Content uploaded by Michael L May
Author content
All content in this area was uploaded by Michael L May on Sep 01, 2022
Content may be subject to copyright.
63
distant regions, as in many birds and large mam-
mals. Alternative de nitions tend to focus on either
ecological or behavioural criteria. The former
emphasize the consequences of migration: move-
ment into spatially distinct habitats or communi-
ties, frequently associated with different phases of
the life cycle of the migrant (Hack and Rubenstein
2001). Corbet (1999), for example, describes migra-
tion as ‘spatial displacement that entails . . . leaving
CHAPTER 6
Migration in Odonata: a case
study of
Anax junius
Michael L. May and John H. Matthews
Overview
Although migration by Odonata has been recognized for well over 100 years, the phenomenon is still poorly
understood. We argue that it may provide substantial new insight into the patterns, mechanisms, and evo-
lution of insect migration in general, for at least two reasons. First, as aquatic/aerial carnivores dragon ies
can broaden our view of migratory insects, most of which are terrestrial herbivores, and of the selective
pressures to which they respond as well as their consequent genetic, physiological, and behavioural adapta-
tions. We expect this to help differentiate common characteristics of migrant animals in general from those
particular to certain groups. Second, because they are large, diurnal insects, they lend themselves to some
techniques of direct observation that are hard to achieve in most other insects. Our focus here is on the
best-studied North American migrant, Anax junius, the common green darner. We rst discuss the behav-
ioural and ecological attributes of migration and provide a brief descriptive overview of evidence for its
occurrence in Odonata. We then describe recent research on migration in A. junius. Large-scale patterns of
movement and the in uence of weather are brie y reviewed. Geographic analysis of genetic structure and
stable and radiogenic isotope composition and use of newly developed radio-tracking techniques has shed
new light on the nature of migration in this species. Developmental phenology indicates the existence of
early (resident) and late (migrant) cohorts at most sites, but genetic analysis does not indicate genetic differ-
entiation of these groups. Apparently environmental cues and physiological responses to photoperiod and
temperature engender migratory behaviour. Successful radio-tracking of individual A. junius has revealed
alternating periods of migration and energy replenishment and responses to wind and temperature similar
to avian migration. Little is known of orientation mechanisms during migration, and this should be a fruit-
ful area of future research. Also, additional observations of reproductive behaviour en route and estimates
of relative reproductive success of migrants and non-migrants should provide more detailed information on
selective advantages and disadvantages and the historical evolution of migratory behaviour.
6.1 Introduction
6.1.1 What is migration?
6.1.1.1 Behavioural and ecological definitions
and attributes
Formulating a meaningful, operational, and com-
prehensive de nition of migration has historically
been dif cult. Few, if any, insects undergo a round-
trip, seasonal passage to and from geographically
Book 1.indb 63Book 1.indb 63 6/27/2008 8:40:12 PM6/27/2008 8:40:12 PM
64 STUDIES IN ECOLOGY
a well-known European dragon y migrant, Libellula
quadrimaculata. They documented large migrations
at long intervals (approximately 10 years), typically
following very large mass emergences, probably
synchronized by delays due to cold spring weather.
They hypothesized that large migratory swarms
may result from non-adaptive movements set off
by optical interaction-synchronization (i.e. indi-
viduals that see others in ight are likely to start
ying themselves) potentiated by internal irrita-
tion due to high trematode parasite loads. Corbet
(1999) supposed that occupation by Anax junius of
northern areas where larvae cannot overwinter may
have originated by swarms of tropical origin y-
ing on prevailing winds toward areas of abundant
rainfall (and hence favourable breeding areas) but
overshooting their intended destination. Some must
have reproduced in northern ponds, but perhaps
less successfully than further south. If such occur-
rences were frequent and lead to substantial tness
reduction, however, strong selection would ensue
for either avoidance of movement into temperate
areas or adaptation to the northern environment.
To the extent that it is adaptive, migration must
allow either exploitation of an ephemeral resource
a nd /or av oi da n ce of p er io d s o f ad ve rs e en vi r on m en -
tal conditions. This is the case for seasonally migra-
tory monarch butter ies (Danaus plexippus), wh ich
breed in the northern USA and southern Canada on
abundant milkweed during summer, then migrate
to speci c refuges in Mexico and California where
conditions are suitable for adult diapause. Many
tropical insects, including dragon y species such as
Hemianax (=Anax) ephippiger, in response to crowd-
ing or drought, y or are transported downwind
toward the intertropical convergence zone (ITCZ),
where reliable rains renew vegetation and aquatic
habitats. In both cases, insects are able to occupy
resource-rich habitats for a portion of their life cycle,
or sometimes for several generations, then move to
more suitable regions when the original situations
deteriorate (Dingle 1996; Corbet 1999).
6.1.2 Migration in dragonfl ies
6.1.2.1 Historical observations
Mass ights of dragon ies have probably attracted
attention for millennia, but the rst written record,
the habitat where emergence took place and moving
to a new habitat where reproduction ensues’. Such
de nitions are heuristic in focusing on adaptive
aspects of migration in the context of life history.
Migration normally functions to move individuals
and populations from an initially suitable habitat
that deteriorates with time to an alternative and
currently more favourable habitat. Indeed, migra-
tion through space may be likened to behaviours
that effectively show migration through time, such
as diapause or the development of drought- resistant
life stages in odonates.
However, for many insects, the origin, ultimate
destination, and tness consequences of migra-
tion are not known in detail, so ecological de ni-
tions are often hard to apply rigorously in practice.
Behavioural criteria, therefore, may be more suit-
able. Dingle (1996, 2006), following Kennedy (1985),
suggests: ‘Migratory behaviour is persistent and
straightened-out movement effected by the ani-
mal’s own locomotory exertions or by its active
embarkation upon a vehicle. It depends on some
temporary inhibition of station keeping responses
but promotes their eventual disinhibition and
recurrence’. In many instances these criteria are
most easily recognized in organisms moving en
masse, and observations of mass movements have
been important in understanding migration in
dragon ies (Russell et al. 1998; Corbet 1999) as well
as other insects, but note that neither de nition
requires that individuals migrate in groups. Mass
movement may simply be our own visual clue that
migration is occurring in a given species.
6.1.1.2 Ecological, genetic, and evolutionary
consequences
We tend to assume that spectacular events like mass
migration must re ect adaptations for movement
from the migrants’ place of origin to their eventual
destination, but this might not always be the case.
Rabb and Stinner (1979) suggested that large-scale
movements of some important crop pests represent
accidental wind-borne transport followed by local
increase on concentrated resources, from which,
however, the migrants have little chance of return-
ing before succumbing to the hazards of winter or of
migration itself. A different non-adaptive scenario
was suggested by Dumont and Hinnekint (1973) for
Book 1.indb 64Book 1.indb 64 6/27/2008 8:40:12 PM6/27/2008 8:40:12 PM
MIGRATION IN ODONATA 65
seasonal movements. Others were thought to be
irruptively sporadic migrants or based on doubtful
reports. Corbet (1999) listed about 40 Anisoptera and
10 Zygoptera (all but one of the latter Ischnura spp.)
worldwide as well-documented migrants. The best-
represented anisopteran genera are Anax (seven
species listed by Corbet) and Tram e a (nine species),
but several others include frequent migrants, for
example, Sympetrum ( ve species) and Diplacodes
(four species, including “Philonomon” luminans).
