ArticlePDF Available


Identifying where introduced animals fit in a food web relative to each other and to endemic species is key for biodiversity conservation planning. Using a multiproxy study of dog feces from eastern Madagascar, we infer that even dogs that spend time in derived grasslands typically eat forest‐derived foods. Regardless of the time that dogs spend in cleared forest, their impacts are likely concentrated on forest‐dwelling prey. If dogs in forests mostly consume threatened endemic animals (and not other introduced animals such as rats), then the exclusion of dogs from protected forests should be a priority. Introduced predators on islands can help control invasive species yet can also contribute to the extirpation and extinction of endemic taxa. The spread of dogs on Madagascar by ~1000 years ago coincided with the introduction of livestock and spread of grazer‐adapted grasslands, and we help evaluate the extent to which modern dogs are part of novel grassland food webs. To infer dog diet, we identified food remains, where possible, and conducted stable isotope ratio analysis for n = 100 modern dog feces collected in derived grassland at varying distances from protected forest edges around Analamazoatra and Andasibe‐Mantadia National Park in eastern Madagascar. Animal remains in feces and the observed range of fecal δ15N values are consistent with dog meals at multiple trophic levels. However, the observed distribution of fecal δ13C values suggest that few dogs in the study area consumed food derived from open C4 grasslands. Existing data suggest that dogs rely primarily on C3 consumers inhabiting forest biomes (forest‐dwelling animals) for their prey, which may include endemics such as tenrecs, Malagasy rodents, and lemurs and introduced rodents such as rats. These findings indicate that dogs are not confined to the anthropogenic niche defined by grazer‐adapted grasslands, but rather use and impact animal food resources associated with protected forests. Higher resolution study of dog diet and mobility can further clarify the potential for dogs to exploit endemic prey, compete with endemic predators, and spread disease across ecotones. L'identification de la place des animaux introduits dans la chaine alimentaire les uns par rapport aux autres et aux espèces endémiques est essentielle pour la planification de la conservation de la biodiversité. En utilisant une étude multiproxy sur les excréments de chiens de l'est de Madagascar, nous en déduisons que même les chiens qui passent du temps dans les prairies dérivées mangent généralement des aliments dérivés de la forêt. Quel que soit le temps que les chiens passent dans la forêt défrichée, leurs impacts sont probablement concentrés sur les proies vivant dans la forêt. Si la majeure partie de l'alimentation des chiens dans les forêts provient d'animaux endémiques menacés (et non d'autres animaux introduits tels que les rats), l'exclusion des chiens des forêts protégées devrait être une priorité. Ny fahalalana manokana mahakasika ny anjara toerana sy fifampiakinany ireo biby samy tsy zanatany sy ireo zanatany eo amin'ny famatsiana sakafo dia tena zava‐dehibe tokoa eo amin'ny fahafahana miaro sy mametraka drafitra ho fiarovana azy ireo sy ny tontolo manodidina azy. Ny fampiasana ny atotam‐pahalalana sy hevitra mahakasika tain'alika any amin'ny ilany atsinanan'i nosy Madagasika, dia ahafahana milaza fa na dia ny alika izay monina eny aminy toerana tsy misy ala aza dia mihinana sakafo vokatra mivatana na akolana avy aminy ala. Na dia eo azy ny fotoana lanin'ny alika (mikarenjy) eny amin'ny toerana tsy misy ala. Ny trindry dia vinavinaina fa mianjerana amin'ny ireo biby fihaza miankina sy miaina ao anaty ala Raha mifototra aminy biby zanatany izay efa ho lany tamingana monina ao anaty ala mantsy ny ankamaroan'ny sakafon'alika, fa tsy amin'ireo biby tsy zanatany natsofoka teo aminy nosy toy ny voalavo, noho izany dia tena laharam‐pahamehana ary tsy azo iodivirana ny fanalana tanteraka ny alika aminy ireny ala voaharo ireny. Identifying where introduced animals fit in a food web relative to each other and to endemic species is key for biodiversity conservation planning. Using a multiproxy study of dog feces from eastern Madagascar, we infer that even dogs that spend time in derived grasslands typically eat forest‐derived foods. Regardless of the time that dogs spend in cleared forest, their impacts are likely concentrated on forest‐dwelling prey. If dogs in forests mostly consume threatened endemic animals (and not other introduced animals such as rats), then the exclusion of dogs from protected forests should be a priority.
Dogs occupying grassy habitat near protected areas in eastern
Madagascar rely on foods from forests
Sean W. Hixon
| Mikhaela Neelin
| Stephanie Chan
| Dominic Mayo
Caitlynn Filla
| Zach J. Farris
| Susan deFrance
| John Krigbaum
| Kim Valenta
Max Planck Institute for Geoanthropology,
Jena, Germany
Department of Natural Resource Sciences,
McGill University, Ste-Anne-de-Bellevue,
Nunavik Hunting Fishing Trapping
Association, Tasiujaq, Canada
School of Environment, McGill University,
Ste-Anne-de-Bellevue, Canada
Department of Natural Resources and
Environmental Studies, University of Northern
British Columbia, Prince George, Canada
Department of Anthropology, University of
Michigan, Ann Arbor, Michigan, USA
Department of Anthropology, University of
Florida, Gainesville, Florida, USA
Department of Health & Exercise Science,
Appalachian State University, Boone, North
Carolina, USA
Sean W. Hixon, Max Planck Institute for
Geoanthropology, Jena, Germany.
Societal Impact Statement
Identifying where introduced animals fit in a food web relative to each other and to
endemic species is key for biodiversity conservation planning. Using a multiproxy
study of dog feces from eastern Madagascar, we infer that even dogs that spend time
in derived grasslands typically eat forest-derived foods. Regardless of the time that
dogs spend in cleared forest, their impacts are likely concentrated on forest-dwelling
prey. If dogs in forests mostly consume threatened endemic animals (and not other
introduced animals such as rats), then the exclusion of dogs from protected forests
should be a priority.
Introduced predators on islands can help control invasive species yet can also con-
tribute to the extirpation and extinction of endemic taxa. The spread of dogs on
Madagascar by 1000 years ago coincided with the introduction of livestock and
spread of grazer-adapted grasslands, and we help evaluate the extent to which
modern dogs are part of novel grassland food webs.
To infer dog diet, we identified food remains, where possible, and conducted sta-
ble isotope ratio analysis for n =100 modern dog feces collected in derived grass-
land at varying distances from protected forest edges around Analamazoatra and
Andasibe-Mantadia National Park in eastern Madagascar.
Animal remains in feces and the observed range of fecal δ
N values are consis-
tent with dog meals at multiple trophic levels. However, the observed distribution
of fecal δ
C values suggest that few dogs in the study area consumed food
derived from open C
Existing data suggest that dogs rely primarily on C
consumers inhabiting forest
biomes (forest-dwelling animals) for their prey, which may include endemics such
as tenrecs, Malagasy rodents, and lemurs and introduced rodents such as rats.
These findings indicate that dogs are not confined to the anthropogenic niche
defined by grazer-adapted grasslands, but rather use and impact animal food
resources associated with protected forests. Higher resolution study of dog diet
Received: 30 May 2022 Revised: 12 August 2022 Accepted: 12 August 2022
DOI: 10.1002/ppp3.10319
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited and is not used for commercial purposes.
© 2022 The Authors. Plants, People, Planet published by John Wiley & Sons Ltd on behalf of New Phytologist Foundation.
