Experiments with myxobacterial aggregates reveal standing waves called rippling patterns. Here, these structures are modelled with a simple discrete model based on the interplay between migration and collisions of cells. Head-to-head collisions of cells result in cell reversals. To correctly reproduce the rippling patterns, a refractory phase after each cell reversal has to be assumed, during which further reversal is prohibited. The duration of this phase determines the wavelength and period of the ripple patterns as well as the reversal frequency of single cells.

Standing wave oscillations of the cell density (rippling) are observed in premature aggregates of developing myxobacteria. Recently the underlying pattern formation mechanism was shown to be based on the interplay between active cell motion and local interactions triggering reversals in the cells' direction of motion. The propagation of information through the system is mediated by the internal state of moving cells rather than by diffusible chemical signals. Discrete cellular automata and coupled-map lattices have been investigated earlier and indicate the importance of a minimum refractory period between subsequent reversals of a cell. In this paper we consider the continuum limit of the process, that yields a set of hyperbolic partial differential equations with a a single discrete time delay. The time delay corresponds to the duration of the mentioned refractory period of the cells. According to linear stability analysis a minimal time delay is required for a wave instability to occur. The results of the continuum model are in reasonable agreement with the findings in the discrete models adding credibility to the earlier studies.

Self-organization processes in multicellular aggregates of bacteria and amoebae offer fascinating insights into the evolution of cooperation and differentiation of cells. During myxobacterial development a variety of spatio-temporal patterns emerges such as counterpropagating waves of cell density that are known as rippling. Recently, several models have been introduced that qualitatively reproduce these patterns. All models include active motion and a collision-triggered reversal of individual bacteria. Here, we present a systematic study of a generalized discrete model that is based on similar assumptions as the continuous model by Igoshin et al (2001 Proc. Natl Acad. Sci. USA 98 14913). We find counterpropagating as well as unidirectional rippling waves in extended regions of the parameter space. If the interaction strength and the degree of cooperativity are large enough, rippling patterns appear even in the absence of a refractory period. We show for the first time that the experimentally observed double peak in the reversal statistics of bacteria in rippling colonies (Welch and Kaiser 2001 Proc. Natl Acad. Sci. USA 98 14907) can be reproduced in simulations of counterpropagating rippling waves which are dominant in experiments. In addition, the reversal statistics in the pre-rippling phase is correctly reproduced.