CURREN T SCIENC E, VOL. 106, NO. 6, 25 MARCH 2014 874
*For correspondence. (e-mail: abhadra@iiserko l.ac.in)
A dog’s day with humans – time
activity budget of free-ranging dogs
Sreejani Sen Majumder1, Ankita Chatterjee1,2
and Anindita Bhadra1,*
1Behaviour and Ecolo gy Lab, Depar tment of Biological Science s,
Indian Institute of Science Education and Research Kolkata,
Mohanpur, Nadia 741 25 2, I ndia
2Present addres s: National Institut e of Biomed ical Ge nomics,
Kalyani 741 251, Ind ia
Free-ranging dogs, Canis lupus familiaris, are an inte-
gral part of the human environment in India and
many other countries. They can serve as the perfect
model system f or unde rstanding t he process of deve -
lopment of t he human–dog relationship that le d to the
domestication of t he wilder ancestors of the dogs a nd
created ‘man’s best frie nd’. Yet, very little is known
about t he e coethology of these animals and all o ur un-
derstanding of dog behaviour is based on studies of
pets reare d by humans. The free-ranging dogs lead a
scaveng ing life, depending on human exce sses for their
survival, and rarely hunt. They are often considere d
as a menace by many people, as dirty animals t hat
bark, bite and s pread rabies. These notions are ofte n
founded on personal biases and little scientific data
exist to eithe r support or ref ute such claims. As part
of an extended study on the behavioural ecology of
free-ranging dogs in India, we ca rried out random
sampling of dog behavio ur through censuses in two
cities and o ne towns hip of India. We use d o ur data
from 1941 sig htings to draw up a time activity budget
of dogs during the part of the day whe n they share t he
streets with humans. Our analysis reveals that dogs
are ge nerally lazy and friendly animals, and their rare
interactions with humans a re typically s ubmissive.
Thus dogs do not usually pose a threat to human well-
being, and proper manageme nt of our refuse and a
tolerant, if not friendly att itude towards dogs can
ensure t heir peaceful co-existence with us.
Keywords: Census, free-ranging dogs, time activi ty
THE dog, Canis lupus familiaris, is known as man’s bes t
friend, and yet, scientific knowledge on the ecoethology
of dogs in their natural habitat is al most non-existent. A
reason for this is pr obably that the presence of unattended
dogs on the streets is forbi dden by law in most wester n-
ized countries, and so, even if such dogs are present, their
activities are interfered by humans, and stabl e social
groups are not formed1. Due to their l ong his tory of do-
mestication dogs have adapted excellently to living with
humans in their homes, isolated from conspecifics. How-
ever, dogs have descended from wolves (Canis lupus
lupus), and like many other canids, they too are capable
of forming stable social groups that are influenced b y the
same factors that affect social organization of wild canid
systems1–5. Dogs that do not have owners and w hose
movements are not li mited by human beings are typically
called free-ranging dogs. These dogs can be interesting
model systems for studying the effects o f d omestication
on their be haviour, a s well as for understanding the evo-
lution of the dog–human relationship in nature.
Free-ranging dogs are a ubiquitous part of the urban
ecology in many developing and under developed coun-
tries like Mexico 6,7, Ecuador8, Zambia9, Zimbabwe1 0,
Italy1 1, India 12, Nepal and Japan1 3. Though dogs in India
have lived outside of human homes for centuries 14, and
have also been used for hunting, they have not undergone
the usual domestication process to become exclusively
pets as i n most developed countries. Dog figurines a nd
remains have been unear thed i n the Indus Valley Civili-
zation1 5 a nd references to dogs can be found in ancient
Indian texts like the Rg Veda, the Puranas, the Maha-
bharata, the Ramayana and the Manu Samhita and i n
many folk tales from across the country. The Agni
Purana classi fies the dog as a village animal, and though
dogs have been considered as outcastes and have been
associated wi th death and evil i n the Hindu culture, the
householder’s daily duty i ncluded feeding the dogs and
outcastes16. The European i nflue nce has introduced pedi-
greed dogs to the homes of the middle class and elite
society, but the India n Native dog (IN dog) or Indian
pariah dog has co ntinued to live on the s treets, depe nding
on garba ge and begged food for sus tenance 17.
