analyses, whereas in the other half the florets and rachises were immediately separated prior to beginning the analyses. Except for this last case, treatment of flower samples was similar to that reported in previous studies (e.g., Loper and Berdel, 1978).
For all analyses, volatiles were collected by enclosing 2 to 3 g of plant material in a glass tube, to which 9.75 µg of 3-methylcyclohexanone were added as an internal standard. Ultra-pure N2 at flow rate of 20 mL/min through a 3-mm diameter freshly desorbed sampling tube packed with 100 mg of Tenax trap was used to sample the headspace. The sampling period was optimized at 30 min at room temperature (25°C).
Volatile composition was determined by gas chromatography (GC) and gas chromatography–mass spectrometry (GC–MS) analyses. For GC analyses, a gas chromatograph equipped with a thermal desorption cold trap injector connected directly to the GC system and a flame ionization detector was used. For the GC–MS, a gas chromatograph with the same cold trap injection system coupled with a mass spectrometry detector was used. For both systems, the capillary column was a Cp Sil 5 CB of 50 m length, 0.32-mm internal diameter and 5-µm film thickness. The oven temperature program was 40°C for 2 min, then increased at 4°C/min to 220°C and held at this last temperature for 20 min. The carrier gas was H2 (1 mL/min) and He (1 mL/min) for GC and GC–MS, respectively. The sampling tube was inserted into the cold trap injector of the GC or GC–MS and, as the trap heated at 240°C for 10 min in the flow of carrier gas (10 mL/min), the volatiles were desorbed and transferred into a fused silica trap cooled at -120°C by liquid N2. Injection into the capillary column was obtained by flash heating (260°C for 1 min) the cold trap. Detection was by a flame ionization detector heated at 260°C and by MS detector (electron impact, 70 eV, interface 230°C). Compounds were identified by comparison of relative retention time and/or mass spectra of published reference compounds (Jennings and Shibamoto, 1980; McLafferty and Stauffer, 1988; Adams, 1995). Quantitative data of each substance were obtained from Cp Sil 5 CB FID data. 3-Methylcyclohexanone was used as a reference for all the identified compounds. In this study, the results will be presented and discussed in terms of classes of compounds and/or total volatile concentration.
|
Results
|
|---|
Total volatile amounts differed by 2.3-fold among genotypes with different flower color (Table 1); this variation is significant (P < 0.05) according to the ANOVA. Ranking showed that volatile emanation decreased as floral color became lighter. In particular, flowers with dark variegations including green (rating 2.7) and brown shades (rating 2.1) emitted significantly more volatiles than lightly colored flowers such as yellow ones (rating 4.3) or very light purple–very light yellow variegated ones (rating 2.4). Uniformly purple flowers (ratings 1.1 and 1.3), typical of the commonly cultivated subsp. sativa types, appeared intermediate in volatile emanation.
The amount of total volatiles emitted by intact racemes was not significantly different (P > 0.05) from those of only florets or only rachises (Table 2)
. The relatively higher emission of the constituent parts relative to the intact inflorescences indicates that some breakage of cells might have occurred because of the manipulation of the samples, causing the possible release of more volatile compounds. Similar trends were observed for the aldehyde, ester, and terpene classes while the concentration of alcohols and ketones were greater in florets than in rachises.
View this table:
[in this window]
[in a new window]
|
Table 2 Concentration of the five most abundant volatile classes and total volatiles in intact racemes, florets and rachises of alfalfa
|
|
Both flowering date and time of the day of flower collection resulted in significant (P < 0.01) variance in volatile emission (Table 3)
. The flowering date x time of day interaction variance was also significant (P < 0.05).
View this table:
[in this window]
[in a new window]
|
Table 3 Analysis of variance of total flower volatile emanation (µg/g fresh weight) over three flowering dates and three times of the day
|
|
On the average, there was a sharp increase of volatile emanation from the first flowering in May to the second in July, which produced twice the amount of the first flowering. Concentration of volatiles at the third flowering in August did not differ from that of July (Table 4)
. Of the three times of the day at which flower sampling took place, the odor emanation was significantly greater at the 1230-h sampling than either 0830- or 1700-h samplings in July (over 35% higher volatile concentration at 1230 h) and August (over 60% higher concentration at 1230 h) (Table 4).
