Benn and Evans seem to be intent on discrediting the meltwater hypothesis on the basis of questionable reasoning and philosophy. We would like to comment critically on their approach.

They state that most Quaternary scientists give little credence to the megaflood hypothesis and that non-specialists need to be protected from its flawed science. Did they conduct a poll to determine that most Quaternary scientists do not believe the hypothesis? Hardly: there are many Quaternarists who embrace these ideas. Surely Benn and Evans are not suggesting that these researchers are 'unfamiliar with the evidence' or that their research is also 'unscientific, unnecessary and inconsistent with the evidence'. As well, is there really a wider public that needs to be protected against our unscientific thoughts? The alternative to Benn and Evan's view is that the megaflood hypothesis is not really flawed; rather it challenges and threatens establishment research. It would not be the first time that establishment figures have railed against a fruitful hypothesis because they found it repugnant.

The reasoning on the dual interpretation of drumlins misrepresents our work and the work of Popper. We do not hold that there is a need for two types of drumlins because there is sorted sediment in one and not in the other. We hold this view because the sedimentary structure and architecture, clast shape and clast lithology in one type are so different from those in the other type. As well, the forms of cavity fill drumlins, spindle, parabolic and transverse asymmetrical, are so unlike classic drumlins that we felt obliged to give them different names—Livingstones and Beverleys (Shaw, 1996). Beverleys, with troughs wrapped around the proximal ends, are almost certainly erosional, but we do not, as Benn and Evans assert, base this argument 'exclusively' on perceived morphology. We consider morphology related to the action of horseshoe vortices, turbulent structures at forward-facing steps, truncation of internal structure, stone lags on the erosional surface and landform associations (e.g., Shaw et al., 2000). We are well aware that the internal structure and composition is less important in deducing the formation of erosional landforms, compared with that of depositional forms. In the case of landforms such as the French River erosional marks cut in granite and gneiss, this point is so obvious it is taken for granted. By contrast, we go to great lengths in the case of hummocks and large-scale fluting to demonstrate that internal structure and surface form are largely independent (Munro & Shaw, 1997; Shaw et al., 2000; Munro-Stasiuk & Shaw, 2002). We certainly do not accept that sedimentary structure, architecture and clast lithology are irrelevant to understanding subglacial landforms in the megaflood hypothesis. Why would much of our work involve sed-imentology if it were so evidently irrelevant to the meltwater hypothesis? Erosional drumlins may contain sorted and stratified sediment. Benn and Evans appear to have overlooked this point and insist that erosional drumlins are till cored. Hence their confusion on the duality of drumlins and their assertion that the megaflood hypothesis for drumlins is unfalsifiable. In reality, details of form, architecture, sedimentology and lithology allow us to distinguish cavity fill drumlins, Livingstones, from erosional drumlins, Beverleys. There are very specific predictions on the characteristics of erosional and cavity fill drumlins and we continue to use these predictions when interpreting landforms. We wonder why Benn and Evans make such a fuss over a matter that we have treated exhaustively in a number of papers.

We do not make our hypothesis unfalsifiable by protecting it from awkward evidence: the hypothesis is easily falsifiable. For example, the hypothesis would be rejected if the properties of the sediment in depositional landforms contradict the specific predictions for cavity fills (e.g. it is aeolian or marine or is deformed into the shape of the landform) or, for the case of erosional land-forms, the hypothesis is rejected if the patterns of defining ero-sional troughs show cross-cutting rather than bifurcating and merging relationships. Rather, we have amended the hypothesis as it became apparent that it was contradicted by certain observations. Thus, we introduced an erosional version because drumlins in bedrock, for example, demanded it. Benn and Evans argue that this is sleight of hand. We suggest that it is sensible. It seems illogical to argue that erosional drumlins and cavity fills cannot both exist. Sedimentologists and geomorphologists describe some bedforms as erosional and some as depositional without concluding that hypotheses on their genesis are unfalsifiable (Allen, 1982). Since Benn and Evans fail to recognize, rather they ignore, certain observations (e.g. the truncation at the land surface, the presence of boulder lags over erosional landforms, the origin of bedrock forms), they are the ones using the ad hoc protection device they accuse us of using. They are protecting their own models from the 'awkward' evidence we present. For instance, they state that in their experience all flutings and drumlins are mantled by glacitectonite or till and thus they can be easily explained by ploughing or deformation processes. They ignore observations that contradict their prediction (e.g. Rains et al., 1993; Munro & Shaw, 1997).

