A telemetric thread tag for tracking seed dispersal
by scatter-hoarding rodents
Ben T. Hirsch ? Roland Kays ? Patrick A. Jansen
Received: 20 October 2011 / Accepted: 17 April 2012
 Springer Science+Business Media B.V. 2012
Abstract The seeds of many tree species are
dispersed more than once, and this secondary seed
dispersal is believed to enhance seedling recruitment.
However, the effectiveness of secondary seed dis-
persal has rarely been assessed because it is difficult to
track seeds until they die or germinate. We describe a
new technique that uses thread tags attached to radio
transmitters (telemetric thread tags) to track long-
distance multistep seed dispersal by scatter-hoarding
rodents. These telemetric thread tags can be turned off
with a magnet and are reactivated when the seed
moves. This method allows for seed tracking with
minimal cache disturbance or distance bias, over long
time spans, multiple seed movements, and with few
effects on animal behavior. We used telemetric thread
tags to track seed dispersal of the palm tree Astrocar-
yum standleyanum in a Neotropical forest, and
achieved near-complete recovery of dispersed seeds
tracked over distances as far as 241 m. We were also
able to record the recovery time and fate of cached
seeds without disturbing caches. Neither the removal
rate nor the dispersal distance differed between seeds
with telemetric thread tags and thread-tagged seeds.
We conclude that telemetric thread tags can be used to
document secondary seed dispersal by scatter-hoard-
ing animals with unprecedented efficacy and preci-
sion. Given the size of these tags relative to the size of
seeds and their dispersers, this method is applicable to
the majority of tree species that are secondarily
dispersed by scatter-hoarding mammals.
Keywords Radio telemetry  Seed tag  Animal
dispersed seeds  Astrocaryum  Agouti
Introduction
The dispersal of plant seeds is a critically important
ecological process that remains poorly understood
because following a seed from its source to the point
of germination or death is notoriously difficult
B. T. Hirsch (&)
School of Environment and Natural Resources, Ohio State
University, 2021 Coffey Rd., Columbus, OH 43210, USA
e-mail: hirschb@si.edu
B. T. Hirsch  R. Kays  P. A. Jansen
Smithsonian Tropical Research Institute, Unit 9100,
Box 0948, DPO AA, 34002-9898 Panama City, Panama
R. Kays
Nature Research Center, North Carolina Museum
of Natural Sciences, 11 W. Jones Street, Raleigh,
NC 27601, USA
P. A. Jansen
Forest Ecology and Forest Management Group,
Wageningen University and Research Centre,
PO Box 47, 6700 AA Wageningen, The Netherlands
P. A. Jansen
Community and Conservation Ecology Group,
University of Groningen, PO Box 11103, 9700 CC
Groningen, The Netherlands
123
Plant Ecol
DOI 10.1007/s11258-012-0054-0
(Chambers and MacMahon 1994; Wang and Smith
2002; Bullock et al. 2006). In particular, secondary seed
dispersal and ultimate seed fate remain insufficiently
documented because existing tracking methods are
biased against longer dispersal distances and higher-
order movements. Many plant seeds are dispersed in a
sequence of two phases that involve different dispersal
agents, a phenomenon known as ?diplochory? or ?two-
stage seed dispersal?(Wenny 1999; Vander Wall and
Longland 2004; Vander Wall et al. 2005). During the
first phase, primary dispersal, seeds are moved away
from the parent by animals, wind, or gravity. During the
following phase, secondary dispersal, these seeds are
moved to new locations by animals or abiotic mech-
anisms (wind and water). For example, when verte-
brates ingest and defecate fruit, viable seeds can be
removed from feces by insects, birds, or mammals and
dispersed again (Estrada and Coates-Estrada 1991;
Forget and Milleron 1991; Levey and Byrne 1993;
Wenny 1999). Scatter-hoarding rodents that bury seeds
in the topsoil for use as food reserves are a particularly
important source of secondary dispersal (Brodin 1993;
Burnell and Tomback 1985). If cached seeds are not
recovered and eaten, scatter-hoarding can yield effec-
tive dispersal (Jansen and Forget 2001). Vander Wall
and Longland (2004) have argued that the combination
of two dispersal mechanisms is often more beneficial to
seeds than single means of dispersal.
Despite the importance of secondary dispersal by
scatter-hoarding rodents, the degree to which rodents
are effective seed dispersers has rarely been measured
due to methodological limitations (Galvez et al. 2009;
Galetti et al. 2010). Most published studies of seed
dispersal focus on the initial dispersal of seeds from
the source plant and do not record additional (aka
secondary) dispersal (Vander Wall et al. 2005, Emm-
erson et al. 2010). In studies where seed caches have
been monitored, it has often been assumed that seeds
that were removed from the soil were eaten, but few
studies have actually determined the location and fate
of those missing seeds (Vander Wall et al. 2005). In
this article, we describe a method to follow seeds
during secondary dispersal, which allows one to
determine the exact fate of these missing seeds.
The current standard method for tracking seed
dispersal by scatter-hoarding mammals is to affix
thread tags (with or without numbered flagging tape
at the distal end) to seeds and retrieve dispersed seeds
by visually searching for the tags in a radius around the
point of release (Forget and Wenny 2005). Alternative
tagging methods for tracking seed dispersal by animals
include radioisotopes (Carlo et al. 2009; Vander Wall
2002a, b; Winn 1989), fluorescent dye (Lemke et al.
