Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
Available online 4 March 2024
2213-2244/© 2024 The Authors. Published by Elsevier Ltd on behalf of Australian Society for Parasitology. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Flea infestation of rodent and their community structure in frequent and
non-frequent plague outbreak areas in Mbulu district, northern Tanzania
Stella T. Kessy
a
,
b
,
d
,
*
, RhodesH. Makundi
b
,
c
, Apia W. Massawe
b
,
c
, Alfan A. Rija
a
a
Department of Wildlife Management, Sokoine University of Agriculture, P.O. Box 3073, CHUO KIKUU, Morogoro, Tanzania
b
The African Centre of Excellence for Innovative Rodent Pest Management and Biosensor Technology Development (ACE IRPM&BTD), Tanzania
c
Institute of Pest Management, Sokoine University of Agriculture, P. O. Box 3110, Morogoro, Tanzania
d
School of Life Science and Bio-Engineering (LiSBE), Nelson Mandela African Institution of Science and Technology, P. O. Box 447, Arusha, Tanzania
ARTICLE INFO
Keywords:
Plague
Flea abundance
Flea community
Flea-rodent interactions
ABSTRACT
Understanding rodent-ectoparasite interactions and the factors driving them is important in understanding the
epidemiology of diseases involving an arthropod vector. Fleas are the primary vector for Yersinia pestis, the
bacteria that causes plague and monitoring of ea population is essential for planning the potential mitigation
measures to prevent the disease outbreak. In this study, we investigated ea abundance, community structure
and the potential factors driving ea infestation in areas with frequent (persistent) and non-frequent plague (non-
persistent) outbreaks. We collected eas from captured rodents in two villages with both forest and farm hab-
itats. We found 352 eas belonging to 5 species with Dinopsyllus lypusus the most abundant overall (57.10%) and
Ctenophthalmus spp. the lowest (1.70%). There were no signicant differences of ea abundance between study
localities, habitats and seasons (p >0.05) but, ea infestation was signicantly positively associated with the
persistent locality and with the short rain season (p <0.05). Further, ea abundance increased signicantly with
rodent body weight (p <0.05). Furthermore, we found eas broadly structured into two communities varying
between the dry, long rain and short rain seasons. These ndings have important implications for public health,
as they may be used to assess and control the risks of plague transmission and other ea borne diseases in the
foci.
1. Introduction
Fleas are bloodsucking insects with signicant implications for
human and animal health worldwide (Bitam et al., 2010). Fleas infest a
wide range of hosts including wild and domestic animals, birds and
human (Durden and Hinkle, 2019; Zając et al., 2020; Zurita et al., 2019).
Infestation is inuenced by environmental and human behavior modi-
cations. For instance, when farmers share their dwellings with live-
stock or have corrals located in close proximity to their homes, it exposes
domestic animals and humans to infestation, leading to the transmission
of ea-borne diseases. Further, activities such as urbanization, defores-
tation, and encroachments into natural habitats, increase interactions
between human and ea-infested environments that may also increase
the risk of exposure to ea-borne pathogens (Gage et al., 2008; Bitam
et al., 2010). Fleas are well known vectors of several illness including
murine typhus caused by Rickettsia typhi, ea-borne spotted fever
caused by Rickettsia felis, cat scratch disease caused by Bartonella hen-
selae, and bubonic plague caused by Yersinia pestis (Krasnov, 2008;
Durden and Hinkle 2019; Sherman 2007). Furthermore, some eas such
as the human eas, act as vector for tape worms (Kandi et al., 2019;
Ramana et al., 2011) and pose signicant public health concerns. In
regions with sporadic ea-transmitted disease outbreaks, such as plague,
the absence of up-to-date information on ea dynamics and host infes-
tation intensies these concerns. Access to ssuch data could inform the
development of strategies to counter potential outbreaks through, for
example, targeting on reducing the population of eas and rodents.
Several factors are known to inuence ea richness and abundance
including; host diversity (Krasnov et al., 2002; Young et al., 2015), host
body condition (Bitam et al., 2010; Krasnov, 2008), host density
(Krasnov et al., 2002; Stanko et al., 2002) and climatic conditions
(Krasnov et al., 2004, 2005). However, it is not clear how such factors
are directly linked to plague persistence especially in regions with
* Corresponding author. Department of Wildlife Management, Sokoine University of Agriculture, P.O. Box 3073, CHUO KIKUU, Morogoro, Tanzania.
E-mail addresses: kessystella78@gmail.com (S.T. Kessy), rmakundi@yahoo.com (RhodesH. Makundi), apiamas@yahoo.com (A.W. Massawe), al.rija10@gmail.
com (A.A. Rija).
Contents lists available at ScienceDirect
International Journal for Parasitology: Parasites and Wildlife
journal homepage: www.elsevier.com/locate/ijppaw
https://doi.org/10.1016/j.ijppaw.2024.100921
Received 5 December 2023; Received in revised form 1 March 2024; Accepted 2 March 2024
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
2
history of disease outbreaks. Thus, understanding ea density, infesta-
tion and community structure in the plague foci may allow us to easily
predict transmission risks of ea borne diseases among co-existing
sympatric hosts. Plague is a zoonotic disease that is largely spread by
eas from rodents to humans (Gage and Kosoy, 2005). The disease
continues to be a public health concern, with over 90% of all reported
human cases worldwide originating from Sub-Saharan Africa and the
Madagascar region (Bertherat and Bertherat, 2019; Vall`
es et al., 2020).
In Tanzania, plague has been reported in several districts including
Lushoto, Karatu and Mbulu and remains a signicant potential health
risk in case of outbreak. Studying host-parasite interactions therefore
may help us to understand the risk of both persistence and outbreak of
plague. The transmission of the bacteria causing plague (Yersinia pestis),
is inuenced by various factors, including ea density in the environ-
ment (Krasnov et al., 2006a; Pham et al., 2009; Tripp et al., 2009).
