2010 SOUTHEASTERN NATURALIST 9(3):413–426
Visitors to Southeastern Hawkmoth Flowers
Sean P. Graham*
Abstract - Despite global concern for the status of animal pollinators, studies on
pollination systems in the southeastern United States are disproportionately low
compared to the diversity of this region. For example, sphingophilous, or hawkmoth-
attracting plants, occur in the southeastern US, but confirmation is lacking
for the large, long-tongued hawkmoths predicted to visit these flowers by previous
researchers. Hymenocallis coronaria (Shoals Spider Lily, or Cahaba Lily), H. occidentalis
(Woodland Spider Lily), Oenothera biennis (Common Evening Primrose),
and O. grandiflora (Large-flowered Evening Primrose) were studied to confirm this
prediction. Manduca rustica (Rustic Sphinx) was confirmed as a frequent visitor to
all four plant species studied. M. sexta (Carolina Sphinx) was confirmed for three of
the four plants. To determine the range of animal visitors to these plants, three of the
plant species were observed during day- and night-observation periods, and total
visitation was compared between these times. For H. coronaria, flower-visitation
rates did not differ between day and night periods. H. occidentalis and O. biennis
were visited significantly more during night hours than during the day. Although
hawkmoths are frequent visitors to H. occidentalis and O. biennis, and are probably
their most efficient pollinators, Archilochus colubris (Ruby-throated Hummingbird)
may also play a role in the pollination of H. coronaria.
Introduction
Recent concern for the loss or decline of animal pollinators and this
phenomenon’s possible impact on plant populations, environmental integrity,
and human welfare, make descriptive studies of pollination biology
necessary (Cane and Tepedino 2001, Kremen and Ricketts 2000, National
Research Council 2007, Watanabe 1994). Even faunal lists of visitors and
pollinators have the potential to advance our understanding of this globally
important mutualism by providing baseline information for more theoretical
or long-term studies (National Research Council 2007). Unfortunately,
especially considering the exceptional biodiversity of this species-rich
region of North America (Odum 2002), studies on southeastern US species
are few (for exceptions see Fenster and Dudash 2001, Irwin and Adler
2006, Motten 1986).
Studies of pollination biology lead to the recognition of consistent
trends among floral traits, including those referred to as pollination syndromes.
Pollination syndromes are suites of floral characteristics which
are apparently adapted to attract specific pollinators, and are classically
thought to represent cases of recurring co-evolution between plants and
pollinators with specific behavioral and morphological proclivities (Proctor
et al. 1996). The pollination-syndrome concept has been criticized
*Auburn University Department of Biological Sciences, 331 Funchess Hall, Auburn
University, AL, 36849; grahasp@auburn.edu.
414 Southeastern Naturalist Vol. 9, No. 3
based on the repeated finding that many plants with apparently specialized
morphologies corresponding to syndromes are actually generalized in their
attractiveness to diverse visitors (Waser et al. 1996). However, whether
due to specialized or generalized attraction patterns, pollination syndromes
do outline predictions and are therefore testable hypotheses. A famous
example of one of these hypotheses is Darwin’s (1862) prediction that the
spectacularly long-spurred Angraecum sesquipedale Thouares (Madagascar
Orchid) co-evolved with an equally incredible long-tongued hawkmoth
(Kritsky 1991). Modern studies revealed several species of long-tongued
hawkmoths pollinate the long-spurred orchid guild in Madagascar (Nilsson
et al. 1987, Wasserthal 1997).
A review of North American hawkmoth flowers listed plants in the southwestern
and southeastern US with characteristics consistent with hawkmoth
pollination, listed known moth visitors and pollinators for numerous species,
and provided a similar, and currently untested prediction (Grant 1983b).
These sphingophilous, or hawkmoth-attracting, plants share a suite of characters
which define a well-known pollination syndrome. This syndrome is
characterized by 1) pale-colored flowers with nocturnal anthesis; 2) heavy,
sweet, nocturnal perfume; and 3) long nectar tubes or spurs (Grant 1983b,
Proctor et al. 1996). Sphingophily is common in the tropics and southwestern
US, and interactions between Manduca moths and Datura species from
this area are well studied (Grant 1983a, b; Raguso and Willis 2005; Raguso
et al. 2003). However, Grant (1983b) concluded this review by highlighting
our lack of information of species in the southeastern US which conform
to this syndrome. Several plant genera of the Southeast, including Hymenocallis
(Amaryllidaceae), Zephyranthes (Amaryllidaceae), Oenothera
(Onagraceae), and Crinum (Amaryllidaceae), have night-blooming, large,
pale flowers with long nectar tubes (>4 cm) and heavy nocturnal perfume.
