pollination by fungus gnats (diptera: mycetophilidae) and self … · 2008. 10. 10. · ca. 7 mm...

Pollination by fungus gnats (Diptera: Mycetophilidae)

and self-recognition sites in Tolmiea menziesii (Saxifragaceae)

P. Goldblatt1, P. Bernhardt

2, P. Vogan

2, and J. C. Manning

3

1B. A. Krukoff Curator of African Botany, Missouri Botanical Garden, St. Louis, Missouri, USA2Department of Biology, St. Louis University, St. Louis, Missouri, USA3National Botanical Institute, Claremont, South Africa

Received December 16, 2002; accepted August 8, 2003Published online: February 3, 2004� Springer-Verlag 2004

Abstract. Observations of insects visiting stronglyprotandrous flowers of Tolmiea menziesii (Saxi-fragaceae) show that Gnoriste megarrhina(Mycetophilidae), a large fungus gnat with a bodyca. 7 mm long and a proboscis ca. 6.5 mm long, isthe primary pollinator. Pollen is transferred toventral portions of the gnat’s thorax while itprobes for an unusually dilute nectar (8.3% to12% sucrose equivalents) produced at the base ofthe floral tube. Grains are transferred to the tworeceptive stigmas on each pistil when it foragesfor nectar on a female phase flower. Pollen loadanalyses indicate that Tolmiea menziesii may bethe only nectar source visited by the gnat while itis in bloom. While hoverflies (Syrphus spp.) andbumblebees (Bombus spp.) also visit T. menziesiiflowers, they do not forage for nectar and usuallybehave as pollen robbers visiting only male phaseflowers after earlier visits by G. megarrhina. Whenthe modes of floral presentation of T. menziesiiare compared to other species pollinated primarilyby micro-dipterans, and mycetophilids in particu-lar, there appears to be a floral convergence insome spring-flowering, perennial herbs of moist,evergreen woodlands and forests. Analyses ofopen (gnat-pollinated) vs. hand-pollinated pistilsshow that manual cross-pollinations areslightly more successful than open pollinations.Fluorescence microscopy confirms two sites of

self-recognition and rejection within the pistil:following controlled self-pollination experimentsalmost half of the pollen tubes produced eitherstop growing once they reach the top of the ovaryand/or grow horizontally. The remaining tubesproduced by self-pollination penetrate ovulesbut seed is never set. This double mode ofself-recognition and rejection is similar to theone described for Heuchera micrantha var. diver-sifolia (Saxifragaceae) but may also confirm anearlier report of post-zygotic rejection made forTolmiea.

Key words: Tolmiea, Saxifragaceae, pollination,nectar, protandry, pollen tube growth, Gnoriste,fungus gnats, Mycetophilidae.

Species of the family Saxifragaceae have servedas model systems for floral biologists since theearly 20th century. For example, they havebeen used repeatedly to determine the role(s)of dicliny, dichogamy and distyly as self-isolation mechanisms (Knuth 1908, Faegriand van der Pijl 1979, Ornduff 1975) whilePrimack (1985) used them as prime examplesof the positive correlation between reductionsin the number of organs per flower and anincreasing floral life span.

Plant Syst. Evol. 244: 55–67 (2004)DOI 10.1007/s00606-003-0067-1

Interpretations of the pollination ecologyof members of the Saxifragaceae have changedsince the 19th century when naturalists listedall insect visitors found on their flowers asevidence of generalist pollination systems.Knuth (1908) reinterpreted some of this infor-mation to suggest that certain species werepollinated primarily by true flies (Diptera),while others depended on butterflies andmoths (Lepidoptera). More recent evidenceof specialized pollination systems in the familyinclude Ornduff (1971, 1975), who noted thatpopulations of Jepsonia heterandra were polli-nated by a combination of small bees andhover flies, and Pellmyr et al. (1996) whoshowed that moths of the genus Greya areprimary pollinators of Lithophragma parviflorumNutt. ex Torr. & A. Gray, Heuchera cylindricaDougl. and Mitella stauropetala Piper. The roleof Greya politella shifts between pollinator andseed parasite (Thompson and Cunningham2002).

Pollen-pistil interactions within the familyalso suggest unusual variability. While hetero-morphic and homomorphic species may bestrongly self-incompatibile, the rejection ofself-pollination is gametophytic, as it occursafter pollen tubes penetrate much of the lengthof the style (Rabe and Soltis 1999). Moreimportant, there may be more than one site ofself-recognition and rejection. In Heucheramicrantha var. diversifolia more than 50% ofthe pollen tubes of self-pollinations stop in thestyle while the remainder stop in the ovaryand/or ovule (Rabe and Soltis 1999).

Pollination and self-isolation mechanismswere studied by Weiblen and Brehm (1996) todetermine why the sympatric and highly inter-compatible species, Tellima grandiflora Purshand Tolmiea menziesii (Pursh) Torr. & A.Gray, rarely hybridized. They concluded thatinterspecific isolation was encouraged, in part,by divergent modes of floral presentation withTellima grandiflora pollinated by beetles whileTolmiea is dependent on bumblebees, addi-tional visits from hover flies (Syrphus sp.) andsolitary bees (Halictus sp.). The authors alsostudied self-incompatibility in T. menziesii and

concluded that there was a late rejectionmechanism as pollen tubes based on self-pollination regularly penetrated ovules.

