Reproductive Ecology of Euglossine Bee-pollinated Orchid Catasetum Integerrimum Hook (Ochidaceae)

Catasetum is a dimorphic and dioecious genus of orchids with a pollinaria release mechanism triggered by pollinator visitation. The reproductive ecology of Catasetum integerrimum Hook (Orchidaceae) was characterized in Central Veracruz, Mexico. For that, it was determined whether oral senescence and fruit production are resulted of pollination eciency in the species, and if so, to what extent present reproductive event constraint the investment in subsequent reproductive event. In the sampled population, ecient pollinaria removal/deposition triggered oral senescence. Eulaema polychroma bee visited earlier and with lower frequency the staminate than pistillate owers. Alleviating pollen limitation increased fruit production but it caused a reduction in ower and fruit production in the subsequent reproductive event. C. integerrimum seemed to exhibit reproductive cost-saving mechanisms linked to the pollinator identity and behavior, whereas pollinator eciency appeared to be mediated by dioecism and dimorphism in oral traits. This study highlights the importance to understand the adaptive signicance of dimorphic oral traits in animal-pollinated plants, stimulating new research avenues on the role of pollinators in maintaining safe reproduction threshold of plants.


Introduction
Neotropical region harbors one orchid genus that display a fascinating adaptation for forcibly attaching the pollinaria onto visiting insects' body by triggering a catapult mechanism. Although catapult mechanism has evolved repeatedly in different angiosperm families, those typically are limited to few Darwin termed "sensitiveness" at this pollination mechanism in Catasetum and he argued that this phenomenon offered a strong evidence for advantages of cross-fertilization and natural selection (Darwin 1862). Regardless of the remarkable pollination mechanism, Catasetum is of particular interest within orchids, due to is one of the few genera of the family in which species are functionally dioecious with a pronounced sexual dimorphism in oral traits (i.e. each individual is able to produce pistillate owers or staminate owers at different times; Romero  Aside from functionally dioecious trait, the owers of the genus produce no nectar as reward, instead, the reward comprises oral fragrance compounds that euglossine bees, especially Euglossa and Eulaema collect and store (Janzen 1981; Romero and Nelson 1986;Whitten et al. 1986; Del Mazo Cancino and Damon 2007; Nunes et al. 2017). In this regard, oral scent in orchid species is of key importance for attracting pollinators and promoting oral visitation constancy through its high speci city (Romero and Nelson 1986). The orchids belonging to Catasetum, Ophrys and Stanhopea genus have been subject of oral fragrance research since 1960s (reviewed by Raguso 2008). Those studies show that scent chemical composition is species-speci c, promoting specialized pollination through using pre-existing biases in the sensory systems of pollinators (Romero and Nelson 1986

Materials And Methods
Field work was carried out during the raining season of 2016-2019. The study area was located in a commercial polyculture of shaded coffee plantation located in Coatepec, Veracruz México (19°28"N, 96°55"W, 1224 m a.s.l). The epiphytic plants of C. integerrimum were located on trunks and branches 2-10 m above ground of trees of Inga jinicuili and Citrus spp. prediction, I recorded the ower longevity of ten pistillate and ten staminate individuals of C. integerrimum in 2016. I randomly selected and tagged with plastic rings three recently opened owers per individual. Of these tagged-owers, the rst ower was bagged with mosquito netting to avoid pollinaria removal/deposition (intact-bagged). The second ower was leave open to natural pollinaria removal/deposition (intact-unbagged). The third ower was hand-pollinated or pollinaria were handremoval (hand-manipulated). I carefully performed pollinaria removal or deposition by using tweezers. Then, I bagged the owers with a mosquito netting to avoid potential effects of oral visitation by pollinators. I surveyed oral longevity daily from the anthesis ( ower opening) to tagged-ower withering.
I used a nested ANOVA for testing differences in oral longevity. In the model, sexual form (pistillate or staminate) was treated as a xed factor and treatment (intact-bagged, intact-unbagged, and handmanipulated) was nested within sexual form (Zar 1999). The response variable was the oral longevity measured in days. The variable was square root transformed to accomplish the parametric assumptions of the analysis (Zar 1999).
Floral display size and pollinator e ciency In fragrance-producing orchids, euglossine bees exhibit a highly specialized pollination e ciency to pick up and deposit male gametes on conspeci c stigmas (Tremblay 1992). To test this prediction, I carried out eld observations on twenty owering individuals Similarly, I used a nested repeated-measures ANOVA for testing differences on fruit production (fruit set; proportion of owers that produced fruits). In the model, type of plant (effectively pollinated or ineffectively pollinated) was the main factor, and treatment (open pollinated or hand-manipulated) was nested in the main factor. Fruit set produced by year was the repeated measure.
