AJB Advance Article published on June 16, 2017, as 10.3732/ajb.1700078.

The latest version is at http://www.amjbot.org/cgi/doi/10.3732/ajb.1700078

RESEARCH ARTICLE

A M E R I C A N J O U R N A L O F B O TA N Y

B R I E F C O M M U N I C AT I O N

Pollinia transfer on moth legs in Hoya carnosa

(Apocynaceae)1
Ko Mochizuki2, Saori Furukawa, and Atsushi Kawakita

PREMISE OF THE STUDY: The genus Hoya (Apocynaceae; Asclepiadoideae) is characterized by a set of complex floral characters unique among the asclepi-
ads, but their role in pollination is poorly understood. Here, we report a new mechanism of asclepiad pollination in the wax plant Hoya carnosa: the polli-
naria are transferred on the legs of medium or large settling moths.
METHODS: Floral visitors and their behavior were observed on Amami-Oshima Island during 2013–2015, and the efficacy of different pollinators was deter-
mined by counting the pollinarium loads on different flower visitors. The floral characters were studied to establish the process of pollination in relation
to pollinator behavior on the flower.
KEY RESULTS: Hoya carnosa was visited by various settling moths at night, but pollinia attachment was confirmed predominantly on the legs of the large
moth Erebus ephesperis (Noctuidae) and less frequently on the legs of the medium-sized moths Bastilla arcuata (Crambidae) and Cleora injectaria (Geome-
tridae). The moths walked around and searched for nectar on the inflorescence, and the corpusculum became clipped to the arolia (adhesive pads on
moth tarsi) as they stepped on the pollinaria between the staminal corona. The downward spherical inflorescence of aggregated flowers with flat, velvety
petals and a slippery corona provides restricted footholds for the visitors, which efficiently leads pollinator legs to the pollinaria.
CONCLUSIONS: Although the pollination system of Hoya is largely unknown, pollination by insect legs may be a major pollination system in this genus
because these basic floral characters are shared among many species.

KEY WORDS arolium; Erebus ephesperis; floral traits; Hoya carnosa; Japan; noctuid pollination; nocturnal pollinator; pollen transfer efficiency; pollination
system; settling moth

Most flowering plants rely on animal pollen vectors for successful
reproduction, and evolutionary specialization to pollinators is
considered one of the mechanisms generating the morphological
diversity of flowers in angiosperms (Grant, 1949; Stebbins, 1970;
Sargent, 2004; Crepet and Niklas, 2009; Armbruster, 2014). Plants
have diverse mechanisms for achieving specialized pollination, and
some plants have specialized not only to certain taxa of pollinators
but also to pollen transfer by specific parts of the pollinator’s body
(Armbruster et al., 2014). Pollen placement on restricted parts of
a pollinator body sometimes occurs in plants with granular pollen
(e.g., Armbruster et al., 1994). However, restricted pollen place-
ment is much more common in orchids and asclepiads, whose pol-
len is packed into sacs known as pollinia; pollinia are subtended by
accessory structures (viscidium in many orchids and corpusculum
1
Manuscript received 23 February 2017; revision accepted 27 April 2017.
Center for Ecological Research, Kyoto University, Otsu, Shiga 520-2113, Japan
2
Author for correspondence (e-mail: ko.mochizuki@ecology.kyoto-u.ac.jp); ORCID id:
0000-0003-2739-5420
https://doi.org/10.3732/ajb.1700078
in asclepiads) that attach precisely to specific parts of the pollinator
body (Harder and Johnson, 2008). For example, Pauw (2006)
showed that a group of sympatric orchid species that all specialized
to a single oil-collecting bee species, Rediviva peringueyi, reduced
interspecific reproductive interference through differences in pol-
linium attachment sites. Another example is the African orchid
Disa chrysostachya, which attaches the pollinia only to the feet of
pollinating birds as the birds perch on the spike-like inflorescence
consisting of numerous small flowers (Johnson and Brown, 2004).
Animal legs, especially those of insects, are commonly used for pol-
linia transfer (Frost, 1965; Ollerton et al., 2003; Peter et al., 2009),
but plants whose pollinia are transferred only on legs are rela-
tively uncommon (Eisikowitch, 1986; Liede and Whitehead, 1991;
Coombs et al., 2009). In this study, we report a new pollination sys-
tem in Hoya carnosa: settling moths walk on the inflorescence to
feed on nectar and transfer pollinia on their legs.
The wax plant genus Hoya R.Br. (Apocynaceae) is the most speciose
genus in the Asclepiadoideae subfamily, comprising more than 300
species distributed throughout Southeast Asia, Melanesia, and

A M E R I C A N J O U R N A L O F B OTA N Y 104(6): 1–8, 2017; http://www.amjbot.org/ © 2017 Botanical Society of America • 1

