Biology Reference
In-Depth Information
2.5 From Stone Soup to Mulligan Stew
h e directed-graph-network model I have presented is obviously incom-
plete, although I believe that it captures the essential elements of emer-
gent complex social behavior. In the sections that follow I discuss the
features of the model that can be altered to better rel ect reality but also
result in the model being perhaps less general and understandable.
2.5.1 Decision Functions Are Not Fixed
h e model assumes that the threshold functions of individual ele-
ments are i xed. However, learning experience af ects response
thresholds and the probability that an individual engages again in a
given task. Another ef ect is to generate greater variance in the distri-
bution of thresholds in a colony
{F}
and af ect the colony's behavioral
dynamics. One example is the response of a colony to odors it has
learned are associated with specii c food sources. Odors that initially
have no ef ect on recruiting experienced foragers recruit large num-
bers of foragers when they are presented inside the nest at er foragers
have experience with them.
2.5.2 Response h resholds Change with Nutrition
Jochen Erber and I performed a set of experiments that demonstrated
dif erences in response thresholds to sucrose between pollen and nec-
tar foragers. I was on sabbatical leave at the Technical University of
Berlin. I had been invited to give a lecture at an international confer-
ence and wanted to discuss some elements of the response-threshold
models I had been working on for 10 years. I needed a demonstration of
a response threshold for my talk. Until this time, I had treated response
thresholds as black-box phenomena that must be true. How else can
one explain behavior? Erber suggested that we collect some foragers at
the entrance of a colony in the Technical University's apiary and test
them using the proboscis extension response (PER) test (Figure 2.7).
He would take pictures. We subjected bees to increasing concentrations