Plants People Planet. 2022;110. 1
and mobility can further clarify the potential for dogs to exploit endemic prey,
compete with endemic predators, and spread disease across ecotones.
diet, dog, feces, feeding ecology, grassland, predator, stable isotope
People manage and clear Madagascar's forests (Harper et al., 2007;
Kaufmann & Tsirahamba, 2006) and, in the process, have introduced
an expanding multitude of plants and animals (Kolby, 2014; Kull
et al., 2012). This ecological transformation is part of a long-term
trend, as people brought significant changes to the island during the
past millennium by contributing to the expansion of grasslands
(Burney, 1987; Domic et al., 2021; Razanatsoa, 2019; Virah-Sawmy
et al., 2016), the proliferation of introduced grazers (Hixon, Douglass,
Crowley, et al., 2021), and the extinction of endemic megaherbivores
(Burney et al., 2004; Crowley, 2010; Hansford et al., 2021). Remote
sensing data from 2018 suggest that over half of the island is covered
by grassland (Zhang et al., 2020), and existing data from functional
ecology and paleoecology suggest that the island's grazer-adapted
grassland biome is a recent phenomenon (Godfrey & Crowley, 2016;
Joseph & Seymour, 2020). The distinction between novel and ancient
grasslands is important, because derived grasslands tend to be ecolog-
ically depauperate (Veldman & Putz, 2011). Today, derived grasslands
in Madagascar's eastern lowlands can form in as little as 30 years
(Joseph et al., 2021; Styger et al., 2007), which has led to an abun-
dance of grasslands even in wet regions (MAP >2000 mm/yr) where
natural savanna is typically scarce (Joseph et al., 2022). The associated
spread of animal husbandry and derived grasslands provides compel-
ling examples of how positive interactions among humans and intro-
duced plants and animals can have compounding negative effects on
endemic biodiversity. This type of synergy exists in other disturbed
island ecosystems such as on Hawai'i, where the frugivory of intro-
duced rodents helps spread invasive plants (e.g., Shiels, 2011).
However, negative interactions may exist also among introduced
taxa, which has motivated the control of pest species through further
species introductions (Messelink et al., 2008; Peacock &
Abbott, 2010). Such attempts at biological control are limited by our
inability to predict ecosystem-wide impacts of introduced species
(Simberloff & Stiling, 1996). On Madagascar, basic research continues
to highlight what may be unexpected aspects of introduced animal
diets (e.g., reliance of livestock on a variety of woody and succulent
plants rather than primarily grass in the arid SW; Feldt et al., 2017;
Kaufmann & Tsirahamba, 2006). Research on the diverse diets of
introduced predators (e.g., dogs, cats, and Indian civet) can clarify the
extent to which they control populations of other introduced taxa
(e.g., rats and mice), versus those of endemic prey and potential
competitors in Madagascar's forests.
Most of Madagascar's modern dogs have African ancestry
(Ardalan et al., 2015), and their spread to the island may have
coincided with the introduction of livestock from mainland Africa. No
other canids inhabited the island prior to the introduction of
domesticated dogs. Radiocarbon dated dog bone recovered from
archaeological and paleontological deposits across Madagascar spans
the past millennium (Hixon, Douglass, Godfrey, et al., 2021), which left
potential for dogs to aid human-led hunts of now-extinct megafauna
(Burney et al., 2020). Today, dogs kill endemic rodents such as the
Malagasy giant jumping rat (Hypogeomys antimena; Sommer &
Hommen, 2000), occasionally stalk lemurs (Brockman et al., 2008),
and help people procure forest bushmeat (including lemurs and
tenrecs; Decary, 1939; Garcia & Goodman, 2003; Gardner &
Davies, 2014). This can contribute to negative interactions between
dogs and endemic predators such as the fosa (Cryptoprocta ferox),
which are harassed by dogs (Kshirsagar et al., 2020; Valenta
et al., 2016) and tend toward avoidance (Farris et al., 2015; Merson
et al., 2019). However, dogs are also known to prey on introduced
animals that can flourish in relatively open habitat (e.g., rodents and
chickens), and dogs rely heavily on plant and animal scraps in human-
derived food waste (Kshirsagar et al., 2020; Valenta et al., 2016).
Additionally, ancient dog bone collagen (primarily from SW
Madagascar) tends to be enriched in
C relative to ancient fosa
collagen (Hixon, Douglass, Godfrey, et al., 2021), which suggests that
dogs introduced during the past millennium tended to consume fewer
forest-derived foods than fosa. It is unclear whether dogs living near
Madagascar's remaining forests today rely primarily on foods from
forest-dwelling taxa or on taxa in open habitat.
Evidence for modern predation by dogs can be observed directly,
identified from gut or fecal remains, and inferred from the stable iso-
tope composition of consumer tissue (e.g., bone and hair), including
feces (Newsome et al., 2007). While direct observation can most
accurately and precisely characterize prey diversity and abundance,
inference based on analysis of intact feces and other consumer tissues
provide an expedient way of estimating aspects of diet integrated
over various periods of time (depending upon tissue turnover rates).
Each approach has strengths and limitations. For example, through
direct observation, one can estimate the number of introduced rats
killed and consumed per dog per week (a parameter necessary but not
sufficient for estimating dog impacts on introduced rat populations),
but it is difficult to estimate whether the killed rats foraged mostly in
open or closed habitat, for they are known to exploit both
(Goodman, 1995). Conversely, through stable isotope analysis of dog
tissue, it is possible to estimate whether most of dog diet comes from
open or closed habitat, yet it is often difficult to make precise esti-
mates of the proportion of diet that comes from a particular taxon
such as introduced rats. For an initial exploration of dog diet in the
study area, we focus on the analysis of feces, which integrate diet
over relatively short periods of time (Salvarina et al., 2013) and have
the advantage of being easier to collect than consumer tissue such as
bone, hair, or blood (Codron et al., 2007).
The precision of prey inference through fecal analysis is limited
by myriad physiological and taphonomic factors including the facts
that identifiable prey remains are typically scarce, and feces are highly
variable in composition. While feces do contain lipids and proteins
from consumed prey, they tend to include large amounts of indigest-
ible food (e.g., cellulose) and dead bacteria. Therefore, diet quality and
digestion strongly influence the composition of bulk feces and the
degree to which the stable isotope content of bulk feces accurately
approximates that of bulk prey tissue (Jones et al., 1979).
The stable carbon isotope (δ
C) content of consumer tissue is
sensitive to the types of plants at the base of the food web (Codron
et al., 2005). Specifically, trees and shrubs that use the C
thesis are significantly depleted in
C relative to C
grasses (see
Hixon, Douglass, Crowley, et al., 2021). While differences in CO
tion between C
and C
plants can explain the majority of δ
C varia-
tion in consumer tissues, multiple variables can account for observed
N variation in plant and consumer tissues (Szpak, 2014). For exam-
ple, soil nitrogen cycling can lead to
N enrichment in plants and their
consumers, particularly in arid areas (Austin & Vitousek, 1998;
Crowley et al., 2011), and consumption of foods at higher trophic
levels can also lead to
N enrichment of consumer tissue
(McCutchan et al., 2003).
To help characterize the diet of dogs living close to remaining for-
ests in eastern Madagascar, we analyzed opportunistically collected
dog feces across an open ecosystem characterized by C
adjacent to protected forest, to test the following hypotheses:
1. Dog fecal δ
C values for samples collected closer to continuous
forest are depleted in
C relative to samples collected from adja-
cent grassland and mixed grass and shrubland. This pattern in fecal
stable isotope data is most parsimoniously explained by a situation
in which dogs both opportunistically rely on forest taxa and move
little between the place of eating and the site of defecating. We
expect a broad range (>10) of dog fecal δ
C values, consistent
with the broad range of ancient dog collagen δ
C values (Hixon,
Douglass, Godfrey, et al., 2021) and the wide range of modern hair
C values in another commensal omnivore (Rattus rattus) studied
in eastern Madagascar (Dammhahn et al., 2017).
2. Dog fecal δ
N values are similar across the relatively mesic forests
and grasslands of eastern Madagascar. Consistent with the
observed omnivory among dogs, we also expect a wide range of
fecal δ
N values (>6) and for feces in many cases to be
enriched in
N relative to the tissue of potential prey
(e.g., tenrecs, rodents, and chickens).
3. Dog fecal stable isotope content matches expectations based on
identifiable prey remains in feces. Specifically, feces with abundant
undigested cellulose reflect meals with C
grass that enrich bulk
feces in
C, and feces with identifiable animal remains (e.g., bone,
hair, claws, and feathers) come from relatively carnivorous meals
that enrich bulk feces in
N. This clarifies the degree to which the
two lines of dietary inference are complementary.