An interpretation of large bee and hover flypollination in Tolmiea menziesii, a species thatextends from coastal central Californiathrough western Oregon, Washington, andBritish Columbia to coastal southern Alaska(Hickman 1993), conflicts with its floral pre-sentation. While large-bodied bees do pollinatesmall flowers, such minute blossoms are usu-ally congested into a head or densely cymoseinflorescence to provide united landing plat-forms (e.g. in Asteraceae), unlike the looseraceme of T. menziesii. Futhermore, this spe-cies fails to produce discernibly sweet floralodors and its flowers are dull colored and havefiliform petals. Specifically, floral presentationin Tolmiea appears to converge with angio-sperms associated with sapromyiophily (sensuFaegri and van der Pijl 1979; Vogel 1954,1979), in which the primary pollen vectors areflies that lay their eggs in fungal fruitingbodies, or dung, or detritus, and/or animalcorpses. The late-acting mode of self-incom-patibility in T. menziesii may also be morecomplicated than the description provided byWeiblen and Brehm (1996). More than one sitein the same pistil may recognize and rejectpollen tubes produced by self-pollination(Rabe and Soltis 1999). An earlier but verylimited study by Sears (1937) suggests thatT. menziesii may reject young embryos pro-duced by self-pollination.

For this reason, we have completed a two-year field and laboratory study, reexaminingthe floral biology of Tolmiea menziesii. Ourresults test the hypotheses that this species ispollinated primarily by bumblebees and hov-erflies and that the majority of ovules containpollen tubes following self-pollination.

Materials and methods

Populations of Tolmiea menziesii (common names:youth-on-age, piggy-back plant) were observed atstudy sites in northwestern Oregon in two fieldseasons in 2001 and 2002 from early May to late

56 P. Goldblatt et al.: Pollination by fungus gnats (Diptera: Mycetophilidae)

June (Table 1). Plants were also examined in thefield for attractants (floral odor, color signals) andrewards (nectar, pollen) as outlined elsewhere(Goldblatt et al. 1995, 1998a, 2000a, 2000b, 2003;Goldblatt and Manning 2000).

Insects observed visiting flowers and contact-ing either the anthers or stigmas were netted orcaptured by hand and killed in jars containingethyl acetate fumes. Captured insects were exam-ined for the presence of pollen first visually, andthen pollen samples were removed from theirbodies using a dissecting needle and mounted onslides in Calberla’s fluid (Ogden et al. 1974). Pollengrains were identified and counted under thecompound microscope after Goldblatt and Bern-hardt (1990).

Nectar volume was measured from 10:00h to12:00h using calibrated 2 lm micropipettes mea-suring 55 mm in length. Micropipettes were used tocarefully probe the base of the floral tube withoutdamaging hypanthium tissue to avoid samplecontamination. The percentage of sucrose equiva-lents in nectar taken from plants growing in thefield was measured in the field using a Bellinghamand Stanley hand-held refractometer (0–50%).

We repeated the hand-pollination experimentsof Weiblen and Brehm (1996) in the laboratoryusing cut stems due to persistent tourist traffic andtrampling at our sites. These flowers, now excludedfrom contact with insects, were either leftuntouched as controls to determine whether mechani-cal self-pollination (autogamy) occurs, or pollen wasapplied by hand to exposed stigmas followingexpansion and separation of the two styles on eachpistil. The two stigmas/flower received either pollenfrom a flower on the same inflorescence (self-pollination, geitonogamous cross) or pollen from aflower of a different plant, assumed to be a distinctgenotype (cross-pollination, xenogamous cross). Atthe end of 24 hours after hand pollination wholeflower were excised and fixed in 3:1 95% ethanol/glacial acetic acid. After one hour fixed flowers weretransferred to 70% ethanol. To observe pollen tube

penetration of female tissue, pistils were excised fromfloral receptacles and soaked in a 10% NaSo3solution for 24 hours at room temperature thentaken through three baths of de-ionized water beforesquashing, staining and observation under epifluo-rescence following Goldblatt and Bernhardt (1990).Pollen tubes in styles were so closely constricted andpressed together that counts were not taken until thepollen tubes were observed to enter the muchexpanded ovary chambers (locules). To compareresults of pollen tube penetration in open (insect-pollinated controls) vs. bagged and hand-manipu-lated cross- and self-pollinations we ran an ANCO-VA, testing the hypothesis that there is no differencebetween the number and length of pollen tubes in anovary regardless of how the pistils were pollinated.

To compare and contrast rates of natural(open) pollinations, we collected nine flowers withseparate, expanded styles, at random, from nineinflorescences of nine separate genets. These flow-ers were fixed and their pistil contents analyzed asabove.

