The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 ›...

The Future of British Odonata

Determining temporal range dynamics from distribution patterns and dispersal

Alyson Pavitt

MSc Conservation Science

Imperial College London, Silwood Park

A thesis submitted in partial fulfilment of the requirements for the degree of Master of

Science and the Diploma of Imperial College London.

PDF processed with CutePDF evaluation edition www.CutePDF.com

http://www.cutepdf.com

© Alyson Pavitt, 2011 [email protected]

Contents

1. Introduction ..................................................................................................................................... 1

1.1. Statement of need ..................................................................................................................... 1

1.2. Range dynamics ........................................................................................................................ 2

1.3. Distribution and dispersal ....................................................................................................... 3

1.4. Aims and objectives .................................................................................................................. 4

1.5. Hypotheses ................................................................................................................................ 4

2. Background ...................................................................................................................................... 7

2.1. Analysis of range dynamics .................................................................................................... 7

2.2. Odonates as a model taxon ................................................................................................... 10

2.3. Spatial distribution patterns .................................................................................................. 12

2.4. Use of traits in understanding species distribution ........................................................... 15

3. Methods and materials ................................................................................................................. 20

3.1. Range dynamics ...................................................................................................................... 20

3.2. Patterns of distribution .......................................................................................................... 21

3.3. Trait data .................................................................................................................................. 22

3.4. Data analysis............................................................................................................................ 23

3.5. Analysis of measurement methods ...................................................................................... 24

4. Results ............................................................................................................................................. 25

4.1. Range dynamics ...................................................................................................................... 25

4.2. Patterns of distribution .......................................................................................................... 26

4.3. Morphometric analysis .......................................................................................................... 27

4.4. Dispersal capacity ................................................................................................................... 32

4.5. Representativeness of museum specimens ......................................................................... 33

5. Discussion ...................................................................................................................................... 34

5.1. Spatial dynamics ..................................................................................................................... 34

5.2. Patterns of Distribution .......................................................................................................... 36

5.3. Dispersal morphometrics ...................................................................................................... 37

5.4. Determining occupancy change ........................................................................................... 38

5.5. Suborder variation .................................................................................................................. 42

5.6. Value of historic collections ................................................................................................... 42

5.7. Recommendations for future research................................................................................. 43

6. Conclusions .................................................................................................................................... 45

References........................................................................................................................................... 46

Appendices ......................................................................................................................................... 62


Acronyms

AOO Area of Occupancy

BAP Biological Action Plan

BDS British Dragonfly Society

BRC Biological Records Centre

EOO Extent of Occupancy

IUCN International Union for Conservation of Nature

NBN National Biological Network

NHM Natural History Museum, London

T1 Time period 1 (1970-1996)

T2 Time period 2 (1997-2008)

VCA Variance Component Analysis


Figures

Page

4.1 The relationship between the fractal dimension of species distribution (D) and cell

occupancy. The change in shape between t1 and t2 is shown for each species,

displayed with the linear regression model for mean species residual D against mean

species cell occupancy (grey line). The dotted lines are species with a positive

residual D change, these species became more aggregated from t1 to t2. The solid

lines are species with a negative residual D change, and these are species which

became sparser across time.

27

4.2 Linear regression models for the relationships between (a) dry wing length and the

logs of abdominal length, (b) thoracic volume and the logs of abdominal length3, and

(c) wing aspect ratio and the logs of abdominal length2. Solid lines show the linear

regression fitted straight to the data, whilst dashed lines show the linear regression

once phylogenetic signal is controlled for.

28

4.3 Relationship between the log of wing aspect ratio and residuals of dry wing length

(controlling for phylogenetic signal) in anisopterans (filled points and solid

regression line) and zygopterans (empty points and dashed regression line).

30

4.4 Difference between the two primary methods of measuring odonate wing length.

This is based on length measurements taken from the right forewing of Coenagrion

puella (see Conrad et al. 2002; Hassall et al. 2008 for complete methodologies).

31

4.5 Distribution of occupied grid cells in the British ranges of (a) Orthopterum cancellatum

(a species increasing in occupancy) and (b) Calopteryx virgo (a species decreasing in

occupancy),for t1 (left) and t2 (right).

39

4.6 Distribution of occupied grid cells in the Britain range of Aeshna isosceles for t1 (left)

and t2 (right).

40


Tables

Page

2.1 Comparison of the two main methods used to measure species range shift.

10

2.2 Variables responsible for determining species distribution patterns, grouped

according to their influences as defined by Guisan & Thuiller (2005).

12

2.3 Comparison of methods used to measure spatial distribution pattern.

13

2.4 Selection of species traits which have been found to correlate significantly with

changes to invertebrate ranges.

16

2.5 Morphological traits used as proxies for dispersal capacity in a range of species.

18

4.1 The number of species which showed cell occupancy or range shift changes. The

binomial sign test shows whether a significantly greater proportion of species

showed a range expansion or northwards shift).

25

4.2 The number of species with more or less aggregated distributions (residual D), the

number of species changing over time (residual D change), and whether there was a

significant proportion with more aggregated or increasingly aggregated ranges

(binomial sign test).

26

4.3 The variance in dry wing length and aspect ratio accounted for by each of the nested

taxonomic scales (for n= 572 specimens).

29

4.4 The variance in dry wing length and aspect ratio accounted for by inter- (between

species and all higher taxonomic levels) and intraspecific (between and within sex)

variation. The binomial sign test shows whether a significantly greater proportion of

variation was accounted for interspecific variation.

30

4.5 The number of species with a positive change index (expansion in overall cell

occupancy from t1 to t2). The binomial proportion test shows whether there is a

significant difference in the proportion of each sub-order showing a cell occupancy

increase. Percentages given as a proportion of n.

32

4.6 Statistical outputs from Welch two sample t-tests investigating differences in wing

length between dry museum specimens and wet specimens (recently collected).

33


2

Abstract

Temporal range dynamics (range shift and change index), dispersal morphometrics

(wing length, wing aspect ratio, and thoracic volume), and distribution pattern

(residual D) were investigated in British Odonata. Initial analyses discounted range

shift from this study due to the absence of evidence that the data were showing

directional shift rather than non-directional expansion. A significant proportion of

change index (Telfer et al. 2002) was described by combining the three dispersal

traits and residual D. Species with increasing range size were those with long, broad

wings, large thoracic volumes and a more aggregated distribution. This morphology

is found in the anisopteran (dragonflies) suborder, which showed a substantially

greater increase in occupancy than the smaller and weaker zygopterans

(damselflies). In a preliminary investigation into the extant representativeness of

museum collections, there was found to be no differences in wing length with

recently caught, wet specimens.

Key words: Dragonfly, damselfly, anisoptera, zygoptera, change index, range shift, fractal

dimension


2

Acknowledgements

I thank my supervisors Dr. Nick Isaac (CEH) and Dr. Simon Leather (Imperial

College) for their comments, support and guidance; Steve Brooks (NHM) for all his

help and advice throughout the project; Gary Powney for his assistance with

statistical analyses; Dr. Christopher Hassall, Dr. Kelvin Conrad and Philip Bowles

for their kindness in making their data available to me; and to everyone else who has

supplied advice or comment on this work. I also extend thanks to the inexhaustible

support, hugs and cake-eating of SW3, and to tea, without which none of this would

have been possible.

Word count: 9,472


1

1. Introduction

1.1. Statement of need

Under the combined force of multiple environmental and anthropogenic drivers,

many species are undergoing extensive change to their global ranges. Past work has

focussed on the impact of rising global temperatures (Houghton et al. 2002), but this

is just one of the many pressures that species are under. The relative impacts of these

drivers are often species specific, limiting the potential for proactive conservation

without extensive, single-species research. Most species lack such detailed

information, and there is neither the time nor the available resources to fill these

knowledge gaps. In species with long historic records, however, it may be possible

to study long term changes in light of more general species traits, with the potential

to build a database of traits which are frequently associated with declining species

ranges. Achieving this requires the in depth analysis of multiple and disparate taxa,

with a heavy reliance on those for which substantial and long-running records

already exist.

The odonates (order Odonata) are one such taxon, with records from amateur

naturalists dating back centuries, although their value in this field has only recently

been recognised (e.g. Hickling et al. 2005; Hassall et al. 2010). Through the study of

their range dynamics and associated traits, this study will investigate the predictive

potential of dispersal morphology and distribution pattern, testing the belief that key

traits can indeed be used to predict species temporal changes. Findings of this study

will also be important for national odonate conservation by providing additional

means of determining species most at risk. Species with chronic declines, or

increasing range fragmentation, may otherwise be overlooked by conservationists

working with small snapshots of present distribution, particularly if they remain

fairly common.


2

1.2. Range dynamics

Focussing on changes to the odonates’ British distributions, this study will utilise

two different range dynamic metrics: range shift and occupancy change (see

Appendix I). Range shift is characterised by a series of low latitude extinctions and

high latitude colonisations (Malcolm et al. 2006; Thomas, 2010). Extensive research

almost exclusively attributes this to shifting climate envelopes under climate change

pressures (Elith & Leathwick, 2009). Past focus has primarily been on birds and

butterflies (Parmesan et al. 1999; Warren et al. 2001; Poyry et al. 2009), but similar

trends are increasingly being found across a wide variety of taxa (Hickling et al.

2006). Drivers of occupancy change, however, are far less certain. Although evidence

suggests that there is an indirect link with climate (Robinet & Roques, 2010), this is

only one of a suite of determinants. Other pressures include habitat loss and

fragmentation, invasive species, and over-exploitation (Gaston, 1996; Korkeamaki &

Suhonen, 2002; Hassall & Thompson, 2008). These are being counteracted in some

instances by improved habitat quality, remediation and conservation work (Willis et

al. 2009), resulting in complex effects and interactions.

Whilst regional scale studies are needed to fully understand range dynamics

(Parmesan, 1996), national or local level research can still be hugely important,

particularly in informing national conservation policy and management. Range shift

poses the most problems for localised monitoring, because it will not incorporate

both northern and southern range margins. Instead it can only be interpreted from

the uni-directional, polewards expansion of the range margin under study.

1.2.1. Study taxa

Macro-invertebrates are particularly useful for studying range dynamics because of

their short generation times (Parmesan, 1999; Thomas et al. 2004). Many also have

long running national data sets (e.g. in Britain) because their ease of collection and

charisma made them appealing to amateur naturalists. This, coupled with its

position at the dynamic margins of many of the species ranges (Wilson et al. 2004;

Hassall et al. 2010), makes Britain an ideal site for entomological range dynamics


3

study. Until recently, research in this field centred on butterfly responses (e.g. Hill et

al. 1999; Wilson et al. 2004; Willis et al. 2009), however odonates are now gaining

recognition as an order of equal, if not greater importance. Not only do odonates

encompass both aquatic and terrestrial habitats (Hickling et al. 2005), they are alos

sufficiently small in number (~6,000 species) to allow a large, representative

proportion to be studied in detail (Clausnitzer et al. 2009).

1.3. Distribution and dispersal

Species range dynamics are heavily influenced by internal distribution patterns (see

Appendix I), and the ability of individuals to disperse between patches of suitable

habitat (Hill et al. 1999). It is important to consider interactions between the two, as

landscape may be more permeable to a sparsely distributed but mobile species, than

to a more sedentary species with an aggregated distribution (Thomas et al. 1992;

McCauley, 2006). This has implications for both the colonisation of habitat beyond

current range boundaries, and the recolonisation of habitat following local

population extinction.

1.3.1 Patterns of distribution

Species are not evenly distributed across their range, but instead form unique

patterns of occupancy and absence through a complex assortment of biotic and

abiotic interactions (Wilcove et al. 1986; Fischer & Lindenmayer, 2007). Past work on

butterflies suggests that a fixed snapshot of this spatial pattern can predict the

species’ range dynamics across time, with more aggregated species showing range

expansions (Wilson et al. 2004). By applying methods used in this study to a new

taxon (Odonata), this project will investigate the value of distribution pattern as a

predictive metric of range dynamics.

1.3.2. Dispersal capacity

Because direct measures of dispersal distance are less accessible as predictors, and

are unknown for a number of species, this study will take three accepted

morphological proxies of dispersal instead: wing length, wing aspect ratio, and


4

thoracic volume (Malmqvist, 2000; Hughes et al. 2007; Meyer et al. 2008). The

relationship between wing length/aspect ratio and flight capacity are well

documented in the literature. Long wings and higher aspect ratios (narrow wings)

have long been associated with economical flight and dispersal (Conrad et al. 2002;

Rundle et al. 2007; Saino et al. 2010). Wing morphology, however, provides little

information on the energetic requirements for flight, and thus its relationship with

thoracic volume, heavily influenced by flight musculature (Pringle, 1949), is an

informative one.

1.4. Aims and objectives

The primary focus of this study is on identifying temporal changes in British

odonates, and whether they show any general trends with dispersal morphology

and distribution pattern. It also aims to investigate the value of museum collections

as sources of this morphological trait data.

With these in mind the objectives of the project are to:

1.4.1. Create an up-to-date picture of how species ranges are changing over time,

both in latitude (range shift) and overall size (occupancy change).

1.4.2. Investigate the potential for both static and dynamic measures of species

distribution pattern to predict future range changes.

1.4.3. Collate and generate dispersal trait data, and investigate their relationships

with range dynamics.