2009; Longland and Clements 1995), and metal or
magnets (Den Ouden et al. 2005). A limitation of the
above methods is that seeds dispersed over longer
distances are less likely to be retrieved because search
effort increases exponentially with search radius. Most
studies omit non-retrieved seeds from the dataset,
leading to an underestimation of the true seed dispersal
distance (Hirsch et al. 2012; Vander Wall et al. 2005).
Given the ecological importance of long-distance
dispersal, there is a need for methods that allow for
tracking seeds without bias against long distances (Cain
et al. 2000; Nathan 2008; Portnoy and Willson 1993).
Radio transmitters offer a long-distance tracking
solution, because they are detectable over many
hundreds or thousands of meters and have recently
been miniaturized to the extent that they can be
attached to seeds. Radio transmitters mounted onto or
inside seeds have been successfully used to track
dispersal of acorns by mice in Spain (Pons and Pausas
2007) and Japan (Sone et al. 2002), and of walnuts by
squirrels and mice in Japan (Tamura et al. 2002). A
major shortcoming of these small transmitters is that
they have short-lived batteries, and therefore a limited
window of time to track seed movement. So far, this
has made transmitters unsuitable for the long-term
monitoring ([2?3 months) needed to determine ulti-
mate seed fate. Other complications include attaching
the transmitter to the seed and designing a transmitter
with an antenna sufficiently long and exposed to
improve detection range.
Here, we describe the use of thread tags with
integrated motion-sensitive miniature transmitters
(i.e. telemetric thread tags) for tracking seed dispersal
by scatter-hoarding rodents. The telemetric thread
tagging technique was designed to meet four criteria:
(1) There should be little or no bias against long-
distance dispersal.
(2) Monitoring should be possible over a time span
long enough to track seeds until they germinate
and establish into seedlings.
(3) Identification of cached seeds should be possible
without disturbing the cache. Opening caches
to visually identify seeds could strongly affect
seed fate since rodents are known to use soil
Plant Ecol
123
disturbance as a cue for finding seeds (Vander
Wall et al. 2003).
(4) There should be little or no effect of the tag on
disperser behavior, either negative or positive.
We report details of the method along with
successful and failed field trials with the rodent-
dispersed palm species Astrocaryum standleyanum in
a Neotropical forest. We describe advantages and
disadvantages of the method, general guidelines as to
when the method can be used, and reasons why we
believe telemetric seed monitoring has the potential to
substantially increase our understanding of plant
dispersal and forest ecology.
Methods
Telemetric thread tracking
The telemetric thread tag consists of a small cylindri-
cal radio transmitter with a 20-cm wire antenna
(Fig. 1). The transmitter is pulled behind the seed in
the same manner as traditional thread tags, which are
the current standard for tracking rodent-dispersed
seeds (Forget and Wenny 2005). A general advantage
of thread tags is that it is possible to identify buried
seeds without disturbing the cache, as soil disturbance
potentially affects subsequent cache and seed fate
(Guimaraes et al. 2005; Murie 1977). Having the
transmitter separated from the seed by a thread makes
it possible to manually turn off the transmitter when a
seed is buried, thus greatly extending battery life, and
making it possible to follow multiple movements of
the same seed over periods of [1 year. All previous
studies that radio-tracked seeds put transmitters
directly onto or inside the seed, and could only track
seeds over short time spans (Pons and Pausas 2007;
Sone et al. 2002; Tamura et al. 2002).
A key aspect of our method is that the transmitters
consume power only when dispersal events happen,
and not during the long time spans during which seeds
remain motionless. We developed two designs for
turning off transmitters when not moving (in collab-
oration with Advanced Telemetry Systems, Isanti,
MN). In the first design, transmitters contained an
internal tip switch that was designed to turn on the
transmitter when the seed was moved. Beta testing of
these first generation transmitters in 2006 showed
them to be too sensitive; they often turned on during
heavy rains or from subtle movements of the leaf litter.
These switches were sensitive to high humidity and
many got stuck over time, thus we do not recommend
this method for future studies.
The second- and third-generation transmitters used
a magnet operated switch inside the transmitter to turn
off radio transmitters until they are ready to be used
(Kenward 2001). These tags were deactivated by
placing the transmitter on top of an 8 9 22 mm
magnet taped to the head of a 10-cm nail that was
pushed into the soil *25 cm from the seed (Fig. 1).
When a seed was moved, the transmitter was pulled
off the magnet, reactivating the radio signal. A small
piece of wire inserted into the second-generation
transmitters provided enough magnetic attraction to
the magnet to ensure that the transmitter did not easily
slip off the magnet during rain storms, but provided
little resistance when an animal pulled the transmitter
Fig. 1 Diagram showing
the telemetric seed tag
(center) connected by wire
to a buried seed (bottom
right). The telemetric seed
tag is laid on top of a magnet
which has been staked into
the ground with a nail.
Drawing by Patricia Kernan
Plant Ecol
123
off the magnet. The second-generation transmitters
weighed 4.10 g, including a 20-cm long wire antenna.