Plant vouchers are housed at the MissouriBotanical Garden herbarium (MO). Insects wereidentified by Scott Fitzgerald (Mycetophilidae),Oregon State University, Corvallis, Oregon;C. Thompson (Syrphidae) U.S.D.A. (Beltsville);R. Thorp (bumblebees), University of California,Davis. Vouchers are deposited respectively at thesethree institutions.

Results

Vegetative/reproductive morphology and nectar

secretion. Tolmiea menziesii, the only speciesof the genus, occurs in damp shady situations,usually close to streams. Plants are perennialwith a tuft of basal leaves and elongated,simple or branched flowering stems that bearterminal racemes of small, spirally arrangedflowers. The dull purple, brownish or occa-sionally greenish perianth is held horizontally

Table 1. Study sites, all in the U.S.A., and voucher information. Herbarium vouchers are housed at theMissouri Botanical Garden, St. Louis

Location Voucher data

Eagle Creek, Hood River Co, OR Goldblatt 11654Tryon Creek, Lake Oswego, Clackamas Co., OR Goldblatt 11655MacLeay Park, Portland, Multnomah Co., OR Goldblatt 11655A

P. Goldblatt et al.: Pollination by fungus gnats (Diptera: Mycetophilidae) 57

and is bilaterally symmetrical (Fig. 1). Theflowers are evidently odorless to the humannose whether smelled as a freshly picked,bunched bouquet or confined to glass vialsfor 15-30 minutes to allow build up of odor.The calyx, corolla, and three stamens areunited in a cylindrical floral (hypanthium)tube ca. 5 mm long. The tube is slit almost tothe base abaxially where it forms a smallpouch. The inner surfaces of the five palegreenish calyx lobes (sepals) have a radiatingpattern of darker, purple-brown veins. Thesepals are unequal, with the adaxial threelarger and obtuse, while the abaxial two areshorter, acute and directed forward. The fourcorolla lobes (petals) alternate with the sepalsbut there is no petal between the abaxial pairof sepals (where the floral tube is slit).

Nectar droplets are visible by 09:00h asglistening spots on the proximal inner walls ofthe perianth tube around the base of the ovary.Nectar quantity per individual flower ranges

from >0.1 to 0.2 ll (Table 2) and mean nectarconcentration ranges from 8.3 to 12.0%sucrose equivalents in two populations sam-pled on three separate days. Three stamens,opposite the upper (adaxial) sepals are unilat-eral, and arch downward to lie horizontally,oriented just above the abaxial pair of sepals.They extend for ca. 2 mm beyond the floraltube and spreading tepals and bear conspicu-ous, bright orange pollen (a few genotypeshave dull yellow pollen). The bilocular, supe-rior ovary is contained entirely within thecorolla tube at anthesis. Bilobed above, theovary consists of two carpels which taperdistally into twin styles with terminal stigmas(Fig. 1).

The fruit is a two-chambered greenish, butultimately light brown, thin-walled, translu-cent capsule that dehisces prior to the matu-ration of the seeds (green, unripe seeds areclearly visible through the open slit at the apexof each capsule). Mature black seeds spillthough the open slits of the capsules as early astwo weeks after fertilization. A mean of 90.4seeds per capsule was counted (n¼ 7 flowersfrom seven separate inflorescences); min–max¼ 42–155).

Floral phenology and life-span. The flow-ering season of all populations of Tolmieastudied begins in early to mid May and lastsuntil about the end of June. Flowers are long-lived, lasting six to eight days but are protan-drous. That is, stamens are visible the daybefore anthers dehisce as soon as the calyxlobes separate. Once dehisced, bright orangepollen is clearly visible on the inner surfaces ofthe anthers. Flowers remain in this exclusivelymale phase for three to four days (n¼ 5 flowers

Fig. 1. Section of a flower of Tolmiea menziesii. Scalebar = 1 mm. Drawn from preserved material andphotographs by J. C. Manning

Table 2. Nectar properties in Tolmiea menziesii populations. Sample size indicates number of flowers (ofdifferent individuals) examined at study site or date

Study site Sample size (n) Nectar

volume ll (n) % sugar (+SD)

Eagle Creek 5/20/02 5 <0.1–1.84 12.0 (3.7)Eagle Creek 5/31/02 5 <0.1–0.21 10.9 (2.6)Tryon Creek 5/24/02 5 <0.1–0.22 8.3 (2.0)


observed). At this time the two styles lieparallel and appressed to one another andare 2–3 mm long, as measured from the top ofthe ovary chamber to the stigma, which lies atthe apex of the style.

On the fourth day of anthesis the stylesdiverge from one another in the vertical planeand elongate, reaching twice their originallength, ca. 6 mm long. They emerge from themouth of the perianth tube and come to lie inthe same plane as the stamens with the stigmasextended almost to the level of the anthers.Stigma to stigma distance is ca. 3 mm. By thetime the styles and stigmas are visible outsidethe tube, all pollen is usually lost from theanthers in the wild as a result of insect activityand the flowers may remain in an exclusivelyfemale phase for another two or three days. Inwild collected plants observations show that bythe time the upper flowers of an inflorescenceare open, the capsular fruits produced by thelower flowers contain ripe seeds that fall fromthe now suberect capsules if they are shaken orinverted.

