DOES BEHAVIOURAL PLASTICITY CONTRIBUTE TO DIFFERENCES IN POPULATION GENETIC STRUCTURE IN WILD RABBIT POPULATIONS IN ARID AND SEMI-ARID AUSTRALIA? Mr Geoffrey Anthony de Zylva – B. App. Sc., Hons. School of Natural Resource Sciences Queensland University of Technology Submitted for the degree of Doctor of Philosophy (Science), in 2007.
151
Embed
DOES BEHAVIOURAL PLASTICITY CONTRIBUTE TO …eprints.qut.edu.au/16437/1/Geoffrey_de_Zylva_Thesis.pdfThe European rabbit, Oryctolagus cuniculus, ... territorial intrusion approach to
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
DOES BEHAVIOURAL PLASTICITY CONTRIBUTE TO DIFFERENCES IN POPULATION GENETIC
STRUCTURE IN WILD RABBIT POPULATIONS IN ARID AND SEMI-ARID AUSTRALIA?
Mr Geoffrey Anthony de Zylva – B. App. Sc., Hons. School of Natural Resource Sciences Queensland University of Technology
Submitted for the degree of Doctor of Philosophy (Science), in 2007.
Keywords
Oryctolagus cuniculus
European Rabbit
Australia
DNA
mtDNA
microsatellite
behaviour
flexible behaviour
genetic variability
metapopulations
genetic bottleneck
Abstract
The European rabbit, Oryctolagus cuniculus, was introduced to Australia in 1859
and quickly became a significant vertebrate pest species in the country across a wide
distribution. In arid and semi-arid environments, rabbit populations exist as
CH5 - GENERAL DISCUSSION.............................................................................................................111 POPULATION GENETICS..........................................................................................................111 BEHAVIOURAL ECOLOGY .......................................................................................................114 PEST MANAGEMENT ISSUES ...................................................................................................119 FUTURE DIRECTIONS OF RESEARCH AND CONCLUSION..........................................................123
APPENDIX 1 – LIST OF ALL RABBIT BEHAVIOURS ...............................................................................125 BIBLIOGRAPHY ..................................................................................................................................127
List of Tables and Figures FIGURE 1.1 - TYPES OF METAPOPULATION ............................................................................................ 6 FIGURE 2.1 – AREAS OF STUDY ............................................................................................................ 27 TABLE 3.1 – MICROSATELLITE PRIMERS ............................................................................................... 35 TABLE 3.2 – PCR AND ELECTROPHORESIS CONDITIONS (TA = ANNEALING TEMPERATURE)................... 38 TABLE 3.3 – POPULATION SAMPLE SIZES AT EACH LOCUS ..................................................................... 39 TABLE 3.4 – NUMBER OF ALLELES PER LOCUS PER POPULATION ........................................................... 41 TABLE 3.5 – MEAN ALLELIC STATISTICS ACROSS ALL LOCI FOR EACH POPULATION.............................. 42 TABLE 3.6 – SIGNIFICANT GENIC DIFFERENTIATION FOR POPULATION PAIRS ACROSS ALL LOCI ............ 43 TABLE 3.7 – MATRIX OF SIGNIFICANT GENIC DIFFERENTIATION BETWEEN POPULATION PAIRS ............. 44 TABLE 3.8 – PAIRWISE POPULATION FST VALUES.................................................................................. 45 TABLE 3.9 – SIGNIFICANCE OF PAIRWISE POPULATION FST VALUES...................................................... 45 FIGURE 3.1 – SORTED MEAN FIS............................................................................................................ 46 TABLE 3.10 – AMOVA SUMMARY TABLE ............................................................................................ 47 FIGURE 3.2 – AMOVA SUMMARY PIE CHART...................................................................................... 47 FIGURE 3.3 – RANDOMISATION OF PHIPT ............................................................................................. 48 FIGURE 3.4 - UPGMA TREE FOR NEI SIMILARITY MATRIX .................................................................... 49 TABLE 3.11 – SPECIES WITH REDUCED GENETIC DIVERSITY .................................................................. 51 TABLE 4.1 – SITE LOCATIONS................................................................................................................ 58 FIGURE 4.1 – VISION FIELD OF VIDEO CAMERA ..................................................................................... 60 TABLE 4.2 – BEHAVIOUR OBSERVED ON VIDEO..................................................................................... 62 TABLE 4.3 - WARREN COUNT DATA ..................................................................................................... 64 FIGURE 4.2 – 2001 RABBIT WEIGHT V SEX (TOTAL CAPTURES) ............................................................ 65 FIGURE 4.3 – 2002 RABBIT WEIGHT V SEX (TOTAL CAPTURES) ............................................................ 66 FIGURE 4.4 – MEAN DECOY WEIGHT .................................................................................................... 67 TABLE 4.4 – MEAN PERCENTAGE COVER............................................................................................... 67 FIGURE 4.5 – MEAN PERCENT COVER COMPARISON BETWEEN YEARS SITE 1 ........................................ 68 FIGURE 4.6 – MEAN PERCENT COVER COMPARISON BETWEEN YEARS SITE 2 ........................................ 69 FIGURE 4.7 – MEAN PERCENT COVER COMPARISON BETWEEN YEARS SITE 3 ........................................ 70 FIGURE 4.8 – SCATTERPLOT OF TOTAL BEHAVIOUR VS NUMBER OF RABBITS ...................................... 72 FIGURE 4.9 – MEAN PLOT OF SUM BEHAVIOUR PER RABBIT PER HOUR................................................ 73 FIGURE 4.10 – AGGRESSIVE BEHAVIOUR SITE 1 CONTROL 2001........................................................... 75 FIGURE 4.11 – AGGRESSIVE BEHAVIOUR SITE 1 EXPERIMENTAL 2001.................................................. 76 FIGURE 4.12 – AGGRESSIVE BEHAVIOUR SITE 2 CONTROL 2001........................................................... 77 FIGURE 4.13 – AGGRESSIVE BEHAVIOUR SITE 2 EXPERIMENTAL 2001.................................................. 78 FIGURE 4.14 – AGGRESSIVE BEHAVIOUR SITE 3 CONTROL 2001........................................................... 79 FIGURE 4.15 – AGGRESSIVE BEHAVIOUR SITE 3 EXPERIMENTAL 2001.................................................. 80 FIGURE 4.16 – AGGRESSIVE BEHAVIOUR SITE 1 CONTROL 2002........................................................... 81 FIGURE 4.17 – AGGRESSIVE BEHAVIOUR SITE 1 EXPERIMENTAL 2002.................................................. 82 FIGURE 4.18 – AGGRESSIVE BEHAVIOUR SITE 2 EXPERIMENTAL 2002.................................................. 83 FIGURE 4.19 – AGGRESSIVE BEHAVIOUR SITE 1 CONTROL 2002........................................................... 84 FIGURE 4.20 – AGGRESSIVE BEHAVIOUR SITE 2 EXPERIMENTAL 2002.................................................. 85 FIGURE 4.21 – 10MIN INTERVAL PLOT SITE 1 2001 ............................................................................... 87 FIGURE 4.22 – 10MIN INTERVAL PLOT SITE 2 2001 ............................................................................... 88 FIGURE 4.23 – 10MIN INTERVAL PLOT SITE 3 2001 ............................................................................... 89 FIGURE 4.24 – 10MIN INTERVAL PLOT SITE 1 2002 ............................................................................... 90 FIGURE 4.25 – 10MIN INTERVAL PLOT SITE 2 2002 ............................................................................... 91 FIGURE 4.26 – 10MIN INTERVAL PLOT SITE 3 2001 ............................................................................... 92 TABLE 4.5 – T TEST SITE 1 CONTROL V EXPERIMENTAL 2001 .............................................................. 93 TABLE 4.6 – T TEST SITE 2 CONTROL V EXPERIMENTAL 2001 .............................................................. 93 TABLE 4.7 – T TEST SITE 3 CONTROL V EXPERIMENTAL 2001 .............................................................. 94 TABLE 4.8 – T TEST SITE 1 CONTROL V EXPERIMENTAL 2002 .............................................................. 94 TABLE 4.9 – T TEST SITE 3 CONTROL V EXPERIMENTAL 2002 .............................................................. 94 TABLE 4.10 – ANOVA ACROSS ALL SITES CONTROL DATA 2001 ......................................................... 95 TABLE 4.11 – ANOVA ACROSS ALL SITES EXPERIMENTAL DATA 2001 ................................................ 96 TABLE 4.12 – T TEST SITE 1 V SITE 3 CONTROL DATA 2002.................................................................. 96 TABLE 4.13 – T TEST SITE 1 V SITE 3 EXPERIMENTAL DATA 2002 ........................................................ 97 TABLE 4.14 – T TEST YEAR COMPARISON AT SITE 1 ............................................................................ 98 TABLE 4.15 – T TEST YEAR COMPARISON AT SITE 3 ............................................................................ 98 FIGURE 4.27 – GENERAL LINEAR MODELLING...................................................................................... 99 TABLE 4.16 – PERCENTAGE OF TOTAL BEHAVIOUR OCCURING IN FIRST 15MINS.................................101 TABLE 4.17 – PROPORTION OF AGGRESSIVE BEHAVIOUR IN FIRST 15 MINS ..........................................101 FIGURE 5.1 – BREAKDOWN OF SOCIAL SYSTEMS DUE TO VARIABLE RESOURCES..................................116 FIGURE 5.2 – RABBIT CALCI VIRUS RELEASE, MITCHELL, 1996............................................................122
STATEMENT OF ORIGINAL AUTHORSHIP
The work contained in this thesis has not been previously submitted for a degree or diploma at any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made. Signature:_______________ Date:___________________