We experimented with two power sources. The first-
and second-generation transmitters had one cylindrical
3-V lithium pin battery (25 mAh, 4.2 9 25.9 mm,
0.55 g; National BR425, New York, NY). These
transmitters experienced high rates of battery failure
([60 % of transmitters of second generation failed
during a 6-month period). A third generation of trans-
mitters used two 1.55 V silver oxide button cells
(80 mAh, 5.4 9 7.9 mm, 1.5 g; Renata 393, Itingen,
Switzerland) and had a low failure rate (2 %) over a
6-month period. The third generation transmitters
attached to magnets without an inserted metal wire in
the transmitter housing, resulting in a lighter transmitter
(3.80 g). The approximate lifespan of a radio transmitter
was 6?8 weeks of continuous transmission. Because the
transmitters were generally turned off within 24 h of
moving they worked in the field much longer
([10 months). As seeds were eaten or censored during
the course of the study, we were able to reuse the
transmitters on different seeds, thus the long lifespan of
the transmitter allowed us to follow more seeds than we
initially had telemetric tags for. These transmitters cost
the same as standard radio collars (*$200 per unit, or
less with bulk discount), thus reusing tags allowed us to
obtain more data without increasing the cost of the study.
Field application
We used telemetric thread tags to track seed dispersal
of the palm A. standleyanum on Barro Colorado
Island, Panama (9100N, 79510W), a 1,560 ha island
covered with tropical moist forest. Each Astrocaryum
tree produces up to 1,500 fruits per year. Each fruit
contains a ca. 9.6 g stone consisting of a large seed
enclosed in woody endocarp (Wright et al. 2010). The
seeds are typically dispersed by the Central American
agouti, Dasyprocta punctata, a 2?4 kg caviomorph
rodent that scatter hoards the stones as food reserves
for periods of food scarcity (Smythe 1978, 1989;
Jansen et al. 2010). Astrocaryum seeds generally
germinate during the wet season, one or more years
after being dispersed (Smyth 1989; Potvin et al. 2003).
Telemetric tags were attached to A. standleyanum
stones (henceforth ??seeds??) with 30 cm of black
nylon-coated stainless-steel leader wire (American
fishing wire: Surflon 1 9 7 black coating) tied to a
7-mm screw eye inserted halfway into the basal tip of
the seed. Screws were inserted opposite to the embryo,
thus the endosperm was hardly penetrated and not
exposed to water or microbes which could have
affected germination. The wire tied to the screw was
difficult to visually detect, and the wire and screw
were relatively difficult for rodents to cut in compar-
ison to traditional thread tags. To facilitate visual
retrieval and individual recognition, we attached a
7 cm piece of pink flagging tape to the wire near the
transmitter.
We tracked the fate of tagged seeds with traditional
manual telemetry in combination with an Automated
Radio Telemetry System (ARTS, Kays et al. 2011).
Most seed tracking was conducted with manual
telemetry, while the ARTS provided a backup for
determining if transmitters were still turned on in the
field. Whereas transmitters in a typical study would
each have a unique radio frequency, seed transmitters
in our study were tuned to one of four frequencies in
the 150?152 MHz range, which reduced the number
of frequencies that needed to be monitored with
ARTS. A drawback to using multiple transmitters with
the same frequency was that locating the transmitters
was more difficult. In some cases, we could hear 20 or
more seed transmitters of the same frequency from one
location, resulting in a chorus of beeping transmitters.
We were still able to manually locate transmitters by
walking toward the strongest signal, which was
typically the closest seed, turning this transmitter off
and then continuing toward the next strongest signal.
We strongly recommend using unique frequencies for
each seed in future studies.
The effect of telemetric thread tags on seed removal
rates and dispersal distance was evaluated by com-
paring fate between seeds with first-generation tele-
metric thread tags, seeds with traditional thread tags,
and untagged seeds. During 2006, we placed ten seed
stations across BCI that each contained five seeds of
each treatment (150 seeds in total). We monitored seed
removal with a motion-triggered camera trap (Silent
Image, Reconyx, Holmen, WI) and then compared
time-to-removal between the three types. We located
as many dispersed seeds as we could and compared
dispersal distance between radio-tagged and thread-
tagged seeds (untagged seeds could not be retrieved).
During 2009 and 2010, we only used telemetric thread
tags (second and third generation tags).
The overall field performance of the third genera-
tion telemetric tags was evaluated for 589 seeds with
Plant Ecol
123
telemetric thread tags in 2010. We placed five tagged
seeds at a time in 52 stations scattered across BCI, and
monitored removal with a motion-triggered camera
trap (RC55 or PC800, Reconyx, Holmen, WI) (Fig. 2).
We recorded the animal species and time of seed
removal for each seed (as in Jansen et al. 2002, 2004,
Jansen and den Ouden 2005, Yasuda et al. 2000,
2005). Seed stations were placed in the field for a
maximum of 8 days. If seeds remained at the end of
8 days, the seeds were replaced with fresh seeds, or the
entire station was canceled. Seeds were removed by
rodents from 32 of 52 stations. Each seed plot was
checked daily and removed seeds were located by
sight or with hand-held radio-telemetry equipment
(Yaesu-VR500, Cypress, CA) to determine dispersal
distance and seed fate. If the seed was found\20 m of
the seed plot, the dispersal distance was measured with
measuring tape, and the direction of movement was
recorded using a precision compass (Sunto KB-14). If
the seed was moved [20 m, the location of the seed
was recorded using a GPS receiver (Garmin 60CSx).
