| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
2School of Botany and Zoology, University of Natal, P. Bag X0 I, Scottsville, Pietermaritzburg 3209, South Africa; 3Department of Botany, University of Cape Town, Rondebosch 7700, South Africa
Received for publication October 27, 2000. Accepted for publication March 8, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
20%), the nectar of M. depressa is almost 400 times as viscous as an equivalent sugar solution. The jelly-like constituent in the nectar may discourage robbing by insects, while also facilitating lapping by rodents. Our findings illustrate the utility of floral syndromes for generating testable predictions about pollination systems.
Key Words: convergent evolution • floral syndrome • Hyacinthaceae • Massonia depressa • nectar • pollen • pollination • rodent • southern Africa
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), and, until now, was thought to be confined to shrubs belonging to three dicotyledon families: Proteaceae, Loasaceae, and Melastomataceae (Lumer, 1980
; Wiens et al., 1983
; Cocucci and Sérsic, 1998
). Here we report the discovery of rodent pollination in a monocotyledon and the first record of pollination by gerbils.
The group of flower visitors known as "nonflying mammals" (to distinguish it from bats) is composed mainly of marsupials, primates, and rodents. Since the first study by Porsch (1934)
, nearly 100 plant species, mainly in the southern hemisphere, have been found to be visited by at least 59 species of nonflying mammal (Carthew and Goldingay, 1997
, and references therein). In only a handful of studies, however, has it been shown that the flowers are actually pollinated by these animals and in even fewer instances has it been shown that they are the primary pollinators and therefore likely to have shaped the evolution of floral traits.
Evidence that a plant is adapted for pollination primarily by nonflying mammals ideally must include: (1) observations of nondestructive feeding on flowers; (2) finding pollen on fur or in the feces of the animals; (3) observed contact between body parts covered with pollen and the stigma; (4) infrequent insect or bird visitation; and (5) reduced fecundity when nonflying mammals are excluded from flowers. In addition, (6) nectar secretion, scent production, and floral anthesis would be expected to be synchronized with the activity periods of nonflying mammals in a plant that is adapted for pollination by these animals.
Plants adapted for pollination by nonflying mammals tend to have robust flowers with dull coloration, copious amounts of nectar, and exserted styles and stamens (Carthew and Goldingay, 1997
). Within this broad "syndrome" of traits, however, there are important differences that reflect the morphology and foraging behavior of the three main nonflying mammal groups. For example, flowers pollinated by primates tend to be unscented and very large to accommodate these relatively heavy flower visitors (Nilsson et al., 1993
; Kress et al., 1994
); flowers pollinated by marsupials are usually situated in the canopy (Turner, 1982
), whereas flowers pollinated by rodents tend to be situated at ground level (Wiens et al., 1983
). Thus, it seems appropriate to consider pollination by rodents distinct from, yet functionally related to, pollination by other groups of nonflying mammals.
The flowers of the South African lily Massonia depressa Houtt. (Hyacinthaceae) captured our attention because they have many features in common with rodent-pollinated proteas, such as being positioned at ground level and lacking the colorful petals that most plants use to attract insects or birds (Fig. 2). The flowers also share stiff anthers, a stigma–nectar distance of 10 mm, a yeasty scent, and a winter flowering period with their Protea counterparts (Wiens et al., 1983
). On the assumption that convergent evolution of floral traits among unrelated plants reflects adaptation to similar pollinators (Faegri and van der Pijl, 1979
; Johnson and Steiner, 2000
), we hypothesized that M. depressa is pollinated by rodents.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
40 h over 6 d, of which 25 h were in daylight and 15 h at night.
For ethical reasons we carried out only live-trapping of rodents, although it is known that grooming by rodents in live traps can significantly reduce the load of pollen on their fur (Wiens et al., 1983
). Peanut butter and rolled oats were used to bait the traps. We laid out 100 traps in four lines with 5 m between each trap at the Vanrhynsdorp site on the nights of 21 and 22 July 1998 and 70 traps in five lines with 5 m between each trap at Nieuwoudtville on the nights of 23 and 24 July 1999.