Copyright 2017 by the Botanical Society of America
2 • A M E R I C A N J O U R N A L O F B OTA N Y
Australia with some species in East Asia, India, and Sri Lanka (Kleijn
and Van Donkelaar, 2001; Wanntorp et al., 2014). Most species are epi-
phytic lianas with opposite succulent leaves and extra-axillary umbel-
late floral clusters (the flowers are produced on the persistent peduncle,
Fig. 1A; Forster et al., 1998). Because they are easy to grow and have
showy flowers, Hoya are popular house plants. The flowers have a pen-
tamerous corolla with a prominent star-shaped gynostegium formed
by fused male and female organs at the center (gs in Fig. 1B; Kunze and
Wanntorp, 2008). The five diamonds in the gynostegium are staminal
corona derived from anthers (cs in Fig. 1B, C; Forster and Liddle, 1992;
Kunze and Wanntorp, 2008), a character unique to the genus Hoya. As
with other asclepiads, the pollen grains of Hoya are packed into the
pollinium, and a pair of pollinia is connected by the caudicles to the
corpusculum, a mechanical clip that grips a pollinator’s body. A pair of
pollinia thus connected by the corpusculum is known as a pollinarium
(Fig. 1D, E). The positional placement of the pollinarium in Hoya is
upside-down compared with other asclepiads (po and cl in Fig. 1C, D;
Wanntorp, 2007). The outer side of the pollinia is framed by the pel-
lucid margin, which is trapped by the guide rails upon pollination. The
pollen tubes then germinate from the pellucid margin and grow
externally within the stigmatic chamber before reaching the stigma
(Fig. 1D; Forster, 1992). Such a pollination mechanism involving pel-
lucid margin is also found in Stapeliae (Meve and Liede, 1994). In as-
clepiads, nectar is typically secreted from the nectary inside the slit
(stigmatic chamber; n in Fig. 1C) into the nectar tube (nt in Fig. 1C);
however, in most species of Hoya, the nectary has lost its function, and
nectar secretion is instead transferred to the secondary nectary on the
surface of the hollow made by the anther skirt, modified anther lobes
fused to the staminal corona (ns in Fig. 1C; Wanntorp and Kunze,
2009). Consequently, nectar is stored in the space between the anther
skirt and the corolla (np in Fig. 1C). These unique floral characters of
the genus Hoya are well understood anatomically, but their adaptive
meaning, the way these characters are involved in pollination, remains
unknown. Despite the large diversity of Hoya in Asclepiadoideae and
their remarkable floral characters, studies on Hoya pollination are sur-
prisingly scarce. The only study on Hoya pollinators, by Forster (1992)
using H. australis, reported that a diurnal skipper butterfly transfers
the pollinaria, but the author also suggested the possibility of nocturnal
pollination because of the increased scent production at dusk.
Hoya carnosa (L.f.) R.Br. is a representative Hoya species that is
distributed in southern Japan, Taiwan, and central to southern China,
favoring humid, shaded forest edges, often in limestone areas
(Yamazaki, 1993). The plant flowers continuously from May to Octo-
ber, with a downward, spherical inflorescence formed by aggregated,
salver-shaped, white-colored, hairy flowers (Fig. 1A, B). The white
color of the corolla provides a good contrast against the pinkish red
staminal corona (Fig. 1B). The pollination biology of H. carnosa is
unknown, but the nocturnal fragrance emission and nectar produc-
tion (Altenburger and Matile, 1988; Matile, 2006) and the whitish
flower color are suggestive of moth pollination (Faegri and van der
Pijl, 1979; Willmer, 2011). However, the flowers lack the nectar spurs
or corolla tubes typical of some moth-pollinated plants.
In this study, we investigated the pollinators of H. carnosa by directly
observing flower visitors during both day and night. To evaluate the
visitors as pollinators, they were captured, and we recorded the pres-
ence/absence of pollinarium attachment, the body part where the pol-
linaria attached, as well as the numbers of corpuscula and pollinia. On
the basis of our observations on pollinaria attachment, visitor behavior,
and the process of pollination, we discuss the functional significance of
the complex floral features of the spectacular plant genus Hoya.
MATERIALS AND METHODS
Study site—Field research was conducted on Amami-Oshima Is-
land, Kagoshima Prefecture, in the Amami Archipelago of southern
Japan. Hoya carnosa is patchily distributed throughout Amami-
Oshima Island, and each patch usually comprises one to several indi-
viduals. For the observations, two patches in Yamato-son (Yamato1:
28.328181, 129.304496 [geographical coordinates in decimal degrees];
Yamato2: 28.350198, 129.387118) and one in Uken-son (28.266855,
129.324864) were selected. Details of the field observations are pro-
vided in online Appendix S1 (see the Supplemental Data with this
article). These patches contained at least five flowering inflores-
cences at a time and plants that produced fruit, indicating suc-
cessful pollination.
Observations of floral characters—To assess how the floral charac-
ters are involved in the pollination process, we studied the floral char-
acters in detail using a stereomicroscope and scanning electron
microscope (SEM, Hitachi TM3000). Although pollination is consid-
ered to be achieved once the pellucid margin of the pollinium is
trapped by the narrow guide rail, there has been only one observation
of this, made in H. australis (Forster, 1992). Because the mechanism
of pollinia deposition in Hoya is still not well understood, we con-
firmed the placement of pollinia deposition in H. carnosa. Among
the flowers collected for the purpose of estimating pollinia transfer
efficiency (below), flowers pollinated naturally were dissected to ob-
serve pollen tube growth. Pollen tubes in asclepiads are visible with a
microscope without staining because a bundle of pollen tubes grows
externally before entering the stigmatic tissue (Sage and Williams,
1995). This information was used to construct the pollination pro-
cess, together with the observed pollinator behavior.
Pollinator observation—Floral visitors were observed on one or
two flowering individuals during both the day and the night during
May–August in 2013–2015. Visitor behavior was observed at a dis-
tance of at least 2 m from the inflorescence so as not to disturb the
visitor. All insect visitors were captured with an insect net and sac-
rificed either with ethyl acetate or by crushing the thorax with a
finger. The taxa of any visitors that an observer could not capture
were recorded. After collection, the insect visitors were identified,
and the presence/absence of pollinaria, the body part where the
pollinaria attached, and the numbers of corpuscula and pollinia
carried on their bodies were determined using a stereomicroscope.
The pollinaria were categorized into pollinaria with two pollinia
(complete pollinaria), pollinaria with one pollinia (half pollinaria),
and pollinaria without pollinia (Coombs et al., 2012; Wiemer et al.,
2011). The detailed placement of the pollinaria attached to the in-
sect body was studied using SEM.
Pollinia transfer efficiency—In asclepiads, pollinaria carried on the
pollinator body contain information on how often pollinia are
picked up and expended (Theiss et al., 2007; Wiemer et al., 2011).
Because the corpusculum remains on the pollinator body even after
the pollinia are used for pollination, the number of corpuscula on a
pollinator body indicates the number of successful pollinaria re-
movals. Thus, the proportion of pollinia removed from the corpus-
culum is an estimate of the rate at which pollinia are expended,
either as a result of successful pollination or loss during transfer
(Wiemer et al., 2011). Based on the number of corpuscula and pollinia
 J U N E 2017 , V O LU M E 104 • M O C H I Z U K I E T A L. — P O L L I N I A T R A N S F E R O N M OT H L E G S I N H OYA • 3

