Flowering Dynamics in Arum italicum (Araceae): Relative Role of Inflorescence Traits, Flowering Synchrony, and Pollination Context on Fruit Initiation

by Marcos Méndez, Anita Díaz
Flowering Dynamics in Arum italicum (Araceae): Relative Role of Inflorescence Traits, Flowering Synchrony, and Pollination Context on Fruit Initiation
Marcos Méndez, Anita Díaz
American Journal of Botany
Start Page: 
End Page: 
Select license: 
Select License



'Departamento de Biologia de Organismos y Sistemas (Unidad de Ecologia), Universidad de Oviedo. E-33071 Oviedo, Spain; and

'Bournemouth University, School for Conservation Sciences, Dorset House, Talbot Campus, Fern Barrow, Poole, BH12 5BB UK

We studied the relative role of inflorescence traits, flowering synchrony, and pollination context for infructescence and fruit initiation in two Spanish populations of Arum italicum, a species in which inflorescences are the pollination unit. In this species, a specialized inflorescence organ. the appendix, is important for pollinator attraction. However, the short floral longevity and the production of mostly one inflorescence per plant make its pollination potentially dependent on strong flowering synchrony and on external factors not controlled by the plant (the pollination context). The flowering period in both sites lasted >3 mo. Day-to-day variation in simul- taneous antheses was high, and 11-50% of antheses occurred on days during which no pollen donor was present. Inflorescence traits, flowering synchrony, and between-plant distance all influenced infructescence and fruit initiation, but their relative importance differed between sites. In one large population, infructescence initiation was positively related to inflorescence traits; in a smaller population infructescence initiation increased with the number of donor inflorescences. In both sites, percentage of fruits initiated per infructescence was dependent on a combination of inflorescence traits, flowering synchrony, and between-plant distance. Plants producing 2-4 inflorescences had higher probability of infructescence initiation and overlapped their antheses with more plants than single-inflores- cence ones.

Key words: Araceae; Arum italicum; between-plant distance; floral display; flowering dynamics; flowering synchrony; inflores-

cence number; pollination context.

Pollen limitation commonly constrains reproductive success in flowering plants (Burd, 1994). Floral biology studies usually test the assumption that floral traits have some adaptive value to pollination and reproductive success (Waser, 1983). Shape or size of floral structures (Campbell et al., 1991), nectar re- wards (Mitchell, 1993), and floral display size (combining both size and number of flowers) (Andersson, 1991) are known to influence rates of pollinator visitation and subse- quent pollen export andlor receipt.

The hypothesis that floral traits play a significant role in pollination and reproductive success assumes either implicitly or explicitly that additional confounding factors are not present or can be safely neglected. However, reproductive success of particular individuals in one population can be more related to the context in which those individuals rewoduce than to their intrinsic properties. For example, several species require dis- assortative mating between members of two groups (e.g., pin- thrum, male-female). In such species, relative abundance of plants in every group (Wyatt and Hellwig, 1979; House, 1992) or the quality of the neighbors as pollen donors (e.g., their flower number; House, 1993) can affect reproductive success. In addition, negative effects of small population size and plant isolation on fecundity have been documented for species with

Manuscript received 8 December 2000; revision accepted 8 March 2001.

The authors thank Juan JosC MCndez, Nieves MCndez. Rosa MenCndez, Jorge Sostres, Rosa Viejo for helping with collection of field data; and Nick Buck, Christopher T. Ivey. Pedro Jordano, JosC M. G6mez, Jan Mikesell, Randall J. Mitchell, Jost Ram6n Obeso, Jeff Ollerton, 0rjan Totland, Helen Young, and several anonymous referees for useful suggestions on earlier drafts. This study was partially funded by a grant from the I1 Plan Regional de Investigaci6n del Principado de Asturias (Espafia) to MM.

"uthor for reprint requests. current address: Department of Plant Ecology, University of Uppsala, Villavagen 14, SE-752 36 Uppsala, Sweden (e-mail: marcos.mendez@ebc.uu.se).

diverse breeding systems (WidCn and WidCn, 1990; House, 1992; Kearns, Inouye, and Waser, 1998). Weather can also be an important determinant of fitness. Low temperature de- creased pollinator activity and seed set in two subarctic species (Bergman, Molau, and Holmgrem, 1996). We will refer to all the abovementioned factors as the flowering context. Despite their disparity they share two characteristics: they potentially influence the pollination and reproductive success and they are not factors on which natural selection can act (as opposed to intrinsic properties such as floral morphology or display).

