For single-state capture recapture data, an m-array which is a (years-1) x years x groups array, where element [i, j, g] contains the number of individuals in group g released in year i and recaptured in year j+1 (by definition no recaptures can occur in year 1). If no groups are specified, this will be a (years-1) x years matrix. The last column contains the number of individuals released in year i and never recaptured.
For multi-state capture-recapture data with s states (including potential unobservable states), an m-array which is a (years-1)*s x (years-1)*2 + 1 x groups array. An element [i, j, g] contains the number of individuals in group g released in year t and state n (i = (t-1)*s + n) and recaptured in year k+1 in state m (j = (k-1)*s + m). If no groups are specified, this will be a matrix. The last column contains the number of individuals released in year t and state s and never recaptured. The labeling of unobserved states starts with the number of observed states + 1.