Except for Hemicordulia, the known species are
either Aeshnidae or Libellulidae, most of which
breed in lentic waters. Not surprisingly, migratory
species characteristically inhabit ephemeral to semi-
permanent ponds that dry up frequently or unpre-
dictably and generally support few or no sh.
These surely are not exhaustive lists, and the
number of known migrants is likely to increase
substantially with more focused study. Moreover,
‘migration’ in general tends to be con ated with
long-distance migration over large spatial scales,
and not all implicated species may engage in
migration over tens of kilometres. Given that cav-
eat, migration appears to be an exceptional life-
history strategy among Odonata in general. On
the other hand, given the evidence (see below) that
migration is a facultative response in A. junius, we
cannot discount the possibility that a small minor-
ity of individuals of mostly non-migratory species
may migrate. Such varying behaviour could evolve
by small adjustments to environmental cues and
might account for occasional reports of species
such as Pachydiplax longipennis amon g mi xed agg re-
gations of migrants (Russell et al. 1998).
6.2 The evidence
6.2.1 Movement patterns: visual observations
The Atlantic coast of North America from Maine at
least to New Jersey, and probably the entire Atlantic
seaboard, is a major migratory route (Shannon 1916).
Other major hypothesized pathways run along the
north shores of Lakes Ontario and Erie and thence
into Ohio, along Lake Michigan, and into central
Illinois, and on a broad front from Minnesota into
eastern Oklahoma. Large swarms have been noted
as well along the coast of the Gulf of Mexico from
to our knowledge, was by Hermann Hagen (1861).
Observations of dragon y migrations in North
America date back at least to the late nineteenth
century (Calvert 1893), and movements along the
east coast and parts of the midwest were recorded
and mapped by Shannon in 1916. Although large
swarms naturally attract attention and are the sub-
ject of the vast majority of anecdotal reports (e.g.
Osburn 1916; Borror 1953; Cook 1991; Daigle 1991;
Glotzhober 1991), migrants may often occur as scat-
tered individuals or small groups. Shannon (1916),
Bagg (1958), Dumont (1977), and Sprandel (2001)
reported movements of this sort.
Beyond intermittent observations of dragon ies
on the move, the role of migration in their life his-
tory soon excited interest. As early as 1929 Calvert
raised the possibility of two emergence groups in
populations of A. junius. Robert Trottier (1966, 1971),
working on A. junius in southern Canada, found
that near Montreal larvae probably did not over-
winter at all, although they were regularly found
during summer, whereas in southern Ontario both
early-emerging and late-emerging cohorts existed.
The former emerged from late June to mid-July
and nished oviposition by mid-August. The lat-
ter, evidently offspring of mature adults that
appeared in April or May before any evidence of
emergence, developed rapidly during the summer
and emerged in late August to September. Adults
then disappeared, presumably having own
southward. These observations clearly imply that
migrations are a normal part of the life cycle that
permits colonization of northern areas. This idea is
supported by reports of apparently annual move-
ments described along the eastern seaboard by
Shannon and on the northern shores of Lake Erie
by Nisbet (1960), Walker (1958), and Corbet (1984).
Trottier’s data also raised the possibility that the
separate emergence cohorts could be genetically
distinct since the adult emergence seasons did not
overlap. A similar pattern is observed on the west
coast of North America (D.R. Paulson, personal
communication).
6.1.2.2 Which dragonflies migrate?
Russell et al. (1998) listed 18 probable North
American migrant Anisoptera, of which nine were
regarded as ‘core’ species that engage in annual
Book 1.indb 65Book 1.indb 65 6/27/2008 8:40:12 PM6/27/2008 8:40:12 PM
66 STUDIES IN ECOLOGY
and Nisbet (1960) con rmed apparent correlations
of Anax migratory ight in New England and along
Lake Erie, respectively, with the passage, in early fall,
of cold fronts that brought winds that could assist
migration. Numerous dragon ies, some known to
be migratory, arrived with a fall weather front on
the coast of Nayarit, in western Mexico (Paulson
2002). Russell et al. (1998) documented additional
instances of the association of mass movements
southward with the passage of cold fronts and of
the arrival of spring migrants with southerly ows
of warm air, but they noted that in some instances
the relation of ights to frontal activity is not clear.
One possible explanation is that migrating Anax
may move beyond frontal systems that initiate
aggregated ight, especially as cold fronts slow or
stall in southern latitudes. Nevertheless, there has
as yet been no strict quanti cation of a correlation of
migration with weather fronts. Only very recently
did Wikelski et al. (2006) quantitatively document
that individual southbound migrant A. junius do y
with light northerly winds (see below).
6.2.3 Reproduction and refueling
The physical and physiological condition of
migrants may be indicative of the adaptive func-
tion of migration. Many insect migrants are pre-
reproductive, with females often pre-vitellogenic;
this presumably assures that when they reach
a favourable destination they retain maximum
reproductive capacity (Johnson 1969; Dingle 1996).
Migrating Odonata, too, are often described as ten-
eral or ‘fresh’, but numerous exceptions are known
(Corbet 1999). Many but not all of these occur dur-
ing what Corbet has distinguished from migra-
tion as ‘seasonal refuge ights’; these would be
considered migratory under our behavioural def-
inition. Corbet (1984) found that the great major-
ity of apparent migrant species in Uganda and
in Ontario, except Sympetrum vicinum, were pre-
reproductive and laden with fat. Odonates that
accumulate fat stores, as indicated by preserved
specimens that are detectably greasy, are typically
those of migratory genera such as Anax, Pantala,
and Tram ea (Paulson 1998).
Among A. junius, and perhaps other species,
along the Atlantic seaboard of North America,
Florida to Texas, along with a few observations of
movement over the Gulf far from land. Numerous
records also exist of migrating dragon ies, princi-
pally A. junius, at hawk watches along mountain
ridges in the eastern USA (K. Soltesz, personal com-
munication; Matthews 2006).
A consequence of migratory behaviour, however,
may be that many individuals, responding to the
same cues for initiation, orientation, and arrest-
ment of ight may aggregate at ‘leading lines’ such
as lake- or seashores (Russell et al. 1998). Corbet
(1999) pointed out that aggregation may, in fact, be
adaptive when reproduction occurs en route or at
the destination, but it is also possible that aggrega-
tion of A. junius is entirely or partly a consequence
of a tendency to y downwind but to avoid being
forced over broad expanses of water. Similar behav-
iour may also account for large accumulations of
dragon ies, especially A. junius, during the fall
at southwardly directed peninsulas such as Point
Pelee, Ontario (Corbet and Eda 1969) and Cape
May, New Jersey (Russell et al. 1998). A number
of accounts suggest that some of these may redir-
ect their ight around the water barrier (Shannon
1916; Russell et al. 1998; Wikelski et al. 2006), but
other may cross expanses of water or remain
trapped until they perish with the onset of winter.
Sympetrum corruptum congregates similarly along
the Paci c coast of Washington and Oregon during
periods of east winds in fall (D. Paulson, personal
communication, based on numerous accounts on
the Internet).