The free-ranging dogs in India have a wide distributi on
ranging from ci ties to forest fringes3,17,18. Typicall y the y
have mongrel characteris tics, with poi nted ears, very
short fur, wolf-like poi nted faces and patch baldness in
their coats. They live in small groups or singly o n s treets
and depend on garbage and human generosity for their
sustenance17. Competition for food is high and fights are
common at garbage dumps, near roadside food stalls, or
when humans occ asio nally offer food to the dogs. Such
fights ar e sometimes a source of irritatio n for people, a nd
this makes dogs unpopular among many humans. They
breed twice a year, once i n autumn and once i n spri ng,
but a given female usually produces one li tter per year
(qualitative observations). Mortality in early life is quite
high, with less than 50% of the pups surviving beyond
the juvenile stage (Paul et al., in preparation). Though
humans are generally tolerant of dogs, dog–human c on-
flict is not uncommon, and a part of the human pop ula-
tion in India is regularly affected by dog bites.
Rabies is a serio us proble m i n India, with an estimated
2 in 1 00,000 people being affected ever y year19. Since
1985, 25,000–30,000 deaths have been reported due to
rabies i n the co untry2 0. In a multi-centric study based in
six anti-rabies clinics, Ichhp ujani et al.21 reported 1248
CURREN T SCIENC E, VOL. 106, NO. 6, 25 MARCH 2014 875
fresh dog bites over a peri od of 18 months. The aggres-
siveness of dogs and their propensity to attack a nd bi te is
often put forth as a justifica tion for culli ng the dog popu-
lation i n cities. Though most repor ted animal bites are by
dogs (91.5%), only about 60% o f these is by free-rangi ng/
stray dogs, while the remaining 40% is by pets2 2. Thus
there is indeed some amount of dog–human conflict on
the streets, but these studies only report the human per-
spective of such conflic t. No studies exist on either the
conflict or c ooperation that dogs receive from humans.
Efficie nt management of a population requires an under-
standing of the behaviour and ec ology of the species, and
in order to mitiga te dog–human conflict in our environ-
ment, scientific understanding of the behaviour o f free-
ranging dogs is necess ary. As dog–human interactions are
maximal during the daylight hours, and so are the inci-
dences of dog-bite23,2 4, we conducted a survey of free-
ranging dogs to draw up their time activity budget duri ng
the human activi ty hours on the streets of India.
We sampled dogs i n three different locations – the
IISER-Kolkata campus at Mohanpur (2294N, 8853E),
West Bengal; the Indian Institute of Science campus at
Bangalore (1298N, 7758E), Karnataka, and the town-
ship of Kalyani (2258N, 8828E), West Bengal. The
three locations were regarded as ‘urban’ co nsideri ng
the definition of urban and r ural India accordi ng to the
Census of India 2001 (ref. 25) and the Natio nal Sample
Survey Organization26 .
Sampling was carried out in the morning (0630–
1030 h), afternoon (1400–1630 h) and evening (1630–
1930 h), when both hu mans and dogs are typically see n
on the streets. We avoided the time between 1030 and
1400 h, as the dogs usually rest in shelters at this time,
avoiding the heat, a nd hence are difficult to find on the
streets ( quali tative observa tions). Though we s ampled
along streets which were mostly lit in the eve ning,
the do gs were often sighted at s pots off the s treets, w here
the li ghting conditions did not allow for accurate obser-
vations. Hence we avoid ed sampling beyond 1930 h.
The observer rando mly picked a road i n the pre-
defined area and started wal king alo ng the same, covering
all byla nes along the road. Whe never a dog was si ghted,
its se x (deter mined by looking at the genitalia), a ge class
(adult or juvenile, determined by the structure of the
genitalia), a nd behaviour a t the time of sighting were
noted. For each dog, only the behaviour seen at the i n-
stance of sighting was recorded. For example, if a dog
was observ ed to be scratching itself and then sniffing
grass, scratching was recorded as the observed behaviour.
