Geographical and host species barriers

royalsocietypublishing.org/journal/rspb
differentially affect generalist and
specialist parasite community structure in
a tropical sky-island archipelago
Research
Pooja Gupta1,2, C. K. Vishnudas3, Uma Ramakrishnan4,
Cite this article: Gupta P, Vishnudas CK,
Ramakrishnan U, Robin VV, Dharmarajan G.
V. V. Robin3,† and Guha Dharmarajan1,†
2019 Geographical and host species barriers 1
Savannah River Ecology Laboratory, University of Georgia, Aiken, SC, USA
2
differentially affect generalist and specialist Warnell School of Forestry and Natural Resources, University of Georgia, Athens, GA 30602, USA
3
Indian Institute of Science Education and Research Tirupati, Mangalam, Tirupati 517507, India
parasite community structure in a tropical sky- 4
National Centre for Biological Sciences, TIFR, Bangalore 560065, India
island archipelago. Proc. R. Soc. B 286:
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

PG, 0000-0002-4241-8208; VVR, 0000-0003-3109-5498; GD, 0000-0001-8500-0429

20190439.
http://dx.doi.org/10.1098/rspb.2019.0439 Understanding why some parasites emerge in novel host communities while
others do not has broad implications for human and wildlife health. In the
case of haemosporidian blood parasites, epidemic wild bird mortalities on
oceanic islands have been linked to Plasmodium spp., but not genera like
Received: 11 March 2019 Haemoproteus. Indeed, Haemoproteus is absent from many oceanic islands.
Accepted: 10 May 2019 By contrast, birds on continental islands share long coevolutionary histories
with both Plasmodium and Haemoproteus, and are thus ideal model systems
to elucidate eco-evolutionary endpoints associated with these parasites in
oceanic islands. Here, we examine eco-evolutionary dynamics of avian
haemosporidian in the Shola sky-island archipelago of the Western Ghats,
Subject Category:
India. Our analyses reveal that compared to Plasmodium, Haemoproteus
Evolution lineages were highly host-specific and diversified via co-speciation with
their hosts. We show that community structure of host-generalist Plasmodium
Subject Areas: was primarily driven by geographical factors (e.g. biogeographic barriers),
ecology, evolution, health and disease and while that of host-specialist Haemoproteus was driven by host species barriers
epidemiology (e.g. phylogenetic distance). Consequently, a few host species can harbour a
high diversity of Plasmodium lineages which, in turn, are capable of infecting
Keywords: multiple host species. These two mechanisms can act in concert to increase
the risk of introduction, establishment, and emergence of novel Plasmodium
avian haemosporidians, Plasmodium,
lineages in island systems.
Haemoproteus, disease emergence, community
structure, India

1. Background
Author for correspondence: Emerging infectious diseases are considered to be one of the greatest challenges
of our times from the perspective of human and wildlife health, as well as
Pooja Gupta
ecosystem function and stability [1,2]. An important driver for the dramatic
e-mail: [email protected] increase in disease emergence over the past several decades is the recent and
rapid spread of parasites outside their native range owing to a myriad of
factors, including global climate change and increased human-mediated trans-
port [1]. Such parasite range expansion can lead to serious epidemics in naive
host populations into which these parasites are newly introduced [3–5].
Avian haemosporidians (Apicomplexa: Haemosporida; Plasmodium and
other related genera such as Haemoproteus—hereafter avian malaria) are a glob-
ally distributed group of vector-borne blood parasites that infect a wide array
†
Contributed equally. of bird taxa [6]. Avian malaria caused by Plasmodium spp. is one of the most
important emerging infectious diseases of wild bird populations globally [7–9].
Electronic supplementary material is available Large-scale mortalities in native wild birds have been well documented owing
to the accidental introduction of Plasmodium spp. and Culex quinquefaciatus into
online at https://dx.doi.org/10.6084/m9.
island bird communities which had no coevolutionary history with these parasites
figshare.c.4509134. (e.g. Hawaii [6,10] and New Zealand [7,8]). However, similar epidemic mortalities

& 2019 The Author(s) Published by the Royal Society. All rights reserved.
 by Haemoproteus spp. have not been recognized. Indeed, while wide elevational range. Thus, the Shola sky-islands offer an 2
Plasmodium spp. are cosmopolitan [11], Haemoproteus spp. only excellent opportunity to better understand the relative impor-

royalsocietypublishing.org/journal/rspb
appears to have colonized some oceanic islands systems (e.g. tance of geographical (e.g. spatial distance and biogeographic
Lesser Antilles [12,13]) and is absent from many others (e.g. gaps), climatic (e.g. elevational gradients), and host species
Hawaii, New Zealand, and French Polynesia [8,11,14]). barriers (e.g. host phylogeny and host ecology) in driving
The reduced ability to colonize some islands by Haemo- evolution of parasite community structure.
proteus spp. versus Plasmodium spp. and, consequently, the In this study, we test whether Plasmodium spp. and Haemo-
lower negative consequences associated with parasite inva- proteus spp., owing to their varying levels of host specialization
sions on native bird communities are likely driven by a differ in terms of: (i) host association patterns, (ii) coevolu-
myriad of factors such as parasite specialization and avian tionary dynamics, (iii) genetic structure, and (iv) global
host/vector community composition. Previous studies indi- phylogenetic structure. We predict that: (i) diversity of hosts
cate that Plasmodium spp. are relatively generalist, infecting infected by a single lineage would be greater for generalist
a wide range of host species, whereas Haemoproteus spp. gen- versus specialist parasites, (ii) generalist parasites would likely
erally exhibit specialist associations and are restricted to coevolve with hosts through host-switching, while specialists

Proc. R. Soc. B 286: 20190439

phylogenetically related host species [11,15 –17], but this pat- would likely co-speciate with their hosts, (iii) genetic structure
tern is not universal [14,18,19]. Such eco-evolutionary of generalist parasites would primarily be influenced by geo-
differences likely affect the ability of generalist parasites, graphy, while specialists would be more affected by host
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

