Amer. J . Bot. 67(4): 550-555. 1980. STUDIES ON THE POLLINATION ECOLOGY OF TIPULARIA DISCOLOR (0RCHIDACEAE)l DENNIS F. WHIGHAM AND MARGARET MCWETHY Chesapeake Bay Center for Environmental Studies, Smithsonian Institution, P.O. Box 28, Edgewater, Maryland 21037 A B S T R A C T Tipularia discolor, a woodland orchid, flowers in mid-summer when reproductive activity is minimal within the herb synusia. Tipularia is insect-pollinated, and artificial crosses showed that seeds are produced after self-pollination, intra-inflorescence pollination, and outcrossing. The single nocturnal pollinator, Pseudaletia unipuncta (Noctuidae), located Tipularia popu- lations within a day or two of anthesis. Pollinators were shown to be capable of utilizing portions of the inflorescences that contained the most nectar. After total nectar resources declined, pollinators were no longer active on the inflorescences, even though flowers and nectar were still available. The mode of pollinator activity seems to be closely related to floral morphology, although the moths are able, early in the flowering phenophase, to successfully obtain nectar without effecting any change in the reproductive status of flowers. ASEXUAL REPRODUCTION is the predominant mode of propogation of perennial herbs in mesophytic teqperate deciduous forests (Ka- wano, 1975; Struik, 1965; Whigham, 1974) and a large percentage of the net annual primary production is allocated to belowground storage structures (Abrahamson, 1979). Accordingly, most would be characterized as K-adapted species (Gadgil and Solbrig, 1972; Harper, 1977). In addition, most produce small num- bers of large seeds and rely primarily on per- ennation by asexual reproduction (Kawano, 1970, 1975; Kawano and Nagai, 1975; Morgan, 1971 ; Stevenson, 1972; Whigham, 1974). Tipularia discolor (Cranefly orchid), a species which is widespread throughout the eastern deciduous forest herbaceous flora (Braun, 1950), is of interest because it is one of the few woodland herbs that produces nu- merous seeds, a characteristic usually asso- ciated with species of early successional hab- itats, yet populations contain few seedlings and asexual reproduction is the primary mode of population maintenance (Whigham, unpub- lished). Tipularia exhibits other life cycle char- acteristics that are of ecological interest. Ti- pularia and Aplectrum hyemale (Auclair, 1972; Adams, 1970) are the only two orchids in east- ern deciduous forests that produce leaves, Received for publication 26 June 1979; revision ac- cepted 9 October 1979. This research was funded by the Smithsonian Institu- tion's Environmental Studies Program. We would like to thank Ann Hall for assistance with the field work and Jay O'Neill, John Falk, Jim Lynch, and Beryl Simpson for reviewing the manuscript. Moth identifications were made through the efforts of Gary Hevel. usually one per plant, that develop in the late summer, persist through the winter, and se- nesce in spring prior to the onset of flowering. Unlike Aplectrum, Tipularia flowers in mid- summer (July-August) when sexual reproduc- tion within the herb synusia is minimal (Taylor, 1974). Tipularia, consequently, flowers when pollinator resources, pollen and nectar, are minimal. This paper reports results of studies of the reproductive ecology of Tipularia. Spe- cific objectives were to determine: (1) the type of breeding system exhibited by Tipularia; (2) the relationship between flowering phenology and pollinator activity; and (3) the relationship between nectar production and pollinator ac- tivity. The research was conducted at the Chesa- peake Bay Center for Environmental Studies (CBCES), a 1,000-ha research facility of the Smithsonian Institution. The CBCES property is located on the watershed of the Rhode River (Correll, 1977), approximately 16 km south of Annapolis, Maryland. Upland vegetation con- sists of a series of mature and successional forests, as well as pasture and cultivated areas. The Tipularia research was performed in a mixed hardwood forest that has not been dis- turbed since the Civil War (1861-65). Dominant overstory species include Quer- cus alba, Fagus grandijblia, Carya glabra, Carya tomentosa, Quercus falcata, Lirioden- dron tulipifera, and Liquidambar styraciflua. Understory