CHAMPION POLLINATORSby Peter Goldblatt, Missouri Botanical Garden, St. Louis, Missouri and

John C. Manning, Compton Herbarium, National Botanical Institute, Kirstenbosch

Southern Africa abounds inplants that have flowers withlong, slender floral tubes and

in an unusual range of animalswith long mouthparts that feed onthe sugar-rich nectar inthe base of the tubes. Theanimals that feed on thenectar, all potential polli-nators of these plants,include sunbirds withlong bills and tongues anda variety of differentgroups of insects withelongated mouthparts.Among these insects arebutterflies and moths,flies, including bee flies(Bombyliidae) andacrocerid flies(Acroceridae), and somebees. These flies and beeshave mouth parts up to8-12 mm long and visitflowers with floral tubesof moderate length,usually in the 10-15 mmrange. Such pollinatorsare not uncommon andthey occur all over theworld.

Southern Africa is alsohome to a special group oflong-proboscid polli-nators: flies with mouth-parts mostly exceeding20 mm, one of only a fewkinds of organismscapable of taking nectarfrom plants with super-long, narrow floral tubes thatexceed 20 mm in length. Such fliesare remarkable among pollinatorsfor their needle-like mouthpartsare longer than their bodies,sometimes up to three times aslong. Many of the plants that theyvisit to forage for nectar haveslender floral tubes exceeding30 mm in length and sometimesup to 70 mm. Clearly, these long-proboscid flies and the plantswith which they are associatedhave co-evolved and form a closeecological mutualism.

Champion among long-proboscid flies is the Cape west

coast species Moegistorhynchuslongirostris which has a tongue aslong as 70 to 80 mm. Several otherflies have tongues up to 40 mm,still amazing when one considers

Philoliche gulosa and P rostrata,and several nemestrinids,including six species of the genusProsoeca, one of Stenobasipteron,and at least two species of the

genus Moegistorhynchus.These flies with super-long mouthparts occurthroughout southernAfrica, their collectiverange extending fromsouthern Namibia to theCape Peninsula in thewest, along themountains of thesouthern Cape andthrough Lesotho andKwaZulu-Natal intoMpumalanga, NorthernProvince and Swaziland.None of these flies isknown from tropicalAfrica, which also seemsto lack plants associatedwith long-proboscid flypollination.

Our interest in long-proboscid fly pollinationbegan with an investi-gation of the taxonomy,biology and evolution ofthe genus Lapeirousia,and subsequently that ofthe genus Gladiolus,both members of theplant family Iridaceae.These genera exhibitmajor radiation andspeciation in southernAfrica, and in both there

has been significant adaptiveradiation for a variety of polli-nation systems. Many species ofboth genera have flowers withparticularly long tubes and secreterelatively large amounts of nectarthat is rich in sucrose. Our workon Lapeirousia gave us the firstindications of the presence ofguilds (see accompanying box) ofplant species using particular long-proboscid fly species to achievetheir pollination.

In Namaqualand, in the interiorwest coast of South Africa, wediscovered that the carmine-redflowers of Lapeirousia silenoides

Active in the spring in the Western Cape and Namaqualand,Prosoeca peringueyi is the sole pollinator of a range of wildflowers with long floral tubes and bright red to purple petals, suchas Lapeirousia silenoides. Photo: J. Manning.

that the fly has no way to retract itsprodigious appendage, but must flywith it extended forward or tuckedloosely below its body. Flies withthese super-long mouthpartsinclude some twelve knownspecies in two fly families, theNemestrinidae (tangle-veined flies- a fairly uncommon animal) andthe Tabanidae (horseflies - short-proboscid species which bothercattle and humans alike as thefemales painfully bite the skin toobtain the blood meal they requireto complete their life cycle).The cast of long-proboscid fliesincludes the two horseflies,

(Springbok painted petals) werepollinated exclusively by onespecies of fly, Prosoeca peringueyi.At particular times of the day andin sunny, warm weather this largefly,which has mouthparts 30 to 40mm long, could be found foragingfor nectar on the flowers of1. silenoides. Its violet-colouredpollen deposited all over the faceof the flies can often be seen withthe naked eye. We soon noticedthat Properingueyi also visitedsimilarly coloured and shapedflowers of other plantsthat grew nearby. Oftenthe flies would visit thebrilliant magenta flowersof Pelargonium incras-satum, a species of theGeraniaceae fairlycommon in Namaqualand.Flies would also visit theflowers of Babianaspecies (Iridaceae) withlong-tubed, dark blue-violet flowers andHesperantha latifolia withdeep red flowers. Itgradually dawned on usthat a wide range ofspring-blooming, intensered, purple or bluishflowers with palecoloured markings,complete absence of scent,and a long floral tubecontaining sugary nectarwere visited solely, butindiscriminately byProsoeca peringueyi.

