Meta-Analysis of the Effectiveness of Marked Wire in Reducing Avian Collisions with Power Lines

Abstract

Collisions of birds with power transmission and distribution lines have been documented for many species, and cause millions of casualties worldwide. Attempts to reduce mortality from such collisions include placing bird flight diverters (i.e., wire markers in the form of, e.g., spirals, swivels, plates, or spheres) on static and some electrified wires to increase their visibility. Although studies of the effectiveness of such devices have yielded contradictory results, the implementation of flight diverters is increasing rapidly. We reviewed the results of studies in which transmission or distribution wires were marked and conducted a meta-analysis to examine the effectiveness of flight diverters in reducing bird mortality. We included in our meta-analysis all studies in which researchers searched for carcasses of birds killed by a collision with wires. In those studies that also included data on flight frequency, we examined 8 covariates of effectiveness: source of data, study design, alternate design (if marked and unmarked spans were alternated in the same line), periodicity of searches for carcasses, width of the search transect, and number of species, lines, and stretches of wire searched. The presence of flight diverters was associated with a decrease in bird collisions. At unmarked lines, there were 0.21 deaths/1000 birds (n =339,830) that flew among lines or over lines. At marked lines, the mortality rate was 78% lower (n =1,060,746). Only the number of species studied had a significant influence on effect size; this was larger in studies that addressed more species. When comparing mortality at marked and unmarked lines, we recommend use of the same time intervals and habitats and standardizing the periodicity of carcass searches.

Full text

Review
Meta-Analysis of the Effectiveness of Marked Wire
in Reducing Avian Collisions with Power Lines
RAFAEL BARRIENTOS,∗† JUAN CARLOS ALONSO,∗CARLOS PONCE,∗AND CARLOS PALAC´
IN∗
∗Departamento de Ecolog´
ıa Evolutiva, Museo Nacional de Ciencias Naturales (CSIC), Jos´
e Guti´
errez Abascal, 2,
E-28006 Madrid, Spain
Abstract: Collisions of birds with power transmission and distribution lines have been documented for
many species, and cause millions of casualties worldwide. Attempts to reduce mortality from such collisions
include placing bird flight diverters (i.e., wire markers in the form of, e.g., spirals, swivels, plates, or spheres)
on static and some electrified wires to increase their visibility. Although studies of the effectiveness of such
devices have yielded contradictory results, the implementation of flight diverters is increasing rapidly. We
reviewed the results of studies in which transmission or distribution wires were marked and conducted a
meta-analysis to examine the effectiveness of flight diverters in reducing bird mortality. We included in our
meta-analysis all studies in which researchers searched for carcasses of birds killed by a collision with wires.
In those studies that also included data on flight frequency, we examined 8 covariates of effectiveness: source
of data, study design, alternate design (if marked and unmarked spans were alternated in the same line),
periodicity of searches for carcasses, width of the search transect, and number of species, lines, and stretches of
wire searched. The presence of flight diverters was associated with a decrease in bird collisions. At unmarked
lines, there were 0.21 deaths/1000 birds ( n=339,830) that flew among lines or over lines. At marked lines,
the mortality rate was 78% lower ( n=1,060,746). Only the number of species studied had a significant
influence on effect size; this was larger in studies that addressed more species. When comparing mortality at
marked and unmarked lines, we recommend use of the same time intervals and habitats and standardizing
the periodicity of carcass searches.
Keywords: bird collision, bird flight diverter, flight frequency, ground-wire marking, power line
Meta-An´
alisis sobre la Eficacia de la Se˜
nalizaci´
on de los Cables para Reducir las Colisiones de Aves contra
Tendidos El´
ectricos
Resumen: La colisi´
on de aves con tendidos el´
ectricos tanto de transmisi´
on como de distribuci´
on ha
sido documentada en numerosas especies y causa millones de muertes en todo el mundo. Los intentos
para reducir la mortalidad causada por dichas colisiones incluyen la colocaci´
on de dispositivos anticolisi´
on
(i. e., marcadores en cables con forma de espiral, dispositivos giratorios, platillos o esferas) en los cables
de tierra, as´
ı como, a veces, en los conductores, para aumentar su visibilidad. Aunque los estudios llevados
a cabo sobre la efectividad de tales medidas han llegado a conclusiones contradictorias, la instalaci´
on de
dispositivos anticolisi´
on est´
a aumentando r´
apidamente. Revisamos los resultados de los estudios en los que
se se˜
nalizaron cables de transmisi´
on o de distribuci´
on y llevamos a cabo un metan´
alisis para examinar la
eficacia de los dispositivos anticolisi´
on a la hora de reducir la mortalidad. Incluimos en nuestro metan´
alisis
todos los trabajos en los que los investigadores realizaron una b´
usqueda de aves muertas tras colisionar
con los cables. En aquellos estudios que adem´
as incluyeron frecuencias de vuelo, examinamos 8 covariables
de la efectividad: origen de los datos, dise˜
no del estudio, dise˜
no alternado (si los vanos se˜
nalizados y no
se˜
nalizados se alternaban en el mismo tendido), periodicidad en la b´
usqueda de cad´
averes, ancho de la
banda de b´
usqueda, y n´
umero de especies, tendidos y tramos muestreados. La instalaci´
on de dispositivos
anticolisi´
on estuvo ligada a un descenso en el n´
umero de aves colisionadas. En los tendidos sin se˜
nalizar,
†Current address: ´
Area de Zoolog´
ıa, Departamento de Ciencias Ambientales, Facultad de Ciencias del Medio Ambiente, Universidad de Castilla-La
Mancha, Avenida Carlos III, s/n, E-45071 Toledo, Spain, email rafael.barrientos@uclm.es
Paper submitted November 3, 2010; revised manuscript accepted February 21, 2011.
893
Conservation Biology, Volume 25, No. 5, 893–903
C
2011 Society for Conservation Biology
DOI: 10.1111/j.1523-1739.2011.01699.x

894 Wire Marking to Reduce Bird Collisions
hubo 0.21 muertes/1,000 aves (n=339,830) que cruzaron los cables. En los tendidos marcados, la mortalidad
fue un 78% inferior (n=1,060,746). S´
olo el n´
umero de especies estudiadas tuvo una influencia significativa
en el tama˜
no del efecto; ´
este fue mayor en aquellos trabajos que estudiaron m´
as especies. Cuando se compare
la mortalidad en tendidos se˜
nalizados y sin se˜
nalizar, recomendamos que se usen los mismos intervalos de
tiempo y h´
abitats y que se estandarice la periodicidad de la b´
usqueda de cad´
averes.
