Conservation and Reintroduction of
in Hainan Province, China
Stephan W. Gale, Jihong Li and Gunter A. Fischer
Kadoorie Farm and Botanic Garden
Hong Kong SAR, China
Phalaenopsis pulcherrima is a perennial lithophytic or terrestrial herb belonging to the family Orchidaceae. Despite
being widespread throughout much of tropical Asia, the species has a very scattered distribution with only ca. 35
conrmed populations, at least some of which are known to have been extirpated. In China, P. pulcherrima occurs
only in the very south of Hainan Province, from where it is known from only six populations. Globally, the species is
threatened by collection for the ornamental plant trade and by habitat degradation.
In 2015, Hainan Wildlife Conservation Bureau, Bawangling National Nature Reserve Administration and Kadoorie
Farm and Botanic Garden joined forces to help conserve the species in Hainan by undertaking applied ecological
research and a large-scale micropropagation and reintroduction programme. This entailed enlisting the support of
a research institute (Hainan University) and a commercial orchid nursery (Dongfang Tengfei Horticulture Biotech
Ltd.). This report details the lessons learnt and key outcomes of this progressive collaborative project. Phalaenopsis
pulcherrima is shown to be:
• Adapted for growth in seasonally dry monsoon forest, within which it is typically conned to exposed
• Made up of geographically isolated populations that mostly comprise few individuals.
• Nectarless and dependent on the pollinating services of one or two bee species (Amegilla zonata
and Nomia punctulata), which it deceives into visiting its owers by imitating the colour signal of
co-blooming, rewarding magnet plants.
• Restricted to growing in sites at which an abundance of rewarding magnet plants occur, such that
pollinator interest in its non-rewarding owers is sustained.
• Highly genetically diverse and capable of gene ow via pollen over distances of at least ca. 350 m.
• Subject to varying levels of protection within range countries, but generally vulnerable to collection
even where it occurs within protected areas.
• Near Threatened globally and Vulnerable in Hainan, according to IUCN Red List Criteria.
• Readily propagated from seed in vitro and easily grown-on in the nursery.
Bawangling National Nature Reserve provided the ideal opportunity for species restoration, given the existence there
of four natural populations, suitable habitat and management infrastructure. Following the collection of open-pollinated
capsules from one these four populations and micropropagation of their seeds, over 2000 seedlings were prepared
for reintroduction into three new host sites in 2017. After monitoring the reintroduced plants for two years, it was
• Axing seedlings to bare rock surfaces using a non-toxic bond is an appropriate means of achieving
• Approximately 800 reintroduced plants survived, eectively doubling the known population of the
species in China.
• The observed survivorship rate of 46.7% compares well with that reported in other published orchid
reintroduction studies conducted worldwide.
• Flowering and fruiting had already occurred by the conclusion of the project, indicating population
A series of ve recommendations are put forward to ensure the long-term success of this project, for the benet of
both P. pulcherrima itself and other similarly threatened orchids in China and the wider region.
This report summarises the content and outcomes of a collaborative project conducted by Kadoorie Farm and Botanic
Garden (KFBG), Bawangling National Nature Reserve Administration (BNNRA) and Hainan Wildlife Conservation
Bureau (HWCB) for the conservation and reintroduction of the threatened lithophytic orchid Phalaenopsis pulcherrima
in Hainan. A Memorandum of Understanding (MOU) signed in March 2015 dened the respective roles and
responsibilities of the three parties in executing the project for a period of up to ve years. In view of the fact that all
responsibilities have duly been completed, this report synthesises the key lessons learnt as of the MOU’s conclusion
in November 2019. Given the signicance of the recovery of local populations of P. pulcherrima for plant conservation
in Hainan, and in recognition of the paucity of scientically-driven plant reintroduction programmes undertaken
regionally, it is hoped that the body of work presented here will serve as a model for the restoration of other threatened
plants throughout South China and in neighbouring countries.
Aims and objectives
The lithophytic orchid Phalaenopsis pulcherrima has declined throughout its range in tropical Asia, primarily as a result
of habitat loss and over-exploitation for the horticultural trade. In China, the species occurs only in Hainan Province,
where local extinctions and declines have been reported. In order to better understand the species’ conservation
status and ensure its recovery in the wild, KFBG, BNNRA and HWCB jointly agreed to undertake ecological research
and coordinate a complementary reintroduction programme. The bulk of this work was carried out at Bawangling
National Nature Reserve (BNNR), where the existence of natural populations, suitable habitat and management
infrastructure together provided the ideal opportunity for species restoration.
The specic aims of the ecological research component were:
• To estimate the historic and current occurrence of Phalaenopsis pulcherrima in Hainan through eld
surveys and interviews with local partners and communities.
• To gauge the nature and level of threats impinging on wild populations of P. pulcherrima in Hainan.
• To characterise the ecological niche of P. pulcherrima in Hainan and infer habitat specicity.
• To ascertain the requirements for successful fruit-set.
• To quantify the spatial genetic structure, degree of genetic variation and level of gene ow within and
between wild populations of P. pulcherrima in Hainan using appropriate genetic markers.
• To derive inferences on the processes of seed dispersal and population establishment.
• To gauge whether the inferred capacity for population growth of the species is being outstripped by the
The specic aims of the reintroduction component were:
• To select one or two sites at BNNR for reintroduction of Phalaenopsis pulcherrima seedlings and
perform a feasibility and risk assessment at each.
• To collect up to 50 naturally formed capsules from as broad a range of mother plants as possible at
existing wild populations close to the selected reintroduction sites.
• To develop propagation and cultivation protocols in order to produce up to 2500 healthy seedlings
under quarantine conditions for reintroduction into the wild.
• To conduct a short-term planting-out trial at one site prior to conducting the full-scale
reintroductions so as to test methods of attachment and assess the adaptability of nursery-grown
plants to in situ conditions.
• To conduct full-scale reintroductions at the selected reintroduction site(s) using up to 1000 plantlets
• To undertake monitoring of plants and populations at regular intervals for a minimum of two years after
reintroduction to ascertain survivorship and performance.
• To derive a practical reintroduction protocol that observes IUCN guidelines for reintroductions and
optimises the likelihood of establishment and long-term ecological functioning.
Pursuant to these aims, the over-arching objectives of this collaborative project were:
• To elucidate key aspects of species biology and identify signicant constraints on population growth so
as to infer the species’ conservation status in Hainan.
• To establish one or two new, self-sustaining and ecologically viable populations of Phalaenopsis
pulcherrima in BNNR using genetically diverse, articially propagated plantlets of local origin.
• To develop inter-institutional rapport conducive to social learning for government ocials, nature
reserve sta, conservation biologists, students and other relevant individuals and organisations on the
eective conservation of this species and its habitat, and publicise outcomes as far as practically
possible to raise public awareness for plant conservation.
• To derive practical recommendations for the conservation and reintroduction of P. pulcherrima that may
serve as a comprehensive model for the conservation and reintroduction of other endangered plants
in Hainan and the wider region.
This report deals with each of these four objectives in turn. In addition to the scientic literature cited, the primary data
reported here are derived from as yet unpublished eld- and laboratory-based studies undertaken by Mr. Zhe Zhang,
a PhD student based at Hainan University who was hired by KFBG from June 2015 to December 2018 specically
to elucidate aspects of the species biology of P. pulcherrima, as well as from applied work carried out by KFBG and
BNNR sta during the MOA period. This report and its recommendations have been compiled with reference to IUCN’s
Guidelines for Reintroduction and Other Conservation Translocations (IUCN/SSC, 2013).
Phalaenopsis is a genus of 45–50 species occurring in tropical and subtropical Asia, New Guinea and Australia. Most
species are epiphytic with short stems, pendulous inorescences and long, adventitious roots that splay laterally and
anchor the plant to tree stems and branches. In this respect, P. pulcherrima is unusual in the genus in that it grows
on the ground, either as a lithophyte attached directly to rock surfaces or as a terrestrial rooted in shallow humus
over coarse, sandy soils (Plate 1-A, 1-B). In accordance with its upright growth habit, P. pulcherrima possesses erect
inorescences and prop-like roots that radiate downwards from the lower stem internodes. The stem readily branches
near the base to produce lateral oshoots, giving rise to dense, clonal clumps (Plate 1-C). Because of this, individual
plants can live for many years.
Taxonomy and botanical description
In addition to its upright, lithophytic habit, Phalaenopsis pulcherrima possesses several other features that are
discordant with many other species of the genus, including its long column-foot and four (as opposed to two) pollinia
borne on a long stipe (Fig. 1). In combination, these characters have led to P. pulcherrima (and a complex of closely
related taxa) being placed in a separate genus, Doritis, in which the species was rst described (Lindley,
1833). However, over-emphasis of such dierences and the articial weighting of pollinia number as a key
character have been discredited in determining homology and resolving monophyly within the
Orchidcaeae, and recent molecular phylogenetic studies have proven Doritis to be nested within Phalaenopsis.
As a result, the species is now treated as a member of Phalaenopsis, in which it is accommodated in section
Esmeralda, a small clade of lithophytic and terrestrial species (Christenson, 2001; Govaerts et al., 2019). An array
of morphologically similar plants found within the geographic range of P. pulcherrima have been distinguished on the
basis of minor variation in the size, outline and colour of oral parts, but these almost certainly represent forms of a
single variable species (Kumar et al., 2018; Govaerts et al., 2019); the species has therefore accrued a large number
of synonyms. Plants in Hainan conform to the typical form.