This pattern was alsoevident to the south in theOlifant's River andBidouw valleys whereProsoeca peringueyicould be found visitingthe flowers of Lapeirousiajacquinii, 1. violacea or1. pyramidalis, differentspecies of Babiana, Sparaxisvariegata (Iridaceae), andPelargonium magenteum, allhaving violet to dark reddishflowers with pale markings and along, nectariferous floral tube incommon. On the BokkeveldPlateau near Nieuwoudtville in theNorthern Cape Province we founda local variation of what wasbecoming a familiar pattern.Purple-flowered and long-tubedLapeirousia oreogena, 1. jacquinii,Babiana framesii and B. sambucinawere being visited by a secondspecies of Prosoeca as yet not

named. And on the Hantamsbergnear Calvinia, in the westernKaroo, Babiana flabellifolia and -Romulea hantamensis, which alsohave long-tubed, purple flowers,were visited exclusively by thisspecies of Prosoeca. The puzzle asto why long-tubed purple ormagenta flowers are so commononly north of the Olifant's Rivervalley was explained. They areadapted for pollination by thesetwo flies which are found onlynorth of Clanwilliam.

probable pollinator of anotherLapeirousia, 1. anceps (by theentomologist, A. J. Hesse, andreported by the German biologist,Stefan Vogel in 1954). We laterfound this fly pollinating not only1. anceps, but several plants withsimilarly shaped and colouredflowers with long floral tubes. NearYzerfontein on the Cape west coastM. longirostris also pollinatesPelargonium stipulaceum, andother species of Iridaceae withcream flowers, including Babiana

tubulosa, Geissorhizaexscapa, Gladiolusangustus and Ixia panic-ulata. The cream-flowered orchid, Disadraconis is also part ofthis guild. In theOlifant's River valleynear Clanwilliam weagain found this fly, thistime pollinating both1. anceps and 1. fabricii,as well as Geissorhizaexscapa and Tritoniacrispa.

A second fly, thetabanid Philolichegulosa, shares this sameseries of plants, butmainly outside the rangeof M. longirostris.Ph. gulosa has slightlyshorter mouthparts thanM. longirostris andpresumably is at a slightdisadvantage incompeting withM. longirostris for nectar.Their ranges overlaponly to a small extent,however, and bothspecies are able to utilizethe same set of plants. Itis notable that the lengthof the floral tube usuallytracks very closely the

length of the visiting insect'smouthparts. The floral tube in theplant species is much longer inpopulations visited by M. longi-rostris than in those visited only byPh. gulosa, graphically illustratingthis phenomenon.

In the mountains of the WesternCape, including the CapePeninsula, we have seen some ofthe same plant species that wefound being visited by Moegistor-hynchus longirostris and/orPhiloliche gulosa, as well as otherswith a very similar appearance,being visited by other flies

With the longest proboscis of any fly, Moegistorhynchuslongirostris is the only insect able to reach into the base of thefloral tubes of plants like Lapeirousia anceps to feed on the amplenectar located there. Moegistorhynchus appears to be attractedparticularly to cream and pink flowers with darker markings.Photo: J. Manning.

A second group of Namaqualandand Cape west coast plant speciesformed another guild with flowersadapted for pollination by differentspecies of long-proboscid flies.They also have long-tubed flowerswith ample amounts of sweetnectar, but the flowers are white topale pink and have red to purplemarkings that seem to function as avisual signal. We first saw the.nemestrinid fly, Moegistorhynchuslongirostris on a species ofLapeirousia, 1. fabricii inNamaqualand. This fly had alreadybeen recorded as a visitor and

including P. rostrata and Prosoeca nitidula. Theseplants evidently belong to a greater MoegistorhynchusPhiloliche pollination guild in the southern Africanwinter-rainfall zone that involves cream to pink-flowered plant species that flower in the spring andearly summer, September to November or December.

