Relationships between the occurrence of misshapen fruit on late-season strawberry in the United Kingdom and infestation by insects, particularly the European tarnished plant bug, Lygus rugulipennis

In experiments in which the European tarnished plant bug, Lygus rugulipennis, was caged on the developing flowers or young fruits of strawberry, the insects caused malformation of the fruits. Another species of capsid, Plagiognathus chrysanthemi, caused similar damage; this species is less numerous than L. rugulipennis on late-season crops of strawberry in UK. Other insects which sometimes occur in large numbers in the flowers of late-season strawberry, i.e. various species of thrips and pollen beetles, did not cause fruit malformation in caging experiments, though thrips sometimes caused discoloration of the fruit. In field experiments where numbers of L. rugulipennis were reduced by the use of insecticides, the amount of misshapen fruit was reduced greatly compared to untreated plots. Correlations between the numbers of L. rugulipennis present at the early stages of fruit development and damage scores for fruit deformity were highly significant. This capsid is likely to be the major cause of fruit malformation in late-season crops of strawberry in the UK.     

Transmission of prunus necrotic ringspot virus using plum pollen and thrips

Prunus necrotic ringspot virus (PNRSV) was transmitted to cucumber but not to peach seedlings after they were dusted with infective plum pollen and caged with 8–10 thrips per seedling for 24 h. When the pollen was taken from three plum trees shown by mechanical inoculation tests to have highly infective flower buds and pollen, the transmission rates to cucumber seedlings were 56% with Thrips tabaci and 66% with a mixture of five thrips species collected from Ageratum houstonianum flowers. However, the transmission rate averaged only 7% when pollen was taken from five other plum trees which had flowers with less infectivity in sap transmission tests. In 1990 T. imaginis, T. tabaci and T. australis, which were present in the mixture of thrips from A. houstonianum, also formed the major part of the thrips population in flowers of the plum trees used as the pollen source.     

The coincidence of thrips and dispersed pollen in PNRSV-infected stonefruit orchards – a precondition for thrips-mediated transmission via infected pollen

Recent laboratory studies have demonstrated that Prunus necrotic ringspot virus (PNRSV) (family Bromoviridae) can be readily transmitted when thrips and virus‐bearing pollen are placed together on to test plants. For this transmission mechanism to result in stonefruit tree infection in the field, PNRSVbearing pollen must be deposited onto surfaces of stonefruit trees on which thrips also occur. In a previous paper, we demonstrated that almost all pollen in a PNRSV‐infected Japanese plum orchard in southeastern Queensland was deposited onto flowers, whereas few grains occurred on leaves and none on stems. Here, we present results of our investigation of thrips species composition, distribution and abundance on stonefruit trees in the same study area as our previous pollen deposition study. We collected a total of 2010 adult thrips from 13 orchards during the 1989, 1991 and 1992 flowering seasons of which all but 14 were in the suborder Terebrantia. Most (97.4%) terebrantian thrips were of three species, Thrips imaginis, Thrips australis and Thrips tabaci. Thrips tabaci as well as species mixtures that included T. imaginis, T. australis and T. tabaci have been shown to transmit PNRSV via infected pollen in laboratory tests. Adult thrips were frequently collected from flowers but rarely from leaves and never from stems. Large and significant differences in numbers of T. imaginis, T. australis and T. tabaci adults in flowers occurred among orchards and between seasons. No factor was conclusively related to thrips numbers but flowers of late‐flowering stonefruit varieties tended to hold more thrips than those of early‐flowering varieties. Our results indicate that the common thrips species present on stonefruit trees in the Granite Belt are also ones previously shown to transmit PNRSV via infected pollen in the laboratory and that these thrips are concentrated in tonefruit flowers where most stonefruit pollen is deposited. These results contribute to mounting circumstantial evidence that stonefruit flowers may be inoculated with PNRSV via an interaction of thrips with virus‐bearing pollen and that this transmission mechanism may be an important cause of new tree infections in the field     

