FITNESS EFFECTS OF KINSHIP DEPEND ON ECOLOGICAL CONTEXT IN THE 

AMERICAN TOAD (ANAXYRUS AMERICANUS) 

 

By 

 

Sara Christine Garnett 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

A DISSERTATION 

 

Submitted to 

Michigan State University 

in partial fulfillment of the requirements 

for the degree of 

 

Integrative Biology—Doctor of Philosophy 

Ecology, Evolutionary Biology, and Behavior—Dual Major 

 

 

FITNESS EFFECTS OF KINSHIP DEPEND ON ECOLOGICAL CONTEXT IN THE 

AMERICAN TOAD (ANAXYRUS AMERICANUS)  

ABSTRACT 

 

 

By 

 

Sara Christine Garnett 

 

Studies of cooperation often ask how variation in kinship impacts the inclusive fitness 

payoffs of altruism. Hamilton’s rule defines this as a function of cooperation's costs and 

benefits, which in principle can vary widely across ecological contexts. In this 

dissertation, I address how kinship influences fitness, how selection balances fitness 

costs and benefits, and how the effects of kinship modulate the effects of other aspects 

of an organism’s environment using the American toad system.  

I first asked whether tadpoles use chemical cues to perceive differences in 

relatedness and whether this variation affects the response of several fitness proxies to 

environmental cues. I found that tadpole growth rates differed in response to cues of 

resource and kinship environment. In another experiment, growth rate differed based on 

cues of relative size, with larger tadpoles outperforming smaller partners. This was 

affected by kinship, at least for smaller tadpoles, who grew more rapidly with a sibling. 

This indicates that chemical cues communicate information necessary for tadpoles to 

perceive aspects of their environment, which interact with relatedness to affect fitness. 

I then investigated whether relatedness influences growth and development in 

experimental groups of tadpoles, and whether other factors – such as density and 

nutrient availability – impact the fitness benefits of grouping with kin. In our experiments, 

groups composed of full-sib kin reached metamorphosis faster and at a larger size than 

mixtures of different sib groups. The benefits of these fitness components were 

 

significant in more competitive, resource-scarce environments, but negligible in less-

competitive, resource-abundant environments.  Kinship and resource abundance have 

an interactive effect on the fitness components we measured.  

Finally, I assessed tadpole aggregation preferences for kin compared to non-kin 

in the presence and absence of predator cues. In the presence of predator cues, 

tadpoles may be more likely to choose kin over non-kin. While increased body mass 

might result in a tadpole being more likely to avoid conspecifics in the presence of 

predator cues, we saw that larger tadpoles potentially increased the probability of 

choosing kin over non-kin. While these results were not significant, indicating that 

predator avoidance is likely not the primary driver of kin aggregation behavior in this 

species, they are suggestive of a kin-selected benefit to grouping. 

Taken together, these results help us understand the contexts in which we might 

expect relatedness to affect fitness, which could further contribute to our understanding 

of the evolution of social behavior. This emphasizes that the fitness benefits of kin-

directed behavior are not identical in all circumstances, and that the ratio of costs to 

benefits may drive the evolution of different strategies depending on the environment. 

 

 

ACKNOWLEDGEMENTS 

 
 

I would like to thank my advisor, Tom Getty, and my committee members (Gary 

Mittelbach, Kay Holekamp, and Weiming Li) for their support, guidance, and feedback 

throughout my graduate career. Other members of the Getty lab (Idelle Cooper, Robin 

Tinghitella, Sarah Bodbyl, Eben Gering, Helen McCreery, Jean Johnson, Michael 

Kuczynski, and Cara Krieg) have provided valuable feedback and support with 

professional development, discussion of ideas, experimental design, and interpretation 

of results. I would also like to thank Eben Gering for assistance with statistical analysis, 

and Helen McCreery, Ava Garrison, Robert Logan, and Brendan Reid for helpful 

comments on drafts of my chapters. Al Manner, Taylor Schoen, Max Grezlik, Rebecca 

Benedict, and Olivia Guswiler worked as my undergraduate research apprentices at 

various points during my graduate career and played key roles in my lab and field 

projects. I am also grateful to Danielle Zoellner for her mentorship while working for 

KBS Academic Programs during my final year, and to the Graduate School, College of 

Natural Sciences, Ecology, Evolutionary Biology, and Behavior Program, Department of 

Integrative Biology, BEACON Center, and the NSF GK-12 program for funding me, my 

work, and my undergraduate assistants at various points during my program. I would 

also like to acknowledge the essential support provided by family and friends, including 

Cameron Manning, Casey terHorst, Michael Buchanan, Melissa Sienko, Mike Walker, 

Eyal Suseela, Caitlin Benallack, Leora Norkin, Tiffany VanPeenen, Daniel VanPeenen, 

Katie Coffman, Hanna Tsuhara, and Andrew Tengwall (and Kitty).  

 

 

iv 

TABLE OF CONTENTS 

 
 

LIST OF TABLES ................................................................................................. vi 
 
LIST OF FIGURES .............................................................................................. vii 
 
INTRODUCTION ...................................................................................................1 
 
CHAPTER ONE: CHEMICAL CUES OF RELATEDNESS MODIFY THE EFFECTS OF 
RESOURCE AVAILABILITY AND RELATIVE SIZE ON FITNESS INDICES 
IN THE AMERICAN TOAD (ANAXYRUS AMERICANUS) ....................................5 
Introduction ......................................................................................................5 
Methods ...........................................................................................................9 
Results ........................................................................................................... 15 
Discussion ..................................................................................................... 22 

 
CHAPTER TWO: THE BENEFITS OF DEVELOPING WITH KIN DEPEND ON 
DENSITY AND NUTRIENT AVAILABILITY IN THE AMERICAN TOAD (ANAXYRUS 
AMERICANUS) ................................................................................................... 26 
Introduction .................................................................................................... 26 
Methods ......................................................................................................... 31 
Results ........................................................................................................... 34 
Discussion ..................................................................................................... 41 

 
CHAPTER THREE: DOES THE PRESENCE OF PREDATOR CUES ALTER 
AGGREGATION PREFERENCES IN AMERICAN TOAD (ANAXYRUS AMERICANUS) 
TADPOLES? ....................................................................................................... 45 
Introduction .................................................................................................... 45 
Methods ......................................................................................................... 48 
Results ........................................................................................................... 51 
Discussion ..................................................................................................... 54 

 
CONCLUSION .................................................................................................... 59 
 
WORKS CITED ................................................................................................... 62 
 

 

v 

LIST OF TABLES 

 
 
Table 1.1. Hypotheses for the underlying relationship between tadpole fitness indices 
and chemical cues in the water. ............................................................................8  
 
Table 1.2. Models considered in the process of choosing the best-fit model by stepwise 
addition of each subsequent term. Each model’s p-value represents the results of a 
likelihood ratio test comparing it to the next simplest model. The best-fit model was 
reached when no additional interactions resulted in a significant improvement .. 16 
 
Table 1.3. Models considered in the process of choosing the best-fit model by stepwise 
addition of each subsequent term. Each model’s p-value represents the results of a 
likelihood ratio test comparing it to the next simplest model. The best-fit model was 
reached when no additional interactions resulted in a significant improvement ..20  
 
Table 2.1. Models considered in the process of choosing the best-fit model by stepwise 
addition of each subsequent term. Each model’s p-value represents the results of a 
likelihood ratio test comparing it to the next simplest model. The best-fit model was 
reached when no additional interactions resulted in a significant improvement ..36  
 
Table 3.1. Hypotheses for the underlying model governing how tadpoles aggregate in 
the presence of predators. ..................................................................................48  
 

 

vi 

LIST OF FIGURES 

 
 

Figure 1.1. A schematic of the experimental design testing how focal tadpoles respond 
to cues of relatedness and food availability. Water source tanks contain tadpoles from 
particular families receiving abundant resources (three pellets per tadpole) or scarce 
resources (one pellet) each day. Focal tadpoles are housed in individual containers 
where they receive three pellets a day. Focal tadpole containers received water from 
one of the source tanks depending on their assigned treatment. ........................ 11 
 
Figure 1.2. A schematic of the experimental design testing how focal tadpoles respond 
to cues of relatedness and size asymmetry. Two tadpoles, one larger in size than the 
other, were placed in the same container separated by a screen. In the kin treatment, 
both tadpoles were from the same clutch. In the non-kin treatment, each tadpole was 
from a different clutch. .........................................................................................14  
 
Figure 1.3. Relationship between focal tadpole growth rate and the cues in the water 
received by that tadpole. Panel A shows growth rates for all focal tadpoles receiving 
water from kin compared with tadpoles receiving water from non-kin. Panel B shows 
growth rate for all focal tadpoles receiving water from tanks with scarce resources 
compared with tadpoles receiving water from tanks with abundant resources. Panel C 
shows growth rates broken down by each combination of kinship and resource cues, as 
well as whether the focal tadpole eventually reached metamorphosis (boxes in gray 
indicate that they did metamorphose, while boxes in white indicate that they did not).
 ............................................................................................................................17  
 
Figure 1.4. Proportion of focal tadpoles reaching metamorphosis based on resource 
cues (water from tanks with scarce or abundant resources) received for (A) tadpoles 
receiving cues from kin and (B) tadpoles receiving cues from non-kin. Error bars 
indicate the standard error of proportion. ............................................................18  
 
Figure 1.5. Relationship between time (in days) and mass (in grams) for tadpoles in a 
pair separated by a screen. Filled circles and solid lines indicate the larger tadpole of a 
pair (L), open circles and dashed lines indicate the smaller tadpole (S). Points in black 
indicate tadpoles paired with a sibling (K), points in gray indicate tadpoles paired with a 
non-sibling (N). Lines show the prediction from the full mixed-effects model (± SE)
 ............................................................................................................................21  
 
Figure 2.1. A sigmoidal relationship between size or condition and fitness, where growth 
results from acquisition of limited resources. Dashed lines indicate the marginal fitness 
benefit that individuals receive from an increase in size or condition given their current 
state. Filled circles and thicker dashed lines indicate the larger individual while open 
circles with thinner dashed lines indicate the smaller individual. If the marginal benefit of 
a resource to a larger sibling is more than twice that of a smaller sibling, the smaller 
sibling should yield that resource (A). When the marginal benefit to the smaller sibling is 

vii 

more than twice that of the larger sibling, the larger sibling should yield instead (B).
 ............................................................................................................................ 27 
 
Figure 2.2. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the pure kin and mixed kin treatments. Lines show the 
prediction from the full mixed-effects model (± SE). Larger points indicate the treatment 
means. ................................................................................................................ 37 
 
Figure 2.3. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the low density and high density treatments. Lines show 
the prediction from the full mixed-effects model (± SE). Larger points indicate the 
treatment means. ................................................................................................ 38 
 
Figure 2.4. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the treatments with and without nutrient enrichment. Lines 
show the prediction from the full mixed-effects model (± SE). Larger points indicate the 
treatment means. ................................................................................................ 39 
 
Figure 2.5. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the pure kin and mixed kin treatments for (A) high density 
with no nutrient addition, (B) high density with nutrient addition, (C) low density with no 
nutrient addition, and (D) low density with nutrient addition. Lines show the prediction 
from the full mixed-effects model. Larger points indicate the treatment means. ..40  
 
Figure 3.1. Experimental setup for tests of tadpole preference. Two choices were 
provided in separate buckets, with airline tubing allowing water to flow from the bucket 
into the testing apparatus. Another piece of airline tubing allowed excess water to flow 
out the other end of the box. 20 mL of tap water or predator cue were added right before 
the tadpole was placed in the testing apparatus, and the side the tadpole chose was 
recorded. .............................................................................................................49  
 
Figure 3.2. Relationship between the probability that the focal tadpole chose kin, non-
kin, or showed no preference and (A) the presence or absence of predator cues and (B) 
the interaction between the presence or absence of predator cues and focal tadpole 
mass (in grams). Black indicates kin, gray indicates non-kin, and orange indicates no 
preference. Error is ± SE. .................................................................................... 53 
 
Figure 3.3. Relationship between the probability that the focal tadpole chose control tap 
water, non-kin, or showed no preference and (A) the presence or absence of predator 
cues and (B) the interaction between the presence or absence of predator cues and 
focal tadpole mass (in grams). Blue indicates control tap water, gray indicates non-kin, 
and orange indicates no preference. Error is ± SE. ............................................. 53 
 
Figure 3.4. Relationship between the probability that the focal tadpole chose control tap 
water, kin, or showed no preference and (A) the presence or absence of predator cues 
and (B) the interaction between the presence or absence of predator cues and focal 

viii 

tadpole mass (in grams). Blue indicates control tap water, black indicates kin, and 
orange indicates no preference. Error is ± SE. .................................................... 54

ix 

 

INTRODUCTION 

Kin can be an organism’s most intense competitors or its most reliable allies, 

depending on the circumstances. The development of kin selection theory and the 

concept of inclusive fitness provided a theoretical framework to help us understand how 

variation in genetic similarity among individuals may lead to differences in the strength 

of competition and cooperation (Gardner and West 2014).  Hamilton’s rule (Hamilton 

1964) provides a simple rule-of-thumb for finding the evolutionary optimum between 

selfishness and cooperation among related individuals. This specifies that a gene for 

some altruistic behavior will spread if the average relatedness between the individual 

performing an action and the recipient of that action multiplied by the benefit gained by 

the recipient due to that action is greater than the cost to the actor of performing that 

action. This concept has provided the basis for a great deal of sibling conflict theory, as 

it provides not only the conditions necessary for altruism but also, if reversed, the 

conditions under which selfishness will prevail.  

 

This last point is critical, as Hamilton’s rule highlights that it is not just relatedness 

that matters in determining whether an action is favored by selection, but also the ratio 

of fitness costs and benefits. A great deal of focus has been placed on how relatedness 

of interacting individuals affects the evolution of cooperation, with less consideration 

being given to how the context in which these interactions take place can alter the costs 

and benefits of cooperation, which is likely a significant part of the story (Bourke 2014). 

