The Curious Case of the Cane Toad (Rhinella marina): An Assessment of Exploratory Behavior and Foraging Success of an Invasive Vertebrate in a Novel Environment by Amanda J. Arner, B.S. A Thesis In BIOLOGY Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCES Approved Dr. Ximena E. Bernal, Ph.D. Chair of Committee Dr. Rachel A. Page, Ph.D. Dr. Richard E. Strauss, Ph.D. Peggy Gordon Miller Dean of the Graduate School August, 2012
85
Embed
The Curious Case of the Cane Toad (Rhinella marina): An ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Curious Case of the Cane Toad (Rhinella marina): An Assessment of Exploratory
Behavior and Foraging Success of an Invasive Vertebrate in a Novel Environment
by
Amanda J. Arner, B.S.
A Thesis
In
BIOLOGY
Submitted to the Graduate Faculty of Texas Tech University in
Partial Fulfillment of the Requirements for
the Degree of
MASTER OF SCIENCES
Approved
Dr. Ximena E. Bernal, Ph.D. Chair of Committee
Dr. Rachel A. Page, Ph.D.
Dr. Richard E. Strauss, Ph.D.
Peggy Gordon Miller Dean of the Graduate School
August, 2012
Copyright 2012, Amanda J. Arner
ii
ACKNOWLEDGMENTS
I would like to take this opportunity to thank my friends, family, advisors and
mentors for helping me through the completion of my Masters of Science. I am truly
blessed to have such wonderful people in my life; without their motivation, support
and proverbial shoulders to cry on, I would not have made it to this point in my career.
Firstly I would like to thank my advisor, Dr. Ximena E. Bernal, for her support
through this process. Her guidance and trust in me and her other graduate students are
apparent and reflected in her advising practices. I would like to that my committee
members Dr. Rachel A. Page and Dr. Rich E. Strauss for their support as well; Dr.
Page for her generosity in allowing me to use her resources and lab space and
equipment for my field research, and Dr. Stauss for his continued advice on statistical
methods and interpretation of biological meaning.
Though my advisors guided me through the process of conducting an original
research project, I cannot claim success without acknowledging my graduate student
peers for their care and support. Graduate school is a time of intrinsic growth and
establishment of individual identity as a scientist and professional, and the following
people have helped shape who I have become over this journey; Lynne Beaty, Maria
Gaetani, Meghan Cromie, Elizabeth Waring, Janice Kelly, Jenny Strovas, and
Priyanka DeSilva. My undergraduate research assistants, Katelyn Jordan and Meagan
Phelps, have been a tremendous help in conducting my laboratory research.
Texas Tech University, Amanda Arner, August 2012
iii
My research projects were monetarily supported in part by three entities, which
I would like to acknowledge as well. Dr. Ximena Bernal, thank you for allowing my
research to benefit from your start-up funding, and allowing me the flexibility to
expand the toad lab to suit my research needs. The Association of Biologists at Texas
Tech University provided me with grant funding to purchase research supplies while
conducting field work in Panama, and provided me with financial assistance to attend
several conferences to present my research. Lastly, I would like to thank the HHMI-
funded Center for the Integration of Science Education and Research for their funding
support through the Graduate Teaching Scholars program. This funding enabled me to
grow as a professional educator as well as a scientist, and has better prepared me for a
future in science education.
Last but certainly not least; I would like to thank my family for their support.
My mother, Jennifer Holcombe, my father and stepmother, Howard and Karen Arner,
and my sister, Megan Bryan – thank you for lending your ears and supporting me
through this difficult time in my life. Finally, I would like to thank my fiancé, Hil
Miller, for just about every single thing he does, especially for enduring the strain of a
long-distance relationship and accepting my fierce desire to remain in Lubbock and
finish my degree, regardless of the consequences.