1.4.3. (a) Investigate variation in different methods of measuring morphology.

1.4.3. (b) Investigate whether traits show the greatest level of variation within or

between species.

1.4.4. Investigate how representative museum specimens are of extant populations.

1.5. Hypotheses

1.5.1. Overall, significantly more species will show a polewards shift at their

northern range margin. The range margins of all southern species will shift


5

northwards. Northern species will show no change at the polewards range margin

due to geographic barriers (e.g. British coastline).

1.5.2. Overall, significantly more species will increase their cell occupancy between

the two time periods as conditions become more suitable under shifting climate

envelopes. Northern species will be the only species to decrease in overall cell

occupancy as fewer areas become suitable for them.

1.5.3. There will be a significant positive correlation between northwards range shift

and change index as species ranges move further into Britain. There will not be a

corresponding movement at the eastern range margin, indicating a directional

polewards shift rather than non-directional range expansion.

1.5.4. Species with sparse distributions will decline in cell occupancy due to local

extinctions which were not recolonised. More aggregated species will remain stable,

or increase in cell occupancy. More aggregated species will also show a greater

northwards range shift.

1.5.5. Species becoming sparser will show a greater decline in cell occupancy as

localised extinctions are not replaced.

1.5.6. There will be a significant relationship between wing length, wing aspect ratio,

and thorax volume, and together these will describe a significant proportion of

variation in occupancy change and range shift.

1.5.7. The combination of species morphology and distribution pattern will describe

the most variance in occupancy change and range shift.

1.5.8. Wing length will not vary between different measurement methods.

1.5.9. The greatest variation in wing morphology will be seen above the level of the

species, rather than being intraspecific.

1.5.10. Wing morphology will not vary between museum specimens and wet

specimens.

1.6 Scope

This project is concerned with tracking temporal changes in the distribution of

odonates within Britain, and identifying whether these changes can be pre-empted

with an understanding of the species’ dispersal capacity and distribution pattern.


6

Two forms of temporal change will be measured; change in the number of grid cells

occupied (occupancy change) and latitudinal movement at the margins of the range

(range shift). Interspecific variation in three morphological dispersal traits (wing

length, wing aspect ratio, and thoracic volume), and in distribution pattern (Wilson

et al. 2004) will be analysed in relation to these range dynamics, both individually

and together. The morphological data, collected from Natural History Museum

(NHM) specimens, will also be compared to more recently collected wet

measurements, in order to investigate the assumption that morphometric data from

these old collections is representative of extant populations.


7

2. Background

2.1. Analysis of range dynamics

Whilst much can be learnt about a species’ ecology from the size and distribution of

its range (Gaston, 2003), this static measure is of limited use in understanding its

long term viability. Species with small ranges (geographically rare), for example,

may be stable and at optimal population size, whilst those with extensive ranges

may be rapidly declining and in need of conservation attention (Gaston, 1994). It is

therefore vital that not only the range size, but also temporal changes to that size, are

understood. This ensures that limited conservation resources are targeted at the

species with the most immediate risk.

2.1.1. Drivers

Whilst many studies investigating species range dynamics have focussed on the

escalating issue of climate driven changes (e.g. Parmesan, 1996; Parmesan & Yohe,

2003; Hickling et al. 2006; Termaat et al. 2010), it is important that this is not treated

as a default driving force of all change. Even where species ranges are shifting in

response to corresponding shifting climate envelopes (Elith & Leathwick, 2009), it is

unlikely that they are responding in isolation from other drivers. Both range shift

and occupancy change are ultimately determined by a unique combination of

internal pressures, such as habitat transformation and altered interspecific

interactions (Parmesan & Yohe, 2003; Hofmann & Mason, 2005). Some of these may

directly or indirectly result from climate change, but this is not always the case.

Some may be caused by population stochasticity generating “noisy biological data”

(Parmesan & Yohe, 2003), or result from other anthropogenic factors such as

pollution (Goffart, 2010) or land conversion. In such cases, range is limited by the

suitability of the habitat, not by temperature tolerances, and must be interpreted as

such. Indeed, Parmesan & Yohe (2003) claim the latter of these to have had more

influence on 20th century range shifts than any climate change effects.


8

2.1.2. Methods of analysis

Metrics of range change have traditionally been based on the difference in species

distribution between two separated time periods of fixed length (e.g. Hickling et al.

2006; Hassall et al. 2010). This convention has become less absolute in recent years,

however, with unequal time periods being used to account for record effort (e.g.

Termaat et al. 2010). Analyses tend to fall under one of two categories, either

measurements of changing range size (“occupancy change”) or shifts to the overall

species range.

Occupancy change

Change in a species’ habitat occupancy can be measured using a number of different

methods which are dictated by the form or quality of the input data: see Gaston &

Fuller (2009) for a summary of these. By putting a value on the actual area of a range,

area of occupancy (AOO) may provide the most information (e.g. Warren et al. 2001),

but only tends to work on the small scale (Termaat et al. 2010). For large scale

monitoring schemes such as species atlases and national databases, change needs to

be calculated from coarser measurements of grid cell occupancy (Tefler et al. 2002;

Hui, 2011). These rely on distribution maps built from binary records of species

presence or absence within a regular system of grid cells, and can therefore only

provide a measure of temporal change in the number of occupied cells at a given

scale. Other methods that have been utilised include extent of occupancy (EOO)

which maps the distance of the outer limits of the distribution (Gaston, 1996). Whilst

this can be useful (e.g. Jetz et al. 2008), it is likely to be of limited value when tracking

temporal changes, particularly in light of species conservation and management.

Whilst the value of range size and occupancy change cannot be overestimated

(Gardenfors et al. 1999), a huge variation in scale (from globally ubiquitous species to

single site endemics) means that no standard monitoring methodology has yet been

implemented (Segurado & Araujo, 2004).


9

Range shift

Because of its association with climate change, species range shift has been

extensively studied, although only a few groups (primarily birds and butterflies)

have been studied in great detail, making generalisations difficult. Recent analyses

by Hickling et al. (2006), however, found the responses of such groups to be

representative of trends seen across a range of less studied taxa. They analysed

British distribution data for 16 taxonomic groups with ranges in the south of Britain.

This included well studied species, such as birds (Aves) and butterflies

(Rhopalocera), and understudied species like millipedes (Diplopoda). Across the

taxa, 84% (of 329 species) were found to shift polewards at their range margin,

averaging 31-60 km between the two time periods studied.

There are two main methods which can be utilised to measure range shift (Table

2.1), the applicability of which vary depending upon the scale of the study and

quality of the available data. Ideally range shift analyses should consider the entire

range of a species, but in reality studies tend to focus on the more manageable local

level, and usually only at one latitudinal range margin. Whilst tracking the mean

range shift would be preferable, encompassing more individuals and taking into

consideration latitudinal clines in morphology and behaviour (Hassall et al. 2008,

2009: Table 2.1), this is unlikely to detect smaller scale changes (Shoo et al. 2006). In

light of this, and despite questions over the representativeness of the method, range

shift continues to be studied predominantly at the range margins (e.g. Hickling et al.

2006; Hassall et al. 2010).


10

Table 2.1: Comparison of the two main methods used to measure species range shift.

Marginal shift Central shift

Method

Shift is measured as change at (or

close to) the margin of the range

(Hickling et al. 2005, 2006; Hassall et al.

2010).

Shift is measured as a change to the

mean/ centre of the range (Mattila et

al. 2011).

Advantages

Detects small changes, making it

useful at the national and local level.

Provides a picture of overall range

change because it considers change to

the spread of a large number of

individuals under a range of different

conditions (Archaux, 2004).

Disadvantages

Cannot conclusively show range shift

is occurring from only a subsection of

the range (i.e. from only one margin).

Dependent on a small number of grid

cells, so is more affected by varying

recorder effort (Shoo et al. 2006).

Based on extreme sightings which

may fall outside the species’ normal

range.

Produces smaller range shift values,

so is unlikely to detect any changes at

the local level (Shoo et al. 2006).

2.2. Odonates as a model taxon

Globally there are around 6,000 known odonate species, 39 of which breed in Britain,

with a further 12 as regular migrants (Corbet & Brooks, 2008). In recent years several

migrants have established permanent colonies in Britain (Merritt et al. 1996; Parr,

2010), leading to disagreement over the exact number of “resident” species, though

the commonly accepted migrants and residents are listed in Appendix II.


11

2.2.1. Life history

Odonata is an ancient order of insects which has remained largely unchanged for

over 300 million years. The extant species are classified into two sub-orders based on

the morphology of their wings: the zygopterans (“similar wing”), small-bodied,

slender “damselflies”, and anisopterans (“dissimilar wing”), the larger bodied

“dragonflies”. Odonates are “morphologically generalised” (Corbet & Brooks, 2008)

insects, with a larval form resembling the adult in all but the presence of fully

developed wings. Almost all species have obligatory aquatic larvae before they

develop into airborne imagos.

2.2.2. A model taxon for range dynamics research

Odonates are one of the best studied insect taxa, with the British Dragonfly Society

(BDS) database holding records dating back as far as 1807 (Hassall et al. 2007).

Despite this, the impact of abiotic interactions on their spatial distribution is only just

beginning to receive attention. Within Britain, work is currently underway to

produce an up to date dragonfly atlas (due for publication by the BDS in 2013), and

the past few years have seen several publications investigating the causes and

impacts of range change in this taxa. Hassall et al. (2010) used range shift analyses to

show a northward shift in all British odonate families, predicting changes to species

distribution which could affect the reliability of odonate larvae as water quality

indicators. Hickling et al. (2005, 2006) found the same trend in data from the

Biological Records Centre (BRC) for both southerly and northerly species. The only

species to be adversely affected were two of the four northern species (Aeshna

caerulea and Leucorrhinia dubia), which showed overall range declines as their

margins move further into Scotland (Hickling et al. 2005).

Despite being less studied than lepidopterans, odonates may have the advantage as

a study group, in that there are sufficiently few species to allow most, if not all, to be

studied individually. Numbering in the few thousand means there is scope for a

more comprehensive species-level analysis, rather than relying on a small subset of

species. This has allowed for the first global assessment of an insect order to be


12

conducted. Clausnitzer et al. (2009) studied and mapped the distributions of over

25% of all known odonates (1,500 species), to identify the order’s threat level. Such

detailed analyses are not possible for groups like Lepidoptera where over 135,000

species have already been identified (Kristensen, 1999).

2.3. Spatial distribution patterns

Species ranges are not homogenous across a landscape, but are aggregated in

accordance with the suitability of the immediate environment and underlying

habitat. This depends on a wide range of factors including inter-specific interactions

such as predation, competition and parasitism (Williams et al. 2001; Begon et al. 2005;

Martinez et al. 2008), as well as interactions with the abiotic environment (Guisan &

Zimmermann, 2000; Huston, 2002; Guisan & Thuiller, 2005: Table 2.2). This

distribution pattern is a key consideration when attempting to understand the

determinants of species distribution and range changes.

Table 2.2: Variables responsible for determining species distribution patterns, grouped

according to their influences as defined by Guisan & Thuiller (2005).

The exact determinants of a species geographic distribution has long been

considered a key question in ecology (Levin, 1992), with little evidence that it will be

resolved any time soon (Hartley et al. 2004; Gelfand et al. 2006). One of the central

problems is that these variables work at different scales, and so are often not

compatible. In a recent botanical study comparing patterns at different spatial scales,

Influence Definition Examples

Regulators “factors controlling species

eco-physiology”

Temperature, soil

composition

Disturbances “perturbations affecting

environmental systems”

Natural or anthropogenic

habitat change

Resources “components that can be

assimilated by organisms”

Energy, water


13

Hartley et al. (2004) noted each scale to be driven by a different key variable. For

example, the “intermediate scale”, used in this analysis to describe national (British)

scale patterns, was determined predominantly by habitat distribution and

persistence. Meanwhile, “fine scale” patterns were found to be more dependent on

sampling consistency and intraspecific competition for space. As expected,

determinants of distribution pattern correlated at similar scales, but become

increasingly dissimilar as the difference between scales increased.

2.3.1. Methods of analysis

It has been proposed that by understanding how distribution patterns correlate with

range changes, predictions of future change could be made based on a present day

snapshot of their distribution (Wilson et al. 2004; Hui, 2011). This could have huge

implications for conservation, were specific patterns of distribution indicative of

species or populations most vulnerable to decline. Numerous methods have been

employed to provide a metric for spatial distribution pattern (Table 2.3), though

many focus on the patterns of individuals within a population, rather than at the

regional level (Davis et al. 2000).

Table 2.3: Comparison of methods used to measure spatial distribution pattern.

Nearest neighbour Ripley’s K (or L) function Scaling patterns of

occupancy (SPO)

Data

Individuals recorded as

point data within a

population, assuming

randomly selected

sample points and spatial

continuity.

Point data, assuming all

points have been selected

within a predefined study

area.

Binary point data on a grid

system based on

presence/absence data,

where distances between

sampling points are the

same (eg/ species atlas).


14

Table 2.3 cont: Comparison of methods used to measure spatial distribution pattern.

Nearest neighbour Ripley’s K (or L) function Scaling patterns of

occupancy (SPO)

Scale

Primarily distance

between individuals,

though recent

developments have

shown it to work at larger

scales (Davis et al. 2000).

