• Coenagrion mercuriale;
  • conservation;
  • density dependence;
  • inverse density dependence;
  • landscape connectivity;
  • metapopulations;
  • Odonata


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Coenagrion mercuriale is one of Europe's rarest and most threatened damselflies. It is listed in the European Community (EC) Habitats and Species Directive and is the only odonate currently given priority status in the UK Biodiversity Action Plan. Information regarding patterns of movement and dispersal in this species is required to guide conservation and management programmes. Management is currently geared towards habitat restoration of isolated subpopulations, with little attention paid to the metapopulation and landscape context.
  • 2
    A multisite mark–release–recapture project was carried out in the valley of the River Itchen in southern England to determine the extent of movement and the factors affecting movement of mature adults of this endangered damselfly. A total of 8708 individuals was marked.
  • 3
    The species was found to be extremely sedentary, with dispersal limited to an area of contiguous habitat. The median net lifetime movement was 31·9 m and 66% of individuals moved less than 50 m in their lifetime. Movements of greater than 500 m were rare and the longest recorded movement was 1·79 km. This makes it the most sedentary odonate that has been studied in the UK.
  • 4
    The highest recapture rates and the lowest movement distances were recorded at the most isolated site. Time between capture and recapture, and day in season had an effect on movement, and individuals travelled further on their first than on subsequent moves. There was no consistent effect of age or sex on distance moved.
  • 5
    There was strong evidence for inverse density-dependent movement, with individuals moving further in low-density than high-density populations. This is the first time that inverse density-dependent movement has, to our knowledge, been observed in a natural population of odonates.
  • 6
    Synthesis and applications. Coenagrion mercuriale, along with many other invertebrate species of conservation concern, lives in a management-dependent mid-successional habitat. However, the species is highly sedentary. Furthermore, patterns of movement and dispersal are strongly affected by landscape structure and population density. This means that it is unable to recolonize isolated sites and requires ‘stepping stone’ habitats to improve its chances of survival in the medium to long term. Suitable habitat management between sites that are beyond the dispersal distance of individuals can be used to connect or reconnect populations. Within existing sites only small sections of habitat should be managed in any one year and new areas should be created close to existing populations. The long-term persistence of C. mercuriale and other invertebrate species requires a landscape approach to management, with connectivity an important part of management planning. It is clear that carefully conducted studies of movement and dispersal are key components in guiding invertebrate conservation strategies.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Movement and dispersal play a fundamental role in the ecology and evolution of species. These processes drive local and metapopulation dynamics, determine the spatial scale of evolutionary change, and dictate the response of organisms to fragmentation and climate change (Dieckmann, O’Hara & Weisser 1999; Clobert et al. 2001; Bullock, Kenward & Hails 2002). Understanding movement and dispersal is becoming increasingly important as landscapes become ever more fragmented and species numbers continue to decline (Baguette, Petit & Quéva 2000; Petersen et al. 2004).

Patterns of movement and dispersal are strongly influenced by the structure of the landscape. Increased habitat fragmentation will lead to an increased mortality rate associated with dispersal, and this can eventually lead to the loss of genes coding for dispersal in isolated populations (Dieckmann, O’Hara & Weisser 1999). If they become isolated, small populations can lose genetic variation through inbreeding and genetic drift, and will become increasingly prone to extinction. Connectivity therefore has a major bearing on the persistence of local and regional populations (Baguette, Petit & Quéva 2000) and on community composition (Reckendorfer et al. 2006).

Most studies of dispersal in insects have concentrated on Lepidoptera, with relatively few studies of Odonata. However, odonates, especially damselflies, are particularly good study organisms. They are large, conspicuous, easily handled and straightforward to mark. They live in inherently patchy environments, because they are restricted to aquatic habitats for larval development, and most of the mature adult life is spent at or near to breeding sites.

In this study we used mark–release–recapture (MRR) methods to measure directly movement of the endangered damselfly Coenagrion mercuriale (Charpentier) (Odonata: Coenagrionidae). Previous studies of C. mercuriale have suggested that most individuals are extremely sedentary, although a few move distances of up to about 1 km (Hunger & Röske 2001; Purse et al. 2003). Our study system was much larger in scale than these previous studies, including several sites with two areas of contrasting landscape structure. The study area was divided into two by a major urban area, with a large area of near-continuous habitat to the south and an area of smaller, more isolated sites to the north. Sites were arranged in a linear series along a river valley; hence movement was effectively constrained to one dimension at the landscape scale.

The aims of the study were to investigate patterns of dispersal of C. mercuriale in an effort to understand the ecology and conservation requirements of this endangered species. We examined dispersal both between sites along the valley of a chalk river and within subsites of the largest stretch of contiguous suitable habitat, which were under three different management regimes. We also examined factors that might potentially influence movement within each site and subsite, notably age, time, day of season, sex and density. In a parallel study, we investigated movement in these populations indirectly using DNA microsatellite markers (Watts et al. 2004).


