Reproductive Ecology of Ocotea Catharinensis, an Endangered Tree Species T

Home , Ocotea, Ocotea catharinensis, Ocotea porosa, Ocotea puberula

Plant Biology ISSN 1435-8603

RESEARCH PAPER Reproductive ecology of Ocotea catharinensis, an endangered tree species T. Montagna1 , J. Z. Silva1, T. G. Pikart2 & M. S. Reis1 1Nucleo de Pesquisas em Florestas Tropicais, Universidade Federal de Santa Catarina, Florianopolis, Santa Catarina, Brazil 2 Laboratorio de Entomologia, Centro de Ciencias^ Biologicas e da Natureza, Universidade Federal do Acre, Rio Branco, Acre, Brazil

Keywords ABSTRACT Conservation; floral biology; mating system; phenology; pollination; seed dispersal. • Ocotea catharinensis (Lauraceae) is an endangered tree species from the Brazilian Correspondence Atlantic Rainforest. Currently, little is known about the reproductive ecology of this T. Montagna, Rodovia Admar Gonzaga, 1346. species. Aiming to propose conservation measures, we described aspects related to Florianopolis, Santa Catarina, Brazil. phenology, floral biology, pollination, seed dispersal and mating system of O. cathari- E-mail: [email protected] nensis. • We conducted phenological observations in 62 individuals for 2 years. In one repro- Editor ductive event, we evaluated nectar production, stigmatic receptivity and pollen viabil- Z.-X. Ren ity. Floral visitors were observed, identified and classified on a scale of pollination effectiveness. Seed dispersers were observed and identified using camera traps. Finally, Received: 19 March 2018; Accepted: 16 May the mating system was evaluated through pollen/ovule ratios, experimental pollination 2018 treatments and genetic analysis with molecular markers. • Ocotea catharinensis presented a supra-annual fruiting pattern with a substantial doi:10.1111/plb.12847 reduction of reproducing individuals from bud phase to ripe fruit phase. Several mechanisms prompting cross-fertilisation were identified, such as attractive, herkogamic and protogynic flowers. The main floral visitors and pollinators were from the Diptera order, and all seed dispersers were birds. The species presented a predomi- nantly outcrossed mixed mating system with significant selfing rate (17.3%). • Although based on restricted evidence, we hypothesised that selfing is an escape mechanism for situations unfavourable to cross-fertilisation. Specifically, for the studied population selfing is a response to reduced population size, which is caused by the non-reproduction of all potentially reproductive individuals and by past exploitation events. Therefore, conservation efforts should be able to enhance population sizes, as well as prevent overexploitation.

et al. 2014). Pollination and seed dispersal ecology studies can INTRODUCTION provide valuable information related to gene flow, such as vec- Knowledge of plant reproductive ecology is critical to predict tors, distances and possible limitations (e.g. Dick et al. 2003; population survival ability and to propose viable conservation Danieli-Silva & Varassin 2013). Mating system studies, in turn, efforts in the case of endangered species (Bernardello et al. examine how individuals from a given species cross to form the 2001; Newton 2007). Studies on reproductive ecology are com- progeny. Consequently, it is possible to estimate, for instance, prehensive, often reporting on aspects related to phenology, outcrossing and selfing rates, number of pollen donors and the floral biology, pollination and seed dispersal ecology and mat- degree of inbreeding in the progeny (e.g. Conte et al. 2008; ing system (e.g. Mariot et al. 2014; Custodio et al. 2017). Such Rogalski et al. 2017). data are fundamental to a complete understanding of repro- All these study fields are deeply connected and provide use- ductive processes of a given species, including possible ecologi- ful knowledge to identify possible factors that can reduce the cal relationships with other organisms, such as pollinators or reproductive ability of a given species. Thus, conservation pro- seed dispersers. posals can best be supported when based on reproductive ecol- Phenological studies refer to recurrent (cyclic) events and ogy studies. For instance, one mating system parameter, their timing in plant species (Newstrom et al. 1994). Such stud- effective variance size, is commonly used to determine how ies allow us to know the onset, peak, ending, intensities and many seed trees (or seed plants) are necessary for seed collec- variations between individuals, populations and years of flow- tion in order to retain a given effective size (Hmeljevski et al. ering and fruiting periods of a given species (e.g. Grogan & 2011; Ferreira et al. 2012). Furthermore, pollination vectors Loveless 2013). Floral biology is concerned with flowering can alert researchers to the need for conservation efforts, espe- (Percival 1965), and studies in this field can provide informa- cially for species with few, or even single, pollinators, such as tion about nectar production, stigmatic receptivity, pollen via- figs (Janzen 1979) or Ocotea porosa (Danieli-Silva & Varassin bility and floral morphology (e.g. Kubitzki & Kurz 1984; Pacini 2013).

926 Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Montagna, da Silva, Pickart & dos Reis Reproductive ecology of Ocotea catharinensis

