American Journal of Botany 78(6): 789-794. 1991.

ALLOCATION OF REPRODUCTIVE RESOURCES WITHIN AND AMONG INFLORESCENCES OF 1 LAVANDULA STOECHAS (LAMIACEAE) Go to table of contents

JAVIER HERRERA Departamento de Biologia Vegetal y Ecologia, Universidad de Sevilla, Apartado 1095, 41080 Sevilla, Spain

Patterns of fruit set were studied in Lavandula stoechas, a Mediterranean shrub commonly occurring in southern Spain. The small, hermaphroditic flowers of this shrub are aggregated into dense, headlike inflorescences and exhibit extensive variations in fecundity. It was shown that as the number of developing fruits in the inflorescence increased, the probability of a flower setting fruit, the size of seeds, and their germinability decreased, most likely because of strong within-inflorescence resource limitation. An experiment was designed to ascertain whether increased fertility in late-opening flowers could be induced through reallocation of reproductive resources between different inflorescences. The experiment consisted of removing half of the inflorescence buds from a set of plants and comparing their fecundity with that of intact individuals. Thinning did not increase the proportion of flowers setting fruit which, in fact, was slightly lower than that of intact individuals (probably due to some reduction of floral display brought about by thinning). Although treated plants produced heavier seeds than controls, results suggest that inflorescences of L. stoechas behave as autonomous modules among which resources cannot be reallocated. Predispersal seed predation by insects accounted on average for a 31% reduction in fruit set. Predation was found to be nonrandomly distributed within inflorescences, with most damage concentrated on late fruits (i.e., those with smaller and less germinable seeds).

Plants often have their flowers aggregated into higher-level units termed inflorescences. Species differ widely in the organization, compactness, and size of their inflorescences (Weberling, 1965), but in some cases flowers are so perfectly integrated and packed that the whole inflorescence may become a pollinator attraction unit by resembling a single flower (pseudanthium), as in the capitula ofthe Compositae (Faegri and Van der Pij 1, 1979). It is accepted that inflorescences represent gamet-packages molded mainly by selection pressures related to pollination, resource distribution, and fixed costs of associated structures (Schoen and Dubuc, 1990). The aggregation of flowers into semiautonomous, physiologically integrated units (Watson and Casper, 1984) sometimes results in a conflict between different floral functions (Wyatt, 1982). For example, many-flowered inflorescences increase pollinator attraction (Willson and Price, 1977; Cruzan, Neal, and Willson, 1988; Palmer, Travis, and Antonovics, 1988), ' Received for publication 2 August 1990; revision accepted 8 February 1991. The author thanks A. N. Andersen, J. Arroyo, H. Dobson, U. Molau, A. G. Stephenson, R. Wyatt, and two anonymous reviewers for helpful comments on earlier drafts of the manuscript. The Estacion Biologica de Donana (C.S.I.C.) provided permission to work in Donana. 789

but as pollination takes place, interovary competition for resources can dramatically affect fruit and seed set (Stephenson, 1979, 1980; Bawa and Webb, 1984; Holtsford, 1985). Thus, the flowers in an inflorescence may have different reproductive values for the plant but be morphologically identical. Among-flower variation in fecundity that occurs in hermaphroditic plants has resulted in a considerable body of literature on `surplus' flowers (Willson and Price, 1977; Stephenson, 1981; Queller, 1983), serial adjustment of reproductive resources (Lloyd, 1980; Lloyd, Webb, and Primack, 1980), flower/fruit ratios (Sutherland, 1986, 1987), and selective fruit abortion (Bookman, 1983; Stephenson and Winsor, 1986). There is evidence that plants can easily reallocate reproductive resources within inflorescences. Following flower or pod thinning, for example, larger seeds can be produced (Maun and Cavers, 1971) or the probability of setting fruit can change (Stephenson, 1979, 1980). Much less is known, however, on the relationship among flowers on different inflorescences. For example, to what extent does the probability of setting a fruit depend on the number of flowers elsewhere on the plant? The goals of the present study were to investigate the within-inflorescence pattern of fe-