ACKNOWLEDGEMENTS This thesis would not have been possible without the support of Peter Mather and John Wilson. Thankyou for your advice, encouragement, and inspirational enthusiasm for ecology. I also wish to recognise the financial support from The School of Natural Resource Sciences QUT, and the Federal Government. Many thanks also to Dave Berman and his team of “Bulloo Warriors” from the Queensland Department of Natural Resources, particularly Michael Brennan, Craig Hunter, Peter Elsworth, and John Conroy – I would be buried in the desert if it weren’t for you blokes. Thankyou to Stanbroke Pastoral Company for access to Bulloo Downs, and thanks to Geoff and Wendy Murrell and all the staff of Bulloo Downs for your hospitality during my field trips. To the various landholders in the Mitchell Region, thankyou for access to your properties during my various pilot trips – I hope the rabbits stay away for many years to come. I owe a huge debt of thanks to those who volunteered their time to drive to the middle of Australia and chase rabbits: Ben de Zylva (who enjoyed the first trip so much he came back for more), Alison Crawford, and Alex Wilson. Our trips to Bulloo Downs would not have been possible without the logistical support team: Jo Chambers, Peter Prentis, and Stephen Craig-Smith – thanks for driving to Cunnamulla. This project would not have been possible without the cast of thousands from QUT, from the admin support to the radiation lab. Nat and Juanita (thanks for the help in the lab), Craig and Danny (thanks for sharing an office with me), and everyone at the Campus Club (thanks for the 12 hour lunches). Special thanks to Grant Hamilton and his efforts to “Show me the bunnies!” Let this thesis serve as an example of “what not to do” – the external factors such as drought and disease necessitated much variation to the original experimental design, to the point that resulted in a fairly limp dataset, containing far too many assumptions. If future students read this, please make sure you have the ability to collect enough data to rigorously test your theory. Check your field sites early, and if it looks like you can’t get the data – find another way to test your theory – or even change your topic altogether. Finally, thankyou to Rebecca, without your help I wouldn’t have made it this far.
Chapter 1 Introduction
Introduction
In many animal species, social behaviour can influence many aspects of life history
characteristics. The interaction, however, is bi-directional in the sense that social
behaviour can be influenced by a species’ characteristics in addition to external
environmental factors. The idea that behaviour patterns are inflexible within species
has been challenged by new research into social systems and genetics. This review
aims to explore the idea that a species' social structure can be influenced by different
environmental factors that it experiences.
Dispersal, Habitat Variability, And Gene Flow
Organisms that live in groups must ultimately decide whether to stay in the natal
territory or to disperse into new areas. Many factors influence the potential for
individuals to disperse successfully, not the least of which, is social organisation. If
an individual is of low rank in a social hierarchy, then dispersal to a new territory
may be a good option if the cost of dispersing is offset by the benefits gained by
reaching a new territory. Dispersal is only effective however, if an individual is able
to survive and reproduce in the new habitat. Three main types of dispersal have been
described (Krebs 1994). Diffusion is the gradual movement of a population across
hospitable terrain that occurs over several generations. Jump dispersal occurs when
individuals move large distances in a single event, usually across areas of
unfavourable habitat. Species introduced to non native areas through human
intervention can be viewed as an assisted form of jump dispersal. Secular dispersal is
a diffusion event that occurs over geological time and usually involves an
evolutionary change in the species across a specified time period; it can also be
associated with continental drift.
Habitat quality has the potential to affect both the social organisation of a species, its
dispersal dynamics, and the interrelationships between the two. A habitat that is
temporally and spatially stable is likely to be used in a different manner to one that is
dynamic. In a study of the red squirrel, Sciurus vulgaris, Lurzs et al. (1997)
examined the effect of habitat variability (temporal and spatial) on dispersal. They
studied squirrel dispersal patterns in a stable habitat with a reliable food supply, and
1
Chapter 1 Introduction
a variable habitat with temporal and spatial differences in food availability. In both
habitats, they observed male-biased dispersal in spring and female-biased dispersal in
autumn. More adults dispersed however, in the variable (66%) than in the stable
(31%) habitat. Large differences were also evident in the extent of site fidelity
between the two squirrel populations. Food availability was the main factor that
affected female dispersal. In contrast, male dispersal was influenced by the
distribution of females with male site fidelity high in the stable habitat, whereas
males tracked the movement of females in the variable habitat. This most likely
occurs because the stable habitat has sufficient resources to satisfy female needs.
Lurz's et al. (1997) data on squirrels suggest that female dispersal patterns are an
adaptive response to the spatial and temporal predictability of food resources.
Dispersal of individuals into potentially new habitat or territory that leads to effective
reproduction can result in gene flow. Different dispersal strategies can therefore lead
to different population genetic structures that are consequences of different
behaviour patterns. Dispersal or migration alone does not constitute gene flow - there
must also be an exchange or transfer of genetic material i.e. reproduction. Gene flow
(or a lack thereof) can lead to population structuring, which is defined as differences
in genetic variation among constituent parts of a species’ natural range, provided the
effect is not counteracted by other evolutionary processes such as a mutation, natural
selection or genetic drift. Gene flow is a major factor which influences population
structure because it determines the extent to which each local population of a species
acts as an independent evolutionary unit. If a large amount of gene flow occurs
among local populations, then the collection of populations evolve together; but if
there is little gene flow each population will tend to evolve independently (Slatkin
1994). A number of theoretical models have been developed which describe gene
flow and its potential effects on population genetic structure.
Modelling Gene Flow
The simplest models are based on the island model of migration which was first
proposed by Wright (1931). In this model, a species' distribution consists of discrete
populations that are geographically separated and are assumed to be large enough
such that genetic drift can be ignored as a population structuring process. Migration
2
Chapter 1 Introduction
is assumed to occur between population islands as a process in which the allele
frequency of the migrants is equivalent to that of the total population and therefore
the amount of migration is measured as the probability that a randomly chosen allele
in any sub-population comes from a migrant (Hartl and Clark 1997).
Two alternative models were developed that address population structure in
continuous rather than discrete population systems: 'Isolation by distance' models,
and 'Stepping stone' models. Sewall Wright was also responsible for the early work
on isolation by distance. The theory is based on the premise that if, in the continuous
distribution of a population, migration of individuals and subsequent interbreeding is
restricted to short distances due to short range dispersal; then remote populations
may be differentiated because of the distance among them (Wright 1943).
The concept of a species' range being large enough such that colonies develop and
exchange genetic information through migration was developed by Kimura and
Weiss (1964). They proposed three types of stepping stone model with increasing
degrees of complexity referred to as 1, 2, and 3 dimensional models.
A one dimensional model is where the colonies are located in a linear fashion.
Migration can only occur between adjacent colonies, that is, for each generation an
individual can migrate 'one step' in either direction. For the other two models, the
array of colonies will increase. The two dimensional model assumes a rectangular
arrangement of colonies, therefore an individual can migrate in four directions. The
third dimensional model introduces a cubic system in which migration can occur in
six directions. It is important to note that the 3rd dimension does not necessarily have
to be of a spatial or habitat capacity, it may simply refer to an attribute of the species
that enables greater variety in life style. Social rank is an example of one factor that
may provide a third dimension to population structure.
The development of methods for estimating gene flow occurred as a corollary to the
theoretical work that developed the models. Direct estimation methods are based on
experiments or field observation which gather measurements of dispersal distances.
The distance estimates can be converted into estimated gene flow based on the
assumption that migrant individuals have the same probability of reproductive
3
Chapter 1 Introduction
success as do residents. Indirect methods, however, are based on mathematical
models which explain interactions of gene flow and other forces to predict how much
gene flow must have been occurring to explain the observed patterns (Slatkin 1994).