Rather than focusing simply on cases where there are dramatic fitness costs to the 

actor, theoretical work has also provided a way to think about more subtle costs of 

cooperation. The optimal amount of some resource to take depends on the marginal 

1 

fitness benefits an individual will receive from taking in more of the resource (Parker 

1989), which could have implications for the cooperation-competition balance.  

 

Individuals should be expected to vary the intensity of resource competition 

according to relatedness. In some cases, smaller siblings will increase inclusive fitness 

by deferring to larger, healthier kin, while in other cases larger siblings increase 

inclusive fitness by allowing smaller kin primary access to food (Parker et al. 1989, 

Yamamura and Higashi 1992). Kin interactions in line with these predictions have been 

seen in a variety of taxa (Willson et al. 1987, Tonsor 1989, Briskie et al. 1994, 

Johnstone and Roulin 2003, Boncoraglio et al. 2009, Madden et al. 2009, File et al. 

2011), and we expect them to prove adaptive in amphibians as well. This provides us 

with novel opportunities to investigate how variation in the environment influences the 

effects of kinship on development and fitness.  

Some amphibian species form aggregations in the larval stage, often composed 

of siblings (Waldman and Adler 1979, Waldman 1982, Waldman 1986, Pfennig et al. 

1993). These provide many benefits, such as anti-predator defense and social foraging 

(Watt et al. 1997, Sontag et al. 2006), but the dynamics become more complicated with 

limited resources and increased potential for competition. A variety of population 

responses have been seen when tadpoles interact at high densities (Steinwascher 

1979, Breden and Kelly 1982, Semlitsch and Caldwell 1982, Travis and Trexler 1986, 

Warner et al. 1991, Newman 1994), but one common trend is for larger tadpoles to 

inhibit the growth of smaller conspecifics when competing for limited resources. This 

presumably involves the use of a waterborne chemical agent and not just increased 

aggression, as we see these inhibitory effects even when tadpoles are only raised in 

2 

water from crowded conditions (Light 1967, Steinwascher 1978, Griffiths et al. 1993). 

There is evidence that chemical cues are used to recognize kin and influence behavior 

in some kin-aggregating species (Waldman 1985, Pfennig and Frankino 1997), and 

some cues seem to provide information about size (Lee and Waldman 2002), so this 

may provide a mechanism for communicating growth information to siblings. Although 

larger tadpoles may simply suppress growth in their competitors, the degree of 

relatedness between conspecifics has the possibility to alter the response. Relatedness 

has been considered in some experiments on growth and development (Hokit and 

Blaustein 1994, Hokit and Blaustein 1997, Saidapur and Girish 2001, Pakkasmaa and 

Aikio 2003, Gramapurohit et al. 2008), and the performance of smaller tadpoles is 

sometimes improved by grouping with kin (Jasienski 1988, Smith 1990). Overall, mixed 

results have been found and more work is necessary to understand how kin recognition 

might interact with group density, growth suppression, and environmental context to 

influence fitness.  

The American toad (Anaxyrus americanus) has been shown to preferentially 

aggregate with kin during the larval stage in both lab and field experiments (Waldman 

and Adler 1979, Waldman 1982). This preference is seen even when tadpoles are only 

exposed to water from related and unrelated individuals (Waldman 1986) and 

disappears when olfaction is blocked, suggesting a waterborne chemical cue is involved 

(Waldman 1985). There is no evidence of active cannibalism in this species (Heinen 

and Abdella 2005), suggesting there must be other less-obvious benefits to kin 

recognition. Toads are also explosive breeders that may reproduce in a variety of water 

bodies ranging in size and permanence, leading to significant natural variation in 

3 

relatedness, density, and resource availability within these breeding ponds. This makes 

them a good system for testing questions of how relatedness influences fitness within 

different contexts and whether chemical cues might mediate these effects. 

 

I am broadly interested in how kinship influences fitness, how selection balances 

fitness costs and benefits, and how the effects of kinship modulate the effects of other 

aspects of an organism’s physical and social environment. In this dissertation, I will 

address several of these questions using the American toad system. In Chapter One, I 

ask whether tadpoles use chemical cues to perceive differences in relatedness and 

whether variation in relatedness affects the response of several fitness proxies to 

environmental cues. In Chapter Two, I investigate whether relatedness influences 

growth and development in experimental groups of tadpoles, and whether other factors 

– such as density and nutrient availability – impact the fitness benefits of grouping with 

kin. Finally, in Chapter Three, I assess tadpole aggregation preferences for kin 

compared to non-kin in the presence and absence of predator cues. Taken together, 

these results help us understand the contexts in which we might expect relatedness to 

affect fitness, which could further contribute to our understanding of the evolution of 

social behavior. 

 

 

4 

CHAPTER ONE: 

CHEMICAL CUES OF RELATEDNESS MODIFY THE EFFECTS OF  

RESOURCE AVAILABILITY AND RELATIVE SIZE ON FITNESS INDICES  

IN THE AMERICAN TOAD (ANAXYRUS AMERICANUS) 

 

Introduction 

 

Animals extract information from their physical and social environment in many 

ways to make behavioral decisions. Many taxa rely on chemical cues and signals to 

communicate information relevant to fitness (Steiger et al. 2011). Chemical 

communication is used to locate mates or appropriate sites for reproduction (Buchinger 

et al. 2015), evaluate reproductive status or mate quality (Johansson and Jones 2007), 

warn conspecifics of predator risk (Schoeppner and Relyea 2009), and signify individual 

or group identity (Wyatt 2010). One aspect of identity that may impact social behavior is 

relatedness, as individuals may behave differently in the presence of kin versus non-kin. 

 

The ability to distinguish kin from non-kin is essential for kin-directed behavior 

(Penn and Frommen 2010), which can be crucial for increasing inclusive fitness. This 

idea holds that an individual’s fitness can be increased by benefitting relatives who 

share genes by common descent in addition to their own direct fitness (Hamilton 1964, 

Gardner and West 2014). Theory predicts that, in some cases, smaller siblings will 

increase inclusive fitness by deferring to larger, healthier kin, while in other cases larger 

siblings increase inclusive fitness by allowing smaller kin primary access to food (Parker 

et al. 1989, Yamamura and Higashi 1992). This pattern has been seen in several taxa 

where individuals may not be able to avoid interacting with relatives (Johnstone and 

Roulin 2003, Bulmer et al. 2007, Madden et al. 2009, File et al. 2011) and might prove 

5 

adaptive in amphibians as well, where some species are similarly dispersal-limited at 

the larval stage. 

Optimizing inclusive fitness in the manner described above requires reliable 

information about the condition, as well as the relatedness, of nearby individuals, and 

amphibians are a system well-suited to communicate this. There is evidence that 

chemical cues are used to recognize kin and influence behavior in some kin-

aggregating species (Waldman 1985, Pfennig and Frankino 1997). Some chemical cues 

also provide information about size (Lee and Waldman 2002), so this may provide a 

mechanism for communicating size information to siblings. Mass at metamorphosis is a 

commonly used proxy for fitness in larval amphibians (Earl and Whiteman 2015), and 

larger size can increase survival once toads enter the terrestrial environment (Sams and 

Boone 2010). Conspecific size may also provide information about whether individuals 

are likely to be related and whether consuming them provides an inclusive fitness 

benefit in species exhibiting cannibalism (Crossland et al. 2011, Dugas et al. 2016). If 

tadpoles vary the intensity or manner of competition based on competitor identity, body 

size will likely influence the fitness costs and benefits of competition. 

 

Another factor that should alter the intensity of competition is resource 

availability. Collins (1979) suggests that when resources are more limited but 

consistently present, we expect to see early-metamorphosing tadpoles emerge at the 

minimum size necessary for metamorphosis, in order to escape competition in the 

aquatic environment. Those that metamorphose later can reach a larger size given less 

competition for the consistent resource pool. When resources are more plentiful, all 

tadpoles should metamorphose at larger sizes, showing no relationship between size at 

6 

metamorphosis and time to metamorphosis (Collins 1979). Being able to perceive the 

resource potential of the environment should allow tadpoles to shift their development in 

a way that maximizes fitness, which may vary depending on the relatedness and size of 

competitors. 

Some evidence suggests that larger tadpoles can suppress the growth of their 

competitors (Light 1967, Steinwascher 1978, Griffiths et al. 1993), but this pattern may 

be affected by relatedness between individuals. In some experiments, the smallest 

tadpoles in a group have been larger at metamorphosis when grouped with kin versus 

non-kin (Jasienski 1988, Smith 1990). Here we propose that larger individuals 

aggregating with kin may be able to avoid producing a signal of growth suppression 

rather than suppressing all conspecifics, or they may produce a signal that affects kin 

behavior in other ways. Individuals may change their foraging behavior when in the 

presence of kin or reduce the amount of time spent on interference or exploitation 

competition. It is important to understand how the relatedness of a group might 

influence the behaviors and fitness strategies we see, and which factors affect the costs 

and benefits of these strategies. This would help us to understand how these different 

contexts influence fitness, allowing us to better predict how individuals are likely to 

respond given their circumstances.  

The American toad (Anaxyrus americanus) has been shown to preferentially 

aggregate with kin during the larval stage in both lab and field experiments (Waldman 

and Adler 1979, Waldman 1982). This preference is seen even when tadpoles are only 

exposed to water from related and unrelated individuals (Waldman 1986) and 

disappears when olfaction is blocked, suggesting a waterborne chemical cue is involved 

7 

(Waldman 1985). There is no evidence of active cannibalism in this species (Heinen 

and Abdella 2005), suggesting there must be other less-obvious benefits to kin 

recognition. American toads are also explosive breeders that may reproduce in a variety 

of water bodies ranging in size and permanence, leading to significant natural variation 

in relatedness, density, and resource availability within these breeding ponds. This 

makes them a good system for testing questions of how relatedness influences fitness 

within different contexts and whether chemical cues might mediate these effects. 

We devised a set of experiments to determine whether differences in relatedness 

would be perceived by tadpoles and result in differences in fitness indices. We also 

inquired whether other aspects of the tadpole’s physical and social environment would 

influence this relationship. We first considered the impacts of differences in kinship on 

growth and survival when perceived resource availability to conspecifics in the shared 

water source differed between treatments. Our second experiment considered whether 

a tadpole’s relative size influenced the effect of kinship on growth and survival. In both 

cases, focal tadpoles received water from these different treatments, but treatment 

groups did not vary in the containers in which they were housed or the amount of food 

they received.  

 

Table 1.1. Hypotheses for the underlying relationship between tadpole fitness indices 
and chemical cues in the water.  

H0: Chemical cues do not impact tadpole growth and survival 
H1: Tadpoles use chemical cues to increase direct fitness 
Prediction 1: Tadpoles grow less rapidly with reduced resource availability to 
conspecifics across treatments regardless of kinship 
Prediction 2: Smaller tadpoles grow less rapidly regardless of kinship 
H2: Tadpoles use chemical cues to increase inclusive fitness 
Prediction 1: Tadpole growth with reduced resource availability to conspecifics varies 
with kinship treatment 
Prediction 2: Smaller tadpoles grow more rapidly when partnered with kin   

8 

 

We considered several possibilities about the underlying influence of waterborne 

chemical cues on tadpole fitness (Table 1.1). If tadpoles are not able to perceive 

differences in the kinship and growth rate of conspecifics based on water alone, we 

should not expect to see any differences in growth rate between treatments. If chemical 

cues in water do impact tadpole fitness, they could simply result in improved 

performance by tadpoles who perceive themselves to be in a superior environment with 

no effect of relatedness, or information about body condition and food availability in 

transferred water could interact with relatedness to lead to differences in growth and 

survival in the presence of kin cues. By manipulating relatedness as well as other 

factors that could influence inclusive fitness, we hoped to determine how the 

interactions between them work to increase tadpole fitness. 

 

Methods 

Experiment 1: Relatedness and Food Levels 

To test our first prediction for each of hypotheses 1 and 2 (Table 1.1), we set up 

an experiment in which we exposed individual focal tadpoles to water from different 

source tanks. These water source tanks consisted of different tadpole clutches receiving 

food at either abundant or scarce levels, while all focal tadpoles received resources at 

the abundant level. This allowed us to test whether potential chemical cues in the water 

could affect focal tadpole growth and development, resulting in different patterns based 

on perceived group relatedness and resource availability. 

We collected amplexed pairs of adult American toads from breeding choruses in 

several locations in southwest Michigan in April 2012. These locations included the 

9 

Experimental Pond Laboratory and Lux Arbor Reserve at the Kellogg Biological Station 

(KBS). Pairs were isolated in buckets between 24 and 48 hours and given a chance to 

lay eggs, at which point we returned them to their capture site. Four pairs laid clutches 

for use in this experiment. Each clutch was split into multiple single-family aquarium 

tanks until needed.  

 

A subset of tadpoles was taken from each clutch and evenly split into two 20-

gallon aquaria of 300 tadpoles each. These tanks corresponded to resource level 

treatments within the experiment, based on Glennemeier and Denver’s food levels 

required for “normal growth” compared to “inducing competition” (Glennemeier and 

Denver 2002). Our abundant resource tanks received three pellets of Zoo Med aquatic 

frog and tadpole food per tadpole, consistent with Glennemeier and Denver’s food 

levels for “normal growth,” and the scarce resource tanks received one pellet per 

tadpole, a level that would be likely to induce competition.  

 

We then set up 240 focal tadpoles for the experiment, taken from the same 

clutch-specific tanks used to populate the treatment aquaria. Each tadpole was placed 

within an individual container containing approximately 465 mL of water from one of the 

aquarium tanks, depending on to which treatment the tadpole was assigned. Tadpoles 

received water from a tank containing either kin (same clutch) or non-kin (different 

clutch) receiving abundant resources (three pellets per tadpole per day) or scarce 

resources (one pellet per tadpole per day) (Fig. 1.1). All focal tadpoles received three 

pellets per day. Other than differing potential cues in water, the conditions were identical 

for all 240 focal tadpoles. We recorded each focal tadpole’s mass, snout-vent length 

(SVL) and Gosner stage (Gosner 1960) and placed it in the appropriate container. 

