Estimations of the population size of Smooth Newts (Lissotriton vulgaris) breeding in a pond in Lincolnshire, England

2009·0.34 MB·English

by  Deeming, D C

Short Communications
SALAMANDRA 45 2 119-124 Rheinbach, 20 May 2009 ISSN 0036-3375
Estimations of the population size of Smooth Newts
(Lissotriton vulgaris) breeding in a pond
in Lincolnshire, England
Denis Charles Deeming
Abstract. The population size of smooth newts (Lissotriton vulgaris) in a small artificial pond was moni-
tored during the spring and summer of 2007 using netting and bottle traps. Individual animals were
identified from digital photographs of their ventral spot patterns. Although 20-25 newts were caught per
trapping, four methods for estimating population size each gave values around 205 newts. The Schnabel
method was considered to give the most precise population estimate. Trapping exhibited a clumped pat-
tern because newts appeared to be leaving the pond only to return later in the season. Methods that count
newts without identifying them run the risk of severely underestimating population sizes, which could
impact on decisions about habitat protection and the survival of local populations.
Key words. Smooth newts, Lissotriton vulgaris, population estimate, clumped trapping pattern.
A pond at Riseholme Park campus of the ings, greenhouses and a stream. The nearest
University of Lincoln is home to smooth other pond to the study pond is around 200
newts and as part of developing skills in m to the south-west and beyond the stream,
practical conservation work a long-term and there is a large lake on campus – smooth
study of the population size was established newts are known to be present in the up-
in the spring of 2007. Although originally in- per part of this lake at a location over 500 m
tended as a means of introducing students to away to the west and separated by a complex
trapping techniques it became clear that the of buildings and roads.
population size of the pond was more than Smooth newts (Lissotriton [Triturus] vul-
was originally expected. Moreover, the re- garis) were caught on a regular basis (approx-
capture patterns of these newts appeared not imately 8-day intervals) from mid-March to
to be random. This short report outlines the mid-July 2007. Two methods were used: the
outcome of this study providing estimates of first involved visiting the pond at dusk and
population size, using a variety of methods, searching the circumference of the pond
and highlights the point that the observed with torches for adult newts, which were net-
number of newts in a pond may not reflect ted and transferred to a water-filled bucket
the population of animals using the habitat. where they were held overnight. The net-
The study pond is set within a horticultur- ting event was between 25-30 minutes and
al unit on Riseholme Park, Lincoln (53º 6’ N, involved on average four people. The second
0º 32’ E). It is oval with a circumference of 5 method involved using ten bottle traps (2 L
m, a surface area of around 8 m2 and has an capacity; see Griffiths & Langton 2003)
approximate maximum depth of  m. Lined set in, and evenly around, the margins of the
with PVC the pond has a well-established pond in the evening and left overnight. The
mixed marginal and aquatic flora. The imme- traps were examined no later than 09.00 h
diate vicinity includes a tarmac road, mown the next morning and any adult newts caught
grass, and flowerbeds. Within a radius of 00 were then examined. Bottle traps were used
m of the pond there are also several build- during the last four weeks of the trapping pe-
© 2009 Deutsche Gesellschaft für Herpetologie und Terrarienkunde e.V. (DGHT)
http://www.salamandra-journal.com 119
Short Communications
riod as the numbers of newts caught began 50
to decline and dusk occurred later. Juvenile 45
newts lack ventral spots and so could not be 40
individually identified. For this reason any 35
jnuovte inniclelus dceadu ginh tt hwe esraem npolete cdo, ubnut.t they were mber2350
Nu
To record ventral spot patterns a newt was 20
lifted out of the water in the palm of the hand 15
and allowed to crawl into a clear plastic 30 10
ml Universal tube and the top was loosely se- 5
cured. This arrangement allowed the newt to 0
70 90 110 130 150 170 190 210
be held securely but without excessive han-
Julian Day
dling and contact with hands (minimising
Fig. 1. Number of newts caught per trapping event,
rapid increases in body temperature or dehy-
plotted as Julian day, in the spring and summer
dration). When held horizontally above the
of 2007.
head the tube allowed digital photographs to
be taken of the ventral surface of the newt.
