|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
REPRODUCTIVE AND DISPERSAL ECOLOGY OF THE INVASIVE CORAL
ARDISIA (Ardisia crenata) IN NORTHERN FLORIDA
MICHAEL J. MEISENBURG
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
2007 Michael J. Meisenburg
This document is dedicated to my sister Marie Gedeon (1951-2006).
To finish this-and to believe in myself enough to do so-I am grateful to my wife
Vasiliki. Without her I would not have journeyed down this path. Most of us need
someone in our lives to coach us on, to prod us along, and to pick us up when we
stumble. She is that person in my life. I will never be all that I can be, or maximize my
potential, but that is okay. My life is richer with her in it, and because of her I will
accomplish more. It is because of her that I am about to finish this journey.
Despite much encouragement from my advisor and my committee, my friends and
my family, and (most of all) my wife, this project took seven years to finish. In the end, I
suppose that it was the impending early retirement of my advisor that finally forced me to
complete what has become my personal 800-lb gorilla. My gorilla has accompanied me
everywhere these last few years, and has been with me day and night. My gorilla taught
me how to incorporate stress into my life, how to exist after nights with little sleep, how
to maintain a 40-lb weight gain, how to abandon hobbies that I once enjoyed, how to
despise writing, and how to avoid tasks and not reach my goals. Soon I will turn in my
completed thesis, and I will lose my hairy companion. I will not miss him.
A special thank you is due to Alison Fox (my advisor) and Randall Stocker
(Alison's husband and one of my committee members) for getting me through this. I
hope they enjoy their travels.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
L IST O F TA B L E S ............ ........ .... ............................... .... ...... ... ............. vii
LIST OF FIGURES ............... .. ................................ viii
A B ST R A C T ................. ................................................................................... ........
1 UNDERSTANDING THE IMPORTANCE OF SEED DISPERSAL FOR
N ONN ATIVE PLAN T SPECIES ........................................... ........................... 1
Introduction ......... ......... .............................. ...............1
Coral A rdisia in Florida .......... ..... ................................... ...... .. ................ .2
Importance and M echanisms of Seed Dispersal..................................... .......... ....... 4
A D iffu se M utualism ........................................................................... .. ..
The Flesh of the Fruit ........................................ .................. ............ .... ..6 6
D ispersal Process by B irds ........................................ ......................................7
Seed Ingestion ................................................................7
Seed D position ............................................. 7
S eed V iab ility ....................................................... 8
Observations of Bird Behavior .................................................. 9
Project G oals .............................................................................................
2 REPRODUCTIVE PHENOLOGY OF Ardisia crenata ........... ......................11
Introduction ........................... .................. 11
M e th o d s ..................................................................12
Site Descriptions ................ ....... .......... ........ 12
Data Collection ....................................................................... ......... ................... 13
D ata A n a ly sis .................................................................................................. 1 5
Results ....................... ...................................16
Flowering and Fruiting Phenology ............ ....................... ......... 16
Fruit Loss ........................ .......................... 16
D isc u ssio n ............. ......... .. .............. .. .......................................................1 7
F ruiting P henology ............................................................17
Fruit Loss ........................ ..... ..................... 17
Longetivity of fruit persistence ................. ..............................................17
Influence of fruit/plant presentation on loss rate...................... ...............18
Seasonality of fruit loss ....................................................... ..... .......... 20
3 NUTRITIONAL CONTENT OF Ardisia crenata FRUITS .................................. 28
In tro d u ctio n ............. ...... ....... ................................. ................ 2 8
M eth o d s ..............................................................................2 9
R results ............. ...... ... .. ............ ... ............................................30
D discussion ............... ........................................................................... 31
4 CONSUMPTION OF ARDISIA CRENATA FRUIT BY BIRDS AND THEIR
ROLE IN SUBSEQUENT SEED DISPERSAL .......................................................38
Introduction..................................... ......................................... ... 38
M eth o d s ............................................................................. 3 9
Field Observations .......... ... ........................ ........ .. .. .. .... .... ... 39
Selection of Birds for Experim ents ................................................... ................. 39
Captive Feeding Trials .............................................. ............. .............. 39
Selection of Native Plants for Experiments........................ .... ...............41
D ata A n a ly sis ................................................................................................. 4 3
Results .......... .................................. ...............43
Field O observations ................. ........... .. .... ........ .. .................43
F e e d in g T ria ls ................................................................................................ 4 4
Discussion ............... ......... .........................45
Field Observations .............. .... ......... .... ... .........45
F e e d in g T ria ls ................................................................................................ 4 9
LIST OF REFERENCES .................................................................... .... ......... ......... ......... 63
BIOGRAPHICAL SKETCH ........... .... ....... .... ................................. ........... 68
LIST OF TABLES
2-1 Rainfall levels for Gainesville, FL (units are inches). It is not known how many
years data collection constitute averages (Gainesville Sun, June 30, 2002)............27
3-1 Data for coral ardisia fruits (means and standard deviations). Percentages of
nutrient data are on a dry matter basis......................................... ...............35
3-2 Nutritional data for fruits available in northern Florida natural areas, grouped
together by season. Fruits from coral ardisia are generally available winter through
summer; values of fall fruits were included as a reference. Data for all species
other than coral ardisia are from W hite (1989). ............................. ...................36
3-3 The means and standard deviations of the fruit nutrient data from Table 3-2.
Although the some coral ardisia fruits remain on the plant for the entire year, most
competition for frugivores occurs winter through summer............... ..................37
4-1 Population statuses of the birds tested. Data are from Stevenson and Anderson
(1994). .............................................................................. 6 1
4-2 Transformed slope values from feeding trials. Significance levels indicate that two
bird species did not significantly choose one fruit species over the other, while the
remaining four bird species preferred control fruits over test (A. crenata) fruits....61
4-3 Ardisia seed germination rates from seeds voided during feeding trials. Control
fruits were those that were manually depulped. ................... ...............................62
LIST OF FIGURES
2-1 Flowering dates of Ardisia crenata........................................................... 23
2-2 Ardisia crenata fruit production dates for three populations. Values on the Y-axes
represent the % of monitored plants at that stage..................................................24
2-3 Fruit losses from Ardisia crenata plants at two sites during 2001. Error bars
represent standard errors. ............................................... .............................. 25
2-4 The loss rates ofArdisia crenata fruits at Putz during 2001. The three lines
represent loss rates for three densities ofA. crenata plants. Error bars show 95%
confidence intervals......... ................................................................ .... .... .... 25
2-5 Probability density function graphs for Ardisia crenata fruit loss rates at three
different densities of plants: minimal (top), moderate (middle), and maximum
(bottom ) ................................................................. ...... ........... 26
4-1 Results from European Starling feeding trials with fruits from Ardisia crenata and
Cinnamomum camphora. The birds showed no taste preference for either fruit
species. Bars represent standard error. ........................................ ............... 55
4-2 Results from gray catbird feeding trials with fruits from Ardisia crenata and
Forestiera godfreyi. The birds showed no taste preference for either fruit species.
B ars represent standard error....................................................................... ....... 56
4-3 Results from Northern Mockingbird feeding trials with fruits from Ardisia crenata
and Vaccinium corymbosum. The birds showed a taste preference for V.
corymbosum. Bars represent standard error... ................... ......................... 57
4-4 Results from American Robin feeding trials with fruits from Ardisia crenata and
Ilex opaca. The birds showed a taste preference for I. opaca. Bars represent
stan d ard erro r ................... ................... ......... ................ ................ 5 8
4-5 Results from Cedar Waxwing feeding trials with fruits from Ardisia crenata and
Ilex x attenuata. The birds showed a taste preference for I. x attenuata. Bars
represent standard errors. ............................................... .............................. 59
4-6 Results from Fish Crow feeding trials with fruits from Ardisia crenata and
Vaccinium corymbosum. The birds showed a taste preference for V. corymbosum.
Bars represent standard error............... ................. ................... .......... 60
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
REPRODUCTIVE AND DISPERSAL ECOLOGY OF THE INVASIVE CORAL
ARDISIA (Ardisia crenata) IN NORTHERN FLORIDA
Michael J. Meisenburg
Chair: Alison M. Fox
Coral ardisia was introduced into the New World from Asia more than 100 years
ago and has become a part of Florida landscapes. It has since been recognized as being
an invasive weed even though it is suspected that consumption of its fruits by possible
seed dispersers is an uncommon event. The purpose of this study was to gain insight into
the reproductive ecology of coral ardisia in its introduced range of northern Florida.
Tracking tagged plants showed flowered through the summer and fruits ripened in
mid-winter. The fruits can persist on the plants for up to one year, with the greatest rate
of fruit loss occurring in late April. Plants in a natural area lost fruits at a faster rate than
those in an urban area, and isolated plants lost fruits quicker than those in denser
populations. The period of greatest fruit loss coincides with the spring migration of birds
through Florida and suggests that some consumption of fruits by birds is occurring.
Fruits were found not to be nutritionally inferior to native fruits, nor were any other
factors found that would suggest a reason why fruits of coral ardisia are rarely eaten.
Despite many hours in the field, observations of birds feeding on coral ardisia fruits were
limited to a single day with a single bird species (gray catbird). Captive feeding trials
found that six species of birds would eat coral ardisia fruits, but often favored the fruits of
native species to those of coral ardisia. Gray Catbirds and European Starlings showed the
greatest acceptance of coral ardisia fruits during captive feeding trials. Cedar Waxwings
defecated seeds after eating fruits, while all other species regurgitated seeds. Seed
germination rates were no different for 5 of 6 birds species tested between seeds
defecated or regurgitated compared to those that were manually depulped. From damage
incurred while manually trying to remove seeds from fruits, Fish Crows significantly
decreased seed germination rates. However, Fish Crows ate the fewest coral ardisia
UNDERSTANDING THE IMPORTANCE OF SEED DISPERSAL FOR NONNATIVE
The human-assisted spread of species into new regions and the subsequent
ecological effects that some of these species have on native ecosystems is recognized as
one of the most serious contemporary threats to biodiversity. Indeed, Wilcove et al.
(1998) regarded introduced species as being second only to habitat loss in terms of
habitat impacts. Whether by accident or intention, humans have been moving organisms
for thousands of years, but the recognition of the negative impacts to natural systems as
well as the magnitude of these impacts are more recent (for a good description see Mack
et al. 2000).
Although many taxa are represented, plants constitute some of the better-known
examples, and within the state of Florida alone, millions of dollars are spent annually to
control nonnative plants. For example, the Florida Department of Environmental
Protection (FDEP) spent 6.3 million dollars in FY2003 to control nonnative weeds. As
large as this expenditure is, it includes neither aquatic plants, with two of the worst
introduced species, water hyacinth (Eichhornia crassipes) and hydrilla (Hydrilla
verticillata), nor money allocated directly to the Melaleuca Program (FDEP 2004). One
plant responsible for much of this expenditure in northern Florida is coral ardisia, Ardisia
crenata Sims. (Myrsinaceae) (FDEP 2004).
Coral Ardisia in Florida
Native to southeastern Asia, coral ardisia grows from one to one and a half meters
tall in shaded to partially-shaded areas. Initially single-stemmed, most plants eventually
produce additional stems and maintain a multi-stemmed status for many years. After
several years, stems begin growing short branches. With leaves growing at the ends, the
initial function of the branches is energy production. After one to two years, a branch's
purpose changes to reproduction as the leaves are replaced by flowers in a cyme. During
the branch's final year, the flowers produce single-seeded drupes. Branches typically
have about 5 to 20 fruits each and plants have from 1 to 10 branches. Branches fall from
plants after the fruits are gone, and the lifespan of each branch is 3 to 4 years. The
branches circle the stem, with the lowest branches being the oldest. Each year, the plant
grows taller and produces a new set of branches, and each ring of branches advances to
the next stage of its cycle. Plants may begin producing branches when about 20 cm tall.
The pattern of continuously regenerating stems makes it impossible to estimate the age of
the plant based on stem size and reproductive status.
