Biology Reference
In-Depth Information
(b)
(a)
10
80
80
Fo
od
Experiment
Model
Alarm
70
70
8
60
60
50
50
6
40
40
4
30
30
20
20
2
10
10
0
0
0
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
Food
Control
Food &
Alarm
Alarm
Group size
Group size
Treatment
Fig. 6.23
Shoal size in banded killifish. (a) Median group size in experiments where the treatments were: food
odour added to aquarium, control, food odour and alarm odour (extract of broken skin of killifish), and alarm
odour only. (b) The variation in observed group sizes across two of the treatments, compared with outcomes of
a simulation which varied an individual's response to neighbours (see text). From Hoare
et al
. (2004). With
permission from Elsevier.
they also do not know which are the informed individuals in the group. Computer
simulations reveal that even a small proportion of knowledgeable individuals (less than
5% in a group of 200) can influence the direction of movement of the whole group and
guide everyone to a target location. Therefore, group decision making can occur without
explicit signals or complex communication (Couzin
et al
., 2005).
Experiments with groups of humans provide evidence for this effect (Dyer
et al
.,
2008). Some individuals (naïve) were given a piece of paper with the instruction 'stay
with the group'; others were given the instruction 'move to a specific target'. The results
showed that a small informed minority could guide a group of naïve individuals to the
target without verbal communication or obvious signalling. When informed individuals
had conflicting information about the target, then the majority dictated the group
direction.
In animal groups, too, individuals who move in specific directions to resources, for
example because they are hungrier (Rands
et al
., 2003) or bolder than the rest (Harcourt
et al
., 2009), become leaders of the group's movements.
A small informed
minority can lead
the group
Voting
In some cases groups reach a consensus about when and where to go by opinion polling
(voting), without any control by leaders (Conradt & Roper, 2005). The pooling of
information is common in human groups. Francis Galton (1822-1911), mathematician
and cousin of Charles Darwin, attended a cattle fair in which people placed wagers on
the weight of an ox. Galton noted down all the wagers (nearly 800). Individual wagers
varied widely but the overall average was remarkably close to the true weight of the ox.
Therefore, pooling information can reduce extreme errors and result in the best overall
solution.