Palabras Clave: colisi´
on de aves, dispositivo anticolisi´
on, frecuencia de vuelo, se˜
nalizaci´
on del cable de tierra,
tendido el´
ectrico
Introduction
Avian collisions with and electrocution by power lines
have been documented since the early 1900s, but it was
not until the 1970s that biologists and engineers began to
realize the extent of these events and to study mitigation
measures (e.g., Bevanger 1998; APLIC 2006; Lehman et al.
2007). The number of power lines is increasing world-
wide at 5% per year (Jenkins et al. 2010). This percentage
applies to both power distribution (generally 2.4 kV to
60 kV) and transmission lines, which carry >69 kV of
electricity (APLIC 2006).
Bird mortality from collisions with power lines and
other electric-utility structures has been documented
for nearly 350 species of birds (Manville 1999). Some
crude estimates of the number of individuals that die
are also available. For instance, bird collisions with
power lines may cause 1 million deaths/year in the
Netherlands (Koops 1994), and in the United States esti-
mates show power lines kill from hundreds to thousands
to >175 million birds per year (Manville 2005, 2009).
Worldwide it is estimated that bird collisions with power
structures, including transmission and distribution lines,
that result in fatalities could approach 1 billion annually
(Hunting 2002).
Until an assessment of the cumulative effects of bird
mortality from power lines is conducted, the magnitude
of such mortalities will remain uncertain (Manville 2009).
Although collisions with power lines are the most im-
portant mortality source for some endangered species
of birds (Manville 2009), few detailed analyses of how
these losses affect trends in population size have been
conducted. Collision-related losses may be equivalent to
9–90%, depending on the species, of the annual number
of individual tetraonids (grouse) harvested by hunting
in Norway (Bevanger 1995). Whereas estimated hunting
harvest of Capercaillies (Tetrao urogallus) was 22,200,
estimated mortality from collisions with power lines was
19,900. In Switzerland ring-recovery data show 25% of
juvenile and 6% of adult White Storks (Ciconia ciconia)
die annually due to collision with and electrocution by
power lines (Schaub & Pradel 2004). Shaw (2009) esti-
mated that in South Africa 30% of Denham’s Bustard (Neo-
tis denhami) are killed annually by collisions with power
lines. Birds with low maneuverability, that is, those
with high wing loading and low aspect (e.g., bustards,
cranes, storks, pelicans, waterfowl, some grouses), are
among the species most likely to collide with power lines
(Bevanger 1998; Janss 2000). Species with narrow visual
fields also have a high probability of colliding with power
lines (Martin & Shaw 2010).
Although efforts to reduce bird collisions are increas-
ing rapidly worldwide, the effectiveness of such measures
has not yet been tested adequately. Results from exami-
nations of the effectiveness of anticollision systems are di-
verse, varying from no reduction of collisions (e.g., Scott
et al. 1972; Janss et al. 1999; Anderson 2002) to a reduc-
tion in collisions (e.g., Alonso et al. 1994; Bevanger &
Brøseth 2001). This heterogeneity may be due to differ-
ences in behavior and morphology of species, habitat
variability, weather, type and number of marking devices
used per length of line, and approaches used to test for
an effect.
The mitigation measures used include placement of
raptor decoys on posts (Janss et al. 1999), marking
static wire to make it more visible, and replacement of
static wire with lighting arrestors at transmission tow-
ers (Beaulaurier 1981; Bevanger & Brøseth 2001). Where
collisions of birds with energized transmission lines are a
problem, lines are sometimes marked with clamp-on de-
vices. However, these devices, which surround the wire
where they are attached, can cause power reductions
and line damage and thus may not be feasible for high-
voltage wires (APLIC 2006). Removing the static wire
would reduce bird mortality, but the wires are needed
to protect conductors from lightning (Beaulaurier 1981).
Because lightning strikes can result in power outages and
line damage and fires, the most common mitigation mea-
sure has been the attachment of spirals, plates, swivels,
or spheres to static wire to increase wire visibility. Collec-
tively, these devices are called bird flight diverters (e.g.,
APLIC 1994; Hebert & Reese 1995; Jenkins et al. 2010).
Despite the general belief that bird flight diverters reduce
bird mortality, results of several studies show they do not
(e.g., Scott et al. 1972; Anderson 2002). However, sev-
eral of these studies had small sample sizes or did not
include statistical tests. Placement of bird flight divert-
ers is expensive (e.g., US$1100–2600/km of marking in
South Africa [Kruger 2001] and €6000 in Spain [Alonso
et al. 2005]), so evidence of their effectiveness is needed.
Narrative reviews of the effectiveness of wire mark-
ing to reduce avian collisions with power lines have
been conducted (APLIC 1994; Bevanger 1994; Jenkins
et al. 2010), mainly through counts of the number of
Conservation Biology
Volume 25, No. 5, 2011

Barrientos et al. 895
studies indicating markers do or do not reduce collisions.
These qualitative reviews do not control for sample size
or variance across studies. They give equal value to pub-
lications with anecdotal data and to those with detailed
experimental designs and large sample sizes and small
variances. Not controlling for sample size can lead to
type II errors (Arnqvist & Wooster 1995), that is under-
estimation of the effects of collisions on population sizes
of birds (Fern´
andez-Duque & Valeggia 1994).
We conducted a meta-analysis of the published
literature and unpublished reports (primarily reports of
private companies) to evaluate whether wire marking
reduces the number of bird collisions with power lines.
A meta-analysis is quantitative and allows for comparison
of results among studies. Meta-analysis weights the value
of different studies on the basis of their sample sizes
and variances and provides a balanced effect for the
studied topic (Arnqvist & Wooster 1995; Gurevitch &
Hedges 2001; Stewart 2010). Meta-analyses have been
used widely in research domains in which available
empirical data provide no clear consensus (Stewart et al.
2005, 2007; Ben´
ıtez-L´
opez et al. 2010). Meta-analyses are
especially valuable when there is a high probability of
incurring type II errors (Arnqvist & Wooster 1995).
Methods
Data
Our meta-analyses included studies that reported on
counts of carcasses associated with marked and un-
marked power lines. We did not include studies that
provided only data on the behavior of birds when ap-
proaching the power lines or only data from marked or
unmarked sections of power lines in space or time. We
did not use data on mortality that was estimated after
correcting for potential biases (e.g., scavenger removal,
estimates of injury, habitat, or observers) because not
all studies correct for such biases. Furthermore, calcula-
tion of mortality after correcting for biases was beyond
the scope of our analyses. Instead, we used the raw data
from carcass counts.
We conducted 2 meta-analyses, the first with data from
all studies that reported carcass counts and the second
with studies that also included counts of birds flying
across the line that were used to calculate collision rates.