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

study species

Coenagrion mercuriale is a habitat specialist and is restricted to two fragmented biotopes in the British Isles (Thompson, Rouquette & Purse 2003). These are small lowland heathland streams emanating from base-rich substrates, and calcareous streams and fens. Since 1985 its distribution has declined by 38% (Purse 2001) and it is classified as rare in the British Red Data Book (Shirt 1987). It is protected under the Wildlife & Countryside Act of 1981, and is the only odonate currently given priority status in the UK Biodiversity Action Plan (HMSO 1994, 1995). It is also threatened in the rest of Europe, particularly at the northern and eastern boundaries of its distribution (Grand 1996), and is protected under the Bern Convention and the European Community (EC) Habitats Directive. The UK is thought to contain a significant proportion of the total global population, although data are lacking from other European countries.

study sites

The study area was located between Winchester and Southampton in Hampshire (southern England). Here, C. mercuriale is found principally on old water meadow carriers and ditches along the flood plain of the River Itchen. The species occurs in three main areas, which from north to south are Mariner's Meadow, Highbridge and the Lower Itchen Complex (LIC). The LIC is separated from Highbridge by approximately 3 km of largely suboptimal habitat, made up of an urban area and sections of intensive agriculture. Mariner's Meadow and Highbridge are also separated by approximately 3 km. An additional site (Twyford Moors) does exist between these two, where C. mercuriale occurs in low numbers, but where we were not able to perform a MRR study. The study sites are described in more detail elsewhere (Rouquette & Thompson 2005).

The LIC is a large area of near-continuous habitat (c. 3·5 km in length); thus, to ensure comparable sampling intensity, it was divided into five subsites of approximately equal length. From north to south these were West Horton, Allington Manor and three subsites (Upper, Middle and Lower) within the Itchen Valley Country Park (IVCP). In total, 7·65 km of ditch were surveyed at the seven sites and subsites, hereafter referred to as sites.

mrr survey

The MRR survey was performed at all seven sites in the summer of 2001. Damselflies were captured with a kite net and the location was recorded using a global positioning system (GPS) calibrated to the UK Ordnance Survey (accurate to approximately ±5 m). Animals were marked by writing a unique alphanumeric code on the left forewing in waterproof ink and by putting a small dab of paint on the thorax. We sampled all sites from approximately 09:30 to 16:00 every day for 42 days from 12 June 2001, except during bad weather, when adult damselflies are not active. This coincided with the peak flight period in this area. Although we did not sample at the very beginning or end of the flight period, a previous study (Purse et al. 2003) did not report any alteration in movement patterns during these periods.

statistical analysis

Distances moved were calculated as the straight-line distance between initial and subsequent captures. When multiple captures of the same individual on the same day were made, only the first capture was included. The movement parameters estimated (modified from Scott 1973) are shown in Table 1.

Table 1.  The movement parameters estimated
dDistance moved between successive captures (m)
tTime between successive captures (days)
vVelocity (m days−1) (d/t)
DCumulative distance moved (m); sum of d for each individual
LNet lifetime movement (m); distance between first and last captures
TTime between first and last capture (days)
VNet velocity (m days−1) (L/T)

A multiple regression was performed to investigate the effects of site (coded as a series of dummy variables), sex, age (midpoint age during movement), day (midpoint day of season during movement), time (t) and order of movement (first movement, second movement and so on) on d. We used a combination of backwards elimination and stepwise procedures to select significant variables. All analyses were carried out on log-transformed distances [log10(d + 1)] as the data were highly skewed. This transformation was successful at improving the distribution and fit of the data. All analyses were performed using SPSS version 11·0.

To investigate whether movement patterns varied within each site, we divided each site into 50 × 50-m sections, and the damselflies marked or recaptured within each section were tabulated. Thus the scale of each section reflected the approximate scale of lifetime movement for the majority of damselflies (see the Results). We tested for differences between sections by running a one-way anova for each site. Density was then calculated as the average number of C. mercuriale seen in each section per day of recording. The effect of density on movement was investigated by running a regression of the log10 mean distance moved by damselflies starting in each section against log10 density, weighted by sample size. We also plotted cumulative distance dispersed for three density categories, low (<1 C. mercuriale per section per day), medium (1–10) and high (> 10).


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

numbers marked and recaptured

In total, 8708 C. mercuriale were marked, 7659 males and 1049 females. Of these, 2523 individuals were recaptured (29·0%) at least once and there were 3727 recapture events. The maximum number of times an individual was recaptured was eight for a male at Mariner's Meadow and the longest time between first and last capture was 29 days for a male at IVCP Upper.

A breakdown of the numbers marked and recaptured at each site is provided in Table 2. It was clear that the Upper and Middle IVCP contained particularly strong populations, although the damselfly was present in reasonable numbers at all sites. The site with the lowest population was West Horton, where it was found in good numbers on one short stretch of stream but was sparsely represented on the rest of the site.