Thus, he predicted that several long-tongued hawkmoth species native to
the Southeast (e.g., Manduca sexta L. [Carolina Sphinx], Manduca rustica
Fabricius [Rustic Sphinx], Agrius cingulatus Fabricius [Pink-spotted Hawkmoth],
and Cocytius antaeus Drury [Giant Sphinx]) visit these species and
transfer their pollen (Grant 1983b).
Here, I test Grant’s (1983b) hypothesis that southeastern hawkmoth
plants are visited by these predicted species of long-tongued hawkmoths.
Hymenocallis coronaria (Le Conte) Kunth (Shoals Spider Lily, or Cahaba
Lily), H. occidentalis (Le Conte) Kunth (Woodland Spider Lily), Oenothera
biennis L. (Common Evening Primrose), and O. grandiflora L'Hér.
(Large-flowered Evening Primrose) were studied because they are among
those genera predicted to conform to the hawkmoth plant syndrome, and
are locally abundant, facilitating flower-visitor observations. One of
these species, H. coronaria, is also a source for conservation concern, and
therefore flower-visitation information is important from a management
perspective. Finally, I compared the range of diurnal and nocturnal visitors
to three of the plant species to estimate the relative importance of hawkmoths
as potential pollinators.
2010 S.P. Graham 415
Methods
Study Areas
Hymenocallis coronaria visitor observations were conducted at Yellowjacket
Shoals (32.87521°N, 84.410031°W) and Hightower Shoals
(32.810056°N, 84.401047°W) on the Flint River in west-central Georgia.
These shoals contain hundreds of H. coronaria individuals, and during peak
bloom (see below), thousands of flowers can be observed. Observations of
H. occidentalis took place in the floodplain forest of Choctafaula Creek in
Tuskegee National Forest, Macon County, AL (32.490048°N, 85.603969°W).
Oenothera biennis were observed along County Road 53 in Macon County,
AL (32.514118°N, 85.610238°W), which is <2 km from the Choctafaula
Creek site. Oenothera grandiflora were observed at three localities: along
Florida State Road 85, Okaloosa County, fl(30.742161°N; 86.564364°W),
along Byrne Lake Landing Road, Baldwin County, AL (30.794004°N,
87.891685°W), and along Alabama State Road 225, Baldwin County, AL
(30.893665°N, 87.855374°W). Both Oenothera species are found sympatrically
in south Alabama, and O. biennis and O. grandiflora are syntopic at the
Byrne Lake Landing Road site.
Studied plants
Hymenocallis coronaria is one of 15 recognized spider lily species in
the southeastern US (Smith and Garland 2003), and is unique in its habitat
preference for large lotic streams (Fig. 1a; Davenport 1996). The flowers of
H. coronaria have a large corona (a membranous staminal cup; Smith and
Garland 2003), with projecting strap-like perianth elements, long (4–6 cm)
stamens, and a 6–8 cm pistil. The corona has a long nectar tube (mean =
5.51 cm, range = 3–7.5 cm, n = 10). It blooms from mid-May through early
June. Individual plants have six to ten flowers, and new flowers open in
mid- to late afternoon (Patrick et al. 1995, Davenport 1996). These wither
the next morning (Patrick et al. 1995) and continue to produce scent intermittently
throughout blooming (S.P. Graham, pers. observ.). A previous
study documented only two visitors to this species (Battus philenor L.
Figure 1. Growth habit and habitat of the plant species observed in this study. a) Hymenocallis
coronaria, Yellowjacket Shoals on the Flint River, Talbot County, GA. b)
H. occidentalis, Tuskegee National Forest, Macon County, AL. c) Oenothera biennis,
Macon County, AL. Inset: close-up of O. biennis at night. d) O. grandiflora, Okaloosa
County, fl. Scale bars = approximately 4 cm.
416 Southeastern Naturalist Vol. 9, No. 3
[Pipevine Swallowtail] and Paratrea plebeja Fabricius [Plebian Sphinx]),
and concluded that animal pollination and clonal propagation (apomixis)
are probably equally important to this species’ breeding system since animal
visitation is rare (Davenport 1996).
Hymenocallis occidentalis has similar flower morphology (Smith and
Garland 2003; Fig. 1b); however, this species occurs in much lower population
densities at a given site, and each plant can be separated from its
nearest neighbor by as much as 50–100 m. It prefers sites in floodplains
or wet woods (Davenport 1996, Smith and Garland 2003). It has a more
widespread geographic distribution throughout southeastern floodplain forests
than H. coronaria (Smith and Garland 2003). Anthesis was observed
for three flowers on two different plants, each occurring approximately 1 h
before sunset. One to three new flowers open each night, resulting in one to
six flowers per plant. These flowers can remain open for at least 48 h before
withering (S.P. Graham, pers. observ.). The nectar tube is longer than that
of H. coronaria (mean = 8.6 cm; range = 4.9–11.2 cm; n = 9). This species
blooms in late July to early August at the Alabama study area, and flowers
remain open throughout the day and produce a heavy, pleasant odor that
can be detected from up to 10 m away from a single plant. Scent could also
be detected during daylight hours (S.P. Graham, pers. observ.). No data on
animal visitors to this plant are currently available.
