Tills are first derivative sediments of bedrock (Shilts 1976) and therefore retain some of the lithological and geochemical characteristics of their original bedrock source. The abundances of source components in till are commonly used to define distinctive two-dimensional spatial patterns of concentration, so-called dispersal fans, plumes or trains (e.g., DiLabio 1990), that decrease immediately down-glacier of the bedrock source. Commonly these dispersal patterns are wedge-, amoeboid- or ribbon-shaped with concentrations elongated parallel to ice flow and decreasing in content inversely with distance from the source (Shilts 1976). The interpretation of two-dimensional dispersal patterns can be confused by: multiple ice-flow directions producing palimpsest or composite trains (Parent et al. 1996; Stea and Finck 2001); topographic influence resulting in preferential erosion or ice streaming (Broster and Huntley 1995); or post-depositional weathering and modification of geochemical content (Shilts 1976; Paulen 2001). This is particularly problematic in sub-alpine terrains such as in northern New Brunswick and other areas of Canada. Such areas are characterized by highly variable topography and ice movements, having experienced glaciation by growth of individual cirque glaciers and their coalescence into larger valley glaciers or ice sheets (e.g., Broster and Huntley 1995). In addition, the bedrock of local Appalachian terrains has been weathered to depths of up to 26 m (Veillette and Nixon 1982). This pre-glacial weathered material was incorporated into the overlying tills that locally exceed 4 m in thickness behind obstructions and along valleys and underlie a veneer of humus commonly 10 cm thick.

Northern New Brunswick is a region of known sulphide mineralization associated with Devonian base-metal (Cu-Pb-Zn-Ag-Mo) porphyry and skarn deposits. Eleven mineral occurrences and a gossan (Rose and Johnson 1990) were identified as potential target sources that are used to examine the effects of multiple ice-flow directions and relative variability between local and regional dispersal of till clasts and matrix geochemistry.

The study area is located in north-central New Brunswick, 20 km southeast of the city of Campbellton (Fig. 1). Access to the study area is provided by provincial highways and an extensive network of secondary and logging roads. New Brunswick is located in the Appalachian region of Canada (Bostock 1970). The province is divided into six main physiographic regions, based on tectonostratigraphic zones and structural elements, with each region subdivided into smaller physiographic elements (Bostock 1970; Rampton et al. 1984). The study area is situated in the Chaleurs Uplands physiographic region, encompassing two subdivisions: the Jacquet Plateau and the Chaleur Coastal Plain (Fig. 2). The Jacquet Plateau is a low, undulating surface with elevations generally less than 300 m, although a few ridges reach elevations of 600 m. Generally, the plateau slopes gently to the north and east and is dissected by ridges and valleys also oriented to the northeast. The Chaleur Coastal Plain, bordering the Baie des Chaleurs, is a moderately drained gently undulating plain, commonly below 70 m elevation, sloping towards the bay (Fig. 2).

The study area is located in the Chaleur Bay Synclinorium tectonostratigraphic zone (Ruitenberg et al. 1977; Fyffe and Fricker 1987; Wilson 2000a; Carroll 2003). Bedrock is gently folded, with limbs dipping 2°– 15°. Fold hinges trend north-east-southwest and plunge gently to the southwest. The oldest rocks in the area are middle to late Ordovician arc-related mafic volcanic rocks and Caradocian slates and cherts of the Balmoral Group (Fig. 3), an inlier of Dunnage Zone rocks exposed in the core of the Popelogan Anticline (Ruitenberg et al. 1977; Williams 1979; van Staal and Fyffe 1995; Wilson 2000b, c). The Balmoral Group is disconformably overlain by late Ordovician to early Silurian Matapedia Group sedimentary rocks, consisting of turbidite deposits. The Matapedia Group is conformably overlain by Silurian sedimentary and volcanic rocks of the Chaleurs Group (Ruitenberget al.1977; Williams 1979; van Staal and Fyffe 1995; Wilson 2000b, c).

