1. Introduction
Orchids (Orchidaceae) are the world’s second-largest family of flowering plants (c. 28,000) [
1]. They are adapted to diverse ecosystems but are highly threatened plants. Recently, The International Union for Conservation of Nature (IUCN) reported that 1636 orchid species were assessed as threatened, five species were extinct, and the state of the other species (c. 94%) remained unknown [
2]. The global orchid decline is mainly attributed to extrinsic factors, such as habitat destruction and fragmentation, over-collecting, and climate change [
3,
4,
5]. Further, intrinsic factors (e.g., life-history) have been recognized by orchid biologists as critical for the decline [
4,
6]. Thus, conservation actions to reduce the extinction risk are urgently needed.
Translocation is widely used as a conservation tool to ameliorate the known threats to orchid species [
4,
7,
8]. However, without understanding the key ecological attributes of the orchids (e.g., relationship with pollinators or mycorrhizal fungi), translocation actions cannot be successful [
4]. Indeed, in many cases of orchid translocations, the pollinator presence at the translocated sites was rarely determined, which could have revealed low recruitment rates to the translocated sites [
4]. In addition, plant–pollinator interactions, between the target species and recipient community, should be fully understood before applying any conservation translocation measures (e.g., reintroduction and reinforcement) [
9] to orchid species [
4,
6,
7,
10].
Slipper orchids (genus
Cypripedium) are one of the most intensively studied orchid groups due to their unique life history traits, such as pollination, and most of them (ca. 90% of species in subfamily Cypripedioideae) were assessed as threatened [
11].
Cypripedium species are non-rewarding orchids and pollinated by deceit, mainly as food mimics, and rarely as nesting-site mimics or brood-site mimics [
12]. Usually, in
Cypripedium species, pollination is accomplished by insect pollinators falling into the pouch-like labellum, escaping through a posterior opening, and visiting another flower of the same species. The pollinators of
Cypripedium species are bees, flies, and occasionally, wasps [
12,
13,
14]. Among the pollinators, bees account for most of their effective visitations [
12,
15].
To conserve
Cypripedium species, understanding the ecology of co-occurring rewarding plants in the community is a prerequisite, although identifying the effective pollinators is important. These orchids employ generalized food deception strategies [
16,
17], and such food-deceptive orchids display general floral signals (e.g., visual or olfactory) that attract various insect visitors [
18,
19]. Further, the spectrum of foraging pollinators visiting
Cypripedium species is generally wide, as they visit various flowers for food (i.e., polylectic) [
20,
21,
22,
23,
24,
25]. Thus, the pollination of
Cypripedium species is likely to be facilitated when rewarding plants around their populations are blooming sufficiently to feed diverse pollinators [
26,
27]. In this case, the understanding of community-wide plant–pollinator interactions is needed to identify co-occurring plants that function to support the populations of potential orchid pollinators [
28]. However, pollination studies on deceptive orchids have rarely attempted to identify effective pollinators or examine plant–pollinator interactions from community perspectives [
21,
29].
Understanding the structure and dynamics of the interactions between plants and pollinators can provide new insight into the ecological context, in which this deceptive orchid is effectively pollinated. In real-world ecosystems, species are interlinked with each other [
30]. Thus, for species conservation, we cannot consider only the pairs of interacting species [
31,
32], as have many previous studies, which focused on the interactions between the target orchid species and their effective pollinators. Network analysis is a tool that has been employed for understanding the structure and dynamics of plant–pollinator interactions, as well as identifying topologically important species within the networks [
33,
34,
35]. For example, metrics at the network level (e.g., nestedness, specialization, and modularity) are used in describing network structure and provide information on structural stability, resilience, and fragility [
33,
36,
37]. At the species level, the unweighted degree (i.e., number of interaction partners) [
38] and centrality (i.e., centrality and closeness centrality betweenness) [
35] have been frequently used to determine the topological importance of each species (e.g., keystone species), which can be achieved with low sampling effort [
39]. Because a pollination network survey generally involves observing all flowering plants and their visitors within a community and detecting changes in those interactions over time, community-wide plant–pollinator interactions can be more clearly understood than pollination surveys at the population level. In addition, the network approach has received growing attention as a tool for understanding the ecological context, since a robust method for comparing networks with different sizes was developed [
40]. However, the network approach has rarely been applied in studies regarding the pollination of deceptive orchids [
41].
Cypripedium guttatum is a critically endangered slipper orchid, especially in the Republic of Korea, although the orchid is one of the most widely distributed slipper orchids in the world [
21]. In the Republic of Korea, only two isolated populations remain, and thus, translocation action is urgently needed. Pollinators (e.g., Halictid bees;
Lasioglossum spp.) visiting
C. guttatum and their interactions with rewarding plants have been comprehensively studied in China [
21]. However, there is no such information for the populations in the Republic of Korea, hindering effective conservation actions for the orchid.
In the present study, we aimed to briefly assess the pollination of the slipper orchid C. guttatum, thus, providing useful insight into the successful translocation of the orchid. To this end, we examined flowering phenology, the effective pollinators of C. guttatum, and the structure and dynamics of community-wide plant–pollinator interactions by applying a network approach.