Coral ardisia is a shrub that has been used for landscaping in Florida for more than
100 years (Royal Palm Nurseries 1900). At least three factors may have contributed to
the popularity of this species. First, it grows well in sites with no direct sunlight, a
condition many landscape plants cannot tolerate. Second, it produces large, bright red
fruits just prior to Christmas that persist for months and contrast with its evergreen
glossy, dark green leaves. Third, the plant seems to rarely suffer insect damage. By 1982
coral ardisia was recognized as escaping cultivation and invading moist woods
In response to the plant's invasion, the Florida Exotic Pest Plant Council has placed
the plant on its Category I list of invasive plants (www.fleppc.org); a designation
indicating that the plant is not only invading natural lands, but altering native plant
communities. In 2003, the FDEP's Upland Weeds Program ranked Ardisia spp. (which
includes both coral ardisia and shoebutton ardisia (Ardisia elliptica)) as the seventh-most
herbicide-treated taxon in the state (FDEP 2004).
While coral ardisia grows well in moist mesicc) sites, it also invades wet hydricc)
woods. Mesic hardwood hammock seems to be the natural community most prone to
invasion (Langeland and Burks 1998). The plant is not known to invade mesic or hydric
pinelands (Dozier 1999).
Coral ardisia fruits seem to be removed from plants only occasionally (Dozier
1999), thus presenting a paradox in attempting to discover dispersal agents. If fruits are
not often eaten, so that the seeds can be subsequently dispersed, then how has it spread
throughout natural areas in the state? Because the plant is still used in landscaping, some
dispersal is anthropogenic, but this does not account for the plant's widespread presence
in natural areas (e.g. remote locations in San Felasco Hammock State Park, Gainesville,
FL, Sam Cole, park biologist, personal communication). Birds are a likely candidate for
non-human dispersers of this plant because coral ardisia does not posses mammalian-
dispersed fruit traits, such as being sweet-tasting and falling to the ground soon after
ripening (Stiles 1980). In addition, while coral ardisia's fruits are brightly-colored and
highly visible in the forest due to this color, most mammals are color blind (Van der Pijl
1972). Fruits of a native congener (A. escallonoides) similarly seem not to be heavily
used by birds, except for during spring migration (John Pascarella, University of Miami,
The Importance and Mechanisms of Seed Dispersal
There are several reasons why seed dispersal is important to plants. First, because
sites suitable for growth vary in space and time, plants must be able to colonize new areas
as conditions change (Howe and Smallwood 1982). Second, dispersal beyond the canopy
is a means of avoiding competition between parents and siblings. Finally, the Escape, or
Janzen-Connell Hypothesis (Janzen 1970, Connell 1971), indicates that dispersal from
the parent is important because seed predators (rodents, insects, or microbes) often
concentrate their efforts under parent plants where food density is highest.
Plants use various strategies for dispersing seeds beyond the range of their
branches, and while the process is an important part of plant population dynamics, it is
especially important if the plants are invasive, exotic species. Certain traits may
influence a particular species' propensity to become invasive, but a species is much less
likely to become a problem (beyond a local scale) without a reliable seed dispersal
vector. Little is known about vertebrate-assisted seed dispersal of invasive plant species
in Florida, and few dispersers have been adequately verified. Identification of dispersal
agents relies heavily on assumptions (e.g. Morton 1982, Cronk and Fuller 1995), and
there often seems to be a disregard of important subtleties (e.g. not distinguishing
between those that eat fruit pulp and those that eat fruit seeds).
Many plants achieve seed dispersal through relatively simple processes such as via
wind or water, but more complicated interactions can occur when plants use vertebrates
to disperse their seeds. Ectozoochoric fruits or seeds attach to animals with hooks, barbs,
or sticky secretions, while animals are usually enticed to eat endozoochoric seeds with a
fleshy fruit meal. Successful seed dispersal occurs only when still viable seed is
dispersed. While seed dispersal may be an important initial step in a plant's life, it is
only the initial step as germination and seedling establishment are also required.
Although fruit ingestion and subsequent seed dispersal of coral ardisia do not seem
to be common events (Dozier 1999), even low frequency events could be important if
plant mortality of seedlings was low. A trait shared by many nonnative plants is a lack of
pathogens and predators, and indeed, insect damage on coral ardisia is rarely observed
(personal observation). In other words, a limited frequency of seed dispersal may not
negatively affect coral ardisia as much as it would a native species that has coevolved
with a suite of pathogens, parasites, and predators, because coral ardisia may suffer lower
rates of mortality predation at the seed and seedling stages.
A Diffuse Mutualism
Tight relationships between a specific fruit and frugivore are unusual, and not
known to occur for invasive plants. Indeed, some non-native plant species have their
seeds dispersed by bird species that did not coevolve with them. If they needed a specific
disperser, introduced plants probably would not become a problem when introduced
beyond their natural range. This type of seed dispersal represents a mutualism because
both participants benefit; the frugivore with a meal and the plant with its seeds dispersed.
However, the mutualism is only fulfilled if a viable seed is moved beyond the range of
the plant's branches.
Following fruit consumption, seeds may be carried away from or dropped from the
parent plant, with the latter resulting in no dispersal. For those seeds that are transported
away from the parent, ingestion may increase (Renne et al. 2001, Bartuszevige and
Gorchov 2006, Figueroa and Castro 2002), decrease (Bartuszevige and Gorchov 2006,
Meyer and Witmer 1998), or have no effect on seed germination rates (Meyer and
Witmer 1998, Izhaki and Safriel 1990, Figueroa and Castro 2002). Seeds dropped under
the parent plant and ingestion decreasing seed germination rates are non-mutualistic
situations where the frugivore benefits but the plant does not.
The Flesh of the Fruit
Endozoochoric fruits typically consist of a digestible outer layer surrounding (at
least) one seed, and in the majority of cases this is a fleshy pericarp consisting of pulp
and skin. Alternatively, arillate fruits possess endozoochoric seeds in which the fruits
open and reveal seeds that are covered in a digestible coating. Arillate seeds may have a
fleshy aril, such as southern magnolia (Magnolia grandiflora), or a dry, waxy aril like
Chinese tallow (Sapium sebiferum).
The nutrient content of fruits is usually assessed by measuring the levels of lipids,
carbohydrates, and protein (Stiles 1980). Summer/early fall fruits tend to be higher in
carbohydrates and water, while fall/winter fruits generally contain higher levels of lipids.
Protein levels in fruits are usually low. Mammals feed more on summer and early fall
fruits, which are often sweet-tasting, while the lipid-rich fruits of fall and winter are
mostly utilized by birds (Stiles 1980).
Most species of temperate fruiting plants in the eastern U.S. set fruit in the fall,
presumably benefiting from migrating birds (Stiles 1980). These birds need energy to
fuel migration, and fruits provide a readily digested source in packages that are easy to
procure. However, fruit set in Florida's natural communities may follow a different
schedule because this geographic region is subject to the selective pressure of a large
over-wintering bird population rather than the passage of fall migrants (Skeate 1987).
While much of the state has a temperate flora and hence most plant species produce fruit
in the fall, it is thought that a greater fruit biomass is produced in the winter months when
birds such as American Robins (Turdus migratorius), Cedar Waxwings (Bombycilla
cedrorum), Gray Catbirds (Dumetella carolinensis), and yellow-rumped warblers
(Dendroica coronata) over-winter in Florida (Skeate 1987).
The Dispersal Process by Birds
For efficient flight it is essential that birds minimize unnecessary weight. One
strategy to accomplish this is for birds to eliminate heavy, undigestible seeds as quickly
as possible. Frugivorous birds can be divided into two groups: gulpers and mashers
(Moermond and Denslow 1985). Gulpers are species that tend to swallow fruits whole,
separate the seeds from the pulp internally, and then generally void the seeds at some
distance from the parent plant (e.g., Northern Mockingbird (Mimuspolyglottos)).
Mashers tend to crush fruits in their bills separating the seeds from the pulp and
swallowing just the pulp (e.g., northern cardinals (Cardinalis cardinalis)). Mashers
typically drop seeds from the canopy of the parent plant. Generally, birds with heavier
conical bills are mashers while those with thinner bills are gulpers.
How a gulper rids itself of seeds is also important relative to plant dispersal. The
seed may be separated from the pulp in the crop with the seed regurgitated, or separation
can occur further along the digestive tract with the seed defecated. The crop is an
enlargement of the esophagus and while it is generally used to store food prior to entering
the gizzard or stomach, it is also used for separating the pulp from the seeds. Birds can
regurgitate seeds that are cleaned of even the most adhering pulp (such as those of coral
ardisia). Murray et al. (1994) found that there is a positive correlation between the
distance over which a seed is dispersed and the length of time that a bird carries the seed.
Since regurgitation is faster than defecation, method of voidance might affect seed
dispersal distance. Meyer and Witmer (1998) found the mean seed defecation time of the
native shrub Viburnum dentatum after fruit consumption by American Robins was 58
minutes versus a mean regurgitation time of 19 minutes. Although these authors studied
both voidance methods for one type of seed using a single bird species, more often a bird
species either regurgitates or defecates seeds based on seed size.
Recognizing that many factors can influence which bird species feed on which
fruits (e.g., time of year ripening occurs, proximity to ground), it can be hypothesized that
as a result of different voidance methods, plant population expansion rates could be
influenced by which bird species tend to feed on the fruits.
The time of the year that ripe fruits are on the plant could influence the direction of
seed dispersal. For example, a plant species whose fruits ripen in the spring may be most
likely to experience a gradual northward population shift due to the spring migration of
millions of birds. It should be remembered that each bird carries seeds a small distance at
a time (depending upon flight speed and duration of seed retention), not for the thousands
of miles of the whole migration route.
A mutualism between bird and plant only exists if a viable seed is dispersed. Seed
viability may be affected by the digestion process, and the severity of damage may
increase with the length of time a seed is retained in the bird's digestive tract (Murray et
al. 1994). Thus, there may be a trade-off for the plant between the distance seeds are
carried prior to defecation (potentially improving seed dispersal), and the proportion of
seed remaining viable (reducing viable seed dispersal).
In addition to frugivorous birds feeding on fruit pulp, granivorous birds may feed
on the seeds of fruits. The granivorous house finch (Carpodacus mexicanus) is
historically a western U.S. species that expanded its range into Florida following a 1940's
introduction into the northeastern U.S. While they are commonly observed feeding on
fruits, often they are actually cracking the seeds and feeding on the entire fruit (skin,
pulp, and seeds). Similarly, with a gizzard that is capable of crushing pecans and acorns,
wild turkeys (Meleagris gallopavo) and other members of the order Galliformes often
digest the seeds passed through their digestive tracts. An observation of these species
feeding on fruits can easily be misinterpreted as seed dispersal rather than the seed
predation that it is.
Observations of Bird Behavior
A conclusive determination of endozoochoric seed dispersal by birds requires
verification that the seeds are ingested, carried away from the parent plant, and voided in
a viable condition. Observation of only fruit or seed consumption does not distinguish
between seed dispersal and seed predation (Meisenburg and Fox 2002). However,
documentation of seedlings distant from the rest of the plant population and in sites
frequented by birds (e.g., under tree roosts, along fence lines) is an indication that bird
dispersal is likely (McDonnell and Stiles 1983).
The goals of this study were to gain a greater understanding of coral ardisia
reproductive phenology and to determine whether some bird species might have a role in
dispersing viable seeds.
The objectives of chapter two were to determining when flowers, unripe, and ripe
fruits first appear on plants, and then to determine their duration. Assessing fruit nutrient
content to compare to native species was the objective of chapter three. Finally, the
objectives of chapter four were to determine if birds would eat the fruit of coral ardisia,
conduct preference trials with other fruits, assess germination rates of voided seeds, and
report on bird feeding activity in coral ardisia stands.
REPRODUCTIVE PHENOLOGY OF ARDISIA CRENA TA
The fruits of coral ardisia have been observed to persist for most of the year on
plants in northern-central Florida (Dozier 1999). For this reason it is often assumed that
the fruits are rarely eaten by frugivores in Florida. Although it is possible that
consumption rates are low, an alternative hypothesis is that fruit production continues for
much of the year and thus gives the appearance that fruits are not removed.