Bedrock is best exposed along river valleys and ridges, but is otherwise covered by a blanket of till. Argillaceous and calcareous sandstones and siltstones of the Chaleurs Group and limestones interlaminated with calcareous shales of the Matapedia Group underlie most of the western part of the study area (Fig. 3). Mafic volcanic rocks, minor tuff, chert, and graptolitic slate of the Balmoral Group are also present (Wilson 2000b, c).

The majority of the eastern part of the study area is under-lain by Chaleurs Group volcanic rocks. Massive amygdaloidal and porphyritic basalts occur in the northeastern and central regions of the study area. Felsic volcanic rocks and associated sedimentary rocks are found in the northeast, as well as to the south. In the north, sandstone, shale, and conglomerate of the Carboniferous Bonaventure Formation are exposed along the Baie des Chaleurs. The central region is underlain mainly by fossiliferous limestone and volcanic boulder conglomerate, breccia, and tuff.

According to Irrinki (1990), Devonian granite, granodiorite, and monzonite stocks are common throughout the study area (Fig. 3). Dykes and sills associated with the stocks comprise dioritic feldspar porphyry, granite porphyry, gabbro, and aplite (Quatermain 1981; Woods 1993).

Currently, there are eleven known mineral occurrences in the study area (sites 1– 11, Fig. 3) and a gossan (site 12) that serve as point sources for glacial dispersal studies.

The Popelogan skarn occurrence is the largest mineral occurrence in the study area, containing 3.1 million tonnes @ 0.3% Cu (Quartermain 1981). The Popelogan Group sulphide mineralization (Cu-Mo-Pb-Zn) is found in the (1) Popelogan North, (2) Popelogan South, (3) Popelogan East, and (4) Popelogan Lake zones (Fig. 3). The Popelogan Lake mineralization occurs as disseminations and stringers hosted by quartz-carbonate veins and stockworks associated with fracture/shear zones (Rose and Johnson 1990). The Popelogan North and South zones are hosted by a skarn formed around apophyses of the Popelogan granodiorite stock and related dykes and sills intruding Silurian limestone. Sulphide mineralization (Cu, Mo, Zn, Pb, W, and As) occurs as disseminations, isolated pockets of massive material, and in quartz-chlorite-carbonate veinlets. The Popelogan East Zone (site 3, Fig. 3) is hosted by a quartz vein system cutting the Popelogan intrusion and hornfels in the contact aureole. It is characterized as a greisen and stockwork vein deposit with mineralization occurring as selvages along quartz veins and as vein fillings.

The Benjamin River Complex comprises four deposits of porphyry-type Cu-Mo mineralization: (5) Benjamin River South A, (6) Benjamin River South B, (7) Benjamin River South C, and (8) Benjamin River South D (Fig. 3). The mineralization is stockwork-type with fine disseminations (chalcopyrite and molybdenite), mineralized narrow fractures, and quartz veinlets, and is related to a Devonian granodiorite intrusive complex intruding Silurian Bryant Point Formation volcanic rocks (Rose and Johnson 1990).

The Goulette Brook Cu-Pb-Zn-Ag deposit (site 9, Fig. 3) occurs as irregular veinlets, disseminations, and small isolated pods in highly altered palagonite tuff of the Goulette Brook Formation (Rose and Johnson 1990). The Lower McNair Brook North (10) and Charlo River (11) deposits (Fig. 3) are copper occurrences; however, insufficient information is presently available to classify them further. Information is also lacking at present, on the Balmoral Gossan (site 12, Fig. 3) that may be the source of some geochemical dispersal trains (e.g., Mo), discussed below.

Local studies of the surficial geology and till geochemistry conducted by Pronk (1986, 1987), Pronk and Parkhill (1988), Lamothe (1992) and Parkhill and Doiron (2004) indicate a main ice-movement to the east followed by movement to the northeast. Pronk et al. (1989) further determined that the Baie des Chaleurs region experienced four major phases of ice flow during the Late Wisconsinan: 1) southeast, 2) east, 3) north-northeast, and 4) multidirectional. The eastward flow was found to be the most dominant trend in striaton and clast provenance studies, whereas pebble fabrics measured in glacial sediments represent the later northeastward flow (Pronket al. 1989). They attributed this to two possibilities: a regional shift in ice-flow from east to northeast, resulting from a change in basal regime under the influence of drawdown in the Baie des Chaleurs, or to a northeast flow representing a separate event within the Late Wisconsinan advance (Pronket al. 1989).

