Expression of trematode-induced zombie-ant behavior is strongly associated with temperature

Abstract

Parasite-induced modification of host behavior increasing transmission to a next host is a common phenomenon. However, field-based studies are rare, and the role of environmental factors in eliciting host behavioral modification is often not considered. We examined the effects of temperature, relative humidity (RH), time of day, date, and an irradiation proxy on behavioral modification of the ant Formica polyctena (Förster, 1850) by the brain-encysting lancet liver fluke Dicrocoelium dendriticum (Rudolphi, 1819). This fluke induces ants to climb and bite to vegetation by the mandibles in a state of temporary tetany. A total of 1264 individual ants expressing the modified behavior were observed over 13 non-consecutive days during one year in the Bidstrup Forests, Denmark. A sub-set of those ants (N = 172) was individually marked to track the attachment and release of infected ants in relation to variation in temperature. Infected ants primarily attached to vegetation early and late in the day, corresponding to low temperature and high RH, presumably coinciding with the grazing activity of potential herbivorous definitive hosts. Temperature was the single most important determinant for the induced phenotypic change. On warm days, infected ants altered between the manipulated and non-manipulated state multiple times, while on cool days, many infected ants remained attached to the vegetation all day. Our results suggest that the temperature sensitivity of the infected ants serves the dual purpose of exposing infected ants to the next host at an opportune time, while protecting them from exposure to high temperatures, which might increase host (and parasite) mortality.

Full text

Behavioral
Ecology
The ocial journal of the
ISBE
International Society for Behavioral Ecology
Behavioral Ecology (2023), XX(XX), 1–9. https://doi.org/10.1093/beheco/arad064
Address correspondence to S.N. Gasque, who is now at the Laboratory
of Virology, Wageningen University and Research, Droevendaalsesteeg 1,
6708 PB Wageningen, The Netherlands. E-mail: simone.gasque@wur.nl.
© The Author(s) 2023. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits
unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
Original Article
Expression of trematode-induced zombie-
ant behavior is strongly associated with
temperature
Simone NordstrandGasque and Brian LundFredensborg
Section for Organismal Biology, Department of Plant and Environmental Sciences, University of
Copenhagen, Thorvaldsensvej 40, 1871 Frederiksberg C, Denmark
Received 19 January 2023; revised 17 July 2023; editorial decision 17 July 2023; accepted 29 July 2023
Parasite-induced modification of host behavior increasing transmission to a next host is a common phenomenon. However, field-
based studies are rare, and the role of environmental factors in eliciting host behavioral modification is often not considered. We
examined the effects of temperature, relative humidity (RH), time of day, date, and an irradiation proxy on behavioral modification
of the ant Formica polyctena (Förster, 1850) by the brain-encysting lancet liver fluke Dicrocoelium dendriticum (Rudolphi, 1819).
This fluke induces ants to climb and bite to vegetation by the mandibles in a state of temporary tetany. A total of 1264 individual
ants expressing the modified behavior were observed over 13 non-consecutive days during one year in the Bidstrup Forests,
Denmark. A sub-set of those ants (N = 172) was individually marked to track the attachment and release of infected ants in re-
lation to variation in temperature. Infected ants primarily attached to vegetation early and late in the day, corresponding to low
temperature and high RH, presumably coinciding with the grazing activity of potential herbivorous definitive hosts. Temperature
was the single most important determinant for the induced phenotypic change. On warm days, infected ants altered between the
manipulated and non-manipulated state multiple times, while on cool days, many infected ants remained attached to the vegeta-
tion all day. Our results suggest that the temperature sensitivity of the infected ants serves the dual purpose of exposing infected
ants to the next host at an opportune time, while protecting them from exposure to high temperatures, which might increase host
(and parasite) mortality.
Key words: context-dependent behavioral manipulation, Dicrocoelium dendriticum, lancet liver fluke, field study, Formica
polyctena, parasite-induced host behavioral manipulation, temperature.
INTRODUCTION
Parasite manipulation of the host phenotype (appearance and/
or behavior) to increase transmission to a new host is a wide-
spread phenomenon across host and parasite phyla (Moore
2002). The phenotypic changes of the infected host may involve
the dispersal of infective propagules to favorable locations and
at favorable times or, among parasites with a complex life cycle,
facilitate an increase in the predation rate of the infected host
by the next host in the life cycle of the parasite (i.e., trophic
transmission) (Combes 2001; Poulin 2007). Thus, the phenom-
enon of host phenotypic manipulation is the result of a strong
selection pressure on the parasite to enhance the otherwise very
small chance of completing its life cycle (Poulin 2007; Froelick
et al. 2021). Many cases of host phenotypic manipulation by
parasites have been reported (Laerty and Shaw 2013), but envi-
ronmental or physiological factors that regulate the expression of
the manipulated behavior in infected hosts remain undescribed
in most cases (Moore 2002; Laerty and Shaw 2013; Hughes
and Liebersat 2018). Field studies of host behavior in relation
to parasite infection are complex and extremely rare since the
parasitized host usually displays alterations of already existing
behaviors, and parasite contribution to those behaviors may be
dicult to quantify under natural conditions (Poulin 1995). The
commonality of trophically transmitted parasites in natural eco-
systems and their potentially great eect on food web proper-
ties (Laerty et al. 2008) calls for a much better understanding
of the environmental factors associated with their transmission.
We studied Dicrocoelium dendriticum, the lancet liver fluke, which
induces a radical but temporary behavioral change in the ant in-
termediate host, making it easy to phenotypically distinguish in-
fected from uninfected individuals. This quintessential example,
therefore, serves as an appropriate model to investigate factors
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Behavioral Ecology
involved in eliciting and maintaining parasite-induced behavioral
manipulation under field conditions.
D. dendriticum has a complex life cycle, including a terrestrial snail
as the first intermediate host, worker ants as the second interme-
diate host, and herbivorous mammals as the definitive host (Figure
1). D. dendriticum relies on trophic transmission for the completion
of its life cycle. Individual larval parasites (metacercariae) migrate
in an act of altruistic kin selection behavior (Criscione et al. 2020)
to the suboesophageal ganglion of the ant (Romig et al. 1980;
Martín-Vega et al. 2018), where they elicit a reversible and radical
behavioral change, which is unique to infected hosts (Botnevik et
al. 2016). Hence, the infected ant climbs and locks its mandibles
to vegetation (Badie and Rondelaud 1988) in a state of tetany that
leaves the ant susceptible to ingestion by an herbivorous mamma-
lian definitive host (Tarry 1969). If not ingested by the herbivorous
host, the mandibles unlock, and the infected ant is free to return to
the forest floor.