Insect foraging. Observations at all studysites totaling over 35 hours yielded threegroups of insects visiting flowers of Tolmieamenziesii (Table 3). From 9:00h until noon thedominant floral foragers were small lightbrown gnats, later identified as Gnoristemegarrhina (Mycetophilidae) (Fig. 2), and occa-sionally hoverflies, Syrphus spp. (Syrphidae).Individual gnats were observed to forage fornectar with their much elongated probosces, ca.6.5 mm in length (Table 4). The foragingpatterns of G. megarrhina included enteringflowers, flying from one plant to another, andvisiting other flowers of the same species. Welimited our netting and capture to no more than10 gnat individuals on any day but many moregnats were seen visiting flowers (Table 3) thanwere captured. A gnat typically crawled overthe outside of a Tolmiea flower before movingto the mouth of the perianth. It then graspedthe calyx or corolla lobes with its legs andinserted its proboscis into the gullet-like floraltube (Fig. 3). Once in this position, a gnat’shead was hidden from view but individuals

could be identified by their two wings whichextended beyond the tepals. Individual visitslasted from 20 to 35 seconds (mean 25.13 SD11.29, n¼ 8).

Individuals of G. megarrhina have a bodyca. 7 mm long (measured from the base of theproboscis to the tip of the abdomen) and aslender proboscis ca. 6.5 mm in length (mean6.55 mm SD 0.18; n¼ 10), which individualsused to probe the floral tube (Table 4). Gnatsshowed no interest in pollen consumption. Theclose ‘‘fit’’ of a gnat’s body to the dimensionsof Tolmiea flowers was notable. When grasp-ing the sepals and placing the head at themouth of the tubular flower, the slenderproboscis is almost exactly the combinedlength of the floral tube plus the calyx lobes,ca. 7 mm (Table 4). At this time the ventralsurface of the gnat’s thorax made contact withthe dehisced anthers or the extended stigmas ofa flower.

All captured gnats from our study sitescarried visible loads of orange pollen on theventral parts of their thoraxes and proximalparts of the legs. The pollen was depositedon their bodies passively as they pushed theirbodies into flowers. Under compoundmicroscopy, pollen of Tolmiea menziesii wasremoved from gnat thoraces and was identi-fied easily by a suite of three characters. Thegrains are small and globose (when rehydrat-ed), have three broad sulci, and a relatively

smooth, thin exine. All 37 flies examined forpollen loads carried < 100 grains each ofthe pollen of T. menziesii pollen (Table 3)exclusively.

In contrast, syrphid flies did not forage fornectar. They alighted on perianth segmentsand probed the anthers with their proboscesfor pollen. All syrphids were captured on malephase flowers and none was seen visitingflowers in female phase when pollen was nolonger present in anthers. Pollen load analysesshowed that five of the nine syrphid specimenscarried Tolmiea pollen on their bodies(Table 3) but pollen washes indicated thatthese were trace amounts consisting of lessthan 10 grains for each fly.


Bumblebees were observed foraging onTolmiea menziesii from 13:00 to 17:00h.However no bumblebee was ever seen visitinga female phase flower. Bumblebees removedpollen from male phase flowers foragingrapidly from one male phase flower to thenext on the same inflorescence, scrapingpollen from anthers before moving to an-other inflorescence. Captured bees carriedmixed loads of pollen on their bodiesincluding Hydrophyllum tenuipes A. Heller

(Hydrophyllaceae) and Rubus sp. noted flow-ering nearby.

Relatively small amounts (>100 grainseach) of Tolmiea pollen were identified on theventral thorax and abdomen of some bees(Table 3). Far larger quantities of Tolmieapollen were identified with both the naked eye(due to their deep orange color) and bymicroscopic analyses of the contents of thebee’s corbiculae (pollen baskets) on the hindlegs. These pollen baskets also contained

Table 3. Insect visitors to Tolmiea flowers. Asterisk (*) indicates more individuals seen but not captured

Study site Date and time Insect taxa No. with hostpollen only

No. with hostand otherpollen

No. with otherspecies orno pollen

Eagle Creek May 23/01:11:30–13:30

Gnoriste megarrhina 10* 0 0

May 28/0111:00–13:00

G. megarrhina 10* 0 0

Parasyrphus macularis 0 1 1

May 20/0212:00–15:00

Bombus sitkensis 0 2 1

Parasyrphus macularis 2 0 1Melastoma mellinum 1 1 1G. megarrhina 2 0 0

May 31/02 Gnoriste megarrhina 10* 0 0Latychaeris obscurus 0 1 0

Tryon Creek June 1/01:16:30–17:30

Bombus flavifrons 0 2 0

G. megarrhina 2 0 0June 21/01:12:00–14:00

Bombus sitkensis 0 2 0

G. megarrhina 3 0 0Parasyrphus macularis 0 0 1

May 24/02:11:00–1:00

Bombus flavifrons 1 1 1

Bombus caligenosus 1 0 0Parasyrphus insolitus 1 0 0Melastoma mellinum 0 0 0

McLeay Park June 20/01:2:30–4:30

G. megarrhina 2 0 0

June 21/01:10:00–11:30

G. megarrhina 2 0 0

Bombus sitkensis 0 2 0June 27/01:1:30–3:00

Melastoma mellinum 0 0 2


pollen of co-blooming, Hydrophyllum andRubus, often in discrete layers, the lowerconsisting of Hydrophyllum mixed with Rubus(both nectar-secreting flowers) while the out-ermost, bright orange-red layer consisted pri-marily of Tolmiea pollen. This suggested thatBombus species foraged on nectariferous