Plague tends to persist in a particular locale or region when multiple
eas capable of transmitting Y. pestis infest hosts susceptible to plague
infection (Eisen and Gage, 2009), thus making the disease more or less
predictable based on known pre-disposing causes. Additionally, re-
searchers have developed statistical models and used ecological data to
predict the occurrence and distribution of plague in various regions. For
instance, Eisen et al., (2007) used a GIS-based model to predict the
habitat suitability for Yersinia pestis, in New Mexico, nding that 30.8%
of the state as suitable plague habitat, Similarly, Neerinckx et al., 2008
used ecological niche modelling (ENM) to predict the potential distri-
bution of plague occurrences across sub-Saharan Africa based on envi-
ronmental variables and occurrence data. They identied elevation,
potential evapotranspiration, mean diurnal temperature range, annual
rainfall, and Normalized Difference Vegetation Index contributing to the
plague occurrences in Sub-Saharan Africa. Furthermore, Poje et al.,
2020 studying ea populations in black-tailed dog burrow in North
America, found that the likelihood of prairie dog burrow being infested
with eas increased with high temperatures, while the prevalence of
infested burrow declined with increased winter precipitation. This, in
turn, impacted the dynamics of plague in prairie dog colonies.
Several studies have reported disease persistence and transmission
conditions in Mbulu districts, Tanzania (Makundi et al., 2008; Ziwa
et al., 2013). High ea diversity and rodent hosts richness, with a
multiple host-ea interaction in different habitats are variables that
contribute to plague persistence in this focus (Makundi et al., 2015). A
more recent study has shown plague bacteria continues to circulate
among susceptible rodents in Mbulu district (Haikukutu et al., 2022),
suggesting potential risks of plague outbreak. These studies suggest that
regular monitoring and updating data on the ea-rodent interactions
and the likely factors driving potential outbreaks and disease persistence
are important to control the disease in these rural communities. This can
be achieved through public awareness campaigns and educational pro-
jects that inform and educate residents about lifestyle practices that
encourage ea-rodent-human interaction. Additionally, community
engagement is crucial, with health ofcers visiting local communities to
identify possible risks and provide valuable guidance as well as devel-
oping strategies that target both ea vector and rodent hosts.
In this study, we aimed to provide current information on the ea
infestation of rodents, their community structure and how infestation
varied between plague persistent and non-persistent foci in Mbulu dis-
trict, northern Tanzania. Specically, we (i) assessed rodent ea abun-
dance in different habitats, seasonality, and localities contrasting in
plague outbreaks, (ii) examined which factors inuence prevalence of
ea infestation, (iii) assessed the effect of habitats, seasons, temperature,
humidity and rodent species traits (sex, sex condition, species ID,
weight) on overall ea abundance and, (iv) assessed how ea commu-
nity structures between localities in different habitats and its potential
hosts. We predicted that ea load would be greater in plague persistent
than non-persistent localities and we predicted that ea abundance and
infestation would be positively associated with seasonality and plague
persistent locality due to available hosts and suitable habitats and
environmental conditions that would provide ea population growth.
Finally, we predicted that ea species would be structured according to
similar resources such as blood meals from host animals, microclimate
conditions, and habitats use, and that some ea species should show host
preferences while others exhibit host sharing pattern between multiple
hosts, providing conducive environment for the disease enzootic
circulation.
2. Materials and methods
2.1. Study area
This study was conducted in two villages, Mongahay (04◦03
′
S, 35◦
26
′
E) and Endesh-Arri (04◦03
′
S, 35◦27
′
E) located in Mbulu District,
Manyara Region in Northern Tanzania from Jan 2019 to Dec 2019
(Fig. 1). The villages were chosen based on the plague outbreak history
and presence of plague pathogen in the rodent population (Makundi
et al., 2008; Ziwa et al., 2013; Mwalimu et al., 2022). Villages with and
without human plague cases were purposefully selected in consultation
with village leaders. Villages with a history of bubonic plague cases were
identied as ’plague persistent’ (Endeshi village), while those without a
history of bubonic plague were identied as ’non-persistent’ (Mongahay
village). Both villages engaged in crop farming and livestock keeping as
their primary economic activities.
The district where the study villages are lies between 1000 and
2400m above sea level and is characterized by bimodal rainfall pattern,
with a long rainy season between March and May, and a short rainy
season between November and January (Nyembo et al., 2021). The short
rain season is characterized by sporadic and light rainfall, which is less
predictable. During the short rain season the mean temperature was on
average 16.84 ◦C (SE =0.13). On the other hand, the long rain season is
characterized by cloudy skies and heavy rainfall, with mean tempera-
ture of 14.79 ◦C (SE =0.12).
2.2. Rodent trapping
Rodents were live trapped using Sherman traps (LFA 7.6 x 8.9 ×23
cm, H.B. Sherman Trap, Inc., Tallahassee, USA) baited with peanut
butter mixed with maize our. Five transect lines with 10 trapping
stations set 10 m apart were established in the natural forest (natural
forest) and farmland (mixed farming) habitats in each village (Kessy
et al., 2023). Traps were left overnight and inspected each morning for
three days. Trapping was conducted every month for 15 months be-
tween Jan 2019 to Dec 2019.
Captured animals were anaesthetized with diethyl ether for immo-
bilization (Palomino et al., 2020). Morphological measurements
(weight, head body length, tail length and ear length) and other char-
acteristics of each captured animal (sex and reproductive status) were
recorded. Sex condition were noted as indicators of reproductive status
of the host species i.e. the position of the testes, vagina and nipples.
Females were classied as virginal perforated (PSN), perforated and
lactating (PLY), virginal closed (CSN), perforated small nipple with
young ones (PSY) and perforated lactating not pregnant (PLN). Males
were classied as scrotal visible (SV) as active males and testes were
abdominal (AN) as non-active male (Makundi et al., 2007). Rodents
were identied to species level using Happold (2013) and conrmed by
sequencing the mitochondrial cytochrome b gene at the Institute of
Vertebrate Biology, Czech Republic.
2.3. Flea collection
Rodents were removed from the holding bag and carefully brushed in
a pan to remove eas. Each bag was thoroughly checked to remove
dislodged eas and the tray was examined carefully with a hand lens to
remove all ectoparasites using a moistened paint brush.