In other host–parasite associations, the behavioral change is also
radical but precedes the inevitable death of the host (e.g., tree-
top disease by baculoviruses and death grip by Ophiocordyceps- or
Pandora-infected ants) (reviewed by Gasque et al. 2019). Thus, in
fungal- and viral-induced behaviors where the spores/viral progeny
can persist in the environment, the host normally dies within
hours/days after the induction of the altered behavior. Contrary to
this, D. dendriticum cannot survive outside the intermediate ant host,
and transmission is dependent on the direct ingestion of infected
ants by herbivorous definitive hosts. The likelihood that an infected
ant is eaten upon initial display of parasite-induced biting behavior
may be very small. The chance of transmission of D. dendriticum
will, however, increase with prolonged exposure to the definitive
host and thus depend on ant host longevity. Thus, it is likely that
natural selection favors increased exposure to the definitive host but
disfavors exposure to pre-mature predation or hazardous environ-
mental conditions, which could kill the ant host and, therefore, the
parasite (Fritz 1982; Parker et al. 2009).
Studies dating back to the 1970s and 1980s reported anecdotal
observations of D. dendriticum-infected ants biting in the morning
and evening, leading to speculations on the possible role of time
of day, temperature, or other environmental factors related to in-
creased transmission success to grazing ruminant definitive hosts,
but this was not investigated further (Tarry 1969; Badie et al. 1973;
Badie 1975; Spindler et al. 1986; Badie and Rondelaud 1988).
The chronobiology and eects of light have been shown to play
a role in the expression of parasite-induced behavioral changes in
other systems. In the case of Ophiocordyceps unilateralis s.l.-infected
carpenter ants climb around solar noon and bite the vegetation in
locations and orientations, which would lead to optimal humidity
conditions for spore formation and spore release (Andersen et al.
2009; Hughes et al. 2011). Another host–parasite system similarly
dependent on trophic transmission to D. dendriticum is the brain-
encysting trematode Microphallus papillorobustus and acanthoceph-
alans infecting gammarids. Infected gammarids seek light and cling
Adult
Unencysted
Metacercaria
Metacercaria
Cercaria
Egg
Miracidium
Figure 1
Lifecycle of Dicrocoelium dendriticum. Eggs are excreted from an herbivorous mammal serving as the definitive host and hatch into a miracidium when ingested
by a terrestrial snail. Mother and daughter sporocysts develop in the hepatopancreas of the snail (not shown), producing the cercaria stage. Cercariae are
released from the snail via slimeballs excreted from the pneumastoma. Ants eat the slimeballs containing infective cercariae, which develop into metacercariae
in the hemocoel of the gaster of the ant. However, one or more cercariae penetrates the lining of the crop upon ingestion and migrates to the suboesophageal
ganglion (depicted in the headcapsule of the ant by arrow; Martín-Vega et al. 2018), where it induces a reversible biting behavior of the ants where the
mandibles are locked to vegetation leaving the ant exposed to ingestion by a suitable definitive host. Following ingestion, the metacercariae excyst in the
duodenum, migrate to the bile ducts of the liver and mature to become adult worms producing large numbers of eggs by sexual reproduction. Original
drawing by Simone Nordstrand Gasque.
Page 2 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Gasque and Fredensborg · Expression of trematode-induced zombie-ant behavior
to submerged macrophytes in the habitat of dabbling ducks which
serve as definitive hosts (Bethel and Holmes 1973; Helluy 1983).
Our previous laboratory study indicated that the biting behavior
of D. dendriticum-infected ants was determined by a negative rela-
tionship between increasing temperature and the probability that
infected ants would bite a leaf (Botnevik et al. 2016).
However, the eect of temperature on D. dendriticum infected
ant’s biting behavior has, so far, not been tested under natural con-
ditions, where ants are exposed to multiple environmental variables
with potential eects on ant behavior.
In this study, we recorded the manipulated phenotype of the
second intermediate host, the European Red Wood Ant, Formica
polyctena, naturally infected with D. dendriticum, concurrent with in
situ measurements of temperature, relative humidity (RH), and
time of day. In addition, we quantified the frequency of changes
to and from the manipulated phenotype in relation to temperature
on individually marked ants on 7 dierent days from August to
October and at two separate sites.
We hypothesized that temperature was the most important factor
in infected ants expressing the manipulated behavior in line with
the previous laboratory study (Botnevik et al. 2016) and with cir-
cumstantial evidence from previous field observations (Badie et
al. 1973; Paraschivescu and Raicev 1980; Spindler 1986). To our
knowledge, this is the first study to individually label and track the
behavior of infected ants in the field, which rendered the option to
recatch individual ants and study their behavioral expression of in-
fections over 24 h, and on longer term.
MATERIALS AND METHODS
Fieldsite and execution of fieldwork
Fieldwork took place in the Bidstrup Forests in Hvalsø, approxi-
mately 45 km Southwest of Copenhagen, Denmark, in 2016–2017.
The forests contain a mixture of hardwood and coniferous trees in
a hilly terrain and the D. dendriticum-F. polyctena system inhabits the
forests (Botnevik et al. 2016).
The aims of the fieldwork were to 1) Investigate the relative con-
tribution of temperature and RH (measured in the microclimate
of infected ants), time of day, date, and an irradiation proxy in
inducing and maintaining the altered phenotype (biting to vegeta-
tion). 2) Track biting behavior of individually marked infected ants
to observe the dynamics of initiation, maintenance, and termina-
tion of the altered phenotype.
Four anthills were included in the observations each day (except
in April 2017, see below) and were from two dierent sites (s and
h). Three anthills; Anthill 1 (55°34ʹ41ʹʹN 11°52ʹ22ʹʹE), Anthill 2
(55°34ʹ40ʹʹN 11°52ʹ22ʹʹE) and Anthill 3 (55°34ʹ40ʹʹN 11°52ʹ23ʹʹE)
were located in a line at approximately 1 m distance from each
other (site s). Anthill 4 was located approximately 300 m from site s
(55°34ʹ37ʹʹN 11°52ʹ43ʹʹE) and termed site h.
Ant observations
Fieldwork was avoided on days with precipitation to standardize the
conditions for ant behavioral studies. Ant observations were con-
ducted from early morning to late afternoon/evening on nine days
from 16 August 2016 until 12 October 2016, and on 3 days from
6 April 2017 to 16 August 2017. On one occasion, ant observa-
tions continued until sunset and continued the following morning,
to cover a 24 h cycle (22 September 2016 to 23 September 2016).
A total of 131.5 h of field observations were conducted from 12
occurrences, with one stretching 2 days, giving a total of 13 obser-
vation days from 16 August 2016 to 16 August 2017.
