The Prosoeca peringueyi (Diptera: Nemestrinidae) pollination guild in southern Africa: long-tongued flies and their tubular flowers
Abstract | Materials and Methods | Results | Discussion | Literature
The Lapeirousia silenoides-type Floral Syndrome
Our observations indicate that plants with flowers conforming to the following syndrome
constitute a guild adapted for pollination by the long-proboscid flies, Prosoeca peringueyi and P. sp. nov.: perianth or hypanthium tube narrow, straight or slightly curved, and 30-60 mm long; tepals or petals short in relation to tube length and pigmented dark purple to crimson, or
sometimes lilac or pale mauve, with nectar guides consisting of white to cream spots and streaks
and areas of darker pigmentation; and exserted and prominent anthers and stigmas that are
presented outside the mouth of the tube so that they will contact the body of any animal that
probes the floral tube. Associated with these features is the production of nectar with a relatively
high sugar concentration and the absence of floral odor (present in two species). The plant taxa
occurring north of 29°S, i.e., in the Richtersveld and southern Namibia, comprise a distinct subset of the guild characterized by the paler-colored lilac or pale mauve flowers. This group comprises Lapeirousia dolomitica subsp. dolomitica, Tritonia marlothii, Pelargonium cortusifolium, and P. crassicaule.
Only Pelargonium sericifolium has no measurable floral nectar and we presume that it is
an example of pollination by deceit (Goldblatt et al., 1995). The flowers closely resemble in color
and shape those of Lapeirousia silenoides and P. magenteum. These species frequently co-occur with P. sericifolium and have ample amounts of nectar of relatively high sugar concentration. Two species of Babiana and one subspecies of L. dolomitica are exceptional here in having sweetly scented flowers. Presence of scent is usually considered to add to the attractiveness of the flowers, but in the case of Babiana we are inclined to consider it a vestigial trait in view of its rarity among members of the guild. Most of the species of Babiana section Babiana, to which the Prosoeca-pollinated species belong, have strongly scented flowers and are bee pollinated (e.g., B. odorata, B. scabrifolia, unpublished observations).
Nectars of moderate sugar concentration and typically sucrose-rich to sucrose-dominant
seem to be characteristic of plants pollinated by active insects such as bees of the family
Anthophoridae and by long-tongued flies (Goldblatt et al., 1995) although not of plants
pollinated by other Diptera (Baker & Baker, 1983, 1990) such as Calliphoridae, Muscidae, and
Tachinidae. The nectar sugar concentrations of flowers pollinated by Prosoeca and other
Nemestrinidae Table 2 are typically somewhat lower than those of bee-pollinated flowers. This may be related to the difficulty of sucking up liquids of higher viscosity, as is the case in long-tongued butterflies (Johnson & Bond, 1994). The low sucrose to hexose ratios in the two species of Pelargonium Table 2 contrasts with the pattern in Iridaceae belonging to the guild. Species of Pelargonium are visited as avidly as any of the Iridaceae. Indeed, on the basis of the frequency of visits, Pelargonium incrassatum appears to be one of the most important nectar sources for P. peringueyi. This leads us to conclude that nectar sugar composition is not a significant factor in the P. peringueyi pollination guild.
Not all species with long perianth tubes and dark purple to crimson flowers belong to the
guild. In some species that have flowers apparently conforming to the guild the lower part of the
perianth tube is so narrow as to effectively prevent the penetration of an insect's tongue. In these
cases nectar is forced into the upper part of the tube where it is accessible to a variety of insects. Long-tongued flies visiting these species may obtain nectar but will not come into contact with the pollen and stigmatic surfaces. This situation has already been described for the long-tubed
species Lapeirousia montana which superficially appears to be a member of a long-tongued insect pollination guild (Goldblatt et al., 1995). The tube is up to 43-55 mm long but is so narrow in the lower half that the nectar is forced upward and lies only 10-15 mm from the mouth of the tube.