vegetation consists of canopy transgressives and Cornus Jlorida, Carpinus caroliniana, Viburnum dentatum, and Lindera benzoin. The herb synusia is diverse, with Claytonia virginica, Dentaria laciniata, Podo- phyllum peltatum, Smilacina racemosa, and April, 19801 WHICHAM A N D MCWETHY-POLLINATION ECOLOGY OF TIPULARIA 55 1 D e s m o d i u m nudigorurn being abundant species. Small populations of Tipularia are distributed throughout the forest. METHODS-Breeding experiments were conducted in 1977 and 1978. Inflorescences, which emerge in July, were located prior to anthesis and the following manipulations per- formed: (I) pollinia were removed from 151 flowers on inflorescences of seven plants (apomixis); (2) pollinia were also removed from 45 flowers on three inflorescences and remaining flowers (53) on the three inflores- cences were used for a study of automatic self- ing and were not manipulated. The three in- florescences were then covered with a nylon bag to exclude pollinators: (3) flowers on eight plants (133 flowers) were self-pollinated (autogamy) and inflorescences were covered with nylon bags; (4) flowers of eight plants (156 flowers) were outcrossed (xenogamy) and in- florescences covered with nylon bags; (5) intra- inflorescence crosses (geitonogamy) were made on five plants (78 flowers) and inflores- cences were covered with nylon bags. Nylon bags were removed when floral parts started to senesce. In addition, all inflorescences with- in a control population of 100 (1977) and 118 (1978) plants were observed throughout the flowering period and the number of pollinated flowers counted on each inflorescence. The activities of pollinators were studied during both flowering seasons. All plants in the marked populations were visited at 2 or 3 day intervals in 1977 and 1-3 day intervals in 1978. During each visit in 1977, we examined all flow- ers on each inflorescence and scored them ac- cording to the following scheme which was designed to permit us to determine when flow- ers were visited and the impact of the polli- nators: (I) flower intact; (2) pollinia removed; (3) pollinia placed on flower but not on stig- matic surface; (4) pollinia placed on stigmatic surfaces. Observations during the 1977 study indicat- ed that pollinators might have removed nectar from some flowers without otherwise affecting them. In 1978, we repeated the study, but also measured the amount of nectar in the spur of each flower. Nectar was estimated from mea- surements of the length of the nectar column in the spur of each flower. The spurs of Ti- pularia flowers are translucent and, nearest the throat, the miniscus of the nectar column is clearly visible. Nectar column length (mm) was converted to p l of nectar using a regression equation that was developed by measuring the length of the nectar column (X) in flowers and then extracting the ,nectar with I-pl micropi- pettes. The linear regression equation was Y = -. 143 + .11475X, where Y is nectar volume (pl). The correlation coefficient for the equa- tion was .93 (n = 50 flowers). The study populations were observed daily in 1977, and it was apparent that the flowers were visited by nocturnal animals. In 1978, we attempted to collect pollinators by sampling for 5-hr periods, on four consecutive nights, during the peak in pollinator activity. The full moon conditions that prevailed during the four nights (August 13-1 6) enabled us to observe and collect pollinators that visited Tipularia inflorescences. Phenological aspects of the marked population were noted at monthly in- tervals between May, 1977 and October, 1978. Observations were increased to a weekly in- terval during the flowering phenophase. RESULTS-Inflorescences began to emerge during the 1st wk of July in 1977 and the 2nd wk in 1978. Twenty-three percent of 100 marked plants flowered in 1977 and 8.9% of 118 in 1978. Only one marked plant flowered in both 1977 and 1978. This observation is con- sistent with other data (Whigham, unpub- lished) indicating that very few plants flower in consecutive years. The species successfully set seed after self pollination and both inter- and intra-inflores- cence crosses. The highest success rates were for artificially selfed flowers, where 91.3 + 3.4% (1 SE) of the flowers set seed. A slightly lower seed set success occurred in inter-inflo- rescence (69.1 t 8.2%) and intra-inflorescence (84.0 t 11.1%) crosses. The lower percentages reflect the abortion of developing fruits due to unknown causes in 1978. Flowers from all in- ter- and intra-inflorescence crosses made in 1978 began to develop fruits, but several abort- ed within two weeks. The low (23.