Pollination Biology of Lapeirousia subgenus Lapeirousia (Iridaceae) in southern Africa; floral divergence and adaptation for long-tongued fly-pollination Abstract | Materials and Methods | Results | Discussion | Literature DISCUSSION The limited published research on pollination ecology within the Iridaceae has tended to emphasize genera with meranthia of gullet flowers such as Iris, Moraea, and Gynandriris (Faegri & van der Pijl, 1979). The pollination ecology of Lapeirousia, however, is far closer to what has been described in the southern African genus Nivenia (Goldblatt & Bernhardt, 1990). In Nivenia floral tubes are not occluded and the stamens and styles are prominantly displayed. The flowers are pollinated by long-tongued bees and nemestrinid flies in the genus Prosoeca. Access to the nectar secreting tube is direct and the insect head and thorax contact the anthers and stigma while the insect hovers or clings to the tepals. The disparity between the length of perianth tube in subgenus Lapeirousia and the shorter length of the mouth parts of the primary pollinators is quite easy to explain. Records of nectar secretion show that species in subgenus Lapeirousia secrete copious amounts of fluid for insect-pollinated flowers, and it is most unlikely that dominant pollinators are ever forced to extend their mouth parts to the base of the tube unless all the nectar has been removed by earlier foragers. More important, Darwin (1862) hypothesized that successful pollination of spurred orchids occurred when orchids evolved floral spurs slightly longer than the tongues of their pollinators, forcing the insect to ram its head down the floral throat, ensuring contact between the insect's head and the orchid's column. This has since been shown experimentally by Nilsson (1988). As in the nectariferous orchids, species in subgenus Lapeirousia "oblige" their pollinators to make contact with the anthers and stigmatic surfaces of the style branches of the flowers that block or at least encircle the entrance to the floral tube. Some members of subgenus Lapeirousia appear to be self-compatible. Spatial isolation between anthers and stigmas is not always well expressed in subgenus Lapeirousia and protandry is weakly developed. Four Lapeirousia species show successful fruit and seed production in the absence of pollinators. This is most likely the result of mechanical autogamy. (We consider the alternative possibility that apomixis takes place most unlikely: the phenomenon is unknown in Iridaceae.) Three of the four species that show self-pollination start flowering during the southern African winter when rain and low temperatures may restrict pollinator activity. Mechanical autogamy by contact between the stigmatic surfaces and the pollen (or apomixis) then becomes a fail-safe mechanism in the absence of dependable pollinators. This has also been described in the late winter-early spring flowering herbs of North America (Schemske et al., 1978) and some terrestrial orchids of southern Australia (Dafni & Bernhardt, 1989). There are two major differences between floral mechanisms in Nivenia versus Lapeirousia. First, fly-pollination in Nivenia appears to be restricted to nemestrinids. In some Lapeirousia species pollination may be dependent on the long-tongued tabanid, Philoliche gulosa. Second, and more important, analysis of floral presentation and observation of floral foragers emphasize that pollination systems in southern African species of subgenus Lapeirousia can be subdivided into a minimum of three syndromes, or perhaps four if the sphinx moth syndrome is regarded as separate from the generalist system that otherwise prevails in those species with the L. divaricata-type flower. Pollination by nectar foraging flies has been treated as a relatively common but unspecialized syndrome in which many fly taxa visit the same flower, and the dispersal of pollen may be shared with co-foraging bees and butterflies (Grant & Grant, 1965; Barth, 1985). We may compare the more classical treatment of myophily with our results. Fly-pollination in Lapeirousia has evolved into such a specialized syndrome that two different modes of floral presentation appear to attract and depend on two different sets of fly genera. Species with the L. silenoides type of flower appear to depend exclusively on two species of flies in one genus, Prosoeca. We also note that plants with the L. silenoides-type flower appear to be restricted to the west coast and adjacent near interior of southern Africa. The species exhibiting the L. fabricii-type of presentation frequently have marginally longer tubes than species that have the L. silenoides-type flower and seem to be pollinated exclusively by Moegistorhynchus longirostris and Philoliche gulosa. Plant species with this flower type occur widely across southern Africa, although