If consumption of fruits is occurring, seasonal variation in fruit loss rate could
implicate certain species as being major consumers of the fruit. For example, Florida is a
corridor to millions of migrating birds every spring and fall, and if these events were
correlated to increased rates of fruit loss from plants, this would suggest consumption by
migrants. If viable seeds are voided, high rates of fruit consumption could lead to high
rates of seed dispersal.
Fruit loss is the detachment of fruit from the plant, and this may be passive (e.g.
senescence of peduncle) or active (i.e. removal by an animal). For the purposes of this
study, mechanisms of fruit loss were not differentiated but the assumption is made that
active removal may have accounted for some of the observed fruit loss.
With regard to fruit loss, several factors were taken into account because they could
influence fruit loss rates. One such factor is habitat (Denslow 1987, Gosper et al 2006),
probably because different bird species (with their specific food preferences and
nutritional needs) occur in different habitats. Study sites were chosen in two habitats
with significant coral ardisia populations: intentionally-planted landscapes and near old
homesites in natural areas with persisting coral ardisia populations.
It is possible that fruit height could influence loss rates. If this was the case, then
certain types of dispersal agents may be implicated. For instance, some fruit-eating
mammals are limited to foraging below a certain height relative to their size (though
woody, coral ardisia stems could only support small mammals). The mammals that I
considered to be potential consumers of ardisia fruit were all mid-sized (mesomammals):
red fox (Vulpes vulpes), gray fox (Urocyon cinereoargenteus), raccoon (Procyon lotor),
and Virginia opossum (Didelphis virginiana), and the assumption was made that they
were unlikely to feed on fruits above 0.6 meters from the ground.
Another factor in fruit loss rate is plant density. Denslow (1987) found that
aggregated red elderberry (Sambucuspubens) plants had lower individual fruit removal
rates than did isolated plants, and she hypothesized that competition among plants for
frugivores led to decreased fruit removal rates for clumped plants. While recording data
in the initial stages of this study, it appeared that the more isolated an ardisia plant was,
the more quickly it lost its fruit. Consequently, I monitored plants for fruits loss across
densities from relatively isolated to growing within dense stands.
The objectives of Chapter 2 were to determine the dates of flower and fruit
production, and the duration of ripe fruit persistence on the plants as influenced by
habitat, branch height, and plant density.
Three sites were selected in Alachua County, Florida. The first was near Rocky
Point Road, Gainesville, on Paynes Prairie State Preserve property (hereafter called
PPSP), the second was east ofNewnan's Lake (on private property owned by University
of Florida Botany Department professor Francis J. Putz) (Putz), and the last site included
several residences around southwestern Gainesville, Florida (City). Sites were selected
on the basis of extensive ardisia infestations and permission to access the property.
The PPSP site was located in a hardwood forest that was principally upland with
seasonal wetlands. Upland trees included live oak (Quercus virginiana), laurel oak
(Quercus hemisphaerica), Southern magnolia (Magnolia grandiflora), American holly
(Ilex opaca), and coral ardisia. Wetlands were primarily red maple (Acer rubrum) and
black gum (Nyssa sylvatica).
The Putz site had experienced disturbance from logging and turpentine production,
and consequently had a successional forest component. It contained live oak, laurel oak,
sweet gum (Liquidambar styraciflua), loblolly pine (Pinus taeda), and American holly,
and bordered a bayhead that contained sweetbay magnolia (Magnolia virginiana),
loblolly bay (Gordonia lasianthus), and black gum. Coral ardisia was found throughout
this site, but was most extensive in the successional forest.
The City site was comprised of ardisia plants in landscaped locations, as well as
plants in undeveloped wooded lots adjacent to the residences. The plants at the
residences were intentional plantings while the plants in the wooded lot appeared to be
free-living and self sustaining populations, probably originating from the intentionally-
Field data were collected to determine reproductive phenology and rate of fruit
loss. Phenological data consisted of recording when flowers, green (unripe) fruits, and
red (ripe) fruits appeared on plants. On each sampling date, each individual plant was
categorized according to the most advanced of the three stages of the flower/fruit cycle
found. Following fruit maturation, periodic fruit counts were done to assess when plants
lost their fruits. There was no distinction between those fruits that were removed by
vertebrates and those that abscised naturally from the parent.
Phenologic data were gathered at the Putz and City sites during 2000 and 2001, and
at PPSP during 2000. Fruit loss was followed at the Putz and City sites during 2001.
Fruit loss was not followed at the PPSP site because a freeze in early January 2001
severely damaged previously tagged fruits and plants.
I monitored plants at the Putz site along on a 100-meter transect that encompassed
several environmental variables (canopy cover, wetland proximity, and plant height and
density). I stratified the transect into 10-meter sections because it was not practical to
monitor all plants on the transect. Within each 10 x 1 meter section I used computer-
generated random numbers to select plants for monitoring purposes. On each selected
plant I chose up to three branches for periodic fruit counts.
Monitored plants in the City site were spread among four residences and an
undeveloped wooded lot. The four residences were separated by a minimum of 100
meters. One residence had plants tagged in both the front and back yards, but no more
than 15 plants were monitored in any one yard.
Sixty plants were initially monitored for fruit loss at the City site, and 100 in the
Putz site. Plants that died were excluded from analysis, and the final number of plants
consisted of 54 at the City site and 91 at the Putz site. Some plant deaths occurred as a
result of trees falling, mowing, whereas others were due to unknown causes.
Two variables were recorded at Putz to determine if they influenced fruit loss rates:
branch height and the density of ardisia surrounding the plant being monitored. Branches
were grouped into two groups: those above and those below 60 cm. I believe that this
value was a reasonable maximum height for mesomammals to reach if eating the fruits.
Plants were reported in three categories of density. "Minimal density" was 5 or
fewer other ardisia plants within a 1-meter radius, "moderate density" was 6-10 plants
within a 1-meter radius, and "maximum density" was >10 plants within a 1-meter radius.
Generally, minimally-dense plants also had few other herbaceous or shrub species
present, and bare soil was exposed on at least half of the 2-meter diameter circle.
Moderately-dense plants usually contained laurel oak seedlings in the circles, and had
little, if any, exposed soil. Ardisia plants in the densest circles usually had no other
herbaceous or shrub species, and contained a considerable layer of forest duff.
All fruit loss data were converted from actual numbers to percentages (as a change
from initial counts) and then plotted as fruit loss over time. I used the PROCMIX
statement in SAS statistical software (SAS Institute, version 8.2). I used logit
transformation on the data transformation because the error variance was not constant,
and logit was the appropriate transformation because the variance was dependent on the
value of Y. Transformation produced linear equations where the slopes and Y-intercepts
were compared for significance.
Another statistical program, MATLAB (MathWorks Inc., version 6), was used to
produce probability density function graphs that displayed the average rate of fruit loss at
any given time for the three densities tested. These graphs are the derivatives of the fruit
Flowering and Fruiting Phenology
Ardisia flowered June through September, and peaked in late July-August (Figure
2-1). There were not sufficient replications for statistical testing, but monitoring multiple
sites for multiple years indicated a trend for later flowering in 2000. In both years, a few
City plants flowered earlier than any other plants.
Similar patterns existed for fruit production, as fruits appeared later in 2000 than in
2001, and emerged sooner on several city plants than other plants (Figure 2-2).
Flowering on individual plants lasted 4 6 weeks, with green fruits becoming clearly
visible 3 4 weeks later. Fruits began ripening in December, approximately four months
after formation. It was mid-late January before all plants had mature fruit (Figure 2-2).
Some fruits remained on the plants for the duration of the 10-month sampling
period, by which time the next season's fruits were developing. Fruit loss at Putz
appeared to be greatest between March and June, while the City site appeared to have a
more constant fruit loss rate (Figure 2-3). The loss rates were significantly different
between sites, with plants at Putz losing fruits more quickly than those in the City (p = <
0.0001). The height of the branch above and below 60 cm did not influence fruit loss rate
(p = 0.478, data not reported), but plant density did (p = 0.0002), with minimally dense
plants experiencing the shortest fruit retention times (Figure 2-4.).
Probability density function graphs for the three densities (Figure 2-5) show
average rate of fruit loss for the three densities tested. The peak of the curve represents
the date of greatest fruit loss, and the end point is the expected date of fruit depletion.
The end points were similar for minimal and moderate densities (161 and 165 days after
ripening, respectively), but was much later (332 days) for the maximum density plants.
The intervals between fruit loss rate peaks were similar between the minimal and
moderate density plants (33 days) and between the moderate and maximum density plants
The results support rejecting the constant fruit production hypothesis, as fruits
ripened in late December and January and were retained on the plants for much of the
year (Figure 2-1). The plasticity shown in the timing of plant reproduction (flowering,
fruit production and maturation) dates among years may be related to water stress. The
spring of 2000 was considerably drier than 2001 in north-central Florida (Table 2-1), and
flowering peaked 1-2 months later in 2000 (Figure 2-1). The notion that water influenced
flowering and fruiting dates is also supported by those City plants that flowered in May
(sooner than any other monitored plants), because those few plants received water from
sprinklers. Furthermore, plants at Putz growing close to the bayhead (where soil may
have been wetter) also flowered earlier. One problem with correlating flowering dates to
rainfall is that it difficult to know when rainfall is most important. Pascarella (1998)
found that while rainfall strongly influenced flowering dates in the native Ardisia
escallonoides in southern Florida, the implicated rainfall occurred during the previous
Longetivity of fruit persistence
Ardisia fruits persisted on the plant for up to ten months without decaying (Figure
2-3). Furthermore, on a few occasions (on unmonitored plants) I found fruits from the
previous year in good condition still on the plant in January with the new crop of ripe
The ability of fruits of many species to resist decay and microbial attack has been
correlated to the presence of secondary compounds (Cipollini and Stiles 1992, Cipollini
and Levey 1997a). These compounds are important because frugivores avoid eating
infected fruits (Travaset et al. 1995, Garcia et al. 1999). However, the use of these
compounds presents a paradox to plants: while vertebrates may select against infected
fruits, they also prefer fruits from plant species that do not incorporate high levels of
secondary compounds (Cipollini and Levey 1997b, Levey and Cipollini 1998).
The low rates of fruit consumption in the field for ardisia fruits by birds and
mammals may be due to relatively high levels of secondary compounds, and the fruit's
apparent resistance to microbial attack supports this hypothesis.
Influence of fruit/plant presentation on loss rate
The lack of a significant difference in fruit loss rates for branches above and below
60 cm may indicate that mesomammalian consumption of ardisia fruits in these sites is
not common. This is supported by the lack of ardisia seeds found in mammal scat (see
this study, Chapter 4).
While plant density was negatively correlated with fruit loss rate (Figure 2-4), a
separate study would be needed to test for cause and effect. For example, ardisia may
grow at lower densities where less tree cover leads to a drier microclimate, or in less
fertile soils. If these conditions occur, they could result in plant stress that then causes a
shorter duration of fruit retention. Without fruitfall traps to distinguish between removal
and abscission, fruit fates could not be differentiated.
Several studies (Manasse and Howe 1983, Sargent 1990) have examined whether
fruit abundance in the immediate vicinity influences removal rate, while Denslow (1987)
studied the effects of plant density on removal rates. In conflicting results, Sargent
(1990) found that greater fruit abundance enhanced removal rates while Manasse and
Howe (1983) found lower fruit abundance led to higher removal rates. Both the Sargent
and Manasse and Howe studies measured the fruit abundance in the immediate area, but
my variable of interest was plant spacing and not fruit abundance. It is possible that both
factors (numbers of fruits and plants in an area) can influence fruit removal rates from
individual plants. Either through attracting more frugivores to the immediate vicinity or
competing among each other for frugivores, the likelihood of an individual fruit being
consumed may be influenced by neighborhood effects. Another possibility is that
increased plant density could obstruct the birds' views while they are searching for fruits.