Thus we obtained data equivalent to instantaneous scan
sampling of the population. For eac h pre- defined area, a
sampling bout lasted for 2–3 h, and all roads in the area
were covered on foot. The da ta were collected between
August 2008 a nd August 2011, in five phases – one sa m-
pling event each in Kalya ni and IISc, and three sampling
events on the IISER-Kolkata Campus. Thus we obtained
a ra ndom sample spread over di fferent seasons a nd areas,
such that it would be re presentative of the population.
The data were sorted according to behaviours, and the n
the behaviours were sorted into vario us cate gories li ke
inactive, mainte nance, vocalizatio ns, interactions, i ndi-
vidual behaviours and others. We kept vocalizations as a
separate category and did not put these under interactions
because for ev ery vocalization r ecorded, we did not know
the context in which it was produced. While voc alizations
are typicall y used for interactions, we did not always
know who these were directed at, or why. In addition, not
every vocali zation needs to be an interaction. Interactio ns
could be with do gs, humans or other animals like cows
and cats. Dogs were seen to be walking b oth solitarily as
well as with other dogs. However, if we did not see any
direct physical interactions, we co nsidered walking to be
an individual activity, as our sampling methodology did
not allow us to discern if the dog happe ned to be present
with other dogs by chance.
The various behaviours that were recorded and catego-
rized under these headings are provided in Table 1. The
data thus sorted were subjected to statistical analysis
using STATISTICA 7.0 and StatistiXL 1.8.
A total of 1941 free-ranging dog si ghtings were re-
corded a nd used in this analysis. For 1308 dogs we could
record the age class and se x, whereas for the r est data
were not available, though the behaviour was recorded.
We used the entire data to draw up the time ac tivity
budget of the free-ranging dogs, a nd the subset of 1308
dogs for more detailed analysis.
We c ompared the five sampling events over the three
locatio ns for all the five behavioural ca tegories (inactive,
active, vocalizations, maintenance and interac tions) to
check if there were significant variatio ns between sa m-
plings and l ocations. There was no significa nt difference
between the five samples (ANOVA: F4,20 = 1.13 4,
P = 0.369), and hence we could conclude that the overall
behavioural profiles of the dogs i n all our samples were
similar. Thus for all further analysis we pooled the data
from all the five samples under these behavioural catego-
The dogs were found in a s tate of rest or inactivity in
52.7% of the si ghtings, which was si gnificantly higher
than the cases in which they were fo und i n various states
of activity ( Figure 1 a;
2 = 5.46, df = 1, P <
0.019). When the do gs were active, they were sighted
most often as walking, ei ther i ndividuall y or with other
dogs. The d ogs spent 15.66% of their time wal king,
which contrib uted to 4 7.7% of the i ndividual activi ties.
Individual activities were divided into the three sub-
categories of walking, maintenance and other activities,
and the dogs did not show these various behaviours in
equal proportions (
2 = 60.49, df = 2, P <
0.0001). The amount of time spent walking was si gnifi-
cantly higher than that spent in maintenance activities
(24.8%) (Figure 1 b;
2 = 23.66, df = 1,
CURREN T SCIENC E, VOL. 106, NO. 6, 25 MARCH 2014 876
Table 1 . De tails o f catego rizat ion of b ehaviour s used for a nalyse s with the various behaviours included in eac h catego ry
Behavioural category Behavioural subcate gory Beha viours include d
Inactive Sleep, laze, sit
Maintenance Groom, scratc h, defe cate, urinate, drink, eat, e at grass, c hew object, fo od search, forage,
sniff garba ge, beg, follow, receive food
Vocaliza tions Bark, growl, howl, angry ba rk
Dog–dog intera ctions Aggressive Attack, c hase, fight, subm it, bite
Affiliat ive Mock bite, play, allogroo m, sniff d og
Indirect Mark, a ngry bark
Dog–huma n inte ractions Affiliat ive Submit, beg, follow, wag tail, receive food
Individual be haviours Stand, alert, watch, run, walk, jump, inspect ob ject, sniff
Figure 1. a, The time activity b udget o f dogs dur ing the hours of hu-
man ac tivity, between 0630 and 1930 h, calculated from 1941 dog
sightings in three lo catio ns, over 5 d iffer ent p hases of obser vatio n
spanning 3 years. b, The proportion o f time of the total ac tivity period
spent in three differe nt kinds of individua l level activities. The free-
ranging dogs spend most o f their active time (47.7%) in wa lking. Dif-
ferent alphabets signify statistic ally signi fica nt differences between the
values d enoted b y t he bars.
