Download Seed Dispersal and Frugivory in Tropical Ecosystems

INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
SEED DISPERSAL AND FRUGIVORY IN TROPICAL
ECOSYSTEMS
K. E. Stoner and M. Henry
Centro de Investigaciones en Ecosistemas, Universidad Nacional Autónoma de México
Keywords: Bats, birds, frugivory, foraging behavior, plant diversity, rodents, seed
germination and establishment, seed limitation, seed shadow, zoochory
Contents
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
1. Introduction
2. The seed dispersal cycle
3. Why seed dispersal matters?
4. Fruit syndromes and fruit attributes attracting dispersers
5. Field methods for studying seed dispersal
6. Concepts and statistical approaches
7. Seed Dispersal and Human Development
Acknowledgements
Glossary
Bibliography
Biographical Sketches
Summary
U
This chapter first presents an overview of the seed dispersal cycle and then focuses
specifically on seed dispersal. The distinctions between primary and secondary seed
dispersal and factors that affect the efficiency and effectiveness of seed dispersal are
presented.
How mutualistic interactions between frugivores and fruits may be
recognized and definitions of fruit syndromes are described. Several methods for the
study of seed dispersal are presented ranging from simple descriptive studies
documenting seed dispersers, to more modern techniques that use genetic markers.
Finally, this chapter concludes by discussing the possible consequences of the
elimination of vertebrate seed dispersers from tropical forests.
1. Introduction
1.1. Definition of Seed Dispersal
Seed dispersal consists of the removal and deposition of seeds away from parent plants.
Plants have evolved several different mechanisms of seed dispersal to achieve dispersal
from the mother plant including anemochory (wind-dispersed), hydrochory (waterdispersed) barochory (gravity-dispersed), autochory (self-dispersal by explosion), and
zoochory (animal-dispersed). Zoochory may be further divided into exozoochory,
where the seeds are attached to the outside of the animal’s body or endozoochory,
when the seeds are swallowed and ultimately dispersed via defecation.
Of these seed dispersal mechanisms, endozoochory is the most important in tropical
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
ecosystems. An estimated 51 to 98 percent of canopy and sub-canopy trees in
Neotropical forests are vertebrate-dispersed; a similar estimate is found for the
Paleotropics, with estimates ranging from 46 to 80 percent. Mammals and birds are the
most important vertebrate groups responsible for seed dispersal in tropical regions. Bats
and primates contain the most frugivorous species among mammals and are recognized
as key taxonomic groups for seed dispersal in tropical forests. Other mammal groups
that are documented to be primary dispersers as well include species from Carnivora,
Rodentia, Proboscoidea, Perissodactyla and Artiodactyla.
1.2. Organization of this Review
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
In this review, we will first explain the seed dispersal cycle and highlight where seed
dispersal fits into this scheme and patterns of primary and secondary dispersal in
tropical regions. We will then discuss the importance of seed dispersal and mutualistic
relationship between plants and their animal-dispersers. This section is followed by a
description of different techniques used to study the phenomena of seed dispersal in
tropical regions with case studies used to represent different methods. We close with a
section that discusses the implications on seed dispersal of human alterations of
vertebrate populations and the consequences this may have for the long-term
maintenance of tropical forest diversity.
2. The Seed Dispersal Cycle
A distinction should be made between the process of seed dispersal (i.e. seeds are
moved from parent plants) and the seed dispersal cycle which consist of a series of
events beginning with seeds produced by parent plants and ending with the adult plant
composition found within the forest (Fig. 1). Seed dispersal is only one step in the seed
dispersal cycle which functions as a “demographic bridge” linking the end of the adult
plant reproductive cycle with the establishment of their offspring.
U
Seed dispersal is first affected by the adult composition of tree species and their fruit
availability. Fruit availability of any one species at a particular time of year depends
upon successful pollination, and the flower and fruiting phenology of the species. Fruit
removal by frugivores depends on many factors including diet choice, nutrition,
competition with other frugivores, visitation and fruit removal rate, and frugivore
behavior. Once seed uptake has occurred, actual seed dispersal depends on seed
passage time in gut, disperser movements, disperser behaviors, scatter hoarding and
seed caching. Seed rain or deposition depends on seed handling time by animals and
seed predation. Germination depends on the availability of light, water and gaps, the
existing seed bank, pathogens and fungi. Germination success will ultimately determine
the seedling distribution within the forest. Seedlings will need to survive the same
threats as seeds undergoing germination, in addition to density-dependent mortality in
order to be recruited into the next generation of saplings and the next generation of adult
plants (Figure 1).
©Encyclopedia of Life Support Systems (EOLSS)
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
Figure 1: Seed dispersal loop figure from Wang and Smith 2002
2.1. Primary Dispersal
U
Primary dispersal consists of the removal of a fruit from a tree and the deposition of
seeds from this fruit in a particular area. In addition to factors that affect frugivore
choice that determine if a fruit is consumed or not, once the fruit is swallowed a series
of factors affect primary seed dispersal and the ultimate fate of the consumed seeds.
These factors include body size, digestive strategies, ranging behavior, and defecation
of the frugivores. Larger animals can swallow bigger seeds than smaller ones. The
time required for seeds to pass through the digestive tract affects the fate of swallowed
seeds in that seeds that spend more time in the digestive tract are generally deposited at
greater distances from the mother plant and frequently consist of one species. Animals
with short gut retention times more often deposit seeds closer to the mother plant and
usually create mixed species seed shadows. Furthermore, the action that occurs in the
gut may promote seed germination success (higher velocity of germination or greater
number germinated) or acids in the stomach may actually destroy the seeds making
them nonviable. The distance that different animals move and their foraging pattern also
affect the fate of primary dispersal. In general, animals that travel widely in a day will
deposit seeds over a greater area than mammals that intensively exploit a smaller day
range moving shorter distances. Finally, defecation patterns also affect seed dispersal,
in that they may be deposited in high-density clumps, singly, or in low density clumps.