Table 2.  Total numbers and percentage of adult C. mercuriale marked and recaptured at each site and movements between sites. Recapture figures refer to recaptures on days subsequent to marking or previous capture
Site MarkedRecapturedMovement events
Mariner's MeadowMales 959 43345·2 793 0 0
Females 185  4725·4  59 0 0
Total1144 48042·0 852 0 0
HighbridgeMales 716 28539·8 450 0 0
Females  63  1117·5  11 0 0
Total 779 29638·0 461 0 0
West HortonMales 251  7128·3 10413 8
Females  14   0 0   0 0 0
Total 265  7126·8 10413 8
Allington ManorMales 637 22635·5 3731022
Females  57   3 5·3   3 0 0
Total 694 22933·0 3761022
IVCP UpperMales2106 65130·9 89814 7
Females 378  3910·3  50 0 0
Total2484 69027·8 94814 7
IVCP MiddleMales2038 44822·0 5773124
Females 232  18 7·8  22 2 0
Total2270 46620·5 5993324
IVCP LowerMales 952 27829·2 3741724
Females 120  1310·8  13 0 2
Total1072 29127·1 3871726
All sitesMales7659239231·235698585
Females1049 13112·5 158 2 2

The proportion of marked individuals recaptured varied among sites, with the highest proportion occurring at Mariner's Meadow (42·0%) and the lowest at IVCP Middle (20·5%). The differences were highly significant (χ2 = 151·1, d.f. = 6, P < 0·001) and were still apparent if the lower five sites were amalgamated before analysis (χ2 = 112·8, d.f. = 2, P < 0·001). Males were significantly (χ2 = 112·4, d.f. = 1, P < 0·001) more likely to be recaptured (31·2%) than females (12·5%), and this pattern was true at all sites (Table 2).

movement patterns

Eighty-five of the recaptured individuals (3·4%) transferred between sites, with two individuals transferring twice to give a total of 87 movement events (Table 2). There were no observed movements between Mariner's Meadow and Highbridge (the two isolated sites) and any other site, but movement was recorded to and from the remaining five sites in the LIC.

Net lifetime movement was defined as the distance from where the animal was first marked to the place where it was last recaptured. The pattern of movement was similar at each site (Fig. 1). The overall median net lifetime movement recorded in this study was 31·9 m (geometric mean = 33·2 m, n= 2523), and 65·7% of individuals moved less than 50 m in their lifetime. However, differences between the sites were also apparent. Damselflies at Mariner's Meadow were the most sedentary, with more than 75% moving less than 50 m in their lifetime, while only about 40% moved that short a distance at Allington Manor. Furthermore, the maximum distance moved by any individual was 554 m at Mariner's Meadow and 406 m at Highbridge, but was 1374 m at IVCP Upper and 1790 m at IVCP Middle.


Figure 1. Percentage distribution of net lifetime movements (L) in 25-m distance categories for (a) Mariner's Meadow, (b) Highbridge and (c) Lower Itchen Complex, all to the same scale. Sample sizes are 480, 296 and 1747 individuals, respectively.

Download figure to PowerPoint

The mean net lifetime movement for each site is illustrated in Fig. 2. The mean was lowest at Mariner's Meadow (25·7 m) and highest at Allington Manor (68·1 m). The differences between the sites were highly significant (one-way anova, F= 23·9, d.f. = 6,2516, P < 0·001). The same pattern of results was evident when we used cumulative distance moved (D) and net velocity (V). Indeed, both were highly correlated with L (D, rs = 0·863, P < 0·001; V, rs = 0·697, P < 0·001) and so were not considered further.


Figure 2. Net lifetime movement (L) at each site (bar shows mean and 95% confidence interval). Net lifetime movement is significantly different across the sites (one-way anova, F= 23·9, d.f. = 6,2516, P < 0·001). Means displaying the same letter are not significantly different at the 5% probability level (Tukey multiple comparison test).

Download figure to PowerPoint

factors affecting movement

A number of variables had a significant effect on C. mercuriale movement (Table 3). The longer the time between consecutive captures, the further the damselfly had travelled (Fig. 3a). There was a quadratic effect of day, with slightly greater movement occurring in the middle of the marking period and a tail-off towards the end of the season (Fig. 3b). Although this was statistically significant, the effect was small. Variation in distance moved was also greater towards the end of the marking period, reflected in the larger confidence intervals, as the number of individuals recorded each day gradually declined.

Table 3.  Significant predictors of mean distance moved (log10) by C. mercuriale derived from a multiple regression model. The F-value and the associated P-value, d.f., R2 and adjusted R2 are shown (*P < 0·05, **P < 0·01, ***P < 0·001). For each variable retained in the model, the P-value derived from t-tests, parameter estimates and standard errors are shown
Model summaryVariabletPParameter estimatesStandard error
F = 73·04Time 15·79*** 0·03560·0023
P = ***Site, Allington Manor  8·36*** 0·2100·025
d.f. = 7,3719Site, Mariner's Meadow−6·05***−0·1150·019
R2 = 0·121Site, IVCP Upper−4·33***−0·07700·0178
Adjusted R2 = 0·119Order of movement−3·34***−0·02990·0089
Day in season  2·00* 0·00620·0031
(Day in season)2−2·79**−0·00020·0001
Constant 43·18*** 1·3740·0318

Figure 3. Effect of (a) time, (b) season and (c) age on mean distance moved (log d). Vertical lines show 95% confidence intervals, which are not included for samples with less than five individuals. Dotted lines on (a) and (b) are partial regression lines derived from a multiple regression model (see text and Table 3 for more details). The effect of age (c) was not significant and so a regression line was not added. Day in season was taken as the midpoint day between capture and recapture, where day 1 was the first day of marking (12 June 2001). Age was taken as the midpoint between first capture and each recapture.