In the second meta-analysis, we evaluated the sensitiv-
ity of the results from the first by using data on flight
frequency (Stewart et al. [2007]; Ben´
ıtez-L´
opez et al.
[2010]).
We searched ISI Web of Science, Scirus, Zoological
Record, and JSTOR for wire-marking studies. We identi-
fied additional studies in the reference lists of pertinent
papers we found in the databases. The number of pub-
lished papers listed on the 4 databases was relatively low,
which may reflect the small sample sizes typical of such
studies. Thus, we searched Google for additional studies
not published in peer-reviewed journals.
For all searches, our search terms were combinations
of the following words or phrases: bird, crane, swan,
raptor, waterfowl, aviation ball, flight diverter, swan di-
verter, ground wire, static wire, marker, power line,
spiral, wire, collision, effectiveness, impact, power-line
marking, and wire marking. We searched for publications
in English, Spanish, German, and French. We contacted
most authors who have worked on this topic over the last
30 years. These authors were our most fruitful source of
data because they provided use with other contacts and
with several unpublished studies that would otherwise
have been inaccessible. We also contacted environmen-
tal departments of electrical companies, managers of state
and federal wildlife agencies, and nongovernmental con-
servation organizations worldwide to obtain unpublished
documents, such as PhD dissertations and public and in-
ternal reports, to increase the number of studies we could
use (Fern´
andez-Duque & Valeggia 1994).
We expected this variety of sources would reduce the
probability of biasing the meta-analysis toward studies
reporting statistically significant results, which are be-
lieved to be published more frequently than those with
results that are not statistically significant (Arnqvist &
Wooster 1995; Stewart 2010). We reveal our sources here
so as to avoid the hidden-publication bias that may be
present in other nonsystematic syntheses (Stewart 2010).
However, we also formally tested potential publication
biases.
If more than one publication presented results from
the same study area and period (e.g., Crowder [2000],
Crowder & Rhodes [2001], Shaw [2009], and Shaw et al.
[2010]), we relied on data from the most complete study.
We extracted the raw data from each study in our meta-
analyses. Thus, if the same results were published in an
abbreviated form (typically, a paper) and in a more de-
tailed form (for instance, in a report or dissertation), we
used the latter because we could obtain the raw data
more easily. When a publication included more than
one study, for instance if it assessed the number of colli-
sions associated with more than one marker in different
line segments or lines (e.g., Anderson 2002) or the same
marker at different intervals in different line segments or
lines (e.g., Koops & de Jong 1982), we treated the experi-
ments as independent (Ben´
ıtez-L´
opez et al. 2010; Gilbert-
Norton et al. 2010). We also treated as independent
studies that tested marker effectiveness with a before-
after-control-impact (BACI) design and with a parallel de-
sign (i.e., marked and unmarked lines studied during the
same time interval). We excluded studies that simulta-
neously tested more than one marker (e.g., bird flight
diverters and strips in the same wire) because their ef-
fects could be cumulative and could have stronger effects
than those experiments in which only a single device was
used.
Conservation Biology
Volume 25, No. 5, 2011

896 Wire Marking to Reduce Bird Collisions
Studies of Carcasses
Because our primary data were the means of 2 groups,
we calculated the ratio of the means to obtain the
“response ratio” (hereafter R) as the effect size
(Borenstein et al. 2009). Following Bevanger (1999), we
controlled the number of avian deaths per power line
length and period of time as
R=MML/MUML =[cML/(kmML ×tML)]
/[cUML/(kmUML ×tUML)],(1)
where MML is the mortality associated with the marked
line, MUML is the mortality associated with the un-
marked line, cML is the number of carcasses found under
the marked line, kmML is the length in kilometers of
the marked line, tML is the number of months during
which carcasses were counted under the marked line,
and cUML, kmUML, and tUML are the respective values
for the unmarked line. We analyzed Rafter log trans-
formation to maintain symmetry in the analysis (Hedges
et al. 1999; Borenstein et al. 2009). Negative lnRvalues
(i.e., ln R<0) indicated a decrease in mortality, and
positive values indicated an increase. The response ratio
is a common metric in meta-analyses of ecological stud-
ies (Hedges et al. 1999). However, because authors of
the studies we analyzed reported only the total number
of dead birds per kilometer and month, the sample size
for every study was 1.0 (J. Gurevitch, personal commu-
nication). Thus, we conducted an unweighted analysis
(i.e., all the variances were 1.0) (Rosenberg et al. 2000;
J. Gurevitch, personal communication). An unweighted
analysis does not allow one to investigate the potential
structure of the data because the weight from every study
is required (Neter et al. 1989).
Carcass Counts and Flight Frequencies
We examined the sensitivity of the results from the first
analysis by conducting a second meta-analysis that in-
cluded only those studies in which both carcasses and
flight frequencies were counted (i.e., number of birds
flying across studied power lines, or sample size). We as-
sessed the difference between probabilities of collision
(i.e., the risk difference [RD]) associated with unmarked
and marked lines. The probability of collision associated
with either line was mortality divided by the total number
of birds crossing the line (Borenstein et al. 2009).
RD =(MML/nML) −(MUML/nUML),(2)
where nML is the number of birds flying across the
marked line and nUML is the number of birds crossing
the unmarked one.
The sample size we used to calculate the weight for ev-
ery study was the total number of birds observed flying
across the studied power lines. Mortality could have been
overestimated in some cases because studies counted
bird crossings during periods that were shorter than
those in which carcasses were counted. Furthermore,
sampling efforts were not always the same in marked
and unmarked sections because the sections were not
always the same length or sampled during the same time
intervals. Thus, we controlled for sampling effort before
calculating nML and nUML. We divided the sample size of
each study by the sampling effort we assessed, ignoring
the reported number of birds crossing marked and un-
marked sections (unless sampling effort was equal). In 4
studies (Brown & Drewien 1995; Crowder 2000; Brauneis
et al. 2003; Lorenzo & Cabrera 2009) that evaluated more
than one device, but did not clearly indicate the sections
of marked lines where flying birds were counted, we av-
eraged the numbers of birds crossing for the different
markers. We did not use data from studies that assumed
the same crossing rates for stretches with and without
markers because this assumption is incorrect (Alonso
et al. 1994; Calabuig & Ferrer 2009). Negative RD val-
ues indicated a decrease in mortality and positive values
indicated an increase. We calculated the variance of RD
with the following formula (Borenstein et al. 2009):
VRD =(MML ×AML)/nML3+(MUML ×AUML)/nUML3,
(3)
where AML and AUML are birds that were alive after cross-
ing marked and unmarked lines, respectively. In applying
this sensitivity analysis we had to discard some studies,
but we improved statistical power because we could use
weighted analyses.
