Flower Phenology and Sexual Maturation: Partial Protandrous Behavior in Three Species of Orchids

RAYMOND L. TREMBLAY*, GRIZEL POMALES-HERNÁNDEZ AND MARÍA DE LOURDES MÉNDEZ-CINTRÓN

Department of Biology, University of Puerto Rico—Humacao, 100 carr. 908, Humacao, P.R., 00791 *Corresponding author: [email protected]

ABSTRACT.—Plants have theoretically multiple alternatives for preventing self pollination and consequently the effect of inbreeding, such as sequential flowering, dichogamy and self–incompatibility to name a few. We investigated the reproductive biology of three sequentially flowering (acropetal) endemic orchids from Puerto Rico. Since sequential flowering is present in the studied species and very rarely (1.0%) is there more than one flower open simultaneously on an inflorescence, we hypothesized that the orchids should be self-compatible and show no effect of protandry (dichogamy). We performed hand self—and cross- pollinations and evaluated whether the species are self-compatible and whether the receptivity to pollination success (fruit set) is influenced by the age of flowers (protandry). We define protandry as pertaining to a hermaphroditic organism that assumes a functional male condition prior to shifting to a functional female state. We found that all three species are self-incompatible. Furthermore, flower age is important for pre- dicting the likelihood of fruits set. Older flowers (6+ days) are significantly more likely to produce fruits (functional protandry). The multiple mechanisms for preventing self-pollination (sequential flowering, dichogamy and self-incompatibility) that are noted for these species suggest that the historical evolutionary processes for preventing inbreeding may be complex. We hypothesized that because multiple mechanisms are present for preventing self-pollination inbreeding depression is likely to be high.

KEYWORDS.—Lepanthes, Puerto Rico, reproductive success, dichogamy, self-incompatible

INTRODUCTION togyny) and solitary flowers (Delaporta and Calderón-Urrea 1993; Walker-Larson Low fruit set in many plant species is a and Harder 2000). consequence of pollinator limitation (Trem- In the Orchidaceae, pollinator limitation blay et al. 2005), or self-incompatibility, is frequently invoked as the cause of low which are perceived as mechanisms to pro- fruit set, but in self-compatible species low mote cross-pollination (Marshall and Fol- fruit set may additionally be attributed to som 1991). Promoting cross-pollination is incompatible self—and geitonogamous pol- believed to increase fitness by increasing linations, resulting in flower or fruit abor- genetic diversity, while self-pollination de- tion (Delaporta and Calderón-Urrea 1993; creases genetic diversity and can result in Walker-Larson and Harder 2000; Tremblay inbreeding depression (Lande and Schem- et al. 2005). In self-compatible orchids a ske 1985; Dudash 1990; Jarne and Charles- number of mechanisms for preventing self- worth 1993). Consequently, many species pollination (Johnson and Edwards 2000) probably evolved complex mechanisms in have been detected, including: sequential order to prevent self-pollination, including flowering (Psychilis spp., Ackerman (1989); self-incompatibility through sporophytic or Malaxis massonii (Ridl.) Kuntze, Aragón gametophytic processes, and through pre- and Ackerman (2001)), dichogamy/pro- visitation mechanisms, including sequen- tandry (Goodyera oblongifolia Raf., Acker- tial flowering, dichogamy (protandry, pro- man (1975); Listera cordata, Ackerman and Mesler (1979); Spiranthes, Catling (1983); ms. received October 2, 2004; accepted August 31, Sipes and Tepedino (1995); Prescottia 2005 stachyodes (Sw.) Lindl., Singer and Sazima 75 76 R. L. TREMBLAY ET AL.