Denslow (1987) tested the effects of plant density on fruit removal rates, and found
that the more isolated a plant was, the greater the fruit loss rate. Her findings agree with
this study, and her interpretation was that competition for frugivores is greatest among
bushes in close proximity to one another. Denslow studied plants growing in forest
clearings along a river where the separate populations were not within sight of each other.
My study differed in that the plants were all within a single population thus giving the
frugivores the ability to choose the fruit from the plant that offered the preferred density
of surrounding vegetation. At the Putz site different animal species may have been
selecting for different plant densities.
There were many observations of fruits that were partially depulped while still on
the plants, suggesting that small rodents (e.g. cotton mice, Peromyscus gossypinus) are
responsible for at least some fruit loss. In a pattern suggesting mouse consumption,
depulping was usually in a circular pattern from the middle of the fruit cluster, where
only the tops of the fruits were depulped. I also found rodent feeding platforms where
ardisia fruits had been depulped, leaving piles of skins and seeds. These observations
seemed more prevalent in the minimal-density plants, and appeared more often in April
and May. This period coincides with a dry period in Florida (Table 2-1), and a
hypothesis is that the mammals were using the fruits as a source of water. After losing
their moisture-retaining pulp and skin, the seeds of many of these fruits shriveled and
became hard, and probably lost viability. Probably due to their small size and ability to
climb, feeding evidence from small mammals was found in all but the tallest ardisia
plants. Investigation of mammalian consumption and the causes of different fruit loss
rates among different plant densities would be interesting areas for further research. Not
only would this mechanism of fruit loss appear no have no effect on long distance seed
dispersal, but introduces a possible source of seed loss.
Another alternative hypothesis is that a bird species that I had not considered is
responsible for the fruit loss in less dense plants. Wild turkey (Meleagris gallopavo) eat
fruit when available (Martin et al. 1951), and could choose fruits from isolated plants
over those in dense clusters to reduce the risk of predation.
Seasonality of Fruit Loss
The increase in fruit loss rate at Putz between March and June (Figure 2-5) may be
due to migrating birds, because this period coincides with spring migration (Stevenson
and Anderson 1994). That the quieter, less human-visited Putz site experienced this
increased rate and not the City site could be due to habitat preferences of shy, frugivorous
bird species bird species (such as gray catbird, Dumetella carolinensis).
Fruit removal rates may also be influenced by the flocking tendencies of different
bird species. At San Felasco Hammock State Preserve just north of Gainesville, FL,
Skeate (1987) censused bird numbers through the year and assessed their usage of native
fruits. Although he considered American Robins (Turdus migratorius) and Cedar
Waxwings (Bombycilla cedrorum) to be common in the winter and spring, their numbers
were erratic. These two highly frugivorous species often travel together in large flocks,
and exhaust local fruit resources before moving on to another area (Skeate 1987, personal
observation). Thus, fruit removal rates from these two species should show steep
declines in relatively short periods of time, such as between my sampling dates. This was
not observed in my study.
Skeate (1987) categorized the native fleshy-fruit producing plants according to
when their fruits matured, and he found only four of 45 species fruited in winter. These
were American holly (Ilex opaca), laurel cherry (Prunus caroliniana), American olive
(Osmanthus americanus), and mistletoe (Phoradendron leucarpum). It is interesting to
note that these four species shared the traits of being evergreen, and having fruits with
high persistence and low rates of spoilage. Ardisia is also a winter-fruiting plant that has
those same qualities. Skeate (1987) speculated that the fruit's ability to persist on the
plant for long periods without spoiling was a coevolved relationship between the plants
and the erratic behavior of the robins and waxwings that feed so heavily on those fruits.
If they are not depleted by wintering birds, fruits of these species will persist into spring
migration times (personal observation).
The ability of coral ardisia to retain its fruits for long periods is a trait that has been
found with other introduced plants (White and Stiles 1992, Bartuszevige and Gorchov
2006). The majority of native species set fruit in the fall, a period that coincides with fall
migration. Due to high consumption rates, these native fruits are considered to be very
important to birds (Stiles 1980) with most usage occurring in late fall and winter. White
and Stiles (1992), and Bartuszevige and Gorchov (2006) also concluded that if not
removed from plants, the fruits of introduced species tended to persist for long periods
without deteriorating. However, this quality may not reflect any taxonomic relationships,
but rather selection by humans for particular landscaping features. For instance, if having
persistent berries is not a common winter trait for plants, humans may select those
species that have this feature. Indeed, this trait could have also increased the use of
ardisia as a landscape plant (Kitajima et al. 2006).
In conclusion, ardisia flowered in the summer, peaking in July or August. There
was some variation among years that was possibly due to water availability. Green fruits
became visible 3 to 4 weeks following flowering, and took about four months to ripen.
All fruits ripened in the winter, and ripening peaked in January. Fruits usually lasted on
the plants through the summer, and some remained into October. Plant density affected
loss rates, with the more densely growing plants retaining fruits for longer periods.
Branch height did not affect loss rates. There were differences in fruit loss rates among
sites, with the natural sites losing fruits more quickly than those in the city. The use of
fruitfall traps to distinguish between active and passive loss would be a useful next study.
City -2000 __2001
Putz __2000 2001
Figure 2-1. Flowering dates ofArdisia crenata.
e- 01 A\0 A\<'P 0011 4p 40 cg
unripe 2000 ripe 2000
. - unripe 2001 .- ripe 2001
S- - unripe-2001 ripe-2001
100 -OO .-
PPSP unripe -- ripe
0" ^N \ \
" v ^^y^ i
Figure 2-2. Ardisia crenata fruit production dates for three populations. Values on the
Y-axes represent the % of monitored plants at that stage.
^ ^ ^A,\
--- Putz ---A--- City
2/18 4/9 5/29
Figure 2-3. Fruit losses from Ardisia crenata plants at two sites during 2001. Error bars
represent standard errors.
....... Minimal -- Moderate Maximum
1/15 1/29 2/28 3/31 5/7 6/1 7/3 7/31 9/1 10/5
Figure 2-4. The loss rates ofArdisia crenata fruits at Putz during 2001. The three lines
represent loss rates for three densities ofA. crenata plants. Error bars show
95% confidence intervals.
-" l- I l i;=A '_- I= J=,- -ii s,_ ,'M- I
curve peak line end
minimal 55 161
moderate 88 165
maximum 114 332
Figure 2-5. Probability density function graphs for Ardisia crenata fruit loss rates at
three different densities of plants; A) minimal, B) moderate, and C)
maximum. The values given at the top of each graph are the coefficients for
the quadratic equation. Values on the X-axis are the number of days since
fruit ripening. Peak curve values are when fruit losses are greatest, and the
end of the line is the estimated last day of fruit persistence.
I E-i -* I -- l F I.I =- L=- : .- 573 ,=1= I: I l J I*
Table 2-1. Rainfall levels for Gainesville, FL (units are inches).
2000 Amount 2001 Amount Average Amount
January 3.17 January 0.80 January 2.60
February 0.69 February 0.91 February 3.27
March 2.13 March 5.36 March 4.11
April 0.92 April 1.01 April 3.48
May 0.51 May 1.36 May 3.72
June 5.78 June 10.83 June 6.64
Total 13.2 20.27 23.82
It is not known how many years data collection constitute averages (Gainesville Sun,
June 30, 2002).
NUTRITIONAL CONTENT OF Ardisia crenata FRUITS
As a means to achieve seed dispersal, many plants have evolved fleshy fruits. Most
fruits adapted for vertebrate consumption consist of seed(s), pulp, and skin, with the pulp
(and to a lesser extent the skin) providing a nutritional reward to dispersers. Although
invertebrates may occasionally be involved (Kaufmann et al. 1991), this type of seed
dispersal is usually carried out by vertebrates.
The nutritional value of fleshy fruits is usually assessed by measuring protein, lipid,
and carbohydrate levels (Stiles 1980), and studies have found that levels of these
nutrients can help to explain the choices made by birds when selecting among the fruits
of different plant species (Johnson et al. 1985, Lepczyk et al. 2000).
Another factor that can influence fruit choices by birds is the seed load. Seed load
is the proportion of fruit mass that is undigestible seed, and not surprisingly, birds may
prefer fruits with lower seed loads (Howe & Vande Kerckhove 1981, Stanley & Lill
2002, Russo 2003).
The research described in Chapter 3 assessed the nutritional content of coral ardisia
fruits and compared the results to other fruit species that represent alternative choices for
birds in northern Florida. If ardisia is obviously nutritionally deficient, it may help to
explain the low rate of fruit consumption by birds.
Fruits were gathered from two populations on the University of Florida campus
near Bivens Arm Lake and Lake Alice during May of 2002. Following collection, fruits
were weighed and measured with calipers (n = 58). Pulp and skin were manually
separated from seeds to calculate seed load and to determine lipid, soluble carbohydrate,
and protein levels of the pulp and skin. Seeds were not included in the nutritional
analyses because they were not digested during the bird feeding trials (Chapter 4). Fruit
skin is commonly not digested, but after examining feces from feeding trials (Chapter 4),
it appeared that much of the skin was at least partially digested. Consequently, fruit skins
were included in the nutrient analyses.
Fruit material was freeze-dried to a constant mass. Freeze-drying was used rather
than heat drying because high temperatures have been shown to influence the results of
nutritional analyses (Mary Beth Hall, Department of Animal Sciences, personal
Moisture content of fruits was calculated using the formula: (dry weight / wet
weight) x 100. Samples for lipid analyses were not ground. Samples for protein and
carbohydrate estimates were ground in a Wiley mill with a 1 mm screen. Material was
mixed together and subsampled for nutrition analyses. Three replicates were used for the
lipid analysis and five replicates each for protein and carbohydrates.
Lipid content was assessed using the ether extraction method in which fruit
material (pulp and skin) is dried, weighed, soaked in ether 6-8 hours, dried, and re-
weighed. The difference in weight is due to the movement of lipids into the ether.
The amount of protein was calculated from nitrogen levels using the Kjeldahl
extraction method (Association of Official Agricultural Chemists 1980) with a boric acid
modification during distillation, and a conversion factor of 6.25 to calculate protein
(Pierce and Haenisch 1947).
Carbohydrate analysis was performed using an 80:20 ethanol/water solution to
remove sugars, which are ethanol-soluble. Enzymatic analysis was performed on the
residues to estimate starch content. Carbohydrates are reported as TESC (Total Ethanol
Soluble Carbohydrates). The carbohydrates measured in this analysis include mono- and
oligosaccharides (Hall et al. 1999).
Unless otherwise specified, data are presented as a percentage of dry matter.
Fruit measurements and nutritional data are presented in Table 3-1. The fruits of
coral ardisia averaged 8.9 mm in diameter (sd = 0.41 mm). The average fresh weight of
whole fruits was 244.4 mg (sd = 60.74) with a mean seed mass of 91.4 mg (sd = 24.73),
resulting in an average seed load of 37.6 %. Pulp and skin of fresh fruits averaged 88.5
percent water content (sd = 2.42). The lipid analysis found 8 percent (sd = 3.61) of dry
matter was due to lipid weight. Crude protein levels were calculated to be 3.7 percent (sd
= 0.09), and carbohydrate levels to be 47.8 percent (sd = 0.55).
Finding nutrient deficiencies may suggest why birds rarely eat the fruits of coral
ardisia. To put the fruit nutritional data into context, I included data for other (native)
species from White (1989) (Table 3-2). White (1989) did not provide intraspecific
statistical data (i.e. means and error terms). The values in Table 3-3 are means for the
native species listed, grouped together by seasonality. White (1989) also did not include
seeds with his nutritional analyses. However, other authors have not made that claim
(e.g. Stiles 1980), and failure to omit seeds may affect nutritional analyses with that part
of the fruit that is not typically digested. Mono- and disaccharide concentrations from
White (1989) were reported in the present study as carbohydrate levels, and the
differences in fruit dry mass are probably attributable to indigestible, structural
When foraging for fruits, birds often have multiple plant species available from
which to choose. Patterns emerge as they consistently choose fruits of some plant species
over others. While the choices can be observed and the preferences reported, explaining
why the selections are made is not easy. To determine some of the factors involved in
fruit choice, birds can be captured and fed artificial fruits where a single variable of
interest is manipulated.
