The analyses of surface wind and pressure at 2 observation times are coupled by using the forecast equations of horizontal momentum as dynamical constraints. A Cressman objective analysis procedure is used to interpolate the observations to a uniform grid. The interpolated data at these times are adjusted in a least squares sense subject to the 2 equations of constraint. Although the basic algorithm is designed to adjust data at 2 time levels, more time levels can be incorporated by using the algorithm in a sequential manner. The primary advantages of this scheme over a conventional analysis are: (1) the wind and pressure field are coupled, (2) observations at earlier and later times are incorporated into a given analysis, (3) the analyses have time continuity and, (4) data void regions can be analyzed if data are available at adjoining time levels. Since the adjustments are made in accord with the forecast equations, the methodology can be used for initialization of prediction models or updating these models by inserting new data in a consistent fashion. The scheme is tested on the well-documented squall line case of June 8, 1966, which occurred in the National Severe Storms Laboratory (NSSL) network in central Oklahoma. The data set consists of the hourly surface observations from the Aviation Network (Service A teletype records). A 3-hr period prior to squall line development is examined in order to see the influence of the larger scale circulation upon the generation and organization of mid-latitude squall lines. The principal results of the study were: (1) the observed wind components must be adjusted by 1–2 m s ?1 (root-mean-square value) in order to satisfy the governing constraints, (2) the gradual build-up of the convergence zone associated with the squall line is better depicted by the adjusted pattern and, (3) the correlation between surface convergence and radar echo is more reasonable when adjusted winds are used. DOI: 10.1111/j.2153-3490.1978.tb00867.x