Carbon allocation to volatiles and other reproductive components in male Ficus carica (Moraceae) 1
- Centre d'Ecologie Fonctionnelle et Evolutive, CNRS, 1919 route de Mende, 34293 Montpellier Cedex 5, France
- Received for publication 9 February 2001.
- Accepted for publication 31 May 2001.
Abstract
Volatile compounds are often mediators of plant–pollinator interactions. Their emission is presumed to be costly, but this cost has seldom been quantified. Figs of Ficus carica (a dioecious species) release volatile compounds when receptive, thus attracting the agaonid wasp Blastophaga psenes. In dioecious fig species, wasps lay eggs inside male figs and pollinate female ones. For a male tree, we estimated carbon allocation to reproduction using the annual growth module (AGM) as the unit of measurement. Over the growing season, leaf and fig carbon exchange and construction costs were measured, as well as carbon investment in stamens, provisioning pollinators, and biosynthesis and release of volatile compounds. Representativity of the tree studied was evaluated by measuring some of these parameters on seven other male fig trees. The results show that 7.6–16.4% of the carbon assimilated by leaves and figs was invested in reproduction. Of the carbon invested in reproduction, pollinator attraction and feeding represented only 0.08–0.12% and 1.84–2.33%, respectively, while pollinator sheltering (fig construction and respiration) represented 97.6–98.0%. In this strict and coevolved plant–pollinator association, the main male reproductive investment was thus in the structures sheltering the associated pollinators.
- Blastophaga psenes
- carbon allocation
- cost of reproduction
- Ficus carica
- Moraceae
- mutualism
- photosynthesis
- reproductive effort
- respiration
- volatile compounds
Resources allocated to reproduction are diverted from somatic growth and maintenance (Williams, 1966⇓). The direct cost of reproductive male function comprises resource expenditure in stamens and often in devices favoring pollen dispersal. Wind-pollinated plants invest more in pollen production than insect-pollinated plants (Raven, Evert, and Curtis, 1976⇓; Charlesworth and Charlesworth, 1981⇓; Whitehead, 1983⇓; but see also Cruden and Lyon, 1985⇓). However, insect-pollinated species also have to invest in pollinator attraction and reward (Simpson and Neff, 1983⇓; Waser, 1983⇓; Dobson, 1994⇓).
In these species, pollinator attraction is based on visual or chemical signals (Kevan and Baker, 1983⇓). While such volatile compounds have been identified in floral fragrances of >400 taxa (Knudsen, Tollsten, and Bergström, 1993⇓), their cost, which depends on the type of compound and the rate and duration of emission, has seldom been evaluated. Production of volatile compounds is indeed generally assumed to be costly (Pichersky et al., 1994⇓; Euler and Baldwin, 1996⇓), but does the cost of fragrance compare to the cost of floral structures attracting pollinators? The latter can constitute a large proportion of reproductive biomass (e.g., 63% in the male individuals of the shrub Oemleria cerasiformis, Rosaceae; Antos and Allen, 1994⇓). Similarly, the cost of pollinator rewards has seldom been measured, except for nectar, which has been shown to represent up to twice the energy invested in seeds (Southwick, 1984⇓; Pyke, 1991). But for other types of rewards, for example, sheltering the pollinators in nursery-pollination systems, the cost has not been evaluated.
Ficus carica, the cultivated fig tree, is a dioecious species. Female trees bear figs that produce seeds when pollinated, and male trees bear figs in which pollinators lay eggs. Both male and female figs are urn-shaped inflorescences (syconia) and flowering is synchronized by sex (Kjellberg et al., 1987⇓). As in most fig species, there is a species-specific mutualism between the plant species and its pollinator (Janzen, 1979⇓; Wiebes, 1979⇓). The obligate pollinator of F. carica (the wasp Blastophaga psenes, Agaonidae) overwinters as larvae in the male figs that developed in autumn. In spring, it leaves these figs and enters the spring crop of male figs, where it lays eggs. In summer, the new generation of adult wasps emerges from the figs and exits the male figs loaded with pollen grains. At this point, female figs are receptive. The wasps enter them, pollinate the flowers, and die without laying eggs. The wasp cycle is maintained through a few individuals leaving some of the male figs late enough in summer for the wasps to find new male figs in the autumn crop and lay eggs inside them (Kjellberg et al., 1987⇓).