Ocotea catharinensis is a long-lived tree species that occurs in flowering (buds and open flowers; Fig. 3) and the fruiting (un- the Brazilian Atlantic Rainforest. This climax species is dis- ripe and ripe fruits) using binoculars. We applied the method tributed between latitudes 19°570 S (Saiter & Thomaz 2014) proposed by Fournier (1974) to score the intensity of pheno- and 30°150 S (Carvalho 1994). O. catharinensis can reach 30 m logical events, attributing 0 for individuals without phenologi- in height and can surpass 100 cm diameter at breast height cal activity, and 1, 2, 3 and 4 for individuals with 1–25, 26–50, (DBH) when adult (Reitz et al. 1978). It was considered as the 51–75 and 76–100% of phenological intensity, respectively. dominant and most abundant tree in the higher stratum of Phenological index to each phenophase was obtained by sum- Ombrophilous Dense Forest (Veloso & Klein 1959; Klein ming all intensity values (6¼0) and averaging by the number of 1980). In spite this dominance, several aspects O. catharinensis individuals presenting phenological activity during a given reproductive ecology remain poorly understood. For instance, month, modifying the proposal of Bencke & Morellato (2002). the literature contains little, to no, information regarding nec- We compared the average DBH between reproductive and tar production, stigma receptivity, pollen viability and floral non-reproductive individuals through a confidence interval visitors. Furthermore, the mating system was only indirectly (CI) (95%) obtained with 1000 bootstrap replicates. Possible estimated (apparent outcrossing rate; Tarazi et al. 2010), and relationships between phenological events, including monthly only one study focused on phenological aspects of this species frequency of individuals presenting buds, open flowers, unripe (Silva et al. 2000). or ripe fruits and climatological data, including monthly aver- This species was intensively exploited for timber production ages of mean temperature, maximum temperature and mini- (Reitz et al. 1978), and it resulted in severe reductions to popu- mum temperature and total monthly rainfall, were investigated lation sizes in Santa Catarina state (Schorn et al. 2012). Because through Spearman correlation. Climatological data were of the exploitation processes and the high fragmentation of the obtained from an official meteorological station 18 km from Atlantic Rainforest (Ribeiro et al. 2009), O. catharinensis cur- the study area. rently figures on the Brazilian List of Endangered Plant Species (MMA 2014) and also on the IUCN Red List (Varty & Guadag- Floral biology nin 1998), classified as vulnerable. Furthermore, a recent study has shown that O. catharinensis presents low recruitment into Floral biology aspects, such as nectar production, stigmatic the seedling category, potentially harming the population receptivity and pollen viability, were evaluated using three dynamics in the long term (Montagna et al. 2018). metallic towers (11–13 m in height) installed near three indi- The endangered condition of O. catharinensis demands stud- viduals in order to gain access to flowers. Due to the steep ies to better understand aspects of its reproductive ecology. Nev- topography at the study site, we were able to install only three ertheless, comprehensive knowledge gained from reproductive towers. Standing crop nectar was evaluated every 2 h, from ecology, although difficult to achieve, is very relevant for the tar- 05:00 to 23:00 h, in ten different flowers from three individuals get species of our study. Therefore, in order to propose feasible per evaluation using a 1 ll microcapillary tube. In each of these conservation actions, we investigated the following aspects of evaluations, which were conducted over 5 days, the air temper- O. catharinensis: (i) reproductive phenology; (ii) floral biology: ature and the presence or absence of floral scent was also nectar production, stigmatic receptivity and pollen viability; recorded. Nectar soluble solids concentration (24 flowers from (iii) floral visitors and pollinators; (iv) seed dispersers; and three individuals) was measured with a portable refractometer (v) mating system. Some of these aspects are described here for only when the volume of nectar extracted was >0.5 ll needed the first time, such as nectar production, stigmatic receptivity, to allow refractometer measurement. pollen viability, floral visitors and mating system. Stigmatic receptivity was first verified using the Sudan III- glycerin test (Dafni 1992) on buds (15 stigmas, five of each individual) and open flowers (15 stigmas, five of each individ- MATERIAL AND METHODS ual). Later, it was observed that stigmatic receptivity remained until stigma oxidation; therefore, we adopted the criterion of Study area stigma oxidation to evaluate stigmatic receptivity. To verify This study was conducted in a protected area in Santa Catarina, whether the species presents dichogamy (protogyny or protan- the Floresta Nacional de Ibirama (Ibirama National Forest – dry), 102 flowers were collected from three individuals (43 fully FNI; 27°01059.0000 S, 49°27020.0000 W). The FNI comprises opened and 59 partially opened) and classified according to 570 ha of Ombrophilous Dense Forest in advanced stage of sec- stigma oxidation, i.e. not oxidised, partially oxidised or oxi- ondary succession (MMA 2008), and the climate is described dised, and anther opening, either open or closed, using a stere- as subtropical humid (Cfa), according to Koppen’s€ classifica- omicroscope (magnification 49). Pollen viability was tested tion. Studied individuals (DBH > 20 cm) were distributed using the aceto-carmine test (Dafni 1992). Pollen grains were across 15.5 ha (350 9 430 m) with a density of 4.1 individu- collected and placed on microscope slides, and then a drop of alsha1. Selective logging of O. catharinensis was reported for aceto-carmine (2%) was added. Using an optical microscope the study site in the 1950s (MMA 2008). (magnification 129), pollen staining was observed in four places on each microscope slide. This procedure was repeated for 15 flowers (five of each individual). Reproductive phenology Phenological observations were carried out from January 2016 Floral visitors and pollinators to December 2017 (except in February 2016 and 2017 and July 2017 due to logistical difficulties). In this period, 62 individu- Floral visitors were observed in two of the three individuals als, with DBH > 20 cm, were observed monthly for the from the metal towers because of the low flowering intensity of

Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 927 Reproductive ecology of Ocotea catharinensis Montagna, da Silva, Pickart & dos Reis one individual. These observations were made over 9 days in individuals), and spontaneous self-pollination, in which buds August and September 2016, from 07:00 to 00:00 h, using the were marked, bagged with voile and after the end of the flower- following methodology. First, only one visitor was observed at ing period, bags were removed (1750 buds among three indi- a time, even if other visitors were in the tree crown. Second, as viduals). Buds were marked and bagged in July 2016 and soon as the visitor interacted with a flower, it was pho- monitored until September 2017, the time of last ripe fruit. We tographed, and the air temperature was recorded. Third, for were only able to implement two pollination treatments, each visitor, the number of visited flowers, time spent in each because even minimal manipulations caused the oxidation of flower and whether the visitor touched or not the reproductive buds. Independence of fruit set and treatments was tested with structures (stigma and anthers) were described. Fourth, when- the Fisher’s exact test. ever possible, visitors were collected with a dip net. Open-pollinated fruits were collected under the crown pro- The collected visitors were observed and photographed with jection of 12 seed trees in October 2017. Fruit collection was a stereomicroscope in order to identify pollen of O. catharinen- carried out at ground level, due to the difficulty of climbing sis attached to their bodies. Furthermore, each visitor was seed trees. Fruits were pulped, and then seeds were sown in ranked on a scale of pollination effectiveness based on whether containers and maintained under shade net (50% shade) with (i) pollen was attached to the body; (ii) the reproductive struc- constant irrigation. At least 40 seeds from 11 seed trees plus tures of flowers had been touched; and (iii) the frequency of eight seeds from one seed tree were sown, totalling 707 seeds. visits was higher than the median of all frequencies of visits. Open-pollinated progeny arrays (hereinafter progeny: offspring Therefore, a visitor was classified as “pollinator” if the three from one seed tree) and seed trees were genotyped with allo- answers were positive; as “potential pollinator” if two answers zyme markers, suitable for the analysis of mating system were positive; “eventual pollinator” if only one answer was pos- (Conte et al. 2008), in maize starch gel (Penetrose 30, 13%). itive; and “visitor” if all answers were negative. We resolved the following enzymatic systems in a Tris-citrate pH 7.5 buffer (Tris 27 gl1 and citric acid 16.52 gl1): 6- phosphogluconate dehydrogenase (6PGDH, Enzyme Commis- Seed dispersers sion 1.1.1.44), a-esterase (aEST, EC 3.1.1.1), diaphorase (DIA, Seed dispersers were observed at tree crown level, aiming to EC 1.8.1.4), malic enzyme (ME, EC 1.1.1.40), phosphogluco- record the primary dispersers, and at ground level, aiming to mutase (PGM, EC 5.4.2.2), malate dehydrogenase (MDH, EC record secondary dispersers. Observations at crown level were 1.1.1.37), peroxidase (PO, EC 1.11.1.7) and shikimate dehydro- â carried out using a camera trap (Bushnell Trophy Cam 8mp ) genase (SKDH, EC 1.1.1.25). installed in one of the metallic towers and programmed to The mating system was analysed under the mixed-mating shoot 30-s videos with minimum intervals of 2 min. The film- model (Ritland & Jain 1981) and the correlated mating model ing started in August 2017, still with unripe fruits in the crown, (Ritland 1989), both implemented in MLTR, version 3.4 (Rit- and was extended until November 2017 when the observed land 2002). The following parameters were estimated at the ^ individual no longer presented ripe fruits. Primary dispersers population level: multilocus (tm – also estimated at progeny ^ were also opportunistically observed, with binoculars and pho- level) and single locus (ts) outcrossing rates, selfing rate ^ ^ tographic camera, along the phenological evaluations. In order (^s ¼ 1 tm), mating among relatives plus selfing rate (1 ts), ^ ^ to record secondary dispersers, about 200 ripe fruits were col- mating among relatives or biparental inbreeding rate (tm ts), lected and deposited grouped in the soil under the crown of multilocus paternity correlation (^r ), and inbreeding coeffi- ^ pðmÞ two individuals of O. catharinensis, about 100 fruits for each cient in maternal generation (Fm). individual. In each place where fruits were deposited, a camera From the aforementioned parameters, we were able to esti- trap was installed with programming identical to that previ- mate other parameters, as follows: effective number of pollen ously mentioned. Ground observations began on October 2017 donors (N^ ¼ 1=^r ) (Ritland 1989) and, according to Seb- ep pðmÞ ^ and lasted until November 2017. benn (2002), the proportions of self-sibs (Pss ¼ ^s), half sibs ^ ^ ^ ^ ^ ^ (Phs ¼ tmð1 rpðmÞÞ) and full-sibs (Pfs ¼ tm rpðmÞ). In addition to the inbreeding coefficient in maternal generation, we Mating system also estimated the total inbreeding coefficient in progenies ^ Three methodologies were used to describe the mating system (F0 ¼ F), using the method described by Weir & Cockerham of O. catharinensis: pollen–ovule ratios (Cruden 1977), experi- (1984) and implemented in FSTAT (Goudet 2002); inbreeding ^ ^ mental pollination treatments and genetic analysis of open-pol- coefficient in progenies owing to selfing (Fs ¼ 0:5 ^sð1 þ FmÞ) linated progeny arrays with molecular markers. Flower buds (Barrett & Kohn 1991); and inbreeding coefficient in progenies were collected from three individuals (five buds each), and five resulting from mating among relatives (F^ ¼ F^ F^ ). tm^ts 0 s anthers of each bud were excised and stored in 500 ll lactic Average coancestry within progenies (Hxy) was estimated acid (85%). Subsequently, the anthers were macerated, and dividing by two the relatedness correlation within progenies as ^ : ^ ^ ^2 ^ ^ ^ ^ pollen count was performed in a Neubauer chamber, using an rxy ¼ 0 250ð1 þ FmÞ½4 s þðtm þ tm s rsÞð1 þ rpðmÞÞ (Rit- optical microscope (magnification 129). The pollen count was land 1989), where ^rs represents the selfing correlation. Based repeated four times for each bud. The number of ovules was on the coancestry coefficient, the effective variance size was ^ : = H^ = ^ = determined from nine flowers (three of each individual) by estimated as NeðvÞ ¼ 0 5 f xy½ðn 1Þ nþð1 þ F0Þ 2ng longitudinally cutting the carpels and analysing them under a (Cockerham 1969), in which n is the sample size (assumed to stereomicroscope (magnification 49). be 17 – average overall seed trees). Finally, we estimated the ^ =^ Two experimental pollination treatments were implemented: minimum number of seed trees as m ¼ NeðrefÞ NeðvÞ (Sebbenn natural pollination, in which flowers were not manipulated, 2002) from which it would be necessary to collect seeds, aiming but marked from the bud phase (1839 buds among three to retain a reference effective population size (Ne(ref)) of 100. An