790

AMERICAN JOURNAL OF BOTANY

cundity in Lavandula stoechas (Lamiaceae) and to gain some knowledge on the extent to which dense inflorescences act like independent reproductive modules. In L. stoechas, the architecture of the inflorescence determines a distinct pattern of flower opening that can be tracked when flowering has ceased, thus making the plant suitable for studying among-flower variations in fecundity and the factors associated with it. In addition, the whole set of inflorescence buds corresponding to a reproductive episode appear simultaneously on the plant just before the onset of flowering, which allows manipulation of overall flower number. Previous studies have demonstrated that neither the rate of nectar secretion (Munoz and Devesa, 1987) nor the number of seeds per fruit (Devesa, Arroyo, and Herrera, 1985) are evenly distributed within inflorescences. MATERIAL AND METHODS

The plant— Lavandula stoechas L. (Lamiaceae) is a xerophytic, aromatic shrub up to 1 m high that commonly occurs in shrubland communities of southern Spain. It grows on a wide range of soil types under dry, sunny conditions at elevations below 1,500 m. Populations differing in floral and inflorescence traits have been given varietal or subspecific status (Devesa, Arroyo, and Herrera, 1985). The results reported here refer to subspecies sampaiana Rozeira, characterized mainly by comparatively short and few-flowered inflorescences. Plants grow and flower in most populations during late winter and early spring (February through May; Herrera, 1986) and are dormant during summer and autumn. A variety of Lepidoptera, Diptera, and Hymenoptera (mostly honeybees) pollinate L. stoechas in southern Spain (Herrera, 1988). The flowers of L. stoechas are hermaphroditic and tubular (Fig. 1), have a dark purple, tubular corolla, and secrete minute amounts of sugar-rich nectar (Herrera, 1985). The pollen : ovule ratio is near 1,000. L. stoechas is self-compatible, but self-pollination seldom if ever takes place because of very strong protandry (Devesa, Arroyo, and Herrera, 1985; Munoz and Devesa, 1987). Inflorescences are headlike, nearly cylindrical aggregations of dichasia at the top of a peduncle, terminated by several showy, sterile, purple bracts (Fig. 2). The dichasia are arranged into four vertical columns at right angles, each column bearing four to five dichasia, and each dichasium bearing five flowers (Fig. 3). The first flower to open in any dichasium is referred to here as a central flower, and those opening later in the sequence

[Vol. 78

are referred to as lateral flowers. An inflorescence bears 60—100 flowers, of which four of the lateral type will be found for each of the central type. Anthesis of central flowers is remarkably synchronous and restricted to the first 10 days of the 40—50 day blooming period of an inflorescence. Within any inflorescence, the anthesis of central flowers never overlaps that of lateral flowers. The ovary develops into zero to four nutlets (seeds, hereafter), each of which is 1 mm long. Upon detachment from the receptacle, seeds remain enclosed within the calyx until dispersal. Ovaries from central flowers set an average of three seeds, while those from lateral flowers set just one seed on average (Devesa, Arroyo, and Herrera, 1985). Regardless of whether they contain any seed or not, calyces are not shed until the inflorescence disintegrates, which occurs several months after seed dispersal. At fruiting, calyces from lateral and central flowers can easily be identified.