Stanley and Lill (2002) fed birds translucent artificial fruits containing either a
single large or no plastic bead (which simulated a seed). The authors found that the birds
showed a strong preference for fruits that contained no bead. The coral ardisia seed load
value of 37.6% was within the range of 10.5% to 68.0% for winter fruits from native
species (Table 3-2), and was very similar to the mean seed load for winter fruits (Table 3-
3). All of the species listed for spring/summer are probably eaten with greater frequency
than is ardisia (personal observation), and all of those species have lower seed loads than
Conversely, poison ivy fruits are produced in winter and readily consumed by birds
despite a seed load of 68.0%, a trait perhaps overridden by its 47.2% lipid value. Many
fruits that ripen in the fall are high in lipids, and because this period coincides with fall
migration it is assumed that this represents an advancement of the bird-fleshy fruit
relationship by meeting the metabolic demands of migration with high-energy, lipid-rich
fruits (Stiles 1980). Thus, high-lipid fruits are some of the most sought after fruits in the
fall. In addition to this temporal-related trend, certain fruit selections may also be on an
avian taxonomic level, such as the high-lipid fruit preference exhibited by thrushes
(Witmer and Van Soest 1998).
Lipid content of coral ardisia fruits (8.0%) falls within the extremes of winter
(0.7% to 45.2%) and spring/summer (0.2% to 19.1%) native fruit species (Table 3-2).
Lipid values for coral ardisia are less than the mean value for winter fruits, but greater
than the mean lipid value for spring/summer fruits (Table 3-3). It is interesting to note
that coral ardisia has a 10-fold greater lipid value than American holly, and the holly is
probably the most preferred native winter species of those listed (personal observation).
My observation of American Robins stripping American hollies of their fruits and
ignoring the equally abundant coral ardisia fruits (Chapter 4) suggests that despite the
higher lipid values of coral ardisia, other factors influence fruit selection.
The protein content in ardisia fruits (3.7%) falls within the extremes of winter
(1.9% to 6.2%) and spring/summer (2.2% to 9.8%) native fruit species (Tables 3-2),
suggesting that low protein levels is not why birds do not consume ardisia fruits.
Similarly, the carbohydrate content in ardisia fruits (47.8%) falls within the range of
winter (0.0% to 52.0%) and spring/summer (38.3% to 83.3%) fruits (Tables 3-2 and 3-3).
Carbohydrate content may not be very useful for predicting winter fruit usage, as
the species with the highest content gallberryy) retains its fruits for months with
seemingly little consumption by birds, and the species with no carbohydrates (poison ivy)
often has its fruits removed quickly. Lepczyk et al. (2000) found that American Robins
preferred high-sugar, low-lipid fruits in summer, but changed their preferences to low-
sugar, high-lipid fruits in autumn.
It is difficult to draw any firm conclusions about the effects of coral ardisia's
nutritional components on avian fruit preferences when compared to native species,
because some of those species are more-or less-preferred over others (though probably all
would be favored over ardisia fruits). For example, of the winter fruits listed in Table 3-
2, greenbrier and gallberry are rarely eaten, whereas American holly and red cedar are
consumed by birds much more frequently (personal observation). Without a ranking of
the preference values for the native species listed in Table 3-2, those data should only be
used to gauge the variability and "average" of fruit nutritional values.
There are no obvious qualities with regard to seed load or nutritional content that
would explain why birds rarely eat coral ardisia fruits. However, the ability of the fruits
to resist decay suggests high levels of secondary metabolites (Chapter 2), which have
been shown to deter fruit consumption (Cipollini and Levey 1997b, Levey and Cipollini
To estimate protein using the Kjeldahl extraction method (Association of Official
Agricultural Chemists 1980), nitrogen is measured and then multiplied with a conversion
factor of 6.25. This conversion factor was derived from animal-based samples, where all
nitrogen is associated with protein. However, while this factor is also used to estimate
protein levels in plant-based samples, plants can contain nitrogen-based secondary
compounds. Thus, it is possible to over-estimate the amount of protein in a sample of
plant material if a 6.25 conversion factor is used. Levey et al. (2000) calculated
conversion factors for fruits of 18 plant species in the southeastern United States,
including coral ardisia. These authors determined that the correct conversion factor for
estimating protein levels in coral ardisia fruits is 6.28 (corrected conversion factors
ranged from 3.11 to 6.77). This value is nearly identical to that for animal-based
samples, and suggests that any secondary compounds present in coral ardisia do not
contain high levels of nitrogen.
Based on the measured characteristics of the fruit size, seed load, and nutrient
content, there is no reason to believe that birds would not eat coral ardisia fruits if
provided no alternatives. Nor are there any obvious characteristics measured here that
would lead us to expect that coral ardisia would not be preferred over many other fruits.
Table 3-1. Data for coral ardisia fruits (means and standard deviations). Percentages of
nutrient data are on a dry matter basis.
fruit diameter (mm)
fruit mass (mg)
seed mass (mg)
seed load (%)
total ethanol-soluble carbohydrates (%)
free glucose (%)
Table 3-2. Nutritional data for fruits available in northern Florida natural areas, grouped
together by season. Fruits from coral ardisia are generally available winter
through summer; values of fall fruits were included as a reference. Data for
all species other than coral ardisia are from White (1989).
Species season fruit seed seed water lipid protein CHO
mass mass load (%) (%) (%) (%)
(mg) (mg) (%)
Ardisia crenata coral ardisia w/s/s/f 244.2 91.4 37.6 88.5 8.0 3.7 47.8
Ilex glabra gallberry w 181.4 3.4 10.5 68.6 0.7 1.9 52.0
Ilex opaca American w 281.7 15.8 23.6 56.0 0.8 4.9 49.2
Juniperus red cedar w 46.3 9.4 25.1 48.4 7.2 3.8 41.7
Rhus coppalina winged w 20.7 12.9 62.3 26.0 15.8 2.9 3.4
Smilax greenbrier w 241.6 35.4 38.1 74.9 0.1 6.2 17.7
Toxicodendron poison ivy w 21.9 14.9 68.0 3.6 47.2 2.0 0.0
Gaylussacia blue s/s 285.2 1.6 5.7 82.8 1.2 2.1 82.3
Morus rubra red s/s 780.1 1.6 3.9 85.6 1.1 5.9 66.4
Phytolacca pokeweed s/s 395.3 8.5 20.5 83.6 1.0 9.8 38.3
Prunus serotina black cherry s/s 613.4 108.9 17.8 77.5 0.4 3.9 62.0
Rubus sand s/s 915.3 2.3 7.7 88.9 0.2 5.8 67.0
Sambucus elderbeny s/s 78.8 2.2 10.7 87.8 19.1 4.9 43.9
Vaccinium highbush s/s 425.5 0.9 3.0 84.3 0.8 4.9 49.2
Table 3-2 continued
Species season fruit seed seed water lipid protein CHO
mass mass load (%) (%) (%) (%)
(mg) (mg) (%)
f 257.8 72.2 39.2 83.3 0.4 5.8 27.5
f 293.5 11.5 10.2 73.8 0.4 4.4 12.5
f 203.2 42.9 21.1 78.8 2.0 4.8 54.8
f 786.2 80.9 20.6 68.9 0.9 2.4 30.7
f 51.2 19.4 37.9 77.5 7.0 8.2 37.3
f 371.1 141.8 38.2 81.5 34.6 11.9 12.0
f 135.9 56.1 41.3 41.7 36.7 5.2 23.6
f 433.4 128.2 29.6 73.4 14.8 3.6 45.6
Table 3-3. The means and standard deviations per season of the fruit nutrient data from
seed load lipids protein carbohydrates
coral ardisia 37.6(5.17) 8.0(3.61) 3.7(0.09) 47.8(0.55)
winter 37.9(22.89) 12.0(18.28) 3.6(1.70) 27.3(23.26)
spring / summer 9.9 (6.85) 3.4 (6.93) 5.3 (2.36) 58.4 (15.39)
fall 29.8 (11.76) 11.7 (16.52) 6.1 (3.09) 28.3 (14.87)
CONSUMPTION OF ARDISIA CRENATA FRUIT BY BIRDS AND THEIR ROLE IN
SUBSEQUENT SEED DISPERSAL
Two traits are evident in coral ardisia populations: fruits persist on the plant for a
long time (up to a year, as per Chapter 2) and the ground beneath mature plants is often
covered with seedlings (up to 300 stems per m2, Kitajima et al 2006). These observations
suggest that rates of both fruit consumption and seed dispersal are low. In spite of this,
the plant has become invasive, signifying that at least some successful seed dispersal is
occurring. In some locations, ardisia's dispersal is probably anthropogenic, but this does
not account for the plant's presence in more remote natural areas.
The fruit's bright red color and subsequent high visibility suggests bird dispersal of
the seeds is likely. Mammals, most of which are color blind, tend to feed on odorous
fruits (Van der Pijl 1972). Furthermore, fruits that utilize mammals for seed dispersal are
usually sweet-tasting (Van der Pijl 1972), and fall to the ground soon after ripening
(Stiles 1980). Using both field observations and laboratory experiments, the intent of the
present study was to evaluate birds as dispersal agents of ardisia.
Field observations consisted of monitoring bird activity when working in ardisia
stands during the collection of plant data (Chapter 2). Laboratory experiments consisted
of: 1) determining if certain frugivorous birds would indeed eat the fruits, 2) measuring
their preference for ardisia fruits compared to another native fruit, and 3) assessing any
effects of consumption on seed viability. If successful, the methods from these feeding
and germination trials would be used to develop a protocol to further assess birds as seed
dispersal agents of other introduced plants.
While collecting the life-history data reported in Chapter 2, approximately 100
hours of field observations were conducted to document birds feeding on the fruits of
both ardisia and native species. Although fruit traits do not suggest mammalian
consumption, it does not mean that it can be ruled out. Consequently mammal scat found
near ardisia stands was examined for ardisia seeds.
Selection of Birds for Experiments
Because coral ardisia fruits ripen in the winter and most persist through the spring
(Chapter 2), the bird species chosen for feeding trials were those that were considered to
be the greatest fruit consumers during this period in northern Florida (Table 4-1). They
included Gray Catbird (Dumetella carolinensis), American Robin (Turdus migratorius),
Cedar Waxwing (Bombycilla cedrorum), and Northern Mockingbird (Mimuspolyglottos).
Two other frugivorous species were tested (European Starling (Sturnus vulgaris) and Fish
Crow (Corvis ossifragus)) because they were already available in captivity where the
feeding trials were conducted. The remaining four native bird species were captured
using mist-nets at several locations in Alachua County.
Captive Feeding Trials
Captive feeding trials were performed at the USDA's National Wildlife Research
Center, in Gainesville, Florida. The European Starlings, Gray Catbirds, and Northern
Mockingbirds were housed and tested in individual cages measuring 18 inches x 18
inches x 18 inches. The Fish Crows were housed and tested in cages measuring 10 feet
long x 10 feet wide x 6 feet high. American Robins and Cedar Waxwings were held in
the larger cages until they acclimated to captivity, and then were placed in the smaller
cages for individual testing.
While in captivity, birds were fed a maintenance diet of moistened Purina kitten
chow, fly pupae, and fruit. Several species of fruits were offered including both wild and
cultivated species. The fruit were never of the same species used as control fruits tested
for that bird species. This was done because fruits contain secondary metabolites that
may affect consumption rates (Levey and Cipollini 1998, Cipollini and Levey 1997a,
Northern Mockingbirds were the quickest to acclimate, and would often begin
feeding within an hour. American Robins and Cedar Waxwings took the longest to
acclimate, and one Cedar Waxwing died after not feeding for the first three days. To
entice birds to eat during acclimation, birds were often given a branch of a native plant
that held fruits. It was found that presenting food that the birds may have been feeding
on-and in a manner that they were accustomed to-helped their acclimatization. Before
undergoing feeding trials, birds had to show that they would feed from cups suspended
from the sides of the cages. The European Starlings and Fish Crows had been in
captivity for some time and were already feeding freely. Birds were housed individually
and could not see each other while in the test cages.