Download figure to PowerPoint

The effect of order of movement on distance moved was also significant (Fig. 4a). Individuals travelled further on their first move than subsequently, and distance declined logarithmically the more moves that were made. Three sites were also included in the regression model; movement was significantly greater at Allington Manor and significantly shorter at Mariner's Meadow and IVCP Upper than the model average (Table 3 and Fig. 4b). This was consistent with the net lifetime movement patterns explained above. Overall these variables had a highly significant effect on distance moved, but the amount of variation explained was relatively small (r2 = 0·121).


Figure 4. Effect of (a) order of movement (first movement, second movement, etc.) and (b) site and sex on mean distance moved (log d). Males are shown with dark bars, females with light bars. Vertical lines show 95% confidence intervals. The solid lines are partial regression lines derived from a multiple regression model (see text and Table 3 for more details). The effect of sex was not significant and so the regression line amalgamates the two sexes.

Download figure to PowerPoint

There was no effect of age on movement (Fig. 3c). The effect of sex was not consistent across the sites (Fig. 4b). Males moved further than females at Mariner's Meadow but the reverse was true at all the other sites.

the effect of c. mercuriale population density on distance moved

The results above illustrate broad-scale differences in movement patterns between sites. However, an analysis of the 50 × 50-m sections revealed that there were highly significant differences in the distance moved in different sections at Mariner's Meadow and in all three subsites in the IVCP, although movement was similar in all parts of Allington Manor (Table 4).

Table 4.  Number of 50 × 50-m sections at each site, the results of one-way anovas to test for differences in distance moved from different sections, and regressions of mean distance moved by damselflies marked in each section against density. All analyses were performed on log-transformed data. One-way anovas were performed separately on movements that started and ended within each section. The F-value, the associated P-value (*P < 0·05; **P < 0·01; ***P < 0·001; NS, not significant) and d.f. are shown. The regression analyses were weighted by sample size and ‘Curve’ indicates whether the best fitting model was linear (L) or quadratic (Q)
SiteNo. of sectionsDifference in distance moved by damselfliesRegressions of mean distance against density
Starting in different sectionsEnding in different sections
Mariner's Meadow 188·18*** 17,8215·34*** 17,82875·83***0·817L
Highbridge  61·38NS  5,4333·31**  5,431 0·91NS0·154L
West Horton  81·86NS  7,882·34*  6,88 5·80NS0·492L
Allington Manor 321·14NS 31,3320·82NS 31,34310·00***0·408Q
IVCP Upper 187·73*** 17,9106·46*** 14,88331·17***0·806Q
IVCP Mid 294·24*** 28,5853·71*** 23,58222·89***0·459L
IVCP Lower 272·54*** 26,3372·04** 25,35411·15**0·308L
All sites1385·14***137,35064·71***127,350993·76***0·578Q

The mean distance moved by damselflies starting in each section was negatively related to density (Fig. 5). A linear regression, weighted by sample size, provided a good fit to the data (r2 = 0·547, F= 166·41, P < 0·001) but a slightly improved fit was achieved by adding a quadratic term (r2 = 0·578, F= 93·76, P < 0·001). A significant effect of density was found at each site except for Highbridge, and the effect was of marginal significance (P = 0·053) at West Horton (Table 4). These two sites had the smallest number of sections, with only six and eight, respectively. At all the remaining sites a linear relationship was significant, but at two of the sites a quadratic relationship provided an improved fit. The strongest relationships were at Mariner's Meadow (linear r2 = 0·817, F= 75·83, P < 0·001) and at IVCP Upper (quadratic r2 = 0·806, F= 31·17, P < 0·001), the two sites with the highest densities of C. mercuriale.


Figure 5. Regression of mean distance moved (log d) against density for each 50 × 50-m section. Density was calculated as the average number of individuals (marks and recaptures) seen in each section per day. The solid line is the weighted regression line (r2 = 0·578, F= 93·76, P < 0·001).

Download figure to PowerPoint

Our results indicated that movement of C. mercuriale was inversely density dependent, with the effect levelling off at higher densities. This was confirmed when we plotted the cumulative distance moved by C. mercuriale for three density categories (Fig. 6). There was a clear separation of each density category, with consistently shorter movements from damselflies in higher density areas. This effect was found at all sites and there was no difference in the response between males and females.


Figure 6. Cumulative distance moved by C. mercuriale in three density categories. The lightest (dashed) line represents sections with a mean of less than 1 C. mercuriale, the medium-weight line 1–10 individuals and the thickest line represents sections with > 10 individuals day−1.

Download figure to PowerPoint


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

numbers recaptured and sex ratio

The percentage of individuals recaptured in this study (29·0%) was remarkably similar to that reported by Purse et al. (2003), who had recapture rates of 29·0% and 30·9% in sites in the New Forest and Pembrokeshire, respectively.