The genus Oenothera contains the highest proportion of sphingophilous
species in North America (Grant 1983b, 1985). Oenothera biennis is the
most widespread and common member of its genus in the eastern US (Cleland
1972). The tall (2 m) plants live in highly disturbed habitats and in
August–October, produce four to 25 new flowers each night (Fig. 1c), which
start blooming approximately 1.5 h after dark and continue opening one at a
time every four to ten minutes. These flowers are closed by 0800 h the next
morning. The 1–2 cm long petals are pale yellow, and the stamens and pistil
protrude only one cm from the corolla. The hypanthium tube averages 4.17
cm long (range = 3.9–4.6 cm, n = 11). Ten hymenopterans, one moth, and
one dipteran have been recorded as visitors to this species’ flowers at a New
Jersey study site (Dickerson and Weiss 1920).
Oenothera grandiflora lives in similar habitats and blooms during the
same time period as O. biennis, but is more locally distributed in the southeastern
United States. Earlier authors considered it endemic to the Mobile
Bay area (Schumacher and Steiner 1993), but it is now found in surrounding
areas as well (Steiner and Stubbe 1984). It has much larger flowers than
O. biennis (up to 10 cm in diameter) and often has a stigma that protrudes
approximately 1.5 cm further than the stamens (Fig. 1d). The hypanthium
tube is longer than in O. biennis, averaging 5.72 cm (range = 5.4–6.3 cm,
n = 10). The flowers of O. grandiflora are closed by mid-morning; however,
the exact timing of their closing was not observed. A similarly sized population
of O. grandiflora seems to produce a much heavier scent than O. biennis
(S.P. Graham, pers. observ.). This species outcrosses extensively, although
genetic analyses have demonstrated that self-pollination and hybridization
with local members of the O. biennis complex is taking place within its
2010 S.P. Graham 417
historical range (Schumacher and Steiner 1993). Currently, no information
regarding visitors to this species is available.
Hymenocallis coronaria, H. occidentalis, and Oenothera grandiflora
were specifically predicted to be sphingophilous in Grant’s review of North
American hawkmoth flowers (1983b). Oenothera biennis was omitted from
the review, probably because it is considered to be primarily self-pollinated
(Cleland 1972).
Observations and collections
Observations of H. coronaria were conducted to determine the range
of visitors during 2006–2009 for a total of 28 h (day and night). Each observation
ranged from one to several hours. Daytime observations were
conducted using binoculars and video camera at a distance of 20 m from
one to three patches of approximately five to 20 H. coronaria plants (≈50–
250 flowers). Night surveys in 2006–2008 were conducted using headlamps,
and a red light was used during 2009. Night surveys were conducted
by standing immediately alongside a patch of spider lilies, scanning the
patch slowly with the head lamp with red light filter. Visitors were defined
as species observed to enter the flower and feed or attempt to feed from the
corona tube’s opening.
For H. occidentalis, 50 observation hours were conducted at individual
plants (day and night) in July–August 2009. Observations took place with
the observer standing 10 m away from the plant, or by viewing video footage
taped remotely using night-shot function. Observations of Oenothera
biennis were conducted from August–September 2009 for a total of 45 h.
These plants were initially checked at all times of the day, night, and morning
to determine flowering phenology at this locality. Cameras were placed
for night and morning observations (flowers are completely closed during
the afternoon). Due to the distance of natural populations of Oenothera
grandiflora from my base of operations, this species was observed for five
person-hours on three nights in September 2009. Comparisons between day
and night visitation were not conducted for this plant.
Attempts were made during observation hours to collect and identify at
least one individual of each insect species observed to visit the above flowers
using a butterfly net, and photographs were taken of subsequent visitors
captured of the same species. Care was taken not to disturb the plants while
capturing visitors. Animal visitors were placed into the following categories
to facilitate analysis: birds, bumblebees, other hymenopterans (e.g., wasps
or hornets), flies, butterflies, hawkmoths (family Sphingidae), and other
moths. Contacts with stamens and/or stigma were noted for all categories
of visitors and many species, and additional notes were also taken (e.g.,
direction of visitor travel). For certain visitor species, attempts were made
to determine individual flower visitation rates by recording the number of
seconds spent feeding at individual flowers. Pollen was noted when grossly
apparent on moths or moth organs; however, no attempt was made to quantify
or identify pollen. Moths were identified by consulting Hodges (1971)
418 Southeastern Naturalist Vol. 9, No. 3
and Covell (1984). All insects collected were prepared and deposited in the
Auburn University entomology collection.
