The popularity of social information systems has been driven by their ability to help users manage, organize and share online resources. Though the research exploring the use of tags is relatively new, two things are widely acknowledged in the research community: (a) tags act as a medium for social collaboration, navigation and browsing and (b) an overall stable equilibrium exists among tag patterns due to the social nature of the tagging process. But there is very little agreement on what causes these stable patterns. In this paper, we take an evolutionary perspective to understand the process of tagging to investigate whether tags act as "way finders" or digital pheromones in social tagging systems. We investigate the existence of tag trails based on a semantic similarity measure among existing tags. We found that over 50% of the resources we evaluated exhibited strong trail patterns. The implications of these patterns for the design and management of social tagging systems...