2.2. Secondary Dispersal
Secondary dispersal consists of the removal of seeds once they have been deposited by
their primary disperser. Spit seeds and dropped, wasted fruit may be exploited by other
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
seed dispersers such as, rodents, deer, and peccaries who may then serve either as
secondary dispersal agents or seed predators. Some invertebrates like ants and dung
beetles may also contribute to secondary dispersal of small seeds, but their effect on
final seed germination and establishment is poorly known compared to that of
mammals. The distance to which mammals secondarily remove seeds relative to
parental trees varies significantly among species. For example, peccaries and tapirs
perform long-distance seed movement (possibly up to 20 km), while small bodied
mammals such as rodents move seeds much shorter distances (5-100 m).
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Most rodents are granivores that prey on seeds; however, the two main foraging
behaviors employed by many rodents, scatter-hoarding and larder-hoarding have
different effects on secondary dispersal. In larder-hoarding seeds are usually buried
more deeply and in fewer locations, while in scatter-hoarding seeds are buried less
deeply and in several locations. Therefore, scatter hoarding often results in effective
secondary dispersal and may contribute to seed survival when one of the following
occurs: (1) the rodent forgets the location of a cache, (2) the rodent has a
superabundance of seeds in several caches, and thus does not return to all of them, or (3)
the rodent suffers mortality and fails to return to a cache. In contrast, larder hoarding
rarely results in effective secondary dispersal because of the low probability of
germination of seeds buried deeply.
3. Why Seed Dispersal Matters?
3.1. Seed Dispersal and Plant Diversity
U
The reward to the animals is quite obvious, food, but what do the plants get out of this
mutualistic relationship? Seeds removed from parent trees escape from densitydependent mortality under their crowns, may colonize open habitats (succession),
and/or experience directed dispersal to appropriate microsites, which allows escape
from predators or enhances the establishment of seedlings. Although these factors are
not mutually exclusive, the Janzen-Connell escape hypothesis is well supported (see
Section 6.1). Recent work in Panama has unequivocally shown that density-dependent
mortality ⎯ and its subsequent effect on seedling recruitment ⎯increases tropical plant
species diversity.
In the tropics, seed dispersal by animals (mostly bats and birds) is also a critical
component of plant regeneration and diversity restoration processes in disturbed areas.
Flying vertebrates are attracted by isolated trees in pastures and open habitats that they
use as perch or feeding roost. This locally fosters seed deposition. Light-demanding
pioneer plant species, adapted to the dry conditions of open habitats, will first emerge
from this seed potential. They will produce a more shaded environment where shadetolerant species will in turn develop and eventually replace pioneer plants.
3.2. Efficiency, Effectiveness, Quantity and Quality
Not all fruits that are removed and ingested by animals result in effective seed dispersal.
The effectiveness of seed dispersal depends on both the quantity of seeds removed and
the quality of dispersal. The main factors that affect the quality of dispersal are the
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
distance moved from the parental tree, and the particular microsite the seed is deposited
in. An efficient disperser (that disperses far or provides a well scattered seed deposition)
may not be an effective disperser in that it does not deposit seeds in a suitable manner
(e.g. clumped) or in a suitable place (e.g. in forest understory some seeds do not
germinate because of lack of light). Some animals regularly result in poor seed dispersal
because of dispersal limitation. Animals contribute to this phenomenon when they
disperse seeds to feeding roosts (bats and birds), resting or sleeping sites (monkeys),
latrines (rhinos and tapirs), caches (rodents) or fruiting trees (monkeys). In some
circumstances, dispersal limitation is considered as a positive feature for the
maintenance of plant diversity (see Section 6.4 for more information).
4. Fruit Syndromes and Fruit Attributes Attracting Disperses
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
The idea that particular groups of fruit characteristics are more attractive to particular
groups of seed dispersers was originally suggested by van der Pijil in 1969 and defined
as fruit syndromes. For example bird-dispersed fruits are often brightly colored and
small, while bat-dispersed fruits are similarly small, they are more often dull or dark in
color. Large fleshy fruits often with reddish or yellow colors, sometimes with husks, or
harder coverings that require manipulation, have been classified as primate-dispersed.
The idea of fruit syndromes has been questioned in recent years. In particular, when
phylogenetic effects are removed, the only fruit characteristic that may be significantly
related to dispersal type is fruit dimension.
4.1. When is Seed Dispersal a Mutualistic Interaction?
U
Zoochory is commonly viewed as a mutualistic interaction whereby animals carry away
seeds and are rewarded by energy in the form of edible fleshy fruits offered by plants. A
true mutualistic interaction only exists between a particular seed species and its
disperser when both receive a benefit from the interaction. Although many tropical
plant species benefit from their seeds being dispersed by vertebrates, no examples exist
in which one plant species is strictly dependant on only one animal disperser, indeed
most animal-dispersed seeds are attractive to many different dispersers (see Section 7
for more information).
4.2. The Dodo Bird and the Tambalacoque Tree: An Example of an Obligate
Mutualism?