The flowers are visited by a variety of bees, Lepidoptera and bombyliid flies, all of which may
accomplish pollen transfer. Likewise, we have found that whereas B. sambucina var.
longibracteata and var. unguiculata do conform to the guild except in their scented flowers, var. sambucina does not. Although the perianth tubes of all three varieties are 30-55 mm long, in var. sambucina the lower part of the tube is effectively blocked and nectar is forced into the top of the cylindrical part of the tube where it is accessible to a variety of insects. The situation in var. sambucina is no surprise because its distribution range is mostly outside that of Prosoeca peringueyi or P. sp. A comparable situation exists in Babiana dregei which has a tube 50-65 mm long, thus longer than the mouthparts of P. peringueyi. In this plant too, the tube is narrowed in the lower 20-30 mm rendering the functional tube length much shorter than the external length, and the nectar is thus available to individuals of P. peringueyi with tongues of moderate length.
Color of the perianth and of the nectar guides may also be misleading. Babiana pauciflora
G. Lewis has purple flowers with tubes 35-45 mm long, but a strong fragrance, bright yellow
nectar guides and a flowering period of June appear to exclude the species from the guild. Babiana
attenuata G. Lewis and B. truncata G. Lewis also have perianth tubes in the 30-45 mm long range but the blue or mauve flowers have yellow nectar guides and, in the case of B. attenuata, fragrant flowers. We do not regard them as guild members.
The Prosoeca peringueyi Pollination Guild
The Prosoeca peringueyi-P. sp. pollination guild (hereafter referred to simply as the P.
peringueyi pollination guild) is unique in its combination of a long floral tube combined with a
distinctive perianth coloring, and its restricted occurrence in so small a part of southern Africa is
striking. While a long-tongued fly pollination syndrome has been described for the Cape Flora
(Whitehead et al., 1987), this covered only plants with cream to pink flowers with dark nectar
guides. The P. peringueyi guild thus deviates significantly from other long-tongued fly pollination systems. There are at least two other guilds of plant species adapted for long-tongued fly pollination in southern Africa. Late spring and early summer flowering species with white to
cream flowers with red nectar guides (e.g., Lapeirousia anceps (L.f.) Ker-Gawl., L. fabricii (Delaroche) Ker-Gawl.) are pollinated by either Moegistorhynchus longirostris (Nemestrinidae), or Philoliche gulosa, or P. rostrata (Tabanidae) (Vogel, 1954; Goldblatt et al., 1995). Summer- and autumn-flowering species with long-tubed, blue, pink, or white flowers (e.g., Nivenia stenosiphon Goldblatt (Iridaceae, Goldblatt & Bernhardt, 1990), Disa oreophila H. Bol. (Orchidaceae, personal observation), Gladiolus microcarpus G. Lewis (Iridaceae, personal observation), and Zaluzianskya microsiphon (O. Kuntze) K. Schum. (Scrophulariaceae, personal observation) are pollinated by Prosoeca ganglbaueri over a wide portion of southern Africa. Both these syndromes are quite distinct from the Prosoeca peringueyi pollination system in the flower color, plant and insect distribution, pollinator identity, and flowering time.
Tongue length is surprisingly variable in Prosoeca peringueyi and somewhat variable in P. sp. The variation corresponds locally to floral tube length of their nectar plants. In the
Kamiesberg Mountains where Hesperantha latifolia, Babiana dregei, and Pelargonium incrassatum are major nectar sources and effective tube length is 20-40 mm, P. peringueyi has a tongue of 20-25 mm long. In other places in Namaqualand where L. silenoides and B. pubescens have floral tubes in excess of 50 mm, individuals of P. peringueyi have tongues 35-40 mm long. In the Olifants River Valley where L. pyramidalis and L. jacquinii are major nectar sources for P. peringueyi and have tubes 35-45 mm long, flies caught on these species have tongues 30-33 mm long. Clearly local variation in tube length in the flowers of food plants is tracked by the pollinator.