Fleas were grouped based on locality, habitat, month, and host
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
3
species and were counted and preserved in 70% ethanol for future
identication. The eas were then processed by adopting a modied
version of the method described in (Philip Samuel et al., 2021). Briey,
each group was exposed to NAOH 10%, dehydrated in various concen-
trations of ethanol (50%, 70%, 95%, absolute), cleared in clove oil,
temporarily mounted using glycerin on a microscopic slide, and exam-
ined under a light microscope using a 10x objective.
To understand how local climatic parameters inuence eas in the
area, rain data were measured and recorded using an ordinary rain
gauge installed outside Mongahay village ofce between Jan 2019 to
Dec 2019. Data were recorded every day, and monthly mean values
were calculated. We also collected atmospheric temperature, and rela-
tive humidity data using data loggers (Thermochron iButtons ®), with
two data loggers placed under trees in each locality. We considered trees
that had dense canopy so that they could give enough shades throughout
the day. The iButton data were downloaded once a month. The monthly
mean values of temperature (⸰C), and humidity (%) were calculated.
2.4. Data analysis
To establish ea abundance, we grouped ea data collected from
each rodent host species and across the sampled sites and tested for
normality using Shapiro test (P <0.05). Flea abundance is used here to
refer to the total number of eas collected for each rodent host and
sampled sites during the sampling period regardless of the species
identity. To assess how ea abundance varied between localities, habitat
types and season, we used the Mann-Whitney-Wilcoxon test to explored
signicant differences of ea load between habitats and localities.
Similarly, the Kruskal-Wallis test was used to assess differences in ea
abundance across ea species and seasons as well as differences of each
ea species across rodent species and habitats in each locality (Npfarm
=farm in non-plague persistent locality, Npforest =forest in non-
persistent locality, Pfarm =farm in persistent locality and Pforest =
forest in persistent locality).
Further, to assess how temperature, humidity, rainfall, and rodent
species traits (sex, sex condition, rodent species, weight, head body
length) inuenced ea abundance, we built a negative binomial
generalized linear mixed model (GLMM) implemented in the ‘lme4’
package. Prior to modelling, we examined the data variables for po-
tential multicollinearity among temperature, rainfall, the weight, head
and body length variables. We subsequently dropped rainfall from the
model and retained temperature as these were highly correlated (r =
0.51) and because temperature is known to inuence ea growth and
development (Cavanaugh and Marshall 1972; Kreppel et al., 2016; Ming
ming et al., 2013). The rst model included sex, sex condition, head and
body length, temperature, weight and humidity as xed factor and ro-
dent species as random factor. The relative inuence of each variable in
the model was evaluated by deleting non-signicant model term in a
backward step-wise process, assessing model variance at each step of the
modelling. The drop1 function was used to delete non-signicant term
along each modelling steps and model signicance assessed using the
Wald test (Bolker et al., 2009). The best model tting the data was
chosen using the Akaike Information Criterion (AIC).
Furthermore, the binomial generalized linear model (GLM) imple-
mented in the MASS package was used to examine the probability of ea
infestation as a function of localities, habitats, seasons and rodent spe-
cies. Flea infestation-referred as presence or absence was treated as a
dependent variable in the model. To understand the relative inuence of
each variable in the model similar procedure as performed above was
followed. Further, the relative risk ratio (RR) of each independent var-
iable was computed from exponentials of coefcients generated from the
best models. To understand how these factors from the best model were
able to predict the ea load and prevalence of ea infestation, we built
prediction models using the ‘predict’ function with the “ggplot2
″
pack-
age. All modelling analyses were performed in R program, version 4.3.1.
Finally, to assesses species interaction and how ea community
structures between localities, habitats and seasons we used cluster
analysis based on a Bray-Curtis similarity matrix of grouped variables
with the program PRIMER v6. To obtain this, abundance matrix data
were rst square root transformed to down weight high abundance data,
Fig. 1. A map of Mbulu district indicating the two study localities, Endeshi-Arri (Persistent locality and Mongahay (non-persistent locality), along with the two study
habitats (Farmland and forests) in each locality.
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
4
normalizing them and creating a resemblance matrix. Further, we
visualized whether ea species clustered based on locality, habitat, and
season using a dendrogram plot.
3. Results
3.1. Abundance of eas in the study area
A total of 352 eas belonging to 5 species were collected, with
Dinopsyllus lypusus being the most abundant species, comprising 57.10%
of the total (n =201), followed by Xenopsylla brasiliensis at 29.26% (n =
103), Nosopsyllus spp. at 8.52% (n =30), Xenopsylla cheopis at 3.41% (n
=12) and Ctenophthalmus spp. at 1.70% (n =6).
A total of 420 individuals belonging to 12 species within family
Muridae were captured. Among all species, Mastomys natalensis had the
highest number of captures compared to other species in different
habitats and localities. Additionally, the short rainy season had higher
number of rodent hosts compared to other seasons. The total number of
rodent hosts for each species across habitats, localities, and seasons is
presented in Table 1.
Flea abundance by ea species across rodent hosts and habitats
revealed that, cultivated land, ea abundance was dominated by
D. lypusus, accounting for 48.26% (n =83) of the total ea population,
followed by X. brasiliensis at 36.63% (n =63), Nosopsyllus spp. at 9.30%
(n =16), and X. cheopis at 5.82% (n =10). Among the rodent species,
Mastomys natalensis had the highest ea abundance at 66.28% (n =114),
followed by Aethomys kaiseri at 24.42% (n =42). In the forest habitat,
D. lypusus was also the most abundant ea species, accounting for
65.56% (n =118) of the total ea population, followed by X. brasiliensis
at 33.89% (n =40), Nosopsyllus spp. at 11.86% (n =14), Ctenophthalmus
spp. at 3.33% (n =6), and X. cheopis at 1.69% (n =2).
The plague persistent locality had the highest ea abundance
71.88% (n =253) compared to non-persistent locality 28.13% (n =99).