Testing lasted for three days. On the first day, birds were offered the fruits of a
native species that would later act as a control fruit during preference trials. The purpose
was to ensure that the birds being tested would eat the fruit of that plant species. On day
two, birds were offered coral ardisia fruits only. If birds failed to consume fruits on
either of these two days, they would not proceed to the choice test and would be released.
In the choice test on the third day, birds were offered equal numbers of test fruits (ardisia)
along with control fruits, and the remaining fruits were counted hourly to determine
which fruit species the birds preferred. The number of birds tested for each species
ranged from five to ten.
During testing, food from the previous night was removed at 0730 hours, and birds
were without food for one hour. Test fruits were placed in cups from 0830 until 1230.
Twenty fruits of each species were placed in the same dish, and counted hourly for a total
of five times (t = 0, 1, 2, 3, 4). If the birds ate all fruits before 1230, no additional fruit
was provided and maintenance diet was placed back into cage. Birds often dropped fruits
during the trials, and many fell outside of the cage. During the hourly counts, these fruits
were returned to the feeding cups and listed as fruits remaining. Failure to return these
fruits would have led to an over-estimation of fruits consumed. For some birds, a video
camera was used to observe fruit handling during feeding trials.
Voided ardisia seeds were gathered to assess germination rates. Seeds were placed
in petri dishes between two sheets of filter paper, wetted with deionized water, and put
into germination chambers. The conditions in these chambers were 12 hours of light per
day with temperature during the light period set at 250 C and 150 C during darkness. No
more than 100 seeds from each bird species were evaluated, and no more than twenty
seeds went into a single petri dish.
Selection of Native Plants for Experiments
The initial protocol called for the fruits of American holly (Ilex opaca) to be used
as the native (control) fruits for preference trials. American holly grew naturally at two
of the field sites and fruited at the same time of the year as ardisia, and thus served as an
alternative fruit choice similar in size and color to ardisia.
Because it was not known if European Starlings would eat American holly fruits,
and because they were mainly used for protocol evaluation, the fruits of camphor trees
(Cinnamomum camphora), sometimes eaten by starlings, were selected for use as the
control fruits for European Starling testing.
American Robins arrived in Alachua County early during the winter of 2000-2001,
which may have been due to very low temperatures earlier than usual. After their arrival
into the county, it became difficult to find fruits on American holly trees. While
American holly fruits were used for the American Robin feeding trials, alternative fruit
species for all other birds had to be found. In selecting alternative control fruits, species
were chosen that shared fruiting periods with ardisia (i.e. winter, spring), and those that I
knew were eaten by the relevant bird species.
The fruits chosen for Cedar Waxwing were those from East Palatka hollies (Ilex x
attenuata). This tree is commonly used for landscaping around North Central Florida,
and like American holly, has red fruits winter through summer if they are not removed by
birds. East Palatka holly is a hybrid between dahoon (Ilex cassine) and American holly.
Fruit choice for Gray Catbirds was Godfrey's privet (Forestiera godfreyi), a native
shrub whose springtime fruits are often eaten by Gray Catbirds (personal observation).
Fruit choice for Northern Mockingbirds and Fish Crows was highbush blueberry
(Vaccinium corymbosum). This commercially-important species also occurs naturally,
and both of these bird species are known to eat the fruits in commercial groves (Mike
Avery, USDA, personal communication).
Data from choice test feeding trials were analyzed using GENMOD in SAS
statistical software (SAS Institute, version 8.2). Data were logit transformed and
Generalized Estimating Equations (GEE) (Agresti 1996) were estimated. These GEE
estimates were the slope estimates of control and test results, and the slopes were then
tested for significance from each other. Significant differences meant that the bird
species being tested chose one fruit species over another.
In December of 2000, a flock of about 40 American Robins come into the Putz site.
The birds gorged themselves on American holly fruits, and most of the birds departed
when the fruits were depleted. A couple of birds remained for several days and fed on
what appeared to be the less desirable fruits of a Smilax spp. and American olive
(Osmanthus americanus) until these fruits, too, were exhausted. There were numerous
ardisia plants with ripe fruits in the area, but no robins were ever seen feeding on ardisia
In the second half of April 2002 at the Putz site, several Gray Catbirds were
observed consuming ardisia fruits. There were about sixteen birds on three acres of
ardisia plants, and each bird appeared to swallow approximately three fruits every six to
fifteen minutes. This was the only time out of two and a half years (three spring
migrations) that Gray Catbirds were observed in the ardisia stands, and Gray Catbirds
were the only species observed to feed on ardisia fruits in the field.
Outside the study sites, I found regurgitated ardisia seeds under a Northern
Mockingbird song perch at Carr Hall on the University of Florida campus, and was told
of a Northern Mockingbird eating fruits during an especially cold winter evening
(Carmine Lanciani, Zoology Department, University of Florida, personal
communication). Finally, the founder of Birdsong Nature Center in Thomasville,
Georgia, reported a single occurrence of a flock of Cedar Waxwings feeding on the fruits
of coral ardisia at the center (Kathleen Brady, Birdsong Nature Center, personal
Mammal scat in the vicinity of the field sites often contained the seeds of other
plants, but none contained ardisia seeds. On a single occasion I did find ardisia seeds in
mammal feces, but not at any study site.
All six bird species consumed native and ardisia fruits during the no choice trials.
Fish Crows were the most reluctant species to consume fruit, where only five often birds
consumed ardisia fruit during no choice trials. Therefore, only five Fish Crows were
used for the preference (choice) trials. Similarly, while five Gray Catbirds passed the no
choice test, one of them ate nothing during the preference trials.
European Starling (Figure 4-1) and gray catbird (Figure 4-2) displayed no taste
preference between control and test fruits (Table 4-2). The remaining four species,
Northern Mockingbird (Figure 4-3), American Robin (Figure 4-4), Cedar Waxwing
(Figure 4-5), and Fish Crow (Figure 4-6), showed significant preference for control fruits
(Table 4-2). These four species ate few ardisia fruits until all the control fruits had gone.
Germination rates for voided seeds from Gray Catbirds, Northern Mockingbirds,
American Robins, and Cedar Waxwings did not differ from manually depulped seeds
(Table 4-3). Consumption by Fish Crows did significantly decrease seed germination (p
< 0.05, Fischer's chi-square test) (Table 4-3), probably due to seed coat damage that was
incurred while attempting to depulp seeds with their beaks. Germination rates for seeds
voided by European Starling were not determined.
All bird species swallowed ardisia fruits whole, and it appeared all species
regurgitated seeds, with the exception of Cedar Waxwings, which defecated seeds. In
addition to swallowing whole fruits, some Fish Crows also attempted to manually
separate fruit pulp from the seed in a technique known as mashing (Moermond and
Deslow 1985). However, because the fruits of ardisia have pulp that tightly adheres to
the seed, they were not very successful. Some fruits had no pulp removed while others
were partially depulped, but none were completely depulped through mashing. The only
seeds that were completely depulped by the Fish Crows were those that were done so in
their crops. Of these, the five birds completely depulped only 29 seeds (Table 3-2), and
of those, a single bird was responsible for 17 seeds.
When the American Robins exhausted the American holly fruits at the Putz site,
they had two choices: either remain and feed on less-preferred fruits, or leave the site to
forage for alternate food sources. Two individuals chose the former alternative, but most
of the birds made the latter choice. Even for the two robins that remained, ardisia fruits
The observation of Gray Catbirds feeding on ardisia occurred during spring
migration, a period when only three native plants were observed to produce fruits. These
include Godfrey's privet, red mulberry (Morus rubra), and blackberry spp. (Rubus sp.).
However, like ardisia, winter-fruiting species retained fruits if they had not been removed
by birds. These included East Palatka holly trees in urban communities where frugivores
were less common and a single American holly whose fruits the birds did not eat. It
appeared that red mulberry and Godfrey's privet in Alachua County did not produce fruit
in 2002, possibly due to a late hard freeze (18.80 F) on February 28
(http://fawn.ifas.ufl.edu/) that may have damaged the plants. Red mulberry produces one
of the most frequently bird-consumed fruits in Alachua County (personal observation),
and I have seen Gray Catbirds eating unripe red mulberry fruits while ignoring abundant
and ripe ardisia fruits on the University of Florida campus. The choice of Gray Catbirds
to feed on ardisia fruits during the spring of 2002 at the Putz site may have been related
to the apparent crop failure of Godfrey's privet and red mulberry in Alachua County. If
crop failure in other plant species forced Gray Catbirds to eat more ardisia fruits than
normal in spring 2002, then my observations may not be indicative of a typical year.
Coral ardisia grows in dense patches of hardwood forest understory, a habitat it
shares with its analogous congener in southern Florida, shoebutton ardisia (A. elliptica).
Similarly, shoebutton ardisia's primary disperser seems to be Gray Catbirds (Koop 2004),
and during spring migration (Tony Koop, personal communication). Gray Catbirds often
remaining in dense low-growing vegetation. They are common in winter in Florida, and
waves of migrants come through northern Florida from mid-April through mid-May
(http://myfwc.com/bba/GRCA.htm). Stragglers linger into June.
Research discussed in Chapter 2 showed that fruits of coral ardisia can be expected
on plants January through October. Fruit loss rates increased in March and continued
through June, with loss rates peaking April through early May. This loss rate is
consistent with populations of Gray Catbirds in northern Florida. Most American Robins
leave Florida in February, with far fewer numbers remaining into March. Cedar
Waxwings begin leaving in April, although stragglers linger into May. Skeate (1987)
found that winter flocks of American Robins and Cedar Waxwings in northern Florida
were nomadic, remaining in an area as long as fruit were present, and then dispersing to
other areas. Based on fruit loss rates, it would not appear that American Robins or Cedar
Waxwings are responsible for much of the coral ardisia fruit loss.
The observations of Cedar Waxwings eating ardisia at Birdsong Nature Center
fruits may be considered atypical behavior. Kathleen Brady reported that their ardisia
plants grow under scattered mature trees, a habitat different from that normally invaded
by ardisia: a closed-canopy forest (personal communication). Lima (1993) argued that
birds choose habitats based upon their escape tactics from predators. He described Cedar
Waxwing's escape strategy as aerial, and my observations of Cedar Waxwings being
attacked by a merlin (Falco columbarius) while mist-netting support this. Based on
Lima's reasoning, Cedar Waxwings might avoid the forest understory where their escape
routes would be compromised.
Kaoru Kitajima (Botany Department, University of Florida) observed that rates of
fruit loss in coral ardisia's native range (Japan) appeared to be greater than those in
northern Florida (personal communication, unpublished data). In Japan, Brown-eared
bulbuls (Hypsipetes amaurotis) are known to consume the fruit of ardisia, but other Asian
forest-interior bird species, such as the pale thrush (Turduspallidus), may consume the
fruits as well (K. Kitajima, personal communication). Brown-eared bulbuls are the
family Pycnonotidae, a family with no native North American representatives. This lack
of closely-related species in North America could help explain ardisia's low fruit
consumption rates in its introduced range, and may be a situation where plant population
and expansion is limited by a lack of suitable seed dispersers.
For many plant species, high mortality-especially at the seedling stage-tends to
swamp out the importance of seed dispersal (Howe 1993, Howe et al. 1985, Schupp
1988, Chapman and Chapman 1995). Thus, while it is generally accepted that seed
dispersal is not as important a stage in the life history of a plant as once thought (Schupp
1995), seed dispersal could be limiting for introduced species. It may be possible that
coral ardisia lacks suitable dispersers because it lacks suitable fruit consumers.