FIGURE 1 Floral structure of Hoya carnosa. (A) Flowering inflorescence. (B) Front view of the flower. (C) Cross section of flower, cut along the line indicated by
white and gray arrows in (A). (D) Front view of a pollinaria. (E) SEM image of the corpusculum. The narrow slit in the corpusculum is a clip that pinches an in-
sect’s body. (F) SEM image of the guiderail and nectar tube. The left guiderail is slightly open. A pollinium is inserted in the right guiderail. (G) SEM image of
dissected pollinated flower preserved in 70% ethanol. Pollen tubes are growing from the pellucid margin into the lower part of the stigma, which forms a
membrane-like structure. as, anther skirt; c, corolla; ca, caudicle; cl, corpusculum (clip); cs, staminal corona; ft, filament tube; g, guide rail; gs, gynostegium; n,
nectary; ns, secondary nectary; np, nectar pool (reward); nt, nectar tube; o, ovules; pe, pellucid margin; po, pollinium; pt, pollen tube; s, stigma; sh, style head.
4 • A M E R I C A N J O U R N A L O F B OTA N Y

remaining on the insect body, the proportion of missing pollinia
for each individual pollinator was calculated as the number of miss-
ing pollinia divided by double the number of corpuscula.
In addition to pollinator observation, 24 flowers were randomly
collected from three inflorescences of one large individual at the
Uken-son site on 19 May 2015. The collected flowers were preserved
in 70% ethanol until observation. The number of pollinaria removed
and the inserted pollinia were counted with a stereomicroscope, and
the pollinia transfer efficiency was calculated by dividing the number
of inserted pollinia by the number of removed pollinia (Wiemer
et al., 2011). All inserted pollinia were confirmed to have grown pollen
tubes into the stigma; thus, they were likely a result of successful pol-
linia deposition (Fig. 1F, G; Forster, 1992). The difference between
the proportion of missing pollinia (based on information from pol-
linators) and the pollinia transfer efficiency (based on information
from flowers) is indicative of the efficiency of the pollination system.
The two values are equivalent when pollinators do not lose the pol-
linia during transfer; however, when pollinators frequently lose the
pollinia, the proportion of missing pollinia becomes much higher
than the pollinia transfer efficiency.
RESULTS
Flower visitors and their behavior—Throughout the observation
period, H. carnosa flowers received only a few visits from the but-
terfly Ideopsis similis during the daytime, but they received many
visits from moths after sunset. Overall, 129 individual visitors
belonging to 23 species of eight insect families were recorded dur-
ing the 51.75-h observation period (Table 1). The visitors were
composed of hawkmoths (3 individuals), crane flies (5 individuals),
and settling moths of the families Arctiidae, Crambidae, Geometri-
dae, Noctuidae, and Pyralidae (121 individuals). The settling moths
gathered on the inflorescences at night, and they walked around,
searching for nectar (video in online Appendix S2). Visitation to an
inflorescence usually started between 1800 and 2000 hours and per-
sisted until midnight (2400 hours). Hawkmoths were observed
only at dusk, hovering in front of inflorescences and foraging for
nectar.
Pollinators, proportion of removed pollinia, and pollinia transfer
efficiency—Despite the broad range of visitor moth taxa, pollinaria
were attached to only 1 of 27 individuals of medium-sized Bastilla
arcuata (Crambidae, wing length =14.24 ± 1.36 mm); 1 of 7 indi-
viduals of Cleora injectaria (Geometridae, wing length = 20.12 ±
1.31 mm); and all 13 individuals of large-sized Erebus ephesperis
(Noctuidae, wing length = 44.24 ± 2.12 mm). They invariably car-
ried pollinaria on the adhesive pads on their tarsi, or arolia (Table 1;
Fig. 2A–D). The average number of corpuscula carried by one E.
ephesperis individual was 8.1 ± 2.3 (mean ± SD, N = 13), and the
average numbers carried by B. arcuata and C. injectaria were 1 and
3, respectively. For E. ephesperis, 88% (68 legs of 13 individuals) of
the legs carried at least one corpusculum, attached only to the arolia
(Table 2). Among the 112 pollinaria carried, 37 were complete pol-
linaria, 34 were half pollinaria, and 41 pollinaria had no pollinia.
Overall, 116 pollinia of 224 were lost from the corpuscula; there-
fore, the proportion of missing pollinia for all individual pollina-
tors was 0.52 (Table 2). Of the 24 flowers collected, 17 (71%) had