like Plasmodium spp., to readily establish in island commu- species barriers, and (iv) phylogenetic structure at global
nities when introduced by natural or anthropogenic factors scales would be lower for the generalist versus specialist para-
[11,20]. However, the taxonomic distinctiveness of host com- sites because geographical range tends to correlate positively
munities on islands may protect them from invasions by with niche breadth [26].
specialist parasites, like Haemoproteus spp., if island commu-
nities consist of species phylogenetically distant to hosts in
the parasite’s native range. Consequently, the colonization
history of avian hosts/vectors, that is specific to each island
2. Material and methods
system, can critically affect the likelihood of colonization (a) Field and laboratory methods
by specialist parasites, such as Haemoproteus spp., but not Field sampling was conducted at 7 –14 sites across four major
generalist ones, such as Plasmodium spp. geographical regions in the southern 600 km mountain range
Understanding the underlying eco-evolutionary mechan- of the Western Ghats (at 100 – 2500 m.a.s.l.) (figure 1; see elec-
isms that influence the colonization and maintenance of tronic supplementary material, table S1). Each geographical
Plasmodium spp. versus Haemoproteus spp. can help elucidate region corresponded to the sky-island group separated by three
the drivers of disease emergence in natural communities. In biogeographic barriers—Chaliyar River valley, Palghat Gap,
this context, continental sky-islands are a fascinating model and Shencottah Gap. Adult birds were captured using mist-
nets during 2011 – 2013 and blood samples were collected from
system because they provide excellent natural laboratories
bird’s ulnar vein in Queen’s lysis buffer, following Robin et al.
for examining parasite eco-evolutionary dynamics. Sky-
[27]. Genomic DNA was extracted using Qiagen blood and
islands are isolated montane forests surrounded by a ‘sea’ tissue extraction kit (Qiagen, Hilden, Germany) and screened
of low-elevation habitat, limiting dispersal of both bird and for haemosporidian infection by amplifying 478 bp of mitochon-
parasite lineages, similar to oceanic islands [21]. Thus, drial cytochrome b gene (cytb) of avian haemosporidian parasites
sky-island bird communities may face many of the same [28] (details in the electronic supplementary material).
eco-evolutionary challenges as their oceanic counterparts.
However, sky-island bird communities, in contrast with
many oceanic counterparts, have generally shared long (b) Phylogenetic analyses
coevolutionary histories with their parasites. Consequently, To assess phylogenetic relationships among the Shola sky-island
the bird communities on continental sky-islands can help haemosporidian parasite lineages, we conducted Bayesian
phylogenetic analyses in MRBAYES [29]. Similarly, we built a
elucidate the potential long-term ecological and evolutionary
host phylogenetic tree based on cytochrome b sequence data
endpoints for oceanic island bird communities where avian
(1143 bp) for bird species from an earlier study [25]. To examine
haemosporidians have recently been introduced.
parasite phylogenetic relationships at the global scale, we
Here, we examine the eco-evolutionary dynamics of avian obtained cytochrome b sequence data from the MalAvi database
haemosporidians in the sky-island archipelago of the [30] (accessed February 2018) and built Bayesian parasite
Western Ghats, southern India. These sky-islands (hereafter phylogenetic trees in MRBAYES [29]. We calculated rarefaction
Shola sky-islands) are high-elevation montane ecosystems curves of expected phylogenetic diversity for host species and
characterized by unique habitats called Sholas, a natural parasite lineages to ensure adequate sampling, as implemented
mosaic of wet, tropical evergreen forests and grasslands, iso- in R-package PDCALC [31] (details in the electronic supplementary
lated by drier lowland habitats [22]. The Shola sky-islands material). All statistical analyses were carried out in R 3.3.3 [32],
harbour remarkable species diversity and endemism driven unless specifically mentioned otherwise.
by geographical complexity at multiple spatial scales
[22,23]. At large spatial scales (i.e. across the Western
Ghats), the deep and wide biogeographic barriers (Chaliyar,
(c) Host–parasite association patterns
We measured the diversity of parasite lineages infecting each
Palghat, and Shencottah gaps; figure 1) have led to avian line- host species and diversity of hosts infected by each parasite line-
age diversification [24,25]. At small spatial scales (i.e. age using the Shannon diversity of interactions index (H2) [33], a
individual mountains), the steep elevational gradient contrib- two-dimensional equivalent of the Shannon index [34]. We built
utes to colonization of sky-islands by both specialist avian null models by randomizing the network interactions (10 000
species restricted to montane habitats and generalists with a times) while maintaining the marginal sums (i.e. sum of
 (a) (b) 3

royalsocietypublishing.org/journal/rspb
12
N Chaliyar
gap

11
Palghat
region I gap
region II
region III
region IV
latitude (°)

Proc. R. Soc. B 286: 20190439

Shola forest
10 Shola grassland
plantations
agriculture
elevation profile
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

0.12 0.4
9
0.3 Shencottah
0.08 gap
prevalence

0.2
0.04
0.1
8
0 0
50 0 50 km
P H

75 76 77 78 0 500 1500 2500

longitude (°) elevation
Figure 1. Map of Western Ghats. (a) Locations of sampling sites (filled circles) in four geographical regions: I (Bababudan and Banasura hills), II (Nilgiri hills),
III (Anamalai-Palni-Highwavies hills), IV (Ashambu hills), corresponding to the major sky-island group separated by three biogeographic barriers—Chaliyar gap,
Palghat gap, and Shencottah gap. Underlying natural (i.e. forest and grassland) versus plantation habitats and 1400 m.a.s.l. isoclines are also depicted. Inset shows
the proportion of individuals infected with Plasmodium spp. and Haemoproteus spp. in each geographical region with their 95% bootstrap confidence intervals;
(b) elevation profile of the Western Ghats along a linear transect connecting the highest elevation points in each geographical region (black transect line in a). (Online
version in colour.)

interactions for each species was kept constant) using R-package each evolutionary event, CORE-PA does not require a priori assign-
VEGAN [35]. We performed two-sided tests of the network metric ment of cost values to compute a cost minimal reconstruction.
value against the distribution of the null model metric values to Run parameters and settings are detailed in the electronic
assess statistical significance. We quantified host specialization supplementary material.
for parasite lineages infecting greater than or equal to two host
species by measuring the phylospecificity index—mean phylo-
genetic distance (MPD) and standardized effect sizes of the
MPD values (SES.MPD) [36,37] using R-package PICANTE [38]
(e) Parasite genetic structure
To test whether parasite genetic structure was influenced by host
(details in the electronic supplementary material).
species barriers, biogeographic gaps (figure 1), and geographical
structure within each biogeographic region, we used a hierarchical
analysis of molecular variance (AMOVA) as implemented in the
(d) Host–parasite coevolutionary dynamics R-package HIERFSTAT [44]. We assessed the statistical significance
We visually assessed phylogenetic congruence between the host of each variance estimate by conducting 1000 randomizations
and parasite phylogenetic trees by constructing a cophylogenetic among species (for FHost/Total ), regions within each species (for
tanglegram using TREEMAP [39]. We then statistically tested for FRegion/Host) and sampling sites within regions (for FSite/Region).
host – parasite phylogenetic congruence by conducting a dis- Furthermore, we tested the relative effects of geographical, cli-
tance-based cophylogenetic analyses in Procrustean Approach matic, and host factors on parasite genetic structure using
to Cophylogeny (PACo) [40], as implemented in R-packages multiple regression on distance matrices (MRM) [45], as
APE and VEGAN [35,41]. We also conducted an event-based cophy- implemented in ECODIST [46]. The geographical factors considered
logenetic analyses, as implemented in JANE [42] and CORE-PA [43], were biogeographic gaps (as a Boolean matrix) and geographical
to determine the type and frequency of different coevolutionary distance (i.e. the great circle distance between sampling coordi-
scenarios, e.g. co-speciation, duplication, host switch, sorting, or nates); climatic factors included elevational distance (i.e. absolute
loss of parasite lineages. While JANE assigns an a priori cost for difference in elevation between sampling sites); host factors
 included host phylogenetic and host ecological distance (measured strong positive association between hosts and parasite 4
as Gower distance between host ecological traits; [47]). Host eco- lineages, with 27 of 29 parasite lineages infecting a lower
logical data included species traits that could affect