Rodents were removed from traps in the early morning (0600) and temporarily placed in plastic bags. Photographs and measurements were taken in order to confirm their identity. The snouts of the rodents protruded through a small hole in the corner of the plastic bag and were swabbed with a small block of fuchsin gelatine (Beattie, 1971
), which was then melted onto a slide and examined microscopically for the presence of pollen. Rodent feces were retrieved from the traps and flowers and crushed in a mixture of water and fuchsin gelatine. A sample of this mixture was mounted on a slide and examined microscopically for the presence of pollen.
Evidence of rodent activity at flowers has been previously obtained from footprints left on glass "smoked plates" (Cocucci and Sérsic, 1998
). We used sheets of stiff white cardboard (270 x 150 mm), which were blackened with soot over a paraffin lamp and placed adjacent to 11 flowering plants of M. depressa at the Nieuwoudtville site on the night of 24 July 1999. Footprints on these plates were compared to those made by captive rodents.
As it is virtually impossible to observe rodents in the field at night (cf. Wiens et al., 1983
), we recorded the foraging behavior of captured rodents. Three gerbils captured earlier in the evening at the Vanrhynsdorp site were released one at a time from 2400 to 0100 into a 100 cm long by 30 cm wide glass tank with a 10 cm deep layer of sand containing two flowering plants of M. depressa. We also added several freshly picked Hibiscus flowers to determine if the rodents visit flowers selectively.
Floral characteristics
Dimensions of ten flowers, each sampled from a different plant, were measured to the nearest 0.5 mm. Patterns of anthesis were determined by marking 33 plants and recording the number of open flowers at 3-h intervals for 24 h.
Daily rhythms in the nectar properties of M. depressa were determined by sampling the standing crop of nectar from ten randomly selected flowers approximately every 2 h over a 24-h period at the Vanrhynsdorp site in 1998 and the Nieuwoudtville site in 1999. The nectar volume was measured with 100 µL capillaries (Drummond Scientific Company, Broomall, Pennsylvania, USA) and nectar concentration was determined with a 0–50% field refractometer (Bellingham and Stanley, Tunbridge Wells, UK). Samples of nectar from five flowers were spotted onto filter paper and air dried for later determination of constituent free sugars using standard high pressure liquid chromatography methods, as detailed by Van Wyk (1993)
.
Viscosity of the nectar was determined from the descent times of a 20-µL sample over a distance of 80 mm in a 100-µL capillary with 0.563 mm radius, using the formulas given by Heyneman (1983)
. The accuracy of this method was verified from timed descent of sugar solutions of known viscosity (Fasman, 1975
).
Breeding system and exclusion of rodents from flowers
To determine the breeding system of M. depressa, we kept ten plants in a pollinator-excluded greenhouse at the University of Cape Town and randomly selected newly opened flowers for one of three treatments: unmanipulated to test for autogamy, pollinated by hand with self- pollen to determine whether plants are self-compatible, and pollinated by hand with pollen from a different plant. The number of flowers that developed fruits and the mass of the fruits were determined for each treatment.
To determine whether rodents are important for seed production in Massonia, we selected 23 pairs of plants with the same number of flower buds at the lowland Vanrhynsdorp site and enclosed one member of each pair in a wire cage with a mesh diameter of 15 x 20 mm that excluded rodents but allowed insects free access to the flowers.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
Soon after its release back into the field at the Nieuwoudtville site in 1999, we observed an individual Namaqua rock mouse (Aethomys namaquensis) feeding on the flowers of several plants of M. depressa at 0700.
Floral morphology and patterns of anthesis
The flowers of M. depressa are situated at ground level and lack any bright pigmentation (Fig. 2). The nectar is situated in a bowl-shaped chamber formed by the fused bases of the anthers (Figs. 2, 4). The chamber is 5.6 mm in depth (SD = 1.3, N = 10), and its entrance measures 9.1 mm (SD = 1.4, N = 10) by 5.9 mm (SD = 1.1, N = 10). The anthers are 18.3 mm in height (SD = 2.8, N = 10), while the stigma is 15.9 mm in height (SD = 5.6, N = 10). The diameter of the inflorescence is 28.8 mm (SD = 8.8, N = 10).