On the habitat types, the forest had the highest ea abundance 51.14%
(n =180) compared to cultivated areas 48.86% (n =172). Also, ea
abundance was highest in the short rain season 59.94% (n =211) than
the long rain season and dry season (22.73%, n =80 and 17.33%, n =
61; respectively). Furthermore, there were signicant differences in ea
abundance between ea species (
χ
2
=11.69, df =4, p =0.02). There
were no signicant differences in ea abundance between localities (W
=1744, p =0.68), habitats (W =2157, p =0.83) and seasons (
χ
2
=
5.04, df =2, p =0.08) (Fig. 2a–c).
The rodent species with the highest ea abundance were Mastomys
natalensis at 32.22% (n =58) and Praomys delectorum at 30.56% (n =55)
(Fig. 3). When assessing how each ea species varied between rodent
species and habitats in each locality; there was a signicant difference in
X. Brasiliensis abundance between rodent species (
χ
2
=25.55, df =11, p
=0.01). A Signicant higher abundance of X. Brasiliensis was observed
on M. natalensis compared to Mus cf. gratus (p =0.03), Grammomys cf.
macmillan (p =0.01) and Lophuromys makundii (p =0.02). However,
there were no signicant difference in X. brasilliensis abundance between
habitats of each locality (
χ
2
=1.03, df =3, p =0.79). Similarly, the
abundance of D. lypusus species varied signicantly between rodent
species (
χ
2
=26.16, df =11, p =0.01). Mastomys natalensis had
signicantly higher abundance of D. lypusus compared to Mus minutoides
(p =0.03), Mus gratus (p =0.02), Lophuromys makundii (p =0.01),
Graphiurus cf. raptor (p =0.03), and Lemniscomys striatus (p =0.01). No
signicant differences were found in D. lypusus abundance between
habitats of each locality (
χ
2
=3.19, df =3, p =0.36). Furthermore, the
abundance of X. cheopis varies signicantly between rodent species (
χ
2
=20.26, df =11, p =0.04). Mastomys natalensis had signicantly higher
abundance of X. cheopis compared to Mus minutoides (p =0.04),
Lophuromys makundii (p =0.02), Graphiurus cf. raptor (p =0.02), Lem-
niscomys striatus (p =0.02), Grammomys cf. macmillan (p =0.01) and
Arvicanthis sp. “Masai Mara”. No signicant differences in X. cheopis
were observed between habitats in the locality (
χ
2
=0.87, df =3, p =
0.83). Moreover, there was a signicant difference in Nosopsyllus spp
abundance between rodent species (
χ
2
=32.31, df =11, p <0.05), with
M. natalensis having higher abundance compared to all other rodent
species (p <0.05). However, there were no signicant difference in
Nosopsyllus spp abundance between habitats in the locality. Addition-
ally, the abundance of Ctenophthalmus spp did not vary signicantly
between rodent species (
χ
2
=17.75, df =11, p =0.08) and between
habitats in the localities (
χ
2
=3.83, df =3, p =0.28).
3.2. Factors inuencing ea abundance
The model results indicated rodent weight was signicantly and
positively correlated with ea abundance (mean =0.02 ±0.004SE, p <
0.05) Fig. 4a). Furthermore, male rodents had higher ea abundance
than females (mean =0.27 ±0.158SE, p =0.09; Fig. 4b).
3.3. Effect of locality, season, habitat and rodent species on probability of
ea infestation
The highest probability of ea infestation was mostly associated with
the plague persistent locality. Similarly, there was a signicant effect of
the short rain season on the probability of higher ea infestation.
(Table 2, Fig. 5).
3.4. Flea community structure in the plague foci
Cluster analysis based on the ea abundance data revealed two
distinct ea community structures based on the habitats. The dendro-
gram plot (Fig. 6) showed that ea species were clustered into two main
groups, group A comprising of four species (Nosopsylla spp., Xenopsylla
Table 1
Number of rodent species captured across localities, habitats (Pfarm =farm in plague persistent locality, Pforest =forest in plague persistent locality, NPfarm =farm in
non-plague persistent locality and NPforest =forest in non-plague forest) and seasons.
Rodent species Localities Seasons
Pfarm (n) Pforest (n) Plague locality (n) NPFarm (n) NPForest (n) Non-plague locality (n) Dry Long rain Short rain
Aethomys kaiseri 17 0 17 6 0 6 12 4 7
Arvicanthis sp. “Masai Mara" 6 0 6 2 0 2 0 0 8
Grammomys cf. macmillani 4 16 20 0 3 3 7 4 12
Graphiurus cf. raptor 0 5 5 0 4 4 0 0 9
Lemniscomys striatus 0 14 14 0 1 1 1 1 13
Lemniscomys zebra 2 0 2 3 0 3 1 0 4
Lophuromys makundii 0 32 32 0 0 0 9 12 11
Mastomys natalensis 113 9 122 54 31 85 49 36 122
Mus cf. gratus 0 2 2 0 1 1 0 2 1
Mus minutoides 8 0 8 1 0 1 6 2 1
Praomys delectorum 0 81 81 0 4 4 29 13 43
Rattus ratus 1 0 1 0 0 0 0 0 1
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
5
cheopis, Dinopsyllus lypusus, and Xenopsylla brasiliensis) and group B
consisting of only one species (Ctenophathalmus spp). Group A had a
ner-scale separation of the two subgroups, with Nosopsylla spp. and
Xenopsylla cheopis clustering together and Dinopsyllus lypusus and Xeno-
psylla brasiliensis forming a separate cluster. Furthermore, ea commu-
nities were structured based on habitat, with some ea species
Fig. 2. Flea abundance in the (a) localities, (b) habitats and (c) seasons. Error bars represent the standard error. There were no statistically signicant differences that
were observed.
Fig. 3. Flea abundance for different ea species across habitat types in each locality and rodent species.
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
6
associated with both forest and cultivated land, while others were
associated with only forest habitats.