One of the most famous stories of what happens to seed dispersal when the disperser
goes extinct comes from the Tambalacoque tree (Sideroxylon grandiflorum) and the
terrestrial Dodo bird (Raphus cucullatus), two species endemic to Mauritius island. The
Dodo went extinct, supposedly in 1681, as sailors and new settlers extensively hunted
this large easy to catch prey for food. Since then, the Tambalacoque tree failed to
regenerate efficiently and eventually underwent a severe population decline. In 1977, an
ornithologist hypothesized that the decline of the Tambalacoque tree was a direct
consequence of the disappearance of its main seed disperser, the Dodo bird. He assumed
that seeds of the Tambalacoque, due to their thick endocarp, needed the abrasive action
of Dodo’s gut to germinate. This elegant example of an “obligate mutualism”, whereby
a particular tree species is entirely dependant upon one effective animal disperser so that
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
its seeds may germinate, has long lasted in the scientific literature and the local
Mauritius folklore. The Dodo history illustrates how the action of humans on a given
animal may indirectly exert detrimental effects on the associated plant species.
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Nevertheless, recent studies, have seriously questioned this scenario, and have identified
more probable causes for the decline of Tambalacoque trees. The early studies have
been strongly criticized, not only for basing overstating conclusions on an invalid
germination experiment, but also for ignoring other potential native seed dispersers
(tortoises and parrots) and overall for underestimating other factors that could have
influenced the regeneration of Tambalacoque trees. The extreme deforestation and
development of human populations on the island, the introduction of non-native plant
competitors (American elm and Chestnut trees) and seed predators (deer, pigs,
monkeys) are the most obvious reasons for the decline of Tambalacoque trees.
Nowadays, the native Mauritius forest is restricted to 1% of its initial size, and the few
last Tambalacoque individuals are hardly preserved in this remnant island ecosystem.
5. Field Methods for Studying Seed Dispersal
5.1. Who Disperses What? Removal Studies and Faecal Analyses
The question “who disperses what” cannot be reduced to “who eats what”. Some
frugivores that strictly consume pulp feed on fruits without removing and dispersing
seeds away from parent plants, but rather drop seeds directly below them. This is the
case of some birds that only pick large-seeded fleshy fruits. Accurate behavioral
observations are required to distinguish frugivores that remove and carry fruits and
seeds away from parent plants from those that leave seeds directly below parent plants.
U
Observing foraging activity of frugivores is the basic method to identify potential
dispersers. This method is known as “removal survey”. Primatologists, for instance,
usually follow primates to report the fruiting trees they visit and establish their diet and
the plants they are likely to disperse. Conversely plant ecologists document the set of
frugivorous species that visit a particular fruiting plant. A removal survey of a fruiting
fig tree in Peru revealed that fig seeds could potentially be dispersed by up to 44 diurnal
frugivores (40 bird species and 4 primate species). The use of photographic traps and
video recording (including infrared recording for nocturnal dispersers) is sometimes
required when the presence of observers keeps dispersers away.
The indirect equivalent of removal surveys are fecal samples. Seeds ingested by
dispersers are later defecated. Examining feces of animals allows one to document
ingested seeds and determine potential dispersers. Fecal samples may be directly
collected from captured animals or collected in feeding and resting roosts, in latrines, or
any other place where the animal of interest spends time. Probably one of the most
original fecal sampling strategies employed are the fecal collectors that have been used
for terrestrial tortoises. These collectors consist of a small screen bucket that is attached
to the bottom rear portion of the shell of tortoises, allowing fecal material released to be
collected all day long.
5.2. What is Dispersed Where? Monitoring Seed Fate
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
In addition to the identification of dispersers, researchers usually want to establish the
fate of seeds. For example, at what distance from their origin are seeds deposited and do
these seeds succeed in germinating and becoming established as seedlings? Various
direct and indirect field methods are used to track seed fate at all stages of the seed
dispersal loop; nevertheless, great variability in technical accuracy exists among seed
types. In particular, it is far easier to track large, conspicuous seeds typically transported
and deposited several meters or tens of meters away from parent plants, than tiny seeds
(0.5mm to 1cm in length only) ingested by dispersers and later defecated hundreds of
meters away, or even further.
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Beyond their dispersal mode, large and small seeds differ in many life history traits,
which in turn affect research methods used to monitor seed fate. For example, small
seeds generally have a very low probability of germination and establishment, while
large seeds do much better. Conversely, the amount of seeds produced by small-seeded
plants often exceeds by orders of magnitude that produced by large-seeded plants.
Because of their size, large seeds are the most easy to monitor as they can be
individually marked for later retrieval. Thus, researchers studying large seeds may
attempt to predict the fate of individual seeds, while those working on small seeds often
attempt to depict general trends in dispersal probabilities of a species. Examples of
methods to work on large-seeded and small-seeded species are given below.
U
Marking methods for large-seeded species mostly have been used to document the fate
of seeds fallen to the ground and cached by rodents. Seed caching behavior reported
from a range of rodents in temperate and tropical regions is also called “scatterhoarding”. Fixed marking techniques have been used to study this phenomenon. Seeds
are attached to a thread that allows them to be followed through their dispersal pathway
from a fixed point. A thread bobbin is used, either attached to the fixed point or inserted
inside the seed. This method has been successfully used to follow seeds up to 50-200m
from their origin. Nevertheless, some dispersers tend to abandon seeds when the thread
gets entangled and hinders effective transportation. This methodological problem
prompted the use of a new free marking technique. This consists of using a labeling
device that leaves seeds unattached to any point and allows them to be freely
transported. Seeds that can be visually retrieved may be marked by notches on the
external envelop, color marks, flags or any other type of identification. Seeds that are
buried by scatter-hoarding animals may be fitted with a 30-200 cm long nylon thread
with a bright color or labeled with a colored tag. Seeds can be relocated from the thread
even though they are cached or buried. More sophisticated methods exist, including the
use of radioisotopes which can be detected from a short (<50 cm) range using a
portable Geiger counter. Miniature radio-transmitters also can be placed inside a seed to
monitor its fate. Transmitters weighing <1g can be detected by a radio-receptor from
several hundreds of meters away. This is probably the most efficient seed monitoring
method currently available, but it is rather expensive and can only be used on large
seeds.