Hydrophyllum and Rubus earlier in the day,before visiting T. menziesii flowers and ignor-ing their sugar-poor nectar.

Flower to fruit conversion ratio. Fruit set inTolmiea menziesii was often high. Open (gnat)pollinated inflorescences showed a conversionrate of 95% flowers carrying capsules (n¼ 10)at one study site. At a second site, perhaps dueto weather or fewer available pollinators,showed 40–60% of flowers on an inflorescenceforming mature capsules.

Fruit set. In the laboratory, all flowers leftunpollinated fell from the flowering stalk bythe eighth day of anthesis. All flowers self-pollinated by hand also fell from the stalk bythe eighth day of anthesis (n¼ 20). All flowerscross- pollinated by hand were retained on thestalk, the ovary rapidly enlarged, and fruitsdeveloped to maturity (n¼ 10) but the resul-tant seeds were not examined for viability. Weconcur with Weiblen and Brehm (1996) thatTolmiea is self-incompatible, is incapable ofmechanical autogamy and requires insect med-iated cross-pollination before seed productionmay occur.

Comparative pollen tube penetration. Wereject the null hypothesis that there is nodifference in the number of pollen tubespenetrating styles and ovules based on whetherpollenation was effected by gnats (open) orvia manipulated cross- or self-pollinations(F (2,18)¼ 9.01, P < 0.01). Hand-manipulatedpollinations produce >50% more pollen tubesin styles than gnat-mediated pollinationsregardless of whether it is a cross- or self-pollination. However, the mean of pollen tubespenetrating ovules is 45% higher in cross- vsself-pollinations. In a self-pollination, 47 of thetubes grow no further than the top of the

Fig. 2. The mycetophilid gnat, Gnoriste megarrhina.Scale bar = 1 mm. Drawn from preserved materialby J. C. Manning

Table 4. Significant floral and insect dimensions of Tolmiea menziesii and Gnoriste megarrhina

Taxon Mean flower/insectbody length (SD)

floral tube/proboscislength mm

Tolmiea ca. 8 mm ca. 5Gnoriste 7.42 mm (SD 0.53, n = 10) mean 6.55 (SD 0.18, n = 10)Bombus sitkensis 11.1 mm (SD 01.34, n = 5) mean 5.1 (SD 0.65, n = 5)


ovary (Table 5). Tubes that fail to growfurther than the top of the ovary fail to showsuch typically aberrant signs of development ascorkscrew or swollen or ruptured tips. Manytubes, however, appear to lose their ‘‘sense ofdirection’’ and grow horizontally within theovary, often entering the base of the oppositestyle (Figs. 4 and 5).

Discussion

We agree with Weiblen and Brehm (1996) thatdifferent modes of floral presentation limitintergeneric hybridization between Tolmieaand Tellima grandiflora. However, bumblebeesand syrphid flies were not primary pollinatorsof Tolmiea at our study sites. Successfulpollination of T. menziesii appears to occurwhen there is passive contact (sensu Bernhardt1996) between the pollen-bearing vector andreceptive stigmas while the insect forages fornectar. Our field observations and pollen loadanalyses indicate that active pollen collectionby syrphid flies is associated typically withunusually small quantities of pollen trans-ported by these insects. Bumblebees, whilemassive collectors of pollen, avoided flowers inthe female phase and we did not observe anyindividuals of either of these insect groups incontact with receptive stigmas of the hostflowers.

Therefore, the primary pollinators of Tol-miea menziesii appear to be the fungus gnat,Gnoriste megarrhina, considering the ‘‘good-ness of fit’’ between insect and flower. Thegnat’s proboscis is almost as long as its bodyand it carries significant loads of Tolmieapollen after visiting several flowers for nectar.More important, gnats visit flowers in bothmale and female phase hours before the arrivalof the first bumblebees. While a zygomorphicperianth is atypical for most flowers pollinatedby Diptera, it is not uncommon (Vogel 1954,Proctor et al. 1996, Larson et al. 2001, Bern-hardt and Goldblatt 2000, Goldblatt et al.2003). We have already mentioned the pres-ence of dull colors and filiform floral organs insapromyiophilous flowers but the character of

Fig. 3. Orientation of Gnoriste megarrhina on theflower of Tolmiea menziesii while foraging for nectar.Scale bar = 1 mm. Drawn from preserved materialand photographs by J. C. Manning

Table 5. Comparative pollen tube penetrations in pistils of Tolmiea menziesii

Cross type Number ofpistils

Ratio of styleswith tubes

Number of pollen tubesin top of ovary

Number of pollen tubespenetrating ovules

Mean S.D. Mean S.D.