If at all possible the photograph would show
the animal lying along the length of the tube
and both the throat and belly spots would be
visible. Preview of photographs ensured that
a good view of the ventral surface was re-
corded before the sex of the newt was noted ncy
ue
and it was released. All newts caught in any Freq
trapping event were released into the water
at the same point in the pond although this
varied between trapping events. It was con-
sidered that the small size of the pond did not
prevent any newt from moving to any loca-
tion within the habitat.
The photographs of the ventral skin pat-
Days between capture and recapture
terns of black spots were examined to deter-
mine whether the individual animal had been Fig. 2. Frequency distribution histogram of the
previously captured or was a new individual. days recorded for individual smooth newts bet-
ween their initial capture and identification and
The pattern of spots was readily identifiable to
their subsequent first recapture.
individuals particularly if both the belly and
throat patterns were compared. Identifica- on the belly did vary through time, often fad-
tion of individuals was done by eye with ref- ing as the season progressed, but intensity of
erence to printed copies of photographs and spots on the throats of male newts showed
all new animals were given a unique number. little sign of fading.
A record of each capture of the different indi- Initial analysis determined whether the
viduals was made for each capture event. population was open or closed using the
Whilst the spot patterns of male newts method outlined by Greenwood & Robin-
were more obvious, visual recognition of fe- son (2006). The proportion marked in each
male newts did not pose a problem. Identifi- sample showed a significant positive cor-
cation was generally easier if pattern of spots relation with successive trapping occasions
were compared on both the belly and throat (Spearman’s rank correlation coefficient, r =
s
of the individuals. Intensity of black pigment 0.900, P < 0.00) and so the population was
120
Short Communications

considered to be effectively closed. There was 300 Estimated population GCP
no significant trend in the numbers caught Total caught GCP
per trapping event i.e. Julian day (r= –0.363, 250
s
P > 0.05) although testing for the total num-
bers of marked newts caught per trapping did mate200
ds8h2u.oew, tP oa < th h0ie.g0 hm0lyo) .vs Thiegmnieisfi nlcatat toneftr neeffffeweeccttts wotofa stai pmnrdoe bf(raGob ml=y Population esti110500
the pond over the course of the trapping pe-
riod – there were fewer newts caught at the
50
start and end of the trapping period (Fig. ),
which reflects changes in distribution of the
0
newts between terrestrial and aquatic habi- 70 90 110 130 150 170 190 210
tats (see Griffiths 984). The two trapping Julian day
techniques may have also contributed to this Fig. 3. Population estimate for the smooth newt
result because bottle-trapping may be less ef- population of GCP (Garden Centre Pond) using
ficient than netting newts but this needs to be the two-sample method plotted against Julian day
investigated. in 2007 together with the cumulative total for
Population size was estimated after each newts captured and identified.
capture event using the two-sample method
of Greenwood & Robinson (2006): newts caught twice, 26 caught three times, 4
Population estimate = {([number caught + newts were caught four times, 6 were caught
] x [total number caught previously + ]) / five times, 3 were caught six times and 2 newts
[number of individuals recaptured+ ]} – . were caught seven times.
Three additional models (Greenwood Except for the first and last values the pop-
& Robinson 2006) were tested for estimat- ulation estimates (Fig. 3) derived from the
ing the population size for newts using the two-sample method were comparable aver-
pond. The pseudo-removal method allows aging 203 newts (SE = 6 and 95% confidence
for behavioural responses to being trapped, limits for the mean of ± 2 newts). Calcula-
the Schnabel method allows for temporal dif- tion of 95% CI for the population estimate
ferences between capture occasions, and the for each trapping showed that these values
Burnham and Overton method allows for were imprecise and another technique was
heterogeneity in catchability amongst indi- required to estimate the newt population.
viduals. However, the total number of newts caught
Over the four-month period (mid-March over time did approach the estimated popu-
to mid-July) 72 individual newts were lation of ~200 (Fig. 3).
identified with more males (N = 0) being Each of the three models yielded com-
caught than females (N = 62). Eighty-six of parable population estimates although their
the identified newts (50%) were recaptured at levels of precision were different (Table ).