%) per- centage of fruit development from flowers that had pollinia removed (no bag placed over in- florescence), indicates that only a small per- centage of flowers on a given inflorescence are successfully pollinated. We anticipated that there would be no fruit development in flowers that were tested for apomixis by removing the pollinia and then covering the inflorescence with a nylon bag. No fruits developed from flowers tested for apomixis in 1977 but, un- fortunately, in 1978 each inflorescence used to test for apomixis was also used to test for au- tomatic selfing and thus had terminal flowers from which we did not remove the pollinia. Fruits developed from 15.9 t 5.0% of the flow- ers used to test for apomixis and 12.8 t 8.9% of those used to test for automatic selfing. After removal of the nylon bags, we noted, however, 552 AMERICAN JOURNAL OF BOTANY [Vol. 67 % FLOWERS Y l S l i E O NUMBCR P L I N T S VlSlTED 2 4 6 B 10 12 4 15 IB 20 ZZ 71 26 D A Y S AFTER JULY 2 6 Fig. I . Results of 1977 population census during the flowering phenophase. Pattern of flower availability is shown in upper graph. Temporal pattern of pollinator ac- tivity is shown in lower graph. Pollinator visitations were based on noting physical changes in the reproductive status of Tipularia flowers (see methods for explanation of pro- cedure). that pollinia had been removed from some flowers used to test for automatic selfing and had, somehow, been transferred to stigmatic surfaces of all flowers that developed into fruits. The data, therefore, suggest that Tipu- laria is not apomictic nor does it automatically self, because no fruits developed from any of the flowers that remained intact. Anthesis began on July 26, 1977, and the last floral part had fallen by August 15. The flow- ering phenophase began later (August 4) in 1978, and lasted for 23 days, which was five days longer than in 1977 (Fig. 1, 2). In both years, all flowers had opened by the 12th day after initial anthesis (Fig. 1, 2). Plants were first visited by pollinators on day 4 after an- thesis in 1977, and on the 5th day in 1978. Visitations continued for 13 (1977) and 15 (1978) days, with pollinator activities peaking on the 13th day after anthesis in 1977 and on the 11 th day in 1978. Interestingly, pollinator visitations ceased while there were still open flowers. In 1977, open flowers were present for 15 days after the last definite pollinator visita- tion. In 1978, open flowers were present for only 3 days after the last recorded pollinator - % FLOWERS VISITEDIA) m - > %FLOWERS VISITED ( B ) u, 2 4 6 8 10 12 14 16 18 20 22 24 26 28 D A Y S A F T E R A U G U S T 2 Fig. 2. Results of 1978 population census during the flowering phenophase. Patterns of flower and nectar avail- ability are shown in upper graph. Pattern of pollinator activity is shown in lower graph (see methods for a de- scription of procedure). Parts of the figure labeled B are visitations based on noting physical changes in reproduc- tive status of Tipularia flowers. Parts labeled A are based on estimated total visits, which is the sum of B plus vis- itations during which nectar was removed, but no change in floral reproductive status was observed. visitations. The maximum percentage of avail- able flowers visited daily by pollinators was between 25% and 45% (Fig. 1, 2). Although each inflorescence was not visited daily, all were visited during the flowering phenophase. Figure 2 shows the amount of nectar avail- able during the flowering phenophase. Maxi- mum nectar availability occurred very early and prior to anthesis of all flowers. Total avail- able nectar then began to decline within the population once flowers began to be visited by pollinators and there were very few pollinator visitations after the large drop (30%) in nectar between days 15 and 16. The weighted mean April, 19801 WHIGHAM A N D MCWETHY-POLLINATION