To increase GPS accuracy, we averaged at least 50
waypoints per seed location. If the seed was cached,
we turned off the transmitter by placing it on a magnet,
and continued to monitor the seed. Higher-order seed
movements (seed removed from caches) were treated
in the same way. Seeds moved frequently after being
first set out, in some cases more than once per day.
After 3?4 months, many seeds were eaten and the
remaining seeds were moved less frequently. We
gradually decreased the amount of manual seed checks
after four months of daily checks, and used the ARTS
to determine if any seeds had moved before manually
checking seeds.
Possible effects of telemetric tags on cache fate
were evaluated by following the fate of caches for a
subset of 46 seeds in 2010 with camera traps to
determine whether rodents used the telemetric seed
tags as cues for finding caches. A camera was mounted
onto a nearby tree or onto a U-shaped metal rebar
pushed into the ground 1.5 m from the seed cache
(Fig. 2), and set to continuously take photos when
animals passed in front. We closely investigated the
series of photos taken when a cache was removed to
check for cueing behavior. If agoutis or other rodents
are able to use the telemetric seed tags as a cue to the
location of buried seeds, we expected that we would be
able to observe this behavior from the photos. In
addition, we verified whether the cameras themselves
influenced the likelihood of a cache being removed by
comparing removal rates between caches with and
without cameras.
Fig. 2 Motion-sensitive camera trap mounted on a U-shaped concrete-reinforcement bar, used to monitor removal of seeds from an
experimental station or from a cache
Plant Ecol
123
Results
Telemetric seed tag efficacy
The telemetric tags did not affect seed removal. The
2006 tests yielded no significant difference in the
proportion of seeds removed between seeds without
tag (92.0 %), with a traditional thread tag (86.7 %),
and with a telemetric thread tag (94.0 %; Chi-square
test v22;145, p = 0.440), nor was there a significant
difference in time-to-removal (Fig. 3a; Kaplan?Meier
survival analysis with log-rank test on lumped data:
n = 145, v22 ? 1:3, p = 0.516). Also, dispersal
distance did not differ between thread-tagged seeds
and radio-tagged seeds (Fig. 3b; n = 72, v21 ? 0:5,
p = 0.482). We cannot evaluate, however, whether
dispersal distance differed between tagged and
untagged seeds, because there was no way to retrieve
untagged seeds.
Rates of seed recovery after dispersal were
extremely high. Of the 589 seeds placed during
2010, 424 (72 %) were removed by animals. Of these
424, we recovered 97 %. Of the 13 non-retrieved
seeds, 11 were lost because rodents chewed through
the wire attaching the transmitter to the tag, and two
seed tags were never found, presumably because of
transmitter failure. Assuming that tag cutting and radio
failure are independent of dispersal distance, the
method captured dispersal with minimal bias against
long distances.
Turning off transmitters during inactive periods
allowed us to monitor seeds over long time spans. We
tracked higher-order movements for a subset of the
424 seeds removed during 2010. We followed 224
cached seeds across multiple seasons, until these either
were eaten or had survived to 1 year. When a seed had
been eaten or a tag cut, we would reuse the transmitter,
thus we were able to monitor far more than 100 seeds.
During 2010, only 2 out of the 100 third generation
transmitters were lost over a span of 6 months. It was
not known whether these transmitters had internal
malfunctions, premature battery failure, were dam-
aged by animals (we observed chew marks of rats on
some transmitters), or were taken so far that we were
unable to detect the radio signal. It is unlikely that
seeds dispersed too far to be detected since signals
were typically detected[400 m from the ground, and
1,000 m from the ARTS towers.
The rate of severed tags was low (2.5 % per
movement), much lower than the proportions reported
in studies with traditional thread tags. However, the
probability of a transmitter being cut off increased
over time, as the same seeds could be handled multiple
times. Ultimately, 60 (26.7 %) of the 224 seeds had the
wire portion of their seed tags cut off, and this
remained a major source of seed loss.
0.0
0.2
0.4
0.6
0.8
1.0
Cu
m
ul
at
ive
 p
ro
po
rti
on
0.1 1 10 100
A 
Removal time (days)
0 10 20 30 40 50 60
0.0
0.2
0.4
0.6
0.8
1.0
Dispersal distance (m)
Cu
m
ul
at
ive
 p
ro
po
rti
on
B 
Fig. 3 Time-to-removal (a) and dispersal distance (b) by rodents
for Astrocaryum standleyanum seeds with telemetric thread tags
(dashed line), with traditional thread tags (solid lines), and
without any tag (dotted line), on Barro Colorado Island, Panama
during 2006. Values shown are inverse Kaplan?Meier survivor-
ship estimates. The differences were not significant
Plant Ecol
123
Dispersal distance
We recorded a total of 1,582 first- and higher-order
seed movements during 2010. The mean movement
distance was 16.8 m (range 0.1?241.3 m), which is
within the normal search radius for seed dispersal
studies using thread tags. However, 280 movements
(17.7 %) exceeded 30 m, and would normally have
resulted in loss of the seed in studies using other
tagging methods (Howe 1990, de Steven 1994).