Our objective was to quantify the cost of production of the various components of male reproduction in Ficus carica (Moraceae), including fragrance emission and pollinator sheltering, and to compare it to resources provided by photosynthesis. Male F. carica trees offer indeed the opportunity to study reproductive components that are usually not quantified in terms of cost. During the receptive phase, figs of F. carica emit volatile compounds to attract their pollinating wasps (Hossaert-McKey, Gibernau, and Frey, 1994⇓; Gibernau et al., 1998⇓). Receptivity of figs is particularly long (2–4 wk; Khadari et al., 1995⇓), and it has been shown that such a long receptivity, rare among angiosperms, is necessary to the maintenance of the mutualistic cycle (Khadari et al., 1995⇓). Moreover, the quantity of volatiles emitted must be high enough for the short-lived pollinators (2–3 d; Kjellberg, Doumesche, and Bronstein, 1988⇓) to detect receptive trees. However, this is less true for F. carica trees, which are usually in populations denser than tropical figs. A strong volatile emission rate may also be of lesser importance in the spring crop of male trees, which probably receive many wasps emerged from autumn figs of the same tree. Male fig trees have additional costs of reproduction since they bear large pseudofruits. Indeed, in contrast to most dioecious species (Antos and Allen, 1994⇓; Obeso, 1997⇓; Nicotra, 1999⇓), male and female figs of F. carica are relatively similar in many characters, including mass. The large male investment in a pseudofruit may be necessary for the function of sheltering and rearing the pollinators (Patiño, Herre, and Tyree, 1994⇓). However, because figs are green for a large period of their development, one can expect that they will contribute through photosynthesis to their construction cost, as is the case for reproductive organs of a large array of species (Bazzaz, Carlson, and Harper, 1979⇓; Blanke and Lenz, 1989⇓).
MATERIALS AND METHODS
Experimental approach
Estimations of resource acquisition and allocation were made at the level of the annual growth module (AGM). An AGM was defined as all the organs developing a given year on the twig formed the previous year. The organs are the figs on the distal part of that twig (they develop in spring) and the current year twigs (one or more), which develop from the terminal buds and bear the current year's leaves and the autumn figs. In May 1999, leaf area and fig biomass of a mean AGM were estimated by measurements on seven south-facing AGMs of an isolated, large, >30-yr-old tree located on the Centre National de la Recherche Scientifique (CNRS) campus in Montpellier, France. Leaf and fig seasonal carbon exchange parameters are less variable between AGMs than is the leaf area/fig mass ratio (see RESULTS). They were analyzed on a subset of the seven AGMs. The seasonal carbon budget was estimated for a mean AGM calculated from the mean leaf area and fig mass of the seven AGMs to which the physiological parameters were applied.
Measurements of leaf area and carbon exchange
Leaf area development was measured nondestructively on two AGMs by drawing the leaves and measuring the drawn area with a leaf area meter (model MK2, Delta-T Devices, Cambridge, UK). Measurements were taken on 15 April, 7 May, 9 June, 21 July, 17 September, and 4 November. Unfolding of the first leaves occurred on 1 April, and most leaves had turned brown or had fallen on 9 November. Leaf area was assumed to be zero at these two dates and to vary linearly between measurements.
