Project J.M. Aylsworth and W.W. Shilts Terrain Sciences Division. Abstract

Transcription

1 Glacial features of the west-central Canadian Shield Project J.M. Aylsworth and W.W. Shilts Terrain Sciences Division Aylsworth, J.M. and Shilts, W.W., Glacial features of the west-central Canadian Shield; & Current Research, Part B, Geological Survey of Canada, Paper 05-18, p , Abstract A small scale map of selected glacial features (eskers, related outwash and meltwater features, rogen moraine, and drift-free areas) was compiled for approximately km2 of the Canadian Shield lying between 60 and 66" N and west of Hudson Bay. It combines data obtained by airphoto interpretation of selected glacial sediment characteristics of 37 1: NTS map areas and surficial geology mapping of 28 additional sheets. This map is presented along with preliminary discussion of regional glacial sedimentation patterns. A widespread, integrated system of trunk and tributary eskers suggests stagnation of the Keewatin Ice Sheet while it still covered a large area. In addition, distribution of constructional glacial landforms suggests that erodibility of bedrock and character of resulting sediment significantly influenced the type and pattern of glacial sedimentation. This influence may ha've been as great a control on the nature and distribution of landforms as glacier dynamics. Les auteurs ont port6 sur une carte B petite Bchelle des formes du relief glaciaire (eskers, plaines alluviales et autres elements form& par les eaux de fonte moraine de Rogen et regions d6pourvues de sediments glaciaires) qui couvrent environ km 1 du Bouclier canadien, entre les 60e et 66e degres de latitude nord, Itouest de la baie dlhudson. 11s ont group6 des donnees obtenues par photo-interprgtation de certaines caracteristiques des sediments glaciaires observees dans 37 regions ca~tographiees b et par cartographie des formations en surface de 28 autres regions. Les auteurs accompagnent la presentation de cette carte dlun expose preliminaire des modeles de sedimentation glaciaire regionaux. Ltexistence dtun reseau etendu et int6gre dteskers principaux et secondaires semble indiquer que ltinlandsis du Keewatin a connu une phase stationnaire au moment oh il recouvrait encore une vaste rkgion. En outre, dtapr8s la repartition des formes construites du relief glaciaire, il semble que la facilite dt6rosion de la roche en place et les caract6ristiques des sediments qui en ont result6e aient profondement influ6 sur la nature et la forme de la sedimentation glaciaire, au point, peut-&re,.de contr6ler la nature et la repartition des formes de relief.9 Itinstar de la dynamique des glaciers.

2 Introduction Preliminary results of a reconnaissance airphoto interpretation of selected glacial features of much of the westcentral part of the Canadian Shield have been combined with more detailed mapping carried out near Hudson Bay by the authors and colleagues. The resulting map reveals distinctive patterns of deposition and erosion that may have considerable significance in interpretation of the history and dynamics of the western part of the Laurentide Ice Sheet. In this paper we present a preliminary small-scale map (Fig. 45.1) and a limited discussion of the meaning of the patterns of deposition and erosion that it reveals. Background Most of the study area was mapped at a general reconnaissance level by Geological Survey of Canada personnel in the late 1950s and early 1960s (Fyles, 1955; Lee, 1959; Craig, 1964, 1965). These early maps were restricted to depicting or symbolizing the most prominent geomorphic features, such as major esker ridges, and only portray rogen moraine, sometimes termed ribbed moraine (Cowan, 1968; Hughes, 1964), in a fraction of the area actually covered by it. No attempt was made to differentiate drift-covered from drift-free areas. These works were compiled onto the Glacial Map of Canada (Prest et al., 1968). Since the Glacial Map of Canada was published, more detailed surficial geology maps (1: and 1: ) have been produced for much of the District of Keewatin. These maps, in published or unpublished form, cover approximately 40% of the area of the present mapping project. For the remaining 60%, major glacial features have been compiled from interpretation of 1: scale air photographs on 1: scale maps under contract to Terrain Analysis and Mapping Services Ltd., Stittsville, Ontario. Information from these maps was plotted at 1:l and further generalized for presentation here (Fig. 45.1). General distribution pattern In Figure 45.1 it is possible to discern four zones of sedimentation patterns that are roughly concentric about the Keewatin Ice Divide: Zone I The innermost zone, characterized by the absence of eskers and rogen moraine (Lundqvist, 1969) occupies southern Keewatin and extends about 50 km on either side of the Keewatin Ice Divide (Lee et al., 1957). The characteristic