Texas Tech University, Amanda Arner, August 2012
iv
TABLE OF CONTENTS
Acknowledgments ......................................................................................................... ii
Abstract .......................................................................................................................... vi
List of Tables .............................................................................................................. viii
List of Figures ................................................................................................................ ix
I. Introduction ................................................................................................................. 1
* Number of times bowl was encountered during all trials Ϯ Number of times bowl was eaten from during all trials Ӕ Number of times bowl was the first eaten from during all trials
Total bowl encounters summed across all trials for both the control (C) and
experimental (E) groups. The number in italics represents the proportion of the total
that each bowl represents Bowls 5 and 6 were encountered more often than other
bowls in both the experimental (E) and control (c) groups.
Texas Tech University, Amanda Arner, August 2012
30
Figure 2.3 Total Path Length, Time Spent in Margin, and Number of Escape Attempts
Texas Tech University, Amanda Arner, August 2012
31
Three variables were significant across trials for both treatment and control groups;
TPL (top), TMAR (middle), and ESC (bottom). Both total path length and number of
escape attempts decreased steadily over time. Time spent in the margin initially
decreased steadily for the experimental group in trials 1-4 before reaching a level
threshold held relatively constant for trials 5 and 6. The control group saw a more
gradual decrease in TMAR over time.
Texas Tech University, Amanda Arner, August 2012
32
Figure 2.4 Latency to Leave Origin and Time to Encounter Bowl
Neither of the variables associated with time changed significantly for either
treatment group as experience with the arena environment increased. LO (above, top)
showed almost no change over time, while the time to encounter a food bowl (bottom)
highly varied per individual tested, independent of treatment group.
Texas Tech University, Amanda Arner, August 2012
33
Figure 2.5 Number of Total and Unique Bowl Encounters
The number of total bowl encounters (top) substantially decreased for both treatment
groups between trials 1 and 6, however there were not substantial decreases between
intermediate trials, or between groups. The change in number of unique bowl
encounters (bottom) was not significant in either treatment group, however there is a
decreasing trend in the control group and an increasing trend in the experimental
group.
Texas Tech University, Amanda Arner, August 2012
34
EXPLORATION AND FORAGING BEHAVIOR OF THE CANE TOAD
(RHINELLA MARINA) FROM AN INVASIVE POPULATION
Introduction
When a species is introduced to a non-native area, individuals either establish a
small, sustained population or die out within a few generations. Occasionally, an
introduced species flourishes in its new environment and the population grows larger
each generation, successfully expanding from the introduction site (Lockwood et al.
2005). Many factors have been suggested to contribute to a species invasive potential
and ultimate success, however introduction effort, or propagule pressure, is the only
factor that is well-studied and consistent across invasive species (Lockwood et al.
2005). Behavioral characteristics may also influence the invasion success of
vertebrates (Holway & Suarez 1999), and behavioral flexibility has been identified as
a likely characteristic of invasive species (Sol 2002; Adamo & Lozada 2009).
Learning ability has been highlighted as a factor that promotes invasion success
(Amiel et al. 2011; Roudez et al. 2007; Sol 2002; Sol et al. 2008). If the ability to learn
is present in an invasive species to a greater extent than in native species that occupy
similar trophic levels, the invasive species could potentially outcompete natives and
receive a fitness advantage. This idea has been tested in a native and invasive species
of crab (Roudez et al. 2007), and similar results were found in native and invasive
crayfish when tested for association between a novel odor and predation risk (Hazlett
et al. 2002). If learning is a potential contributor to invasive success, then this ability
CHAPTER III
Texas Tech University, Amanda Arner, August 2012
35
should be present in the native populations prior to introduction. It is unlikely that
learning evolves in the introduced populations separate from the native populations,
due to the short time span over which introduced species often become invasive.