The basic method to document the dispersal of small seeds ingested and defecated by
animals is the use of seed traps. Seed traps are collectors that typically take the form of
1m² plastic or cloth sheets stretched horizontally, round funnels with a collecting
pocket, or recipients of any shape and size placed on or near the forest floor. The aim of
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
seed traps is to measure seed rain, i.e. to sample seeds that are deposited in a given site.
Instead of giving detailed insights about where seeds are dispersed, it permits to
measure how many seeds arrive in a given site. This approach is more efficient for small
seeds that can be very abundant (e.g. a deposition rate of 130 seeds per 1 m² per month
for bat-dispersed seeds in a mature Neotropical rainforest). Using this information, it
has been estimated that 50 traps of 1 m² set during several months are enough to
document the seed rain of the most common species in this habitat. Seed traps are on the
contrary less efficient to document the dispersal of large seeds as these are much less
abundant and many species have supra-annual fruiting phenology. Only large-scale
surveys involving more than 200 traps for periods of several years can provide reliable
information on the dispersal of large seeds.
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Similarly, sampling the soil seed bank is usually more conclusive when applied to
smaller, more abundant seeds. The soil seed bank refers to the reserve of ungerminated
seeds accumulated in the soil. This reserve can be estimated by documenting variations
of seed abundances from soil samples at different depths. This method can be used to
establish past successions of plant-dispersers assemblages for up to several thousands of
years (depending on the depth of soil samples). The limitation of seed traps and soil
samples is that it is impossible to determine with precision which animal species
dispersed which seeds. It only gives indications on the spatial variations of seed rain.
5.3. Germination Experiments
U
With rare exceptions, the process of seed dispersal by animals implies the consumption
of edible pulp and the regurgitation or defecation of seeds. Seeds normally withstand
this treatment imposed by effective dispersers. Pulp removal and gut treatment is even
thought to favor, or be a prerequisite for successful seed germination for some species.
Seeds may however, be damaged or crushed by teeth or altered by the acid environment
of the gut when processed by less effective dispersers. Therefore, documenting the
effects of teeth and gut treatment on germinability is highly advisable when one needs
to identify the effective dispersers of particular plant species. This is achieved through
germination experiments.
The standard procedure that is used in most germination experiments consists of the
following: (i) collect seeds dropped or defecated by dispersers, either in the field or in
captivity, (ii) place seeds to germinate on a humid substrate with conditions of light,
temperature and humidity close to those estimated to be optimal in the field, (iii)
compare the observed germination pattern to that obtained from a set control seeds that
were not handled or ingested by dispersers. The germination pattern can be described by
two main parameters (Figure 2): germinability (percentage of seeds that germinate) and
germination rate or velocity (time lapsed to reach, for instance, 50% of all
germinations). Germination protocols may be improved by testing additional treatments.
Two series of control seeds may be tested, namely with and without pulp (pulp removed
by the observer). Ingested seeds also may be tested with and without fecal material
surrounding them. Making germination experiments closer to natural conditions make
results more tractable. It is therefore recommended to repeat experiments not only in
controlled laboratory conditions, but also in the natural field environment.
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Figure 2. Example of germination curves of ingested and control seeds
More that 350 germination experiments have been recently reviewed to document the
general effects of gut treatments across a wide range of plants and disperser species.
Seed ingestion generally increases substantially germinability and germination rate, but
this positive effect is up to four times greater in seeds dispersed by birds and bats than
by reptiles and non-flying mammals. This difference is partly attributed to shorter gut
passage time in flying animals. Flying is energetically costly, and birds and bats have
developed physiological adaptations to speed up the digestive process and rapidly
empty their gut load. Consequently, seeds are exposed to the action of digestive acids
for shorter periods of time (5 to 30 min in some Neotropical bats). Seed size also
influences gut effect. Larger seeds are more likely to withstand the gut treatment than
smaller seeds as the latter tend to be retained longer in the gut, and display a thinner
coat layer for protection against external abrasion. Finally, dry seeds (seeds devoid of
pulp) generally undergo negative gut treatments. Unlike fleshy seeds that evolved with
the action of frugivores that eat pulp, most dry seeds appeared not to have developed
adaptations to survive gut treatments.
U
6. Concepts and Statistical Approaches
Many crucial issues have been or are currently debated in the field of plant-disperser
interactions. Several of these issues are presented herein, as well as their respective
methods for research. The first important distinction to clarify when assessing statistical
approaches used for evaluating plant-disperser interactions is to determine if we are
addressing this question from the animal’s point of view or from the plant’s point of
view. In the former case, researchers will think in terms of “fruit resource availability”
and will study how the pattern of fruit production will affect the distribution and activity
of animals. In the latter case, biologists will think in terms of “disperser availability”
and elucidate the effects of animals’ activity patterns on the patterns of seed dispersal
and the regeneration process. Integrative studies depicting the overall plant-animal
interaction process are scarce.
6.1. Seed Shadow and the Janzen-Connel Hypothesis
The seed shadow refers to the pattern of spatial deposition of seeds relative to parent
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
plants. A thorough knowledge of the seed shadow is often a prerequisite to the
understanding of plant-disperser interactions. A basic characteristic of this pattern is the
sharp decline in the density of seed deposition with increasing distance to parent plants.