Within-season timing of events can also influence the fitness of particular individuals in a population. Variation in pollina- tion success can be a result of weak flowering synchrony in hermaphrodites (Augspurger, 1981; Guitiin and Shnchez, 1992), as ~$11 as variation in floral sex ratio in monoecious (Le Corff, Agren, and Schemske, 1998) or gynodioecious (Williams, Kuchenreuter, and Drew, 2000) species. The pol- lination context, such as pollinator visitation rate (Totland, 1994) or weather (Totland, 1994; Bergman, Molau, and Holm- grem, 1996), may also vary over short time scales. Vagaries of the pollination context and its temporal variation can have strong consequences for reproductiveperformance of species producing one or few flowers. Long floral longevity (Primack, 1985) or high flowering synchrony (Rathcke and Lacey, 1985) of single- or few-flowered species can be considered as ad- aptations to reduce the effect of the pollination context in re- productive success.

Previous studies have shown that external factors can over- ride the effects of floral traits on fitness (Herrera, 1993; Mitch- ell, 1994; O'Connell and Johnston, 1998), although pollination context has been an understudied topic. Arum italicurn Miller (Araceae) is a good system to study how interactions among floral traits, flowering synchrony, and pollination context af-

TABLE1. Summary statistics for the inflorescence traits, flowering synchrony, and pollination context variables and for infructescence and fruit

initiation variables in two populations of Arum italicum. Range numbers beginning with -1 or -3 indicate that "donors" were present 1 or

3 d previous to anthesis, respectively.

Vanable Mean il so Range N


Appendix length (mm)

Number of "donors"

Number of "competitors"

Appendix length of nearest "donor" (mm)

Distance to nearest "donor" (m)

Fruits initiated per infructescence (%) AvilCs Appendix length (mm) Temperature excess ("C) Number of "donors" Number of "competitors" Appendix length of nearest "donor" (mm) Distance to nearest "donor" (m) Fruits initiated per infructescence (%)

a Mode, i.e., the most frequently occumng measurement in a data set (Zar, 1999), is given. For ordinal variables mode is biologically more

meaningful than mean when data distribution is asymmetrical.

day. We also counted the number of inflorescences in female (f) and male (m) phases. From these counts we obtained the daily variation in the number of newly open (female phase) inflorescences and daily variation in the number of "donors" (i.e., m) and putative "competitors" (i.e.,f-1) that each newly open inflorescence faced. In Ablaneda, the large population size made it im- practical to examine all the plants. Instead, we monitored all the plants within a 25 X 1.5 m plot (>200 plants, -40-50% of the plants at this site). Arum iralicum is fully self-compatible (A. Diaz, personal observation), which makes all pollen donors suitable as mates independent of genetic relatedness to the focal inflorescence.

For each newly open inflorescence we measured the following variables:

(a) appendix length, from its tip to the joint with the stipe, to the nearest millimeter; (b) appendix length of its nearest "donor" (an estimate of neigh- bor quality as pollen donor); (c) distance to nearest "donor," to the nearest centimeter. In Ablaneda, we searched for the nearest "donor" both within and outside the plot monitored. When inflorescences opened on days when no "donors" were in bloom, we noted the time elapsed from the most recent day in which "donors" had bloomed. The data on variables (b) and (c) above for these cases refer to the most recent day in which "donors" were present (usually the previous day; Table 1). Final sample size (inflorescence number) is shown in Table 1 and corresponds to 82 plants in Ablaneda (a subsample of all the flowering plants within the plot monitored) and 43 in Avilts (all the flowering plants).

For a subsample of 27 inflorescences in AvilCs, we estimated the heat production by the appendix using two digital thermometers. One of them, fixed with plastic tape to the appendix, measured the appendix temperature. The other, situated on one leaf, measured the ambient temperature. Temper- atures to the nearest 0.l0C were recorded every hour from the beginning of anthesis in the evening until next morning, when heat production stopped. We calculated the difference between appendix and ambient temperatures and used the highest difference recorded (temperature excess hereafter) as an es- timate of heat production by the appendix.