6.2.2 Weather and climate
Weather profoundly affects when and how migrants
travel. Many insects, even strong iers like migra-
tory locusts that actively maintain a constant ight
heading, nevertheless actually move mostly pas-
sively with prevailing winds. Dumont (1977, 1988)
and Dumont and Desmet (1990) presented evidence
that Anax ephippiger migrations are mainly of this
type. Other tropical migrants like Pantala avescens
probably fall into the same category. Some of these
may y at great height (Corbet 1984). Even species
that closely follow distinct routes may take advan-
tage of favourable winds created by particular wea-
ther patterns, as migrating birds do. Bagg (1958)
Book 1.indb 66Book 1.indb 66 6/27/2008 8:40:12 PM6/27/2008 8:40:12 PM
MIGRATION IN ODONATA 67
there is a trend towards still higher fat stores in fall
migrants (Figure 6.1b). Notably, spring migrants
in New Jersey have much less fat, as expected if
they were near the end of their migratory jour-
ney. Likewise, the ovary size of spring migrants
appears to be about as large as in non-migrants,
whereas in fall migrants the mean is signi cantly
less, although maximum values are similar. This
would be expected if spring females were freshly
mature whereas fall females included a mixture
of mature and immature individuals. From radio-
tracking data on individual A. junius (see below),
substantial time during fall migration apparently
evidence suggests that sexual maturity and
fat stores increase progressively as the season
advances, so that by mid-October most individuals
are mature (Figure 6.1a). We have also seen pairs
in tandem among migrants in early September. It
appears likely that individuals initiate migration
soon after emergence but mature en route.
Most migrants, regardless of taxon, accumulate,
and, when possible, replenish, energy stores before
or during migration (Dingle 1996). Fat content in
A. junius is higher than in most dragon ies (<10%;
Anholt et al. 1991; M.L. May, unpublished results)
even among presumed non-migrants, although
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Early
Sept.
Late
Sept.
Early
Oct.
Early
Nov.
Fat
Ovary
(a)
(b)
Time period
Mass of fat or ovary as
proportion of total mass
Late
Oct.
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.00
0.05
Spring
migrant
Resident Fall
migrant
Fat in males
Fat in females
Ovary
Seasonal status
Mass of fat or ovary as
proportion of total mass
Figure 6.1 (a) Change with time in fat content
and ovary mass as a proportion of body mass
in fall migrant
A. junius.
Fat content increases
markedly from early September to mid-October,
then decreases slightly, suggesting progressive fat
accumulation during early fall, possibly followed
by reduction as migrants from farther north
are collected later. Ovary size increases, then
stabilizes, as the proportion of mature females
increases with time; no females were collected
in November, when remaining migrants are very
few. (b) Change in fat content and ovary mass as
a proportion of body mass with seasonal status in
A. junius
in New Jersey. Fat content is lowest in
putative spring migrants and highest in putative
fall migrants while ovary size is high in spring
migrants and residents but lower in fall migrants.
These patterns are expected if the former have
already migrated a long distance and are mature
while the latter are still accumulating fat and
include many immature individuals.
Book 1.indb 67Book 1.indb 67 6/27/2008 8:40:13 PM6/27/2008 8:40:13 PM
68 STUDIES IN ECOLOGY
locus (J.H. Matthews, G. Mynhardt, and A. Cognato,
unpublished results). Even with this data-set, how-
ever, no population structure was evident. Indeed,
the additional larvae only contributed a few new
haplotypes, and the resulting haplotype network
was very similar to that of Freeland et al. (2003),
with no signi cant geographic patterning.
Similar patterns have been seen with other spe-
cies in which many individuals migrate or disperse
long distances, such as European eels (Anguilla
anguilla). Early studies using coarse mitochondrial
and allozyme approaches found no geographic
patterns in species such as A. anguilla and mon-
arch butter ies (Danaus plexippus), leading to the
presumption that such vagile species were pan-
mictic (Brower and Jeansonne 2004). Development
of multilocus techniques using neutral markers
such as nuclear microsatellite DNA has markedly
improved molecular approaches to address such
questions, nally applying a tool suitable to the
task at hand. Thus, nine microsatellite loci were
able to parse the high gene- ow rates of A. anguilla,
estimating the Fst value (a standard measure of
inter-population gene ow; Wright 1951) to be as
low as 0.0002 (Wirth and Bernatchez 2001), nearing
the theoretical limits of Fst. Microsatellite loci have
recently been developed for A. junius (Matthews
et al. 2007a), signi cantly increasing their value as a
model for insect migration. Clearly, without devel-
oping markers for monarch butter ies that have a
similar degree of resolution to microsatellite loci,
we cannot make comparable inferences about gene
ow and population differentiation.
Biomolecular studies of movement analyse intrin-
sic proxy markers in individuals (Webster et al.
2002). Mark–release–recapture methods epitomize
acquired proxy markers, but new methods of study-
ing naturally acquired markers have been devel-
oped since the early 1990s. The application of stable
and radiogenic isotopic ratios to movement studies
assumes that an environmental source such as prey
items or ambient water contains recoverable infor-
mation about the origin or path of movement. Most
such studies have focused on birds, although stable
isotopes have proven highly effective at measur-
ing the scale of continental movement by monarch
butter ies and distinguishing separate migration
routes (e.g. Dockx et al. 2004).
is spent during days with little southward advance
that are thought to be devoted to feeding.
6.2.4 The scale and adaptive basis of
Anax
junius
movement
The study of why a species migrates requires a
description of the movement itself; the latter is a
prerequisite for untangling the former. Molecular
techniques have a long history as a means of
describing the geographic component of popula-
tions across generations. However, odonates are a
group with few such applications to date. Species
with relatively limited vagility have proven amen-
able to molecular approaches (Watts et al. 2004,
2007), whereas attempts to discern major patterns
of movement in odonates suspected of high rates
of long-distance movement have had little or no
resolution, even on a multi-continental scale (e.g.
Artiss 2004).
The rst signi cant attempt to understand the
scope of A. junius movement used a single mito-
chondrial locus, with samples of adults and a few
larvae taken over a multi-year period from eastern
Canada to Hawaii (Freeland et al. 2003). A nested
clade analysis of haplotypes (Templeton 2004)
showed a spiderlike pattern, with about half of all
individuals falling into one of two near-identical
haplotypes, with other lineages radiating out from
this core. No signi cant spatial pattern was found
that was related to this haplotype network, and
Freeland et al. (2003) concluded that high rates
of movement were blurring the ability to resolve
population structure.
Several confounding issues may have reduced the
effective power of this analysis. The collection site
of adults potentially capable of moving hundreds or
thousands of kilometres may not be a very mean-
ingful predictor of their natal pond. Indeed, with-
out additional information, a population structure
study based on adults may not be able to explain
geographic clustering. In a currently unpublished
study Matthews, Mynhardt, and Cognato focused
on a single 500-km transect of larvae from eastern
Texas to northern Arkansas, collecting enough indi-
viduals at each site over a narrow temporal win-
dow to make both intra- and inter-site comparisons
based on the sample mitochondrial DNA (mtDNA)
Book 1.indb 68Book 1.indb 68 6/27/2008 8:40:13 PM6/27/2008 8:40:13 PM
MIGRATION IN ODONATA 69
of the gene pool within eastern North America.
Together, these data suggest (1) that the mixed
swarms re ect largely random-origin assortments
of individuals, (2) that these individuals are mat-
ing and ovipositing during large-scale movement,
and (3) that mating and ovipositing are probably
occurring multiple times for each individual en
route. Eastern North America may reasonably be
described as a single population.
