Search

Grasses exhibit a great variety of inflorescence forms and these appear homoplasious when mapped onto cladograms. The overall pattern is sufficiently complex that it is difficult to analyze inflorescence evolution. We have reduced the complexity of the problem by examining one group of grasses, the panicoid 'bristle clade,' which exhibits a less complex pattern of variation. The clade is morphologically defined by inflorescences bearing both spikelets and sterile bristles and is monophyletic in both morphological and molecular phylogenetic analyses. We have constructed a chloroplast DNA phylogeny of the three main genera, which finds three well-supported clades, two comprising species placed in Setaria and one of Pennisetum + Cenchrus. In this tree Cenchrus is monophyletic, but both Setaria and Pennisetum are paraphyletic. Developmental morphology of these groups is very similar at early stages. Changes in axis ramification, primordial differentiation, and axis elongation account for most variation in mature inflorescence morphology. Characters derived from comparisons of developmental sequences were optimized onto one of the most parsimonious trees. Most developmental characters were congruent with the molecular phylogeny except for three reversals in the subclade containing S. barbata, S. palmifolia, and two accessions of S. poiretiana. Changes in just a handful of developmental events account for inflorescence evolution in the bristle clade, and similar changes may account for inflorescence diversity in the grasses as a whole. Key words: Cenchrus; development; inflorescence; morphology; Pennisetum; phylogeny; Poaceae; Setaria.

Variation in inflorescence development patterns is a central factor in the evolutionary ecology of plants. The genetic architectures of 13 traits associated with inflorescence developmental timing, architecture, rosette morphology, and fitness were investigated in Arabidopsis thaliana, a model plant system. There is substantial naturally occurring genetic variation for inflorescence development traits, with broad sense heritabilities computed from 21 Arabidopsis ecotypes ranging from 0.134 to 0.772. Genetic correlations are significant for most (64/78) pairs of traits, suggesting either pleiotropy or tight linkage among loci. Quantitative trait locus (QTL) mapping indicates 47 and 63 QTL for inflorescence developmental traits in Ler x Col and Cvi x Ler recombinant inbred mapping populations, respectively. Several QTL associated with different developmental traits map to the same Arabidopsis chromosomal regions, in agreement with the strong genetic correlations observed. Epistasis among QTL was observed only in the Cvi x Ler population, and only between regions on chromosomes 1 and 5. Examination of the completed Arabidopsis genome sequence in three QTL regions revealed between 375 and 783 genes per region. Previously identified flowering time, inflorescence architecture, floral meristem identity, and hormone signaling genes represent some of the many candidate genes in these regions.

Maternal reproductive success was examined in Styrax obassia (Styracaceae), a bumble-bee pollinated mass-flowering tree in a cool-temperate deciduous forest in northern Japan. The effects of flower number on the success of individual flowers at three levels (inflorescence, individual, and population) were considered. During 1995 and 1996, variations in size, light availability to branches, floral display size, and fruit set were monitored in 37 out of 211 individual S. obassia trees in a 4-ha forest plot. In addition, the locations of the 211 trees in this plot were mapped and the number of inflorescences in each tree was counted. A multiple regression analysis showed that flower number per inflorescence and inflorescence number per individual had negative effects on fruit set, and inflorescence number of aggregated clumps of flowering trees, tree size, and light resource had positive effects on fruit set although significant level were marginal. It is concluded that pollinator attraction may occur not at the individual tree level, but at the level of a clump of flowering trees. It is also suggested that geitonogamy increased with inflorescence number of tree and inflorescence size and that resource limitation was related to the light condition and variation of tree size. Key words: bumble bee; display size; fruit set; geitonogamy; masting; size-structured population; spatial distribution; Styracaceae; Styrax obassia.

Tassel and ear primordia were collected from greenhouse-grown specimens of the Mexican maize landrace Chapalote and prepared for scanning electron microscopic (SEM) examination. Measurements of inflorescence apices and spikelet pair primordia (spp) were made from SEM micrographs. Correlation of inflorescence apex diameter with number of spikelet ranks showed no significant difference between tassels and ears, except at the two-rank level where the ear apical meristem had a significantly smaller diameter than corresponding two-ranked tassels. Within individual inflorescences, spp in different ranks enlarged at comparable rates, although the rates from one ear to the next along the stem differed. In both tassels and ears, spp divide to form paired sessile and pedicellate spikelet primordia when the spp is [approximately equal to]150 [[micro]meter] wide; ear axes are significantly thicker than tassel axes at the time of bifurcation. The similarities in growth between ear and tassel primordia lend further support to the hypothesis that both the maize tassel and ear are derived from a common inflorescence pattern, a pattern shared with teosinte. Inflorescence primordial growth also suggests that a key character difference between teosinte and maize, distichous vs. polystichous arrangement of spikelets, may be related to size of the apical dome and/or rate of primordium production by the apical meristem. There appears to be more than a single morphological event in the shift from vegetative to reproductive growth. The evocation of axillary buds (ears) is independent of, and temporally separated from, the transition to flowering at the primary shoot apex (tassel). Key words: development; evolution; flowering; inflorescence; maize; Poaceae; Zea.