The sex ratio at the breeding sites was strongly male biased, as found in most odonate species (Cordoba-Aguilar 1994; Stettmer 1996; Stoks 2001a), and is probably brought about through a combination of developmental, survival and behavioural differences between the sexes. There is no difference in the pattern of emergence of the two sexes in C. mercuriale (Purse & Thompson 2003). However, it is believed that, as female odonates take longer to mature, survivorship during this period is lower for females (Banks & Thompson 1985; Bennett & Mill 1995; Stoks 2001b) and their pattern of behaviour once mature is also different. Females will only visit oviposition sites when ready to mate; they will subsequently leave the area and will not return until a new batch of eggs has matured. Males, on the other hand, will tend to arrive at the breeding sites first and remain for much longer periods.

movement patterns

There were no direct movements of marked damselflies between the two isolated northern sites and the remainder, but several movements occurred between the sites in the area of contiguous habitat. Movements over 500 m were rare (1·3% of individuals) and there were only three movements greater than 1000 m (0·1%). This would suggest that populations in the northern sites are ecologically isolated but that those in the southern sites could be considered to be one large population. These findings are encouraging, as they are remarkably well supported by analysis of DNA microsatellite markers. Watts et al. (2004) found that damselflies from the five areas of contiguous habitat were genetically similar but that samples from Highbridge and Mariner's Meadow showed significant genetic differentiation. Indeed, there was a significant correlation between genetic differentiation and geographical distance and evidence for isolation by distance even over the short distances present in our study area. Within the area of continuous habitat, genetic samples showed a pattern of positive autocorrelation over short distances, but showed isolation by distance over a distance of about 1000 m.

The maximum net lifetime movement recorded in this study (1790 m) was longer than that recorded previously (1060 m; Thompson & Purse 1999; Purse et al. 2003). However, Purse et al. (2003) measured dispersal from patches of only 300 m in length. In our study system the maximum possible between patch dispersal distance was approximately 9 km, and the LIC provided a near-continuous patch 3·5 km in length. The scale of our study system was clearly much greater than the scale over which C. mercuriale typically moves, providing us with increased confidence in the accuracy of our results and particularly of the tail of the dispersal distribution.

Coenagrion mercuriale is a weak flier and a poor disperser compared with other odonates (Garrison 1978; Bennett & Mill 1995; Stettmer 1996; Schutte, Reich & Plachter 1997). It is the smallest of the blue damselflies found in the UK, and distance moved and dispersal probability have been reported to increase with increasing species size for a range of odonates (Conrad et al. 1999; Angelibert & Giani 2003).

All movement parameters measured were lowest at Mariner's Meadow. As sites become more isolated, dispersing individuals become less likely to find suitable habitat. Thus the mortality rate associated with dispersal increases, and this can eventually lead to the loss of genes coding for dispersal in these populations (Dieckmann, O’Hara & Weisser 1999). Thus theory predicts that movement will be lower at Mariner's Meadow and at Highbridge, and this is supported by our genetic analysis (Watts et al. 2004). However, Mariner's Meadow and Highbridge also contained the smallest lengths of suitable habitat, and so long-distance within-patch movements were inevitably missing. Mariner's Meadow also contained some of the highest density sections of C. mercuriale, along with IVCP Upper, and movement was lowest from these two sites. This could provide further evidence of inverse density-dependent movement. Finally, habitat quality was good at Mariner's Meadow and IVCP Upper and correspondingly bad at Allington Manor (J. R. Rouquette, unpublished data). Coenagrion mercuriale may simply be moving away from areas of less suitable habitat and staying in the most suitable habitat (Rouquette & Thompson 2005).

factors affecting movement and dispersal

The length of time between captures had the greatest effect on distance moved, and this is consistent with a previous study of C. mercuriale (Purse et al. 2003). It has also been reported for Calopteryx splendens (Schutte, Reich & Plachter 1997), and Enallagma cyathigerum (Garrison 1978).

As far as we are aware, this is the first study of mature adult odonates to observe an effect of day of season on local movement patterns, although the effect was small. The two most likely causes are weather and phenotype. Weather is known to affect odonate activity; damselflies are more likely to be on the wing in good weather (Angelibert & Giani 2003) but could be dispersed over long distances in windy weather (Corbet 1999). Damselfly phenotype also varies over the course of the flight period. For example, body size at emergence declines over the course of the season in C. mercuriale (Purse & Thompson 2003). If movement was correlated with body size then this would explain the pattern in our data, although Purse (2001) found no evidence for this.

Coenagrion mercuriale moves further on its first movement than subsequently, and movement declines with each subsequent move. Order of movement is highly correlated with age (rs = 0·642, P < 0·001) but provides a better fit to the data. Indeed, if order of movement is removed from the multiple regression, age is added in its place. This pattern is different to that seen in Sympetrum danae (Michiels & Dhondt 1991) and many butterflies, where it is common for females in particular to move increasing distances with age or number of moves (Warren 1987; Bergman & Landin 2002).

This study has only examined movement patterns in mature adults. Dispersal by tenerals (newly emerged immature adults), usually by means of a maiden flight, has been suggested to be the most important dispersive phase for some odonates (Anholt 1990; Corbet 1999), although Conrad et al. (1999) and Angelibert & Giani (2003) found no difference in dispersal between immature and mature adults in a number of pond-dwelling odonates. Tenerals were not marked in our study because of the risk of damaging this protected species, and retaining individuals until their wings harden may alter behaviour upon release. It is also possible that movement occurs through the process of larval drift. However, odonates are not significant components of the larval drift fauna and C. mercuriale prefers slow-flowing marginal habitats (Thompson, Rouquette & Purse 2003; Rouquette & Thompson 2005) and so is unlikely to drift far. Furthermore, an analysis of the direction of movement of C. mercuriale in our study sites provided little evidence to suggest that damselflies moved upstream to counteract the affect of drifting downstream (J. R. Rouquette, unpublished data). Thus the ecology of the species combined with our knowledge of the genetic structure of the Itchen Valley populations would suggest that larval or teneral dispersal is generally limited and if occurring probably follows the same pattern as mature adult dispersal.

