Oecologia (2007) 154:411–421
DOI 10.1007/s00442-007-0834-8
123
BEHAVIORAL ECOLOGY
Contact networks and transmission of an intestinal pathogen
in bumble bee (Bombus impatiens) colonies
Michael C. Otterstatter · James D. Thomson
Received: 19 March 2007 / Accepted: 23 July 2007 / Published online: 23 August 2007
© Springer-Verlag 2007
Abstract In socially living animals, individuals interact
through complex networks of contact that may inXuence
the spread of disease. Whereas traditional epidemiological
models typically assume no social structure, network theory
suggests that an individual’s location in the network deter-
mines its risk of infection. Empirical, especially experimen-
tal, studies of disease spread on networks are lacking,
however, largely due to a shortage of amenable study sys-
tems. We used automated video-tracking to quantify net-
works of physical contact among individuals within
colonies of the social bumble bee Bombus impatiens. We
explored the eVects of network structure on pathogen trans-
mission in naturally and artiWcially infected hives. We
show for the Wrst time that contact network structure deter-
mines the spread of a contagious pathogen (Crithidia
bombi) in social insect colonies. DiVerences in rates of
infection among colonies resulted largely from diVerences
in network density among hives. Within colonies, a bee’s
rate of contact with infected nestmates emerged as the only
signiWcant predictor of infection risk. The activity of bees,
in terms of their movement rates and division of labour
(e.g., brood care, nest care, foraging), did not inXuence risk
of infection. Our results suggest that contact networks may
have an important inXuence on the transmission of patho-
gens in social insects and, possibly, other social animals.
Keywords Epidemiology · Social insects ·
Disease ecology · Infection
Introduction
Interaction networks are pervasive in biological systems.
Food webs, neural nets, and metabolic pathways, in particu-
lar, have garnered much attention (May 2006; Montoya
et al. 2006; Proulx et al. 2005); yet, the primary motivation
for studying networks has often been to understand better
the spread of disease (Keeling and Eames 2005; Newman
2003). Conventional epidemiological theory (Anderson and
May 1991) assumes that hosts are randomly interacting
entities, such that every susceptible individual is equally
likely to encounter, and acquire infection from, an infected
individual. Network theory (Newman et al. 2006), in con-
trast, assumes that contagious pathogens propagate via the
highly structured interaction networks that exist among
socially living organisms; thus, risk of infection varies
among hosts according to their location in the social net-
work. Although theoretical advances (Meyers 2007;
Newman et al. 2006) have provided a Wrm basis for the
study of contact network epidemiology, empirical, especially
experimental, studies of disease spread on networks are lack-
ing, largely due to a shortage of amenable study systems.
Although there is a clear connection between contact
networks and the spread of sexually transmitted diseases in
humans (Friedman et al. 1997; Klovdahl 1985; Liljeros
et al. 2003; Neaigus et al. 2001), the relevance of network
structure for the spread other diseases, particularly those of
non-human animals, is uncertain. Recent work suggests
that disease transmission might vary with social network
structure in wildlife (brushtail possums, Corner et al. 2003;
African buValo, Cross et al.
2004). It is ironic, though, that
almost no studies have investigated the contact network
epidemiology of highly social species, such as the social
insects. A notable exception is the study by Naug
and Smith (2006), which showed that, in honey bees
Communicated by Nathan Sanders.
M. C. Otterstatter (&) · J. D. Thomson
Department of Ecology and Evolutionary Biology,
University of Toronto, 25 Harbord Street,
Toronto, ON, Canada M5S 3G5
e-mail: michael.otterstatter@utoronto.ca