Reliability represents a central problem in animal signal-use.  According to basic theory signalers may benefit from signaling unreliably, but this destabilizes signaling systems, because receivers will ignore unreliable signals, ultimately rendering signaling useless.  But how unreliable can signals be, before this happens?  I review a simple model and experiment that addresses this question (McLinn & Stephens 2006).  The model suggests that environmental uncertainty places an important constraint on the level of uncertainty that receivers will tolerate.  For example, when uncertainty is high, receivers can benefit from partially reliable signals. 

      In contrast, however, receiver may be very intolerant of reduced reliability when uncertainty is low. An experimental study from my laboratory supports these basic claims.  These models assume that receivers can easily discriminate signals.  To address this limitation, I extend these results to consider imperfectly discriminable signals using signal detection theory. Signals must be more reliable when receivers discriminate imperfectly, but this relationship depends on environmental uncertainty, because receivers bias their sensory thresholds to reflect the most likely outcomes.  Finally, I will discuss implications of these results for game theoretical signaling models.

Suggested reading:

• McLinn, C. M, and D. W. Stephens (2006) What makes information valuable: signal reliability and environmental uncertainty? Animal Behaviour 71:1119-1129 download 430kb pdf