Can describe patterns

between individuals at

multiple scales (e.g.

Wiegand & Moloney,

2004).

Can describe patterns at

multiple scales, but works

particularly well at larger

landscape scales (e.g.

Wilson et al. 2004)

Methodology

Compare the distance

between closest data

points with the expected

distance to (a) just the

nearest neighbour (Clark

& Evans, 1954) or (b)

every other point (e.g.

David et al. 2000).

Sums records within a

given radial distance

around each data point

(Ripley, 1977).

Compare summed

occupancy at two different,

nested scales (e.g. Wilson

et al. 2004).

Results

Identifies whether a

population is clumped,

uniform or randomly

distributed. Populations

will show uniform

distributions when all

individuals are a similar

distance apart (e.g. Smith

& Murphy, 1982).

Populations with

aggregated distributions

will have more occupied

cells in close proximity to

the central data point.

Sparse or randomly

distributed species will

have similar numbers of

records at all distances

from it (Dixon, 2002).

Relative to occupancy at

the fine (small) grid cell

scale, species with

aggregated distributions

show a lower occupancy at

the coarse (large) grid cell

scale, than those which are

sparsely distributed.


15

Whilst the nearest neighbour method and Ripley’s K function serve important roles

in understanding the distribution patterns of individuals within a population, this

cannot readily be scaled up for landscape level studies of range dynamics. In order

to generate a meaningful measure of species distribution pattern, extensive data sets

across large spatial scales are needed (e.g. species atlas and national surveys). Data is

usually presented in the form of binary values of grid cell occupancy (presence/

absence), rather than treating each individual sighting as its own data point. The

limited amount of data available at this scale means that these finer analyses are not

applicable, instead, coarser methods like SPO need to be utilised (e.g. Wilson et al.

2004).

2.4. Use of traits in understanding species distribution

It is possible to consider each species as a system comprising of a unique set of trait

variables (see Table 2.4). In understanding the relationships between these traits and

species distribution, it may be possible to predict how a species (or species with a

similar set of traits) will respond to change. In an ideal situation, each species would

be studied in comprehensive detail, however limited resources make this impossible.

If traits are found to significantly correlate with distribution change, as is suggested

by numerous studies (Dukes & Mooney, 1999; Hill et al. 1999; Holt, 2003; Rundle et

al. 2007; Poyry et al. 2009; Termaat et al. 2010), then they have the potential to predict

under-studied species most at risk. This will allow for better management and the

focussing of limited conservation resources. They may also be used to roughly

predict future responses to climate change events, enabling a proactive conservation

management approach.


16

Table 2.4: Selection of species traits which have been found to correlate significantly with


Trait Reference

Morp

hological traits

Body size Body size positively correlated with

dispersal and range shift in odonates.

Angelibert &

Giani (2003)

Wing aspect

ratio

Aspect ratio was positively correlated

with dispersal capacity in odonates.

Hassall et al.

(2008)

Wing: body

length ratio

Wing: body length ratio was positively

correlated with range size in odonates.

Rundle et al.

(2007)

Wing length/

span

Wing length/ span were positively

correlated with dispersal, range expansion

and range size in butterflies and odonates.

Rundle et al.

(2007); Poyry et

al. (2009)

Wing loading Wing loading negatively correlated with

dispersal capacity in odonates.

Hassall et al.

(2008)

Eco

logical traits

Agricultural

tolerance

Butterflies found to tolerate agricultural

land showed greater range size and

expansion.

Poyry et al.

(2009)

Breeding habitat Butterflies breeding in the forest edge

showed a greater northwards expansion

than those in open grasslands.

Poyry et al.

(2009)

Diet breadth Food plant number was positively

correlated with range size in butterflies.

Slove & Janz

(2011)


17

Table 2.4 cont: Selection of species traits which have been found to correlate significantly with


Trait Reference

Eco

logical traits

Distribution

size

Distribution size was positively correlated

with range expansion in butterflies.

Poyry et al.

(2009)

Growth rate Growth rate was positively correlated with

range expansion in butterflies.

Bale et al. (2002)

Habitat

preference

Beetle species in coniferous forests had

larger ranges than those in broadleaf.

Species of odonate in grasslands and lotic

waterways had much smaller ranges, and

range shifts than in other habitats.

Bense (1995):

Clausnitzer et al.

(2009)

Habitat

tolerance

There was a positive correlation between the

number of habitats occupied and range shift

in odonates.

Hill et al. (1999)

Larval host

plant

Species of butterfly found on woody plants

showed a greater northwards expansion

than those feeding on herbaceous plants or

grasses.

Poyry et al.

(2009)

Mobility Dispersal capacity was positively correlated

with range shift in odonates.

Hill et al. (1999)

Threat status

(red list)

Non-threatened species of butterfly showed

poleward shifting ranges, whilst threatened

(red list) species remained stationary.

Mattila et al.

(2011)


18

2.4.1. Dispersal ability

It is important to consider the capacity for individuals to disperse through the

landscape, particularly in light of growing threats such as climate change. Actual

measures of dispersal distance are rare however, so morphological traits tend to be

relied upon as proxies (e.g. Table 2.5).

Table 2.5: Morphological traits used as proxies for dispersal capacity in a range of species.

Morphology Dispersal capacity Examples

Aspect ratio (wing shape) Higher aspect ratio (narrow

wings) is associated with

migratory species with long

dispersal distances.

Odonates (Conrad et

al. 2002); Birds (Saino

et al. 2010)

Body size (proxies inc.

abdomen length and wing

length; Cordero, 1994)

Larger species have longer

(odonates & multiple species)

or shorter (butterflies)

dispersal distances.

Butterflies (Mattila et

al. 2011); Multiple

species (Jenkins et al.

2007); Odonates (Allen

et al. 2010)

Wing length Long wings are associated

with migratory species with

long dispersal distances.

Birds (Rayner, 1988);

Odonates (Conrad et

al. 2002)

2.4.2. Problems with the use of traits

With a few exceptions (e.g. Hassall & Thompson, 2008; Hassall et al. 2008, 2009) most

invertebrate studies have focussed on butterflies. This large bias means that many

conclusions drawn are largely based on the work from one or two taxa, and thus will

be hugely influenced by taxon-specific characteristics. This emphasises a need for

further research in a more representative range of taxa since many traits, particularly

those focussed on the morphology of the species, are likely to be fairly taxon specific.

For example, wing aspect ratio (the length/ width of the wing) has been found to


19

correlate positively with dispersal capacity in odonates (Hassall et al. 2008), but

negatively in some butterfly studies (Hill et al. 1999). Body size is another such

example. Recent research on one of the smaller British odonates, Ischnura pumilio,

suggests a positive relationship between species size, dispersal distance and range

shift (Allen et al. 2010). Work on butterflies suggests that the opposite is true, with

larger species being more prone to range contractions (Mattila et al. 2011).

Due to the heritable nature of species traits, their use can be statistically invalidated

if species relatedness isn’t considered and controlled for. One of the core

assumptions for statistical analysis is that data points are independent from one

another; comparing species which share a common ancestor breaches this

assumption. Furthermore, some species will be monophyletic to the exclusion of

others. This problem has been dealt with in a number of ways, the commonest of

which is through the use of independent contrast scores (Rundle et al. 2009),

generated from comparative analyses such as Comparative Analysis by Independent

Contrast (CAIC) (Purvis & Rambaut, 1995). This allows taxa to be compared

independent of their phylogenetic relatedness.

Although frequently used in the literature, direct body measurements should be

treated with caution because of the huge number of other variables which confound

them. Because the body size of individual insects is so heavily influenced by

variables such as temperature and diet (Ikeya et al. 2002; Nijhout et al. 2006; Parker &

Johnston, 2006), even minor variations in conditions may have a noticeable impact.

Individual physiological plasticity is not the only issue; whilst wings remain largely

unchanged, body size will vary substantially between “wet” and dry specimens

because of the water content of the body. Indeed, it has been estimated that insects

can lose up to 20% of their length during post-senescence drying (Schoener & Janzer,

1968). Subsequently, the conditions and duration of storage will have a significant

bearing on measurements obtained. Body mass measurements are also problematic,

particularly in species which eat as adults, as they will be highly dependent on the

food availability and time from emergence (Johansson, 2003).


20

3. Methods and materials

3.1. Range dynamics

3.1.1. Data sources

Data on species’ distributions within Britain were obtained from National Biological

Network (NBN) Ordnance survey maps of cell occupancy (for n=37 species),

primarily from records held with the BDS. Cell occupancy was grouped into two

time periods of approximately equal record size: 1 Jan. 1970- 31 Dec. 1996 (t1: n=

309,830 records), and 1 Jan. 1997- 31 Dec. 2008 (t2: n= 312,571 records). Cell

occupancy values were taken for 10 x 10 km grid cells at t1 (t1_10) and t2 (t2_10), and

for 100 x 100 km grid cells at t1 (t1_100) and t2 (t2_100) (see Appendix I).

3.1.2. Change index

Occupancy change was measured using the “change index” method (see Appendix

I) developed by Telfer et al. (2002). T1 and t2 (at 10 km scale) were converted to

proportions of the total survey area, logit transformed, and a linear regression model

was fitted. Species change index was taken from standardised residuals of this

model (see Telfer et al. 2002 and Telfer, 2003 for complete methodology). This

method assumes that recorders report all species within each grid cell equally, and

controls for the effects of variation in geographic coverage and recorder effort by

restricting analyses to grid cells recorded in both time periods.

3.1.3. Range shift

Using records of cell occupancy, range shift was calculated to investigate the

directionality of species range movements in response to climate change. Range shift

was quantified as the difference in the northern range margin between t1 and t2 on

the 10 km scale. For each time period, the mean northing was taken from the species’

top ten occupied grid cells within Britain, and the distance (in km) between them

was used to generate overall range shift (Warren et al. 2001; Hickling et al. 2005,

2006). This is the most frequently used range shift metric (Thomas & Lennon, 1999;

Warren et al. 2001; Hickling et al. 2005, 2006; Hassall et al. 2010) because it only


21

requires 20 grid cells to be occupied at each time period (i.e. is not hugely data

hungry). As with the change index method, it also accounts for variation in

geography and recorder effort by using only grid cells recorded in both time periods.

3.2. Patterns of distribution

Work on butterflies suggests that a species distribution pattern can be indicative of

its temporal range dynamics (Wilson et al. 2004). This could have important

implications for conservation if a stationary snapshot can be used to interpret species

trends without several decades’ worth of distribution data.

A metric for distribution pattern (fractal dimension) was calculated using methods

derived from Wilson et al. (2004). “Fractal dimension” (D) was defined as “a

descriptive measurement of spatial aggregation over a narrow range of scales”

(Wilson et al. 2004), rather than being a measurement of the true mathematical fractal

dimension at infinite scales. For each species the number of occupied cells at t1_10

and t1_100 were individually summed and converted to AOO. The log10 of cell side

length (a measure of scale) was plotted against the log of AOO, and D was calculated

from the resulting slope using the formula:

D= 2-(y2-y1)/(x2-x1)

where x is the log10 of the cell side and y is the log of AOO. D2 was calculated in the

same manner using the number of occupied cells at t2_10 and t2_100.

These were standardised for cell occupancy by taking residuals from D and D2

regressed against the logs of t1_10 and t2_10 respectively. Positive values were more

aggregated than expected for the relative occupancy, whilst negative values were

sparser. Temporal change in these residuals was calculated using the formula:

Residual D change= (residual D2- residual D)/log(t2_10/t1_10)


22

3.3. Trait data

3.3.1. Data sources

Trait data was obtained from several sources (listed in Appendix I). Abdominal

lengths were obtained from a meta-analysis of the field guide literature (Bowles,

2010). Measurements of wing length (using the computer package ImageJ; Rasband,

1997-2007) were obtained for Calopteryx splendens (n=129), Coenagrion puella (n=193),

Erythromma najas (n=123), and Pyrrhosoma nymphula (n=444) (Hassall, pers.comm.); for

complete collection methods refer to Hassall et al. (2008, 2009). Additional wing

lengths (using calipers) were obtained for C. puella (n=770) (Conrad, pers.comm) to

enable a comparison of measurement methods; see Conrad et al. (2002) for complete

methods.

3.3.2. Morphometrics

In addition to those collected from the literature, morphological measurements were

taken from dry specimens held in NHM collections. This enabled conclusions to be

drawn on the representativeness of museum specimens to extant populations.

The dry length and width (at the widest point) of each wing, and the three

dimensions of the thorax, were taken from specimens (for n=37 species: ♀=8, ♂=8 for

each species) using digital callipers (0.1 mm precision). Specimens which had

incomplete labels, damaged wings or thorax, or were recorded as “bred” were

omitted. One specimen for each species was re-measured to estimate measurement

error. Given wing symmetry and the significant correlation between fore- and

hindwings in NHM specimens (Appendix III), the left hindwing will be used across

this study as a proxy for all wings. Hindwings are preferable as their shape shows

greater intraspecific variation, and is thus more likely to inform species ecology.

3.3.3. Controlling for allometry

Direct body measurements are unlikely to be independent of overall body size. To

identify these allometric relationships, the three dispersal morphometrics were

regressed against a specific abdomen length measure: dry wing length against


23

abdomen length, aspect ratio against abdomen length2 and thoracic volume against

abdomen length3. Where strong correlations were identified, regression residuals

were used in the analyses.

