Variation in Floral Sex Allocation and Reproductive Success in Sequentially Flowering Inflorescence of Corydalis remota var. lineariloba （Fumariaceae）

In hermaphroditic plants, female reproductive success often varies among different positions within an inflorescence. However, few studies have evaluated the relative importance of underlying causes such as pollen limitation, resource limitation or architectural effect, and few have compared male allocation. During a 2-year investigation, we found that female reproductive success of an acropetally flowering species, Corydalis remota Fisch. ex Maxim. var. lineariloba Maxim. was significantly lower in the upper late developing flowers when compared with the lower early flowers. Supplementation with outcross pollen did not improve female reproductive success of the upper flowers, while removal of the lower developing fruits significantly increased female reproductive success of the upper flowers in both years, evidencing resource limitation of the upper flowers. Female production in upper flowers was greatly improved by simultaneous pollen supplementation of the upper flowers and removal of the lower fruits, suggesting that, when resources are abundant, pollen may limit the female reproductive success of the upper flowers. The less seed mass in the upper flowers didn＇t increase in all treatments due to architecture. In the upper flowers, ovule production was significantly lower and the pollen ： ovule ratio was significantly higher. These results suggest that male-biased sex allocation in the upper flowers may lead to increased male reproductive success, whereas the lower flowers have higher female reproductive success.     

An Examination of the Function of Male Flowers in an Andromonoecious Shrub Capparis spinosa

The pollen donor and pollinator attractor hypotheses are explanations for the functions of the male flowers of andromonoecious plants. We tested these two hypotheses in the andromonoecious shrub Capparis spinosa L. （Capparaceae） and confirmed that pollen production and cumulative volume and sugar concentration of nectar do not differ between male and perfect flowers. However, male flowers produced larger anthers, larger pollen grains and smaller ovaries than perfect flowers. Observations on pollinators indicated that two major pollinators （Xylocopa valga Gerst and Proxylocopa sinensis Wu） did not discriminate between flower morphs and that they transferred pollen grains a similar distance. However, there were more seeds per fruit following hand pollination with pollen from male flowers than from perfect flowers. Individuals of C. spinosa with a larger floral display （i.e. bearing more flowers） received more pollen grains on the stigma of perfect flowers. Female reproductive success probably is not limited by pollen. These results indicate that male flowers of C. spinosa save resources for female function and that they primarily serve to attract pollinators as pollen donors.     

Fitness consequences of altering floral circadian oscillations for Nicotiana attenuata

Ecological interactions between flowers and pollinators are all about timing.Flower opening/closing and scent emissions are largely synchronized with pollinator activity,and a circadian clock regulates these rhythms.However,whether the circadian clock increases a plant's reproductive success by regulating these floral rhythms remains untested.Flowers of Nicotiana attenuata,a wild tobacco,diurnally and rhythmically open,emit scent and move vertically through a 140° arc to interact with nocturnal hawkmoths.We tethered flowers to evaluate the importance of flower positions for Manduca sexta-mediated pollinations;flower position dramatically influenced pollination.We examined the pollination success of phase-shifted flowers,silenced in circadian clock genes,NaZTL,NaLHY,and NaTOCi,by RNAi.Circadian rhythms in N.attenuata flowers are responsible for altered seed set from outcrossed pollen.     

Elevation-related variation in the population characteristics of distylous Primula nivalis affects female fitness and inbreeding depression

The population characteristics of distylous species are highly sensitive to stochastic natural selection pressure.Therefore,populations growing under different environmental conditions may vary in floral morph ratios,potentially affecting female fitness and leading to inbreeding depression.However,the variation in offspring quality among populations as a result of inbreeding depression is poorly understood in distylous species.This study investigates variations in plant density,seed mass,seed viabilityfemale fitness,and post-dispersal inbreeding depression in both sexual morphs(long-styled and shortstyled plants)of the distylous Primula nivalis that were subjected to different pollination treatments along an elevational gradient from 1657 to 2704 m a.s.l.Population characteristics(morph plant density and ratio)and fruit set were significantly affected by sexual morph and elevation.Plant density and fruitset frequencies were lower for short-styled than for long-styled plants at 2704 m a.s.l.The seeds from the cross-pollinated flowers of both morphs were higher in quality than those of self-pollinated flowers.The female fitness of seeds from cross-pollinated flowers of both morphs was higher than that of seeds from open-pollinated and self-pollinated flowers.The female fitness of seeds from long-styled flowers was higher than that of seeds from short-styled flowers at all elevations.Inbreeding depression increased with elevation among plants with short-styled flowers but not among those with long-styled flowers.Variation in the elevation-dependent mating system might influence female fitness and affect inbreeding depression in both floral morphs.In conclusion,the low quality of seeds from short-styled flowers at high elevations might decrease short-styled flower frequency,affecting population characteristics.     