they appear to be most frequent in the southwest and west of the subcontinent. Among the species with this type of flower, L. anceps stands out in its remarkable range of perianth tube lengths, 20-76 mm. The pattern suggests that only populations in the west and north of its range with tubes 45-76 mm long are pollinated by M. longirostris, because nectar is accessible to only this fly, given the length of its proboscis. Populations with shorter tubes, as little as 20-30 mm in the south of its range, cannot be pollinated by this fly species, which does not occur in this part of the range of L. anceps. The pollinator(s) for these short-tubed plants must be some other, presumably long-tongued fly, possibly Philoliche gulosa which does occur here and has a proboscis 20-33 mm long. Evidently, tube length is extremely labile and may respond rapidly to selection by pollinators. Despite the segregation of Lapeirousia species into three pollination guilds, the majority of species in this subgenus secrete sucrose-rich/sucrose-dominant nectar regardless of the major pollinators. This adds a new dimension to the analytic work and categorization of nectar by Baker & Baker (1983, 1990). In their earlier treatments of myophily the Bakers found that fly-pollinated flowers tended to be weak in sucrose, like the flowers pollinated by short-tongued bees. However, when the Bakers analyzed fly flowers they concentrated on taxa pollinated by short-tongued flies, such as the Muscidae, Syrphidae, and Phoridae. It now appears that just as flowers pollinated by long-tongued bees are usually rich in sucrose (Baker & Baker, 1983, 1990), flowers pollinated by long-tongued flies are also sucrose producers. Perhaps pollination by large-bodied physically active insects that maintain wing movement while feeding (e.g., nemestrinids, Philoliche, sphinx moths and some anthophorids (Goldblatt & Bernhardt, 1990) requires an emphasis on sucrose instead of hexose rewards that is independent of insect order. Although certain Lapeirousia species may be pollinated by only one or two fly species, the degree of dependency in this insect-flower relationship is not shared to the same degree by the flies. As in so many proposed cases of co-adaptation, the Lapeirousia flowers appear to have become modified for pollination by specific flies to a greater extent than flies have become modified for Lapeirousia flowers. The evidence for this unequal relationship is presented in the polytrophic foraging behavior of flies at field sites and confirmed by the results of pollen load analyses. It is more likely that the Lapeirousia species with L. silenoides-type and L. fabricii-type flowers belong to broader guilds, encouraging the partitioning of long-tongued fly pollination into more than one syndrome in the southern African flora. Observations on flowers in other genera and families visited by long-tongued flies, e.g., Babiana and Hesperantha (Iridaceae), Pelargonium (Geraniaceae), suggest that floral presentation for pollination by Prosoeca species or by Moegistorhynchus and Philoliche shows a high degree of convergence in such floral characters as color patterns and tube or spur length. The types of floral presentation in southern African plants pollinated by different fly genera and species may ultimately prove to be as diverse as, yet distinct from, pollination guilds in other parts of the world, such as members of the neotropical flora that are pollinated by straight-billed hummingbirds versus those plants taxa pollinated by hermit hummingbirds with curved bills (Feinsinger & Colwell, 1978; Feinsinger et al., 1985). Pollen-load analysis of the bee-pollinated members of the Lapeirousia divaricata group suggests community dynamics typical of flowers pollinated by long-tongued bees in other parts of the world. These anthophorids and long-tongued bees in the Apidae (e.g., Bombus, Euglossa) often show foraging strategies in which individual bees balance visits to nectarless flowers [e.g., Dianella (Phormiaceae), Echeandia (Anthericaceae), Schrankia (Fabaceae), Acacia (Fabaceae), Hibbertia (Dilleniaceae) that offer copious pollen with visits to plants that produce copious nectar, but from which pollen is not collected actively (Bernhardt, 1989, 1990; Bernhardt & Montalvo, 1979; Bernhardt, 1995). The presence of the pollen of nectarless Cyanella and Hermannia on female anthophorids collected on flowers of the L. divaricata-type suggests that community pollination by long-tongued bees in southern Africa may not be significantly different from those syndromes cited above for the floras of Central America and southern Australia, and the woodlands and prairies of North America (Schemske et al., 1978). Members of subgenus Lapeirousia now join an expanding list of plant taxa in which segregated pollen flow is due in part to ethological