Racemose inflorescences were investigated organographically and ontogenetically in three species of Campylotropis. An additional phyllome was found to be borne at the base of the pedicel in C. hirtella and C. giraldii but not in C. polyantha. Organographic observations revealed that the additional phyllome is the prophyll on the reduced lateral branch borne in the axil of the phyllome on the central inflorescence axis, and that the additional phyllome subtends the pedicel terminating in a flower. The flower presents its abaxial side to the additional phyllome in accordance with the organographic observations. Ontogenetic observations, moreover, revealed the presence of a rudimentary apex of the reduced branch in earlier stages of development. On the basis of such evidence, the central inflorescence axis of Campylotropis is interpreted as being composed of reduced lateral branches each of which usually bears only one flower in the additional phyllome axil. The inflorescence structure is kept in spite of the loss of the additional phyllome. In comparison with allied genera, the inflorescence of Campylotropis is regarded as a reduced pseudoraceme. An evolutionary trend in the inflorescence structure of the genus is inferred. Key words: Campylotropis; inflorescence; Leguminosae; phyllome.

Natural populations of Thymelaea hirsuta have previously been shown to comprise four distinct sexual morphs: males, females, protogynous individuals, i.e., first female then male, and protandrous individuals, i.e., first male then female. The objective of the present study has been to confirm the genetic basis of this sexual tetramorphism by quantifying morph ratios in the open-pollinated progeny of the four sexual phenotypes growing in a natural population. All four phenotypes were recovered in the progeny of each morph. All observed plants displayed a single sexual phenotype, thus confirming the genetic basis of the tetramorphism. The progeny sex ratios indicate that the genetic determination of sex in this species may be influenced by cytoplasmic factors, while the observed levels of functional female fertility suggest a near-dioecious system. The evolutionary significance of this tetramorphism as a transitional stage in the evolution of dioecy is discussed.

We characterized the development of a tobacco morphological mutant, puzzle box, previously obtained by selection for resistance to an inhibitor of an enzyme in the polyamine biosynthetic pathway. Puzzle box plants are shorter than wild type with smaller leaves and an irregular leaf plastochron. The diameter and shape of the puzzle box shoot apex are similar to wild type. In puzzle box plants, the terminal flower develops but not the cymose inflorescence. This flower has a variable phenotype with more than five sepals, five petals, and five stamens. The organs in the fourth whorl are produced by carpel-like primordia and contain tissue biochemically similar to wild type transmitting tissue. These organs form a cylinder within which additional floral organs are produced. Some cultures of puzzle box buds and excised cylinders produce additional floral organs. The inflorescence and floral programs may be expressed together in puzzle box.

A peculiar and often discussed feature in the genus Campanula is the presence and function of the pollen-collecting hairs (PCHs), covering the style and the dorsal surface of the stigmatic lobes on the pistil. The hairs sweep up the pollen as the anthers dehisce, and this 'stylar brush' serves as a secondary pollen presentation mechanism, facilitating the transfer of pollen from the flower to visiting insects. The adaptive significance of the hairs in relation to pollination has been variously interpreted by several authors during the past two centuries, but a satisfactory explanation for the floral mechanism is still lacking. The aim of the present investigation is to elucidate the function of the PCHs by experimental manipulations of the male phase. Comparisons are made 1) among species with various degrees of protandry and different mating systems, and 2) between annuals and perennials. The results clearly show that the hairs serve not only as pollen-collectors, but also as an important and efficient mechanism controlling stigma spreading and promoting cross-pollination. A tactile stimulation of the hairs causes a shortening of the length of the male phase and accelerates stigma maturation. The more frequently the hairs are treated, the shorter the male phase. Similar trends were found in all species regardless of differences in the degree of protandry, life span, or mating system. Geographically well-separated populations in two species showed differences in length of the male phase, depending on mating system and/or weather conditions. The relevance of the method used in this study compared to pollinator activity and its effect on the PCHs is discussed. The consequences of the retraction of the hairs with respect to self-pollination are briefly discussed.