royalsocietypublishing.org/journal/rspb
diversity of hosts (binomial p , 0.001; electronic supplemen-
haemosporidian infection dynamics and were collected from pub- tary material, table S5) and 23 of 24 host species being
lished sources [48], as well as field observations by V.V.R. and
infected by a lower diversity of parasites than expected by
C.K.V. (see electronic supplementary material, table S9).
chance (binomial p , 0.001; figure 2; electronic supplementary
material, table S6).
(f ) Global parasite phylogenetic structure Furthermore, phylogenetic host specificity analyses for
To test whether specialist versus generalist parasites were phylogen- parasite lineages infecting multiple host species revealed
etically more clustered across the global haemosporidian phylogeny, higher host specialization for Haemoproteus spp. (MPDw
we calculated the nearest neighbour phylogenetic distance (DKN) mean ¼ 0.132, CI ¼ 0.038, 0.248) compared to Plasmodium
within Plasmodium spp. and Haemoproteus spp. lineages. We pro- spp. lineages (MPDw mean ¼ 0.358, CI ¼ 0.246, 0.443). While
duced a null distribution of DKN values by randomizing (1000 four of seven Haemoproteus spp. lineages showed higher phylos-
times) tip labels across the global phylogeny and calculated the prob-
pecificity (based on their significant SES.MPD values), none of
ability of obtaining a simulated DKN value  observed DKN value.

Proc. R. Soc. B 286: 20190439

the Plasmodium spp. lineages had higher host specificity than
We tested the overall significance (i.e. across lineages within each
parasite genus) using the exact binomial test in R.
expected by chance (electronic supplementary material, table
S7). Thus, Haemoproteus spp. were highly host specialized,
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

with most lineages infecting one or a very few phylogenetically

clustered hosts, compared to Plasmodium spp., as observed in
3. Results and discussion other biogeographic regions [11,15–17,52].
Interestingly, high prevalence and diversity of Plasmodium
(a) Parasite prevalence patterns spp. lineages were recovered from a single host species—
We sampled 1177 birds belonging to 28 species (including 14
T. merula. Based on existing genetic data and plumage-
endemics), representing almost the entire Shola sky-island
based taxonomy, T. merula is known to harbour cryptic
bird community (except two species, see electronic sup-
species diversity, with overlapping ranges of resident and
plementary material, table S2) and found 24 species (490
migratory races [25,53], which may explain why it was
birds) infected with haemosporidians (41.6% prevalence;
infected by diverse haemosporidian lineages. Additionally,
figure 1). Plasmodium spp. was found at a prevalence of
T. merula harboured some widespread and pathogenic hae-
13.6% (across 19 bird species), while Haemoproteus spp. had
mosporidian lineages, which may underscore its role as a
a prevalence of 68.9% (across 20 bird species; electronic sup-
potential reservoir host in the Shola sky-island bird commu-
plementary material, table S2). Haemosporidian prevalence
nity. Among the eight Plasmodium spp. lineages infecting T.
varied across species, with Turdus merula as a key host species
merula, one was a generalist, while others were restricted to
for Plasmodium spp. infection (29% prevalence) and Zosterops
T. merula and two lineages matched FANTAIL01 and
palpebrosus for Haemoproteus spp. (77.1% prevalence) infec-
GRW06 (Plasmodium elongatum) (electronic supplementary
tion. Rarefaction analyses revealed that our sampling was
material, table S3). While FANTAIL01 is relatively less
adequate to recover the observed parasite phylogenetic diver-
common, GRW06 is globally widespread and often virulent
sity (electronic supplementary material, figure S1). Among
in naive bird hosts [54]. Moreover, out of the three Haemopro-
the 47 parasite lineages recovered, a majority of Plasmodium
teus spp. lineages detected in T. merula, one matched and two
spp. (10 of 18) and Haemoproteus spp. (24 of 29) lineages
were 99% similar to Haemoproteus minutus, a widespread
were novel and unique to the Shola sky-islands (electronic
European lineage of Turdus spp. While Haemoproteus minutus
supplementary material, table S3), indicating that many
is relatively benign for native European birds, lethal out-
haemosporidian lineages are generally restricted to a single
breaks have been recorded for naive captive parrots in
biogeographic region and characterized by local diversification
Europe [55,56]. Previous studies have also shown that Eura-
as suggested by Ellis et al. [49–51].
sian blackbird and other thrushes (Turdus spp.) generally
serve as key reservoir hosts for Plasmodium spp. infections
(b) Host–parasite association patterns with high prevalence and diversity in continental commu-
Plasmodium spp. and Haemoproteus spp. differed markedly in nities; and contribute to high spillover risk to naive host
terms of host – parasite associations, with two Plasmodium communities when introduced to islands (such as in Azores
spp. lineages infecting a greater diversity of hosts than [57], Robinson Crusoe [58], and New Zealand [8]). Thus,
expected by chance (P_MSP02: observed H2 ¼ 1.748; T. merula could be a potential key reservoir host in the
expected H2 ¼ 0.806; p ¼ 0.016; P_MSP03: observed H2 ¼ Western Ghats with several virulent lineages.
2.246; expected H2 ¼ 0.795; p , 0.001; figure 2; electronic
supplementary material, table S4). However, patterns of
generalist host –parasite associations were not statistically (c) Host–parasite coevolutionary dynamics
significant across all Plasmodium spp. lineages (binomial We found no evidence of significant cophylogenetic congru-
p ¼ 0.058). Additionally, while host individuals were not sus- ence between hosts and Plasmodium spp. phylogenies (PACo,
ceptible to a greater diversity of Plasmodium spp. lineages m 2 ¼ 5.297, p ¼ 0.640), but there was significant cophyloge-
than expected by chance (binomial p ¼ 0.340), it is important netic congruence between host and Haemoproteus spp.
to note that a disproportionately high diversity of Plasmodium phylogenies (PACo, m 2 ¼ 7.39, p ¼ 0.047; see also electronic
spp. lineages (7 of 18) were recovered from a single host supplementary material, figures S2 and S3). Cophylogenetic
species—T. merula (observed H2 ¼ 1.715; expected H2 ¼ analysis with JANE revealed significant topological congru-
0.978; p ¼ 0.046; figure 2; electronic supplementary material, ence between host and Plasmodium spp. or Haemoproteus
table S6). By contrast, for Haemoproteus spp., there was a spp. phylogenies (optimal inferred reconstruction cost lower
 support % association 5
0.5 . 0.7 . 0.9 . 0 1 2 4 8 16 0 0.3 1 3 9 27

