Congratulations are due to Ruth King who had the paper "Semi-Markov Arnason-Schwarz Models" accepted by Biometrics and is available online as early view (http://onlinelibrary.wiley.com/doi/10.1111/biom.12446/abstract).

This paper develops new multi-state models for capture-recapture(-recovery) data that is able to efficiently incorporate memory into the process. Individuals observed within a study are recorded as in a given (discrete) state which vary over time. However, the observation process is imperfect so that the state of the individuals is frequently unknown. Traditional models assume that the transition process between states is memory-less (i.e., first-order Markovian) but this is often biologically unrealistic. For example, the length of time an individual is in a state of "hungry" will typically be a function of when it last ate. The semi-Markov model is defined by specifying a distribution on the dwell-time distribution for each state (i.e., how long an individual remains in a given state). An efficient model-fitting algorithm is described that approximates the semi-Markov model by a first-order Markov model using a state aggregation approach, taking into account the presence of unknown state observations. The model is applied to house finch data where state refer to the presence/absence of conjunctivitis and where we show that there is some evidence that the dwell-time state for the absence of conjunctivitis is not memoryless.