Data analysis
Mean number of visitors to plant patches (H. coronaria) or individual
plants (H. occidentalis, O. biennis) per hour per observation period for all
visitor categories were compared between day observations (dawn to dusk)
and night observations (dusk and night) using an independent-samples t-test
(i.e., the same plants were not observed on consecutive nights in most cases).
Data were normalized by increasing each mean visitation rate/observation
period by a factor of one, followed by log transformation . However, to assist
in interpretation, untransformed means are presented in the figures. Mean
number of visits per hr/observation period for each visitor category were
also calculated for each plant species, but these data were not compared statistically.
Statistical analyses were performed using SPSS, with α = 0.05.
Results
Flower Visitors
Visitor rates for the plant species observed during night and day periods
are summarized in Table 1. For Hymenocallis coronaria, visitation rates
did not differ significantly between day (from dawn up to dusk) and night
(dusk and night) observations (t1, 15 = 0.599; P > 0.05; Fig. 2a). Daytime
visitors were small Bombus sp. (bumblebees), Archilochus colubris L.
(Ruby-throated Hummingbird), and various unidentified butterflies. Most
butterflies visited too briefly to be identified or collected. Hawkmoths were
common nighttime visitors; other moth species visited occasionally. Signifi-
cantly more nighttime visits occurred at H. occidentalis (t1, 14 = 3.857, P =
0.002; Fig. 2b). One fly attempted to feed on the nectar of H. occidentalis
on one occasion during the day, whereas hawkmoths and other moths were
common visitors to this species at night. Oenothera biennis was also visited
significantly more often at night (t1,17 = 5.808, P < 0.0001; Fig. 2c). Small
Table 1. Summary of diurnal vs. nocturnal visitation patterns for visitors to the studied plants.
Total number of visits recorded indicated, and mean number of visits/hour/observation period
indicated in parentheses. For H. coronaria, mean number of visits to plant patches are presented;
for H. occidentalis and O. biennis, means are for individual plants. Day = dawn until
dusk; Night = dusk until dawn.
Hymenocallis Hymenocallis Oenothera
coronaria occidentalis biennis
Visitor categories Day Night Day Night Day Night
Bumblebees 36 (3.43) 3 (0.18) 0 0 4 (0.28) 0
Other Hymeopeterans 2 (0.19) 0 0 0 0 0
Fly 0 0 1 (0.07) 0 0 0
Ruby-throated Hummingbird 22 (2.1) 2 (0.11) 0 0 1 (0.07) 0
Butterflies 5 (0.48) 1 (0.05) 0 0 1 (0.07) 0
Moths 0 21 (1.24) 0 11 (0.37) 0 5 (0.16)
Hawkmoth 2 (0.19) 79 (4.65) 0 22 (0.74) 4 (0.29) 132 (4.24)
2010 S.P. Graham 419
hawkmoths (possibly Dolba hyloeus Drury [Pawpaw Sphinx]) were day
(dawn) visitors to O. biennis, and three bumblebees, one hummingbird, and
one butterfly were also observed visiting this plant during the day. Hawkmoths
were abundant visitors to O. biennis at night, while other moths were
occasional visitors.
Bumblebees, Ruby-throated Hummingbirds (Fig. 3a–b), hawkmoths
(Fig. 3c–d), and other moths (Fig. 3e) were observed to contact the anthers
and stigma of H. coronaria, and hawkmoths and other moths frequently
contacted the anthers and stigma of H. occidentalis, O. biennis, and
O. grandiflora (Table 2). However, the rapid wing beats and large wingspan
and body of Ruby-throated Hummingbirds and hawkmoths resulted in more
frequent contact with these floral organs compared to other visitors.
There appeared to be interesting trends in the timing of visitation and
behavior of the various hawkmoth species (Table 1). The diurnal Hemaris
thysbe Fabricius (Hummingbird Clearwing) was observed in the evening
and also during midday visiting H. coronaria. Dolba hyloeus visited both
H. coronaria and H. occidentalis at dusk, consistent with previous observations
that this moth is a dusk flier (Hodges 1971). Paratrea plebeja
and Manduca rustica visits began ≈1 h after dark at H. coronaria patches
and continued as late as 0100 h the following morning. Manduca rustica
usually visited H. occidentalis 1–2 h after dark, and no visits were
recorded by any moths after this time at these plants. Large hawkmoths
(M. rustica and M. sexta) arrived at O. biennis as soon as the first flowers
opened (≈2100 h), and P. plebeja begin visiting shortly thereafter. Paratrea
plebeja and M. sexta were also captured visiting O. biennis shortly
before sunrise, suggesting that visitation occurs intermittently throughout
the night in this species. Hawkmoths tended to visit H. coronaria plants
from downstream to upstream (n ≥ 20 observations).