On each field day, the vegetation within a perimeter of 2 m
from each anthill (3 m from Anthill 4 due to the dispersed vege-
tation at this site) was examined for ants biting to vegetation at a
minimum of six times every day. A temperature/humidity -logger
(OM-EL-USB-2, Omega Engineering Inc.) was attached with two
plastic strips to a 1-m long marker stick and placed in the ground
at the same height and in close proximity to the majority of ants
biting to vegetation at the first observation time point (Figure 2B).
Temperature and RH were recorded continually by the temper-
ature/humidity-logger in intervals of 2 min. For the analysis, the
median temperature and median RH were used for the minutes
that did not have a recorded temperature and RH linked to it. For
Anthill 2, the mean values from the surrounding anthills (1 and 3)
were used for each timepoint. Five factors that potentially influ-
enced the proportion of infected ants biting to vegetation were in-
cluded in the analysis of ant activity: temperature, RH, date, time
of day (day length divided in three equal groups of the total day
length), and an irradiation-proxy (calculated from ocial records
of sunrise and sunset, www.soltider.dk), indicating the level of solar
irradiance that biting ants could be exposed to.
(a) (b) (c) (d)
Figure 2
Photographs of Formica polyctena ants expressing the behavioral alteration induced by Dicrocoelium dendriticum in the Bidstrup Forests, Denmark, in 2016 and
2017. A, shows an ant expressing the biting behavior (on probably Stellaria holostea), and in B a temperature/humidity- logger is seen placed in close proximity
to an ant expressing the altered behavior, biting onto cocksfoot grass, Dactylis glomerata. C with a number tag glued to the dorsal part of gaster (individual B6,
also biting onto a piece of Dactylis glomerata). In D, an assembly of 6 infected ants were found under a big leave (undefined Arctium species), 4 of these with a
number tag on (R33, R63, R76, and a B-tag not readable from this angle).
Page 3 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Behavioral Ecology
Thus, concurrent measures of temperature, RH, and the
number of ants biting the vegetation were recorded at each of the
minimally six observation time points per field day. In August–
October 2016, 172 ants were individually marked (after the final
observation time point of the day, of these 126 ants were observed
again at the same anthills as they were at when tagged) by num-
bered and colored tags (Swienty A/S) glued to the dorsal part
of gaster of the ants observed biting to vegetation (see Figure
2C,D, in-depth description of procedure in supplementary, data
in dataset S3, and further analysis in the section “Persistence of
biting behavior”).
Ants biting to vegetation received a touch stimulus, which induced
slight movement in ants harboring D. dendriticum (Paraschivescu and
Raicev 1980; Gasque and Fredensborg, personal observation) to
distinguish them from infections by the entomopathogenic fungus
Pandora formicae (de Bekker et al. 2018), which also inhabits the
Bidstrup Forests (Małagocka et al. 2015, 2017). P. formicae initially
induces a similar behavior, but infected ants remain motionless.
Orientation of infected ants and infection
patterns
On two separate days (15 September 2016 and 22 September
September 2016, recorded, respectively in the afternoon and in the
morning), the orientation of infected ants biting onto grass blades
was noted. It was noted for each recording whether the ant was
facing downwards toward the roots or facing upwards to the tip of
the grass blade (see Figure 2C).
One hundred and seventy-four ants expressing biting behavior
were collected on dierent dates (Supplementary Table S2) from
the four observed anthills and transported back to the laboratory
in individual 15 mL centrifuge tubes for dissection and verification
of infection. Eighty ants collected randomly by the use of a hand-
kerchief placed on the nest (Botnevik et al. 2016) were used as a
negative control. Ants were stored at 5 °C until they were meas-
ured from mandible to the tip of gaster to the nearest 1 mm and
dissected. Dissections were performed in a 0.9% saline solution
under a Leica MZ12 stereomicroscope (×20), and the presence and
number of metacercariae in the gaster were recorded (for the first
two groups). For the last group collected on 16 August 2017, dissec-
tions were conducted only to verify the infection status of the ants.
Statistical analysis
The number of ants biting onto the vegetation at the dierent ob-
servations were considered independent observations since the
probability of expressing parasite-induced behavior was deter-
mined by the environmental factors at the time of observation
(Supplementary Dataset S1). The number of ants expressing the
altered phenotype fluctuated considerably between sampling days
at a given anthill from a few individuals to 116. We, therefore, used
the proportion of ants biting to vegetation (the number of ants
biting at said time point relative to the maximum number of ants
observed biting to vegetation at the same anthill that day) on each
day as the dependent variable. Temperature, RH, date, an irradi-
ation proxy, and time of day (interval) were used as explanatory
variables.
We do not have an individual ID on the majority of the ants
(only the 126 tagged ants mentioned earlier). Therefore, it was
not possible to include individual ID as an independent variable
in the analysis. With a proportional response variable, the data fol-
lowed a binomial distribution, and we used logistic regression in a
generalized linear model (GENMOD, SAS®) to assess the associ-
ations (see codes used in Supplementary Appendix 1).
Initially, all five explanatory variables were included in the
model (Supplementary Table S1). By including or excluding
date in the model, it was possible to assess the significance level
of the other variables operating within and among dates, respec-
tively. The significance levels of all the variables were compared
between the two runs (including date: Supplementary Table S1
vs. excluding date: Table 1), to estimate which factor was the one
explaining most of the expressed parasite-induced behavior of
the ants within a day as well as across observation dates. By com-
paring the models with AIC, BIC, Deviance, Pearson chi-square,
and more, the dierent tests pointed at the two dierent models
of being the best-fitting model. So, to further analyze which factor
had the biggest influence on the expression of the proportion of
infection in our data set, we performed a regression analysis based
on the Random Forest method in R (v 4.1.2). As the date itself
was not regarded correctly, we instead included it in the format of
year, month, and days per year as separate columns utilizing the
R packages Dplyr (v 1.0.8) and Lubridate (v 1.8.0). The initially
run included all the variables in Supplementary Dataset S1, be-
sides the two used to calculate the proportion of infection (infec-
tion and maximum), as they obviously would skew the test. Lastly,
excluding all the factors used to calculate irradiation proxy. We
utilized the R package Boruta (v 8.0.0) to perform feature selec-
tion and ranking as a regression to predict the proportion of in-
fection, based on 100 iterations of 20,000 trees. R package Caret
(v 6.0-94) was utilized to split train and test datasets, and also train
and test the models to predict the proportion of infection. Train
and test splitting separated the data randomly, with 30% of the
data retained as a testing set. We trained the model with 5-fold
cross-validation repeated 10 times, utilizing forests with 2000 trees
and 4 tuning length options. Model performance was evaluated
with R2 and RMSE.
The number of ants maintaining their attachment throughout
the day (Supplementary Dataset S2) was used as a proportion to
the maximum of ants at each anthill in relation to first; date, ant-
hill, average RH and average temperature, and later solely average
temperature and average RH at each anthill per day, in both cases,
a generalized linear model with binomial errors (GENMOD) was
chosen and tested in SAS®.