928 Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Montagna, da Silva, Pickart & dos Reis Reproductive ecology of Ocotea catharinensis effective population size of 100 is expected to avoid inbreeding the average DBH of non-reproductive individuals depression in the short term (five generations; Frankham et al. (DBH = 39.9, CI 31.8–48.9 cm). The reproductive individual 2014). Confidence intervals (95%) for each parameter were with lower DBH presented 22 cm, indicating that individuals obtained from 1,000 bootstraps within progenies. with DBH near this value have potential to reproduce. Spear- man correlation between phenological events and climatological data was statistically significant, but only for bud and RESULTS flower phases against monthly averages of mean temperature, maximum temperature and minimum temperature (Table S2). Reproductive phenology Phenological observations were sufficient to evaluate a com- Floral biology plete reproductive event, which lasted 19 months, from June 2016 (first buds) up to December 2017 (last ripe fruits; Fig. 1). Average standing crop nectar production was 0.09 ll and Bud phase occurred between June and October 2016, while recorded only from 05:00 up to 09:00 h (Table S3). In spite of flowering lasted 5 months (August to December 2016). Unripe the low average, standing crop nectar production reached max- fruit phase started in October 2016 and lasted 12 months (until imum volumes of 0.79 and 0.62 ll (Table S3). Pearson correla- October 2017) and, finally, ripe fruits were recorded for tion between mean temperature and standing crop nectar 4 months (August to November 2017). A second flowering per- production was statistically significant (r = 0.72 P < 0.05). iod (June to September 2017) was recorded at the same time as Flowers presented a sweet scent throughout all evaluations, but the bud phase, but in less than 10% of individuals. At the end this was more conspicuous during the hottest hours. Nectar of September 2017, all these flowers had aborted (Fig. 1a). soluble solids concentration averaged 6.5° Brix (SD = 0.9° In total, 43 individuals (70% of 62) presented some kind of Brix). reproductive structure. Nevertheless, only 21 individuals (34% Sudan III-glycerin test (Dafni 1992) revealed 100% of stig- of 62) produced ripe fruits. Besides the reduction of reproduc- matic receptivity before and after anthesis. Stigmas remained ing individuals, phenological index also reduced from bud receptive until oxidation; nevertheless, it was not possible to (peak at 2.4; Fig. 1b) to ripe fruits phase (peak at 1.4; Fig. 1e). determine the average time between anthesis and oxidation. Phenological index also presented a high variation between Based on the observations in fully opened and partially opened individuals, as evidenced by the SD values for each phenophase flowers, it was possible to describe the occurrence of protog- (Fig. 1b–e). Reproductive individuals presented average DBH yny. All flowers with open anthers presented oxidised (76) or of 48.5 cm (CI 43.2–54.1 cm), not statistically different from partially oxidised (13) stigmas, while all flowers with non-oxidised stigmas had closed anthers (13) (Fig. 2). Pollen had 100% viability according to the aceto-carmine test (Dafni 1992). Pollen grains were also highly sticky, easily adhering to tweezers and to other pollen grains, forming pollen aggregates. The flowers are bisexual, radially symmetric, with six tepals, six nectaries, six introrse stamens, three extrorse stamens and three staminodes (Fig. 3). Each anther has four valves. The six nectaries are kidney-shaped, positioned at the base of the filament of the extrorse stamens, in pairs. According to the proposal of Schmid (1988), the nectaries can be classified as filament nectaries. The stamens surrounding the stigma present extrorse dehiscence (Fig. 3); therefore, there is a spatial separation between stigma and anthers, characterising herkogamy.

Floral visitors The observations were conducted for 63 h 40 min, including 44 h 25 min of diurnal observations and 19 h 15 min of noc- turnal observations. In this period, 220 visits of insects to the species flowers were recorded. Insects from the Diptera order represented 79.5% of all visits, followed by Lepidoptera (13.2%), Hymenoptera (4.2%), Coleoptera (2.2%), Neuroptera (0.5%) and Dermaptera (0.5%). The highest frequency of visits occurred from 10:00 to 15:00 h, with predominance of Diptera. In this period, it was possible to see several insects visiting flowers at the same time. A smaller peak of visitation was observed from 18:00 to 22:00 h, and during this period, the visitors were exclusively from the Lepidoptera order (Figure S1). Fig. 1. Reproductive phenology of Ocotea catharinensis in percentage of The insects were classified into 70 morphospecies, 34 Diptera individuals presenting a given phenophase (a). Phenological index to each (48.6%), 23 Lepidoptera (32.9%), six Hymenoptera (8.6%), phenophase: buds (b), flowers (c), unripe fruits (d) and ripe fruits (e). Bars five Coleoptera (7.1%), one Dermaptera (1.4%) and one Neu- represent SD. roptera (1.4%) (Table S1). Regarding pollination effectiveness,

Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 929 Reproductive ecology of Ocotea catharinensis Montagna, da Silva, Pickart & dos Reis