least once and the time between initial cap- Given the lower confidence intervals it was
ture and the first capture varied 5 to 07 days judged that the Schnabel method provided
with a median of 27 days (Fig. 2). More male the most precise method for estimating pop-
newts were recaptured (6 versus 25 females), ulation size in this situation.
but they showed no significant difference in A dispersion index was calculated for
the time between capture and first recapture each capture event following the method of
(medians = 27 days for both genders, Mann Fowler et al. (998) to test whether trapping
Whitney test, P > 0.05). Median number of of individuals was random. The number of
captures per newt was 2 with thirty-eight newts for the first trapping was noted. For
121
Short Communications
Tab. 1. Estimates and 95% confidence intervals for population size for Lissotriton vulgaris utilising the
Riseholme pond during the breeding season. Values are based on three different methods (Greenwood
& Robinson 2006).
Population estimate 95% confidence intervals
Pseudo-removal method 204 73 - 25
Schnabel method 202 90 - 25
Burnham & Overton method 20 68 - 252
the second trapping the number of newts other systems of photographing belly spots
caught previously was recorded together (see Griffiths & Langton 2003). For male
with the number of new newts. At the third smooth newts in particular, patterns of throat
trapping the number of newts trapped that spots proved more consistent over time than
were caught the first week were noted, as those of belly spots, which changed in inten-
were those caught during the second trap- sity (but not pattern), probably according to
ping together with the new individuals. This breeding condition.
method continued for each trapping event so Seasonal change in newts captured in
that mean and variance values for the vari- the pond were comparable to that shown by
ous trapping cohorts could be calculated. Griffiths (984). This small study showed
The dispersion index was calculated for each that the population of newts dependent on
trapping event by dividing the mean by the the pond was much higher than would be
variance and χ2 was calculated by multiply- considered if individual trapping events were
ing the value by the degrees of freedom (n–). considered. Four different methods showed
Therefore, for example, for trap event five a that the newt population of the pond was
total of 23 newts were caught: none were re- just over 200 newts. The consistency of esti-
captured from the first capture, 2 were recap- mates showed that this relatively small pond
tured from the second capture, 5 were recap- was of local significance to this population
tured from the third capture, none from the of smooth newts. However, casual observa-
previous capture and 6 new newts were cap- tions would suggest that the newt population
tured. Mean newt number per capture event was rather limited; for any single trapping
was 4.6 and the variance was 44.6, χ2 = 39.0 event only around 20 individuals were typi-
for 4 degrees of freedom. Chi-squared values cally caught, but this is only 0% of the esti-
were then plotted against degrees of freedom mated population. Such a discrepancy could
to determine whether the values fell relative lead to a significant misinterpretation of the
to the 95% confidence zone of random disper- importance of the pond for smooth newts. In
sal (Fowler et al. 998). For two-thirds of the this instance, a professional ecological survey
trapping events the numbers caught were not of this same pond in 2004 only observed -2
randomly distributed between the weeks of smooth newts. If this result is typical of simi-
capture but showed clumped dispersal (Fig. lar ponds for this and other newt species then
4). Only towards the end of the trapping sea- the implication for conservation of habitat is
son was the capture of newts more random important. Simple counts of newts by torch-
according to previous trapping events. ing, netting or trapping could suggest that
The technique of identifying individu- the population dependent on the habitat is
als from digital photographs of their spot- much lower than actually uses the pond over
ting patterns worked well in this study. Al- the course of a breeding season. Underesti-
though potentially time consuming for large mation of a population could impact upon
samples, the method of holding the newt assessment of habitat quality and the need
in a plastic tube was seen to be practically for its protection of various breeding sites
useful and also more welfare friendly than (see Beebee & Grayson 2003).
122
Short Communications
turing novel newts decreased as the season
Clumped progressed even though trapping techniques
remained constant. This suggests that the
animals are leaving the pond for terrestrial
habitats for extended periods of time over
the breeding season (but are not apparently
q
hi-s migrating away to other ponds), similar to
C smooth newts in German ponds (Wedde-
ling et al. 2004), and so are not present to
be caught within the aquatic habitat. In this
Random
study it was not possible to determine how far
the newts travelled when they left the Rise-
Regular
holme pond and on-going investigations will
ascertain the distances that the newts travel
DF and where they seek refuge during the day.