landscape of Zone I comprises till plains with areas of low till hummocks and virtually no oriented depositional features. It is almost completely devoid of glaciofluvial deposits, except for minor outwash in some valleys. The location of the ice divide (Fig. 45.2), which forms the axis of the region, is clearly defined by striation orientations. Zone 2 Zone 2 is km wide, surrounds Zone 1 on three sides, and is characterized by the presence of well developed rogen moraine and esker-outwash systems. Rogen moraine comprises sinuous ridges trending roughly perpendicular to ice flow (Cowan, 1968; Shilts, 1977) and occurs in linear belts or trains parallel to major directions of ice movement. The trains of rogen ridges radiate from the Keewatin Ice Divide like the spokes of a wheel. In the areas between rogen trains, drift forms featureless or drumlinized or fluted till plains (see GSC maps ; 8, ; ; 2, ). Individual drumlins may occur within rogen moraine trains and individual large ridges may be fluted. Where rogen moraine is best developed, its distribution is little influenced by topography, occurring both in depressions and on uplands, but where the belts become more widely separated, down ice from the ice divide, rogen moraine generally occurs preferentially in depressions. Distribution of rogen moraine about the Keewatin Ice Divide is not uniform. It is rare west of Dubawnt Lake and east of Baker Lake (indicated by? on inset map, Fig. 45.1) where large areas are devoid of glacial deposits or are covered by a featureless till veneer, and northwest of Dubawnt Lake where an extensive drumlin field occurs. The northern boundary of the southwestern-most field of rogen moraine is remarkably distinct against the featureless till plain west of Dubawnt Lake. In general the down-ice margin of the zone of rogen moraine is abrupt although isolated, small, linear trains occur in the inner half of the next zone. Eskers also begin near the boundary between Zone 1 and Zone 2 and, like the rogen moraine trains, radiate from the region of the Keewatin Ice Divide. A typical esker system begins as a series of hummocks or short segments which pass downstream into continuous large eskers joined by areas of outwash or meltwater channels. Along a section measured across the southwestern part of Zone 2, eskers are spaced approximately 13 km apart, with spacing varying from 2 to 27 km. Throughout most of the area eskers are sharp ridged, up to 40 m high, with occasional conical knobs projecting well above the average elevation of the esker crest. Along their length they may be periodically interrupted by bulges where the single ridge splits into multiple ridges which coalesce downstream. Above marine limit they are commonly flanked by outwash in terraces disrupted by kettle lakes. North of Thelon River eskers are associated with prominent outwash terraces and the tops of the eskers are in some places flat, planed off by meltwater flowing on a stagnant ice floor into which the esker ridge was frozen (Shilts, 1984). Below marine limit, eskers are commonly reworked by wave action in the offlap phase of the Tyrrell Sea. This has had the effect of subduing their relief to the extent that they rarely project more than 10 m above the adjacent terrain. Sonar profiles show them to retain their relief and sharp crest where they are submerged in deep lakes (W.W. Shilts, unpublished data). Zone 3 Zone 3, which lies west of Zone 2, is characterized by an intricate dendritic pattern of eskers, and continuous drift cover. It forms a 200to 300 km-wide belt of glacial sediment, commonly till, which thins to a discontinuous veneer in the outer part of the zone. Large esker systems are known to extend into Hudson Bay and wave-reworked eskers, bearing erratics common to the Keewatin mainland, occur within the Bay on Coats Island (Shilts, 19.82), suggesting that an area similar to Zone 3 may lie east of Zone 2, submerged by Hudson Bay. Within Zone 3 rogen moraine occurs only as isolated, short, narrow, linear trains, lying in depressions and closely associated, in space, if not in time or genesis, with eskers. Within the rogen areas, individual ridges are much smaller and depart considerably in shape from the "classical" rogen moraine (Lundqvist, 1969) of Zone 2 (S. Paradis, personal communication, 1985). Compared to the 13 km spacing of eskers in Zone 2, spacing between eskers decreases to approximately 8 km (3-15 km range) 200 km downstream (southwest) from the

3 Esker (segment : continuous) Rogen mora ine Drift-f ree zone (,80% bedrock outcrop) " WO WO~ Scale /,.... / ~-- ; ;:...-- \ ~\ 1 / / I ----r I : / / " C--7 I? : ; 1~/!.,..-, \ I I I_...- ~ t, \ I 1/ 1..._./ n/.,... I \.ij,l ~ I l I \zone1/ 3 \ Zone 2 \I I Zo ne \ ' I \ '-J --..1_ : \ ) Zones of Distribution..., 'J 'J Figure 45.1 Distribution of selected glacial features on the west-centr al Canadian Shield.