(2001a); Erythrodes arietina (Rchb. F. & manipulation (apomixis). No fruits were Warm.) Ames, Singer and Sazima (2001b); set unless flowers were hand pollinated. Sauroglossum elatum Lindl., Singer (2002); Mesadenella cuspidata (Lindl.) Garay, Singer Determination of the breeding system (2002); Notylia nemorosa Barb. Rodr., Singer and Koehler (2002); Warford (1992); Man- Cross pollination vs. self pollination.— niella cypripedioides Salazar, T. Franke, Flowers of the three species of Lepanthes Zapfack & Beenken, Salazar et al. (2002)), studied do not show any evidence of self- temporal variation in pollinarium size (Bul- pollination, but to determine if autogamous bophyllum weddelii (Lindl.) Rchb. f., B. ipane- pollinations produce viable fruits, we per- mense Hoehne and B. involutum Borba, formed a series of hand self-pollinations. Semir and F. Barros, Borba and Semir We self-pollinated 3-day old flowers of L. (1999a); Trigonidium obtusum Lindl., Singer rupestris (n = 33) and L. woodburyana (n = (2002)), and change in position of lip and 50), and 6-day old flowers for L. rubripetala column after pollinarium removal or after a (n = 14). few days if pollinarium is not removed Determination of protandry.—To test if (Sosa and Rodríguez-Angulo 2000). protandry is present in Lepanthes we made We investigated whether or not self- cross-pollinations at different time intervals incompatibility is present in three sequen- of the flower age from one to eight days. tially flowering tropical species of Lepanthes We compared the percentage of fruit pro- (Orhcidaceae). In addition we determined duction per pollination for two age treat- whether their flowers are protandrous. ments (1-5 days and 6+ days) and among Protandry is a dichogamous mechanism; it the three species. We used contingency is defined as the maturation of the male tables to evaluate the null hypothesis that (androecium) structure before the matura- the pattern was random among days and tion of the female (gymnoecium) structures species. (Kearns and Inouye 1993). RESULTS

MATERIALS AND METHODS Flower development and fruit phenology From the time we noticed flower bud set, We determined the breeding system of these took 11 to 15 days to open in all three three endemic species of Lepanthes from species. Open flowers lasted up to 10 days Puerto Rico: L. woodburyana Stimson (53 in L. rupestris and L. woodburyana, and 11 individuals), L. rupestris Stimson (56 indi- days in L. rubripetala when not manipu- viduals), and L. rubripetala Stimson (35 in- lated. However, flowers manipulated dividuals). Each species was grown in a TM through pollinia removal or pollination re- Wardian case (Orchidarium Case )inthe duced the life span of open flowers. In gen- laboratory at the same temperature and hu- eral, removal of pollinia caused flowers to midity (18-24°C, 45-100% humidity). close and wither within 1 to 2 days, more- The phenology of flower buds, flowers over, pollination (deposition of pollinia (including anthesis), and fruits for each onto the stigma) resulted in flower closure plant was recorded. The age of flower the next day (Table 1). Mean life span (until buds, flowers and viable fruits along each dehiscence) of fruits was estimated at 31 inflorescence was recorded daily. We con- and 39 days for L. rubripetala and L. rupes- sidered a fruit viable as soon as we noticed tris, respectively. None of the fruits that the expansion of the ovary. As a control we were set aborted (abscission of fruit in early segregated four specimens of Lepanthes stages of development). rupestris and L. rubripetala and L. woodburyana to determine the effect of flower Hand self and cross-pollinations longevity on non-manipulated flowers, the effect of pollinia removal on flower longev- Hand self pollination did not produce ity and the ability to self-fertilize without any fruits for any of the three species of PROTANDRY IN ORCHIDS 77

TABLE 1. Flower phenology of Lepanthes flowers. Time (in days with s.e., sample size in parentheses).

Characteristics L. rubripetala L. rupestris Development of flower bud to anthesis 14.6 ± 0.3 (104) 11.7 ± 0.2 (150) Flower lifespan without manipulation 6.9 ± 0.5 (32) 5.8 ± 0.4 (36) Flower lifespan with pollinia removal only 1.2 ± 0.1 (11) 1.4 ± 0.1 (46) Flower lifespan when flower is pollinated 1.0 ± 0.0 (55) 1.0 ± 0.0 (61) Fruits life span 31.2 ± 2.2 (23) 39.3 ± 1.3 (34) orchids. On the other hand, fruit set by the flowers of Habenaria orbiculata (Pursch) cross-pollinations (irrespective of flower Torr. last about 12.7 days in the field. How- age) was 23.2% for L. rubripetala, 23.1% for ever, flowers of species such as Dendrobium L. rupestris and 41.1% for L. woodburyana. crumenatum Sw. and Sobralia macrantha Lindl. last only one day, while an extreme The effect of flower age on fruit set case are the flowers of Grammatophyllum multiflorum Lindl. which last nine months The number of hand cross-pollination in greenhouses. In Lepanthes flower longev- varied among species and was dependent ity is in the range of 6 – 8 days. When flow- on flower production. Three hundred sev- ers are manipulated (pollinaria removal enty-three pollinations were performed in and pollinaria deposited) flower age is L. rupestris, 142 in L. rubripetala and 192 in L. drastically reduce to one day. This phe- woodburyana. The pattern of fruit set and nomenon of shorten flower life span when flower age was consistent among species. flowers are manipulated is well known in Older flowers (6+ days) had significantly orchids. Given the small size of the flowers, higher fruit set in all three species as com- a reduction in the life span of open flowers pared from the expected proportion (Table may not simply be due to flower manipu- 2). Under the null hypothesis, the expected lation but also to actual physical damage to number of fruits (observed in brackets) delicate sexual structures. were all significantly lower than observed The effect of protandry has been ob- fruit set, 12 (17) for L. rubripetala, 22 (31) for served in orchids. In Goodyera oblongifolia L. rupestris, 38 (56) for L. woodburyana. pollen deposition is prevented by the prox- imity of the column to the lip in young DISCUSSION flowers. As flowers age the column be- comes exposed and thus receptive to pollen Flower longevity is often short (a few deposition (Ackerman 1975). In the genus days) in plants (Primack 1985). In orchids, Spiranthes, outcrossing is promoted by se- Primack (1985) noted that these have char- quential flowering but also by protandry acteristically long-lived flowers and that (Catling 1983; Sipes and Tepedino 1995).