Two of the large, long-tongued hawkmoth species predicted by Grant
(1983b) as possible visitors to southeastern hawkmoth flowers were confirmed as visitors to the study plants (Table 2). Manduca rustica were
captured twice visiting H. coronaria, twice visiting H. occidentalis, once
visiting O. biennis, and once visiting O. grandiflora; large hawkmoths with
morphology consistent with this species and presumed to be M. rustica were
also observed numerous additional times visiting each of these species.
Figure 2. Mean visitation rates between night (dusk through dawn; black bars) and
day (dawn up to dusk; white bars) for all visitors to Hymenocallis coronaria patches
(a), Hymenocallis occidentalis plants (b), and Oenothera biennis plants (c). n = number
of observation periods.
420 Southeastern Naturalist Vol. 9, No. 3
Manduca sexta was captured twice visiting O. biennis and once visiting O.
grandiflora, and a hawkmoth with morphology consistent with this species
was videotaped visiting H. coronaria (Fig. 3c).
Pollination mechanisms
The far-exerted stigma and widely-spaced anthers of Hymenocallis
contacted the larger visitors (hummingbirds and hawkmoths) on the wings
and abdomen as they oriented into the flower to feed (Fig. 3b–c; Table 2).
Hawkmoths often hovered at the flowers while feeding, and during this time,
their wings thoroughly agitated the stamens and stigma. Occasionally these
insects simply landed at the flowers to feed. Pollen was not grossly apparent
Figure 3. Visitors to southeastern hawkmoth flowers. a) Ruby-throated Hummingbird
hovering over H. coronaria. b) Ruby-throated Hummingbird feeding from H. coronaria.
c) Large hawkmoth (probably M. sexta) visiting H. coronaria. d) Paratrea
plebeja visiting Oenothera biennis. e) Unidentified moth visiting H. coronaria.
f) Plusiodonta compressipalpis (Moonseed Moth) robbing nectar from H. occidentalis.
Images a, b, c, and e recorded at Hightower Shoals, Talbot County, GA, and d
and f taken in Macon County, AL. Scale bars for a–d are approximately 4 cm; scale
bars for e–f are approximately 2.5 cm.
2010 S.P. Graham 421
Table 2. Summary of animal visitors to southeastern hawkmoth flowers, presented in ascending order of presumed pollination effectiveness; large species which
contacted stamens and stigma most often, visited most often, and visited each flower longest are presumed the most efficient pollinators.* indicates pollinator
species predicted by Grant (1983b). H.c. = Hymenocallis coronaria, H.o. = Hymenocallis occidentalis, O.b. = Oenothera biennis, and O.g. = Oenothera grandiflora.Day = dawn until dusk; Dusk = period between sunset and total darkness; Night = total darkness until dawn.
Stigma Anther
Common name Latin name Day Dusk Night H.c. H.o. O.b. O.g. contact contact Mean visit time per flower
Rustic Sphinx* Manduca rustica x x x x x x x 2 sec.; n = 4 plant visits
Carolina Sphinx* M. sexta x x x x x x
Banded Sphinx Eumorpha fasciatus x x x x
Pandorus Sphinx E. pandorus x x x x
Ruby-throated Hummingbird Archilochus colubris x x x x x x 2.39 sec.; n = 4 patch visits
Plebian Sphinx Paratrea plebeja x x x x x x 1.18 sec.; n = 1 patch visit
Pawpaw Sphinx Dolba hyloeus x x x x x
Hummingbird Clearwing Hemaris thysbe x x x x x 2.42 sec.; n = 2 patch visits
Banded Tussock MothA Halysidota tessalaris x x
Bilobed Looper MothA Autographa biloba x x
Common Looper MothA A. precationis x x
Tobacco Budworm MothA Heliothis virescens x x x
Sharp-stigma Looper MothA Agrapha oxygramma x x x
BumblebeesB Bombus spp. x x x x
Butterflies x x x
Gulf Fritillary Agraulis vanillae x x
Other Hymenopterans x x
Flies x x
AMoths occasionally observed to contact stigma and anther.
BBumblebees rarely observed to contact stigma and anther.
422 Southeastern Naturalist Vol. 9, No. 3
on the body of most visitors to either Hymenocallis species. Oenothera
biennis’ stamens and pistil are equally exerted from the corolla, protrude
only about one cm, and are less widely spaced than in Hymenocallis. The
proboscides of Manduca rustica, M. sexta, and P. plebeja were sufficiently
long for individuals to hover and feed from flowers without contacting the
stamens or stigma with the wings, head, or body in many cases. However, the
sticky pollen threads of newly-opened flowers adhered to the visiting moth’s
proboscis, and individuals of these species were observed moving between
plants with large accumulations of pollen attached to this organ (see photos
in Gregory 1963–1964). Hawkmoths contacted the anthers and stigma of
O. grandiflora with their proboscis, head, abdomen, wings, and legs, and
pollen accumulations were noted on these areas on the insects.