ECOLOGY OF TIPULARIA 553 estimated amount of nectar production was .98 t .08 (1 SE) pl per flower and ranged from a high of 1.35 -+ .47 p1 per flower to a low of .53 t .22 pI per flower. As mentioned in methods, it seemed that potential pollinators were capable of removing nectar without effecting any change in floral reproductive status. We attempted to clarify this question by comparing the number of flow- ers that showed any of the observed morpho- logical changes, as well as had significant (.23-.25 or more) changes in nectar vol- ume (Method A, Fig. 2), with the number of flowers that only had observable morphologi- cal changes (Method B, Fig. 2). The two cal- culations differed, with the largest discrepancy found on the seventh day after anthesis. On that day, volumetric changes in nectar indi- cated that the reproductive status changed on only 18% of the 138 flowers that were visited. A large drop in total nectar also occurred on that day (Fig. 2). Differences between the methods (Fig. 2) of determining visitations then declined and averaged approximately 40% thereafter. Three potential pollinator species, all of them moths, visited Tipularia inflorescences during each of the four observation periods. Two smaller geometrid moths (Protoboarmia porcelaria and Xanthorhoe ferrugata) were frequently seen, but only occasionally visited Tipularia inflorescences. When they did alight on a flower, they did not spend more than 2 or 3 sec and they flew away without visiting any other flowers. Examination of flowers after they had been visited, showed that the smaller moths did not effect any changes in the repro- ductive status of the flowers and, based on visual observations before and after visita- tions, they did not appear to remove nectar from spurs. Large noctuid moths (Pseudaletia unipuncta) were observed each night, and their flight pattern suggested they recognized Ti- pularia inflorescences. Pseudaletia, typically, was observed flying about 0.5-1 m above the ground. When individuals flew within approx- imately 3-5 m of an inflorescence, their pattern of flight suddenly altered, and they moved im- mediately to the inflorescence. Pseudaletia moths alighted on sepals and petals at the open throats of flowers. The heads then were in- serted into the throat regions of flowers while nectar was removed. The moths' wings were continually in motion while flowers were vis- ited, and they were able to move quickly from one flower to the next on a given plant. Each moth only visited a few flowers on an inflo- rescence before flying away. Visual observa- tions of Tipularia inflorescences immediately after the visitation showed, however, that nec- tar was removed from all flowers visited, while pollinia remained intact on some visited flow- ers. D~SCUSSION-Although orchid flowers are adapted for insect pollination (van der Pijl and Dodson, 1966; Vogel, l978), Hagerup (1952) and Kirchner (1922) have shown that insect pollination of temperate orchids is not oblig- atory and that autogamy occurs in several species. Tipularia does not appear to be au- togamous nor is outcrossing obligatory. Apo- mixis and automatic selfing in Tipularia can't be excluded, but we believe that it, at most, occurs only infrequently. Of the flowers that were being tested for apomixis and automatic selfing, fruits only developed on those that, for some unknown reason, had pollinia on the stig- matic surface. Self-pollination resulted in successful seed production, but was only effected through the action of a pollinator. Seeds were also pro- duced by intra-inflorescence crosses and out- crosses. The potential to produce seeds inde- pendent of the source of pollinia would be of selective advantage to Tipularia because there are so few pollinators present during the flow- ering phenophase (only one species observed during this study). Few other temperate or- chids seem to exhibit this type of restricted pollinator-plant interaction. Stevenson (1972) found that Goodyera puhescens, a species with a flowering phenology similar to Tipularia was pollinated by Bomhus during daytime and that the species was not autogamous, although he did not perform artificial crosses. Ackerman (1975) also has cited Bomhus pollination of Goodyera. As noted (Hagerup, 