TABLE 2. Hand cross-pollinations of three species of orchids with different flower age categories (in days). Numbers in parenthesis are observed proportion. test performed to determine if fruit set is flower age inde- pendent.

Hand Cross Pollination L. rubripetala L. rupestris L. woodburyana Flower No Fruit Fruit No Fruit Fruit No Fruit Fruit Age 1–5 75 (52.8%) 16 (11.2%) 222 (59.5%) 55 (14.7%) 77 (40.1%) 23 (11.9%) 6+ 34 (23.9%) 17 (11.9%) 65 (17.4%) 31 (8.3%) 36 (18.8%) 56 (29.2%) Chi Square 4.55 6.22 28.4 P-value 0.033 0.013 < 0.001 78 R. L. TREMBLAY ET AL.

Protandry in Spiranthes is mechanical, as in not sufficiently effective for preventing G. oblongifolia, and the time for male matu- self-pollination. ration to female stage last from 5 to 30 days We would predict that inbreeding de- (depending on the species of Spiranthes). pression (mating among related individu- Furthermore, the probability of cross- als) may be high in Lepanthes because of the pollination is influenced by the behavior of multiple mechanism that are apparently foraging insects that move up the inflores- present to prevent self-pollination, thus cence (acropetal). A similar pattern of mating with unrelated individuals (ex- flower phenology along the inflorescence ample, from other populations) should re- and pollinator behavior was also found in sult in higher fitness. This hypothesis does Prescottia stachyodes (Singer and Sazima rely on the fact that self-incompatibility, se- 2001a). In Erythodes arietina the male phase quential flowering, and protandry were lasts 1 – 2 days and is a mechanical function evolutionarily selected and not random of the column (Singer and Sazima 2001b) as processes. A complicating issue is that se- in Sauroglossum elatum and Notylia nemorosa quential flowering has evolved numerous from Brazil (Singer 2002; Singer and times in the Orchidaceae and may be linked Koehler 2002). In many of these species the to other evolutionary processes such as pol- viscidium dries out in older flowers and the linator attraction and not to minimize self- pollinaria are consequently not removed if pollination. further visits occur (Singer and Koehler Previous population genetic studies us- 2002). Therefore, in some species of orchids ing allozymes in Lepanthes showed that flowers do go through fully separate male populations are usually more inbred (Fis) and female phases. than one would expected from a random In Lepanthes the shift to receptive female mating system and rarely at Hardy- is clearly not as abrupt as in the previous Weinburg equilibrium (Tremblay and Ack- examples. The change in receptivity is one erman 2001; Carromero, Tremblay, and of proportion of increase fruit set as flowers Ackerman, unpublished data). Conse- age, with a more pronounced effect of quently, the opportunity for breeding with flower age in L. woodburyana as compared closely related individuals is present and to the other two species. Temporal overlap may be a common event (Tremblay and in the male and female flower function is Ackerman 2001, 2003). present in many species (Bertin 1993; Bertin Recent phylogenetic analyses suggest and Newman 1993; Navarro 1997; Fetscher that the most likely sister groups to Lepan- 2001; Offord 2004). Interestingly, the gen- thes are Frondaria, Trichosalpix, Zootrophion, eral explanation for the presence of di- Acianthera and Lepanthopsis (Pridgeon, So- chogamy in plants is to reduce geitonoga- lano, and Chase 2001; Pridgeon and Chase mous pollinations in self-compatible 2001). The relationships between the six plants; however, dichogamy is as common genera are poorly resolved in the consensus in self-compatibility and self-incompa- tree. However, the flowers on inflores- tibility species (Bertin 1993). cences of Lepanthopsis, Trichosalpix, Zoo- It is interesting to note that in Lepanthes, a trophion, Acianthera and Frondaria (mono- group of over 800 neotropical species, al- typic) usually open simultaneously; most all species have sequential flowers. If although single flower species are present the process of sequential flowering is a pri- in some of the genera (Luer 1986, 1991, mary mechanism for reducing self- 1997). Unfortunately we have no informa- pollination then why should plants be self- tion in regards to reproductive biology of incompatible and protandry (at least these five sister genera and thus we do not partially) be present? Which one of the yet know if these species are self- mechanisms, self-incompatibility, sequen- incompatible and if flower age is important tial flowering or protandry was the first to for fruit production. In addition, we do not evolve? It seems that since self-incom- know if simultaneous flowering is derived patibility is present, one might argue that or ancestral. sequential flowering and protandry were Our results show that flowers of the PROTANDRY IN ORCHIDS 79 three species of Lepanthes are self-incom- Borba, E. L., and S. Semir. 