TABLE 1. Insect visitor during the diurnal (27.83 h) and nocturnal (23.92 h) observation, and the absence and presence of pollinia on the body.

Captured No. individuals Forewing length Pollinaria

Order Family Species (no. observed) carrying pollinaria mm, mean ± SD) attachment site
Lepidoptera Noctuidae
Erebus ephesperis 13 (18) 13 44.24 ± 2.12 Arolia
Bastilla arcuata 1 0 24.11
Ecliptopera umbrosaria 1 0 16.95
Arctiidae
Cyana unipunctata 1 0 10.19
Crambidae
Bradina erilitoides 27 1 14.24 ± 1.36 Arolia
Argotera posticalis 1 0 9.48
Calamotropha paludella 1 0 17.7
Glaucocharis unipunctalis 1 0 4.48
Glyphodes onycinalis 1 0 9.14
Glyphodes actorionalis 2 0 15.07 ± 1.73
Maruca vitrata 1 0 13.23
Paracymoriza fuscalis 9 0 8.37 ± 0.65
Paracymoriza prodigalis 12 (13) 0 7.64 ± 1.13
Paraloidea sp.2 1 0 5.54
Paraloidea sp.1 1 0 6.28
Piletocera aegimiusalis 21 0 9.28 ± 0.71
Geometridae
Cleora injectaria 7 1 20.12 ± 1.31 Arolia
Ptychopseustis pallidochrealis 1 0 7.7
Nymphalidae
Ideopsis similis 3 0 46.78 ± 1.39
Pyralidae
Tegulifera faviusalis 10 0 7.3 ± 0.76
Sphingidae
Macroglossum sp.1 2 0 20.88 ± 0.3
Macroglossum sp.2 1 0 27.43
Diptera Tupilidae Tupilidae 1 (5) 0 not measured
 J U N E 2017 , V O LU M E 104 • M O C H I Z U K I E T A L. — P O L L I N I A T R A N S F E R O N M OT H L E G S I N H OYA • 5

FIGURE 2 Pollination of Hoya carnosa. (A) Erebus ephesperis visiting and feeding for nectar on the inflorescence. The moth is inserting its proboscis into
the nectar pool. (B) The legs holding the edge of the corolla. Yellow pollinaria can be seen at the tip of the leg (white arrow). (C) The leg stepping on
the pollinaria (black arrow). (D) SEM image of the tip of the leg of E. ephesperis with pollinaria (bottom side). Two corpuscula are attaching to the aro-
lium of the leg. Pellucid margins of the pollinia are aligned in the same direction. a, arolium of the leg; ca, caudicle; cl, corpusculum; l, leg; pe, pellucid
margin; po, pollinium.
6 • A M E R I C A N J O U R N A L O F B OTA N Y

TABLE 2. The number of corpuscula attached to each leg. Values in parentheses are the number of pollinia remaining.