The majority of seed species are dispersed over short distances, leading to intricate
mechanisms of local competition and predation pressures on seeds in the recruitment
and regeneration processes of plants. In the early 1970’s, Janzen and Connel developed
independently a theory on the relation between seed shadow and seed survival, and its
implications for patterns of plant diversity. The essence of the Janzen-Connel
hypothesis is that the probability of seed survival is inversely related to seed density. It
indeed states that the high seed and seedling density around parent plants enhances the
activity of predators, pathogens and herbivores, as well as the intraspecific competition
during the process of seedling establishment. As a consequence, it is thought that
evolution has favored the development of seed dispersal mechanisms in order to
“escape” the locally high seed mortality around parent plants.
U
The Janzen-Connel hypothesis has influenced generations of plant ecologists, and is still
the main focus of many studies on seed dispersal. It is recognized to be the theory that
best accounts for the role of animal dispersers. It has however some limitations and
cannot be applied to all systems. For example, small shrubby plants producing bird- or
bat-fruits, have such low seed production and dispersal distances are so large that it
makes no sense to attempt measuring any Janzen-Connel effect around parent plants.
The Janzen-Connel effect mostly concerns plants with large-seeded crops and/or shortdistance dispersal. Second, the seed shadow generated by many dispersers, although
conforming to the decline of seed density with distance, may display features that cancel
the Janzen-Connel effect. For example, dispersers may deposit seeds in habitats
unsuitable for germination, or accumulate them in feeding roosts where they are also
subject to pathogens. Such a “contagious” dispersal pattern does not favor seed survival.
On the contrary, a random deposition pattern (i.e. seeds deposited in a random way
rather than concentrated at particular points) is generally considered a desirable feature.
Finally, within the forest, the seed shadows produced by several trees may overlap,
rendering it difficult to really establish the regeneration processes and assess the role of
dispersers in helping seeds to escape from local mortality.
6.2. Modeling Seed Shadows
When the seed shadow cannot be directly monitored in the field, indirect mathematical
modeling may be used. The modeling approach is a projection of the expected seed
shadow given the foraging behavior of dispersers. This method requires a thorough
knowledge of the way dispersers search for food resources, choose and process fruits,
and move within landscapes among their foraging areas and their feeding and sleeping
roosts. First and foremost, all details of the seed production pattern of the target plant
species are needed, including the location of productive individuals, the duration and
size of fruit crops, the fruiting rate, and the number of seeds per fruit. Subsequently, the
activity of dispersers should be described using: abundance, frequency and duration of
visits to the plant, feeding rate and type of fruit processing (ingest or regurgitate seeds),
and post-feeding movements with the seeds (distance and preferred habitats). In the case
of dispersers that ingest seeds, additional parameters are needed such as gut load
(number of seeds swallowed at a time), gut retention time (time lapse between seed
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
ingestion and defecation), and the conditions of defecation (in flight, at perch, in
latrines, at feeding roosts, or random). Integrating all of this information will help in
producing a model of seed dispersal and seed shadow in simple systems. In cases where
several types of dispersers use the same fruit resource, this modeling should be repeated
separately for each of the main species of disperser. However, this results in more
variability in the accuracy of the final picture of seed shadow. In any case, it is
recommended to assess the predictive power of the model using independent seed rain
surveys in the field.
6.3. Spatial Patterns of Seed Deposition and Seedling Establishment
U
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
In situations where parent plants are abundant over a study area, studies of seed
shadows are replaced by more global and powerful analyses aiming at establishing the
links between the spatial distribution of parent plants and that of seeds and seedlings.
Such analyses are called “spatial point pattern analyses”, because they identify all
studied elements (plants, seeds, seedlings) to points, and establish the spatial
relationships between theses series of points. More specifically, the central question
addressed by these analyses is whether the deposition of seeds or the establishment of
seedlings is spatially independent from the location of parent plants. If no spatial
relationship is observed, seeds and seedlings will tend to be randomly distributed. If on
the contrary an attractive (or repulsive) effect links seed and seedling distribution to
parent plant distribution, the seeds and seedlings will be aggregated around (or
separated from) parent plants (Figure 3).
Figure 3. Example of spatial point pattern analysis.
Comparing the spatial deposition of seeds to the spatial establishment of seedlings helps
in assessing the effectiveness of dispersers and to determine if dispersal limitation is
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
occurring. If a repulsive effect results in seedlings establishing away from parent plants
while seeds are actually mostly deposited around parent plants, then one might conclude
that seed dispersers do not optimize the regeneration of plants and that dispersal
limitation is occurring.
6.4. Dispersal Limitation: “Winning By Forfeit”
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
A new field of investigation in ecology of seed dispersal has emerged in the mid-1990s,
under the name of recruitment limitation. The fact that plants are usually limited in all
the stages of their recruitment (or regeneration) process is thought to have important
implications in the maintenance of plant diversity. A given plant may succeed to
establish in a given site simply because superior competitors have failed to do it before.
Such a chance effect, whereby plants “win by forfeit”, permits the possibility of poor
competitors, and very rare species to be maintained in the ecosystem together with
superior competitors. This mechanism is theoretically enhanced in systems, like tropical
forests, where plants are recruitment-limited, in that they fail to reach and establish in
all suitable places.
Recruitment limitation has three main components: seed source limitations (insufficient
number of seeds), dispersal limitation (inefficient dispersal), and establishment
limitation (failure of seeds and seedlings to survive and develop). Critical issues
currently debated are the assessment of the relative importance of these three limiting
steps in the recruitment processes. Dispersal limitation, due to short dispersal distances
or spatially contagious seed deposition, is often regarded as the main cause of
recruitment limitation. In particular, large seeded plants dispersed by monkeys or
terrestrial vertebrates are much more dispersal-limited than small-seeded plants ingested
and dispersed by flying vertebrates (birds and bats). The latter plants produce great
numbers of seeds that are defecated in a variety of sites. They may however suffer from
greater establishment limitation, as the probability that small seeds germinate and
develop is low compared to larger seeds.