The orientation data of ants biting onto grass were analyzed by
chi-square test compared to the equal chance of orientation to ei-
ther direction.
Table 1
Results of generalized linear model examining the eects of
environmental factors and time on the expression of biting
behavior in Dicrocoelium dendriticum-infected Formica
polyctena given as the proportion of ants displaying the
behavior at each observation in relation to the maximum
number of ants observed displaying biting behavior on that day.
“Irradiation-proxy” indicates the solar position on the sky and
“Time of day” representing an interval as day length is divided
into three groups.
Factor Den df Estimate χ2P-value
Temperature 437 −0.1090 120.94 <0.0001
Relative humidity 437 0.0145 26.17 <0.0001
Time of day 437 0.0544 0.50 0.4786
Irradiation-proxy 437 −0.0012 3.28 0.0703
Page 4 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Gasque and Fredensborg · Expression of trematode-induced zombie-ant behavior
Graphs were made in GraphPad Prism (v 9.5.1.), besides
Supplementary Figure S1, which was generated in R and improved
in Inkscape (v 1.2), and Supplementary Figure S2 was generated in
Excel (v 2208).
RESULTS
Factors determining parasite-induced behavior
modification
Temperature (Den df = 437, χ2 = 120.94, P < 0.0001) was strongly
associated with the proportion of infected ants expressing the ma-
nipulated phenotype (Table 1).
As seen from the estimates from the maximum likelihood
parameters in the test excluding the date variable, temperature
explains most of the variation in our dataset (Supplementary
Dataset S1), as a 1 °C increase makes 11% of the ants let go of
their biting (estimate −0.1090, SE ± 0.0106) and when including
date (Supplementary Table S1) in the test the same accounts for
14% of the ants (estimate −0.1384, SE ± 0.0268). Both estimates
are comparable as they fall within the same standard error. With
the comparison between the date-included and date-excluded
runs, temperature was the only variable of high importance in all
tested models, which kept a high χ2 value, and highly significant
P-value (P < 0.0001). In the excluded-date run, relative humidity
(Den df = 437, χ2 = 26.17, P < 0.0001) was highly associated
with the proportion of ants expressing the D. dendriticum infection.
However, RH was not significantly related to ant-biting behavior
within date, and the significant eect of RH observed across dates
was likely an artifact of a negative correlation between RH and
temperature with the latter being the main explanatory factor of
ant-biting behavior (see above). Irradiation-proxy or time of day
(by interval) showed no significant relationship to the induced be-
havioral modification in either of the runs. In a Random Forest
regression modeled to predict the proportion of infection, the
temperature was ranked as the most important factor for predic-
tion (Supplementary Figure S1). The trained model could pre-
dict roughly half of the proportion of infection in the test set
(R2 = 0.5297, RMSE = 0.2029), while a linear regression between
the proportion of infection and temperature presented an R2 of
~0.25 in the full dataset.
Persistence of biting behavior
The number of switches to and from biting behavior was strongly
and positively related to the daily temperature range (dier-
ence between maximum and minimum temperatures) in indi-
vidually marked infected ants across the observation days (Linear
Regression: F = 61.1, df = 1, P < 0.001; Figure 3). Thus, the larger
the dierence between maximum and minimum temperature, the
less persistent the manipulated biting behavior.
Labeled ants that were used in a 24-h study from 22 September
2016 to 23 September 2016 showed that of the numbered ants
biting to vegetation in the evening (N = 30); 96.67% were observed
in the same state at the same location the following morning.
Seasonality
The number of biting ants counted on each observation day varied
greatly (Supplementary Figure S2). However, a seasonal dierence
was observed from summer 2016 to summer 2017 when plotting
the maximum number of infected ants counted on each day from
the two sites (Figure 4). In April 2017, up to five times, the number
of ants biting the vegetation were observed in comparison to the
other recorded months at the same site.
All-day attachment
At the end of September/start of October 2016, we observed that
a great majority (up to 1/3) of the ants stayed in the biting state for
the whole day of observations (Figure 5), in contrast to observations
made earlier in the season (where none remained attached during
all the observation times of the day). When analyzed, the date
showed positive correlation to the proportion of the maintenance
of biting (Num df = 4, χ2 = 20.26, P = 0.0004). Considering only
the influence of the abiotic factors (average temperature and av-
erage RH for the individual anthills), average temperature showed
the highest correlation to the maintenance of biting during the
day (Num df = 1, χ2 = 17.05, P < 0.0001), whereas average RH
showed no significant relationship to the proportion of ants biting
to vegetation.
Orientation of infected ants and infection
patterns
None of the 80 ants randomly collected from the anthills were in-
fected with D. dendriticum metacercaria (Supplementary Table S2).
A majority of the ants (70%) expressing infection by biting on to
vegetation were facing downwards (n = 144, df = 1, χ2 = 23.36,
P < 0.0001). Combining all dissections, a metacercariae prevalence
of 97% (N = 174) was found in ants collected when biting to the
vegetation in the field.
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0510
Daily temperature range (Max.- Min. temp. (°C))
15 20 25
Mean no. of changes in biting behavior
12 40
16
28 33
11 39 34
30
42
8
8
Figure 3
Number of changes in biting behavior (mean ± SEM) of individually
marked Formica polyctena ants infected with Dicrocoelium dendriticum in
relation to the observed temperature range (dierence between maximum
and minimum temperature) on seven (anthill 1-3, closed circle), or five
observation days (anthill 4, open circle) between August and October
2016, in the Bidstrup Forests, Denmark (see Supplementary Dataset S3:
“GasqueFredensborg_2023_S3_Individually_marked_ants”). The number
of individually marked ants observed per site per day is indicated at
each error bar. A significant positive relationship was observed between
temperature range and the number of changes to the manipulated behavior
for both sites (anthill 1-3, R2 = 0.77, anthill 4, R2 = 0.98, combined R2 = 0.86
(regression line: mean number of changes in behavior = 0.484 + (0.0957 ×
temperature range)), all P < 0.001).
Page 5 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Behavioral Ecology
DISCUSSION
There is a need for studies on how parasitic manipulation of an-
imal behavior responds to changes in environmental conditions
(Loreto et al. 2018). However, it is inherently challenging to study
the eects of individual environmental factors on animal beha-
vior under natural conditions, as many of them cannot be con-
trolled and/or may co-vary. In this study, we succeeded to monitor
the biting behavior of the ant host in relation to individual en-
vironmental factors under natural conditions. We usually ob-
served the greatest numbers of ants in the biting state in the early
morning and in the late afternoon/evening, similar to previous
studies (Supplementary Figure S2) (Badie et al. 1973; Badie 1975;
Spindler 1986; Badie and Rondelaud 1988; Manga-González et
al. 2001). However, our analysis clearly demonstrated that temper-
ature, not the time of day itself, explained the switch to a biting
state observed in the manipulated phenotype (biting to vegetation).