Pollination disruption by European honeybees in the Australian bird-pollinated shrubGrevillea barklyana (Proteaceae)

European honeybees (Apis mellifera) were less efficient pollinators ofGrevillea barklyana than nectar-feeding birds. Nectar-collecting honeybees did not contact reproductive parts of flowers. Pollen-collecting honeybees preferentially visited malestage flowers but rarely visited female-stage flowers. Fruit set on caged inflorescences that allowed access to honeybees but excluded birds was reduced by more than 50% compared to inflorescences that were visited by both types of visitors. Further, fruit set on caged inflorescences was less than on bagged inflorescences that excluded both birds and honeybees, indicating that pollen removal by bees decreased opportunities for delayed autonomous selfing in the absence of birds. Although fruit set was not pollen-limited at the study site, pollen removal by honeybees would decrease fruit set in small populations where birds are scarce. In addition, pollen removal by honeybees would reduce opportunities for outcrossing and reproductive success through male function. Although honeybees have been in Australia for insufficient time to have exerted selection on floral traits, evolutionary shifts in response to these animals are likely to occur in the future.     

Effects of pollen quality and quantity on pollen limitation in Crataegus monogyna (Rosaceae) in NW Spain

Pollen limitation occurs when plants produce less fruits and/or seeds than they would with adequate pollen receipt. If the addition of cross-pollen to stigmas increases fruit/seed production, it is interpreted as an evidence of pollen limitation. Much of the limitation may be associated with the quality rather than quantity of pollen; however, most studies do not discriminate between the two, which may lead to misinterpretation of the results. We studied the effects of quality and quantity of pollen on the reproduction of a northern Spanish population of Crataegus monogyna. The treatments included self- and cross-pollination, and supplementation to open and bagged flowers. The response variables considered were number of pollen grains per stigma, pollen tubes per style, and initial and final fruit set. In the Cantabrian range, C. monogyna requires insect pollinators to set fruit and is partially self-incompatible. We found that the number of pollen tubes did not differ between cross- and self-pollination treatments; however, self-pollinated flowers set less fruits than flowers that received pure cross-pollen or were supplemented with both cross- and self-pollen. The experimental design allowed us to infer qualitative rather than quantitative pollen limitation. Comparison of the number of pollen grains and tubes, and initial and final fruit set among pollination treatments suggested post-zygotic embryo selection against selfed progeny.     

Insect-induced wind pollination of the palm Chamaedorea pinnatifrons and pollination in the related Wendlandiella sp.

In Peruvian Amazonia pollination ecology of two palms, Chamaedorea pinnatifrons (Jacq.) Oerst. and Wendlandiella sp. was studied from October 1988 to January 1990. Both palms are dioecious understory species that form locally dense populations.In C. pinnatifrons both sexes flower synchronously during the dry season. Prior to anthesis, the pendulous male inflorescence is inhabited by numerous thrips (Thysanoptera) and Ptiliidae (Coleoptera). Staminate flowers open by a small basal slit between the petals. At anthesis pollen is shed and the movements of the insects inside the flowers trigger pollen release in small clouds. Thus, the powdery pollen becomes airborne and finally air currents act as a vector, carrying pollen to the inconspicuous female plants, which usually are not visited by insects. The term insect induced wind pollination is suggested for this pollination mode. Wendlandiella flowers during four months in the dry season. Male and female plants were not visited by insects. The dry condition of the pollen indicates that anemophily is the pollination mode in Wendlandiella. Fructification is rare but the plants show intense vegetative propagation.The significance of anemophily in the tropical lowland rain forest is discussed.     