Open 6 0.66 29.83 30.35 27.83 28.66Cross 9 0.83 62.56 41.02 58.22 43.24Self 6 0.78 67.29 46.68 32.14 53.70


minute quantities of nectar with low levels ofdissolved sugars could be added to this floralsyndrome.

Tolmiea menziesii now joins the small butgrowing list of plants pollinated by fungusgnats in the Mycetophilidae and/or the related

Fig. 4. Pistil self-pollinated by hand showing how the pollen tubes from each of the twin styles have stoppedgrowing within the top half of the ovary. Note that approximately five tubes in the right hand bundle continueto grow downward. Brighly colored dots are the swollen heads of the transparent epidermal cells (cleared in10% sodium sulphide solution). Scale x 75

Fig. 5. Pistil self-pollinated by hand showing how the pollen tubes from each of the twin styles stop growingdownward, change direction and grow toward each other. Note the approximately four tubes continuing togrow downward. Scale x 75


Sciaridae (Larson et al. 2001). Pollination byfungus gnats falls into two interrelated sys-tems. Pollination by deceit is presumed in thegenera Aristolochia (Aristolochiaceae, Stebbins1971), Heterotropa (Asaraceae: Sugawara1988), Corybas and Pterostylis (Orchidaceae:Dafni and Bernhardt 1990, Bernhardt 1995).In Heterotropa and Corybas pollination isbelieved to occur by brood site deception andonly pregnant females are exploited. Pterosty-lis curta is interpreted as an example of sexualmimesis exploiting only male gnats. Tolmieamenziesii does not express either mode ofpollination-by-deceit.

The second pollination system by fungusgnats offers an edible reward in the form ofminute quantities of nectar. This has beenrecorded for Acianthus (Orchidaceae: Dafniand Bernhardt 1990), Scoliopus (Liliaceae s.l.:Mesler et al. 1980), and Listera cordata (L.)R. Br. (Orchidaceae: Ackerman and Mesler1979). What is particularly significant aboutthis reward system for gnats is that Listeracordata and Scoliopus also occur in the PacificNorthwest. Scoliopus, Listera cordata, andnow Tolmiea are probably components of asyndrome more common in this region thananticipated. Pollination by fungus gnats mayrepresent a significant reproductive trend with-in some floristic alliances within the PacificNorthwest. We suggest that fungus gnat pol-lination may be particularly advantageous toherbs confined to the forest floors and lightgaps of mesic, cool, fungus-rich forests dom-inated by evergreen conifers. It is interesting tonote that in Australia, Acianthus, Corybas andPterostylis spp. are also associated with cool,wet, late winter-early spring flowering periodswithin shaded-to partially shaded woodlandsand heathlands (Bishop 1996).

Most flies of the families of suborderNematocera of the Diptera, to which theMycetophilidae belong, are characterised byprimitive suctorial mouth-parts mounted on avery short proboscis. Larson et al. (2001)observed that they are generally restricted toopen flowers with exposed nectar or to flowerswith short tubular perianths. Flowers visited

by nematocerans include, in particular, somemembers of the Apiaceae, Brassicaceae,Euphorbiaceae, Rosaceae, Saxifragaceae, andsome Asteraceae distributed through Green-land, North America and New Zealand (Cru-den 1972, Primack 1983, Shaw and Taylor1986, Proctor et al. 1996). Nematocerans areconsidered less important as dependable pollenvectors than their frequency of visitation mightsuggest. However this suborder contains manynocturnal species and observations of flowersat night are needed because these may greatlyincrease our recognition of their prevalence aspotential pollinators. The ancient origin ofNematocera and their consumption of nectaras a carbohydrate fuel for flight places them ascontenders for early pollinators of the firstangiosperms (Willemstein 1987). Obviously,Gnoriste megarrhina is a most important excep-tion to this rule when one considers the sheerlength of its proboscis, particularly in relationto its body. Indeed, this insect’s body is alsotwice the length of the diminutive Mycomyaspecies identified as pollinators of Pterostylisand Corybas R. Br. (Bernhardt 1995, and seereview in Dafni and Bernhardt 1990).

In turn, the dull-colored, zygomorphicand tubular flowers of Tolmiea menziesii aredistinctive within the Saxifragaceae althoughvegetative growth in this species stronglyresembles such genera as Heuchera, Mitella,and Tellima, to which it is so closely related(Soltis and Kuzoff 1995). Molecular analysisshows Tolmiea to belong to the Heucherinaclade of the family (Soltis et al. 2001). Withinthat clade Tolmiea is distinguished fromrelated genera by its floral zygomorphy, threestamens per flower, five unequal calyx lobes,and four filiform petals. In contrast, itsclosest relatives have radially symmetrical,pentamerous flowers with laminate petals. Itis apparent that taxonomists have placedTolmiea in a separate genus because its flowerstructure appears so divergent from relatedSaxifragaceae despite its inter-compatibilitywith Tellima.

