Understanding plant reproductive diversity

Flowering plants display spectacular floral diversity and a bewildering array of reproductive adaptations that promote mating, particularly outbreeding. A striking feature of this diversity is that related species often differ in pollination and mating systems, and intraspecific variation in sexual traits is not unusual, especially among herbaceous plants. This variation provides opportunities for evolutionary biologists to link micro-evolutionary processes to the macro-evolutionary patterns that are evident within lineages. Here, I provide some personal reflections on recent progress in our understanding of the ecology and evolution of plant reproductive diversity. I begin with a brief historical sketch of the major developments in this field and then focus on three of the most significant evolutionary transitions in the reproductive biology of flowering plants: the pathway from outcrossing to predominant self-fertilization, the origin of separate sexes (females and males) from hermaphroditism and the shift from animal pollination to wind pollination. For each evolutionary transition, I consider what we have discovered and some of the problems that still remain unsolved. I conclude by discussing how new approaches might influence future research in plant reproductive biology.     

Pollination by fungus gnats in four species of the genus Mitella (Saxifragaceae)

The first example of pollination by fungus gnats in the eudicots is reported. The genus Mitella (Saxifragales) is characteristically produces minute, inconspicuous, mostly dull-coloured flowers with linear, sometimes pinnately branched, petals. To understand the function of these characteristic flowers, we studied the pollination biology of four Mitella species with different floral traits and different sexual expression: dioecious M. acerina , gynodioecious M. furusei var. subramosa , and hermaphroditic M. stylosa var. makinoi and M. integripetala. Flower-bagging experiments showed that wind pollination did not occur in the dioecious and gynodioecious species. Two years of observations of flower visitors at six study sites in Japan revealed that the principal pollinators of all four Mitella were specific species of fungus gnats (Mycetophilidae), which landed on the flowers with their long spiny legs settling on the petals. Characteristically, numerous pollen grains were attached to the fungus gnats in specific locations on the body. Although, on average, 1.3–2.6 fungus gnats visited each inflorescence per day, the fruit set of both bisexual and female flowers exceeded 63%. These results suggest that fungus gnats are highly efficient pollinators of Mitella spp., and that Mitella flowers are morphologically adapted to pollination by fungus gnats.  © 2004 The Linnean Society of London, Botanical Journal of the Linnean Society , 2004, 144 , 449–460.     

Floral scent in dioeciousSalix (Salicaceae)—a cue determining the pollination system?

Floral scents of male and female inflorescences of three dioeciousSalix species were collected by head-space adsorption, and analysed by GC-MS. InSalix caprea andS. cinerea 1,4-dimethoxy benzene was the main compound, and male and female scents showed a high degree of resemblance. No dominant compound was found inS. repens and malefemale scent similarity was low. Floral scent inSalix is likely a strong orientation cue, guiding pollinators between male and female plants ensuring pollen transfer and pollination. We suggest that a high degree of male-female floral scent resemblance is coupled to a high degree of insect pollination. Floral scent does not promote reproductive isolation betweenS. caprea andS. cinerea.     

Experimental documentation of natural hybridization inCactaceae: Origin of Lloyd's Hedgehog Cactus,Echinocereus ×lloydii

The origin ofEchinocereus ×lloydii Britt. & Rose, pro sp. (Lloyd's Hedgehog Cactus) was investigated using comparative morphology, cytology, biochemistry, and particularly, artificial hybridization. Numerous artificial crosses between the putative parentsE. coccineus Engelm. (a species of claret-up cactus) andE. dasyacanthus Engelm. (Texas Rainbow Cactus) were successful, resulting in the production of hundreds of seeds with hybrid embryos. The F₁ hybrid progeny (i.e., syntheticE. ×lloydii) grew to sexual maturity in about four and one-half years, whereupon successful backcrosses and F₂ generation hybrids were also obtained. The known F₁ hybrids closely approximated naturalE. ×lloydii. The fertility of these syntheticE. ×lloydii was high, like their natural counterparts. The populations ofE. ×lloydii in Pecos County, Texas are inferred to have originated as the result of natural interspecific hybridization. It is assumed thatE. ×lloydii or similar plants may arise wherever the parental taxa grow sympatrically.     

ALLOZYME DIVERSITY AND GENETIC IDENTITY IN SCHIEDEA AND ALSINIDENDRON (CARYOPHYLLACEAE: ALSINOIDEAE) IN THE HAWAIIAN ISLANDS

Genetic diversity of allozymes, genetic identity based on allozyme variability, and phylogenetic relationships were studied with respect to breeding system diversity, population size, and island age in 20 of the 29 species of Schiedea and Alsinidendron (Caryophyllaceae: Alsinoideae), a monophyletic lineage endemic to the Hawaiian Islands. Average levels of genetic variability in Schiedea and Alsinidendron were comparable to or higher than those found in other Hawaiian lineages for which equivalent data are available [Bidens, Tetramolopium, and the silversword alliance (Asteraceae: Madiinae)] and similar to average values for species of dicots. Allozyme variability was strongly dependent on breeding system, which varies widely in the Hawaiian Alsinoideae. Species with autogamous breeding systems showed very low variability, measured as the number of alleles per locus, percent polymorphic loci, and mean heterozygosity per locus. Outcrossing hermaphroditic and dimorphic species (those with gynodioecious, subdioecious, and dioecious breeding systems) showed significantly higher genetic variability. Small population size was associated with lower values for all measures of genetic variability. Nearly half of the species occurring in small populations are also autogamous; thus, both factors may have influenced levels of genetic variability in these species. Founder effect was apparent in one species (Schiedea adamantis), which occurs in a single large population, has a gynodioecious breeding system but a very low genetic variability. Island age appeared to have little effect on genetic variability. Slightly lower values of genetic variability for species occurring on Kaua'i and O'ahu result primarily from the occurrence of autogamous Alsinidendron species on those islands. Values for Nei's genetic identity for different species pairs were 0.201–0.942, a far greater range than in Bidens, the silversword alliance, and Tetramolopium. Using UPGMA clustering, there was only moderate support for relationships detected through cladistic analysis. Nei's unbiased genetic identity (I) was greatest among species with outcrossing breeding systems, which for the most part clustered together. Nei's genetic identities for self-fertilizing species were low, indicating that these species are less similar to one another and to outcrossing species, regardless of their affinities based on cladistic analysis. Parsimony analysis of allele frequency data supported two clades also found in phylogenetic analyses using morphological and molecular data. Clades recognized in parsimony analysis of allele frequencies were those lineages containing selfing species, indicating that conditions favoring fixation of alleles occurred in ancestral species. In contrast, maintenance of high genetic diversity in outcrossing species interferes with recognition of phylogenetic relationships using allozyme variability.     