Effects of indiscriminate foraging by tropical hummingbirds on pollination and plant reproductive success: experiments with two tropical treelets (Rubiaceae)

Summary In cloud forest at Monteverde, Costa Rica, two common treelets (Palicourea lasiorrachis and Cephaelis elata, both Rubiaceae) depend simultaneously on one hummingbird population (Lampornis calolaema) for pollination. Both species are distylous and self-incompatible. In laboratory experiments, we examined possible effects of indiscriminate foraging by hummingbirds among flowers of both species, as observed in the field, on pollination of Palicourea. In each of 35 trials, captive L. calolaema probed 2 flowers from pin plants of Palicourea followed by 20 thrum flowers of the same species, with either 0, 2, or 10 Cephaelis flowers intervening. We assessed pollen transfer by staining and counting pin pollen tubes growing in thrum styles; counts of 0, 1, or 2 pollen tubes relate directly to seed output (0, 1, or 2 seeds per fruit, respectively). Intervening Cephaelis flowers sharply reduced pollen receipt by thrum flowers of Palicourea and reduced some aspects of pollen dispersal from pins as well, thereby curtailing maternal and paternal reproductive potential of Palicourea. Such effects of interspecific pollen loss on reproductive output may lead to strong competition among some, though not all, combinations of plant species pollinated by L. calolaema or of other plant combinations that share animal pollinators.     

Increased control of thrips and aphids in greenhouses with two species of generalist predatory bugs involved in intraguild predation

The combined release of species of generalist predators can enhance multiple pest control when the predators feed on different prey, but, in theory, predators may be excluded through predation on each other. This study evaluated the co-occurrence of the generalist predators Macrolophus pygmaeus Rambur and Orius laevigatus (Fieber) and their control of two pests in a sweet pepper crop. Both predators consume pollen and nectar in sweet pepper flowers, prey on thrips and aphids, and O. laevigatus is an intraguild predator of M. pygmaeus. Observations in a commercial sweet pepper crop in a greenhouse with low densities of pests showed that the two predator species coexisted for 8 months. Moreover, their distributions in flowers suggested that they were neither attracted to each other, nor avoided or excluded each other. A greenhouse experiment showed that the predators together clearly controlled thrips and aphids better than each of them separately. Thrips control was significantly better in the presence of O. laevigatus and aphid control was significantly better in the presence of M. pygmaeus. Hence, combined inoculative releases of M. pygmaeus and O. laevigatus seem to be a good solution for controlling both thrips and aphids in greenhouse-grown sweet pepper. The predators are able to persist in one crop for a sufficiently long period and they complement each other in the control of both pests. This study also provides further evidence that intraguild predation does not necessarily have negative effects on biological control.     

Plant size,geitonogamy and seed set in Ipomopsis aggregata

Summary We used powdered fluorescent dyes to estimate receipt of self vs. outcross pollen in the self-incompatible species Ipomopsis aggregata (Polemoniaceae). Flowers on small and large plants received equal amounts of outcross pollen, whereas flowers on large plants received more self pollen, so the proportion of self pollen delivered through geitonogamy increased with plant size. In natural populations emasculation of all flowers on a plant raised average seed set per flower from 5.19 to 6.99 and also raised fruit set, though not significantly. From these results one expects a negative correlation between plant size and seeds per flower. The opposite trend was observed in a sample of plants in the field, suggesting that deleterious effects of geitonogamy on female fecundity in large plants can be overruled by other factors such as size-related fruit or seed abortion. Results are discussed in relation to the evolution of gynodioecy.     

