Content uploaded by Gergely Horváth
Author content
All content in this area was uploaded by Gergely Horváth on Mar 02, 2015
Content may be subject to copyright.
TISCIA 39, 9-15
9
ASSESSMENT OF RIVERINE DRAGONFLIES (ODONATA:
GOMPHIDAE) AND THE EMERGENCE BEHAVIOUR OF THEIR
LARVAE BASED ON EXUVIAE DATA ON THE REACH OF THE
RIVER TISZA IN SZEGED
G. Horváth
Horváth, G. (2012): Assessment of riverine dragonflies (Odonata: Gomphidae) and the emergence
behaviour of their larvae based on exuviae data on the reach of the river Tisza in Szeged. — Tiscia
39, 9-15
Abstract. Abundance, phenology, sex ratio, emergence pattern, mortality and larval emergence
behaviour of riverine dragonflies (Odonata: Gomphidae) were studied at the Lower-Tisza reach at
Szeged (168–173 rkm) during the emergence period in 2011. Three 20 meter long sampling sites
were chosen and searched systematically for exuviae, dead specimens and dragonfly wings, which
were left behind by bird predators. At the studied reach of the river two species form stable
populations: G. flavipes and G. vulgatissimus. G. flavipes was much more abundant than G.
vulgatissimus. Exuviae indicated the excess of females in the G. vulgatissimus population (altough
there were no significant difference between sexes), while in the case of G. flavipes the number of
individuals in both sexes were almost the same. G. vulgatissimus started to emerge first as a ’spring
species’, while G. flavipes started to emerge about a month later showing the characteristics of a
’summer species’. The rate of mortality in the G. flavipes population during emergence was slight
and quite normal compared to the abundance of the species. Selection of emergence support of G.
flavipes showed that the significant majority of the larvae chose soil, but this could have been
caused by the notable minority of other types of substrates at the sampling sites. The distance
crawled by the larvae from the water-front to the emergence site differed significantly between the
two species, G. vulgatissumus crawled further, and in the case of G. flavipes the effect of the
measured background variables to the distance had not been proven.
Key words: Gomphus flavipes, G. vulgatissimus, collections of exuviae, abundance, emergence
pattern, sex ratio.
G. Horváth, Department of Ecology, University of Szeged, H-6726 Szeged, Közép fasor 52.,
Hungary
Introduction
If we want to examine Odonata (in this
particular case Gomphidae) populations, there are
three different methods to carry out the work:
The most difficult way is the imago based
examination because of the excellent manoeuvring
skills and hiding behaviour of the adults. The
collection of larvae is also not easy, sampling can be
problematic in large watercourses. In the case of
these two methods there is another disadvantage in a
conservational point of view, as imagines and larvae
may die during the collection.
Knowing all those, the most reliable and the
simplest method to define the population size and
emergence specificity of riverine dragonflies is the
regular quantitative collection of their exuviae. The
best time to estimate the accurate size of the
Gomphidae population is during the emergence
period (Suhling and Müller – cit. Farkas et al.
2012a). Beside species composition and abundance,
exuviae provide information about phenology,
pattern of emergence, sex ratio, mortality during the
emergence, we could make statement about
morphological features or even about the moult-
strategy of the larvae (Berzi-Nagy 2011, Farkas et al.
10 TISCIA 39
2009, 2011, 2012a,b; Jakab 2006). The application
of the method is highly recommended, because it
does not require the collection of living animals so it
is not objectionable in a conservational point of view
either. Furthermore, exuviae of Anisopterans remain
intact for a long time, even under unsuitable weather
conditions (Jakab 2006) so there is no need for daily
collection.
During the past few years, many studies were
published about the riverine dragonfly populations of
the Upper- and Middle-Tisza regions (Bánkuti et al.
1997; Mátyus 2006 – cit. Berzi-Nagy 2011; Berzi-
Nagy 2011; Farkas et al. 2009; Farkas et al. 2012a;
Jakab 2006). Nevertheless, this present study is the
first to discuss the Gomphid assemblages of the
Lower-Tisza region.