10 

Containers were set up in blocks of 120 on May 26 and 27. By this point tadpoles had 

reached developmental stage 25 (Gosner 1960), when independent feeding begins. 

The date a focal tadpole was added to its container was considered its start date for 

measuring the larval period. Sixty tadpoles were assigned to each treatment – receiving 

water from kin who were given abundant resources, kin who were given scarce 

resources, non-kin who were given abundant resources, or non-kin who were given 

scarce resources.  

 

Figure 1.1. A schematic of the experimental design testing how focal tadpoles respond 
to cues of relatedness and food availability. Water source tanks contain tadpoles from 
particular families receiving abundant resources (three pellets per tadpole) or scarce 
resources (one pellet) each day. Focal tadpoles are housed in individual containers 
where they receive three pellets a day. Focal tadpole containers received water from 
one of the source tanks depending on their assigned treatment. 
 
 

Focal and aquarium tank tadpoles were given their assigned food daily. Focal 

 

tadpoles’ mass, SVL, and Gosner stage were recorded weekly. We also changed the 

water in the focal containers during the weekly measurement. Tadpoles were removed 

11 

from the experiment once they had reached metamorphosis, defined as Gosner stage 

42 (at least one forelimb visible), or when they died, whichever came first. After 54 days, 

112 tadpoles reached metamorphosis, while 128 did not survive to metamorphosis. 

While some of this mortality may be the result of smaller tadpoles reacting poorly to 

being handled, such individuals are also likely to face increased mortality in natural 

populations (Heinen and Abdella 2005); while high, these mortality levels are likely 

within the range of variation tadpoles may experience depending on their pond. 

 

Statistical Analysis 

After determining that there was no significant nonlinearity in the relationship 

between mass and date for the focal tadpoles, we found each tadpole’s growth rate 

over the course of the experiment by regressing tadpole mass on date and taking the 

slope. We used the lme4 package (Bates et al. 2015) in R (R Core Team 2017) to test 

whether kinship and resource cues in water affected tadpole growth rate. To do this, we 

fitted our response variable of tadpole growth rate to fixed effects of family, kinship cue 

(kin or non-kin), resource cue (abundant or scarce resources), and whether the 

individual ultimately reached metamorphosis, as well as interactions between these 

main effects. We also included the water source tank and the source tank of each focal 

tadpole as random intercepts in the models, to account for differences among tanks that 

might otherwise obscure the relationships of interest to our study. We built our possible 

models by stepwise addition of each main effect and interaction. Each time a new term 

was added, we compared the new model to one without the additional term using 

likelihood ratio tests until we had identified the simplest model capturing significant 

12 

variance in the data. We used the normal approximation for determining parameter-

specific p-values. In addition, we used Fisher’s exact test to evaluate differences 

between treatments in the proportion of tadpoles surviving to metamorphosis. 

  

Experiment 2: Relatedness and Size Asymmetry 

To test our second set of predictions, we set up an experiment in which we 

paired two different-sized tadpoles, either from the same clutch or two different clutches. 

This allowed us to test whether potential chemical cues in the water could affect tadpole 

growth and development, differing based on perceived kinship and relative size 

asymmetry. Resources were held constant across treatments in this experiment.  

We collected adult toad pairs in late April and early May of 2013 as described for 

the first experiment, resulting in a total of eight clutches from both the KBS Experimental 

Pond Laboratory and Lux Arbor Reserve.  

 

To determine whether kinship and relative size interact to influence our fitness 

indices, we created containers to house two tadpoles in the same box while separating 

them by a screen. This allowed visual and waterborne cues to travel back and forth, 

giving tadpoles the opportunity to grow, develop, or change feeding habits in response 

to differences in cues if either individual had a fitness benefit to doing so, but prevented 

the larger tadpole from simply stealing food from the smaller tadpole. Boxes were set up 

in blocks of 60, for a total of 240 tadpoles, on May 28 and 29. Each tadpole’s mass, 

SVL, and Gosner stage was measured before being added to one side of a container 

containing 600 mL of water. Tadpoles were paired with a sibling (kin treatment) or an 

individual from another clutch (non-kin treatment), with one tadpole designated as the 

13 

larger individual compared to its smaller box partner (Figure 2). All tadpoles received 

three pellets of food daily.  

 

Figure 1.2. A schematic of the experimental design testing how focal tadpoles respond 
to cues of relatedness and size asymmetry. Two tadpoles, one larger in size than the 
other, were placed in the same container separated by a screen. In the kin treatment, 
both tadpoles were from the same clutch. In the non-kin treatment, each tadpole was 
from a different clutch.  
 
 

We measured mass, SVL, and Gosner stage for each block of tadpoles every 

 

other day for approximately three weeks, to capture initial growth trajectories for each 

tadpole. After that point, tadpoles were checked daily and measured and removed if 

they had reached metamorphosis (Gosner stage 42). The first tadpole reached 

metamorphosis on June 22, while the final metamorph emerged on July 23. After 56 

days, 221 tadpoles reached metamorphosis. 17 did not survive, and two survived the 

experiment but did not metamorphose.  

 

Statistical Analysis 

We tested whether kinship and relative size affected tadpole growth over time 

using the lme4 package in R to fit the response variable of tadpole mass to the day that 

measurement was taken. We also looked at fixed effects of relative size (larger or 

smaller) and kinship (kin or non-kin), as well as interactions between these main effects. 

Random effects of individual tadpole ID and family were also included in the models to 

14 

account for variation in individual growth trajectories or differences among families that 

might otherwise obscure the relationships among the fixed effects. As in the previous 

experiment, we used stepwise addition of each main affect and interaction to build our 

possible models for consideration. We used likelihood ratio tests to compare each new 

model to one without the additional term until we determined the simplest model that 

captured significant variance in the data. We then used the normal approximation for 

determining parameter-specific p-values. In addition, we used Fisher’s exact test to 

evaluate differences between treatments in the proportion of tadpoles surviving to 

metamorphosis. 

 

Results 

Experiment 1: Relatedness and Food Levels 

 

Our best-fit model of tadpole growth rate included whether the tadpole survived 

to metamorphosis, family, kinship cue, and resource cue. We expected that whether a 

tadpole lived to metamorphosis may be an indicator of some underlying quality, which 

could then influence how quickly that tadpole grew over the course of the experiment. 

The model also included interactions between reaching metamorphosis and family, and 

family and kinship cue (Table 1.2). This was a significant improvement over the next 

simplest model (χ2 = 7.88, p = 0.049), indicating that the included parameters predict 

observed variation in tadpole development. All models included focal tadpole source 

tank and the water source tank as random intercepts. 

 

 

15 

Table 1.2. Models considered in the process of choosing the best-fit model by stepwise 
addition of each subsequent term. Each model’s p-value represents the results of a 
likelihood ratio test comparing it to the next simplest model. The best-fit model was 
reached when no additional interactions resulted in a significant improvement.  
 

Model 

Growth Rate ~ Reached Metamorphosis + Family + Resource Cue + 
Kinship Cue + Family:Kinship Cue + Reached Metamorphosis:Family + 
(1|Tadpole Source Sank) + (1|Water Source Tank) 
Growth Rate ~ Reached Metamorphosis + Family + Resource Cue + 
Kinship Cue + Family:Kinship Cue + (1|Tadpole Source Sank) + (1|Water 
Source Tank) 
Growth Rate ~ Reached Metamorphosis + Family + Resource Cue + 
Kinship Cue + (1|Tadpole Source Sank) + (1|Water Source Tank) 
Growth Rate ~ Reached Metamorphosis + Family + Resource Cue + 
(1|Tadpole Source Sank) + (1|Water Source Tank) 
Growth Rate ~ Reached Metamorphosis + Family + (1|Tadpole Source 
Sank) + (1|Water Source Tank) 
Growth Rate ~ Reached Metamorphosis + (1|Tadpole Source Sank) + 
(1|Water Source Tank) 
Growth Rate ~ 1 + (1|Tadpole Source Sank) + (1|Water Source Tank) 

p-value 
0.049 

0.032 

0.89 

0.049 

0.007 

1.16e-14 

 

 
 

We saw that the growth rates of isolated tadpoles with plentiful food were 

influenced by the cues of kinship (Fig. 1.3A) and resource abundance (Fig. 1.3B) they 

received in water. The patterns became clearer when examined by treatment 

combination. Tadpoles receiving cues of abundant resources had lower growth rates 

over the experiment than those that received cues of scarce resources (Fig. 1.3B). 

Consistent with our second hypothesis (Table 1.1), this difference was particularly 

pronounced when comparing tadpoles that did and did not survive to metamorphosis, 

especially for those receiving cues from kin (Fig. 1.3C, t = -3.09, p = 0.002). We also 

saw focal tadpoles that eventually reached metamorphosis had higher growth rates 

compared to those that did not (Fig. 1.3C, t = 2.28, p = 0.02). We did not see an effect 

of kinship cues on growth rates when looking at tadpoles that reached metamorphosis, 

16 

but for tadpoles that failed to metamorphose we found a kinship effect consistent with 

our second hypothesis (Table 1.1) that they grew more rapidly when receiving cues 

from kin with scarce resources than from non-kin with scarce resources (Fig. 1.3C, t = -

2.64, p = 0.008).  

  

Figure 1.3. Relationship between focal tadpole growth rate and the cues in the water 
received by that tadpole. Panel A shows growth rates for all focal tadpoles receiving 
water from kin compared with tadpoles receiving water from non-kin. Panel B shows 
growth rate for all focal tadpoles receiving water from tanks with scarce resources 
compared with tadpoles receiving water from tanks with abundant resources. Panel C 
shows growth rates broken down by each combination of kinship and resource cues, as 
well as whether the focal tadpole eventually reached metamorphosis (boxes in gray 
indicate that they did metamorphose, while boxes in white indicate that they did not).  
 
 

While growth rates did not differ significantly between kinship treatments for 

those that reached metamorphosis, we did see differences among treatments in the 

proportion of tadpoles that reached metamorphosis (Fig. 1.4). Overall, a greater 

17 

proportion of tadpoles survived to metamorphosis when receiving water from tanks with 

scarce than abundant resources (p = 0.027). Of the tadpoles in the abundant resource 

cue treatment, those receiving water from kin metamorphosed in similar proportions to 

those in the scarce resource cue treatment (Fig. 1.4A, p = 0.36), while those in the non-

kin treatment were less likely to reach metamorphosis than those receiving cues of 

resource scarcity (Fig. 1.4B, p = 0.044).  

 
 

 

Figure 1.4. Proportion of focal tadpoles reaching metamorphosis based on resource 
cues (water from tanks with scarce or abundant resources) received for (A) tadpoles 
receiving cues from kin and (B) tadpoles receiving cues from non-kin. Error bars 
indicate the standard error of proportion.  
 
 
Experiment 2: Relatedness and Size Asymmetry 

 

Our best-fit model of tadpole mass contained time, relative size, family, and 

kinship as predictors, as well as interactions between size and family, size and time, 

time and family, family and kinship, and time, size, and kinship (Table 1.3). This was a 

18 

significant improvement over the previous model (χ2 = 4.69, p = 0.09), indicating that the 

included parameters predict observed variation in tadpole development. All considered 

models included random effects of focal tadpole identity on slope and intercept. 

 

Tadpoles showed differences in growth in response to visual and chemical cues 

of relative size, with larger tadpoles outperforming smaller partners. At the beginning of 

the experiment, there was a significant size gap between tadpoles assigned to the 

“larger” and “smaller” treatments (t = -10.44, p = 0.00) but not between kinship 

treatments (t = -0.22, p = 0.82). We saw significant increases in mass over the initial 

period of the experiment for larger tadpoles (Fig. 1.5, t = 25.76, p = 0.00), with kinship 

treatment having no effect on growth patterns for these individuals (Fig. 1.5, t = -0.08, p 

= 0.94). Smaller tadpoles grew less rapidly than their larger partners overall (Fig. 1.5, t = 

-2.72, p = 0.006).  

Contrary to what we saw in larger tadpoles, kinship treatment had a significant 

effect on smaller tadpoles, with individuals paired with non-kin growing even more 

slowly than those paired with kin (Fig. 1.5, t = -2.14, p = 0.03). Smaller tadpoles growing 

more rapidly if their partner was a sibling supports our second hypothesis (Table 1.1).   

 

Over 90% of tadpoles in this experiment reached metamorphosis. Larger 

tadpoles were more likely than smaller tadpoles to reach metamorphosis (p = 0.003). 

Tadpoles paired with kin did not differ between size treatments in the proportion 

reaching metamorphosis (p = 0.11), while tadpoles paired with a non-sibling were more 

likely to reach metamorphosis if they were the larger of a pair (p = 0.029).  

 
 
 

19 

Table 1.3. Models considered in the process of choosing the best-fit model by stepwise 
addition of each subsequent term. Each model’s p-value represents the results of a 
likelihood ratio test comparing it to the next simplest model. The best-fit model was 
reached when no additional interactions resulted in a significant improvement.  
 

Model 

Mass ~ Time + Relative Size + Family + Kinship + Relative Size:Family + 
Time:Relative Size + Time:Family + Family:Kinship + Time:Relative 
Size:Kinship + (1 + Time|Tadpole ID) 
Mass ~ Time + Relative Size + Family + Kinship + Relative Size:Family + 
Time:Relative Size + Time:Family + Family:Kinship + (1 + Time|Tadpole 
ID) 
Mass ~ Time + Relative Size + Family + Kinship + Relative Size:Family + 
Time:Relative Size + Time:Family + (1 + Time|Tadpole ID) 
Mass ~ Time + Relative Size + Family + Kinship + Relative Size:Family + 
Time:Relative Size + (1 + Time|Tadpole ID) 
Mass ~ Time + Relative Size + Family + Kinship + Relative Size:Family + 
(1 + Time|Tadpole ID) 
Mass ~ Time + Relative Size + Family + Kinship + (1 + Time|Tadpole ID) 
Mass ~ Time + Relative Size + Family + (1 + Time|Tadpole ID) 
Mass ~ Time + Relative Size + (1 + Time|Tadpole ID) 
Mass ~ Time + (1 + Time|Tadpole ID) 
Mass ~ 1 + (1 + Time|Tadpole ID) 

p-value 

0.09 

0.04 

0.003 

9.08e-07 

1.37e-10 

0.332 

<2.2e-16 
<2.2e-16 
<2.2e-16 

 
 

20 

 

Figure 1.5. Relationship between time (in days) and mass (in grams) for tadpoles in a 
pair separated by a screen. Filled circles and solid lines indicate the larger tadpole of a 
pair (L), open circles and dashed lines indicate the smaller tadpole (S). Points in black 
indicate tadpoles paired with a sibling (K), points in gray indicate tadpoles paired with a 
non-sibling (N). Lines show the prediction from the full mixed-effects model (± SE).  
 