royalsocietypublishing.org/journal/rspb
MONMER
MONCAC

POMHOR

EUMALB

TURMER
ZOODAU
Out-

GARDEL
MONJER

SHOMAJ
SHOALB
SHOASH

ANTRUF
HYPLEU

SAXCAP
CYOPAL
MONFAI
COLELP

PYCJOC
ZOSPAL
PELRUF

ZOOCIT
ALCPOI

IOLIND

FICNIG
group

P_SAXCAP01
P_EUMALB02
0.1
P_TURMER07
P_MSP03
P_MSP02
P_TURMER01

Proc. R. Soc. B 286: 20190439

P_MONFAI02
P_TURMER03
P_TURMER05
P_TURMER04
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

P_GARDEL02
P_MONCAC01
P_MSP04
P_TURMER02
P_MSP01
P_ZOOCIT02
P_TURMER06
P_ANTRUF01
H_MSP03
H_COLELP01
H_SHOMAJ03
H_ZOSPAL04
H_MSP05
H_POMHOR01
H_MSP07
H_MONCAC03
H_MONFAI02
H_MONMER02
H_ZOSPAL03
H_MSP06
H_CYOPAL01
H_MSP02
H_ZOODAU01
H_MONCAC01
H_MSP01
H_TURMER03
H_TURMER02
H_TURMER01
H_MSP04
H_ALCPOI05
H_ALCPOI06
H_ALCPOI04
H_IOLIND01
H_PYNJOC01
H_HYPLEU01
H_HYPLEU03
H_HYPLEU02

Figure 2. Host association matrix for avian haemosporidians in the Shola sky-islands. (Left) Bayesian phylogenetic tree of Plasmodium spp. (blue) and Haemoproteus
spp. (red) lineages based on cytochrome b gene sequence data, with Leucocytozoon spp. as outgroups. Bayesian posterior probability support values are colour
coded. (Top) Bayesian phylogenetic tree of Shola sky-island bird species. See electronic supplementary material, tables S2 and S3 for details on tree tip
labels. The network matrix represents the heat map of abundance and distribution of each Plasmodium spp. and Haemoproteus spp. lineage, ranging from
cool blues/reds (low abundance) to warm blues/reds (high abundance), respectively. White circles in the coloured cells indicate significance of the network
metric value against null expectations. Triangles depicted on the edges of the matrix indicate significant values of Shannon diversity of interactions (two-
tailed test), circles show non-significance and dashes indicate an absence of infection. (Online version in colour.)
 (a) (b) 6
0.4

royalsocietypublishing.org/journal/rspb
parasite phylogenetic distance
0.6 0.3

F statistic
0.4 0.2

0.1
0.2

0
0

Proc. R. Soc. B 286: 20190439

FHost/Total FRegion/Host FSite/Region Phylo Eco Biogeo Geo Elev
Dist Dist Gap Dist Dist
Figure 3. Biogeographic structuring of Shola sky-island haemosporidian lineages. (a) AMOVA representing the effects of host species barriers (FHost/Total ), biogeo-
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

graphic regions within host species (FRegion/Host), and geographical site within each biogeographic region (FSite/Region) on parasite genetic structure for Plasmodium
spp. (blue) and Haemoproteus spp. (red). (b) MRM analysis representing the effects of host phylogenetic distance (Phylo Dist), host ecological distance (Eco Dist),
biogeographic gaps (Biogeo Gap), geographical distance (Geo Dist), and elevational distance (Elev Dist) on parasite phylogenetic structure for Plasmodium spp. (blue)
and Haemoproteus spp. (red). Filled symbols indicate F-values significantly different from random expectation with their 95% bootstrap confidence intervals. (Online
version in colour.)

than expected by chance; p , 0.001; electronic supplementary with old host evolutionary histories and many endemic
material, figures S4 and S5). However, CORE-PA revealed co- host radiations, parasites likely benefit by establishing host-
speciation for Haemoproteus spp., with inferred co-speciation specialized associations and diversify by co-speciation
events significantly greater than expected by chance ( p ¼ rather than adapting a generalist strategy and having more
0.05), while other host-switching, sorting, or duplication opportunities for host-switching, as suggested earlier [19,61].
events did not differ significantly from random expectations.
For Plasmodium spp., none of the events occurred significantly
more than expected by chance (electronic supplementary (d) Parasite genetic structure
material, table S8). AMOVA revealed that parasite genetic differentiation between
Overall, as expected, our cophylogenetic analyses host species was low for Plasmodium spp. (FHost/Total ¼ 0.073,
revealed a signal of host –parasite congruence mediated by p ¼ 0.045) and high for Haemoproteus spp. (FHost/Total ¼ 0.688,
co-speciation for specialist Haemoproteus spp., but lack of con- p ¼ 0.001; figure 3; electronic supplementary material, table
gruence for the generalist Plasmodium spp. The significant S10). We found a significant effect of biogeographic gaps,
role of co-speciation versus host-switching in the evolution- within host species on the genetic structure of Plasmodium
ary history of Haemoproteus spp. in the Shola sky-islands is spp. (FRegion/Host ¼ 0.208, p ¼ 0.004) but not Haemoproteus spp.
in contrast with previous studies that recognize host-switch- (FRegion/Host ¼ 0.031, p ¼ 0.464). However, there was significant
ing as the dominant coevolutionary mechanism [59–61]. parasite genetic structure between sampling sites within biogeo-
Our study suggests that coevolutionary mechanisms under- graphic regions for both Plasmodium spp. (FSite/Region ¼ 0.079,
lying diversification of avian haemosporidians are likely p ¼ 0.007) and Haemoproteus spp. (FSite/Region ¼ 0.113, p ¼
more complex than has been anticipated earlier. Employing 0.018; figure 3; electronic supplementary material, table S10).
a probabilistic approach such as approximate Bayesian com- Furthermore, multiple regressions on distance matrices
putation represents a useful future direction for an improved (MRM) analyses showed that Plasmodium spp. parasite genetic
understanding of avian haemosporidian diversification as distance was significantly associated with biogeographic gaps
has been proposed recently [62]. (B ¼ 0.229, t ¼ 4.686, p ¼ 0.003) and geographical distance
The specialist strategy of Haemoproteus spp. and history of (B ¼ 0.094, t ¼ 4.425, p ¼ 0.002) but not with host phylogenetic
co-speciation may have facilitated its diversification in the (B ¼ 0.020, t ¼ 0.809, p ¼ 0.592), ecological (B ¼ 0.023, t ¼
Shola sky-island bird community. For example, three special- 0.656, p ¼ 0.609), or elevational distance (B ¼ 20.027,
ist lineages—MONCAC03, MONFAI02, and MONMER02— t ¼ 21.101, p ¼ 0.360; figure 3; electronic supplementary
showed signals of co-speciation and have co-diversified material, table S11). Alternatively, Haemoproteus spp. parasite
with their endemic hosts Montecincla cachinnans, Montecincla genetic distance was significantly associated with host phy-
fairbanki, and Montecincla meridionalis, respectively (see elec- logenetic (B ¼ 0.059, t ¼ 18.157, p ¼ 0.014), ecological (B ¼
tronic supplementary material, figure S5). Our results 0.164, t ¼ 44.794, p ¼ 0.001), and elevational distance (B ¼
further strengthen the patterns of local diversification of 0.053, t ¼ 16.614, p ¼ 0.037), but was not affected by biogeo-
avian haemosporidians observed in other tropical bird graphic gaps (B ¼ 0.017, t ¼ 3.114, p ¼ 0.558) or geographical
communities [49,51]. Broadly, empirical data from other distance (B ¼ 0.001, t ¼ 0.389, p ¼ 0.909; figure 3; electronic
host–parasite systems suggest that parasites tend to be supplementary material, table S11).
host-specialists in species-rich communities [63]. Similarly, From an eco-evolutionary perspective, parasites are intrin-
in the highly diverse Shola sky-island bird communities sically tied to their hosts and may be affected by host
 (a) (b) 7