Evidence of Dichogamy in Santalum album L.

Flowering, fruit set, embryological development, and pollination trials were Investigated In Santslum album L. Each ovary may have three to four ovules. Mlcrosporogenesls and megasporogenesls In the same flower were synchronized at the earlier stages of flower development. However, at anthesls, when pollen was mature, the magaspore had developed only to the stage of a one- to two-nucleus embryo sac. As the eight-nucleus embryo sac developed, some mamelon cells began to undergo programmed cell death, forming holes Into which the eight-nucleus embryo sacs extended, becoming ＂N＂ or ＂S＂ shaped. The development from a two-nucleus embryo sac to a matured eight-nucleus embryo sac lasted up to 10 d. Fruit-set from open pollination was less than 2%. The endosperm develops prior to division of the zygotic embryo and one to three embryos and endosperms were formed In the same fruit. A mature seed usually germinates to produce one seedling; however, two and three seedlings from one seed were also observed, albeit at a low frequency. Pollination trials showed that no seed sets when Inflorescences were covered with s bag; however, artificial pollination could Improve fruit set. Our pollinaUon trials and embryological studies proved that the flower of S. album Is dlchogamous and fruit set has high heterozygosity.     

Potential Autonomous Selfing in Gesneria citrina （Gesneriaceae）, a Specialized Hummingbird Pollinated Species with Variable Expression of Herkogamy

Species with mixed mating systems often demonstrate variable expression of breeding system characteristics and thus represent the opportunity to understand the factors and mechanisms that promote both outcrossed and selfed seed production. Here, we investigate variation in levels of herkogamy （variation in stigma-anther separation distance） in a Puerto Rican population of hummingbird pollinated Gesneria citrina Urban. There is significant variation in herkogamy levels among individuals of this species and stigma-anther separation is negatively associated with the ability to set fruits and seeds in the absence of pollinators. The variation in levels of herkogamy may represent a mechanism to ensure the production of some self-fertilized progeny in the absence of hummingbird pollinators. We also describe a novel breeding system in G. citrina, where stamens elongate over time to reach stigma height, but stamen elongation is accelerated by pollination. These results suggest that once the flowers are pollinated, stamen elongation may favor increased pollen removal and siring success, while the reduction in stigma-anther distance no longer imposes the risk of interference between male and female functions. We discuss our findings of breeding system variation in the context of pollination system evolution in an island setting （Antillean islands）.     

Influences of Cross Pollination on Pollen Tube Growth and Fruit Set in Zuili Plums （Prunus salicina）

Zuili plum （Prunus salicina L.） trees usually set fruit poorly, although they produce high quality fruit. To elucidate the causes of the poor fruit set, pollen tube growth into pistils and fruit set percentage were investigated after cross-, self- and open-pollination. Ovule development in Zuili pistils was also investigated. Pollen tube penetration into the ovules via the obturator and micropyle was best when Zuili pistils were pollinated by cv. Black Amber （P. domestica） pollen grains, although cross-pollinations with Hongxinli and Miili （P. salicina） pollen were more effective than self- and open-pollination. The fruit set percentage was also highest in pistils pollinated with Black Amber pollen grains. Morphological observation of Zuili pistils revealed that the trees produce ＂double pistils＂, developing two ovaries from a basal pistil, at a rate as high as 28%. In such abnormal pistils, most ovules were lacking an embryo sac or were entirely degenerated. The percentage of normally developed ovules was 24.3% and 8.9% in normal and double pistils, respectively. From these results, we conclude that the main causes of poor fruit set of Zuili plums are a lack of effective cross-pollination and the production of high percentages of double pistils in which normally developed ovules are scarcely formed.     