After anthesis, we estimated pollination success as infructescence and fruit initiation (sensu Stephenson, 1981), i.e., whether, and how many, flowers in the inflorescence initiated their fruit development. Thirty days after anthesis, we scored any inflorescence appearing dry and shriveled as unsuccessful and any inflorescence showing fruit development as an infructescence. We scored inflorescence status after 30 d because bagged inflorescences (i.e., unpolli- nated) dried and abscised by that time in a previous assay; despite being self- compatible, we never observed self-pollination in A. italicum. We excluded inflorescences from analyses if abscission occurred <21 d after anthesis. This early abscission, related to the presence of an unidentified pathogen, made it not possible to discern whether growth of fruits or ovules had occurred. For every infructescence, we analyzed fruits and recorded whether they developed (contained seeds or expanded ovules) or not (dry, brown fruits with unde- veloped ovules).

Statistical analyses-We considered two successive components of female fecundity: (1) the probability of infructescence initiation and (2) the percent- age of fruit initiation within an infructescence, i.e., number of fruits initiated per flower, expressed as a percentage. First, the relative influence of variables measured on the probability of infructescence initiation was identified by means of logistic regression (Hosmer and Lemeshow, 1989). We tested the significance of variables using the likelihood-ratio test (Trexler and Travis, 1993).

Second. we identified variables that influenced the percentage of fruit ini- tiation within an infructescence using multiple linear regression. We treated two discrete variables (number of "donors" and number of "competitors") as continuous for these analyses (Sokal and Rohlf, 1981), and we modeled "rain" as a dummy variable (Zar, 1999). No transformation was applied to these or any other variable, because the analysis of residuals did not show significant departures from the assumptions of the regression. No strong co- linearity was found between independent variables, according to the variance inflation factor (all values below 10) or condition index (all values below 30) (Philippi, 1993).

Disagreements exist on the best method (enter, forward, backward, or step- wise) of selecting variables in multiple regression. We ran logistic and linear regressions using both enter and stepwise methods, as performed by the sta- tistical package SPSS-PC 4.0 (Norusis, 1990). For logistic regression, we got identical results by using either enter or stepwise methods; for linear regres- sion, the stepwise method gave more conservative (is., less significant vari- ables) results than the enter method. In the following, we will give results using the enter method, signaling differences with stepwise method when necessary.

Throughout the text, values are given as means t 1 SD.


Flowering dynamics-The duration of the flowering period in both populations was similar (Fig. 1). Flowering lasted 119 d (10 February-8 June) in Ablaneda and 107 d in AvilCs (16 Febmary-2 June). In both sites, high variation in the number of newly open inflorescences was observed between succes- sive days (Fig. 1). Rain coincided with the anthesis of 29.2% (N = 106) of inflorescences in Ablaneda and of 32.8% (N = 58) in Avil6s.

TABLE 2. Results of logistic regression analysis for the probability of infructescence initiation by A. italicurn as a function of inflorescence traits, flowering synchrony, and pollination context variables measured in two populations. The coefficient B (t 1 SE), the partial correlation coefficient (R), a measure of the influence on the dependent variable (eB),change in model fit when removed (AG), and level of significance are given.

Population and variable BZlSE

Ablaneda (N = 106 inflorescences)


Appendix length

Distance to nearest "donor"


Number of "donors"

Number of "competitors"

Appendix length of nearest "donor" Avilts (N = 58 inflorescences)


Number of "donors"


Distance to nearest "donor"

Appendix length

Number of "competitors"

Appendix length of nearest "donor"

The overall percentage of infructescence initiation was 73.6% (N = 106) in Ablaneda and 65.5% (N = 58) in AvilCs. These percentages were not significantly different (G test: GI = 1.163, P = 0.281). Other summary statistics are presented in Table 1. We found that 11 % of inflorescences in Ablaneda (N = 106) and 50% in AvilCs (N = 58) opened on days when no "donor" was blooming. These percentages were signifi- cantly different (G test: G, = 29.169, P <0.0001). From those inflorescences, 67% (N = 11) initiated fruit in Ablaneda and 48% (N = 29) did in Avilts. These percentages did not differ significantly (G test: GI = 1.174, P = 0.279).

