HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Armstrong, J. E. Search for Related Content PubMed Articles by Armstrong, J. E. Agricola Articles by Armstrong, J. E.

Fringe Science: are the corollas of Nymphoides (Menyanthaceae) flowers adapted for surface tension interactions?

Behavior, Ecology, Evolution and Systematics Section, Campus Box 4120—Biological Sciences, Illinois State University, Normal, Illinois 61790-4120 USA

Received for publication May 31, 2001. Accepted for publication October 2, 2001.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Attractive features of flowers are adaptations for biotic interactions, and a few floral adaptations are for interactions with the physical environment. Marginal corollar appendages of Nymphoides (Menyanthaceae) can be membranous, a fringe of trichomes, or a ruffle. Although clearly enhancing display, a fringed corollar margin might function by generating a significant upward force through surface tension, an interaction adaptive in an aquatic environment. The force needed to dunk flowers with an intact corollar fringe and those whose fringe had been trimmed showed a significant difference. The fringe added a mean of 10.4% to the floral mass, but the upward force generated increased by nearly 50%, a significant difference from the predicted change based upon buoyancy alone. A correlation between plant form and type of corolla margin supports the surface-tension hypothesis. The membranous and ruffled corollar margins were found in species whose flowers had less risk of contacting the water's surface.

Key Words: aquatic plants • corolla • floral adaptations • floral form • Menyanthaceae • Nymphoides • surface tension

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Diverse floral forms are adaptations for biotic interactions. Flowers have such conspicuous suites of characteristics that, in many cases, the type of pollen vector can be inferred or predicted from such floral syndromes. Abiotic dispersal of pollen involves adaptations for interacting with the physical component of the environment. Many wind-pollinated flowers have large, feathery stigmatic surfaces (Faegri and van der Pijl, 1979 ) and other elaborations for pollen sorting and capture (Niklas, 1992 ). Hydrophilic flowers have floating masses of filiform pollen and other adaptations for pollination at the two-dimensional water-air interface (Cox, 1985, 1988 ). Anemophilous and hydrophilic flowers generally lack the attractants, rewards, and forms associated with biotic pollen vectors. Among zoophilic flowers, examples of floral adaptations for interacting with the abiotic environment are lacking, although pedicel modifications have been demonstrated to reduce wind drag on a large corolla (Etnier and Vogel, 2000 ). Concluding that all conspicuous, attractive features are adaptations for biotic interactions is thus reasonable.

Many species of Nymphoides (Menyanthaceae) have a habit similar to waterlilies, with rhizomatous plants rooted in the substrate, floating leaves connected to the rhizome by long petioles, and an emergent axillary inflorescence. The flowers of Nymphoides have conspicuous, attractive, UV-reflecting corollas (Ornduff, 1969 ) that clearly function to attract insect pollinators. The white or yellow corollas have a short tube and five rotate lobes with a conspicuous marginal appendage, a feature shared with other genera in Menyanthaceae, Goodeniaceae, and with some Hydrophyllaceae and Campanulaceae. The marginal corollar appendage among Nymphoides species ranges from a thin, membranous outgrowth to a fringe of long trichomes to a more robust, mostly fused, ruffled margin. These appendages make the corolla lobes at least two times broader contributing significantly to floral display. The adaptive value of enhanced displays has been demonstrated in many studies, and given the taxonomic occurrence of these corollar appendages and their presence in nonaquatic species, the most parsimonious assumption is that they arose to enhance floral display. This does not preclude either a shift in function or a dual function.

When Nymphoides flowers with trichome-fringed corollas were pulled beneath the water's surface, they reemerged completely dry and functional. If the water level rose around an inflorescence tethered to its rhizome, the corolla lobes of open flowers bent upward into a valvate, bud-like configuration. The fringe from adjacent corolla lobes met and entrapped an air bubble, not unlike a plastron of aquatic insects (Hinton, 1976 ), which traps a bubble of air for respiration within a mat of nonwettable hairs. This observation prompted an investigation of the corollar fringe as an adaptation to an aquatic environment.

Floating is the most common interaction between organisms and the water-air interface, a function of the organism's density (mass by volume) generating an upward buoyancy to overcome the pull of gravity (mass by acceleration). Surface tension interaction also generates an upward force, so negligible in most cases that it is ignored. However, a few organisms generate a significant upward force from surface tension interactions. Insects like water striders can actually walk and jump on the water surface. Vogel (1988) calculated a walk-on-water index (WOWI) as the ratio of the force holding the organism up to the force pulling the organism down. To overcome the force generated by gravity (approximately gravity [g] times an organism's density [p], times it's volume or length [l] cubed), an upward force is generated by the organism's contact perimeter (l) with the water's surface times the surface tension (). Combining terms after canceling length in the numerator produces a unitless index (Vogel, 1988 ); the formula is

(1)

Because surface tension, gravity, and the density of the study organism are all essentially constants, the variable term is the square of the length, the perimeter of contact with the water surface (Vogel, 1988 ). Given the mass of an organism, one can calculate the perimeter needed for the organism to walk on water, or conversely, given a perimeter one can calculate the mass that can be supported by surface tension. Clearly as the mass of the organism increases, the perimeter of contact must increase greatly for walking on water to remain possible. This interaction also requires nonwettable surfaces, so the cuticular layer of plants is preadapted for surface tension interaction. Like many plant surfaces, the flowers of Nymphoides have a conspicuous waxy cuticle.