royalsocietypublishing.org/journal/rspb
Proc. R. Soc. B 286: 20190439
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

0.50
DKN

DKN
0.50

0.05 0.05
5 10 15 5 15 25
lineage number lineage number
Figure 4. Global phylogenetic structure based on nearest neighbour phylogenetic distance (DKN). Bayesian phylogenetic trees for (a) Plasmodium spp. and (b)
Haemoproteus spp. lineages based on cytochrome b gene sequence data obtained from the MalAvi database and endemic (closed circles) and non-endemic
(open circles) lineages recovered from the Shola sky-islands. Leucocytozoon spp. were used as an outgroup. Inset shows the observed (circles) and expected
(line) nearest neighbour phylogenetic distance (DKN) for each Shola sky-island haemosporidian lineage. (Online version in colour.)

phylogeography. Thus, given the effect of biogeographic gaps isolation, a finding consistent with earlier studies [16,52].
in the Western Ghats on host phylogeographic structure, we Interestingly, Haemoproteus spp. populations were structured
expected to find a similar phylogeographic structure among by elevation compared to Plasmodium spp., indicating a higher
the parasite lineages. Indeed, at large spatial scales, Plasmodium probability of elevational spread by Plasmodium spp., which
spp. lineages revealed a phylogeographic structure across the has critical implications from the perspective of disease
biogeographic gaps. Surprisingly, Haemoproteus spp. structure emergence in novel climatic niches.
was not affected by biogeographic gaps, suggesting that these Broadly, our results provide interesting insights into how
parasites tend to track their hosts closely and have likely colo- hosts may be analogous to islands from the perspective of para-
nized their hosts before genetic divergence of the hosts. It was site colonization [64]. For instance, in the case of Plasmodium
especially surprising that even host species (e.g. Sholicola spp. spp., biogeographic gaps influenced parasite genetic structure,
and Montecincla spp.) that showed deep genetic divergence indicating that host communities in each sky-island group
(approx. 4–5 Ma; [24]) across the biogeographic gaps were served as islands. By contrast, host phylogenetic and ecological
infected by similar Haemoproteus spp. lineages across their differences constrained the dispersal of Haemoproteus parasites,
range. This could likely occur owing to differences in mutation thus characterizing each host species as islands.
rates of parasites compared to their hosts. Additionally, an
open and interesting question remains regarding the role of
the dipteran vectors in facilitating dispersal of Haemoproteus (e) Global parasite phylogenetic structure
spp. lineages across the biogeographic gaps. We found that phylogenetic clustering in Plasmodium spp.
Within a biogeographic region, we found that Plasmodium lineages from the Shola sky-islands did not differ from a
spp. lineages were shared more among geographically closer random sample of lineages from the global parasite pool at the
hosts and did not show any host phylogenetic or ecological community or lineage level (mean DKN ¼ 0.533, p ¼ 0.272; two
constraints, coherent with their generalist strategy and a charac- of 18 lineages had DKN lower than expected; figure 4; electronic
teristic that likely contributes to its role as an emerging parasite supplementary material, table S12). By contrast, Haemoproteus
in novel bird communities. By contrast, specialist Haemoproteus spp. lineages showed strong phylogenetic clustering at both
spp. lineages were shared more among closely related hosts community and lineage level (mean DKN ¼ 0.281, p ¼ 0.002;
(phylogenetically and ecologically), despite their geographical 14 of 29 lineages had DKN less than expected; figure 4; electronic
 supplementary material, table S13). Overall, Haemoproteus spp. elevational range expansion and were more likely to 8
lineages had a significantly higher chance of being clustered emerge when introduced to novel environments. In the

royalsocietypublishing.org/journal/rspb
compared to Plasmodium spp. lineages (b + s.e. ¼ 1.958 + 0.86, Shola sky-islands, this is an especially troubling finding as
odds ratio ¼ 7.086; z ¼ 2.277, p ¼ 0.023). high-elevation habitats harbour a higher number of endemic
In line with our expectations, the generalist Plasmodium host species, which are also more likely to have evolved with
spp. lineages were widely interspersed across their global avian haemosporidian parasites (for example, see [8]).
phylogeny, whereas specialist Haemoproteus spp. lineages Overall, our results reveal that the higher likelihood of
were phylogenetically more clustered. This suggests that Hae- emergence in novel host communities by Plasmodium spp.
moproteus spp. have likely diversified in the Western Ghats, versus Haemoproteus spp. was likely driven by two interrelated
owing to the relatively old origin [65] and the deep evolution- mechanisms. First, there are a few Plasmodium spp. lineages
ary history of Western Ghats endemic avian hosts [25] such as that can infect a diverse array of host species without being
Sholicola spp. and Montecincla spp., which diverged from constrained by host phylogenetic/ecological similarity, and
their most recent ancestor about 11– 12 Ma and later diversi- thus, these lineages could emerge rapidly when introduced
fied on the Shola sky-islands about 4–5 Ma [24]. The lack of into a novel host community. Second, a few host species har-