Leaf photosynthesis was measured on the six dates of area measurements as well as on 31 May. Photosynthesis and local light level were measured with a portable gas exchange system (Li-Cor 6400, Li-Cor, Lincoln, Nebraska, USA) on all the leaves (or ten randomly sampled leaves when an AGM possessed more than ten leaves) for each of the two AGMs. Leaves were kept at their natural inclination and azimuth during measurements. Each set of measurements was repeated 2–6 times a day between 0730 and 1530 solar time. Parameters inside the chamber matched outside conditions. They varied with time of the day and season. For 15 April and 13 July, respectively, they were: temperature 13° and 34.8°C, relative humidity 79% and 25%, chamber CO2 concentration 342 and 353 μmol/mol, internal CO2 concentration 261 and 214 μmol/mol. Average photosynthetic rate and light level per AGM were calculated for each set of measurements, allowing establishment of a relationship between AGM light intensity and photosynthesis.
Night respiration rate of leaves was measured on 20 April, 7 June, 21 September, 22 October, and 4 November on the two AGMs used for measures of photosynthesis. Measurements were taken early in the morning on ten randomly chosen leaves per AGM (AGMs had been kept in the dark, with black bags, since the previous evening). Variation of respiration rate between two dates of measurement was assumed to be linear.
For each measurement date, the data from the two AGMs were combined and fitted with Rabinowitch's asymptotic function (F: Pn = (a × b × I)/(a × I + b) − Rd, where Pn is the mean net photosynthesis rate of the AGM, I the mean light intensity reaching the AGM, a the photosynthetic light efficiency, b the light saturated rate of CO2 assimilation [a and b obtained from the model], and Rd the dark respiration) (Thornley, 1976⇓). This relationship was then used to estimate daily and seasonal AGM photosynthesis. For each measurement date, the daily course of the ratio R between each AGM light level and the light level recorded at a meteorological station 300 m away was fitted with the function providing the highest coefficient of determination, generally a linear function. For each hour of the day, R was assumed to vary linearly between the measurement dates. Using R, AGM leaf area, and the function F, photosynthesis of each AGM was calculated for every hour with detectable light intensity and for every day over the whole season.
To estimate the amount of carbon invested in the leaves and eventually lost as litter, ten leaves were harvested on 5 and 17 September, 23 October, and 12 November. Their C content was determined with a C-H-N analyzer (model EA 1108, Carlo Erba Instruments, Milan, Italy). To estimate the carbon invested in perennial vegetative biomass, we measured the dry mass of the twig formed in 2000 on the seven AGMs studied in May 1999.
Measurements of reproductive components: fig production and activity
Emission of volatile compounds
For the spring crop, emission of volatile compounds and change in fig size of all figs of four AGMs were measured five times (once a week) from 7 April (fig diameter = 25–30 mm) to 5 May (1 wk after pollinator visits). Volatile compounds were collected in situ by the adsorption-desorption headspace technique (Turlings et al., 1991⇓; Grison, Edwards, and Hossaert-McKey, 1999⇓). Each AGM (including 4–8 figs) was enclosed in a polyethylene terephtalate (Nalophan, Kalle Nalo GmbH, Wursthüllen, Wiesbaden, Germany) bag. Air was pushed into the bag by a pump (400 mL/min) through a charcoal filter. It was then drawn out of the bag (300 mL/min) over a Porapak Q filter (25 mg, 80–100 mesh) to trap the volatile compounds. On each day of measurement, collection of volatile compounds lasted from 0900 to 1700. On 5 May, measurements were also done from 1700 to 0900 the following day. Filters were eluted with 150 μL of dichloromethane. Four micrograms of nonane and 4 μg of dodecane were added as internal standards. Solutions were then analyzed by gas chromatography, using a CP-9003 (Chrompack, Middelburg, The Netherlands) chromatograph (column EC-1, length: 30 m, internal diameter: 0.25 mm, film thickness: 0.25 μm, carrier gas: helium, on-column injector, oven temperature program: 50°–250°C, 5°C/min), and compounds were identified using a Hewlett-Packard (Palo Alto, California, USA) gas chromatograph (HP 5890)-mass spectrometer (MS: HP 5870, column: 20 m, ID: 0.20 mm, film thickness: 0.1 μm, carrier gas: helium, oven temperature program: 50°–200°C, 3°C/min).