We recorded 88 open flowers (2.7 flowers per plant) on the 33 marked plants when observations commenced at 0700, which increased to 118 open flowers (3.57 flowers per plant) by 0700 the following day. Floral anthesis occurred mainly during the afternoon; of the 30 flowers that opened during the observation period, 16 (53%) opened in the 3-h period between 1300 and 1600. The flower opening patterns were significantly nonrandom (Table 2).
|
), increased from 3.16 ± 4.7 mg at 1500 to 14.22 ± 8.8 mg at 2400. The sugars consisted of sucrose (42%), glucose (23%), fructose (32%), and xylose (3%).
|
), was found to be almost 400 times that of a solution of the equivalent sugar concentration (Fasman, 1975
) (0.6544 ± 0.115 Pa.s for Massonia nectar, N = 6, vs. 0.00165 Pa.s for 16% sucrose solution at 20°C).
Breeding system and exclusion of rodents from flowers
Production of seeds by Massonia plants clearly requires visits by pollinators, as no fruit set occurred in flowers from which pollinators were excluded (N = 29). The average fruit set was 38.5% in self-pollinated flowers (N = 26) vs. 82.4% in cross-pollinated flowers (N = 34). This difference is highly significant (
2 = 10.5 with Yates correction, P = 0.001). In addition, the average mass of fruits formed as a result of self-pollination (0.19 g, SD = 0.17, N = 10) was significantly less than those formed as a result of cross-pollination (0.29 g, SD = 0.12, N = 28) (t = 2.18, P = 0.035, two-tailed).
Exclusion of rodents from flowers at the Vanrhynsdorp site resulted in a significant decline in the number of plants that set fruit, as well as the number of fruits and seeds per plant (Table 3). Only 4.3% of plants in the rodent-excluded group set fruit vs. 30.4% in the control group, and an average of only 1.95 seeds were produced in rodent-excluded plants vs. 20.0 seeds in the control plants (Table 3). Honey bees were observed to collect pollen from flowers of M. depressa on approximately ten occasions at the Vanrhynsdorp site, but no insect visitors were observed at the Nieuwoudtville site. No visits by sunbirds were recorded at either site. Sunbirds ignored flowering plants of M. depressa while foraging on flowers of Microloma sagittata (Asclepiadaceae) just a few meters away.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Two of the rodent species found to carry pollen of M. depressa have been previously recorded as vectors of the pollen of Protea species (Wiens et al., 1983
), but the capture of gerbils with M. depressa pollen loads was unexpected and is apparently the first example of pollination by these animals. Wiens et al. (1983)
captured five individuals of the gerbil Tatera afra with very small amounts of pollen originating from Protea laurifolia, but there was no evidence that the gerbils effected pollination of the flowers of this species, which is morphologically adapted for pollination by birds.
Massonia pollen was present in very large amounts in the feces of the captured rodents, and, in some cases, made up the bulk of the solid matter. This pollen was presumably ingested only after grooming, as captive rodents made no attempt to feed directly on pollen in the anthers. Previous studies have shown that rodents can extract protein from Protea pollen (van Tets, 1997
). By flowering in the middle of winter, when food resources are scarce, flowers of M. depressa, like their Protea counterparts, may become a valuable source of energy and protein for rodents. Nevertheless, nectar and pollen are probably such a minor and spatially restricted element of the diet of rodents that morphological and behavioral adaptations for flower feeding are highly unlikely (Wiens et al., 1983
).