Data Analyses
We loaded effect sizes and variances to carry out all
analyses with MetaWin (version 2.0, Sinauer Associates,
Sunderland, MA, USA; Rosenberg et al. 2000). We first
assessed the effect of wire marking for the entire data
set with random-effects modeling, which allows for
the possibility that studies differ in sampling error (as
fixed-effects models do) and in random variation in
effect sizes (Gurevitch & Hedges 2001). Random-effects
models are more appropriate for analysis of ecological
data because numerous complex interactions are likely
to result in heterogeneity among studies (Pullin &
Stewart 2006). We calculated 95% confidence intervals
(CIs) (bias-corrected bootstrap, 999 iterations) for each
effect size (Rosenberg et al. 2000). If the 95% CI did not
overlap zero, then effects were significant at p<0.05.
We calculated the total heterogeneity, QT, to analyze
whether the variance among effect sizes was greater than
expected due to the sampling error (Rosenberg et al.
2000). This variable was a weighted sum of squares, and
we tested it against a χ2distribution with n–1 df. A signif-
icant QTimplies that other explanatory variables should
be investigated (Rosenberg et al. 2000). We estimated the
percentage of variation in effect sizes explained by each
Conservation Biology
Volume 25, No. 5, 2011

Barrientos et al. 897
predictor as QM/QT,whereQMis the variance explained
by the model from every predictor (Rosenberg et al.
2000).
In analysis of the subset of studies that included data
on flight frequency, we calculated differences in effect
sizes for the set of variables with random-effects models.
We evaluated the homogeneity of results for the set of
variables. We evaluated the “source” of information, that
is, the differences between studies from peer-reviewed
journals (journal) and other sources (unpublished). We
evaluated the variable “design,” which differentiated be-
tween BACI designs and those in which marked and un-
marked lines (or line sectors) were studied in the same
time interval. We used the variable “alternate” to test
whether studies in which spans were marked in alternat-
ing order (i.e., marked-unmarked-marked) affected effect
size; birds may have flown into unmarked spans to avoid
marked spans (Alonso et al. 1994; Crowder 2000). For
categorical data structure, we tested QMagainst a χ2dis-
tribution with n–1 df. We calculated cumulative effect
sizes for every group; effects were significant at p<0.05
when the 95% CI did not overlap zero. A significant QM
indicated there were differences among cumulative ef-
fects for the designated groups, whereas a significant QE
implied some heterogeneity among effect sizes was not
explained by the model (Rosenberg et al. 2000).
We also selected 5 continuous variables: “periodicity,”
mean number of carcass searches per month; “strip,”
total width searched on both sides of the line in meters;
“species,” number of bird species recorded; “lines,” num-
ber of power lines included in the study; and “stretches,”
number of marked or unmarked stretches of power line,
independent of the number of spans (i.e., the length
of line between 2 consecutive posts) contained in ev-
ery stretch. We determined the relation between effect
size and every continuous variable with weighted least-
squares regression. A significant QM(or regression coef-
ficient) implied the independent variable explained sig-
nificant variation in the effect sizes, and a significant QE
implied some heterogeneity among effect sizes was not
explained by the model (Rosenberg et al. 2000).
We explored the possibility of publication bias with
3 different approaches: construction of a funnel plot of
sample size versus effect size; use of Spearman rank corre-
lation between the standardized effect size and the stan-
dardized variance across studies (statistical significance
indicates large effect sizes are more likely to be published
than small effect sizes); and evaluation of the Rosenthal
(1979) fail-safe number, which is the number of non-
significant, unpublished, or missing studies that would
need to add to the meta-analysis to lose the statistical
significance of the results. If the fail-safe number is >5
times the sample size plus 10 ([sample size +10] ×5), it
is reasonable to conclude the results are robust regarding
publication bias. The 3 methods were used with the meta-
analysis that included flight-frequency data. We used the
fail-safe number in the meta-analysis of studies in which
only carcass counts were used because these studies did
not provide sample sizes, an essential parameter for the
other 2 methods.
Results
Twenty-one studies, including 52 separate wire-marking
experiments, met our selection criteria (Supporting Infor-
mation). Wire marking reduced bird mortality at p<0.05
(i.e., 95% CI did not overlap zero; −1.42 to −0.47). The
test for overall heterogeneity was not significant (QT=
51.00, df =51, p=0.47).
We selected 11 of the 21 studies, including 15 sep-
arate wire-marking experiments because data on flight
frequencies were collected (Supporting Information). Of
these 15 experiments with flight frequencies, results of
7 were published in peer-reviewed journals and 8 were
in unpublished reports, 3 of the experiments had BACI
designs, and 11 had parallel designs. Of these 11, 5 had
alternate designs and 6 had continuous designs (Fig. 1).
Overall collision rates were 0.21/1000 bird crossings at
unmarked lines and 0.05/1000 crossings at marked lines
(Fig. 1). With the exception of the few studies with BACI
designs, wire marking reduced bird mortality by 55–94%
(overall 78%; Fig. 1). The test for overall heterogeneity
was significant (QT=69.27, df =14, p<0.001), which
implies other explanatory variables should be investi-
gated. Among the 8 variables, only number of species
was significant at p<0.05 (Table 1 & Fig. 2), and the
effect size of the studies was larger for those in which
more species were present. All variables showed hetero-
geneity among effect sizes that were not explained by
their respective models (Table 1).
The fail-safe number for the meta-analysis of carcass
counts was 751, which is >5 times the sample size plus
10 (i.e., 310). The scatter plot derived from the sensitiv-
ity analysis did not show publication bias. The plot was
funnel shaped with a large opening at the smallest sam-
ple sizes. The fail-safe number for sensitive analysis was
393, which is 5 times larger than the sample size plus
10 (i.e., 125). Accordingly, Spearman rank correlation
of standardized effect size and the standardized variance
was not significant (Rs=−0.10, p=0.73).
Discussion
Effectiveness of Wire Marking
Our results suggest that marking static wires reduces the
number of bird casualties at power lines. However, colli-
sion risk was generally low even at unmarked lines. We
did not compare the relative efficiency of various types of
markers (color, shape) or evaluate the density of marking
devices on the wire. Few studies have been conducted on
Conservation Biology
Volume 25, No. 5, 2011

898 Wire Marking to Reduce Bird Collisions
Figure 1. Rate of avian collisions with power lines for the studies included in the meta-analysis (overall) and for 3
categorical variables, including studies reporting carcass counts and flight frequencies (black, unmarked power
lines; gray, marked power lines; numbers above bars, number of experiments), and percent reduction in collisions
for every category of study (striped, right axis) (∗, significant at alpha <0.05; ns, alpha ≥0.05; significant
values are based on 95% bootstrapped confidence intervals). The categorical variables are (1) source (study
published in a journal or unpublished), (2) BACI study design (before-after-control-impact) or study design in
which marked and unmarked lines (or line sectors) were monitored simultaneously (parallel), and (3) parallel
study designs with spans marked alternately (alternate) or continuously (continuous).
the effectiveness of different markers for different species
or habitats.