Wrights FST statistic is the best known of these methods; it is a measure of the
correlation between genes in a sub-population relative to the entire population
(Wright 1951). The model states that in an island model at equilibrium,
FST = 1/1 + 4Nm where N = the effective population size, and m = immigration rate.
There is no way to gain a separate estimate for the terms N and m, however, by
solving for Nm the formula is transformed to Nm = 1/4 (1/ FST - 1). FST can be
calculated easily from allele frequency data. By solving the equation, one gains an
estimate of gene flow for the population under study.
Distinct advantages and disadvantages are associated with the direct and indirect
methods of gene flow estimation, which are discussed by Slatkin (1994). Direct
estimates can reveal certain aspects of the dispersal mode such as the life stage most
common for dispersal, and the environmental conditions most conducive to dispersal.
The disadvantage of the direct methods is that they are limited by the size of the
project, and it can be difficult to gather information regarding any long distance
dispersal or dispersal under abnormal environmental conditions. Indirect methods are
able to incorporate any effects of variation in dispersal and average out the
differences over time. The major disadvantage however, is that the methods rely on
assumptions regarding allele frequencies, and these assumptions cannot always be
tested independently.
The use of indirect methods to measure dispersal, and in particular, to estimate gene
flow using Nm has been accepted practise for many years. More recently however,
conjecture has grown regarding the validity of the formula. Whitlock and McCauley
(1999) argue that in many cases FST does not equate to the formula 1/(4Nm + 1),
because the formula is based on several assumptions that are violated in most natural
systems. The five critical assumptions are:
1) The alleles at the loci are selectively neutral and are not linked to selected loci.
2) The rate of mutation is not high relative to the rate of migration.
3) All populations are created equal, with a constant number of individuals and
equal contributions to the migrant pool.
4
Chapter 1 Introduction
4) Migration is random (no spatial structure).
5) The system is in equilibrium between migration and genetic drift.
Measurement of genetic variation from genetic data is a valid use of FST, however, it
is clear that estimates of dispersal and gene flow based on F statistics should be
viewed with care. Estimates may be correct within a few orders of magnitude, and
should be performed only in situations when the biological question depends on
estimating migration rates among populations where 'errors' associated with the
estimate can be relatively large (Whitlock and McCauley 1999).
Metapopulations
Following on from the ideas on dispersal developed with Island and Stepping-Stone
models, came the concept of metapopulations. The term itself is used to define a set
of local populations that interact via individuals moving among populations (i.e.
dispersing). The characteristic feature of a metapopulation is that local populations
are dynamic and will undergo phases of extinction, and subsequent recolonisation
from other populations within the system; this is also referred to as turnover. Several
kinds of metapopulation were characterised and summarised in Harrison (1991)
(figure 1.1).
5
Chapter 1 Introduction
Closed circles represent habitat patches; filled = occupied, unfilled = vacant. Dashed lines show the boundaries of populations. Arrows indicate migration (colonisation). A. Levins-type metapopulation. B. Mainland-island/source-sink metapopulation. C. ‘Patchy population’. D. Non-equilibrium metapopulation. E. An intermediate combination of B and C.
Figure 1.1 Types of Metapopulations (reproduced from Harrison, 1991)
The Levin's (1969) model of metapopulations (figure 1.1A) was the first step in
developing theories behind newer models. It is based on the scenario where a set of
conspecific populations exists in a balance, at the regional level, between extinction
and colonisation. This model most closely resembles the island and stepping stone
models of migration. Mainland-island and Source-sink metapopulations (figure 1.1B)
6
halla
This figure is not available online. Please consult the hardcopy thesis available from the QUT Library
Chapter 1 Introduction
occur when there is one large central patch that is resistant to extinction, with
peripheral patches that undergo periods of extinction and subsequent recolonisation
by migrants from the main patch. There is a distinct difference however, between
these types of metapopulations with respect to the outlying patches. Island habitats
are simply smaller versions of mainland habitats, whereas sinks are qualitatively
different from sources, being unsuitable in some way for survival and reproduction
(Harrison 1991). This type of metapopulation has also been referred to as a 'Core-
satellite system' (Hanski and Gilpin, 1991).
The patchy population (figure 1.1C) describes systems where habitat patches exhibit
spatial and temporal variation, however, there are also large amounts of dispersal
among patches, which effectively makes the group of populations a single interacting
unit. There is little opportunity for extinction to occur in a system like this because of
the high rates of dispersal. A non-equilibrium population (figure 1.1D) is
diagrammatically similar to the basic Levin's model (figure 1.1A) except that the
recolonisation process does not occur. If there is a lack of migration (recolonisation),
then when a patch becomes extinct, it will remain so. It represents a population
system of species in regional decline.
The main factors affecting localised extinction rates are usually stochastic in nature,
and include demographic, genetic, environmental, and catastrophic processes/events
(Shaffer 1981; Harrison 1991). Random changes in birth and death rates represent
demographic factors. These are most likely to have the greatest effect on small
populations or those in regional decline that are below a population size threshold
(Ebenhard 1991). Obviously, threshold levels will vary among species. Genetic
stochasticity concerns the loss of heterozygosity through drift effects and inbreeding
– the net result being a reduction in fitness, and increased probability of extinction. A
genetic effect, like a demographic effect, is more likely to occur in small populations,
however, it will definitely be more pronounced in a population that is newly small,
and is not conditioned to undergoing periods of population flux.
Environmental stochasticity and catastrophes are probably the most important causes
of local extinction because they can affect populations of varying sizes (Harrison
1991). Variations in environmental characters such as food availability and weather
7
Chapter 1 Introduction
conditions may affect the entire range of patches in a region, yet not all populations
are likely to go extinct. This observation led to the idea that certain patches are
effectively refuges that enable survival through adverse environmental conditions;
either by providing basal nutritional requirements or by providing better quality
shelter sites that, in some species, will facilitate a period of torpor until conditions
are more conducive to reproduction and dispersal (Harrison, 1991). In some
instances, larger patches will be better suited for use as refugia simply due to size
and ability to ‘absorb’ adverse conditions better than smaller patches – this would be
commonly observed in mainland-island metapopulations. Catastrophic events such as
flood, drought, and fire usually cause widespread extinctions in metapopulations.
While survival may be higher in larger patches, this will depend to a large extent on
the species in question and the nature of the catastrophe.
A metapopulation can persist only when colonisation follows extinction events.
Colonisation can be defined as starting with the arrival of a propagule (the migrants)
and ending when the extinction probability of the population no longer depends on
the initial state of the propagule (Ebenhard 1991). While the process could be viewed
simply as dispersal from an occupied patch, the migrant individuals must move
through inhospitable habitat in order to colonise the extinct patch. This process will
present its own set of problems. The success of the propagule will depend on the
probability of finding a suitable patch, and effectively reproducing once there.
Differences in dispersal rates among sex and age classes are most common in
polygamous species and in long-lived species with many litters per female (Hansson
1991). Other important observations on dispersal made by Hansson (1991) are that
dispersal distances appear to be longer in poor environments and habitat specialists
are more affected by boundaries than habitat generalists. Thus the ability of a species
to survive the dispersal phase through harsh environments will enhance its ability to
function as a metapopulation. Individuals in the colonised patch will have a higher
probability of extinction in the new habitat, than if they dispersed within the natal
patch. Ebenhard (1991) presents data which suggests the best colonisers will be large
propagules with potential for rapid increase in variable habitats or with a low
mortality in stable habitats. Dispersing propagules may also reach patches with
extant populations, and while this is not considered a colonisation event in the
strictest terms, it does have some important ramifications for metapopulation
8
Chapter 1 Introduction
dynamics. The migrant individuals may offer the opportunity for gene flow to occur,
possibly reducing the chance of inbreeding and any associated deleterious effects
(Gilpin, 1987). Migrants arriving successfully in an occupied patch may also be of
benefit to the local population if it is in decline, for whatever reason, by boosting the
species abundance in the patch – an occurrence referred to as the 'rescue effect'
(Hanski, 1991). An alternative scenario, however is that the migrant individuals may
not integrate with the local population at all, and instead develop their own
independent breeding group which would be detrimental to the original declining
population.
Severe fluctuations in population size (where periods of small population sizes
occur) can reduce allelic diversity and heterozygosity levels in a population. It is an
effect commonly referred to as a genetic bottleneck (Hedrick 1999). Elephant seals
and African cheetahs are two examples of species which have low levels of genetic
diversity that can be explained by historical bottlenecks (Bonnell and Selander 1974;
O'Brien et al. 1987). Random genetic drift caused by migration of a few individuals
to a new patch from an established subpopulation can create a bottleneck known as a
founder effect (Hartl 1997). The classic examples of founder effects occur in
instances where species have been introduced (translocated) through human activities
into completely new habitats (eg. Bufo marinus, the cane toad, and Oryctolagus
cuniculus, the European rabbit in Australia).