Other interactions
A few moths were observed to rob nectar from H. coronaria by probing
their proboscis between the perianth elements from underneath the corona.
Plusiodonta compressipalpis (Noctuidae) Guenée (Moonseed Moth) were
observed robbing nectar from H. occidentalis on several occasions by inserting
their proboscis into holes in the corolla tube near its base (Fig. 3f).
Several other flowers were found with the characteristic bore holes used
by this moth. Oenothera grandiflora’s stamens and pistil are exerted further
than in O. biennis, and prevented small noctuid moths from contacting
them; no stigma contact was observed from moths other than hawkmoths in
this species. However, small moths were frequent visitors to the nectar tube
opening, where they probably rob nectar.
Discussion
In this study, I confirmed Grant’s prediction that southeastern plant genera
which exhibit characteristics consistent with the hawkmoth pollination
syndrome are primarily visited by sphingid moths. One of Grant’s predicted
hawkmoth species (Manduca rustica; Fig. 6) was observed to frequently
visit Hymenocallis coronaria, H. occidentalis, Oenothera biennis, and
O. grandiflora. Another predicted species, M. sexta, was confirmed as a
visitor to O. biennis and O. grandiflora, and H. coronaria (Fig. 3c). Unfortunately,
although videotaping plants was extremely useful for documenting
visitation rates of different visitor categories, positive species identification
was impossible for many recorded visits (e.g., color and morphology useful
in identification was not discernable). I am confident that I have not determined
the total number of visiting hawkmoth species to these plants.
Several other animal groups with a corresponding pollination syndrome
were also observed to frequently visit H. coronaria. Although it could be
argued that they are unable to transfer the pollen of these plants, at least
one (the Ruby-throated Hummingbird) is probably an important pollinator
of this species. There was no difference in day vs. nighttime visitation in
H. coronaria, and hummingbirds frequently fed at this species and probably
transfer pollen between individual plants and possibly even different populations.
Hummingbirds visited during the evening, and since the flowers open
2010 S.P. Graham 423
before dark, it is possible that they often transfer pollen before moths do.
This observation supports the view that pollination syndromes are often less
predictive due to the generalized pollination patterns of most plant species
(e.g., Waser et al. 1996).
For H. occidentalis and O. biennis, hawkmoths visited most frequently
and contacted the stigma and anthers of these species far more frequently than
any other visitor. This result is probably also true for O. grandiflora, whose
flowers also close during the day. Despite the potential for alternative visitors
in one of the hawkmoth plants studied, the syndrome is highly predictive
for the species taken together. Many of the same hawkmoth species visited
these plants despite large differences in blooming period, geography, habitat,
density, and phylogeny (Table 1). This pattern is similar to findings from the
western US, where a guild of hawkmoths visit plants from diverse phylogenetic
ancestry and convergent flower morphology (e.g., Grant 1983b). Additional
experiments are needed to determine the relative importance of these visitors
to southeastern hawkmoth plants, and to confirm whether any of them actually
transfer pollen between individuals of these plants.
The transfer of pollen by large hawkmoths (Manduca rustica, M. sexta,
and Paratrea plebeja) and hummingbirds might explain published patterns
of gene flow in H. coronaria. Markwith and Scanlon (2007) hypothesized
downstream gene flow in H. coronaria, predicting stream current generally
carries seeds and genetic material downstream. However, they found no
evidence of this, and could not exclude zoochory or animal pollination as a
possible explanation for their results, despite lack of evidence for frequent
animal visitation. This paper provides evidence for frequent animal visitation
in Hymenocallis and the potential for frequent pollen transfer among plants.
Most hawkmoths visited patches of H. coronaria moving upstream, probably
following scent trails in air currents carried downstream. This direction
of pollen transfer possibly counteracts downstream movement of gene flow
by hydrochory. The migrations of either hawkmoths or hummingbirds are
capable of long-range pollen transfer (Raguso and Willis 2003, Williamson
2001). It is likely that hawkmoths—particularly M. rustica—are responsible
for movement of genetic material among shoals or even among drainages,
since hummingbirds have largely completed migration when H. coronaria is
blooming (Williamson 2001).
The different visitor types between the two Hymenocallis species are
likely linked to their habitat preferences and density. H. coronaria can attain
very high population densities and cover an entire shoal of a large river. In
this respect, they are essentially a meadow of flowers that can elicit visitation
from large numbers of generalist visitors that are feeding among other
plants in the area. H. occidentalis are less numerous, and individual plants
are scattered and are therefore only visited by more specialized visitors. It is
also interesting to note that the longer nectar tube length of H. occidentalis
may exclude visitation from short-tongued bees and lepidopterans. However,
this feature does not explain the lack of visitation by hummingbirds in
this species, which were present in the area when they bloom. Interestingly,
H. coronaria is one of the only North American Hymenocallis with a yellow
424 Southeastern Naturalist Vol. 9, No. 3
spot on the staminal cup (Smith and Garland 2003), which may be a derived
feature to attract visually oriented diurnal feeders. Perhaps these and other
features (e.g., shorter nectar tube, earlier diel bloom time) became adaptive
during the evolutionary shift to the more open shoal habitat of H. coronaria.
