Garvine and Monk [1974] have contributed to the body of brackish water through the tank. The arrow at the top inknowledge of river plumes. Their results fill in many details of dicates the convergence line of the front. Ahead of this point some earlier work done at our laboratory in which the outflow the brackish (darker) water has increased its thickness. (At the of the Nid river in Trondheimsfjord was treated as a problem , same time, a very thin layer of freshwater that escaped from in hydraulics.The side spreading of the lighter river water acts like a density front for which the (two-dimensional) front velocity computed by means of the Bernoulli equation yields a densimetric Froude number Fr = 2 [Moshagen, 1972]. The degree to which shear stresses enter into the propagation velocity is a matter of speculation. Benjamin [1968] implies that it is important, at least for small-scale laboratory experiments. This is not necessarily true on the length scale associated with large river plumes.Considering the difficulties of measuring the thickness and velocity, which tend to be highly unstationary near a front, the values of 1.7 and 3.3 appear to support the purely dynamical computation. These values have nothing to do with the problem discussed by $tommel and Farmer [1952], as is suggested by Garvine and Monk.The Nid river plume was computed by considering the time history of a cross-sectional slice of the river in a quasi-onedimensional model. By preserving the river momentum and allowing the river to spread sideways with Fr = 2 the width of

No abstract