The wide-mesh cages that we used to exclude rodents from flowers would have also excluded birds and bats, but sunbirds were present only at the Nieuwoudtville site and ignored flowers of M. depressa, despite foraging on other plants in close proximity. Wiens et al. (1983)
reported "rare and inconsequential" sunbird visits to rodent-pollinated Protea species, but flowers of Cajophora coronata (Loasaceae) in South America are frequently pollinated by at least five bird species, in addition to rodents (Cocucci and Sérsic, 1998
). Pollination by bats has not been recorded in the western half of southern Africa and is most unlikely to occur in M. depressa, as there are no flower-visiting bats in the Succulent Karoo region. The Egyptian fruit bat (Rousettus aegyptiacus) has been recorded from some large caves in the Cape region, but it is not known to occur in the Succulent Karoo.
Nectar characteristics
The presence of the pentose sugar xylose in the nectar of M. depressa, albeit at a concentration of just 3%, is interesting because it has been previously reported as a common constituent of rodent-pollinated Protea species in South Africa (Van Wyk and Nicolson, 1995
). The ecological significance of xylose, however, remains unclear. Birds and honey bees show strong aversion to xylose, but it is willingly consumed by the Namaqua rock mouse (Aethomys namaquensis; Johnson, Van Tets, and Nicolson, 1999
). Also, in contrast to birds, Namaqua rock mice show up to 97% absorption of xylose and probably metabolize the sugar with the aid of intestinal bacteria (Johnson et al., 1999
).
The jelly-like constituent in the nectar of M. depressa is probably a mucopolysaccharide (unpublished data). The remarkable viscosity of the nectar (almost 400 times greater than a solution of equivalent sugar concentration) may serve to prevent robbing by insects. Although the nectar is very exposed, it did not appear to be fed on by insects, which may reflect the difficulty they have of drawing viscous solutions through their tubular mouthparts (cf. Heyneman, 1983
). An alternative, though not mutually exclusive, explanation is that rodents prefer to feed on viscous nectar because it is more easily lapped into their mouths with their tongues. Bat-pollinated Musa spp. (Musaceae) have jelly-like nectar, while their bird-pollinated congeners have watery nectar (Nur, 1976
). Polysaccharide mucilage has been reported in the nectar of Musa paradisiaca L. var. sapientum Kuntze (Fahn and Benouaiche, 1979
). There is no evidence that the nectar of rodent-pollinated proteas is particularly viscous, however Cocucci and Sérsic (1998)
describe the nectar of a putatively rodent-pollinated South American plant, Bromelia urbaniana (Bromeliacae), as "jelly-like."
The rodent pollination floral syndrome
The flowers of M. depressa show striking convergence to rodent-pollinated proteas. Convergent traits include: dull-colored, robust flowers positioned close to the ground with exserted wiry anthers and a stigma–nectar distance of 10 mm; a yeasty odor; copious amounts of easily accessible nectar secreted in the evening; and winter flowering. Other South African monocotyledons, such as Androcymbium pulchrum Schltr. & Krause, Androcymbium eucomoides (Jacq.) Willd. (Colchicaceae), and Eucomis humilus Bak. (Hyacinthaceae), have similar traits that fit the rodent pollination syndrome, suggesting that this mode of pollination may be more widespread than previously believed.
The concept of floral syndromes (Faegri and van der Pijl, 1979
) has come under criticism for being typological and lacking the flexibility to accommodate the generalized nature of many pollination systems (cf. Waser et al., 1996
). But here we have shown that convergent syndromes can be used successfully to develop predictive hypotheses about which animals have shaped the evolution of floral traits. We emphasise that all hypotheses based on floral syndromes must be tested through field observations and experiments.
