Mathematical manuscripts recently excavated from tombs sealed in China around the beginning of the unification of the Empire (end of the 3rd century BCE) provide new evidence about mathematical activity in early imperial China. In fact, and more precisely, they provide evidence of a new type, when compared to writings handed down through the written tradition. The thesis for which this paper aims to argue, with respect to the history of mathematics in ancient China, is that these sources allow us to perceive a diversity of mathematical cultures in early imperial China. These writings mainly, if not only, deal with mathematical problems and procedures solving them —in contemporary words, “algorithms.” This feature, as well as others, indicates that these cultures had much in common, and were thus by no means impervious to each other. However, they also show key differences. It is by focusing on how they document “knowledge at work” that these differences can be brought to light. This is the second issue, this one related to historiography, for which I shall argue in the presentation. The source material and topic dealt with will require that I discuss what I mean by a “mathematical culture” and that I consider the question of which aspects of a given culture can be documented by these sources.