4. Discussion
This study aimed to understand the pattern of ea abundance be-
tween localities, habitat type, and season. Flea abundance was found to
be similar between localities and seasons. However, the study found that
ea infestation was mostly associated with the plague persistent locality
and the short rain season. Furthermore, ea abundance was found to
have a signicant positive correlation with rodent weight. In addition,
ea community was structured into two distinct groups.
We did not nd signicant difference in ea abundance between the
localities, despite the hypothesis that the plague persistent locality
would have higher ea abundance. This observation seems to contradict
the hypothesis that high ea abundance in persistent localities increases
the risk of bubonic plague. However, it is important to note that the
study found that the probability of ea infestation was signicantly
higher in the plague persistent locality, indicating that the risk of plague
pathogen spreading may still be elevated in this locality. One possible
explanation for the lack of signicant difference in ea abundance be-
tween the localities could be differences in the ea species assemblage
and level of infestation among different hosts. Even if the ea abundance
is similar, the composition of ea species and the levels of infestation on
individual host species could still be important determinant of disease
persistence, consistent with the available literature (Eisen et al., 2012).
Moreover, we did not nd any signicant differences of ea abundance
between seasons, but we observed that rodents were more frequently
infested with eas during the short rain season. This observation may be
attributed to the warmer and more humid conditions during the short
rain season, which create a favourable environment for ea develop-
ment and survival, leading to increased infestation in rodents. These
ndings align with previous studies which have shown that warmer and
humid condition promote ea development and survival, leading to
higher ea abundance (Krasnov et al., 2001; Kreppel et al., 2016; Sharif
1949; Mboera et al., 2011; Ngeleja et al., 2017), Importantly, such
condition has also been associated with an elevated incidence of human
plague in some of the plague foci. For example, Debien et al. (2010)
reported that precipitation resulted in higher ea abundance and an
increased incidence of plague in Lushoto, Tanzania.
Further, we found a positive association between rodent weight and
ea abundance. We also found a positive association between male ro-
dents and ea abundance, which is often attributed to their larger body
size (Moore and Wilson, 2002), but this relationship was not signicant
in our study area. Mostly, male rodents tend to have higher ea abun-
dance due to their larger body size, ample blood supply, weaker im-
munity and less grooming ability (Eads and Hoogland, 2016; Kiffner
et al., 2013). In addition, larger rodents tend to have higher activity
levels, which could increase their exposure to eas in the environment
(Krasnov et al., 2006b). However, different species can vary from these
patterns, and more studies are necessary to better understand relation-
ship between rodent weight and ea-borne diseases enabling more in-
sights into their specic host-ea relationships.
Furthermore, we found two communities of eas in the foci, sug-
gesting the ea community structures were inuenced by the seasons,
habitats types and hosts present in these habitats. These results are
consistent with studies elsewhere which have shown strong ea-habitats
(Brinkerhoff, 2008), host-habitat relationships (Krasnov et al., 2006)
and environmental factors (Chotelersak et al., 2015). In the present
study, the rst ea community included Dinopsyllus lypusus,
X. brasilliensi, X.cheopis and Nosopsyllus spp. which were found in both
Fig. 4. Plots showing predicted effect of rodent traits on ea abundance, based on nal best tting generalized linear mixed model with a negative-binomial
function. The plots (a) indicates that rodent weight increased with ea abundance. The gray shade in the plots represents the strength and direction of the corre-
lation, with the width of the shade indicating the 95% condence interval (CI) around the estimated effect. Furthermore, plot (b) indicates that male rodents are more
likely to have higher ea abundance compared to female rodents, but this association was not statistically signicant. The bars are 95% condence intervals of
the effects.
Table 2
Effect size with standard errors (±SE) and relative risk ratio (RR) of localities
and seasons on the probability of ea infestation, from the nal best tting
Generalized Linear Model (GLM) model.
Predictors Estimate (SE) RR RR 95% CI z-value p-value
(Intercept) −2.09 (0.49) 0.12 0.05–0.32 −4.23 <0.001
Locality
Plague persistent 1.01 (0.45) 2.75 1.15–6.58 2.27 0.02
Season
Long rain −0.16 (0.57) 0.84 0.27–2.63 −0.29 0.77
Short rain 0.98 (0.51) 2.69 0.98–7.36 1.93 0.05
Non-persistent locality and dry season were dened as reference.
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
7
farm and forest habitats. The species prefer rodents as primary hosts. For
example, X.cheopis, Dinopsyllus lypusus, and X. brasilliensis prefer R. ratus
as a primary host (Msangi, 2019), but they can also infest other rodent
species (Palazzo, 2011; Trivedi, 2003). However, a ne scale separation
observed in this group may be connected to the habitat types, seasons
and/or other factors in the foci, further studies would be needed to
conrm this hypothesis. Dinopsyllus lypusus, and X. brasilliensis, have
been reported as potential vector of plague among sylvatic rodents
(Ziwa et al., 2013) and the species were found on commensal rodent
such as Ratus rattus and M. natalesnsis, and on wild rodents such as
P. delectorum and L. makundii (Fig. 2a), indicating a ea-host-habitat
association. In addition, other studies have revealed that Xenopsylla
species is primarily an efcient vector of plague to humans (Zhang et al.,
2015; Hinnebusch et al., 2017). Moreover, M. natalensis is a social spe-
cies that nests in burrows and occasionally associates with other wild
rodent species (Coetzee 1975); given that Y.pestis is still circulating in
this species in the foci (Haikukutu et al., 2022), the diverse ea infes-
tation on M.natalensis may be contributing to plague persistence in the
foci and possibly inuencing spreading of Y.pestis between other rodent
species and/or other hosts sharing these habitats. In the second ea
community, Ctenophthalmus spp. were found to be the only species. The
species was only present in forest habitats, suggesting a strong associa-
tion with areas characterized by vegetation, such as grassy and wooded
environments. Additionally, their presence in rodent burrows and nests
reinforces their connection to habitats where these particular hosts were
commonly located. The species was found on host M. natalensis and
P. delectorum which was consistent with previous ndings conducted in
the same study area (Haule et al., 2013). The state of the forest habitat
supporting more diverse ea species compared to the farms, and pres-
ence of some of ea species infesting multiple rodent hosts that includes
Fig. 5. Plot showing the predicted effect of locality and season on the probability of ea infestation, based on the nal best-tting generalized linear model with a
binomial function. The analysis aimed to identify the factors that strongly inuence ea infestation. The strongest predictors of ea infestation were plague persistent
localities and short rain seasons. The probability of infestation on these predictors was found to be statistically signicant (p <0.05), suggesting a higher likelihood of
ea infestation in this locality and season. The bars are 95% condence interval of the effects.