View this table:
[in this window]
[in a new window]
|
Table 4 Total volatile emanation measured on alfalfa flowers collected at three times of the day during the first three flowerings of the season. Some climatic parameters recorded at the same time of flower collection are also reported
|
|
Examining the five main classes of volatiles (representing about 90–95% of total), the odor burst observed around 1200 h involved aldehydes, alcohols, and ketones both in July and August, whereas esters had a similar trend only in August, and terpenes only in July (data not reported).
Air temperature slightly increased from the first flowering date to the second and third (on average, across sampling times, about 2°C). Mean relative humidity showed the lowest value in July, whereas solar radiation had an opposite trend, recording the highest value in July. The evolution of volatiles during the season was similar to that of temperature, increasing from May to July and August. The simple correlation coefficient between volatile amount and temperature, computed on the date mean values reported in Table 4, was r = 0.99 (P = 0.015).
The trend of volatile emanation during the day observed in July and August, with a peak corresponding to the sampling at 1230 h, was similar to the daily pattern of the radiation. The radiation showed the same pattern also on 28 May when, however, the volatile amount did not show a peak at midday. When considering only the July and August flowerings, the simple correlation coefficients between volatile concentration and solar radiation, computed on values of individual sampling hours, was r = 0.84 (P = 0.036). The correlation was not significant when the May flowering date was included (r = 0.41, P = 0.267).
|
Discussion
|
|---|
The clear trend of greater volatile emanation from dark- than from light-colored flowers fully supports our previous findings obtained with fewer genotypes (Tava and Pecetti, 1997). A wide range of volatile emanation among alfalfa clones was reported by Loper (1976) who also suggested that total flower aroma can be genetically influenced. However, information was not given by that author on the flower color of each individual genotype. Loper et al. (1974) indicated that a certain relationship exists between bee attractiveness by alfalfa clones and volatile amount. A 10-fold range in volatile production was observed among 28 clones by Loper et al. (1974) but, contrary to the present study, there was no trend of differences among flower colors that encompassed, though with a predominance of purple-flowered clones, the whole range of colors present in alfalfa. Kauffeld and Sorensen (1971) observed wide differences in bee attractiveness among alfalfa clones and found that, in general, light colors were preferred to dark ones by bees in tests on flowers presented in sealed containers. If discrimination is also carried out by the bees in the portion of the spectrum discernible by the human eye, in addition to ultraviolet wavelengths, as suggested by Clement (1965) in a study on flower color attractiveness, greater volatile emanation by darker flowers could bear an evolutionary significance to compensate for their lower visibility in a plant stand. Intrinsic differences among flowers with different colors in the density per flower (or per inflorescence) of petal cells responsible of volatile emanation could be matter of investigation.
The similarity in volatile amount and composition of intact inflorescences versus florets and rachises taken separately supports the use of whole inflorescences in studies on volatiles rather than single florets the handling of which might also influence the measurements.
This study confirmed the presence of temporal patterns of flower volatile emanation with a clear seasonal effect between spring and summer flowerings, and a strong time-of-day effect during the two summer flowerings. The seasonal effect we observed did not agree with Loper and Berdel (1978) who reported that volatile concentration increased from mid-March to early June and then decreased until mid-September. Loper and Berdel (1978) also reported that volatile amount fluctuations seemed independent of daily maximum temperatures which increased continuously during the whole sampling period. In our case, the thermal levels appeared somewhat related with the seasonal changes of volatile emanation, inasmuch as an increase of about 2°C from May to July and August was associated with a doubling of aroma concentration. It has been reported (Hatanaka, 1993) that the enzymatic activity leading to the formation of volatile aldehydes is closely related to temperature and incident solar radiation, showing a maximum in the summer months. In the present study, the aldehydes indeed showed an increase of concentration in summer relative to spring (data not reported). Other abundant classes of volatiles (viz. alcohols and ketones) also increased from the first to the July and August flowerings, suggesting that a similar external influence, as indicated by Hatanaka (1993), might also affect other biosynthetic pathways.