9
Chapter 1 Introduction
Behavioural Diversity And Genetic Determination
Individual differences in behaviour can influence differences in dispersal strategy.
Dispersal, however, represents only one type of behaviour and there are a great
diversity of potential behaviours, many of which have a genetic component.
Evidence for the genetic determination of dispersal/movement behaviour is
widespread, one example occurs in fruit fly larvae. The larvae occur either as
‘rovers’, which move a long way to find food, or ‘sitters’, which forage in a more
restricted area; the polymorphism is determined by alleles at a cyclic GMP-
dependent protein kinase gene (Partridge and Sgro 1998). It is clear that selection is
able to act upon genes controlling behaviour; in the case of the fruit fly larvae, where
‘rovers’ may have an advantage in crowded populations, ‘sitters’ may have an
advantage in low density populations. Studies on the genetics of behaviour led
Alcock (1984) to the following generalisations:
1. Single allelic differences can influence behavioural differences among
individuals.
2. Artificial selection for certain behavioural traits can be highly effective in
altering the behaviour of a population over time.
3. Physiological effects that are determined by genetic differences among
individuals are responsible for their distinctive behavioural characteristics.
4. Differences in the genetic and physiological characteristics of populations of the
same species may be related to variation in ecological pressures operating in
different areas.
The fact that selection can act on genes which determine behaviour, therefore means
that behaviour can evolve through this process, like any other trait which is under
natural selection.
Group Living, Cooperation, And Sociality
While evolution of social behaviour has been studied extensively, much of the
earliest work focused on how evolution of behavioural strategies were of benefit to
the group. Tinbergen (1964) suggested that groups of 'capable' individuals survive,
while those containing inferior individuals do not, and therefore cannot reproduce
effectively. He was essentially arguing for group selection influencing the fitness of
10
Chapter 1 Introduction
individuals. This theory was opposed by Williams (1966) who suggested that clutch
size and many social interactions enhance individual fitness. Williams argued social
behaviours evolved for the benefit of the individual, not the group. Altruistic
behaviours however, which involve the act of sacrificing ones personal fitness for
the benefit of others, does not appear to fit his argument easily. Hamilton (1964a,b),
in discussing the evolution of altruism, raised the issue of inclusive fitness when he
suggested that individuals can pass copies of their genes to future generations by
assisting the reproduction of close relatives (indirect fitness) as well as via their own
reproductive efforts (direct fitness). Hamilton described a model that allowed for
interactions between relatives which affect fitness. Species which act 'altruistically'
may evolve behaviours so that individuals maximise their inclusive fitness and this
implies a limited restraint on selfish competitive and self-sacrificing behaviours
(Hamilton 1964a,b). Hamilton’s theory can more easily explain the evolution of
altruistic behavioural patterns such as cooperative breeding and coloniality.
The bell minor, Manorina melanophrys, is an example of a species that breeds
cooperatively. Individuals have never been observed breeding unassisted and
individual helpers, even breeders, often give aid to a number of breeding pairs within
a breeding season (Painter et al. 2000). The species has a multi-tiered behavioural
and social organisation, which was observed in studies by Clarke (1984, 1989) and
Clarke & Fitzgerald (1994). There are three levels of social organisation: colony,
coterie, and the nest contingent. The colony is a geographically discrete collection of
up to 200 individuals that communally defend an area against both inter- and
intraspecific avian competitors. The coterie is a group within the colony that contains
one or more breeding pairs. While helpers may aid more than one pair within a
coterie, they do not interact with members from other coteries except in territorial
defense. The third level of social organisation is the nest contingent, which consists
of individuals that assist at the nest as well as the breeding pair(s). Painter et al.
(2000) found (using microsatellite analysis) significant differences between coteries
in a high density colony, which resulted from related individuals associating
preferentially with each other. They also showed that individuals helping at the nest
were close relatives of the breeders, thus supporting models of kin selection for the
evolution of altruism in this bird. The classic examples of kin selection, however,
occur in eusocial insects including bees, wasps, and ants. In eusocial societies, a
11
Chapter 1 Introduction
queen produces all the offspring, and an army of sterile workers that share most of
their genes in common with their siblings. The evolution of eusocial systems is
complicated however, by ploidy differences between the two sexes. Females are
diploid and males haploid, a situation which changes the argument about altruism
when genetic relationships between offspring and parents are considered.
The main reason for the evolution of social behaviour is that natural selection has
influenced the frequencies of genes that give rise to such displays. Some species will
have certain evolutionary adaptations that favour the adoption of sociality while
others will not. Ultimately, it is natural selection or genetic drift acting on random
mutations that cause social behaviour to evolve in species, and consequently there
are several advantages and disadvantages. Through the selective process each
condition will affect individual fitness in a different manner for each species.
Costs and benefits of social behaviour (from Alcock, 1984)
Benefits:
Reduction in predator pressure by better detection and/or repulsion.
Improved foraging efficiency for large game or spatially and temporally clumped
resources.
Better defence of limited resources (space and food) against other groups of
conspecifics.
Enhanced care of offspring through communal feeding and protection.
Costs:
• Competition within the group for food, mates, nest sites and materials.
• Risk of infection by contagious diseases and parasites.
• Exploitation of parental care by conspecifics.
• Increased risk that conspecifics will kill progeny.
For species that have evolved solitary lifestyles, the costs may be greater than any
benefits gained from social living. Conversely, for species that live in social
communities the costs may be equalised or bettered by the benefits of the behaviour.
A good example occurs in two closely related species of freshwater fish, Lepomis
12
Chapter 1 Introduction
macrochris and Lepomis gibbosus (bluegill and pumpkinseed sunfish), studied by
(Gross and MacMillan 1981). The bluegill sunfish exhibits social behaviour during
the breeding season when males construct nests close together to form a colony.
Formation of the colony results in reduced pressure from the primary predators of
their eggs, which are catfish and snails, because males cooperate in colony defense.
The advantage of sociality to the bluegill is reduced by factors such as conspecific
interference and predation of eggs, as well as disease (fungi) that can spread through
dense colonies. The pumpkinseed sunfish however, lives a solitary life due in part to
the evolution of large, strong mouthparts designed for crushing snails and deterring
other potential predators. Colonial nesting is not advantageous to the pumpkinseed
sunfish because predation is not as great a problem as it is for the bluegill sunfish.
Resource Defence
The use of caches (food storing) usually occurs in species that exhibit territorial
behaviour and the act of creating the cache is a potentially costly exercise. Roberts
(1979) argued that adaptations are likely to arise that will reduce costs and/or
increase benefits - and that territoriality is one such adaptation in this sense because
it reduces the amount of competitors that are able to gain access to stored resources.
When food is clumped spatially, aggressiveness can be expected to increase because
the cost of defending an area is small compared to the benefit of access to a large
share of the resource (Grant 1993). If food is clumped temporally, aggression levels
may be expected to fall because any time spent defending is simultaneously time
away from resource utilisation (Trivers 1972; Wells 1977; Robb and Grant 1998).
The mountain lion (Puma concolor), is one organism in which intraspecific
aggression is known to occur (Pierce et al. 1998). In this instance the food resource is
not clumped temporally, but instead the social class of females with kittens utilise the
resource at an earlier time than other social classes. Adult females usually have
overlapping home ranges that also overlap within the range of one or more males.
Females tend to reduce confrontations through a system of mutual avoidance,
however, it is not uncommon for males to kill other males, females, juveniles and
kittens (Seindensticker et al. 1973). As mountain lions are known to cache food and
have overlapping home ranges, Pierce et. al. (1998) suggest that females with kittens
that visit the cached prey at different times to the other social classes, could obtain
13
Chapter 1 Introduction
fitness benefits by further minimising the probability of contact with other mountain
lions. In this example, one could argue that the resource is clumped spatially, in the
cache location, however, aggressiveness does not increase (with respect to the
suckling females) due to the differential timing of feeding events.
In some cases, resource clumping (spatially) is the result of an organism actively
caching food. Smith and Reichman (1984) in their review of caching by birds and
mammals limited their discussion to the movement of potential food from one
location to another for consumption at a later date. Not all species cache food, those
that do are found predominantly in temperate rather than tropical areas. This is most
likely because food resources are more predictable temporally in tropical areas which
probably negates the need to cache food. The high temperatures and humidity of a
tropical environment may also promote spoiling of cached food, which reduces the
efficiency of the method as a means of survival through unfavourable conditions
(Smith and Reichman 1984). Species that are known to cache food, will do so in one
of two ways. They can create a horde cache, or many scatter caches. The evolution of
caching behaviour is a method of resource defence. Horde caches can be effective
methods of food storage when the individual is able to defend the cache from
competitors. If an active defence is not feasible however, then scattering several
caches across a home range can be a viable alternative provided the organism is
capable of remembering all cache locations. Many species have been shown to
possess the ability to remember the location of caches and distinguish which are used
and unused (Wrazen and Wrazen 1982; Sherry 1984; Smith and Reichman 1984).