The WOWI equation suggests an explanation of the antidunking function provided by corollar fringe. A fringe greatly increases the corolla perimeter without greatly increasing the mass, so it could generate a significant upward force. The fringe increases the corollar perimeter by two times the length of the trichomes times the number of trichomes less the length of the fringeless perimeter (Fig. 1). With many long, narrow, closely spaced trichomes, the increase in perimeter (l) is considerable. The membranous or ruffled corolla margins would not significantly increase the corolla perimeter because of their continuous, although slightly irregular, margin.

View larger version (27K): [in this window] [in a new window]   Fig. 1. Diagram of plain (trimmed) and fringed corollas to show the increase in perimeter produced by the maginal trichomes. The perimeter of the trimmed corolla is ten times the length of the corolla lobe margin (L). Trichomes increase the corolla perimeter by twice their mean length (2H), times the number of trichomes, less L. Trichomes are spaced 4–5/mm and are many times longer than wide

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Several species of Nymphoides were purchased from retail sources. The spellings of specific epithets varied, and specimens from several sources were misidentified. Specimens used in this study were of three basic morphs, free-floating rafts, Nymphoides indica and N. cristata; rooted rhizomatous plants with emergent/floating leaves and stout inflorescences, N. peltata (Gmel.) Kuntze; and rooted, rhizomatous plants with floating leaves and near surface inflorescences, N. geminata (Fig. 2). Grown in large plastic tubs outside the Illinois State University greenhouse facility, clonal growth was aggressive providing an ample number of flowers.

View larger version (23K): [in this window] [in a new window]   Fig. 2. Corollar appendages and plant form in Nymphoides. Species with fringed corollas were rooted rhizomatous plants with floating leaves and inflorescences near the surface (center). Species with membranous corollar margins grow as free-floating rafts with inflorescences arising from nodes just below the raft of leaves (left). Species with robust, ruffled corollar margins were rooted rhizomatous plants with emergent leaves and inflorescences borne on stout petioles and pedicels (right)

To determine the structure of the corollar margin, flowers and buds of N. indica (Armstrong 1170, ISU) and N. geminata (Armstrong 1161, ISU) were killed in a formalin : ethanol : acetic acid fixative and transfered to 50% ethanol. Buds were dehydrated to 95% ethanol to facilitate dissection. Floral parts were dehydrated, critical point dried, and prepared for observation on a scanning electron microscope by mounting on stubs and sputter-coating with gold-palladium.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
The corollar fringe of N. geminata consists of a row of trichomes along the margin of the corolla lobes. The mean length of the trichomes is 2.2 mm (± 0.2 SD) with 4–5 trichomes per linear millimeter of corolla margin (Fig. 1). Corolla lobes are 2.5–2.7 mm wide at the median and average 12.8 mm in length, so the trichomes increase the width of the corolla lobes by >150% and the perimeter of corolla (l) by 16.5–18.8 times. Although trichome length was easily measured, the functional change in perimeter may differ because observations suggest the trichomes function in clusters or clumps, like bristles in a paint brush. Thus, the contact perimeter of a fringed corolla with the water-air interface would be less than the measured perimeter, but still greater than that of a fringeless corolla.

The corollar appendage of N. indica consists of a thin membranous layer some 1.1–1.4 mm wide. The corolla lobes of this species are 6.0–6.4 mm long and 2.0–2.3 mm wide at the median, so the membranous appendage increases the width of the corolla lobes by just >100%.

Both appendages arise somewhat late in the floral ontogeny, initially appearing as a thin continuous ridge of epidermal tissue. A similar ridge develops adaxially over the median vascular bundle. These ridges appear earlier in the ontogeny of N. indica than in N. geminata, just as zonal growth beneath the anthers and corolla lobes is beginning to form a corolla tube. This ridge continues marginal growth and enlargement leading up to anthesis. In N. geminata, a series of trichome primordia appear along the ridge just after the corolla tube development has begun. Thereafter the trichomes develop individually. The corollar margin of N. indica remains a continuous layer upon which no trichomes appear.

Flowers of N. indica and N. cristata, which have a membranous corollar margins, did not survive the dunking. The corolla adhered to the water surface and although the flower floated, some of the corolla lobes remained trapped beneath the surface. Water filled the corolla tube and wetted anthers and pollen. Flowers of N. geminata, which have a more robust ruffled margin, did not survive the dunking. The more robust corolla lobes only flex slightly. After dunking, the corolla tube remained filled, and anthers and pollen were wetted. The flowers of N. geminata, which have a fringe of trichomes, survived the dunking fully functional because of an air bubble trapped within the corolla. Submerged flowers bobbed back to the surface and reopened. Since only one form of the corollar appendage appeared to have any antidunking function, only flowers of N. geminata were subjected to further manipulation. Stigmatic function would appear to survive dunking, but observations suggest insects may avoid landing on flowers partially flooded with water.