According to literature two Gomphidae species,
the River Clubtail (Gomphus flavipes) and the
Common Clubtail (Gomphus vulgatissimus) were
expected to occur with abundance big enough to
form stable populations along this region. In the light
of the current Gomphidae based works my goals
were to reveal the sex-ratio and emergence
characteristics (phenology, emergence pattern) of the
riverine dragonflies that inhabit the reach of the river
Tisza near Szeged. I also examined the emergence
behaviour (the distance larvae crawled from the
waterfront and correlation with the variables, and
substrate preference) and mortality during the
emergence.
Materials and Methods
Study sites and sampling
Sampling was carried out at the bank of the river
Tisza within the administrative territory of Szeged
(between 168-173 river kilometers).
I chose 3 different sampling sites, one on the left
bank [L.I. (46°12'46.71"N, 20° 7'42.43"E)], and two
on the right bank of the river [R.II. (46°14'25.53"N,
20°8'57.53"E); R.III. (46°13'31.74"N, 20°8'23.20"E)],
each site was 20 m long.
Each sampling site differed from the others in
their characteristics. L.I. site was sunny, cover of
vegetation was low, and the inclination angle of the
riverbank was little. The R.II. site was shaded the
whole day, cover of vegetation was low, and the
inclination angle of the riverbank was high. The
R.III. site had sunny and always shaded parts too,
cover of vegetation was relatively high, the
inclination angle of the riverbank was medium
compared to the other two sites.
Sampling was performed between 6 May and 18
August in 2011 twice a week, usually in the third and
the fifth day.
I checked the bank of the river carefully twice
(the soil and the vegetation) in a 4-5 m zone from the
water-front. I recorded the emergence support and
the distance crawled from the water-front to the
emergence support, then I collected the exuviae with
tweezers and stored them in boxes in dry conditions.
To study the substrate-preference I determined 8
different support-types (artificial objects in the
watercourse, dead fallen leaves, exuviae, green
leaves, objects washed up by the river, roots, soil and
thin branches).
To study the mortality during the emergence, I
recorded data of individuals that were captured by
predators (wings near the exuviae indicate bird
predation), or died during the emergence due to other
reasons (e.g. abnormal moulting or abnormal wing
decompression). To determine total mortality, I also
paid attention to young adult dragonflies that were
damaged. These individuals would not live enough
to mate, they usually die shortly after emergence. I
did not count these imagines in the total number of
individuals, because in most cases I found their
exuviae next to them.
Processing of the exuviae took place at the
laboratory of the Department of Ecology of the
University of Szeged. Identification of the specimen
to species and gender level had been carried out with
a stereomicroscope. I used the keys and descriptions
of Askew (1988), Gerken and Sternberg (1999) and
Raab et al. (2006). To separate the sexes I used the
work of Berzi-Nagy (2011).
The water level and water and air temperature
data came from the on-line database
www.vizadat.hu, data of the measure station in
Szeged (173,6 river kilometer) were used as
background variables.
Statistical analysis
PAST (Hammer et al. 2001) and R (R
Development Team 2009) softwares were used to the
statistical analysis of the dataset.
To the comparison of the sex ratio of G.
vulgatissimus and G. flavipes, χ2 test was used.
To compare the distance crawled from the
waterfront to the place of emergence by G.
vulgatissimus and G. flavipes larve, Kruskal-Wallis
test was used.
The number of G. vulgatissimus larvae was so
low that if the data of this species were used by
ANOVA and linear regression the results would be
quite questionable, so during the following statistical
methods I used only the data of G. flavipes exuviae.
Linear regression was used to examine the
relationship between the amount of emerged G.
flavipes specimens and that of captured by birds.
TISCIA 39 11
The analysis of the substrate-preference of G.
flavipes was carried out with one-way ANOVA, and
Tukey-test was used to the pairwise comparisons.
Multiple linear regression was used to reveal the
connection between the distance crawled by G.
flavipes larvae and the background variables (water
level, water temperature and air temperature) and
between the number of the exuviae and the
background variables. The best model had been
chosen with Stepwise models election based on
Akaike information criterion (AIC).
Results
Abundance of species
During the examination period, 1217 exuviae
were found. Thirty two (2,6%) of them were G.
vulgatissimus, 1183 (97,2%) G. flavipes and 2
(0,2%) were Green Snaketail (Ophiogomphus
cecilia). The 1217 exuviae come from 3 study sites,
so the average number on a 20 meter long study site
is 406. In the case of G. vulgatissimus exuviae this
number is 11 and the G. flavipes is 394. In the case
of O. cecilia, there is no point talking about
population density, because of their low number.