Fig. 6. Dendrogram of ea species showing two main groups (AB) of ea community based on farm and forest habitats.
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
8
the susceptible species may encourage the potential of epizootic cycle of
disease transmissions between rodent species. Alarmingly, these ea
species have the ability to harbor other zoonotic pathogens such as
Bartonella and Richettsia typhi (Leulmi et al., 2014; Occhibove et al.,
2022) highlighting the need for more studies on ea borne pathogen
pattern and their role as pathogen vector in the foci.
5. Conclusion
Our ndings provide insight into the complex interactions between
ea communities, rodent host species and environmental factors in the
plague foci. The observed ea vector associating with sylvatic host, its
ability to harbor other zoonotic pathogens inuences the relevance of
extending our study to a broader disease transmission dynamic within
the foci. Our data about these ecosystems, provide opportunities for
potential strategies to targeted public health interventions that can
lower risks of bubonic plague and other ea borne diseases in these rural
communities.
Ethical approval
Ethical clearance was obtained from Sokoine University of Agricul-
ture Ref. no DPRTC/R/126/182/38, Manyara region Ref. no FA.262/
347/01/H/247, Mbulu district Ref. no AB.323/381/01/’B’/9. Animal
handling followed the guidelines of the American Society of Mammol-
ogists (ASM) for the use of wild mammals in research and education
(Sikes & Animal Care and Use Committee of the American Society of
Mammologists, 2016).
Author contributions
STK designed, conducted eld data collection, data analysis and
wrote original draft manuscript. AAR analyzed the data and reviewed
the original drafts. RHM and AM reviewed the manuscripts. AAR, RHM
& AM supervised the research. All authors read and approved the nal
version of the manuscript for submission.
Data availability
All data used in this analysis can be obtained from the corresponding
author upon request.
Funding
The study was funded by the African Centre of Excellence for Inno-
vative Rodent Pest Management and Biosensor Technology Develop-
ment (ACE IRPM&BTD) ACE II–Credit number 5799–TZ at Sokoine
University of Agriculture, Morogoro, Tanzania.
Declaration of competing interest
The authors declare that they have no conict of interest.
Acknowledgements
Many thanks to the community leaders and local people of Endesh
and Mongahay villages in Mbulu district for allowing us to conduct this
study. Thanks to the technical staffs for the assistance in eld trapping
and animal processing. We also extend our thanks to Professor Josef
Bryja, Institute of Vertebrate Biology, Czech Republic for rodent species
identication.
References
Bertherat, E., Bertherat, ´
E., 2019. Plague around the world in 2019. Wkly. Epidemiol.
Rec. 25, 289–292.
Bitam, I., Dittmar, K., Parola, P., Whiting, M.F., Raoult, D., 2010. Fleas and ea-borne
diseases. Int. J. Infect. Dis. 14, e667–e676.
Bolker, B.M., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens, M.H.H.,
White, J.S.S., 2009. Generalized linear mixed models: a practical guide for ecology
and evolution. Trends Ecol. Evol. 24 (3), 127–135.
Brinkerhoff, R.J., 2008. Habitat-associated differences in ea assemblages of striped
skunks (Mephitis mephitis). Comp. Parasitol. 75, 127–131.
Cavanaugh, D.C., Marshall, J.R.J.D., 1972. The inuence of climate on the seasonal
prevalence of plague in the Republic of Vietnam. J. Wildl. Dis. 8 (1), 85–94.
Chotelersak, K., Apiwathnasorn, C., Sungvornyothin, S., Panasoponkul, C., Samung, Y.,
Ruangsittichai, J., 2015. Correlation of host specicity, environmental factors and
oriental rat ea abundance. Southeast Asian J. Trop. Med. Publ. Health 46, 198.
Coetzee, C.G., 1975. The biology, behaviour, and ecology of Mastomys natalensis in
southern Africa. Bull. World Health Organ. 52 (4–6), 637.
Debien, A., Neerinckx, S., Kimaro, D., Gulinck, H., 2010. Inuence of satellite-derived
rainfall patterns on plague occurrence in northeast Tanzania. Int. J. Health Geogr. 9,
60.
Durden, L.A., Hinkle, N.C., 2019. Fleas (siphonaptera). In: Medical and Veterinary
Entomology. Elsevier, pp. 145–169.
Eads, D.A., Hoogland, J.L., 2016. Factors that affect parasitism of black-tailed prairie
dogs by eas. Ecosphere 7, e01372.
Eisen, R.J., Borchert, J.N., Mpanga, J.T., Atiku, L.A., MacMillan, K., Boegler, K.A.,
Montenieri, J.A., Monaghan, A., Gage, K.L., 2012. Flea diversity as an element for
persistence of plague bacteria in an East African plague focus. PLoS One 7, e35598.
Eisen, R.J., Gage, K.L., 2009. Adaptive strategies of Yersinia pestis to persist during inter-
epizootic and epizootic periods. Vet. Res. 40, 1.
Eisen, R.J., Reynolds, P.J., Ettestad, P., Brown, T., Enscore, R.E., Biggerstaff, B.J.,
Cheek, J., Bueno, R., Targhetta, J., Montenieri, J.A., Gage, K.L., 2007. Residence-
linked human plague in New Mexico: a habitat-suitability model. Am. J. Trop. Med.
Hyg. 77 (1), 121–125.
Gage, K.L., Kosoy, M.Y., 2005. Natural history of plague: perspectives from. Annu. Rev.
Entomol. 50, 505–528.