That our literature review included unpublished re-
ports, dissertations, and papers published in peer-
reviewed journals allowed us to increase the number
of studies included in the meta-analysis and to compare
results between published and unpublished studies. As-
suming positive results are more likely to be published,
use of all possible data sources reduces the probability
of overestimating positive results (Fern´
andez-Duque &
Valeggia 1994).
On the basis of studies reporting only carcass counts,
we inferred that marked lines are associated with a reduc-
tion in the number of collisions. Fifty-six percent of the
studies we examined included estimates of bird density.
Of these, we included in our second meta-analysis only
those in which flight frequencies in relation to marker ef-
fectiveness were analyzed. Bird densities around power
lines may not be associated with the magnitude of col-
lision risk because species with large home ranges and
species that make daily forays between roosting and feed-
ing habitats often cross power lines more frequently than
species with small home ranges (e.g., most sedentary
passerines), which makes them more likely to collide
with wires.
When marked and unmarked spans are alternated,
birds could fly into unmarked spans more often, presum-
ably to avoid marked ones (Alonso et al. 1994; Crowder
2000). However, we found no evidence of this. The ab-
sence of an effect of alternate marking could be due to
the fact that the most common reaction of birds when
approaching marked spans adjacent to unmarked spans
is to fly higher rather than to change direction (Morkill &
Anderson 1991; Savereno et al. 1996). The effect size
was smaller in studies with one or only a few species,
probably because these (e.g., Morkill & Anderson 1991;
Sudman 2000) studies focused on species that rarely re-
spond to marking and thus are most likely to collide with
wires (i.e., cranes and waterfowl) (Bevanger 1998; Janss
2000). All continuous variables had significant residual
error variances, which implies that not all heterogene-
ity among effect sizes was explained by their respective
models (Rosenberg et al. 2000).
Testing Differences among Marker Traits
Some researchers have statistically tested differences
among marker characteristics. Scott et al. (1972) found
no evidence that 2 types of devices, clipped strips
and tapes, reduce bird casualties. Brown and Drewien
Conservation Biology
Volume 25, No. 5, 2011

Barrientos et al. 899
Table 1. Summary of results of random models used to analyze
differences in effectiveness of markers to prevent bird collisions with
power lines.
Variableadf QMb(p) QEc(p) QM/QTd
Source 1, 13 0.12 63.70 0.00
(0.72) (<0.00)
Design 1, 12 0.04 62.20 0.00
(0.85) (<0.00)
Alternate 1, 9 0.49 27.85 0.02
(0.48) (<0.00)
Periodicity 1, 13 1.79 66.27 0.03
(0.18) (<0.00)
Strip 1, 12 2.62 64.71 0.04
(0.11) (<0.00)
Number of species 1, 12 6.85 70.63 0.09
(<0.01) (<0.00)
Number of lines 1, 13 1.45 64.43 0.02
(0.23) (<0.00)
Number of stretches 1, 13 0.80 62.54 0.01
(0.37) (<0.00)
aKey: source, peer-reviewed journal or unpublished; design, BACI
(before-after-control-impact) design or marked and unmarked lines
(or line sectors) studied simultaneously; alternate, stretches of wire
marked in alternating order or in a continuous form; periodicity,
mean number of carcass searches per month; strip, total width of
search strip; number of species, total recorded per study; number of
lines, power lines included in the study; number of stretches, total
marked and unmarked stretches of power line studied.
bHeterogeneity explained by the model or between-group heterogene-
ity.
cResidual error variance or within-group heterogeneity.
dFraction of the total heterogeneity explained by the model.
(1995) found slightly fewer mortalities associated with
damper devices compared with plates, although reduc-
tion in mortality of segments marked with these devices
compared with unmarked segments was significant for
both dampers and plates. Janss and Ferrer (1998) found
that white spirals and flapper flight diverters signifi-
cantly reduce mortality, whereas black stripes do not.
However, habitat differences (crops vs. wetlands) may
have influenced their results. Anderson (2002) found that
flappers are more effective than flight diverters. Stake
(2009) found that small flight diverters and large divert-
ers both reduce the number of bird collisions, but that
large diverters are ineffective at some sites. Calabuig and
Ferrer (2009) found that cross-shaped markers are more
effective than spirals of different colors. Calabuig and Fer-
rer (2009) found that spirals of different colors (white,
orange, and yellow) similarly reduce mortality relative to
unmarked spans. Spirals appear to be more durable than
flapper flight diverters (Calabuig & Ferrer 2009). How-
ever, all these studies have low statistical power, which
could lead to type II errors (Arnqvist & Wooster 1995).
Some researchers have evaluated the effects of marking
at different intervals, but did not statistically analyze their
results (Koops & de Jong 1982; Anderson 2002). Other
researchers have examined the effectiveness of markers
of different sizes, but did not make statistical comparisons
among treatments (Koops & de Jong 1982).
Wire marking with standard flight diverters may not re-
duce the number of collisions of crepuscular or nocturnal
birds. Waterfowl and nocturnal migrants are among the
birds most prone to collisions with wires (reviewed in
Drewitt & Langston 2008). Only a few studies of the ef-
fectiveness of new types of flight diverters (e.g., FireFly di-
verters) have been conducted (Pilo et al. 1994; Yee 2007;
Murphy et al. 2009). Furthermore, many collisions may
occur during the day, when visibility is high (Drewitt &
Langston 2008). Martin and Shaw (2010) suggest that
wire marking may have limited success for bird species
with narrow visual fields, such as bustards, storks, and
cranes. Thus, it is possible that no single type of marker
will be equally effective with all bird species or in all situa-
tions, which suggests investigations of nonvisual devices
are needed.
Guidelines for Marking-Efficiency Experiments
The large differences in wire-marking techniques con-
strained our ability to infer whether this method reduces
bird collisions. We expect that the number of studies
on this topic may increase substantially over time given
the increasing demand for fewer and smaller effects of
human actions on the environment and increasing use of
marking devices on power lines worldwide (APLIC 1994,
2006; Manville 2009).
Four improvements in the design of studies of wire-
marking effectiveness may help determine which mark-
ing techniques are the most effective. First, we recom-
mend collecting data on carcass counts and flight fre-
quency for the same length of time and at the same
time of year at marked and unmarked wire segments.