In the case of G. vulgatissimus, there was no big
difference between the number of the individuals at
the three study sites. However, in the case of G.
flavipes, the number of the exuviae in the L.I. site
exceeded the combined number of the individuals of
the R.II. and R.III. (Fig. 1.).
Figure 1. Distribution of G. vulgatissimus (dark grey columns) and
G. flavipes (light grey columns) exuviae between the sampling
sites.
Sex ratio
From the 32 G. vulgatissimus exuviae there were
19 (59,4%) female and 13 (40,6%) male specimens. In
the case of the 1183 G. flavipes exuviae, there were
590 (49,8%) female and 591 (49,9%) male, while the
sex of 2 individuals were uncertain. There was no
significant difference between the ratio of sexes either
in case of G. vulgatissimus (χ2=0,0008; df=1; p=0,98)
or G. flavipes (χ2=1,125; df=1; p=0,29).
Pattern o f the emergen ce
G. vulgatissimus started to emerge on the 6th
May. The emergence of G. flavipes started on the
25th May.
The pattern of the emergence is fundametally
different between the two Gomphid species (Fig. 2.).
In the case of G. vulgatissimus the whole population
emerged within a month (19 days), the EM50 value
(the time needed for the 50% of the population to
emerge) is 4 days, the curve of the emergence is
steep – the species act as a ’spring species’. In the
case of G. flavipes, however, the emerging of the
total population took more than two months (72
days), the EM50 value is 17 days – the species act as
a ’summer species’.
If we examine the pattern of the emergence on
the basis how many exuviae had been found each
day, two peaks can be seen in the case of G. flavipes
(Fig. 3.).
Figure 2. The emergence curve of G. vulgatissimus and G. flavipes
at the investigated reach of the river Tisza at Szeged in 2011
(EM50 :the time needed for the 50% of the population to emerge;
▲–G. vulgatissimus ■–G. flavipes).
Figure 3. The emergence pattern of G. vulgatissimus and G.
flavipes at the investigated reach of the river Tisza at Szeged in
2011 (▲–G.vulgatissimus ■–G.flavipes).
12 TISCIA 39
Mortality of G. flavipes during the
eme rgence
Total mortality of G. flavipes during the
emergence period was 5,58%. According to the
literature (Farkas et al. 2011, 2012b) this is a normal
value by this abundance of the species. Larvae
consumed by predators (4,9%) make up the largest
proportion of the total value, and within predation
birds are liable for the most consumed larvae
(4,48%). These specimens can be easily
distinguished from those that were consumed by
other, unknown predators (0,42%) in most cases.
When birds eat the emerging dragonflies they leave
the uneatable wings of the insects behind, so if the
wings are found nearby the exuviae that refers to
bird predators. Linear regression shows that there is
a significant and positive correlation between the
amount of emerged dragonflies and mortality caused
by birds (β=0,52; F=15,4; df=1 and 12; p=0,002,
n=12).
The remaining proportion of mortality was
caused by abnormal moulting (0,34%) and the
abnormal decompression of the wings (0,34%).
Substrate perefe rence of G. flavipes
I used data of 763 specimen of G. flavipes
exuviae to the examination of substrate preference,
because the original support could be identified
without doubts by that many exuviae (in the other
420 case the exuviae were found lying on their back,
or due other reasons I could not identify the original
support). According to one-way ANOVA there is a
significant difference between the support choice
(F=7,832; df=7,16; p<0,0003), the majority of the
larvae chose soil as an emergence support (Table 1.).
Tukey’s pairwise comparison shows (Table A.1.)
that soil is significantly differ from any other
supports and between other support types there are
no significant difference in terms of preference.