Table 1. Mating system estimates for a population of Ocotea catharinensis. recorded interacting with the fruits at the ground; however, this species did not swallow, pulp or transport any fruit. estimates (confidence parameter interval) Mating system Number of seed-trees [total number of 11 [197] offsprings] The results indicate that O. catharinensis presents a predomi- ^ Multilocus outcrossing rate (tm) 0.827 (0.696/0.953) nantly outcrossed mixed mating system. The species had, on aver- ^ Single-locus outcrossing rate (ts) 0.741 (0.574/0.966) age, 1808 pollen grains (SD = 1693) and one ovule per flower. Selfing rate (^s) 0.173 (0.047/0.304) Thus, the pollen/ovule ratio was 1808, closer to that obtained by ^ Mating among relatives and selfing (1 ts) 0.259 (0.034/0.426) Cruden (1977) for facultative xenogamy (796.6 87.7) than to ^ ^ Mating among relatives (tm ts) 0.086 (0.029/0.150) the ratio for xenogamy (5859.2 963.5). The 1750 buds submit- ^ Multilocus paternity correlation (rpðmÞ) 0.247 (0.051/0.432) ^ ted to spontaneous self-pollination yielded five unripe fruits but Effective number of pollen donors (Nep)4 ^ no ripe fruits. Natural pollinated buds (1839) formed 38 unripe Self-sibs proportion (Pss) 0.173 (0.047/0.304) ^ fruits and one ripe fruit. Fruit set was not significantly associated Half-sibs proportion (Phs) 0.623 (0.429/0.865) = ^ with treatments according to Fisher’s exact test (P 0.88), Full-sibs proportion (Pfs) 0.204 (0.044/0.333) although a trend of higher fruit production was observed in indi- Inbreeding in maternal generation (^F ) 0.017 ( 0.200/0.170) m viduals submitted to natural pollination treatment. Inbreeding in progeny (^F ) 0.146 (0.078/0.237) 0 Mating system estimates obtained from progeny analysis are Inbreeding in progeny from selfing (^F ) 0.088 (0.021/0.154) s given in Tables 1 and 2. The multilocus outcrossing rate esti- Inbreeding in progeny from mating among 0.058 (0.008/0.125) ^ ^ mate statistically differed from unity (tm = 0.827), thus deter- relatives (Ftmts) ^ mining a mixed-mating system and a significant selfing rate Coancestry within progeny (Hxy) 0.203 (0.136/0.249) ^ (^s = 0.173). Mating among relatives rate (^t ^t = 0.086) was Effective variance size (NeðvÞ) 2.220 (1.864/3.096) m s Number of seed-trees for seed collection 45 not statistically significant, unlike the multilocus paternity cor- ^ =^ ^ = ^ (m ¼ 100 NeðvÞ) relation (rpðmÞ 0.247). Based on rpðmÞ, the average effective number of pollen donors for each progeny array was estimated to be four. Multilocus outcrossing rate estimates presented three morphospecies were classified as “pollinator”, 25 as “po- high variation among progenies, ranging from 0.452 up to tential pollinator”, 35 as “eventual pollinator” and seven as 1.091 (1). “visitor” (Table S1). All three pollinators belong to the Diptera On average, progenies were composed of 17.3% self-sibs, order: Lauxaniidae sp. 05, Sciaridae sp. 01 and Syrphidae sp. 62.3% half-sibs and 20.4% full-sibs. Inbreeding coefficient in 01. Another five morphospecies presented pollen attached to maternal generation was not significant (F^ = 0.017), unlike the ^ m ^ their bodies, but these were not classified as pollinators because inbreeding in progenies (F0 = 0.146). The partition of F0 of the low frequency of visits (Table S1). resulted in a significant inbreeding in progenies from selfing ^ (FS = 0.088) and a non-significant inbreeding in progenies from mating among relatives (F^ = 0.058). Coancestry Seed dispersers ^ tmts within progenies (H = 0.203) was higher than expected in xy ^ Five bird species were recorded dispersing O. catharinensis fruits, completely panmictic progenies (half-sibs, Hxy = 0.125). In four of them at the tree crown (primary dispersers) and the order to retain effective size of 100 individuals, seeds must be other at ground level (secondary disperser). Primary dispersers collected from 45 seed trees. were Carpornis cucullata (Contigidae), Turdus albicollis, T. flavipes (Turdidae), and Selenidera maculirostris (Ramphasti- dae). The secondary disperser species was Geotrygon montana DISCUSSION (Columbidae). All these species were recorded/sighted swallow- ing O. catharinensis seeds. A rat species (Cricetidae family) was Reproductive phenology Ocotea catharinensis presented a supra-annual fruiting pattern, ^ a characteristic already described for this (Silva et al. 2000) and Table 2. Multilocus outcrossing rate (tm) and its respective standard devia- other Lauraceae species (Wheelwright 1986). The reduction in tion (SE) for each evaluated seed-tree of Ocotea catharinensis. reproducing individuals from bud phase to ripe fruits phase ^ ^ seed-tree [progeny size] tm SE (tm) and the high variation in flowering and fruiting intensities between individuals were also described for Lauraceae species, c19 [16] 1.091 0.001 including O. catharinensis (Wheelwright 1986; Silva et al. c64 [16] 1.038 0.065 2000). The occurrence of these features, i.e. reduction of repro- c18 [24] 1.028 0.031 ducing individuals and high variation in phenological indices, cp3 [20] 0.998 0.032 can represent a serious restriction to population effective size. c57 [17] 0.887 0.074 The start of flowering was linked to a drop in the tempera- c62 [24] 0.883 0.079 tures, mean, maximum and minimum, as evidenced by Spear- c58 [24] 0.861 0.093 man correlations. Similar correlations were found for c46 [10] 0.820 0.105 O. porosa, which presented bud phase negatively correlated r15 [15] 0.649 0.110 = c23 [15] 0.624 0.151 with mean temperature (r 0.51; Bittencourt 2007). Never- c66 [10] 0.452 0.414 theless, the confirmation of this pattern requires longer phenological observations since flowering can start in different

930 Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Montagna, da Silva, Pickart & dos Reis Reproductive ecology of Ocotea catharinensis

Lauraceae species (Kubitzki & Kurz 1984; Danieli-Silva & Varas- sin 2013), and herkogamy. Nevertheless, none of these traits was efficient in completely preventing self-fertilisation, as evidenced by the significant estimated selfing rate. Standing crop nectar production was detected only from 05:00 up to 09:00 h. After this period, increases in mean temperature and frequency of floral visitors, especially dipterans, were observed. If nectar is produced after 09:00 h, these observations suggest that it is evaporated due to the increase in temperature or is collected by visitors. It should be emphasised that the volume of nectar produced by O. catharinensis flowers can be considered high in comparison to other Lauraceae species. For instance, the amount of nectar produced by Laurus azorica was too small to be extractable with 1 ll micropipettes (Forfang & Olesen 1998). A similar situation was described for L. nobilis (Pacini et al. 2014). The production of nectar represents a large investment of energy (Pyke 1991; Pacini & Nepi 2007); therefore, the secretion of nectar in O. catharinensis can be linked to a dependence on biotic vectors to perform pollination. Regarding pollen presentation, although not evaluated, stickiness may result from the presence of pollenkit, an adhe- sive substance commonly present in pollen grains of Lauraceae Fig. 2. Schematic illustration representing sampled flowers (open and par- species (Hesse & Kubitzki 1982; Pacini et al. 2014). tially open) and the condition of their anthers (open or closed) and stigmas According to Kubitzki & Kurz (1984), wide open flowers (oxidized, partially oxidised or not oxidised). with synchronous dichogamy and nectar production would represent a basic state in Lauraceae. In synchronised dicho- seasons for O. catharinensis, including spring and summer, as gamy, self-fertilisation is not possible because all flowers of a well as winter (Silva et al. 2000). given individual are simultaneously in the same sexual phase, with receptive stigma and closed anthers or with wilted stigma and open anthers – protogyny; nevertheless, the occurrence of Floral biology receptive stigmas and open anthers is synchronised between The floral characteristics described for O. catharinensis demon- individuals, allowing cross-fertilisation to be mediated by polli- strated that the species uses several mechanisms to promote nators (Kubitzki & Kurz 1984). From these hermaphroditic cross-fertilisation. The flowers presented typical traits of gener- nectar-producing ancestors, two lines of specialisation have alised pollination, such as radially symmetrical flowers and evolved, according to the same authors: (i) species with low exposed reproductive organs (Fenster et al. 2004). Furthermore, pollen/ovule ratios, hermaphroditic flowers, synchronous flowers produced nectar and scent, which are attractive to insects dichogamy, but no nectar production, and (ii) dioecious spe- (Fægri & van der Pijl 1979). Flowers also presented barriers to cies, with high pollen/ovule ratios and nectar production. self-fertilisation, such as protogyny, already described for other Although we were not able to describe the occurrence of