Relative influence of inflorescence traits, flowering synchrony and pollination context on infructescence and fruit initiation-At Ablaneda, the logistic regression analysis showed a positive relationship, as indicated by the sign of the partial correlation coefficient (R), between appendix length and the probability of infructescence initiation (Table 2). No other variables were significantly related to the probability of


infructescence initiation in this population. In the multiple re- gression analysis, percentage of fruits initiated per infructesc- ence was positively related to number of "donors," appendix length, and appendix length of nearest "donor" (Table 3). Dis- tance to nearest "donor" was negatively related to fruit initi- ation (Table 3). In the more conservative regression analysis using the stepwise method, only number of "donors" and ap- pendix length were significantly related to fruit initiation (Ta- ble 3). Sign and magnitude of coefficients of partial correlation for these two variables were similar in both models.

At AvilCs, the logistic regression analysis revealed that the relationship between number of "donors" and the probability of infructescence initiation was significantly positive (Table 2). For the subsample in which temperature excess was measured (N = 27), the appendix length of nearest "donor" was nega- tively and significantly related to the probability of infructesc- ence initiation (results not shown). Number of "donors" was only marginally significant (P > 0.05). None of the variables considered explained a significant amount of variance in per-

TABLE 3. Results of multiple-regression analysis for the percentage of fruit initiation within an infructescence as a function of inflorescence traits, flowering synchrony, and pollination context variables measured in two populations of A. italicurn. ANOVA results, coefficient of determination, R2, and partial regression coefficient, B (-C 1 SE), are given for the enter method. The t value and level of significance are given for both enter and stepwise (step) methods.

Population and variable B+lSE t (enter) P (enter) t (step) P (step)

Ablaneda (F,,,, = 4.436, P = 0.001, R2 = 0.36) Constant -17.76 ? 22.34 Appendix length 0.78 t 0.29 Number of "donors" 4.10 2 1.69 Distance to nearest "donor" -2.36 t 1.03 Appendix length of nearest "donor" 0.61 t0.29 Rain -9.30 t 9.15 Number of "competitors" 0.28 t 1.60

Aviltsa (F,, ,, = 15.565, P = 0.0001, R2 = 0.91) Constant -44.65 t 38.49 Temperature excess 10.08 -C 1.22 Number of "competitors" -14.16 t 2.56 Appendix length of nearest "donor" -1.25 -C 0.47 Number of "donors" 7.12 t 3.05 Appendix length 0.64 i0.42 Distance to nearest "donor" -2.64 2 2.04 Rain 10.46 t 9.79

aAnalysis only includes data for the subsample in which heat production was estimated.

TABLE4. Comparison between single and multi-inflorescence plants in two populations of A. italicurn. (A) Mean number i1 SD (sample size) of plants flowering on the same day. Results of ANOVAs testing differences between groups of plants are also given. (B) Percentage of plants initiating infructescences (sample size in pa- rentheses). Results of a binomial test for differences between groups of plants are also given. Only plants in which fruit initiation could be determined for all the inflorescences produced were in- cluded in the analyses.

A) No. of plants flowering on the same day Ablaneda 4.3 + 3.4 (27) 7.7 i3.9 (17)  <0.01  
AvilCs  1.0 2 0.8 (18)  1.8 i 1.2 (17)  <0.05  
B) Percentage of plants initiating infructescences   
Ablaneda  66.7% (27)  94.1% (17)  0.071  
AvilCs  41.2% (17)  88.2% (17)  <0.001  

centage of fruits initiated, using multiple regression. However, for the subsample of inflorescences in which heat production was estimated (N = 19 inflorescences initiating fruits), tem- perature excess and number of "donors" showed a positive relationship with fruit initiation (Table 3). In addition, number of "competitors" and appendix length of nearest "donor" were significantly negatively related to percentage of fruit ini- tiation (Table 3). Again, stepwise multiple regression was more conservative and showed only a significant relationship between temperature excess and number of competitors and fruit initiation (Table 3), although both sign and partial cor- relation coefficients were similar in both models.