1952; Acker- man and Mesler, 1979), some temperate zone orchids are capable of autogamy, but most ap- pear to be insect pollinated. Nilsson (1978a) found that 28 species of lepidopterans, pri- marily Noctuidae, visited Platanthera chlor- antha in Sweden. Moth-pollination of Pla- tanthera has also been cited by Stoutamire (1944) and Smith and Snow (1976). Epipactis palustris, in Sweden (Nilsson, 1978b), was vis- ited by a broad spectrum of insects. Of the species examined by Nilsson, the mode of pol- lination of Platanthera was somewhat similar to that observed in this study. Platanthera was also pollinated by moths and there was a spe- cific relationship between size of moths' heads and morphology of the flower reproduction parts. Flowers of both Platanthera and Tipu- laria frequently are robbed of nectar (Fig. 2), seemingly without any effective change in the reproductive status of the flower. It seems like- 554 AMERICAN JOURNAL O F BOTANY [Vol. 67 DAYS AFTER F I R S T NECTAR MEASURED Fig. 3. Comparison of temporal availability of nectar in Tipularia inflorescences. Values represent mean daily percentages of nectar present in quarters of eleven inflo- rescences examined in 1978. Quarters represent, in as- cending order beginning with the 1st quarter, sections of the inflorescence. ly that a moth could remove nectar from full Tipularia spurs without effecting changes in the flower. When nectar volume is less, the moth, which has to probe deeper in the spur for nectar, touches the viscid disc of the pol- linia with its eyes. When the moth's head moves further into the flower, the pollinia, now attached to the moth's eye, are pulled from the anther sac and moved to the region of the stig- matic surface, which is located immediately posterior to the anther sac. The stigmatic sur- face is very sticky and pollinia that contact it remain when the moth pulls away from the flower. Nilsson (1978a, 1978b) also found that deposition of pollinia in insect eyes is a com- mon mode of pollen transfer in temperate or- chids. As shown in Fig. 2, moths more fre- quently effect changes in the floral reproductive status after nectar begins to decline in the Ti- pularia population. Once pollinia were trans- ferred to the stigmatic surface, almost all of those flowers developed fruits (93.5 t 2.9% in 1977 and 70.7 k 8.7% in 1978). Because of the interdependent nature of the moth and Tipularia breeding system, pollina- tor patterns at the population level are of in- terest. Fig. 1, 2 show that pollinators located Tipularia flowers within a day or two of the beginning of anthesis. Pollinator activity in- creased sharply and then declined, even though open flowers and nectar were available for a much longer period of time. Willson and Bertin (1979) have provided additional infor- 2 4 6 8 10 12 14 16 18 20 DAYS AFTER FIRST NECTAR Fig. 4. Comparison of temporal patterns of pollinator visitations to portions of Tipularia inflorescences. All val- ues represent mean daily percent of visitations to quarters of eleven inflorescences examined in 1978. Quarters rep- resent, in ascending order beginning with 1st quarter, sec- tions of the inflorescence. Visitations calculated as de- scribed for A in Fig. 2. mation suggesting that Pseudaletia unipuncta exploits nectar for only a portion of the flow- ering period. They noted that Pseudaletia was a common pollinator of Asclepias for only 1 wk out of the total flowering period. Pseuda- letia was seemingly able to respond to the total amounts of nectar available and, based on our technique for identifying visitations, ceased foraging for Tipularia nectar after nectar pro- duction declined below some minimum level. The production of nectar occurred acropetally on the inflorescence (Fig. 3) in synchrony with anthesis that also started at the base of the inflorescences. Initial pollinator visits were to the lower half of the inflorescences (Fig. 4), where almost all of the nectar was located (Fig. 3). After approximately 5 days, nectar was equally divided throughout the inflorescences (Fig. 3) and there was also indiscriminate use of the inflorescences by the pollinators (Fig. 4). After the 10th day, flowers on the upper half of the inflorescence contain most of the nectar, although the total amount