Probabilistic flooding is a simple to implement, information dissemination scheme which is known to alleviate the broadcast storm problem. Simplicity of implementation is a critical requirement of network protocol design in nanonetworks and this renders probabilistic flooding a good candidate information dissemination solution. In this paper we examine the possible application of probabilistic flooding in nanonetworks using simulations and analysis. Our simulation results on two and three dimensional grid structures verify the existence of phase transition phenomena and indicate that the desired rebroadcast probabilities are affected by the network size, the transmission model and the node density with the latter parameter being the most critical. A mechanism for estimating the node density is thus required. For nanonetworks employing molecular communications we are able to derive a node density estimation scheme which is based on synchronous transmission of the network nodes. A linear parametric model of the unknown parameter is derived, which is used to derive an iterative estimation scheme based on online parameter identification techniques from adaptive control theory. The estimation scheme can be used as a baseline to develop an adaptive probabilistic flooding scheme for nanonetworks.