Consequences for single-vs. multi-inflorescence plants- Percentage of multi-inflorescence plants (range: 2-4 inflorescences) was 63.9% (N = 122) in Ablaneda and 51.1% (N = 43) in AvilCs. In multi-inflorescence plants, inflorescences were produced sequentially and their antheses were separated by 33.7 -+ 8.3 d (range: 7-56 d, N = 43) in Ablaneda and

31.6 2 8.3 d (range: 18-52 d, N = 25) in AvilCs. Production of several inflorescences by the same plant extended the flow- ering period both at the plant and the population level (Fig. 1). At the population level, a second peak of flowering was apparent.

At both sites, multi-inflorescence plants overlapped their flowering with significantly more plants than single-inflores- cence ones (Table 4A). In addition, probability of infructesc- ence initiation was significantly higher for multi-inflorescence plants in AvilCs (Table 4B) and close to significantly higher in Ablaneda (Table 4B).


Specialized inflorescence structures, as appendices or nutri- tious staminodia, play a role in pollination of many Araceae (e.g., Knoll, 1926; Bown, 1988; Vogel, 1990; Young, 1990; Gottsberger and Silberbauer-Gottsberger, 199 1 ; Gibernau et al., 1999). Most previous studies on the importance of such structures for pollinator attraction have been qualitative (e.g., Lack and Diaz, 1991), but there is some quantitative evidence of the positive influence of size or number of those structures on pollinator attraction (Young, 1990; Gibernau et al., 1999). In the present study, appendix length was positively related to both the probability and amount of fruit initiation. In addition, fruit initiation increased with temperature excess, a component of floral display understudied until now in Araceae. A previous paper also showed a positive relationship between appendix length and pollinator visits in this specie5 (MCndez and Obeso, 1992).

Some evidence exists regarding the influence of pollination context, such as between-plant distance or weather, in pollen transfer or reproductive success in Araceae. Fruit set decreased as distance to male inflorescences increased in Arisaema tri- phyllum (Rust, 1980; but see Bierzychudek, 1982) and Amorphophallus johnsonii (Beath, 1996). Also, cool weather made midges pollinating Arum nzaculatum torpid (Kite, 1995). In A. italicum, fruit initiation decreased with increasing between- plant distance in Ablaneda, although this result was not con- sistent between regression models. On the other hand, any dai- ly variations in pollinator abundance or activity due to rain seemed to have little effect on infructescence initiation. Tem- perature could be a more important factor, but it was not mea- sured in this study. ~ver~thing

else being the same, we ex- ~ectedthat infructescence or fruit initiation would increase for inflorescences having "donors" with long appendices, because these should attract more pollinators (MCndez and Obeso, 1992) and, consequently, be better sources of pollinators for newly opened inflorescences. This was the result for Ablaneda, but the opposite effect occurred in AvilCs. In all cases, the effect of appendix length of nearest "donor" was weak and its significance changed among regression models. Further ob- servations should be made in order to test the importance of this aspect of "donor quality" for fruit initiation.

Flowering synchrony affected reproductive success in Arum ~naculatum(Ollerton and Diaz, 1999). In the same way, in the present study number of "donors" was positively related to either infructescence or fruit initiation at both sites studied. This was not surprising, because of the high daily variation in number of "donors." However, some ability of buffering against this variability seemed to be present in A. italicurn because infructescence initiation also occurred on days in which no "donors" were in bloom. This fact is opposed to findings of Beath (1996) for Amorphophallus johnsonii. In the latter species, no successful fertilization occurred unless pollen was transferred on the same day of anthesis. Results in A. italicum inflorescences could be due either to carryover of viable pollen on pollinators for 1 or 2 d or to pollen arrival from other populations. Unfortunately, data on pollen longev- ity in this species are lacking to test this.

The relative effect of inflorescence traits, flowering syn- chrony, and pollination context was population specific. The present study was correlative: thus, it is difficult to disentangle which population characteristics were responsible for such dif- ferences. One tentative explanation is offered here that should be tested bv means of future ex~erimental mani~ulation. In Ablaneda, large population, pribability of oveiapping an- theses was higher than in AvilCs (89 vs. 50%, see RESULTS), and this could mean both a lower uncertainty in the Ablaneda pollination context and a lower influence of number of "do- nors" on infructescence initiation. This, combined with a low- er average appendix length, could give better opportunities for the effect of inflorescence traits to become apparent.