Phalaenopsis pulcherrima (Lindl.) J.J.Sm., Repert. Spec. Nov. Regni Veg. 32: 366 (1933). Type: Vietnam, Turon (=
Danang), Finlayson 521 (Wall. Cat. 7348) (holotype, K-W).
Doritis pulcherrima Lindl., Gen. Sp. Orchid. Pl.: 178 (1833).
Phalaenopsis esmeralda Rchb.f., Gard. Chron., n.s., 2: 582 (1874).
Phalaenopsis esmeralda var. albiora Rchb.f., Otia Bot. Hamburg.: 35 (1877).
Phalaenopsis antennifera Rchb.f., Gard. Chron., n.s., 11: 398 (1879).
Phalaenopsis esmeralda var. candidula Rolfe, Lindenia 6: 49 (1890).
Phalaenopsis esmeralda var. rubra Stein, Orchid.-Buch: 506 (1892).
Phalaenopsis mastersii King & Pantl., J. Asiat. Soc. Bengal, Pt. 2, Nat. Hist. 66: 591 (1897).
Phalaenopsis esmeralda var. punctulata Cogn., Chron. Orchid. 2(1): 2 (1904).
Doritis pulcherrima var. caerulea Fowlie, Orchid Digest 33: 241 (1969).
Doritis pulcherrima f. alba O.Gruss & Roeth, Orchidee (Hamburg) 50: 671 (1999).
Doritis pulcherrima f. albiora (Rchb.f.) Roeth & O.Gruss, Orchidee (Hamburg) 50: 671 (1999).
Doritis pulcherrima f. caerulea (Fowlie) O.Gruss & Roeth, Orchidee (Hamburg) 50: 671 (1999).
Phalaenopsis pulcherrima f. alba (O.Gruss & Roeth) Christenson, Phalaenopsis: a monograph: 232 (2001).
Phalaenopsis pulcherrima f. albiora (Rchb.f.) Christenson, Phalaenopsis: a monograph: 233 (2001).
Phalaenopsis pulcherrima f. caerulea (Fowlie) Christenson, Phalaenopsis: a monograph: 233 (2001).
Doritis pulcherrima f. apiculata Aver., Lindleyana 22(2): 13 (2009).
Doritis pulcherrima var. laotica O.Gruss, Orchidee (Hamburg) 60: 440 (2009).
Doritis pulcherrima f. nivea Aver., Lindleyana 22(2): 14 (2009).
Doritis pulcherrima f. purpurea Aver., Lindleyana 22(2): 13 (2009).
Doritis boulbetii Telepova, Rhône-Alpes Orchid. 50: 7 (2013).
Doritis pulcherrima f. cinnabarina Telepova, Rhône-Alpes Orchid. 50: 11 (2013).
Phalaenopsis pulcherrima f. nivea (Aver.) O.Gruss & W.E.Higgins, Phalaenopsis J. 27(1): 15 (2017).
Phalaenopsis pulcherrima f. purpurea (Aver.) O.Gruss & W.E.Higgins, Phalaenopsis J. 27(1): 15 (2017).
Perennial lithophytic or terrestrial herb with a short, upright stem that branches to produce basal oshoots. Stem
enclosed in overlapping sheaths at base. Roots emerging from lower stem internodes, wiry, terete, enclosed in a
silvery velamen. Leaves 3–7, distichous, clustered towards stem apex, succulent, oblong, 4–13 × 1–3.5 cm, apex
obtuse to acute, articulated and clasping the stem at base. Scape emerging from lower stem node opposite a leaf,
erect, 12–42 cm long, racemose, bearing 4–28 owers that open from the base 1–3 at a time. Flowers widely spaced,
about 2.5 cm across, variable in colour, the sepals and petals ranging from white to pink or purple, the lip pink to
dark purple tinged orange-brown on the side lobes and lobules of the mid-lobe. Pedicel and ovary spreading, slightly
curved, 1.4–2 cm long. Dorsal sepal ovate-elliptic, 8–15 × 5–7 mm, apex obtuse. Lateral sepals ovate-triangular, 8–15
× 6–9 mm, apex obtuse, adnate to the column-foot at base. Petals obovate, 8–15 × 5–7 mm, apex rounded. Lip
attached at right angle to apex of column-foot, 12–16 mm long, 3-lobed, very narrow at base, broader
across mid-lobe; side lobes erect, linear, 3–4 mm long; mid-lobe deexed, thickened, ligulate, ridged and
with prominent lobules at base; lobules erect, suborbicular, 5–6 × 4–5 mm, broadly rounded. Column stout,
6–8 mm long, narrowly winged, with a long pendulous column-foot up to 8 mm long at base; stigma large, concave;
rostellum prominent, long; anther cap globose; pollinarium comprising 4 yellow, waxy, globose pollinia attached to a
single linear stipe and an ovate viscidium. Capsule fusiform, 2–3 cm long, ridged.
Fig. 1. Schematic illustration of Phalaenopsis pulcherrima indicating key vegetative and oral parts.
Phalaenopsis pulcherrima occurs primarily in the seasonal Asian tropics, from northeast India and Myanmar, through
Thailand to Laos, Cambodia and Vietnam, but it has also been recorded in parts of the aseasonal wet tropics,
in Peninsular Malaysia, Sumatra and Borneo (Fig. 2). The species has a marginal occurrence in China, with six
populations known in the south of Hainan Province. Despite this wide range, P. pulcherrima has a very scattered
distribution, typically being conned to open sandstone and granitic shelves in sparse, deciduous woodland, often
along streams or rivers. Based on herbarium data available from the Global Biodiversity Information Facility (GBIF.
org, 2019), the species has been recorded from ca. 35 populations, with an estimated Extent of Occurrence of ca. 3.7
million km2 and Area of Occupancy of ca. 112 km2. However, at least 15 of these records are historic (>50 years old)
and it is not known how many of these populations persist to today. Comprehensive range-wide surveys to conrm
current occurrence and abundance were far beyond the scope of the present project.
Fig. 2. Global distribution of Phalaenopsis pulcherrima (indicated in red) and approximate occurrence of its six known populations in
Hainan (red dots).
Phalaenopsis pulcherrima is a warm- to hot-growing species that is supremely adapted to growth in a
seasonally dry climate – its clustered succulent leaves can withstand exposure to strong sunshine, high temperature
and extended periods of drought, and its eshy roots ax the plant to bare rock and are enclosed in a spongy
velamen that resists desiccation and soaks up water during periods of rain or inundation (Plate 1-D). Its distribution
is essentially coincident with the occurrence of mixed deciduous or monsoon forest, of which it forms a typical
component. This forest habitat is the dominant vegetation type throughout the region where mean annual rainfall
ranges from 1200–1800 mm and the dry season persists for 4–6 months (Rundel, 1999). The species has been
recorded from an elevational range of 100–1200 m a.s.l. (Christenson, 2001).
Its six populations in southern Hainan (Fig. 2, Table 1) grow either lithophytically on open sandstone shelves in
disturbed monsoon forest (populations DL, JFL, WX and YJ) or as a terrestrial in the understorey of secondary scrub
composed of regenerating native species and planted trees, such as Acacia mangium and Pinus latteri (populations
DLH, LD and WX). In this sense, the species appears to be tolerant of disturbance and may, to an extent, even
benet from it: with more than ca. 700 individuals, by far the largest population is that at WX, where plants grow both
lithophytically at the margins of intact though disturbed monsoon forest and as a terrestrial on eroded soils in relatively
open patches of secondary forest with planted P. latteri. The other populations in Hainan are considerably smaller, with
up to ca. 25 individuals growing in scattered clusters. The elevational range of the species in Hainan is ca. 250–800 m
Table 1. Site features of the six populations of Phalaenopsis pulcherrima in Hainan.
* Population codes follow Fig. 2.
Four of the populations in Hainan (DL, DLH, WX and YJ) lie within BNNR. Climate data available online
(Worldweatheronline, 2019) for the ten-year period from January 2009 to December 2018 indicate that BNNR has a
mean annual temperature of 25.2oC, a mean annual maximum temperature of 27.1oC and a mean annual minimum
temperature of 23.6oC. Mean annual rainfall is ca. 1390 mm, with most rain falling during the southeast monsoon,
which in Hainan generally lasts from June to October. Mean annual humidity ranges from 71–87%.
Clarifying the means by which threatened species such as Phalaenopsis pulcherrima achieve pollination is key to
evaluating their regenerative capacity and is therefore an important step in ensuring that natural populations can
recover following decline and that articially established populations can persist following reintroduction. In addition,
conrming the degree to which pollination is brought about through selng, either by self-pollination of the same ower
(autogamy) or vector-mediated pollen transfer between owers belonging to the same individual (geitonogamy),
as opposed to outcrossing (xenogamy), provides insights into species ecology in terms of inferred dependence on
pollinators to facilitate pollen ow, the risk of inbreeding depression, resilience to founder eects and bottlenecks,
expected genetic diversity and the likely scale of population-level genetic structuring.