Gage, K.L., Burkot, T.R., Eisen, R.J., Hayes, E.B., 2008. Climate and vector borne
diseases. Am. J. Prev. Med. 35, 436–450.
Haikukutu, L., Lyaku, J.R., Lyimo, C., Kasanga, C.J., Kandusi, S.E., Rahelinirina, S.,
Rasoamalala, F., Rajerison, M., Makundi, R., 2022. Plague in Tanzania: rst report of
sylvatic plague in Morogoro region, persistence in Mbulu focus, and ongoing
quiescence in Lushoto and Iringa foci. IJID Regions 4, 105–110.
Happold, D.C.D. (Ed.), 2013. Mammals of Africa. Volume III: Rodents, hares and rabbits.
Bloomsbury Publishing, London, UK, p. 784.
Haule, M., Lyamuya, E.E., Hang’ombe, B.M., Matee, M.I., 2013. Investigation of Fleas as
Vectors in the Transmission of Plague during a Quiescent Period in North-Eastern,
Tanzania.
Hinnebusch, B.J., Bland, D.M., Bosio, C.F., Jarrett, C.O., 2017. Comparative ability of
Oropsylla montana and Xenopsylla cheopis eas to transmit Yersinia pestis by two
different mechanisms. PLoS Neglected Trop. Dis. 11, e0005276 https://doi.org/
10.1371/journal.pntd.0005276.
Kandi, V., Koka, S.S., Bhoomigari, M.R., 2019. Hymenolepiasis in a pregnant woman: a
case report of Hymenolepis nana infection. Cureus 11.
Kessy, S.T., Makundi, R.H., Sabuni, C., Massawe, A.W., Rija, A.A., 2023. Rodent
Abundance, Diversity and Community Structure in a Bubonic Plague Endemic Area,
Northern Tanzania. Mammalia.
Kiffner, C., Stanko, M., Morand, S., Khokhlova, I.S., Shenbrot, G.I., Laudisoit, A.,
Leirs, H., Hawlena, H., Krasnov, B.R., 2013. Sex-biased parasitism is not universal:
evidence from rodent–ea associations from three biomes. Oecologia 173,
1009–1022.
Krasnov, B., Khokhlova, I., Shenbrot, G., 2002. The effect of host density on ectoparasite
distribution: an example of a rodent parasitized by eas. Ecology 83, 164–175.
Krasnov, B.R., 2008. Functional and Evolutionary Ecology of Fleas: a model for
ecological parasitology. Cambridge University Press.
Krasnov, B.R., Khokhlova, I.S., Fielden, L.J., Burdelova, N.V., 2001. Effect of air
temperature and humidity on the survival of pre-imaginal stages of two ea species
(Siphonaptera: pulicidae). J. Med. Entomol. 38, 629–637.
Krasnov, B.R., Mouillot, D., Shenbrot, G.I., Khokhlova, I.S., Poulin, R., 2004.
Geographical variation in host specicity of eas (Siphonaptera) parasitic on small
mammals: the inuence of phylogeny and local environmental conditions.
Ecography 27, 787–797.
Krasnov, B.R., Shenbrot, G.I., Khokhlova, I.S., Poulin, R., 2005. Diversication of
ectoparasite assemblages and climate: an example with eas parasitic on small
mammals. Global Ecol. Biogeogr. 14, 167–175.
Krasnov, Boris R., Shenbrot, G.I., Mouillot, D., Khokhlova, I.S., Poulin, R., 2006a.
Ecological characteristics of ea species relate to their suitability as plague vectors.
Oecologia 149, 474–481.
Krasnov, Boris R., Stanko, M., Khokhlova, I.S., Moˇ
sanský, L., Shenbrot, G.I., Hawlena, H.,
Morand, S., 2006b. Aggregation and species coexistence in eas parasitic on small
mammals. Ecography 29, 159–168.
Krasnov, Boris R., Stanko, M., Miklisova, D., Morand, S., 2006. Habitat variation in
species composition of ea assemblages on small mammals in central Europe. Ecol.
Res. 21, 460–469. https://doi.org/10.1007/s11284-005-0142-x.
Kreppel, K.S., Telfer, S., Rajerison, M., Morse, A., Baylis, M., 2016. Effect of temperature
and relative humidity on the development times and survival of Synopsyllus
fonquerniei and Xenopsylla cheopis, the ea vectors of plague in Madagascar.
Parasites Vectors 9, 1–10.
Leulmi, H., Socolovschi, C., Laudisoit, A., Houemenou, G., Davoust, B., Bitam, I.,
Raoult, D., Parola, P., 2014. Detection of Rickettsia felis, Rickettsia typhi, Bartonella
S.T. Kessy et al.
International Journal for Parasitology: Parasites and Wildlife 23 (2024) 100921
9
species and Yersinia pestis in eas (siphonaptera) from Africa. PLoS Neglected Trop.
Dis. 8, e3152.
Makundi, R.H., Massawe, A.W., Borremans, B., Laudisoit, A., Katakweba, A., 2015. We
are connected: ea–host association networks in the plague outbreak focus in the
Rift Valley, northern Tanzania. Wildl. Res. 42, 196–206.
Makundi, R.H., Massawe, A.W., Mulungu, L.S., Katakweba, A., Mbise, T.J., Mgode, G.,
2008. Potential mammalian reservoirs in a bubonic plague outbreak focus in Mbulu
District, northern Tanzania. 2007. Mammalia 72, 253–257. https://doi.org/
10.1515/MAMM.2008.038.
Makundi, R.H., Massawe, A.W., Mulungu, L.S., 2007. Breeding seasonality and
population dynamics of three rodent species in the Magamba Forest Reserve,
Western Usambara Mountains, north-east Tanzania. Afr. J. Ecol. 45, 17–21. https://
doi.org/10.1111/j.1365-2028.2006.00667.
Mboera, L.E., Mayala, B.K., Kweka, E.J., Mazigo, H.D., 2011. Impact of climate change
on human health and health systems in Tanzania: a review. Tanzan. J. Health Res. 13
(5).
Moore, S.L., Wilson, K., 2002. Parasites as a viability cost of sexual selection in natural
populations of mammals. Science 297, 2015–2018, 1979.