1999b. Reproductive sys- patible and a tendency to increase female tems and crossing potential in three species of Bul- receptivity as flower age. The partial bophyllum: (Orchidaceae) occurring in Brazilian ‘campo rupestre’ vegetation. Plant Systematics and protandrous behavior and sequential flow- Evolution 217:205-214. ering may be a relic character prior to the Catling, P. M. 1983. Pollination of northeastern North development of self-incompatibility. How- American Spiranthes (Orchidaceae). Canadian Jour- ever, this assumes that self-incompatibility nal of Botany 61:1080-1093. is more efficient as a mechanism for pre- Dellaporta, S. L., and A. Calderón-Urrea. 1993. Sex venting self-pollination than protandry and determination in flowering plants. The Plant Cell 5:1241-1251. sequential flowering. A further assumption Dudash, M. R. 1990. Relative fitness of selfed and out- is that inbreeding depression (reduce fit- crossed progeny in a self-compatible, protandrous ness) is present when mating among re- species, Sabatia angularis L. (Gentianaceae): a com- lated individuals, otherwise there would be parison in three environments. Evolution 44:1129- no fitness advantage for any of the previ- 1139. ous mechanisms. Because of the small ef- Fetscher, A. E. 2001. Resolution of male-female conflict fective populations size of Lepanthes the op- in an hermaphroditic flower. Proceeding of the Royal Soceity of London, Series B. 268:525-529. portunity for inbreeding is likely to be high Jarne, P., and D. Charlesworth. 1993. The evolution of (Tremblay and Ackerman 2001) and thus the selfing rate in functionally hermaphrodite mechanisms for promoting outcrossing plants and animals. Annual Review of Ecology and may be advantageous. Systematics 24:441-466. Johnson, S. D., and T. J. Edwards. 2000. The structure Acknowledgments.—We would like to and function of orchid pollinaria. Plant Systematics and Evolution 222:243-269. give our appreciation to Mario Blanco and Kearns, C. A., and D. W. Inouye. 1993. Techniques for Gustavo Romero plus two anonymous re- Pollination Biologists. Niwot, Colorado: University viewers for comments and constructive dis- Press of Colorado. cussions on the subject. We appreciate the Lande, R., and D. W. Schemske. 1985. The evolution of financial support from the AMP program self-fertilization and inbreeding depression in to GPH and MLMC. plants. I. Genetic models. Evolution 39:24-40. Luer, C. 1986. Icones Pleurothallidinarum I: Systemat- ics of the Pleurothallidinae (Orchidaceae). Mono- LITERATURE CITED graphs in Systematic Botany from the Missouri Botani- cal Garden 15. Ackerman J. D. 1975. Reproductive biology of Good- Luer, C. 1991. Icones Pleurothallidinarum VIII: Sys- yera oblongifolia (Orchidaceae). Madroño 23:191-198. tematics of Lepanthopsis, Ocotomeria subgenus Ackerman, J. D. 1989. Limitations to sexual reproduc- Pleurothallopsis, Restrepiella, Restrepiosis, Salpistele tion in Encyclia krugii (Orchidaceae). Systematic and Taegueia. Monographs in Systematic Botany from Botany 14:101-109. the Missouri Botanical Garden 39. Ackerman, J. D., and M. R. Mesler. 1979. Pollination Luer, C. 1997. Icones Pleurothallidinarum XV: System- biology of Listera cordata (Orchidaceae). American atics of Trichosalpix. Monographs in Systematic Journal of Botany 66:820-824. Botany from the Missouri Botanical Garden 64. Aragón, S., and J. D. Ackerman. 2001. Density effects Marshall, D. L, and M. W. Folsom. 1991. Mate choice on the reproductive success and herbivory of Ma- in plants: an anatomical to population perspective. laxis massonii. Lindleyana 16:3-12. Annual Review of Ecology and Systematics 22:37-63. Bertin, R. I. 1993. Incidence of monoecy and di- Navarro, L. 1997. Is the dichogamy of Salvia verbenaca chogamy in relation to self-fertilization in angio- (Lamiaceae) an effective barrier to self- sperms. American Journal of Botany 80:557-560. fertilization? Plant Systematics and Evolution 207: Bertin, R. I., and C. M. Newman. 1993. Dichogamy in 111-117. angiosperms. The Botanical Review 59:112-152. Offord C. A. 2004. An examination of the reproductive Blanco, M. A., and G. Barboza. 2005. Pseudocopula- biology of Telopea speciossissima (Protaceae) with tory pollination in Lepanthes (Orchidaceae: Pleroth- emphasis on the nature of protandry. International allidinae) by fungus gnats. Annals of Botany 95:763- Journal of Plant Sciences 165:73-83. 772. Pridgeon, A. M., and M. W. Chase. 2001. A phyloge- Borba, E. L., and S. Semir. 1999a. Temporal variation netic reclassification of Pleurothallidinae (Orchida- in pollinarium size after its removal in species of ceae). Lindleyana 16:235-271. Bulbophyllum: a different mechanism preventing Pridgeon, A. M., R. Solano, and M. W. Chase. 2001. self-pollination in Orchidaceae. Plant Systematics Phylogenetic relationships in Pleurothallidinae and Evolution 217:197-204. (Orchidaceae): Combined evidence from nuclear 80 R. L. TREMBLAY ET AL.