2.1 |Sample collection and analysis
The subset of eastern humid forests on Madagascar that have been
protected by Analamazoatra and Andasibe-Mantadia for >30 years
are rich in biodiversity (Goodman, 2010). A cleared area of mixed
grass and shrubland separates these forests from the highland grass-
lands 20 km to the east (Bond et al., 2008). Part of the mixed grass
and shrubland is a Ramsar Wetland Site (Torotorofotsy), and there is a
proposal to keep some of the unprotected forest in the area as a natu-
ral resource reserve (Zahamena Ankeniheny; Figure 1). The vegetation
in the study area's cleared forest has not been the subject of detailed
vegetation survey. However, invasive C
grasses often dominate
cleared habitat, and some C
shrubs and native taxa (e.g., bamboos)
can also thrive in secondary stands (Binggeli, 2003). Madagascar
includes a diverse clade of endemic C
grasses that may be present in
sheltered pockets of otherwise cleared habitat, but these grasses tend
to flourish in high mountains and forest understory (Hackel
et al., 2018; Solofondranohatra et al., 2018). Where C
grasses are
abundant, the reliance of consumers on grassy ecosystems in general
cannot be readily distinguished based on consumer δ
C values.
FIGURE 1 Study region illustrating dog fecal sample locations
relative to national parks (Analamazoatra and Andasibe-Mantadia,
established in 1970 and 1989, respectively), the Ramsar Site Wetland
Site of Torotorofotsy (recognized in 2005), and the proposed natural
resource reserve of Zahamena Ankeniheny (outlined in green). Only
the sample locations marked in orange are >1 km from a settlement.
The inset includes green areas of Madagascar that outline remaining
primary vegetation (Royal Botanical Gardens, Kew). Both maps were
produced over a Google Satellite base.
Over three hundred modern dog feces samples were collected
opportunistically on forest margins and in the derived grassland sur-
rounding Analamazoatra and Andasibe-Mantadia National Parks dur-
ing June and July 2018. Of these samples, 100 were selected for
analysis by aggregating randomly selected subsets from collection
areas at different distances to the forest edge. All samples were col-
lected within 1 km of roads, and all but six come from within 1 km of
human settlement.
After collection, each fecal sample was placed in a plastic 50-ml
tube and immersed in 95% ethanol. The location of each fecal sample
was recorded using a handheld GPS. After 24 h, the ethanol from
each sample was decanted and transferred to a sterile 50-ml tube
containing indicating silica beads topped with a sterile Kimwipe to
promote desiccation. Fecal samples were then transported to the
United States for analysis. As ethanol is an organic compound, it is
important to note that its use in the preservation of biological samples
can influence the stable isotope values of some tissues, but the
magnitude of this effect on δ
C and δ
N tends to be small (1)
(Barrow et al., 2008; Sweeting et al., 2004). The effect of preservative
treatment on fecal stable isotope values likely varies according to
heterogeneous content of the feces, and existing research with
orangutan feces found no effect of 99.5% ethanol treatment on fecal
C values and only a minor effect on fecal δ
N values (0.3)
(Tsutaya et al., 2021).
At the University of Florida and University of California at Santa
Barbara, all fecal samples were visually inspected, and analysts noted
the presence of plant material, animal material (bones, hair, feathers,
and claws), and human trash (foil, textile, and wax). Analysis of the ani-
mal remains present in the feces included taxonomic identification
whenever possible. The fecal samples contain highly fragmentary ani-
mal bones, portions of animal keratin (e.g., tenrec quills and claw cov-
erings), and at least one bird feather. To identify the taxa represented,
bone and keratin fragments were examined using a binocular
microscope and/or a Dinoscope and compared to modern skeletal
specimens of known taxa. Due to fragmentation, many specimens
could only be identified to higher level taxonomic categories
(e.g., Mammalia and Vertebrata); however, some specimens could be
identified as tenrecs (species undetermined), rodents, and possibly
reptiles. Additionally, the feces were X-rayed at the University of
Florida's College of Veterinary Medicine to explore screening for
dense remains of prey. Because most feces included at least some
plant material, samples with particularly abundant plant material were
also noted, which made this measure semiquantitative.
Following the approach of Codron et al. (2007), desiccated fecal
samples were dried and homogenized, and the <1 mm fraction was
collected for stable isotope analysis at the Human Paleoecology and
Isotope Geochemistry Laboratory at the University of California at
Santa Barbara. Stable isotope (δ
C and δ
N) and elemental (weight
%C and %N) analysis was conducted at the Yale Analytical and Stable
Isotope Center (YASIC) using a Thermo Delta Plus Advantage IRMS.
The standard deviation across standard δ
C and δ
N measurements
was 0.2and 0.1, respectively (see Method S1 for details regard-
ing data analysis).
3.1 |Fecal dataset trends
Most fecal samples included visible plant remains (80% of total) and
animal tissue (63%; Dataset S1). A minority (21%) included particularly
abundant plant material, which tended to make the dried feces loosely
consolidated. Hair tended to be the most abundant animal tissue
(in 44% of total), but bone, claws, and a feather were also present
(Figure 2). Bone and hair tend to be associated, for 73% of the sam-
ples containing bone are part of the subset that also includes hair. Ani-
mal remains in a minority of the feces (19%) could be identified at
some level, and the majority of these (63%) were identified as belong-
ing to mammals. Three samples included recognizable bones of ten-
recs or rodents, and reptiles may also be present in four samples.
Additionally, a minority of samples (9% of total) included human trash
(foil, wax, and textiles).
The ranges of both δ
C and δ
N values were >10(12.0
and 12.9, respectively): δ
C values (n =100, m =26.8) ran-
ged from 28.8to 16.8, and δ
N values (n =100, m =7.5)
ranged from 2.6to 15.5(Figure 2). Distributions of fecal weight
% C:N, δ
C, and δ
N data are positively skewed and failed Shapiro
Wilk tests for normality, with test statistics (W) > 0.7 and p< 0.001.
There is a strong positive relationship between fecal δ
C and δ
values (n=100, r
=0.40, p< 0.0001). A 5increase in fecal δ
values is matched by a 1increase in fecal δ
N values, but the
former cannot reliably predict the latter (r
=0.08). Weight % C:N
values (n=100, m =10.1) ranged between 4.0 and 28.6.
There is no correlation between fecal δ
C values and distance to
forest (n=100, r
=0.16, p=0.11; Figures 2and 3). The presence of
animal remains and abundant plant remains best explains variability in
fecal δ
Cvalues(Model1,AIC=451.6; Table 1and Method S1), yet
the magnitude of the effect of these factors on fecal δ
C values is
consistently small (<1). When controlling for the abundance of plant
remains, Model 1 suggests that the δ
C values of feces containing ani-
mal remains (n=48) are insignificantly lower than those of feces that
do not contain animal remains (n=52, p=0.06). The presence of
abundant plant remains has the opposite association, with significantly
higher δ
C values in feces containing abundant plant remains (n=21)
than in feces not containing abundant plant remains (n=79, p=0.04).
Generally, δ
C data from feces with abundant plant material have a
larger midspread (interval including 50% of the data) than those from
feces without abundant plant material (6[27.5 to 21.5]as
opposed to 1[27.3 to 26.3], respectively).
Both the presence of animal remains and distance from forest
edge best explain variability in fecal δ
N values (Model
2, AIC =406.9), yet the magnitude of the effect of these variables on
fecal δ
N values is again consistently small (< 1). Model 2 suggests
that, among fecal samples with animal remains present (n=48), those
collected relatively far from forest edges have insignificantly higher
N values (p=0.07). Though the magnitude of the gradient is small,
feces collected further from the forest edge are generally relatively
enriched in
N (Figures 2and 3;n=100, r
=0.28, p=0.006).
The presence of animal remains, trash, and abundant plants and
interactions between the presence of trash and abundant plants best
explains variability in fecal C:N values (Model 3, AIC =595.8;
Table 1). Model 3 suggests that the C:N values of feces containing
trash (n=9) tends to be significantly greater than those without trash
(n=91, p=0.01).
3.2 |Comparisons with plants and animals
The distribution of dog fecal δ
C values overlaps primarily with the
distribution of δ
plants (Method S1 and Dataset S2).
