In this paper I examine Harry Collins’s influential writing on tacit knowledge. In particular I turn my attention to his recent book, Tacit and Explicit Knowledge [Collins 2010], or TEK, which is arguably the most complete and systematic statement of what he means by the term “tacit knowledge”. As well as examining tacit knowledge as elaborated in this contribution, I draw out an underlying tension in Collins’s major contributions to the sociology of scientific knowledge in general between the realism underlying his notion of “tacit knowledge” and the constructivism underlying his other well-known concept, “the experimenters’ regress” (as for instance, elaborated in his well-known book Changing Order [Collins 1992]). In order to make this argument I pay particular attention to an aspect of his writings on tacit knowledge which I think is worthy of closer examination: namely the sorts of empirical support claimed for the features and properties of tacit knowledge to which he attends. In short I ask questions concerning some of the specific empirical examples and the conclusions he draws from them.