 
 
 
 
 
 
 

21 

Discussion 

 

We found evidence that tadpoles were able to perceive differences in the kinship 

and size of conspecifics based on water exposure alone, consistent with our predictions 

that chemical cues are sufficient to influence tadpole fitness. In our first experiment, we 

found that cues indicating conspecifics experiencing scarce resources resulted in more 

rapid growth of well-fed focal individuals than cues indicating abundant resources for 

conspecifics (agreeing with Table 1.1, H1, P1). We also found differences between 

kinship cue treatments in tadpoles who died before reaching metamorphosis (Table 1.1, 

H2, P1). In the second experiment, larger tadpoles grew more rapidly than paired 

smaller tadpoles. While kinship did not affect the growth rates of larger tadpoles, smaller 

tadpoles performed better when their partner was a sibling (Table 1.1, H2, P2). In both 

experiments, we saw variation in proportion surviving to metamorphosis, with tadpoles 

in the treatment resulting in slower growth being less likely to metamorphose when 

receiving cues from non-kin. This indicates that kinship can determine how the 

perceived environment influences fitness, consistent with our second hypothesis (Table 

1.1).  

It might seem counterintuitive that tadpoles in the first experiment grew more 

rapidly when receiving cues of scarcity, but the treatments refer to the resource levels in 

the tanks providing the focal tadpole’s water. All focal tadpoles were fed at the abundant 

resource level, although they may use some type of cue in the water to inform them 

about resource availability to conspecifics. We hypothesize that tadpoles in the scarce 

resource cue treatment perceive their resource patch to be of higher quality than the 

surrounding environment. This perceived difference might lead them to harvest these 

22 

abundant local resources as quickly as possible, as personal survival in a “harsher” 

environment may trump any benefits of sharing those resources with kin (Pfennig et al. 

1993).  

 

Another possibility is that tadpoles in the abundant resource source tanks are 

producing more waste with more food and the resulting water is of poor quality. 

However, the differences between kinship treatments we saw for tadpoles that did not 

survive to metamorphosis may be suggest they may not be simply failing to thrive. 

Tadpoles in most treatment combinations showed similar patterns of growth over time 

regardless of whether they reached metamorphosis, with a tendency for slower-growing 

tadpoles to fail to metamorphose. Those receiving water from kin in the abundant 

resource treatment, on the other hand, did not show an increase in mass over time over 

the course of the experiment, despite starting out at similar sizes as tadpoles who failed 

to metamorphose in other treatments (Fig. 1.3C). This is consistent with the hypothesis 

that tadpoles who are smaller and unlikely to reach metamorphosis may receive a 

greater inclusive fitness benefit from leaving resources for kin instead of competing for 

resources and attempting to reach metamorphosis themselves. 

 

Although it may seem costly to be around kin in these circumstances, we found a 

fitness benefit. While an overall greater proportion of tadpoles reached metamorphosis 

when receiving water from tanks with scarce than abundant resources (54% vs. 39%), 

we saw that there was no significant difference in survival to metamorphosis between 

resource cue conditions when looking only at tadpoles in the kin cue treatments (Fig. 

1.4). In natural populations, the proportion of tadpoles that reach metamorphosis when 

around kin may be further increased by siblings who were unlikely to reach 

23 

metamorphosis themselves competing less intensely for available resources. This is 

consistent with previous work showing that there is an inclusive fitness benefit to being 

around kin even for those individuals who are unlikely to successfully compete (Pfennig 

1997, Dugas et al. 2016).  

 

In the second experiment, we saw that larger tadpoles in a pair did not appear to 

be affected by kinship treatment, but smaller tadpoles grew more rapidly when their 

partner was a sibling (Fig. 1.5). A chemical mechanism seems likely; as tadpoles were 

separated by screens, this cannot simply be due to larger tadpoles competing more 

intensely with unrelated individuals for available food. This suggests that previously 

seen patterns of larger tadpoles suppressing the growth of smaller tadpoles (Light 1967, 

Steinwascher 1978, Griffiths et al. 1993) may be sensitive to specific environmental 

factors rather than being a uniform phenomenon.   

If larger tadpoles do not uniformly suppress the growth of smaller tadpoles, this 

raises questions about how this can be avoided between kin. It is possible that tadpoles 

may be able to avoid releasing the chemical cues responsible for growth suppression 

when surrounded by kin, but another alternative is that smaller tadpoles are able to 

avoid responding to sibling cues. Other toad species produce a greater variety of 

bufotoxins, not just greater amounts, as the density of conspecifics increases (Bókony 

et al. 2016). If variation in bufotoxin composition has a genetic component, kin who are 

more genetically similar may also produce more similar toxins, which may have an 

immune function (Bókony et al. 2018). Tadpoles are unlikely to respond to compounds 

similar to “self,” suggesting a mechanism for avoiding suppression by larger kin.  

24 

 

In addition to these changes in early growth patterns, we also saw some variation 

in the proportion of tadpoles reaching metamorphosis across treatments. For tadpoles 

paired with a non-sibling, they were more likely to reach metamorphosis if they were the 

larger rather than the smaller of the pair. In both experiments, tadpoles showed greater 

differences between treatments in survival to metamorphosis when the cues they 

received came from non-kin. The fact that this pattern exists given only exposure to 

water and not direct interaction between individuals indicates that waterborne chemical 

cues are an important component of fitness benefits from kin aggregations. 

 

Tadpole growth and development varies with differences in environmental 

conditions; these experiments indicate that their perceptions of their environment are 

influenced by chemical cues in the water. Physical interactions with conspecifics and 

the surrounding habitat are necessary for understanding the full picture, but we have 

seen that exposure simply to water from different conditions communicates information 

about relatedness and relative size of conspecifics as well as resource availability. 

 

These differences influence the proportion of tadpoles surviving to 

metamorphosis as well as altering how quickly they get there, both of which affect 

fitness. Responding to variation in the environment requires tadpoles to be able to 

perceive this variation; understanding the role of chemical cues allows us to make 

sense of how variation in more complex environmental scenarios results in changes in 

fitness.   

 
 

 

25 

THE BENEFITS OF DEVELOPING WITH KIN DEPEND ON DENSITY AND 

NUTRIENT AVAILABILITY IN THE AMERICAN TOAD (ANAXYRUS AMERICANUS) 

CHAPTER TWO: 

 

Introduction 

Siblings can be an organism’s most intense competitors or its most reliable allies, 

depending on the circumstances. The development of kin selection theory and the 

concept of inclusive fitness provided a theoretical framework to help us understand the 

shifting interplay of competition and cooperation among individuals with more or less 

genetic similarity (Gardner and West 2014).  Hamilton’s rule (Hamilton 1964) provides a 

framework for us to think about how the evolutionary optimum between selfishness and 

cooperation among related individuals might be reached.  Since its formulation, most 

studies have focused on how the relatedness of interacting individuals impacts the 

evolution of cooperation, without much consideration to variation in the contexts in 

which these interactions occur.  More recently, researchers have begun to explore how 

context-dependencies can alter the costs and benefits of cooperation, which appears to 

be an important part of the story (Bourke 2014).  

Parker et al. (1989) showed how changes in the relative marginal benefits of a 

resource can change the level of conflict between siblings. Differences in size among 

siblings can be an important factor in these competitive relationships. When the benefits 

of increasing condition or size show diminishing returns (a concave fitness function, as 

in Jensen’s inequality in economics, Fig. 2.1A) and the marginal fitness benefits of the 

resource to the larger (or better condition) full sibs are less than twice the marginal 

benefits to the smaller (poor condition) sib, the larger full sib will increase its inclusive 

fitness more if the smaller individual is able to increase in size (Fig. 2.1B) 

26 

Figure 2.1. A sigmoidal relationship between size or condition and fitness, where growth 
results from acquisition of limited resources. Dashed lines indicate the marginal fitness 
benefit that individuals receive from an increase in size or condition given their current 
state. Filled circles and thicker dashed lines indicate the larger individual while open 
circles with thinner dashed lines indicate the smaller individual. If the marginal benefit of 
a resource to a larger sibling is more than twice that of a smaller sibling, the smaller 
sibling should yield that resource (A). When the marginal benefit to the smaller sibling is 
more than twice that of the larger sibling, the larger sibling should yield instead (B).  

 

Extending that logic, accelerating benefits of resource acquisition in very poor condition 

or size create a convex fitness function. This creates the possibility that the marginal 

fitness benefit of a resource to the larger sib is more than twice the benefit to the 

smaller sib. Given an average relatedness of 0.5 between full sibs, according to 

Hamilton’s rule, the smaller sib should then yield resources to the larger (Fig. 2.1A). 

This idea seems to be the foundation for the recent theoretical analysis by Thompson et 

al. (2017), showing that dominant sibs are sometimes selected to exploit weaker sibs 

because they are less resistant to exploitation than are unrelated individuals. If the 

fitness function is concave at low size or condition and convex at high size or condition, 

the resulting sigmoid fitness function (see Smallwood 1996 Fig. 5 and references) could 

select for the small sib to defer to the larger sib early on and then the larger sib might 

27 

defer to the smaller later. Patterns that look like “deferral” may be the result of dominant 

individuals selectively exploiting their weaker relatives or individuals competing less 

intensely for (or yielding) resources to a relative. Both could result in increased inclusive 

fitness for all individuals and should be common among sibs of different size. 

Unfortunately, measuring the marginal fitness benefit of a resource to siblings of 

differing size is challenging, particularly for seasonally-breeding vertebrates like the 

ones that inspired Parker’s model, and where any given unit of the resource results in a 

relatively small increase in fitness.  Consequently, many well-known examples of kin-

directed altruism center on cases where the costs are clear-cut and dramatic, as in 

cases of cannibalism or when an individual puts itself at risk to warn relatives (Sherman 

1977, Pfennig et al. 1993, Joseph et al. 1999, Pfennig 1999, Beavis et al. 2007). There 

are, however, many conditions in which individuals must account for more subtle costs 

of cooperation when interacting with kin, especially given limited methods for minimizing 

direct competition (Lambin et al. 2001). These scenarios require different strategies to 

evolve to maximize inclusive fitness. Theory predicts that individuals might adjust how 

intensely they compete for resources according to how related they are; for example, 

smaller siblings will increase inclusive fitness by deferring to larger, healthier kin, while 

in other cases larger siblings increase inclusive fitness by allowing smaller kin primary 

access to food (Parker et al. 1989, Yamamura and Higashi 1992). Studies from a 

variety of taxa do in fact show sibling interactions conforming to these predictions 

(Willson et al. 1987, Tonsor 1989, Briskie et al. 1994, Johnstone and Roulin 2003, 

Boncoraglio et al. 2009, Madden et al. 2009, File et al. 2011). We predict similar 

conditions should occur in amphibians, which would offer novel opportunities to 

28 

determine how environmental heterogeneity can induce and modulate kin effects on 

development.  

Because larval amphibians cannot leave their ponds until metamorphosis, they 

are likely to interact with other individuals present in the same water body.  Depending 

on population densities and distributions, local conspecifics might originate from one 

clutch of eggs, or from clutches of many different females. Some species of larval 

amphibians form aggregations, often composed of siblings (Waldman and Adler 1979, 

Waldman 1982, Waldman 1986, Pfennig et al. 1993), which may further structure the 

relatedness of individuals that interact during development. These aggregations provide 

many benefits, such as anti-predator defense and social foraging (Watt et al. 1997, 

Sontag et al. 2006), but their costs are more complicated, particularly under varying 

levels of local resource competition.  

A variety of population responses have been seen when tadpoles interact at high 

densities (Steinwascher 1979, Breden and Kelly 1982, Semlitsch and Caldwell 1982, 

Travis and Trexler 1986, Warner et al. 1991, Newman 1994), but one common trend is 

for larger tadpoles to inhibit the growth of smaller conspecifics when competing for 

limited resources, presumably through use of a waterborne chemical signal (Light 1967, 

Steinwascher 1978, Griffiths et al. 1993). Relatedness has sometimes been considered 

in experiments looking at growth and development (Hokit and Blaustein 1994, Hokit and 

Blaustein 1997, Saidapur and Girish 2001, Pakkasmaa and Aikio 2003, Gramapurohit et 

al. 2008), but mixed results have been found, and it is not always clear how kin 

recognition mechanisms might interact with density and growth suppression.  

29 

The American toad (Anaxyrus americanus) is one species that shows evidence 

of aggregating with kin in its larval stage, which is likely facilitated by an olfactory 

mechanism (Waldman and Adler 1979, Waldman 1982, Waldman 1986). Unlike some 

other species that recognize kin, there is no evidence that American toads actively 

cannibalize one another (Heinen and Abdella 2005), suggesting that they must receive 

other benefits from being able to distinguish kin from non-kin. A. americanus are 

explosive breeders, with mating and egg-laying taking place within a relatively small 

window in the spring, who use water bodies ranging from small temporary pools to large 

permanent ponds. This means that there can be a great deal of natural variation in 

tadpole density, and potentially relatedness, within a pond. American toads are 

therefore well suited for testing questions of how relatedness influences fitness and 

whether this relationship is context-dependent.  