Learning is expected to be favored when resources in an environment are
predictable within the individual’s lifetime, but are not predictable between
generations (Stephens 1993). In a novel environment, such as those encountered by
the original individuals of an introduced or invasive species, resources are unknown
and therefore unpredictable. Because information about resources and environmental
factors is unknown, learning about the environment through exploration is a key
component of successful colonization (Russell et al. 2010). Therefore, we would
predict that there is a strong selective pressure on initial individuals of an introduced
species to learn about the environment, and those who do so will have higher survival
and reproductive success. The capacity to learn has been shown to have a genetic
component (Mery & Kawecki 2002), meaning this ability will be passed on to the
offspring of those individuals who are successful. Learning comes at a cost to
individuals who no longer need the ability (Mery & Kawecki 2004), so there is a
potential tradeoff between learned and innate behaviors depending on predictability of
the environment, consistent with Stephen’s learning model.
To elucidate possible differences in learning ability between cane toads in
native and invasive populations, we investigated exploratory and foraging behavior in
individuals from the invasive range in South Florida. The findings from chapter 2
suggest that cane toads from the native range change their foraging behavior with
Texas Tech University, Amanda Arner, August 2012
36
increased experience in a novel environment, and modulate their movements and
space use to more efficiently find food. If selection for the ability to learn is higher in
the invasive range, then cane toads from the invasive population should demonstrate
higher learning capabilities than the native population in Gamboa, Panama.
Methods
Cane toads were purchased from Carolina Biological Supply, who collected
them from areas in the South Florida invasive populations. Toads were housed at an
ACUC-approved animal facility at Texas Tech University (Lubbock, Texas), in
groups of 3-5 individuals in 50-gallon cattle tanks. Toads were kept under conditions
equivalent to those found in their native range of the tropics (80-85ºF, RH 85 %,
12L:12D cycle). Before the experiment, toads were removed from group housing and
housed individually to control their food intake. Individual housing bins were similar
to those used in the Gamboa experiment (Chapter 2); 24-liter plastic bins with 1-3cm
of peat moss substrate, a water container and a hide (ceramic flower pot). Toads were
randomly assigned to one of two treatment groups, control (no food in bowls) and
experimental (food in bowls).
Experimental Arena Design
The exploratory arena was developed based on concepts derived from
traditional maze designs such as the radial arm maze (Olton et al. 1977), open field
maze (Walsh & Cummins 1976) and the Morris water maze (Bilbo et al. 2000). The
arena consisted of a circular plastic wading pool 244 cm diameter x 46 cm tall,
Texas Tech University, Amanda Arner, August 2012
37
extended to 76 cm in height. The arena floor was covered with a mixture of 3-5cm
peat moss. Three types of pre-configured block formations were randomly placed at
equal spacing along the margin of the arena to promote exploration by creating a
spatially complex environment (Figure 1). Six plastic bowls (6cm tall x 15cm in
opening diameter) were placed in the arena in a randomized block design. Each bowl
contained one mealworm (2.5-4 cm long) for the experimental group and remained
empty for the control group. Infrared lights with a 30-ft range (Clover electronics) and
a high-resolution outdoor security camera (Supercircuits model # PC88WR) were
mounted on the ceiling approximately 160cm above the arena floor, and were used to
record all trials. Because cane toads are known to be visual foragers (Robins & Rogers
2004), LED lights were placed on the ceiling to give off an ambient light within the
range of natural moonlight (0.30 + 0.06 lux). Individuals began the trial from a
preselected point along the edge of the arena, which was held constant across all trials,
treatments, and individuals.
Experimental Procedure
Toads were moved into the arena in their individual hide structure from their
cage, which serves as a familiar origin point for trials. Each toad was tested for 60
minutes between the hours of 1430 and 1900, during their dark cycle. Trials were
video-recorded for further analysis. The number of mealworms eaten and location of
bowls eaten from was recorded for each trial. Although toad are not known to use
chemical cues when foraging (Martof 1962) the arena was sprayed liberally with aged
Texas Tech University, Amanda Arner, August 2012
38
water between trials to account for any chemosensory cues left behind by the previous
trial.