The daily pattern of volatile amount that we observed in the second and third flowering, with a peak of production around midday, is similar to that reported by Loper and Waller (1970) under greenhouse conditions, by Loper and Lapioli (1971) in growth chambers, and by Loper and Berdel (1978) in the field. The study by Loper and Lapioli (1971) on the physiology of flower volatile emanation indicated that this is an active metabolic process with large fluctuations within short times (hours) and that it is a photoperiodically induced phenomenon. Our data suggest, at least for the summer period, that a certain relationship may exist between solar radiation and volatile production.
The sharp increase of volatiles in summer and the burst of aroma around midday in this season might be related to greater nectar secretion or higher concentration of sugars in the nectar. Shuel (1952) reported solar radiation as a factor affecting nectar secretion in red clover (Trifolium pratense L.). The higher concentration of volatiles might thus signal an effective reward for insects, as also suggested by Loper and Waller (1970). The insects visiting alfalfa fields in northern Italy during the two summer flowerings in July and August (including honey bees and various species of wild bees) were remarkably more numerous than during the first flowering in early June, and their activity was maximum between 1000 and 1600 h (Marletto et al., 1985). Loper and Berdel (1978) reported a certain coincidence between the peak of volatile emission and the period of maximum insect visits. The same environmental conditions leading to maximum volatile emanation are likely to be optimal for the foraging activity of pollinating insects. The activity of honey bees (Louveaux, 1984), bumble bees, genus Bombus, (Pouvreau, 1984), and leafcutter bees, Megachile rotundata, (Guy et al., 1969) is closely related to temperature and light levels.
This research integrated the information on volatile emanation with other factors such as flower color and environmental conditions. It provided evidence that a relationship may exist between flower color and the amount of volatiles produced. The possible effect of this relationship on insect visitation and seed set in cultivars with different frequencies of flower colors deserves further investigation. The observed variation in volatile concentration indicated that the summer flowerings are likely more attractive to insects than the first flowering in mid-spring. In this sense, the common agronomic practice in the evaluation environment of carrying out the seed production on the second flowering is also biologically sound.
In this kind of study, a multiple approach is conceivable which considers the physiology of volatile formation and emanation from the plant and the physiology of aroma perception and the ecoethology of foraging by the insect. Breeders should also be involved in these investigations if the practical consequences in terms of pollination and, hence, seed yield are to be assessed.
|
ACKNOWLEDGMENTS
|
|---|
We express our gratitude to Dr. Doug Johnson, USDA-ARS, Utah State University, for providing useful material for the classification of flower color, to Maddalena Carli for her assistance in flower collection and classification, and to Lamberto Borrelli for providing the climatic data used in this study.
|
NOTES
|
|---|
Research supported by the Italian Ministry of Agricultural Policy through the special project "Foraggicoltura Prativa".
Received for publication January 6, 1999.
|
REFERENCES
|
|---|
- Adams P.R. Identification of essential oil components by gas chromatography/mass spectroscopy. Carol Stream, IL: Allured Publ. Corp, 1995.
- Barnes D.K. A system for visually classifying alfalfa flower color. USDA-ARS Agric. Handbook No. 424. Washington, DC: U.S. Gov. Print. Office, 1972.
- Buttery R.G., Kamm J.A., Ling L.C. Volatile components of alfalfa flowers and pods. J. Agric. Food Chem. 1982;30:739-743.
- Clement W.M., Jr. Flower color, a factor in attractiveness of alfalfa clones by honey bees. Crop Sci. 1965;5:267-268.[Free Full Text]
- Dafni A., Neal P.R. Size and shape in floral advertisement: measurement, concepts and implications. In: Richards K.W., ed. Proc. Intl. Symp. on Pollination from Theory to Practise, 7th, Lethbridge, AB. 23–28 June 1996. Acta Horticulturae 437. Leuven, Belgium: ISHS, 1997:121-140.