Msangi, S., 2019. The surveillance of rodent populations in East Africa in relation to
plague endemicity. Tanzan. J. Sci. 1, 8–20.
Mwalimu, C.D., Mgode, G., Sabuni, C., Msigwa, F., Mghamba, J., Nyanga, A.,
Mohamed, A., Kwesi, E., Nagu, N., Kishimba, R., John, L., 2022. Preliminary
investigation and intervention of the suspected plague outbreak in Madunga, Babati
District-Tanzania. Acta Trop. 233, 106566 https://doi.org/10.1016/j.
actatropica.2022.106566.
Neerinckx, S.B., Peterson, A.T., Gulinck, H., Deckers, J., Leirs, H., 2008. Geographic
distribution and ecological niche of plague in sub-Saharan Africa. Int. J. Health
Geogr. 7 (1), 1–12.
Ngeleja, R.C., Luboobi, L.S., Nkansah-Gyekye, Y., 2017. The effect of seasonal weather
variation on the dynamics of the plague disease. Int. J. Math. Math. Sci. 2017, 25.
Nyembo, L.O., Larbi, I., Rwiza, M.J., 2021. Analysis of spatio-temporal climate
variability of a shallow lake catchment in Tanzania. Journal of Water and Climate
Change 12, 469–483.
Occhibove, F., McKeown, N.J., Risley, C., Ironside, J.E., 2022. Eco-epidemiological
screening of multi-host wild rodent communities in the UK reveals pathogen strains
of zoonotic interest. Int J Parasitol Parasites Wildl 17, 278–287.
Palazzo, S., 2011. Nosopsyllus fasciatus [WWW Document]. Animal Diversity Web. URL.
https://animaldiversity.org/accounts/Nosopsyllus_fasciatus/, 5.16.23.
Palomino, P.S., Montenegro, O.L., Mesa-Gonz´
alez, E., 2020. Anesthesia of sigmodontine
rodents under eld conditions. J. Vet. Med. 69–83.
Pham, H.V., Dang, D.T., Tran Minh, N.N., Nguyen, N.D., Nguyen, T.V., 2009. Correlates
of environmental factors and human plague: an ecological study in Vietnam. Int. J.
Epidemiol. 38, 1634–1641.
Philip Samuel, P., Govindarajan, R., Krishnamoorthi, R., Venkatesh, A., 2021. A rapid
protocol for clearing, staining, and mounting of Arthropoda: trombiculidae,
Pediculidae and Pulicidae. North West J Zool 17.
Poje, J.E., Rocke, T.E., Samuel, M.D., 2020. Impacts of environmental conditions on eas
in black-tailed prairie dog burrows. J. Vector Ecol. 45 (2), 356–365.
Ramana, K.V., Rao, S.D., Rao, R., Mohanty, S.K., Wilson, C.G., 2011. Human dipylidiasis:
a case report of Dipylidium caninum infection from Karimnagar. Online J. Health
Allied Sci. 10.
Sharif, M., 1949. Effects of constant temperature and humidity on the development of the
larvae and the pupae of the three Indian species of Xenopsylla (Insecta:
siphonaptera). Philos. Trans. R. Soc. Lond. B Biol. Sci. 233, 581–633.
Sherman, D.M., 2007. Tending Animals in the Global Village: a Guide to International
Veterinary Medicine. John Wiley & Sons.
Stanko, M., Miklisov´
a, D., Goüy de Bellocq, J., Morand, S., 2002. Mammal density and
patterns of ectoparasite species richness and abundance. Oecologia 131, 289–295.
Tripp, D.W., Gage, K.L., Montenieri, J.A., Antolin, M.F., 2009. Flea abundance on black-
tailed prairie dogs (Cynomys ludovicianus) increases during plague epizootics.
Vector Borne Zoonotic Dis. 9, 313–321.
Trivedi, J., 2003. Xenopsylla cheopis [WWW Document]. Animal Diversity Web. URL.
https://animaldiversity.org/accounts/Xenopsylla_cheopis/, 5.16.23.
Vall`
es, X., Stenseth, N.C., Demeure, C., Horby, P., Mead, P.S., Cabanillas, O.,
Ratsitorahina, M., Rajerison, M., Andrianaivoarimanana, V., Ramasindrazana, B.,
2020. Human plague: an old scourge that needs new answers. PLoS Neglected Trop.
Dis. 14, e0008251.
Young, H.S., Dirzo, R., McCauley, D.J., Agwanda, B., Cattaneo, L., Dittmar, K.,
Eckerlin, R.P., Fleischer, R.C., Helgen, L.E., Hintz, A., 2015. Drivers of intensity and
prevalence of ea parasitism on small mammals in East African savanna ecosystems.
J. Parasitol. 101, 327–335.
Zając, Z., Kulisz, J., Wo´
zniak, A., 2020. Flea communities on small rodents in eastern
Poland. Insects 11, 894.
Zhang, Y., Dai, X., Wang, Q., et al., 2015. Transmission efciency of the plague pathogen
(Y. pestis) by the ea, Xenopsylla skrjabini, to mice and great gerbils. Parasites Vectors
8, 256. https://doi.org/10.1186/s13071-015-0852-z.
Ziwa, M.H., Matee, M.I., Hang’ombe, B.M., Lyamuya, E.F., Kilonzo, B.S., 2013a. Plague
in Tanzania: an overview. Tanzan. J. Health Res. 15.
Ziwa, M.H., Matee, M.I., Kilonzo, B.S., Hang’ombe, B.M., 2013b. Evidence of Yersinia
pestis DNA in rodents in plague outbreak foci in Mbulu and Karatu Districts,
northern Tanzania. Tanzan. J. Health Res. 15.
Zurita, A., Callej´
on, R., García-s´
anchez, ´
A.M., Urdapilleta, M., Lareschi, M., Cutillas, C.,
2019. Origin, evolution, phylogeny and taxonomy of Pulex irritans. Med. Vet.
Entomol. 33, 296–311.
S.T. Kessy et al.