Fig. 3. Ocotea catharinensis. (a) Partially open flower (ST, oxidized stigma; #, extrorse stamens). (b) Open flower (TE, tepals; *, nectaries; +, introrse stamens). (c) Flowering sequence of O. catharinensis: 1, 2, and 3, buds; 4, partially open flower already with oxidised stigma; 5 and 6, open flowers; 7, open flower with non- functional stamens, stigma, and nectaries.

Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 931 Reproductive ecology of Ocotea catharinensis Montagna, da Silva, Pickart & dos Reis synchronised dichogamy, O. catharinensis seems to better fit Seed dispersers into the ancestral group. However, even if the species exhibits synchronised dichogamy, this mechanism was not completely All the birds identified as seed dispersers of O. catharinensis efficient in preventing self-fertilisation. were also recorded feeding on fruits of several other forest species. Fruits from 35 and 11 species were consumed by T. albicollis and T. flavipes, respectively (Castro et al. 2012). Carpornis Floral visitors cucullata was recorded feeding on fruits from 45 species (Fran- Flowers of Lauraceae species are visited by insects belonging to cisco & Galetti 2002), while S. maculirostris was recorded feed- several orders, such as Hymenoptera (Kubitzki & Kurz 1984), ing on fruits from at least 40 species (Galetti et al. 2000). Coleoptera (Forfang & Olesen 1998; Dupont 2002), Lepidoptera Geotrygon montana feeds on fruits from Euterpe edulis (Are- (Pacini et al. 2014), Diptera (Nobrega et al. 2016), Hemiptera caceae; Galetti et al. 2015) and Margaritaria nobilis (Euphor- (Souza & Moscheta 1999) and Thysanoptera (Danieli-Silva & biaceae; Cazetta et al. 2008). Varassin 2013). Nevertheless, hymenopterans seem to be the more At the study site, seedlings from O. catharinensis present an frequent and diverse floral visitors to Lauraceae species. The pre- aggregated spatial pattern; however, this pattern is independent of dominance of Diptera as major floral visitors, herein described the spatial distribution of reproductive individuals (Montagna for O. catharinensis, was also recorded to O. duckeii (Nobrega et al. 2018). Furthermore, a large quantity of deteriorated seeds et al. 2016). Pollinator type is often related to nectar sugar com- were found rotten and infested by insect larvae under the crown position, either sucrose or hexoses (Baker & Baker 1983; Pacini & of the few individuals that had high seed production. These facts Nepi 2007), and Diptera insects are linked to sucrose-poor/hex- highlight the importance of seed dispersers in transporting seeds ose-rich nectar (Baker & Baker 1983; Wolff 2006). to appropriate sites for germination and seedling establishment. It The pollination effectivity of dipterans is limited by their is worth noting that neither of the species cited in the literature as body size, i.e. small insects carry small pollen loads, but their O. catharinensis seed dispersers (Brachyteles arachnoides and Pipile number can compensate for this (Fægri & van der Pijl 1979). jacutinga) were recorded in the present study. Moreover, Diptera is considered the second most important order of flower-visiting and -pollinating insects (Larson et al. Mating system 2001). According to visiting behaviour and frequency and the presence or absence of pollen attached to the body, three Mating system estimates for O. catharinensis are consistent dipterans were classified as pollinators of O. catharinensis: with the estimates already obtained for other Lauraceae species. Lauxaniidae sp. 05, Sciaridae sp. 01 and Syrphidae sp. 01. Addi- Ocotea porosa presented fruit set of 5% under spontaneous tionally, from the 25 potential pollinators, 19 belong to the self-pollination treatments (Danieli-Silva & Varassin 2013). Diptera order. Insects belonging to Lauxaniidae, Sciaridae and Based on molecular markers, O. tenera was classified as pre- Syrphidae families are recognised as pollinators of several dominantly outcrossed (Gibson & Wheelwright 1996), as was angiosperms (see Herrera 1987; Ollerton & Liede 1997; Larson Cryptocarya moschata (de Moraes & Monteiro 2002). High et al. 2001). Also, insects of these families can carry pollen over rates of outcrossing are also common to several tropical forest long distances, such as 200 m (Wratten et al. 2003) or even species (Ward et al. 2005). 400 m (Rader et al. 2011). Nevertheless, it should be stressed A large portion of the progeny arrays was composed of ^ ^ that the mentioned studies were conducted in open-field envi- full-sibs (Pfs = 20.4%) and self-sibs (Pss = 17.3%). These pro- ronments, which present fewer barriers to insect flights com- portions resulted from the significant paternity correlation pared to forests. Furthermore, flight distance distributions are (full-sibs – same father and same mother) and also from the strongly leptokurtic, i.e. shorter flights are much more com- significant selfing rate (self-sibs). Self-fertilisation acts by mon than longer flights (Herrera 1987; Wratten et al. 2003). increasing the frequency of homozygotes, essentially reducing Lepidoptera was the second most frequent and diverse order genetic variability (Wright 1931). Thus, in the long term, self- of flower-visiting insects of O. catharinensis. Nevertheless, lepi- fertilisation can be detrimental to the adaptive potential of the dopterans do not seem to be efficient pollinators for the follow- species through the loss of variability (Frankham 2005). On the ing reasons. None of captured insects presented pollen other hand, the described self-compatibility can be understood attached to their body, and only three insects visited the flowers as an escape for situations in which the production of seeds more than once. Furthermore, lepidopterans often acted as and progenies is maintained, even if it is not possible to cross. nectar thieves, inserting their long proboscides when clinging This possibility is discussed by other authors for other predom- to the flower pedicel or tepals without touching anthers or inantly outcrossed species (Nazareno & dos Reis 2012; stigma. None of collected hymenopterans presented pollen Danieli-Silva & Varassin 2013). attached to their body, and five of the six identified morphos- Two main lines of evidence support the hypothesis that self- pecies visited flowers only once. Therefore, in spite of their fertilisation is an escape mechanism for situations unfavourable recognised importance as pollinators (Fægri & van der Pijl for cross-fertilisation. The first is related to the population size. 1979), hymenopterans cannot be considered effective pollina- Ocotea catharinensis populations from Santa Catarina went tors of O. catharinensis. Finally, among the other sampled through intensive timber exploitation, especially in the last cen- insects, it is worth mentioning Chrysopidae spp. 01. This mor- tury (Reitz et al. 1978). This exploitation process caused severe phospecies was classified as a potential pollinator because it reductions in the size of O. catharinensis populations. According presented pollen attached to its body. Insects belonging to the to the last estimates for Santa Catarina, the species occurs at an Chrysoperla genus are described as pollenophagous when adults average density of 5.86 individualsha1 (DBH > 10 cm; Lingner (Li et al. 2008). Thus, pollination could occur by chance while et al. 2013). However, in the past, higher densities were recorded, these insects feed on O. catharinensis pollen. such as 23.9 individualsha1 (DBH > 12.7 cm; Veloso & Klein