The pollen-collecting hairs (PCHs) of Campanula have been a subject of intense debate for the past two centuries. Although several morphological studies have been made on these hairs, detailed comparative studies among species are still lacking, their function and adaptive significance being an unsolved question. The present study comprises two microscopy techniques: scanning electron microscopy and confocal scanning laser microscopy. The aim of the present study is to elucidate: 1) the variation in morphology of the PCHs, 2) the variation in presence/absence of the PCHs by the time of spreading of the stigmatic lobes, 3) the variation in the retractive mechanism of the PCHs, and 4) the correlation between pollination and the retraction of the PCHs. In several species PCHs of various lengths are found. Despite the variations in length of the hairs, the same retractive mechanism is found in all species studied. In most species the hairs retract into basal cavities within the style late in anthesis. The cells into which the hairs retract differ in length among species. Pollen grains are often found within the cavities together with the retracted hairs, a mechanism considered to prevent self-pollination. Pollen germination within the cavities was not observed. In a few species, the PCHs are still present at stigma receptivity. Differences in the shape and size of the cells surrounding the PCHs are documented. The diameter of the pits and the pollen grains vary among species. Other types of hairs on the style are recognized in some species, being of various lengths. These other types do not retract at stigma development and should not be regarded as pollen-collectors. They possibly facilitate for visiting insects to reach the nectar glands, present at the top of the ovary.

Flowers of the organ number (meristic) mutant clavata1-1 of Arabidopsis thaliana (Brassicaceae) were studied to examine timing and patterns of floral organogenesis as compared to the wild type. All clavata1-1 flowers examined had four- instead of two-loculed gynoecia; half showed increased numbers of stamens; and 10% formed increased numbers of sepals. An inflorescence plastochron index was used to establish the timing of developmental events during flower organogenesis. clavata1-1 flowers initiate faster but grow more slowly than in the wild type. The stages of sepal and stamen initiation were prolonged compared to those of the wild type. Although gynoecial initiation was not prolonged, the preceding stage was and it was characterized by a proliferation of meristematic cells above the initiating stamens. The clavata1-1 flower apex did not become wider than that of the wild type until after the establishment of the gynoecium. We propose that clavata1-1 is a heterochronic mutant, where flower organ number increases are due partly to prolongation of organ initiation stages.

Green dragon (Arisaema dracontium; Araceae) is a perennial woodland herb capable of switching gender from year to year. Small flowering plants produce only male flowers but when larger they produce male and female flowers simultaneously. Distinct male and monoecious phenotypes (referred to hereafter as plants) share a single underlying cosexual genotype. Four populations in southern Louisiana were sampled to determine frequencies and size distributions of male and monoecious plants, and to determine the relationship of plant size with male and female flower production in monoecious plants. Male plants were significantly smaller than monoecious plants and made up 34%-78% of flowering plants within populations. Flower number (average = 120) was weakly positively correlated with size. Monoecious plants produced an average of 169 flowers (90 female) and had 100% fruit set, with individual berries containing an average of 2.5 ovules and 1.3 filled seeds. Male flower number was negatively correlated, and female flower number positively correlated, with basal stem diameter. Extrapolation of regression slopes suggested that green dragon should become completely female at a size 20% larger than the largest plant observed in this study. A simple model of inflorescence development is presented to illustrate how the reproductive system of green dragon is related to that of jack-in-the-pulpit (A. triphyllum), which exhibits a more distinct switch between male and female phenotypes.

Using both multivariate and univariate regression techniques, I measured selection acting through female reproductive success in two hermaphroditic species with precise pollen placement but different pollinators: hummingbird-pollinated Lobelia cardinalis and bumblebee-pollinated L. siphilitica. Six traits were analyzed in two populations of L. cardinalis and one population of L. siphilitica: flower number, mean number of flowers open per day, inflorescence height, number of days in flower, median-flower date and nectary-stigma distance. In another study it was found that female reproductive success in one population of L. cardinalis was much less pollen limited than in the other two populations, and it was therefore expected that selection of female reproductive traits in thsi population would be weaker. In the univariate analyses correlations caused nearly all traits to have significant directional selection coefficients. However, in the multivariate analyses no traits in L. siphilitica experienced directional or quadratic selection. Selection acted differently in the two L. cardinalis populations. The less pollen-limited population experienced positive directional selection on flower number and median-flower date, while in the other L. cardinalis population there was positive directional selection on flower number and nectar-stigma distance and both positive directional and positive quadratic selection on height. The functional significance of floral traits in these two species and the probable effect of increased sample sized are discussed.