4 / rfield 9-===. probable ice recessional position km... extent 0 f g 1 a c i a 1 1 a k e s extent of T y r r e l l Sea...~..._.... Figure Glacial geology features discussed in the text. section that was measured in Zone 2. Although numbers of Structural Province (Fig. 45.3). The long trains of relatively eskers and their tributaries increase in Zone 3, the overall continuous drift cover that extend northwestward along size (height and width) decreases in this zone. lowlands developed in the crystalline bedrock of the Beartslave province, are probably-trains of debris eroded and In the outer part of Zone 3 the esker systems become more discontinuous and disorganized, and esker distributaries transported from the poorly consolidated sedimentary rocks of Thelon Basin (Fig. 45.3). are common. Short esker segments are commonly oriented in various directions and areas of crevasse -fillings are associated with disruption of the esker system. Esker pattern Within the northwestern portion of this zone a prominent ice recessional position (Fig. 45.2) is marked by a chain of ice-contact deltas or fans, other outwash features, and esker distributaries which extends parallel to and north of the Back River. In the vicinity of this feature, short parallel esker segments abound between the major systems. Zone 4 The outermost zone of the Canadian Shield part of the study region is characterized by extensive bedrock outcrops that are nearly bare of drift. Although the transition from Zone 3 to 4 is abrupt in the southern part of the study area, long trains of drift project into major lowlands along the northern part of the boundary. The abrupt transition from drift to no drift in the south parallels 11 O0 longitude and then curves northeastward south of the Eastern Arm of Great Slave Lake. It corresponds roughly to the eastern boundary of the Fort Smith Belt, a northward-trending zone of radioactive gneisses and intrusive rocks (Charbonneau, 1980; Fig. 45.3). The transition zone north of Great Slave Lake corresponds with the eastern edge of the Beartslave The eskers radiating outwards from the Keewatin Ice Divide consist of networks df tributaries which form dendritic patterns similar to a simplified Horton system. These esker systems can be traced as far as 600 km and continuous lengths of individual eskers can be traced for up to 75 km. In many places, discontinuous segments are linked by meltwater channels, usually scoured through drift to bedrock or filled by outwash. Tributary eskers join the main esker preferentially from the left; that is, from the north on those eskers deposited by eastward-flowing meltwater east of the Keewatin Ice Divide and from the south on those eskers deposited by westwardflowing meltwater west of the divide (Fig. 45.1). This observation is best illustrated by the esker system trending northwestward from Dubawnt Lake where, in a distance of 300 km, 9 tributaries join from the south and none from the north. Similarly, between its origin near Yathkyed and Henik lakes and Hudson Bay most tributaries of the Maguse River esker system join the trunk stream on its north side. In most cases the orientation of the tributary esker changes abruptly as it approaches the main esker and the tributary joins the main esker at right angles.