Right legs Left legs

Proportion of
Insect taxa Fore Middle Hind Fore Middle Hind Total pollinia missing
Erebus ephesperis 1(1) 1(1) 2(3) 2(1) 1(1) 2(3) 9(10) 0.44
E. ephesperis 2(0) 2(0) 0 3(0) 1(0) 0 8(0) 1.00
E. ephesperis 2(0) 2(0) 2(2) 2(0) — — 8(2) 1.00
E. ephesperis 3(5) 0 1(2) 2(2) 1(2) 0 7(11) 0.21
E. ephesperis 2(2) 2(2) — 2(3) 2(4) 1(0) 9(11) 0.39
E. ephesperis 1(1) 1(0) 2(3) 2(1) 2(0) 2(3) 10(8) 0.60
E. ephesperis 2(1) 2(1) 2(1) 3(3) 2(2) 3(4) 14(12) 0.57
E. ephesperis 2(4) 1(2) 1(2) — 2(3) 2(4) 8(15) 0.06
E. ephesperis 2(2) 0 3(3) 2(2) 1(2) — 8(9) 0.44
E. ephesperis 1(2) 2(4) 3(4) 0 1(0) — 7(10) 0.29
E. ephesperis 2(0) 1(0) 3(1) 2(3) 2(0) 1(0) 11(4) 0.82
E. ephesperis 0(0) 1(1) 1(1) 2(2) 1(2) 1(1) 6(7) 0.25
E. ephesperis 1(2) 1(1) 1(0) 2(2) — 2(4) 7(9) 0.36
Bradina erilitoides 1(2) 0 0 0 0 0 1(2) 0.00
Cleora injectaria 0 0 1(2) 0 0 2(3) 3(5) 0.17
Note: The dash indicates that legs were lost during capture.