Inflorescence and floral ontogeny are described in the mimosoid Acacia baileyana F. Muell., using scanning electron microscopy and light microscopy. The panicle includes first-order and second-order inflorescences. The first-order inflorescence meristem produces first-order bracts in acropetal order; these bracts each subtend a second-order inflorescence meristem, commonly called a head. Each second-order inflorescence meristem initiates an acropetally sequential series of second-order bracts. After all bracts are formed, their subtended floral meristems are initiated synchronously. The sepals and petals of the radially symmetrical flowers are arranged in alternating pentamerous whorls. There are 30-40 stamens and a unicarpellate gynoecium. In most flowers, the sepals are initiated helically, with the first-formed sepal varying in position. Petal primordia are initiated simultaneously, alternate to the sepals. Three to five individual stamen primordia are initiated in each of five alternipetalous sectorial clusters. Additional stamen primordia are initiated between adjacent clusters, followed by other stamens initiated basipetally as well as centripetally. The apical configuration shifts from a tunica-corpus cellular arrangement before organogenesis to a mantle-core arrangement at sepal initiation. All floral organs are initiated by periclinal divisions of the subsurface mantle cells. The receptacle expands radially by numerous anticlinal divisions in the mantle at the summit, concurrently with proliferation of stamen primordia. The carpel primordium develops in terminal position by conversion of the floral apex.

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

The order of floral initiation and subsequent organogeny of Erigeron philadelphicus L. (Asteraceae: Astereae) was found to deviate from the acropetal pattern generally reported for the Asteraceae. Light micrographs show periclinal divisions in the first, second, and deeper subsurface layers of cells on the flanks of the inflorescence apex as the earliest evidence of floral initiation. Scanning electron microscope micrographs indicate that the disk flowers appear first and arise as small protuberances approximately one-third of the way up the previously and undifferentiated highly convex inflorescence apex. A succession of disk flowers arises acropetally in a complex anthotaxy characterized by about 21 dextrorse and 12-15 sinistrorse parastichies (although this pattern is obscured at the apex). After one to three disk flowers have been initiated in each parastichy, the first ray flower initials can be seen to initiate in sites proximal to the oldest and largest disk flowers. Additional ray flowers then initiate basipetally following the dextrorse parastichies established by the disk flowers. Overall floral initiation on the inflorescence apex proceeds acropetally for the disk flowers and basipetally for the ray flowers until the available space is filled. Floral development adheres to the same plan -- proceeding bidirectionally on the inflorescence meristem with the oldest and most complete flowers of both types located on the equator established at initiation.

Fruit distribution among individuals in populations of three orchid species was analyzed. Fruit set was low in the three species. The distribution of fruits in the non-autogamous species was skewed to the right, with high variance: mean ratios, and a moderate to high proportion of fruitless plants. Larger inflorescene size did not necessary result in an increase in percent fruit set but tended to increase the probability to produce at least one fruit. Pollinator-mediated selection on inflorescene size through fruit production may be limited by the low overall level of visitation and the resulting uncertainty of pollination at the individual level.

Inflorescence ontogeny and morphology of the det mutant of Pisum sativum L. were investigated using scanning electron microscopy. This mutation causes the production of a limited number of axillary flowers followed by the formation of an apparent terminal flower slightly offset from the vertical. Our study indicates that the apparent terminal flower arises from an axillary meristem. The terminal meristem senesces and differentiates hairs, forming a rudimentary stub in the same manner as axillary meristems of conventional (Det) and det plants. Thus the dramatic effect of the det gene on inflorescence architecture results from early apical arrest rather than conversion of the terminal meristem to a flower as implied by the symbol det. Thus mutant will be valuable in elucidating regulation of apical arrest.

Floral development in Florex and Ottawa cultivars of red clover (Trifolium pratense L.: Leguminosae) was examined by scanning electron microscopy. No differences between the two cultivars were found. The terminal inflorescence is initiated in the axil of the penultimate bract before the final bract is initiated. After initiation of the final bract, the remnant apical dome is transformed to become the least mature part of the inflorescence dome. Subsequent inflorescences are initiated laterally in basipetal sequence. Inflorescence development is zygomorphic. This leads to an unusual pattern of floret initiation, the oldest florets resting basally and proximal to the penultimate bract. Florets develop with zygomorphic symmetry, each whorl of floral organs developing unidirectionally from the abaxial side. Initiation of the adaxial organ of each whorl is delayed until the abaxial organ of the succeeding whorl has been initiated. Thus there is overlapping development of the whorls of organs. The antepetalous stamens arise in close association with their respective petal primordia. As development proceeds, the corolla tube and the staminal tube exhibit basal zonal growth. In the mature flower, above the distal zone of fusion of the keel petals, marginal cells project and interlock, producing a pollination mechanism that can be sprung by the pollinator.

