Sustainable harvesting and conservation of Laelia furfuracea, a rare epiphytic orchid from Oaxaca, Mexico

Many epiphytic orchids are harvested in Mexico for different purposes. Laelia furfuracea is one of the most intensively traded species. Its inflorescences are used as ornaments during the December festivities. We investigated the effect of severing the flowering pseudobulb. This is the traditional technique frequently used by collectors at the study area. We wished to investigate its effects on the production of new pseudobulbs, as well as on their size and flowering probability. Also, we examined the survival probability and growth of individuals that had fallen on the ground to evaluate their potential as trading resources. Inflorescence collection did not affect the production of new pseudobulbs the following season. However, it affected the size of these pseudobulbs, as well as their flowering probability. Yet, the direction of this effect was not consistent between years. Nearly six percent of all L. furfuracea plants at the study site were found on the ground. Over 80 percent of them survived for at least two years, although most of them showed pseudobulb loss over that period of time. We conclude that harvesting of flowering pseudobulbs may be sustainable in terms of its effects on plant performance, at least in the short term. The active management of plants that have fallen on the ground may reduce the harvesting pressure on natural populations. Harvesting of flowering pseudobulbs may diminish some aspects of plant performance, but its effects need to be evaluated with complete life cycle data and take into account interannual variation in vital rates.

| 143 OROZCO-IBARROLA et AL. out in different species and places are sustainable (Dutra-Elliott, 2014;Emeterio-Lara, 2019). The effects of harvest depend on the following factors, among others: (a) the nature of the harvest unit, that is, whole plants, plant fragments, or inflorescences (Mondragón, 2009); (b) the rate at which natural populations can regenerate the lost biomass (Emeterio-Lara, 2019); and (c) the frequency and intensity with which populations are harvested (Dutra-Elliott, 2014). In orchids, when the harvest units are plant fragments, population numbers are not necessarily affected in a direct manner. However, these fragments usually carry inflorescences. Thus, this form of harvest results in a reduced reproductive input of the affected plants. In addition, plant severing may result in a lowered growth rate or a decreased survival probability of the severed plants (Emeterio-Lara, 2019;Parra-Tabla, Vargas, Naval, Calvo, & Ollerton, 2011).
Irrespective of whether whole plants or plant fragments are the harvest unit, an additional aspect that remains to be investigated is the potential role of individuals that have fallen on the ground to satisfy at least some of the commercial demand for this type of plants.
In another group of epiphytes, the bromeliads, it has been suggested that naturally fallen individuals, which may survive on the ground for up to 1.5 yr, could be a significant source of plant material whose exploitation may reduce the harvest pressure on natural populations (Mondragon & Ticktin, 2011;Toledo-Aceves, Hernández-Apolinar, & Valverde, 2014). In epiphytic orchids, this form of exploitation has also been reported and may play an important role in this respect (Cruz-García et al., 2015), allowing harvesters to take advantage of living plant material that otherwise would be lost over time.
However, the length of time they can remain alive-and therefore potentially exploitable-on the ground is unknown.
In the Mexican state of Oaxaca, several epiphytic orchids are harvested and marketed. This region is exceptional for its high biodiversity, endemism, and cultural/ethnic richness (Ordoñez & Rodríguez-Hernández, 2008), and the Orchidaceae family is one of the most diverse, with 733 species reported (Salazar, 2012;Solano-Gómez, Martínez-Ovando, Martínez-Feria, & Gutiérrez-Caballero, 2016). Dutra-Elliott (2014) identified 19 orchid species being traded between March 2010 and April 2011 in the central market of Oaxaca city (the "Central de Abastos" in the state's capital city). Of these, 35,790 specimens were Laelia furfuracea Lindl. pseudobulbs with inflorescence ( Figure 1a). In a similar study, Cruz-García, Lagunez-Rivera, Chavez-Angeles, and Solano-Gómez (2015) recorded 36 orchid species being sold in the street markets of the village of Tlaxiaco, located in the western part of the state of Oaxaca.
The latter study identified Laelia furfuracea as the most intensively traded species, with 7,834 items (whole plants, plant fragments, or inflorescences) being traded between October 2011 and January 2012, during its flowering months. Also, Molina-Luna, Arellanes-Cancino, and Martínez (2015) reported that L. furfuracea was the most intensively traded species (orchids and bromeliads considered) during the Christmas season in markets within the central region of Oaxaca. Ticktin et al. (2020) also identified L. furcuracea as one of the most important species of the orchid trade in Mexico.
The harvesting and commercialization of non-timber forest products (NTFP), which include orchids, are an important source of income for millions of people around the world (Schmidt, Mandle, Ticktin, & Gaoue, 2011). In the rural areas of many developing countries, family income is supplemented by the harvesting and trading of NTFP, which reduces poverty and supports livelihood diversification. As mentioned above, plants and inflorescences of L. furfuracea are important NTFP in Oaxaca. They are intensively extracted from their habitat and traded mostly in local traditional markets.
Their use is mainly ornamental and ritual. In particular, temples and churches in many Oaxacan villages are adorned with L. furfuracea flowers during the December festivities (Cruz-García et al., 2015;Dutra-Elliott, 2014;Molina-Luna et al., 2015). One form of harvest is the extraction of whole plants from their natural populations.
F I G U R E 1 (a) Laelia furfuracea plant in the field established on a Quercus sp. tree. (b) and (c) Example of pseudobulb severing practiced by plant collectors and applied in the "extraction" experiment in this study Alternatively, local people cut flowering pseudobulbs in half, removing the inflorescence, and leaving the rest of the plant behind (Cruz-García et al., 2015;Figure 1b,c). In the latter case, the two meristems located at the base of the pseudobulb are left intact, which allows the development of a new pseudobulb from which a new inflorescence may eventually grow (Cruz-García et al., 2015). This is the traditional technique commonly used to harvest L. furfuracea and other epiphytic orchids in the region (Cruz-García et al., 2015). Thus, although the extracted inflorescence is prevented from making a reproductive contribution to population growth, at least the harvested plant remains in its habitat, that is, its exploitation does not add to natural mortality, and it can make a reproductive contribution to population growth in future years through the production of new pseudobulbs.
The aim of this study was twofold: to investigate (a) how the severing of the flowering pseudobulb in L. furfuracea affects the production of new pseudobulbs, and how it impacts their flowering probability in the next growing season; and (b) whether fallen L. furfuracea plants are a realistic resource for pseudobulb and inflorescence harvest in terms of their probability of survival and growth in the short term (one or two years). It has previously been suggested that the intense harvesting pressure on L. furfuracea may be seriously depleting its natural populations (Acosta-Castellanos, 2002; Anonymous, 1991;Halbinger & Soto-Arenas, 1997). Thus, the knowledge about the response of individual plants to harvesting is an important step toward the design of sustainable harvesting regimes and conservation programs.