932 Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Montagna, da Silva, Pickart & dos Reis Reproductive ecology of Ocotea catharinensis

1959). Thus, pollen availability can be affected by the fact that large quantity of deteriorated seeds, i.e.rottenandinfestedby the populations nowadays are smaller than in the past. It is worth insect larvae, was observed under the crown of the few individu- noting that O. catharinensis occurs with a density of 4.1 individuals that presented high seed production. Therefore, the conserva- alsha1 in the study site. Furthermore, not all potentially repro- tion of birds, the main seed dispersers of O. catharinensis,isalso ductive individuals are actually reproducing, reducing even more important to allow seedling establishment. Regarding ex situ con- the population size. Significantly lower outcrossing rates were servation, the collection of seed from 45 seed trees is suggested in estimated for Symphonia globulifera (Clusiaceae) in disturbed order to retain an effective size of 100 individuals. The estimate sites with lower densities of reproductive individuals (Aldrich & of effective variance size can also be used in combination with Hamrick 1998), as well as for Carapa procera (Meliaceae) in other reference effective sizes, if necessary. logged plots when compared to populations from undisturbed This report represents the first attempt to understand several plots (Doligez & Joly 1997). aspects of the reproductive ecology of O. catharinensis. Thus, Second, the multilocus outcrossing rate was highly variable we firmly believe that this kind of study should be carried out among evaluated seed trees. This result demonstrates that self- in other populations and continuously in order to better pre- ing rates are not uniform in seed trees. Therefore, in addition dict reproductive bottlenecks and, consequently, to propose to effective size effects, other factors, such as crown position in better and more effective conservation efforts. Nevertheless, the forest stratum, distance to other reproducing individuals our study allowed inferences regarding in situ and ex situ con- and position in the landscape, edge or core, may influence the servation measures for O. catharinensis. estimated outcrossing rates. Edge effect, for instance, enhanced selfing rates in individuals of Copaifera langsdorffi (Fabaceae; ACKNOWLEDGEMENTS Tarazi et al. 2013). Finally, we highlight the high number of flowers produced We thank the Instituto Chico Mendes de Conservacß~ao da Biodi- when compared to the number of ripe fruits. Although abor- versidade (ICMBio) and the researchers of Nucleo de Pesquisas tion of flowers and fruits is a common phenomenon in plants em Florestas Tropicais for support with the fieldwork. We also (Bawa & Webb 1984), it is really surprising that only one ripe thank the Laboratorio de Fisiologia do Desenvolvimento e Gen- fruit arose from 1839 buds under natural pollination treat- etica Vegetal for providing the necessary infrastructure required ment. As several factors can explain abortion of flowers and for the genetic analysis performed in this work. This study was fruits, such as pollen and resources limitation (Bawa & Webb funded by the Fundacß~ao de Amparo a Pesquisa e Inovacß~ao do 1984), only further research can clarify this question in Estado de Santa Catarina (FAPESC; grant numbers 11939/2009, O. catharinensis. However, a better understanding regarding 18868/2011-9 and TR2013-3558), the Coordenacß~ao de Aper- flower and fruit abortion could indicate limiting factors to fruit feicßoamento de Pessoal de Nıvel Superior (CAPES) to TM and production, valuable knowledge for the conservation of an JZS, and the Conselho Nacional de Desenvolvimento Cientıfico endangered species with low recruitment into the seedling cate- eTecnologico (CNPq) to MSR (304724/2010-6). We also thank gory (Montagna et al. 2018). David Martin for editing the manuscript.

Implications for conservation SUPPORTING INFORMATION Although the species presented several mechanisms to promote Additional Supporting Information may be found online in the cross-fertilisation, a high portion of the progenies was com- supporting information tab for this article: posed of full-sibs and self-sibs as a result of correlated matings Figure S1. Number of insects visiting the flowers of Ocotea and selfing. In the long term, this condition can harm the catharinensis per insect order and daytime hour. adaptive potential of the species through the loss of genetic Table S1. Floral visitors of Ocotea catharinensis. N, number variability. Furthermore, selfing and correlated matings can be, of visits; Fr, relative frequency (%); T/V, average time spent per for the studied population, a direct result of the low density of visit (min:seg); F/V, average number of visited flowers per visit; reproductive individuals, which is caused by past logging activ- P, pollen attached to the insect body; TF, insect touched repro- ities and by the fact that not all potentially reproductive indi- ductive structures of flowers; F>M, frequency of visits was viduals are actually reproducing. Consequently, efforts to higher than the median of all frequencies of visits; Class., classi- enhance population sizes are desirable, as well as to continue fication; PO, pollinator; PP, potential pollinator; PE, eventual restricting exploitation of O. catharinensis. In this sense, the pollinator; V, visitor. establishment of local seed collection areas plus the production Table S2. Spearman correlations between phenological and planting of seedlings in natural remnants of Ombrophilous events of Ocotea catharinensis (monthly frequency of indivi- Dense Forest should be promoted. duals presenting buds, open flowers, unripe or ripe fruits) and Sticky pollen, when combined with protogynic, herkogamic climatological data (monthly averages of mean temperature, and nectar-producing flowers, is a clear signal indicating that maximum temperature, and minimum temperature and total O. catharinensis pollination largely depends on insects. Therefore, monthly rainfall). *P < 0.1; **P < 0.05. any effort to create a new protected area or establish new popula- Table S3. Standing crop nectar production (lL) of Ocotea tions, as well as to conserve the existent populations, should take catharinensis per day and daytime hour and mean temperature into account the presence of pollinating insects. As mentioned, a (°C) per daytime hour.

Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands 933 Reproductive ecology of Ocotea catharinensis Montagna, da Silva, Pickart & dos Reis