5 %----~ ' I L, < " I \ \ Thelon Basin 1 & r- 98' western limit of Canadian Shield r " ' BearISlave Structural Province km L A < \ 2;... ::::::::::::::::::%:>:::A... Dubawnt Group: sedimentary 8 volcanic rocks... Figure Bedrock geology features discussed in the text. In some locations clusters of short, parallel esker ridges occur between trunk eskers, and distributary ridges fan out from trunk eskers. One such cluster occurs along the coast south of Chesterfield Inlet. There are several areas with similar clusters in eastern Keewatin and somewhat similar features, although more disorganized and associated with crevasse fillings, occur near the western edge of Zone 3. The eastern clusters may mark the position of the ice front during slowing of glacier melt-back and thus may correlate with similar positions northwest of the Keewatin Ice Divide (see above, Zone 3) (Fig. 45.2). Implication of patterns Even at the reduced scale of these figures. the resolution of major geomorphic features is c&siderably enhanced over the only previous compilation, the Glacial Map of Canada (Prest et all, 1968). The following are only preliminary comments based on an initial analysis of data currently on hand. Esker systems In an earlier paper, Shilts (1984, p. 218) described in some detail the characteristics of a typical esker of the western Canadian Shield; esker systems were described as reflecting 'I... what appears to have been a fully integrated Horton system of tributaries and trunk streams, regularly bifurcating upstream into lower order tributaries until the deposits disappear near the Keewatin Ice Divide. This pattern may be interpreted in at least three ways: I) The whole system may have functioned subglacially in sub-ice tunnels extending from the centre of a thin, stagnant glacier to its retreating margins, which lay at one time some km away. Although this model has some merit, the very size of the ice sheet at the inception of esker deposition would seem to argue against it, the thicker ice near the divide being too plastic at the base to maintain open tunnels. In addition, it is hard to image how the Horton system could have developed so fully within a solid mass of ice; topographic irregularities at the base of the ice would have exercised greater influence on esker trends than is evident from the Horton pattern. 2) The esker may have been deposited by streams flowing in short tunnels near the margin of the glacier, continuity being maintained by up-ice migration of the heads of the tunnels by melting (St-Onge, 1984, p. 274). Although this model is more compatible with observed sedimentation features and probable dynamic conditions in the retreating ice, it does not explain well the Horton pattern of tributaries. It is hard to imagine how a subglacial tunnel would bifurcate regularly as it melted up ice without some external control. 3) The most attractive model of glacial meltwater drainage in this region is presently one in which an integrated system of drainage channels developed on the surface of the glacier, the meltwater plunging to the base of the glacier to flow in a subglacial tunnel the last few kilometres of its course before issuing from the

6 retreating glacier front (similar to the model suggested by St-Onge, 1984, p. 273). This system would have developed quite late in the glacial cycle, when most of the glacier was below the equilibrium line. The tunnels near the ice edge would have extended themselves headward by melting, as in the preceding model, but their headward migration would have followed roughly the traces of the surface drainage, thus accounting for the regular bifurcation uostream. ' This hvbrid model best explains both 'the Horton drainagk pattern and the manifest evidence of subglacial origin of esker sediments." The dendritic pattern of the eskers provides the most compelling evidence of the stagnant nature of the ice sheet. It is difficult to conceive of a dendritic meltwater system developing and maintaining itself on the surface of an active ice sheet; it is similarly difficult to conceive of tributary eskers joining the main esker at right angles within vigourously moving ice. The reason for the preferred direction of approach of tributaries (from the left) is unknown. Perhaps it reflects the influence of Coriolis force on the movement of water on a large featureless plain -the surface of the ice sheet (a hypothesis that requires most of the ice cover over the area of this study to be below the equilibrium line). Disruptions in the esker pattern are thought to indicate temporary halts in the retreat of the ice front, possibly the result of climate deterioration. The most prominent of these zones lies in an arc north of Back River (Fig. 45.2) and is accompanied by an extensive outwash complex. When areas of numerous disorganized short esker segments, crevasse fillings, and esker distributaries, which occur in the outer portion of Zone 3, are linked, they seem to extend this marginal position in a single arc across the western part of the study area. Clusters of esker distributaries and short parallel esker segments in eastern Keewatin may also record positions of temporary halts of the ice front. The most prominent of them may be contemporaneous with the western position, the rapid retreat of the ice which fronted the more than 150 m deep water of the Tyrrell Sea, accounting for the asymmetry of their locations with respect to the ice divide. The height and width of an esker ridge is roughly proportional to the amount of debris supplied- to its conduit while it is flowing near or at the base of a glacier. Assuming that basal conduit length and duration of meltwater flow are similar for most eskers in this region, the diminishing size of the eskers with distance from the ice divide may reflect a regularly decreasing amount of basal debris in the glacier. 