Fossil inflorescences and flowers of Mauldinia mirabilis gen. et sp. nov. are described from the early Cenomanian Elk Neck beds of northeastern Maryland, U.S.A., and are assigned to the family Lauraceae. Specimens are exceptionally well preserved and provide the earliest evidence of trimerous floral organization and endosperm in angiosperms. Inflorescences are compound and consist of elongated axes bearing distinctive, spirally arranged laterl, bilobed cladode-like units. Each lateral unit typically bears five sessile bisexual flowers on the adaxial surface. Flowers have a perianth of three small outer and three larger inner tepals, and an androecium of nine fertile stamens in three whorls. In addition, there is an inner fourth whorl of three dorsiventrally flattened staminode-like structures, and each of the three fertile inner stamens has an associated pair of staminode-like appendages with clavate-sagittate heads. Anthers dehisce by two valves that are hinged distally. The gynoecium consists of a superior, unilocular carpel with a single anatropous, pended ovule. Floral organization in M. mirabilis is the same as that of many extant Lauraceae, and the unique inflorescence structure can also be interpreted in terms of inflorescence patterns in extant taxa. Mauldinia mirabilis is the first unequivocal documentation of Lauraceae from the Cretaceous and provides further evidence for considerable diversity among magnoliid dicotyledons at an early phase in angiosperm diversification.

The enlarged inflorescence (curd) of cauliflower and broccoli provide not only a popular vegetable for human consumption, but also a unique opportunity for scientists who seek to understand the genetic basis of plant growth and development. By the comparison of quantitative trait loci (QTL) maps constructed from three different [F.sub.2] populations, we identified a total of 86 QTL that control eight curd-related traits in Brassica oleracea. The 86 QTL may reflect allelic variation in as few as 67 different genetic loci and 54 ancestral genes. Although the locations of QTL affecting a trait occasionally corresponded between different populations or between different homeologous Brassica chromosomes, our data supported other molecular and morphological data in suggesting that the Brassica genus is rapidly evolving. Comparative data enabled us to identify a number of candidate genes from Arabidopsis that warrant further investigation to determine if some of them might account for Brassica QTL. The Arabidopsis/Brassica system is an important example of both the challenges and opportunities associated with extrapolation of genomic information from facile models to large-genome taxa including major crops.

The morphology of the banana inflorescence has been described as a long terminal raceme, a terminal panicle, and a serial flower assemblage. Mclean Thompson proposed that the banana inflorescence was an aerial stem with spirally arranged leaves in the axils of which grow cones of flowers. The observations of Abraham Fahn contest this hypothesis and favor another hypothesis also proposed by Thompson in which the flower-groups are cincinni. Our research has examined inflorescence development in Musa velutina. Musa is monecious, containing unisexual flowers arranged in hands. The number of flowers per hand varies. Each hand develops in the axil of a primary bract and consists of two rows of flowers. The entire inflorescence is made up of a definite number of female hands, a small number of transitional hands, and an indefinite number of male hands. Based on his observations of Musa acuminata and Musa balbisiana, Fahn suggests that the sequence of flower initiation begins with the primordium farthest right in the adaxial row. This is followed by the development of the primordium directly adjacent to the first. The third is the fight most primordium in the abaxial row. He suggests the sequence continues to alternate, adaxial to abaxial, throughout the hand. Our research does not wholly agree with these findings. Many of the hands we studied demonstrate the sequence of flower initiation beginning from the center of the hand and progressing outward. In many instances the right-most flower appears to be younger than the flower adjacent to it. While we have observed the adaxial-abaxial arrangement of flowers we have not observed development of flower primordia following an alternating sequence between the rows.