Study site and field procedures— Field work was conducted at the Reserva Biologica de Donana (Donana National Park, southern Spain), a sandy, coastal area with a Mediterranean climate where shrubland is the dominant vegetation (Rivas-Martinez et al., 1980). Rainfall, which is highly seasonal, averages 600 mm/yr. At the study site, a population of L. stoechas grows on old, stabilized sand dunes, intermingled with other sclerophyllous shrub species. In February 1989, ten similarly sized L, stoechas individuals, each bearing 22—74 inflorescence buds, were selected and tagged. Plants were then randomly assigned to one of the following manipulations: 1) every second inflorescence bud was carefully clipped (thinning treatment); or 2) no manipulation (control). At that time, plants in the population bore a complete set of half-sized buds (Fig. 4). After flowering, and slightly before the onset of seed dispersal (June), four heads were collected at random from each plant, and a number of calyces therein sampled for the presence of ripe, brown nutlets. Sampling — To test for within-inflorescence variation in fecundity, sampling of calyces was stratified so as to include all of the central flowers (10—15) and 15 lateral flowers per head. In total, fruit production was monitored in 975 calyces (from 377 central and 598 lateral flowers) from 40 inflorescences and ten plants. In L. stoechas, calyces may contain either only black and unenlarged (presumably unpollinated) ovules and/or aborted, pale-yellow seeds (i.e., the flower has failed to set fruit), or one

June 1991]

HERRERA—ALLOCATION OF RESOURCES IN LAVANDULA

791

1 mm

O 10mm

Figs. 1—4. Morphology of flowers and inflorescences in Lavandula stoechas. 1. Side view of flower. 2. Mature inflorescence. 3. Schematic representation of the horizontal section of a spike showing the arrangement of dichasia around the axis. Larger circles represent flowers that open first. Black circles represent secondary flower buds that invariably abort. The complete set of flowers is shown in only two adjacent dichasia for simplicity. 4. Inflorescence in bud.

to four brown, viable seeds. Additionally, close examination of apparently viable seeds under a dissecting microscope revealed that some had been bored and their content removed by predators (parasitic hymenopteran larvae; J. Herrera, personal observation). Invariably, the larvae destroyed all seeds within attacked calyces. When examining fruit set per inflorescence, a separate record was kept of the proportion of

ovaries setting either viable or bored, unviable seed.

Estimates of fecundity — Fruit set was defined as the number of calyces containing at least one brown (preyed or not) seed, divided by the number of flowers. The predation rate was defined as the ratio of preyed fruits to total fruit number. Effects of inflorescence removal,

792

[ Vol. 78

AMERICAN JOURNAL OF BOTANY

W ithin-inflorescence variability in the fecundity of thinned and control L. stoechas plants (in each case, N = 5) a

TABLE 1.

TABLE 2.

Results from analyses of variance for effects of inflorescence removal (Treatment ), type of flower ((Flower), and plant identity (`Plant) on fecundity in L. stoechas

Variables

Flower' type

Fruit set

Seed mass

Predation

81.4(3.3) 72.0 (3.3) 76.7 (2.4)

3.9(0.1) 3.1 (0.1) 3.5 (0.1)

12.7(4.4) 43.9 (6.7) 28.3 (4.7)

Untreated plants C L Overall Thinned plants C L Overall

65.7 (4.8) 72.3 (2.9) 69.0 (2.8)

4.2 (0.1) 3.7 (0.1) 3.9 (0.1)

18.3 (7.1) 49.8 (6.8) 34.1 (5.5)

a

Values are means for 20 inflorescences, and numbers in brackets are standard errors. Fruit set and predation rate are percentages. Seed masses are in mg. See Table 2 for the significance of differences. b C, central; L, lateral flowers.