苏门白酒草的繁殖特征与入侵性的关系以及某些进化含义

Conyza sumatrensis （Retz.） E. Walker, a member of Asteraceae, is a highly invasive species. However, its reproduction biology remains poorly known. To understand the role of reproductive traits in successful invasion of the species, we studied several traits of its reproductive system： the miniature capitulum and gynomonoecious sexual system, the biology and phenology of capitula and florets, pollen/ovule ratio, the mating system （selfcompatibility）, flower visitors, physical traits and dispersal potential of achenes, germination potential of achenes from manually pollinated capitula, and the association of these traits with invasiveness. Our study showed that the reproductive traits of autonomous seed production, versatile mating system of self- and cross-pollination, and generalized pollination system might contribute to the species＇ successful invasive capability. The invasiveness was further enhanced by the high and rapid production of achenes, as well as the high percentage, rapid germination rate and high dispersal capability of achenes. It was concluded that in annual or winter-annual weeds, autonomous seed production contributed significantly to the invasiveness of the species.     

GNOM-LIKE 2, Encoding an Adenosine Diphosphate-Ribosylation Factor-Guanine Nucleotide Exchange Factor Protein Homologous to GNOM and GNL1, is Essential for Pollen Germination in Arabidopsis

In flowering plants, male gametes are delivered to female gametophytes by pollen tubes. Although it is important for sexual plant reproduction, little is known about the genetic mechanism that controls pollen germination and pollen tube growth. Here we report the identification and characterization of two novel mutants, gnom-like 2-1 （gnl2-1） and gn12-2 in Arabidopsis thaliana, in which the pollen grains failed to germinate in vitro and in vivo. GNL2 encodes a protein homologous to the adenosine diphosphate-ribosylation factor-guanine nucleotide exchange factors, GNOM and GNL1 that are involved in endosomal recycling and endoplasmic reticulum-Golgi vesicular trafficking. It was prolifically expressed in pollen grains and pollen tubes. The results of the present study suggest that GNL2 plays an important role in pollen germination.     

A palynotaxonomic study of the genus Filipendula （Rosaceae）

Pollen grains from 15 species （18 taxa） of the genus Filipendula were examined with light and scanning electron microscopy. It was revealed that the pollen grains are isopolar, tricolporate, with scabrate or scabrate-microechinate surface. The pollen morphology was compared with the conventional classification systems of the genus by different authors, and supported Shimizu＇s system （1961）, in which the genus was divided into three subgenera. The monotypic subgen. Hypogyna is characterized by pollen lacking fastigium and thickened costae colpi. The other monotypic subgen. Filipendula differs from others by pollen having larger grain, larger pore size, longitudinally elliptic fastigium and thickened costae colpi. The largest subgen. Ulmaria is distinguished by pollen having rounded or latitudinally elliptic fastigium and thickened costae colpi. Sectional classification was not supported by the pollen morphology due to insufficient variability. Key words Filipendula, palynology, systematics.     

Variation of pollination and resource limitation in a low seed-set tree, Liriodendron chinense (Magnoliaceae)

Pollen limitation and resource limitation have been documented as the major factors responsible for plants commonly producing more ovules than seeds, but few studies have examined pollen deposition directly in natural populations at different sites and times. We investigated the causes of low seed set in four populations of Liriodendron chinense (Magnoliaceae), an insect‐pollinated endangered tree endemic to southern China, over 2–3 years. One pistil potentially produces two ovules. The number of pistils per flower varies among populations, but in three of the four populations the variation in a given population was not significantly different among years. Overall, populations with higher pistil numbers tend to set more seeds per flower, but a positive correlation between pistil numbers and seed production per flower was observed in only one of the four populations. The numbers of pollen grains deposited per stigma varied from 0 to 60. The proportion of pollinated stigmas per flower ranged from 44% to 88% among populations and years. The numbers of pollen grains deposited per stigma and the percentages of pollinated stigmas were significantly different between populations, and two populations showed significant differences between years. A positive correlation between stigmatic pollen load and seed set was sought in ten population‐by‐year combinations but, in a given population, high stigmatic pollen loads did not always result in high seed set. Examination of pollen deposition, pistil and seed production over several sites and years showed that in addition to pollination, other factors such as resource or genetic loads were likely to limit the (lower than 10%) seed set in L. chinense. It appears that small, isolated populations experience severe pollination limitation; one population studied had seed/ovule ratios of 0.84% and 1.88% in 1995 and 1996. © 2002 The Linnean Society of London, Botanical Journal of the Linnean Society, 2002, 140 , 31–38.     