Cueing
Seed caches monitored with remote cameras were
removed more slowly than normal seeds (average time-
to-removal for non-camera caches = 7.8 days, camera
caches = 19.7 days, Kruskal?Wallis v2 = 52.14, p =
\0.001). We were able to determine the species of
animal that recovered the seed from 92.5 % of caches.
In 12.5 % of these cases, the photo sequences of cache
retrieval suggested that agoutis sometimes cued on the
flagging tape for finding cached seeds. In these cases,
agoutis approached the flagging tape, and either
followed the attached wire toward the location of the
cached seed (n = 10), or started digging in soil nearby
the flagging tape (n = 7). This behavior was very
distinct from the typical seed discovery seen in the other
cases where animals tune in on the seed or digging
traces; in those cases the animal typically walked
straight toward the cached seed, and dug immediately
above the location of the seed.
Discussion
How effective are telemetric seed tags for tracking
seed dispersal?
Traditional seed dispersal studies are limited because
dispersing seeds are notoriously difficult to track,
resulting in bias against long-distance dispersal and
secondary seed movements (Wang and Smith 2002).
In this article, we have described seed tracking with
telemetric thread tags as a method for studying seed
dispersal by scatter-hoarding rodents without these
biases. Field tests in Panama indicated that the method
fulfilled all four of our performance criteria: (1) the
transmitters allowed near-complete retrieval of dis-
persed seeds with minimal bias against longer
distances, (2) the thread design allowed continued
monitoring of cached seeds without disturbance of the
caches, 3) a simple motion-sensitive triggering mech-
anism allowed the monitoring of seeds over long time
spans, and (4) tag effects on seed fate were minor and
all related to the optional flagging.
The telemetric thread tags allowed us to quickly
and accurately find seeds regardless of dispersal
distance, including seeds that had been dispersed
[200 m, which would have been nearly impossible to
retrieve with traditional tagging methods. Retrieval
with minimal distance bias is a major advantage of
telemetric tags compared to other tags. Studies using
non-telemetric tags invariably involve searching the
area around the location of seed dispersal up to a
certain radius, typically \30 m, as search effort
increases exponentially with distance (Howe 1990,
de Stevens 1994). The inherent bias against dispersal
distances exceeding this radius is a major problem that
is poorly dealt with (Hirsch et al. 2012). In our trials,
17.7 % of the dispersal events exceeded 30 m, and
would normally have resulted in loss of the seed in
studies using traditional tagging methods. One poten-
tial limitation of these telemetric seed tags is that our
hand-held radio-tracking equipment had a maximum
detection range of *400 m. In our case, even the
seeds that were dispersed the furthest (almost 250 m)
still remained within this detection range. If
an extreme long-distance dispersal event occurred
(Higgins et al. 2003), we may not have been able to
discover the seed with hand-held equipment (although
we almost certainly would have detected the seed
through the ARTS system). For this reason, research-
ers using telemetric thread tags may need to use
caution in interpreting results when seeds disappear
from the system.
Our final transmitter design was robust, with a
maximum 2 % failure rate over 6 months, similar to
previous seed-tracking studies that used telemetry
(0?4.5 %; Pons and Pausas 2007; Sone et al. 2002;
Tamura et al. 2002). The most common source of seed
loss was due to animals cutting the wire attaching the
transmitter to the seed. Even with wire cutting, the rate
of seed loss was much lower than studies using
traditional non-telemetric seed tags (up to 100 %
Forget and Wenny 2005), yet wire cutting remains a
significant problem for tracking seeds that are handled
by rodents multiple times. We saw no evidence of any
pattern or bias in seeds with cut wires, thus we do not
Plant Ecol
123
expect cutting to bias dispersal distance estimates.
However, we are unable to conclusively eliminate the
possibility that agoutis preferentially cut seeds they
intended to disperse particularly far.
All previous telemetric studies of seed movement
used transmitters directly mounted onto, or inside, the
seed or nut (Pons and Pausas 2007; Sone et al. 2002;
Tamura et al. 2002). Our integration of the transmitter
into a thread tag has several advantages to direct
mounting. Because the transmitter was not buried
during seed caching, we could determine the identity
of the seed without disturbing the cache, avoiding soil
disturbance that rodents use as cue for cache finding
(Guimaraes et al. 2005; Murie 1977). It also allowed
us to turn the transmitter off when seeds were not
moving, which saved battery power. Frequent deac-
tivation allowed us to use small, lightweight tags for
monitoring seeds over many months to determine their
ultimate fate; germination or death. We checked seeds
once per day during the initial 4 months of our study,
but different study systems may require more or less
frequent seed monitoring, depending on the number of
seeds being tracked and their rate of movement.
Camera-trap footage of cache recovery suggested
that agoutis occasionally used flagging tape as a cue
for finding cached seeds. The videos revealed that the
animals cued on the pink-flagging tape rather than the
black wire, transmitter, or transmitter antenna. After
noticing the flagging, some agoutis appeared to use the
wire thread to guide them to cached seed, even though
the wire appeared difficult for rodents to see. If the
visual cue provided by the flagging tape was elimi-
nated, we expect that animals would not have found
the seeds using only the wire. Although cueing on
flagging had a minor effect on the results, we
recommend that future studies should aim to eliminate
the flag, thus minimizing this problem. Finding
transmitters without flagging is more difficult for field
workers, but still possible. A unique serial number
written onto or inserted inside the transmitters along
with unique transmitter frequencies can be used as to
identify these seeds when no flagging is present. Our
finding that some animals cue on flags to find seeds
also suggests that thread tags may have influenced
removal rates and survival times of caches in past
studies.