Table 1. Substrate types chosen by G. flavipes larvae at the reach
of the river Tisza in Szeged.
support type
number of
individuals
percentage of
individuals
fallen leaves
14
1,83
exuviae
5
0,66
roots
6
0,79
artificial object
10
1,31
washed up objects
7
0,92
soil
687
90,04
branches
12
1,57
green leaves
22
2,88
Total
763
100
Distance crawled from waterfront, number
of exuviae and connecti on with
background variables
According to the distance data crawled by
larvae, there is a significant difference between G.
vulgatissimus and G. flavipes populations (Kruskal-
Wallis-test: H=13,35; Hc=13,35; p<0,0005). G.
vulgatissimus crawl greater distance
(horizontal+vertical) from the waterfront than the
other species, although there was no vertical
movement observed of G. vulgatissimus specimens.
(Figure 4. and Table 2.).
Water level, water temperature and air
temperature (investigated their effects individually)
have no significant effect on the distance crawled by
G. flavipes larvae (Linear regression: β=-0,24;
F=0,35; df=7,3; p=0,79; n=12), and they have no
significant effect if we assume interaction between
them either (Linear regression: β= 0,07; F=1,11; df=
7,3; p=0,51; n=12). None of the models were
supported by the AIC.
Nevertheless, marginally significant positive
connection was found between water temperature
and the number of exuviae (Linear regression:
β=0,52; F=14,1; df= 1,11; p=0,003; n=13).
Figure 4. Distance crawled by the larvae of G. flavipes and G.
vulgatissimus from the waterfront to the emergence substrate at
the reach of the river Tisza in Szeged.
Discussion
At the investigated reach of the river Tisza G.
vulgatissimus and G. flavipes seem to form stable
populations. Although there is a huge difference
between the abundance of the two species this
phenomenon seems to be normal along the river
Tisza and every other places where the two species
occur together. (Jakab and Dévai 2008). In 2011 at
Szeged the abundance of G. flavipes was 36 times
bigger than G. vulgatissimus, many authors inform
about a similar result, nevertheless, the differences in
TISCIA 39 13
abundance quoted by these papers are greatly
variable. According to Jakab (2006) at the reach
between Tiszafüred and Tiszacsege in 2001 the
abundance of G. flavipes was 8 times greater, during
the following years the abundance of G. flavipes was
much more greater than the abundance of G.
vulgatissimus: 11 times greater in 2002; 23 times
greater in 2003 and 26 times greater in 2012 (Farkas
et al. 2012a). At Vásárosnamény in 2008 the
abundance of G. flavipes was 2 times greater than the
abundance of G. vulgatissimus (Farkas et al. 2008).
As we can also see the differences increase toward
the south greatly, it could be possible that the
southern regions of the river Tisza can provide better
conditions for the populations of G. flavipes, as this
species, in his paper Berzi-Nagy (2011) made the
same conclusions.
In the case of O. cecilia it is quite sure that the
species has no stable population at the investigated
reach of Tisza. This species, as well as the Small
Pincertail (Onychogomphus forcipatus), the fourth
occurring Gomphid in Hungary, prefers small rivers
and streams with high oxygen level and moderate
flow (Raab et al. 2006). The two specimens might
have drifted from the river Maros, where they form
populations (Jakab and Dévai 2008).
The result of the investigation shows that in
2011 at Szeged there was no significant difference
between the ratio of sexes either in case of G.
vulgatissimus or G. flavipes. In the case of G.
flavipes the number of individuals in both sexes are
almost the same. Although, for the subgenus
Anisoptera it is general that the number of females is
higher than the number of males (Berzi-Nagy 2011;
Farkas et al. 2009; Jakab 2006) similar result may
occur (Jakab 2006). It is also an example that the sex
ratio of a certain species differ at the same reach of a
river between years (Corbet 1999 – cit. Farkas et al.
2009), so it is possible that next years the sex ratio of
the G. flavipes also will change.
In 2011 at Szeged the G. vulgatissimus acted as
a ’spring species’ (the species emerged strongly
synchronized within a short time) the G. flavipes as a
’summer species’. (the emergence was less
synchronized and stretched in time) This
phenomenon can be observed at other regions of the
river Tisza during the last decade as well. According
to Berzi-Nagy (2011) the emergence pattern of G.
flavipes showed the trait of a ’spring species’ at the
Middle-Tisza region near Szolnok. Jakab (2006)
reported the same phenomenon from that region, but
during his three year long investigation the pattern of
the emergence of G. flavipes varied between years,
too, and the differences were significant. Similarly to
sex ratio, the pattern of emergence can vary between
years, and also between different reaches of one
certain river. Variability could be caused by water
temperature: lower water temperature in winter and
higher one during summer caused more synchron-
ized emergence (Suhling 1995 – cit. Jakab 2006).