A higher flowering synchrony would apparently be advan- tageous in A. italicurn (MCndez, 1998). A similar result was obtained by Ollerton and Diaz (1999) for A. inaculatum. However, flowering synchrony was not high in A. italicum (maximum of 11 simultaneous antheses in Ablaneda, a population in which 256 antheses were recorded; see Fig. 1) or other Araceae studied (Dieffenbachia longispatha-Young, 1988, 1990; Philodendron solimoesense-Gibernau et al., 1999). At present it is not known which proximate cues or ultimate mechanisms can be responsible for that asynchrony. Benefits of an increased flowering synchrony in A. italicum could be overridden by an increase in competition for pollen, as indi- cated by the negative effect of number of "competitors" on fruit initiation found in AvilCs.

Another way to cope with environmental uncertainty is to produce several inflorescences (Burd, 1995). But despite its benefits for multi-inflorescence plants, this is a limited possi- bility, because inflorescence number in A. italicum is size de- pendent (MCndez and Obeso, 1993). Production of additional inflorescences could potentially be dependent on the pollina- tion success of the first inflorescence but this possibility could not be explored with the data available. On the other hand, an increased inflorescence longevity could also buffer against a variable pollination context. Although many Araceae inflores- cences exhibit a short functional period (Gibernau et al., 1999; Ollerton and Diaz, 1999), extended flowering also exists (Wada and Uemura, 2000). To what extent phylogenetic con- straints (Ollerton and Diaz, 1999) or a high maintenance cost (Primack, 1985) of an energetically expensive floral display affect inflorescence longevity in Araceae should be addressed by future studies.

Finally, although we have mainly discussed the evidence concerning Araceae, the interaction among floral characters, flowering synchrony, and pollination context can be important for fitness in many other species producing one or a few flow- ers (Totland, 1994) or requiring disassortative mating between members of two groups (Wyatt and Hellwig, 1979; House, 1992). The pollination context could potentially affect selec- tive pressures on flowering synchrony or floral characters (Schemske and Horvitz, 1989) differently in populations of varying densities or sex ratios, as well as influence genetic neighborhoods (Young, 1988). This will be a fruitful topic for future studies.


ANDERSSON,S. 1991. Floral display and pollination success in Achillea ptar-

nzica (Asteraceae). Holarctic Ecology 14: 186-191. AUGSPURGER,

C. K. 1981. Reproductive synchrony of a tropical shrub: ex- perimental studies on effects of pollinators and seed predators on Hybanthus pr~nifolius (Violaceae). Ecology 62: 775-788.

BEATH, D. D. N. 1996. Pollination of Amorphophallus johnsonii (Araceae) by carrion beetles (Phaeochrous amplus) in a Ghanaian rain forest. Journal of Tropical Ecology 12: 409-418.

BERGMAN,F?, U. MOLAU, AND B. HOLMGREN. 1996. Micrometeorological impacts on insect activity and plant reproductive success in an alpine environment, Swedish Lapland. Arctic and Alpine Research 28: 196-202.

BIERZYCHUDEK,1982. The demography of jack-in-the-pulpit. a forest pe-

F? rennial that changes sex. Ecological Monographs 52: 335-351. BOWN, D. 1988. Aroids. Plants of the Arum family. Timber Press, Portland. Oregon, USA. BURD. M. 1994. Bateman's principle and plant reproduction: the role of pol- len limitation in fruit and seed set. Botanical Review 60: 83-139. BURD, M. 1995. Ovule packaging in stochastic pollination and fertilization environments. Evol~~rion

49: 100-109.

CAMPBELL,D. R.. N. M. WASER, M. V. PRICE, E. A. LYNCH,AND R. J. MITCHELL. 1991. Components of phenotypic selection: pollen export and flower corolla width in Ipo~nol~sis

aggregata. Evolution 45: 1458- 1467. CRUDEN,R. W., AND S. M. HERMANN-PARKER.

1977. Temporal dioecism: an alternative to dioecism? Evolution 31: 863-866. DAFNI,A. 1984. Mimicry and deception in pollination. Anncial Review oj Ecology and Systematics 15: 259-278.