ß~ REFERENCES and fragmented Amazonian rainforest. Molecular Klein R.M. (1980) Ecologia da flora e vegetacao do Ecology, 12, 753–764. Vale do Itajaı. Sellowia, 32, 164–369. Aldrich P.R., Hamrick J.L. (1998) Reproductive domi- Doligez A., Joly H.I. (1997) Mating system of Carapa Kubitzki K., Kurz H. (1984) Synchronized dichogamy nance of pasture trees in a fragmented tropical forest procera (Meliaceae) in the French Guiana Tropical and dioecy in Neotropical Lauraceae. Plant System- mosaic. Science, 281, 103–105. forest. American Journal of Botany, 84, 461–470. atics and Evolution, 147, 253–266. Baker H.G., Baker I. (1983) Floral nectar sugar con- Dupont Y.L. (2002) Evolution of apomixis as a strategy Larson B.M.H., Kevan P.G., Inouye D.W. (2001) Flies stituents in relation to pollinator type. In: Jones C. of colonization in the dioecious species Lindera and flowers: taxonomic diversity of anthophiles and E., Little R. J. (Eds), The handbook of experimental glauca (Lauraceae). Population Ecology, 44, 293–297. pollinators. The Canadian Entomologist, 133,439–465. pollination biology. Scientific & Academic Editions, Fægri K., van der Pijl L. (1979) The principles of pollina- Li Y., Meissle M., Romeis J. (2008) Consumption of Bt New York, USA, pp 117–141. tion ecology, 3rd edn. Pergamon Press, Oxford, UK. maize pollen expressing cry1 Ab or Cry3Bb1 does Barrett S.C.H., Kohn J.R. (1991) Genetic and evolu- Fenster C.B., Armbruster W.S., Wilson P., Dudash not harm adult green lacewings, Chrysoperla carnea. tionary consequences of small population size in M.R., Thomson J.D. (2004) Pollination syndromes PLoS One, 3, e2909. plants: implications for conservation. In: Falk D. A. and floral specialization. Annual Review of Ecology, Lingner D.V., Schorn L.A., Vibrans A.C., Meyer L., I., Holsinger K. E. (Eds), Genetics and conservation of Evolution, and Systematics, 35, 375–403. Sevegnani L., Gasper A.L., De Sobral M.G., Kruger€ rare plants. Oxford University Press, New York, Ferreira D.K., Nazareno A.G., Mantovani A., Bitten- A., Klemz G., Schmitt R., Junior C.A. (2013) Fitosso- USA, pp 3–30. court R., Sebbenn A.M., Reis M.S. (2012) Genetic ciologia do componente arboreo/arbustivo da Flor- Bawa K.S., Webb C.J. (1984) Flower, fruit and seed analysis of 50-year old Brazilian pine (Araucaria esta Ombrofila Densa em Santa Catarina. In: abortion in tropical forest trees: implications for the angustifolia) plantations: implications for conserva- Vibrans A. C., Sevegnani L., Lingner D. V., De Gas- evolution of paternal and maternal reproductive tion planning. Conservation Genetics, 13, 435– per A. L. (Eds), Inventario florıstico florestal de Santa patterns. American Journal of Botany, 71, 736– 442. Catarina, Floresta Ombrofila Densa. Edifurb, Blume- 751. Forfang A.S., Olesen J.M. (1998) Male-biased sex ratio nau, Brazil, pp 159–248. Bencke C.S.C., Morellato L.P.C. (2002) Comparacß~ao and promiscuous pollination in the dioecious island Mariot A., Mantovani A., Bittencourt R., Reis M.S. de dois metodos de avaliacß~ao da fenologia de plan- tree Laurus azorica (Lauraceae). Plant Systematics (2014) Aspectos da biologia reprodutiva de Drimys tas, sua interpretacß~ao e representacß~ao. Revista Brasi- and Evolution, 212, 143–157. brasiliensis Miers (Winteraceae) em Floresta Ombrofila leira de Botânica, 25, 269–275. Fournier L. (1974) Un metodo cuantitativo para la Mista, Sul do Brasil. Ciência Florestal, 24,877–888. Bernardello G., Anderson G.J., Stuessy T.F., Crawford medicion de caracterısticas fenologicas en arboles. MMA. (2008) Plano de Manejo da Floresta Nacional de D.J. (2001) A survey of floral traits, breeding Turrialba, 24, 422–423. Ibirama – Informacß~oes Gerais. Available from http:// systems, floral visitors, and pollination systems of Francisco M.R., Galetti M. (2002) Aves como potenci- www.icmbio.gov.br/portal/images/stories/imgs-uni the angiosperms of the Juan Fernandez islands ais dispersoras de sementes de Ocotea pulchella Mart. dades-coservacao/flona_ibiramaaa.pdf%3E (acessed (Chile). Botanical Review, 67, 255–308. (Lauraceae) numa area de vegetacß~ao de cerrado do 19 February 2018). Bittencourt R. (2007) Caracterizacß~ao da estrutura gen- sudeste brasileiro. Revista Brasileira de Botânica, 25, MMA. (2014) Portaria Ministerio do Medio Ambiente etica interna e aspectos da autoecologia de uma pop- 11–17. no 443, de 17 de dezembro de 2014. Lista Nacional ulacß~ao natural de imbuia (Ocotea porosa - Frankham R. (2005) Genetics and extinction. Biological Oficial de Especies da Flora Ameacßadas de Extincß~ao. Lauraceae). M.Sc. Thesis. Universidade Federal de Conservation, 126, 131–140. Available from http://www.mma.gov.br/estruturas/ Santa Catarina, Florianopolis, Santa Catarina, Brazil. Frankham R., Bradshaw C.J.A., Brook B.W. (2014) ascom_boletins/_arquivos/83_19092008034949.pdf Carvalho P.E.R. (1994) Especies florestais brasileiras: Genetics in conservation management: revised rec- (acessed 15 January 2018). recomendacß~oes silviculturais, potencialidades e uso da ommendations for the 50/500 rules, Red List criteria Montagna T., Lauterjung M.B., Candido-Ribeiro R., da madeira. EMBRAPA/CNPF, Colombo, Brazil. and population viability analyses. Biological Conser- Silva J.Z., Hoeltgebaum M.P., da Costa N.C.F., Ber- Castro E.R., Corrêa Cortes^ M., Navarro L., Galetti M., vation, 170,56–63. nardi A.P., Reis M.S. (2018) Spatial genetic struc- Morellato L.P.C. (2012) Temporal variation in the Galetti M., Laps R.R., Pizo M.A. (2000) Frugivory by ture, population dynamics and spatial patterns in abundance of two species of thrushes in relation to toucans (Ramphastidae) at two altitudes in the the distribution of Ocotea catharinensis Mez. from fruiting phenology in the Atlantic rainforest. Austral Atlantic Forest of Brasil. Biotropica, 32, 842–850. southern Brazil: implications for conservation. Ornitolhology, 112, 137–148. Galetti M., Bovendorp R.S., Guevara R. (2015) Defau- Canadian Journal of Forest Research, 48, 506–516. Cazetta E., Zumstein L.S., Melo-Junior T.A., Galetti M. nation of large mammals leads to an increase in seed de Moraes P.L.R., Monteiro R. (2002) Taxas de cruza- (2008) Frugivory on Margaritaria nobilis L.f. predation in the Atlantic forests. Global Ecology and mento em uma populacß~ao natural de Cryptocarya (Euphorbiaceae): poor investment and mimetism. Conservation, 3, 824–830. moschata Nees (Lauraceae). Biota Neotropica, 2,1–10. Revista Brasileira de Botânica, 31,303–308. Gibson J.P., Wheelwright N.T. (1996) Mating system Nazareno A.G., dos Reis M.S. (2012) Linking phenol- Cockerham C.C. (1969) Variance of gene frequencies. dynamics of Ocotea tenera (Lauraceae), a gynodioe- ogy to mating system: exploring the reproductive Evolution, 23,72–84. cious tropical tree. American Journal of Botany, 83, biology of the threatened palm species Butia erios- Conte R., Reis M.S., Mantovani A., Vencovsky R. 890–894. patha. Journal of Heredity, 103, 842–852. (2008) Genetic structure and mating system of Goudet J. (2002) FSTAT, a program to estimate and test Newstrom L.E., Frankie G.W., Baker H.G. (1994) A Euterpe edulis Mart. populations: a comparative gene diversities and fixation indices (version 2.9.3.2). new classification for plant phenology based on analysis using microsatellite and allozyme markers. Grogan J., Loveless M.D. (2013) Flowering phenology flowering patterns in lowland tropical rain forest Journal of Heredity, 99, 476–482. and its implications for management of big-leaf trees at La Selva, Costa Rica. Biotropica, 26, 141–159. Cruden R.W. (1977) Pollen-ovule ratios: a conservative mahogany Swietenia macrophylla in Brazilian Ama- Newton A.C. (2007) Reproductive ecology and genetic indicator of breeding systems in flowering plants. zonia. American Journal of Botany, 100,2293–2305. variation. In: Newton A. C. (Ed), Forest ecology and Evolution, 31,32–46. Herrera C.M. (1987) Components of pollinator “qual- conservation: a handbook of techniques. Oxford Custodio T., Comtois P., Araujo A.C. (2017) Repro- ity”: comparative analysis of a diverse insect assem- University Press, Oxford, UK, pp 235–284. ductive biology and pollination ecology of Triplaris blage. Oikos, 50,79–90. Nobrega S.R., Fortunato M.E.M., Quirino Z.G.M. gardneriana (Polygonaceae): a case of ambophily in Hesse M., Kubitzki K. (1982) The sporoderm ultra- (2016) Biologia reprodutiva, visitantes florais e dis- the Brazilian Chaco. Plant Biology, 19, 504–514. structure in Persea, Nectandra, Hernandia, Gomor- pers~ao de Ocotea duckei Vattimo-Gil (Lauraceae) na Dafni A. (1992) Pollination ecology: a practical tega and some other Lauralean genera. Plant Reserva Biologica Guaribas, Paraıba, Brasil. Gaia Sci- approach. Oxford University Press, Oxford, UK. Systematics and Evolution, 141, 299–311. entia, 10, 252–261. Danieli-Silva A., Varassin I.G. (2013) Breeding system Hmeljevski K.V., Reis A., Montagna T., Reis M.S. Ollerton J., Liede S. (1997) Pollination systems in the and thrips (Thysanoptera) pollination in the endan- (2011) Genetic diversity, genetic drift and mixed Asclepiadaceae: a survey and preliminary analysis. gered tree Ocotea porosa (Lauraceae): implications mating system in small subpopulations of Dyckia ibi- Biological Journal of the Linnean Society, 62, 593–610. for conservation. Plant Species Biology, 28,31–40. ramensis, a rare endemic bromeliad from Southern Pacini E., Nepi M. (2007) Nectar production and pre- Dick C.W., Etchelecu G., Austerlitz F. (2003) Pollen Brazil. Conservation Genetics, 12, 761–769. sentation. In: Nicolson S.W., Nepi M., Pacini E. dispersal of tropical trees (Dinizia excelsa: Fabaceae) Janzen D.H. (1979) How to be a fig. Annual Review of (Eds), Nectaries and nectar. Springer, Dordrecht, the by native insects and African honeybees in pristine Ecology and Systematics, 10,13–51. Netherlands, pp 167–214.