|
FOOTNOTES |
|---|
4 Author for reprint requests (e-mail: JohnsonSD{at}nu.ac.za
.).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Carthew S. M. R. L. Goldingay 1997 Non-flying mammals as pollinators. Trends in Ecology and Evolution 12: 104-108[CrossRef]
Cocucci A. A. A. N. Sérsic 1998 Evidence of rodent pollination in Cajophora coronata (Loasaceae). Plant Systematics and Evolution 211: 113-128[CrossRef][ISI]
Dafni A. 1992 Pollination ecology: a practical approach. Oxford University Press, Oxford, UK
Faegri K. L. van der Pijl 1979 The principles of pollination ecology. Pergamon, Oxford, UK
Fahn A. P. Benouaiche 1979 Ultrastructure, development and secretion in the nectary of banana flowers. Annals of Botany 44: 85-93
Fasman G. D. 1975 Handbook of biochemistry and molecular biology, vol. 1, Physical and chemical data. CRC Press, Cleveland, Ohio, USA
Heyneman A. J. 1983 Optimal sugar concentrations of floral nectars—dependence on sugar intake efficiency and foraging costs. Oecologia 60: 198-213[CrossRef][ISI]
Johnson S. A. I. G. van Tets S. W. Nicolson 1999 Sugar preferences and xylose metabolism of a small mammal pollinator, the Namaqua rock mouse (Aethomys namaquensis). Physiological and Biochemical Zoology 72: 438-444[CrossRef][ISI][Medline]
Johnson S. D. K. E. Steiner 2000 Generalization versus specialization in plant pollination systems. Trends in Ecology and Evolution 15: 140-143
Kress W. J. G. E. Schatz M. Andrianifahanana H. S. Morland 1994 Pollination of Ravenala madagascariensis (Strelitziaceae)) by lemurs in Madagascar: evidence for an archaic coevolutionary system. American Journal of Botany 81: 542-551[CrossRef][ISI]
Lumer C. 1980 Rodent pollination of Blakea (Melostomataceae) in a Costa Rican cloud forest. Brittonia 32: 512-517[CrossRef][ISI]
Nilsson L. A. E. Rabakonandrianina B. Pettersson R. Grünmeier 1993 Lemur pollination in the Malagasy rainforest liana Strongylodon craveniae (Leguminosae). Evolutionary Trends in Plants 7: 49-56
Nur N. 1976 Studies on pollination in Musaceae. Annals of Botany 40: 167-177
Porsch O. 1934 Säugetiere als Blumenausbeuter und der Frage der Säugetierblume 1. Biologia Generalis 10: 657-685
Turner V. 1982 Marsupials as pollinators in Australia. In J. A. Powell and R. J. Richards [eds.], Pollination and evolution, 55–66. Royal Botanic Gardens, Sydney, Australia
Van Tets I. G. 1997 Extraction of nutrients from Protea pollen by African rodents. Belgian Journal of Zoology 127: 59-65[ISI]
Van Wyk B.-E. 1993 Nectar sugar composition in southern African Papilionoideae (Fabaceae). Biochemical Systematics and Ecology 21: 271-277[CrossRef][ISI]
Van Wyk B.-E. S. W. Nicolson 1995 Xylose is a major nectar sugar in Protea and Faurea. South African Journal of Science 91: 151-153[ISI]
Waser N. M. L. Chittka M. V. Price N. M. Williams J. Ollerton 1996 Generalization in pollination systems, and why it matters. Ecology 77: 1043-1060[CrossRef]
Wiens D. J. P. Rourke 1978 Rodent pollination in southern African Protea spp. Nature 276: 71-73[CrossRef][Medline]
Wiens D. J. P. Rourke B. B. Casper E. A. Rickart T. R. LaPine C. J. Peterson A. Channing 1983 Nonflying mammal pollination of southern Africa Proteas: a noncoevolved system. Annals of the Missouri Botanical Garden 70: 1-31
This article has been cited by other articles:
|
|
 |
|
  C. Kleizen, J. Midgley, and S. D. Johnson Pollination Systems of Colchicum (Colchicaceae) in Southern Africa: Evidence for Rodent Pollination Ann. Bot., November 1, 2008; 102(5): 747 - 755. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Pauw Floral syndromes accurately predict pollination by a specialized oil-collecting bee (Rediviva peringueyi, Melittidae) in a guild of South African orchids (Coryciinae) Am. J. Botany, June 1, 2006; 93(6): 917 - 926. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. OLLERTON, S. D. JOHNSON, L. CRANMER, and S. KELLIE The Pollination Ecology of an Assemblage of Grassland Asclepiads in South Africa Ann. Bot., December 1, 2003; 92(6): 807 - 834. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
