I.—River Terrace Profiles.
A Diagram of the development of river terraces published recently by Lobeck and another produced by Hills* draw attention to a necessary deduction from Davis's hypothesis of terrace-development during restrained downcutting. The abandonment of unmatched terraces alternating in level on opposite sides of a valley calls for a to-and-fro swinging of a river across the valley during slow degradation; and generally it is assumed that a whole meander belt so swings.†† Lobeck's diagram shows this process as cutting successive flood-plain floors which slope alternately right and left across the valley; and the conclusion that something of the nature of slip-off slopes must be left during combined across-valley migration and downward cutting is inescapable. It is very doubtful, however, whether these floors are usually sloping plains which are smooth for their whole width except for the usual swells and swales and other inequalities of flood-plains. If such is the case, the treads of broad terraces that result from swinging may be expected to have continuous slopes of measurable declivity away from the valley sides, these being, of course, distinct from and beneath any surface slopes that have resulted from the accumulation of valley-side talus and fans on the terraces. Every terrace tread of this kind would be itself a polygenetic terrace, as defined by Chaput. It does not appear that any precise measurements are on record which would prove that terrace treads commonly slope in this manner; but the hypothesis is worth testing, and examination of terrace treads with the object of detecting any systematic across-valley slopes may be recommended to observers in New Zealand.
In most cases it may be expected, however, that a meander belt migrating sideward and downward instead of developing a smooth slope (polygenetic terrace) would carry down with it a strip as wide as the belt itself, which would have a floor approximately horizontal in transverse profile, while at the same time leaving behind on the slip-off slope of abandoned flood-plain strips an irregular flight of low terraces with level treads. Should the river not degrade in
[Footnote] * A. K. Lobeck, Geomorphology, p. 238, 1939; E. S. Hills, Physiography of Victoria, p. 109, 1940.
[Footnote] † W. M. Davis, River Terraces in New England, Bull. Mus. Comp. Zool. Harvard, vol. 38, pp. 281–346, 1902.
well-defined free-sweeping meanders, however, but be either confined to a restricted channel or spread in a braided course over a somewhat wider strip, it is possible that in some cases the minor slip-off slope terraces on a future major terrace tread may be lower and more numerous, thus producing a closer approximation to a smooth slope.
Fig. 1.—Diagram illustrating the development of river terraces by a degrading: river swinging from side to side of its valley.
In any case only exceptionally broad terraces may be expected to exhibit on their treads either minutely terraced or smooth transverse slopes, and where valley-side terraces are relatively narrow they may be expected to preserve only parts of the floors that were developed as level plains, these being either parts of the meander belts or perhaps of rather broad strips that may have been occupied by channels in braided patterns. It would appear, therefore, that level across-valley profiles will be most often found on terrace treads, and, indeed, the treads are shown level in all Davis's diagrams. In such cases the hypothetical slip-off slopes, either smooth or minutely terraced, down which successive lateral migrations have taken place, have been destroyed by later swings of the river, as suggested in Fig. 1.
The diagram drawn by Hills is apparently based on that of Lobeck, of which it may be regarded as a simplified version. It frankly deduces strong slip-off slopes on terrace treads. The photograph of terraces in the valley of an Australian river which accompanies the diagram on the same page fails, however, to support the slip-off slope deduction, for the indication of level afforded by standing water on the terrace in the foreground of the view makes it very clear that the slope of the tread is not towards the valley axis, as shown in the diagram, but rather away from it, though the actual departure from the horizontal may be explained as due to the presence of an abandoned stream channel.