Clearly, southeastern Hymenocallis offer an excellent opportunity for comparative
research in pollination biology.
Oenothera biennis exhibits many features consistent with hawkmoth pollination,
and is visited frequently by hawkmoths, yet Grant (1983b) omitted
this plant from his list of sphingophilous North American species. This omission
was presumably due to genetic work that determined this species as being
primarily self-pollinating and producing clonal seeds (Cleland 1972, Gregory
1964). If this is indeed the case, the results of this study are surprising and suggest
that a plant with a highly specialized, self-pollination breeding system has
dispensed with the benefits of outcrossing via frequent hawkmoth visitation.
Despite exhibiting derived features associated with self-pollination (smaller
flowers, equal stamen and style lengths, clonal seed lines), the flowers of
O. biennis still produce nectar, perfume, and viscid pollen threads similar to
their sphingophilous relatives which encourage hawkmoth visitation (S.P.
Graham, pers. observ.). This discrepancy begs one of two possible explanations:
either O. biennis achieves more cross-pollination than previously
assumed (e.g., Cleland 1972), or that in this area, clonal plants are robbed by
insect visitors which do the plant no pollination service yet take the rewards.
The abundant visitation this plant receives suggests that the plant receives
some benefit, as do many other facultative self-pollinators (Proctor et al.
1996). Cleland (1972) suggested the possibility that O. biennis occasionally
outcrosses, and that the rampant hybridization between true-breeding clonal
lines in this complex is probably brought about by pollinators.
Oenothera grandiflora has many apparently ancestral features consistent
with cross-pollination (larger flowers, longer hypanthium tube, and larger,
unequally exerted stamens and pistil), and is also visited frequently by
hawkmoths. Visits from these insects probably provide the hybridization
mechanism between O. biennis and O. grandiflora reported by researchers
interested in the genetics of this group (Schumacher and Steiner 1993). Both
plants are common near the Baldwin County, AL study sites, and the same
species of hawkmoths visit both plants. The O. biennis complex and O. grandiflora provide yet another comparative opportunity for pollination studies,
and future research on these plants would benefit from an extensive amount
of genetic information already available (e.g., Cleland 1972, Schumacher
and Steiner 1993, Steiner and Stubbe 1984).
Further studies on these and other sphingophilous species in the southeastern
US (e.g ., Crinum, Zephyranthes, and Ipomoea spp.) are encouraged
and will likely uncover similar fascinating interactions. The results of this
study provide evidence that hawkmoths are probably important pollinators
for plants in the southeast which exhibit the hawkmoth plant syndrome, although
other visiting species are possibly important in some cases. Similar
studies are recommended by naturalists interested in our southeastern flora,
and attempts should be made to determine pollination patterns of additional
2010 S.P. Graham 425
understudied species, which will simultaneously provide information about
the plants and their visitors.
Acknowledgments
I thank K. Gray and R. Birkhead for assistance with observations and collecting,
and S. Hermann, C. Guyer, R. Birkhead, and the Guyer lab for discussions and
suggestions that much improved this research and earlier drafts of this manuscript.
K. Barrett provided statistical suggestions. I also thank N. Burkett-Cadena and S.
Collins for their assistance with insect preparation and curation, and C. Hansen for
his assistance in viewing herbarium specimens. I thank C. Guyer for allowing me the
freedom to pursue this project, the editorial board of Southeastern Naturalist, Robert
Raguso, and two anonymous reviewers for their suggestions which considerably
improved this document. S.P. Graham is supported by NIH grant # R01-A149724 to
T. Unnasch.
Literature Cited
Cane, J.H., and V.J. Tepedino. 2001. Causes and extent of declines among native
North American invertebrate pollinators: Detection, evidence, and consequences.
Conservation Ecology 5:1–7.
Cleland, R.E. 1972. Oenothera: Cytogenetics and Evolution. Academic Press, London,
UK and New York, NY. 370 pp.
Covell, C.V. 1984. Moths of Eastern North America. Virginia Museum of Natural
History Special Publication Number 12. Martinsville, VA. 496 pp.
Darwin, C. 1862. On the various contrivances by which British and foreign orchids
are fertilized by insects. John Murray, London, UK.
Davenport, L.J. 1996. The Cahaba Lily: Its distribution and status in Alabama. Journal
of the Alabama Academy of Science 67:222–233.