Thus, A. junius population structure is shallow.
In addition to the isotopic evidence mentioned
above, a northern spring movement is also implied
by the uniformity of genetic diversity across North
America. Repeated major southbound movements,
would eventually result in the gradual loss of gen-
etic diversity over most of the northern part of the
range (as in Hanski et al. 2004) unless compensated
by an in ux of alleles from elsewhere. When cou-
pled with a northbound movement, however, espe-
cially if by a different generation of individuals,
both genetic diversity and spatial homogeneity
would be maintained. Moreover, since reproduc-
tion and ovipositing are coupled during the move-
ment, long-distance displacement in A. junius is
not made in a pre-reproductive/teneral adult stage
seen in many migrant lepidopterans. This pat-
tern generally goes against the ‘classic’ views of
insect migration summarized in Dingle (1996), in
which large-scale movement precedes or follows
reproduction, focusing physiological resources
on movement and reproduction in series rather
than contemporaneously. In fact, the coupling of
migration and reproduction may make A. junius
migration a novel form of large-scale movement.
The combination of reproduction and long-
distance movement suggests that whereas A. junius
ight can safely be described as migration (Dingle
1996), this migration is not to a speci c locality.
Rather, movement is almost certainly from habi-
tats declining in seasonal quality to more southern
habitats with better average quality (but with a sub-
stantial variance in quality), suggesting migration
as ecological movement, occurring at a landscape
scale. The genetic evidence in particular suggests
that A. junius movement spreads reproductive risk
across multiple water bodies, minimizing threats
f ro m pr e dat io n, i nt ra - a nd in te r- sp ec i c competition,
and droughts. Given that A. junius larvae are most
A handful of studies have gone further and
combined molecular genetic and isotopic ratio
techniques to the same set of individuals, gener-
ating synergies between analytical techniques
with high power over large regions for much less
effort than required for direct tracking of individu-
als via telemetry or large-scale banding networks
(e.g. Chamberlain et al. 1997; Clegg et al. 2003). As
with isotopic ratio studies, these few multi-method
studies have focused overwhelmingly on birds.
By applying two isotopic ratio techniques with
microsatellite-based population genetics to a set
of 180 adult A. junius collected during a 6-week
window in the fall of 2005, Matthews (2007c) was
able to demonstrate continental-scale movement by
individual adults, some of which traveled over 2800
km before capture. Moreover, these distances were
measured as net north–south distances traveled—
displacement—rather than, as might be generated
by telemetry studies, the length of the actual route
followed. Mean southbound movement before col-
lection was over 900 km. Although collection sites
were all near coastal areas between 45 and 19°N
latitude, more than 80% of individuals originated
between 36 and 49°N, with more than 95% of indi-
viduals showing isotopic evidence of southbound
movement. Genetic and radiogenic isotopic ratio
evidence indicated that individuals at each collec-
tion site were ‘mixed swarms’, originating from
multiple latitudes, parent populations, and largely
inland sites although collection often occurred in
or within sight of tidal wetlands.
Genetic data also indicated that near-panmictic
conditions prevail, with continental-scale Fst val-
ues in the order of 0.04 for this group. Traditional
population genetic estimates of spatial organiza-
tion (e.g. a Mantel test) showed no isolation by
dista nce. Bayesian population clu steri ng of ind ivid-
uals into ‘populations’ based on microsatellite pat-
terns (Pritchard 2000) again found no geographic
pattern among southbound migrants. Moreover,
a large study of the population structure of some
500 individuals, including 300 larvae, found an
Fst value of 0.02 over the same spatial scale, a pat-
tern also found when focusing only on the subset
of larvae, at continental, regional, and single-pond
scales (Matthews 2007d). That is, populations in
single ponds contain a nearly random sample
Book 1.indb 69Book 1.indb 69 6/27/2008 8:40:13 PM6/27/2008 8:40:13 PM
70 STUDIES IN ECOLOGY
Caledon, Ontario. The spring-emerging group has
been referred to as residents or non-migrants, and
the fall-emerging group is normally described as
migrants.
Resident larvae take up to 1 year between ovi-
position and emergence, with a diapause or qui-
escent period during the winter months. Thus,
development proceeds through part of the fall
before stopping or slowing and begins again the
following late winter or spring. The migrant larvae
develop over a much shorter period spanning only
a few months between late spring and late summer
or early fall (Figure 6.2). These patterns have been
trac k ed at se vera l site s (su mm a r i zed in Cor bet 19 99),
including one locality as far south as 30°N lati-
tude (although at least one site at 42°N apparently
lacked bimodal emergence; M.L. May, unpublished
results). Generally, the spring cohort emerges earl-
ier and the fall cohort later with decreasing latitude
(J.H. Matthews, unpublished results). Moreover,
ponds with bimodal size distributions have been
abundant in shallow, largely lentic wetlands that are
often ephemeral and susceptible to even small shifts
in precipitation on a regional scale, a bet-hedging
strategy seems singularly appropriate as an adap-
tive basis for large-scale movement. Indeed, a large
body of theoretical literature on dispersal polymor-
phisms would predict such patterns (e.g. Roff 1994).
In a more speculative vein, the patterns and basis of
A. junius migration may be representative of many
taxa with a relatively non-dispersive aquatic larval
stage and a highly vagile reproductive adult stage,
such as many aquatic insects and amphibians.
6.2.5 The evolutionary ecology of dispersal
phenology and phenotype
Although Calvert (1929) and Walker (1958) made
scattered observations suggesting the presence of
more than one emergence group, Trottier (1971)
rst clearly described distinct but sympatric spring-
emerging and fall-emerging cohorts of A. junius in
Dispersal
Adult behavior
Larval
development
Dispersal
Oviposition
Dispause
Winter Spring Fall Winter
Locally dispersing larvae
Long-distance dispersing larvae
Summer
Egg Egg
Emergence Emergence
Oviposition
Figure 6.2 Generalized larval life-history tracks for
A. junius
. ‘Migrant’ fall-emergent larvae are shown with a dotted line (long-distance-
dispersing larvae). ‘Resident’ or ‘non-migrant’ spring-emergent larvae are shown with a dashed line (locally dispersing larvae). Note that
residents complete their life cycle in just under 1 year, undergo winter diapause, and are active as mature adults before emergence of
migrants. The latter complete development in 4–6 months without diapause. Both patterns have been observed between latitudes 45 and
19°N in eastern North America, but the clear separation of adult fl ight seasons is not always evident. After Trottier (1971).
Book 1.indb 70Book 1.indb 70 6/27/2008 8:40:13 PM6/27/2008 8:40:13 PM
MIGRATION IN ODONATA 71
(e.g. Hopper 2001; De Block and Stoks 2003). For any
particular plastic trait, a particular phenotype—
whether one of several qualitative alternatives (e.g.
phenotype A or B) or a point on a continuum of pos-
sibilities (e.g. phenotypic expression ranging from
1 to 10)—is cued by some environmental trigger.