one or more pollinaria removed. Of the120 pollinaria (240 pollinia)
originally produced by the 24 flowers, 28 (23%, 1.2 pollinia per
flower on average) were removed, but only four pollinia were de-
posited to the three flowers. The transfer efficiency was 0.14, which
is much smaller than the proportion of pollinia missing on pollina-
tor bodies.
Floral characters and the process of pollination—The moths vis-
ited the inflorescence after sunset (around 1900) and then searched
for nectar, which accumulates at the basal part of the anther skirt
(np in Figs. 1C, 2A). When foraging on the scattered nectar succes-
sively, medium or large moths struggle to find a foothold due to the
flat, velvety petals and the slippery staminal corona. Consequently,
medium or large moths hold on to either the edge of the flowers
(Fig. 2B) or the center part of the gynostegium (Fig. 2C), which has
a different texture from the surrounding staminal corona. When moths
try to place their legs and grip the gynostegium, they frequently step
on the pollinaria, led by the slippery staminal corona, and the pol-
linaria is removed as the corpusculum clips to the arolium of the moth’s
leg (Fig. 2D). Similarly, pollinia are deposited when moths with at-
tached pollinaria step on the guide rail, which is located immediately
below the corpusculum, and the pellucid margin of the pollinia is
trapped by the guide rail (pollinia itself is located outside Fig. 1F). After
the pollinia are successfully inserted to the guide rail, pollen tubes
start growing from the pellucid margin to the stigma (Fig. 1G).
DISCUSSION
Hoya carnosa was found to have a specialized pollination system in
which the pollinaria are attached only to the legs of medium or
large moths. The ways in which asclepiads and orchids attach pol-
linia to a pollinator’s body are remarkably diverse, and pollinator
legs are often used for pollinia transfer. In orchids, pollinia transfer
on legs is known in the genera Corycium, Disperis, Disa, and Habe-
naria (Johnson, 2005; Pauw, 2006; Peter et al., 2009). In asclepiads,
pollinia transfer on legs occurs in at least 12 genera pollinated
mostly by hymenopteran insects; however, in some genera, such as
Asclepias, pollinia are attached to the mouthparts in addition to
the legs (Frost, 1965; Morse, 1981; Forster, 1992; Betz et al., 1994;
Vieira and Shepherd, 1999; Ollerton et al., 2003; Shuttleworth and
Johnson, 2008, 2009; Wang et al., 2011; Nakahama et al., 2013).
Pollinia placement restricted to the legs is known only in four spe-
cies of asclepiads that have dipteran or hymenopteran pollinators:
Calotropis procera, Gomphocarpus physocarpus, Philibertia gilliesii,
and Sarcostemma viminale (Eisikowitch, 1986; Liede and White-
head, 1991; Coombs et al., 2009; Fernández et al., 2009; Cocucci et al.,
2014). To date, the only example of specialized pollinia transfer on
moth legs is in the South African orchid Habenaria epipactidea,
which attaches the pollinaria to the forelegs of the hawkmoth
(Peter et al., 2009). Our observations thus provide the first example
of an asclepiad species specialized to pollinia transfer on the legs
of moths.
The scarcity of pollinarium attachment to smaller moths and the
frequent attachment to large moths suggest pollination predomi-
nantly by the large moth Erebus ephesperis. To our knowledge,
E. ephesperis is the largest settling (nonsphingid) moth pollinator of
angiosperms ever reported. Three years of observation of a pollina-
tor of H. carnosa indicate that E. ephesperis was consistently the
most effective pollinator in the Amami-Oshima population. Erebus
ephesperis is widely distributed throughout Japan, Taiwan, and In-
dia, within the range of H. carnosa; thus, E. ephesperis is likely the
main pollinator of H. carnosa in other populations, although other
large moth species may also be pollinators in other parts of its
range. Why the pollinaria do not attach to smaller moths is unclear,
but the hinge of the corpusculum may need to be strongly stepped
on to pinch the arolium. Therefore, large-sized moths would be
more capable of removing the pollinaria. Because pollinaria attach-
ment was observed only on the arolium, and the large butterfly
I. similis, which lacks arolia, did not carry pollinaria, the absence/
presence of an arolium is also an important factor determining
pollinaria attachment.
The set of floral characters unique to the genus Hoya seems to
function effectively in pollinaria removal and pollinia insertion in
H. carnosa. In many asclepiads, nectar is located inside the guide
rails (but some genera such as Asclepias and Aspidonepsis secrete
nectar from the corona cups; Ollerton et al., 2003), which lead the
pollinator’s mouthparts to the reproductive organs (i.e., stigma and
pollinaria). However, in most Hoya plants, including H. carnosa,
the position of nectar secretion is transferred to the newly acquired,
 J U N E 2017 , V O LU M E 104
large secondary nectary beneath the anther skirt, which is a struc-
ture unique to Hoya (Wanntorp and Kunze, 2009). Such a separa-
tion of nectary and reproductive organs results in pollinaria not
attaching to the mouthparts. Indeed, hawkmoth species visited the
flowers and foraged for nectar, but they never carried pollinaria
(Table 1). On the other hand, nonsphingid moths settled on the
inflorescence and foraged scattered nectar, which led to leg place-
ment on the pollinaria and guide rails. Therefore, we hypothesize
that the shift in the nectar production site in the genus Hoya is an
adaptation to pollination involving insect legs. Additionally, the
shape of the inflorescence, providing little foothold, may also be an
adaptation to pollinia transfer by legs. This set of floral characters is
shared by many Hoya species, although the flowers of Hoya are di-
verse in terms of the color combination of the corolla and gynoste-
gium. The preserved floral structure and the diversity of floral
characters suggests that the basic pollination system involving the
use of pollinator legs is shared among species, while different pol-