- Dobson H.E.M. Flower volatiles in insect biology. In: Bernays E.A., ed. Insect-plant interactions, Vol. V. Boca Raton, FL: CRC Press, 1994:47-81.
- Guy P., Lecomte J., Écalle C. Utilization d'insectes pour la pollinisation de la luzerne en enceintes fermées. Ann. Amélior. Plantes 1969;19:333-343.
- Hatanaka A. The biogeneration of green odour by green leaves. Phytochem. 1993;34:1201-1218.
- Henning J.A., Peng Y.S., Montague M.A., Teuber L.R. Honey bee (Hymenoptera: Apidae) behavioral response to primary alfalfa (Rosales: Fabaceae) floral volatiles. J. Econ. Entomol. 1992;85:233-239.[ISI]
- Jennings W., Shibamoto H. Qualitative analysis of flavor and fragrance volatiles by glass capillary gas chromatography. Qualitative flavor. New York: Academic Press, 1980.
- Kauffeld, N.M., and E.L. Sorensen. 1971. Interrelations of honey bee preference of alfalfa clones and flower color, aroma, nectar volume, and sugar concentration. Kansas Agric. Exp. Stn. Res. Bull. No. 163.
- Kauffeld N.M., Sorensen E.L., Painter R.H. Stability of attractiveness of alfalfa clones by honey bees under varying locations, seasons, and years. Crop Sci. 1969;9:225-228.
- Loper G.M. Differences in alfalfa flower volatiles among parent and F1 plants. Crop Sci. 1976;16:107-110.
- Loper G.M., Berdel R.L. Seasonal emanation of ocimene from alfalfa flowers with three irrigation treatments. Crop Sci. 1978;18:447-452.[Abstract/Free Full Text]
- Loper G.M., Lapioli A.M. Photoperiodic effects on the emanation of volatiles from alfalfa (Medicago sativa L.) florets. Plant Physiol. 1971;49:729-732.
- Loper G.M., Waller G.D. Alfalfa flower aroma and flower selection by honey bees. Crop Sci. 1970;10:66-68.[Abstract/Free Full Text]
- Loper G.M., Waller G.D., Berdel R.L. Olfactory screening of alfalfa clones for uniform honey bee selection. Crop Sci. 1974;14:120-122.[Abstract/Free Full Text]
- Louveaux J. L'abeille domestique dans ses relations avec les plantes cultivées. In: Pesson P., Louveaux J., eds. Pollinisation et productions végétales. Paris: INRA, 1984:527-555.
- Marletto F., Manino A., Barbero R. Indagini sui pronubi in coltivazioni di erba medica per la produzione di seme. Apicolt. Mod. 1985;76:95-102.
- McLafferty F., Stauffer D. Wiley register of mass spectral data, 4th edition New York: Wiley, 1988.
- Piano E., Valentini P., Pecetti L., Romani M. Evaluation of a lucerne germplasm collection in relation to traits conferring grazing tolerance. Euphytica 1996;89:279-288.
- Pouvreau A. Biologie et écologie des bourdons. In: Pesson P., Louveaux J., eds. Pollinisation et productions végétales. Paris: INRA, 1984:595-630.
- Shuel R.W. Some factors affecting nectar secretion in red clover. Plant Physiol. 1952;27:95-110.[Free Full Text]
- Smith D.L. The seed industry. In: Hanson A.A., et al. , ed. Alfalfa and alfalfa improvement. Madison, WI: ASA, CSSA, SSSA, 1988:1023-1036.
- Tava A., Pecetti L. Volatiles from Medicago sativa complex flowers. Phytochem. 1997;45:1145-1148.
- Viands D.R., Sun P., Barnes D.K. Pollination control: mechanical and sterility. In: Hanson A.A., et al. , ed. Alfalfa and alfalfa improvement. Madison, WI: ASA, CSSA, and SSSA, 1988:931-960.
TOP
ABSTRACT
INTRODUCTION