Thus, low temperature facilitated the biting behavior while higher
temperatures reduced it (Table 1; Figures 3 and 5). We also found
a significant positive relationship between the daily temperature
range and the number of times infected ants switched biting be-
havior between the manipulated and non-manipulated phenotype
300
250
200
150
Maximal expression
of infection
Date
100
50
0
16/08/2016
19/08/2016
07/09/2016
12/09/2016
15/09/2016
22/09/2016
23/09/2016
26/09/2016
30/09/2016
12/10/2016
06/04/2017
30/04/2017
16/08/2017
site s
site h
Figure 4
Bar chart representing the seasonal variation in the expression of Dicrocoelium dendriticum-induced biting behavior of Formica polyctena from two sampling sites
(site s (anthill 1-3) and anthill 4, from site h) on 13 observation days (16 August 2016 to 16 August 2017) within 1 year at the Bidstrup Forests, Denmark. The
bars represent the maximum counted number of ants expressing the manipulated state at any of the counting time points during each date as a total for both
the sites (for site h: one number, for site s: maxima for anthill 1-3 added together) (see Supplementary Dataset S1: “GasqueFredensborg_2023_S1_Abiotic_
factors”). In 2 days of April, at the h site, no ants were observed.
100
80
60
40
20
0
100
80
60
40
20
0
15/9/2016
22/9/2016
26/9/2016
30/9/2016
12/10/2016
64.82
24.44
16.64 17.29
13.69
21.43
32.43
77.81
78.33
Average RH (%)
Average temperature (°C)
Percentage of attachment (%)
78.84
68.45
8.64
9.6
3.74
0
Date
Figure 5
The percentage of maintained all-day attachment of Dicrocoelium dendriticum infected Formica polyctena ants (%, filled dark gray line) to the total of ants observed
at all the anthills per day, throughout the observation time points of the assigned days (15 September 2016 to 12 October 2016), in relation to the average
temperature (°Celsius) (dashed light gray line) and the average relative humidity (RH, %) (dotted line) measured in the microclimate of the ants that day (see
Supplementary Dataset S2: “GasqueFredensborg_2023_S2_All-day_attachment”). The exact values of the three parameters for each day are given near the
lines. On earlier dates (16 August 2016 to 15 September 2016), the percentage of all-day attachment was also 0%, as shown for 15 September 2016. (N total;
15/9: N = 166, 22/9: N = 187, 26/9: N = 177, 30/9: N = 168 and 12/10: N = 74).
Page 6 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Gasque and Fredensborg · Expression of trematode-induced zombie-ant behavior
(Figure 3). This indicates that the persistence of the biting beha-
vior was greatly reduced when ants experienced large variations
in the temperatures during a day. The results from this field study
therefore corroborate with results from a previous laboratory-
based study, in which temperature determined the probability
that infected ants were found biting to grass leaves (Botnevik et al.
2016).
In further support of temperature as the main driver of the in-
duced biting behavior, the proportion of infected ants maintaining
all-day attachment significantly increased from the middle of
September until the middle of October concurrent with a decrease
in daily average temperature (Figure 5). Thus, in October, infected
ants were more likely to remain attached to the vegetation the en-
tire day when the maximum temperature was 11.5 °C. Termination
of the parasite-induced behavior, therefore, seems to be determined
by temperature, and not by time of day, RH, or a proxy for irra-
diation. Similarly, as indicated by our 24-h study, ants presumably
remain attached all night. It is from late evening to early morning
that we measured the lowest temperatures. In April 2017, there
were up to five times as many ants observed exhibiting the manipu-
lated phenotype than in other months (Figure 4), suggesting that
infected ants survive the winter in the nest and continue to express
the manipulated behavior the following spring as transmission be-
tween snail and ant host primarily takes place during late spring
and summer (Tarry 1969; Manga-González and González-Lanza
2005). Interestingly, late May to early June is when female roe
deer express the highest total daily activity time (56.7% of the day,
Cederlund 1989) in central Sweden (similar latitude to our study).
The largest proportion of infected ants biting may therefore, not
only coincide with the daily activity of potential definitive hosts
(Stache et al. 2012) through the indication by temperature but po-
tentially also be synchronized with the seasonal activity of poten-
tial other definitive hosts (Cederlund 1989). Temperature seems to
be an appropriate indicator of parasite transmission success and,
thus, parasite fitness, as it relates to both host longevity and the en-
counter possibility with potential definitive hosts. Hence, biting to
vegetation could be lethal to ants and, thereby the D. dendriticum
metacercariae at peak temperatures. In addition to host mortality,
temperature also provides an indicator for the most opportune daily
transmission window to potential definitive hosts since low temper-
ature coincides with the crepuscular activity of herbivorous defini-
tive hosts, for example, the European roe deer (Stache et al. 2012).
The exact mechanism linking temperature to climbing and
biting behavior of D. dendriticum-infected ants remains unknown.
However, it is known that one or few metacercariae lodge at the
suboesophagal ganglion in the ant host (Romig et al. 1980; Martín-
Vega et al. 2018, see arrow in Figure 1), which controls the initiation
of mandibular abductor and adductor movement (Chapman 2013).
Changes in temperature might aect parasite or host secretion of
molecules, an inflammatory response or the mechanical pressure of
the parasite on mandibular nervous tissue may in turn provoke a
biting behavior at low-temperature conditions. Penetration of mus-
cles by the fungus O. kimflemingiae may lead to the hypercontractions
seen in the mandibles of infected carpenter ants (Mangold et al.
2019). However, in contrast to behavior-manipulating tremat-
odes and viruses (termed neuroparasites; Hughes and Liebersat
2018), hypocrealean fungi do not invade the CNS of the living
host (Fredericksen et al. 2017), whereas recent evidence suggests
that entomophthorelean species can infiltrate the CNS tissue while
the host is still alive (Elya et al. 2018, 2023). This all suggests that
similar behaviors may be regulated by dierent underlying mech-
anisms across parasite phyla (Mangold and Hughes 2021).
In accordance with previous studies, we also found that most
of the infected ants were facing head downwards (Badie and
Rondelaud 1988; Manga-González et al. 2001). The reason is un-
known, although we wonder if the final orientation is the result of
a method for seeking out the optimal conditions of the attachment
for the trematode-host or to provide shading at least for the head
part, where the SOG-attached metacercaria is located (Gullan and
Cranston 2010).