Pollination Rates and Pollen Tube Growth in a Vulnerable Plant,Liriodendron chinense (Hemsl.) Sarg. (Magnoliaceae)

The limited seed production of insect-depended plant, Liriodendron chinense was once considered to be pollen-limited and insufficient cross pollination. In this study, we counted pollen grains deposited on stigmas in three populations in Guizhou, Hunan and Zhejiang provinces of China respectively. Over 61.9% stigmas were pollinated. From 1994 to 1996, the mean number of pollen grains on each stigma ranged from 4.4 to 42.6, much more than ovules(2) in each pistil. Based on observations of three years, both the pollination rate and pollen quantity on stigmas significantly affected seed set. When flowers opened without stamens dehiscencing at the early stage of anthesis, stigmas received considerable quantity of pollen grains. Pollen grains from different sources were able to germinate and pollen tube growth rates were not greatly variable. It is very likely that pollen grains arrived first would fertilize eggs. Since only several pollen tubes went through the stylar canal, the potential pollen competition may exist. In this case, there would be strong selection on floral syndrome which benefit early insect visits. Pollen grains from the early visits would have a greater chance to fertilize ovules than those from later visits, which implies that cross pollination is the predominant breeding system of this plant. The conclusion was also confirmed by following four artificial experiments. Three treatments, including flowers bagged, netted or with the perianth removed, all reduced seed set clearly, but flowers with the stamens removed (emasculation)had a higher seed production by open pollination. As the rates of deposited stigmas in three populations were 6～8 times more than full seed set, we considerthat lower seed production in this plant may not mainly be due to pollen limitation.     

Ovule pre-emption and pollen limitation in a self-fertile perennial herb (Blandfordia grandiflora,Liliaceae)

The extent, frequency and causes of pollenlimited seed production were examined in partially selffertile populations of Blandfordia grandiflora for 2 years. Percentage seed set of open-pollinated plants (50–57%) did not differ within or between years, and was about 19% less than experimentally cross-pollinated plants (70–75%). Floral visits by honeybees did not differ through the flowering season and the number of pollen grains deposited on stigmas within 12 h of flowers opening exceeded the number of ovules per flower, indicating that the quality rather than the quantity of pollination limited seed set. Pollen limitation was caused by concurrent self- and cross-pollination and the subsequent abortion of some selfed ovules due to inbreeding depression. Natural seed set (55%) was intermediate between selfed (43%) and crossed (75%) flowers and was not increased when flowers that had been available to pollinators for 24 h were hand cross-pollinated, suggesting that ovules were already fertilized. Similarly, experimental pollination with both cross and self pollen within 24 h of flowers opening did not increase seed set relative to natural seed set, indicating that both cross- and self-fertilizations had occurred. In contrast, when selfing followed crossing by 48 h, or vice versa, seed set did not differ from crossed-only or selfed-only flowers, respectively, indicating that ovules were pre-empted by the first pollination. Collectively, these results indicate that under natural conditions self pollen pre-empts ovules, rendering them unavailable for cross-fertilization. This selfing reduces fecundity by 50%, as estimated from the natural production of cross seeds when selfing was prevented. Consequently, selection should favour floral traits, such as increased stigma-anther separation or protandry, that reduce interference between male and female functions that leads to selfing.     

Differential self‐fertilization rates in response to variation in floral traits within inflorescences of Aquilegia buergeriana var. oxysepala (Ranunculaceae)

To assess variation in the proportion of self‐fertilized seeds among flowers within inflorescences and the relationship between floral traits and the rate of self‐fertilization, the proportion of self‐fertilized seeds among individual flowers was estimated using ten microsatellite markers in self‐compatible plants of Aquilegia buergeriana var. oxysepala. Within‐inflorescence variation in floral traits, such as the duration of the male and female phases, flower size, herkogamy and the number of pollen grains and ovules in two natural populations, were investigated. The first flower in an inflorescence produced more seeds and a higher proportion of self‐fertilized seeds than the second flower. The higher proportion of self‐fertilized seeds in the first flowers was accompanied by a higher number of pollen grains and ovules in the bud stage and the female phase. These results indicate that the high proportion of self‐fertilized seeds in the first flowers in an inflorescence may be due to the high number of remaining pollen grains in the female phase. This suggests that variation in floral traits within inflorescences affects seed quality and quantity among flowers within inflorescences.     