was less (Fig. 2), and pollinators began selectively to utilize the upper halves of inflorescences (Fig. 4). The few visitations that occurred near the end of the flowering period (Fig. 1, 2) were restricted to the upper flowers (Fig. 4) where all of the nectar was found (Fig. 3). Pollinators therefore not only respond to overall levels of nectar availability within the flowering cycle of Ti- pularia, but also seem able selectively to utilize those specific areas of a given inflorescence where potential rewards are greatest. April, 19801 WHIGHAM AND MCWETHY-POLLINATION ECOLOGY OF TIPULARIA 555 LITERATURE CITED ABRAHAMSON, W. G. 1979. Patterns of resource allo- cation in wildflower populations of fields and woods. Amer. J. Bot. 66: 71-79. ACKERMAN, J. D. 1975. Reproductive biology of Good- yera ohlongifolia (Orchidaceae). Madrofio 23: 191- 198. ACKERMAN, J. D., A N D M. R. MESLER. 1979. Pollination biology of Listera cordata (Orchidaceae). Amer. J . Bot. 66: 820-824. ADAMS, M. S. 1970. Adaptations of Aplectrum hyemale to the environment: Effects of preconditioning on net photosynthesis. Bull. Torrey Bot. Club 97: 219-224. AUCLAIR, A. N. 1972. Comparative ecology of the orchids Aplectrum hyemale and Orchis spectahilis. Bull. Tor- rey Bot. Club 99: 1-10. BRAUN, E. L. 1950. Deciduous forests of eastern North America. Hafner Publ. Co., New York. CORRELI~, D. L. 1977. An overview of the Rhode River watershed program. In: D. L. Correll [ed.], Wa- tershed research in Eastern North America, CBCES, Smithsonian Institution, Edgewater, Maryland, p. 100-120. G A D G I ~ ~ , M., A N D 0 . T . SOLBRIG. 1972. The concept of r and K selection: Evidence from wildflowers and some theoretical considerations. Amer. Nat. 106: 14- 31. HAGERUP, 0. 1952. Bud autogamy in some northern or- chids. Phytomorphology 2: 51-60. HARPER, J . L . 1977. Population biology of plants. Aca- demic Press, New York. KAWANO, S. 1970. Species problems viewed from pro- ductive and reproductive biology. I. Ecological life histories of some representative members associated with temperate deciduous forests in Japan-A prelim- inary discussion. J. Coll. Lib. Arts, Toyama Univ. 3: 181-213. . 1975. The productive and reproductive biology of flowering plants. 11. The concept of life history strategy in plants. J . Coll. Lib. Arts, Toyama Univ. 8: 51-86. , AND Y. NAGAI. 1975. The productive and repro- ductive biology of flowering plants. I. Life history strategies of three wild Allium species in Japan. Bot. Mag. (Tokyo) 88: 281-318. KIRCHNER, 0. 1922. Uber selbstbestaiibung bei den Or- chideen. Flora 1 15: 103-129. MORGAN, M. D. 1971. Life history and energy relation- ships of Hydrophyllum appendiculatum. Ecol. Mono- gr. 41: 329-349. NII~SSON, L. A. 1978a. Pollination ecology and adaptation in Platanthera chlorantha (Orchidaceae). Bot. Noti- ser 131: 35-51. . 1978b. Pollination ecology of Epipactis palustris (Orchidaceae). Bot. Notiser 131: 355-368. SMITH, G. R., ANDG. E. SNOW. 1976. Pollinationecology of Platanthera (Hahenaria) ciliaris and P. blephar- iglottis (Orchidaceae). Bot. Gaz. 137: 133-140. STEVENSON, J. C. 1972. Evolutionary strategies and ecology of Goodyera and Spiranthes species (Orchi- daceae). Ph.D. Thesis, Univ. North Carolina, Chapel Hill. STRUIK, G. J. 1965. Growth patterns of some native an- nual and perennial herbs in southern Wisconsin. Ecology 46: 401-420. STOUTAMIRE, W. P. 1974. Relationships of the purple- fringer orchids Platanthera psycodes and P. gran- diJlora. Brittonia 26: 42-58. TAYLOR, F. G. , JR. 1974. Phenodynamics of production in a mesic deciduous forest. In: H. Lieth Led.], Phe- nology and seasonality modeling, p. 237-254. Spring- er-Verlag, N.Y. VAN DER PIJI-, L., AND C. H. DODSON. 1966. Orchid flowers: their pollination and evolution. Coral Gables, Florida. VOGEL, S. 1978. IV. Floral ecology. Report on the years 1974 (73) to 1978. In: H. Ellenberg, K . Esser, H. Merxmuller, E. Schnepf, and H. Ziegler [eds.], Pro- gress in botany, p. 453-481. Springer-Verlag, New York. WHIGHAM, D. F. 1974. An ecological life history study of Uvularia perfoliata L. Amer. Midl. Nat. 91: 353- 359. WIISON, M. F., AND R. I. BERTIN. 1979. Flower-visitors, nectar production, and inflorescence size of Asclepias syriaca. Can. J. Bot. 57: 1380-1388.