Fig. 4. Chi-squared values calculated from disper- In conclusion, this small study of a pond
sion indices per trapping event (equal to DF +1) used by smooth newts in Lincolnshire sug-
plotted against degrees of freedom (DF). Points gests that the population is closed, but indi-
lying in the area of the graph above the grey area
vidual newts do not occupy the pond contin-
exhibit clumped dispersion and points within the
uously over the breeding season. Individual
grey area exhibit random dispersion.
trapping events would have severely under-
estimated the population size of newts using
this small pond. Different methods for esti-
That the pond was a habitat for a closed
mating population size varied in precision
population of newts means that its degrada-
but were consistent in their population esti-
tion or loss would severely impact on this
mates. Care should be taken in estimating the
population of animals. The nearest pond may
size of newt populations and photographing
provide an alternative habitat, but it is over a
ventral spot patterns allows for relatively easy
stream and, albeit quiet, road and is small-
identification of individuals that would pro-
er than the original pond. Baker & Hal-
vide realistic estimates of population sizes.
liday (999) showed that in an agricultural
landscape, which matches that of Riseholme
and its environs, the maximum distance be-
Acknowledgements
tween ponds travelled by smooth newts was
400 m. The upper part of the lake at Risehol- Thanks go to the various people who assisted in
me is at least 500 m distant and there would the trapping and recording of the newts in this
be a significant level of human infrastruc- study. These include: K. Deeming, E. Deeming
ture to negotiate for newts to move to this and R. Deeming, R. Blackman, J. Glossop, H.
other habitat. Whilst smooth newts may not Hind, S. Bennett, D. Suty, C. Barnard and S.
Brown. Many thanks to R. Griffiths for initial
be a threatened species the loss of this pond
advice, and to P. Eady and G. Baggott for their
would mean a significant loss of their habitat
interest and critical comments on a previous ver-
and the ecosystem that it supports.
sion of this manuscript.
The newts in this study were not trapped
in a random or uniform manner and it seems
that they are utilising the aquatic habitat as References
cohorts. One possibilty to explain this is that
trapping or netting of newts was inefficient Baker, J.M.R., & T.R. Halliday (999): Amphibi-
at recapturing newts at the start of the trap- an colonization of new ponds in an agricultural
ping season. However, the likelihood of cap- landscape. – Herpetological Journal, 9: 55-63.
123
Short Communications
Beebee, T., & R. Grayson (2003): Site assessment Griffiths, R.A., & T. Langton (2003): Catching
and protection. – pp. 95-06 in Gent, T. & S. and handling. – pp. 33-43 in Gent, T. & S. Gib-
Gibson (eds.): Herpetofauna Workers’ Hand- son (eds.): Herpetofauna Workers’ Handbook.
book. – Joint Nature Conservation Committee, – Joint Nature Conservation Committee, Pe-
Peterborough. terborough.
Fowler, J., L. Cohen, & P. Jarvis (998). – pp. Greenwood, J.J.D., & R.A. Robinson (2006):
63-66. Practical Statistics for Field Biology, 2nd General census methods. – pp. 87-85 in Suth-
Edition. – John Wiley & Sons, Chichester. erland, W.J. (ed.): Ecological Census Tech-
Griffiths, R.A. (984): Seasonal behaviour and niques, 2nd Edition. – Cambridge University
intrahabitat movements in an urban popula- Press, Cambridge.
tion of smooth newts, Triturus vulgaris (Am- Weddeling, K., M. Hachtel, U. Sander, & D.
phibia: Salamandridae). – Journal of Zoology, Tarkhnishvili (2004): Bias in estimation of
London, 203: 24-25. newt population size: a field study at five ponds
using drift fences, pitfalls and funnel traps. –
Herpetological Journal, 4: -7.
Manuscript received: 18 December 2007
Author’s address: Denis Charles Deeming, Department of Biological Sciences, University of Lincoln,
Riseholme Park, Lincoln, LN2 2LG, United Kingdom, E-Mail: [email protected]
124