P < 0.0001) , as well as the time spent in other activi ties
like standing, watching, sniffing, etc. considered toge ther
(Figure 1 b;
2 = 18.01, df = 1, P < 0.0001). Thus
it can be c oncluded that walking was the most co mmon
individual activity displayed by the dogs. There was no
significant differe nce between the propor tion of time
spent in maintenance behaviours like grooming, scra tch-
ing, foraging, etc. and the pooled be haviours i n the
‘others’ category (Fi gure 1 b;
2 = 0.39, df = 1,
P = 0.535).
We po oled all ki nds of vocalizations like bar k, growl,
howl and angr y bark under the category of vocalizations,
which comprised only 3.3% of the activities of the dogs.
Interesti ngly, all interactions recorded, including those
with dogs and humans claimed only 10.9% of the total
23% of the ac tive time of the dogs. Co nsidering these two
categories together, the dogs spent only about 14% of
their total time in any kind of active i nterac tions with
each other or with humans, whether through actual physi-
cal interactions or thro ugh vocali zations. This was si gni-
ficantly lower than the total time spe nt in other
behaviours when the do gs were not sitting idle or res ting
2 = 144.9764, df = 1, P < 0.0001) . Of the i nter-
actions recorded, 84.7% were with other dogs, which was
significantly higher than the prop ortion of interactions
seen with humans (
2 = 101.505, df = 1, P <
0.0001), and onl y two cas es of chasing a calf were recor-
ded (Fi gure 2 a). Of the 32 interactio ns seen with humans
(0.13% of all interactions), none was a ggressive, and 16
were in fact submissive interactions like tail wagging,
submitting and begging for food.
We categorized all instances of interactions between
dogs into aggressi ve, affilia tive or indirect interactions.
Attack, chase, fight, submit and bite were listed under
aggressive interactions; mock bi te, play, allogroom and
sniff dog were categorized as affiliative interactions;
mark and angr y bark (very loud bar k wi th a n alert body
posture) were included in the category of indirect interac-
tions. We also noted that dogs produce three other kinds
of vocali zations – bark, growl and howl, which we did
not use in the cate gory of interactions as dogs can pro-
duce these sounds without havi ng other dogs in the vici n-
ity (for example, when they are in pain), a nd we had no
records of the conte xt i n which the vocalizations had
been recorded. The three kinds of interactions did not
occur i n eq ual freq uencies (Fi gure 2 b;
110.029, df = 2, P < 0.0001), with affiliative i nteractions
comprisi ng 65.21% of all interactions, which was si gnifi-
cantly higher than the two other categories of interactions
taken together (
2 = 10.12, df = 1, P = 0.001).