The disparity between the length of floral tube in plants with the Lapeirousia silenoides-
type flower and the shorter length of the mouthparts of the sole pollinators is quite easy to
explain (Goldblatt et al., 1995). Records of nectar secretion show that these flowers secrete
ample amounts of fluid for insect-pollinated flowers and it is most unlikely that dominant
pollinators are ever forced to extend their mouthparts to the base of the tube unless all the nectar
has been removed by earlier foragers. More importantly, Darwin (1877) hypothesized that
successful pollination of spurred orchids occurred when orchids evolved floral spurs slightly
longer than the tongues of their pollinators, ensuring maximum contact between the insect's head
and the orchid's column by forcing the insect to ram its head down the floral throat. This has
since been shown experimentally by Nilsson (1988). As in the nectariferous orchids, species with
the L. silenoides-type flower force their pollinators to make maximum contact with the anthers
and stigmatic surfaces that block or at least encircle the entrance to the floral tube.
The combined geographical ranges of all the species pollinated primarily by Prosoeca
peringueyi and P. sp. Fig. 11 is rather greater than the ranges so far recorded for the two fly species Fig. 12 . To the north in southern Namibia P. peringueyi may simply not have been collected yet. To the south, essentially below the 33rd parallel and west of the 19th north-south parallel, the guild is represented mainly by L. jacquinii. This species is known to be autogamous (Goldblatt et al., 1995), although it is actively pollinated by both species of Prosoeca within their ranges. Presumably autogamy has enabled it to extend its range outside that of its facultative pollinators. The only other species that occurs outside the ranges of the two flies is Pelargonium
magenteum. The reasons for its wider distribution to the east of the range of P. eringueyi and P. sp. are unknown.
The potential for pollen contamination from other members of the same guild is greatly
increased for plant species that share a single specialized pollinator species. One strategy to
enhance segregated gene flow is differential placement of pollen on the insect body. This strategy
is developed in Orchidaceae (Dressler, 1968a, b; Manning & Linder, 1992), Scrophulariaceae
(Steiner & Whitehead, 1988, 1990), and various other families (Grant, 1994). In flowers of
species belonging to the P. peringueyi pollination guild four mutually exclusive sites have been identified. In this guild the contamination coefficient is never below 0.5, suggesting that there is a threshold of pollination efficiency which determines the number of species that can effectively use the same loading site. In view of mechanical constraints on flower architecture in the genera involved, there appears to be a definite limit to the number of species that can enter the guild at any locality. Pollination contamination may thus be a significant factor in influencing species packing in specialist systems. Locally, the niche offered by pollination by Prosoeca peringueyi or P. sp. appears to become saturated by the presence of more than three co-flowering species. The presence of four similarly adapted species is rare.
Differential placement of pollen on an insect's body demands precise orientation of the
pollinator relative to anther position. Floral zygomorphy facilitates this and we suggest that this
is an important factor in favoring genera with predominantly zygomorphic flowers.
The Lapeirousia silenoides pollination syndrome appears to have evolved in five different
lineages in Lapeirousia subgenus Lapeirousia, a taxon that includes just 21 species (Goldblatt et al., 1995). A comparably polyphyletic evolution of the syndrome appears to have occurred in Babiana and in Pelargonium as species with this syndrome in these genera are taxonomically isolated.
Origin of the Prosoeca peringueyi pollination guild
The Prosoeca peringueyi pollination guild involves at least six genera of plants in two families and two species of Prosoeca (Nemestrinidae). Although the plant species belonging to the guild and their pollinators are restricted to western southern Africa, all of the genera, both plant and insect, extend beyond the range of the guild. Other species of Prosoeca, both long- and short-tongued, visit flowers of various colors including white, cream, pink, lilac, blue and yellow (Johnson, 1992; Goldblatt et al., 1995; pers. obs.). P. peringueyi has been observed visiting species outside the guild for nectar that is pale lilac or pink and green in color. In addition the northern taxa of the guild are lilac or pale mauve and some have been confirmed to be pollinated by Prosoeca peringueyi. The original determinant for the characteristic crimson or purple color in the guild was thus apparently not directed by the innate preference of P. peringueyi for that particular flower color, and was in consequence presumably plant-directed.