Dickerson, E.L., and H.B. Weiss. 1920. The insects of the evening primroses in New
Jersey. Journal of the New York Entomological Society 28:32–74.
Fenster, C.B., and M.R. Dudash. 2001. Spatiotemporal variation in the role of hummingbirds
as pollinators of Silene virginica. Ecology 82:844–851.
Grant, V. 1983a. Behavior of hawkmoths on flowers of Datura meteloides. Botanical
Gazette 144:280–284.
Grant, V. 1983b. The systematic and geographical distribution of hawkmoth flowers
in the temperate North American flora. Botanical Gazette 144:439–449.
Grant, V. 1985. Additional observations on temperate North American hawkmoth
flowers. Botanical Gazette 146:517–520.
Gregory, D.P. 1963–1964. Hawkmoth pollination in the genus Oenothera. Aliso 5:
357–419
Hodges, R.W. 1971. The Moths of North America North of Mexico: Fascicle 21,
Sphinogoidea. E.W. Classey Limited and R.B.D. Publications, Inc., London,
UK. 158 pp.
Irwin, R.E., and L.S. Adler. 2006. Correlations among traits associated with herbivore
resistance and pollination: Implications for pollination and nectar robbing
in a distylous plant. Ecology 93:64–72.
Kremen, C., and T. Ricketts. 2000. Global perspectives on pollination disruptions.
Conservation Biology 14:1226–1228.
Kritsky, G. 1991. Darwin’s Madagascan hawkmoth prediction. American Entomologist.
37:206–210.
426 Southeastern Naturalist Vol. 9, No. 3
Markwith, S.H., and K.C. Parker. 2007. Conservation of Hymenocallis coronaria
genetic diversity in the presence of disturbance and a disjunct distribution. Conservation
Genetics 8:949–936.
Markwith, S.H., and M.J. Scanlon. 2007. Multiscale analysis of Hymenocallis coronaria
(Amaryllidaceae) genetic diversity, genetic structure, and gene movement
under the influence of unidirectional stream flow. American Journal of Botany
94:151–160.
Motten, A.F. 1986. Pollination ecology of the spring wildflower community of a
temperate deciduous forest. Ecological Monographs 56:21–42.
National Research Council. 2007. Status of Pollinators in North America. The National
Academies Press, Washington, DC. 307 pp.
Nilsson, L., A.L. Johnsson, L. Ralison, and E. Randrianjohany. 1987. Angraecoid
orchids and hawkmoths in Central Madagascar: Specialized pollination systems
and generalist foragers. Biotropica 19:310–318.
Odum, E. 2002. The southeastern region: A biodiversity haven for naturalists and
ecologists. Southeastern Naturalist 1:1–2.
Patrick, T.S., J.R. Allison, and G.A. Krakow. 1995. Protected Plants of Georgia.
Georgia Department of Natural Resources, Social Circle, GA.
Proctor, M. P. Yeo, and A. Lack. 1996. The Natural History of Pollination. Timber
Press, Portland, OR. 479 pp.
Raguso, R.A., and M.A. Willis. 2003. Hawkmoth pollination in Arizona’s Sonoran
Desert: Behavioral responses to floral traits. Pp. 43–65, In C.L. Boggs, W.B.
Watts, and P.R. Ehrlich (Eds). Butterflies: Ecology and Evolution Taking Flight.
University of Chicago Press, Chicago, IL. 756 pp.
Raguso, R.A., and M.A. Willis. 2005. Synergy between visual and olfactory cues
in nectar feeding by wild hawkmoths, Manduca sexta. Animal Behaviour
69:407–418.
Raguso, R.A., C. Henzel, S.L. Buchmann, and G.P. Nabhan. 2003. Trumpet flowers
of the Sonoran Desert: Floral biology of Peniocereus cacti and sacred Datura.
International Journal of Plant Science 164:877–892.
Schumacher, E., and E. Steiner. 1993. Cytological analysis of complex-heterozygotes
in populations of Oenothera grandiflora (Onagraceae) in Alabama. Plant
Systematics and Evolution 184:77–87.
Smith, G.L., and M.A. Garland. 2003. Nomenclature of Hymenocallis taxa (Amaryllidaceae)
in the southeastern United States. Taxon 52:805–817.
Steiner, E., and W. Stubbe. 1984. A contribution to the population biology of Oenothera
grandiflora L’Her. American Journal of Botany 71:1293–1301.
Waser, N.M., L. Chittka, M.V. Price, N.M. Williams, and J. Ollerton. 1996. Generalization
in pollination systems, and why it matters. Ecology 77:1043–1060.
Watanabe, M.E. 1994. Pollination worries rise as honey bees decline. Science
265:1170.
Williamson, S.L. 2001. A Field Guide to Hummingbirds of North America. Houghton
Mifflin Harcourt, New York, NY. 275 pp.