U
All steps of recruitment limitation can be monitored by surveying seed arrivals and
seedling establishment in small plots (or seed traps) typically 0.25 to 1m² in size.
Following the concept of recruitment limitation, it must be assumed that each plot is a
suitable microsite for seedling establishment. First, seed source limitation is measured
as the proportion of microsites that would have remained empty of seeds if no more
than one seed had been deposited per microsite. This index ranges from 0% (more seeds
observed than microsites available) to 100% (no seeds observed). Dispersal limitation is
then calculated as the ratio between the number of microsites that effectively remained
empty, and the number of microsites that would have remained empty with a one-seedper-site distribution. Finally, establishment limitation is measured as the proportion of
microsites that received seeds where no seedling established.
Although dispersal limitation is supposed to favor plant diversity in undisturbed natural
systems, it may also have detrimental repercussions on the recruitment processes of
plants if it is extended beyond a certain threshold. In disturbed areas deserted by
dispersers, dispersal limitation would theoretically reach its maximum, and the normal
dispersion processes would be disrupted. This dispersal failure would preclude the
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
reproduction and re-colonization of plants that require dispersers. The links between
dispersal limitation and the maintenance of tropical plant diversity is still being
investigated, and there is no doubt that animal dispersers play a key role in this dynamic
process.
6.5. Seed Sowing Experiments and Disperser Exclusion Experiments
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Experimental setups where one or several stages of seed dispersal are controlled
constitute a powerful tool to investigate all components of recruitment limitation.
Some studies have monitored the consequences of massive additions of seeds in plots to
assess how seed source limitation influences plant recruitment. In half of these cases,
plants appeared seed limited because adding more seeds increased population size.
However, with excessive seed additions, plants enter in intraspecific competition and
cannot further increase their density. Similarly, excluding or introducing dispersers from
study plots, using fences or enclosures, helps understand their contribution to plant
recruitment. In an experiment consisting in depleting the activity of bats in a small
forest portion in French Guiana, bat-dispersed seeds declined sharply in abundance, and
became 30% to 75% less diverse.
6.6. Mutualistic Networks: Assessing the Stability of Seed Dispersal Systems
U
The increasing number of long-term, locally intensive, surveys of plant-animal
interactions stimulated the development of methods integrating all organisms known to
interact in a single analysis. Mutualistic networks analysis is such an example. Derived
from topology theories, network analyses build complex diagrams linking all plants and
dispersers that interact, and estimate the consequence of extinctions of some animal or
plant species on the overall cohesion and stability of the network.
Figure 4. Example of a mutualistic seed dispersal network involving birds and plants in
a Neotropical forest
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
To build a seed dispersal mutualistic network for a given study area, the first step
consists of making a census of all seed dispersal interactions involving on one hand
plants, and on the other hand animals. Each plant and animal species is graphically
represented by a dot. Each plant-animal interaction is then depicted by a line linking the
corresponding dots (see Figure 4 for an example). On the network graph, a generalist
frugivore will be linked to many plants, and a specialist to few plants. Similarly, plants
attracting many dispersers will be graphically connected to many animal dots.
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Such plant-animal networks can theoretically adopt one of three kinds of structure. First,
networks can be random, in that interactions link plants and animals without any
particular pattern. On the contrary, networks can be compartmentalized, i.e. organized
into independent subsets of species among which few or no interactions exist. Recent
reviews of plant-animal networks, focusing on seed dispersal or on pollination, have
revealed that the third type of network, termed “nested” is actually the most common
system in nature. In a nested network, a core of generalist animals interact with a core of
plants, altogether accounting for the bulk of all interactions in the system. This central
interaction core is surrounded by additional, peripheral, specialist interactions. It is
thought that the interaction core in nested networks drives the evolution of the whole
system and promotes the maintenance of biodiversity.
U
Mutualistic networks are an interesting approach to the study of seed dispersal because
they allow the simulation of effects of successive extinctions on the whole system. The
disappearance of some plant species may foster the disappearance of the frugivores they
are associated with, which may in turn threaten the maintenance of other plants that they
disperse, and so on. In highly disturbed systems, extinctions in chain may fragment the
interaction network and provoke its collapse or its simplification into a handful of
interconnected species. The “stability” or “robustness” of a network against species
extinctions is largely dependant on its structure. Nested networks, because of their tight
interaction core involving multiple connections, are theoretically more stable than other
types of network. Mathematical methods assessing nestedness of networks are expected
to become attractive tools in management and conservation biology. In addition to
providing readily a measure of stability of a whole interaction system, they also permit
us to focus conservation efforts on the plant-animal subset that is most critical for the
maintenance of a particular target, or threatened species (animal or plant).
6.7. Genetic Techniques
Genetic techniques provide the top of the art in ecology of plant-animal interactions and
are becoming more common every day. Using genetics, one can exactly determine the
maternal plant from which a particular seed originated and was dispersed. This leads to
powerful surveys in which one can accurately determine the fate of each individual seed
(and seedling) from its origin to its final deposition site. Establishing kinship among
established plants (both seedlings and/or adults) allows one to make backward
deductions about seed movements many years after seeds were effectively dispersed.