As mentioned above, resource defence through territoriality is one of the factors that
is considered to have led to the evolution of group living in many species. The cost
of sharing the resource within a territory with conspecifics is balanced by the benefits
gained from exclusive access to the resource, whether it is cached or distributed
naturally. Furthermore, the benefits in terms of fitness and selection are increased if
the members of the group belong to the same deme. If an individual is not dominant
or not producing offspring, then by participating in group behaviour, it may
contribute to the survival and breeding success of closely related individuals and thus
increase the likelihood of a small percentage of its own genes being passed to the
next generation through relatives (inclusive fitness).
14
Chapter 1 Introduction
Territory defence, however, is not the only contribution to group living that an
individual can make. Other activities that benefit the collective include predator
avoidance/warnings, collection of food, and rearing of young. If the group consists of
individuals that are not related, there may still be benefits associated with
participating in predator warnings and avoidance, as well as access to communal
food resources. Social hierarchies develop as a consequence of group living, and
therefore many species that live in groups (though not all) exhibit this structure to
varying degrees. In a study of the crane, Grus grus, foraging in cereal farmland
Alonso et al. (1995) found that birds left more resource-rich patches earlier than
expected and at higher intake rates than in poor patches, although they stayed longer
when in larger flocks. The results suggest that cranes may change their foraging
behaviour according to their expected energy balance. In this instance, cranes benefit
from group association by gaining greater food intake, and better avoidance of
predators.
Behavioural Flexibility
Behaviour can be modified by the environment, and clear-cut relationships between
energy requirements, resource distribution, and social systems can often be
demonstrated (Pough et al. 1989). An animal with a large mass will have much
greater energy requirements relative to a small animal. To obtain the necessary
resources, the animal may have to search widely across their home range - the area
in which they live to find their food and shelter. One might expect to find the size of
home range correlated with size/mass of an animal, but this does not take into
account the possibility of habitat patchiness. An animal may utilise a resource that
occurs in an uneven distribution, if so, then the size of the home range will be a
reflection of the quality of the habitat (in terms of the resource in question). Forest
duikers (small African antelope) have been shown to be more active in habitats of
high quality, although differences in home range have been observed between the
blue and red species (Bowland and Perrin 1995). Bowland (1995) found that core
areas in the home range of both duiker species were usually associated with bed sites.
Blue duiker home ranges and core areas however, were fixed year round with no
overlap between neighbours, while home ranges and core areas of red duikers
15
Chapter 1 Introduction
overlapped extensively. Temporal separation in red duikers is suggested between
some individuals and not others, which means there may be occasions where red
duiker individuals come in contact whilst using the overlapping home range. If
contact is occurring between red duikers, then passive or tolerance behaviour may
occur - which may manifest itself simply as non-recognition (ignoring). The fact that
red duiker home ranges overlap, suggests an absence of territoriality, however the
blue duiker appears to behave conversely with strictly defined home ranges.
Therefore, one might expect to observe aggressive, territory defence behaviour in
blue duikers.
The European rabbit (Oryctolagus cuniculus), is another species that exhibits
territorial behaviour patterns and group living attributes. The population
demographics and abundance of the rabbit make it a useful study species to further
examine theories of behavioural ecology and population genetics.
The European Rabbit
The European Rabbit (Oryctolagus cuniculus) is believed to have evolved in
southern France and Spain. While the species may have been widely distributed
throughout western Europe during pre-historic times including the Pliocene and
Pliestocene; glacial activity 3000 years ago confined rabbit populations to only
warmer southern refuge areas in Europe (Corbet 1986; Flux 1994). Thus populations
historically must have been exposed to large fluctuations in size and demography.
The European rabbit, however, is also a pest and game species, and natural
distributions are often in close association with humans (Flux 1994). Consequently,
rabbit populations were established by humans across much of the European
continent, South America, New Zealand, and Australia.
While domestic rabbits were present on the first fleet which arrived in Australia in
1788, the wild European rabbit was first introduced to the Australian mainland by
Thomas Austin, a keen sportsman and member of an acclimatisation society. The
role of the societies was to facilitate the emigration of settlers from the United
Kingdom to Australia; one method employed was to introduce game species. The
first wild rabbits were introduced at Geelong in 1859 and were maintained in
16
Chapter 1 Introduction
enclosures, but later some were deliberately released into the wild or escaped
(Williams et al. 1995).
Further deliberate releases were made in South Australia and New South Wales; and
by 1900, the rabbit “front” had entered parts of Western Australia, Queensland and
the Northern Territory. The spread of rabbit populations continued at various rates,
the result being the current distribution in which most areas south of the Tropic of
Capricorn are populated, and rabbit populations north of this line generally consist of
small, scattered populations in suitable habitats (Rolls 1984; Stodart and Parer 1988;
Myers et al. 1994).
The great success of rabbit colonisation in Australia can be attributed to a number of
factors:
• Lack of predators and parasites,
• favourable climate and soils,
• human activity, and
• efficient reproductive biology.
When the European rabbit was first released into Australia, few natural predators
were present in sufficient numbers or possessed the ability to significantly reduce
population growth, thus relaxing one of the ecological constraints present on rabbit
populations in their natural habitats in Europe. The Australian climate and landscape
also facilitated rabbit colonisation because the winter season is not as harsh as that of
Europe, indeed many areas of the Australian continent experience a Mediterranean
type climate all year round. In many parts of Australia, soils are composed of sandy
loams ideal for burrowing which also sustain the growth of suitable feed. The rabbit
also proved to be a better competitor in Australia than many native burrowing
herbivorous species such as the bilby (Macrotis lagotis), and thus found ready-made
burrows in many instances.
The single most important factor which led to the successful colonisation of
Australia by rabbits, however, was the actions of humans. Initially the rabbit spread
along riparian systems, following watercourses, but its spread was greatly aided by
17
Chapter 1 Introduction
the pastoral activities of the early European settlers. The clearing of forest for the
growth of grain crops and raising of cattle made vast areas of land available to rabbit
populations that were previously inaccessible and/or unsuitable. Thus, the rabbit
spread to a variety of different habitats, although the degree to which rabbit
populations utilise specific habitat types depends largely on the type of vegetation
present.
The vegetation suitable for rabbits can be classified into five categories (Williams et
al, 1995).
1. Shrub (scrub and bracken thickets) either with or without an overstorey of trees.
2. Patches of dense scrub interspersed with patches of grassland in various
proportions.
3. Savanna woodland with extensive grassland.
4. Grasslands of varying vegetation density
5. Short or sparse grass with varying extents of bare ground.
As ground cover levels decrease, accordingly there is an increase in the size and
structure of warren systems; in the most open of environments, the rabbit will rely
heavily on underground shelter. A rabbit that has colonised a new area, however, will
not dig a new warren, unless the area consists of sandy soils (Cowan 1987a).
Usually, the colonisers live in depressions under logs or rocks (termed a squat).
Females dig shallow burrows in the squat in order to raise a litter – called a stop –
they are usually well concealed to avoid detection from predators. If further tunnels
are excavated within the stop for successive generations of litters, the stop can be
referred to as a warren (Mykytowycz et al. 1960).
Generally, rabbits are largely nocturnal animals, and only emerge from warrens
between one to three hours before dusk, returning just before dawn. Typically, they
will engage in a period of grazing, followed by socialising, on or around the warren
until dark, at which point they will venture further a field (Williams et al. 1995).
Rabbits remain above ground for the duration of the night, although Fullager (1981)
reported that presence of predators will cause them to retreat to their warrens.
18
Chapter 1 Introduction
Group size varies from between two to ten individuals, and within the groups,
typically, independent dominance hierarchies exist for males and females (Williams
et al. 1995). Males compete to gain access to females, and females compete to gain
access to nesting sites. Consequently, male aggression occurs near females, and
female aggression occurs near nesting sites (Cowan, 1987a). A female living as the
sole female in a social group will have greater longevity and greater reproductive
success than will females competing in the same group (Cowan 1987b), which may
account for the evolution of female dominance hierarchies – and the fact that they
will attack individuals attempting to establish in their territory (Parer 1982).