Gender specialization in <Emphasis Type="Italic">Palicourea demissa</Emphasis> (Rubiaceae), a distylous,hummingbird-pollinated treelet

Distyly has been interpreted as a mechanism that promotes cross-pollination among conspecific plants and as one of the routes leading to the evolution of dioecy. In one of the possible evolutionary pathways, pollinators may disrupt intermorph pollen flow, and, as a consequence, floral morphs may gradually specialize as either male or female (functional dioecy). Natural patterns of pollen deposition and fruit and seed production were estimated in Palicourea demissa (Rubiaceae) and used as parameters to assess functional gender differences between floral morphs. Pollen flow was asymmetrical in P. demissa. Long-styled flowers were more effective than short-styled flowers in pollen deposition towards compatible stigmas, whereas short-styled flowers were more effective in legitimate pollen receipt. Accordingly, short-styled plants produced more fruits and viable seeds than long-styled plants. The contributions of male and female function to the potential functional gender were equivalent in both morphs. However, the realized functional gender deviated significantly from the potential functional gender in both morphs, in which short-styled plants were more successful through their female function, but long-styled plants through their male function. If pollinators disrupt the complementarities of pollen transfer between the two morphs (asymmetric pollen flow), the expression of a more profitable gender is expected in each morph. Thus, our results support the hypothesis that dioecy may evolve in distylous populations through the gradual specialization of each morph as either male or female.     

REPRODUCTIVE STRATEGY OF CLEISTES DIVARICATA (ORCHID ACEAE)

In Cleistes divaricata, a hinged anther dispenses a sequence of loosely aggregated pollen tetrad masses onto the dorsal thorax of Bombus and Megachile workers. This mechanism allows more than one bee to receive pollen from a single flower. To determine the effectiveness of this strategy, plants in bud were caged and hand pollinated using pollen from sequential releases. Capsules produced from a flower's first three pollen dumps were significantly larger and heavier and contained more seeds than those produced by later pollen deposits with fewer tetrads. Decreasing pollen dosage did not affect fruit set, capsule development, individual seed weight, or percentage of healthy looking embryos per fruit. Fruit set for artificial pollinations was 93% for pollen dumps 1–3 and 97% for dumps 4–18. As set of naturally pollinated marked plants was only 38%, low frequency of pollinator visits apparently limited seed production. Efficacy of infrequent visits is maximized, however, because a flower's first pollen releases contain the most pollen, but later low-dosage dumps can produce some seeds. An unexpected cost was a higher investment of pericarp per seed in low-dosage capsules (130 ng pericarp seed^-1 fresh weight) than in high-dosage pollinations (only 27 ng pericarp seed^–1). Flower fading in plants receiving low pollen dosage was as slow as that in caged, unpollinated controls, thus increasing opportunity for additional pollination.     

The distribution of immature thrips among flowers of faba beans in commercial crops and experimental plots

Numbers of eggs and larvae of thrips {Kakothrips pisivorus (Westw.)) infesting pistils of various stocks of faba beans (Viciafaba L.) were recorded in different environments. By the end of flowering in spring beans, 25 times as many flowers were infested with, on average, twice as many thrips, as at the onset of flowering in winter beans. Within each of two seasons, levels of infestation varied up to 30-fold with location. At a single location, some bean stocks were much more heavily infested (greater proportions of flowers contained thrips and/or greater numbers of thrips per infested pistil) than others. Levels of infestation were not clearly associated with the closed-flower character (cf). Stocks with lush growth were more heavily infested than those with either generally restricted growth or with terminal inflorescences (ti). Those with white flowers, long pods or cytoplasmic male-sterility were more infested than those which contrasted for these three characters. The path of the pollen tubes was blocked by eggs of thrips in less than 0–5% of flowers. Infestation had no apparent effect on the fertilisation of flowers.     