A. junius presumably follow the qualitative model
with two alternative states. Trottier (1971) found
that the rate of development of late-instar larvae
could be manipulated across a wide range of val-
ues by ambient temperature, but he also found
that spring- and fall-emerging larvae showed dif-
ferent categories of responses to temperature. A
similar pattern was found by Corbet (1957) with
A. imperator. Moreover, Calvert (1929, 1934) found
that an A. junius larva bound for spring emergence
entered diapause when raised from an egg indoors
where it was exposed to relatively constant room-
temperature conditions rather than the harshness
of a Pennsylvania winter. All of these leads suggest
that emergence phenology is determined early in
development and that temperature has little or no
cueing effect for phenotype, even if temperat ure can
play a role in determining relative rates of devel-
opment within phenotypic states. Corbet (2003)
suggested that aeshnids at intermediate to high
latitudes may regulate emergence phenology via
photoperiod. Given the especially wide latitudinal
range of A. junius, photoperiod would be a very rea-
sonable and reliable cue for dispersal phenotype.
Matthews (2007e) tested this hypothesis with
split-sibling groups of A. junius exposed to con-
sta nt, increasing, and decreasi ng photoper iod from
the egg in an environmental chamber. The increas-
ing photoperiod treatment group (corresponding
to the fall-emerging larvae, with eggs laid between
the spring equinox and summer solstice) devel-
oped signi cantly (P < 0.001) more rapidly than the
equinoctial constant (light/dark period of 12 h:12 h)
and decreasing photoperiod (the latter correspond-
ing to eggs laid between the summer solstice and
fall equinox) groups. These results suggest that
there is little if any maternal effect on emergence
phenology, but they do not rule out the possibility
of a gene × environment effect, which might re ect
signi cant lineage differences in plasticity. Indeed,
several families did show signi cant differences in
their responses.
seen as far south as 19°N latitude (J.H. Matthews,
unpublished results). Another distinct emer-
gence ‘bump’ well before most spring-emergers
mature has also been observed in a few instances
(Wissinger 1988), suggesting that some larvae
‘destined’ for a fall emergence entered a temper-
ature-triggered quiescent state until the following
spring, perhaps as a result of early cold-water tem-
peratures, a pattern also seen in Anax imperator in
Europe (Corbet 1957).
Debate about the major spring- and fall-
emergence groups has centred on two points. First,
how are these emergence phenologies maintained
over large spatial scales? Second, what is the adap-
tive strategy maintaining this bimodal strategy?
The basis for determining emergence (and thus
dispersal) phenology and phenotype is arrived at
more easily. Direct and indirect evidence discussed
above suggests that fall-emergers migrate south
and spring-emergers migrate north, although only
limited est imates of the scale of north-bound move-
ment are available currently (Matthews 2007c).
Thus, the traditional terms resident and migrant
for describing these groups are misleading and
inaccurate. Moreover, since at least the time of
Trottier (1971), these groups have been described as
separate ‘populations’, even when referring to indi-
viduals belonging to different cohorts in the same
pond. Assuming that ‘population’ implies meas-
urably restricted gene ow between these groups,
the work of Freeland et al. (2003), Matthews (2007d),
and Matthews (2007e) suggests that these cohorts
are not different populations. Indeed, Matthews
(2007d) measured multilocus gene ow between
spring- and fall-emergers within and between 500
adult and larval samples collected over 27° of lati-
tude across eastern North America, nding Fst
values in the order of approximately 0.02 between
all categories, without regard to geographical
structuring. Although based on a much coarser
means of genetic resolution via a nested clade ana-
lysis, Freeland et al. (2003) found a similar pattern,
leading them to attribute variation in emergence
phenology to phenotypic plasticity rather than a
genetically determined phenotype. Conceivably, a
maternal effect could also be involved.
Phenotypic plasticity is widespread in nature and
has been identi ed in a handful of odonate species
Book 1.indb 71Book 1.indb 71 6/27/2008 8:40:13 PM6/27/2008 8:40:13 PM
72 STUDIES IN ECOLOGY
reaching adulthood in both shallow (ephemeral)
and deep (more permanent) water bodies, with
northern spring-emerging larvae (‘residents’) act-
ing as habitat specialists for permanent water bod-
ies (Matthews 2004). Several theorists predict just
such a system for the evolution of movement poly-
morphisms (Roff 1994). However, these adaptive
explanations remain to be tested.
6.2.6 Behaviour en route
6.2.6.1 Orientation
Migratory monarch butter ies are known to
use time-compensated sun-compass orientation
(Mouritson and Frost 2002), and other butter ies
may be able to use magnetic orientation (Srygley
et al. 2006). Nothing, however, is known with cer-
tainty about mechanisms of long-distance orienta-
tion in dragon ies. Typical routes of aggregated
migrants follow obvious landmarks for much
of their length, and several writers, especially
Shannon (1916) and Dumont and Hinnekint (1973),
have emphasized the importance of visual land-
marks in orienting migration. On the other hand,
as already noted, accumulation along leading lines
need involve no more than downwind ight and
avoidance of open water. Furthermore, dragon ies
may sometimes pass over long stretches of water,
such as Delaware Bay or Lake Erie (Corbet 1984;
Root 1912), or featureless plains without obvious
landmarks. Corbet (1984), reporting on movements
of A. ephippiger, a well-documented Old World
migrant, on the plains of East Africa, suggested that
their strikingly constant ight heading was due to
sun-compass orientation.
We suggest that landmarks are probably used
at least as secondary aids to navigation as sug-
gested above but note that for consistent southerly
movement along a coastline, for example, some
mechanism of choosing one of two directions
must exist. Likewise, it appears that individual
Anax can distinguish northerly from southerly
winds (see below). Certainly sun-compass orien-
tation is possible given the known capabilities of
other insects. On the other hand if, as we suppose,
a speci c narrow destination is not necessary for
A. junius, it is possible that positive phototaxis
toward the sun or the brighter southern sky in fall
In essence, any individual egg could develop
along the slow spring-emergence pathway or the
fast fall-emergence route based on the photo-
period experienced by the egg or a very early
instar. This result ts neatly with the genetic data
(Matthews 2007d) showing no signi cant differ-
ences between spring-emerging and fall-emerging
larvae. Temperature may modulate the rate of
development within the limits set by photoperiod
and, presumably, determine the direction of adult
migration as well. This model also implies that the
small emergence bump noted by Wissinger (1988)
may re ect mismatches between temperature
modulation and photoperiod cues, leading a few
individuals to develop rapidly before ambient tem-
peratures triggered diapause, halting development
in the nal or near- nal instar.
Although not tested experimentally, no evidence
indicates an effect of larval crowding on migra-
tory predilections, as is well known in locusts
(reviewed by Dingle 1996). For example, no effect
on development or evidence to suggest early
migration has been observed in populations in
which the early-emerging cohort is much larger
than the late-emerging larvae (M.L. May, unpub-
lished results).
The second issue mentioned above regarding
the adaptive basis of emergence phenology and
phenotype is much harder to test. The key ques-
tion here is, why maintain two and only two
cohorts? One theory is based on the observation
that there appears to be a latitudinal relation-
ship between the climate-normal timing of rain
and emergence phenology across eastern North
America (Matthews 2007a, 2007b). This relation-
ship suggests that long-distance migration may be
timed in the fall to arrive when the largest number
of suitable habitats are available; for example, lake
and wetland water levels in south Florida generally
peak in October and, at least in larger lakes, remain
high until February (Abtew et al. 2006), consistent
with arrival of probable northern migrants and
timing of local emergence (Paulson 1999). Northern
movement is in turn set to coincide with the coinci-
dence of appropriate temperatures for larval devel-
opment with high spring water levels. In a closely
related hypothesis, northern fall-emerging larvae
(‘migrants’) may be habitat generalists, capable of
Book 1.indb 72Book 1.indb 72 6/27/2008 8:40:14 PM6/27/2008 8:40:14 PM
MIGRATION IN ODONATA 73
day always succeeded a night that was colder than
the previous night, whereas non-migrating days
had on average warmer nights leading up to them.