All these long-proboscid flies have'a randomforaging behaviour. They will visit flowers of thedifferent members of their respective guilds quiteindiscriminately, and thus carry pollen of severalspecies on their bodies, a feature easy to determinewith the naked eye by the heavy deposits of differ-ently coloured pollen, and by microscopic analysis ofpollen removed from their bodies. The problem ofpollen clogging (see accompanying box) however, doesnot occur. This is because at sites where plant speciesuse a particular fly for pollination each species hasanthers so positioned that pollen is consistentlydeposited on a different part of the fly's body. Forexample, at sites in Namaqualand, pollen ofLapeirousia silenoides is deposited on the frons(front part of the head) of Prosoeca peringueyi, that ofPelargonium on the lower thorax and of Babianadregei or B. pubescens on the upper thorax. A similarpattern is repeated in guilds using Moegistorhynchusand Philoliche.

Our studies of the biology of Gladiolus in southernAfrica has led us to identify two more completelyseparate guilds of plants species using quite differentlong-proboscid fly species to achieve their pollination.In the Drakensberg of KwaZulu-Natal and Lesotho thefly, Prosoeca ganglbaueri, with mouth parts 20 to35 mm long, is active in the late summer and earlyautumn and appears to visit, indiscriminately, long-tubed cream or pink flowers, usually with darker -.,

Moegistorhynchus Iongirostris about to insert its elongatedproboscis into the slender tube of lxia paniculata. Photo: J, Manning.

Pollination by flies with extremely long mouthparts hasbeen called rhinomyophily, a tough name to rem!ilmber. "The term includes pollination by all flies with slender,sucking mouthparts, simply to distinguish it from themore general myophily, or pollination by flies withshort mouthparts. Sapromyophily, pollination by flieswith lapping mouthparts is a completely differentpollination syndrome, typically associated with rank-smelling flowers without tubes and readily accessiblenectar. Long-proboscid flies feed primarily on nectar andthey may visit almost any flower that offers this reward.However, it is only from flowers with very long tubesthat an adequate supply of nectar is assured. Otherinsects are prevented from exploiting this resourcebecause their mouthparts are simply too short. Nectar inshort-tubed flowers is available to a host of insects and isreadily depleted by common insect visitors includinghoney-bees, butterflies and wasps. It is then in the bestinterests of the energetic needs of insects with longermouthparts to confine their foraging to flowers with longtubes because they benefit from the lack of competitionfor this food resource. Long-tubed flowers produce agenerous food supply and store it in sites that are onlyaccessible to a few highly specialized visitors.

Only those plant species with very long-tubed flowersactually benefit from the visits of the flies because asthese insects forage for nectar they brush against theanthers. Pollen is then dusted onto their bodies and iscarried to another flower where it can, in turn, be

brushed against the stigma, so accomplishing cross-pollination, essential for the life cycle of the plant to becompleted. Insects often learn rapidly to recognizeimportant sources of food and the form and colour of aparticularly valuable source of food becomes imprintedon their memory. A consequence of this is that plantswith flowers closely resembling the primary nectarsource or sources will also be visited with somefrequency, particularly if they also offer a reward.Thus it is typical of long-proboscid fly pollinationsystems for the several plant species in a particular areathat use the same insect for their pollination to haveflowers of similar colour, and sometimes shape andoverall appearance. These plants are said to form a guild,that is, a group of organisms using a particular resourcein a similar way, in this case a highly specific pollinator.