To measure the proportion of the volatile compounds in the bag that were trapped and detected, we performed the following experiment. Variable quantities of linalool (the main compound in fig fragrance; Gibernau et al., 1997⇓) or limonene (another, minor, compound of the fig blend; see RESULTS) were taken in micropipettes, which were weighed and inserted in a bag similar to the one used to trap the fig fragrance. Air was pulled out of the bag through a Porapak Q filter by a sucking pump at a rate of 300 mL/min. Upstream of the micropipettes, the bag had an air entrance just sufficient to prevent deflation of the bag. After 3–4 h of collection, the micropipettes were weighed and the molecules trapped in the filter eluted with dichloromethane. Quantity of linalool in the solution could thus be compared to that evaporated in the bag. The same experiment was also done three times with limonene. The results did not much differ from those obtained with linalool. The data obtained were fitted to the closest function (least square method) and the final relationship was Qr = exp(0.528ln(X) + 2.4494), with Qr the total quantity released (in micrograms) and X the quantity estimated from the peak area on the chromatogram (in micrograms). This relationship was then applied to all compounds of the fig blend. For example, if the peak of the compound on the chromatogram indicated a mass of 10 μg, this meant that 52 μg had actually been released into the bag. Compound recovery was thus 19% for that quantity. This result can be compared to the study of Raguso and Pellmyr (1998)⇓, where for different compounds, with entrance and exit airflows of 500 mL/min, percentages of recovery averaged 10%. Because of the difference between entrance and exit airflow rates (entrance: 400 mL/min, exit: 300 mL/min), 75% of the compounds in the bag could be trapped on the Porapak Q filter and we added another correcting factor of 4/3.
Carbon mass emitted through the volatile compounds was calculated using their chemical formulae. Knowing the diameter of the figs at each measurement date and using a regression between diameter and volume (established on a separate set of figs), carbon emitted was averaged over the four AGMs and expressed per volume of fig. Since emission rate is not simply related to climatic variables, interpolation between dates of measurements was done using a linear function. Quantity of CO2 respired for the construction of each volatile compound was estimated from its carbon content using the method of Vertregt and Penning de Vries (1987)⇓.
For the autumn crop, diameters of all figs of seven AGMs were measured once every 2 wk from 11 August to 3 November. To establish a relationship between fig diameter and dry mass, diameter, volume, and dry mass of 20 figs of different AGMs of this tree were measured on 23 August. This relationship was then used to estimate fig mass evolution over the season. Volatile compounds were collected in situ on four of the seven AGMs on 24 August. Collection lasted 24 h but yielded a low amount of volatiles. On 1 September, we then cut branches from other male trees at the same site, put the section into water, and enclosed the branches in bags. In each of the two bags, only figs of a given diameter class (10–15 mm, 40 figs; or >15 mm, 24 figs) were kept on the branches. Collection lasted 4 h. The phenological stage of each fig was noted. For each diameter class, carbon emitted was expressed per fig. For each of the seven studied AGMs, carbon emission over the season was calculated based on the growth curve of figs and their number on the AGM. Mean AGM values of fig dry mass and volatile emissions are the average of the values for these seven AGMs.
Fig carbon exchange
Photosynthetic light response curve of figs (syconia) was measured with a portable gas exchange system (Li-Cor 6200, Li-Cor, Lincoln, Nebraska, USA) on four figs on 20 April, 19 May, and 19 June (spring crop), and three figs on 20 October (autumn crop). Measurements were taken in the morning; light intensity was manipulated using shadecloths of different densities that were progressively added. Photosynthesis was expressed per unit of external fig area. Fig area was derived from a diameter–area regression, established on 20 figs of the tree in spring and autumn. Mean light response curves for each day were fitted to Rabinowitch's model (Thornley, 1976⇓; see above). Fig net carbon exchange for the two AGMs was calculated using the three parameters evaluated on each day of measurement and the mean fig area. The mean branch light level calculated with the leaf measurements was used. We assumed the three parameters to vary linearly between dates of measurement. The calculation of fig net photosynthesis and dark respiration was applied from the beginning of fig development (20 March for the spring crop or 7 August for the autumn crop) up to fig maturation (30 June, spring crop) or fall of leaves (8 November, autumn crop).