| The study species
Laelia furfuracea ( Figure 1a) is known in the study region with the common names "lirio morado" (in Spanish), "gihtsl" (in the Mixtec language) or "ita ndeka morada" (in Mixtec-Spanish) (Halbinger & Soto-Arenas, 1997). It is a small epiphyte with clustered oblongovoid pseudobulbs. At the apex of each pseudobulb grows a single straight, rigid leaf. From the pseudobulb apex emerges a short raceme that bares 1-3 (exceptionally 4-5) rose to rose-pink flowers that are 7-9 cm wide ( Figure 1a). These pleasantly scented flowers are produced from October to January and last for about 3 weeks (Halbinger & Soto-Arenas, 1997; and personal observation). Each plant (a group of pseudobulbs with a common genetic origin) may produce a single raceme (i.e., inflorescence) per season from the apex of the leading pseudobulb. Some plants may have several "growing fronts," or leading pseudobulbs, and may thus produce several inflorescences in the same season.
This species is listed on the Mexican red list of endangered species (NOM-059-SEMARNAT-2010) under the category "subject to special protection" (Trujillo-Segura, 2019), due to the fact that it is exploited by local people. Some authors even consider it vulnerable or threatened because it is thought to be overexploited (Acosta-Castellanos, 2002;Anonymous, 1991). Here, we refer to it as "rare" due to its restricted geographic distribution and its high habitat specificity (Acosta-Castellanos, 2002;Halbinger & Soto-Arenas, 1997).