Toads were first introduced to the arena in a series of training trials, in which
each toad had 60 minutes to explore the arena. During this time, toads in the
experimental group were given the opportunity to find the bowls in the arena that
offered a food reward. Toads in the control group also encountered bowls; however no
food reward was presented. The end of the training period was marked by the
individual encountering a food bowl and successfully eating a mealworm. If a toad
did not eat by the fifth he was not tested in further trials. Since toads in the control
group were not offered a food reward, the training period for these individuals ended
when they encountered an empty food bowl. Control group toads were fed one
mealworm in their housing bin every other day, after their trial was completed.
Toads were tested for five trials after initially finding and eating a mealworm
during training. After the fifth trial, toads were tested for a sixth trial in which the
bowls were moved to new locations. We scored whether the toad first visited a
location that previously held a bowl, or the new location of the bowl for this trial. This
trial allowed us to determine if the toads were using the relative location of the food
within the environment or were associating the food bowls themselves with
mealworms despite their location in the arena.
The behavioral software package Ethovision XT (version 8.0) was used to
analyze each video-recorded trial at the rate of 1 fps (frame per second). Unless
otherwise stated, all variables were calculated using the program’s built in analysis
Texas Tech University, Amanda Arner, August 2012
39
functions. Videos were recorded to an external DVR box (Supercircuits model
#DMR80U) and imported into Ethovision for analysis. All statistical analyses were
conducted in SPSS (version 19) and critical p values are reported at a 95% confidence
level unless otherwise stated.
To determine if specific bowls were preferentially visited during the main
experimental trials, the frequency of visits to each bowl was calculated and compared
to the null hypothesis that if toads were visiting bowls randomly, then each bowl
should be eaten out of an equal number of times. The likelihood of eating out of a
specific bowl first during each trial was also calculated, based on bowls that had been
eaten out of first during previous trials.
Experiment #1 – Changes in Exploration with Increased Experience
To determine how cane toads modify their movement in the arena as
experience increases, the following variables were measured: time spent moving
(TM), total path length (TPL), time spent in the margin, or outer 25% of the arena
(TMAR), latency to leave the origin (LO), time to encounter a bowl (TB), and number
of escape attempts (ESC). Latency to leave origin and number of escape attempts were
recorded directly from videos by an observer. If toads are changing their behavior as
experience in the arena increases, then we expect a reduction of time allocated to
behaviors that are considered ‘exploratory’ (i.e. total time moving), and an increase of
behaviors that indicate familiarity with the environment (i.e. decreased thigmotaxis).
Due to sample size restrictions, data collected from this experiment were not
analyzed for statistical differences across trials or between groups. Correlations among
Texas Tech University, Amanda Arner, August 2012
40
variables were taken into account when describing the overall trends in behavior, and
highly correlated variables (rho ≥ 0.90) were selectively removed from analysis.
Experiment #2 – Changes in Foraging Behavior with Increased Experience
To determine if foraging behaviors changed across trials in the experimental
group, serving as a proxy for learning about the location of food resources in the
environment, we measured bowl encounters during each trial. A bowl encounter is
defined as a toad physically touching a bowl or coming within a three-centimeter
radius of the bowl for greater than two seconds. The total number of bowl encounters
(BE), as well as the number of unique bowl encounters (how many of the six bowls
available in the arena were encountered by the toad), were scored for each trial.
We hypothesized that toads in the experimental group will learn that food is available
in the bowls in the arena, and will continuously seek out food during trials. Thus, we
predicted that toads in this group will increase bowl encounters and unique bowl
encounters over time. Four additional variables related with food consumption were
recorded in the experimental group: Time from leaving the origin to eating first
mealworm (TE), path length from leaving the origin to eating first mealworm (PLE),
tortuosity, or curvature, of the path to eat first mealworm (T), and total number of
mealworms eaten per trial (MLS). PLE was calculated using point coordinate data
exported from Ethovision, and TE and MLS were directly measured by an observer
from the videos. Tortuosity was calculated by dividing the length of the curve (total
path length) by the Euclidean distance between its ends (Benhamou 2004). We
predicted that if toads are modifying foraging behavior in the arena due to availability
Texas Tech University, Amanda Arner, August 2012
41
of food resources, they should take less time to find food, cover a shorter distance to
reach food bowls, and eat a greater number of mealworms as experience with the
arena increases (i.e. as information about the location of food resources is learned).