We investigated whether relatedness would influence growth and development in 

experimental groups of tadpoles, and whether other factors – such as density and 

nutrient availability – would impact the fitness benefits of grouping with kin. We 

predicted that tadpoles in less competitive environments (e.g. lower density, higher 

nutrient availability) would metamorphose at their theoretically optimal size (Collins 

1979), regardless of how long they develop, as they are not constrained by competition 

for resources. In contrast, we expected tadpoles reared in competitive, resource-limited 

conditions would exhibit a trade-off between rate and size at metamorphosis, since 

early-emerging individuals will leave comparatively more competitive environments than 

late developing counterparts, assuming resource productivity is constant. If tadpoles do 

not suppress the growth of siblings or spend less time engaged in interference 

30 

competition with them, we expect that tadpoles in a pure kin environment will 

outperform individuals in mixed groups. Such a benefit from grouping with kin may even 

counteract the costs that come from potentially increased competition when density is 

higher or resources are more limited. It is also possible that more competitive 

environments will cause the costs of deferring to kin to outweigh the benefits, causing 

kin instead to behave more selfishly in order to maximize direct fitness.  

As a result, we predicted that (1) there will be a difference in size at 

metamorphosis between kin and mixed treatments, (2) tadpoles will be larger at 

metamorphosis in low density and higher nutrient availability treatments, and (3) the 

effect of kinship will vary depending on density and resource availability. By 

manipulating relatedness, resource availability, and population density in one 

experiment, we hoped to determine not just whether these factors matter individually, 

but how the interactions between them affect fitness.  

 

Methods 

To test these predictions, we set up a cattle tank experiment in which we 

manipulated the relatedness and density of tadpole groups and the nutrient availability 

in the tank. Adult American toads were collected during breeding choruses in late April 

of 2016. We collected pairs in amplexus from several locations at the Kellogg Biological 

Station (KBS) Experimental Pond Laboratory and Lux Arbor Reserve in southwest 

Michigan. Each pair was isolated in a bucket for approximately 24 to 48 hours until they 

finished laying their eggs, at which point they were returned to the location at which they 

were captured. We then transferred each clutch to a 20-gallon aquarium tank, until the 

31 

tadpoles reached a Gosner developmental stage of at least 26 (Gosner 1960), after 

which they are free-swimming and able to feed independently. 

 

We then set up 18 cattle tanks outdoors at the Pond Laboratory. Each tank 

contained approximately 75 L of water, with each tank being randomly assigned to 

treatments (pure or mixed kin, high or low tadpole density, and with or without nutrient 

enrichment). Each tank was seeded with pond water to promote the growth of 

periphyton to serve as a food source for tadpoles. In addition to this basic tank setup, 

half of the tanks received augmented nutrient levels (NaNO3-N 160 µg L-1, KH2PO4 10 

µg L-1) as described in Wood and Richardson 2009. We added an additional nutrient 

pulse every two weeks in order to create and maintain a nutrient-enriched environment 

with more plentiful resources. Tanks were populated with either pure kin – containing 

tadpoles from only one sibship – or mixed kin – containing tadpoles from two sibships. 

High density tanks contained 100 tadpoles, while low density tanks contained 50 

tadpoles. These densities are within the range seen in natural tadpole populations 

based on personal observation. A total of 1350 tadpoles were introduced to the cattle 

tanks.  

 

Before starting the experiment, we recorded each tadpole’s mass, snout-vent 

length (SVL) and Gosner stage and placed it in the appropriate tank. All tadpoles in a 

given tank were introduced on a single day between May 31 and June 4. The date on 

which tadpoles were added was considered the start date for that particular tank, and 

larval period was measured from that date. We monitored each tank daily, removing 

tadpoles once they had reached metamorphosis, defined as Gosner stage 42 (at least 

one forelimb visible). Metamorphosed tadpoles had their mass, SVL, and Gosner stage 

32 

recorded, as well as the number of days that had passed since they were added to the 

tank. Size at metamorphosis is a common fitness proxy in studies of larval amphibians 

(Earl and Whiteman 2015), and larger size also increases toad survival in the terrestrial 

environment (Sams and Boone 2010). After 105 days, 1148 tadpoles reached 

metamorphosis. 196 did not survive, and six survived the experiment but did not 

metamorphose; these individuals were excluded from assessments of mass at 

metamorphosis and time to metamorphosis.  

 

Statistical Analysis 

 

We used the lme4 package (Bates et al. 2015) in R (R Core Team 2017) to test 

the effects of kinship and other environmental factors on the mass of emerging 

tadpoles. Specifically, we fitted our response variable (mass at metamorphosis) to fixed 

effects of time to metamorphosis, density, nutrient addition treatment, and kinship, as 

well as interactions between these main effects. We also included family effects as 

random intercepts in the models, to account for any differences among clutches that 

might otherwise obscure the relationships of interest to our study. We built our possible 

models by stepwise addition of each main effect and interaction. Each time a new term 

was added, we compared the new model to one without the additional term using 

likelihood ratio tests until we had identified the best model capturing significant variance 

in the data. We considered this model to be reached when no additional terms resulted 

in a significant improvement to the model. We used the normal approximation for 

determining parameter-specific p-values. 

 

33 

Results 

Predictors of larval development 

Our best-fit model of mass at metamorphosis contained time to metamorphosis, 

density, nutrient addition, and kinship, as well as two-way interactions between time to 

metamorphosis and kinship, density and nutrient treatment, and a three-way interaction 

between time to metamorphosis, kinship, and nutrient treatment (Table 2.1). This was a 

significant improvement over the next simplest model (χ2 = 9.79, p = 0.007), indicating 

that all the included parameters are predictive of observed variation in tadpole 

development. In all models, family was included as a random effect. 

 

Effects of kinship on development 

In pure kin treatments, tadpoles that took longer to reach metamorphosis were 

also larger at metamorphosis than those with shorter larval periods (Fig. 2.2, t = 9.63, p 

= 0.00). We saw the opposite relationship in tanks containing mixed kin groups (Fig. 2.2, 

t = -5.19, p = 2.03e-7). The kinship treatment on its own did not predict size at 

metamorphosis independently of these interactions with larval period (t = 0.75, p = 

0.45), and/or density and nutrient availability (see below).  

 

Effects of density on development 

Tadpoles in the low density treatment were larger at metamorphosis than 

tadpoles raised at high density (Fig. 2.3, t = 2.98, p = 2.8e-3), but we did not find that 

adding an interaction with larval period and density treatment significantly improved our 

model of size at metamorphosis (χ2 = 1.25, p = 0.26).  

34 

 

Effect of nutrient availability on development 

The relationship between larval period and size at metamorphosis was 

significantly weaker in nutrient enriched treatments (Fig. 2.4, t = -2.87, p = 4.1e-3), 

resulting in late-metamorphosing tadpoles reaching similar sizes at metamorphosis 

regardless of the nutrient treatment. 

 

 

Combined effects of kinship, density, and resource availability on development 

We also found that the effect of kinship varied depending on the context. In 

addition to the overall relationship between time to metamorphosis and kinship 

treatment, we found an additional interaction between these variables and the nutrient 

treatment. At high densities with no nutrient addition we saw the greatest difference 

between pure kin and mixed kin treatments (Fig. 2.5A). Tadpoles grouped with siblings 

accumulated mass faster, and also reached metamorphosis earlier, than counterparts in 

mixed kin tanks. Tadpoles also see a benefit in increased size at metamorphosis in the 

low density tanks with no nutrient addition treatment (Fig. 2.5C). While significant 

differences between pure and mixed kinship treatments were seen with no nutrient 

addition (Fig. 2.5A, Fig. 2.5C), we found no significant difference between kinship 

treatments when more resources were available (Fig. 2.5B, Fig. 2.5D, t = 0.44, p = 

0.66). We did not find any significant interactions between density and these other 

factors. 

 

 

35 

Table 2.1. Models considered in the process of choosing the best-fit model by stepwise 
addition of each subsequent term. Each model’s p-value represents the results of a 
likelihood ratio test comparing it to the next simplest model. The best-fit model was 
reached when no additional interactions resulted in a significant improvement.  
 

Model 

p-value 

Mass ~ Time to Metamorphosis + Density + Nutrient + Kinship + 

0.007 

Kinship:Time to Metamorphosis + Density:Nutrient + Kinship:Time to 

 

Metamorphosis:Nutrient + (1|Family) 

Mass ~ Time to Metamorphosis + Density + Nutrient + Kinship + 

0.07 

Kinship:Time to Metamorphosis + Density:Nutrient + (1|Family) 

Mass ~ Time to Metamorphosis + Density + Nutrient + Kinship + 

3.24e-

Kinship:Time to Metamorphosis + (1|Family) 

05 

Mass ~ Time to Metamorphosis + Density + Nutrient + Kinship + 

0.09 

(1|Family) 

Mass ~ Time to Metamorphosis + Density + Nutrient + (1|Family) 

0.04 

Mass ~ Time to Metamorphosis + Density + (1|Family) 

Mass ~ Time to Metamorphosis + (1|Family) 

Mass ~ 1 + (1|Family) 

 

1.51e-

15 

<2.2e-

16 

 

36 

Figure 2.2. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the pure kin and mixed kin treatments. Lines show the 
prediction from the full mixed-effects model (± SE). Larger points indicate the treatment 
means. 

 

37 

Figure 2.3. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the low density and high density treatments. Lines show 
the prediction from the full mixed-effects model (± SE). Larger points indicate the 
treatment means. 
 

 

38 

Figure 2.4. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the treatments with and without nutrient enrichment. Lines 
show the prediction from the full mixed-effects model (± SE). Larger points indicate the 
treatment means. 
 

 

39 

Figure 2.5. Relationship between time to metamorphosis (in days) and mass at 
metamorphosis (in grams) for the pure kin and mixed kin treatments for (A) high density 
with no nutrient addition, (B) high density with nutrient addition, (C) low density with no 
nutrient addition, and (D) low density with nutrient addition. Lines show the prediction 
from the full mixed-effects model. Larger points indicate the treatment means.  
 

 

 

 

40 

Discussion 

 

We found strong evidence that 1) kinship affects aspects of development that are 

likely important for the fitness of A. americanus tadpoles and 2) the nature of these 

effects is highly context-dependent. First, we found an overall trend for tadpoles to 

reach larger sizes over a longer larval period in the pure kin treatment, while time did 

not predict size at metamorphosis in mixed kin treatments (Fig. 2.2). The fact that the 

duration of the larval period is not a strong predictor of size at metamorphosis in mixed 

kin treatments is consistent with Collins’s (1979) predictions if interference competition 

is playing a larger role there than in pure kin treatments. Collins (1979) suggests that 

this pattern can emerge because at any given time, increased interference competition 

results in larger, more competitive tadpoles metamorphosing as smaller, less 

competitive tadpoles are metamorphosing. As a result, we do not see any overall trend 

in size at metamorphosis over time. This could result from chemical suppression, where 

larger tadpoles suppress the growth of smaller individuals, or physical interference, 

where individuals are directly interfering with access to resources. While further study is 

needed to distinguish between these (non-exclusive) explanations, our results suggest 

that, overall, tadpoles may be experiencing less interference competition from siblings 

than from unrelated individuals.  

 

Additionally, we found that tadpoles were larger at metamorphosis in the low 

density treatments and that individuals metamorphosing earlier were larger in the 

nutrient addition treatment, but this difference disappeared with longer larval periods. 

Results from the nutrient-enriched treatments were consistent with the predictions of 

Collins (1979). Tadpoles whose tanks did not receive additional nutrient enrichment 

41 

were larger at metamorphosis when they took longer to develop, while larval period did 

not predict size at metamorphosis in tanks with higher nutrient availability (Fig. 2.4). 

When resources are limited, we expect that the metamorphosis and departure of rapidly 

developing individuals will free up resources that allow remaining tadpoles to grow to 

larger sizes. In contrast, if resource levels are consistently high, and do not limit growth 

or development, there should be little difference in the size at metamorphosis of early 

vs. late-developing tadpoles. Overall, tadpoles raised at low density reached 

metamorphosis at larger body sizes than those raised at high density, but with similar 

increases in mass over time (Fig. 2.3). 

 

While these overall trends are visible, we also found that environmental context 

influences the degree to which kinship drives differences in size at metamorphosis, 

which is a common proxy for fitness in developing tadpoles. We saw the greatest 

differences between kinship treatments in conditions where competition is likely more 

intense, given that more tadpoles are making use of less abundant resources (Fig. 

2.5A, Fig. 2.5C). The comparative benefits of kinship disappear, however, in the nutrient 

enriched treatments (Fig. 2.5B, Fig. 2.5D), particularly at low density. When resource 

competition is less intense, refraining from interference competition to benefit kin does 

not provide enough of an additional fitness benefit to elicit cooperative strategies. 

Instead, strategies that maximize direct fitness are favored.  

Previous work has found that larger tadpoles might suppress the growth of 

smaller tadpoles using chemical inhibition, a form of interference competition that can 

operate even without physical contact between individuals (Light 1967, Steinwascher 

1978, Griffiths et al. 1993, Crossland and Shine 2012, Bókony et al. 2016). Based on 

42 

our results, chemical inhibition may be occurring in mixed sibship groups, where larger 

tadpoles do not receive a fitness benefit from deferral to smaller ones. This may explain 

why we see that time to metamorphosis is typically a poor predictor of size at 

metamorphosis in mixed kin groups. In pure kin treatments, by contrast, larger tadpoles 

may avoid suppressing the growth of smaller kin. This would mirror previous findings 

from other species, in which smaller individuals also benefit more from being grouped 

with kin (Jasienski 1988, Smith 1990).  

Much work looking at kin discrimination in larval amphibians has been directed at 

species where there is strong potential for cannibalism. The advantage of such systems 

is that the costs and benefits considered under Hamilton’s rule are clear. One tadpole 

receives a meal, while the cannibalized tadpole loses its life, bringing its direct fitness to 

zero. In these systems, it is clear that selection can promote kin discrimination, thereby 

minimizing the inclusive fitness costs of cannibalizing kin, while taking advantage of the 

benefits of cannibalizing non-related individuals (Pfennig 1997). In contrast to this 

scenario, A. americanus seems to better represent a case where the cost-benefit ratio is 

not so straightforward, which is likely most common in nature. It is therefore important to 

understand how and whether kin selection operates in these situations, and whether the 

resulting outcomes are a matter of degree, with kin discrimination acting more weakly 

but in the same manner, or kind, with less drastic costs leading to kin discriminating 

behaviors acting in a different way. It is also likely that the costs-benefits ratio will be 

more dependent on environmental context when direct fitness costs are not as drastic, 

leading to the variation we see across treatments. 