| The study site
This study was carried out in the Mixteca region, within the Mexican state of Oaxaca (municipality of Santo Domingo Yanhuitlán, Nochixtlán district, Figure S1). The specific area where this research was conducted has been designated as a conservation site by the municipal authorities; thus, no NTFP extraction has been permitted since 2010 (although we believe there is some low-intensity extraction which is unlikely to have affected our experimental plants). The altitude at this site varies between 2,380 and 2,580 m a.s.l. The climate is temperate, subhumid, with the majority of the precipitation falling during the summer months (June to September) (INEGI, 2010).
The mean annual rainfall is 420 mm, and the mean annual temperature is 16.9°C, with January being the coldest month (monthly mean:

| Fieldwork
The study consisted of two sections:

| Inflorescence extraction
During December 2016, 70 flowering L. furfuracea plants were selected at the study site and allocated randomly to one of two treatments: control (N = 45) or extraction (N = 25). To locate the plants, we did random walks at the study site covering an altitudinal interval from 2,380 to 2,580 m a.s.l., and all the plants observed during these walks were included in the study (with the exception of those that were inaccessible or badly damaged by herbivory). All the plants were tagged and the number of live pseudobulbs was counted; the pseudobulb bearing the inflorescence was measured (length and diameter). In the same way, in December 2017, a different set of 75 flowering plants selected as described above were tagged and randomly allocated to one of two treatments (control, N = 45; and extraction, N = 30). They were followed up to December 2018, recording the same variables as previously described. With these data, we wished to evaluate the effect of the removal of the flowering pseudobulb on the probability of producing a new pseudobulb the following growing season, as well as on the size of the new pseudobulb, and its flowering probability.

| Survival and growth of fallen individuals
Between August 2016 and January 2017, random walks were carried out at the study site within the altitudinal interval from 2,380 to 2,580 m a.s.l.; 36 phorophytes (most of them Quercus liebmanii) bearing a considerable number of L. furfuracea plants were located during these walks. All L. furfuracea individuals established on these phorophytes, as well as all individuals that were found fallen on the ground below the canopy of these trees, were tagged, their number of live pseudobulbs counted, and the size of the largest pseudobulb measured (the largest pseudobulb is usually the youngest, that is, the one that represents most closely the current state of the plant). L. furfuracea plants store water and nutrients in their pseudobulbs, so pseudobulb number and size may be related to their survival probability and growth potential. In addition, the presence of inflorescences was noted. We had no information regarding the length of time these individuals had been on the ground, but they were alive at the moment of tagging. Individuals were allocated to one of five size-stage categories (Table 1). Two years later (between September 2018 and January 2019), individuals were relocated to check whether they were still alive, in which case their number of pseudobulbs was counted and the size of the largest pseudobulb measured.

| Statistical analyses
To evaluate the effect of extraction on the probability of plants pro- Regarding the fallen plants, we calculated the proportion of the total number of plants in each category (see Table 1 for category definition) that were found on their phorophytes and the proportion that were found on the ground. We built a GLM with a binomial error distribution, the dependent variable being survival, and the independent variables being position (on the ground or on the phorophyte) and number of pseudobulbs. Additionally, we estimated the growth of each individual as its number of pseudobulbs in 2018 minus its number of pseudobulbs in 2016, and we built an additional GLM with a Poisson error distribution (first testing for overdispersion and co-linearity) to test the effect on plant growth of category (with 4 levels, as no seedlings were found on the ground), position (on the phorophyte versus on the ground), and the interaction between them, with initial number of pseudobulbs as a covariable.
Generalized linear models were carried out in R (version 3.5.2).  Figure 2); however, the effects of treatment and year on the probability of producing a new pseudobulb were not significant, neither was the effect of the covariables (Table 3a).  (Table 2). Again, the effect of treatment, year, and the interaction between them was significant (Table 3c). Also, the diameter of the focal pseudobulb (which was introduced as a covariable) positively influenced the size of the new pseudobulb, while the number of pseudobulbs did not affect it (Table 3c).