Mykytowycz’s (1958, 1959, 1960) studies of an experimental rabbit population in
Canberra, Australia, provided extremely useful data on social behaviour and
dominance hierarchies, that expanded the work of Southern (1948) studying a
population in the United Kingdom; and provided the baseline of social behaviour that
many researchers have used in subsequent studies. The population was enclosed, but
all individuals were identified and marked prior to introduction to the study area,
therefore social interactions were able to be recorded at the individual level. When
the top ranked male was experimentally removed from the population, all remaining
males attempted to improve their position, however, the second ranked male always
succeeded in this contest. When the original top ranked male was returned to the
population, there was prolonged and severe fighting, with the loser downgraded to
the lowest rank in the group. Similar experiments with the female hierarchy did not
produce the same aggressive results (Mykytowycz, 1958).
In the first year, the study obtained evidence that dominance hierarchies, and
therefore social behaviour patterns, had a clear link to survival. The offspring of the
dominant pair had greater survival than those of subordinates; and the dominant pair
was also able to breed more frequently (Mykytowycz, 1959). During the second year,
the survivors of the first breeding season formed several groups each with a distinct
dominance hierarchies. Although the groups were of mixed parentage, the offspring
of the original dominant female were always dominant, and those of subordinate
females were generally also subordinate. Again, the offspring of dominant pairs had
greater survival rates because of breeding earlier in the season under better resource
(food and nests) conditions (Mykytowycz, 1960; Henderson, 1979).
19
Chapter 1 Introduction
While the dominant male in a group will have first choice and access to females, it is
not always possible for him to guard two females at the same time. Therefore, it is
not uncommon for the dominant male to sire only about 60% of the litters in a group
(Daly 1981) – the remainder of the litters being sired by subordinate males through
promiscuous matings. This occurs, in part, due to female synchrony of the oestrus
period (Parer and Fullagar 1986). The fact that populations generally live in groups
creates situations in which the dominance hierarchies (combined with environmental
factors) inhibit reproduction below the highest physiological capacity (Mykytowycz
and Fullagar, 1973).
Territory defence of the group is usually conducted by males, and the territorial
boundaries are a reflection of the size of the home range of the dominant male
(Williams et al. 1995). Mykytowycz and Gambale (1965) studied a 45 acre area
containing three populations, and found that dispersal between warrens only occurred
during non-breeding seasons; the study also reinforced the importance of warrens for
survival and the tendency for group living. Food resources (eg. grazing patches),
however, are typically spread over a large area, and therefore often cannot be
defended adequately. If population density is high, several social groups may occur
in large warrens, while at the other end of the scale, a single group may utilise
several warrens provided population density is low (Wood 1980; Fullagar 1981;
Williams et al. 1995).
Australian populations of the European rabbit, particularly those in arid
environments, have existed as metapopulations – frequently undergoing periods of
extinction and recolonisation (Parer and Fullager, 1986). Like many
metapopulations, the regional persistence of the rabbit in Australia has often relied
on certain patches acting as refuges during times of unfavourable conditions (eg.
drought). As a result of this pattern, rabbit population numbers have fluctuated
accordingly. Such a population dynamic effectively pushes the population through a
genetic bottleneck whenever a large population size fluctuation occurs. Similarly,
introductions of diseases such as myxomatosis and rabbit calici virus, whilst not
eradicating the species, have caused great perturbations to population size and hence
have probably resulted in significant genetic bottleneck effects.
20
Chapter 1 Introduction
In a study of rabbit populations in the East Anglia region of Britain, Surridge et al.
(1999) found that local populations were genetically distinct from one another and
had small effective population sizes. It is thought the distinction occurs due to the
combined effects of their natural social structure and random genetic drift acting on
bottlenecked populations after exposure to myxomatosis. She argued that the genetic
structure observed in East Anglia represented recent events rather than historical
influences (Surridge et al. 1999). On the other hand, Queney et al. (2000) and Zenger
(2003) found no evidence of genetic bottlenecks in rabbit populations in northern
France and Australia respectively. They argued that levels of genetic diversity in
rabbit populations in Europe may not have been affected by disease outbreaks
causing high mortality, and rapid population expansion following a population crash
can limit the effect of the crash on the population genetic structure. In another study
of rabbit populations in East Anglia, Webb et al. (1995) showed that population
genetic structure was influenced by social organisation. In particular, the natal
dispersal pattern where females exhibit philopatry, and males disperse to new social
groups before the start of the new breeding system results in detectable differences to
population genetic structures.
Small effective population sizes have also been observed in some wild rabbit
populations in Australia. Daly (1979) suggested this was influenced by a
combination of social structuring (i.e. dominant individuals providing the majority of
genetic information to subsequent generations) and habitat heterogeneity. Studies
conducted in Britain, focused on populations that exist in largely stable
environmental conditions, which facilitate the development of stable social groups.
In Australia, however, habitats where rabbits are found are not always of the best
quality in terms of resource availability, and therefore a significant amount of habitat
heterogeneity may occur. Rabbit population genetic structure in arid Australia differs
from that in semi-arid and mediterranean systems. Fuller et al. (1996) examined
rabbit populations in an arid region of south western Queensland (1600km2) and
reported that significant gene flow occurred across large geographic areas because no
significant genetic differences were observed among populations (panmixia). It was
suggested that environmental fluctuations had caused frequent localised
extinction/recolonisation events leading to homogenising gene flow. The study was
21
Chapter 1 Introduction
extended to an examination of a semi arid region 500km east of the arid region,
where a significant difference in population structuring was observed (Fuller et al.
1997). While populations in the western system (arid) essentially function as a
panmictic unit, the eastern system (semi-arid) exhibited distinct population
structuring. The structure was hypothesised to be related to the pattern of distribution
of good quality habitats, which can be described as more 'patchy' in the semi arid
compared with the arid regions. The fact that one system was essentially panmictic
while the other was genetically structured over small geographic distances, suggested
that rabbits may also be influenced by other factors that result in variations in
population gene flow. The cause of this dichotomy was hypothesised to be a
combination of spatial and temporal variation in three primary resources – food,
nests, and mates.
Hamilton (2003) examined long term connectivity levels among local rabbit
populations and found they are influenced by the spatial distribution of resources and
other habitat factors. Hamilton developed a habitat heterogeneity model using
specific population parameters representative of the eastern semi-arid region. The
validity of model assumptions was assessed by regressing model output against
independent population genetic data, which could explain over 80% of the variation
in the structured genetic data set (Hamilton, 2003).
Cowan and Garson (1985) studied the social structure of two wild rabbit populations
in England (Oxfordshire and Northumberland) that were exposed to different
environmental conditions. One population was located on a chalk hill and the other
was located on a sand hill. Discrete social groups were only evident at the chalk hill
site, where females competed for burrows, and male territorial behaviour was
observed due to clumped female dispersal (Cowan and Garson 1985). The sand hill
population had more rabbits than the chalk hill site and growth rates were negatively
correlated with density. The authors concluded that scramble competition for food
occurred at the sand hill site while contests occurred for nests and mates in the chalk
hills. In this instance it is clear that the sand hill habitat has more abundant resources
than the chalk hills, and thus was able to support larger population sizes; although
due to the large numbers and lack of resource clumping, there was no fitness
advantage to being 'social'. Variation in social structure due to habitat parameters has
22
Chapter 1 Introduction
also been observed in other species, such as the brushtail possum in New Zealand
(Jolly et al. 1999; Taylor et al. 2000), which, like the rabbit in Australia, is a
significant introduced pest.
While it is widely accepted that rabbit social organisation and dispersal potential can
influence the genetic structure and patterns observed, it is unclear whether the degree
of organisation in rabbit social systems in arid and semi-arid Australia is a response
to differences in the extent of habitat heterogeneity in the region. Models of dispersal
and gene flow in Australian populations of the European rabbit, based on habitat
characteristics, have been developed and can account for over 80% of genetic
variation (Hamilton, 2003). It is not known however, if rabbit behavioural flexibility
contributes to the remaining 20% of genetic variation, or what (if any) are the
potential social structure consequences of density effects in arid and semi-arid
environments. Fuller et al. (1996, 1997) completed the initial research of rabbit
population genetics in arid and semi-arid Australia; which was followed by
Hamilton’s (2003) research on connectivity. Therefore, the next step in an holistic
approach to understanding, and ultimately managing rabbit populations in Australia
is to study the relationship between what type of social systems are present and
variation in habitat characteristics. The specific questions the current project will
focus on in order to address the main objectives are listed below:
• Do patterns of genetic structuring vary with differences in major habitat
attributes?
• Do aggressive/territory defence behaviour patterns vary with differences in
availability of key resources?
The questions will be addressed using genetic marker studies in areas with different
habitat attributes and behavioural experiments under varied environmental conditions
to quantify difference in aggression patterns. The observation experiments will aim
to determine if rabbits behave differently when exposed to different habitat
conditions in Australia; while the genetic assessments will aim to present evidence
that variable social systems have distinct effects in terms of population gene flow.