The pollination biology of Phalaenopsis pulcherrima was investigated by Jin et al. (2012) at three sites in
Hainan. Flowering was found to be broadly coincident with the monsoon rains, commencing in June or
July and lasting until October. Hand pollination experiments revealed the species to be capable of
producing capsules following both articial self-pollination and cross-pollination, although the rate of fruit-set (i.e.
the proportion of owers that produce a fruit) was signicantly higher in the latter (42.3% versus 73.9%). In contrast,
unmanipulated owers that were simply bagged to prevent access to potential pollinators failed to set fruit. These
results demonstrate that the species is capable of self-pollination but facultatively outcrossing and dependent on an
insect vector for pollen transfer. Whereas unpollinated owers remained fresh for up to ten days, the sepals and petals
of pollinated owers withered within three days.
Flowers of P. pulcherrima were found to be nectarless, suggesting that pollinating insects are not rewarded for their
service. Non-rewarding owers of this sort are termed deceptive, because the mechanism by which they attract
pollinators involves emitting false signals that mimic an environmental cue that would otherwise entice the insect
towards a particular reward, such as a food source, shelter or a mate (Jersákova et al., 2006). This mechanism is
hypothesised to have evolved to promote outcrossing, since a deceived insect is less likely to visit another non-
rewarding ower on the same inorescence and will therefore travel further before alighting on another ower of
the same species (i.e. favouring xenogamy over geitonogamy), thereby reducing inbreeding. Observations of oral
visitors in Hainan revealed the blue-banded bee Amegilla nigritar (Hymenoptera: Apoidea: Anthophorini) to be the sole
eective pollinator of P. pulcherrima (Jin et al., 2012). These bees were commonly seen ying around the orchid’s
forest habitat foraging for food, and their size and shape was found to be a perfect match for the opening created
by the placement of the lip and column of a P. pulcherrima ower at full anthesis. However, only approximately one
in six of these bees visited owers of P. pulcherrima; more frequently, they were observed visiting owers of other
plants in the vicinity, including Lasianthus species (Rubiaceae), Leptodermis species (Rubiaceae), Polygala species
(Polygalaceae) and Helicteres species (Malvaceae). All of these co-blooming species were found to oer nectar to
their visitors, and the colour of their owers mostly ranged from pink to purple.
Taken together, these ndings are interpreted as an example of generalised food deception, in which a foraging bee
is lured to the orchid’s non-rewarding owers by visual cues (i.e. oral colour) that mimic various co-blooming owers
of other species that oer a food reward. Functionality of this system depends on an abundance of other rewarding
plant species that ower simultaneously in the vicinity, such that the interest of foraging bees is sustained even when
they occasionally visit a non-rewarding orchid (Fig. 3). Because bees learn to discriminate non-rewarding owers from
rewarding ones (Menzel & Müller, 1996), the level of pollination of deceptive owers is usually low (Jersákova et al.,
2006). Mean natural fruit-set at the three study populations in Hainan was found to range from just 10.9–13.1% (Jin et
Further work carried out by Zhang and co-workers in Hainan has since conrmed that P. pulcherrima is indeed
incapable of spontaneous autogamy or agamospermy, but they found no statistical dierence in fruit-set between
articial self- and cross-pollination treatments (89.2% versus 90.4%; Zhang et al., in review). However, the proportion
of capsules that contained viable seeds did dier among pollination treatments: whereas a mean of 96.2% of seeds
from cross-pollinated capsules contained an embryo, only 16.1% of seeds from self-pollinated capsules did so and
13.3% of self-pollinated capsules contained no seeds at all. This reveals that the species is not fully self-fertile, due
either to inbreeding depression or partial self-incompatibility. In addition to the species’ generalised food deceptive
mating strategy, this feature is further expected to promote the prevalence of outcrossed progeny and therefore
generate greater population-level genetic diversity (Zhang et al., in review).
More recently, Zhang et al. (in prep.) identied two bees as primary pollinators of P. pulcherrima in Hainan: Amegilla
zonata (Plate 1-E, 1-F), suggesting that reports of A. nigritar in the earlier study by Jin et al. (2012) could have been
misidentications, and Nomia punctulata (Hymenoptera: Apoidea: Halictidae), both of which were routinely observed
Fig. 3. Pollination of Phalaenopsis pulcherrima is achieved through deception of foraging blue-banded bees, Amegilla zonata. The
opening at the centre of P. pulcherrima owers created by the placement of the lip mid-lobe, its lobules and the column perfectly
matches the size of the bee (A). A visit typically lasts only 2–3 seconds before the bee detects the absence of nectar, but this is
long enough for the bee to inadvertently dislodge the pollinarium from beneath the anther cap and rostellum, gluing the entire
structure to its head via the sticky viscidium. The pollinarium is thereby carried o by the bee (B), with pollination occurring when
the bee visits a second ower, usually some distance from the rst. The success of this outcrossing system depends on the
presence of co-blooming, rewarding species, such as Helicteres angustifolia (C) and Hedyotis consanguinea (D), the owers of
which act as magnets, attracting bees to the similarly coloured owers of the non-rewarding orchid.
foraging on and pollinating owers of the rewarding Helicteres angustifolia and Hedyotis consanguinea
(Rubiaceae) in the orchid’s habitat (Plate 1-G, 1-H). Analysis of reectance spectra in bee colour space
(Dyer, 2006; Dyer et al., 2012) suggests that bees perceive the colour of the pink- and purple-owered
morphs of P. pulcherrima as identical to that of the owers of these sympatric species. Moreover, pollination success
was found to be signicantly greater at sites with a higher frequency of these two rewarding species, suggesting that
these plants serve as so-called ‘magnet’ plants in attracting pollinators to the deceptive orchid (Zhang et al., in prep.; Fig. 3).
These studies highlight that ecological viability in P. pulcherrima depends upon the abundance of co-blooming
magnet plants, as well as on the occurrence of an eective pollinator. Whereas H. angustifolia has a wide distribution
throughout East and Southeast Asia (Tang et al., 2007), H. consanguinea has a rather more restricted range, being
known only from the southeastern provinces of China (Chen et al., 2011). Similarly, whereas Amegilla zonata has
been reported from India, Myanmar, Sri Lanka, Thailand, China, Malaysia and Indonesia (Engel, 2007), a range
largely sympatric with that of P. pulcherrima, N. punctulata is known from only scattered sites in subtropical and
temperate East Asia (Astafurova & Pesenko, 2005). Reintroduction of P. pulcherrima is therefore most likely to
succeed within its known range at sites with intact local oras that have suered little disturbance, since this is where
these essential components of the orchid’s habitat are most likely to be found.
Population genetics and gene ow
Genetic diversity allows natural populations to respond to changes in their environment and withstand demographic
or other stochastic challenges, such as colonisation bottlenecks, inbreeding and drift. This is particularly important
for rare species that have suered decline because the amount of remaining genetic diversity will, to a large extent,
determine their capacity to persist, recover and adapt. Outcrossing species such as Phalaenopsis pulcherrima are
expected to have relatively high levels of genetic diversity as compared to those that predominantly self-pollinate,
because most new individuals recruited into a population are likely to be the product of two genetically distinct
parents. However, insect-pollinated species usually suer more substantial losses of genetic variation following habitat
fragmentation, because pollinator populations are also impacted by environmental degradation and any resulting
restriction in the mobility of the pollinating insect among remaining habitat fragments will inevitably compromise pollen
ow. Given its scattered population structure and history of decline in Hainan, we sought to elucidate the level of
genetic diversity and patterns of gene ow within and among populations at BNNR.
Three physically separated sub-populations at population WX (see Fig. 2 and Table 1) were sampled for genetic
analysis in June 2015. The rst sub-population (WX-1) lies at ca. 500 m a.s.l., occupies an area of ca. 140 × 120
m and comprises 56 individuals; the second sub-population (WX-2) lies at ca. 300 m a.s.l., occupies an area of ca.
80 × 220 m and comprises 39 individuals; the third sub-population (WX-3) lies at ca. 200 m a.s.l., occupies an area
of ca. 40 × 100 m and comprises 60 individuals. WX-1 and WX-2 occur in partial shade under mixed Pinus latteri
plantation and secondary scrub, whereas WX-3 occurs in full sun on an exposed rocky shelf beside a stream. The
three sub-populations are separated by dense woodland that is unsuitable for the orchid, with intervening distances of
>100 m (Fig. 4). All 155 adult plants at the three sub-populations were labeled and mapped, and a small leaf sample
was collected from each. To enable comparison of genetic diversity measures for the background, multigenerational
population with those for the following generation of F1 recruits, a further 1105 samples were collected from seedlings
raised through micropropagation from ten capsules harvested from the three WX sub-populations in November 2015.
Genomic DNA was extracted and amplied with 15 polymorphic microsatellite markers (representing 15 specic loci)
specically developed for this species (Zhang et al., in review). Genetic diversity, ne-scale genetic structure, paternity
assignment and dispersal estimates were calculated using standard software packages.
Two important measures of genetic diversity are allelic richness (the absolute number of alleles present in a population
at any one particular locus) and heterozygosity (the proportion of heterozygotic individuals within the population,
measured on a scale of 0 to 1). Results of our analyses revealed high genetic diversity: the mean number
of alleles present at each of the 15 loci was 6.8 across the three WX sub-populations and 5.0 among the
seedlings, whereas observed heterozygosity was 0.656 and 0.580 respectively (Zhang et al., in review).