Proc. R. Soc. B 286: 20190439

phylogenetic clustering among the Plasmodium spp. lineages bour a high diversity of Plasmodium spp. lineages, and thus
suggests that these parasites are a random sample of their invasion of such hosts into a novel bird community will be
global phylogenetic pool and remain unconstrained by host associated with the introduction of multiple parasite lineages,
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

phylogeny, further highlighting their potential as emerging increasing the likelihood of spillover to native hosts. Conse-
parasites in novel host communities. quently, Plasmodium spp. lineages were globally widespread,
reiterating their increased potential for colonization and emer-
gence in novel host communities. Elucidating the underlying
4. Conclusion ecological and evolutionary factors that contribute to the
rapid emergence of some parasites (e.g. Plasmodium spp.) but
We present one of the first comprehensive investigations of
not others (e.g. Haemoproteus spp.) has critical implications for
avian haemosporidian dynamics in the Indian subcontinent
an improved understanding of emerging infectious diseases.
(see also [66]) by sampling almost the entire bird community
in an important biodiversity hotspot. Here, we addressed the Data accessibility. The parasite genetic data generated in the current
differential effects of geographical, climatic, and host species study have been deposited in the National Center for Biotechnology
barriers in shaping generalist and specialist haemosporidian Information (NCBI) under the following GenBank Accession
parasite community structure. Our results reveal that, in a Numbers: MK493368–MK493401. Other datasets supporting this article
continental island system with long host –parasite coevolu- have been provided as part of the electronic supplementary material.
tionary history, there were several novel haemosporidian Authors’ contributions. P.G., V.V.R., and G.D.: conceived and designed the
study; C.K.V. and V.V.R.: coordinated and collected field data; P.G.:
parasite lineages, endemic to the Shola sky-islands. Plasmo- molecular laboratory work and sequence alignments; P.G. and
dium spp. and Haemoproteus spp. clearly differed in terms of G.D.: data analyses, drafted the manuscript; P.G., C.K.V., U.R.,
their host diversity, with higher host specialization in the V.V.R., and G.D.: revised and edited the manuscript. All authors
case of the latter but not in the former. Consequently, there gave final approval for publication.
was a strong signal of co-speciation in the coevolutionary Competing interests. We declare we have no competing interests.
history of Haemoproteus spp., but not in Plasmodium spp. Funding. Financial support was provided from the US Department of
Energy Financial Assistance Award no. DE-EM0004391 to the Uni-
These parasites also differed dramatically in terms of their
versity of Georgia Research Foundation. This work was also
emerging infectious disease risk, with sharing of generalist funded by the Ramanujan Fellowship from the Department of
Plasmodium spp. lineages among multiple host species Science and Technology, India, to G.D., National Geographic Society
primarily constrained by geographical factors such as geo- Research and Exploration grant to V.V.R., DAE award to study
graphical proximity, whereas specialist Haemoproteus spp. Indian Biogeography, Ramanujan Fellowship and NCBS internal
funding to U.R.
lineages were more influenced by host species factors such
as host phylogeny, host ecology, and climatic factors driven Acknowledgements. We thank: Permits—Forest Departments of Kerala
(R. Rajaraja Varma, V. Gopinathan, T.M. Manoharan, and B.S.
by elevation. Critically, our analyses revealed that Plasmodium Corrie) and Tamilnadu (R. Sunderaraju); GIS data—M. Arasumani;
spp. were less affected by climatic gradients (i.e. elevation), laboratory work—Anuradha Chandramouli; comments on manu-
indicating that these parasites had a higher likelihood of script—Farah Ishtiaq.

References
1. Rogalski MA, Gowler CD, Shaw CL, Hufbauer RA, 3. Dunn AM, Hatcher MJ. 2015 Parasites and biological community. Proc. Natl Acad. Sci. USA 103,
Duffy MA. 2017 Human drivers of ecological and invasions: parallels, interactions, and control. Trends 3165– 3170. (doi:10.1073/pnas.0506889103)
evolutionary dynamics in emerging and Parasitol. 31, 189–199. (doi:10.1016/j.pt.2014.12. 6. Valkiŭnas G. 2005 Avian malaria parasites and other
disappearing infectious disease systems. Phil. 003) haemosporidia. Boca Raton, FL: CRC Press.
Trans. R. Soc. B 372, 20160043. (doi:10.1098/rstb. 4. Tompkins DM, White AR, Boots M. 2003 Ecological 7. Schoener ER, Banda M, Howe L, Castro IC, Alley MR.
2016.0043) replacement of native red squirrels by invasive greys 2014 Avian malaria in New Zealand. N. Z. Vet. J. 62,
2. Preston DL, Mischler JA, Townsend AR, Johnson PTJ. driven by disease. Ecol. Lett. 6, 189 –196. (doi:10. 189–198. (doi:10.1080/00480169.2013.871195)
2016 Disease ecology meets ecosystem science. 1046/j.1461-0248.2003.00417.x) 8. Niebuhr CN, Poulin R, Tompkins DM, Bensch S,
Ecosystems 19, 737– 748. (doi:10.1007/s10021-016- 5. Lips KR et al. 2006 Emerging infectious disease and Križanauskienė A, Hasselquist D. 2016 Is avian
9965-2) the loss of biodiversity in a Neotropical amphibian malaria playing a role in native bird declines in New
 Zealand? Testing hypotheses along an elevational 21. McCormack JE, Huang H, Knowles LL. 2009 Sky 35. Oksanen J et al. 2016 Vegan: community ecology 9
gradient. PLoS ONE 11, e0165918. (doi:10.1371/ islands. In Encyclopedia of islands (eds R Gillespie, package. R Package version 2.4-1. https//CRAN.R-

royalsocietypublishing.org/journal/rspb
journal.pone.0165918) D Clague), pp. 839– 843. Berkeley, CA: University of project.org/package=vegan
9. Marzal A, Garcı́a-Longoria L, Cárdenas Callirgos JM, California Press. 36. Webb CO. 2000 Exploring the phylogenetic structure
Sehgal RNM. 2015 Invasive avian malaria as an 22. Robin VV, Nandini R. 2012 Shola habitats on sky islands: of ecological communities: an example for rain
emerging parasitic disease in native birds of Peru. status of research on montane forests and grasslands in forest trees. Am. Nat. 156, 145–155. (doi:10.1086/
Biol. Invasions 17, 39 –45. (doi:10.1007/s10530- southern India. Curr. Sci. 103, 1427–1437. 303378)
014-0718-x) 23. Myers N, Mittermeier RS, Fonseca GAB, Kent J. 2000 37. Poulin R, Krasnov BR, Mouillot D. 2011 Host
10. Warner RE. 1968 The role of introduced diseases in Biodiversity hotspots for conservation priorities. specificity in phylogenetic and geographic space.
the extinction of the endemic Hawaiian avifauna. Nature 403, 853–858. (doi:10.1038/35002501) Trends Parasitol. 27, 355–361. (doi:10.1016/j.pt.
Condor 70, 101–120. (doi:10.2307/1365954) 24. Robin VV, Vishnudas CK, Gupta P, Rheindt FE, 2011.05.003)
11. Clark NJ, Clegg SM, Lima MR. 2014 A review of Hooper DM, Ramakrishnan U, Reddy S. 2017 Two 38. Kembel SW, Cowan PD, Helmus MR, Cornwell WK,
global diversity in avian haemosporidians new genera of songbirds represent endemic Morlon H, Ackerly DD, Blomberg SP, Webb CO. 2010
(Plasmodium and Haemoproteus: Haemosporida): radiations from the Shola Sky Islands of the Western Picante: R tools for integrating phylogenies and