23
Chapter 2 Experimental Design and Methodology
Experimental Design And Methodology
Recent population genetic studies in arid and semi-arid Queensland identified
regions where different genetic structures were present based on variation in the
mitochondrial DNA genome. Fuller et al. (1996, 1997) identified a region in far
western Queensland that showed high levels of gene flow among rabbit populations
resulting in effective panmixus over large areas (>1000 km2). The same study also
identified a region 600km east of the western panmictic zone that exhibited much
lower levels of gene flow, resulting in population genetic structuring at much smaller
spatial scales. The basis of the observed differences in genetic structure was
suggested to be variation in the levels of habitat heterogeneity in terms of vital
resources – food and nesting sites.
For the most part, the behaviour of rabbits has been overlooked as a component
which could contribute to an explanation for patterns of population genetic structure
observed in arid and semi-arid Australia. The purpose of the present study is to
investigate the potential that rabbits may be capable of flexible territorial behaviour
patterns depending on the amount and distribution of favourable habitat. If rabbits
adjust their aggressive behaviour in response to differences in habitat conditions in
arid Australia, then their well documented social/territorial defence may be relaxed
in times of abundant resources and result in subsequent population explosions and
consequent dispersal events.
The first part of this study will assess the genetic structure of O.cuniculus in arid and
semi-arid Queensland using highly variable nuclear DNA markers (microsatellites);
and compare the results with those from previous studies (Fuller et al. 1996, 1997)
that examined patterns in mitochondrial DNA and more conservative nuclear
allozyme markers (Fuller 1995). This comparison is necessary because the
mitochondrial genome is maternally inherited and the predominant dispersing sex in
rabbits is known to be the male, which could result in female only genetic
structuring. A solution to this problem is to examine variation in nuclear DNA in
combination with mitochondrial DNA, however, at present, the only use of nuclear
DNA in studies of rabbit populations in arid and semi-arid south west Queensland
has been via allozyme electrophoresis, where variation was limited as a result of
24
Chapter 2 Experimental Design and Methodology
functional constraints on coding sequences and potential loss of genetic diversity
levels due to past bottlenecks.
The second part of the study will assess rabbit behavioural flexibility. The initial
design was to conduct field experiments at sites in the arid and semi-arid regions
identified by Fuller et al. (1996, 1997) that were used for assessing population
genetic structure to test the potential for differences in aggressive behaviour
associated with habitat difference. Due to the effect of recent releases of rabbit calici
virus (1996), however, the study sites (especially those in the eastern semi-arid zone)
no longer have rabbit populations large enough to study. Even in the western arid
zone, sites from which genetic material was sampled in the past now also have very
low rabbit numbers due to calici virus, extreme drought, and control efforts of
property owners. New study sites were located within the panmictic arid zone
identified by Fuller (1995) – and were used subsequently for the behaviour
component of this project.
In order to determine if differences exist in the relative levels of aggression and
territorial response depending on resource availability, it was originally planned to
conduct field experiments in both habitat types (arid and semi-arid); this was not
possible for the previously listed reasons. Therefore, the only remaining option for
the behavioural component of the project was to examine levels of aggression and
territoriality in the same arid sites under high and low resource availability
conditions.
Description of Study Sites
The study sites were located within two major regions of south western Queensland.
The regions were identified in the studies by Fuller (1995) and Fuller et al. (1996,
1997) and are broadly identified in Figure 2.1.
The arid zone (referred to as ‘western’) is located in far south western Queensland,
and is centered on a large cattle property called Bulloo Downs, 28° 31.62’ S 142°
57.63’ E (owned by Stanbroke Pastoral Co.). The property is 1,093,500 hectares in
size, located 120km south west of Thargomindah, and on the edge of the “Channel
25
Chapter 2 Experimental Design and Methodology
Country”. Land in this area is susceptible to floods which result from rains either in
situ (average rainfall is about 200mm/year) or further upstream in the catchment
areas for the Bulloo River that runs through the middle of the property and drains
into the Bulloo Lakes in the south west corner. The property is large enough that
several landforms exist, however, the systems used in this study were confined to
sandy hills separated by claypans. Few trees occur on the property except in areas
adjacent to channels and waterholes and pasture growth is dependant on rainfall or
floodwaters especially in the winter months, therefore, rabbit numbers fluctuate
extensively depending on the frequency of rainfall.
The semi-arid zone (referred to as ‘eastern’) is located in a 200km radius around the
township of Mitchell in the Maranoa district (approximately 500km west of Brisbane
and 600km east of Bulloo Downs). Properties around Mitchell are much smaller than
in western Queensland therefore more were included for sampling in order to cover
the same geographic area. The region consists of cleared pastoral areas interspersed
with remnant dry sclerophyll forest which is known to be unfavourable for rabbits.
26
Chapter 2 Experimental Design and Methodology
East
West
Figure 2.1 – Areas of study, West and East.
Population Sampling
An assessment of patterns of genetic variation between the regions required sampling
of DNA from target populations in both regions. During pilot trips in 1999 and 2000,
however, it became apparent that due to situations beyond the control of the project
(ie drought and disease), there were insufficient populations to allow appropriate
genetic sampling in both regions. While the best design for the project consisted of
collecting new samples under present day conditions, the next best option, was to use
historical tissue samples that were collected between 1993 and 1995 by Fuller (1995)
and were stored at -80°C. It is acknowledged that these samples were collected by
Fuller, under different environmental conditions, however, using the historical
samples was the only way to achieve any assessment of microsatellite genetic
variation in the regions of interest during the timeframe of this project.
27
Chapter 2 Experimental Design and Methodology
The following describes the collection of tissue samples from the arid and semi-arid
zones (Fuller, 1995):
In the arid zone, approximately one hundred adult animals were sampled humanely
from 3 sites on Bulloo Downs; in the semi-arid zone, a minimum of thirty adult
rabbits were taken from eight sites centered around the township of Mitchell.
Animals were dissected within 30 minutes from time of death and a small piece of
liver tissue was collected from each animal and samples stored in cryoware vials
(Nalgene Co.) on liquid nitrogen. On return to the laboratory all samples were stored
at –80°C until used for genetic analysis.
In the arid zone, there are many water bores on Bulloo Downs and populations were
sampled at three of these sites named: Ponto, Thurloo, and Willala.
The semi arid region consists of smaller properties, and eight sample sites were
named after the properties on which they were found: Alice Downs, Bowann,
Claravale, Currawong, Glenalba, Glenlea, Polworth, and Verniew. A minimum
sample size of 30 was set as the target, though this was not always attained.
Consequently, sample sizes were uneven across the data set.
Genetic Methods
Genetic variation, in the form of multiple alleles at genetic loci exists in most natural
populations (Hartl 1988), but the methods used to sample variation will depend on
the species and the specific question to be answered. Allozyme electrophoresis was
used as the genetic analysis standard for many years due to its ease of use, speed,
cost, and results (distinguishable loci with codominant alleles that can be scored
unambiguously). The disadvantage of allozymes is that they are functional gene
products, and so are limited by past and present functional constraints on the gene.
Microsatellites consist of tandem repeats of very short nucleotide motifs; the repeat
array is usually 10-50 copies of a sequence that is 1-10bp long (more commonly 2-
5bp long). Unlike allozymes which are the product of coding DNA, microsatellite
loci are randomly distributed and present at high frequency in eukaryotic non-coding
28
Chapter 2 Experimental Design and Methodology
regions and because most are not constrained, they usually show high levels of
genetic variation. The function (if any) and evolutionary significance of
microsatellites is unknown.
Variability at microsatellite loci is derived from the number of repeats of the motif
sequence; each variation is considered a different allele and the alleles can be
discriminated on the basis of size using electrophoretic methods. New alleles are
formed through any genetic mutation which results in a net loss or gain of repeats.
As microsatellite analysis is a method which relies on size discrimination, any
mutation that does not lead to a difference in the number of repeats will not be
identified. Microsatellites can be classified depending on the repeat motif length (di,
tri, tetra nucleotide) and continuity (perfect or interrupted) eg. (GT)18 is a
dinucleotide continuous repeat and (TG)3CG(TG)17 is interrupted. The type, length
and continuity appear to affect the rate of mutation and levels of allelic variation.
Interruption in the core sequence seems to stabilise the array, such that loci with pure
repeat sequences are more polymorphic than those with interruptions. Levels of
allelic diversity are correlated with repeat length - loci with longer repeats are more
polymorphic than loci with short repeats.
Ten different microsatellite loci were trialed throughout the course of this project,
however, only five were able to optimised for local laboratory conditions. One of the
five was found to be a nested repeat of another loci, therefore, microsatellite
variation was assayed using four polymorphic microsatellite loci developed
specifically for use on O.cuniculus in European populations; Sat3 and Sat5
developed by Mougel et al. (1997) and SOL28 and SOL30 developed by Rico et al.