FZGRI,K., AND L. VAN DER PIJL. 1979. The principles of pollination ecology, 3rd ed. Pergamon Press. Oxford, UK. FRIDLENDER,A. 1999. Observations sur la biologie de 1'Arum cylindraceurn Gasp. (Araceae) en Corse. Actu Botanica Gallica 146: 297-309. GIBERNAU, AND A. DEJEAN. 1999. Beetle pol-

M., D. BARABE, P. CERDAN, lination of Philodendron solimoesense (Araceae) in French Guiana. International Journal of Plant Sciences 160: 1135-1143.

GOTTSBERGER, 1991. Olfactory and G., AND I. SILBERBAUER-GOTTSBERGER. visual attraction of Eriocelis emarginata (Cyclocephalini, Dynastinae) to the inflorescences of Plzilodendron sellounz (Araceae). Biotropica 23: 23-28.

GUITIAN,J., AND J. M. SANCHEZ. 1992. Flowering phenology and fruit set of Petrocoptis gi-andijlora (Caryophyllaceae). International Journal of Plant Sciences 153: 409-412.

HERRERA,C. M. 1993. Selection on floral morphology and environmental determinants of fecundity in a hawk moth-pollinated violet. Ecological Monographs 63: 25 1-275.

HOSMER.D. W., AND S. LEMESHOW. 1989. Applied logistic regression. Wiley. New York, New York, USA. HOCSE, S. M. 1992. Population density and fruit set in three dioecious tree species in Australian tropical rain forest. Jo~tnzal of Ecologj~ 80: 57-69. HOUSE, S. M. 1993. Pollination success in a population of dioecious rain forest trees. Oecologia 96: 555-561.

KEARKS. C. A,, D. W. INOUYE,AND N. M. WASER. 1998. Endangered mu- tualisms: the conservation of plant-pollinator interactions. Annual Re- view of Ecology and Systematics 29: 83-1 12.

KITE, G. C. 1995. The floral odour of Arum rnaculatutn. Biochemical Systenz- atics and Ecology 23: 343-354.

KNOLL,E 1926. Insekten und Blumen. Experimentelle Arbeiten zur vertie- fung unserer kenntnisse iiber die Wechselbeziehungen Zwischen Pflanzen und Tieren. IV. Die Arum-bliitenstande und ihre Besucher. Abhandlungen der zoologisch-botanischen Gesellschaft in Wien 12: 379-481.

LACK. A. J.. AKD A. D~Az. 1991. The pollination of Arum maculatun~ L.-a historical revjew and new observations. Watso~zia18: 333-342.

LE CORFF, J.. J. AGREN,AND D. W. SCHEMSKE. 1998. Floral display, polli- nator discrimination, and female reproductive success in two monoecious Begonia species. Ecology 79: 1610-1619.

MEEUSE, B. J. D. 1975. Thermogenic respiration in aroids. Atznclal Review of Plant Physiology 26: 117-126. MENDEZ, M. 1997. Sources of variation in seed mass in Arunz italicum. In- ternational Journal of Plant Sciences 158: 298-305.

MENDEZ, M. 1998. Modification of phenotypic and functional gender in the monoecious Arum italicunz (Araceae). American Journal of Botany 85: 225-234.

MENDEZ, M., AND J. R. OBESO. 1992. Influencia del osmdforo en la pro- duccion de infrutescencias en Arum iralicunz Miller (Araceae). Anales del Jardfiz Botunico de Madrid 50: 229-237.

MENDEZ, M., AND J. R. OBESO. 1993. Size-dependent vegetative and repro- ductive allocation in Arcirn italicurn (Araceae). Canadian Jo~irrzal of Botany 71: 309-314.

MITCHELL,R. J. 1993. Adaptive significance of Iponzopsis aggregata nectar production: observation and experiment in the field. Evolution 47: 25-


MITCHELL,R. J. 1994. Effects of floral traits. pollinator visitation, and plant size on Ipornopsis aggregatu fruit production. American Nat~iralist 143: 870-889.

NORUSIS,M. 1990. SPSS introductory statistics student guide. SPSS Inc., Chicago, Illinois, USA. O'CONNELL,L. M.. AND M. 0. JOHNSTON. 1998. Male and female pollination success in a deceptive orchid, a selection study. Ecology 79: 1246-1260.