云南草寇花序内不同部位的性分配

两性花植物花序内的性分配常存在差异,资源竞争、结构效应、交配环境(雌雄异熟、传粉者定向访花行为等)或授粉不均匀等几种假说可以解释这种现象。为验证上述假说,该研究以云南草寇两种表型(雄先熟型和雌先熟型)为材料,分析了其花序内不同部位(基部、中部和顶部)的每花花粉数、胚珠数、花粉/胚珠比、结实率和结籽率,花序内传粉者的定向访花行为,以及人工辅助授粉和去花处理对结实率和结籽率的影响。结果表明:两种表型花序内每花花粉数不随部位而变化,每花胚珠数、结实率和结籽率由基部到顶部依次降低,每花花粉/胚珠比由基部到顶部依次增加,表明顶部花存在偏雄的性分配。人工辅助授粉后,结实率、结籽率仍由基部到顶部依次降低,表明授粉不均匀假说不能解释云南草寇花序内不同部位结实率、结籽率的差异。去除基部和中部花后,顶部花人工辅助授粉条件下的结实率、结籽率与基部花人工辅助授粉条件下的结实率、结籽率无差异,表明云南草寇花序内不同部位结实率、结籽率的差异主要由资源竞争引起。雌先熟表型每花花粉数、花粉/胚珠比高于雄先熟表型,表明两种表型存在性分配差异。传粉者主要先访问云南草寇基部的花,然后向顶部移动。云南草寇花序内顶部偏雄的性分配可能是由资源竞争和传粉者定向访花造成的。     

Pollination ecology of Yucca elata

The pollination biology of a population of 250 Yucca elata (Liliaceae) plants was studied in southern New Mexico. Yucca elata and the prodoxid yucca moth Tegeticula yuccasella have a mutualistic association that is essential for the successful sexual reproduction of both species. However, a wide range of other invertebrate species visit flowers during the day and at night. Our aim was to quantify the role of yucca moths and other invertebrate visitors in pollination and fruit set, using manipulative field experiments. Inflorescences were bagged during the day or night (N=12 inflorescences) to restrict flower visitors to either nocturnal or diurnal groups. Yucca moths were active exclusively nocturnally during the flowering period and thus did not visit inflorescences that were unbagged during the day. None of the 4022 flowers exposed only to diurnal visitors set fruit, whereas 4.6% of the 4974 flowers exposed only to nocturnal visitors (including yucca moths) produced mature fruit. The proportion of flowers producing fruit in the latter treatment was not significantly different from unbagged control inflorescences. In a series of experimental manipulations we also determined that: (1) flowers opened at dusk and were open for two days on average, but were only receptive to pollen on the first night of opening; (2) pollen must be pushed down the stigmatic tube to affect pollination; and (3) most plants require out-cross pollination to produce fruit. The combination of these results strongly suggests that yucca moths are the only species affecting pollination in Y. elata, and that if another species was to affect pollination, it would be a rare event.     

Effectiveness of nectarivorous birds and honeybees as pollinators of Banksia spinulosa (Proteaceae)

Abstract The effectiveness of nectarivorous birds and honeybees (Apis mellifera) as pollinators of Banksia spinulosa (Proteaceae) was investigated. Birds visited inflorescences in the early, mid and late flowering seasons. In contrast, honeybees visited only on days in the late flowering period when maximum temperatures exceeded 15°C. Self pollen remained on pollen presenters of flowers for up to 5 days in the early and mid periods. In the late period, when honeybees visited inflorescences, self pollen was removed within 2 days. Pollen removal was similar for caged (birds excluded) and open inflorescences in the late period, indicating that most pollen was removed by honeybees. In the early and mid periods, honeyeaters pollinated 22% and 27% of flowers on open inflorescences, respectively. In the late period, when both birds and bees visited inflorescences, 64–73% of flowers on open inflorescences were pollinated. Foraging by honeybees resulted in pollen deposition as 38% of flowers on caged inflorescences were pollinated. Throughout the flowering season a similar number of pollen grains was deposited per stigma. There were 3.0–3.7 pollen grains per stigma on open inflorescences in the late period, although only 2.0 grains per stigma on caged inflorescences. In the early and mid periods, fewer caged than open inflorescences produced fruits, indicating the importance of honeyeaters to reproductive success at these times. In contrast, in the late period when honeybees visited inflorescences, fruit-set was similar on caged and open inflorescences. Overall, these results indicate that honeybees were effective pollinators of B. spinulosa.     