Due to the complex glacial history and variable physiography, the study area has a variety of glacial sediments and land-forms. Basal till blankets most of the study area and occurs as three different end-members: melt-out, lodgement, and deformation tills (sensu stricto, Dreimanis 1976) . The till-types form a continuum as melt-out till grades downward into lodgement and/or deformation till. At some sites deformation till occurs at the surface as a thin veneer overlying bedrock that can be differentiated from ablation till, which occurs as a veneer of coarse, less compact, striated fragments of local bedrock with the finer grain-size material winnowed away. These deposits were sampled where basal till was not present.

Only one stratigraphic till unit was observed, although the colour and texture vary from site to site, reflecting the characteristics of the underlying bedrock. The texture varies from sandy loam overlying igneous bedrock, to silty clay loam overlying sedimentary bedrock. Where exposed in section, till thickness is seen to grade from a thin veneer (<0.5 m) at elevations above 450 m, to till blankets (>2 m) at elevations below 400 m. Till veneer overlying bedrock is common in the southeast, while thicknesses exceeding 4 m occur in low lying areas in the western part of the study area.

Stratified glaciofluvial, glaciolacustrine and glaciomarine deposits are present along most river valleys and in coastal areas, overlain by post-glacial fluvial and colluvial sediments. Colluvial deposits occur as a thin layer of broken and weathered bedrock clasts in a poorly sorted, sandy matrix along most slopes of the area. In relatively flat-lying areas, a well-developed soil profile with humus cover typically comprises the upper 1 m of most surficial deposits.

Surficial mapping of the study area was carried out during the 1999 field season for the New Brunswick Department of Natural Resources and Energy. Ice-flow indicators (striations, grooves, and rat-tails) were recorded from bedrock outcrops encountered during the sampling and mapping program. (Fig. 2)

During field traverses, 171 samples were collected (Fig. 4) for lithological and geochemical analyses. Geochemical data were supplemented by re-analyzing an additional 157 archival samples saved from previous government surveys (Lamothe 1986; Pronk 1986, 1987; Pronk and Parkhill 1988: Fig. 4). Using topographic maps (1:50 000 and 1: 20 000) and aerial photos (1:35 000 and 1:12 500), samples were collected to form an average sampling density of one sample per four square kilometres. Sample density was increased to between 50 to 100 m spacing in the area of the Popelogan porphyry-skarn occurrences (Figs. 3, 4).

Samples were collected from hand-dug pits, road cuts, and natural exposures. Sample pits were approximately 0.5 m in diameter and ranged in depth from 0.2 to 0.8 m. Approximately 3 to 5 kg of till matrix was collected from the C- horizon for geochemical analysis. Basal till was the desired sample medium; however, ablation till was sampled where basal till could not be found within a 100 m radius of the designated sample site. Field duplicates were collected at random (1 in 20 sites) to evaluate the precision of the sampling procedure. From 25 to 100 pebble-sized (2– 10 cm diameter) clasts were collected from the C-horizon at each sample site for lithological identification. Fragments of larger clasts were included to avoid any sampling bias.

Measurements of striae orientation at 40 sites indicate ice movements in three general directions: northeast, east, and southeast, with the most dominant trend towards the northeast. Only one outcrop, located approximately 10 km to the east of the study area, indicated a cross-cutting relationship with striations showing all three ice movements (Fig. 1). This cross-cutting relationship indicates relative ages of the ice movements, from oldest to youngest, of: 1) 160°, 2) 110°, and 3) 010°, which supported the earlier findings of Pronk et al. (1989).