Dependent variable

Source

Fruit set

Treatment Flower Plant

Seed mass

Treatment Flower Seedinessa

Predation rate

Treatment Flower Plant

Total

Total

Total

df

F

1 1 8

P

3.755 0.897 1.907

0.057 0.347 0.073

1 1 3 74

12.073 29.846 0.341

0.001 <0.001 0.796

1 1 8

0.775 26.299 2.632

0.382 <0.001 0.014

69

69

Refers to whether the seed was in a fruit with one, two, three, or four seeds. a

flower position, and individual plant on fecundity were tested by analysis of variance (ANOVA) with the factor `plant' nested under treatment. Calculations were performed on arcsine-transformed data (Sokal and Rohlf, 1981). Seed mass was determined by weighing batches of five air-dried seeds to the nearest 0.1 mg on an electronic balance. Only good, not preyed seeds were used. In all, 40 batches were weighed for each flower type (central or lateral), 20 per treatment (control or thinning), and five for each seediness score (i.e., whether the seed came from a fruit with either one, two, three, or four seeds). After being log-transformed, seed mass was analyzed by ANOVA with treatment and seediness as the main effects. Because of the difficulty of gathering an adequately large sample, seeds from different heads and plants had to be pooled. The viability of seeds from central and lateral flowers was also estimated and their germination rates compared by sowing 300 nutlets of each type on moist filter paper inside petri dishes (100 nutlets per dish) kept at 14—20 C. The seeds were examined daily under a dissecting microscope for the emergence of the radicle. RESULTS Mean number of heads exposed to pollinators after manipulation was 33.8 per plant (SE = 5.8, range 22—55) for controls, and 19.4 per plant (SE = 4.0, range 11—32) for thinned individuals. Among controls, central flowers set fruit at a significantly higher rate than lateral flowers (t = 2.368, P = 0.023), while this pattern vanishes if thinned individuals are considered (t = 0.831, P = 0.411). Maximum dis-

tance between any of the experimental plants was 30 m. Percent fruit set ranged from 81.4% (for central flowers on controls) to 65.7% (for central flowers on thinned individuals), suggesting relatively higher rates of insect visitation to intact plants during the earliest phases of flowering (Table 1). Overall, intact plants set more fruit than thinned plants, although differences are not statistically significant. Plant identity had no effect on percent fruit set (Table 2). Seeds developing from central flowers were significantly heavier (Tables 1, 2) and exhibited a higher germination rate (Fig. 5) than those developing from lateral flowers. Thinning inflorescences significantly increased the mass of seeds produced by treated plants (Table 2). In both treated and control plants the probability of a fruit being attacked by predators was more than two times higher among lateral than among central flowers (Table 1), and the differences are statistically significant (Table 2). Across all plants and flower types, predation ranged from 12.7% to 49.8%, and averaged 31% of fruits. Thinning had no significant effect on the rate of fruit predation, while the identity of plants showed a moderately significant effect. DISCUSSION Fruiting—In intact L. stoechas plants, those flowers that open first and are directly attached to the inflorescence axis (Fig. 3) set fruit at a higher rate than late-opening flowers (Devesa, Arroyo, and Herrera, 1985; this study, Table 1). This is most unlikely to result from changes in pollen deposition rates through the flowering

June 1991]

HERRERA—ALLOCATION OF RESOURCES IN LAVANDULA

5

10

15

20

DAYS FROM SOWING Fig. 5. Germination of seeds from fruits on different positions of the inflorescence in Lavandula stoechas. C, fruits from central; L, from lateral flowers.

season, since Munoz and Devesa (1987) have demonstrated that removing the buds of central flowers causes the lateral ones to increase their fruiting rate. Thus, low fertility in lateopening flowers results from resource rather than pollen limitation. Moreover, our finding that average seed size (and germinability) decreases from early to late fruits is also indicative of resource exhaustion and of increased within-inflorescence embryo competition for resources as pollination proceeds. Consequently, variation in fruiting is undoubtedly the outcome of a peculiar pattern of resource partitioning determined by inflorescence architecture. This would be similar to what has been reported for Catalpa (Stephenson, 1979, 1980) and A sclepias ( Willson and Price, 1977; Wyatt, 19 81; Bookman, 1983), where early competition among ovaries within inflorescences is an important determinant of eventual levels of fruit set. Inflorescence bud removal affected the fruit set of early-opening flowers on treated plants (Table 1). Probably, the reduction in floral display brought about by inflorescence thinning resulted in decreased pollen deposition during the earliest phases of flowering. Additionally,