43 

 

These more subtle cases are also important to consider because frequent 

interactions with kin or the ability to discriminate kin from non-kin do not automatically 

result in cooperation. If kin tend to be more similar to one another in the types of 

resources they use or their ability to take advantage of and hold those resources, we 

may see more intense competition between relatives than non-relatives, particularly if 

diverse resources are not available for niche partitioning (Martin and Garnett 2013, 

Foster and Briffa 2014). Related individuals may also be targeted for aggression if they 

are less likely to resist kin (Thompson et al. 2017), and individuals with low fitness 

potential may even benefit from an inclusive fitness perspective by being cannibalized 

(Pfennig 1997, Dugas et al. 2016). The present study confirms the importance of 

considering how environmental factors influence the costs and benefits of competing or 

cooperating with kin and how this is likely to shape the resulting behaviors. 

Our findings suggest that complex interactions between population structure and 

local environments are likely to shape the evolution of both kin discrimination, and the 

ways it is used to maximize inclusive fitness. The effects likely operate not only in 

American toads, but also in other species that can interact with both relatives and non-

relatives during development. We suggest that further studies that isolate, and 

manipulate, both relatedness and ecological contexts will offer powerful insights into the 

evolution of kin recognition, cooperation, and their modulation of inclusive fitness.  

 

 

44 

DOES THE PRESENCE OF PREDATOR CUES ALTER AGGREGATION 

PREFERENCES IN AMERICAN TOAD (ANAXYRUS AMERICANUS) TADPOLES? 

CHAPTER THREE: 

  

Introduction 

 

Predation acts as a strong pressure on many aspects of animal biology, including 

selecting for warning coloration (Stevens and Ruxton 2012) or camouflage (Merilaita et 

al. 2017), inducing and selecting for changes in body shape or size (Price et al. 2015), 

altering foraging behavior and activity level (Lima and Dill 1990), and influencing other 

aspects of phenotype and behavior (Caro 2005). One aspect that can be strongly 

shaped by predation pressure is grouping behavior. Animals may come together into 

groups of varying size to benefit from increased vigilance and alarm calls or increased 

ability to defend against potential predators (Edmunds 1974). In addition to these 

strategies, there are other ways that joining larger groups may result in a fitness benefit 

for an individual animal. 

 

 One theoretical explanation is that aggregations may form as “selfish herds” 

(Hamilton 1971), in which an individual reduces its own predation risk at the expense of 

others in the group. Through this dilution effect, any one individual is less likely to be 

attacked by a predator due to larger numbers of potential prey. It is also possible to 

achieve a “cooperative group,” in which an increase in prey density decreases predator 

success (Milinski 1979). Understanding the functions of animal groups may help us 

further understand how individuals make the choices they do in response to predation. 

 

Larval amphibians are an excellent system for investigating questions about 

responses to predation pressure. Tadpoles of many species face the possibility of 

predation from a variety of sources, and their responses include changes in their 

45 

behavior (Skelly and Werner 1990, Skelly 1992, Relyea 2004), life history (Skelly and 

Werner 1990, Relyea 2007), and morphology (Relyea and Hoverman 2003, Relyea 

2004, Schoeppner and Relyea 2008). Many of these antipredator defenses or 

avoidance behaviors are mediated by chemical cues (Kats et al. 1988, Fraker et al. 

2009, Schoeppner and Relyea 2005, Schoeppner and Relyea 2009) which may vary in 

specificity depending on the prey and predator species involved.  

 

The American toad, Anaxyrus americanus, is a species that is particularly well-

suited for asking these questions. Due to toxins in the skin it becomes increasingly 

unpalatable to predators throughout the tadpole stage (Brodie et al. 1978, Formanowicz 

and Brodie 1982), which may result in them being increasingly protected as they 

develop and get larger. It also prefers to aggregate with kin in lab and field experiments 

(Waldman and Adler 1979, Waldman 1982), identifying kin based on chemical cues in 

the water (Waldman 1985, Waldman 1986). In other species that show similar 

unpalatability and kin preferences, tadpoles school more cohesively in a single sibship 

group and in the presence of predator cues (Watt et al. 1997). We might therefore 

expect that in the American toad a “cooperative group” response to predation might be 

even more advantageous in a pure sibship group. 

 

If predators must learn that toad tadpoles are unpalatable, aggregating with kin 

may provide an additional benefit beyond the dilution effects that come from using a 

group to minimize predation risk. With kin groups, a predator that consumes a tadpole 

learns that it is distasteful and avoids consuming the original tadpole’s siblings, which 

tend to share genes by common descent. The idea of inclusive fitness holds that an 

individual’s overall fitness is more than just its direct genetic contribution to the next 

46 

generation (Hamilton 1964, Gardner and West 2014); a tadpole can still increase its 

inclusive fitness if being consumed by a predator increases the probability that its 

relatives survive. While predators should learn to avoid tadpoles regardless of the 

kinship composition of the group, the increased inclusive fitness benefit if the group is 

composed of relatives could select for kin aggregation in these tadpoles. If predators fail 

to learn (or learn imperfectly) that these tadpoles are unpalatable, we may still expect to 

see tadpole aggregation patterns more in line with the selfish herd model. 

 

We designed an experiment to investigate the process underlying how American 

toad tadpoles respond to predator cues (Table 3.1). By setting up a choice experiment 

where tadpoles were provided with water free of tadpoles, cues from a group of kin, and 

cues from a group of non-kin, we hoped to gain an understanding of tadpole 

aggregation patterns in the presence and absence of predator cues. Under the “selfish 

herd” model, we expect to see an increased preference for aggregating with tadpoles 

rather than remaining alone in open water when predator cues are present, with no 

difference between kin and non-kin. In the “cooperative group” model, given the 

possibility of predators learning to avoid unpalatable tadpoles and increased inclusive 

fitness benefit to tadpoles that group with kin, we expect to see a preference for related 

tadpoles in the presence of predator cues. Larger tadpoles may be at decreased risk of 

predation (Travis et al. 1985), possibly due to an increase in toxins across development 

(Brodie et al. 1978), which may lead to weaker aggregation preferences in larger 

tadpoles. This experiment may help us further our understanding of the patterns of 

schooling we see in the lab and field and how these aggregations contribute to tadpole 

fitness. 

47 

 
Table 3.1. Hypotheses for the underlying model governing how tadpoles aggregate in 
the presence of predators.  

H0: Tadpole aggregation behavior does not change in response to predators 
H1: Tadpole aggregation in the presence of predators follows the “selfish herd” 
model 
Prediction: Focal tadpoles will be more likely to choose to join groups of other 
tadpoles, with no preference for kin vs. non-kin, rather than remaining alone in open 
water 
H2: Tadpole aggregation in the presence of predators is kin-selected 
“cooperative group” 
Prediction: Focal tadpoles will be more likely to choose to join groups of kin over non-
kin, rather than remaining alone in open water 

 

Methods 

We collected pairs of adult American toads in amplexus from several locations in 

southwest Michigan in May 2014. These locations included the Experimental Pond 

Laboratory at the Kellogg Biological Station (KBS). Pairs were isolated in buckets 

between 24 and 48 hours and given a chance to lay eggs, at which point we returned 

them to their capture site. Nine clutches were laid for use in this experiment. Each 

clutch was split between two 10-gallon aquarium tanks, so we would later be able to 

investigate preferences for kin independent of familiarity. Tadpoles were maintained in 

these tanks until they reached developmental stage 25 as laid out by Gosner (1960), 

when independent feeding begins.  

Our testing apparatus was made up of a plastic storage container 13.5 cm x 35 

cm x 50 cm, divided in half by a partition extending three-quarters of the length of the 

box. Identical 5-gallon buckets provided water to either side via 3/16-inch airline tubing. 

An outflow hole at the end of the container opposite the buckets allowed for continued 

water flow during the trials (Fig. 3.1). 

 

48 

Figure 3.1. Experimental setup for tests of tadpole preference. Two choices were 
provided in separate buckets, with airline tubing allowing water to flow from the bucket 
into the testing apparatus. Another piece of airline tubing allowed excess water to flow 
out the other end of the box. 20 mL of tap water or predator cue were added right before 
the tadpole was placed in the testing apparatus, and the side the tadpole chose was 
recorded.  
 
 

We recorded each tadpole’s mass, snout-vent length (SVL), and Gosner stage 

 

before beginning preference trials. Tadpoles were provided with three sets of choices in 

randomized order in each portion of the experiment: kin vs. non-kin, kin vs. control, and 

non-kin vs. control. The “control” bucket contained only tap water, while other buckets 

contained water from one of the aquaria containing tadpoles. The water in the “non-kin” 

stimulus came from an unmixed family of tadpoles other than the one from which the 

focal tadpole originated, as opposed to a mixture of families. For a given sample of 

water from a particular tank, we set up the bucket and selected our focal tadpoles from 

the same family for a “kin” treatment or a different family for the “non-kin” treatment. We 

49 

know from previous experiments that water from other tadpoles is sufficient to cause 

changes in development and behavior.  

 

In this experiment, we used the water bug of the genus Belostoma (Hemiptera) 

as our predator (Swart and Taylor 2004). Predators were collected from ponds at the 

Experimental Pond Laboratory. We obtained predator cues by placing belostomatid 

individuals in a container with five tadpoles and using water from this container, which 

contained any potential odors or cues from the predator itself as well as cues of tadpole 

consumption, in our trials.  Shortly before introducing the focal tadpole to the 

experiment, we added 20 mL of tap water (in control trials) or 20 mL of a predator cue 

(in predator trials) to the test area. A tadpole was then added to the testing apparatus 

and their movement monitored for up to one minute. If a tadpole swam more than two 

inches up one of the unique sides of the apparatus, we recorded which side they chose 

and how long it took them. If they did not choose a side before the end of one minute, 

we recorded “no preference” for that tadpole.  

In total, 181 tadpoles were tested in at least one of the treatment combinations. 

For the kin vs. non-kin choice, 117 tadpoles were tested in both the presence and 

absence of predator cues, 13 only with predator cues, and 21 only without predator 

cues. For the kin vs. control choice, 87 were tested in both treatments, 39 only with 

predator cues, and 48 only without predator cues. For non-kin vs. control, 78 were 

tested in both treatments, 39 only without predator cues, and 49 only without predator 

cues. 

 

 

50 

Statistical Analysis 

We used a multinominal logistic regression analysis to determine whether there 

were significant differences in which choices tadpoles made depending on treatment. 

We used the nnet package (Venables and Ripley 2002) in R (R Core Team 2018) to 

test whether the presence of predator cues affected the probability of choosing each 

stimulus. We first fit our response variable (which stimulus each tadpole chose, or 

whether it showed no preference) to predator treatment for each of the three 

comparisons (kin vs. non-kin, kin vs. control, and non-kin vs. control). We also fit a 

version of the models including an interaction with focal tadpole body mass and 

predator treatment. We generated p-values using the Wald chi-square test.  

 

Results 

 

We asked whether the addition of predator cues to our experimental system 

altered the probability that focal tadpoles would choose one stimulus over the other in 

three sets of comparisons: kin vs. non-kin, tap water vs. non-kin, and tap water vs. kin. 

In all three comparisons, we saw that tadpoles were significantly less likely to show no 

preference compared to choosing one side over the other (Kin vs. Non-Kin: z = -3.51, p 

= 0.0004; Control vs. Non-Kin: z = -3.67, p = 0.0002; Control vs. Kin: z = -2.55, p = 

0.011), and this did not change with the addition of predator cues (Kin vs. Non-Kin: z = -

0.54, p = 0.59; Control vs. Non-Kin: z = -0.23, p = - 0.82; Control vs. Kin: z = -0.27, p = 

0.79). 

 

While we saw trends in the probability that tadpoles exhibited certain 

preferences, we observed high variation and few trends were significant. When 

51 

presented with a choice between water from kin and water from non-kin in the absence 

of a predator cue, tadpoles were slightly more likely to choose non-kin (Fig. 3.2A, z = 

0.47, p = 0.64), but were more likely to choose kin when exposed to a predator cue (Fig. 

3.2A, z = -0.73, p = 0.46). Tadpoles chose a control stimulus of plain water more 

frequently than non-kin under control conditions (Fig. 3.3A, z = -0.98, p = 0.33), but the 

presence of a predator cue increased the probability of choosing non-kin (Fig. 3.3A, z = 

1.26, p = 0.21). Tadpoles chose kin over plain water in both control (Fig. 3.4A, z = 0.48, 

p = 0.63) and predator (Fig. 3.4A, z = 0.78, p = 0.44) conditions.  

 

When considering mass as well as presence of predator cues in our model, we 

saw additional trends. In the choice between kin and non-kin in control conditions, 

tadpoles were more likely not to exhibit a preference as their mass increased (Fig. 3.2B, 

z = 1.32, p = 0.19), while in the predator treatment larger tadpoles were more likely to 

associate with kin than with non-kin or exhibit no preference (Fig. 3.2B). We saw a 

similar pattern where larger tadpoles were more likely not to have a preference between 

non-kin and tap water in the control treatment (Fig. 3.3B, z = 1.10, p = 0.27), although 

with this choice we saw that larger tadpoles were more likely to choose tap water over 

unrelated conspecifics (Fig. 3.3B). In the final set of comparisons, we again saw larger 

tadpoles being more likely to exhibit no preference between kin and tap water in the 

control treatment (Fig. 3.4B, z = 0.59, p = 0.56). In the predator cue treatment, the 

probability of choosing kin remained constant across mass (Fig. 3.4B, z = -0.38, p = 

0.71), while the likelihood of choosing tap water increased as mass did (Fig. 3.4B). 

Although none of these trends were significant, some of the patterns suggest interesting 

dynamics which we discuss below. 

