MTS markets are an example of quote driven electronic order book markets for Government securities. Proposals are firm, immediately executable and aggregated in a limit order book. In this paper we analyse the evolution of the interaction between the order book and order flow by exploring the determinants of the flows of limit and market orders over three years. We are able to test several hypotheses coming from theoretical models. This is, to our knowledge, one of the first empirical test of these hypotheses based on Government bond data. We find that market and limit orders show positive autocorrelation and clustering as in Biais et al. (1995). No activity is clustered as well. This diagonal effect could also explain the positive impact of book depth near the quote on the flows of new limit orders. On the contrary, depth beyond the second best price seems to play no role in book dynamics. Furthermore, increasing competition, measured as an increase in the number of operators, has a negative effect on orders. In addition, we find mild support to the theoretical prediction of a positive effect of book depth on order aggressiveness. Best spread, instead, behaves consistently with theoretical models: a larger best spread encourages new limit orders inside-thequote.