Long-proboscid flies have a random foragingbehaviour, visiting flowers of the different members oftheir respective guilds quite indiscriminately, and thuscarrying pollen of several species on their bodies. Theproblem of pollen clogging, that is the deposition offoreign pollen on the stigma of a flower, thus preventingpollen of the right species reaching the stigmatic surfaceand eventually effecting fertilization, however, does notoccur. This is because at sites where plant species use aparticular fly for pollination each species has anthers sopositioned that pollen is consistently deposited on adifferent part of the fly's body.

markings, of Gladiolus microcarpus, Hesperanthascopulosa, Zaluzianskya microsiphon(Scrophulariaceae) and species of Orchidaceaeincluding Disa oreophila. At sites along theMpumalanga Escarpment Pr. ganglbaueri or two otherlong-proboscid flies, Pr. robusta and Stenobasipteronwiedmannii, visit and appear to pollinate several moreplants with long floral tubes including several speciesof Gladiolus and Watsonia (Iridaceae), some orchids,and a few members of the mint family, includingOrthosiphon species. Another plant that appears touse a long-proboscid fly for its pollination is the large,pink-flowered Nerine angustifolia (Amaryllidaceae)which, although it doesn't have a long floral tube, hasstamens over 20 mm long that brush the underside ofa long-proboscid fly as it feeds on nectar at the base ofthe tepals. Other species at these sites with similarlycoloured, long-tubed flowers include the orchid, Disaamoena, which also appears to be adapted for polli-nation by long-proboscid flies. At all these sites too,pollen is deposited by each member of the guild on adifferent part of the fly's body. The problem of pollenclogging of the stigma thus appears to be a primaryconsideration in long-proboscid fly pollination and afactor limiting the local occurrence of two or morespecies that use the same sites of pollen depositionon a fly.

So far some sixty plant species have been identifiedthat appear to be pollinated exclusively by one or nomore than two species of long-proboscid fly. A reviewof southern African plant species with similar flowersto those now known to be pollinated by long-proboscid flies brings the list of species potentiallypollinated by these flies to at least 150 species, andthere are certainly more. This makes pollination bynemestrinid and tabanid flies with mouthp~rtsexceeding 15 mm one of the more significant polli-nation systems in the subcontinent, and one of >"particular importance in families such as the.· .Geraniaceae and lridaceae. It is surprising then, thatuntil recently this pollination system was virtuallyunknown. ;

A disturbing aspect of long-proboscid fly polli-nation is that it is a highly specialized system.Numerous plant species depend on just one or twoorganisms for their pollination and thus for theirultimate survival. Such organisms have been termedkeystone species for good reason. The conservation ofplants depending on long-proboscid flies for theirpollination depends ultimately on the survival ofthese flies. Unfortunately the survival of the flies isnot assured. Tabanid flies have the larval stages oftheir life cycle in wetlands, not always close to thesites where the plants that they feed upon grow. Thefemales are also believed to require a blood meal froma large mammal before she can lay eggs. Farming andurban development activities can damage and destroywetlands and reduce habitat for the mammals onwhich tabanid flies feed. No wonder then that at somesites where species depend on Philoliche for theirpollination one can find populations of plants that nolonger produce seeds. Their specialized pollinator islocally extinct.

Unlike the tabanid flies, the life cycles of thesouthern African nemestrinids, Moegistorhynchus,Prosoeca, and Stenobasipteron are completely

Like Moegistorhynchus, the eastern escarpment fly,Stenobasipteron wiedmannii is attracted to pale pink flowers.One is seen here pollinating the rare Gladiolus macneilii.Photo: ]. Manning.

unknown. The few nemestrinid flies whose life cyclesare known have larvae parasitic on other insectsincluding locusts. With such complex life cyclesnemestrinid flies are probably especially vulnerable toenvironmental disturbance and loss of habitat. Againwe have found populations of plants with flowersadapted for pollination by both Moegistorhynchus andProsoeca that set no seed at all in years when we havestudied them. It is not yet known whether theemergence of the flies from their pupae had failed tocoincide with the flowering time of the floral guildsthey feed upon or whether they were extinct locally.Nevertheless, we are convinced that in the long term,all these specialized long-proboscid flies will becomeeliminated from much of their original ranges, andthat several of the plant species which depend onthese flies for pollination will soon be eliminatedlocally because of the absence of their pollinators.Some plant species with small ranges will probablyface total extinction.

It is extremely important, then, that we learn aboutthe life histories of such specialized pollinators.Until this information is available, reserves for theconservation of plants with specialized pollinationsystems will have to be made as large as possible andmust include a variety of habitats so that the complexwebs that allow the lives of these organisms to besustained will be preserved. ®