and plastid DNA sequences. American Journal of Goodyerinae orchids from southeastern Brazil. An- Botany 88:2286-2308. nals of Botany 88:989-997. Primack, R. B. 1985. Longevity of individual flowers. Sipes, S. D., and V. J. Tepedino. 1995. Reproductive Annual Review of Ecology and Systematics 16:15-37. biology of the rare orchid, Spiranthes diluvialis: Salazar, G. A., T. Franke, L. Zapfack, and Beenken. breeding system, pollination, and implications for 2002. A new species of Manniella (Orchidaceae, conservation. Conservation Biology 9:929-938. Cranichideae) from western Tropical Africa, with Sosa, V., and M. Rodríguez-Angulo. 2000. Pollination notes on protandry in the genus. Lindleyana 17:239- and natural history of Mormodes tuxtlensis 246. (Catasetinae: Orchidaceae). Lindleyana 15:53-58. Singer, R. B. 2002a. The pollination biology of Sauro- Tremblay, R. L., and J. D. Ackerman. 2001. Gene flow glossum elatum Lindl. (Orchidaceae: Spiranthinae): and effective population size in Lepanthes (Orchi- moth-pollination and protandry in neotropical Spi- daceae): A Case for Genetic Drift. The Biological ranthinae. Botanical Journal of the Linnean Society Journal of the Linnean Society 72:47-62. 138:9-16. Tremblay, R. L., and J. D. Ackerman. 2003. The genetic Singer R. B. 2002b. The pollination mechanism in Tri- structure or orchid populations and its evolution- gonidium obtusum Lindl. (Orchidaceae: Maxillarii- ary importance. Lankersteriana 7:87-92. nae): sexual mimicry and trap-flowers. Annals of Tremblay, R. L., J. D. Ackerman, J. K. Zimmerman, Botany 89:157-163. and R. Calvo. 2005. Variation in sexual reproduc- Singer, R. B., and S. Koehler. 2002. Notes on the pol- tion in orchids and its evolutionary consequences: lination biology of Notylia nemorosa (Orchidaceae): a spasmodic journey to diversification. The Biologi- do pollinators necessarily promote cross pollina- cal Journal of the Linnean Society. 84:1-54. tion? Journal of Plant Research 116:19-25. Walker-Larson, J., and L. D. Harder. 2000. The evolu- Singer, R. B., and M. Sazima. 2001a. The pollination tion of staminodes in angiosperms: patterns of sta- mechanism of three sympatric Prescottia (Orchida- men reduction, loss, and functional re-invention. ceae: Prescottinae) species in southeastern Brazil. American Journal of Botany 87:1367-1384. Annals of Botany 88:999-1005. Warford, N. 1992. Pollination biology: the reciprocal Singer, R. B., and M. Sazima. 2001b. Flower morphol- agreement between Notylia and Euglossa viridis- ogy and pollination mechanism in three sympatric sima. American Orchid Society Bulletin 61:884-889.