Field observations indicate that Gray Catbirds may be the most important avian
consumers of ardisia fruits. Some fruit may also be eaten by Northern Mockingbirds and
The lack of ardisia seeds in mammal scat supports earlier findings in this study
(Chapter 2) where low-growing branches did not lose fruits at a greater rate than those
that grew higher. However, the one instance of finding mammal scat containing ardisia
seeds indicates that at least one species will consume fruits from this plant. Dozier
(1999) reported ardisia seeds in mammal feces, and what she called "vomit piles." On
several occasions I also found piles of seeds (not in feces), but in my opinion they
appeared to be the result of masticating a mouthful of fruits from which the pulp was
sucked and the seeds spit out en masse. Furthermore, these were found within ardisia
stands, and not at a distance away from the plants where one might enough time to have
passed for a toxin to trigger vomiting. Dozier speculated that the raccoons made the
vomit piles, and I have no reason to disagree with that hypothesis. It is not known what
effect these actions have on seed dispersal of coral ardisia.
No bird species preferred the fruits of ardisia over control fruits. While four of six
chose control fruits over ardisia, the remaining two species (European Starling and gray
catbird) showed no choice preference for either test or control fruits. Thus, laboratory
feeding trials support field observations that that most birds prefer native fruits over those
The predictive ability of the feeding trials may be limited because they were
conducted with a single species of control fruit. Birds often have several choices
available when foraging for fruits, and they make their choices based on what is
available, including more than just the two choices that I presented. Furthermore, I chose
control fruits with the knowledge that the bird species being tested would eat that fruit,
conceivably biasing selection for control fruits over ardisia fruits. However, the feeding
trials showed that all species could eat coral ardisia fruits.
Northern Mockingbirds, Cedar Waxwings, and Fish Crows did not begin to
consume ardisia fruits until nearly all control fruits were exhausted (Figures 4-3, 4-5, and
4-6, respectively). Thus, while these three species exhibited significant differences in
fruit preference (Table 4-2), the values for those preference slopes could have been
different if more fruits were used in the feeding trials. For example, if enough fruits were
used of each species to meet the caloric requirements for the four hours of the feeding
trial, these three species may not have eaten any of the coral ardisia fruits. Alternatively,
if only five fruits were used for each fruit species, the metabolic demands of the birds
might have forced them to consume all ten fruits in the first hour of the experiment. A
fruit count at the end of the hour would have given the impression that there was no
preference for either fruit species.
In an attempt to get fruits with similar caloric values, fruits were chosen that were
of similar size as those of ardisia. However, the choice of highbush blueberry did not
meet these criteria. Probably as a result of selective breeding by humans, the fruits of this
species were larger than those of ardisia. Although the smallest fruits were used in
feeding trials, they were still larger than the test fruits. Their large size could have
resulted in less ardisia fruits eaten by Northern Mockingbirds and Fish Crows, especially
if the blueberries met the caloric requirements for the five-hour test.
Wheelwright (1985) found that the gape width of birds was correlated to the size of
the fruits they were able to swallow, with the wider-gaped species being able to swallow
larger-diameter fruits. If birds tried to swallow fruits of a diameter larger or similar to
their gape width, they were often unsuccessful and dropped fruits to the ground.
While the average diameter of coral ardisia fruits was measured and found to be 8.9
mm (sd = 0.41), the diameters of control fruits were not determined. The numbers of
fruit dropped during feeding trials were not counted, but if fruits were a bit too large for
the smallest of the birds tested, then the dropping and replacement of ardisia fruits may
have affected the results by prolonging the period that coral ardisia fruits remained.
Cedar Waxwings were the smallest of the birds tested, and their control fruits were
noticeably smaller than those of coral ardisia. Cedar Waxwings were the species that
dropped the greatest number of fruits.
Using their bill, some birds mash fruits to remove seeds prior to swallowing, a feat
not possible with coral ardisia fruits. Coral ardisia fruit pulp adheres very strongly to the
seed, a quality that may limit consumption to gulpers (see Chapter 1). Fish Crows
appeared to be both mashers and gulpers. Initially, they tried mashing the fruits, but
when unsuccessful with separating the seed from the pulp, some became gulpers and
swallowed fruits whole. Only a single bird was able to completely clean seeds of pulp
before regurgitating them, and seed survivability was compromised when damaged
during mashing (Table 4-3).
The experiences with Fish Crows suggest two negative consequences to plants with
tightly-adhering fruit pulp. First, potential seed dispersal agents may be limited if fruit-
mashing birds, such as Fish Crows, choose mashable fruits for consumption. When
monitoring fruit loss rates (Chapter 2), I found some fruits with angled puncture wounds
on the top and bottom of the fruit. The wounds suggested a heavy-beaked bird sampled
and rejected the fruit. Heavy beaked birds are those that mash fruits (Moermond and
Denslow 1985). Second, the tight-adhering pulp can result in seed mortality, as the Fish
Crows damaged the seed while attempting to separate seed from pulp. Birds may lack
the ability to control the method of seed voidance, as it seems that with the difficulty in
separating seed from pulp in the crop, they would have allowed fruits to pass through
their digestive tract where the seed would have been cleaned of digestible material.
Seed handling and depulping strategies were determined by observing the voided
seeds and assessed using video footage. Defecated seeds were in feces, and were stained
by the skin. Regurgitated seeds were very clean, with no stains or pulp remaining, with
the exception of those regurgitated by Fish Crows. The video camera was useful to
watch birds trying to swallow (and perhaps taste) the fruits, but because voidance
happens so quickly, it was not always clear whether seeds were regurgitated or defecated.
A plastic or glass section of the cage may have allowed better video recording than the
In addition, if determining the duration of seed retention were an objective, fruits should
have been fed to the birds singly to be certain that the voided seed was from a certain
fruit. Video footage showed birds eating several fruits over a period of time before
ridding any seeds.
While it was not one of the objectives, better data concerning seed voidance
strategies and seed retention times may have been helpful to further assess bird species as
agents of seed dispersal, and would help to project plant population expansions.
The methods evaluated in this chapter appear to be capable of evaluating different
species of birds as seed dispersal agents. The field observations suggest birds in the
family Mimidae (especially gray catbird, but also Northern Mockingbird) are likely the
greatest consumers of coral ardisia fruits. Captive feeding trials suggest Gray Catbirds as
coral ardisia fruit consumers. None of the bird species that swallowed whole fruit
negatively affected seed germination rates.
Coral ardisia was introduced into the New World over 100 years ago and has
remained a part of Florida landscaping. It has since been recognized as being an invasive
weed, and has done so despite that fact that consumption of its fruits by possible seed
dispersers is an uncommon event. The purpose of this study was to gain insight into the
reproductive ecology of coral ardisia in its introduced range of northern Florida.
It was found that plants flower through the summer and fruits ripen in mid-winter.
The fruits can persist on the plants for up to one year, with the greatest rate of fruit loss
occurring in late April. Plants in a natural area lost fruits at a faster rate than those in an
urban area, and isolated plants lost fruits quicker than those in denser populations. The
period of greatest fruit loss coincides with the spring migration of birds through Florida,
and suggests that some consumption of fruits by birds is occurring. Fruits were found not
to be nutritionally inferior to native fruits, nor were any other factors found that would
suggest a cause as to why fruits of coral ardisia are rarely eaten. Despite many hours in
the field, observations of birds feeding on coral ardisia fruits was limited to a single day
with a single bird species (gray catbird). Captive feeding trials found that six species of
birds would eat coral ardisia fruits, but often favored the fruits of native species to those
of coral ardisia. Gray Catbirds and European Starlings showed the greatest acceptance of
coral ardisia fruits during captive feeding trials. Cedar Waxwings defecated seeds after
eating fruits, while all other species regurgitated seeds. Seed germination rates were no
different for 5 of 6 birds species tested between seeds defecated or regurgitated compared
to those that were manually depulped. From damage incurred while manually trying to
remove seeds from fruits, Fish Crows significantly decreased seed germination rates.
However, Fish Crows ate the fewest coral ardisia fruits.
Even if the dispersal of coral ardisia fruit by birds to beyond the edge of a stand is a
rare event, in the long-term these infrequent events may have a significant cumulative
effect. Coral ardisia seedlings have been shown to have a very high survival rate in field
studies (22 44%; Lindstrom 2002), so the dispersal of relatively few viable seeds into
suitable habitats could result in the establishment of new populations.
To test the likelihood of this further it would be necessary to use fruit traps under
the monitored plants so that the mechanisms of fruit loss could be better discerned. Now
that it is known which are the most likely species of bird to eat coral ardisia fruits and the
conditions under which coral ardisia fruits are most likely to be eaten, it might be
worthwhile to conduct further studies similar to those of Bartuszevige and Gorchov
(2006), in their study of the spread ofLonicera mackii by birds. That could include, for
example, capturing sufficient numbers of Gray Catbirds in and around coral ardisia
stands to evaluate if any of them voided coral ardisia seeds. Also the time between
ingestion and voidance could be measured and compared to how far the birds are likely to
move during that time. Alternatively, it might be interesting to establish why coral
ardisia fruit are not more appealing to native birds, either by examining fruit
characteristics such as the content of secondary compounds or by a more realistic
presentation of fruit on whole coral ardisia plants to birds in larger aviaries.
In conclusion, we have advanced from not knowing whether bird dispersal of coral
ardisia was possible to being able to say that it could occur. Given the number of
invasive plants for which bird dispersal is claimed but which lack sufficient evidence
(Meisenburg and Fox 2002), there are many of other applications of these types of study
that could be useful in improving our understanding of, and predictions about, the spread
of invasive plant species.
- Ardisia crenata
-*- Cinamwomum camphora
1 2 3 4
Figure 4-1. Results from European Starling feeding trials with fruits from Ardisia
crenata and Cinnamomum camphora (n=4). The birds showed no preference
for either fruit species. Bars represent standard error.
- Forestiera godfreyi
1 2 3 4
Figure 4-2. Results from gray catbird feeding trials with fruits from Ardisia crenata and
Forestiera godfreyi (n=5). The birds showed no preference for either fruit
species. Bars represent standard error.
-- Ardisia crenata
-- Vaccinium corymbosum
1 2 3 4
Figure 4-3. Results from Northern Mockingbird feeding trials with fruits from Ardisia
crenata and Vaccinium corymbosum (n=5). The birds showed a preference
for V corymbosum. Bars represent standard error.
25 flex opaca
0 1 2 3 4
Figure 4-4. Results from American Robin feeding trials with fruits from Ardisia crenata
and Ilex opaca (n=10). The birds showed a preference for I. opaca. Bars
represent standard error.
-- Ardisia crenata
- Ilex x attenuata
0 1 2 3 4
Figure 4-5. Results from Cedar Waxwing feeding trials with fruits from Ardisia crenata
and Ilex x attenuate (n=10). The birds showed a preference for x attenuata.
Bars represent standard errors.
- Ardisia crenata
- Vaccinium corymbosum
1 2 3 4
Figure 4-6. Results from Fish Crow feeding trials with fruits from Ardisia crenata and
Vaccinium corymbosum (n=5). Bars represent standard error.
Table 4-1. Population statuses of the birds tested. Data are from Stevenson
and Anderson (1994).
Alachua County status
Fish Crow Fairly common to common in summer, but uncommon to
abundant in winter (irregular)
American Robin Abundant in winter, majority leave before April
Gray catbird Uncommon in winter, abundant mid-April to early May
Northern Mockingbird Common to abundant year-round resident, numbers are reduced
away from humans and in wooded communities
Cedar Waxwing Common to abundant from late winter through spring
European Starling Common to fairly common year-round, with population
increasing from wintering individuals that depart February
Table 4-2. Transformed slope values from feeding trials. Significance levels indicate
that two bird species did not significantly choose one fruit species over the
other, while the remaining four bird species preferred control fruits over test
(A. crenata) fruits.
ardisia fruits native fruits level of significance
European Starling -0.86 -0.65 0.5961
Gray catbird -1.06 -1.03 0.3566
Northern Mockingbird 0.0* -1.83 <.0001
American Robin -0.69 -3.46 <.0001
Cedar Waxwing -0.85 -29.68 **
Fish Crow -0.86 -4.77 <.0001
* = Did not differ significantly from zero.