Proc. R. Soc. B 286: 20190439

new insights from molecular data. Int. J. Parasitol. Ghats, India. BMC Evol. Biol. 17, 31. (doi:10.1186/ ecology. Bioinformatics 26, 1463–1464. (doi:10.
44, 329–338. (doi:10.1016/j.ijpara.2014.01.004) s12862-017-0882-6) 1093/bioinformatics/btq166)
12. Svensson-Coelho M, Ricklefs RE. 2011 Host 25. Robin VV, Vishnudas CK, Gupta P, Ramakrishnan U. 39. Charleston MA, Robertson DL. 2002 Preferential host
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

phylogeography and beta diversity in avian 2015 Deep and wide valleys drive nested switching by primate lentiviruses can account for
haemosporidian (Plasmodiidae) assemblages of the phylogeographic patterns across a montane bird phylogenetic similarity with the primate phylogeny.
Lesser Antilles. J. Anim. Ecol. 80, 938 –946. (doi:10. community. Proc. R. Soc. B 282, 20150861. (doi:10. Syst. Biol. 51, 528–535. (doi:10.1080/
1111/j.1365-2656.2011.01837.x) 1098/rspb.2015.0861) 10635150290069940)
13. Fallon SM, Bermingham E, Ricklefs RE. 2005 Host 26. Slatyer RA, Hirst M, Sexton JP. 2013 Niche breadth 40. Balbuena JA, Mı́guez-Lozano R, Blasco-Costa I,
specialization and geographic localization of avian predicts geographical range size: a general Naisbit R, Alvarez N. 2013 PACo: a novel procrustes
malaria parasites: a regional analysis in the lesser ecological pattern. Ecol. Lett. 16, 1104– 1114. application to cophylogenetic analysis. PLoS ONE 8,
antilles. Am. Nat. 165, 466–480. (doi:10.1086/ (doi:10.1111/ele.12140) e61048. (doi:10.1371/journal.pone.0061048)
428430) 27. Robin VV, Sinha A, Ramakrishnan U. 2010 Ancient 41. Paradis E, Schliep K. 2019 ape 5.0: an environment
14. Fecchio A, Wells K, Bell JA, Tkach VV, Lutz HL, geographical gaps and paleo-climate shape the for modern phylogenetics and evolutionary analyses
Weckstein JD, Clegg SM, Clark NJ. 2019 Climate phylogeography of an endemic bird in the Sky in R. Bioinformatics 35, 526– 528. (doi:10.1093/
variation influences host specificity in avian malaria islands of southern India. PLoS ONE 5, e13321. bioinformatics/bty633)
parasites. Ecol. Lett. 22, 547– 557. (doi:10.1111/ele. (doi:10.1371/journal.pone.0013321) 42. Conow C, Fielder D, Ovadia Y, Libeskind-Hadas R.
13215) 28. Hellgren O, Waldenström J, Bensch S, Heligren O, 2010 Jane: a new tool for the cophylogeny
15. Beadell JS, Covas R, Gebhard C, Ishtiaq F, Melo M, Waldenstrom J, Bensch S, Hellgren O, Waldenström reconstruction problem. Algorithms Mol. Biol. 5, 16.
Schmidt BK, Perkins SL, Graves GR, Fleischer RC. J, Bensch S. 2004 A new PCR assay for simultaneous (doi:10.1186/1748-7188-5-16)
2009 Host associations and evolutionary studies of Leucocytozoan, Plasmodium, and 43. Merkle D, Middendorf M, Wieseke N. 2010 A
relationships of avian blood parasites from West Haemoproteus from avian blood. J. Parasitol. 90, parameter-adaptive dynamic programming
Africa. Int. J. Parasitol. 39, 257 –266. (doi:10.1016/ 797 –802. (doi:10.1645/GE-184R1) approach for inferring cophylogenies. BMC Bioinf.
j.ijpara.2008.06.005) 29. Ronquist F, Huelsenbeck JP. 2003 MrBayes 3: 11, S60. (doi:10.1186/1471-2105-11-S1-S60)
16. Fecchio A et al. 2017 Host community similarity and Bayesian phylogenetic inference under mixed 44. Goudet J. 2005 HIERFSTAT, a package for R to
geography shape the diversity and distribution of models. Bioinformatics 19, 1572–1574. (doi:10. compute and test hierarchical F-statistics. Mol. Ecol.
haemosporidian parasites in Amazonian birds. 1093/bioinformatics/btg180) Notes 5, 184–186. (doi:10.1111/j.1471-8286.2004.
Ecography (Cop.) 41, 505–515. (doi:10.1111/ecog. 30. Bensch S, Hellgren O, Perez-Tris J. 2009 MalAvi: a 00828.x)
03058) public database of malaria parasites and related 45. Lichstein JW. 2007 Multiple regression on distance
17. Ishtiaq F et al. 2007 Prevalence and diversity of haemosporidians in avian hosts based on matrices: a multivariate spatial analysis tool. Plant Ecol.
avian hematozoan parasites in Asia: a regional mitochondrial cytochrome b lineages. Mol. Ecol. 188, 117–131. (doi:10.1007/s11258-006-9126-3)
survey. J. Wildl. Dis. 43, 382 –398. (doi:10.7589/ Resour. 9, 1353–1358. (doi:10.1111/j.1755-0998. 46. Goslee S, Urban D. 2015 Package ‘ ecodist ‘. R Top. Doc.
0090-3558-43.3.382) 2009.02692.x) 47. Gower JC. 1971 A general coefficient of similarity
18. Svensson-Coelho M, Blake JG, Loiselle BA, Penrose 31. Nipperess DA, Matsen FA. 2013 The mean and and some of its properties. Biometrics 27,
AS, Parker PG, Ricklefs RE. 2013 Diversity, variance of phylogenetic diversity under rarefaction. 857–871. (doi:10.2307/2528823)
prevalence, and host specificity of avian Methods Ecol. Evol. 4, 566– 572. (doi:10.1111/2041- 48. Wilman H, Belmaker J, Simpson J, de la Rosa C,
plasmodium and haemoproteus in a western 210X.12042) Rivadeneira MM, Jetz W. 2014 EltonTraits 1.0:
Amazon assemblage. Ornithol. Monogr. 76, 1 –47. 32. R Development Core Team R. 2016 R: a language species-level foraging attributes of the world’s birds
(doi:10.1525/om.2013.76.1.1) and environment for statistical computing. R and mammals. Ecology 95, 2027. (doi:10.1890/13-
19. Moens MAJ, Pérez-Tris J. 2016 Discovering potential Foundation for Statistical Computing, Vienna, 1917.1)
sources of emerging pathogens: South America is a Austria. See http://www.r-project.org. 49. Ellis VA, Sari EHR, Rubenstein DR, Dickerson RC,
reservoir of generalist avian blood parasites. 33. Rzanny M, Voigt W. 2012 Complexity of Bensch S, Ricklefs RE. 2018 The global
Int. J. Parasitol. 46, 41 –49. (doi:10.1016/j.ijpara. multitrophic interactions in a grassland ecosystem biogeography of avian haemosporidian parasites is
2015.08.001) depends on plant species diversity. J. Anim. Ecol. characterized by local diversification and
20. Ewen JG, Bensch S, Blackburn TM, Bonneaud C, 81, 614– 627. (doi:10.1111/j.1365-2656.2012. intercontinental dispersal. Parasitology 146,
Brown R, Cassey P, Clarke RH, Pérez-Tris J. 2012 01951.x) 213–219. (doi:10.1017/S0031182018001130)
Establishment of exotic parasites: the origins and 34. Blüthgen N, Fründ J, Vazquez DP, Menzel F. 2008 50. Clark NJ. 2018 Phylogenetic uniqueness, not
characteristics of an avian malaria community in an What do interaction network metrics tell us about latitude, explains the diversity of avian blood
isolated island avifauna. Ecol. Lett. 15, 1112 –1119. specialization and biological traits? Ecology 89, parasite communities worldwide. Glob. Ecol.
(doi:10.1111/j.1461-0248.2012.01833.x) 3387 –3399. (doi:10.1890/07-2121.1) Biogeogr. 27, 744–755. (doi:10.1111/geb.12741)
 51. Fecchio A et al. 2019 Avian host composition, local Parasitol. 120, 372–380. (doi:10.1016/j.exppara. switching in avian malaria parasites. Syst. Biol. 53, 10
speciation, and dispersal drive the regional 2008.09.001) 111–119. (doi:10.1080/10635150490264987)