As in the analysis for the first experiment, we did not have a sample size large
enough to support statistical analyses across trials or between groups. Correlations
were accounted for a priori, and variables that were highly correlated were excluded
from analysis based on biological meaning to the system.
Experiment #3 – The Role of Spatial and Visual Cues Involved in Learning
The purpose of the final experiment was to determine if cane toads in the
experimental group were using the relative spatial location of the food, or simply
associating the presence of food with bowls despite their location in the arena. During
the final testing trial we moved the bowls to new locations that did not previously
contain bowls (Figure 2). We predict that if toads are using spatial cues from the
environment to located food, then they should visit locations that previously held food
bowls before successfully locating a bowl in its new location. If the toads, however,
are associating the visual cue of the bowl with the presence of food, then we expect
the toads to move to the new bowl locations first before visiting the old bowl
locations, if at all.
In this experiment we recorded whether each toad first visited an old bowl
location or a new bowl location after leaving the origin. A visit to a bowl location was
scored if the toad entered the zone where the bowl was previously located and
remained static for two seconds or more, the same criterion used to determine bowl
Texas Tech University, Amanda Arner, August 2012
42
encounters during all other trials. We predict that toads in the control group, having
learned no association between bowls and food items, will visit old and new locations
with equal probabilities, while toads in the experimental treatment will preferentially
visit bowl locations depending on the specific cues used.
Results
We tested a total of 19 toads between September 5th and November 20th 2011
(Control = 8, Experimental = 11), with SVL 98.24 + 21.42mm and mass 104.31 +
6.33g. Due to methodological issues, three of the toads in the Experimental group
were unable to be used for analysis. Each toad in the control group completed 7 trials,
while six of the eight toads in the experimental group ate during the first trial. One
toad only ate once during the third trial, and one toad did not eat at all in the arena,
resulting in a final sample size of 14 (Control = 8, Experimental = 6). Due to the small
sample size and high variance associated with behavioral research, general trends seen
during trials are described here but no statistically analyze was performed. Toads in
both the control and experimental treatments behaved similarly during the training
trial when the arena was novel (Table 1), thus allowing us to draw conclusions about
their behavior based on experience gained in the arena.
Toads showed a preference for visiting certain bowls in the arena, as seen in
the previous study. The visitation preferences, however, differed slightly between the
two populations (Table 2). Bowl five was still a preferred bowl, with 26% of the total
bowl visits, but bowl two was visited most often, making up 29% of the total bowl
visits. Of bowl visits resulting in a successful meal, bowl five comprised 47% of all
Texas Tech University, Amanda Arner, August 2012
43
food visits and was the first bowl to be eaten from in a trial 58% of the time. The
second most-visited bowl for meals and first meals in a trial was bowl two, with 24%
of successful food visits and 23% of first meal visits.
Exploratory Behavior and Experience
Similar to the findings of the previous experiment, strong trends were seen in
both the total path length and time spent in the margin (Figure 2). Total path length
shows a strong decreasing trend for both experimental and control groups, with a
plateau beginning at trial 3. Conversely, time spent in the margin shows an increasing
trend for both groups across all trials. Toads in neither the control nor experimental
group tried to escape during trials, which may provide valuable insights for
interpreting their behavior in the arena. There were no changes in the latency to leave
the origin for either group as experience with the arena increased (Figure 3), which is
consistent with the findings of the first experiment.
Foraging Behavior and Experience
For both the time to eat the first mealworm and the tortuosity of the path to
reach such mealworm there is a strong decreasing trend during trials 1-5. Trial 6,
however, exhibits higher variation without fitting the trend from previous trials. There
were no apparent trends in the number of mealworms eaten per trial with increased
experience in the arena. For the experimental group, the total number of bowls
encountered and the number of unique bowls encountered seemed to decrease slightly
over time, however this trend may represent random variation in the small sample size.