This behaviour makes sense if migrating after a
decrease in night-time temperatures ensures that
individuals ordinarily migrate when cold, north-
erly winds aid their southward migration.
Three individuals changed their migration
route by more than 120° when hitting an ocean
barrier (e.g. the Delaware Bay), clearly reorienting
in response to landmarks. Two actually ew out
over the bay for about 5 km, then turned north
to spend the night on the New Jersey side, about
30 km north of the capture point; 2–4 days later
they crossed the upper end of the bay, as shown in
Figure 6.4. Such behaviour mirrors that of numer-
ous songbirds and small hawks upon reaching
this same spot during fall migration (e.g. Wiedner
et al. 1992).
might suf ce for southward orientation. If so, this
might again be a useful model for other insects
with similar migratory strategy. Of interest would
be to determine whether orientation in spring is
based only on reversal of the response to similar
cues as during fall and, in any case, how this inter-
generational change is mediated. Also unresolved
is whether some other odonate species might have
more stringent navigational requirements (e.g.
A. ephippiger, which may depend on nding desert
oases; Dumont and Desmet 1990) and species that
sometimes migrate at night (Feng et al. 2006).
6.2.6.2 Flight patterns
Virtually nothing is known about individual deci-
sion rules and trajectories of migratory insects,
because it is very dif cult to follow such small
organisms over vast distances. Recently, however,
Wikelski et al. (2006) were able to attach micro-
radio transmitters to 14 A. junius and follow them
during fall migration for up to 12 days. Although
the presence of the transmitters may have affected
behaviour, individuals migrated up to 140 km per
day, and two were observed foraging apparently
normally.
These observations provide unique insight into
individual behaviour. A. junius alternated distinct
stopover periods with active migration and on
average migrated about every 3 days. The average
advance of 13 migrating individuals was approxi-
mately 60 km (12 km/day), but daily movement
ranges (Figure 6.3a) exhibited a trimodal distri-
bution: short-range and omni-directional and
medium or long-range and directional. Average
direction was within a few degrees of due south.
Dragon ies only migrated when wind speeds
were low, independent of wind direction, although
win ds on d ays when th e dr agon ies migrated were,
on average, more northerly than on non-migratory
days, thus indicating, as suggested above, that
A. junius have at least a rudimentary directional
sense. They did not seem to compensate strongly
for wind drift as individual migration paths only
differed by approximately 20° on average from
wind direction (Figure 6.3b). Thus ight paths
often zig-zagged on consecutive days. Daily high
temperatures did not differ signi cantly between
migrating and non-migrating days, but a migrating
12
(a)
(b)
9
6
3
01.00
–60 –40 –20
Flight direction in relation to wind
direction (degrees)
020406080
10.00
Stopover Migration
100.00
Ave. daily distance (km)
FrequencyFrequency
4
3
2
1
0
Figure 6.3 (a) Daily movement of radio-tagged
A. junius
;
movements greater than 5 km were assumed to represent migration
whereas shorter movements were thought to occur during periods
when the dragonfl ies forage to replenish energy stores and are thus
relatively sedentary. (b) Flight direction relative to wind direction of
radio-tagged
A. junius
; zero deviation represents due south. Note
that the modal direction is quite close to south and that none of the
movement vectors deviated more than 90° from south. Both from
Wikelski
et al.
(2006).
Book 1.indb 73Book 1.indb 73 6/27/2008 8:40:14 PM6/27/2008 8:40:14 PM
74 STUDIES IN ECOLOGY
establishes that oviposition occurs during south-
ward ight, although strong indirect and anecdotal
evidence also indicates that much reproduction
occurs at or near the southern terminus of migra-
tion (Paulson 1999; D.R. Paulson, personal com-
munication; J.H. Matthews, personal observation).
In fact, A. junius is incapable of adult diapause as
far as we know (unlike some tropical Anisoptera,
including probable migrants; Corbet 1999). A. junius
are generalized predators at all life stages, so adults
can obtain food widely as long as ying insects are
available, unlike specialized plant feeders such as
monarch butter ies or Oncopeltus bugs, and larvae
can develop in frequently encountered, but unstable,
aquatic habitats. Moreover, the aquatic larval habi-
tat provides a refuge from thermal extremes. These
characteristics, coupled with high adult mobility,
allow Anax to reproduce in a succession of sites as it
moves progressively southward. Repeated ovipos-
ition both spreads risk to offspring from drought and
predation and allows adults to employ a bet-hedging
strategy against cumulative mortality risks during
migration. We expect that other organisms with low
6.3 Comparisons with other
airborne migrants: what can
dragonfl ies teach us?
A. junius differs in several respects from most bet-
ter-known animal migrants. Unlike birds and mam-
mals, they do not provide parental care beyond
oviposition and both sexes are able to reproduce
at intervals shorter than the duration of migration
throughout most of their adult lives. Unlike migra-
tory insects that have been closely studied hereto-
fore, they apparently do not exhibit an oogenesis
ight syndrome, although we cannot discount the
possibility that maturation might be somewhat
slowed during migration. As already noted, many
southbound migrants are evidently reproductively
mature. Tandem linkage is seen from southern
Ontario southward (J.H. Matthews and M.L. May,
independent personal observations), and one of us
(JHM) observed oviposition by apparent migrants
throughout Atlantic and Gulf coastal regions in the
USA in September and in Veracruz state, Mexico,
as late as mid October. Thus direct observation
5
3
3
9
53
~
~~
~~
~~
~~~~
~~~~
~~~~
~~~~
~~~~
~~~~
~~~~
~~
3
3
3
3
6
6
(10)
(10)
(10)
2
4
4
4
4
7
7
1
1
12
1
4Trenton
NY
Atlantic
Ocean
Atlantic City
Delaware
Bay
DE
MD
N
100 km
PA
Philadelphia NJ
Figure 6.4 Flight tracks of 13 radio-tagged
A. junius
in Fall
2005 in New Jersey. Each line represents a separate individual;
numbers depict days since tagging, numbers in parentheses show
maximum number of days individuals were tracked, broken lines
indicate uncertainty about which day individuals conducted their
migratory fl ight. Note that four individuals changed direction
near the coastline of Delaware Bay, apparently avoiding fl ying far
over water. One individual (black line) initially fl ew over Delaware
Bay for about 5 km, then turned north and stopped for the night
about 25 km north; it was re-found on day 3 in Maryland and is
assumed to have crossed the bay well north of its capture point.
From Wikelski
et al.
(2006).
Book 1.indb 74Book 1.indb 74 6/27/2008 8:40:14 PM6/27/2008 8:40:14 PM
MIGRATION IN ODONATA 75
Artiss, T. (2004) Phylogeography of a facultatively migra-
tory dragon y, Libellula quadrimaculata (Odonata:
Anisoptera). Hydrobiologia 515, 225–234.
Bagg, A.M. (1958) Fall emigration of the dragon- y, Anax
junius. Maine Field Naturalist 14, 2–13.
Borror, D.J. (1953) A migratory ight of dragon ies.
Entomological News 64, 204–205.
