• d u k a s a n d rat c l i f f esummarized in three books at the end of the twentieth century (Balda et al.
1998; Dukas 1998a; Shettleworth 1998; see also Shettleworth 2009). Much
new research building on these foundations has been conducted in the past
decade, and some of this exciting work is presented here.
Cognition can be divided into several interrelated and perhaps inseparable
components. “Perception” is defined as the translation of environmental sig-
nals into neuronal representations. “Learning” is the acquisition and mainte-
nance of neuronal representations of new information. “Working memory”
comprises a small set of neuronal representations active over some short dura-
tion. “Attention” refers to the neuronal representations activated at any given
time. “Long-term memory” consists of passive, stored representations of pre-
viously learned information. Finally, “decision making” involves the determi-
nation of action given the available information about relevant environmental
features and experience (Platt 2002; Dukas 2004a).
At the mechanistic level, cognitive traits are determined and modulated by
genes, neurons, and hormones. At the functional level, cognitive abilities are
evolved features that are shaped by natural selection based on their relative
contribution to animal fitness and that are subject to a variety of mechanis-
tic and phylogenetic constraints (Dukas 1998b). Since research pertinent to
cognitive ecology is conducted in several scientific disciplines, we have to
integrate a vast amount of knowledge from all these fields in order to under-
stand the ecology and evolution of animal cognition (Ratcliffe et al. 2006).
The following chapters represent a variety of attempts at achieving this chal-
lenging goal.
Our definition of “cognition” at the start of this chapter focuses on the
acquisition, retention, and use of information. Organisms can perform rather
well only by perceiving and responding to external information. Most or all
animals, however, also acquire and retain information, and in chapter 2, Du-
kas discusses the ecological and evolutionary significance of this ability. The
mechanistic research on animal learning and memory is typically conducted
under the necessary controlled laboratory conditions using a few animal spe-
cies whose ecology and behavior in the wild are not well known (e.g., the do-
mestic chicken, Gallus gallus domesticus, and the Norway rat, Rattus norvegicus).
A notable exception is the honeybee, Apis mellifera, which has been studied
extensively in the field (Seeley 1996; von Frisch 1967). The honeybee is an
ideal model system for integrating mechanistic knowledge on genes, neurons,
and hormones with whole-animal information on behavior and ecology, as
Fahrbach and Dobrin elegantly illustrate in chapter 3.
Part II presents new work on established and emergent research programs
relating cognition to avian ecology. Song learning in birds is one of the few ex-
ceptional cases where a central behavior has been thoroughly examined from