3.3.4. Sample size

A power analysis was used to determine the minimum number of specimens which

needed to be measured in order to generate a representative sample size. The

analysis was based on four species obtained from previous research (see section 3.3.1

Hassall pers. comm.), and used the formula:

n= (z2 x σ2) / k2

where z is the error, taken at 1.96 at 95% confidence interval; σ is the standard

deviation; and k is the margin of error, defined as the difference between the actual

mean and the estimated mean.

k= µ x p

where µ is the mean of the sample data; and p is precision, the accepted variance of

the data around the mean (+/- 5% in this study) (Hayek & Buzas, 1997).

3.4. Data analysis

3.4.1. Controlling for phylogenetic signal

Many statistical analyses assume independence between data points, however this

assumption doesn’t hold true in across-species analyses because of varying degrees

of phylogenetic relatedness. Given that some species will be monophyletic to the

exclusion of others, this also means that shared trait properties may only exist

because of a close shared ancestry.

To ensure data point independence, this study controlled for phylogenetic signal

using the package “caper” in R 2.11 (Orme et al. 2011). Unless otherwise stated,


24

phylogenetic signal was controlled for in all analyses. A phylogenetic tree

(Appendix IV) was compiled for this analysis in FigTree v1.3.1 (Rambaut &

Drummond, 2009) from several sources (Artiss et al. 2001; Von Ellenrieder, 2002;

Saux et al. 2003; Fleck et al. 2008; Pilgrim & Von Dohlen, 2008; Dumont et al. 2010).

3.4.2. Variance component analysis (VCA)

Variance in residual dry wing length and aspect ratio were analysed across a

number of nested taxonomic scales using VCA. The levels used in increasing order

were: sex, species, genus, family and suborder. This was carried out in R 2.11 (R

Development Core Team, 2011) using the code:

NHM.varcomp<-varcomp(lme(value~1,random=~1|

suborder/family/genus/species/Sex),1)

where “value” is trait of interest; in this case, residual dry wing length and wing

aspect ratio. This required the installation of “nlme” and “ape” libraries (Paradis et

al. 2004; R Development Team, 2011). These analyses did not control for

phylogenetic signal.

3.5. Analysis of measurement methods

Comparisons were drawn between two different measurement methods for

Coenagrion puella right forewing length using a Welch two sample t-test. The two

data sets were collected using the computer software package ImageJ (Hassall et al.

2008) and calipers to 0.1 mm precision (Conrad et al. 2002).

A subset of NHM specimens (♂=3, ♀=3) was selected at random from the four

species noted in Section 3.1.1. Measurements were made to the first cross vein of the

left fore- and hindwings, allowing direct comparison with the wet specimen data

from Hassall et al. (2008, 2009). Differences in length between the two data sets were

investigated using Welch two sample t-tests, enabling conclusions to be drawn on

the representativeness of dried museum specimens.


25

4. Results

4.1. Range dynamics

This study found there to be an even split between the number of species showing a

decline in grid cell occupancy (n=19) and those showing an increase (n=18).

Following the removal of Aeshna isosceles and Coenagrion hastulatum which had

insufficient cell occupancies to comply with the minimum requirements of the range

shift metric (>20 grid cells: Hickling et al. 2006), there was also no significant

difference in the number of species moving north or south (Table 4.1). Even after

removing northern and widespread species which may be restricted by geographic

barriers, there was still not a significant proportion of species expanding their

ranges, or moving polewards at their northern extent (Table 4.1).

Table 4.1: The number of species which showed cell occupancy or range shift changes. The

binomial sign test shows whether a significantly greater proportion of species showed a range

expansion or northwards shift.

All Southern Northern Widespread

n species 37 22 4 11

Change

index

Increase 18 (49%) 14 (64%) 1 (25%) 3 (27%)

Decrease 19 (51%) 8 (36%) 3 (75%) 8 (73%)

Sign test: p-value 1 0.29 0.63 0.23

n species 35 21 3 11

Range

shift

Northwards 18 (52%) 9 (43%) 1 (33%) 8 (73%)

Southwards 15 (43%) 12 (57%) 2 (67%) 1 (9%)

Sign test: p-value 1 0.66 1 0.23

Although northwards and eastwards range shift did not strongly correlation (f=0.46,

df=33, p=0.64), a significant proportion of species showed corresponding outwards

(north and east shift) or inwards (south and west shift) movements at the north and

east range margins (binomial sign test: p=0.029). Furthermore, change index shows a


26

strong positive correlate with both northwards and eastwards range shift (f=7.90,

df=33, p=0.0016 and f=5.58, df=33, p=0.0082 respectively). This was sufficient to

question the degree to which the range “shift” metric was actually recording range

shift, and not just non-directional, marginal range expansion. In light of this doubt,

range shift was discounted from subsequent analyses and this study instead focused

on change index as the primary measurement of range dynamics.


Controlling for cell occupancy at t1, a significant proportion of species were found to

be sparser than expected (Table 4.2). When considered across time, however, species

were equally likely to become more aggregated as they were to become sparser.

Table 4.2: The number of species with more or less aggregated distribution (residual D), the

number of species changing over time (residual D change), and whether there was a

significant proportion with more aggregated or increasingly aggregated ranges (binomial

sign test).

All

species

(More)

Aggregated

(More)

sparse

Sign test:

p value

Residual D 37 12 (32%) 25 (68%) 0.047

Residual D change 37 18 (49%) 19 (51%) 1

It was hypothesised that species with sparser distributions (negative residual D)

would decline in cell occupancy over time (negative change index), but this was not

found to be the case (f=0.77, df=35, p=0.47). Neither was there a correlation between

residual D change and change index (f=0.11, df=35, p=0.90), however a significant

proportion of the species did show a positive relationship between the two (Figure

4.1; binomial sign test: p<0.001).


27

4.3. Morphometric analysis

4.3.1. Data independence

Linear regressions in the presence and absence of a phylogenetic signal control are

shown for the three dispersal traits and their respect abdomen length metrics in

Figure 4.2. The significant correlations for dry wing length and thoracic volume

against their respective abdomen metrics justify the use of residuals (f=95.74, df=35,

p<0.001 and f=41.46, df=35, p<0.001 respectively; Appendix V). Wing aspect ratio

showed no correlation with abdomen length2 once phylogenetic signal was

2 3 4 5 6 7 8

0.4

0.6

0.8

1.0

1.2

1.4

1.6

log(cell occupancy)

D

Figure 4.1: The relationship between the fractal dimension of species distribution (D) and

cell occupancy. The residual D change between t1 and t2 is shown for each species,

displayed with the linear regression model for mean species D against mean species cell

occupancy (fine black line). The dotted lines are species with a positive residual D change,

these species became more aggregated from t1 to t2. The solid lines are species with a

negative residual D change; these became sparser across time.


28

controlled for (f=0.018, df=35, p=0.99; Appendix V). This indicates that aspect ratio

does not show allometric scaling, and thus does not need to be transformed further.

4.3.2. Body morphometrics

Residuals of dry wing length and thoracic volume were shown to be significantly

correlated (f=137.30, df=35, p<0.001); species with relatively longer wings also had

relatively larger thoraces. Considered together, these two traits also showed a

(a) (b)

3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4

2.8

3.0

3.2

3.4

3.6

3.8

4.0

log(abdomen length (mm))

log(d

ry w

ing length

(m

m))

9 10 11 12 13

3

4

5

6

7

8

log(abdomen length^3 (mm3))

log(t

hora

cic

volu

me (

mm

3))

6.0 6.5 7.0 7.5 8.0 8.5

0.8

1.0

1.2

1.4

1.6

1.8

log(abdomen length^2 (mm2))

log(w

ing a

spect

ratio(m

m2))

Figure 4.2: Linear regression models for the relationships between (a) dry wing length and

the logs of abdominal length, (b) thoracic volume and the logs of abdominal length3, and

(c) wing aspect ratio and the logs of abdominal length2. Solid lines show the linear

regression fitted straight to the data, whilst dashed lines show the linear regression once

phylogenetic signal is controlled for.

(c)


29

significant negative relationship with wing aspect ratio (f=6.28, df=33, p<0.001).

Controlling for both phylogenetic and allometric signals, this study found species

with longer wings and larger thoraces to have significantly broader (lower aspect

ratio) wings.

4.3.3. Level of morphometric variance

Whilst this study incorporates a control for phylogenetic signal, it is important to

ensure that the assumption of negligible intra-specific trait variation holds true in

order to justify across-species comparisons. VCAs for residual dry wing length and

wing aspect ratio both show most variation at the level of the suborder (Table 4.3).

Table 4.3: The variance in dry wing length and aspect ratio accounted for by each of the

nested taxonomic scales (for n= 572 specimens).

% variance accounted for by taxonomic scale

Taxonomic level Residual dry wing length Wing aspect ratio

Suborder 36.88 65.60

Family 33.61 24.94

Genus 13.53 3.28

Species 6.42 1.10

Sex 4.53 1.60

Within 5.03 3.46

Overall, significantly more variation is accounted for at, or above, the level of the

species (see Table 4.4: sign test), with only 9.56% (residual dry wing length) and

5.06% (aspect ratio) accounted for at an intraspecific scale.


30

Table 4.4: The variance in dry wing length and aspect ratio accounted for by inter- (between

species and all higher taxonomic levels) and intraspecific (between and within sex) variation.

The binomial sign test shows whether a significantly greater proportion of variation was

accounted for interspecific variation.

% variance accounted for by taxonomic scale

Level of variation Residual dry wing length Wing aspect ratio

Interspecific 90.44 94.92

Intraspecific 9.56 5.06

Sign test: p-value <0.001 <0.001

Differences between suborders account for the majority of the variation seen in wing

aspect ratio (65.6%), with anisopteran wings significantly broader (lower aspect

ratio) than those in zygopterans (t test: t=-9.15, df=16, p<0.001). For example,

regressions of aspect ratio against residual dry wing length show very different

trends between the suborders (Figure 4.3: zygoptera: f=15.22, df=13, p<0.001;

anisoptera: f=1.45, df=20, p=0.26).

-0.2 0.0 0.2 0.4 0.6

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

residual dry wing length

log

(win

g a

sp

ect

ratio

)

Figure 4.3: Relationship between the log of wing aspect ratio and residuals of dry wing

length (controlling for phylogenetic signal) in anisopterans (filled points and solid

regression line) and zygopterans (empty points and dashed regression line).


31

4.3.4. Measurement errors

NHM measurements were accurate to an average of 0.12 mm. Measurements taken

from the thorax were less accurate than those from the wings; with mean accuracies

of 0.20 mm and 0.086 mm respectively. Thoracic depth showed the greatest

variation, with a mean difference of 0.28 mm across all species.

When incorporating data from numerous sources, the method of collection can also

have implications for the accuracy and comparability of the data. Within this study,

absolute wing length was shown to vary significantly between the two predominant

methods of measuring odonate wings (Figure 4.4: t test: t= 46.38, df= 355, p<0.001).

Figure 4.4: Difference between the two primary methods of measuring odonate wing

length. This is based on length measurements taken from the right forewing of Coenagrion

puella (see Conrad et al. 2002; Hassall et al. 2008 for complete methodologies).

Calipers ImageJ

16

18

20

22

24

26

measurement method

rig

ht

fore

win

g le

ng

th (

mm

)


32

4.4. Dispersal capacity

Dispersal traits and their interactions were not sufficient to explain a significant

amount of the variation seen in change index (f=0.93, df=10, p=0.53). The inclusion

of residual D, however, significantly increased the explanatory power of the model

(f=3.244, df=21, p=0.0064). Following model simplification, the minimum adequate

model included all dispersal traits and residual D, as well as the interactions

between residual dry wing length, residual thoracic volume and residual D (f=6.30,

df=28, p<0.001). This described 54% of the variation seen in change index.

This study found species with longer, broader wings, larger thoraces, and relatively

aggregated distributions to be showing the greatest range expansions. This

morphology is most commonly seen in anisopterans, which have significantly longer

wings (t test: t=-4.048, df=26, p<0.001), larger thorax volumes (t test: t=-2.16, df=21,

p=0.043), and broader wings (t test: t=-9.15, df=16, p<0.001) than the zygopterans.

The positive link between anisopteran-like morphology and range expansion is

further reflected in the substantially higher proportion of anisopterans expanding

their ranges (Table 4.5: 64% compared to 27% of anisopterans).

Table 4.5: The number of species with a positive change index (expansion in overall cell

occupancy from t1 to t2). The binomial proportion test shows whether there is a significant

difference in the proportion of each sub-order showing a cell occupancy increase. Percentages

are given as a proportion of n.

Species n species Expanding

All 37 18 (49%)

Anisoptera 22 14 (64%)

Zygoptera 15 4 (27%)

Prop.test: p-value 0.061


33

4.5. Representativeness of museum specimens

There was no significant difference in wing length between museum specimens (n=6

for the species in Section 3.1.1) and measurements from wet field specimens (Table

4.6; Appendix VI).

Table 4.6: Statistical outputs from Welch two sample t-tests investigating differences in

wing length between dry museum specimens (NHM) and wet specimens (recently collected).