For instance, some studies with BACI designs moni-
tored marked and unmarked lines for different lengths
of time (e.g., Anderson 2002; de la Zerda & Rosselli
2003). This could produce biases in the data because
flight frequencies are not constant throughout the year;
frequencies differ among and within seasons (e.g., spring
migration). Even when data are collected at the same
time of year, flight frequencies can vary. For example,
Alonso et al. (1994) recorded 2.9 times more flights in
December–April before wires were marked compared
with December–April after wires were marked. Conse-
quently, we recommend recording flight frequencies and
number of carcasses simultaneously.
Second, we suggest studying marked and unmarked
lines in areas with similar vegetation and topography,
the use of similar lengths of time spent searching for
carcasses, and searching transects of equal lengths and
widths (Bevanger 1999). For instance, bird collisions with
power lines are more frequent for lines that cross wet-
lands and where lines are between feeding and roosting
areas (Scott el al. 1972; McNeil et al. 1985; Ferrer & Janss
1999). Monitoring of different lengths of lines in different
land-cover types or bird habitats could drive differences
between marked and unmarked lines or among different
Conservation Biology
Volume 25, No. 5, 2011

900 Wire Marking to Reduce Bird Collisions
Figure 2. In a meta-analysis of
studies of effectiveness of wire
marking in reducing bird
collisions with power lines,
relation between effect size and
(a) mean number of carcass
searches per month (periodicity),
(b) total width of transect
searched on both sides of the
power line (strip), (c) number of
species recorded per study, (d)
number of power lines examined
per study, and (e) number of
marked and unmarked stretches
of power line monitored per
study (stretches). Each point
represents an experimental case.
markers (e.g., Janss & Ferrer 1998) because different bird
species have different habitats and not all species have
the same probability of collision (Bevanger 1998; Janss
2000; Martin & Shaw 2010).
Third, we recommend standardizing the periodicity
of carcass searches and the search strip width, at least
within every study. Although both variables must be
constant to support detailed comparisons among studies
Conservation Biology
Volume 25, No. 5, 2011

Barrientos et al. 901
Figure 2. (continued)
(Bevanger 1999), periodicity in the studies we examined
was sometimes fairly different even between marked
and unmarked lines within a study (e.g., Koops & de
Jong 1982). In general, we think the frequency of car-
cass searches should be determined on the basis of the
species’ body size because body size is correlated with
the removal rate of carcasses by scavengers. Larger car-
casses generally remain in the field longer than smaller
ones and are more easily located (Ponce et al. 2010).
Moreover, the carcass removal rate varies among habitats
and density and type of scavengers (Bevanger 1994), so
the periodicity of carcass searches and the length and
width search strip should be defined according to the
target species. Few researchers analyzed the distance
from the power line at which carcasses were found
(but see Frost 2008; Shaw et al. 2010). Ideally, carcass-
disappearance studies in which similar protocols are ap-
plied should be carried out in each study area prior to
studies of marking efficiency (e.g., Pelayo & Sampietro
1994; Onrubia et al. 1996).
Fourth, we recommend researchers compare the effec-
tiveness of currently available commercial markers used
to reduce bird collision. Due to the heterogeneity of mark-
ers used in the studies in our meta-analysis, we could not
compare effectiveness of different types of devices (e.g.,
flight diverters, aviation balls, flappers) or device color,
differences between categorical or continuous measures
of device size, or differences among spacing of devices
(e.g., every 5, 10, 20 m). Few conclusions about effective-
ness can be drawn from experiments in which the life
expectancy or color fading of different commercial de-
vices was examined (Hunting 2002). The optimal density
of markers or the effectiveness of using specific colors
over others has not been explored (Hunting 2002).
Acknowledgments
R. Harness, B. Parshalle, M. Rogen, A.M. Manville, J.A.
Lorenzo, L. Spiegel, J. Kreuziger, D. Frost, M. de Lucas,
Conservation Biology
Volume 25, No. 5, 2011

902 Wire Marking to Reduce Bird Collisions
M. Ferrer, J. Smallie, M. Stake, T. Lislevand, J.J. Ramos,
S. Smallwood, J. Shaw, F.J. Purroy, and J.M. ´
Alvarez
provided valuable information for this review. J. Gure-
vitch kindly resolved several doubts on meta-analysis
design, J.D. Ib´
a˜
nez helped us with MetaWin use, L.
de Neve provided translations of papers in Dutch, and
S. Young reviewed the English. J. Faaborg, M. McCarthy,
E. Fleishman, A.M. Manville, and an anonymous reviewer
improved a first draft. Financial support for this study
was provided by project CGL2008-02567 of the Direcci´
on
General de Investigaci´
on, Spanish Ministry of Science and
Innovation.
Supporting Information
Raw data and effect sizes calculated from studies used in
the meta-analysis are available online (Appendixes S1 &
S2). The authors are responsible for the content and func-
tionality of these materials. Queries (other than absence
of the material) should be directed to the corresponding
author.
Literature Cited
Alonso, J. C., J. A. Alonso, and R. Mu˜
noz-Pulido. 1994. Mitigation of
bird collisions with transmission lines through groundwire marking.
Biological Conservation 67:129–134.
Alonso, J. C., C. A. Mart´
ın, and C. Palac´
ın, coordinators. 2005. Proyecto
de medidas preventivas, correctoras y compensatorias de la afecci´
on
delaM-50ydelaAutopistadePeajeR-2alapoblaci
´
on de avutardas y
otras aves esteparias de la IBA Talamanca-Camarma, y al LIC Cuenca
de los r´
ıos Jarama y Henares. Museo Nacional de Ciencias Naturales
(CSIC), Madrid.
Anderson, M. D. 2002. The effectiveness of two different marking de-
vices to reduce large terrestrial bird collisions with overhead elec-
tricity cables in the eastern Karoo, South Africa. Report 1. Karoo
Large Terrestrial Bird Powerline Project, Eskom, Johannesburg.
APLIC (Avian Power Line Interaction Committee). 1994. Mitigating bird
collisions with power lines: the state of the art in 1994. Edison
Electric Institute, Washington, D.C.
APLIC (Avian Power Line Interaction Committee). 2006. Suggested
practices for avian protection on power lines: state of the art
in 2006. Edison Electric Institute, Washington, D.C., and the
California Energy Commission, Sacramento, California.
Arnqvist, G., and D. Wooster. 1995. Meta-analysis: synthesizing research
findings in ecology and evolution. Trends in Ecology Evolution
10:236–240.
Beaulaurier, D. L. 1981. Mitigation of bird collisions with transmission
lines. Bonneville Power Administration, Portland, Oregon.
Ben´
ıtez-L´
opez, A., R. Alkemade, and P. A. Verweij. 2010. The impacts
of roads and other infrastructure on mammal and bird populations:
a meta-analysis. Biological Conservation 143:1307–1316.