For a s ubset of the da ta (1308 do gs), we had complete
records of the sex and age classes. For the rest, either the
CURREN T SCIENC E, VOL. 106, NO. 6, 25 MARCH 2014 877
Figure 2. a, The distr ibution of interactions with dogs, humans and ot her animals, calcula ted o ut of the total number of interac tions of any kind
shown by the dogs in the 1941 sig htings. Most interactions recorded were bet ween d ogs. b, The distribution of inte ractions betwee n do gs into the
categor ies of affiliative, aggressive a nd indirect. Most interactio ns record ed were affiliative in nature. Different a lphabet s denote statistically sig-
nificant diffe rences between the values de noted by the bars.
sex could not be recorded as the genitalia were not visible
during the sampling, or the age class could not be a ccu-
rately de termined. There were a total of 711 females, of
which 168 were juve niles and 597 males, of which 86
were juveniles. The behavioural profiles considering the
proportion of all behaviours of the four kinds of dogs in
the population, namely adult females, adult males, juve-
nile females and juvenile males did not vary from eac h
other (Kruskal–Wallis test:
2 = 0.199, df = 3, p = 0.978).
When the dogs were ca tegori zed according to either age
or sex, we did not see any significant di fferences between
the sexes (Mann–Whi tney U test, U = 850, df = 41,41,
P = 0.934) or between the age classes (Mann–Whitney
U test, U = 892, df = 41,41, P = 0.638). We carried out
similar a nalysis for each category of be haviours sepa-
rately, and found significant differences of interaction
patterns between the sexes a nd between the a ge class es.
Males were more aggressive than females ( Fisher’s exact
test, P = 0.005), and adults were more aggressive than
juveniles (Fis her’s exact test, P = 0.001).
The free-ranging dogs i n India coexist with humans i n
every possible habi tat, and yet, they are often considered
as a menace by many people because of their scave nging
habit, the territorial fights that often e nsue between dog
groups a nd beca use of occasional dog–human conflict
that leads to people being attac ked and bitten by dogs.
Though there is no dearth of dog lovers in the country,
dogs are faced by the challenge of interacting with per-
haps a larger number of people who are intolera nt of
them, and consider the m to be aggressive, unfriendly
animals that should be removed from the streets1 9. This
aversion towards dogs is a socio-cultural pheno menon
that has very deep roots, going b ack to at least three thou-
sand years1 6. Our sa mpling study in two urban habitats
and one se mi-urban habitat in India covered an area o f
approximately 768.5 acres a nd spanned over different
seasons. In five phases, we recorded 1941 dog sightings
during the time of the da y when humans are usually active,
which included both adults and j uveniles of both sexes,
and was thus representative o f the population at large.
Our analysis reveale d that the d ogs are inactive for
over half of the day, either slee ping, lazing or just sitting.
Considering the fact that we sa mpled only during the time
when dogs could ac tually be see n on the streets, and were
not hiding i n shelters, this is actually an underestimate. If
our sa mpling had spanned the entir e day and included the
time that do gs spend resting in their hideouts, the propor-
tion of time spent i nactive would have been higher. These
results match the obs ervations on free-ranging dogs in
Berkeley, Cali fornia, USA, in which r epeated samplings
were carried out in a 48 ha residential area for 7 months27.
In this study, 1243 si ghtings were made on about 50
unique free-ranging dogs, w hich were found to be r esting
in 44. 4% of the si ghtings. This s tudy also reported that
free-ranging dogs were most abundant i n the early morn-
ings and la te afternoo ns, with the percentage of dogs
found to be resting increasi ng with temperature, for an
observed temperature range of 9–29C. Though we did
not record the tempera ture during our sampling, the aver-
age temperature range during our observations was
8–36C, considering all the time periods and the three
locatio ns covered, wi th the mean tempera ture ranging
from 18C to 30C (www.wunder ground.com/history).
Given that we did not sample ver y earl y in the morning
and i n the middle o f the day, the actual temperature
would have been higher than the minimum and lower
than the maximum, and hence closer to the mean range.
When the dogs were not resting, they were most often
seen to be wal king. Since our sampling was base d on
random sightings, we di d not have any method for re-
cording the p urpose of this walking. Dogs typically wal k
in search o f food, a nd also for marking their territories.