Response variables were transformed in order to achieve the parametric assumptions of the analyses.
Counts (number of in orescences or owers) were square root transformed and fruit set was arcsinesquare root transformed.
All statistical analyses were run using generalized linear modeling with StatView and SuperAnova (Abacous Concepts 1996).
Floral display size and pollinator e ciency ¾ The C. integerrimum owers opened in an acropetalous sequence from the base to the top, reaching its maximum oral display size after 1-2 days. C. integerrimum individuals produced either pistillate or staminate owers arranged in one or two in orescences produced by plant. There were not differences in the number of in orescences produced between staminate and pistillate plants (Means ± S.E., 1.30 ± 0.15, N = 10 pistillate individuals and 1.40 ± 0.16, N = 10 staminate individuals, U' = 45, p = 0.70). Nonetheless, staminate individuals produced a higher number of owers than pistillate individuals (Means ± S.E., 6.30 ± 1.12, N = 10 pistillate individuals and 16.90 ± 0.23, N= 10 staminate individuals, U' = 45, p = 0.70).
C. integerrimum owers were visited by Eulaema polychroma (Friese) (Euglossini, Apidae). Males of E. polychrome approached, alighted at the labellum (occasionally at the column), and entered into the owers, scratching the interior surface of ower with their anterior tarsi. In staminate owers, the scratching behavior caused that bees contacted the rostellum lateral extensions, triggering the ejection of pollinaria and attaching them on the thorax of the bee. After forcible pollinaria attachment, the male bees ew away of the plant. In pistillate owers, the scratching behavior did not cause a catapult-out effect, so that bee ew to the next unoccupied and unvisited ower on the same plant (from the base to the top). Bees hovered around staminate or pistillate owers, but they only approached and entered into unvisited owers during the study period.
A total of 106 oral visitation events were observed in C. integerrimum. Of these, 86% were observed in pistillate individuals and 14% in staminate individuals. The waiting time for a given pistillate or staminate ower to be visited by bees was statistically different (Chi = 9.83, p = 0.002). Staminate owers were visited earlier by bees than pistillate owers in the study population (means ± S.E.; 189.75 ± 13.20, N = 91, pistillate owers and 103.13 ± 15.36, N = 15, staminate owers; Fig. 1). All oral visitation events caused that pollinaria were forcefully attached to the body of bees, promoting an early departure of bees from staminate plants. In pistillate plants, a relatively low number of bees were observed carrying pollinaria. Nonetheless, all pollinaria-carrying bees lodged the pollen mass on the stigma of pistillate owers. As a result, pistillate owers that were natural pollinated initiated and achieved capsule production at the end of fruiting season (near to 15% of tagged owers). The remainder pistillate owers wilted two days after being visited by a bee and not being pollinated. ). In C. integerrimum, the reduced life-time of staminate owers after removing pollinaria were similar to the observed in C. uncatum (77% in C. integerrimum and 70% in C. uncatum). Nonetheless, pistillate owers of C. uncatum lasted longer time without wilting than the pistillate owers of C. integerrimum after pollination (68% in C. integerrimum and 83% in C. uncatum). Furthermore, the oral longevity of pistillate owers was reduced to two days after them being visited and non-pollinated by bees.
Early ower senescence induced by pollination may be associated with the reallocation of resources to fruit production by the plant. Clayton and Aizen (1996) mentioned that reduced oral longevity in manual pollinated owers was associated with a resource reallocation mechanism to developing fruits in Chloraea alpina. In Broughtonia lindenii was suggested that staminate ower wilting caused a reallocation of resources to fruit production, whereas pistillate ower wilting avoided re-visitation of already pollinated owers (Vale et al. 2011). Vega and Marques (2014) recognized that oral longevity changes were determined by a trade-off between maintenance costs of oral attractiveness and allocation of resources to fruit production in several Epindendrum species. In C. integerrimum, the oral senescence may be accounted for saving-costs mechanism associated with oral maintenance rather than avoidance of re-visitation of owers. The above because oral visitors exhibited an immediate rejection to land on a ower that was already visited, so that such ower ended to wither in the following days. In C. uncatum was reported the preference of pollinators for visiting non-pollinated and unvisited owers (Milet-Pinheiro et al. 2015). Thus, the use of short-lived scent marks by bees to avoid re-visitation of owers that were already visited by themselves and/or other bees may be possible in the genus and requires future research.