No consistent differences between the sexes were found in this study, and this is consistent with previous work on C. mercuriale (Purse 2001). Sex differences in movement patterns in other damselflies is equivocal, although where present it is usually females that move further (Bennett & Mill 1995; Conrad et al. 2002; Angelibert & Giani 2003).

the effect of c. mercuriale population density on distance moved

One of the most interesting findings of this study is that C. mercuriale movement is inversely density dependent. Habitat quality and C. mercuriale density are correlated, and it is likely that the presence of conspecifics is used as an indicator of good habitat quality. Martens (2000) showed that tandem pairs of C. mercuriale landed preferentially on leaves where a single motionless male in the typical vertical position of a tandem male had been placed.

Inverse density-dependent dispersal has not, to our knowledge, been observed previously in natural populations of odonates, although a manipulation experiment has hinted at this behaviour. Michiels & Dhondt (1991) used a large outdoor field cage to study the non-territorial dragonfly Sympetrum danae. Male movement was inversely density dependent, using female density as the cue, but females showed the reverse behaviour.

Inverse density-dependent movement has been reported in some butterfly studies (Gilbert & Singer 1973; Brown & Ehrlich 1980; Kuussaari, Nieminen & Hanski 1996; Menendez, Gutierrez & Thomas 2002) and in a study of bush crickets (Kindvall et al. 1998). All of these studies have shown a tendency for individuals to move further when conspecific density is low and/or to have a greater propensity to emigrate from such areas. Conversely, immigration is more likely to patches with a high density of conspecifics. This attraction to conspecifics and emigration from small populations could be because of the need to find mates, avoid inbreeding or find high-quality habitat. Matter & Roland (2002) reported that the immigration of male butterflies was related positively to aspects of habitat quality and to female density in a manipulation experiment where they were able to tease apart these two factors. Odendaal, Turchin & Stermitz (1988) argued that male aggregations in many insect species could be a side-effect of mate-finding behaviour; the tendency of males to chase all conspecifics can lead to the incidental formation of aggregations, regardless of the presence of females.

Reduced individual fitness or population growth rate at low population size or density is generally referred to as an Allee effect (Allee et al. 1949; Courchamp, Clutton-Brock & Grenfell 1999; Stephens & Sutherland 1999) and has been reported in some studies (Kuussaari et al. 1998; Menendez, Gutierrez & Thomas 2002; Liebhold & Bascompte 2003). We do not know if mating success is reduced in our lower density areas, but by increasing movement in these areas individuals may be enhancing the chances of securing a successful mating, thereby increasing fitness and reducing the chance of an Allee effect (Kindvall et al. 1998). Such behaviour is likely to have profound consequences for the population dynamics of the region. In areas of continuous habitat, this behaviour will lead to aggregation and increased competition for mates. In areas of patchy habitat, small populations in relatively isolated areas are more likely to go extinct as a result of higher emigration rates. Increased emigration would also make surviving populations more susceptible to other causes of extinction, such as inbreeding depression, genetic drift, Allee effects and stochastic environmental processes.

conservation implications

The limited dispersal capability shown by C. mercuriale has implications for its conservation and management. The species requires slow to medium flowing channels, with shallow margins and abundant emergent vegetation (Rouquette & Thompson 2005). In most of the areas that C. mercuriale occupies in the UK, this represents a successional phase that will not last without active management. Indeed, C. mercuriale has been lost from sites that have become choked with vegetation or shaded (Purse 2001). It is therefore imperative that management works are carried out and that they are tailored to the scale of movements observed. In other words, only small sections of stream should be managed in any one year and new areas should be created close to existing populations.

It has been suggested that insects living in successional habitats should show dispersal ability related to the life span of the habitat (Southwood 1962). This does not seem to be the case with C. mercuriale and its limited dispersal is likely to be one of the factors causing its decline. However, this may be a reflection of past landscape stability created through years of traditional land management rather than the situation that currently prevails.