While Tolmiea menziesii is an obligate out-crossing species, observations of fruit set and


analyses of open (gnat) pollinated pistils showhigh rates of cross-pollination in situ. Incontrast hand-pollination experiments indicatethat self-isolation mechanisms between pollenand pistils require clarification and comparisonwith earlier studies on this species and othermembers of the same family. While there is nosignificance between the sheer numbers ofpollen tubes penetrating style tissue followinga self-pollination vs. a cross-pollination, theactual number of tubes that enter ovulesfollowing cross-pollination is dramaticallyhigher. This is similar to hand-pollinationexperiments on the allied Heuchera micranthavar. diversifolia in which most pollen tubesproduced by self-pollination stop growingonce they reach the base of the style. The fewremaining tubes enter the ovules but, as inT. menziesii, no fruits develop. Rabe and Soltis(1999) concluded that self-incompatibility inH. micrantha lay between gametophytic andlate-acting incompatibility.

Are we actually looking at two, overlap-ping, self-incompatibility sites in the pistil ofTolmiea menziesii? A second interpretation ispossible following review of a much earlierexperiment by Sears (1937) who self-pollinateda single plant of T. menziesii. The author alsonoted that, while the majority of ovules in aself-pollinated pistil were not fertilized, the fewthat were developed a limited number ofembryo cells (four-nine) and 20-60 endospermnuclei (see Sears 1937; Plate I, Fig. 7). If Searswas correct then self-isolation in T. menziesiimay not be based on two, ‘‘overlapping’’ self-incompatibility sites. It would be based onone, partial gameophytic mechanism in theovary followed by a post-zygotic (embryoniclethal?) rejection during early phases of seeddevelopment. Controlled, hand-manipulatedpollinations obviously produce more tubesin styles and more tubes penetrating ovulesthan open pollinations effected by gnats(Table 5).

We thank Scott Fitzgerald at Oregon StateUniversity, Corvallis, C. Thompson at USDA

Beltsville, and R. Thorp at the University ofCalifornia, Davis, for help with identification ofinsects and helpful comments on insect biology.

References

Ackerman J. D., Mesler M. R. (1979) Pollinationof Listera cordata (Orchidaceae). Amer. J. Bot.66: 820–824.

Bernhardt P. (1995) Notes on the anthecology ofPterostylis curta (Orchidaceae). Cunninghamia4: 1–8.

Bernhardt P. (1996) Anther adaptations in animalpollination. In: D’Arcy W.G., Keating R. C.(eds.) The anther; form, function and phylogeny.Cambridge University Press, Cambridge, U.K.,pp. 192–220.

Bernhardt P., Goldblatt P. (2000) The diversity ofpollination mechanisms in the Iridaceae ofsouthern Africa. In: Wilson K., Weston P.(eds.) Monocots: systematics and evolution.CSIRO, Melbourne, pp. 301–308.

Bishop T. (1996) Field guide to the orchids of NewSouth Wales and Victoria. University of NewSouth Wales Press, Sydney, Australia.

Cruden R. W. (1972) Pollination biology ofNemophila menziesii (Hydrophyllaceae) withcomments on the evolution of oligolectic bees.Evolution 26: 373–389.

Dafni A., Bernhardt P. (1990) Pollination ofterrestrial orchids of southern Australia and theMediterranean region: systematic, ecological andevolutionary implications. In: Hecht M. K.,Wallace B., Macintyre R. J. (eds.) Evolutionarybiology. Plenum, New York, pp. 193–252.

Faegri K., van der Pijl L. (1979) The principles ofpollination ecology, 3rd edn. Pergamon Press,New York, New York, USA.

Goldblatt P., Bernhardt P. (1990) Pollinationbiology of Nivenia (Iridaceae) and the presenceof heterostylous self-compatibility. Israel J. Bot.39: 93–111.

Goldblatt P., Manning J. C. (2000) The long-proboscid fly pollination system in southernAfrica. Ann. Missouri Bot. Gard. 87: 146–170.

Goldblatt P., Manning J. C., Bernhardt P. (1995)Pollination biology of Lapeirousia subgenusLapeirousia (Iridaceae) in southern Africa: floraldivergence and adaptation for long-tongued fly-pollination.Ann.Missouri Bot.Gard. 82: 517–534.


Goldblatt P., Manning J. C., Bernhardt P. (1998a)Floral biology of bee-pollinated Gladiolus speciesin southern Africa. Ann. Missouri Bot. Gard. 85:492–517.

Goldblatt P., Manning J. C., Bernhardt P. (2000a)Adaptive radiation of pollination systems inSparaxis Ker Gawler (Iridaceae: Ixioideae).Adansonia ser. 3, 22: 57–70.

Goldblatt P., Manning J. C., Bernhardt P. (2000b)Adaptive radiation of pollination mechanisms inthe African genus Ixia (Iridaceae: Crocoideae).Ann. Missouri Bot. Gard. 89: 564–577.