The emergence pattern of G. flavipes shows two
peaks, but this cannot be explained with weather
conditions, because these peaks do not coincide with
the highest air temperatures. So in this case, cohort
splitting seems to be the best explanation to the
emergence pattern as cohort splitting and unsuitable
weather conditions can cause a long-drawn
emergence period too. The reason of cohort splitting
is that females lay eggs during the entire emergence
period and some larvae winter in the final (F0)
larval, while some in the penultimate (F-1) larval
stage. Those that winter in F-1 stage will emerge a
few days or weeks later and they cause the second
peak in the emergence pattern.
The fact that more than 90% of the larvae chose
soil as emergence support does not necessarily mean
that there is a specific attachment to the soil as
substrate. 92% of the 763 G. flavipes larvae emerged
within 2 meters from the waterfront and 78% of
them within 1,5 metres. There is a possibility that the
over-representation of the soil has caused this
phenomenon, as during most of the emergence
period there were no – or was in very low proportion
of – other substrates in the first 1,5-2 meter zone
from the waterfront. Former studies (Farkas et al.
2009, 2011) claim that in the case of G. flavipes
larvae there is no substrate-specific attachment, but
they choose supports that are available within a
certain distance from the waterfront. This idea is
supported by the observation that larvae that chose
Table 2. The mean, standard deviation and maximum values of distances (horizontal, vertical and total) crawled by the larvae of G.
flavipes and G. vulgatissimus.
Species
Horizontal distance
Vertical distance
Total distance
N
Mean±SD
Max.
Mean±SD
Max.
Mean±SD
Max.
G.flavipes
104 ± 63
420
1 ± 6
72
105 ± 62
420
763
G.vulgatissimus
174 ± 114
506
0
0
174 ± 114
506
26
14 TISCIA 39
green leaves (second most frequently chosen
emergence support) did not crawl further than the
mean distance but most of them emerged in late June
and July, when the vegetation had grown in this zone
too (68% of them emerged within 2 metres; 68 %
emerged in July and 54% of them emerged within 2
metres in July).
During the emergence period in 2011 there were
66 G. flavipes exuviae, larvae or young imagines
found consumed by predators or wounded mortally
at the investigated reach of the river. This proportion
at this density is quite normal (Farkas et al. 2011,
2012a,b). Due to the emergence strategy of the
species [larvae emerge close to the waterfront and
the entire process takes 15-59 minutes, which is very
short compared to other Anisoptera taxa (Farkas et al
2012b)] the major factor for mortality is predators,
especially birds as common blackbird (Turdus
merula) and white wagtail (Motacilla alba) (personal
observation and Farkas et al. 2012a,b).
Results of the present study shows that G.
vulgatissimus larvae crawl greater distances from the
waterfront than G. flavipes larvae, as there is a strong
significant difference between the crawled distance
(from the waterfront to the emergence support) of the
two species. This seems to be general along the river
Tisza (Farkas et al. 2009, 2011, 2012a,b), and the
explanation is that G. vulgatissimus starts emerging
in late April or early May when greater fluctuations
of the water-level is possible, while in late May or
early June, when the G. flavipes starts the
emergence, there is a less chance of the fluctuation
of the water level (Farkas et al. 2012b).
The background variables could have a strong
effect to the emergence of the Gomphidae species:
Berzi-Nagy (2011) claims that the level and
temperature of water could influence the rate of
synchronization and the timing of emergence.
Moreover, according to former studies (Farkas et al.
2009) water level has a positive and water
temperature has a negative effect on the crawled
distance. In the case of this present study, the fact
that none of the background factors showed to effect
the distance, might be due to the low sample size.
Aknowledgement
My sincere thanks go to Judit Márton and Róbert
Gallé for the English language corrections and for
their indispensable professional guidance and useful
insights that helped me in data processing.
References
Askew, R. R. (1988): The Dragonflies of Europe. – Harley Books,
Colchester, 291 pp.