Species

Forewing Hindwing

T df P t df p

Calopteryx splendens -0.94 5 0.39 -0.89 5 0.41

Coenagrion puella -1.93 5 0.11 -0.11 5 0.92

Erythromma najas 0.86 5 0.43 -1.47 5 0.2

Pyrrhosoma nymphula 0.041 5 0.97 1.25 5 0.26


34

5. Discussion

5.1. Spatial dynamics

Past work has shown most British odonate species to be responding positively to

climate change, exhibiting widespread range expansions and poleward shifts (e.g.

Hickling et al. 2005, 2006; Hassall et al. 2010). This study was unable to substantiate

either of these trends (Table 4.1). Furthermore, findings cast doubt on whether the

range shift metrics can be considered exclusively representative of shifting species

range, independent of general occupancy change.

Disparity between these and previous findings may arise from differences in the

time periods covered. Whilst this study encompasses all years from 1970-2008

(622,401 records), past studies such as Hickling et al. (2005, 2006) have used two

short time periods of equal length set several years apart. The use of two unequally

sized, adjacent time periods is not extensive in the literature, despite having some

advantages over the conventional method. A longer first time period can account for

lower rates of recording in earlier years (Hill et al. 2002; Wilson et al. 2004; Matilla et

al. 2008; Termaat et al. 2010), whilst consecutive time periods ensures that all data is

utilised (Lenoir et al. 2008; Termaat et al. 2010).

5.1.1. Range shift

Range shift was measured in accordance with methods from previous odonate range

studies (Hickling et al. 2005). Whilst working with extreme records of presence (ten

north-most occupied cells) risks generating trends that do not represent the range as

a whole, this method was less data “hungry” than viable alternatives. Hassall &

Thompson (2010) proposed basing shift on changes at the 95th quantile to account for

extreme sightings, but this requires species to occupy over 45 grid cells at each time

period. Under this proposed method, around a quarter of species (9/37) would have

been excluded, as opposed to the two which this study removed.


35

When studies focus on a subsection of the range margin, shift can only be inferred

from the latitudinal dynamics of this area. Most studies, however, fail to consider

corresponding longitudinal dynamics, which may lead to premature conclusions

that latitudinal movements reflect directional range shift. Whilst there was no

correlation in the size of northwards and eastwards range shift, a significant

proportion of species were found to move outwards (north and east) or inwards

(south and west) at both north and east range margins (p=0.029), and both directions

showed strong positive correlations with change index. Whilst it may be that species

are just exhibiting corresponding latitudinal and longitudinal shifts, the significant

number of species showing this trend suggests that the two are related, even though

the distances moved are not proportional. This lack of correlation between north and

east range shift could be explained by factors such as unequal dispersal barriers (e.g.

British coastline). Apparent non-directionality in range margin movement was

sufficient to cast doubt on the representativeness of “range shift” as an unambiguous

measure of shift in a species range. It was for this reason that the study focussed

predominantly on changes in range occupancy.

5.1.2. Change index

Changes in occupancy depend on a species-specific combination of factors, of which

climate change is just one. This makes any conclusions about the relative impacts of

different drivers, including climate change, difficult to interpret. For example,

population contractions or declines may result from habitat destruction or

fragmentation, as is likely to be the case for the declining Coenagrion mercuriale which

has seen significant habitat loss in recent years (JNCC, 2007). Meanwhile, the

migration of species into previously occupied habitats could result from improving

water quality (Gibbins & Moxton, 1998; Environment Agency, 2008), an increase in

habitat (e.g. garden ponds and gravel pits: Mill et al. 2010), or the adoption of aquatic

pesticide buffer zones around waterways. The introduction of targeted habitat

restoration schemes, such as that for Aeshna isosceles, also corresponds with an

increase in cell occupancy.


36

Whilst the drivers themselves may be unknown, what can be drawn from range

change metrics is an indication of how species are responding to the combined

pressures they are under. Although this limits potential for proactive conservation,

understanding how species are changing is still important to ongoing management.

Indeed, changes to a species’ overall occupancy is considered so indicative of species

threat and stability that declining AOO is one of the criteria used by the

International Union for Conservation of Nature (IUCN) in their species Red List

(Gardenfors et al. 1999; Daguet et al. 2008).


Significantly more species (68%) were sparser than expected for a given cell

occupancy. It can be speculated that species freshwater habitat preferences may

account for some of this, with substantially more species found in discrete lentic

(“still”) systems than lotic (“flowing”) systems (Aguilar et al. 1986; Brooks, 2005).

Rivers and streams (lotic) provide a network of permanent or semi-permanent

dispersal corridors linking neighbouring grid cells, whilst lentic waterways (e.g.

lakes and ponds) are often more geographically isolated and ephemeral (Angelibert

& Giani, 2003; Hof et al. 2006).

Lotic waterways function as the predominant dispersal pathway for lotic species

(Ward & Mill, 2007; Chaput-Bardy et al. 2008), mostly through imago dispersal, but

this is also likely to be enhanced by some nymphal drift (Beukema, 2002; Chaput-

Bardy et al. 2008). Lentic species, however, must rely wholly on terrestrial imago

dispersal, and are therefore more likely to establish local metapopulations around

key source sites. Whilst these lentic species are considered better dispersers on the

whole (Corbet, 1999; Clausnitzer et al. 2009), the isolation and unpredictability of

potentially suitable habitat may be such, that there is limited capacity for the

establishment of new sustainable populations far from a dispersal source. In support

of this, of the six obligatory lotic species recognised in this study (Calopteryx

splendens, Calopteryx virgo, Cordulegaster boltonii, Gomphus vulgatissimus, Platycnemis


37

pennipes and Coenagrion mercuriale) (Bowles, 2010), the only one to be more sparsely

distributed than expected was C. mercuriale. This is unsurprising given its rarity and

designation as a UK BAP (Biological Action Plan) species threatened by “isolation

and habitat scarcity” (JNCC, 2007; BDS, 2011). Whilst no conclusions can be drawn

without further investigation, this does suggest that when cell occupancy is

controlled for, lotic species are more likely to display aggregated distributions than

lentic species.

5.3. Dispersal morphometrics

With direct measures of dispersal distance lacking for most species, morphological

traits act as indirect indicators. Because past studies are often relied upon as source

of this data, it is vital that the exact methodologies and their implications are

understood. As shown here with wing length, different measurement methods can

result in significantly different values for seemingly identical traits. Wing lengths

obtained using the computer software ImageJ (Rasband, 1997-2007) were

significantly smaller than those obtained from hand measuring with calipers (Figure

4.4). Because ImageJ works with scanned images of the wing, the very base is

omitted from analysis and is therefore not comparable to other methods which

measure wing length to the base.

Residuals were taken for both dry wing length and thoracic volume to control for

body size, as both strongly correlated with metrics of abdomen length (body size

proxy). Whilst generating thoracic volume residuals from abdomen length3 may rely

on the use of an artificial metric, cubing the length values did ensure that allometric

scaling was corrected for in the right dimension. In general, measurements of

abdomen length are prone to inaccuracies, particularly in species with telescopic

abdomens like the odonates (Nijhout et al. 2006), however it was used in this study

because it was the most readily available proxy for body size (Serrano-Meneses et al.

2007). The value of these wing metrics in across-species comparisons was validated

using VCAs. In trait analyses, within-species variation is often assumed to be absent

but this is rarely the case (Ives et al. 2007). VCAs can serve an important function in


38

identifying the taxonomic or population level where most variation occurs. In this

study there was negligible intraspecific variation for both wing measurements, with

most variance found above the species level (Table 4.4). This increases the certainty

that it is interspecific differences, and not overlapping intraspecific variation, which

result in the trends observed in this study.

5.4. Determining occupancy change

5.4.1. Distribution pattern change

Although there was no significant correlation between the sizes of change index and

residual D change, a significant number of species (>86%) showed a positive

relationship between the two (Figure 4.1). This may arise from species declining due

to internal localised extinctions (Thomas, 2000; Tscharntke et al. 2002; e.g. Figure

4.5(b)), and increasing due to step-wise expansion from previously occupied cells

(e.g. Figure 4.5(a)). For example, improvements in river management and water

quality have seen populations of species such as Libellula fulva and Platycnemis

pennipes expanding their ranges along the lengths of riverine systems (Corbet &

Brooks, 2008).


39

(a)

(b)

Figure 4.5: Distribution of occupied grid cells in the British ranges of (a) Orthopterum

cancellatum (a species increasing in occupancy) and (b) Calopteryx virgo (a species

decreasing in occupancy),for t1 (left) and t2 (right).


40

Of the five species showing the reverse trend, becoming sparser with increasing cell

occupancy, the most notable was Aeshna isosceles. This rare species (Endangered on

the IUCN British Red List; Daguet et al. 2008) is highly sensitive to pollution, with a

British range confined to Norfolk and north-east Suffolk, where it breeds exclusively

in ditches supporting the aquatic plant Stratiotes aloides (water soldier) (Corbet &

Brooks, 2008). Because of the small number of occupied cells, single sightings far

beyond the species’ acknowledged range will disproportionately influence the

results. Three notably extreme records, each based on a single sighting, account for

over 11% of all occupied grid cells at t2 (Figure 4.6). Such unexpected sightings may

be explained by vagrant individuals transported by extreme weather, coupled with

an increase in recorder effort and species awareness since its appearance in the 1995

UK Biodiversity Strategy.

Figure 4.6: Distribution of occupied grid cells in the Britain range of Aeshna isosceles for

t1 (left) and t2 (right).


41

5.4.2. Distribution and dispersal

Neither distribution pattern (residual D), nor dispersal morphometrics (residual dry

wing length, wing aspect ratio, and residual thoracic volume) were sufficient in

describing change index when considered alone. The descriptive power of the

model, however, was significantly enhanced when the two were considered together

(accounting for 54% of the variation seen in change index). This reinforces the

importance of both dispersal capacity and underlying population distribution in

determining a species’ ability to colonise new areas.

Species with the greatest range expansion were dominated by those with long, broad

wings and large thoracic volumes, in keeping with previously published findings on

invertebrate dispersal (e.g. Malmqvist, 2000; Rundle et al. 2007). Long, broad wings

maximise surface area for efficient flight, whilst a large thoracic volume (indirect

measure of muscle mass; Schilder & Marden, 2004) is indicative of the greater power

needed to generate and sustain the wing beat frequency of these larger wings

The assumption that sparse distributions will always be declining towards extinction

(Wilson et al. 2004) does not hold true, because it fails to consider both variation in

the dispersal ability of individuals, and in the inherent sparsity of the habitat. As a

sole measure of species threat this metric makes the erroneous assumption that all

odonate species start with a relatively aggregated distribution. For example, lentic

habitats, as previously discussed in Section 4.2, are likely to be more distributed

across the landscape than lotic habitats.

When considered with weak dispersal capacity (small, narrow wings and small

thoracic volume) however, those species with sparser distributions were indeed

those with declining range occupancies. New areas are easier to colonise if there is a

concentrated source of potential colonists in close proximity, particularly in species

with the greatest capacity for dispersal. Conversely, the landscape is less permeable

to species with limited dispersal ranges and which are already sparsely distributed

(either through underlying habitat distribution, or through local extinctions which


42

have not been recolonised). Amongst these declining species are the four British

Coenagrion damselflies (C. hastulatum, C. mercuriale, C. puella and C. pulchellum). This

genus has been in decline here since the 1950’s, with two species becoming

nationally extinct between 1953-1983 (Freeland & Conrad, 2002). They have similar

morphologies, with small, delicate thoraces and relatively short, narrow wings; all

are sparsely distributed and all, with the exception of C. mercuriale, favour lentic

habitats. Their limited capacity for recolonisation, and the inherent isolation of their

primary habitats, makes this genus particularly vulnerable to habitat destruction and

fragmentation processes (Daguet et al. 2008; Lowe et al. 2008; BDS, 2011).

5.5. Suborder variation

In light of the morphological associations with occupancy change, the significant

difference in dispersal morphometrics between suborders could have crucial

implications for their respective abilities to respond to changing conditions. The

phenotype found in British anisopterans (significantly longer, broader wings and

larger thoracic volume) was that associated with range expansion, suggesting that

they are responding far better to changes in the past four decades than zygopterans.

This corresponds with extensive evidence that anisopterans perform longer

migrations (Ruppell, 1989), are better colonists (comprising over 75% of recent

British colonists; Appendix II), and exhibit faster flight across greater distances than

the more manoeuvrable but weaker zygopterans (Wakeling & Ellington, 1997).

5.6. Value of historic collections

In order for analyses to have any explanatory powers, a sufficient number of

specimens are needed to represent the population or species under study. When

species are rare, cryptic or difficult to find, collecting enough data points may be

difficult. Under these circumstances, natural history collections can provide an

invaluable source of data. These specimens will have been dried and are often

decades, if not centuries old, calling into question their value in representing extant

populations (Doucet & Hill, 2009). Although only preliminary, findings from this

study show the wing morphology of odonate NHM specimens to be


43

indistinguishable from those of recently caught, wet specimens. These results do, of

course, need to be treated with caution, and no claims can be made regarding their

representativeness of other morphological traits or taxonomic groups. Odonate

wings have little water content and so are unlikely to shrink or warp to any

significant degree when they dry. The same cannot be said for the body, particularly

the abdomen, where results are likely to be more variable.

5.7. Recommendations for future research

Whilst this study investigated general trends in odonate distribution, and looked at

the value of distribution pattern (residual D) and dispersal morphology in

describing these changes, no attempt was made to explore potential drivers of this

change. Future research needs to focus on pulling apart the relative impacts of

climate and habitat change in order to guide pre-emptive conservation action.