Although these measures of genetic diversity were higher for the background population as compared to the
seedlings, the dierences were not statistically signicant, especially when the limited number of capsules sampled
for this study is taken into account. Therefore, there was no evidence of a genetic bottleneck at WX, nor of loss of
genetic diversity from one generation to the next. Rather, these ndings suggest that, despite the history of decline
and fragmentation reported for P. pulcherrima in BNNR (and in Hainan as a whole), high genetic variation has been
retained (or recuperated) at the population level, with most individuals at WX being heterozygotic and all sampled loci
having multiple alleles. This almost certainly reects the species’ predominantly outcrossing pollination system, which
tends to promote xenogamous mating among unrelated individuals.
Furthermore, analysis of the partitioning of genetic diversity in space revealed most variation to lie within sub-
populations (99.3%) as opposed to among them (0.7%), with low dierentiation between the three sub-populations
(Zhang et al., in review). This suggests that high genetic variation occurs even at the scale of comparatively small
and spatially isolated populations and that gene ow among disjunct patches can be signicant. As for most orchids,
P. pulcherrima has minute dust-like seeds that are dispersed by the wind, a feature often interpreted as an adaptive
measure aording the capacity for gene ow over relatively large distances (Phillips et al., 2012). However, in the
case of P. pulcherrima, pollinator-mediated gene ow is also likely to be an important factor in minimising genetic
dierences between populations, because the species’ deceptive mating system that relies on the foraging behaviour
of nectar-feeding bees to disperse pollen over relatively wide distances could potentially connect even geographically
Subsequent analysis of the spatial distribution of all adult plants mapped at the three sub-populations at WX
demonstrated signicant clumping of individuals; this phenomenon was most pronounced at ne to intermediate
scales (i.e. <40 m distance classes; Zhang et al., in review). On the other hand, signicant ne-scale genetic
structuring was detected at even shorter distance intervals (i.e. <10 m), with signicant negative relatedness (i.e.
greater genetic distance) being detected at larger distance intervals (>150 m). This structuring indicates greater
kinship among individuals situated closer to one another. This can be interpreted as evidence of geographic clumping
of related individuals (especially full and half siblings), which would be expected where seeds from the same capsule
predominantly land and germinate close to the mother plant. Indeed, statistical estimates of the relative contribution
of pollen and seed to total gene ow revealed seed dispersal to be more limited than pollen dispersal. This conrms
that, at the landscape level, genes are more eectively distributed by bee-mediated pollen ow than they are by wind-
mediated seed dispersal.
The genetic data also revealed maximum outcrossing rates for the mother plants from which the ten capsules were
harvested, suggesting that all seedlings were products of outcrossing. Moreover, all seedlings derived from the same
capsule were conrmed to share the same pollen parent, and therefore to be full siblings. Paternity analysis matched
all seedlings from four mother plants to four distinct pollen parents in the population with greater than 95% condence.
Since the location of each individual had been mapped, we were able to estimate the geographic distance between
the two parents, revealing pollen dispersal distances of between 113.6–345.6 m, with a mean (±SD) distance of 272.7
m (±108.4 m). Moreover, the pollen donor was located in a dierent sub-population to the mother plant in each of
these four cases. Genetic connectivity among sub-populations therefore appears to be maintained by bees that forage
over wide distances and across habitat types that are unsuitable for the orchid (Fig. 4; Zhang et al., in review).
This is good news for the recovery of natural populations of P. pulcherrima following decline and for the persistence of
articially established populations, since it demonstrates that the species is capable of gene ow over relatively wide
distances and among disjunct sites. Given that P. pulcherrima naturally colonises geographically isolated rock outcrops,
this capacity is probably an adaptive feature of the species’ ecological strategy. Taken together, the
results of our pollination experiments and genetic analyses provide evidence that reintroduction is a
reasonable approach for the conservation of the species at BNNR, because reintroduced plants are likely
to integrate genetically with the existing natural populations through the services of locally abundant pollinating bees.
More generally, these ndings underscore the necessity of an otherwise intact ecosystem in providing the ecological
support that the species needs to ensure pollination and gene ow.
Fig. 4. Genetic analysis of all adult plants of Phalaenopsis pulcherrima in three sub-populations at WX (WX-1, WX-2 and WX-3)
and of 1105 F1 seedlings raised from ten capsules harvested from plants at the same site revealed large pollen ow distances.
Seedlings from four capsules were matched to four pollen parents situated in dierent sub-populations to the four mother plants,
with either parent being separated by dense forest habitat in which the orchid does not grow. The four conrmed pollen ow
distances are shown below the bees. These ndings indicate that the orchid’s pollinators are capable of traversing distances of up
to 345.6 m to bring about pollination and thereby connect spatially isolated individuals.
Because of its elegant habit, wide variation in the colour of its attractive owers and its close phylogenetic relationship
with other Phalaenopsis species, P. pulcherrima has been extensively used in the development of new horticultural
hybrids for the ornamental plant trade. Historically, this has placed intense harvesting pressure on wild populations
to supply demand from breeders and growers. Although the species is now widely established in cultivation, wild-
collected plants continue to appear in trade at markets in Laos (authors, pers. obs.), South China (Gale et al., 2019),
Thailand (Santi Watthana, pers. comm.) and Vietnam (Vuong Truong, pers. comm.). Although data from markets in
South China (Gale et al., 2019) and from online traders in Vietnam (Vuong Truong, pers. comm.) suggest it is not now
commonly encountered in trade, even small numbers of plants are likely to be indicative of signicant local declines,
given the small size of most natural populations. The species is therefore still regarded as vulnerable to decline
throughout its range due to collection.
Phalaenopsis pulcherrima has also declined due to degradation and conversion of its habitat. The species’
monsoon forest habitat has been heavily logged throughout the Indochina region in the past three decades
(Campbell, 2009), and encroachment as a result of re, urban development and the expansion of agriculture,
forest plantations and goat herding have been cited as causes of the orchid’s decline in Thailand (Somran Suddee,
pers. comm.) and Vietnam (Leonid Averyanov, pers. comm.). However, it has not been possible to accurately quantify
the extent of the losses incurred due to these threats because the abundance of the species prior to impact was not
documented. Nevertheless, given the low level of natural fruit-set observed in Hainan, it is likely that the rate of decline
globally has outstripped the species’ natural capacity for recruitment and recovery. At the very least, it is known that
the species’ global area of occupancy has been reduced through extirpation of individual sites, and that some of the
35 sites conrmed with herbarium data (GBIF.org, 2019) have been lost.
Anecdotal reports of the status of the species in the wild in Hainan indicate that its known populations there have also
suered. The populations at DL, JFL, LD and YJ are all reported to comprise fewer individuals now as compared to
10–20 years ago (Chen Qing, pers. comm.). This is thought to be due primarily to the collection of plants for sale and
use in horticulture, as well as the result of habitat degradation. Such threats to plants in the wild persist, even where
populations occur in protected areas.
Although no orchids are aorded statutory protection in China (State Council of China, 1999), the harvesting of any
plant growing within a protected area is illegal (State Council of China, 1994). As such, the four known Phalaenopsis
pulcherrima populations that lie within BNNR (DL, DLH, WX and YJ), as well as population JFL in Jianfeng Ling Forest
Reserve, are all protected by law. However, owing to the fact that it does not occur within a protected area, population
LD is vulnerable to collection and habitat degradation.
The species receives varying levels of protection elsewhere in its range. All native orchids are protected by law in
Vietnam, whether they occur on protected land or not, under Government Decree No. 06/2019/ND-CP 2019, which
governs the management of endangered, precious and rare forest plants and animals. Similar safeguards are in place
in Thailand, where the harvesting of native wild orchids is punishable by substantial prison terms and nes under the
Wildlife Preservation and Protection Act 2019. In contrast, the 1996 Forest Law of Lao PDR and its 2007 amendment
prohibit the removal of biodiversity only from ‘total protection zones’ within Conservation Forest. Lao populations of
P. pulcherrima are therefore subject to legal protection only where they lie within certain designated areas. Orchids
receive a similar level of protection in Malaysia, where the National Parks Act 1980 restricts collection only within
All P. pulcherrima range countries are signatories of the Convention on International Trade in Endangered Species of
Wild Fauna and Flora (CITES). Under this agreement, signatories are obliged to monitor and regulate cross-border
trade in restricted species. All species of orchids are included and are assigned to one of two Appendices, depending
on the severity of perceived endangerment. Phalaenopsis pulcherrima is placed under Appendix II, which stipulates
that trade in wild-collected specimens may only proceed if ocial export and import permits are rst obtained from
both source and destination countries. CITES asserts that acquisition of a valid export permit requires documentation
(in the form of a non-detriment nding) conrming that the stated trade will not adversely impact wild populations
of that species. Given the species’ declining status and the typically small, isolated nature of its populations, it is
questionable whether a non-detriment nding could reasonably be issued for wild specimens of this species.
Populations of Phalaenopsis pulcherrima that occur beyond the boundaries of protected areas therefore remain vulnerable
to collection and habitat disturbance in certain range countries, including China. In this respect, population LD remains
particularly at-risk. However, the species is also known to be subject to poaching as a result of weak enforcement even
where legal safeguards are in place. Wild specimens continue to be encountered at plant markets across Southeast Asia,
and in some cases this trade is known to involve plants that have been illegally sourced or tracked across international
borders (Li et al., 2018; Gale et al., 2019). Adequate on-the-ground enforcement is therefore essential to ensure that
populations of P. pulcherrima (and indeed other commercially valuable species) remain secure. This may take the form of
patrols and monitoring of particular populations, as well as the development of site management plans to provide eective
oversight of protected habitats (to exclude grazing animals, minimise re risk and encourage the regeneration of a diverse
local ora and fauna, for example) for the persistence of this and other threatened species at the landscape level.