Comparative studies of each of the guild genera suggests that pollination by long-tongued
flies is apomorphic. Of the guild members only Babiana has flowers in which dark blue or violet color is the plesiomorphic condition. In addition, even short-tubed species of this genus secrete fairly large amounts of nectar. Also Babiana has flowers with a wide gullet which makes access to the pollinator mouthparts easier than the narrow-tubed flowers in the other guild genera, for example Lapeirousia and Pelargonium. These factors suggest to us that the first steps in the development of the Prosoeca peringueyi pollination system were through the genus Babiana. Significantly, both Babiana and Lapeirousia, the two genera which individually have the most species in the guild and together comprise 54% of the guild, are largely developed in arid habitats. It may be that the origin of the syndrome in Babiana and its subsequent development in species of Lapeirousia was a consequence of higher species richness in these genera in the western part of southern Africa. The predominant dark purple flower color in Babiana would explain the characteristic floral coloring in the L. silenoides pollination syndrome, unknown in other Prosoeca pollination guilds. Subsequently species in other genera could be expected to enter the guild in response to the reproductive benefits derived from these pollinators.
The advantages to the plant species of a dedicated pollinator are obvious and include increased pollination success and decreased pollen contamination and loss. Pollination success in one population of Lapeirousia pyramidalis subsp. regalis that we investigated was 45% (SD +
25%; n = 23). To the pollinator, however, the energetic rewards of floral specialization are
important. For large active insects that hover while foraging the energy demands are likely to be
high. Long-tongued nemestrinids are capable of feeding from short-tubed flowers but these are
smaller and hold far less nectar than that typically present in the long-tubed flowers in the guild.
In addition, the flies are in competition with other insects which can obtain the nectar in short-
tubed flowers. Long-tubed flowers can contain large amounts of nectar which cannot be collected
by short-tongued insects. They are therefore an attractive energy source for insects able to
We speculate that there is reduced pressure to darken the flower color in the northern members of the guild. This more arid region suports both fewer plants and fewer pollinators and more facultatively generalist pollination system might be favored. This is borne out by the greaty extended flowering periods of Pelargonium cortusifolium and P. crassicaule, which although peaking in August and September are prolonged far beyond the flight period of Prosoeca peringueyi. During this time they are presumably visited by other insects, perhaps bees and bee flies. Significantly, both species have short anthers which will contact visitors of a range of shapes and sizes. Possibly a threshhold diversity is necessary before selection pressures become strong enough to favor such specialist pollination systems.
Evolutionary implications of the Prosoeca peringueyi pollination guild
The Prosoeca peringueyi pollination guild appears as distinct as other highly specific
pollination systems in the African subcontinent, including those involving sunbirds, I>Nectarinia
species (Rebelo, 1987), oil-collecting bees in the family Melittidae (Steiner & Whitehead, 1990,
1991a), resin-collecting bees in the family Megachilidae (Steiner & Whitehead, 1991b; Armbruster & Steiner, 1992), other guilds of long-tongued flies (Goldblatt et al., 1995; Johnson & Steiner, 1995; Manning & Goldblatt, 1995), and the butterfly, Aeropetes tulbaghia (Johnson & Bond, 1994). Where such systems occur they contribute to the particular floral characteristics of various plant communities.
The recognition of the Prosoeca peringueyi pollination syndrome and the way it functions
is the key to understanding the presence of a series of species with unusual dark purple or
crimson flowers with long floral tubes in the flora of coastal and near interior western southern
Africa. The ecological niche presented by these two flies is so specific that it will allow fly-
pollinated and nonfly-pollinated members of the same genus flowering at more or less the same
time to coexist with little or no hybridization. The diversity of pollinatioin systems there is one
of the reasons why some 30 species of Babiana (50 % of the total species) and 19 species of
Lapeirousia (48% of the total) co-occur in this area.
Like the members of the Aeorpetes tulbaghia pollination guild (Jonhson & Bond, 1994),
there is a very strong similarity in floral morphology between members of the Prosoeca
peringueyi guild. This suggests strong selection for floral conformity. This may be a characteristic of guilds in which the pollinator is an insect that is not flower constant.
Prosoeca peringueyi and P. sp. nov. may be regarded keystone species. Such species are
defined operationally as those that, by their effective disappearance from a system result
(directly or indirectly) in the virtual disappearance of several other species. The extinction of
either P. peringueyi or P. sp. nov., but especially the former, will result in significant decreases in seed set in many of the species in the P. peringueyi guild, prevent outcrossing, and may lead to their ultimate extinction.
<< BACK | TOP | NEXT >>