Plants' reproductive success are closely linked with the attractiveness of its oral display to attract effective pollinators. In this study, C. integerrimum displayed a staminate-biased ratio, attracting a single species of pollinator in the sampled population. The observed oral display size and the biased sex ratio concurred with the reported in other Catasetum species (C. ochraveum, Romero  In the study population, male individuals of Eulaema polychroma visited the pistillate and staminate owers of the species. Nonetheless, bees visited earlier the staminate owers but in low frequency than pistillate owers. Del Mazo Cancino and Damon (2007) recognized that staminate owers produced a more diverse fragrance chemistry composition than pistillate owers (15 compounds of oral oils in pistillate owers and 29 compounds of oral oils in staminate owers). Thus, fragrances produced by staminate owers may contribute to explain the early arrival of male bees to those owers. Whilst, the forced launching of pollinaria towards bees during its oral visitation contributed to explain the early departure and low visitation frequency of bees to staminate plants. The aversive behavior of bees caused by the forced attachment of pollinaria is coincident with the reported for several Catasetum species (Janzen 1981 The pollinator e ciency of E. polychroma in C. integerrimum owers depends on matching the morphology and behavior of bees to the oral morphology and ower opening in C. integerrimum. In this sense, bees responded to oral morphology in the species, contacting the sexually structures of owers. Together with oral morphology, C. integerrimum owers opened in acropetalous sequence, matching with the bees' activity pattern within the in orescence in pistillate plants (i.e., successive visits to owers from the base to the top). In staminate plants, the aversive behavior of bees associated with the forced attachment of pollinaria avoided successive visitation of owers, mitigating pollen mass wastage within an in orescence. Thus, bees carrying pollinaria were potentially able to pollinated a distant plant. Overall, there is a matching between C. integerrimum oral traits and behavior of E. polychroma as pollinator. Furthermore, the observed high pollinaria deposition e ciency by euglossine bees in C. integerrimum concurs with the reported by Janzen (1981), who observed that the two bees bearing pollinaria were able to deposited them on the stigmas of C. maculatum owers. In the same study, Janzen (1981) mentioned that only 9% of bees were returned to the same plant, suggesting that bees move over long distances. Johnson et al. (2005) observed a trend accumulation of pollinaria on a moth's proboscis by multiple visits to owers, causing a reduction in pollinaria export, as well as obstruction of pollinaria deposition on stigmas of Disa copperi owers. To my knowledge, Catasetum plants do no experience reproductive costs caused by an excess of pollinaria accumulation on bees during oral visitation. Thus, forced attachment of pollinaria to pollinators' body by catapult mechanism seems to play an important role in reducing reproductive costs caused by pollinators return to the same staminate plant. Tremblay (1992) suggested that low taxonomic number of e cient pollinators per orchid species could be related to its role in diminishing the costs of failure pollen transfer. In this regard, the suitability of E. polychroma as pollinator of C. integerrimum seems to be mediated by oral trait dimorphism and dioecious nature of C. integerrimum. In other words, the oral trait associated with attachment of polliaria to pollinators by catapult mechanism avoided reproductive costs associated with pollen mass wastage by interrupting stereotyped behavior of bees during its foraging within a staminate plant. In pistillate plants, bees bearing pollinaria were able to deposit them on the stigmas of owers in every observed case.
Regardless of the e ciency in depositing pollinaria by bees, C. integerrimum exhibited a low fruit production by pollen limitation that was able to be released through manual pollinations. Nonetheless, the gained advantage in the fecundity by pollen limitation release seemed to lead to an over allocation of resources for fruiting that constrained the next reproductive event by reducing the production of reproductive structures. These results are consistent with the reported by several authors, who have documented that alleviating pollen limitation caused a high fruit production but it constrained future growth and/or reproduction in some tropical orchids (Tipularia discolor, Snow and Whigham 1989; Epidendrum ciliare, Ackerman and Montalvo 1990; Dendrobium monophylum, Bartareau 1995).
While pollen limitation has been a topic of major research interest, little is known about ecological and evolutionary consequences of a potential maximum fruit production threshold in angiosperms. Speci cally, sexual reproduction may confer tness bene ts and opportunities of adaptation, but it also may incur tness costs for reproductive plants, as well as a low quality of produced offspring. Hence, this study highlights the importance to understand that the current pollen-limitation concept may be an oversimpli cation of the optimal threshold of sexual reproduction in animal-pollinated plants, so that it will require a concept re-evaluation at theoretical and procedural level.
The ndings of this study showed that oral senescence and fruit production were resulted from pollen vector e ciency in picking up and depositing male gametes on conspeci c stigmas, as well as in avoiding male gametes wastage mediated by oral traits in C. integerrimum. Although C. integerrimum exhibited a fruit production increase after alleviated pollen limitation, a reduction in both ower number and fruit set production were observed in the subsequent reproductive event. Together, these results were consistent with the predicted hypothesis that C. integerrimum would exhibit traits and cost-saving mechanism associated with its main pollinators' identity and behavior. The prevalent ecological conditions seem to favor the maintenance of low fruit production in C. integerrimum plants, so they may be able to adjust its tness function across entire lifespan.