Discussion: Inferred subglacial meltwater origin of lakes on the southern border of the Canadian Shield

Abstract

In their paper, the authors, R. Gilbert and J. Shaw , add small "finger lakes" in narrow valleys incised through an up-icefacing cuesta to the lengthening list of landforms, large and small, that they and their like-minded colleagues ascribe to erosion or deposition by high-discharge subglacial drainage. This list has included, previously, drumlins, erosion marks, subglacial erosion ("s-") forms, Rogen moraine, fluvial dunes, tunnel channels, a moraine, and regional landform assemblages. Often in these works the glaciofluvial flooding argument is so convincingly based on the field evidence, and counters the glacial argument so effectively, that one feels the conceptual ground shift and one's stance upon it changed. Occasionally, however, as in Shaw and Gilbert (1990) and the present paper, the argument seems to be based on little more than that Other features have been explained by this process, so why not this feature too? In the present paper, the authors give an excellent detailed description of the distribution and form of long, narrow lake basins incised into a limestone cuesta facing up-ice over the Precambrian shield of southeastern Ontario. They liken them to the New York Finger Lake valleys and those in the Niagara cuesta of Ontario, which have been viewed previously as of glacial origin. "Thus, preferential erosion by fast-flowing ice and the presumed absence of subglacial tributaries convincingly account for the valley forms of the Finger Lakes and other reentrant or through-valleys" (p. 1634). However, they caution: "Any agent that might have preferentially eroded through-valleys could take the place of glaciers in the above discussion" (p. 1634). Any agent, that is, except subaerial fluvial erosion, which they discount on grounds going back to Playfair's law. " . . . if large volumes of subglacial meltwater were to be channelized across escarpments" (p. 1634, original emphasis), valleys would be cut similar, in many respects, to glacial troughs. Thus is the subglacial meltwater hypothesis introduced: not solely from evidence but with much supposition. " . . . tributaries would cut by the same floods anastomose (Shaw 1988; Brennand and Shaw 1994), and " . . .esker ridges [with dendritic patterns and accordant junctions] recordonly the most powerful flood events" (Brennand 1994, p. 9). The argument continues: both through-valley and reentrant valley regions (Finger Lakes and Niagara cuesta) lie astride major flood routes proposed by Shaw and Gilbert (1990). This would be fine if the latter reconstruction met with wide independent acceptance, but it seems unfair to seek such support from it when that paper so summarily dismisses the widely accepted glacial history of, for example, the western borderlands of Lake Ontario. The present argument has, nevertheless, gained momentum, albeit cheaply, leading to: "We suggest that meltwater floods may well explain the prominent valleys we describe here" (p. 1635, emphasis added). This seems strangely shy, but may recognize the thinness of the logical ice at this point. Furthermore, flow was across "a wide area upstream [of the escarpment] " (p. 1635), yet tunnel channels and an esker in one of them lie upstream as well as downstream of the escarpment (Brennand and Shaw 1994, p. 507), indicating channelled, not sheet, flow. "Nearly equal spacing" (p. 1635) of the valleys in question suggests to the authors that linearly concentrated escape of sheetfloods over the escarpment was hydrodynamically, rather than morphologically or structurally, controlled, as was the origin of fluting on the tops of escarpment promontories ("noses") (p. 1635). We are left in the dark, however, as to how this control works. Rather, we leap ahead to: "As the flow streamed around the escarpment and noses into the valleys . . ." (p. 1635). Given the valleys, we can follow the argument into counterrotating vortices, a residual medial ridge in one valley and the localization of valley-floor basins. But, it is exasperating to miss how a high-discharge, high-velocity sheetflood becomes patterned into longitudinal counterrotating vortices when it meets an escarpment. This is not to say that the phenomenon is incredible, but be unlikely" (p. 1635), yettunnel channels said to have been we are entitled t o more by -way of explanation than the . van Dyke reference. More accessible than this and charac-