Climate change investigations need to move on from focussing on gradual

temperature change to encompass other variables. Maximum temperature

thresholds, the frequency of extreme weather (such as droughts and floods), and

changing humidity will all effect species distribution and survival potential.

Although this study does not dispute the validity of range shift metrics, it does

question the readiness to attribute all polewards range movement to climate change

induced range shift. There is the risk that polewards shift may be concluded

prematurely, particularly when there is no corresponding investigation into

longitudinal movements, or the possibility that trends may result from non-

directional occupancy change. Further investigation is needed in this area to ensure

that no rash conclusions are drawn which may overlook vital factors or driving

forces of change.

Despite being one of the best studied taxa in Britain, there remain many gaps in our

knowledge of simple odonate life history. Although it has fallen from conservation

favour in recent years, there is a pressing need for more targeted natural history

research, not only from data collected by amateur naturalists and enthusiasts, but


44

also from controlled scientific study and experimentation. For example, there is

limited information on many species’ dispersal capacity (this study is only able to

take morphology as a proxy). For a number of species, such information is only

speculative, or based on a small number of observations.

Further research is also needed into the value of museum specimens, an area only

briefly touched upon in the scope of this project. Whilst preliminary findings do

suggest that historic collections are representative of extant species morphometrics,

it is pertinent to note that this was only carried out on a subsample of specimens for

four species stored at NHM. It would be important to extend this to other species

and other collections, ensuring a spatially and temporally representative sample set.

It is known that some physical traits shrink when specimens are dried, however

there has been little research into different drying or storage methods, and whether

interspecific variation remains representative after drying.


45

6. Conclusions

Models are increasingly being used to predict large scale changes in species

distribution, particularly in light of climate change. Because there is insufficient data

on many taxa, the few that have been extensively studied, and which have long-

standing historic records, are being investigated for generally applicable trends.

Using a well studied group of British odonates, this study shows that the

combination of dispersal morphology (dry wing length and aspect ratio, and

thoracic volume) and distribution pattern can be used as an indicator of changes in

range occupancy over time.

Species showing the biggest range expansions were predominantly those with long,

broad wings and large thoracic volumes, particularly when showing aggregated

distributions. This was linked to suborder differences, with most anisopterans

(generally large winged and large bodied) exhibiting range expansions, whilst most

zygopterans (smaller, weaker dispersers) showed a decline. Given the importance of

obtaining reliable morphometrics, this study also investigated the value of museum

specimens as sources of morphometric data. Measurements obtained from NHM

were not found to differ from the variation seen in extant populations of the species,

drawing the, albeit preliminary, conclusion that these collections remain

representative of extant populations.

Whilst there were not a significant number of species with decreasing cell

occupancies, neither was there the overwhelming number of increasing species that

past studies have found (e.g. Hickling et al. 2005). This study therefore, urges caution

in prematurely writing off the British odonates as “safe” under predicted

environmental and climatic changes. It also highlights the risk of automatically

concluding a climate change induced range shift from the dynamics of a single range

margin.


46

References

Aguilar, J., Dommanget, J. & Prechac, R. (1986) A Field Guide to the Dragonflies of

Britain, Europe and North Africa. London, Collins.

Allen, K. & Thomas, D. (2010) Movement characteristics of the scarce blue- tailed

damselfly, Ischnura pumilio. Insect Conservation and Diversity, 3 (1), 5.

Altermatt, F. (2010) Climatic warming increases voltinism in European butterflies

and moths. Proceedings of the Royal Society B: Biological Sciences, 277 (1685), 1281-1287.

Angelibert, S. & Giani, N. (2003) Dispersal characteristics of three odonate species in

a patchy habitat. Ecography, 26 (1), 13.

Archaux, F. (2004) Breeding upwards when climate is becoming warmer: no bird

response in the French Alps. Ibis, 146, 138-144.

Artiss, T., Schultz, T., Polhemus, D. & Simon, C. (2001) Molecular phylogenetic

analysis of the dragonfly genera Libellula, Ladona, and Plathemis (Odonata:

Libellulidae) based on mitochondrial cytochrome oxidase I and 16S rRNA sequence

data. Molecular Phylogenetics and Evolution, 18 (3), 348-361.

Bale, J., Masters, G., Hodkinson, I., Awmack, C., Bezemer, T., Brown, V., Butterfield,

J., Buse, A., Coulson, J., Farrar, J., Good, J., Harrington, R., Hartley, S., Jones, T.,

Lindroth, R., Press, M., Symrnioudis, I., Watt, A. & Whittaker, J. (2002) Herbivory in

global climate change research: direct effects of rising temperature on insect

herbivores. Global Change Biology, 8 (1), 1.

BDS. (2010) Norfolk Biodiversity Action Plan: Norfolk Hawker. British Dragonfly Society.

Report number: Species Action Plan 27.

BDS. Southern Damselfly. [Online] Available from: http://www.british-

dragonflies.org.uk/species/southern-damselfly [Accessed 22/8/2011].


47

Begon, M., Townsend, C. & Harper, J. (2005) Ecology: From Individuals to Ecosystems.

4th edition. Oxford, Blackwell Science.

Bense, U. (1995) Longhorn beetles: illustrated key to the cerambycidae and vesperidae of

Europe. Germany, Margraf Verlag, Wiekersheim.

Beukema, J. (2002) Changing istribution patterns along a stream in adults of

Calopteryx haemorrhoidalis (Odonata: Calopterygidae): a case of larval-drift

complensation? International Journal of Odonatology., 5 (1), 1-14.

Bowles, P. (2010) Correlates of distribution and decline in UK species. Masters of Science.

Imperial College London.

Brittain, J. (2008) Mayflies, biodiversity and climate change. In: Hauer, F., Standford,

J. & Newell, R. (eds.) International Advances in the Ecology, Zoogeography, and

Systematics of Mayflies and Stoneflies. University of California Press, pp. 1-14.

Brooks, S. (ed.) (2005) Field Guide to the Dragonflies and Damselflies of Great Britain and

Ireland. Dorset., British Wildlife Publishing.

Chaput-Bardy, A., Lemaire, C., Picard, D. & Secondi, J. (2008) In-stream and

overland dispersal across a river network influences gene flow in a freshwater insect,

Calopteryx splendens. Molecular Ecology, 17, 3496-3505.

Clark, P. & Evans, F. (1954) Distance to nearest neighbor as a measure of spatial

relationships in populations. Ecology, 35 (4), 445-453.

Clausnitzer, V., Kalkman, V., Ram, M., Collen, B., Baillie, J., Bedjanic, M., Darwall,

W., Dijkstra, K., Dow, R., Hawking, J., Karube, H., Malikova, E., Paulson, D., Schutte,

K., Suhling, F., Villanueva, R., von Ellenrieder, N. & Wilson, K. (2009) Odonata enter

the biodiversity crisis debate: The first global assessment of an insect group.

Biological Conservation, 142, 1864-1869.

Conrad, K. (Personal communication, 2011).


48

Conrad, K., Wilson, K., Harvey, I., Thomas, C. & Sherratt, T. (1999) Dispersal

characteristics of seven odonate species in an agricultural landscape. Ecography, 22,

524-531.

Corbet, P. (1999) Dragonflies: Behaviour and Ecology of Odonata. NY, Cornell University

Press.

Corbet, P. & Brooks, S. (2008) Dragonflies. The New Naturalist Library, , Collins.

Corbet, P., Suhling, F. & Soendgerath, D. (2006) Voltinism of Odonata: a review.

International Journal of Odonatology., 9 (1), 1-44.

Cordero, A. (1994) Reproductive allocation in different-sized adults Ischnura

graellsii (Rambur) (Zygoptera: oenagrionidae). Odonatologica, 23, 271-276.

Daguet, C., French, G. & Taylor, P. (2008) The odonata red list data for Great Britain.

Species Status 11, Peterborough, Joint Nature Conservation Committee.

Davis, J., Howe, R. & Davis, G. (2000) A multi-scale spatial analysis method for point

data. Landscape Ecology, 15, 99-114.

DEFRA. (2011) Chemical river water quality. [Online] Available from:

http://www.defra.gov.uk/statistics/environment/inland-water/iwfg07-chrivqual/

[Accessed 26/4/2011].

Dixon, P. (2002) Ripley's K function. In: El-Shaarawi, A. & Piegorsch, W. (eds.)

Encyclopedia of Environmetrics. Volume 3 series. Chichester, UK, John Wiley & Sons

Ltd, pp. 1796-1803.

Doucet, S. & Hill, G. (2009) Do museum specimens accurately represent wild birds?

A case study of carotenoid, melanin, and structural colours in long-tailed manakins

Chiroxiphia linearis. Journal of Avian Biology, 40 (2), 146-156.


49

Dukes, J. & Mooney, H. (1999) Does global change increase the success of bioloical

invaders? Trends in Ecology and Evolution, 14 (4), 135-139.

Dumont, H., Vierstraete, A. & Vanfleteren, J. (2010) A molecular phylogeny of the

Odonata (Insecta). Systematic Entomology, 35, 6-18.

Elith, J. & Leathwick, J. (2009) Species distribution models: ecological explanation

and prediction across space and time. Annual Review of Ecology, Evolution, and

Systematics, 40, 677.

Environment Agency. (2008) Water Resources in England and Wales- Current State and

Future Pressures. Bristol, Environment Agency.

Ferreras-Romero, M. & Corbet, P. (1999) The life cycle of Cordulegaster boltonii

(Donovan, 1807) (Odonata: Cordulegastridae) in the Sierra Morena Mountains

(southern Spain). Hydrobiologia, 405, 39-48.

Fischer, J. & Lindenmayer, D. (2007) Landscape modification and habitat

fragmentation: a synthesis. Global Ecology and Biogeography, 16 (3), 265.

Fleck, G., Ullrich, B., Brenk, M., Wallnisch, C., Orland, M., Bleidissel, S. & Misof, B.

(2008) A phylogeny of anisopterous dragonflies (Insecta, Odonata) using mtRNA

genes and mixed nucleotide ⁄ doublet models. Journal of Zoological Systematics and

Evolutionary Research, 46 (4), 310-322.

Freeland, J. & Conrad, K. (2002) Genetic similarity within and among populations of

the Variable and Azure damselflies (Coenagrion pulchellum and C. puella).

Hydrobiologia, 479, 69-73.

Gardenfors, U., Rodriguez, J., Hilton-Taylor, C., Hyslop, C., Mace, G., Molur, S. &

Poss, S. (1999) Draft guidelines for the application of IUCN Red List Criteria at

national and regional levels. Species 31/32, 58-70.

Gaston, K. (1994) Rarity. Population and Community Biology Series. , Springer.


50

Gaston, K. (1996) Species-range-size distributions: patterns, mechanisms and

implications. Trends in Ecology and Evolution, 11 (5), 197-201.

Gaston, K. (2003) The Structure and Dynamics of Geographic Ranges. Oxford, Oxford

University Press.

Gaston, K. & Fuller, R. (2009) The sizes of species' geographic ranges. Journal of

Applied Ecology, 46, 1-9.

Gelfand, A., Silander, J., Wu, S., Latimer, A., Lewis, P., Rebelo, A. & Holder, M.

(2006) Explaining species distribution patterns through hierarchical modeling.

Bayesian Analysis, 1, 41-92.

Gibbins, C. & Moxton, J. (1998) Calopteryx splendens (Harris) at edge of range sites

in North East England. Journal of British Dragonfly Society, 14, 33-45.

Goffart, P. (2010) Southern dragonflies expanding in Wallonia (south Belgium): a

consequence of global warming? BioRisk, 5, 109-126.

Guisan, A. & Thuiller, W. (2005) Predicting species distribution: offering more than

simple habitat models. Ecology Letters, 8, 993-1009.

Guisan, A. & Zimmermann, N. (2000) Predictive habitat distribution models in

ecology. Ecological Modelling, 135, 147-186.

Hartley, S., Kunin, W., Lennon, J. & Pocock, M. (2004) Coherence and discontinuity

in the scaling of species' distribution patterns. Proceedings of the Royal Society B:

Biological Sciences, 271, 81-88.

Hassall, C. (Personal communication, 2011).

Hassall, C. & Thompson, D. (2010) Accounting for recorder effort in the detection of

range shifts from historical data. Methods in Ecology and Evolution, 1 (4), 343-350.


51

Hassall, C. & Thompson, D. (2008) The effects of environmental warming on

Odonata: a review. International Journal of Odonatology., 11 (2), 131-153.

Hassall, C., Thompson, D. & Harvey, I. (2008) Latitudinal variation in morphology in

two sympatric damselfly species

with contrasting range dynamics (Odonata: Coenagrionidae). European Journal of

Entomology, 105 (5), 939.

Hassall, C., Thompson, D. & Harvey, I. (2009) Variation in morphology between core

and marginal populations of three British damselflies. Aquatic Insects, 31 (3), 187-197.

Hassall, C. (2010) The impact of climate-induced distributional changes on the

validity of biological water quality metrics. Environmental Monitoring and Assessment,

160 (1), 451.

Hassall, C., Thompson, D., French, G. & Harvey, I. (2007) Historical changes in the

phenology of British Odonata are related to climate. Global Change Biology, 13 (5), 933.