793

flowers on thinned plants developed in fruits with seeds that were heavier than those from flowers on controls (Tables 1, 2). My suggestion is that such an increase in seed mass reflects a within-inflorescence compensation for the lower fruit set of central flowers. However, plants do not appear to reallocate resources among inflorescences, as indicated by the fact that bud removal did not increase percent fruit set of treated plants. In this regard, it is remarkable that the fecundity of late-opening flowers on thinned plants remained virtually identical to that of equivalent flowers on controls (Table 1). The compact inflorescences of L. stoechas probably behave to a great extent as isolated modules of resource allocation. Explanations for the lack of response of lat eral flowers to inflorescence removal should relate to an interaction between growth and reproduction. The formation of new foliage in L. stoechas is simultaneous to flowering, and it was observed that thinned plants exhibited an unusually vigorous growth (unpublished data). Leaf buds nearest to the removed inflorescences might outcompete the remaining inflorescences for the resources not expended on flower, fruit, and seed production. Thus, a compensation for the loss of reproductive potential among flowers on the same inflorescence could often occur (for example, through increased seed mass), while between-inflorescence compensation is likely to be impeded through interactions with vegetative growth.

Predation — L. stoechas predators apparently discriminate between fruits according to their position on the inflorescence. Many plant species experience high levels of predispersal seed predation (i.e., Janzen, 1971; Andersen 1988), and it is not uncommon to find that predation is nonrandomly distributed among the available fruits. Differential predation can operate in other species on the basis of phenology (Augspurger, 1981) or fruit seediness (Zimmerman, 1980; Herrera, 1984). The mechanism by which early fruits of L. stoechas mostly escape predation while late fruits are heavily preyed is unknown, but whatever the proximate cause, it results in the destruction of fruits with fewer, smaller, and less germinable seeds (Table 1; Fig. 5). This is similar to what happens in Bartsia alpina (Scrophulariaceae; Molau, Eriksen, and Knudsen, 1989). Investigations are currently in progress to determine if differential predation could exert a selection pressure on inflorescence architecture and/or promote phenological escape in this shrub.

794

AMERICAN JOURNAL OF BOTANY

LITERATURE CITED ANDERSEN, A. N. 1988. Insect seed predators may cause far greater losses than they appear. Oikos 52: 337340. AUGSPURGER, C. K. 1981. Reproductive synchrony of a tropical shrub: experimental studies on effects of pollinators and seed predators on Hybanthus prunifolius (Violaceae). Ecology 62: 775-788. BAWA, K. S., AND C. J. WEBB. 1984. Flower, fruit, and seed abortion in tropical forest trees: implications for the evolution of paternal and maternal patterns. A merican Journal of Botany 71: 736-751. BOOKMAN, S. S. 1983. Costs and benefits of flower abscission and fruit abortion in A sclepias speciosa. Ecology 64: 264-273. CRUZAN, M. B., P. R. NEAL, AND M. F. WILLSON. 1988. Floral display in Phyla incisa —consequences for male and female reproductive success. Evolution 42: 505515. DEVESA, J. A., J. ARROYO, AND J. HERRERA. 1985. Con tribucion al conocimiento de la biologia floral del genero Lavandula L. A nales Jardin Botanico Madrid 42: 165-186. FAEGRI, K., AND L. VAN DER PUL. 1979. The principles of pollination ecology, 3d ed. Pergamon, Oxford. HERRERA, C. M. 1984. Selective pressures on fruit seediness: differential predation of fly larvae on the fruits of Berberis hispanica. Oikos 42: 166-170. HERRERA, J. 1985. Nectar secretion patterns in southern Spanish Mediterranean shrublands. Israel Journal of Botany 34: 47-58. . 1986. Flowering and fruiting phenology in the coastal shrublands of Donana, south Spain. V egetatio 68: 91-98. . 1988. Pollination relationships in southern Spanish Mediterranean shrublands. Journal of Ecology 76: 274-287. HOLTSFORD, T. P. 1985. Nonfruiting hermaphroditic flowers of Calochortus leichtlinii (Liliaceae): potential reproductive functions. A merican Journal of Botany 72: 1687-1694. JANZEN, D. H. 1971. Seed predation by animals. A nnual Review of Ecology and Systematics 2: 465-492. LLOYD, D. G. 1980. Sexual strategies in plants. I. An hypothesis of serial adjustment of maternal investment during one reproductive session. New Phytologist 86: 69-79. , C. J. WEBB, AND R. B. PRIMACK. 1980. Sexual strategies in plants. II. Data on the temporal regulation of maternal investment. New Phytologist 86: 81-92. MAUN, M. A., AND P. B. CAVERS. 1971. Seed production and dormancy in Rumex crispus. II. The effects of removal of various proportions of flowers at anthesis. Canadian Journal of Botany 49: 1841-1848. MOLAU, U., B. ERIKSEN, AND J. T. KNUDSEN. 1989. Pre -