Shaw (this volume, Chapter 4) does not present imaginings as fact as Benn and Evans assert. More than any other modern hypothesis on subglacial bedforms, our work is grounded on fieldwork, experiment and imagery. We believe that whereas our facts are based on observation, much of the literature on subglacial deformation is based on modelling. We also believe that the most powerful explanations of these bedforms will come ultimately from a combination of both approaches.

William of Ockham's tenet—'entities should not be multiplied unnecessarily'—is very helpful in this discussion because the multiplication of entities has been necessary. Analogues from modern glaciated areas cannot explain Rogen moraine. Nor can they explain streamlined forms in bedrock and loess in the Scablands, lying beyond the limits of the Cordilleran Ice Sheet. Obviously we need additional entities to those from modern glacial environments. Hence, the Scablands streamlined forms are excellent analogues for erosional drumlins and fluting in bedrock and it is proper to use them in a hypothesis stating that some subglacial landforms resulted from megafloods.

The processes Benn and Evans champion as generally applicable fail to explain the characteristics of many bedforms. For example, Tulaczyk et al. (2001) point out that their ploughing method does not explain drumlin patterns. We invite Benn and Evans to explain the topology of a ploughing mechanism that bifurcates around the upstream ends of drumlins and does not cross-cut downstream crescentic troughs. It is topologically impossible that ploughing can behave in this way. Indeed, Tulaczyk et al. (2001, p. 64) write: 'Whilst this (the carving ofinter-mediate grooves by ploughing) does not explain the pattern of all bedforms such as drumlins and Rogen moraine, it may explain the observed form of megalineations'. Italics added. Chris Clark freely admits that this is a problem for their hypothesis and is the reason why they only deal with megalineations and not with drumlins. It is worth noting that Tulaczyk et al. (2001) argue very strongly against the ideas of Boulton (1987), yet Benn and Evans present the contradictory views of these authors as supporting their conclusions. Such inconsistency detracts from their arguments.

The modern glacial analogue approach is incapable of explaining either Rogen moraine or the scale of megalineations. Ploughing by ice keels cannot explain the form and pattern of drumlins. The megaflood hypothesis is attractive because it is not faced by these difficulties. More to the point, such difficulties make it necessary to introduce new entities; the megaflood hypothesis presents such new entities. There is no quarrel with William of Ockham here.

The comments on melt-out till and the generation of meltwa-ter for megafloods set up a red herring. We have been at pains to point out that the melt-out till precedes the megaflood and is commonly a remnant within erosional drumlins (e.g. Shaw et al., 2000). To demonstrate this, we present some of the most detailed field sketches of fluting sediment together with structural and fabric data. A stone lag, interpreted to have been produced by flood erosion, lies on an erosional surface, truncating melt-out till and diapiric mélange. It is this erosional surface that defines the fluting. The melt-out till preceded the megaflood that eroded the drumlin and meltwater involved in the formation of the till was probably of little consequence to the flood which originated far to the north. Of course, as pointed out since 1982, the water for melt-out till was released slowly; we present estimates of thousands of years for melt-out till formation (Shaw et al.,

2000). Once again, we are misrepresented and the arguments we make for a supraglacial origin for the megaflood discharge are overlooked.

Obviously, there is a major difference in perception between Benn and Evans and us. They consider that the megaflood hypothesis flies in the face of a huge body of mainstream research. In our defence, there is no known observation that contradicts the hypothesis. Nor does this hypothesis violate any fundamental principle in science. It might be incompatible with mainstream research, but the same can be said of any new paradigm. In answer to their assertion that our work is unscientific and unnecessary, our only response is—we do not think so. It would help if Benn and Evans were more specific where they write that our work is inconsistent with the evidence. Again, there are no known observations that contradict the megaflood hypothesis.

0 0

Post a comment