Hayek, L. & Buzas, M. (1997) How many field samples? In: Surveying Natural

Populations. New York, Colombia University Press, pp. 66-83.

Hickling, R., Roy, D., Hill, J. & Thomas, C. (2005) A northward shift of range margins

in British Odonata. Global Change Biology, 11 (3), 502.

Hickling, R., Roy, D., Hill, J., Fox, R. & Thomas, C. (2006) The distributions of a wide

range of taxonomic groups are expanding polewards. Global Change Biology, 12 (3),

450-455.

Hill, J. (1999) Climate and habitat availability determine 20th century changes in a

butterfly's range margin. Proceedings of the Royal Society B: Biological Sciences, 266

(1425), 1197.


52

Hill, J., Thomas, C., Fox, R., Telfer, M., Willis, S., Asher, J. & Huntley, B. (2002)

Responses of butterflies to twentieth century climate warming: implications for

future ranges. Proceedings of the Royal Society B: Biological Sciences, 269, 2163-2171.

Hof, C., Brandle, M. & Brandl, R. (2006) Lentic odonates have larger and more

northern ranges than lotic species. Journal of Biogeography, 33 (1), 63-70.

Hofmann, T. & Mason, C. (2005) Habitat characteristics and the distribution of

Odonata in a lowland river catchment in eastern England. Hydrobiologia, 539, 137-

147.

Holt, R. (2003) On the evolutionary ecology of species' ranges. Evolutionary Ecology

Research, 5, 159-178.

Houghton, J., Ding, Y., Griggs, D., Noguer, M., van der Linden, P., Dai, X., Maskell,

K. & Johnson, C. (2002) Climate Change 2001: The Scientific Basis. Contribution of

working group I to the third assessment report of the Intergovernmental Panel on

Climate Change. Cambridge University Press., 57 (8), 267.

Hughes, C., Dytham, C. & Hill, J. (2007) Modelling and analysing evolution of

dispersal in populations at expanding range boundaries. Ecological Entomology, 32

(5), 437.

Hui, C. (2011) Forecasting population trend from the scaling pattern of occupancy.

Biological Modelling, 222, 442-446.

Huston, M. (2002) Introductory essay: critical issues for improving predictions. In:

Scott, J., Heglund, P., Morrison, M., Haufler, J., Raphael, M., Wall, W. & Samson, F.

(eds.) Predicting Species Occurrences: Issues of Accuracy and Scale. Covelo, California,

Island Press, pp. 7-21.

Ikeya, T., Galic, M., Belawat, P., Nairz, K. & Hafen, E. (2002) Nutrient-dependent

expression of insulin-like peptides from neuroendocrine cells in the CNS contributes

to growth regulation in Drosophila. Current Biology, 12 (15), 1293-1300.


53

Ives, A., Midford, P. & Garland, T. (2007) Within-species variation and measurement

error in phylogenetic comparative methods. Systematic Biology, 56 (2), 252-270.

Jenkins, D., Brescacin, C., Duxbury, C., Elliott, J., Evans, J., Grablow, K., Hillegass,

M., Lyon, B., Metzger, G., Olandese, M., Pepe, D., Silvers, G., Suresch, H.,

Thompson, T., Trexler, C., Williams, G., Williams, N. & Williams, S. (2007) Does size

matter for dispersal distance? Global Ecology and Biogeography, 16, 415-425.

Jetz, W., Sekerciogly, C. & Watson, J. (2008) Ecological correlates and conservation

implications of overestimating species geographic ranges. Conservation Biology, 22

(1), 110-119.

JNCC. (2007) Second Report by the United Kingdom under Article 17 on the

Implementation of the Directive from Janurary 2001 to December 2006. Peterborough,

JNCC. Report number: S1044- Coenagrion mercuriale- Southern damselfly.

Johansson, F. (2003) Latitudinal shifts in body size of Enallagma cyathigerum

(Odonata). Journal of Biogeography, 30, 29-34.

Korkeamaki, E. & Suhonen, J. (2002) Distribution and habitat specialization of

species affect local extinctions in dragonfly Odonata populations. Ecography, 25 (4),

459-465.

Kristensen, N. (ed.) (1999) Lepidoptera, moths and butterflies. Volume 1: Evolution,

systematics, and biogeography. Handbuch der Zoologie. Eine Naturgeschichte der Stämme

des Tierreiches / Handbook of Zoology. A Natural History of the phyla of the Animal

Kingdom, Berlin, Walter de Gruyte.

Legendre, P. (1993) Spatial autocorrelation: trouble or new paradigm? Ecology, 74 (6),

1659-1673.

Lenoir, J., Gegout, J., Marquet, P., Ruffray, P. & Brisse, H. (2008) A significant

upward shift in plant species optimum elevation during the 20th century. Science,

320, 1768-1771.


54

Levin, S. (1992) The problem of pattern and scale in ecology. Ecology, 73, 1943-1967.

Lowe, C., Harvey, I., Thompson, D. & Watts, P. (2008) Strong genetic divergence

indicates tha congeneric damselflies Coenagrion puella and C. pulchellum (Odonata:

Zygoptera: Coenagrionidae) do not hybridise. Hydrobiologia, 605, 55-63.

Malcolm, J., Liu, C., Neilson, R., Hansen, L. & Hannah, L. (2006) Global warming

and extinctions of endemic species from biodiversity hotspots. Conservation Biology,

20 (2), 538.

Malmqvist, B. (2000) How does wing length relate to distribution patterns of

stoneflies (Plecoptera) and mayflies (Ephemeroptera)? Biological Conservation, 93,

271-276.

Marchetti, K., Price, T. & Richman, A. (1995) Correlates of wing morphology with

foraging behaviour and migration distance in the genus Phylloscopus. Journal of

Avian Biology, 26, 177-181.

Marden, J. (1994) From damselflies to pterosaurs: how burst and sustainable flight

performance scale with size. American Journal of Physiology, 266, R1077-R1084.

Martinez, J., Martinez, J., Zuberogoitia, I., Zabala, J., Redpath, S. & Calvo, J. (2008)

The effects of intra- and interspecific interactions on the large-scale distribution of

cliff-nesting raptors. Ornis Fennica, 85, 13-21.

Mattila, N., Kaitala, V., Komonen, A., Paivinen, J. & Kotiaho, J. (2011) Ecological

correlates of distribution change and range shift in butterflies. Insect Conservation and

Diversity, (1), 8.

McCauley, J. (2007) The role of local and regional processes in structuring larval

dragonfly distributions across habitat gradients. Oikos, 116 (1), 121.

McCauley, S. (2006) The effects of dispersal and recruitment limitation on

community structure of odonates in artificial ponds. Ecography, 29 (4), 585.


55

McPeek, M. (1989) Differential dispersal tendencies among Enallagma damselflies

(Odonata) inhabiting different habitats. OIKOS, 56, 187-195.

Menendez, R. (2007) How are insects respondin to global warming? Tijdschrift Voor

Entomologie, 150, 355-365.

Meyer, C., Frund, J., Lizano, W. & Kalko, E. (2008) Ecological correlates of

vulnerability to fragmentation in Neotropical bats. Journal of Applied Ecology, 45 (1),

381.

Mill, P., Brooks, S. & Parr, A. (2010) Dragonflies (Odonata) in Britain and Ireland. In:

Maclean, N. (ed.) Silent Summer: The State of Wildlife in Britain and Ireland. Cambridge

University Press, pp. 471-494.

Morin, X., Viner, D. & Chulne, I. (2008) Tree species range shifts at a continental

scale: new predictive insights from a process- based model. Journal of Ecology, 96 (4),

784-794.

Nijhout, H., Davidowitz, G. & Roff, D. (2006) A quantitative analysis of the

mechanism that controls body size in Manduca sexta. Journal of Biology, 5, 16.

Norberg, U. (1995) How a long tail and changes in mass and wing shape affect the

cost for flight in animals. Functional Ecology, 9 (1), 48.

Opdam, P. & Wascher, D. (2004) Climate change meets habitat fragmentation:

linking landscape and biogeographical scale levels in research and conservation.

Biological Conservation, 117 (3), 285.

Orme, D., Freckleton, R., Thomas, G., Petzoldt, T., Fritz, S. & Isaac, N. (2011) Caper:

Comparative Analyses of Phylogenetics and Evolution in R. [R Package Verson 0.4/r73].

Ott, J. (2010) Dragonflies and climate change- recent trends in Germany and Europe.

BioRisk, 5, 253-286.


56

Par, A. (2010) Monitoring of Odonata in Britain and possible insights into climate

change. BioRisk, 5, 127-139.

Paradis, E., Claude, J. & Strimmer, K. (2004) APE: analyses of phylogenetics and

evolution in R language. Bioinformatics, 20, 289-290.

Parker, J. & Johnston, L. (2006) The proximate determinants of insect size. Journal of

Biology, 5 (5), 15.

Parmesan, C. (2006) Ecological and evolutionary responses to recent climate change.

Annual Review of Ecology, Evolution, and Systematics, 37, 637-669.

Parmesan, C. (1996) Climate and Species Range. Nature, 382 (6594), 765.

Parmesan, C. (1999) Poleward shifts in geographical ranges of butterfly species

associated with regional warming. Nature, 399 (6736), 579.

Parmesan, C. & Yohe, G. (2003) A globally coherent fingerprint of climate change

impacts across natural systems. Nature, 421 (6918), 37.

Paynter, Q. & Bellgard, S. (2011) Understanding dispersal rates of invading weed

biocontrol agents. Journal of Applied Ecology, 48 (2), 407-414.

Pilgrim, E. & Von Dohlen, C. (2008) Phylogeny of the Sympetrinae (Odonata:

Libellulidae): further evidence of the homoplasious nature of wing venation.

Systematic Entomology, 33, 159-174.

Poyry, J., Luoto, M., Heikkinen, R., Kuussaai, M. & Saarinen, K. (2009) Species traits

explain recent range shifts of Finnish butterflies. Global Change Biology, 15 (3), 732.

Pringle, J. (1949) The excitation and contraction of the flight muscles of insects.

Journal of Physiology, 108, 226-232.


57

Purse, B. & Thompson, D. (2003) Emergence of the damselfly Coenagrion mercuriale

and Ceriagrion tenellum (Odonata: Coenagrionidae), at their northern range

margins, in Britain. European Journal of Entomology, 100, 93-100.

Purvis, A. & Rambaut, A. (1995) Comparative analysis by independent contrasts

(CAIC): an Apple Macintosh application for analysing comparative data. CABIOS, 11

(3), 247-251.

R Development Core Team. (2011) R: A Language and Environment for Statistical

Computing. [2.13] Vienna, Austria, R Foundation for Statistical Computing.

Rambaut, A. & Drummond, A. (2009) FigTree. [v1.3.1] University of Edinburgh,

Edinburgh, Institute of Evolutionary Biology.

Rasband, W. (1997-2007) ImageJ. [Online] Available from: http://rsb.info.nih.gov/ij/

[Accessed 3/5/2011].

Rayner, J. (1988) Form and function in avian flight. Current Ornithology, 5, 1-66.

Ripley, B. (1977) Modelling spatial patterns. Journal of the Royal Statistical Society: B,

39, 172-212.

Robinet, C. & Roques, A. (2010) Direct impacts of recent climate warming on insect

populations. Integrative Zoology, 5, 132-142.

Rundle, S., Bilton, D., Abbot, J. & Foggo, A. (2007) Range size in North American

Enallagma damselflies correlates with wing size. Freshwater Biology, 52, 471-477.

Ruppell, G. (1989) Kinematic analysis of symmetrical flight manoeuvres of Odonata.

Journal of Experimental Biology., 144, 13-42.

Saino, N., Rubolini, D., Van Hardenberg, J., Ambrosini, R., Provenzale, A., Romano,

M. & Spina, F. (2010) Spring migration decisions in relation to weather are predicted


58

by wing morphology among trans- Mediterranean migratory birds. Functional

Ecology, 24 (3), 658.

Saux, C., Simon, C. & Spicer, G. (2003) Phylogeny of the dragonfly and damselfly

order Odonata as inferred by mitochondrial 12S ribosomal RNA sequences. Annals of

the Entomological Society of America, 96 (6), 693-699.

Schilder, R. & Marden, J. (2004) A hierarchical analysis of the scaling of force and

power production by dragonfly flight motors. Journal of Experimental Biology., 207,

767-776.

Schoener, T. & Janzer, D. (1968) Notes on environmental determinants of tropical

versus temperate insect size patterns. The American Naturalist, 102 (925), 207-224.

Segurado, P. & Araujo, M. (2004) An evaluation of methods for modelling species

distributions. Journal of Biogeography, 31, 1555-1568.

Serrano-Meneses, M., Azpilicueta-Amorin, M., Szekely, T. & Cordoba-Aguilar, A.

(2007) The development of sexual differences in body size in Odonata in relation to

mating systems. European Journal of Entomology, 104, 453-458.

Shoo, L., Williams, S. & Hero, J. (2006) Detecting climate change induced range

shifts: Where and how should we be looking? Australian Ecology, 31, 22-29.

Slove, J. & Janz, N. (2011) The Relationship between Diet Breadth and Geographic

Range Size in the Butterfly Subfamily Nymphalinae- Study of Global Scale. PLoS

ONE, 6 (1), e16057.