The mating system ofElatostema (Urticaceae) in relation to morphology: A comparative study

To investigate why some species are dioecious and others monoecious, a data set of 14 morphological characters in 106 species ofElatostema was regressed onto mating system using partial multiple regression (controlling for subgeneric classification to avoid spurious associations). Morphology and mating system are significantly related (p = 0.05), and dioecious species tend to have large stipules, woody habit, and large leaves. The correlation of dioecy with large stipules is reported for the first time, but this character, like woodiness, is an indicator of plant size, suggesting a functional relationship of large plants (of the under-canopy) with dioecy.     

GENETIC EVIDENCE FOR MULTIPLE ORIGINS OF DIOECY IN THE HAWAIIAN SHRUB WIKSTROEMIA (THYMELAEACEAE)

Dioecy is unusually common in the Hawaiian Islands, yet little is known about the evolutionary biology of this breeding system. A native shrub, Wikstroemia, has an unusually diverse array of breeding systems: two forms of dioecy, cryptic and morphological dioecy, as well as hermaphroditism (perfect flowers). The existence of two forms of dioecy is significant for three reasons: 1) the presence of cryptic unisexuals that are functionally unisexual, but retain the appearance of hermaphroditism in both sexes, is strong evidence for the ancestral status of hermaphroditism; 2) the production of nonfunctional pollen, by female cryptic unisexuals, is a new instance of a phenomenon which has previously been reported for a few other species; 3) the two forms of dioecy are morphological markers which are useful in hybridization studies for tracing the genetic basis of their inheritance. Crosses were made between cryptically unisexual individuals (C), between morphologically unisexual individuals (M), and between the two types of unisexuality. The offspring of crosses between individuals with the same sex type usually resulted in offspring with that sex type, but most of the progeny of between-sex type crosses were, unexpectedly, perfect-flowered hermaphrodites. These results show that genetic control of sex determination is not homologous in all populations, suggesting that dioecy has evolved at least twice in Hawaiian Wikstroemia. The genetic data further suggest that males are the heterozygous sex.     

THE RELATIONSHIP BETWEEN A FLOWER GALL MIDGE PSEUDAS-PHONDYLIA SP. (DIPTERA: CECIDOMYIIDAE) AND ITS HOST PLANT ACTINIDIA VALVATA*

Abstract Kiwifruit plants, Actinidia sp., are native to subtropical China. The flower-bud gall of A. valvata, which is induced by an undescribed gall midge in the genus Pseud as phond ylia, is valued by the pharmaceutical industry. When studying the biology of the Actinid ia/Pseud as phond ylia interaction in Central-south China we found evidence suggesting that under certain circumstances the gall insect modifies the reproductive mode of the dioecious host plant. Surveys and field experiments in the National Hupingshan Natural Reserve showed a high frequency of galled trees. The density of galled trees varied among valleys and among trees within the valleys. In two valleys, 92% and 75%, respectively, of all trees were attacked, while in a third valley no trees were attacked. When infested, staminate tree only produced galls, whereas pistillate plants produced normal fruits as well as galls. Gall shape differed between male and female trees. Trees with galls tended to produce more fruits than treea without galls. We speculate that this is one of a few documented examples of an insect that induces androdioecy in an otherwise functionally dioecious plant.     

Stamens are not essential as an attractant for pollinators in females of cryptically dioecious Thalictrum pubescens Pursch. (Ranunculaceae)

 It has been hypothesized that females of some dioecious species maintain stamens that produce sterile pollen as a means of attracting pollinators to promote greater seed set. However, this hypothesis has rarely been tested. This paper examines the role of pollinators in the maintenance of sterile stamens in the cryptically dioecious species, Thalictrum pubescens. Wind was found to contribute to pollination; branches of female T. pubescens enclosed in cheesecloth to exclude insects but not wind were still able to set seed. Therefore, females may not need the stamens for pollination. In 1994, insects were found to discriminate against emasculated female flowers, but this discrimination did not lead to a significant difference in either the amount of pollen received or seed set. In 1995, emasculation was combined with the addition of pollen to determine if emasculated females were pollen limited. No significant effect of pollen addition or emasculation was found. Emasculated branches in the pollen-addition treatment appeared to have slightly lower seed set than non-emasculated branches that received pollen, indicating that the slight reduction in seed set was caused by emasculation rather than pollen limitation. Since stamens do not appear to be maintained as pollinator attractants, other explanations, such as genetic constraints, must be investigated. Received: 4 February 1997 / Revision accepted: 20 May 1997