Our analysis indicates that the eect of RH on D. dendriticum-
infected ants expressing biting behavior in the field is smaller than
that of temperature in line with previous studies, conducted under
controlled laboratory conditions where no interaction between rel-
ative humidity and biting behavior could be detected (Botnevik et
al. 2016).
We indirectly assessed the eect of solar irradiation, by including
a proxy for solar position in the analysis, on the biting behavior of
ants infected with D. dendriticum. We did not find any evidence that
the solar position had an eect on biting behavior, similar to a study
on the eect of artificial light (Botnevik et al. 2016). Light plays an
important role in other systems, where parasites initially induce the
same attachment to foliage, such as fungi from the genera Pandora
and Ophiocordyceps. O. unilateralis s.l., makes the infected Camponotus
leonardi ants attach approximately 25 cm above the forest floor,
which should be the optimal microclimate for the post-mortem
development of the stalk and subsequent spore release (Andersen
et al. 2009; Hughes et al. 2011). Transition to the so-called death
grip in naturally infected Camponotus ants is synchronized to solar
noon in the field (Hughes et al. 2011; Will et al. 2020), and com-
parisons between O. camponoti-atricipis-infected Camponoti atricipis
ants in shaded and control areas indicated a strong positive pho-
totactic influence (Andriolli et al. 2019). These results combined
suggest that illumination is an essential cue for fungus-infected ants
to locate the optimal microclimate for progeny development and
dispersal (Chung et al. 2017). Light has a pivotal role, not only
in fungal systems but also for the expression of other parasite-
induced behavioral alterations. Studies on aquatic gammarids in-
fected with larval trematodes or acanthocephalans, demonstrated
positive phototaxis of infected individuals, presumably increasing
predation of infected gammarids by duck final hosts (Bethel and
Holmes 1973; Helluy 1983). In that system, temperature had no
eect on parasite-induced host behavior manipulation since light
and not temperature presumably determined the encounter possi-
bility with the definitive host (Labaude et al. 2020). For both of the
baculoviral-induced phenotypic changes (hyperactivity and tree-top
disease) the expressions are influenced by light (Kamita et al. 2005;
van Houte et al. 2014; Bhattarai et al. 2018). Therefore, in future
field studies, it would be relevant to test the eect of light inten-
sity on the timing of behavioral alteration in ants infected with D.
dendriticum, including a broader range of seasons.
For most host–parasite systems, the underlying mechanisms
of induced parasitic manipulation, are to a great extend un-
known. The same accounts for the D. dendriticum-For mica ant host
sytem. By comparing similarly expressed behavioral manipula-
tions across phyla, we speculate that the parasite-encoded protein
tyrosine phosphatase (PTP) could be a potential regulating factor
in the D. dendriticum case. PTP is linked to a change of move-
ment and positive phototaxis, which is induced in hosts by sev-
eral manipulating parasites. This enzyme has been found to play
Page 7 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Behavioral Ecology
a role in the induction of hyperactivity in caterpillars when in-
fected with baculoviruses (Kamita et al. 2005; Katsuma et al. 2012;
van Houte et al. 2012). Fungal-encoded PTP is also upregulated
during the manipulated state (deathgrip) of Ophidiocordyceps-infected
ants (de Bekker et al. 2015, 2017; Will et al. 2020). Since the in-
itial part of the Ophiocordyceps- and Dicrocoelium-induced behaviors
in the ant host (the at minimum negative geotaxism (or positive
phototaxis in Ophiocordyceps case) and the attachment to the vege-
tation by mouthparts) are similar to each other, PTP could have
evolved convergently (Will et al. 2020). Currently, it is unknown
whether the phenotypical similarities of infected host behavior
have evolved independently across a taxonomically diverse range
of parasites, or if they have evolved to exploit the same ancient
host trait of biting to vegetation while sleeping (Lovett et al. 2020).
In any case, investigating the possible role of PTP and other actors
in the manipulated state of ants infected with Dicrocoelium would
be very interesting to include in future quantitative studies such as
transcriptomic or proteomic approaches (Biron et al. 2005, 2006;
Małagocka et al. 2015; de Bekker et al. 2017).
CONCLUSION
This study provided rare field evidence on the eects of environ-
mental factors on parasite-induced behavioral changes in the host.
We found that temperature was the driving factor influencing
parasite-induced ant-biting behavior in the field. We propose that
temperature sensitivity is an adaptation to increase transmission
to the definitive mammalian host, while at the same time pro-
tecting the intermediate host from exposure to lethal temperatures.
Investigating the biochemical underlying mechanisms and com-
paring these to other parasites inducing similar behavior in ants
and other insects will be an important next step.
SUPPLEMENTARY MATERIAL
Supplementary material can be found at http://www.beheco.
oxfordjournals.org/
We thank Per Moestrup Jensen for fruitful theoretical discussions and for
statistical assistance, the management at Bidstrup Forests for access to
fieldsites, Pedro Beschoren da Costa for cross-validation of GLM test results
by application of the Random Forest method and Vera I. D. Ros and Julia
Friman for comments on earlier drafts of the manuscript. Lastly, we also
thank two anonymous reviewers whose constructive criticism improved the
quality of the manuscript.
FUNDING
This work was supported by Villum Fonden (grant number 00007457).
DATA AVAILABILITY
Analyses reported in this article can be reproduced using the data provided
by Gasque SN, Fredensborg BL. 2023a (Dataset S1, Abiotic factors), Gasque
SN, Fredensborg BL. 2023b (Dataset S2, All-day attachment) and Gasque
SN, Fredensborg BL. 2023c (Dataset S3, Individually_marked_ants).
Handling Editor: RobinTinghitella
REFERENCES
Andersen SB, Gerritsma S, Yusah KM, Mayntz D, Hywel-Jones NL, Billen
J, Boomsma JJ, Hughes DP. 2009. The life of a dead ant: the expression
of an adaptive extended phenotype. Am Nat. 174:424–433.
Andriolli FS, Ishikawa NK, Vargas-Isla R, Cabral TS, de Bekker C, Baccaro
FB. 2019. Do zombie ant fungi turn their hosts into light seekers? Behav
Ecol. 30:609–616. doi:10.1093/beheco/ary198.
Badie A. 1975. Cycle annuel d’activité des fourmis parasitées par les
métacercaires de Dicrocoelium lanceolatum (Ruddolphi 1819). Ann Rech
Vet. 6:259–264.
Badie A, et Rondelaud D. 1988. Les fourmis parasitées par Dicrocoelium
lanceolatum, Rudolphi en limousin. Les relation avec le support végétal.
Rev Méd Vét. 6:629–633.
Badie A, Vincent M, Morel-Varieille C, Rondelaud D. 1973. Cycle de
Dicrocoelium dendriticum (Rudolphi, 1819) en Limousin:Ethologie des
fourmis parasitees par les metacercaires. Comptes Rendus des Se´ ances
de la Socie´ te´ de Biologie 167:725–727.