Bánkuti, K., Dévai, Gy., Miskolczi, M. (1997): Faunistical data on
dragonfly (Odonata) exuvia from the active floodplain of
river Tisza between Tiszabercel and Gávavencsellő (NE-
Hungary). – Studia odonatol. hung. 3:43-47.
Berzi-Nagy L. (2011): A sárgás szitakötő (Gomphus flavipes
flavipes Charpentier, 1825) jellemzése és Tisza-menti
populációinak összehasonlító elemzése (Characterization of
river clubtails (Gomphus flavipes flavipes Charpentier, 1825)
and comparative analysis of their populations along the river
Tisza). – Debreceni Egyetem, Hidrobiológiai Tanszék,
Debrecen, 50 pp.
Farkas, A., Jakab, T., Dévai, Gy. (2009): Assessment of riverine
dragonfly (Odonata: Gomphidae) on the basis of exuviae on
the reach of the river Tisza at Vásárosnamény. – Acta Biol.
Debr. Oecol. Hung. 20:65-78.
Farkas, A., Jakab, T., Dévai, Gy. (2011): Emergence behaviour of
riverine dragonfly (Odonata: Gomphidae) larvae along the
Tisza river system based on exuviae surveys. – Acta Biol.
Debr. Oecol. Hung. 26:53-66.
Farkas, A., Jakab, T., Dévai, Gy. (2012a): Emergence
characteristics of riverine dragonflies (Odonata: Gomphidae)
along the river Tisa between Tiszacsege and Tiszafüred. –
Acta Biol. Debr. Oecol. Hung. 27:39-50.
Farkas, A., Móra, A., Dévai, Gy. (2012b): Mortality during
emergence in Gomphus flavipes and G. vulgatissimus
(Odonata: Gomphidae) along the Danube. – Acta Biol. Debr.
Oecol. Hung. 28:65-82.
Gerken, B., Sternberg, K. (1999): Die Exuvien Europäischer
Libellen (Insecta Odonata) (Exuviae of european dragonflies
(Insecta Odonata)). – Arnika & Eisvogel, Höxter & Jena, VI
+ 354 pp.
Hammer, Ø., Harper, D. A. T., Ryan, P. D. (2001): PAST:
Paleontological Statistics Software Package for Education
and Data Analysis. – Paleontologica Electronica 4/1: 1-9 pp.
http://palaeo-electronica.org/2001_1/past/issue1_01.htm
Jakab T. (2006): A Tisza-tó és a Közép-Tisza szitakötő-
fajegyütteseinek (Insecta: Odonata) összehasonlító elemzése
(Comparative analysis of dragonfly assemblages (Insecta:
Odonata) of Tisza-lake and Middle-Tisza region). – Doktori
Értekezések 23. Debreceni Egyetem Kossuth Egyetemi
Kiadója, Debrecen, 131 pp.
Jakab, T., Dévai, Gy. (2008): The occurrence of the riverine
dragonfly-species (Odonata: Gomphidae) in Hungary
according to data of larvae and exuviae. – Acta Biol. Debr.
Oecol. Hung. 18: 53-65.
Raab, R., Chovanec, A., Pennerstorfer, J. (2006): Libellen
Österreichs (Dragonflies of Austria). – Springer, Wien/New
York. 355 pp.
R Development Core Team (2009): R: A language and
environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. URL
http://www.Rproject.org
TISCIA 39 15
Appendix
Table A.1
. Tukey’s pairwise comparisons for the substrate types (* marks the significant differences).
fallen leaves
exuviae
roots
artifical ob.
washed ob.
soil
branches
green leaves
fallen leaves
−
1
1
1
1
0,001*
1
1
exuviae
0,1053
−
1
1
1
0,0009*
1
1
roots
0,0936
0,0117
−
1
1
0,0009*
1
1
artifical ob.
0,0468
0,0585
0,0468
−
1
0,0009*
1
1
washed ob.
0,0819
0,0234
0,0117
0,0351
−
0,0009*
1
1
soil
7,877
7,982
7,97
7,923
7,959
−
0,001*
0,0011*
branches
0,0234
0,0819
0,0702
0,0234
0,0585
7,9
−
1
green leaves
0,0936
0,199
0,1873
0,1404
0,1756
7,783
0,117
−