The assumption of herbivory among dogs gives a maximum estimate of
food web reliance by dogs (Figure 4a), yet the estimated median
FIGURE 2 Map insets around study area showing the presence of different materials visible in dog feces (colored points in (a) and (c)) and
C and δ
N values from bulk feces (color in (b) and (d) graduated according to categories with equal sample size). Stable isotope data are given
relative to internationally accepted standards. Spatial tendencies in each small multiple are illustrated through colored surfaces in each case
(generated using a Bayesian approach in the IsoMemo application AverageR). Because feces come from sampled clusters, the points are displaced
to form gridded clusters. The inset maps are placed over a Google Satellite base map, and notes on material present in the fecal samples are
provided in pie charts to the left.
FIGURE 3 Dog fecal δ
C data (a) and δ
data (b) according to distance to the forests of
Analamazoatra and Andasibe-Mantadia National
Parks, with samples collected >1 km from
settlements marked in yellow.
C value of dog bulk diet in this case (n=100, m =26.0)falls
below the Suess-corrected C
plant δ
C value (n=61, m =14.0)
by more than 10. The assumption that dogs are secondary con-
sumers makes the estimated median δ
C value of their diet
(m =30.0) almost indistinguishable from the median Suess-
corrected C
plant δ
Cvalue(m=30.5). Approximately 7sep-
arates the median estimated dog diet δ
Nvalue(n=100, m =6.3)
from the median C
plant δ
N value (n=135, m =0.8).
Roughly approximated bulk diet stable isotope values (as inferred
from available data from feces, hair, and bone) suggest that modern
dog diet around Andasibe is depleted in
Cby10relative to
ancient dog diet in southwest Madagascar and is also depleted in
relative to the ancient diets of fosa and cattle in the
southwest (Figure 5, Method S1, and Datasets S3 and S4). Estimated
dog diet δ
C values are also lower than those of introduced mice
(Mus musculus) in the eastern study region. However, the dietary δ
values estimated from the dog fecal δ
C values are comparable with
those of a variety of forest-dwelling taxa in the region, including
endemic rodents, tenrecs, and lemurs. The central tendency of the
estimated isotopic composition of dog diet is comparable to that of
both introduced black rats (R. rattus) and the insectivorous shrew
tenrec (Microgale pusilla).
TABLE 1 Description of linear models discussed in the text
Model terms
C Model 1
AIC =451.6
N Model 2
AIC =406.9
C:N Model 3
AIC =595.8
Distance 0.56 0.81 0.60
Abundant Plants 0.89 0.46 0.75
Animals 0.80 0.66 0.84
Trash 0.52 0.59 0.93
Animals Distance 0.12 0.38 0.18
Trash Abundant Plants 0.11 0.10 0.54
Notes: See Table S1 for model outputs. Each model includes the terms highlighted in grey. Within each dataset, glmulti considers all possible models and
ranks them according to an information criterion (AIC in our case). The model-averaged importance of each term within each model reflects the support of
the variable across all possible versions of the model. These values are given in corresponding cells.
Distance=distance between sample collection point and a protected area, Abundant Plants=presence of abundant plants in feces,
Animals=presence of animal prey in feces, Trash=presence of trash in feces.
FIGURE 4 (a) Histograms showing C
, CAM, and C
plant δ
C data relative to dog fecal δ
C data corrected to be comparable with plant
isotope data assuming that dogs are primarily herbivorous (light gray) and carnivorous (dark gray). Means of δ
C values are marked with bold
vertical lines. (b) Histograms showing C
plant δ
N data (collected near the study site) relative to dog fecal δ
N data corrected to estimate the
N values of the corresponding diet. Means are marked with bold horizontal lines, and dashed blue lines approximate trophic level
(TL) estimates on the δ
N scale.
Fecal data suggest that dogs in the study region are omnivores that
primarily rely on C
-based food webs, with a relatively small reliance
on C
grasslands. The fecal δ
C data are inconsistent with the expec-
tation that dogs closer to the forest have feces relatively depleted in
C. However, the observed range of fecal δ
N values are consistent
with the expectation of omnivory. The correspondence between fecal
remains and stable isotope data is generally weak and highlights the
strengths and limitations of each approach to dietary inference.
4.1 |Dietary inference
The distribution of dog fecal δ
C values suggests that very few dogs
relied on meals from C
grasslands. Given the wide distribution of
CAM epiphytes, it is unlikely that they significantly
C enriched or
depleted the diet of omnivores in the study area. If anything, the con-
sumption and defecation of relatively indigestible tissue (presumably
dominated by C
grasses) does seem to enrich dog feces in
C, but
the magnitude of this effect is minor (Model 1). The reliance of mod-
ern dogs on C
-based food webs contrasts with existing data from
ancient bone, which suggest that both ancient dogs and zebu cattle in
southwest Madagascar relied to a greater extent on food webs based
on C
grasslands and CAM succulents. This contrast likely follows
from differences in vegetation composition (with xerophytic plants
more abundant in the southwest) and attests to dietary flexibility
in dogs.
The variability of dog fecal δ
C values is poorly explained by dis-
tance to forest, and dogs with relatively high fecal δ
C values occur
on forest margins. There are multiple plausible explanations for the
absence of a correlation between fecal δ
C values and distance to
forest. Dog meals in the derived grassland that covers most of the
study region may come from food webs based on C
shrubs, and reli-
ance on food webs involving C
grasses may similarly deplete dog
feces in
C. Also, dogs can traverse the study region in a day and can
live in villages on the margin of protected areas. Consequently, move-
ment could easily explain why feces relatively depleted in
C could
be widespread in a C
grassland and conversely why some feces
enriched in
C are present on the margin of C
forests. Additional
behavioral studies and proxies of mobility (e.g., radio collar or Sr iso-
tope work) could help resolve this possibility.
The distribution of dog fecal δ
N values is consistent with meals
at multiple trophic levels (Figure 4b). The widespread plant remains in
dog feces may be somewhat unexpected, yet this is consistent with
modern observations of dog diet (Sueda et al., 2008) and genetic evi-
dence that domestic dogs have experienced selection for starch diges-
tion (Axelsson et al., 2013). Animal remains present in feces
FIGURE 5 Stable isotope data from dogs and other introduced taxa (warm colors) and endemic taxa that are potential prey or competitors
(cool colors). Analyzed tissues include feces (dogs in present study), fur (all other taxa in top frame, modern and from eastern Madagascar:
Modern E), and subfossil bone collagen (ancient specimens in bottom frame all from SW Madagascar: Ancient SW), and all data are corrected
to approximate the isotopic composition of diet (Dataset S4). For groups that include multiple species (e.g., tenrecs in dark blue), the animal icon
and lines mark the weighted mean and standard deviation for the group. Each species has its own mean and standard deviation given as relatively
faint points and lines (when n> 1). Note that both δ
C and δ
N data from modern specimens within 150 km of the eastern study area (top
frame) are shown against only the δ
C data from ancient specimens from the southwest (spanning the past 3000 years).
(particularly bones and claws) indicate some level of carnivory and fur-
ther suggest that at least some dog meals included rodents, tenrecs,
or reptiles. Despite the significant tendency for dogs further from the
forest edge to have feces enriched in
N (Model 2; Figure 3), there is
little evidence to support the idea that dog meals further from the for-
est included more animal tissue. This is because (1) the magnitude of
the fecal δ
N gradient (Figure 3b) is small relative to the overall range
of fecal δ
N values and (2) animal remains, if anything, were more
widespread in feces collected close to the forest edge (Figure 2c). The
depletion of feces in
N closer to the forest could easily be the prod-
uct of most dog diets close to the forest edge coming from relatively
mesic habitat with lower δ
N baseline. This issue of baseline may also
help explain why dog feces enriched in
C (indicating greater reliance
on dry-adapted C
plants) tend to be enriched in
Existing data suggest that dogs and introduced rats have meals
that are isotopically similar. However, the taxonomic resolution of the
dietary inference possible through fecal δ
C and δ
N values is lim-
ited by both the variable composition of feces, limited separation
among potential prey taxa, and missing data from taxa that may con-
tribute significantly to dog diet (e.g., chickens and reptiles; Kshirsagar
et al., 2020). This is particularly true for the broad ranges of rat hair
stable isotope values, which will complicate any attempt to estimate
the proportion of rats in predator diet based solely on bulk tissue
C and δ
N values. Though introduced rats and dogs have diets
that are isotopically similar, dog fecal stable isotope signatures could
easily follow from combinations of tissue from introduced mice and
rats. Alternatively, existing data suggest that the majority of dog meals
could have also included tenrec tissue. The scarce identifiable animal
remains present in feces make it difficult to distinguish among these
possibilities, but complementary prior information regarding diet
(e.g., through behavioral or molecular studies and stable isotope
values from multiple consumer tissues) could improve the resolution
of dietary inference. The details of dog diet are an important area of
future research given that they determine whether the impacts of
dogs on particular endemic taxa are positive or negative.
