Since Holland (1993) introduced the concept of tags as a possible cooperation forming mechanism in evolving system (among other things) a number of tag models with intriguing, and potentially very useful, properties have been advanced. However there is currently little detailed understanding of the underlying processes that produce these results. Specifically it is not know what (if any) are the necessary conditions for tag systems to produce high levels of cooperation. We identify what appears to be a necessary condition that previous tag models implicitly contained. We formulate our hypothesis by detailed analysis of the previous models and then explore and test it (in simulation) with a new model that demonstrates for the first time the importance of high mutation applied to tags. In order to import MABS techniques into engineering of MAS these kinds of analysis will be required. We need to understand the necessary conditions of desirable emergent properties not just existence proof...