A flooding algorithm is an indispensable and fundamental network control mechanism for achieving some tasks, such notifying all nodes of some information, transferring data with high reliability, getting some information from all nodes, or to reserve a route by flooding the messages in the network. It is applied to grasp the network topology in a sensor network or an ad-hoc network, or to retrieve content information by mobile agent systems. A flooding algorithm has the advantages of robustness and optimality by parallel processing of messages. However, the flooding mechanism has a fundamental disadvantage: it increases the processing time until the flooding control is finished and causes congestion in the network. In this paper, we propose and evaluate methods for producing a more efficient flooding algorithm by adopting the growth processes of primitive creatures, such as molds or microbes, to produce a simple flooding control mechanism.