4.2 |Implications
The introduction of dogs on Madagascar was associated with the
spread of grasslands and introduced grazers like zebu cattle
1000 years ago. However, many dogs that currently live in derived
grassland close to modern forests have little reliance on foods derived
from C
grasses. This increases the potential for negative interactions
between dogs and endemic prey and potential competitors. Dogs
could easily contribute to habitat fragmentation by harassing endemic
prey and aiding human-led hunts (Brockman et al., 2008; Garcia &
Goodman, 2003; Kshirsagar et al., 2020), but the extent to which dogs
themselves kill animals versus scavenge animal remains is poorly
known. Higher resolution study of dog diet can further help clarify the
extent to which dogs control populations of other introduced taxa
that also rely on the island's forests (e.g., rats).
As with past research involving rats in the study area, we observe
a mismatch between the fecal δ
C data and the sites where feces
were collected. Specifically, Dammhahn et al. (2017) previously used
soil and plant samples in the study region to note that food webs in
open habitat (anthropogenic steppe and agricultural fields) tend to be
enriched in
C relative to those in closed forests. However, we note
that dog fecal δ
C values are independent of the distance that
separates the feces from the nearest protected primary forest. Thus,
as with rats, we suggest that the evident mobility of dogs between
open and closed habitats increases their potential to share zoonotic
diseases and parasites with endemic taxa. In SW Madagascar, dogs
and other introduced animals are known to harbor ectoparasites with
bacterial pathogens such as Rickettsia spp. (Ehlers et al., 2020).
Future pathogen and stable isotope analyses may clarify whether
isotopic traces of mobility among introduced animal populations are
positively associated with disease transmission. Greater detail in our
understanding of dog feeding ecology will help wildlife management
teams assess when, where, and how it is best to control dog
populations in and around forests for the conservation of endemic
We are grateful to Madagascar National Parks and Association
Mitsinjo and VOI for permission to conduct research in this region
and to Heritiana Randrianatoandro for assistance with Malagasy
translation. Open Access funding enabled and organized by Projekt
DEAL. WOA Institution: Max-Planck-Gesellschaft Consortia Name :
Projekt DEAL.
We declare we have no conflicts of interest.
S.W.H., K.V., J.K., and Z.F. planned and designed the research. K.V.,
M.N., S.C., D.M., and C.F. conducted fieldwork, and S.W.H., K.V., and
S.F. conducted labwork. S.W.H. and K.V. wrote the manuscript, and all
coauthors commented on the manuscript.
All data are provided in Datasets S1S4, and all details regarding data
collection and analysis are provided in the manuscript and Method S1.
Sean W. Hixon
Ardalan, A., Oskarsson, M. C., Van Asch, B., Rabakonandriania, E., &
Savolainen, P. (2015). African origin for Madagascan dogs revealed by
mtDNA analysis. Royal Society Open Science,2(5), 140552. https://doi.
Austin, A. T., & Vitousek, P. (1998). Nutrient dynamics on a precipitation
gradient in Hawai'i. Oecologia,113(4), 519529.
Axelsson, E., Ratnakumar, A., Arendt, M.-L., Maqbool, K., Webster, M. T.,
Perloski, M., Liberg, O., Arnemo, J. M., Hedhammar, Å., & Lindblad-
Toh, K. (2013). The genomic signature of dog domestication reveals
adaptation to a starch-rich diet. Nature,495(7441), 360364. https://
Barrow, L. M., Bjorndal, K. A., & Reich, K. J. (2008). Effects of preservation
method on stable carbon and nitrogen isotope values. Physiological
and Biochemical Zoology,81(5), 688693.
Binggeli, P. (2003). Introduced and invasive plants. In S. M. Goodman &
J. P. Benstead (Eds.), The natural history of Madagascar (pp. 257268).
The University of Chicago Press.
Bond, W. J., Silander, J. A. Jr., Ranaivonasy, J., & Ratsirarson, J. (2008). The
antiquity of Madagascar's grasslands and the rise of C4 grassy biomes.
Journal of Biogeography,35(10), 17431758.
Brockman, D. K., Godfrey, L. R., Dollar, L. J., & Ratsirarson, J. (2008). Evi-
dence of invasive Felis silvestris predation on Propithecus verreauxi at
Beza Mahafaly Special Reserve, Madagascar. International Journal of
Primatology,29(1), 135152.
Burney, D. A. (1987). Late Holocene vegetational change in central
Madagascar. Quaternary Research,28(1), 130143.
Burney, D. A., Burney, L. P., Godfrey, L. R., Jungers, W. L., Goodman, S. M.,
Wright, H. T., & Jull, A. T. (2004). A chronology for late prehistoric
Madagascar. Journal of Human Evolution,47(1), 2563.
Burney, D. A., Hume, J. P., Randalana, R., Andrianaivoarivelo, R. A.,
Griffiths, O., Middleton, G. J., Rasolondrainy, T., & Radimilahy, C.
(2020). Rock art from Andriamamelo cave in the Beanka protected
area of western Madagascar. The Journal of Island and Coastal Archae-
Codron, D., Codron, J., Lee-Thorp, J. A., Sponheimer, M., de Ruiter, D., &
Brink, J. S. (2007). Stable isotope characterization of mammalian
predatorprey relationships in a South African savanna. European Jour-
nal of Wildlife Research,53(3), 161170.
Codron, D., Codron, J., Sponheimer, M., Lee-Thorp, J. A., Robinson, T.,
Grant, C., & De Ruiter, D. (2005). Assessing diet in savanna herbivores
using stable carbon isotope ratios of faeces. Koedoe,48(1), 115124.
Crowley, B. E. (2010). A refined chronology of prehistoric Madagascar and
the demise of the megafauna. Quaternary Science Reviews,29(19),
Crowley, B. E., Thorén, S., Rasoazanabary, E., Vogel, E. R., Barrett, M. A.,
Zohdy, S., Blanco, M. B., McGoogan, K. C., Arrigo-Nelson, S. J.,
Irwin, M. T., Wright, P. C., Radespiel, U., Godfrey, L. R., Koch, P. L., &
Dominy, N. J. (2011). Explaining geographical variation in the isotope
composition of mouse lemurs (Microcebus). Journal of Biogeography,
38(11), 21062121.
Dammhahn, M., Randriamoria, T. M., & Goodman, S. M. (2017). Broad and
flexible stable isotope niches in invasive non-native Rattus spp. in
anthropogenic and natural habitats of central eastern Madagascar.
BMC Ecology,17(1), 113.
Decary, R. (1939). La chasse et le piégeage chez les indigènes de
Madagascar. Journal Des Africanistes,9(1), 341.
Domic, A. I., Hixon, S. W., Velez, M. I., Ivory, S. J., Douglass, K. G.,
Brenner, M., Curtis, J. H., Culleton, B. J., & Kennett, D. J. (2021). Influ-
ence of late Holocene climate change, megafaunal extinction, and
human occupation on terrestrial and aquatic ecosystems in SW
Madagascar. Frontiers in Ecology and Evolution,9.
Ehlers, J., Krüger, A., Rakotondranary, S. J., Ratovonamana, R. Y.,
Poppert, S., Ganzhorn, J. U., & Tappe, D. (2020). Molecular detection
of Rickettsia spp., Borrelia spp., Bartonella spp. and Yersinia pestis in
ectoparasites of endemic and domestic animals in southwest
Madagascar. Acta Tropica,205, 105339.