4. Discussion
Assessing pollination, a key ecological attribute, is essential for the conservation of critically endangered Slipper orchids (e.g.,
C. guttatum). Because
Cypripedium species are mostly highly interlinked with co-occurring species, the examination of plant–pollinator interactions from network perspectives can provide new insight into the translocation of the orchids. We demonstrated that
C. guttatum in the Republic of Korea was only pollinated by
L.
virideglaucum over the observation period, and the effective pollinator was determined as a high-priority keystone species for community conservation, in terms of the maintenance of the target orchid and co-occurring species. Given that the identification of a keystone plant or pollinator in a network can be achieved with low sampling effort [
39], well-connected species (i.e., keystone pollinator) persist over time [
34]. Since
C. guttatum and
L.
virideglaucum were in the same module, the pollination of
C. guttatum is likely to be maintained over time. Further, by examining daily networks, we demonstrated that diverse rewarding plants for the effective pollinator provided an ecological context for effective pollination of the deceptive orchid, which was postulated by previous researchers [
26,
27]. We cannot conclude whether the higher nestedness in the orchid network necessarily enabled the orchid to be pollinated due to the low sampling effort in the network survey. However, given that nestedness increases with network complexity (i.e., the number of interactions) [
33], the results of this study, at least, suggest that high plant and pollinator richness can increase the chance that the deceptive orchid will be pollinated.
Network metrics can provide important insight into species conservation, in terms of community persistence [
33,
36,
37]. However, given the limited resources, including cost, labor, and time needed for managing the conservation projects of endangered species, it may be difficult to allocate the majority of the resources to a network survey that requires high sampling intensity [
34]. In the present study, the network metrics (nestedness, specialization, and modularity) calculated in a single habitat may be inaccurate due to the short survey time (4 d). Low sampling effort can lead to the misinterpretation of network metrics, even though the null model for the network was used [
51,
52]. Instead, comparing the metrics between the networks can lend meaningful insight into the ecological factors affecting the network properties, as in the study by Song et al. [
40]. When employing translocation measures, finding a habitat that is ecologically suitable is the most important. Especially a deceptive orchid, such as that of
C. guttatum, needs a specific ecological context for successful pollination, as shown in the daily network patterns in the present study (
Figure 6 and
Table 6). In this regard, if
C. guttatum is translocated beyond its original habitat, investigating the plant–pollinator network in several experimental sites and comparing the metrics among the networks for each site will provide novel insight into understanding specific network structures that enable the orchid to be pollinated.
Although the foraging spectrum of the pollinators visiting
Cypripedium species is generally diverse [
20,
21,
22,
23,
24,
25], pollinators visiting
C. guttatum may have a weak color preference for flowers.
C. guttatum, as a food-deceptive orchid, is pollinated by three sweat bee species (e.g.,
Lasioglossum virideglaucum,
L.
clypeinitens, and
L.
sauterum) in China [
21]. In this study,
C. guttatum was only pollinated by
L.
virideglaucum, and the other species (e.g.,
L.
clypeinitens, and
L.
sauterum) present in China were not observed. This may be due to the low diversity of sweat bees that fit into the posterior opening size of the flower or the low number of
C. guttatum individuals (i.e., four individuals). However, given that
L.
virideglaucum consistently pollinated
C. guttatum flowers over two countries (i.e., Korea and China), the pollinator is likely to have a preference for the specific colors of both the orchid and co-occurring flowers. Bänziger et al. [
21] reported that
L.
virideglaucum visited co-occurring flowers with white, purple, and yellow colors, which is consistent with our results in the
C. guttatum module (
Figure 5). The color preference of the sweat bee can be explicitly tested using artificial flowers that have the same shape and color as the flowers of
C. guttatum in a future study.
Our study may provide some useful insight into translocation actions for
C. guttatum in the Republic of Korea. For example,
A.
gemmifera was determined to be a keystone species that can provide essential pollen and nectar for
L.
virideglaucum. These findings revealed that such a key species supporting the orchid pollinator could be identified by a network study, as reported by Phillips et al. [
28]. Thus, when selecting a suitable habitat for the survival of
C. guttatum, whether two keystone species (e.g.,
A.
gemmifera and
L.
virideglaucum), which highly contribute to the pollination of the orchid, are present should be considered first. Given that the effective pollinator
L.
virideglaucum is only distributed in Gangwon Province and Hallasan in Jeju Island in Korea [
53], and the distribution of
C. guttatum is restricted above 1000 m a.s.l. or at relatively high latitudes worldwide (Bänziger et al. [
21] and this study), one should consider Gangwon and Jeju, around 1000 m a.s.l., as suitable habitats for the reintroduction or assisted colonization of the orchid. Overall, despite the short term of this survey, the results suggest that network analysis can be a useful conservation tool when investigating suitable habitats for food-deceptive orchids, by finding an ecological context in which the orchids can be pollinated.