linator taxa are involved in the pollination of different species of
Hoya (Wanntorp and Forster, 2007).
Before our study, the only information on Hoya pollinators was
provided by Forster (1992), who reported that the pollinaria of
H. australis attach to both the mouthparts and legs of its pollinator,
the diurnal skipper butterfly Ocybadistes walkeri sothis (although
the author suggested that nocturnal pollinators may also be in-
volved). The nectary of H. australis is considered to still be func-
tional (Wanntorp and Kunze, 2009); therefore, pollinaria may
attach to the mouthparts of the pollinators when pollinator insects
forage for nectar and to legs when they struggle on the slippery
flowers. Further examples of the pollination system of Hoya may
shed light on the relationship between nectary position and special-
ization toward pollinia transfer on pollinator legs. Such relation-
ships between nectar position and pollination have been noticed
in earlier observations in another asclepiad genus, Sarcostemma
(Kunze and Liede, 1991). Although not discussed, the pollinia
placement on pollinator legs is probably due to the dislocation of
the nectary to the corona lobes (Coombs et al., 2009). Therefore, a
shift in the nectar-secreting site may be strongly connected to the
evolution of pollination in many asclepiad genera. By contrast,
consistency in nectary position results in highly conservative polli-
naria placement as in Ceropegia and other Stapeliae (Meve and
Liede, 1994; Ollerton et al., 2009).
The discrepancy between the proportion of missing pollinia and
pollinia transfer efficiency suggests the frequent loss of pollinia
during transfer. Pollinaria transfer on leg tips is likely to be ineffi-
cient, because the risk of accidental loss of pollinaria may be much
higher compared with pollinaria transfer on mouthparts due to the
greater frequency with which legs come into contact with nonstig-
matic objects. The number of corpuscula attached to a single leg
was at most three in H. carnosa (Table 2), which is small when con-
sidering the large agglomerations of pollinaria that accumulate on
pollinator mouthparts in some asclepiads (Coombs et al., 2012).
Why pollination involving legs evolved is thus unknown, but the
lower capacity of pollinaria attachment on legs may be partly com-
pensated for by the six legs of insects. Unlike many other asclepiad
plants, in Hoya, pollinia should be trapped at the pellucid margin
by the narrow slit formed by the guide rails. Therefore, pollinia de-
position also seems to happen very rarely, but it may be compen-
sated for by pollinator behavior: they usually stay for several hours
and keep moving on an inflorescence. On the other hand, such be-
havior increases the risk of geitonogamy. Although not explicitly
• M O C H I Z U K I E T A L. — P O L L I N I A T R A N S F E R O N M OT H L E G S I N H OYA • 7
tested, the pollinia stick to the corpusculum soon after removal
but tend to be separated from the corpusculum more easily as they
dry; thus, pollinia deposition is more likely to occur with time
(K. Mochizuki, personal observation). Such reconfiguration of pol-
linia is suggested to reduce geitonogamous self-pollination (Harder
and Johnson, 2008). Further study is needed to understand how
plants overcome these disadvantages and to reveal the evolutionary
conditions that drove the specialization for pollinia transfer on pol-
linator legs.
ACKNOWLEDGEMENTS
We are grateful to the Amami Wildlife Conservation Center for
assistance during field research. The authors thank Dr. Nina
Sletvold, Dr. Jeff Ollerton, and an anonymous reviewer for many
helpful comments on our manuscript. This work was supported by
a Japan Society for the Promotion of Science grant to A.K. (grant
nos. 15H04421 and 24770018) and a Grant-in-Aid for JSPS Fellows
to K.M. (grant no. 15J01117).
LITERATURE CITED
Altenburger, R., and P. Matile. 1988. Circadian rhythmicity of fragrance emis-
sion in flowers of Hoya carnosa R. Br. Planta 174: 248–252.
Armbruster, W. S. 2014. Floral specialization and angiosperm diversity:
Phenotypic divergence, fitness trade-offs and realized pollination accuracy.
AoB Plants 6: plu003.
Armbruster, W. S., M. E. Edwards, and E. M. Debevec. 1994. Floral character
displacement generates assemblage structure of Western Australian trigger-
plants (Stylidium). Ecology 75: 315–329.
Armbruster, W. S., X. Q. Shi, and S. Q. Huang. 2014. Do specialized flowers
promote reproductive isolation? Realized pollination accuracy of three sym-
patric Pedicularis species. Annals of Botany 113: 331–340.
Betz, R. F., R. D. Struven, J. E. Wall, and F. B. Heitler. 1994. Insect pollinators of
12 milkweed (Asclepias) species. In R. G. Wickett, P. D. Lewis, A. Woodliffe,
and P. Pratt [eds.], Proceedings of the Thirteenth North American Prairie
Conference: Spirit of the land, our prairie legacy, 1992, 45–60. Corporation
of the City of Windsor, Windsor, Ontario, Canada.
Cocucci, A. A., S. Marino, M. A. Baranzelli, P. Wiemer, and A. Sérsic. 2014.
The buck in the milkweed: Evidence of male–male interference among pol-
linaria on pollinators. The New Phytologist 203: 280–286.
Coombs, G., C. I. Peter, and S. D. Johnson. 2009. A test for Allee effects in the
self-incompatible wasp-pollinated milkweed Gomphocarpus physocarpus.
Austral Ecology 34: 688–697.
Coombs, G., A. P. Dold, E. I. Brassine, and C. I. Peter. 2012. Large pollen loads
of a South African asclepiad do not interfere with the foraging behaviour or
efficiency of pollinating honey bees. Naturwissenschaften 99: 545–552.
Crepet, W. L., and K. J. Niklas. 2009. Darwin’s second “abominable mystery”:
Why are there so many angiosperm species? American Journal of Botany
96: 366–381.
Eisikowitch, D. 1986. Morpho-ecological aspects on the pollination of
Calotropis procera (Asclepiadaceae) in Israel. Plant Systematics and
Evolution 152: 185–194.
Faegri, K., and L. van der Pijl. 1979. The principles of pollination ecology.
Pergamon Press, Oxford, UK.
Fernández, V. A., L. Galetto, and J. Astegiano. 2009. Influence of flower func-
