In the past information statistical analyses have been used for the description of the temporal organization of behaviour. In general, these analyses are based upon data collected with the sequential method. This method is known to limit the general applicability of information statistical analyses. In the present report a new model, based upon data collected with the frequential method, is presented. In this model too it is assumed that animals behave according to a MARKOV process. In the present model, however, the assumption of a MARKOV process implies that information (entropy) of a certain behaviour is a function of time, and not a function of "order in time". The key-advantage of this analysis is that it can be assessed to data collected during relatively short lasting observation sessions; using a computerized time recording system with a resolution of 0.1 s, for instance, a 60 min. session provides 36000 sampling points, viz. sufficient data for filling completely a transition matrix. This, in turn, implies that the analysis allows the calculation of the maximum error of the entropy (information) using a purely analytical method. Due to the ideal combination of a high sampling frequency and a relatively short collecting time, this analysis can be used to describe the individual relationships within social group structures in great detail. In the present report the applicability of this analysis is illustrated by assessing it to a single set of data collected in a study o spontaneously occurring changes in a group of 4 Java monkeys (Macaca fascicularis). the analysis deals with data collected in 17 sessions, one hour each, and recorded on separate days. the new analysis provided a detailed and quantitative description of dynamic changes, which could not be detected with the classical frequency analysis. The new analysis also provided a new dimension to the description of dynamic changes which were detected with the frequency analysis.