Brower, A.V.Z. and Jeansonne, M.M. (2004) Geographical
populations and “subspecies” of new world monarch
butter ies (Nymphalidae) share a recent or igin a nd are
not phylogenetically distinct. Annals of the Entomological
Society of America 97, 519–523.
Calvert, P.P. (1893) Catalogue of the Odonata (dragon ies)
of the vicinity of Philadelphia, with an introduction to
the study of this group of insects. Transactions of the
American Entomological Society 20, 152–272.
Calvert, P.P. (1929) Different rates of growth among ani-
mals with special reference to the Odonata. Proceedings
of the American Entomological Society 68, 227–274.
Calvert, P.P. (1934) The rates of growth, larval develop-
ment and seasonal distribution of dragon ies of the
genus Anax (Odonata: Aeshnidae). Proceedings of the
American Philosophical Society 73, 1–70.
Chamberlain, C.P., Blum, R.T., Holmes, R.T., Feng, X.,
Sherry, T.W., and Graves, G.R. (1997) The use of isotope
tracers for identifying populations of migratory birds.
Oecologia 109, 132–141.
Clegg, S.M., Kelly, J.F., Kimura, M., and Smith, T.B. (2003)
Combining genetic markers and stable isotopes to
reveal population connectivity and migration patterns
in a Neotropical migrant, Wilson’s warbler (Wilsonia
pusilla). Molecular Ecology 12, 819–830.
Cook, C. (1991) Editors comments. Argia 3(4), 14.
Corbet, P.S. (1957) The life-history of the emperor dragon-
y, Anax imperator Leach (odonata: Aeshnidae). Journal
of Animal Ecology 26, 1–69.
Corbet, P.S. (1984) Orientation and reproductive condition
of migrating dragon ies (Anisoptera). Odonatologica
13, 81–88.
Corbet, P.S. (1999) Dragon ies: Behaviour and Ecology of
Odonata. Cornell University Press, Ithaca, NY.
Corbet, P.S. (2003) Reproductive behaviour of Odonata:
the history of a mystery. International Journal of
Odonatology 6, 185–193.
Corbet, P.S. and Eda, S. (1969) Odonata in southern
Ontario, Canada, in August 1968. To mb o 12, 4–11.
Daigle, J.J. (1991) A late summer collecting trip to Texas.
Argia 3(4), 8.
De Block, M, and Stoks, R. (2003) Adaptive sex-speci c
life-history plasticity to temperature and photo-
period in a damsel y. Journal of Evolutionary Biolog y 16,
9 8 6 – 9 9 5 .
parental care costs and generalized and reasonably
abundant energy resources, especially other aquatic
insects in unstable lentic habitats, may employ simi-
lar strategies.
On the other hand, individual migratory behav-
iours and decision rules are strikingly similar to
those proposed for songbirds, and may represent
a general migration strategy for directional long-
distance travel of self-propelled aerial organisms
(Wikelski et al. 2006). These studies should be rep-
licated and extended, both geographically and to
include spring migration, but they can provide
independent con rmation of movement patterns
inferred from stable isotope markers as well as
more details about behaviour, possibly includ-
ing reproductive behaviour and success. In any
event, they highlight characteristics that are likely
to be shared very widely among migrants across
taxa because of pervasive effects of limitations on
energy stores in small, ying organisms and the
consequent needs to replenish internal supplies
and take advantage of external power sources; that
is, favourable winds.
Acknowledgements
Our sincere thanks to the many people who have
helped and encouraged us during this work, includ-
ing those, too numerous to mention individually,
who sent observations of dragon y migrations.
MLM especially acknowledges Bob Barber, Joanna
Freeland, Bob Russell, and Ken Soltesz for earl-
ier collaboration and Martin Wikelski and Dave
Moskowitz for introducing him to the art and
science of radio-tracking. Funded in part by New
Jersey Agricultural Experimentation Project 08190,
supported by State and U.S. Hatch Act Funds
to MLM.
References
Abtew, W., Huebner, R.S., and Ciuca, V. (2006) Chapter
5: hydrology of the South Florida environment,
pp. 5-1–5-85. 2006 South Florida Environmental Report.
South Florida Water Management District, West Palm
Beach, FL.
Anholt, B.R., Marden, J., and Jenkins, D. (1991) Patterns
of mass gain and sexual dimorphism in adult dragon-
ies. Canadian Journal of Zoology 69, 1156–1163.
Book 1.indb 75Book 1.indb 75 6/27/2008 8:40:14 PM6/27/2008 8:40:14 PM
76 STUDIES IN ECOLOGY
Matthews, J.H. (2004) Report on Anax junius eme rgence in
Caledon, Ontario, in 2003. Ontario Odonata 5, 12–14.
Matthews, J.H. (2006) Tiny migrants and the man who
chased them. Hawk Mountain News 105, 18–21.
Matthews, J.H. (2007a) Wetland invertebrates and climate
change: what happens when the pond is half-empty?
Wings 30(2), 5–10.
Matthews, J.H. (2007b) Anthropogenic climate change
impacts on ponds: a thermal mass perspective. In Ott,
J. (ed.), Odonate Conservation: Climate Change Impacts and
Monitoring Strategies, pp. 183–207. Pensoft Publishers,
So a.
Matthews, J.H. (2007c) Continental-scale migration by
a North American dragon y revealed by population
genetic and stable and radiogenic isotopes. In Research
in Motion: Patterns of Large-scale Migration in Dragon ies
and Birds, pp. 7–32, PhD Dissertation, University of
Tex as.
Matthews, J.H. (2007d) Large-scale dragon y migration
by Anax junius (Odonata: Aeshn idae) results in genetic
homogeneity despite temporally and spatially discrete
populations. In Research in Motion: Patterns of Large-
scale Migration in Dragon ies and Birds, pp. 33–61, PhD
Dissertation, University of Texas.
Matthews, J.H. (2007e) Phenotypic plasticity deter-
mines migration strategy across large spatial scales.
In Research in Motion: Patterns of Large-scale Migration
in Dragon ies and Birds, pp. 62–91, PhD Dissertation,
University of Texas.
Matthews, J.H., Boles, T., Parmesan, C., and Juenger, T.
(2007) Isolation and characterization of nuclear micro-
satellite loci for the common green darner dragon y
Anax junius (Odonata: Aesh nidae) to constrai n patterns
of phenotypic and spatial diversity. Molecular Ecolog y
Notes 7, 845–847.
Mouritsen, H. and Frost, B.J. (2002) Virtual migration in
tethered ying monarch butter ies reveals their orien-
tation mechanisms. Proceedings of the National Academy
of Science USA 99, 10162–10166.
Nisbet, I.C.T. (1960) Notes on the migration of dragon-
ies in southern Ontario. Canadian Field Naturalist 74,
150–153.
Osburn, R.C. (1916) A migratory ight of dragon ies.
Journal of the New York Entomological Society 24, 90–92.
Paulson, D. (1998) Specimens tell story of migration—but
no longer? Argia 9(4), 22.
Paulson, D.R. (1999) Dragon ies (Odonata: Anisoptera) of
southern Florida. Occasional Papers of the Slater Museum
no. 57.
Paulson, D.R. (2002) Odonata records from Nayarit and
Sinaloa, Mexico, with comments on natural history
and biogeography. Odonatologica 31, 359–370.