Floral biology and pollination mechanisms of four Mexico-endemic Fuchsia species with contrasting reproductive systems

Aims The evolution of the outstanding variation of reproductive systems in angiosperms has been considered an important driver of lineage diversification. Closely related hermaphroditic and dioecious species with biotic pollination provide the opportunity to study and compare traits related to pollinator attraction and their consequences on reproductive components. A higher predictability of pollination syndromes is expected in dioecious species, which are dependent on pollinators, than in self-compatible hermaphroditic taxa. Dioecious species may suffer pollen limitation depending on the quality of floral rewards and the kind and abundance of pollinators, whereas no pollen limitation is expected in hermaphroditic species with autonomous self-pollination. Additionally, in the absence of pollen limitation, more or better seeds are expected in dioecious species, according to the sexual specialization hypothesis.Methods In natural populations of the hermaphroditic Fuchsia fulgens and Fuchsia arborescens and dioecious Fuchsia parviflora and Fuchsia obconica, all endemic to Mexico, we first described flower phenology, flower production and longevity and nectar volume and concentration. Then, we evaluated the correspondence between floral visitors and pollination syndromes. In hermaphrodite plants, we determined the level of herkogamy and the potential for autonomous self-pollination. Finally, we evaluated the effect of pollen limitation on fruit set and seed number and assessed seed germination for all species.Important findings In contrast to our prediction, dioecious species did not show a higher correspondence between pollination syndromes and floral visitors than did hermaphrodites; however, male flowers exhibited a higher correspondence than female flowers. No pollen limitation was detected in dioecious species, for which visitation rate did not differ between male and female flowers. The hermaphroditic F. fulgens showed pollen limitation for seed number, despite the presence of autonomous selfing. Fruit set from autonomous pollination was higher in F. arborescens, which showed a lower level of herkogamy compared with F. fulgens. Finally, dioecious species produced fewer but heavier seeds compared with hermaphrodite species. Although Fuchsia is classified as an outcrossing genus, both hermaphroditic species showed autonomous self-pollination. The heavier but lower number of seeds per fruit in dioecious species may be related to the more efficient resource allocation expected from sexual specialization. This could play an important role in the evolution of dioecy; however, a comparative phylogenetic approach is required to confirm this hypothesis.     

Bateman’s principle and plant reproduction: The role of pollen limitation in fruit and seed set

Bateman’s principle states that male fitness is usually limited by the number of matings achieved, while female fitness is usually limited by the resources available for reproduction. When applied to flowering plants this principle leads to the expectation that pollen limitation of fruit and seed set will be uncommon. However, if male searching for mates (including pollen dissemination via external agents) is not sufficiently successful, then the reproductive success of both sexes (or both sex functions in hermaphroditic plants) will be limited by number of matings rather than by resources, and Bateman’s principle cannot be expected to apply. Limitation of female success due to inadequate pollen receipt appears to be a common phenomenon in plants. Using published data on 258 species in which fecundity was reported for natural pollination and hand pollination with outcross pollen, I found significant pollen limitation at some times or in some sites in 159 of the 258 species (62%). When experiments were performed multiple times within a growing season, or in multiple sites or years, the statistical significance of pollen limitation commonly varied among times, sites or years, indicating that the pollination environment is not constant. There is some indication that, across species, supplemental pollen leads to increased fruit set more often than increased seed set within fruits, pointing to the importance of gamete packaging strategies in plant reproduction. Species that are highly self-incompatible obtain a greater benefit relative to natural pollination from artificial application of excess outcross pollen than do self-compatible species. This suggests that inadequate pollen receipt is a primary cause of low fecundity rates in perennial plants, which are often self-incompatible. Because flowering plants often allocate considerable resources to pollinator attraction, both export and receipt of pollen could be limited primarily by resource investment in floral advertisement and rewards. But whatever investment is made is attraction, pollinator behavioral stochasticity usually produces wide variation among flowers in reproductive success through both male and female functions. In such circumstances the optimal deployment of resources among megaspores, microspores, and pollinator attraction may often require more flowers or more ovules per flower than will usually be fertilized, in order to benefit from chance fluctuations that bring in large number of pollen grains. Maximizing seed set for the entire plant in a stochastic pollination environment might thus entail a packaging strategy for flower number or ovule number per flower that makes pollen limitation of fruit or seed set likely. Pollen availability may limit female success in individual flowers, entire plants (in a season or over a lifetime), or populations. The appropriate level must be distinguished depending on the nature of the question being addressed.     