It is an accepted hypothesis that large transport distances (10– 100s km) are required to comminute most rock types to clay-size fragments or so called "terminal grades" (Dreimanis and Vagners 1971). In the study area known mineral occur-rences are reflected as anomalous point source geochemical and clast dispersal patterns in till overlying the known sources or immediately down-ice of them. Examination of dispersal patterns indicates derivation of local material and relatively short transport distances of 1 to 5 km for till geochemical material and less than 10 km for clast transport. These distances are in agreement with those generally found elsewhere in New Brunswick, likely due to erosion under cold-based ice. (e.g., Balzer and Broster 1994; Munn et al. 1996: Stumpf et al. 1997; Parkhill and Doiron 2004). Thus, much of the clay content in the till is probably due to incorporation of pre-glacial residual soils and transported sediment, of which the latter would have been prominent along major valleys.

Observations of striations and streamlined landforms support the conclusions of Pronk et al. (1989) of three main directions of ice-flow across the study area. The earliest movement, to the southeast (at around 168°), is rarely preserved in the study area and is only recorded at a few sites. This glacial flow direction represents initial glacial erosion from the advance of local Appalachian ice (Pronk et al. 1989) that transported some glacial debris to locations southeastward of their source. The major eastward flow (110°) is likely correlative to the Jacquet and Baie des Chaleurs flow patterns of Ramptonet al. (1984). Eastward movement has been reported (Chalmers 1886, 1887; Gauthier 1982a, b; Rampton et al. 1984; Pronk 1986, 1987; Bobrowskyet al. 1987; Pronk and Parkhill 1988; Pronk et al. 1989; Lamothe 1992; Parkhill and Doiron 2004) as the dominant direction of flow through the area. This middle event eroded much of the striation evidence of the earlier event, and dispersed eastward some of the previously eroded drift. A later ice-flow direction recognized in the study area is flow toward the northeast (010° to 060°). Striations indicating glacial movement in this latter direction are found predominantly in the northeastern part of the study area (Fig. 1). This event probably represents a shift in flow direction across the study area due to drawdown into the Baie des Chaleurs (Pronk et al. 1989). While this flow phase may have been less erosive than the two earlier flows it did produce new dispersal patterns and dilute or re-direct previous concentrations in some areas obliquely to the previous direction of flow transport. This re-direction of dispersed material occurred at different magnitudes across the study area and under different degrees of erosion, thus developing hybrid tills (cf. Stea and Finck 2001) and palimpsest dispersal trains modified by the later event (Parent et al. 1996). It may be that till was not deposited and re-eroded by the subsequent movements, but that the drift was retained in the ice mass while ice flow was redirected, possibly due to a change in the location of ice-growth or ice drawdown (Stumpf et al. 2000). However, the overall result is general eastward and northward dispersal patterns, albeit of different magnitudes for different areas.

Comparison of dispersal patterns across the Popelogan skarn occurrences indicates that some anomalous geochemical concentrations in the study area represent local up-ice sources of unknown mineralization. Dispersal trains are recognizable for only a few kilometres, indicating that where a geochemical dispersal pattern exists, a source is close. However, if the source is buried under till it may be difficult to find due to the complexity of previous ice-movement directions across the area. The best indicators of transport directions are till clasts from known sources, since these can be easily identified in the field and are not susceptible to hydromorphic re-dispersal. Even with subsequent changes in direction of glacier flow, interpretation of local movement vectors is easily made from measurement of dispersal patterns from known point sources.

Till clast , granulometric, and geochemical dispersal patterns in the study area mainly reflect the middle eastward and later north and northeastward ice-flow directions recorded by striae. In the southeast corner of the study area eastward-trending striations are dominant, in contrast with northward-oriented clast dispersal trains recorded in the same area (Figs. 6, 7). These data suggest that eastward-moving ice eroded and transported material, but that deposition also occurred during later northward movement. This interpretation may explain the north-south orientated geochemical dispersal patterns (Figs. 9, 10) centered about the Benjamin River occurrences (Fig. 3).

Evidence indicates that in areas of multiple ice-flow events: (1) indicators of glacial erosion (striations and streamlined landforms) can be biased towards earlier, highly erosive events (palimpsests) and that (2) till clast dispersal patterns can be more useful than striae or till geochemical dispersal patterns for identifying buried exploration targets (Broster and Huntley 1995).