Flowers of Nymphoides geminata had a mean mass (± 1 SD) of 145.7 mg (± 27.6 mg). Trimming the fringe from the corolla reduced their mean mass to 130.6 mg (± 25.2 mg). Thus, the corollar fringe increased the mass of the flowers 11.5%. Flowers with trimmed corolla lobes required a mean of 3.89 g (± 0.64 g) to be dunked. Fringed flowers required a mean of 5.73 g (± 0.67 g) to be dunked, an increase on average of 47%, which is significantly different from both the fringeless flowers and the predicted change based solely upon the change in mass (Fig. 3). Flowers whose fringe had been removed reemerged from the dunking with the androecium wetted and with water in the corolla tube.

View larger version (30K): [in this window] [in a new window]   Fig. 3. Paired bars show mean mass (in grams) required to dunk flowers with both fringed (dark) and trimmed (white) corollas. Bar on right shows predicted mean mass to dunk trimmed flowers based upon buoyancy alone. Ninety-five percent confidence limits are shown

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Because the fringe increases the corollar perimeter theoretially by 16–18 times, the WOWI for the upward force generated by the fringe should be some 200–300% greater. However, the surface tension interaction generated only one-quarter to one-sixth the calculated force. Working backwards using the WOWI, the functional perimeter of the corolla must be increased by 68%, far less than the actual lengths of the fringe hairs, a result consistent with the observation that trichomes worked in clumps.

This antidunking adaptation could be beneficial in one or more of three ways. Plants living in a shallow-water habitat (15–20 cm) probably experience rapid fluctuations in water depth, but because water depth tends to increase rapidly and subside more gradually, the advantage conferred by antidunking would seem minimal for 1-d duration flowers. Loss of a few flowers would seem of little consequence with regard to the total blooms over a growing season. Inflorescences of Nymphoides geminata positioned flower buds at or beneath the water surface. At anthesis flowers were barely emergent rising above the water level by 1–2 cm. Flower buds developed just below the surface. Under these circumstances, a flower bud opening just at or just below the surface would generate a considerable upward force exerting a pull on any slack in the pedicel. Lastly, the fringe allows these flowers to remain quite functional when in full contact with the water's surface. Flowers of N. geminata could be pushed 1 cm below the level surface of the water without either completely closing the flower or breaking the surface tension. These flowers could support the mass of any typical pollinator while floating. Larger Hymenopteran pollinators observed visiting the smaller flowers of N. indica sometimes dunked the flower, ending the visit.

The correlation between floral form and plant form in a limited sampling of Nymphoides species offers some support for the surface tension interaction hypothesis. In other species, the flowers were held aloft above the surface of the water by either a floating ramet or a robust inflorescence (Fig. 2). A free-floating ramet supported by the buoyancy of a raft of leaves simply rises and falls with the water level maintaining its flowers 2–3 cm above the surface. The axillary inflorescences arise from a node just beneath a floating leaf. Species, e.g., N. peltata, with larger, more robust petioles and pedicels held both leaves and inflorescences several centimeters aloft. Clearly, these species have no need of an antidunking adapation.

These results suggest the corollar fringe of some species of Nymphoides provides an antidunking adaption while still functioning to augment display. Such double function pollination adaptions may be one of the means by which floral form evolves (Stebbins, 1988 ). It remains for field studies to confirm or reject this hypothesis. If this surface tension interaction is confirmed, it adds a novel means by which zoophilic flowers interact with the physical environment.

	FOOTNOTES

 
¹

² jearmstr{at}ilstu.edu .

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Cox P. A. 1985 Noodles of the tide. Natural History 94: 36-40

Cox P. A. 1988 Hydrophilous pollination. Annual Review of Ecology and Systematics 19: 261-280

Etnier S. A. S. Vogel 2000 Reorientation of daffodil (Narcissus: Amaryllidaceae) flowers in wind: drag reduction and torsional flexibility. American Journal of Botany 87: 29-32[Abstract/Free Full Text]

Faegri K. L. van der Pijl 1979 The principles of pollination ecology, 3rd ed. Pergamon Press, Oxford, UK

Hinton H. E. 1976 Plastron respiration in bugs and beetles. Journal of Insect Physiology 22: 1529-1550[CrossRef]

Niklas K. J. 1992 Plant biomechanics: an engineering approach to plant form and function. University of Chicago Press, Chicago, Illinois, USA

Ornduff R. 1969 Neotropical Nymphoides (Menyanthaceae): Meso-American and West Indian species. Britonnia 21: 346-352

Stebbins G. L. 1988 Flowering plants: evolution above the species level. Belknap Press of Harvard University Press, Cambridge, Massachusetts, USA

Vogel S. 1988 Life's devices: the physical world of animals and plants. Princeton University Press, Princeton, New Jersey, USA

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Armstrong, J. E. Search for Related Content PubMed Articles by Armstrong, J. E. Agricola Articles by Armstrong, J. E.