Floral sex allocation and reproductive success within inflorescences of Hosta ventricosa,a pseudogamous apomict

Aims Within inflorescences of sexually reproducing hermaphrodites, the production of ovules, fruits and seeds commonly declines from basal (early-opening) to distal (late-opening) flowers, while pollen production remains constant or only changes slightly, with the result that distal (late-opening) flowers become functionally male. However, few empirical studies have specifically examined whether or not changes in allocation to pollen production actually lead to changes in the number of seeds sired, a more direct measure of male fitness. In pseudogamous apomicts, fitness depends on the number of seeds produced; thus, a contrasting pattern of variation in the pollen-to-ovule (P/O) ratio within inflorescences might be expected.Methods We investigated floral sex allocation and reproductive success within racemes of Hosta ventricosa, a pseudogamous apomictic hermaphrodite possessing flowers that open acropetally. We quantified variations in pollen number, ovule number, the P/O ratio and fruit and seed production, from 2007 to 2011, among flowers within racemes of four populations of H. ventricosa in southwest China. Ecological causes for fruit and seed production were evaluated by observing patterns of pollen deposition, flower removal and supplemental pollination.Important findings Pollen number, ovule number and the P/O ratio declined from basal-to-distal positions in all sampled populations (years). Fruit and seed production decreased distally in most populations (years). Low fruit and seed set of distal flowers was not due to pollen limitation because pollen deposition never declined distally and the low fruit and seed set of distal flowers remained even after supplementary pollination was provided. The flower-removal experiment indicated that inter-fruit competition for resources among flowers was common. The low P/O ratio of distal flowers in H. ventricosa might be favored because they were unable to obtain fitness by donating pollen and siring seeds on other plants. Our study may help to understand the adaptive significance of sex allocation among flowers within inflorescences of sexually reproducing hermaphrodites.     

Breeding systems in Clivia (Amaryllidaceae): late‐acting self‐incompatibility and its functional consequences

Late‐acting (ovarian) self‐incompatibility, characterized by minimal or zero seed production following self‐pollen tube growth to the ovules, is expected to show phylogenetic clustering, but can otherwise be difficult to distinguish from early‐acting inbreeding depression. In Amaryllidaceae, late‐acting self‐incompatibility has been proposed for Narcissus (Narcisseae) and Cyrtanthus (Cyrtantheae). Here, we investigate whether it occurs in the horticulturally important genus Clivia (Haemantheae) and test whether species in this genus experience ovule discounting in wild populations. Seed‐set results following controlled hand pollinations revealed that Clivia miniata and C. gardenii are largely self‐sterile. Self‐ and cross‐pollinated flowers of both species had similar proportions of pollen tubes entering the ovary, and those of C. gardenii also did not differ in the proportions of pollen tubes that penetrated ovules, thus ruling out classical gametophytic self‐incompatibility acting in the style, but not early inbreeding depression. Flowers that received equal mixtures of self‐ and cross‐pollen set fewer seeds than those that received cross‐pollen only, but it was unclear whether this effect was a result of ovule discounting or interactions on the stigma. The prevention of self‐pollination by the emasculation of either single flowers or whole inflorescences in wild populations did not affect seed set, suggesting that ovule discounting is not a major natural limitation on seed production. Flowers typically produce one to three large fleshy seeds from approximately 16 available ovules, even when supplementally hand pollinated, suggesting that fecundity is mostly resource limited. The results of this study suggest that Clivia spp. are largely self‐sterile as a result of either a late‐acting self‐incompatibility system or severe early inbreeding depression, but ovule discounting caused by self‐pollination is not a major constraint on fecundity. © 2014 The Linnean Society of London, Botanical Journal of the Linnean Society, 2014, 175 , 155–168.     

Gender variation in Actinidia deliciosa,the kiwifruit

Summary Flower and fruit characters were measured in ten female, five male and five fruiting male selections of A. deliciosa var deliciosa (A. Chev) Liang and Ferguson. Flowers from female vines had functional pistils, which contained many ovules. Stamens appeared to be fully developed but produced only empty pollen grains. Flowers from male vines had functional stamens that produced high percentages of pollen grains with stainable cytoplasmic contents. Pistils did not contain ovules and were generally small with vestigial styles. Fruiting male vines had both staminate and bisexual flowers. Staminate flowers were similar to those found on strictly male vines. Bisexual flowers produced ovules and stainable pollen. Pistils were smaller than in pistillate flowers. Although the three flower sexes differed in style length, ovary dimensions and ovules per carpel, staminate and bisexual flowers were similar in number of flowers per inflorescence, stamen filament length, pollen stainability, inflorescence rachis length and carpel number, and differed from pistillate flowers in these characters. The three flower sexes had similar sepal and petal numbers. The fruit of fruiting males were considerably smaller than those of females. Low ovule number appears to be the major factor limiting fruit size in the fruiting males studied. Prospects for developing hermaphroditic kiwifruit cultivars through breeding are discussed.