Although Phalaenopsis pulcherrima has not previously been Red-Listed, sucient data is now available for a full
global assessment under IUCN Criterion B (geographic range; IUCN, 2012). In China, the species is presently
regarded as Critically Endangered (CR; Ministry of Ecology and Environment of the People’s Republic of China and
the Chinese Academy of Sciences, 2019), and given that Hainan is the only province of China in which it occurs, this
can be considered a regional assessment for the island. However, ndings from the analyses described in this report
provide a clearer picture of species biology, status and threats, and so an updated regional assessment for Hainan
is also presented (IUCN, 2003). Since a reasonable estimate of population size in Hainan is available (see Table 1),
Criterion D (very small or restricted population) is applied in addition to Criterion B.
With an Area of Occupancy (AOO) of ca. 112 km2, P. pulcherrima may provisionally be considered EN. However, on
the strength of the analyses conducted for this report, it cannot reasonably be concluded that the global population
is severely fragmented because the species has, in principle at least, been shown to have a strong capacity for gene
ow among isolated (sub-)populations. With more than 35 populations documented across its range, of which at
least some are known to be reasonably well protected, it is unlikely that its global population comprises fewer than
ten threatened locations. Whilst the species is not regarded as having undergone extreme uctuation in the size or
structure of its global population, the species is inferred to have suered (and is projected to continue to suer) decline
in its area of occupancy; area, extent and/or quality of habitat; and in the number of locations, sub-populations and
mature individuals. Despite its restricted AOO, the species does not therefore meet the minimum number of sub-
criteria to be evaluated as falling in any category of threat (CR, EN or VU). In view of its limited range and declining
population trend, P. pulcherrima is instead assessed as Near Threatened, NT (B2 b: ii,iii,iv,v).
Regional assessment for Hainan
The species’ six populations on Hainan generate an Extent of Occurrence (EOO) of ca. 1363.3 km2 and an AOO of ca.
24 km2. The species is not severely fragmented, but it is known from fewer than ten locations. There is sucient evidence
to infer that the species has suered decline in its area of occupancy; area, extent and/or quality of habitat; and in the
number of locations, sub-populations and mature individuals. Its population size is estimated to amount to fewer than
1000 mature individuals. Since the species is resident on the island, and because it may reasonably be presumed for
now that this regional population does not currently experience any signicant immigration of propagules from elsewhere
in the range, P. pulcherrima may be regarded as Vulnerable VU (B1, B2 a, b: ii,iii,iv,v; D1) at the regional scale.
Summary biological characteristics
• Phalaenopsis pulcherrima is a long-lived perennial lithophytic or terrestrial herb adapted to growth in
the seasonally dry monsoon forest belt of tropical Asia.
• The species is known from at least 35 sites, including six on the island of Hainan. Most populations
comprise just a few scattered plants.
• The species achieves pollination through generalised food deception of bees foraging for nectar. An
abundance of co-blooming, rewarding magnet plants with a similar ower colour are critical to the
success of this system.
• Genetic diversity is maintained through outcrossing and long-range pollen ow, but fruit-set is low.
• Populations are threatened by collection and habitat destruction, even where they occur within
• Globally, P. pulcherrima is assessed as NT. Within Hainan, the species is assessed as VU.
In view of the inferred history of decline of Phalaenopsis pulcherrima in Hainan, the reintroduction of nursery-
raised seedlings into suitable host sites to bolster the species’ population in China was an express objective of this
project. Given the ndings of the ecological research component of this work, particularly in terms of conrming
the presence of eective pollinators and revealing landscape-level gene ow among existing populations in BNNR,
the establishment of new populations was deemed a viable course of action. Two key premises in undertaking this
work were that reintroduced plants should be raised from seeds harvested from open-pollinated wild plants growing
within BNNR and that new hosts sites should be selected within BNNR on the basis of their potential for long-term
Population WX in BNNR was selected as the seed source because the high number of individuals there meant that
removing the required number of capsules would have proportionately less impact as compared to the other, smaller
populations. Forty-two open-pollinated capsules were collected from the site in November 2015 and a further ten were
collected in November 2016. To ensure representation of as broad range of background genetic diversity as possible,
scattered mother plants across the entire site were chosen. Undehisced capsules judged to be mature based on
their outward appearance and texture were carefully removed from the mother plants using scissors and immediately
placed in paper bags (Plate 2-A). All 52 capsules were passed to the micropropagation facility at Dongfang Tengfei
Horticulture Biotech Ltd., a commercial orchid nursery based in southern Hainan, for processing. This company was
contracted by KFBG to produce 2500 seedlings from the capsules.
Micropropagation and hardening
The capsules were surface-sterilised with hypochlorous acid and then opened with scalpel and forceps at a clean
bench. Inspection of the contents revealed the seeds in 27 capsules to be immature with under-developed embryos
and therefore incapable of germination; unfortunately, these had to be discarded. Seeds from the remaining 25
capsules were sown onto a modied Vacin and Went medium (Vacin & Went, 1949) in sterilised glass jars, and these
were incubated under a 10 hour light/14 hour dark cycle at room temperature. Seeds from ten capsules (designated
CC1, HC2, HC3, HC5, IC5, IC6, IC7, AMP1, AMP3 and AMP8) germinated within the rst 2–4 weeks (Plate 2-B).
Those in the remaining 15 capsules failed to germinate altogether. After approximately 24 weeks, seedlings that had
grown from the germinated seeds were large enough to be transferred to fresh medium (i.e. sub-culture
1; Plate 2-C). A second sub-culture was conducted after another 12–15 weeks (sub-culture 2; Plate 2-D),
followed by a nal 12–15 weeks’ incubation. Seedlings were therefore ready for deasking after
approximately one year of in vitro growth, by which time they had attained about 3–5 cm in height and bore two to four
leaves (Plate 2-E). In total, 4071 seedlings were produced (Table 2). Of these, 1171 were made available for use in
the population genetics experiments described above, leaving 2900 for hardening in the nursery.
The seedlings were removed from the glass jars and their roots were carefully washed to remove all traces of the agar
medium. They were then planted individually in sphagnum moss in plastic pots and hardened for two months in an
80% shade house (Plate 2-F). Thereafter, the seedlings were transferred rst to a plastic-sheeted hot house for the
remainder of the winter and then to an open nursery under 70% shade. Over 800 seedlings were lost throughout this
process due to fungal infection and physical damage sustained during a typhoon. Approximately 2000 seedlings were
ready for planting out by summer 2017.
Table 2. Number of seedlings raised in vitro from ten open-pollinated capsules collected at population WX.
Four potential reintroduction sites were identied by BNNR sta and these were jointly inspected by KFBG and
BNNR sta in November 2016. All were deemed to be suitable on account of overall site features (open, seasonally
inundated, rocky shelves onto which the orchid could be directly rooted) and vegetation type (sparse monsoon forest).
Several typical associates of Phalaenopsis pulcherrima were found at the four sites, including the magnet plants
Helicteres angustifolia and Hedyotis consanguinea, suggesting that the orchid’s deceptive mating system could be
functional there. This was further evidenced by the presence of natural P. pulcherrima populations nearby. Three of
the four potential reintroduction sites were eventually selected on account of ease of access, which made monitoring
plant performance that much more practical (Table 3; Plate 2-G, 2-H, 3-A). It was agreed that the location of the three
sites would be kept secret to minimise the risk of disturbance and poaching.
Table 3. Features of three sites at BNNR selected for reintroduction of Phalaenopsis pulcherrima.
In order to evaluate the proposed method of attaching the seedlings to the rocky substrate and to assess
their capacity to adapt to in situ conditions straight out of the pot, a subset of 311 seedlings were collected from the
nursery and transported to site RDLH in July 2017 for a small-scale reintroduction trial (Plate 3-B). This trial was
intentionally scheduled for the onset of summer rains so as to promote root growth. By this stage, the seedlings were
about 5–7 cm tall with three to six leaves and copious wiry roots (Plate 3-C). The plants were taken out of their pots,
all sphagnum was removed from around their roots and a small blob of non-toxic bond was applied to the base of the
stem (Plate 3-D). The plants were then pressed rmly onto the bare rock surface and held in place for about 1 minute
whilst the bond hardened (Plate 3-E). The bond used in this trial was a solvent-free, water-based glue sold under the
name ECO-BOND (Plate 3-F). As far as practicable, the plants were planted at microsites deemed to be appropriate
for the growth of the species (crevices in bare rock with some shade); 150 of them were tagged with a plastic label
bearing a unique identication number to facilitate monitoring. The condition of each plant was recorded at the time
of planting, then one month later in August 2017 and again in September 2017 (Plate 3-G). On each site visit, plants
were judged to be either 1: weak/declining; 2: stable; 3: strong/increasing; 0: dead; or X: washed away. To make this
assessment, the condition of the whole plant (including roots, stem and leaves) was taken into account. Thus, a plant
with fewer leaves as compared to the previous observation might still be regarded as having undergone net growth
if its root network had expanded and become attached to the rock on which it had been glued. A higher proportion
of plants being scored as either 2 (stable) or 3 (strong/increasing) was taken as an indication of good overall
performance at the site. Plants were scored as X (washed away) if the bond (sometimes together with the label) was
still present on the rock surface but the plant was missing, suggesting a failure of the planting method itself.