We were also concerned that agoutis might have
learned to associate the cameras with cached seeds.
We found the opposite effect, and it appears that the
cameras may have scared the agoutis away from the
caches. This result is unexpected given the widespread
use of remote cameras (with and without cached
seeds) in the study area from 2008 through 2010.
Despite this effect, we found that the cache cameras
were an effective method to record the species which
removed cached seeds, and the behavior of the animals
when retrieving cached seeds.
Broader potential
Since most seeds are small, and can take many months
to reach their ultimate fate, the size and lifespan of
radio transmitters has limited their use in the tracking
of seed movement. However, the size of radio tags
continues to drop, with various companies now
offering transmitters that weigh just 0.2 g, ushering
in a new era of study of small animals, such as bees and
dragonflies (Wikelski et al. 2006, 2010). We believe
that the study of seed dispersal is also ripe for
discovery with miniature tags. Seed size in the Plant
Kingdom ranges from the 1-lg dust seeds of orchids to
the 20-kg seeds of the double coconut Lodoicea
seychellarum (Harper 1977). Based on our experi-
ences with the telemetric thread tag, we argue that
many more plant species can be radio tracked than is
currently believed.
The accepted guideline for placing a radio trans-
mitter on an animal is that the transmitter should not be
[5?10 % of an animals? body weight (Murray and
Fuller 2000, Kenward 2001), but there is no reason for
applying this rule to the weight of seeds themselves,
which are being voluntarily moved by animals much
larger than themselves. Our telemetric thread tags
added 40 % weight to the seeds of A. standleyanum,
but only 0.1 % to the weight of the agoutis that carried
the seeds, and this did not appear to affect their
behavior.
To estimate how widespread this technology might be
used for studying animal dispersed seeds, we linked the
fruit species consumed by two common scatter-hoarding
mammals in Central Panama; spiny rat (Proechimys
semispinosus) and agouti (Adler 1995; Aliaga-Rossel
et al. 2008) to the corresponding seed sizes (Wright et al.
2010). A majority of the plant species consumed by these
animals have seeds C0.5 g, the mass of the smallest
available radio transmitter (spiny rat = 50.1 %, agouti =
77.5 %). The maximum weight of seeds taken by these
two species far exceeds that of our seed transmitters
Plant Ecol
123
(spiny rat =13.3 g, agouti = 63.7 g). For mammals
that carry relatively heavy seeds, adding additional
weight to a seed should not noticeably change the
animals? behavior. Indeed, Mun?oz and Bonal (2008)
found that rodents regularly moved seeds up to 70 % as
heavy as their own body mass.
We experimented with several different methods
for attaching the transmitters to seeds and conserving
battery power. We recommend experimenting with
threads, glues, and other techniques to find the right
solution for a given seed-mammal combination. In the
case of seeds ingested by frugivores, gut passage time
and seed dispersal can be measured by placing seed-
shaped transmitters inside fresh fruit (Mack and
Druliner 2003). In each case, the key test will be
how the animals handle the modified seeds relative to
unmodified ones, a relatively simple test to make with
modern camera traps.
Conclusion
The telemetric thread tag method can be used to track
dispersing seeds at distances far greater than tradi-
tional techniques, identify cached seeds without
disturbing the cache site, document secondary dis-
persal, and identify ultimate seed fate. Combining
telemetric thread tags with remote cameras can be
used to document the exact time of removal, and the
species which removed the seed (Jansen and Den
Ouden 2005). These techniques can be used in various
combinations to investigate a wide variety of ecolog-
ical questions in a large number of biological systems.
Given the importance of studying secondary dispersal,
we believe that our methods represent an important
methodological step to help understand the dynamics
of seed dispersal and forest regeneration (Vander Wall
et al. 2005).
Acknowledgments We thank Tom Garin (ATS), Daniel
Obando, Alejandro Ortega and Meg Crofoot for technical
support; Lieneke Bakker, Reyna Bell, Chris Carson, Willem-Jan
Emsens, Matt McElroy, Veronica Pereira, Torrey Rodgers,
Sumana Serchan, Michiel Veldhuis, Brian Watts, and Veronica
Zamora Guttierez for field assistance; the Smithsonian Tropical
Research Institute for facilities and administrative support. Joe
Wright allowed us access to unpublished data on seed weight,
and brought up the original idea for radio-tracking seeds with
ARTS. We thank two anonymous reviewers for their helpful
comments and suggestions. We also thank Patricia Kernan for
the figure artwork. This study was supported by funding from
the National Science Foundation (NSF-DEB 0717071 to RWK)
and the Netherlands Organization for Scientific Research
(Grants: W85-239 and 863-07-008 to P.A.J.).