Goldblatt P., Manning J. C., Bernhardt P. (2003)Observations on the floral biology of Melasphae-rula (Iridaceae: Crocoideae); is this monotypicgenus pollinated by march flies (Diptera: Bibion-idae)? Ann. Missouri Bot. Gard. 92(in press).

Hickman J. C. (ed.) (1993) The Jepson manual:higher plants of California. Univ. CaliforniaPress, Berkeley, CA.

Johnson S. D., Steiner K. E. (1994) Efficientpollination of the mass-flowering Cape orchidDisa obtusa Lindl. (Orchidaceae) by Bibio turneriEdwards (Diptera: Bibionidae). African Entom.2: 64–66.

Knuth P. (1908) Handbook of flower pollinationbased upon Hermann Muller’s work ‘‘Thefertilisation of flowers by insects’’. Transl. J. R.Ainsworth Davis. Clarendon Press, Oxford,United Kingdom.

Larson B. M. H., Kevan P., Inouye D. W. (2001)Flies and flowers: taxonomic diversity of antho-philes andpollinators.Canad.Entom. 133: 439–465.

Mesler M. R., Ackerman J. D., Lu K. L. (1980)The effectivness of fungus gnats as pollinators.Amer. J. Bot. 67: 564–567.

Ogden E. C., Raynor G. S., Hayers J. V., Lewis D.M. (1974) Manual of sampling airborne pollen.Hafner Press, London, United Kingdom.

Ornduff R. (1971) The reproductive system ofJepsonia heterandra. Evolution 25: 300–311.

Ornduff R. (1975) Complimentary roles of halictidsand syrphids in the pollination of Jepsoniaheterandra (Saxifragaceae). Evolution 29(2):371–373.

Pellmyr O., Thompson J. N., Brown J. M.,Harrison R. G. (1996). Sources of variation inpollinator contribution within a guild: the effectsof plants and pollinator factors. Oecologia 107:595–604.

Primack R. B. (1983) Insect pollination in the NewZealand mountain flora. New Zealand J. Bot. 21:317–333.

Primack R. B. (1985) Longevity of individualflowers. Ann. Rev. Ecol. Syst. 16: 15–37.

Proctor M., Yeo P., Lack A. (1996) The naturalhistory of pollination. Timber Press, Portland,OR.

Rabe A. J., Soltis D. E. (1999) Pollen tube growthand self-incompatibility in Heuchera micranthavar. diversifolia (Saxifragaceae). Int. J. Plant Sci.160: 1157–1162.

Sears E. R. (1937) Cytological phenomena con-nected with self-sterility in the flowering plants.Genetics 22: 1662–1682.

Shaw D. C., Taylor R. J. (1986) Pollination ecologyof an alpine fell-field community in the NorthCascades. Northwest Sci. 60: 21–31.

Soltis D. E., Kuzoff R. K. (1995) Discordancebetween nuclear and chloroplast phylogenies inthe Heuchera group (Saxifragaceae). Evolution49: 727–742.

Soltis D. E., Kuzoff R. K., Mort M. E., Zanis M.,FishbeinM., Hufford L., Koontz J., ArroyoM. K.(2001) Elucidating deep-level phylogenetic rela-tionships in Saxifragaceae using sequences for sixchloroplastic and nuclear DNA regions. Ann.Missouri Bot. Gard. 88: 669–693.

Stebbins G. L. (1971) California pipevine: a lighttrap for unwary flies. California Native Pl. Soc.Newslett. 7: 4–5.

Sugawara T. (1988) Floral biology of Heterotropatamaensis (Aristolochiaceae) in Japan. PlantSpecies Biol. 3: 77–12.

Thompson J. N., Cunningham B. M. (2002)Geographic structure and dynamics of coevolu-tionary selection. Nature 417: 735–738.

Vogel S. (1954) Blutenbiologische Typen als Ele-mente der Sippengliederung. Bot. Studien 1:1–338.

Vogel S. (1979) Pilzmuckenblumen als Pilzmime-ten. [Fungus-gnat flowers mimicking fungi].Flora 167: 329–398.

Weiblen G. D., Brehm B. G. (1996) Reproductivestrategies and barriers to hybridization betweenTellima grandiflora and Tolmiea menziesii (Saxi-fragaceae). Amer. J. Bot. 83(7): 910–918.

Willemstein S. C. (1987) An evolutionary basis forpollination ecology. E. J. Brill/Leiden UniversityPress, Leiden.


Addresses of the authors: Peter Goldblatt(e-mail: [email protected]), B. A. Kruk-off Curator of African Botany, Missouri BotanicalGarden, P. O. Box 299, St. Louis, Missouri 63166-0299, USA. Peter Bernhardt (e-mail: bernhap2@

slu.edu), Patrick Vogan, Department of Biology,St. Louis University, St. Louis, Missouri 63103,USA. John C. Manning (e-mail: [email protected]), National Botanical Institute,P. Bag X7, Claremont 7735, South Africa


pollination by fungus gnats (diptera: mycetophilidae) and self … · 2008. 10. 10. · ca. 7 mm...

Documents