PDF PDF

PDF

[Vol. 78

dispersal seed predation in Bartsia alpina. Oecologia 81: 181-185. MUfvoz, A., AND J. A. DEVESA. 1987. Contribucion al conocimiento de la biologia floral del genero Lavandula L. II. Lavandula stoechas L. subsp stoechas. A nales Jardin Botanico Madrid 44: 63-78. PALMER, M., J. TRAVIS, AND J. ANTONOVICS. 1988. Seasonal pollen-flow and progeny diversity in A mianthium muscaetoxicum — ecological potential for multiple mating in a self-incompatible, hermaphroditic perennial. Oecologia 77: 19-24. QUELLER, D. C. 1983. Sexual selection in a hermaphroditic plant. Nature 305: 706-707. RIVAS-MARTINEZ, S., M. COSTA, S. CASTROVIEJO, AND E. VALDES. 1980. Vegetacion de Donana (Huelva, Espana). Lazaroa 2: 5-189. SCHOEN, D. J., AND M. DUBUC. 1990. The evolution of inflorescence size and number. A. gamete packaging strategy in plants. A merican Naturalist 135: 841-857. SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry, 2d ed. W. H. Freeman, San Francisco. STEPHENSON, A. G. 1979. An evolutionary examination of the floral display of Catalpa speciosa (Bignoniaceae). Evolution 33: 1200-1209. 1980. Fruit set, herbivory, fruit reduction, and the fruiting strategy of Catalpa speciosa (Bignoniaceae). Ecology 61: 57-64. . 1981. Flower and fruit abortion: proximate causes and ultimate functions. A nnual Review ofEcology and Systematics 12: 253-279. , AND J. A. WINSOR. 1986. Lotus corniculatus regulates offspring quality through selective fruit abortion. Evolution 40: 453-458. SUTHERLAND, S. 1986. Patterns of fruit-set: what controls fruit-flower ratios in plants? Evolution 40: 117-128. . 1987. Why hermaphroditic plants produce many more flowers than fruits: experimental tests with A gave mckelveyana. Evolution 41: 750-759. WATSON, M. A., AND B. B. CASPER. 1984. Morphogenetic constraints on patterns of carbon distribution in plants. A nnual Review of Ecology and Systematics 15: 233258. WEBERLIING, F. 1965. Typology of inflorescences. Botanical Journal of the Linnean Society 59: 215-221. WILLSON, M. F., AND P. W. PRICE. 1977. The evolution of inflorescence size in A sclepias (Asclepiadaceae). Evolution 31: 495-511. WYATT, R. 1981. The reproductive biology of A sclepias tuberosa, II. Factors determining fruit set. New Phytologist 88: 375-385. . 1982. Inflorescence architecture: how flower number, arrangement, and phenology affect pollination and fruit set. A merican Journal of Botany 69: 585-594. ZIMMERMAN, M. 1980. Reproduction in Polemoniumn: pre-dispersal seed predation. Ecology 61: 502-506.

Introduction Material and methods The plant Study site and field procedures Sampling Estimates of fecundity Results Discussion Fruiting Predation Literature cited Figure 1-4: Morphology of flowers Figure 5: Germination of seeds Table 1: Inflorescence variability Table 2: Removal and fecundity