Farris, Z., Gerber, B., Karpanty, S., Murphy, A., Andrianjakarivelo, V.,
Ratelolahy, F., & Kelly, M. (2015). When carnivores roam: Temporal
patterns and overlap among Madagascar's native and exotic carni-
vores. Journal of Zoology,296(1), 4557.
Feldt, T., Antsonantenainarivony, O., & Schlecht, E. (2017). Feed selection
on dry rangelands in southwestern Madagascar: Implications for rumi-
nant nutrition in view of ecological and social challenge. Journal of Arid
Garcia, G., & Goodman, S. M. (2003). Hunting of protected animals in
the Parc National d'Ankarafantsika, north-western Madagascar.
Oryx,37(1), 115118.
Gardner, C. J., & Davies, Z. G. (2014). Rural bushmeat consumption within
multiple-use protected areas: Qualitative evidence from southwest
Madagascar. Human Ecology,42(1), 2134.
Godfrey, L., & Crowley, B. (2016). Madagascar's ephemeral palaeo-grazer
guild: Who ate the ancient C4 grasses? Proceedings of the Royal Society
B: Biological Sciences,283, 20160360.
Goodman, S. M. (1995). Rattus on Madagascar and the dilemma of pro-
tecting the endemic rodent fauna. Conservation Biology,9(2), 450453.
Goodman, S. M. (2010). Biological research conducted in the general
Andasibe region of Madagascar with emphasis on enumerating the
local biotic diversity. Biodiversity, exploration, and conservation of the
natural habitats associated with the Ambatovy project. Malagasy
Hackel, J., Vorontsova, M. S., Nanjarisoa, O. P., Hall, R. C., Razanatsoa, J.,
Malakasi, P., & Besnard, G. (2018). Grass diversification in
Madagascar: In situ radiation of two large C3 shade clades and
support for a Miocene to Pliocene origin of C4 grassy biomes. Jour-
nal of Biogeography,45(4), 750761.
Hansford, J. P., Lister, A. M., Weston, E. M., & Turvey, S. T. (2021). Simul-
taneous extinction of Madagascar's megaherbivores correlates with
late Holocene human-caused landscape transformation. Quaternary
Science Reviews,263, 106996.
Harper, G. J., Steininger, M. K., Tucker, C. J., Juhn, D., & Hawkins, F.
(2007). Fifty years of deforestation and forest fragmentation in
Madagascar. Environmental Conservation,34(4), 325333. https://doi.
Hixon, S., Douglass, K. G., Godfrey, L. R., Eccles, L., Crowley, B. E.,
Rakotozafy, L., Clark, G., Haberle, S., Anderson, A., Wright, H. T., &
Kennett, D. J. (2021). Ecological consequences of a millennium of
introduced dogs on Madagascar. Frontiers in Ecology and Evolution,9,
Hixon, S. W., Douglass, K. G., Crowley, B. E., Rakotozafy, L., Clark, G.,
Anderson, A., Haberle, S., Ranaivoarisoa, J. F., Buckley, M.,
Fidiarisoa, S., Mbola, B., & Kennett, D. J. (2021). Late Holocene spread
of pastoralism coincides with endemic megafaunal extinction on
Madagascar. Proceedings of the Royal Society of London B,288,
Jones, R. J., Ludlow, M., Troughton, J., & Blunt, C. (1979). Estimation of
the proportion of C3 and C4 plant species in the diet of animals from
the ratio of natural 12C and 13C isotopes in the faeces. The Journal of
Agricultural Science,92(1), 91100.
Joseph, G. S., Rakotoarivelo, A. R., & Seymour, C. L. (2021). How expansive
were Malagasy Central Highland forests, ericoids, woodlands and
grasslands? A multidisciplinary approach to a conservation conundrum.
Biological Conservation,261, 109282.
Joseph, G. S., Rakotoarivelo, A. R., & Seymour, C. L. (2022). Tipping points
induced by palaeo-human impacts can explain presence of savannah in
Malagasy and global systems where forest is expected. Proceedings of
the Royal Society B,289(1971), 20212771.
Joseph, G. S., & Seymour, C. L. (2020). Madagascan highlands: Originally
woodland and forest containing endemic grasses, not grazing-adapted
grassland. Proceedings of the Royal Society B,287(1937), 20201956.
Kaufmann, J. C., & Tsirahamba, S. (2006). Forests and thorns: Conditions
of change affecting Mahafale pastoralists in southwestern
Madagascar. Conservation and Society,4, 231261.
Kolby, J. E. (2014). Stop Madagascar's toad invasion now. Nature,
509(7502), 563563.
Kshirsagar, A. R., Applebaum, J. W., Randriana, Z., Rajaonarivelo, T.,
Rafaliarison, R. R., Farris, Z. J., & Valenta, K. (2020). Human-dog rela-
tionships across communities surrounding Ranomafana and Andasibe-
Mantadia National Parks, Madagascar. Journal of Ethnobiology,40(4),
Kull, C. A., Tassin, J., Moreau, S., Ramiarantsoa, H. R., Blanc-Pamard, C., &
Carrière, S. M. (2012). The introduced flora of Madagascar. Biological
Invasions,14(4), 875888.
McCutchan, J. H., Lewis, W. M., Kendall, C., & McGrath, C. C. (2003). Vari-
ation in trophic shift for stable isotope ratios of carbon, nitrogen, and
sulfur. Oikos,102(2), 378390.
Merson, S., Dollar, L., Tan, C., & W Macdonald, D. (2019). Activity patterns
of sympatric living exotic and endemic carnivores (the fosa) in Western
Madagascar's deciduous forests. Journal of Zoology,307(3), 186194.
Messelink, G. J., van Maanen, R., van Steenpaal, S. E., & Janssen, A. (2008).
Biological control of thrips and whiteflies by a shared predator: Two
pests are better than one. Biological Control,44(3), 372379. https://
Newsome, S. D., Martinez del Rio, C., Bearhop, S., & Phillips, D. L. (2007).
A niche for isotopic ecology. Frontiers in Ecology and the Environment,
5(8), 429436.
Peacock, D., & Abbott, I. (2010). The mongoose in Australia: Failed intro-
duction of a biological control agent. Australian Journal of Zoology,
58(4), 205227.
Razanatsoa, E. (2019). Impact of human land-use and rainfall variability in
tropical dry forests of southwest Madagascar during the late
Salvarina, I., Yohannes, E., Siemers, B. M., & Koselj, K. (2013). Advantages
of using fecal samples for stable isotope analysis in bats: Evidence
from a triple isotopic experiment. Rapid Communications in Mass Spec-
trometry,27(17), 19451953.
Shiels, A. B. (2011). Frugivory by introduced black rats (Rattus rattus) pro-
motes dispersal of invasive plant seeds. Biological Invasions,13(3),
Simberloff, D., & Stiling, P. (1996). Risks of species introduced for biologi-
cal control. Biological Conservation,78(12), 185192.
Solofondranohatra, C. L., Vorontsova, M. S., Hackel, J., Besnard, G.,
Cable, S., Williams, J., Jeannoda, V., & Lehmann, C. E. R. (2018). Grass
functional traits differentiate forest and savanna in the Madagascar
central highlands. Frontiers in Ecology and Evolution,6, 184. https://doi.
Sommer, S., & Hommen, U. (2000). Modelling the effects of life-history
traits and changing ecological conditions on the population dynamics
and persistence of the endangered Malagasy giant jumping rat (Hypo-
geomys antimena). Animal Conservation Forum,3, 333343. https://doi.
Styger, E., Rakotondramasy, H. M., Pfeffer, M. J., Fernandes, E. C., &
Bates, D. M. (2007). Influence of slash-and-burn farming practices on
fallow succession and land degradation in the rainforest region of
Madagascar. Agriculture, Ecosystems & Environment,119(34), 257
Sueda, K. L. C., Hart, B. L., & Cliff, K. D. (2008). Characterisation of plant
eating in dogs. Applied Animal Behaviour Science,111(12), 120132.