OLLERTON:J., AND A. D~Az. 1999. Evidence for stabilising selection acting on flowering time in Arum mac~ilatum (Araceae): the influence of phy- logeny on adaptation. Oecologia 119: 340-348.

PHILIPPI.T. E. 1993. Multiple regression: herbivory. In S. M. Scheiner and

J. Gurevitch [eds.]. Design and analysis of ecological experiments, 183-

210. Chapman and Hall, New York, New York, USA. PRIMACK,R. B. 1985. Longevity of individual flowers. Annual Review of Ecology and Systematics 16: 15-37. RATHCKE, B.. AND E. P. LACEY. 1985. Phenological patterns of terrestrial plants. Annual Revie\$, of Ecology and Systematics 16: 179-214.

RUST,R. W. 1980. Pollen movement and reproduction in Arisaema rriphyl- 6m. Bulletifz of tlze Torrey Botatzical Club 107: 539-542.

SCHEMSKE,D. W., AND C. C. HORVITZ. 1989. Temporal variation in selection on a floral character. Evolution 43: 461-465. SOKAL, R. R., AND E J. ROHLF. 1981. Biometry, 2nd ed. W. H. Freeman, New York, New York, USA. STEPHENSON,

A. G. 1981. Flower and fruit abortion: proximate causes and ultimate functions. Annual Review of Ecology and Systematics 12: 253-


TOTLAND,0. 1994. Intraseasonal variation in pollination intensity and seed set in an alpine population of Ranunculus acris in southwestern Norway. Ecography 17: 159-165.

TREXLER,J. C., AND J. TRAVIS.1993. Nontraditional regression analyses. Ecology 74: 1629-1637. UEMURA, S., K. OHKAWARA,

G. KUDO,N. WADA,AND S. HIGASHI. 1993. Heat-production and cross-pollination of the Asian skunk cabbage Symplocarpus renifolius (Araceae). American Journal of Botany 80: 635-

640. VOGEL, S. 1990. The role of scent glands in pollination: on the structure and function of osmophores. Amerind, New Delhi, India.

WADA, N., AND S. UEMURA. 2000. Size-dependent flowering behavior and heat production of a sequential hermaphrodite, Symplocarpus renifolius (Araceae). Anzerican Journal of Botany 87: 1489-1494.

WASER, N. M. 1983. The adaptive nature of floral traits: ideas and evidence. In L. Real [ed.], Pollination biology, 242-285. Academic Press, London, UK.

WIDEN, B., AND M. WIDEN. 1990. Pollen limitation and distance-dependent fecundity in females of the clonal gynodioecious herb Glechoma he- deracea (Lamiaceae). Oecologia 83: 191-196.

WILLIAMS,C. E, M. A. KUCHENREUTER, AND A. DREW. 2000. Floral di- morphism, pollination, and self-fertilization in gynodioecious Geranium richardsonii (Geraniaceae). American Journal of Botany 87: 661-669.

WYATT, R., AND R. L. HELLWIG. 1979. Factors determining fruit set in het- erostylous bluets, Houstonia caerulea (Rubiaceae). Systematic Botany 4: 103-1 14.

YAFUSO,M. 1993. Thermogenesis of Alocasia odora (Araceae) and the role of Colocasiomyia flies (Diptera: Drosophilidae) as cross-pollinators. Environmental Entomology 22: 601-606.

YOUNG, H. J. 1986. Beetle pollination of Dieffenbachia longispatha (Araceae). American Journal of Botany 73: 931-944. YOUNG, H. J. 1988. Neighborhood size in a beetle pollinated tropical aroid: effects of low density and asynchronous flowering. Oecologia 76: 461-


YOUNG,H. J. 1990. Pollination and reproductive biology of an understory neotropical aroid. In K. S. Bawa and M. Hadley [eds.], Reproductive ecology of tropical forest plants, 151-164. UNESCO and Parthenon, Par- is, France.

ZAR,J. H. 1999. Biostatistical analysis, 4th ed. Prentice-Hall, Upper Saddle River, New Jersey, USA.

  • Recommend Us