References
Adler GH (1995) Fruit and seed exploitation by Central
American spiny rats, Proechimys semispinosus. Stud
Neotrop Fauna Environ 30:237?244
Aliaga-Rossel E, Kays RW, Fragoso JMV (2008) Home-range
use by the Central American agouti (Dasyprocta punctata)
on Barro Colorado Island, Panama. J Trop Ecol 24:
367?374
Brodin A (1993) Low rate of loss of willow tit caches may increase
adaptiveness of long-term hoarding. Auk 110:642?645
Bullock JM, Shea K, Skarpaas O (2006) Measuring plant dis-
persal: an introduction to field methods and experimental
design. Plant Ecol 186:217?234
Burnell KL, Tomback DF (1985) Steller?s jays steal gray jay
caches: field and laboratory observations. Auk 102:417?419
Cain ML, Milligan BG, Strand AE (2000) Long-distance seed
dispersal in plant populations. Am J Bot 87:1217?1227
Carlo TA, Tewksbury J, del Rio CM (2009) A new method to
track seed dispersal and recruitment using 15 N isotope
enrichment. Ecology 90:3516?3525
Chambers JC, MacMahon JA (1994) A day in the life of a seed:
movements and fates of seeds and their implications for
natural and managed systems. Annu Rev Ecol Syst
25:263?292
De Steven D (1994) Tropical tree seedling dynamics: recruit-
ment patterns and their population consequences for three
canopy species. J Trop Ecol 10:369?383
Den Ouden J, Jansen PA, Smit R (2005) Jays, mice and oaks:
predation and dispersal of Quercus robur and Q. petraea in
north-western Europe. In: Forget P-M, Lambert JE, Hulme
PE, Vander Wall SB (eds) Seed fate: predation, dispersal
and seedling establishment. CABI publishing, Walling-
ford, pp 223?240
Emmerson L, Facelli JM, Chesson P, Possingham H (2010)
Secondary seed dispersal of Erodiophyllum elderi, a
patchily distributed short-lived perennial in the arid lands
of Australia. Austral Ecol 35:906?918
Estrada A, Coates-Estrada R (1991) Howler monkey (Alouatta
palliata), dung beetles (Scarabaeidae) and seed dispersal:
ecological interactions in the tropical rain forest of Los
Tuxtlas, Mexico. J Trop Ecol 7:459?474
Forget P-M, Milleron T (1991) Evidence for secondary seed
dispersal by rodents in Panama. Oecologia 87:596?599
Forget P-M, Wenny DG (2005) A review of methods used to
study seed removal and secondary seed dispersal. In: For-
get P-M, Lambert JE, Hulme PE, Vander Wall SB (eds)
Seed fate: predation, dispersal and seedling establishment.
CABI publishing, Wallingford, pp 379?394
Galetti M, Donatti CI, Steffler C et al (2010) The role of seed
mass on the caching decision by agoutis, Dasyprocta
leporina (Rodentia: Agoutidae). Zoologia 27:472?476
Galvez D, Kranstauber B, Kays RW, Jansen PA (2009) Scatter
hoarding by the Central American agouti: a test of optimal
cache spacing theory. Anim Behav 78:1327?1333
Plant Ecol
123
Guimaraes PR, Gomes BZ, Ahn YJ, Galetti M (2005) Cache
pilferage in red-rumped agoutis (Dasyprocta leporina)
(Rodentia). Mammalia 69:427?430
Harper JL (1977) Population biology of plants. Academic Press,
New York
Higgins SI, Nathan R, Cain ML (2003) Are long-distance dis-
persal events in plants usually caused by nonstandard
means of dispersal? Ecology 84:1945?1956
Hirsch BT, Visser M, Kays R, Jansen PA (2012) Quantifying
seed dispersal kernels from truncated seed-tracking data: A
test using Neotropical palm seeds with radio-transmitters.
Methods Ecol Evol. doi:10.1111/j.2041-210X.2011.00183
Howe HF (1990) Seed dispersal by birds and mammals: impli-
cations for seedling demography. In: Bawa KS, Hadley M
(eds) Reproductive ecology of tropical forest plants. Man
and the biosphere. Parthenon/UNESCO, Rome, pp 191?218
Jansen PA, den Ouden J (2005) Agents of post-dispersal seed
removal: video monitoring behavior, seed predation and
dispersal. In: Forget P-M, Lambert JE, Hulme PE, Vander
Wall SB (eds) Seed fate: predation, dispersal and seedling
establishment. CABI publishing, Wallingford, pp 363?378
Jansen PA, Forget PM (2001) Scatterhoarding and tree regen-
eration. In: Bongers F, Charles-Dominique P, Forget PM,
The?ry M (eds) Nouragues: dynamics and plant-animal
interactions in a neotropical rainforest. Kluwer Academic
Publishers, Dordrecht, pp 275?288
Jansen PA, Bartholomeus M, Bongers F et al (2002) The role of
seed size in dispersal by a scatterhoarding rodent. In: Levey
D, Silva WR, Galetti M (eds) Seed dispersal and frugivory:
ecology, evolution and conservation. CABI publishing,
Wallingford, pp 209?225
Jansen PA, Bongers F, Hemerik L (2004) Seed mass and mast
seeding enhance dispersal by a neotropical scatter-hoard-
ing rodent. Ecol Monogr 74:569?589
Jansen PA, Elschot K, Verkerk PJ, Wright JS (2010) Seed
predation and defleshing in the agouti-dispersed palm As-
trocaryum standleyanum. J Trop Ecol 26:1?8
Kays R, Tilak S, Crofoot M et al (2011) Tracking animal
location and activity with an automated radio telemetry
system in a tropical rainforest. Comput J. doi:10.1093/
comjnl/bxr072
Kenward RE (2001) A manual for wildlife radio tagging. Aca-
demic Press, San Diego
Lemke A, von der Lippe M, Kowarik I (2009) New opportuni-
ties for an old method: using fluorescent colours to measure
seed dispersal. J Appl Ecol 46:1122?1128
Levey DJ, Byrne MM (1993) Complex anti-plant interactions:
rain forest ants as secondary dispersers and post-dispersal
seed predators. Ecology 74:1802?1812
Longland WS, Clements C (1995) Use of fluorescent pigments
in studies of seed caching by rodents. J Mammal 76:
1260?1266
Mack AL, Druliner G (2003) A non-intrusive method for mea-
suring movements and seed dispersal in cassowaries.