Sweeting, C. J., Polunin, N. V., & Jennings, S. (2004). Tissue and fixative
dependent shifts of δ13C and δ15N in preserved ecological material.
Rapid Communications in Mass Spectrometry,18(21), 25872592.
Szpak, P. (2014). Complexities of nitrogen isotope biogeochemistry in
plant-soil systems: Implications for the study of ancient agricultural
and animal management practices. Frontiers in Plant Science,5, 288.
Tsutaya, T., Ogawa, N. O., Nomura, T., Shimizu, M., Ohkouchi, N., &
Kuze, N. (2021). Carbon and nitrogen stable isotopic offsets between
diet and hair/feces in captive orangutans. Primates,62(6), 945954.
Valenta, K., Gettinger-Larson, J. A., Chapman, C. A., & Farris, Z. J. (2016).
Barking up the right tree: Understanding local attitudes towards dogs
in villages surrounding Ranomafana National Park, Madagascar can
benefit applied conservation. Madagascar Conservation & Development,
11(2), 8790.
Veldman, J. W., & Putz, F. E. (2011). Grass-dominated vegetation, not
species-diverse natural savanna, replaces degraded tropical forests on
the southern edge of the Amazon Basin. Biological Conservation,
144(5), 14191429.
Virah-Sawmy, M., Gillson, L., Gardner, C. J., Anderson, A., Clark, G., &
Haberle, S. (2016). A landscape vulnerability framework for identifying
integrated conservation and adaptation pathways to climate change:
The case of Madagascar's spiny forest. Landscape Ecology,31(3), 637
Zhang, M., Huang, H., Li, Z., Hackman, K. O., Liu, C., Andriamiarisoa, R. L.,
Ny Aina Nomenjanahary Raherivelo, T., Li, Y., & Gong, P. (2020). Auto-
matic high-resolution land cover production in Madagascar using
sentinel-2 time series, tile-based image classification and Google Earth
Engine. Remote Sensing,12(21), 3663.
Additional supporting information can be found online in the Support-
ing Information section at the end of this article.
How to cite this article: Hixon, S. W., Neelin, M., Chan, S.,
Mayo, D., Filla, C., Farris, Z. J., deFrance, S., Krigbaum, J., &
Valenta, K. (2022). Dogs occupying grassy habitat near
protected areas in eastern Madagascar rely on foods from
forests. Plants, People, Planet,110.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Madagascar’s biota underwent substantial change following human colonization of the island in the Late Holocene. The timing of human arrival and its role in the extinction of megafauna have received considerable attention. However, the impacts of human activities on regional ecosystems remain poorly studied. Here, we focus on reconstructing changes in the composition of terrestrial and aquatic ecosystems to evaluate the impact of human land use and climate variability. We conducted a paleoenvironmental study, using a sediment record that spans the last ∼1,145 years, collected from a lakebed in the Namonte Basin of southwest Madagascar. We examined physical (X-ray fluorescence and stratigraphy) and biotic indicators (pollen, diatoms and micro- and macro-charcoal particles) to infer terrestrial and aquatic ecosystem change. The fossil pollen data indicate that composition of grasslands and dry deciduous forest in the region remained relatively stable during an arid event associated with northward displacement of the Intertropical Convergence Zone (ITCZ) between ∼1,145 and 555 calibrated calendar years before present (cal yr BP). Charcoal particles indicate that widespread fires occurred in the region, resulting from a combination of climate drivers and human agency during the entire span covered by the paleorecord. Following settlement by pastoral communities and the disappearance of endemic megafauna ∼1,000 cal yr BP, grasslands expanded and the abundance of trees that rely on large animals for seed dispersal gradually declined. A reduction in the abundance of pollen taxa characteristic of dry forest coincided with an abrupt increase in charcoal particles between ∼230 and 35 cal yr BP, when agro-pastoral communities immigrated into the region. Deforestation and soil erosion, indicated by a relatively rapid sedimentation rate and high K/Zr and Fe/Zr, intensified between 180 and 70 cal yr BP and caused a consequent increase in lake turbidity, resulting in more rapid turnover of the aquatic diatom community. Land use and ongoing climate change have continued to transform local terrestrial and freshwater ecosystems during the last ∼70 years. The current composition of terrestrial and aquatic ecosystems reflects the legacy of extinction of native biota, invasion of exotic species, and diminished use of traditional land management practices.
Full-text available
Recently expanded estimates for when humans arrived on Madagascar (up to approximately 10 000 years ago) highlight questions about the causes of the island's relatively late megafaunal extinctions (approximately 2000-500 years ago). Introduced domesticated animals could have contributed to extinctions, but the arrival times and past diets of exotic animals are poorly known. To conduct the first explicit test of the potential for competition between introduced livestock and extinct endemic megafauna in southern and western Madagascar, we generated new radiocarbon and stable carbon and nitrogen isotope data from the bone collagen of introduced ungulates (zebu cattle, ovicaprids and bushpigs, n = 66) and endemic megafauna (pygmy hippopotamuses, giant tortoises and elephant birds, n = 68), and combined these data with existing data from endemic megafauna (n = 282, including giant lemurs). Radiocarbon dates confirm that introduced and endemic herbivores briefly overlapped chronologically in this region between 1000 and 800 calibrated years before present (cal BP). Moreover, stable isotope data suggest that goats, tortoises and hippos had broadly similar diets or exploited similar habitats. These data support the potential for both direct and indirect forms of competition between introduced and endemic herbivores. We argue that competition with introduced herbivores, mediated by opportunistic hunting by humans and exacerbated by environmental change, contributed to the late extinction of endemic megafauna on Madagascar.
Full-text available
Introduced predators currently threaten endemic animals on Madagascar through predation, facilitation of human-led hunts, competition, and disease transmission, but the antiquity and past consequences of these introductions are poorly known. We use directly radiocarbon dated bones of introduced dogs (Canis familiaris) to test whether dogs could have aided human-led hunts of the island’s extinct megafauna. We compare carbon and nitrogen isotope data from the bone collagen of dogs and endemic “fosa” (Cryptoprocta spp.) in central and southwestern Madagascar to test for competition between introduced and endemic predators. The distinct isotopic niches of dogs and fosa suggest that any past antagonistic relationship between these predators did not follow from predation or competition for shared prey. Radiocarbon dates confirm that dogs have been present on Madagascar for over a millennium and suggest that they at least briefly co-occurred with the island’s extinct megafauna, which included giant lemurs, elephant birds, and pygmy hippopotamuses. Today, dogs share a mutualism with pastoralists who also occasionally hunt endemic vertebrates, and similar behavior is reflected in deposits at several Malagasy paleontological sites that contain dog and livestock bones along with butchered bones of extinct megafauna and extant lemurs. Dogs on Madagascar have had a wide range of diets during the past millennium, but relatively high stable carbon isotope values suggest few individuals relied primarily on forest bushmeat. Our newly generated data suggest that dogs were part of a suite of animal introductions beginning over a millennium ago that coincided with widespread landscape transformation and megafaunal extinction.
Full-text available
Reconstructing the dynamics and drivers of late Quaternary megafaunal extinctions requires direct radiometric date series that are assessed within probabilistic statistical frameworks. Extinction chronologies are poorly understood for many tropical regions, including Madagascar, which had a diverse, now-extinct Holocene large vertebrate fauna including a “megaherbivore” guild of endemic hippopotami and elephant birds. Madagascar's megaherbivores likely played vital roles in regulating ecosystem structure and nutrient cycling, but few direct dates are available for megaherbivore specimens identified to species level, with uncertainty over when and why different representatives of this guild disappeared. Here, we conduct a new investigation into Malagasy megaherbivore extinction dynamics, including 30 new AMS dates and 63 audited published dates. We use Gaussian-resampled inverse-weighted McInerny (GRIWM) analysis to estimate species-specific extinction dates for three elephant bird species (Aepyornis hildebrandti, Mullerornis modestus, Vorombe titan), eggshell representing Aepyornis or Vorombe, and two hippo species (Hippopotamus lemerlei, H. madagascariensis), and to estimate extinction dates for megaherbivore communities in different biomes. Megaherbivores persisted for millennia after first human arrival. Extinction date estimates vary significantly between biomes, with disappearance from dry deciduous forest over a millennium earlier than other biomes, possibly reflecting local variation in megaherbivore population densities or human pressures. However, megaherbivore communities including all elephant bird and hippo species persisted elsewhere across Madagascar until ∼1200-900 bp, when they collapsed suddenly. Extinctions are closely correlated in time with intensive conversion of forests to grassland at ∼1100-1000 bp, probably associated with a shift to agro-pastoralism and representing a radical change in sustainability of prehistoric human interactions with biodiversity.