Bethel WM, Holmes JC. 1973. Altered evasive behavior and responses to
light in amphipods harboring acanthocephalan cystacanths. J Parasitol.
59(6):945–956.
Bhattarai UR, Li F, Katuwal Bhattarai M, Masoudi A, Wang D. 2018.
Phototransduction and circadian entrainment are the key pathways in the
signaling mechanism for the baculovirus induced tree-top disease in the
lepidopteran larvae. Sci Rep. 8:17528. doi:10.1038/s41598-018-35885-4.
Biron DG, Marché L, Ponton F, Loxdale HD, Galéotti N, Renault L,
Jolyland C, Thomas F. 2005. Behavioural manipulation in a grasshopper
harbouring hairworm: a proteomics approach. Proc R Soc B. 272:2117–
2126. https://www.jstor.org/stable/30047795
Biron DG, Ponton F, Marché L, Galéotti N, Renault L, Demey-Thomas E,
Poncet J, Brown SP, Jouin P, Thomas F. 2006. “Suicide” of crickets har-
bouring hairworms: a proteomics investigation. Insect Mol Biol. 15:731–
742. doi:10.1111/j.1365-2583.2006.00671.x.
Botnevik C, Jensen AB, Fredensborg BL. 2016. Relative eects of tempera-
ture, light, and humidity on clinging behavior of metacercariae-infected
ants. J Parasitol. 102:495–500.
Cederlund G. 1989. Activity patterns in moose and roe deer in a north bo-
real forest. Ecography 12:39–45.
Chapman RF. 2013. The insects: Structure and Function. 5th edition.
Cambridge: Cambridge University Press.
Chung T-Y, Sun P-F, Kuo J-I, Lee Y-I, Lin C-C, Chou J-Y. 2017. Zombie
ant heads are oriented relative to solar cues. Fungal Ecology 25:22–28.
Combes C. 2001. Parasitism: The ecology and evolution of intimate inter-
actions. Chicago (IL): University of Chicago Press.
Criscione CD, van Paridon BJ, Gilleard JS, Goater CP. 2020. Clonemate
cotransmission supports a role for kin selection in a puppeteer par-
asite. Proc Natl Acad Sci USA. 117(11):5970–5976. doi:10.1073/
pnas.1922272117.
de Bekker C, Ohm RA, Loreto RG, Sebastian A, Albert I, Merrow M,
Brachmann A, Hughes DP. 2015. Gene expression during zombie
ant biting behavior reflects the complexity underlying fungal para-
sitic behavioral manipulation. BMC Genomics. 16:620. doi:10.1186/
s12864-015-1812-x.
de Bekker C, Will I, Das B, Adams RMM. 2018. The ants (Hymenoptera:
Formicidae) and their parasites: eects of parasitic manipulations and
host responses on ant behavioral ecology. Myrmecological News 28:1–24.
doi:10.25849/myrmecol.news_028:001.
de Bekker C, Will I, Hughes DP, Brachmann A, Merrow M. 2017.
Daily rhythms and enrichment patterns in the transcriptome of the
behavior-manipulating parasite Ophiocordyceps kimflemingiae. PLoS One.
12:e0187170. doi:10.1371/journal.pone.0187170.
Elya C, Lavrentovich D, Lee E, Pasadyn C, Duval J, Basak M, Saykina V, de
Bivort B. 2023. Neural mechanisms of parasite-induced summiting beha-
vior in “zombie” Drosophila. eLife. 12:e85410. doi:10.7554/eLife.85410.
Elya C, Lok TC, Spencer QE, McCausland H, Martinez CC, Eisen M.
2018. Robust manipulation of the behavior of Drosophila melanogaster
by a fungal pathogen in the laboratory. Elife. 7:e34414. doi:10.7554/
eLife.34414.
Fredericksen MA, Zhang Y, Hazen ML, Loreto RG, Mangold CA, Chen
DZ, Hughes DP. 2017. Three-dimensional visualization and a deep-
learning model reveal complex fungal parasite networks in behavior-
ally manipulated ants. Proc Natl Acad Sci USA. 114:12590–12595.
doi:10.1073/pnas.171167311.
Fritz SR. 1982. Selection for host modification by insect parasitoids.
Evolution. 36:283–288.
Froelick S, Gramolini L, Benesh DP. 2021. Comparative analysis of hel-
minth infectivity: growth in intermediate hosts increases establishment
rates in the next host. Proc R Soc B. 288:20210142. doi:10.1098/
rspb.2021.0142.
Page 8 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023

Gasque and Fredensborg · Expression of trematode-induced zombie-ant behavior
Gasque SN, Fredensborg BL. 2023a. Expression of trematode-induced
zombie-ant behavior is strongly associated with temperature. Behav Ecol.
doi:10.5061/dryad.44j0zpckp.
Gasque SN, Fredensborg BL. 2023b. Expression of trematode-induced
zombie-ant behavior is strongly associated with temperature. Behav Ecol.
doi:10.5061/dryad.np5hqc007.
Gasque SN, Fredensborg BL. 2023c. Expression of trematode-induced
zombie-ant behavior is strongly associated with temperature. Behav Ecol.
doi:10.5061/dryad.j9kd51cj0.
Gasque SN, Poinar GO, Fredensborg BL. 2018. Ant (hymenoptera:
formicidae) parasitism by neoneurinae wasps (hymenoptera: braconidae)
in Denmark. Ent Medd. 86:27–30. ISSN: 0013-8851
Gasque SN, van Oers MM, Ros VID. 2019. Where the baculoviruses lead,
the caterpillars follow: baculovirus-induced alterations in caterpillar be-
haviour. Curr Opin Ins Sci. 33:30–36.
Gullan PJ, Cranston PS. 2010. The Insects: An Outline of Entomology, 4th
edition. Chichester (UK): Wiley-Blackwell.
Helluy S. 1983. Host–parasite relations of the trematode Microphallus
papillorobustus (Rankin, 1940). II—Changes in the behavior of Gammarus
intermediates hosts and localization of metacercaria. Ann Parasit Hum
Comp. 58:1–17.
Hughes DP, Andersen S, Hywel-Jones NL, Himaman W, Bilen J, Boomsma
JJ. 2011. Behavioral mechanisms and morphological symptoms of
zombie ants dying from fungal infection. BMC Ecol. 11:13.
Hughes DP, Liebersat F. 2018. Neuroparasitology of parasite–insect associ-
ations. Annu Rev Entomol. 63:471–487.