Carbon invested in figs
To estimate the amount of carbon invested in figs and dropped from the tree as litter at the time of abortion or maturation, 7–24 figs of the following categories were harvested: spring crop (aborted and mature) and autumn crop (aborted and mature). Their dry mass was measured and C content determined with a C-H-N analyzer.
Efficiency of pollinator visits
In 1999 and 2000, 40 figs of FC6 were randomly sampled. Their state (visited by pollinators or aborting) was noted.
Investment in pollen and pollinators
On 23 June, five figs were harvested before stamen dehiscence had occurred. Stamens were removed from the fig, their dry mass was measured, and C concentration determined with a C-H-N analyzer. Carbon dioxide respired for stamen construction was estimated using the method of Vertregt and Penning de Vries (1987)⇓. Autumn crop figs do not produce any pollen. Pollinators were harvested from emergent figs and weighed, and their carbon content was estimated with a C-H-N analyzer. The mean number of pollinators produced per fig was taken from Abdellouahad (1991)⇓. To estimate CO2 released by wasp respiration, because the developing wasps feed inside the tiny gall and are inactive, we used the data of Tartes, Kuusik, and Vanatoa (1999)⇓ for diapausing Pieris brassicae pupae (60 μL of CO2 per gram of fresh mass and per hour). We assumed their growth curve to be linear.
Intertree variability
In June 2000, leaf area and fig number and mass were measured on seven AGMs on the south-facing part of the tree studied in 1999 (Ficus carica 6 called FC6 for) and on all leaves of three randomly chosen AGMs of seven other male trees growing a few hundred meters away (called M2 to M8). For all these trees, in spring, carbon exchange was measured at 1500 μmol photons·m−2·s−1 (light source: 6400–02 light emitting diode; Li-Cor, Lincoln, Nebraska, USA) on the second youngest fully expanded leaf of each of the three AGMs chosen and under different light conditions on four figs (FC6) or one fig per tree (M2–M8). Respiration measurements were made on 7 June for FC6 and on 8 August for trees M2–M8.
RESULTS
Leaf area development and carbon exchange
Leaf area increased during April and May, peaked at 19 dm2 (12.9 g dry matter) in June and sharply decreased at the end of September (Fig. 1). Mean AGM daily photosynthesis peaked in early June and declined steadily thereafter. Respiration rate was maximal early in the season (Fig. 1). The nonlinearity of the curve comes from the variation of the leaf area and the variability of the number of night hours. A typical leaf light response curve used to model seasonal photosynthesis is shown in Fig. 2. Daily leaf photosynthesis and respiration were integrated over the whole season (Table 1). Leaf carbon content (42%) did not change between mature and senescent leaves at either date (F1,25 = 3.40, P = 0.08). Total carbon investment in leaves was 5.4 g C/AGM.