It is clear from direct measurements of dispersal described here and indirect genetic analysis (Watts et al. 2004) that the two populations in the north of the study area are isolated from each other and from the southern population, even though the distance between Mariner's Meadow and Highbridge is about the same as the distance between IVCP Lower and West Horton. This also illustrates that suitable habitat management between sites that are beyond the dispersal distance of individuals can be used to connect or reconnect populations. Another important finding of this study is that movement is inversely density dependent. One effect of this is that populations on small isolated sites will be more likely to go extinct, as a larger proportion of individuals will attempt to emigrate. Thus landscape connectivity becomes even more important. The long-term persistence of C. mercuriale in the Itchen Valley and elsewhere requires a landscape approach to management. New habitat should be created between the existing sites to reconnect the extant populations, and connectivity should be a key component of all management planning.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Pearl Chung, Jason Farthers, David Giles, Nina Graham, Sam Jacobs, Sam Jones, David Lock, Ken Monro, Viv Owens, Lisa Parker, Jac Pearson, Rachel Remnant, Angie Squires, Gaya Sriskanthan, Debs Stickley, Pete Taylor, Jules Tipper, Carri Westgarth and Kerry Woodbine for assistance with the fieldwork. We thank Tim Sykes for his enthusiasm and help at all stages of the project and Ian Harvey for statistical advice. The work was funded by the Itchen Sustainability Study Group, the Environment Agency, the Natural Environment Research Council (grant no. NER/A/S/2000/01322), Bovis Homes Ltd, Persimmon Homes Ltd and Wimpey Homes Ltd. In addition, J. R. Rouquette was supported by the Environment Agency, English Nature, the Countryside Council for Wales and the University of Liverpool. Coenagrion mercuriale is protected under Schedule 5 of the UK Wildlife & Countryside Act (1981) and all work was carried out under licence from English Nature.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Allee, W.C., Emerson, A.E., Park, O., Park, T. & Schmidt, K.P. (1949) Principles in Animal Ecology. Saunders, Philadelphia, PA.
  • Angelibert, S. & Giani, N. (2003) Dispersal characteristics of three odonate species in a patchy habitat. Ecography, 26, 1320.
  • Anholt, B.R. (1990) Size-biased dispersal prior to breeding in a damselfly. Oecologia, 83, 385387.
  • Baguette, M., Petit, S. & Quéva, F. (2000) Population spatial structure and migration of three butterfly species within the same habitat network: consequences for conservation. Journal of Applied Ecology, 37, 100108.
  • Banks, M.J. & Thompson, D.J. (1985) Emergence, longevity and breeding area fidelity in Coenagrion puella (L.) (Zygoptera: Coenagrionidae). Odonatologica, 14, 279286.
  • Bennett, S. & Mill, P.J. (1995) Pre- and post-maturation survival in adults of the damselfly Pyrrhosoma nymphula (Zygoptera: Coenagrionidae). Journal of Zoology, 235, 559575.
  • Bergman, K.O. & Landin, J. (2002) Population structure and movements of a threatened butterfly (Lopinga achine) in a fragmented landscape in Sweden. Biological Conservation, 108, 361369.
  • Brown, I.L. & Ehrlich, P.R. (1980) Population biology of the checkerspot butterfly, Euphydryas chalcedona. Structure of the Jasper Ridge colony. Oecologia, 47, 239251.
  • Bullock, J.M., Kenward, R.E. & Hails, R.S. (2002) Dispersal Ecology. Blackwell Publishing, Oxford, UK.
  • Clobert, J., Danchin, E., Dhondt, A.A. & Nichols, J.D. (2001) Dispersal. Oxford University Press, New York, NY.
  • Conrad, K.F., Willson, K.H., Harvey, I.F., Thomas, C.J. & Sherratt, T.N. (1999) Dispersal characteristics of seven odonate species in an agricultural landscape. Ecography, 22, 524531.
  • Conrad, K.F., Willson, K.H., Whitfield, K., Harvey, I.F., Thomas, C.J. & Sherratt, T.N. (2002) Characteristics of dispersing Ischnura elegans and Coenagrion puella (Odonata): age, sex, size, morph and ectoparasitism. Ecography, 25, 439445.
  • Corbet, P.S. (1999) Dragonflies: Behaviour and Ecology of Odonata. Harley Books, Colchester, UK.
  • Cordoba-Aguilar, A. (1994) Adult survival and movement in males of the damselfly Hetaerina cruentata (Odonata: Calopterygidae). Florida Entomologist, 77, 256264.
  • Courchamp, F., Clutton-Brock, T. & Grenfell, B. (1999) Inverse density dependence and the Allee effect. Trends in Ecology and Evolution, 14, 405410.
  • Dieckmann, U., O’Hara, B. & Weisser, W. (1999) The evolutionary ecology of dispersal. Trends in Ecology and Evolution, 14, 8890.
  • Garrison, R.W. (1978) A mark–recapture study of imaginal Enallagma cyathigerum (Charpentier) and Argia vivida Hagen (Zygoptera: Coenagrionidae). Odonatologica, 7, 223236.
  • Gilbert, L.E. & Singer, M.C. (1973) Dispersal and gene flow in a butterfly species. American Naturalist, 107, 5872.
  • Grand, D. (1996) Coenagrion mercuriale (Charpentier, 1840). Background Information on Invertebrates of the Habitats Directive and the Bern Convention. Part II. Mantodea, Odonata, Orthoptera and Arachnida (eds P.J.V.Helsdingen, L.Willemse & M.C.D.Speight), Vol. 80, pp. 245253. Council of Europe Publishing, Strasbourg, France.