By the end of this three-month trial, the ecacy of the reintroduction method and the performance of the plantlets
were evaluated as satisfactory: survivorship was high (92.7%), a signicant proportion of plants exhibited stable or net
growth (80.7%) and none of the plantlets had been washed away. The trial was therefore regarded as a success and
full reintroductions were initiated using all remaining seedlings.
In September 2017, 300 seedlings were planted out at site RYJ. In June 2018, a second cohort of 207 seedlings
was planted at RDLH (giving a total of 518 reintroduced seedlings at that site), a second cohort of 218 seedlings
was planted at RYJ (giving a total of 518 reintroduced seedlings at that site), and 665 seedlings were planted at site
REL27 (giving a total of 1701 seedlings reintroduced across all three sites). A proportion of the plants at each site
were labeled and their performance was monitored at 3- or 6-monthly intervals until July 2019 by applying the same
scoring system used in the trial (Plate 4-A, 4-B, 4-C, 4-D, 4-E). Summary statistics for the full reintroduction are shown
in Table 4.
These statistics reveal the performance of seedlings at site RDLH to be better than that at the other two sites:
survivorship, mean seedling state and the proportion of seedlings in either state 2 or 3 remained consistently higher,
and losses due to wash-away were consistently lower, at all monitoring intervals. Losses due to wash-away were
particularly high at sites RYJ and REL27, suggesting a relatively high incidence of failure of the planting method at
these two sites. This may be because the seedlings were not held in place long enough for the bond to dry when they
were planted out. Overall, mortality was highest and the state of the plants was poorest at REL27.
Across all ve cohorts and at all three sites, the overall performance of reintroduced seedlings was satisfactory: on
average, 46.7% of seedlings survived and 27.4% were stable or increasing at the end of the 24-month monitoring
period. Root growth of surviving plants was good, with the majority being rmly rooted to the substrate and therefore
capable of recovery even if their leafy shoots had suered during the process of adaptation to wild conditions (Plate
4-F). Therefore, despite the low mean state calculated for all three sites after 24 months (1.1), there
are grounds for cautious optimism that individuals will continue to accrue biomass and that the new
populations will stabilise (Plate 4-G). In any case, these gures indicate that, of the 1701 seedlings
planted out at the three reintroduction sites, approximately 794 survived as of July 2019, and that 466 of these
were stable or increasing. Given an estimated population size of approximately 750 individuals in the wild in Hainan
prior to the start of this project (Table 1), this collaborative eort has therefore eectively doubled the population of
Phalaenopsis pulcherrima in China.
Table 4. Summary performance of 652 monitored Phalaenopsis pulcherrima seedlings reintroduced at three sites in
BNNR over a two-year period.
* The state of each seedling was assessed as either 1 (weak/declining), 2 (stable), 3 (strong/increasing), 0 (dead) or X (washed
away). The proportion of seedlings scored as 2 or 3 provides an indication of overall cohort/site performance.
To place these gures in a broader context and so help evaluate the success of this reintroduction eort,
we sourced published examples of orchid reintroductions conducted worldwide by searching Google
Scholar and Web of Science databases for peer-reviewed studies using the search terms “orchid
reintroduction” and “orchid restoration”. Only those studies that made use of nursery-grown seedlings (as opposed
to mericlonal stock or wild plants) and which carried out post-reintroduction monitoring for at least 12 months were
retained. Figures for survivorship reported in the resulting body of literature are summarised in Table 5. Although very
few orchid reintroduction case studies have been published (and presumably more successes have been documented
than failures), the highest survivorship rate reported for a nursery-grown orchid reintroduced into the wild is that
for the terrestrial Eulophia cullenii in India (70.1% after two years; Decruse et al., 2013), and the lowest is that for
the epiphyte Epidendrum nocturnum in USA (7.0% after one year; Stewart, 2008). The survivorship achieved for P.
pulcherrima in Hainan (46.7% over two years) falls about midway among these gures and is above the average for
all studies combined (ca. 45.6%), suggesting solid execution in terms of site selection, technical methods and plant
performance. To the best of our knowledge, our project represents the rst attempt to reintroduce a tropical lithophyte.
Table 5. Reported survivorship of nursery-raised orchids reintroduced into the wild as compared to that achieved for Phalaenopsis
pulcherrima in Hainan.
In reviewing orchid conservation projects undertaken in Australia, Reiter et al. (2016) proposed four time-bound criteria
for evaluating the success of orchid reintroductions. Firstly, they argue that, within the rst three years, success should
be measured by the survival of 50–200 individuals at any one site, since this represents the minimum number required
to guarantee population viability in a long-lived, outbreeding species in terms of reproductive potential and genetic
diversity. The three populations of P. pulcherrima established in BNNR meet this criterion (RDLH: 286 individuals,
RYJ: 197 individuals, REL27: 195 individuals). Secondly, reintroduced plants should begin to ower and set fruit within
the rst seven years. Of the 320 surviving labeled plants, 18 (5.6%) owered within the 24-month monitoring period
(Plate 4-H). Based on this gure, we estimate that approximately 45 of the surviving 794 reintroduced plants owered
over this period. Moreover, one plant at RYJ was pollinated naturally and produced a capsule in 2019. This bodes
well for the ecological viability of the newly established populations, well in advance of the time-frame proposed by
Reiter et al. (2016). Thirdly, natural recruitment of new seedlings should be conrmed within 15 years and, lastly, the
reintroduced population ought to become self-sustaining thereafter, with recruitment equaling or exceeding mortality
in the long-term. Ongoing monitoring is needed to ascertain whether the three newly established P. pulcherrima
populations can meet these nal two criteria over the coming ten years or more.
Social learning for plant conservation
Restoring populations of endangered plants not only requires scientic knowledge of species biology but also
understanding of the practical steps that need to be taken for change on the ground. This latter aspect will almost
invariably require coordination and sharing among diverse actors. Far from being an optional extra, failure to reach out
and engage with government, landowners, volunteers and citizen scientists, botanic gardens, green groups, schools,
local communities and so on can curtail the translation of scientic knowledge into conservation action. Despite a
growing body of literature on orchid population ecology, seed biology and propagation, comparatively few orchid
reintroduction and restoration programmes have been undertaken precisely for this reason (Gale et al., 2018). The
successful planning and execution of the reintroduction of Phalaenopsis pulcherrima in Hainan owes a great deal to
the constructive relationship established among four key stakeholders.
Eective conservation management is an applied, multi-stakeholder process that takes place through social learning.
In the case of Phalaenopsis pulcherrima in Hainan, four key parties have come together to make population
persistence possible (Fig. 5). Two government bodies (HWCB and BNNRA) have played a vital role in oering
access to habitats and plants and in guaranteeing the protection of reintroduced populations through monitoring and
enforcement (Plate 5-A, 5-B). As a scientically-driven but goals-oriented NGO, KFBG has provided oversight in
designing protocols, fast-tracking research and applying results (Plate 5-C, 5-D). The involvement of Hainan University
has been invaluable in undertaking this research, analysing data and interpreting the results (Plate 5-E). Finally,
Dongfang Tengfei Horticulture Biotech Ltd. provided pivotal know-how and facilities in producing the seedlings, without
which reintroduction would not have been possible (Plate 5-F). In view of the valuable lessons learnt and outcomes
achieved, this partnership is regarded as exemplary for the design of future orchid restoration programmes. Execution
of the following recommendations stemming from this work will rest with BNNRA in ongoing dialogue with KFBG, but
also ideally in partnership with civil society, given that buy-in among corporations and the general public will be critical
in coming to terms with the threat posed by poaching and the commercial exploitation of wild orchids.
• Phalaenopsis pulcherrima can be readily propagated from seed. However, judging the maturity of open-
pollinated capsules can be dicult.
• Seedlings cultured in vitro can be deasked after one year and hardened within a further 4–6 months.
Nursery cultivation is straight-forward.
• Site selection is a crucial consideration to ensure the species’ habitat requirements are met.
• Attaching seedlings to bare rock surfaces using a non-toxic bond is a viable means of reintroducing
P. pulcherrima. However, care should be taken to select suitable microsites and hold plants in place
long enough to allow the bond to dry.
• Reintroduction success can be adequately ascertained through repeat observations and assessment
of plant condition of individually labeled plants over 24 months.
• Survivorship rates achieved for three newly established populations in BNNR compare favourably to
those reported for similar projects conducted on other orchid species around the world.
• The onset of owering and fruiting among reintroduced plants within two years bodes well for ecological
Fig. 5. Successful restoration of Phalaenopsis pulcherrima in Hainan, as measured ultimately by the formation of self-sustaining
reintroduced populations, has come about as a result of constructive interaction among four key stakeholders that enabled the
execution of six key processes.
Much has been achieved in the ve-year timespan of this project, but ultimate success will depend upon the long-
term viability of the reintroduced populations in BNNR, as well as on the continued presence of Phalaenopsis
pulcherrima in Hainan. The following ve recommendations are put forward to ensure this outcome is realised, as far
as is practicably possible. Implementation of these recommendations stands to benet other plant species that face
similar threats, as well as the habitats in which they grow. Given the paucity of applied orchid conservation projects
conducted to date, it is hoped that the approaches developed and lessons learnt for the reintroduction of this species
in Hainan will serve as a model for future restoration projects focusing on other orchids and indeed other endangered
plants, both in China and throughout the region.