Table 1. Annual growth module (AGM) carbon budget over the whole growing season: quantity of carbon assimilated by leaves and figs, respired by the different components of the AGM, carbon stored in leaves, figs, stamens, pollinators, and allocated to biosynthesis and emission of volatile compounds (in carbon grams per AGM)
Fig. 1. Seasonal variation in leaf area, day net photosynthesis, and night respiration for a mean annual growth module (AGM)
Fig. 2. Typical light dependence of net carbon exchange for leaves and figs. Data for leaves are the average of 10 leaves on each of two annual growth modules (AGM) on 31 May. Data for figs are measurements on one fig on 19 May
Fig production and carbon exchange
Number of figs produced per AGM was 1.9 (SD = 0.7) in spring and 5.4 (SD = 2.0) in autumn. Changes in fig dry mass over the season showed periods of dry mass increase, which were due both to an increase in the number of figs and to the growth of the figs on the AGM during maturation (Fig. 3). Periods when dry mass diminished were due to fig fall, either because of abortion or of shedding of mature figs after pollinator emergence. About 60% of the figs were visited by wasps in each crop, and the others were aborted. Variation in mass production refers strictly to the spring crop up to 1 June and mainly to the autumn crop after 1 August (Fig. 3). The carbon exchange light response for a typical fig in spring is shown on Fig. 2. At moderate and high light intensities, gross photosynthesis can compensate for respiration. Seasonal fig gross photosynthesis and respiration were estimated for the two crops (Table 1). In the spring crop, carbon content in aborted and mature figs was 43.4 and 44.8%, respectively (F1,32 = 14.39, P < 0.05). Total fig dry mass per AGM was 3.3 g (1.1 g from aborted figs and 2.2 g from mature figs); thus, total carbon investment in figs in spring was 1.5 g for the AGM. In autumn, fig mass was 5.5 g per AGM (3.3 g from visited figs and 2.2 g from aborted figs). Percentage of carbon in dry mass was 44.1 and 44.6% in mature and aborted figs, respectively (F1,9 = 4.50, P = 0.06). Thus, investment per AGM in fig construction was 2.4 g C in autumn. Mature stamen dry mass and carbon content were 0.204 g per fig and 47%, respectively, for figs of the spring crop (the only ones to have pollen). Carbon investment in stamens was thus 0.109 g per AGM (Table 1).
Fig. 3. Seasonal variation in fig dry mass production (daily net rate of growth or loss) and in daily rate of emission of volatile compounds. Values per mean annual growth module (AGM)
Seasonal variation in fig volatile emissions
In the spring crop, volatile emission by all figs of the AGM was quite synchronized, whereas in autumn, release of volatiles lasted longer, with two main periods of release, due to two cohorts of figs (Fig. 3). In this last crop, some early figs can mature before the weather is too cold, and in this case, pollinators emerge and the fig drops from the AGM at the same time as other figs are receptive or maturing. In spring, seven compounds were found in fig emissions: limonene (C10H16, molecular weight, M, = 136.2 g/mol, represented 2% of the total mass of compounds), benzyl alcohol (C7H8O, M = 108 g/mol, 1%), linalool oxide (C10H17O, M = 153 g/mol, 9%), linalool (C10H18O, M = 154.2 g/mol, 74%), two kinds of branched octanes: “octane 1” (C8H18, M = 114 g/mol, 1%) and “octane 2” (10%), and a branched nonene (C9H19, M = 123 g/mol, 3%). Quantities of carbon invested in compound biosynthesis are given in Table 1. Even if the total number of maturing figs per AGM is higher in the second crop than in the first one (5.4 compared to 2.9), higher quantities are emitted in spring than in autumn (1.1 compared to 0.6 mg C). This difference does not appear to affect the efficiency of pollinator visits, which is similar in the two crops (60%).
Pollinator construction and respiration costs
In the spring crop, the mean number of visited figs per AGM was 1.14 (60% of 1.9), and we considered a mean number of pollinators per fig of 403 (Abdellouahad, 1991⇓). In the autumn crop, 3.2 figs were visited per AGM. Mean number of pollinators per fig was 242 (Abdellouahad, 1991⇓). Mean pollinator dry mass was 0.11 mg, and carbon content was 54.7%. Total carbon investment in pollinator feeding (respiration and insect carbon mass) was calculated for each crop (Table 1).
Intertree variability
For all trees, inter-AGM variability is lower for all the gas exchange parameters than for the structural parameters (Table 2). It legitimates the sampling of a higher number of AGMs for these parameters in the main study of FC6. Leaf area of a mean AGM as well as fig dry mass per unit leaf area were significantly different for FC6 than for the other trees (F1,29 = 21.09, P < 0.05 and F1,29 = 4.24, P = 0.05, respectively), in contrast to all other parameters (Table 2). As a consequence, for the estimation of the seasonal budget, we did not use the data from FC6 alone, but instead incorporated the large intertree variation in AGMs structure. A range of the components of the seasonal budget was calculated with one extreme given by the data for FC6 and the other one given by the leaf area/fig mass ratio of M2 and the seasonal variation of the physiological characteristics of FC6 (not available for the M2–M8 trees).