934 Plant Biology 20 (2018) 926–935 © 2018 German Society for Plant Sciences and The Royal Botanical Society of the Netherlands Montagna, da Silva, Pickart & dos Reis Reproductive ecology of Ocotea catharinensis

Pacini E., Sciannandrone N., Nepi M. (2014) Floral Saiter F.Z., Thomaz L.D. (2014) Revis~ao da lista de efforts in the insect-pollinated Neotropical tree Copai- biology of the dioecious species Laurus nobilis L. especies arboreas do inventario de Thomaz & Mon- fera langsdorffii (Fabaceae). Heredity, 110,578–585. (Lauraceae). Flora: Morphology Distribution, Func- teiro (1997) na Estacß~ao Biologica de Santa Lucia: o Varty N., Guadagnin D.L. (1998) Ocotea catharinensis. tional Ecology of Plants, 209, 153–163. mais importante estudo fitossociologico em florestas In: The IUCN Red List of Threatened Species 1998: Percival M. (1965) Floral biology. Pergamon Press, montanas do Espırito Santo. Boletim do Museu de e.T33982A9819827. Available from https://doi.org/ Oxford, UK. Biologia Mello Leit~ao, 34, 101–128. 10.2305/iucn.uk.1998.rlts.t33982a9819827.en (ac- Pyke G.H. (1991) What does it cost a plant to produce Schmid R. (1988) Reproductive versus extra-reproduc- cessed 29 January 2018). floral nectar? Nature, 350,58–59. tive nectaries-historical perspective and terminological Veloso H.P., Klein R.M. (1959) As comunidades e Rader R., Edwards W., Westcott D.A., Cunningham recommendations. Botanical Review, 54,179–227. associacßoes~ vegetais da mata pluvial atlântica do sul S.A., Howlett B.G. (2011) Pollen transport differs Schorn L.A., Gasper A.L., Meyer L., Vibrans A.C. do Brasil. II. Dinamismo e fidelidade das especies among bees and flies in a human-modified land- (2012) Sıntese da estrutura dos remanescentes flo- em associacßoes~ do municıpio de Brusque, SC. Sel- scape. Diversity and Distributions, 17, 519–529. restais em Santa Catarina. In: Vibrans A. C., Seveg- lowia, 10,9–124. Reitz R., Klein R.M.R.M., Reis A. (1978) Projeto nani L., Gasper A. L., Lingner D. V. (Eds), Inventario Ward M., Dick C.W., Gribel R., Lowe A.J. (2005) To madeira de Santa Catarina. Sellowia, 28–30,11– florıstico florestal de Santa Catarina, Diversidade e self, or not to self.. a review of outcrossing and pol- 320. Coinservacß~ao dos remanescentes florestais. Edifurb, len-mediated gene flow in neotropical trees. Hered- Ribeiro M.C., Metzger J.P., Martensen A.C., Ponzoni Blumenau, pp 125–142. ity, 95, 246–254. F.J., Hirota M.M. (2009) The Brazilian Atlantic For- Sebbenn A.M. (2002) Numero de arvores matrizes e Weir B.S., Cockerham C.C. (1984) Estimating F-Statis- est: how much is left, and how is the remaining for- conceitos geneticos na coleta de sementes para reflo- tics for the analysis of population structure. Evolu- est distributed? Implications for conservation. restamentos com especies nativas. Revista do Insti- tion, 38, 1358–1370. Biological Conservation, 142, 1141–1153. tuto Florestal, 14, 115–132. Wheelwright N.T. (1986) A seven-year study of individual Ritland K. (1989) Correlated matings in the partial Silva A., Aguiar I.B., De Schoffel E.R. (2000) Fenologia variation in fruit production in tropical bird-dispersed selfer Mimulus guttatus. Evolution, 43, 848–859. reprodutivadeCanela-Preta(Ocotea catharinensis Mez tree species in the family Lauraceae. In: Estrada A., Ritland K. (2002) Extensions of models for the estima- - Lauraceae) no Parque Estadual da Cantareira, S~ao Fleming T.H. (Eds), Frugivores and seed dispersal.Junk tion of mating systems using n independent loci. Paulo (SP). Revista do Instituto Florestal, 12,77–88. Publishers, Dordrecht, the Netherlands, pp 19–35. Heredity, 88, 221–228. Souza L.A., Moscheta I.S. (1999) Morfo-anatomia da Wolff D. (2006) Nectar sugar composition and volumes Ritland K., Jain S. (1981) A model for the estimation of flor de Ocotea puberula (Rich.) Nees (Lauraceae). of 47 species of Gentianales from a southern Ecuado- outcrossing rate and gene frequencies using n inde- Acta Scientiarum, 21,343–348. rian montane forest. Annals of Botany, 97,767–777. pendent loci. Heredity, 47,32–52. Tarazi R., Mantovani A., Reis M.S. (2010) Fine-scale Wratten S.D., Bowie M.H., Hickman J.M., Evans A.M., Rogalski J.M., Reis A., Rogalski M., Montagna T., Reis spatial genetic structure and allozymic diversity in Sedcole J.R., Tylianakis J.M. (2003) Field boundaries M.S. (2017) Mating system and genetic structure natural populations of Ocotea catharinensis Mez. as barriers to movement of hover flies (Diptera: Syr- across all known populations of Dyckia brevifolia:a (Lauraceae). Conservation Genetics, 11, 965–976. phidae) in cultivated land. Oecologia, 134, 605–611. clonal, endemic, and endangered rheophyte brome- Tarazi R., Sebbenn A.M., Kageyama P.Y., Vencovsky R. Wright S. (1931) Evolution in mendelian populations. liad. Journal of Heredity, 108, 299–307. (2013) Edge effects enhance selfing and seed harvesting Genetics, 16,97–159.