** = SAS could not calculate due to its low value (there were no fruits after the first
Table 4-3. Ardisia seed germination rates from seeds voided during feeding trials.
Control fruits were those that were manually depulped.
89.7 wholly depulped
87.5 partially depulped-intact
55.6 partially depulped-
damaged seed coat
* FICR frugivory significantly decreased seed germination rate at the 0.05 level.
LIST OF REFERENCES
Agresti A. 1990. Categorical Data Analysis. Wiley & Sons, New York.
Association of Official Agricultural Chemists. 1980. Official Methods of Analysis, 13th
Ed., Association of Official Analytical Chemists, Arlington.
Bartuszevige, A. M., and D. L. Gorchov. 2006. Avian seed dispersal of an invasive
shrub. Biological Invasions 8:1013-1022.
Chapman, C. A., and L. J. Chapman. 1995. Survival without dispersers: Seedling
recruitment under parents. Conservation Biology 9:675-678.
Cipollini, M. L., and D. J. Levey. 1997a. Antifungal activity of Solanum fruit
glycoalkaloids: Implications for frugivory and seed dispersal. Ecology 78:799-809.
Cipollini, M. L., and D. J. Levey. 1997b. Why are some fruits toxic? Glycoalkaloids in
Solanum and fruit choice by vertebrates. Ecology 78:782-798.
Cipollini, M. L., and D. J. Levey. 1997c. Secondary metabolites of fleshy vertebrate-
dispersed fruits: Adaptive hypotheses and implications for seed dispersal. The
American Naturalist 150:346-372.
Cipollini, M. L., and E. W. Stiles. 1992. Relative risks of microbial rot for fleshy fruits -
significance with respect to dispersal and selection for secondary defense.
Advances in Ecological Research 23:35-91 1992.
Cipollini, M. L., and E. W. Stiles. 1993. Fruit rot, antifungal defense, and palatability of
fleshy fruits for frugivorous birds. Ecology 74:751-762.
Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion
in some marine animals and in rain forest trees. Pages 298-310 in P. J. den Boer
and G. R. Gradwell, eds. Dynamics of populations. Proceedings of the Advanced
Study Institute on Dynamics of numbers in populations, Oosterbeek, 1970. Centre
for Agricultural Publishing and Documentation, Wageningen.
Cronk, Q. C. B., and J. L. Fuller. 1995. Plant Invaders. Chapman and Hall, London.
Denslow, J. S. 1987. Fruit removal rates from aggregated and isolated bushes of the red
elderberry (Sambucus pubens). Canadian Journal of Botany 65:1229-1235.
Dozier, H. 1999. Plant introductions and invasion: history, public awareness, and the
case ofArdisia crenata. Ph.D. Thesis. University of Florida.
FDEP. 2004. Upland Invasive Exotic Plant Management Program Fiscal Year 2003-
2004 Annual Report. Florida Department of Environmental Protection,
Figueroa, J. A., and S. A. Castro. 2002. Effects of bird ingestion on seed germination of
four woody species of the temperate rainforest of Chiloe Island, Chile. Plant
Garcia, D., R. Zamora, J. M. Gomez, and J. A. Hodar. 1999. Bird rejection of unhealthy
fruits reinforces the mutualism between juniper and its avian dispersers. Oikos
Gosper, C. R., G. Vivian-Smith, and K. Hoad. 2006. Reproductive ecology of invasive
Ochna serrulata (Ochnaceae) in south-eastern Queensland. Australian Journal of
Hall, M. B., W. H. Hoover, J. P. Jennings, and T. K. Miller Webster. 1999. A method
for partitioning neutral detergent-soluble carbohydrates. Journal of the Science of
Food and Agriculture 79:2079-2086.
Howe, H. F. 1993. Specialized and generalized dispersal systems-where does the
paradigm stand? Vegetatio 108:3-13.
Howe, H. F., E. W. Schupp, and L. C. Westley. 1985. Early consequences of seed
dispersal for a neotropical tree (Virola-surinamensis). Ecology 66: 781-791.
Howe, H. F., and Smallwood. 1982. Ecology of seed dispersal. Annual Review of
Ecology and Systematics 13:201-228.
Howe, H. F., and G. A. Vande Kerckhove. 1981. Removal of wild nutmeg (Virola
surinamensis) crops by birds. Ecology 62:1093-1106.
Izhaki, I., and U. N. Safriel. 1990. The effect of some Mediterranean scrubland
frugivores upon germination patterns. Journal of Ecology 78:56-65.
Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. Am.
Johnson, R. A., M. F. Willson, J. N. Thompson, and R. I. Bertini. 1985. Nutritional
values of wild fruits and consumption by migrant frugivorous birds. Ecology
Kaufmann, S., D. B. McKey, M. Hossaert-McKey, and C. C. Horvitz. 1991.
Adaptations for a 2-phase seed dispersal system involving vertebrates and ants in a
hemiepiphytic fig (Ficus microcarpa: Moraceae). American Journal of Botany
Kitajima, K., A. M. Fox, T. Sato, and D. Nagamatsu. 2006. Cultivar selection prior to
introduction may increase invasiveness: Evidence from Ardisia crenata. Biological
Koop, A. L. 2004. Differential seed mortality among habitats limits the distribution of
the invasive non-native shrub Ardisia elliptica. Plant Ecology 172:237-249.
Langeland, K. A., and K. C. Burks. 1998. Identification & Biology of Non-Native
Plants in Florida's Natural Areas. University of Florida, Gainesville.
Lepczyk, C. A., K. G. Murray, K. Winnett-Murray, P. Bartell, E. Geyer, and T. Work.
2000. Seasonal fruit preferences for lipids and sugars by American Robins. Auk
Levey, D. J., H. A. Bissell, S. F. O'Keefe. 2000. Conversion of nitrogen to protein and
amino acids in wild fruits. Journal of Chemical Ecology 26:1749-1763.
Levey, D. J., and M. L. Cipollini. 1998. A glycoalkaloid in ripe fruit deters consumption
by Cedar Waxwings. Auk 115:359-367.
Lima, S. 1993. Ecological and evolutionary perspectives on escape from predatory
attack: a survey of North American birds. The Wilson Bulletin 105:1-47.
Lindstrom, G. M. 2002. Life history characteristics of Ardisia crenata growing in natural
areas in north Florida. M.S. Thesis. University of Florida.
Mack R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. A. Bazzaz.
2000. Biotic invasions: Causes, epidemiology, global consequences, and control.
Ecological Applications 10:689-710.
Manasse, R. S., and H. F. Howe. 1983. Competition for dispersal agents among tropical
trees: influences of neighbors. Oecologia 59:185-190.
Martin, A. C., H S. Zim, and A. L. Nelson. 1951. American Wildlife & Plants, A Guide
to Wildlife Food Habits; the Use of Trees, Shrubs, Weeds, and Herbs by Birds and
Mammals of the United States. Dover Publications, New York.
McDonnell, M. J., and E. W. Stiles 1983. The structural complexity of old field
vegetation and the recruitment of bird-dispersed plant species. Oecologia 56:109-
Meisenburg, M. J., and A. M. Fox. 2002. What role do birds play in dispersal of
invasive plants? Wildland Weeds 5:8-14.
Meyer G. A., and M. C. Witmer. 1998. Influence of seed processing by frugivorous
birds on germination success of three North American shrubs. American Midland
Moermond T. C., and J. S. Denslow 1985. Neotropical frugivores: patterns of behavior,
morphology, and nutrition with consequences for fruit selection. Pages 865-897 in
P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgely, and N. G. Buckley, eds.
Neotropical Ornithology. American Ornithologists' Union Monographs, 36.
American Ornithologists' Union, Washington, DC.
Morton, J. F. 1982. Plants Poisonous to People in Florida and Other Warm Areas.
Southeastern Printing Company, Stuart.
Murray K. G., S. Russel, C. M. Picone, K. Winnett-Murray, W. Sherwood, and M. L.
Kuhlmann. 1994. Fruit laxatives and seed passage rates in frugivores:
consequences for plant reproductive success. Ecology 75:989-994
Pascarella, J. B. 1998. Hurricane disturbance, plant-animal interactions, and the
reproductive success of a tropical shrub. Biotropica 30:416-424. or Resiliency
and response to hurricane disturbance in a tropical shrub, Ardisia escallonioides
(Myrsinaceae), in south Florida. American Journal of Botany 85:1207-1215
Pierce, W. C., and E. L. Haenisch. 1947. Quantitative Analysis, 2nd Ed. Wiley, New
Renne, I. J., T. P. Spira, and W. C. Bridges, Jr. 2001. Effects of habitat, age and passage
through birds on germination and establishment of Chinese tallow tree in coastal
South Carolina. Journal of the Torrey Botanical Society 128: 109-119.
Royal Palm Nurseries. 1900. Annual Mail-order Catalogue. Oneco.
Russo S. E. 2003. Responses of dispersal agents to tree and fruit traits in Virola
calophylla (Myristicaceae): implications for selection. Oecologia 136:80-87.
Sargent, S. 1990. Neighborhood effects on fruit removal by birds: A field experiment
with Viburnum dentatum (Caprifoliaceae). Ecology 71: 1289-1298.
Schupp, E. W. 1988. Factors affecting post-dispersal seed survival in a tropical forest.
Skeate, S. T. 1987. Interactions between birds and fruits in a northern Fflorida hammock
community. Ecology 68: 297-309.
Stanley M. C., and A. Lill. 2002. Importance of seed ingestion to an avian frugivore: An
experimental approach to fruit choice based on seed load. Auk 119:175-184.
Stevenson, H. M., and B. H. Anderson. 1994. The Birdlife of Florida. University Press
Stiles, E. W. 1980. Patterns of fruit presentation and seed dispersal in bird-disseminated
woody plants in the eastern deciduous forest. The American Naturalist 116:670-
Travaset, A., M. F. Willson, and J.C. Gaither, Jr. 1995. Avoidance by birds of insect-
infested fruits of Vaccinium ovalifolium. Oikos 73:381-386.
van der Pijl, L.. 1972. Principles of Dispersal in Higher Plants. Springer-Verlag, Berlin,
Wheelwright, N. T. 1985. Fruit size, gape width, and the diets of fruit-eating birds.
Ecology 66: 808-818.
White, D. W. 1989. North American bird-dispersed fruits: ecological and adaptive
significance of nutritional and structural traits. Ph.D. Thesis. Rutgers University.
White, D. W., and E. W. Stiles. 1992. Bird dispersal of fruits of species introduced into
eastern North America. Canadian Journal of botany 70: 1689-1696.
Witmer, M. C., and P. J. Van Soest. 1998. Contrasting digestive strategies of fruit-eating
birds. Functional Ecology 12: 728-741.
Wilcove, D. S., D. Rothstein, J. Dubow, A. Phillips, and E. Losos. 1998. Quantifying
threats to imperiled species in the United States. Bioscience 48:607-615.
Wunderlin, R. P. 1982. Guide to the Vascular Plants of Florida. University Press of
Michael Meisenburg was born in North Tonawanda, New York, and graduated
from Starpoint Central High School in 1982. After nearly 6 years in the Air Force, he
moved to Port Orange, Florida, and soon landed a job spraying aquatic herbicides and
algaecides. It was during this influential period in his life that he learned the joy that
comes from killing invasive plants. He graduated from UF in 1999 with his B.S. in
wildlife ecology and conservation, and currently works as a biologist for the Center for
Aquatic and Invasive Plants at UF conducting research on killing invasive plants.
Michael is chair of the Control and Evaluations Committee of the Florida Exotic Pest
Plant Council and president of Alachua Audubon Society. He has been married for 10
years to Vasiliki, and they have many animals together. Michael is an avid fisherman,
gardener, birder, naturalist, and invasive plant killer.