Table 2. Intertree variability of mean annual growth module (AGM) leaf area, fig mass, and fig and leaf activity in the year 2000 growing season (1 SD). FC6 is the male tree analyzed in detail and M2–M8 are seven other male trees growing on the same campus
Seasonal carbon budget
Fluxes of carbon to the different components are represented in Fig. 4, in percentage of total carbon assimilated by fig and leaf gross photosynthesis. The most important carbon sink is leaf respiration (51.8–53.1%). Total investments in reproduction (fig construction and respiration, synthesis of volatile compounds, pollinator respiration) represent 7.6–16.4% of the assimilated carbon (6.7–13.0% and 0.9–3.4% through leaf and fig photosynthesis, respectively). A similar percentage is invested in leaf carbon mass lost as litter (7.6–7.8%). A quarter to a third of the assimilated carbon (24.4–31.5%) is therefore probably used for growth and respiration of perennial structures (stems, trunk, and roots) and/or stored for later growth or reproduction. Among the allocations to the different reproductive components, main costs are fig construction (5.5–10.8%, lost as litter) and fig respiration (1.9–5.2%). Included in fig construction is the allocation to stamens (0.16–0.72%). Carbon invested in pollinators represents 0.07–0.10% for their respiration and 0.11–0.21% for their body carbon mass. Investment in volatile compounds represents only a small part of the carbon assimilated (0.01–0.02%). Expressed in percentage of the carbon invested in reproduction, fig construction and respiration represent 61.9–70.3% and 25.1–31.7%, respectively, carbon allocated to pollinators 1.9–2.4%, to stamen carbon mass 2.0–4.4%, and to volatile compounds synthesis 0.08–0.12%.
Fig. 4. Fluxes of carbon to vegetative and reproductive compartments (the annual growth module (AGM) is schematized on the figure), integrated over the growing season (100% is the total carbon assimilated by leaves and figs). Bold values refer to the FC6 tree, other values to a tree combining the structural parameters of the M2 tree and the physiological parameters of the FC6 tree
DISCUSSION
The carbon invested in reproduction in male F. carica trees was found to vary between 7.6% (FC6 tree) and 16.4% (M2 tree) of the total assimilated carbon. For Fragaria vesca (wild strawberry), Jurik (1983)⇓ found values of total reproductive effort (RE) between 7.4% and 33.2%. However, RE is more commonly estimated by the ratio of reproductive to total biomass produced during the year (Abrahamson, 1975⇓; Piñero, Sarukhán, and Alberdi, 1982⇓; Obeso, 1997⇓; Verdú and García-Fayos, 1998⇓; Antos and Allen, 1999⇓). To calculate this ratio for the tree FC6 we measured the dry mass of the twigs produced in 2000 (15.9 g per AGM) to complete the leaf and fig dry mass data presented in the RESULTS section. We obtained an RE of 23.5%, an overestimate because the biomass exported to trunk and roots was neglected. For F. vesca, this biomass ratio was between 10.5% and 46.9% (Jurik, 1983⇓). For the male individuals of Oemleria cerasiformis, a dioecious shrub, the average RE estimated as reproductive per leaf biomass was 5.1% (Antos and Allen, 1999⇓). Allocation of biomass to reproduction is highly variable between species (1–60%) or even genotypes, and depends largely on environmental conditions (Bazzaz et al., 1987⇓). It generally increases with the age or the size of the plant (Piñero, Sarukhán, and Alberdi, 1982⇓; Samson and Werk, 1986⇓), is higher in perennial compared to annual species (Bazzaz et al., 1987⇓), and is higher in female than in male individuals of dioecious species (Wallace and Rundel, 1979⇓; Antos and Allen, 1999⇓).