The development of the inflorescence in Setaria italica (foxtail millet), S. verticillata, S. viridis, and S. grisebachii was examined. The inflorescences of the different species differed in numbers of orders of branching, with S. italica having 2 or more orders as compared to one order in the other three species. All species initiate spikelets paired with bristles, but many of the spikelets are suppressed during development. In S. grisebachii pairs of spikelets and bristles appear together on the branches, in S. verticillata and S. viridis a single spikelet at maturity is surrounded by a series of bristles, and in S. italica three or four spikelets and at least twice as many bristles appear on the primary branches. Differences between the species appear early in development and have the potential, along with density of branches in the spike, to be major determinants of yield. The different mature morphologies of the infiorescences are similar to various mutant phenotypes of maize, such as idsl, ts6, and ral & 2, which raises the prospect of using genes identified from maize to identify differences in inflorescence morphology in a comparative manner across species groups. The present study is a step towards relating advances in model plant genetics to an understanding of natural morphological differentiation in plants.

Mature staminate flowers of giant ragweed (Ambrosia trifida L.; Asteracea, tribe Heliantheae) are about 2 mm long. Each of the five fused petals is traversed for most of its length by a prominent polyacetylene-filled reservoir (PAR). Par's have a biseriate epithelium, and except for their short length of about 1.5 mm, seem identical to the much longer PAR's of leaf and stem. Study of sections of resin-embedded flowers at different development stages showed that corolla PAR's appear with a full complement of epithelial cells, and occupy their full relative length, in flowers only 150 [micrometer] in length. As the flower enlarges, the original epithelial cells also enlarge and elongate, and secrete into the lumen, which opens simultaneously for its entire length when the flower is nearly 300 [micrometer] long. PAR's in mature flowers have only their original epithelial cells, which remain alive. Corolla PAR's seem anatomically to be representative of the canals and ducts that have been described elsewhere in Asteraceae.

The development of the inflorescence and flowers are described for Gymnotheca chinensis Decaisne (Saururaceae), which is native only to southeast China. The inflorescence is a short terminal spike of about 50-70 flowers, each subtended by a small bract. There are no showy involucral bracts. The bracts are initiated before the flowers, in acropetal order. Flowers tend to be initiated in whorls of three which alternate with the previous whorl members. No perianth is present. The flower contains six stamens, and four carpels fused in an inferior ovary containing 40-60 ovules on four parietal placentae. Floral symmetry is dorsiventral from inception and throughout organ initiation. Flora organs are initiated in the following order: 1) median adaxial stamen, 2) a pair of lateral common primordia which bifurcate radially to produce two stamen primordia each, 3) median abaxial stamen, 4) a pair of lateral carpel primordia, 5) median adaxial carpel, 6) median abaxial carpel. This order of initiation differs from that of any other Saururaceae previously investigated. The inferior ovary results from intercalary growth below the level of stamen attachment; the style elongates by intercalary growth, and the four stigmas remain free. The floral structure of Gymnotheca is relatively advanced compared to Saururus, but its assemblage of specializations differs from that of either Anemopsis or Houttuynia, the other derived genera in the Saururaceae.

Rhododendron nuttallii T. W. Booth (Ericaceae) was used to derive concurrent timetables for megagametophyte, pollen tube and early postfertilization development from anthesis through 3 wk after pollination, based on timed collections of self-pollinated pistils. Stages of development were determined for over 33,500 cleared ovules, including, for selected collection dates, stages on different portions of the placenta. Pollen tube information was obtained by fluorescence microscopy of pistil squashes stained with aniline blue. Because of the very large number of ovules observed, it was possible to recognize a much more closely graded series of stages in megagametophyte development than is usually the case. While a range of stages occurred on all days, development progressed steadily from a majority of functional spores and 2-nucleate gametophytes on the day of anthesis to mostly a late zygote-primary endosperm stage at 18 days, and some 2-celled endosperm stages at 21 days, after pollination. At all times the most advanced stages, including first pollen tube entries, occurred on the outer surface of the lower half of the placenta, and the youngest on the inner surface of the uppermost portion. Fertilizable ovules were not found in any frequency until 8 days after pollination (then in only about 34% of the ovules); a few fertilized ones were seen after 10 days but constituted less than 5% until 12 days after pollination, thereafter increasing to about 60%. Fertilization occurred in any one of three morphologically recognizable stages distinguished by position and state of fusion of polar nuclei. Pollen germinated on the stigma 1-2 hr after pollination, and pollen tubes grew at a rate of about 1-1.25 cm/day, reaching the top of the ovary in 8-9 days with the first ovule entries seen after 10 days. There was a close correlation between megagametophyte development and pollen tube growth, with large numbers of functionally mature ovules not being found until pollen tubes had reached the ovary. While nuclei within ovules could not be distinguished in the squashes, three gametophyte stages that could be recognized -- unelongated, elongated either without or with a pollen tube -- were tallied for almost 29,000 ovules. The progression in these general stages corresponded well with that documented in more detail from cleared ovules. Unpollinated pistils showed a similar progression of gametophyte stages until the time fertilization would start to occur, after which there was continued accumulation of functionally mature ovules. A variety of abnormally developed and/or collapsing(ed) ovules or gametophytes were seen; collectively, they averaged over 8.6% of all ovules.