To safeguard all Phalaenopsis pulcherrima populations in Hainan, enhanced statutory protection
through the addition of all wild orchids to the State Key Protected Plants List is encouraged. In so
doing, even populations that fall outside protected areas (such as population LD) will be legally
protected from disturbance and collection, as indeed will other orchids, which are currently the focus
of intense collecting pressure internationally.
To ascertain whether the three reintroduced populations stabilise and become self-sustaining, ongoing
monitoring is advised. Under the terms of the MOU signed by HWCB, BNNRA and KFBG in 2015, post-
project, 6-monthly monitoring is requested of BNNRA sta for a further three years after the nal joint
monitoring survey. This means that a further six surveys should be conducted (in December 2019, June
2020, December 2020, June 2021, December 2021 and June 2022). Data for each plant is not needed,
but it is expected that general observations are made on the overall condition of the populations, including
the occurrence of owering, fruiting and seedling recruitment. Observations made should be documented
and shared among all three parties.
To help raise awareness of the need for plant conservation more generally, HWCB, BNNRA and KFBG
are encouraged to publicise the lessons and outcomes of this project to peers and the general public
through traditional and online media, and in the course of workshops and conferences.
In view of the vulnerability of Phalaenopsis pulcherrima and other orchids to commercial exploitation, all
stakeholders (HWCB, BNNRA, KFBG, Hainan University and Dongfang Tengfei Horticulture Biotech Ltd.)
are encouraged to publicise the threat posed by wildlife trade to the public at large, and to seek ways to
mitigate its deleterious impacts. The time and cost invested in reintroducing P. pulcherrima in Hainan
alone cannot easily be scaled up to aid the restoration of the many other species aected by trade.
Stemming the problem at source must therefore be seen as the preferred approach.
Given the many other plants that have suered decline and which also face an uncertain future, all
stakeholders are encouraged to identify other priority species and to apply their experience, leverage
and know-how to assist in their recovery, using this collaborative work on Phalaenopsis pulcherrima as
We are immensely grateful to Yanni Mo at HWCB, Xiangming Wu and Xiaojiang Hong at BNNRA, Xiqiang Song at
Hainan University and Darren Chea at Dongfang Tengfei Horticulture Biotech Ltd. for their invaluable contribution
to this collaborative project. Sta at BNNR, especially Chen Qing, Zhiwei Liu, Donghua Yang and Jinyuan Jiang,
are acknowledged for their generous and enthusiastic support in the eld. Zhe Zhang, who conducted his PhD on
the ecology of Phalaenopsis pulcherrima in Hainan, is thanked for his devotion to his studies and for allowing us
to include some of his ndings in this report. Colleagues at KFBG, notably Mang Lung Cheuk, Pankaj Kumar and
Kelvin Yeung, are thanked for their assistance in processing data and preparing the gures used in this report. Andy
Brown, Bosco Chan, Philip Lo and Helen To provided technical advice and logistic support throughout the lifespan
of the project. Several orchid specialists, including Yibo Luo, Leonid Averyanov, Somran Suddee, Santi Watthana,
Vuong Truong and Ong Poh Teck, provided information on the biology, ecology and legal status of P. pulcherrima in
dierent parts of the species’ range.
Agamospermy – A form of asexual reproduction in which seeds develop from unfertilised ovules.
Allele – A variant form of a gene responsible for a particular character.
Allelic richness – A measure of genetic diversity indicating the number of dierent alleles present at a particular locus
within a sampled population.
Area of Occupancy (AOO) – A measure of a species’ range based on the number of grid cells of xed dimensions
actually occupied by all individuals of that species. To quantify distribution for a Red List assessment, a grid cell of
2 × 2 km is typically used.
Autogamy – Self-fertilisation, usually in reference to plants capable of self-pollination.
Colonisation bottleneck – A signicant reduction in genetic diversity associated with the establishment of a new
population by just one or a few individuals. Overcoming the reduced tness suered as a result of this may require
enhanced gene ow with nearby populations.
Column-foot – An extension of the base of the column present in some orchids.
Deceptive pollination – A plant mating strategy in which owers attract pollinating insects by sending a false signal
that imitates certain rewarding conditions, such as food, shelter or a mate.
Deasking – The process of removing micropropagated seedlings from glass asks for hardening.
Drift – Random change in the frequency of alleles in a population due to chance eects associated with a small
population size or restricted opportunities for mating.
Epiphyte (adj. epiphytic) – A plant that grows on another plant for support or to occupy a niche provided by that
plant, but not a parasite.
Extent of Occurrence (EOO) – A measure of a species’ range based on the area enclosed by a polygon drawn
around all individuals of that species. In quantifying distribution for a Red List assessment, EOO will usually be greater
F1 generation – The rst generation of ospring produced by a parental generation.
Geitonogamy – Pollination of a ower by pollen from another ower belonging to the same (or a genetically identical)
Generalised food deception – A plant mating strategy in which non-rewarding owers exploit the non-specic food-
seeking behaviour of their pollinators.
Hardening – The process by which laboratory-grown plants are adapted to nursery or outdoor conditions.
Heterozygosity (adj. heterozygous, heterozygotic) – The condition of having two (or more) distinct versions of
a gene (i.e. allele) present at any given locus in a single genotype. High heterozygosity denotes greater genetic
Inbreeding depression – The production of ospring from the mating of genetically similar or identical individuals,
often resulting in reduced tness as a result of increased homozygosity.
Lithophyte (adj. lithophytic) – A plant that grows on stone.
Locus (pl. loci) – A specic, xed position on a chromosome where a particular gene or genetic marker is located.
Magnet plant – A plant species that functions ecologically to attract pollinators to another sympatric plant species.
Microsatellite – A piece of repetitive DNA that occurs in an organism’s genotype and from which the degree of overall
genotypic variability of that individual can be measured. Based on this measure, the level of genetic similarity among
dierent individuals can be inferred.
Non-rewarding ower – A ower that does not oer a reward (such as nectar or other food-source) to visiting insects
in return for their pollinating service.
Pollinarium – The functional unit of pollen transfer in orchid pollination, consisting of two or more pollinia together
with certain accessory structures, such as stipe and viscidium in some cases.
Pollinium (pl. pollinia) – An agglutination of pollen grains into a single unit.
Reintroduction – The process of returning individuals of a species to part of their original range from which they have
Restoration – The process of returning degraded populations or habitats to a self-sustaining and ecologically viable
Sub-culture – The process by which micropropagated seedlings are transferred from one in vitro growing medium to
another in order to allow further growth and development.
Sympatric – Of two or more species that occur within the same area.
Terrestrial – A plant that grows on the ground with its roots in the soil or leaf litter.
Velamen – A spongy outer layer of dead cells on the roots of certain orchids (especially epiphytic and lithophytic
species) that absorbs and retains moisture.
Xenogamy – Pollination of a ower by pollen from another genetically distinct individual (i.e. cross-pollination).
Astafurova, Y.V., Pesenko, Y.A. 2005. Contributions to the Halictid fauna of the eastern palaearctic region:
subfamily Nomiinae (Hymenoptera: Halictidae). Far Eastern Entomologist 154: 1–16.
Campbell, I.C. 2009. The Mekong – Biophysical Environment of an International River Basin. Academic Press,
Chen, T., Zhu, H., Chen, J., Taylor, C.M., Ehrendorfer, F., Lantz, H., Funston, A.M., Pu, C. 2011. Rubiaceae. Pp.
57–368 in: Wu, Z.-Y., Raven, P.H., Hong, D.-Y. (eds.), Flora of China Vol. 19. Science Press, Beijing and Missouri
Botanical Garden Press, St. Louis.
Christenson, E.A. 2001. Phalaenopsis – A Monograph. Timber Press, Portland.
Decruse, S.W., Gangaprasad, A., Seeni, S., Menon, V.S. 2003. Micropropagation and ecorestoration of Vanda
spathulata, an exquisite orchid. Plant Cell, Tissue and Organ Culture 72: 199–202.
Decruse, S.W., Reny, N., Shylajakumari, S., Krishnan, P.N. 2013. In vitro propagation and eld establishment of
Eulophia cullenii (Wight) Bl., a critically endangered orchid of Western Ghats, India through culture of seeds and
axenic seedling-derived rhizomes. In Vitro Cellular & Developmental Biology – Plant 49: 520–528.
Dyer, A.G. 2006. Discrimination of ower colours in natural settings by the bumblebee species Bombus terrestris
(Hymenoptera: Apidae). Entomologia Generalis 28: 257–268.
Dyer, A.G., Boyd-Gerny, S., McLoughlin, S., Rosa, M.G.P., Simonov, V., Wong, B.B.M. 2012. Parallel evolution of
angiosperm colour signals: common evolutionary pressures linked to hymenopteran vision. Proceedings of the
Royal Society B: Biological Sciences 279: 3606–3615.
Engel, M.S. 2007. A new Amegilla of the zonata group from Malaysia and Thailand (Hymenoptera: Apidae).
Transactions of the Kansas Academy of Science 110: 16–22.
Gale, S.W., Fischer, G.A., Cribb, P.J., Fay, M.F. 2018. Orchid conservation: bridging the gap between science and
practice. Botanical Journal of the Linnean Society 186: 425–434.
Gale, S.W., Kumar, P., Hinsley, A., Cheuk, M.L., Gao, J., Liu, H., Liu, Z.-L., Williams, S.J. 2019. Quantifying the trade
in wild-collected ornamental orchids in South China: diversity, volume and value gradients underscore the primacy of
supply. Biological Conservation 238: 108204.