Full-text available
Domestic dogs (Canis lupus familiaris) are estimated to be one of the most globally abundant invasive carnivores that threaten wildlife. Madagascar is home to large populations of free-roaming dogs and is a highly diverse and anthropogenically threatened environment, making it one of the world's top conservation priorities. Comparatively little is known about human-dog relationships in developing countries such as Madagascar. We surveyed non-dog owners and dog owners visiting free mobile veterinary clinics in their communities around Ranomafana National Park (RNP) and Andasibe-Mantadia National Park (AMNP) to understand human-dog relationships, gain insight on free-roaming dog behavior, and to assess the feasibility of humane population control measures. Amongst dog owners, the vast majority of respondents reported owning their dog for protection and a significant number had dogs for companionship. Our results indicate that free-roaming (owned, unconfined) dogs may be an underappreciated threat to endemic wildlife in the National Parks of Madagascar, as nearly half of dog owners reported that their dog killed at least one wild animal a month. Most dog owners in surveyed communities approve of spay/neuter/vaccine programs and state that they would use them if freely available, indicating that veterinary intervention can be an important tool in humanely controlling free-roaming dog populations in these regions.
Full-text available
Madagascar, one of Earth’s biodiversity hotpots, is characterized by heterogeneous landscapes and huge land cover change. To date, fine, reliable and timely land cover information is scarce in Madagascar. However, mapping high-resolution land cover map in the tropics has been challenging due to limitations associated with heterogeneous landscapes, the volume of satellite data used, and the design of methodology. In this study, we proposed an automatic approach in which the tile-based model was used on each tile (defining an extent of 1° × 1° as a tile) for mapping land cover in Madagascar. We combined spectral-temporal, textural and topographical features derived from all available Sentinel-2 observations (i.e., 11,083 images) on Google Earth Engine (GEE). We generated a 10-m land cover map for Madagascar, with an overall accuracy of 89.2% based on independent validation samples obtained from a field survey and visual interpretation of very high-resolution (0.5–5 m) images. Compared with the conventional approach (i.e., the overall model used in the entire study area), our method enables reduce the misclassifications between several land cover types, including impervious land, grassland and wetland. The proposed approach demonstrates a great potential for mapping land cover in other tropical or subtropical regions.
Models aimed at understanding C 4 -savannah distribution for Australia, Africa and South America support transition to forest at high mean annual precipitation (MAP), and savannah grasslands of Madagascar have recently been reported to be similarly limited. Yet, when savannah/grassland presence data are plotted against MAP for the various ecosystems across the Malagasy Central Highlands, the relationship does not hold. Furthermore, it does not always hold in other sites on other continents. Instead, in high-rainfall savannahs, palaeo-human impacts appear to have selected a fire-adapted habitat, creating tipping points that allow savannah persistence despite high rainfall, suppressing forest return. We conducted the largest systematic literature review to date for global evidence of palaeo-human impacts in savannahs, and conclude that impacts are widespread and should be incorporated into models aimed at understanding savannah persistence at elevated precipitation, particularly as more palaeodata emerges. Building on existing studies, we refine the MAP savannah relationship at higher MAP. Palaeoanthropogenic impact can help explain inconsistencies in the savannah/forest boundary at higher MAP, and points to a key role for palaeoecology in understanding systems. Including these effects presents a crucial change to our understanding of factors determining global savannah distribution, supporting a human hand in much of their formation.
Estimating stable isotopic offset values is crucial for dietary reconstructions. Although research into stable isotope ecology of wild nonhuman primates is increasing overall, only a minority of studies involve laboratory experiments. This study is the first to report the carbon and nitrogen stable isotopic offset values in hair and feces of orangutans. During an experiment lasting 1 week, the weight of each consumed food item was recorded for each of six captive Bornean orangutan (Pongo pygmaeus) individuals. The food, hair, and fecal samples were collected for a few days, and their stable carbon and nitrogen isotope ratios were measured using an elemental analyzer/isotope ratio mass spectrometer. Subsamples of feces were treated with ethanol during the preservation process. Monte Carlo analyses showed that the 95% confidence intervals (CIs) of the carbon and nitrogen offset values between hair and diet were +0.9‰ to +3.9‰ and +2.3‰ to +4.5‰, respectively. The 95% CIs of the carbon and nitrogen offset values between feces and diet were −3.7‰ to −0.9‰ and +0.3‰ to +2.7‰, respectively. The effect of ethanol treatment on the stable isotope ratios of feces was unclear and inconclusive. The computed offset values of hair in captive orangutans are similar to those reported in other nonhuman primates, although those of feces showed greater interspecies variations. The offset values estimated in this study contribute to isotopic studies into the feeding ecology of free-ranging orangutans who are critically endangered in most wild settings.
Madagascar is the world's highest profile conservation priority. Yet the natural habitat of its largest biome, the 200,000 km² Malagasy Central Highlands (MCH) remains one of the most prominent and longstanding debates in conservation: are MCH grasslands natural and ancient, coevolved with endemic grazers; or new and anthropogenic, representing degraded forest and woodland? Although patterns and processes that delimit ecosystem boundaries are central to conservation science and planning, even prominent conservation organisations take polar views on the MCH natural habitat, underscoring the urgent need for clarity. Misconceptions impact conservation of critically endangered forest and woodland fauna and flora, rare endemic grasses poorly adapted to fire and pastoralism, and human and ecosystem resilience. In this review, the first of its kind for the MCH, we determine natural habitat by assessing literature spanning ecology, evolutionary biology, palaeoanthropology, palaeoecology, molecular studies, zoology and local knowledge. By separating pre human-settlement habitat spanning 150 ka, from post-settlement impacts encompassing 2 ka, evidence is strong and compelling. Pre-settlement habitat mosaics of forest-woodland-ericoid-grassland have vacillated with climate. Post-settlement, multiple lines of evidence affirm conversion of woodland, forest and ericoid habitat to grassland c. 1 ka, with anthropogenic spread of dominant fire- and ungulate grazing-adapted grass species. Grasses were likely present in woodland understories, associated with ericoid systems, or occurred in far smaller patches, dominated by species adapted to disturbances different to those found today. Refining understanding of this nuanced habitat-mosaic provides a platform for facilitating habitat-appropriate patterns and process that enhance conservation of rare animals, grasses, forest and ericoids.
Long considered a consequence of anthropogenic agropastoralism, the origin of Madagascar’s central highland grassland is hotly disputed. Arguments that ancient endemic grasses formed grassland maintained by extinct grazers and fire have been persuasive. Consequent calls to repeal fire-suppression legislation, burn protected areas, and accept pastoralism as the ‘salvation’ of endemic grasses mount, even as the International Union for Conservation of Nature (IUCN) declares 98% of lemurs face extinction through fire-driven deforestation. By analysing grass data from contemporary studies, and assessing endemic vertebrate habitat and feeding guilds, we find that although the grassland potentially dates from the Miocene, it is inhospitable to endemic vertebrates and lacks obligate grazers. Endemic grasses are absent from dominant grassland assemblages, yet not from woodland and forest assemblages. There is compelling evidence that humans entered a highland dominated by woodland and forest, and burned it; by 1000 current era (CE), grass pollens eclipsed tree pollens, reminiscent of prevailing fire-induced transformation of African miombo woodland to grassland. Endemic grasses are survivors from vanished woody habitats where grassy patches were likely small and ephemeral, precluding adaptive radiation by endemic vertebrates to form grazing guilds. Today forests, relic tapia woodland, and outcompeted endemic grasses progressively retreat in a burning grassland dominated by non-endemic, grazing-adapted grasses and cattle. 1.