isolation (Grant, 1994). In subgenus Lapeirousia ethological isolation (sensu Grant, 1994) appears to be based on two factors. First, as in Aquilegia (Ranunculaceae), different pollinators may be restricted to different plant taxa as a partial consequence of mechanical isolation (Grant, 1971), as bees and moths are probably unable to forage successfully on the flowers of Lapeirousia species in which tube length far exceeds tongue length. Second, ethological isolation must also be based in part on flower constancy as differing modes of floral presentation, featuring diverse color patterns and scent production, produce different responses in the foraging behavior of local pollinators that are polytrophic and/or polylectic. That is, flies and bees are not expected to respond to the same forms of floral advertisement due to their different visual and olfactory senses (Barth, 1985). Observations of interspecific hybridization in sympatric and co-blooming populations of Lapeirousia fabricii and L. jacquinii, and of L. silenoides and L. verecunda respectively, indicate that when members of the different pollination guilds are sympatric, ethological mechanisms are sometimes insufficient to prevent interspecific pollination, as evidenced by the recent discovery of scattered hybrids between members of each of the above pairs. In fact, species belonging to different guilds seldom co-occur and when they do they usually flower at different times. These observations suggest that floral divergence is not a result of selection for prepollination isolating mechanisms. Postpollination isolating mechanisms may operate in Lapeirousia jacquinii and L. violacea. These species belong to the same guild and have been seen to be visited by the same fly individuals. Despite this, no F1 hybrids have been found after three years of fieldwork and it seems likely that biochemical recognition and rejection of interspecific pollens may be a more important form of interspecific isolation within some Lapeirousia species belonging to the same pollination guild. However, in other instances interspecific pollen recognition may offer incomplete isolation, as L. jacquinii and L. pyramidalis subsp. regalis have the same mode of floral presentation and hybrids between the two have been recorded at one site (Goldblatt & Manning, unpublished). Information on pollination systems in subgenus Lapeirousia may now be combined with a cladistic analysis to help determine the evolution of pollination syndromes, extensively discussed by Goldblatt & Manning (in mss). The picture that emerges from that study Figure 5 indicates most strongly that the dependence of a Lapeirousia species on a particular guild of long-tongued pollinators has originated several times. While it is true that some sister species (e.g., L. dolomitica and L. violacea or L. divaricata and L. spinosa) share the same pollination syndrome, this sharing appears to be an exception within the current phylogenetic tree. From the combined data we conclude that floral evolution in subgenus Lapeirousia is extremely labile and probably reflects a rapid response to the relative diversity of potential vectors within a given geographic area. It is possible to predict with some degree of confidence the ancestral mode of pollination within subgenus Lapeirousia. We know that in the outgroup, subgenus Paniculata, the majority of species bear relatively short-tubed flowers with pollination types most similar to the L. divaricata-type in subgenus Lapeirousia. Goldblatt (1990) reported that two species, L. erythrantha (Klatt) Bak. and L. avasmontana Dinter (both tropical members of subgenus Paniculata), are actively visited by bees, wasps, and diurnal Lepidoptera, and L. sandersonii Bak. of the subgenus is visited predominantly by diurnal Lepidoptera (Manning, unpublished). This is consistent with our observations here (Table 2) for L. azurea (Eckl. ex Bak.) Goldblatt, L. fastigiata (Lam.) Ker-Gawl., and L. neglecta Goldblatt & Manning, and Scott Elliot's (1891) report for L. corymbosa (all subgenus Paniculata), also most commonly visited by bees. Therefore, short-tubed and funnel-shaped flowers dependent on Hymenoptera and Lepidoptera may be basal to subgenus Lapeirousia, and it seems far more parsimonious to infer that flowers with long perianth tubes and associated nectar guides are ultimately derived from short-tubed flowers with simple nectar guides. In light of this, the terminal position of the species pair, L. divaricata-L. spinosa, primarily pollinated by bees, and nested within a clade of long-tongued fly-pollinated taxa, evidently represents a reversal to an ancestral pollination strategy. This emphasizes the extreme degree of adaptive radiation exhibited within subgenus Lapeirousia and the impact of pollinators on floral morphology and biochemistry. << BACK | TOP | NEXT >>