GBIF.org. 2019. Global Biodiversity Information Facility Home Page. Available at: http://www.gbif.org (accessed 8
Govaerts, R., Bernet, P., Kratochvil, K., Gerlach, G., Carr, G., Alrich, P., Pridgeon, A.M., Pfahl, J., Campacci, M.A.,
Holland-Baptista, D., Tigges, H., Shaw, J., Cribb, P., George, A., Kreuz, K., Wood, J. 2019. World Checklist of
Orchidaceae. Facilitated by Royal Botanic Gardens, Kew. Available at: http://wcsp.science.kew.org (accessed 11
IUCN. 2003. Guidelines for Application of IUCN Red List Criteria at Regional Levels Version 3.0. IUCN Species
Survival Commission, Gland and Cambridge.
IUCN. 2012. IUCN Red List Categories and Criteria Version 3.1. IUCN Species Survival Commission, Gland and
IUCN/SSC. 2013. Guidelines for Reintroductions and Other Conservation Translocations Version 1.0. IUCN Species
Survival Commission, Gland.
Jersáková, J., Johnson, S.D., Kindlmann, P. 2006. Mechanisms and evolution of deceptive pollination in orchids.
Biological Reviews 81: 219–235.
Jin, X., Li, D., Ren, Z., Xiang, X. 2012. A generalized deceptive pollination system of Doritis pulcherrima (Aeridinae:
Orchidaceae) with non-recongured pollinaria. BMC Plant Biology 12: 67.
Kumar, P., Gale, S.W., Pedersen, H.Æ., Phaxaysombath, T., Bouamanivong, S., Fischer, G.A. 2018. Additions to the
orchid ora of Laos and taxonomic notes on orchids of the Indo-Burma region. Taiwania 63: 61–83.
Li, J., Gale, S.W., Kumar, P., Zhang, J., Fischer, G.A. 2018. Prioritizing the orchids of a biodiversity hotspot for
conservation based on phylogenetic history and extinction risk. Botanical Journal of the Linnean Society 186: 473–
Lindley, J. 1833. The Genera and Species of Orchidaceous Plants. Ridgways, London.
Menzel, R., Müller, U. 1996. Learning and memory in honeybees: from behavior to neural substrates. Annual Review
of Neuroscience 19: 379–404.
Ministry of Ecology and Environment of the People’s Republic of China and the Chinese Academy of Sciences.
2019. The China Biodiversity Red List of Higher Plants. Available at: http://www.mee.gov.cn/gkml/hbb/bgg/201309/
t20130912_260061.htm (accessed 8 October 2019).
Ortega-Larrocea, M.P., Rangel-Villafranco, M. 2007. Fungus-assisted reintroduction and long-term survival of two
Mexican terrestrial orchids in the natural habitat. Lankesteriana 7: 317–321.
Phillips, R., Dixon, K.W., Peakall, R. 2012. Low population genetic dierentiation in the Orchidaceae: implications for
the diversication of the family. Molecular Ecology 21: 5208–5220.
Reiter, N., Whiteld, J., Pollard, G., Bedggood, W., Argall, M., Dixon, K., Davis, B., Swarts, N. 2016. Orchid
reintroductions: an evaluation of success and ecological considerations using key comparative studies from Australia.
Plant Ecology 217: 81–95.
Rundel, P.W. 1999. Forest Habitats and Flora in Lao PDR, Cambodia, and Vietnam. WWF Indochina Programme
Segovia-Rivas, A., Meave, J.A., González, E.J., Pérez-García, E.A. 2018. Experimental reintroduction and host
preference of the microendemic and endangered orchid Barkeria whartoniana in a Mexican tropical dry forest.
Journal of Nature Conservation 43: 156–164.
Smith, Z.F., James, E.A., McLean, C.B. 2007. Experimental reintroduction of the threatened terrestrial orchid Diuris
fragrantissima. Lankersteriana 7: 377–380.
State Council of China. 1994. Regulations of the People’s Republic of China on Nature Reserves. Available at: www.
gov.cn/fg/2005-09/27/content_70636.htm (accessed 22 July 2019).
State Council of China. 1999. State Key Protected Plants List. Available at: www.forestry.gov.cn/yemian/minglu1.htm
(accessed 22 July 2019).
Stewart, S.L. 2008. Orchid reintroduction in the United States: a mini-review. North American Native Orchid Journal
Tang, Y., Gilbert, M.G., Dorr, L.J. 2007. Sterculiaceae. Pp. 302–330 in: Wu, Z.-Y., Raven, P.H., Hong, D.-Y.
(eds.), Flora of China Vol. 12. Science Press, Beijing and Missouri Botanical Garden Press, St. Louis.
Vacin, E.F., Went, F.W. 1949. Some pH changes in nutrient solutions. Botanical Gazette 110: 605–613.
Worldweatheronline. 2019. World Weather Online Home Page. Available at: https://www.worldweatheronline.com/
(accessed 29 August 2019).
Wright, M., Cross, R., Dixon, K., Huynh, T., Lawrie, A., Nesbitt, L., Pritchard, A., Swarts, N., Thomson, R. 2009.
Propagation and reintroduction of Caledenia. Australian Journal of Botany 57: 373–387.
Wu, K., Zeng, S., Lin, D., Teixeira da Silva, J.A., Bu, Z., Zhang, J., Duan, J. 2014. In vitro propagation and
reintroduction of the endangered Renanthera imschootiana Rolfe. PLoS one 9: e110033.
Yam, T.W., Tay, F., Ang, P., Soh, W. 2011. Conservation and reintroduction of native orchids of Singapore – the next
phase. European Journal of Environmental Sciences 1: 38–47.
Zeng, S., Wu, K., Teixeira da Silva, J.A., Zhang, J., Chen, Z., Xia, N., Duan, J. 2012. Asymbiotic germination,
seedling development and reintroduction of Paphiopedilum wardii Summerh., an endangered terrestrial orchid.
Scientia Horticulturae 138: 198–209.
Zhang, Z., Gale, S.W., Li, J.-H., Fischer, G.A., Ren, M.-X., Song, X.-Q. In review. Pollen-mediated gene ow ensures
connectivity among spatially discrete sub-populations of Phalaenopsis pulcherrima, a tropical food-deceptive orchid.
BMC Plant Biology.
Plate 1. A. Phalaenopsis pulcherrima growing lithophytically in central Laos. B. Phalaenopsis pulcherrima growing as a terrestrial in
coarse, shallow soil in Hainan. C. Phalaenopsis pulcherrima can form sizeable clonal clumps through lateral branching of the stem.
D. The species’ succulent leaves and velamen-enclosed roots are adaptations to its dry, exposed habitat. E. The blue-banded
bee Amegilla zonata carrying pollinia of P. pulcherrima on its head in BNNR. F. Amegilla zonata visiting a ower of P. pulcherrima
in BNNR. G. Amegilla zonata visiting a ower of Helicteres angustifolia in BNNR. H. Amegilla zonata visiting a ower of Hedyotis
consanguinea in BNNR.
Plate 2. A. A large clump of Phalaenopsis pulcherrima bearing capsules. B. Prolic germination of P. pulcherrima seeds in vitro.
C. Seedlings of P. pulcherrima ready for sub-culture 1. D. Seedlings of P. pulcherrima following sub-culture 2. E. Seedlings of P.
pulcherrima ready for deasking. F. Micropropagated seedlings of P. pulcherrima undergoing hardening in the nursery. G. View of
reintroduction site RDLH. H. View of reintroduction site RYJ.
Plate 3. A. View of reintroduction site REL27. B. Micropropagated Phalaenopsis pulcherrima seedlings ready for reintroduction. C.
Seedlings of P. pulcherrima being prepared for planting-out. D. BNNR and KFBG sta applying non-toxic bond to the base of P.
pulcherrima seedlings. E. Reintroduced seedling glued in place at RDLH. F. ECO-BOND, the glue used to ax the seedlings onto
rocks at the reintroduction sites. G. BNNR sta checking labels of reintroduced seedlings during a monitoring survey.
Plate 4. A. Reintroduced seedling assessed as weak/declining (state 1). B. Reintroduced seedling assessed as stable (state
2). C. Reintroduced seedling assessed as strong/increasing (state 3). D. Reintroduced seedling assessed as dead (state 0). E.
Reintroduced seedling assessed as washed away (state X). F. Healthy reintroduced seedling with copious roots rmly attached to
the rock substrate, despite a reduction in the size of its leafy shoot. G. Cluster of well-established reintroduced seedlings, some
showing vigorous lateral branching, at RDLH. H. Reintroduced seedling owering for the rst time two years after planting-out at
Plate 5. A. Jinyuan Jiang and Xiaojiang Hong with Jihong Li and Stephan Gale surveying plants at population WX. B. BNNRA sta
from Dong Liu Forest Station and Chen Qing planting out seedlings at RDLH. C. Jihong Li, Gunter Fischer and Stephan Gale in
BNNR. D. Jihong Li with Donghua Yang at RYJ. E. Hainan University sta, including Xiqiang Song and Zhe Zhang, with Yi-bo Luo
from the Institute of Botany, Chinese Academy of Sciences, and KFBG sta, during a visit to Danzhou Campus. F. Darren Chea in
his nursery at Dongfang Tengfei Horticulture Biotech Ltd.