Bevanger, K. 1994. Bird interactions with utility structures: collision and
electrocution, causes and mitigating measures. Ibis 136:412–425.
Bevanger, K. 1995. Estimates and population consequences of tetraonid
mortality caused by collisions with high tension power lines in
Norway. Journal of Applied Ecology 32:745–753.
Bevanger, K. 1998. Biological and conservation aspects of bird mortality
caused by electric power lines. Biological Conservation 86:67–76.
Bevanger, K. 1999. Estimating bird mortality caused by collision and
electrocution with power lines: a review of methodology. Pages
29–56 in M. Ferrer and G. F. E. Janss, editors. Birds and power lines:
collision, electrocution and breeding. Quercus, Madrid.
Bevanger, K., and H. Brøseth. 2001. Bird collisions with power lines: an
experiment with ptarmigan (Lagopus spp.). Biological Conservation
99:341–346.
Borenstein, M., L. V. Hedges, J. P. T. Higgins, and H. R. Rothstein. 2009.
Introduction to meta-analysis. Wiley, West Sussex, United Kingdom.
Brauneis, W., W. Watzlaw, and L. Horn. 2003. Das Verhalten von
V¨
ogeln im Bereich eines ausgew¨
ahlten Trassenabschnittes der
110 kV-Leitung Bernburg-Susigke (Bundesland Sachsen-Anhalt). Flu-
greaktionen, Drahtanfl¨
uge, Brutvorkommen. ¨
Okologie der V¨
ogel
25:69–115.
Brown, W. M., and R. C. Drewien. 1995. Evaluation of two power line
markers to reduce crane and waterfowl collision mortality. Wildlife
Society Bulletin 23:217–227.
Calabuig, C. P., and M. Ferrer. 2009. An´
alisis de la eficacia y la vida ´
util
de la se˜
nalizaci´
on anticolisi´
on “salvap´
ajaros” en l´
ıneas de transporte
de energ´
ıa el´
ectrica. Red El´
ectricadeEspa
˜
na, Sociedad An´
onima
Unipersonal, and Consejo Superior de Investigaciones Cient´
ıficas,
Seville.
Crowder, M. R. 2000. Assessment of devices designed to lower the
incidence of avian power line strikes. PhD thesis. Purdue University,
West Lafayette, Indiana.
Crowder, M. R., and O. E. Rhodes. 2001. Avian collisions with power
lines: a review. Pages 139–168 in R. G. Carlton, editor. Avian
interactions with utility and communication structures. Work-
shop Proceedings. Electric Power Research Institute, Palo Alto,
California.
De la Zerda, S., and L. Rosselli. 2003. Mitigaci´
on de colisi´
on de aves con-
tra l´
ıneas de transmisi´
on el´
ectrica con marcaje del cable de guarda.
Ornitolog´
ıa Colombiana 1:42–62.
Drewitt, A. L., and R. H. W Langston. 2008. Collision effects of wind-
power generators and other obstacles on birds. Annals of the New
York Academy of Sciences 1134:233–266.
Fern´
andez-Duque, E., and C. Valeggia. 1994. Meta-analysis: a valuable
tool in conservation research. Conservation Biology 8:555–561.
Ferrer, M., G. F. E. Janss, editors. 1999. Birds and power lines. Collision,
electrocution and breeding. Quercus, Madrid.
Frost, D. 2008. The use of ‘flight diverters’ reduces mute swan Cygnus
olor collision with power lines at Abberton Reservoir, Essex, Eng-
land. Conservation Evidence 5:83–91.
Gilbert-Norton, L., R. Wilson, J. R. Stevens, and K. H. Beard. 2010. A
meta-analytic review of corridor effectiveness. Conservation Biology
24:660–668.
Gurevitch, J., and L. V. Hedges. 2001. Meta-analysis: combining the re-
sults of independent experiments. Pages 347–369 in S. M. Scheiner
and J. Gurevitch, editors. Design and analysis of ecological experi-
ments. Oxford University Press, New York.
Hedges, L. V., J. Gurevitch, and P. S. Curtis. 1999. The meta-analysis
of response ratios in experimental ecology. Ecology 80:1150–
1156.
Hebert, E., and E. Reese (editors). 1995. Avian collision and elec-
trocution: an annotated bibliography. Publication P700–95-001.
California Energy Commission, Sacramento.
Hunting, K. 2002. A roadmap for PIER research on avian collisions
with power lines in California. Technical report P500–02-071F.
California Energy Commission, Public Interest Energy Research
(PIER) Program, Sacramento.
Janss, G. F. E. 2000. Avian mortality from power lines: a morphologic
approach of a species-specific mortality. Biological Conservation
95:353–359.
Janss, G. F. E., and M. Ferrer. 1998. Rate of bird collision with power
lines: effects of conductor-marking and static wire-marking. Journal
of Field Ornithology 69:8–17.
Janss, G. F. E., A. Lazo, and M. Ferrer. 1999. Use of raptor models to
reduce avian collisions with powerlines. Journal of Raptor Research
33:154–159.
Conservation Biology
Volume 25, No. 5, 2011

Barrientos et al. 903
Jenkins, A. R., J. J. Smallie, and M. Diamond. 2010. Avian collisions
with power lines: a global review of causes and mitigation with a
South African perspective. Bird Conservation International 20:263–
278.
Koops, F. B. J. 1994. Collision victims of high-tension lines in The
Netherlands and effects of marking. Pages 51–57 in First technical
sessions on power lines and the environment. Red El´
ectrica de
Espa˜
na, Madrid.
Koops, F. B. J., and J. de Jong. 1982. Vermindering van draadslachtof-
fers door markering van hoogspanningsleiden in de omgeving van
Heerenveen. Elektrotechniek 60:641–646.
Kruger, R. 2001. A risk based approach to the cost of implementing
raptor mitigation measures on Eskom distribution networks in South
Africa. Pages 229–246 in R. G. Carlton, editor. Avian interactions
with utility and communication structures. Workshop proceedings.
Electric Power Research Institute, Palo Alto, California.
Lehman, R. N., P. L. Kennedy, and J. A. Savidge. 2007. The state of
the art in raptor electrocution research: a global review. Biological
Conservation 135:459–474.
Lorenzo, J. A., and M. Cabrera. 2009. Estudio de la mortalidad de aves
por los tendidos el´
ectricos en los ambientes no esteparios de Lan-
zarote y Fuerteventura, y valoraci´
on de las medidas correctoras in-
staladas en Fuerteventura. Memoria final. Sociedad Espa˜
nola de Or-
nitolog´
ıa, BirdLife, and Empresa Nacional de Electricidad, Sociedad
An´
onima, Tenerife, Spain.