Population size,pollinator visitation and fruit production in the deceptive orchid Calypso bulbosa

It has been proposed that in non-rewarding animal-pollinated plants the pollination intensity should decrease with increasing population size and should increase with increasing local abundance of reward-producing plants. To test these hypotheses, we examined how population size, local abundance of Salix caprea, and tree cover were related to pollen removal and fruit production in 16 populations of the deceptive, early-flowering and bumblebee-pollinated orchid Calypso bulbosa in northern Sweden in 3 consecutive years. To determine whether fruit production was limited by pollinator visitation, supplemental hand-pollinations were performed in three populations in 3 years. Finally, to examine whether increased fruit production was associated with a reduction in future flower production, vegetative growth or survival, supplemental hand-pollination was repeated for 5 years in one population. The levels of pollen export, pollen deposition, and fruit set of C. bulbosa varied considerably among years and among populations. The proportion of plants exporting pollen was negatively related to population size, and positively related to density of S. caprea and to tree cover in 1 of the 3 years. In the other 2 years, no significant relationship was detected between proportion of plants exporting pollen and the latter three variables. In no year was there a significant relationship between fruit set and population size, density of S. caprea and tree cover. There was substantial among-year variation in the extent to which fruit production was limited by insufficient pollen deposition and in the amount of weather-induced damage to flowers and developing fruits. Fruit set was consistently higher in hand-pollinated than in open-pollinated plants, but this difference was statistically significant in only one of 3 years. Supplemental hand-pollination in 5 consecutive years increased cumulative fruit production 1.8 times, but did not affect flower production, plant size, or survival. Tree cover was negatively correlated with the incidence of frost damage in 1 year. The results indicate that life-time seed production may be pollen limited in C. bulbosa, and that variation in population size and local abundance of the early-flowering, nectar-producing S. caprea can only partly explain the extensive variation in pollinator visitation among populations of this species.     

The effects of the floral infection by a bacterial pathogen in a dioecious plant,Mallotus japonicus (Euphorbiaceae)

Among all the different organs of a plant, flowers might have one of the most dynamic microbial communities, since many microbes are transmitted during flowering by insects and pollen. However, little is known about how these microbes affect floral characteristics and plant reproduction. Among the microbes transmitted to flowers, pathogens may have highly negative effects on plants' fitness. In this study, we investigated whether a bacterial pathogen, Erwinia mallotivora, occurs on flowers of the host plant Mallotus japonicus, and whether the transmission of the pathogen to flowers can result in systemic infection and/or reduction of fruit production. The pathogen has been reported to infect through leaves, while its ecology on flowers is unknown. We first confirmed the presence of the pathogen on flowers, indicating possible transmission by visitors or pollen. Then, we showed that the bacteria can infect the plant through flowers by inoculating the pathogen to both male and female flowers. Interestingly, the symptoms on leaves appeared earlier on the female plants than on the males. Besides, the inoculation significantly decreased fruit set of the female plants. Our results suggest a higher cost of infection in a female than in a male once the pathogen infected flowers. Although the effects of pathogen infection to flowers have rarely investigated in wild plants, it would be an interesting topic for future study if such sexual differences in the infection cost can cause sexual conflict and intraspecific adaptation load.     

Are nectar-robbers mutualists or antagonists?