52 

 

 

 

Figure 3.2. Relationship between the probability that the focal tadpole chose kin, non-
kin, or showed no preference and (A) the presence or absence of predator cues and (B) 
the interaction between the presence or absence of predator cues and focal tadpole 
mass (in grams). Black indicates kin, gray indicates non-kin, and orange indicates no 
preference. Error is ± SE. 
 

 

Figure 3.3. Relationship between the probability that the focal tadpole chose control tap 
water, non-kin, or showed no preference and (A) the presence or absence of predator 
cues and (B) the interaction between the presence or absence of predator cues and 
focal tadpole mass (in grams). Blue indicates control tap water, gray indicates non-kin, 
and orange indicates no preference. Error is ± SE. 
 

 

53 

Figure 3.4. Relationship between the probability that the focal tadpole chose control tap 
water, kin, or showed no preference and (A) the presence or absence of predator cues 
and (B) the interaction between the presence or absence of predator cues and focal 
tadpole mass (in grams). Blue indicates control tap water, black indicates kin, and 
orange indicates no preference. Error is ± SE. 
 

 

Discussion 

 

We found that tadpoles in our experiment were more likely to choose one 

stimulus over the other than remaining in their starting position, regardless of whether 

predator cues were present (Fig. 3.2A, Fig. 3.3A, Fig. 3.4A). As most of our other 

results were not statistically significant, we were unable to find support for either of our 

hypotheses governing how tadpoles might respond to the presence of predators (Table 

3.1). The results we did see show several interesting trends that suggest relatedness to 

a group may influence how tadpoles respond to predators if these patterns are indeed 

reflective of the underlying processes. 

 

There are several possible reasons why we were unable to detect significant 

differences in preference between treatments. The simplest possibility is that we were 

unable to capture significant patterns because those patterns do not exist. Our 

54 

experimental setup may have failed to capture the ways tadpoles behave in response to 

predators, or it may be that tadpoles do not use the types of chemical cues we provided 

to inform their antipredator responses. Our previous work has shown that differences in 

chemical cues are sufficient to cause differences in tadpole growth, but chemical cues 

alone may not exactly reproduce patterns seen when tadpoles may also interact 

physically with one another (previous chapters). While it is likely that our experimental 

setup provided tadpoles with some relevant information in the water, there may be other 

aspects of their environment or social interactions necessary for antipredator responses 

in the field. The location of the outflow tube (Fig. 3.1) may have also weakened the 

strength of predator cues the tadpoles received.  

 

 Even if our focal tadpoles were able to distinguish between kin and non-kin 

given our stimuli, they may require stronger antipredator cues to shape their responses. 

In this experiment, the focal tadpole could potentially identify the presence of a predator, 

but all stimulus water came from tanks where the tadpoles were not experiencing 

predator cues. In previous work, Bufo bufo tadpoles schooled more cohesively when 

with kin in the presence of predator cues (Watt et al. 1997). Tadpoles may choose 

which group to swim toward in response to a predator cue, but there may be stronger 

pressure to do so if the groups are also exhibiting a predator response. The rest of the 

tadpoles in the group may also need to be exposed to evidence of a predator to result in 

the most advantageous collective response.  

 

It is also possible that the identity of the predator will influence tadpole response. 

The presence of toxins in the skin of Anaxyrus americanus tadpoles renders them 

unpalatable to many common predators (Brodie et al. 1978, Formanowicz and Brodie 

55 

1982). If a predator must learn to avoid toad tadpoles through experience, schooling 

with kin would be beneficial from an inclusive fitness perspective. If the predator tries to 

eat a tadpole and learns to avoid other toad tadpoles (the first tadpole’s kin) as a result, 

there is still an inclusive fitness benefit to the predated tadpole. The water bug pierces 

the body of its prey and consumes the insides, bypassing the toxic skin altogether; it 

has been suggested that predators that feed in this way may not find tadpoles 

unpalatable, unlike those that bite or taste their prey, such as many vertebrate predators 

(Peterson and Blaustein 1992, Jara and Perotti 2009). If tadpoles are capable of 

distinguishing between different types of predators, the water bug would not impose the 

same selective pressure toward kin aggregation. It is unclear whether tadpoles are 

capable of that fine-scale predator discrimination or whether they simply respond to 

cues of conspecific stress, but, if so, it may result in the lack of significance we see in 

this experiment.  

 

We did see several patterns that suggest it could be advantageous to school with 

other tadpoles, particularly kin. Given a choice between tap water and non-kin, tadpoles 

were more likely to swim toward the tap water under the control scenario with a switch 

to preferring non-kin in the presence of predator cues. Tadpoles were more likely to 

choose kin over tap water in both scenarios, but the likelihood of making that choice 

increased in the predator treatment. When given a choice between kin and non-kin, 

tadpoles were slightly more likely to choose non-kin in the control but preferred kin 

when exposed to predator cues. Together, these results, while not statistically 

significant, suggest that there is some benefit to grouping with other tadpoles when 

predators are present, and choosing kin is preferable when they are available. 

56 

 

 In addition to these results, we also saw some potentially interesting (although 

not statistically significant) patterns when looking at how tadpole body mass influenced 

the probability of choosing each option. In all control scenarios, the probability that a 

focal tadpole did not choose a side increased with body mass. In the predator cue 

treatments, we saw that increased body mass correlated with increased probability of 

choosing tap water while probability of choosing non-kin declined. Probability of 

choosing kin over tap water was fairly constant across body mass, while increasing 

body mass actually correlated with increased likelihood of choosing kin over non-kin. In 

general, this suggests that larger tadpoles, who may be at decreased risk of predation 

from at least some predators (Travis et al. 1985), do not receive as great a benefit from 

schooling with conspecifics in response to predators, but we still saw larger individuals 

choosing kin. If this reflects an underlying pattern, this suggests that there may be 

additional benefits to aggregating with kin beyond those outlined in the “selfish herd” 

hypothesis (Hamilton 1971).  

 

Predation is likely not the most significant selective pressure driving the tendency 

for toad tadpoles to associate with kin, but it has the potential to be part of the story. 

Individuals derive a direct benefit from forming groups in response to predation 

pressure, as it reduces the likelihood that any individual is chosen, but there are 

additional aspects of American toad biology that may provide benefits for choosing kin 

over non-kin. To better test our hypotheses in future experiments, it would be beneficial 

to expose stimulus tadpoles to predator cues as well as testing whether the type of 

predator alters tadpole response. The physical complexity of the testing environment 

could also prove important, as it may provide alternate refuge for tadpoles in dangerous 

57 

situations. If the patterns we saw hold true, it is not enough to know whether an 

individual is exposed to predation to predict its prospects for survival. Understanding 

how relatedness modifies the effects of various selective pressures on fitness is 

necessary for a more complete understanding of how individuals navigate their physical 

and social environment and make choices influencing their behavior and survival.  

 

 

58 

 

CONCLUSION 

I found that tadpole growth rates differed in response to cues of resource and 

kinship environment. For tadpoles that did not reach metamorphosis, the combination of 

kinship and resource cues was important in predicting growth rates. Tadpoles receiving 

cues from kin in an environment with abundant resources showed evidence consistent 

with deferral. In another experiment, tadpole growth rate differed based on cues of 

relative size, with larger tadpoles outperforming smaller partners. This relationship was 

affected by kinship, at least for smaller tadpoles, who grew more rapidly if their partner 

was a sibling. In both experiments, tadpoles were equally likely to metamorphose 

across treatments when receiving cues from kin, while proportion surviving to 

metamorphosis was lower in some treatments receiving cues from non-kin. This 

indicates that chemical cues communicate information necessary for tadpoles to 

perceive aspects of their environment, which interact with relatedness to affect fitness. 

 

In a cattle tank experiment, I found strong evidence that kinship affects aspects 

of development that are likely important for fitness in American toad tadpoles and that 

these effects were context-dependent. Tadpoles grew more over the course of the 

experiment when grouped with siblings, but the greatest differences between kinship 

treatments were seen under conditions where competition had the potential to be more 

intense. Comparative benefits of kinship disappeared in nutrient enriched treatments, 

when cooperating with kin might not provide enough of an additional fitness benefit to 

select for such strategies. 

59 

 

In the final experiment, I saw that tadpoles given a choice between two stimuli 

(two of tap water, kin cues, and non-kin cues) were more likely to choose one side over 

the other rather than showing no preference. In the absence of predator cues, we saw 

that tadpoles may be more likely to choose tap water over non-kin, kin over tap water, 

and non-kin over kin. Increasing body mass also likely increased the probability of 

exhibiting no preference. In the presence of predator cues, tadpoles may be more likely 

to choose non-kin over tap water, continue to prefer kin over tap water, and choose kin 

over non-kin. While increased body mass might result in a tadpole being more likely to 

choose tap water in the presence of predator cues, we saw that larger tadpoles 

potentially increased the probability of choosing kin over non-kin. While many of these 

predator cue results were not significant, indicating that predator avoidance is likely not 

the primary driver of kin aggregation behavior in this species, they are suggestive of a 

kin-selected benefit to grouping beyond the selfish herd. 

Tadpoles grow and develop differently in response to variation in environmental 

conditions; these experiments indicate that their perceptions of their environment are 

influenced by chemical cues in the water. Physical interactions with conspecifics and 

the surrounding habitat are necessary for understanding the full picture, but we have 

seen that exposure simply to water from different conditions communicates information 

about relatedness and relative size of conspecifics as well as resource availability. 

 

These differences influence the proportion of tadpoles surviving to 

metamorphosis as well as altering how quickly they get there, both of which affect 

fitness. Responding to variation in the environment requires tadpoles to be able to 

perceive this variation; understanding the role of chemical cues allows us to make 

60 

sense of how variation in more complex ecological contexts results in changes in 

fitness.   

The way kinship modifies the effects of environmental conditions on fitness likely 

operates not only in American toads, but also in other species that can interact with 

relatives and non-relatives during development. We suggest that further studies that 

isolate, and manipulate, both relatedness and ecological contexts will offer powerful 

insights to the evolution of kin recognition, cooperation, and their modulation of inclusive 

fitness. Our findings confirm the importance of considering how environmental context 

influences the subtle costs and benefits of competing or cooperating with kin. They 

suggest that complex interactions between population structure and local environments 

are likely to shape the evolution of both kin discrimination, and the ways it is used to 

maximize inclusive fitness. Considering these factors in isolation is likely to lead to an 

incomplete understanding of the natural world, and these results thus represent an 

important contribution to enhanced understanding of how populations evolve in different 

contexts.  

 

 

61 

 

 

 

 

 

 

 

 

 

 

 

 

 

WORKS CITED 

62 

 
 

WORKS CITED 

Bates, D., M. Maechler, B. Bolker, and S. Walker. 2015. Fitting linear mixed-effects 
models using lme4. Journal of Statistical Software 67(1): 1-48. 
doi:10.18637/jss.v067.i01. 
 
Beavis, A. S., D. M. Rowell, and T. Evans. 2007. Cannibalism and kin recognition in 
Delena cancerides (Araneae: Sparassidae), a social huntsman spider. Journal of 
Zoology 271: 233-237.  
 
Bókony, V., A. M. Móricz, Z. Tóth, Z. Gál, A. Kurali, Z. Mikó, K. Pásztor, M. 
Szederkényi, Z. Tóth, J. Ujszegi, B. Üveges, D. Krüzselyi, R. J. Capon, H. Hoi, and A. 
Hettyey. 2016. Variation in chemical defense among natural populations of common 
toad, Bufo bufo, tadpoles: the role of environmental factors. J. Chem. Ecol. 42: 329-338.  
 
Bókony, V., B. Üveges, A. M. Móricz, and A. Hettyey. 2018. Competition induces 
increased toxin production in toad larvae without allelopathic effects on heterospecific 
tadpoles. Func. Ecol. 32(3): 667-675.  
 
Boncoraglio, G., M. Caprioli and N. Saino. 2009. Fine-tuned modulation of competitive 
behaviour according to kinship in barn swallow nestlings. Proc. R. Soc. B 276: 2117-
2123.  
 
Bourke, A. F. G. 2014. Hamilton’s rule and the causes of social evolution. Philo. Trans. 
R. Soc. Lond. B. 369(1642): 20130362.  
 
Breden, F. and C. H. Kelly. 1982. The effect of conspecific interactions on 
metamorphosis in Bufo americanus. Ecology 63(6): 1682-1689.  
 
Briskie, J. V., C. T. Naugler and S. M. Leech. 1994. Begging intensity of nestling birds 
varies with sibling relatedness. Proc. R. Soc. Lond. B 258(1351): 73-78.  
 
Brodie, E. D., Jr., D. R. Formanowicz, Jr., and E. D. Brodie, III. 1978. The development 
of noxiousness of Bufo americanus tadpoles to aquatic insect predators. Herpetologica 
34: 302-306.  
 
Buchinger, T. J., M. J. Siefkes, B. S. Zielinski, C. O. Brant, and W. Li. 2015. Chemical 
cues and pheromones in the sea lamprey (Petromyzon marinus). Frontiers in Zoology 
12: 32.  
 
Bulmer, E., P. Celis and D. Gil. 2007. Parent-absent begging: evidence for sibling 
honesty and cooperation in the spotless starling (Sturnus unicolor). Behavioral Ecology 
19: 279-284.  
 

63 

Caro, T. 2005. Antipredator defenses in birds and mammals. University of Chicago 
Press. 
 
Collins, J. P. 1979. Intrapopulation variation in the body size at metamorphosis and 
timing of metamorphosis in the bullfrog, Rana catesbeiana. Ecology 60(4): 738-749.  
 
Crossland, M. R., M. N. Hearnden, L. Pizzatto, R. A. Alford, and R. Shine. 2011. Why 
be a cannibal? The benefits to cane toad, Rhinella marina [=Bufo marinus], tadpoles of 
consuming conspecific eggs. Animal Behaviour 82(4): 775-782.  
 
Dugas, M. B., L. McCormack, A. Gadau, and R. A. Martin. 2016. Choosy cannibals 
preferentially consume siblings with relatively low fitness prospects. The American 
Naturalists 188(1): 124-131.  
 
Earl, J. E. and H. H. Whiteman. 2015. Are commonly used fitness predictors accurate? 
A meta-analysis of amphibian size and age at metamorphosis. Copeia 103(2): 297-309.  
 