royalsocietypublishing.org/journal/rspb
assembly of avian malaria parasites in South 56. Ortiz-Catedral L et al. 2019 Haemoproteus minutus 61. Fecchio A et al. 2018 Diversification by host
American birds. Mol. Ecol. (doi:10.1111/mec.15094) is highly virulent for Australasian and South switching and dispersal shaped the diversity and
52. Olsson-Pons S, Clark NJ, Ishtiaq F, Clegg SM. 2015 American parrots. Parasit. Vectors 12, 40. (doi:10. distribution of avian malaria parasites in Amazonia.
Differences in host species relationships and 1186/s13071-018-3255-0) Oikos 127, 1233–1242. (doi:10.1111/oik.05115)
biogeographic influences produce contrasting 57. Hellgren O, Križanauskienė A, Hasselquist D, Bensch 62. Alcala N, Jenkins T, Christe P, Vuilleumier S. 2017
patterns of prevalence, community composition and S. 2011 Low haemosporidian diversity and one key- Host shift and cospeciation rate estimation from co-
genetic structure in two genera of avian malaria host species in a bird malaria community on a mid- phylogenies. Ecol. Lett. 20, 1014–1024. (doi:10.
parasites in southern Melanesia. J. Anim. Ecol. 84, Atlantic island (São Miguel, Azores). J. Wildl. Dis. 1111/ele.12799)
985–998. (doi:10.1111/1365-2656.12354) 47, 849– 859. (doi:10.7589/0090-3558-47.4.849) 63. Krasnov BR, Shenbrot GI, Khokhlova IS, Mouillot D,
53. Rasmussen PC, Anderton JC. 2005 Birds of South 58. Martı́nez J, Vásquez RA, Venegas C, Merino S. 2015 Poulin R. 2008 Latitudinal gradients in niche
Asia: the ripley guide. Volume 2: attributes and Molecular characterisation of haemoparasites in breadth: empirical evidence from haematophagous
status. Washington, DC, and Barcelona, Spain: forest birds from Robinson Crusoe Island: is the ectoparasites. J. Biogeogr. 35, 592 –601. (doi:10.

Proc. R. Soc. B 286: 20190439

Smithsonian Institution and Lynx Edicions. Austral thrush a potential threat to endemic birds? 1111/j.1365-2699.2007.01800.x)
54. Palinauskas V, Ziegyte R, Iezhova TA, Bernotiene R, Bird Conserv. Int. 25, 139–152. (doi:10.1017/ 64. Kuris AM, Blaustein AR, Alio JJ. 1980 Hosts as Islands.
Valki G. 2016 Description, molecular S0959270914000227) Am. Nat. 116, 570–586. (doi:10.1086/283647)
Downloaded from https://royalsocietypublishing.org/ on 10 August 2024

characterisation, diagnostics and life cycle of 59. Ricklefs RE, Outlaw DC, Svensson-Coelho M, 65. Praveen Karanth K. 2015 An island called India:
Plasmodium elongatum (lineage pERIRUB01), the Medeiros MCI, Ellis VA. 2014 Species formation by phylogenetic patterns across multiple taxonomic groups
virulent avian malaria parasite. Int. J. Parasitol. 46, host shifting in avian malaria parasites. Proc. Natl reveal endemic radiations. Curr. Sci. 108, 1847–1851.
697–707. (doi:10.1016/j.ijpara.2016.05.005) Acad. Sci. USA 111, 14 816–14 821. (doi:10.1073/ 66. Ishtiaq F, Bowden CGR, Jhala YV. 2017 Seasonal dynamics
55. Palinauskas V, Valkiūnas G, Bolshakov CV, Bensch S. pnas.1416356111) in mosquito abundance and temperature do not
2008 Plasmodium relictum (lineage P-SGS1): effects 60. Ricklefs RE, Fallon SM, Bermingham E. 2004 influence avian malaria prevalence in the Himalayan
on experimentally infected passerine birds. Exp. Evolutionary relationships, cospeciation, and host foothills. Ecol. Evol. 8040–8057. (doi:10.1002/ece3.3319)