As exploiters of plant-pollinator mutualisms, nectar-robbers remove rewards (nectar) without providing pollination services. Though one might expect nectar-robbing to be costly to plants, it may instead benefit plants by indirectly increasing pollen dispersal. I investigated the direct effects of nectar-robbing bees (Xylocopa californica) on floral rewards and behaviors of pollinators visiting desert willow (Chilopsis linearis) and indirect effects of robbing on the reproductive success of the plant. Nectar-robbers reduced nectar; while unrobbed and robbed flowers were equally likely to contain nectar, nectar volumes were smaller in robbed flowers with nectar. Apis mellifera (honeybees), ineffective pollinators in terms of pollen deposition, avoided robbed flowers. In contrast, Bombus sonorus (bumblebees), effective pollinators, did not avoid robbed flowers. While bumblebees tended to spend less time in robbed flowers, the time that they spent in flowers was not correlated with pollen deposition. Using powder mimicking pollen, I found that on some days, powder was dispersed farther or to more flowers from robbed flowers, indicating that robbing may sometimes benefit plants by increasing male reproductive success. Powder movement suggested that the effect of robbing on male reproductive success ranged from costly to beneficial. The outcome for flowers that were marked early each morning was a function of prevalence of robbing and abundances of effective pollinators, but not a function of spatial variability among trees in prevalence of robbing or the abundance of ineffective honeybees. Unlike powder dispersal, female reproductive success, measured by fruit set and the number of pollen tubes growing in styles, was not affected by robbing. Thus, robbers did not reduce plants female reproductive success either directly by damaging flowers or indirectly by reducing pollen deposition by pollinators. Overall, this study indicates that nectar-robbers were not often costly to plants, and sometimes even benefited plants.     

Transmission of sowbane mosaic virus by Thrips tabaci in the presence and absence of virus-carrying pollen

When pollen of sowbane mosaic sobemovirus (SoMV)-infected Chenopodium amaranticolor was dusted onto C. amaranticolor and C. quinoa test seedlings which were then infested with 5–10 adult Thrips tabaci, SoMV was transmitted to 25%C. amaranticolor and 88% of C. quinoa plants. Five hours access of T. tabaci to pollen-contaminated C. quinoa seedlings was sufficient for SoMV transmission, but 1 h was not. SoMV was also transmitted when T. tabaci was mixed with pollen of C. amaranticolor and then placed on C. quinoa test seedlings. Further, plant-to-plant transmission of SoMV occurred in the absence of virus-carrying pollen when T. tabaci was caged on infected non-flowering C. amaranticolor or C. quinoa for 1–8 days and then adults transferred to C. quinoa test seedlings. This is the first time that thrips have been shown to be a vector of SoMV, and that a virus outside the ilarvirus group is reported to be transmitted using pollen and thrips.     

Determinants of seed production in Geranium maculatum

Summary We experimentally examined factors limiting seed production in two populations of the perennial woodland herb Geranium maculatum in central Illinois, USA. To test the pollinator-limitation hypothesis, we compared the seed production of plants whose flowers were supplementarily pollinated with outcross pollen to that of control plants receiving natural pollination only. To test if fruit production by early flowers suppresses fruit and seed formation by late flowers, a third group of plants was prevented from producing seed from the first 50% of the flowers to open (stigmas were excised at flower opening). Finally, to test if seed maturation and flower initiation are correlated with photosynthetic capacity, we performed a defoliation experiment in which either the stem leaves within the inflorescence, the stem leaves below the inflorescence, or the rosette leaves were removed during late flowering. Plants that reccived supplemental pollination produced 1.5–1.6 times more seeds than control plants. We found no difference between hand-pollinated plants and controls in mortality, flowering frequency or number of flowers produced in the year following the experiment. In both control and hand-pollinated plants, the fruit set and total seed production of early flowers were more than twice as high as those of late flowers. In one of the two populations, plants whose early flowers were prevented from setting seed produced significantly more seeds from their late flowers than did control plants. Seed predation was low and did not differ between early and late flowers. Leaf removal did not affect seed number or size in the year of defoliation, nor did it reduce survival or flower production in the subsequent year. This suggests that the plants were able to compensate for a partial defoliation by using stored resources or by increasing photosynthetic rates in the remaining leaves. Taken together, the results demonstrate that both pollinator activity and resource levels influence patterns of seed production in G. maculatum. While seed production was pollinatorlimited in both populations, a seasonal decline in resource availability was apparently responsible for the low seed production by late flowers.