Platanthera stricta Lindley is entomophilous and can produce seed via facultative self-pollination and intraracemic and interracemic pollination. Capsule production is pollinator-limited and seed set may be pollen-limited. In experimental plants capsules produced via self- and intraracemic pollination contained fewer seeds with normally developed embryos than did capsules produced via interracemic pollination. The inflorescence of Platanthera stricta is fragrant and is attractive to a wide array of anthophilous insects. It is pollinated by a diverse assemblage of short-tongued insects. The primary pollinators are Eustroma fasciata B. and McD. (Lepidoptera: Geometridae), Bombus flavifrons Cresson and B. melanopygus Nylander (Hymenoptera: Apidae), an undescribed species of Greya (Lepidoptera: Prodixidae), and several species of Empis, Rhamphomyia, and Anthepiscopus longipalpis Melander (Diptera: Empididae). Small amounts of glucose are present on the raceme. The extrafloral glucose may retain small polinators on the inflorescence until they locate the floral spur aperture.

The inflorescence of Fagus is generally considered to be a determinate one, i.e., an axillary dichasium, in contrast to those of most genera in the family, which are indeterminate, dichasial, or simple catkins. To understand the relationship between the two types, ontogenetic development of the inflorescence of Fagus crenata was investigated. The early developmental stages are similar in both the male and the female inflorescences. At first, the inflorescence is oval-shaped, then a swelling forms at the distal side of it. Subsequently, another swelling forms at the proximal side. The more or less conspicuous residual part of the primary inflorescence axis remains between the two swellings. The inflorescence becomes heart-shaped and the first flower forms at the summit of each swelling. Subsequently, higher-ordered flowers form dichasially in the male inflorescence, and the cupule valves differentiate in the female one. This organogenetic manner suggests that the inflorescence of Fagus is an indeterminate one, consisting of two dichasia arranged alternately on the primary axis. The scale leaves surrounding the inflorescence were also given a new interpretation. They were considered to be stipules of the bracts, because sometimes they constitute a continuous structure, together with an inconspicuous swelling between them. A proliferous-type monstrous cupule was interpreted as supporting evidence for the hypothesis.

One explanation for low fruit sets in plants with hermaphroditic flowers is that total flower production by a plant is controlled primarily by selection through male function. This male function hypothesis presupposes that success in pollen donation increases more strongly with flower number than does seed set. I tested this prediction by measuring male and female components of reproductive success as functions of flower number in natural populations of the self-incompatible, perfect flowered plant, Ipomopsis aggregata. Fruit set in this hummingbird-pollinated plant averaged 4.9 to 40.3% across the 4 years of study. Both the total amount of pollen donated and the total amount received, as estimated by movement of fluorescent powdered dyes, increased linearly with number of flowers on a plant. Total seed production, however, increased disproportionately quickly because plants with larger floral displays were more likely to set at least one fruit. An estimate of the functional femaleness of a plant, based on pollen donation and seed production, increased with flower number. These results do not support the male function hypothesis.

Sex allocation models predict that cosexuality is stabilized by high allocation to attractive structures in pollen-limited species or by high allocation to shared structures that contribute to both genders. High investment in unilateral fixed costs favor the evolution of dioecy or gender change. With these predictions in mind, I studied sexual mass allocation at flowering in the monoecious Arum italicum (Araceae) and compared it with information available for its sex labile relative Arisaema dracontium. In A. italicum, 68% of biomass was allocated to structures believed to be involved in pollinator attraction and capture. This allocation pattern contrasts with that of Arisaema dracontium, in which 70% of biomass was allocated to scape, considered to be a unilateral fixed cost. The importance of attractive structures in A. italicum was further supported by a disproportionate increased allocation, in larger inflorescences, to the appendix (an attractive structure) compared to fertile flowers. In addition, an increase in inflorescence mass involved a disproportionate increase in mass allocation to male, rather than female, flowers. This pattern also contrasts with a size-related gender change from male to female in Arisaema species. These findings were consistent with sex allocation model predictions and shed light on the evolution of sex lability in Arisaema species. Key words: accessory structures; allometry; Araceae; Arisaema; Arum italicum; sexual allocation; sexual lability.