Kamita SG, Nagasaka K, Chua JW, Shimada T, Mita K, Kobayashi M,
Maeda S, Hammock BD. 2005. A baculovirus-encoded protein tyro-
sine phosphatase gene induces enhanced locomotory activity in a lepi-
dopteran host. Proc Natl Acad Sci USA. 102:2584–2589. doi:10.1073/
pnas.0409457102.
Katsuma S, Koyano Y, Kang WK, Kokusho R, Kamita SG, Shimada T.
2012. The baculovirus uses a captured host phosphatase to induce en-
hanced locomotory activity in host caterpillars. PLoS Pathog. 8:e1002644.
doi:10.1371/journal.ppat.1002644.
Labaude S, Cézilly F, De Marco L, Rigaud T. 2020. Increased temperature
has no consequence for behavioral manipulation despite eects on both
partners in the interaction between a crustacean host and a manipulative
parasite. Sci Rep. 10:11670. doi:10.1038/s41598-020-68577-z.
Laerty KD, Allesina S, Arim M, Briggs CJ, De Leo G, Dobson AP,
Dunne JA, Johnson PTJ, Kuris AM, Marcogliese DJ, et al. 2008.
Parasites in food webs: the ultimate missing links. Ecol Lett. 11:533–546.
doi:10.1111/j.1461-0248.2008.01174.x.
Laerty KD, Shaw JC. 2013. Comparing mechanisms of host manipulation
across host and parasite taxa. J Exp Biol. 216:56–66.
Loreto RG, Araujo JPM, Kepler RM, Fleming KR, Moreau CS, Hughes
DP. 2018. Evidence for convergent evolution of host parasitic manipula-
tion in response to environmental conditions. Evolution. 72:2144–2155.
doi:10.1111/evo.13489.
Lovett B, Macias A, Stajich JE, Cooley J, Eilenberg J, de Fine Licht HH,
Kasson MT. 2020. Behavioral betrayal: How select fungal parasites
enlist living insects to do their bidding. PLoS Pathog 16:e1008598.
doi:10.1371/journal.ppat.1008598.
Małagocka J, Grell MN, Lange L, Eilenberg J, Jensen AB. 2015.
Transcriptome of an entomophthoralean fungus (Pandora formicae) shows
molecular machinery adjusted for successful host exploitation and trans-
mission. J Invertebr Pathol. 128:47–56. doi:10.1016/j.jip.2015.05.001.
Małagocka J, Jensen AB, Eilenberg J. 2017. Pandora for micae, a specialist ant
pathogenic fungus: new insights into biology and taxonomy. J Invertebr
Pathol. 143:108–114.
Manga-González MY, González-Lanza C. 2005. Field and experimental
studies on Dicrocoelium dendriticum and dicrocoeliasis in northern Spain. J
Helminthol. 79:291–302.
Manga-González MY, González-Lanza C, Cabanas E, Campo R. 2001.
Contributions to and review of dicrocoeliosis, with special refer-
ence to the intermediate hosts of Dicrocoelium dendriticum. Parasitology.
123:91–114.
Mangold CA, Hughes DP. 2021. Insect behavioral change and the po-
tential contributions of neuroinflammation—a call for future research.
Genes. 12:465. doi:10.3390/genes12040465.
Mangold CA, Ishler MJ, Loreto RG, Hazen ML, Hughes DP. 2019.
Zombie ant death grip due to hypercontracted mandibular muscles. J
Exp Biol. 222:jeb200683. doi:10.1242/jeb.200683.
Mapes CR, Krull WH. 1952. Studies on the biology of Dicrocoelium
dendriticum (Rudolphi, 1819) looss, 1899 (Trematoda: Dicrocoeliidae),
including its relation to the intermediate host, Cionella lubrica (Müller).
VII. The second intermediate host of Dicrocoelium dendriticum. The
Cornell Veterinarian 42:603–604.
Martín-Vega D, Garbout A, Ahmed F, Wicklein M, Goater CP,
Colwell DD, Hall MJR. 2018. 3D virtual histology at the host/para-
site interface: visualization of the master manipulator, Dicrocoelium
dendriticum, in the brain of its ant host. Sci Rep. 8:8587. doi:10.1038/
s41598-018-26977-2.
Moore J. 2002. Parasites and the behavior of animals. Oxford (UK):
Oxford University Press. p. 315.
Paraschivescu D, Raicev C. 1980. Experimental ecological investigations
on the tetany of the species Formica pratensis complementary host of
the trematode Dicrocoelium dendriticum. Travaux du Muse´ um d’histoire
naturelle “Grigore Antipa” 22:299–302.
Parker GA, Ball A M, Chubb JC, Hammerschmidt K, Milinski M. 2009.
When should a trophically transmitted parasite manipulate its host?
Evolution 63:448–458.
Poulin R. 1995. “Adaptive” changes in the behaviour of parasitized ani-
mals: a critical review. Int J Parasitol. 25:1371–1383.
Poulin R. 2007. Evolutionary ecology of parasites. Princeton (NJ):
Princeton University Press.
Romig T, Lucius R, Frank W. 1980. Cerebral larvae in the second inter-
mediate host of Dicrocoelium dendriticum (Rudolphi, 1819) and Dicrocoelium
hospes Looss, 1907 (Trematodes, Dicrocoeliidae). Z Parasitenkd.
63:277–286.
Spindler E, Zahler M, Loos-Frank B. 1986. Behavioural aspects of ants
as second intermediate hosts of Dicrocoelium dendriticum. Z Parasitenkd.
72:689–692.
Stache A, Heller E, Hothorn T, Heurich M. 2012. Activity patterns of
European roe deer (Capreolus capreolus) are strongly influenced by in-
dividual behavior. Fol. Zoologica. 62:67–75.
Tarry DW. 1969. Dicrocoelium dendriticum: the life cycle in Britain. J
Helminthol. 43:403–416.
van Houte S, Ros VID, Mastenbroek TG, Vendrig NJ, Hoover K, Spitzen
J, van Oers MM. 2012. Protein tyrosine phosphatase-induced hyperac-
tivity is a conserved strategy of a subset of baculoviruses to manipulate
lepidopteran host behavior. PLoS One. 7:e46933. doi:10.1371/journal.
pone.0046933.
van Houte S, van Oers MM, Han Y, Vlak JM, Ros VID. 2014. Baculovirus
infection triggers a positive phototactic response in caterpillars to in-
duce “tree-top” disease. Biol Lett. 10:20140680.
Will I, Das B, Trinh T, Brachmann A, Ohm R, de Bekker C. 2020.
Genetic underpinnings of host manipulation by Ophiocordyceps
as revealed by comparative transcriptomics. G3. 10:2275–2296.
doi:10.1534/g3.120.401290.
Page 9 of 9
Downloaded from https://academic.oup.com/beheco/advance-article/doi/10.1093/beheco/arad064/7250300 by guest on 13 October 2023