Edmunds, E. 1974. Defence in animals. Longman, London.  
 
File, A. L., G. P. Murphy, and S. A. Dudley. 2011. Fitness consequences of plants 
growing with siblings: reconciling kin selection, niche partitioning and competitive ability. 
Proc. R. Soc. B. 279: 209-218. 
 
Formanowicz, D. R., Jr., and E. D. Brodie, Jr. 1982. Relative palatabilities of members 
of a larval amphibian community. Copeia 1982(1): 91-97.  
 
Foster, N. L. and M. Briffa. 2014. Familial strife on the seashore: aggression increases 
with relatedness in the sea anemone Actinia equine. Behav. Processes 103: 243-245.  
 
Fraker, M. E., F. Hu, V. Cuddapah, S. A. McCollum, R. A. Relyea, J. Hempel, and R. J. 
Denver. 2009. Characterization of an alarm pheromone secreted by amphibian tadpoles 
that induces behavioral inhibition and suppression of the neuroendocrine stress axis. 
Hormones and Behavior 55: 520-529.  
 
Gardner, A. and S. A. West. 2014. Inclusive fitness: 50 years on. Philos. Trans. R. Soc. 
Lond. B. 369(1642): 20130356.  
 
Glennemeier, K. A. and R. J. Denver. 2002. Role for corticoids in mediating the 
response of Rana pipiens tadpoles to intraspecific competition. Journal of Experimental 
Zoology 292: 32-40.  
 
Gosner, K. L. 1960. A simplified table for staging anuran embryos and larvae with notes 
on identification. Herpetologica 16(3): 183-190.  
 

64 

Gramapurohit, N. P., B. A. Shanbhag and S. K. Saidapur. 2008. Kinship influences 
larval growth and metamorphic traits of Bufo scaber in a context-dependent manner. 
Journal of Herpetology 42(1): 39-45. 
 
Griffiths, R. A., J. Denton and A. L.-C. Wong. 1993. The effect of food level on 
competition in tadpoles: interference mediated by protothecan algae? Journal of Animal 
Ecology 62(2): 274-279.  
 
Hamilton, W. D. 1964. The genetical evolution of social behaviour. I. Journal of 
Theoretical Biology 7: 1-16.  
 
Hamilton, W. D. 1971. Geometry for the selfish herd. Journal of Theoretical Biology 31: 
295-311. 
 
Heinen, J. T. and J. A. Abdella. 2005. On the advantages of putative cannibalism in 
American toad tadpoles (Bufo a. americanus): is it active or passive and why? Am. Midl. 
Nat. 153: 338-347. 
 
Hokit, D. G. and A. R. Blaustein. 1994. The effects of kinship on growth and 
development in tadpoles of Rana cascadae. Evolution 48(4): 1383-1388.  
 
Hokit, D. G. and A. R. Blaustein. 1997. The effects of kinship on interactions between 
tadpoles of Rana cascadae. Ecology 78(6): 1722-1735.  
 
Jara, F. G. and M. G. Perotti. 2009. Toad tadpole responses to predator risk: 
ontogenetic change between constitutive and inducible defenses. Journal of 
Herpetology 43(1): 82-88. 
 
Jasieński, M. 1988. Kinship ecology of competition: size hierarchies in kin and nonkin 
laboratory cohorts of tadpoles. Oecologia 77(3): 407-413.  
 
Johansson, B. G. and T. M. Jones. 2007. The role of chemical communication in mate 
choice. Biol. Rev. Camb. Philos. Soc. 82(2): 265-289.  
 
Johnstone, R. A., and A. Roulin. 2003. Sibling negotiation. Behavioral Ecology 14(6): 
780-786.  
 
Joseph, S. B., W. E. Snyder, and A. J. Moore. 1999. Cannibalizing Harmonia axyridis  
(Coleoptera: Coccinellidae) larvae use endogenous cues to avoid eating relatives. 
Journal of Evolutionary Biology 12: 792-797.  
 
Kats, L. B., J. W. Petranka, and A. Sih. 1988. Antipredator defenses and the 
persistence of amphibian larvae with fishes. Ecology 69: 1865-1870.  
 
Lambin, X., J. Aars and S. B. Piertney. 2001. Dispersal, intraspecific competition, kin 
competition and kin facilitation: a review of the empirical evidence. Dispersal (eds J. 

65 

Clobert, E. Danchin, A. A. Dhondt, J. D. Nichols), pp. 110-122. Oxford University Press, 
Oxford.  
 
Lee, J. S. F. and B. Waldman. 2002. Communication by fecal chemosignals in an 
archaic frog, Leiopelma hamiltoni. Copeia 2002(3): 679-686.  
 
Light, L. E. 1967. Growth inhibition in crowded tadpoles: intraspecific and interspecific 
effects. Ecology 48(5): 736-745.  
 
Lima, S. L. and L. M. Dill. 1990. Behavioral decisions made under the risk of predation: 
a review and prospectus. Canadian Journal of Zoology 68: 619-640. 
 
Madden, J. R., H. P. Kunc, S. English, M. B. Manser and T. H. Clutton-Brock. 2009. 
Calling in the gap: competition or cooperation in littermates’ begging behaviour? Proc. 
R. Soc. B 276: 1255-1262.  
 
Martin, R. A. and S. C. Garnett. 2013. Relatedness and resource diversity interact to 
influence the intensity of competition. Biol. Journal of the Linnean Soc. 110(3): 689-695. 
 
Merilaita, S., N. E. Scott-Samuel, and I. C. Cuthill. 2017. How camouflage works. Philos. 
Trans. R. Soc. Lond. B. Biol. Sci. 372(1724): 20160341.  
 
Milinski, M. 1979. Can an experienced predator overcome the confusion of swarming 
prey? Animal Behaviour 27: 1122-1126.  
 
Newman, R. A. 1994. Effects of changing density and food level on metamorphosis of a 
desert amphibian, Scaphiopus couchii. Ecology 75(4): 1085-1096.  
 
Pakkasmaa, S. and S. Aikio. 2003. Relatedness and competitive asymmetry – the 
growth and development of common frog tadpoles. Oikos 100: 55-64. 
 
Parker, G. A., D. W. Mock and T. C. Lamey. 1989. How selfish should stronger sibs be? 
The American Naturalist 133(6): 846-868.  
 
Penn, D. J. and J. G. Frommen. 2010. Kin recognition: an overview of conceptual 
issues, mechanisms and evolutionary theory. In: Kappeler, P. (eds) Animal Behaviour: 
Evolution and Mechanisms. Springer, Berlin, Heidelberg. 
 
Peterson, J. A. and A. R. Blaustein. 1992. Relative palatabilities of anuran larvae to 
natural aquatic insect predators. Copeia 1992: 577-584.  
 
Pfennig, D. W. 1997. Kinship and cannibalism. BioScience 47(10): 667-675. 
 
Pfennig, D. W. 1999. Cannibalistic tadpoles that pose the greatest threat to kin are most 
likely to discriminate kin. Proc. R. Soc. B 266: 57-61.  
 

66 

Pfennig, D. W. and W. A. Frankino. 1997. Kin-mediated morphogenesis in facultatively 
cannibalistic tadpoles. Evolution 51(6): 1993-1999.  
 
Pfennig, D. W., H. K. Reeve, and P. W. Sherman. 1993. Kin recognition and 
cannibalism in spadefoot toad tadpoles. Animal Behaviour 46(1): 87-94.  
 
Price, S. A., S. T. Friedman, and P. C. Wainwright. 2015. How predation shaped fish: 
the impact of fin spines on body form evolution across teleosts. Proc. R. Soc. B 282: 
20151428. 
 
R Core Team. 2017. R: A language and environment for statistical computing. R 
Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. 
 
R Core Team. 2018. R: A language and environment for statistical computing. R 
Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. 
 
Relyea, R. A. 2004. Fine-tuned phenotypes: tadpole plasticity under 16 combinations of 
predators and competitors. Ecology 85(1): 172-179.  
 
Relyea, R. A. 2007. Getting out alive: how predators affect the decision to 
metamorphose. Oecologia 152: 389-400.  
 
Relyea, R. A. and J. T. Hoverman. 2003. The impact of larval predators and competitors 
on the morphology and fitness of juvenile treefrogs. Oecologia 134: 596-604.  
 
Saidapur, S. K. and S. Girish. 2001. Growth and metamorphosis of Bufo melanostictus 
tadpoles: effects of kinship and density. Journal of Herpetology 35(2): 249-254.  
 
Sams, E. and M. D. Boone. 2010. Interactions between recently metamorphosed green 
frogs and American toads under laboratory conditions. Am. Midl. Nat. 163(2): 269-279. 
 
Schoeppner, N. M. and R. A. Relyea. 2005. Damage, digestion, and defence: the roles 
of alarm cues and kairomones for inducing prey defences. Ecology Letters 8: 505-512.  
 
Schoeppner, N. M. and R. A. Relyea. 2008. Detecting small environmental differences: 
risk-response curves for predator-induced behavior and morphology. Oecologia 154: 
743-754.  
 
Schoeppner, N. M. and R. A. Relyea. 2009. Interpreting the smells of predation: how 
alarm cues and kairomones induce different prey defences. Functional Ecology 23: 
1114-1121.  
 
Semlitsch, R. D. and J. P. Caldwell. 1982. Effects of density on growth, metamorphosis, 
and survivorship in tadpoles of Scaphiopus holbrooki. Ecology 63(4): 905-911. 
 

67 

Sherman, P. W. 1977. Nepotism and the evolution of alarm calls. Science 197(4310): 
1246-1253. 
 
Skelly, D. K. 1992. Field evidence for a cost of behavioral antipredator response in a 
larval amphibian. Ecology 73(2): 704-708.  
 
Skelly, D. K. and E. E. Werner. 1990. Behavioral and life-historical responses of larval 
American toads to an odonate predator. Ecology 71(6): 2313-2322.  
 
Smallwood, P. D. 1996. An introduction to risk sensitivity: the use of Jensen’s inequality 
to clarify evolutionary arguments of adaptation and constraint. American Zoologist 
36(4): 392-401.  
 
Smith, D. C. 1990. Population structure and competition among kin in the chorus frog 
(Pseudacris triseriata). Evolution 44(6): 1529-1541.  
 
Sontag, C., D. S. Wilson and R. S. Wilcox. 2006. Social foraging in Bufo americanus 
tadpoles. Animal Behaviour 72: 1451-1456.  
 
Steiger, S., T. Schmitt, and H. M. Schaefer. 2011. The origin and dynamic evolution of 
chemical information transfer. Proc. R. Soc. B 278: 970-979.  
 
Steinwascher, K. 1978. Interference and exploitation competition among tadpoles of 
Rana utricularia. Ecology 59(5): 1039-1046.  
 
Steinwascher, K. 1979. Competitive interactions among tadpoles: responses to 
resource level. Ecology 60(6): 1172-1183.  
 
Stevens, M. and G. D. Ruxton. 2012. Linking the evolution and form of warning 
coloration in nature. Proc. Biol. Sci. 279(1728): 417-426.  
 
Swart, C. C. and R. C. Taylor. 2004. Behavioral interactions between the giant water 
bug (Belostoma lutarium) and tadpoles of Bufo woodhousii. Southeastern Naturalist 
3(1): 13-24.  
 
Thompson, F. J., M. A. Cant, H. H. Marshall, E. I. K. Vitikainen, J. L. Sanderson, H. J. 
Nichols, J. S. Gilchrist, M. B. V. Bell, A. J. Young, S. J. Hodge and R. A. Johnstone. 
2017. Explaining negative kin discrimination in a cooperative mammal society. PNAS 
114(20): 5207-5212.  
 
Tonsor, S. J. 1989. Relatedness and intraspecific competition in Plantago lanceolata. 
The American Naturalist 134(6): 897-906.  
 
Travis, J. and J. C. Trexler. 1986. Interactions among factors affecting growth, 
development and survival in experimental populations of Bufo terrestris (Anura: 
Bufonidae). Oecologia 69: 110-116.  

68 

 
Venables, W. N. and B. D. Ripley. 2002 Modern Applied Statistics with S. Fourth 
Edition. Springer, New York. ISBN 0-387-95457-0 
 
Waldman, B. 1982. Sibling association among schooling toad tadpoles: field evidence 
and implications. Animal Behaviour 30: 700-713.  
 
Waldman, B. 1985. Olfactory basis of kin recognition in toad tadpoles. Journal of 
Comparative Physiology A 156: 565-577. 
 
Waldman, B. 1986. Preference for unfamiliar siblings over familiar non-siblings in 
American toad (Bufo americanus) tadpoles. Animal Behaviour 34: 48-53.  
 
Waldman, B. and K. Adler. 1979. Toad tadpoles associate preferentially with siblings. 
Nature 282: 611-613.  
 
Warner, S. C., W. A. Dunson and J. Travis. 1991. Interaction of pH, density, and priority 
effects on the survivorship and growth of two species of hylid tadpoles. Oecologia 88: 
331-339.  
 
Watt, P. J., S. F. Nottingham, and S. Young. 1997. Toad tadpole aggregation behaviour: 
evidence for a predator avoidance function. Animal Behaviour 54: 865-872.  
 
Willson, M. F., W. G. Hoppes, D. A. Goldman, P. A. Thomas, P. L. Katusic-Malmborg 
and J. L. Bothwell. 1987. Sibling competition in plants: an experimental study. The 
American Naturalist 129(2): 304-311.  
 
Wood, S. L. R. and J. S. Richardson. 2009. Impact of sediment and nutrient inputs on 
growth and survival of tadpoles of the Western Toad. Freshwater Biology 54: 1120-
1134.  
 
Wyatt, T. 2010. Pheromones and signature mixtures: defining species-wide signals and 
variable cues for identity in both invertebrates and vertebrates. J. Comp. Physiol. A 196, 
685-700.  
 
Yamamura, N. and M. Higashi. 1992. An evolutionary theory of conflict resolution 
between relatives: altruism, manipulation, compromise. Evolution 46(4): 1236-1239.  
 
 

 

 

69