(1994).
Behavioural Methods
Studies of behaviour patterns under laboratory conditions are abundant for insects
and many invertebrate species; a considerable amount of behavioural research has
also been conducted on vertebrate species with the emphasis generally on species of
commercial value or ecological significance. The exact design used in studies,
however, depends largely on the species in question and the nature of the research.
29
Chapter 2 Experimental Design and Methodology
Historically, studies of behaviour have been made from direct observation of the
organism usually involving a code system and direct recording onto paper. More
recently, behaviour observation studies have utilised a variety of approaches
including video, night vision, infra-red and motion sensors. The great advantage with
video technologies, is that a permanent record of the behaviour in question is created.
The observation tape can be played many times which enables multiple analyses and
post hoc correction of interpretive mistakes.
Examples of video technology use in behavioural ecology are widespread. Bozinovic
and Vasquez (1999) used video cameras to study foraging behaviour of the Degu
Octodon degus, a diurnal rodent found in semi-arid Chile; and observed a time-
minimizing foraging behaviour. In this instance, video cameras were used to measure
the overall time budget of the animal, including frequency and duration of patch
visits, and food gathering events. Widowski and Duncan (1996) used overhead video
cameras to study the behaviour of laying hens in response to high and low frequency
flourescent light sources.
In some cases, a large amount of raw video footage is generated in order to observe
the species over a long period, however, it may not be necessary to use the
continuous footage. In a study that examined foraging behaviour of stabled horses,
the subjects were video recorded between 19:00 h and 12:00 h for the duration of the
experimental period (horses were allowed outside for 7 h per day for exercising and
grazing) and behavioural data were collected by time sampling the video tape every 2
min (Winskill et al. 1996).
Experiments which focus on a specific aspect of behaviour require the relevant
patterns to be defined prior to study and will vary accordingly. Greaves and
Wedderburn (1995) used the definitions of lying, standing and grazing when
studying the ability of goats and sheep to affect rates of soil erosion. In a study which
examined the behaviour of fish in response to the presence of a trolling line,
Akiyama et al. (1995) used the definitions of appearance into the camera view,
approach to the lure, attack on the lure, touching the lure, being hooked, and
captured. Behaviour of the European rabbit has been studied extensively in closed
populations and has resulted in the characterisation of many behaviour patterns in
Figure 4.18 – Aggressive behaviour per rabbit at Site 2 for experimental rabbits in April 2002 (n=6). This site was severely affected by dingo activity.
83
84
April 2002 Site3 Control Avg Total Behaviours per Rabbit(avg tot ind. = 29.8; avg tot behav/rabbit = 35.35)
This figure is not available online. Please consult the hardcopy thesis available from the QUT Library
Chapter 5 Discussion and Conclusions
Future Directions of Research and Conclusion
Although the European rabbit is well researched, there are several research
opportunities available to address the knowledge gaps that have been identified in
this project, which would assist management and conservation efforts in the rabbit’s
introduced and native ranges respectively.
The drought conditions and effectiveness of the calci virus program in the semi-arid
region forced the modification of the experimental design used here so that
behaviour experiments were conducted solely in arid conditions on population sizes
at low levels. An interesting study would be to conduct identical experiments in the
semi-arid and arid regions when population sizes are relatively large. However, this
is unlikely to be possible for at least ten to twenty years and possibly longer, given
the current effects of drought and successful implementation of management
programs targeting rabbit refuge areas (Berman, 2004).
Further research opportunities also include quantifying the population size (or
Resource / Individual level) at which dominance hierarchies may break down – i.e.
under what environmental conditions and population densities does the “short-
circuit” of normal rabbit social systems occur? The question may be answered
through the use of computer simulation modelling and direct population size
estimations in the field when threshold densities are achieved.
The aim of this research project was to assess whether wild rabbits adjust their
behavioural patterns as a response to variation in environmental factors, that leads to
observable differences in population genetic structure. Consequently there are two
major outcomes of this project:
1. A difference in population genetic structure was observed at the individual level
between arid and semi-arid regions which supports the findings of Fuller (1995)
and Fuller et. al. (1996, 1997) that identified regional differences using
maternally inherited markers.
2. Differences in the aggressive response to known vs unknown rabbits were
identified in parts of the arid region, which together with the effects of habitat
123
Chapter 5 Discussion and Conclusions
heterogeneity and connectivity (Hamilton, 2003) may explain the observed
differences in population genetic structure.
A major outcome of this study would be if the findings could be utilised to improve
management strategies, particularly those reliant on biological vectors, in countries
where Oryctolagus cuniculus is a significant pest species. Additionally, the outcomes
from this project may assist in better conservation practices in the native range of
southern Europe where the rabbit is an endangered species.
124
APPENDIX 1 Description of rabbit behaviours (modified from Webb, 1988) 1. Grazing Down: Feeding with head lowered near vegetation. 2. Grazing Up: Head raised away from vegetation but still chewing. 3. Resting Alert: Head up with ears erect, but not chewing. 4. Alert: Sitting upright with front legs raised off ground, ears erect. 5. Resting: Inactive with ears flattened, eyes partially or wholly closed. Lying with
legs tucked beneath, lying on side with white belly fur exposed. 6. Grooming: Licking or scratching the fur. Rabbits may flick the front paws
rapidly up and down (“air box”) before grooming the head and ears. 7. Moving: Either slow hops while feeding, or rapid running in response to
disturbance from people, predators etc. 8. Chasing: One individual rapidly pursuing another. The chasing animal may
attempt to bite the fleeing animal if it gets close enough. 9. Displacement: One individual moves toward another resulting in the latter
moving away. Sometimes accompanied by a threat with the head thrust forward and ears flattened.
10. Sexual Following: Male follows female at a slow pace, often stops to sniff the ground where the female been.
11. Circling: Male hops around female. Often accompanied by behaviours 12 and 13. (NOTE: Circling used in this report refers to local animals circling the cage in which the decoy animal was located.)
12. Urine Spray: Male sprays urine over another individual while leaping over or past it. Usually target animal is a female, occasionally a subordinate male.
13. Tail Flagging: Individual hops with rather stiff looking hind legs and raises tail to expose white underside. This behaviour is performed by both sexes during aggressive interactions, although more commonly by males. It is also seen when males are circling females.
14. Tail Wagging: Tail lowered so that black topside is visible and wagged rapidly from side to side. Performed by females towards courting males and towards their own young.
15. Bowing: One individual lowers head and flattens ears as another approaches. Usually performed by females toward males which then proceed to sniff, groom, and chin mark on the female’s head or move around behind the female and attempt to mount. Occasionally performed by a subordinate individual to a higher ranking animal of the same sex, and also by juveniles to adults.
16. Chin Marking: Rubbing the chin over an object, releasing a secretion from the chin (sub mandibular) gland (Myktowycz 1968)
17. Paw Scrapping: Rapid scratching of the ground with the fore paws. Can be either to expose roots during foraging or it is seen before males chin mark, defecate, or urinate during patrolling and territorial marking. Also performed by males during agnostic encounters (see below).
18. Parallel Running and Paw Scrapping: Males (and sometimes females) of neighbouring social groups run in parallel along the territory boundary, occasionally stopping to paw scrape.
19. Fighting (“Aggressive Leaping”): Two individuals simultaneously jump towards each other. They pass in the air, land, and then repeat the process in the opposite direction. Usually these jumping fights are brief, 1-5 leaps. Mainly seen
125
between males on territory boundaries, occasionally between two females. Individuals from the same group sometimes interact in this way if they have come into close contact “unintentionally” eg. If one of them is engaged in a rapid chase.
20. Fighting (involving close contact): Individuals locked together in combat comprising vigorous scratching with the hind legs and biting.
126
Bibliography Bibliography
Akiyama, S., Yasuda, K., Arimoto, T., Tawara, Y. (1995). Underwater observation
of fish behaviour to trolling line. Nippon Suisan Gakkaishi-Bulletin of the
Japanese Society of Scientific Fisheries. 61(5): 713-716.
Alcock, J. (1984). Animal Behaviour: An Evolutionary Approach. Sunderland,
USA, Sinauer Associates.
Alcock, J. (1998). Animal Behaviour: An Evolutionary Approach.6th Edition.
Boston, USA, Sinauer Associates.
Alonso, J. C., Alonso, J. A., Bautista, L. M., Munoz, P.R. (1995). “Patch use in
cranes: A field test of optimal foraging predictions.” Animal Behaviour
49(5): 1367-1379.
Berman, D. (2001). Personal Communication.
Berman, D. (2004). Bulloo bunnies bite the dust. Mulga Line Newsletter Dec