Smith, D. & Murphy, J. (1982) Spatial relationships of nesting golden eagles in

central Utah. Raptor Research, 16 (4), 127-132.

Stevens, V., Turlure, C. & Baguette, M. (2010) A meta-analysis of dispersal in

butterflies. Biological Reviews, 85 (3), 625.


59

Taylor, P. & Merriam, G. (1995) Wing morphology of a forest damselfly is related to

landscape structure. Oikos, 73 (1), 43.

Telfer, M. (2003) Change index: a measure of change in range size that is

independent of changes in survey effort. In: Proceedings 13th International Colloquium

European Invertebrate Survey, Leiden 2-5 September 2001. pp. 107-110.

Telfer, M., Preston, C. & Rothery, P. (2002) A general method for measuring relative

change in range size from biological atlas data. Biological Conservation, 107 (1), 99.

Termaat, T., Kalkman, V. & Bouwman, J. (2010) Changes in the range of dragonflies

in the Netherlands and the possible role of temperature change. BioRisk, 5, 155-173.

Thomas, C. (2000) Dispersal and extinction in fragmented landscapes. Proceedings of

the Royal Society B: Biological Sciences, 267, 139-145.

Thomas, C., Bodsworth, E., Wilson, R., Summons, A., Davies, Z., Musche, M. &

Conradt, L. (2001) Ecological and evolutionary processes at expanding range

margins. Nature, 411, 577-581.

Thomas, C. & Lennon, J. (1999) Birds extend their ranges northwards. Nature, 399,

213.

Thomas, D., Cameron, A., Green, R., Bakkenes, M., Beaumont, L., Collingham, Y.,

Erasmus, B., de Siqueira, M., Grainger, A., Hannah, L., Hughes, L., Huntley, B., van

Jaarsveld, A., Midgley, G., Miles, L., Ortega-Huerta, M., Peterson, A., Philips, O. &

Williams, S. (2004) Extinction risk from climate change. Nature, 427 (6970), 145.

Thomas, C. (2010) Climate, climate change and range boundaries. Diversity

Distributions, 16 (3), 488.

Thomas, C., Thomas, J. & Warren, M. (1992) Distributions of occupied and vacant

butterfly habitats in fragmented landscapes. Oecologia, 92 (4), 563.


60

Tobin, P., Nagarkatti, S., Loeb, G. & Saunder, M. (2008) Historical and projected

interactions between climate change and insect voltinism in a multivoltine species.

Global Change Biology, 14, 951-957.

Tscharntke, T., Steffan-Dewenter, I., Kruess, A. & Thies, C. (2002) Characteristics of

insect populations on habitat fragments: a mini review. Ecological Research, 17, 229-

239.

Von Ellenrieder, N. (2002) A phylogenetic analysis of the extant Aeshnidae

(Odonata: Anisoptera). Systematic Entomology, 27, 437-467.

Wakeling, J. & Ellington, C. (1997) Dragonfly flight: II. Velocities, accelerations and

kinematics of flapping flight. Journal of Experimental Biology, 200, 557-582.

Ward, L. & Mill, P. (2007) Long range movements by individuals as a vehicle for

range expansion in Calopteryx splendens (Odonata: Zygoptera). European Journal of

Entomology, 104, 195-198.

Warren, M., Hill, J., Thomas, J., Asher, J., Fox, R., Huntley, B., Roy, D., Telfer, M.,

Jeffcoate, S., Harding, P., Jeffcoate, G., Willis, S., Greatorex-Davies, J., Moss, D. &

Thomas, D. (2001) Rapid responses of British butterflies to opposing forces of climate

and habitat change. Nature, 414 (6859), 65.

Wiegand, T. & Moloney, K. (2004) Rings, circles, and null-models for point pattern

analysis in ecology. Oikos, 104, 209-229.

Wikelski, M., Moskowitz, D., Adelman, J., Cochran, J., Wilcove, D. & May, M. (2006)

Simple rules guide dragonfly migration. Biological Letters, 2 (3), 329-325.

Wilcove, D., McLellan, C. & Dobson, A. (1986) Habitat fragmentation in the

temperature zone. In: Soule, M. (ed.) Conservation Biology. MA, Sunderland, pp. 237-

256.


61

Williams, I., Jones, H. & Hartley, S. (2001) The role of resources and natural enemies

in determining the distribution of an insect herbivore population. Ecological

Entemology, 26, 204-211.

Willis, S., Hill, J., Thomas, C., Roy, D., Fox, R., Blakeley, D. & Huntley, B. (2009)

Assisted colonization in a changing climate: a test-study using two U.K. butterflies.

Conservation Letters, 2 (1), 46-52.

Wilson, R., Thomas, C., Fox, R., Roy, D. & Kunin, W. (2004) Spatial patterns in

species distributions reveal biodiversity change. Nature, 432 (7015), 393.


62

Appendices

Appendix I: Variables used in the analysis of species range dynamics.

Name Definition Source #of

species

Range variables

Cell occupancy The number of grid cells a species was

recorded in during the time periods t1

(1970-1996) or t2 (1997-2008).

Source data

from NBN

37

Change index

Metric for the relative change in the cell

occupancy between t1 and t2 (see

Methods; Telfer et al. 2002).

Source data

from NBN

37

D/ D2 Fractal dimension of species

distribution, calculated from 10 km and

100 km grid cell occupancy at t1 (D) and

t2 (D2) (see Methods; Wilson et al. 2004).

Source data

from NBN

37

Distribution

pattern

The level of population aggregation

within a species’ range, measured using

residual D.

37

Occupancy

change

Change in cell occupancy from t1 to t2,

measured using change index.

37

Range shift Directional movement at the northern or

eastern margins of a species’ British

range, based on the 10 north- or east-

most occupied grid cells (see Methods;

Hickling et al. 2006).

Source data

from NBN

37


63

Appendix I cont: Variables used in the analysis of species range dynamics.


species

Range variables

Residual D/

Residual D2

The residuals of D and D2 regressed

against cell occupancy at t1 and t2

respectively (10 km cells) (see Methods;

Wilson et al. 2004).

Source data

from NBN

37

Residual D

change

Change between residual D and

residual D2 (see Methods).

Source data

from NBN

37

Trait variables

Abdominal

length

Midpoint between the minimum and

maximum abdominal lengths, with

measurements taken from base to tip.

Bowles,

2010

37

Dry wing

length

Mean length of the left hindwing from

the tip to the base of the wing plug.

Measured using 0.1 mm digital calipers.

NHM 37

Residual

thoracic

volume

Residuals of thoracic volume (length x

width x depth) regressed against the

abdominal length3.

Source data

from NHM

37

Residual dry

wing length

Residuals of dry wing length regressed

against abdominal length.

Source data

from NHM

37

Wing aspect

ratio

Shape of the left hindwing, expressed as

the ratio between dry wing length and

width (taken at the widest point of the

wing).

Source data

from NHM

37


64

Appendix I cont: Variables used in the analysis of species range dynamics.


species

Trait variables

Wing length

(ImageJ)

Length of the left hindwing (unless

otherwise specified) from tip to the first

cross vein from the base. Measured

using the computer software package

ImageJ (Rasband, 1997-2007).

Hassall et al.

2008,2009

4

Wing length

(Calipers)

Length of the forewing from tip to the

base. Measured using 0.1 mm digital

calipers

Conrad et

al. 2002

1


65

Appendix II: List of all species commonly found in Britain (from Corbet & Brooks, 2008)

and freshwater habitat preferences.

Species name Described

Residents

Aeshna caerulea Azure hawker Strom, 1783

Aeshna cyanea Southern hawker Muller, 1764

Aeshna grandis Brown hawker Linnaeus 1758

Aeshna isosceles Norfolk hawker Muller, 1767

Aeshna juncea Common hawker Linnaeus, 1758

Aeshna mixta Migrant hawker Latreille, 1805

Anax imperator Emperor dragonfly Leach, 1815

Brachytron pratense Hairy hawker Muller, 1764

Calopteryx splendens Banded demoiselle Harris, 1782

Calopteryx virgo Beautiful demoiselle Linnaeus, 1758

Ceriagrion tenellum Small red damselfly De Villers, 1789

Coenagrion hastulatum Northern dragonfly Charpentier, 1825

Coenagrion mercuriale Southern damselfly Charpentier, 1840

Coenagrion puella Azure damselfly Linnaeus, 1758

Coenagrion pulchellum Variable damselfly Vander Linden, 1825

Cordulegaster boltonii Golden-ringed dragonfly Donovan, 1807

Cordulia aenea Downy emerald Linnaeus, 1758

Enallagma cyathigerum Common blue damselfly Charpentier, 1840

Erythromma najas Red-eyed damselfly Hansemann, 1823

Erythromma viridulum* Small red-eyed damselfly Charpentier, 1840

Gomphus vulgatissimus Common club-tail Linnaeus, 1758

Ischnura elegans Blue tailed damselfly Vander Linden, ,1823

Ishnura pumilio Scarce blue-tailed

damselfly

Charpentier, 1825

Lestes dryas Scarce emerald damselfly Kirby, 1890

Lestes sponsa Emerald damselfly Hansemann, 1823

Leucorrhinia dubia White-faced darter Vander Linden, 1825


66

Appendix II cont: List of all species commonly found in Britain (from Corbet &

Brooks, 2008) and freshwater habitat preferences.

Species name Described

Residents

Libellula depressa Broad-bodied chaser Linnaeus, 1758

Libellula fulva Scarce chaser Muller, 1764

Libellula quadrimaculata Four-spotted chaser Linnaeus, 1758

Orthetrum cancellatum Black tailed skimmer Linnaeus, 1758

Orthetrum coerulescens Keeled skimmer Fabricius, 1798

Platycnemis pennipes White- legged damselfly Pallas, 1771

Pyrrhosoma nymphula Large red damselfly Sulzer, 1776

Somatochlora arctica Northern emerald Zetterstedt, 1840

Somatochlora metallica Brilliant emerald Vander Linden, 1825

Sympetrum danae Black darter Sulzer, 1776

Sympatrum sanguineum Ruddy darter Muller, 1764

Sympetrum striolatum Common darter Charpentier, 1840

Migrants

Anax junius Green darner Drury, 1770

Anax parthenope Lesser emperor Selys, 1839

Anax ephippiger Vagrant emperor Burmeister, 1839

Coenagrion lunulatum Irish damselfly Charpentier, 1840

Crocothemis erythraea Scarlet darter Brulle, 1832

Lestes barbarous Southern emerald Fabricius, 1798

Lestes viridis Willow emerald damselfly Vander Linden, 1825

Pantala flavescens Wandering glider Fabricus, 1798

Sympatrum flaveolum Yellow-winged darter Linnaeus, 1758

Sympetrum fonscolombii Red-veined darter Selys, 1840

Sympetrum meridionale Southern darter Selys, 1841

Sympetrum vulgatum Vagrant darter Linnaeus, 1758

*although this species is now a confirmed British resident, it is too recent a colonist to be included in

these analyses.


67

Appendix III: Linear regression models showing highly significant positive correlations

between mean wing lengths and aspect ratios (length/width) for each of the four wings. Data

was taken from specimens held in NHM collections.

f d.f p

Wing

length

left vs right hindwing 4.364 35 <0.001

left vs right forewing 9.433 35 <0.001

left forewing vs hindwing 1.104 35 <0.001

right forewing vs hindwing 8.743 35 <0.001

Wing

aspect

ratio

left vs right hindwing 6.0573 35 <0.001

left vs right forewing 57 35 <0.001

left forewing vs hindwing 1.0092 35 <0.001

right forewing vs hindwing 41.8 35 <0.001


68

Appendix IV: Phylogenetic tree for the resident British odonates (n=37 species), synthesised

from published morphological (Von Ellenrieder, 2002) and molecular (Artiss et al. 2001;

Saux et al. 2003; Fleck et al. 2008; Pilgrim & Von Dohlen, 2008; Dumont et al. 2010)

phylogenetic analyses.


69

Appendix V: Linear models for dry wing length, thoracic volume and wing aspect ratio

regressed against abdomen lengths. The abdomen length metrics used were abdomen length

(mm), abdomen length3 (mm3) and abdomen length2 (mm2) respectively. Model outputs are

given for models with and without phylogenetic signal controlled for.

Appendix VI: Differences in left wing lengths between dry museum specimens (collected in

the time period 1905-1953) and “wet” (recently caught) specimens (collected in 1999; see

Hassall et al. (2008,2009)).

Body metric

With phylogenetic

control

Without phylogenetic

control

F Df p F df P

Dry wing length 95.74 35 <0.001 107.40 35 <0.001

Thoracic volume 41.46 35 <0.001 81.14 35 <0.001

Wing aspect ratio 0.0044 35 0.99 6.19 35 0.018

20

22

24

26

28

30

32

34

left

fore

win

g length

(mm

)

15

16

17

18

19

20

17

18

19

20

21

22

23

24

15

20

25

30

museum wet

18

20

22

24

26

28

30

32

Calopteryx splendens

left

hin

dw

ing length

(mm

)

museum wet

14

15

16

17

18

19

Coenagrion puella

museum wet

15

16

17

18

19

20

21

22

Erythromma najas

museum wet

14

16

18

20

22

24

26

28

Pyrrhosoma nymphula

The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 ›...

Documents

Transcript of The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 ›...