| Survival and growth of fallen individuals
During our initial survey (August 2016 to January 2017), the category with the highest number of fallen individuals was R1, followed by NR1 and NR2 (see Table 1 for category definition). A very small number of R2 individuals (Reproductive 2, plants with more than 100 pseudobulbs) were found on the ground (Figure 3). Almost 6% of all the individuals counted in the survey area were found on the ground (Figure 3). It is worth remembering that the length of time these individuals had been on the ground is unknown.
The percentage of individuals surviving after two years was significantly higher among the plants established on phorophytes than in those fallen on the ground (Figure 4, Table 4). Also, the percentage surviving increased with plant size (i.e., survival was higher in larger categories, especially among fallen individuals; Figure 4). The effect of the number of pseudobulbs at the start of the survey on the probability of surviving to 2018-2019 was also significant (Table 4) (Table 5). The effect of the number of pseudobulbs at the start of the observation period was significant (Table 5); plants with a larger number of pseudobulbs grew more slowly than those with fewer pseudobulbs. The interaction between position and category was significant mainly due to the fact that R1 plants (i.e., reproductive plants with < 100 pseudobulbs) lost a large number of pseudobulbs over time when they were on the ground (Table 5, Figure 6).   Note: Significant effects are in bold numbers. Abbreviation: CI (95%), 95% confidence interval.

| D ISCUSS I ON
during the two study years could have promoted a contrasting response to harvesting.
It is interesting that the size of new pseudobulbs behaved similarly to their probability of flowering: In 2016, severed plants produced larger pseudobulbs than control plants, whereas in 2017, the opposite pattern was observed (Table 2). It may well be that the two response variables are correlated and/or are affected by the same environmental factors, and thus, the smaller pseudobulbs produced in 2016 (in the control plants) had a lower flowering probability than the larger pseudobulbs produced in 2017, that is, the size of the pseudobulb may determine its ability to flower (Jacquemyn, Rein, & Jongejans, 2010;Pfeirer, Wolfgang, & Gottfried, 2006). Also, the effect of harvesting was similar on the two response variables, which suggests again that they may be correlated. This shows the importance of yearly environmental variation as a crucial driver of the behavior of the harvested individuals and points to the difficulty in setting standard quotas of inflorescence extraction. If harvesting promotes large new pseudobulbs and increased flowering probability in one year, but exactly the opposite the following year, it is difficult to come up with clear, unequivocal advice as to how best to exploit this species. Only a longer term experiment would provide the information necessary to understand this phenomenon in more detail.
In addition, the effect of harvesting and yearly environmental variation on pseudobulb size and flowering probability may also reflect on the different components of population dynamics.
An evaluation of individual performance, as essential as it is, is no substitute for the type of information that may be derived from a that whole-plant extraction is indeed more detrimental compared to pseudobulb harvest, which is also a conclusion reached by other authors (Ticktin et al., 2020). Emeterio-Lara (2019) Figure 6). Additionally, the survival of fallen individuals was lower than the survival of plants on phorophytes (Figure 4, Table 6). This suggests that fallen individuals were in the process of dying, which may take over two years. As we do not know for how long these fallen individuals had been on the ground prior to our observa- In the Mixteca region, it is known that local people collect fallen L. furfuracea individuals and cultivate them in their backyards to eventually use or sell their inflorescences (Cruz-García et al., 2015). We suggest this ex situ management does not affect population density, given that fallen individuals, although capable of remaining alive on the ground for a number of years, most likely do not make a contribution to population numbers. Although we did observe some fallen individuals produce ripe fruits, we estimate that the probability of the resulting seeds dispersing and arriving at microsites adequate for germination and establishment is low.
In conclusion, inflorescence collection in L. furfuracea may be sustainable when carried out through pseudobulb severing, at least in terms of its effects on plant survival and growth after one year.
However, the harvesting of inflorescences does affect further inflorescence production, as well as the size of new pseudobulbs, although the direction of this effect is not consistent between years.
Nearly six percent of all L. furfuracea plants found at the study site were on the ground. The probability of these plants surviving for at least two years is high and increases with plant size, although they tend to lose pseudobulbs over time.
We conclude that in addition to the results we present on individual plant performance in response to harvesting, it is necessary to carry out detailed demographic analyses to investigate the effect of exploitation and other aspects of management on population numbers. Notwithstanding, our results add significantly to the knowledge on epiphytic orchids subject to management and offer important insights that may aid in their conservation.

ACK N OWLED G M ENTS
We are grateful to the Yanhuitlán community for allowing us to work on their land. We thank IRUMA (Impulso Rural y Manejo Ambiental

D I SCLOS U R E
The corresponding author confirms on behalf of all authors that there have been no involvements that might raise the question of bias in the work reported or in the conclusions, implications, or opinions stated.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available from the Dryad Digital Repository: https://doi. org/10.5061/dryad.rn8pk 0p71 (Orozco-Ibarrola et al., 2020).