Photosynthesis of fig syconia contributes 12–21% of the total carbon invested in reproduction (17–34% of the carbon necessary for their construction or 49–66% of the carbon necessary for their respiration). We do not know any data about photosynthesis specifically by male structures, but Bazzaz, Carlson, and Harper (1979)⇓ studied 15 species of temperate trees and estimated that 2.3–64.5% of the carbon required for seed construction was contributed by photosynthesis of the reproductive structures.
A fig released 700 μg (in spring) or 100 μg (in autumn) of compounds in 2–4 wk, with a maximum emission rate of 4.6 μg/h (in spring) or 1.2 μg/h (in autumn). This rate is quite low compared to that by the large inflorescences of phytelephantoid palm trees, which emit up to 7000 μg/h of volatile compounds during a few days (Ervik, Tollsten, and Knudsen, 1999⇓), but greater than the rate of emission of compounds released by some Ophrys flowers, which can emit 0.001–0.008 μg/h during 12 h (Borg-Karlson, 1990⇓). Investment in volatile components by flowers can be compared to investments in other compounds attracting or rewarding insects. Young Ochroma pyramidale produce nectar, which was estimated to cost 1% of the energy invested in leaf construction (O'Dowd, 1979⇓). In the plant–herbivore–carnivore interaction studied by Dicke and Sabelis (1988)⇓, the plant released terpenes when attacked by the herbivore, and carnivores were attracted by these terpenes. Total biosynthetic cost of this emission was 0.001% of leaf construction per day. For our male F. carica trees, the carbon invested in volatiles represented 0.08–0.25% of the leaf carbon. Fig investment in volatile compounds is thus in the range found for other kinds of chemical attractants.
This experiment suggests that for male fig trees, attracting and feeding pollinators represent a very small part of the total cost of reproduction, and so does the sensu stricto male function of the fig, which is the production of pollen. Considered together, these investments, which are part of the sensu lato male function, involved 4.5–6.4% of the carbon allocated to reproduction in these male individuals. The remaining 93.6–95.5% were allocated to pollinator sheltering, which involves fig syconia construction and respiration. Few studies have considered the biomass or energy allocated to the different structures and activities implied in reproduction. Antos and Allen (1994)⇓ studied the dioecious shrub Oemleria cerasiformis. In this species, structures contributing to pollinator attraction (petals and hypanthium) represent 63% and 50% of the biomass allocated to reproductive structures in male and female individuals, respectively. The case of the male fig differs from that of Oemleria cerasiformis and most flowers, since pollinators grow inside them. The walls of figs are quite thick, probably helping to protect pollinators from changes in external temperature or humidity and from parasites (Neeman and Galil, 1978⇓; Patiño, Herre, and Tyree, 1994⇓; Kerdelhué, Rossi, and Rasplus, 2000⇓).
In this study, we measured direct costs of reproduction. Resources allocated to reproduction generally diminish further reproduction and growth (Reznick, 1985⇓; Poot, 1997⇓, but see also Dudash and Fenster, 1997⇓ and Nicotra, 1999⇓), thus implying an indirect cost, which was not considered in this study. Dicke and Sabelis (1988)⇓ and Euler and Baldwin (1996)⇓ suggest the existence of another kind of indirect cost: the emission of volatile compounds can imply a risk of attraction of parasitic insects. Parasitic insect species have been seen on figs, some of them at the time of volatile emission, but the attractiveness of the chemical signal to them, if any, has not been studied.
In the very unusual breeding system of F. carica, we showed that the main reproductive cost of male individuals consists of the structures for the pollinators' sheltering. Compared to this investment, the cost of volatile production, as well as that of pollen production, is very low. These conclusions question the reproductive costs in the female trees. Indeed male and female figs perform different functions and therefore male and female fig trees may have distinct patterns of resource acquisition and allocation.
![[FIXME: no text available for sidebars:pub-links:email-alerts-free in any of these
languages: en-us]](/local/img/btn_free.gif)