Mourera fluviatilis from northern South America is a spectacular member of the Podostemaceae (river-weeds). Its raceme-like inflorescences are up to 64 cm long and have 40-90 flowers arranged in two opposite rows. Inflorescence development starts with the initiation of a double-sheathed (dithecous) bract in a terminal position. All lateral bracts (again dithecous) are initiated in basipetal order along the two flanks of the inflorescence. Each gap between two neighboring bracts contains a single flower. The flowers are bisexual, each with a whorl of 16-20 ligulate tepals and 14-40 stamens, which are arranged in one or two whorls. Floral development starts with the formation of a girdling primordium rim around a two-lobed primordial gynoecium. Stamen and tepal initiation is centrifugal on the girdling primordium. The anthers are introrse or extrorse, depending on stamen position. Seedlings develop two entire, threadlike cotyledons, followed by forked filamentous leaves, which arise from the plumular pole. The radicular pole of the hypocotyl develops into a claw-shaped holdfast that fixes the young plant to the rock. The developmental morphologies of Mourera fluviatilis and other members of the Mourera group (including Lonchostephus and Tulasneantha) fit well with the Podostemoideae bauplan known from other New World genera, such as Apinagia and Marathrum. Key words: basipetal bract inception; inflorescence; morphogenesis; Mourera fluviatilis; Podostemaceae; polystemonous flowers; seedling morphology; terminal leaf; water plants.

Comparative developmental studies of inflorescences and flowers in Proteaceae were examined within a phylogenetic and functional context to better understand the modifications and the evolutionary diversification of inflorescence architectures. Taxa among Grevilleoideae ([approximately]38 genera) have a unique feature of two flowers per common bract, and the plesiomorphic condition appears to have been a simple raceme. From this condition, there has been diversification of inflorescence types including: floral number proliferations and reductions (from 10,000+ to one), compound branching, spikes, umbels, capitula and spadices. The diverse inflorescence morphologies are due to minor changes in the timing of different processes within metamers, principally among the internodes. These differences produce spatial changes among the flower-pairs on an axis. In certain lineages, the changes in spatial relationships are associated with modifications in floral development, particularly the processes associated with zygomorphy, and create a functional association among flowers and novel inflorescence morphologies. Changes in floral zygomorphy appear as successive elaborations on a homoplasious inflorescence deviation of whorled phyllotaxis, among taxa in Gevuininae and Banksieae. In Lambertiinae and Stenocarpinae, changes in floral morphology appear to phylogenetically preempt successive changes in inflorescence morphology, there being a direct correspondence between increased floral zygomorphy and reduction of flower number to one per inflorescence axis. Minor developmental changes in components of inflorescences can produce complex morphologies, and interpretations of morphological diversification of flowers should take into account all aspects of inflorescence development.

Within the tribes Ingeae and Acacieae (Fabaceae, Mimosoideae) there is a diverse array of inflorescence-morphologies. Nonetheless, all these types are variations on a common theme: this theme is a series of metamers each possessing one to several axillary unit-inflorescences. The unit-inflorescence is always a pedunculate raceme. There are few structural modifications of the fertile metamers that contribute to the diversity of forms: the number and arrangement of axillary buds are two examples. Rather, most of the diversity in form can be explained by various heterochronies affecting the leaves and axillary bud-systems of the reproductive metamers. These heterochronies are illustrated, and it is shown that some are synapomorphic of higher-level monophyletic groups.

Previous studies of reproductive structures in the helobial monocotyledons (Alismatidae) indicate that partitioning between flower and inflorescence is not always clear (i.e. Lilaea, Scheuchzeria) and this may be the result of ancestral, unisexual modules coming together to form flowers and/or inflorescences. Later evolutionary changes may have included the inflorescence becoming involved or mixed in with vegetative growth. Substitution of vegetative buds for flowers is the simplest version, and there can be additional modifications to the growth behaviour of the inflorescence such as horizontal growth and dorsiventrality. In the Alismataceae and Limnocharitaceae the derivation of stolon-like structures from inflorescences is obvious: vegetative features have been incorporated into structures which are recognizably inflorescences. In the Hydrocharitaceae the inter-relationships between the inflorescence and the vegetative body are much less well-defined. We have previously suggested for Hydrocharis, where a single axillary complex can contain both inflorescence and stolons, that the stolon is basically a sterile inflorescence, and features of the inflorescence have become incorporated into the vegetative body. Here we will explore this theme further for the Hydrocharitaceae, using information from within and outside the family.