CURREN T SCIENC E, VOL. 106, NO. 6, 25 MARCH 2014 878
Often they seem to be wal king randomly, the purpose of
which ca n be r evealed only through detailed behavioural
observations on focal individuals and groups. Interaction
rates were found to be quite low, and all recorded instances
of i nteraction with humans were submissive. Thus, this
analysis does not support the general notion of free-
ranging dogs being aggressive, unfriendly animals that
are a constant source of nuisance to people on the streets
Dogs bark and howl, often producing a chor us reminis-
cent of their wolvine ancestry, and this makes them score
low with many humans. Many encounters between do gs
are often interrupted by people who chase them away,
often by throwi ng stones or dousing them with wa ter.
However, dogs were sighted produci ng some sound in
only 65 cases, which was 3.34% of the total observations.
Thus, the perception of dogs as noisy and aggressive
creatures that prese nt a threat to human well-being is
quite ill-founded and biased. However, i t is true that
many dogs in the Indian streets are rabid, and do g bites
do occur, though these are no t regular incidents as per-
ceived by some28. Dogs are efficient scavengers, and are
responsible for removal of a large part of our garbage
from the streets (Anandar up Bhadra, unpublished data).
Though we need detailed ob servational data for a better
understanding of the behavioural ecolo gy o f the free-
ranging dogs, this preliminary s tudy sugges ts that the
general perception of these dogs as a nuisance is quite
flawed. We would like to argue that the sol ution to do g–
human conflict is not c ulling, but efficient manage ment
of garbage and rabies in the country, and a positive attitude
towards the animals that are otherwise know n to be man’s
1. Ca fazzo, S., Valsecc hi, P. , Bonanni, R. a nd Na toli, E., Domina nce
in re lation to age, sex, and competitive co ntexts in a group of free-
ranging do mestic dogs . Behav. Eco l., 2010 , 21(3 ), 443–45 5;
2. Bona nni, R., Cafa zzo, S., Valsecchi, P. and Natoli, E., Effect of
affiliative and ago nistic relations hips on leade rship behaviour
in free- ranging do gs. An im. Behav., 2010, 79(5), 98 1–991 ;
doi:10.1 016/j.anbeha v.2010.0 2.021.
3. Pa l, S. K., Ghosh, B. and Roy, S., Dispersal behaviour of free-
ranging do gs (C anis f ami liaris) in relat ion to a ge, sex, seaso n a nd
dispersal d istance. App l. Anim. Behav. Sci., 1998, 61(2), 123–132;
doi:10.1 016/S0168-1591 (98)0 0185-3.
4. Pa l, S., Urine marking by free- ranging dogs (Ca nis familiaris ) in
relation to sex, s eason, place and posture. App l. Anim. Behav. Sci.,
2003, 80 (1), 45– 59.
5. Car r, G. and Macdonald, D., The soc iality of so litary foragers: a
model based on reso urce dispers ion. Anim. Behav ., 1986, 34(5),
6. Ortega-Pac heco, A., Rodriguez-Buenfil, J. C., Bolio- Gonzalez, M.
E., Sa uri-Arceo, C. H., Jiménez-Coe llo, M . a nd For sberg, C. L., A
survey of do g pop ulations in urban and rural areas of Yucatan,
Mexico. A nthrozoös, 2007, 20 (3), 261–274.
7. Daniels, T. J. and Bekoff, M., Popula tion and social biology of
free-ra nging dogs, Canis familiaris. J. Mammal., 1989, 70(4) ,
8. Kr uuk, H. and Snell, H., P rey selectio n by feral dogs from a pop u-
lation of marine iguanas (Amblyrhynchus cristatus) . J. A ppl. E col.,
1981, 18 (1), 197 –204.
9. De, B. K., A dog eco logy study in an urba n a nd a s emi-rur al area
of Zambia. Onde rstepoo rt J. Vet. Res., 19 93, 60 (4), 437–443.
10. But ler, J., Toit, J. Du. and Bingham, J., Free-ranging do mestic
dogs (Can is familiaris) as p redator s and prey in rural Zimbabwe :
threats o f co mpetition and disease to lar ge wild carnivores. Biol.
Conserv., 2004, 115(3), 369– 378.