'Particularly in Zone 4, the small size and low frequency of eskers and the paucity of drift suggest that the ice was relatively clean. This implies 1) that there was little erosion of the generally hard crystalline bedrock of Zone 4 and 2) that most debris carried by ice through the region was derived from the inner zones and was mostly deposited before reaching Zone 4. If the assumption that the well developed dendritic esker systems developed in stagnant ice is correct, then the stagnant ice sheet would have covered zones 2 and 3. Ice over Zone I may have remained active while the outer zones were stagnant, although eventually it too would have become stagnant, thinning until only remnant ice blocks remained in lake basins. The absence of well developed, dendritic, esker systems over southern and western parts of Zone 4 may indicate that retreat of the ice sheet within this area was accomplished by back melting of active ice, but, alternatively, may only reflect the paucity of basal debris in stagnant ice. Much of the northern portion of Zone 4 contains well developed dendritic esker systems. The reason that eskers were developed in the northern part of Zone 4 may be that dispersal trains of debris from the east penetrated along the lowlands and provided sediment to the meltwater streams in this area. To the south, a lack of debris in the ice may have prevented the development of esker ridges. A third alternative suggested to the authors is that the lack of eskers in Zone 4 may be related to retreat of the ice front in a proglacial lake. This does not seem to be likely because eskers are well developed below marine limit east of the ice divide, and they occur both below and above lacustrine limits west of the divide (Fig. 45.2). Rogen moraine Rogen moraine is confined to a well defined belt around the Keewatin Ice Divide. Thus, distribution of fields of rogen moraine appears to be directly related to the position of the ice divide, and, therefore, to the glacier dynamics in the region adjacent to the ice divide. In Sweden, Lundqvist (19691, observed that rogen moraine occurred only above marine limit and then only in valleys or other depressions. In contrast, on the western Canadian Shield rogen moraine occurs with equal frequency above and below marine limit, and where best developed, is found both in depressions and on topographically positive areas. Furthermore, both rogen moraine and drumlins or flutings occur in long, narrow trains radiating in the direction of ice flow outward from the region of the ice divide. The two types of landform trains pass laterally one into the other, and individual trains commonly can be traced back to specific outcrops of particular bedrock lithologies (Shilts, 1977) or specific areas of coarse unconsolidated sediments. In general, rogen moraine is composed of coarse, bouldery or gravelly debris that is relatively undeformable, even when saturated. Conversely, at the few sites checked in eastern Keewatin, trains of drumlins or flutings appear to be underlain by more clay- and silt-rich drift that deforms readily when saturated. It is possible that the dynamic conditions in glacier ice near the ice divide were such that either drumlins or rogen moraine could be formed where a sufficient subglacial thickness or basal load of debris was available. Because of the lateral alternation of one feature with the other, and their compositional peculiarities, it is possible that wherever the basal debris load reached a sufficient amount, trains of either drumlins or ribbed moraine formed, with the type of feature formed dependent largely on the local physical characteristics of the debris. The reason for the abrupt down-ice termination of the rogen patterns is not known. Drumlins are most abundant in Zone 2 but continue to occur through Zone 3; rogen moraine, on the other hand, is concentrated in Zone 2 and occurs only sporadically and in depressions in the other zones. Drift cover The lack of drift and apparent lack of erosion within Zone 4 may reflect the resistant nature of the underlying bedrock, the dynamics of the ice, or a combination of both. A prolific source of debris is available in the easily eroded sedimentary rocks of the Thelon Basin. This debris, however, was largely depleted before the ice passed from Zone 3. Possibly the crystalline rock that underlies the Beartslave Province and Fort Smith Belt was largely resistant to erosion and so did not yield much autocthonous debris. Although the absence of drift in Zone 4 probably reflects the resistant nature of the underlying bedrock, it is possible that ice dynamics may have been influenced by local topography, particularly south of Great Slave Lake where the relief is rugged and the trend of valleys is perpendicular to the direction of movement of ice. As the elevation of this area is 300 m higher than that at the centre of the ice sheet and