Texas Tech University, Amanda Arner, August 2012
44
The control group showed a sharp decrease in both total and unique bowl encounters
between trials 1 and 2, but then a more shallow or non-existent decrease for the
remainder of the experiment.
Use of Spatial and Visual Cues to Locate Bowls
During the final trial of the experiment we moved the bowls to different
locations to determine if toads were using the relative location of the food in the arena
or associating the bowls with the presence of food. Of the six toads in the
experimental group, four visited old bowl locations first, one visited a new bowl
location first, and one did not encounter any bowls or bowl locations during the last
trial. In contrast, in the control group, one toad went to an old bowl location first, three
toads went to new bowl locations first, and four toads did not encounter any bowls or
bowl locations during this trial.
Discussion
The toads in the laboratory experiment acted both similarly and differently
from the toads in the field experiment. Toads in both experiments showed a marked
decrease in exploratory behavior, as defined by movement, as experience with the
arena increased. Interestingly, toads in the laboratory showed an increase in time spent
in the margin over trials probably due to their marked decrease in movement. In
general, toads in the control group would leave the origin, hop across the arena, and sit
on a cinder block or brick for the remainder of the trial. This marked decrease in
overall movement suggests a general lack of motivation to explore. Toads in the
Texas Tech University, Amanda Arner, August 2012
45
experimental group exhibited similar behaviors, except that they would visit food
bowls during the beginning of the trial then move to a block and remain stationary
until the trial was over.
The most interesting point to note, I think, is the fact that none of the toads
ever tried to escape from the exploratory arena. This raises the question of whether the
differences in behavior between the two populations are due to their native vs.
invasive origins or to the effects of captivity. By being kept under laboratory
conditions the toads’ decision-making processes could be affected when it comes to
spatial use. The difference in escape behavior highlights an unexpected confounding
effect of laboratory housing, which we cannot tease apart in this experiment.
Conducting this experiment with wild-caught toads in Florida would help disentangle
the effects of captivity from the population of origin for looking at differences in
learning abilities.
Overall, this experiment uncovers signs of potential differences in behavior
between the native and invasive populations. For instance, the toads in the laboratory
decreased their movement much more rapidly than the toads from the field experiment
and did not attempt to escape. One toad in the control group failed to even leave the
origin point during the sixth trial, and many toads simply did not move around as
much as the toads in the field experiment. These findings are contradictory to our
prediction that toads from the invasive range would be faster learners and show a more
rapid decrease in exploration and increase in foraging success than toads from the
Texas Tech University, Amanda Arner, August 2012
46
native range. Again, this could be an artifact of being in captivity, which would need
further testing to be conclusive.
Texas Tech University, Amanda Arner, August 2012
47
Figure 3.1 Experimental Arena Setup for Laboratory Experiment
The experimental arena in the laboratory at Texas Tech University was designed to be
an exact replica of the arena used during the field experiment in Gamboa, Panama. For
substrate, peat moss was used instead of leaf litter and other organic materials
collected from the field.
Texas Tech University, Amanda Arner, August 2012
48
Table 3.1 Observation Values for Behavior During First Experimental Trial
Control Experimental
1st trial Mean SE Mean SE p TM 600.98 190.81 693.69 440.89 0.95 TPL 4372.91 1602.39 5148.15 2595.33 0.50 TMAR 2005.65 438.82 1676.68 836.93 0.36 LO 213.21 65.26 149.32 78.54 0.06 TB 142.02 139.78 472.14 508.26 0.11 BE 16.50 6.19 19.83 12.98 0.70 UBE 3.00 0.93 4.50 1.38 0.04 ESC 0.00 0.00 0.00 0.00 - *p values found using a t-test or Mann-Whitney U test, where appropriate Observations for the exploration variables did not differ significantly between the
control and experimental groups for the first trial, indicating that the toads had similar
behaviors in an environment where they had no previous experience.
Texas Tech University, Amanda Arner, August 2012
49
Table 3.2 Degree of Preference for Bowls Encountered