11. Boita ni, L., Wo lf and dog c ompetition in Italy. Act a Zool. Fenn.,
1983, 17 4, 259– 264.
12. Pa l, S., Population ecology of free-ranging urba n dogs in West
Bengal, India. Acta Therio l., 2001, 46(1), 69–7 8.
13. Kato , M. and Yamamo to, H., S urvey of the stray dog p opulat ion
and the health educa tion program on the prevention of do g bites
and dog-acquired infections: a comparative stud y in Nepa l and
Okayama Prefecture, J apan. Acta Med. Okayama, 2003, 57(5),
14. Bol lée, W., Gone to the Dogs in Ancient India, Issue 2 , Verlag der
Bayerischen Akademie d er Wissensc haften, 2 006.
15. Osada , T. and Uesugi, A. (eds), Occasiona l Paper 10: Ling uistics ,
Archaeology and the Human Past, Nakanishi Pr inting C o Ltd ,
Kyoto, Japan, 2011.
16. Debro y, B., Sarama and her Children: T he Dog in Indian Myth,
Penguin Books India, 2008.
17. Vanak , A. T. and Gompper, M. E., Dietary niche separation bet-
ween sympat ric free-ranging do mestic dogs and Indian foxes in
Central India. J. Mammal., 2 009, 90 (5), 1058–1065.
18. Vanak , A. T., Thak er, M. and Go mpper, M. E., Exper imental exa-
mination of behavioural inte ractions between free-rang ing wild
and domestic canids. Behav. Ecol. Sociobi ol., 2009, 64(2) , 279–
287; doi:10.1007/s002 65-009-0 845- z.
19. Suda rshan, M. K. et al., Assessing the b urden of human rabies in
India: res ults of a nat ional multi- center epidemiologica l survey.
Int. J. Infect. Dis., 2007, 11(1), 29–35.
20. Suda rshan, M. K., Assessing b urden o f rab ies in India. WHO
sponsored national multi-ce ntric rabies survey. Association for
Prevention and Control of Rabies in Ind ia, 2004, pp. 44–45.
21. Ichhpujani, R. L. et al., Epidemiology of animal bites and rabies
cases in India . A mu lticent ric s tudy. J. Commun. Dis., 2008, 40(1) ,
22. Suda rsan, M. K. et al., An ep idemiological study of animal bites
in India: results of a W HO sponsored na tional multi-centric r abies
survey. J . Commu n. Dis., 2006 , 38( 1), 32–39.
23. Khok har, A., Mee na, G. S. and Me hra, M., Pro file of dog bite
cases attendin g M.C .D. dispens ary at Alipur, Delhi. Indian J.
Commun. Med., 2003, XXVIII(4), 157–16 0.
24. Bajwa, M . A., A njum, A. and Manzoo r, S., Dog b ites; assessment
of burden, de terminants and public services ava ilable for their vic -
tims in district Lahore. Prof. Med. J., 20 12, 19 (5), 700–709 .
25. censusofindia.go v.in
26. Conce pts and Defin itions used in NSS, National Sa mple Survey
Organisa tion Golden J ubilee Publication, India, 2000, p p. 5–6.
27. Ber man, M. and Dunbar, I., The social behaviour of free-ranging
suburba n dogs. A ppl. A nim. Et hol., 1983, 10(1), 5–17.
28. Har ris, G., Where streets are thronged with strays baring fangs .
The New York T imes, 6 August 2012.
ACKNOW LEDGEMEN TS. This wor k was c arried out entirely in the
field, and no ani mals were harmed during the obse rvations. We thank
Manabi Paul, IIS ER-Ko lkata, for help with ana lysis. We also thank the
Council o f Scie ntific and Industria l Research, New De lhi; Ind ian
Nationa l Sc ience Acade my, Ne w Delhi and IISER-Kolka ta, fo r provid-
ing funds. We tha nk Prof. Raghavendra Gadagkar (IISc, Bangalore),
for use ful co mments that helped improve the manuscr ipt.
Received 13 J une 2013 ; revised acce pted 20 February 20 14