A first step consists in genotyping, i.e. characterizing the genetic identity of, all
potential maternal plants of the target species in the study area. As this type of genetic
census can be a tedious task, it is normally restricted to trees or plants that are not too
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
abundant. DNA is extracted from leaves, and particular genetic sequences are identified.
A similar procedure can be performed on seeds and seedlings. In most fleshy fruit
plants, the DNA is extracted from the seed endocarp, which is a tissue initially produced
by the maternal plant and hence contains the genetic identity of the maternal plant. This
is the simplest way to match mothers and offspring. If one needs to study kinship among
adult plants, further DNA analyses are needed to find which individuals are the most
probable ancestors of younger individuals.
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Genotyping plants in a study area not only helps investigate the usual questions related
to seed shadow, recruitment limitation and spatial pattern, but also lays the bases of a
variety of more original and complex issues. One of them is the effect of landscape
configuration on the seed rain and genetic structure of plant populations. A landscape is
often composed of various habitat types and geologic formations that are not all suitable
for dispersers. Some constitute natural barriers that impede the movements of frugivores
and the flow of seeds they disperse. Consequently, some plant populations may be
subdivided into sub-populations among which exchanges of genetic material via seed
dispersal is constrained. Such fractioned populations are more prone to suffer from
disturbances, especially in rare species. Using genetics is the most efficient way to
identify deficient seed flows in these populations and adapt conservations strategies.
Conversely, by genotyping seeds collected from fecal samples, one can back-track the
recent foraging movements of dispersers among maternal fruiting plants. This allows
the indirect documentation of behavior of dispersers.
U
The use of genetics also has allowed the discovery of unexpected patterns like the
effects of long-distance seed dispersal events. With a frequency that is probably largely
underestimated, the dispersal of some seeds across distances much larger than usual, for
instance by birds, may lead by chance to the colonization of distant or remote areas by
some plant populations. These events are mostly driven by chance rather than by
predictable ecological mechanisms, and may greatly influence the dynamic of tropical
ecosystems. The accumulation of such examples forced biologists to reconsider the role
of chance or hazard in shaping the abundance and distribution of plant species at local,
regional and even continental scales. This chance effect is also termed “ecological
drift”. Seed dispersal by animals is more and more recognized as a mechanism subject
to strong ecological drift in any kind of ecosystem.
7. Seed Dispersal and Human Development
Many vertebrate animals in tropical regions are seriously threatened by human
development and the consequent deforestation and fragmentation that results. Hunting
and the illegal pet trade also seriously threaten many medium and large-sized
vertebrates in tropical areas. The end result will be forests which contain fewer
individuals and fewer species. How the elimination of large vertebrates may affect
tropical forest dynamics has only recently been explored. Evidence suggests that the
absence of large and medium-sized vertebrate dispersers will result in reduced seed
movement for animal-dispersed species with very large seeds, reduced seed predation
by granivorous vertebrates for large-seeded species, and a change in the composition of
the seedling and sapling cohorts. Most avenues of research suggest that the elimination
of vertebrate seed dispersers will result in changing tropical forest composition and
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
structure. The long-term consequences of this are unknown.
Acknowledgements
We thank M. Quesada for the invitation to contribute to this Encyclopedia, A. Valencia
y H. Ferreira for technical support ant the Centro de Investigaciones en Ecosistemas for
logistical support.
Glossary
Density-dependent
mortality:
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Dispersal
limitation:
increased mortality of seed and seedlings due to locally high
individual densities. Density-dependent mortality could be
related to effects of competition, attraction of predators and
pathogens.
failure of seeds to reach all suitable microsites for germination
and establishment due to inefficient seed dispersion. Dispersal
limitation occurs, for instance, when dispersers deposit seeds
over short distances, or in sites unsuitable for germination.
ecological mechanism whereby the effect of “chance”, i.e.
unpredictable random events, have a preponderant role in
shaping natural plant or animal communities. It is thought that
seed dispersal is subject to strong ecological drift in many
ecosystems, making plant regeneration processes diversity
maintenance very unpredictable at the local scale.
seed dispersal where seeds are swallowed and ultimately
dispersed via defecation.
seed dispersal where seeds are attached to the body of animals,
or transported externally.
Any animal that consumes fruit as a regular part of their diet.
Frugivores may disperse seeds when they swallow seeds whole
or spit them without damaging, or they may be seed predators,
when they chew and damage the seeds.
This argues that the mortality of seeds and seedlings will be
much higher around parents and other conspecifics than farther
away. The argument has been used to explain the evolution of
dispersal and the maintenance of species diversity.
in the seed dispersal jargon, a microsite suitable for seed
germination corresponds to the small-scale location where seeds
are deposited and where local environmental conditions are met
for these seeds to successfully develop. Microsites are plots
often arbitrarily defined by size 0.25- to 1-m², which
corresponds to the range within which neighboring saplings may
influence each others’ survival through competition for water,
light or nutrients.
plants of the same families have greater chance to produce fruits
sharing common characteristics (shape, size, color, chemical
composition, etc.) and attracting similar dispersers. This gives
the impression that some fruit characteristics are associated to
Ecological drift:
Endozoochory:
Exozoochory:
Frugivore:
U
Janzen-Connell
hypothesis:
Microsite:
Phylogenetic effect:
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
Radioisotope:
Recruitment
limitation:
Seed rain:
Seed shadow:
Soil seed bank:
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Zoochory:
particular groups of dispersers, while it may be a simple effect
of plant relatedness, also called Phylogenetic effect.
radioactive variant of a given chemical element, differing from
the current form by the number of neutrons in the nucleus.
failure of a plant species to establish in all suitable microsites. A
limitation in plant recruitment could result from insufficient
production of seeds (seed source limitation) or from inefficient
dispersal (seed dispersal limitation).
flow of seeds dispersed and deposited into a given site
the spatial distribution of seed deposition around parent plants.
the reservoir of viable seeds accumulated in the soil and that has
the potential to germinate and contribute to local processes of
vegetation regeneration.
seed dispersal by animals
Bibliography
Bascompte J. and Jordano P. (2006). The structure of plant-animal mutualistic networks. Ecological
networks (eds. M. Pascual and J. Dunne), 143-159. Oxford University Press. [This reviews the way plantanimal networks, including dispersal networks, are structured].