7 isostatic depression at the time would have accentuated this difference, the ice sheet may have been thin enough over this area to be frozen to its bed, preventing erosion. Another possible explanation of the lack of evidence of erosion is that ice within the valleys oriented transverse to the direction of ice movement, may have been largely stationary so that most regional ice flow occurred above the base of the ice sheet. If this was the case, little erosion could have occurred. Conclusions 1. The pattern of esker distribution described herein strongly suggests that the last stages of Wisconsin Glaciation west of Hudson Bay consisted of the backwasting of a large, thin, stagnant ice sheet centred on the District of Keewatin. The diameter of this relatively dormant ice mass may have exceeded 1500 km. 2. The pattern of rogen moraine distribution is primarily associated with the configuration of the Keewatin Ice Divide. The trains of rogen moraine and associated drumlins that radiate from the ice divide region are also associated with areas of outcrop of specific bedrock lithologies or unconsolidated sediment, suggesting that the erodibility of the substrate and characteristics of the eroded material in transport have a strong influence on the nature and distribution of resulting landforms. 3. Large areas where drift cover is thin or absent may be resistant to glacial erosion and may also be located beyond the zone of deposition of easily entrained bedrock debris (principally in the Thelon Basin) which lies near or on the Keewatin Ice Divide. 4. Finally, although the regional patterns of drift deposition and landform development may be related to some extent to regional dynamic conditions at the base of the Keewatin ice sheet when it was actively flowing, there is little doubt that the lithology and topography of outcrops over which the glacier passed exerted an important influence on its deposits. References Charbonneau, B.W. 1980: The Fort Smith radioactive belt, Northwest Territories; Current Research, Part C, Geological Survey of Canada, Paper 80-LC, p Cowan, W.R. 1968: Ribbed moraine: Till-fabric analysis and origin; Canadian Journal of Earth Science, v. 5, p Craig, B.G. 1964: Surficial geology of east-central District of Mackenzie, Geological Survey of Canada, Bulletin 99, 41 p. 1965: Glacial Lake McConnell, and the surficial geology of parts of Slave River and Redstone River mapareas, District of Mackenzie; Geological Survey of Canada, Bulletin 122, 33 p. Fyles, 3.G. 1955: Pleistocene features; Geological notes on central District of Keewatin, Northwest Territories by G.M. Wright; Geological Survey of Canada, Paper 55-17, p Hughes, O.L. 1964: Surficial geology, ~ichicun-~ania~iskau maparea, Quebec; Geological Survey of Canada, Bulletin 106, 20 p. Lee, H.A. 1959: Surficial geology of southern District of Keewatin and the. Keewatin Ice Divide, Northwest Territories; Geological Survey of Canada, Bulletin 51, 42 p. Lee, H.A., Craig, B.G., and ~ ~les, 3.G. 1957: Keewatin Ice Divide; Geological Society of America Bulletin, v.68, no.12, pt.2, p (abstract). Lundqvist, : Problems of the so-called Rogen moraine; Sveriges geologiska undersokning, Ser C NR 648 (Arsbok 64 NR 51, 32 p. Prest, V.K., Grant, D.G., and Rampton, V.N. 1968: Glacial Map of Canada; Geological Survey of Canada, Map 1253A, 1: scale. St-Onge, D.A. 1984: Surficial deposits of the Redrock Lake area, District of Mackenzie; & Current Research, PartA, Geological Survey of Canada, Paper 84-IA, p Shilts, W.W. 1977: Geochemistry of till in perennially frozen terrain of the Canadian Shield - application to prospecting; Boreas, 6, p : Quaternary evolution of the Hudson/James Bay Region; Naturaliste Canadien, v. 109, p : Esker sedimentation models, Deep Rose Lake map-area, District of Keewatin; & Current Research, Part B, 'Geological Survey of Canada, Paper 84-lB, p

8 GEOLOGICAL SURVEY OF CANADA PAPER 85-1 B COMMISSION GEOLOGIQUE DU CANADA ETUDE 85-1 B CURRENT RESEARCH PART B RECHERCHES EN COURS Issued in two sections1publiee en deux volumes: pages andlet pages

9 @ Minister of Supply and services Canada 1985 Available in Canada through authorized bookstore agents and other bookstores or by mail from Canadian Government Publishing Centre Supply and Services Canada Ottawa, Canada KIA 0S9 and from Geological Survey of Canada offices: 601 Booth Street Ottawa, Canada KIA OE rd Street N.W., Calgary, Alberta T2L 2A7 100 West Pender Street Vancouver, British Columbia V6B lr8 (mainly B.C. and Yukon) A deposit copy of this publication is also available for reference in public libraries across Canada Cat. No. M44-85/IBE Canada: ISBN Other countries: $1: (for both volumes) Price subject to change without notice Cover Folding in the Neruokpuk Formation along the Firth River. These illustrations appear in report 27 by D.K. Norris (p ). Plissement dans la formation de Neruokpuk le long de la rivihre Firth. Ces illustrations apparaissent dans le rapport 27 par D.K. Norris (p ).