tionality and pollination system on the pollen size–pistil length relationship.
Organisms, Diversity & Evolution 9: 75–82.
Forster, P. I. 1992. Pollination of Hoya australis (Asclepiadaceae) by Ocybadistes
walkeri sothis (Lepidoptera: Hesperiidae). Australian Entomologist 19:
39–42.
Forster, P. I., and D. J. Liddle. 1992. Taxonomic studies of the genus Hoya R.Br.
(Asclepiadaceae) in Papuasia. Austrobaileya 3: 627–641.
8 • A M E R I C A N J O U R N A L O F B OTA N Y
Forster, P. I., D. J. Liddle, and I. M. Liddle. 1998. Diversity in the genus Hoya
(Asclepiadaceae—Marsdenieae). Aloe 35: 44–48.
Frost, S. W. 1965. Insects and pollinia. Ecology 46: 556–558.
Grant, V. 1949. Pollination systems as isolating mechanisms in angiosperms.
Evolution 3: 82–97.
Harder, L. D., and S. D. Johnson. 2008. Function and evolution of aggregated
pollen in angiosperms. International Journal of Plant Sciences 169: 59–78.
Johnson, S. D. 2005. Specialized pollination by spider-hunting wasps in the
African orchid Disa sankeyi. Plant Systematics and Evolution 251: 153–160.
Johnson, S. D., and M. Brown. 2004. Transfer of pollinaria on birds’ feet: A new
pollination system in orchids. Plant Systematics and Evolution 244: 181–188.
Kleijn, D., and R. Van Donkelaar. 2001. Notes on the taxonomy and ecology of
the genus Hoya (Asclepiadaceae) in Central Sulawesi. Blumea 46: 457–483.
Kunze, H., and S. Liede. 1991. Observations on pollination in Sarcostemma
(Asclepiadaceae). Plant Systematics and Evolution 178: 95–105.
Kunze, H., and L. Wanntorp. 2008. Corona and anther skirt in Hoya
(Apocynaceae, Marsdenieae). Plant Systematics and Evolution 271: 9–17.
Liede, S., and V. Whitehead. 1991. Studies in the pollination biology of
Sarcostemma viminale R.BR. sensu lato. South African Journal of Botany 57:
115–122.
Matile, P. 2006. Circadian rhythmicity of nectar secretion in Hoya carnosa.
Botanica Helvetica 116: 1–7.
Meve, U., and S. Liede. 1994. Floral biology and pollination in stapeliads—New
results and a literature review. Plant Systematics and Evolution 192: 99–116.
Morse, D. H. 1981. Modification of bumblebee foraging: The effect of milk-
weed pollinia. Ecology 62: 89–97.
Nakahama, N., R. Miura, and T. Tominaga. 2013. Preliminary observations
of insect visitation to flowers of Vincetoxicum pycnostelma (Apocynaceae:
Asclepiadoideae), an endangered species in Japan. Journal of Entomological
Science 48: 151–160.
Ollerton, J., S. D. Johnson, L. Cranmer, and S. Kellie. 2003. The pollination
ecology of an assemblage of grassland asclepiads in South Africa. Annals of
Botany 92: 807–834.
Ollerton, J., S. Masinde, U. Meve, M. Picker, and A. Whittington. 2009. Fly
pollination in Ceropegia (Apocynaceae: Asclepiadoideae): Biogeographic
and phylogenetic perspectives. Annals of Botany 103: 1501–1514.
Pauw, A. 2006. Floral syndromes accurately predict pollination by a special-
ized oil-collecting bee (Rediviva peringueyi, Melittidae) in a guild of South
African orchids (Coryciinae). American Journal of Botany 93: 917–926.
Peter, C. I., G. Coombs, D. F. Huchzermeyer, N. Venter, A. C. Winkler, D.
Hutton, L. A. Papier, A. P. Dold, and S. D. Johnson. 2009. Confirmation
of hawkmoth pollination in Habenaria epipactidea: Leg placement of pol-
linaria and crepuscular scent emission. South African Journal of Botany 75:
744–750.
Sage, T. L., and E. G. Williams. 1995. Structure, ultrastructure, and histochem-
istry of the pollen-tube pathway in the milkweed Asclepias exaltata L. Sexual
Plant Reproduction 8: 257–265.
Sargent, R. D. 2004. Floral symmetry affects speciation rates in angiosperms.
Proceedings of the Royal Society, B, Biological Sciences 271: 603–608.
Shuttleworth, A., and S. D. Johnson. 2008. Bimodal pollination by wasps and
beetles in the African milkweed Xysmalobium undulatum. Biotropica 40:
568–574.
Shuttleworth, A., and S. D. Johnson. 2009. Specialized pollination in the
African milkweed Xysmalobium orbiculare: A key role for floral scent in the
attraction of spider-hunting wasps. Plant Systematics and Evolution 280:
37–44.
Stebbins, G. L. 1970. Adaptive radiation of reproductive characteristics in an-
giosperms I: Pollination mechanisms. Annual Review of Ecology Evolution
and Systematics 1: 307–326.
Theiss, K., S. Kephart, and C. T. Ivey. 2007. Pollination effectiveness on co-
ocurring milkweed (Asclepias; Apocynaceae, Asclepiadoideae). Annals of
the Missouri Botanical Garden 94: 505–516.
Vieira, M. F., and G. J. Shepherd. 1999. Pollinators of Oxypetalum
(Asclepiadaceae) in southeastern Brazil. Revista Brasileira de Biologia 59:
693–704.
Wang, D. K., S. H. Zhai, B. Wang, and G. F. Sun. 2011. Floral structure and
pollination in relation to fruit set in Cynanchum otophyllum schneid. 2011
IEEE International Conference on Systems Biology 179–185.
Wanntorp, L. 2007. Pollinaria of Hoya (Marsdenieae, Apocynaceae)—
Shedding light on molecular phylogenetics. Taxon 56: 465–478.
Wanntorp, L., and P. I. Forster. 2007. Phylogenetic relationships between
Hoya and the monotypic genera Madangia, Absolmsia, and Micholitzia
(Apocynaceae, Marsdenieae): Insights from flower morphology. Annals of
the Missouri Botanical Garden 94: 36–55.
Wanntorp, L., M. Grudinski, P. I. Forster, A. N. Muellner-Riehl, and G. W.
Grimm. 2014. Wax plants (Hoya, Apocynaceae) evolution: Epiphytism
drives successful radiation. Taxon 63: 89–102.
Wanntorp, L., and H. Kunze. 2009. Identifying synapomorphies in the flowers
of Hoya and Dischidia —Toward phylogenetic understanding. International
Journal of Plant Sciences 170: 331–342.
Wiemer, A. P., A. N. Sérsic, S. Marino, A. O. Simões, and A. A. Cocucci. 2011.
Functional morphology and wasp pollination of two South American ascle-
piads (Asclepiadoideae-Apocynaceae). Annals of Botany 109: 77–93.
Willmer, P. 2011. Pollination and floral ecology. Princeton University Press,
Princeton, New Jersey, USA.
Yamazaki, T. 1993. Asclepiadaceae. In K. Iwatsuki, T. Yamazaki, D. E. Boufford,
and H. Ohba [eds.], Flora of Japan, vol. 3a, 168–183. Kodansha Scientific,
Tokyo, Japan.