Fenner M. (2000) Seeds: The ecology of regeneration in plant communities. 2nd edition (ed. M. Fenner),
125–165. CAB, Wallingford, England. . [A comprehensive book reviewing different factors affecting
regeneration, including seed dispersal].
Forget P.-M., Lambert J.E., Hulme P.E. and Vander Wall S.B. (2005). Seed fate: Predation, dispersal,
and seedling establishment, 410 pp. Cambridge, MA, USA: CABI Publishers. [This book provides a
comprehensive review of several of the factors affecting seed dispersal].
Forget P.-M. and Vander Wall S. B. (2001). Scatter-hoarding rodents and marsupials: Convergent
evolution on diverging continents. Trends in Ecology & Evolution 16, 65–67. [This work shows how
different organisms contribute similar roles in seed dispersal in different tropical regions].
Gautier-Hion A., Duplantier J.-M., Quris R., Feer F., Sourd C., Decoux J.-P., Dubost G., Emmons L.,
Erard C., Hecketsweiler P., Moungazi A., Roussilhon C. and Thiollay J.-M. (1985). Fruit characters as a
basis of choice and seed dispersal in a tropical forest vertebrate community. Oecologia 65, 324–337. [A
review of frugivory and fruit syndromes in an African forest].
U
Herhey D.R. (2004). The widespread misconception that the Tambalacoque or Calvaria Tree absolutely
required the Dodo bird for its seeds to germinate. Plant Science Bulletin 50, 105–108. [A critical review
of the Dodo-Tambalacoque mutualism hypothesis].
Hurtt G. C. and Pacala S.W. (1995). The consequences of recruitment limitation: reconciling chance,
history and competitive differences between plants. Journal of Theoretical Biology 176, 1–12. [This work
shows the links between dispersal limitation and plant diversity].
Stoner K.E., Riba-Hernández P., Vulinec K. and Lambert J.E. (2007). The role of mammals in creating
and modifying seed shadows in tropical forests and some possible consequences of their elimination.
Biotropica 39, 316–327. [A comprehensive review of the role of mammals in seed dispersal in tropical
forests].
Stoner K.E., Vulinec K., Wright S.J. and Peres C.A. (2007). Hunting and plant community dynamics in
tropical forests: a synthesis and future directions. Biotropica 39, 385–392. [This work synthesizes
information about the impacts of loosing mammal seed dispersers in tropical forests].
Wang B.C. and Smith T.B. (2002). Closing the seed dispersal loop. Trends in Ecology & Evolution 17,
379–385. [A review of recent advances in fundamental and applied ecology of seed dispersal, including
recommendations for experimental designs and future directions].
©Encyclopedia of Life Support Systems (EOLSS)
INTERNATIONAL COMMISSION ON TROPICAL BIOLOGY AND NATURAL RESORCES - Seed Dispersal and Frugivory in
Tropical Ecosystems - K. E. Stoner and M. Henry
Biographical Sketches
Kathryn E. Stoner received the BS and MA degree from the University of Michigan and the PH.D. from
the University of Kansas in 1993. She worked as Co-Director of Palo Verde Biological Station for the
Organization for Tropical Studies in Costa Rica and worked teaching a variety of tropical biology field
courses in Costa Rica for several universities (Pennsylvania State University, Duke University, University
of Costa Rica). In 1998 she participated as the Invited Scientist during a field course in Costa Rica for
the Special Interest Group of the Smithsonian Institute. During the last 10 years she has worked as a
Researcher for the Centro de Investigaciones en Ecosistemas for the Universidad Nacional Autónoma de
México (UNAM). She has participated as a subject editor of Biotropica since 2005. In 2004 she received
the Sor Juana Ines de la Cruz award for outstanding excellence in research from UNAM. She has worked
on different themes within tropical ecology in Costa Rica, Mexico and Brazil. Her interests include the
effect of forest fragmentation on primates and bats and their consequences on forest regeneration
(including seed dispersal and pollination), the evolution of color vision in primates, and bat-plant
pollinator interactions.
U
N
E
SA
SC
M
O
PL
-E
O
E
LS
C
H
S
AP
TE
R
S
Mickaël Henry completed the BS degree through an agreement between the Université de Rennes
(France) and the Université du Québec à Rimouski (Canada), the MA degree from the Université de
Sherbrooke (Canada) and the Ph.D. from the Université Paris VI in 2005. His doctoral research focused
on the functional consequences of landscape disturbances on the diversity of bats and of the seeds they
disperse in tropical forest ecosystems. Since 2002, he has been involved in various research projects in
French Guiana, Panama and Mexico in collaboration with the Muséum National d'Histoire Naturelle
(Paris), the Centre National de la Recherche Scientifique (CNRS, France), the Universität Ulm
(Germany), and the Universidad Nacional Autonoma de Mexico (UNAM). He is currently engaged in
postdoctoral research at the Centro de Investigaciones de Ecosistemas (UNAM). His current research
focuses on the ecology of interactions between frugivorous and nectarivorous bats and their plant
resources.
©Encyclopedia of Life Support Systems (EOLSS)