J Field Ornithol 74:193?196
Mun?oz A, Bonal R (2008) Are you strong enough to carry that
seed? Seed size/body size ratios influence seed choices by
rodents. Anim Behav 76:709?715
Murie JO (1977) Cues used for cache-finding by agoutis
(Dasyprocta punctata). J Mammal 58:95?96
Murray DL, Fuller MR (2000) A critical review of the effects of
marking on the biology of vertebrates. In: Boitani L, Fuller
TK (eds) Research techniques in animal ecology: contro-
versies and consequences. Columbia University Press,
New York, pp 15?64
Nathan R (2008) Long-distance dispersal of plants. Science
313:786?788
Pons J, Pausas JG (2007) Acorn dispersal estimated by radio-
tracking. Oecologia 153:903?911
Portnoy S, Willson MF (1993) Seed dispersal curves: behavior
of the tail of the distribution. Evol Ecol 7:25?44
Potvin C, Cansari R, Hutton J, Caisamo I, Pacheco B (2003)
Preparation for propagation: understanding germination of
giwa (Astrocaryum standleyanum), wagara (Sabal mauri-
tiiformis), and eba (Socratea exorrhiza) for future culti-
vation. Biodivers Conserv 12:2161?2171
Smythe N (1978) The natural history of the Central American
agouti (Dasyprocta punctata). Smithson Contribut Zool
257:1?52
Smythe N (1989) Seed survival in the palm Astrocaryum
standleyanum: evidence for dependence upon its seed
dispersers. Biotropica 21:50?56
Sone K, Hiroi S, Nagahama D et al (2002) Hoarding of acorns by
granivorous mice and its role in the population processes of
Pasania edulis (Makino). Makino Ecol Res 17:553?564
Tamura N, Hashimoto Y, Hayashi F (2002) Optimal distances
for squirrels to transport and hoard walnuts. Anim Behav
58:635?642
Vander Wall SB (2002a) Masting in animal-dispersed pines
facilitates seed dispersal. Ecology 84:3508?3516
Vander Wall SB (2002b) Secondary dispersal of Jeffrey pine
seeds by rodent scatter hoarders: the roles of pilfering,
recaching, and a variable environment. In: Levey D, Silva
WR, Galetti M (eds) Seed dispersal and frugivory: ecology,
evolution and conservation. CABI publishing, Walling-
ford, pp 193?208
Vander Wall SB, Longland WS (2004) Diplochory: are two seed
dispersers better than one? Trends Ecol Evol 19:155?161
Vander Wall SB, Beck MJ, Briggs JS et al (2003) Interspecific
variation in the olfactory abilities of granivorous rodents.
J Mammal 84:487?496
Vander Wall SB, Kuhn KM, Beck MJ (2005) Seed removal,
seed predation, and secondary dispersal. Ecology 86:
801?806
Wang BC, Smith TB (2002) Closing the seed dispersal loop.
Trends Ecol Evol 17:379?385
Wenny DG (1999) Two-stage dispersal of Guarea glabra and G.
kunthiana (Meliaceae) in Monteverde, Costa Rica. J Trop
Ecol 15:481?496
Wikelski M, Moskowitz D, Adelman JS, Cochran J, Wilcove
DS, May ML (2006) Simple rules guide dragonfly migra-
tion. Biol Lett 2:325?329
Wikelski M, Moxley J, Eaton-Mordas A et al (2010) Large-
range movements of Neotropical orchid bees observed via
radio telemetry. PLoS ONE 5:e10738
Winn AA (1989) Using radionuclide labels to determine the
post-dispersal fate of seeds. Trends Ecol Evol 4:1?2
Wright JS, Kitajima K, Kraft NJB et al (2010) Functional traits
and the growth-mortality trade-off in tropical trees. Ecol-
ogy 91:3364?3674
Plant Ecol
123
Yasuda M, Miura S, Hussein NA (2000) Evidence for hoarding
behaviour in terrestrial rodents in Pasoh Forest reserve, a
Malaysian lowland rain forest. J Trop For Sci 12:164?173
Yasuda M, Miura S, Ishii N et al (2005) Fallen fruits and ter-
restrial vertebrate frugivores: a case study in a lowland
tropical rain forest in Peninsular Malaysia. In: Forget P-M,
Lambert JE, Hulme PE, Vander Wall SB (eds) Seed fate:
predation, dispersal and seedling establishment. CABI
publishing, Wallingford, pp 151?174
Plant Ecol
123