  • HMSO (1994) Biodiversity: The UK Action Plan. HMSO, London, UK.
  • HMSO (1995) Biodiversity: The UK Steering Group Report, Vol. 2. Action Plans. HMSO, London, UK.
  • Hunger, H. & Röske, W. (2001) Short-range dispersal of the southern damselfly (Coenagrion mercuriale, Odonata) defined experimentally using UV fluorescent ink. Zeitschrift für Ökologie und Naturschutz, 9, 181187.
  • Kindvall, O., Vessby, K., Berggren, A. & Hartman, G. (1998) Individual mobility prevents an Allee effect in sparse populations of the bush cricket Metrioptera roeseli: an experimental study. Oikos, 81, 449457.
  • Kuussaari, M., Nieminen, M. & Hanski, I. (1996) An experimental study of migration in the Glanville fritillary butterfly Melitaea cinxia. Journal of Animal Ecology, 65, 791801.
  • Kuussaari, M., Saccheri, I., Camara, M. & Hanski, I. (1998) Allee effect and population dynamics in the Glanville fritillary butterfly. Oikos, 82, 384392.
  • Liebhold, A. & Bascompte, J. (2003) The Allee effect, stochastic dynamics and the eradication of alien species. Ecology Letters, 6, 133140.
  • Martens, A. (2000) Group oviposition in Coenagrion mercuriale (Charpentier) (Zygoptera: Coenagrionidae). Odonatologica, 29, 329332.
  • Matter, S.F. & Roland, J. (2002) An experimental examination of the effects of habitat quality on the dispersal and local abundance of the butterfly Parnassius smintheus. Ecological Entomology, 27, 308316.
  • Menendez, R., Gutierrez, D. & Thomas, C.D. (2002) Migration and Allee effects in the six-spot burnet moth Zygaena filipendulae. Ecological Entomology, 27, 317325.
  • Michiels, N.K. & Dhondt, A.A. (1991) Characteristics of dispersal in sexually mature dragonflies. Ecological Entomology, 16, 449459.
  • Odendaal, F.J., Turchin, P. & Stermitz, F.R. (1988) An incidental-effect hypothesis explaining aggregation of males in a population of Euphydryas anicia. American Naturalist, 132, 735749.
  • Petersen, I., Masters, Z., Hildrew, A.G. & Ormerod, S.J. (2004) Dispersal of adult aquatic insects in catchments of differing land use. Journal of Applied Ecology, 41, 934950.
  • Purse, B.V. (2001) The ecology and conservation of the southern damselfly (Coenagrion mercuriale). PhD Thesis. University of Liverpool, Liverpool, UK.
  • Purse, B.V. & Thompson, D.J. (2003) Emergence of the damselflies, Coenagrion mercuriale and Ceriagrion tenellum (Odonata: Coenagrionidae), at their northern range margins in Britain. European Journal of Entomology, 100, 9399.
  • Purse, B.V., Hopkins, G.W., Day, K.J. & Thompson, D.J. (2003) Dispersal characteristics and management of a rare damselfly. Journal of Applied Ecology, 40, 716728.
  • Reckendorfer, W., Baranyi, C., Funk, A. & Schiemer, F. (2006) Floodplain restoration by reinforcing hydrological connectivity: expected effects on aquatic mollusc communities. Journal of Applied Ecology, 43, 474484.
  • Rouquette, J.R. & Thompson, D.J. (2005) Habitat associations of the endangered damselfly, Coenagrion mercuriale, in a water meadow ditch system in southern England. Biological Conservation, 123, 225235.
  • Schutte, G., Reich, M. & Plachter, H. (1997) Mobility of the rheobiont damselfly Calopteryx splendens (Harris) in fragmented habitats (Zygoptera: Calopterygidae). Odonatologica, 26, 317327.
  • Scott, J.A. (1973) Convergence of population biology and adult behaviour in two sympatric butterflies, Neominois ridingsii (Papilionoidea: Nymphalidae) and Amblyscirtes simius (Hesperioidea: Hesperiidae). Journal of Animal Ecology, 42, 663672.
  • Shirt, D.B. (1987) British Red Data Books. 2. Insects. Nature Conservancy Council, Peterborough, UK.
  • Southwood, T.R.E. (1962) Migration of terrestrial arthropods in relation to habitat. Biological Reviews, 37, 171214.
  • Stephens, P.A. & Sutherland, W.J. (1999) Consequences of the Allee effect for behaviour, ecology and conservation. Trends in Ecology and Evolution, 14, 401405.
  • Stettmer, C. (1996) Colonization and dispersal patterns of banded (Calopteryx splendens) and beautiful demoiselles (C. virgo) (Odonata: Calopterygidae) in south-east German streams. European Journal of Entomology, 93, 579593.
  • Stoks, R. (2001a) Male-biased sex ratios in mature damselfly populations: real or artefact? Ecological Entomology, 26, 181187.
  • Stoks, R. (2001b) What causes male-biased sex ratios in mature damselfly populations? Ecological Entomology, 26, 188197.
  • Thompson, D.J. & Purse, B.V. (1999) A search for long-distance dispersal in the southern damselfly Coenagrion mercuriale (Charpentier). Journal of the British Dragonfly Society, 15, 4650.
  • Thompson, D.J., Rouquette, J.R. & Purse, B.V. (2003) Ecology of the Southern Damselfly, Coenagrion mercuriale. Conserving Natura 2000 Rivers Ecology Series No. 8. English Nature, Peterborough, UK.
  • Warren, M.S. (1987) The ecology and conservation of the heath fritillary butterfly, Mellicta athalia. II. Adult population structure and mobility. Journal of Applied Ecology, 24, 483498.
  • Watts, P.C., Rouquette, J.R., Saccheri, I.J., Kemp, S.J. & Thompson, D.J. (2004) Molecular and ecological evidence for small-scale isolation by distance in an endangered damselfly, Coenagrion mercuriale. Molecular Ecology, 13, 29312945.