Manville, A. M., II. 1999. The ABC’s of avoiding bird collisions at com-
munication towers: the next steps. Pages 85–104 in R. G. Carlton,
editor. Avian interactions with utility and communication struc-
tures. Workshop proceedings. Electric Power Research Institute,
Palo Alto, California.
Manville, A. M., II. 2005. Bird strikes and electrocutions at power lines,
communication towers, and wind turbines: state of the art and state
of the science – next steps toward mitigation. Bird conservation
implementation in the Americas. Pages 1051–1064 in C. J. Ralph
and T. D. Rich, editors. Proceedings 3rd international Partners in
Flight conference 2002. General technical report PSW-GTR-191. U.S.
Department of Agriculture Forest Service, Pacific Southwest
Research Station, Albany, California.
Manville, A. M., II. 2009. Towers, turbines, power lines, and buildings:
steps being taken by the U.S. Fish and Wildlife Service to avoid or
minimize take of migratory birds at these structures. Pages 262–272
in T.D. Rich, C. Arizmendi, D. Demarest, and C. Thompson, edi-
tors. Tundra to tropics: connecting birds, habitats, and people. Pro-
ceedings 4th international Partners in Flight conference. Partners in
Flight.
Martin, G. R., and J. M. Shaw. 2010. Bird collisions with power lines: fail-
ing to see the way ahead? Biological Conservation 143:2695–2702.
McNeil, R., J. R. Rodr´
ıguez, and H. Ouellet. 1985. Bird mortality at
a power transmission line in North Eastern Venezuela. Biological
Conservation 31:153–165.
Morkill, A. E, and S. H. Anderson. 1991. Effectiveness of marking pow-
erlines to reduce sandhill crane collisions. Wildlife Society Bulletin
19:442–449.
Murphy, R. K., S. M. McPherron, G. D. Wright, and K. L. Serbousek.
2009. Effectiveness of avian collision averters in preventing migra-
tory bird mortality from powerline strikes in the central Platte River,
Nebraska. Nebraska Game and Parks Commission, U.S. Fish and
Wildlife Service, and University of Nebraska, Kearney.
Neter, J., W. Wasserman, and M. H. Kutner. 1989. Applied linear regres-
sion models. 2nd edition. Irwin, Homewood, Illinois.
Onrubia, A., F. J. Purroy, and A. Ubeda. 1996. La eficacia de las se˜
nales
de visualizaci´
on sobre colisi´
on y electrocuci´
on de aves en tendidos
el´
ectricos de Madrigal de las Altas Torres (´
Avila). Pages 265–271 in
J. Fern´
andez Guti´
errez and J. S´
anz-Zuasti, editors. Conservaci´
on de
las aves esteparias y su h´
abitat. Junta de Castilla y Le´
on, Valladolid,
Spain.
Pelayo, E, and F. J. Sampietro. 1994. Estudio del impacto sobre las aves
del tendido el´
ectrico de alta tensi´
on La Sierna-Pe˜
naflor. Valoraci´
on
de la eficacia de la se˜
nalizaci´
on con espirales, y aproximaci´
on a
la incidencia de mortalidad real sobre aves. Sociedad Espa˜
nola de
Ornitolog´
ıa/BirdLife and Red El´
ectrica de Espa˜
na, Madrid.
Pilo, B., A. Kumar, S. Oommen, and K. R. Vinod. 1994. Anti-bird collision
strobe lights: field experiments on Indian birds. 22/WP 8. Bird Strike
Committee Europe, Vienna.
Ponce, C., J. C. Alonso, G. Argando˜
na,A.G.Fern
´
andez, and M. Car-
rasco. 2010. Carcass removal by scavengers and search accuracy
affect bird mortality estimates at power lines. Animal Conservation
13:603–613.
Pullin, A. S., and G. B. Stewart. 2006. Guidelines for systematic review
in conservation and environmental management. Conservation Bi-
ology 20:1647–1656.
Rosenberg, M. S., D. C. Adams, and J. Gurevitch. 2000. MetaWin: statis-
tical software for meta-analysis Version 2. Sinauer Associates, Sun-
derland, Massachusetts.
Rosenthal, R. 1979. The ‘‘file drawer problem’’ and tolerance for null
results. Psychological Bulletin 86:638–641.
Savereno, A. J., L. A. Savereno, R. Boettcher, and S. M. Haig. 1996. Avian
behavior and mortality at power lines in coastal South Carolina.
Wildlife Society Bulletin 24:636–648.
Schaub, M., and R. Pradel. 2004. Assessing the relative importance of
different sources of mortality from recoveries of marked animals.
Ecology 85:930–938.
Scott R. E., L. J. Roberts, and C. J. Cadbury. 1972. Bird deaths from
power lines at Dungeness. British Birds 65:273–286.
Shaw, J. M. 2009. The end of the line for South Africa’s national bird?
Modelling power line collision risk for the Blue Crane. MS thesis.
University of Cape Town, Cape Town.
Shaw, J. M., A. R. Jenkins, J. J. Smallie, P. G. Ryan. 2010. Modelling
power-line collision risk for the Blue Crane Anthropoides par-
adiseus in South Africa. Ibis 152:590–599
Stake, M. M. 2009. Evaluating diverter effectiveness in reducing avian
collisions with distribution lines at San Luis National Wildlife Refuge
Complex, Merced County, California. CEC-500-2009-078. Public In-
terest Energy Research (PIER) Program, Ventana Wildlife Society,
and California Energy Commission, Sacramento, California.
Stewart, G. B., C. F. Coles, and A. S Pullin. 2005. Applying evidence-
based practice in conservation management: lessons from the first
systematic review and dissemination projects. Biological Conserva-
tion 126:270–278.
Stewart, G. B., A. S. Pullin, and C. F. Coles. 2007. Poor evidence-base
for assessment of windfarm impacts on birds. Environmental Con-
servation 34:1–11.
Stewart, G. 2010. Meta-analysis in applied ecology. Biology Letters
6:78–81.
Sudman, S. R., coordinator. 2000. Das Anflugverhalten von
¨
uberwinternden, arktischen Wildg¨
ansen im Bereich von markierten
und unmarkierten Hochspannungsfreileitungen am Niederrhein.
Naturschutzzentrum im Kreis Kleve e.V. Kranenburg-N¨
utterden.
Sterna, Kranenburg-N¨
utterden, Germany.
Yee, M. L. 2007. Testing the effectiveness of an avian flight diverter for
reducing avian collisions with distribution power lines in the Sacra-
mento Valley, California. CEC-500-2007-122. Public Energy Related
Environmental Research Program, California Energy Commission,
Sacramento.
Conservation Biology
Volume 25, No. 5, 2011