Dingle, H. (1996) Migration: the Biology of Life on the Move.
Oxford University Press, Oxford.
Dingle, H. (2006) Animal migration: is there a com-
mon migratory syndrome. Journal of Ornithology 147,
212–220.
Dockx, C., Brower, L.P., Wassenaar, L., and Hobson, K.
(2004) Do North American Monarch butter ies migrate
to Cuba? Stable isotope and chemical tracer techniques.
Ecological Applications 14, 110 4–1114.
Dumont, H.J. (1977) On migrations of Hemianax ephippi-
ger (Burmeister) and Tramea basilaris (P .de Beauvois) in
west and north-west Africa in the winter of 1975/1976
(Anisoptera: Aeshnidae, Libelludiae). Odonatologica 6,
13–17.
Dumont, H.J. (1988) Hemianax ephippiger (Bunneister) in
the northern Algerian Sahara in winter (Anisoptera:
Aeshnidae). Notulae Odonatologica 3, 20–22.
Dumont, H.J. and Hinnekint, B.O.N. (1973) Mass migra-
tion in dragon ies, especially in Libellula quadrimacu-
lata L.: a review, a new ecological approach and a new
hypothesis. Odonatologica 2, 1–20.
Dumont, H.J. and Desmet, K. (1990) TransSahara and
transMediterranean migratory activity of Hemianax
ephippiger (Burmeister) in 1988 and 1989 (Anisoptera:
Aeshnidae). Odonatologica 19, 181–185.
Feng, H.-K., Wu, K.-M., Ni, Y.-X., Cheng, D.-F., and Guo,
Y.-Y. (2006) Noctur nal migration of dragon ies over the
Bohai Sea in northern China. Ecological Entomology 31,
511–520.
Freeland, J., May, M.L., Lodge, R., and Conrad, K.F. (2003)
Genetic diversity and widespread haplotypes in a
migratory dragon y, the common green darner (Anax
junius). Ecological Entomology 28, 413–421.
Glotzhober, R.C. (1991) Ohio dragon y survey produces
interesting observations: “stinging dragon ies” and
migrating swarms. Argia 3(4), 13–14.
Hack, M.A. and Rubenstein, D.I. (2001) Migration. In
Levin, S.A. (ed.), Encyclopedia of Biodiversity, vol. 4,
pp. 221–233. Academic Press, San Diego, CA.
Hagen, H. (1861) Über Insektenzuge. Stettin Entomologische
Zeitschrift 22, 73–83.
Hanski, I., Erlahti, C., Kankare, M., Ovaskainen, O., and
Siren, H. (2004) Variation in migration propensity
among individuals maintained by landscape struc-
ture. Ecology Letters 7, 958–966.
Hopper, K.R. (2001) Flexible antipredatory behavior in a
dragon y species that coexists with different predator
types. Oikos 93, 470–476.
Johnson, C.G. (1969) Migration and Dispersal of Insects by
Flight. Methuen & Co., London.
Kennedy, J.S. (1985) Migration, behavioral and ecological.
Contributions in Marine Science (suppl.) 27, 5–26.
Book 1.indb 76Book 1.indb 76 6/27/2008 8:40:14 PM6/27/2008 8:40:14 PM
MIGRATION IN ODONATA 77
Trottier, R. (1971) Effect of temperature on the life-cycle of
Anax junius (Odonata: Aeshnidae) in Canada. Canadian
Entomologist 103, 1671–1683.
Walker, E.M. (1958) The Odonata of Canada and Alaska. Vo l.
II. The Anisoptera—Four Families. University of Toronto
Press, Toronto.
Watts, P.C., Rouquette, J.R., Saccheri, I.J., Kemp, S.J., and
Thompson, D.J. (2004) Molecular and ecological evi-
dence for small-scale isolation by distance in an endan-
gered damsel y Molecular Ecology 13, 2931–2945.
Watts, P.C., Rousset, F., Saccheri, I.J., Leblois, R.L., Kemp,
S.J., and Thompson, D.J. (2007) Compatible genetic
and ecological estimates of dispersal rates in insect
(Coenagrion mercuriale: Odonata: Zygoptera) popula-
tions: analysis of ‘neighbourhood size’ using a more
precise estimator. Molecular Ecology 16, 737–751.
Webster, M.S., Marra, P.P., Haig, S.M., Bensch, S., and
Holmes, R.T. (2002) Links between worlds: unravelling
migratory connectivity. Trends in Ecology and Evolution
17, 76–82.
Wiedner, D.S., Kerlinger, P., Sibley, D.A., Holt, P. Hough,
J., and Crossley, R. (1992) Visible morning ight of
Neotropical landbird migrants at Cape May, New
Jersey. Auk 109, 500–510.
Wikelski, M., Moskowitz, D., Adelman, J.S., Cochran, J.,
Wilcove, D.S., and May, M.L. (2006) Simple rules guide
dragon y migration. Biology Letters 2, 325–329.
Wirth, T. and Bernatchez, L. (2001) Genetic evidence against
panmixia in the European eel. Nature 409, 1037–1040.
Wissinger, S.A. (1988) Life history and size structure
of larval dragon y populations Journal of the North
American Benthological Society 7, 13–28.
Wright, S. (1951) The genetical structure of populations.
Annals of Eugenics 15, 323–354.
Pritchard, J.K., Stephens, M., and Donnelly, P.J. (2000)
Inference of population structure using multilocus
genotype data. Genetics 155, 945–959.
Rabb, R.L. and Stinner, R.E. (1979) The role of insect
dispersal and migration in population processes.
Proceedings of the USDA/NASA Workshop on Radar,
Insect Population Ecology and Pest Management,
Wallops Island, Virginia. NASA Conference Publication
2070, 3–16.
Roff, D.A. (1994) Habitat persistence the evolution of
wing dimorphism in insects. American Naturalist 144,
772–798.
Root, F.M. (1912) Dragon ies collected at Point Pelee and
Pelee Island, Ontario, in the summers of 1910 and 1911.
Canadian Entomologist 44, 208–209.
Russell, R.W., May, M.L., Soltesz, K.L., and Fitzpatrick,
J.W. (1998) Massive swarm migrations of dragon ies
(Odonata) in eastern North America. American Midland
Naturalist 140, 325–342.
Shannon, H.J. (1916) Insect migrations as related to those
of birds. Scienti c Monthly 3, 227–240.
Sprandel, G. (2001) Fall dragon y (Odonata) and butter y
(Lepidoptera) migration at St. Joseph Peninsula, Gulf
County, Florida. Florida Entomologist 84, 234–238.
Srygley, R.B., Dudley, R., Oliveira, E.G., and Riveros, A.J.
(2006) Experimental evidence for a magnetic sense
in Neotropical migrating butter ies (Lepidoptera:
Pieridae). Animal Behaviour 71, 183–191.
Templeton, A.R. (2004) Statistical phylogeography: meth-
ods of evaluating and minimizing inference errors.
Molecular Ecology 13, 789–809.
Trottier, R. (1966) The emergence and sex ratio of Anax
junius (Odonata: Aeshnidae) in Canada. Canadian
Entomologist 98, 794–798.
Book 1.indb 77Book 1.indb 77 6/27/2008 8:40:14 PM6/27/2008 8:40:14 PM
Book 1.indb 78Book 1.indb 78 6/27/2008 8:40:15 PM6/27/2008 8:40:15 PM
