This paper applies Posterior Cramer-Rao Bound theory to the SLAM problem to measure the information supplied by different sensor modalities over time. Range-only, bearing-only and full range-bearing sensors were considered, as well as the gain in information achieved by using multiple sensors in centralized co-operative SLAM. An efficient recursive formula was used to compute the bound for a set of simulated scenarios, and its validity verified by comparing the bound with the second-order error performance of Fast SLAM 2.0 and the EKF.