The Southern Alps of New Zealand constitute part of an orogenic belt forming on the western margin of the Pacific plate due to collision along the Pacific-Australian plate boundary. The topography of the mountain range is asymmetrical with a steep inboard zone adjacent to the plate boundary where very high rainfall causes rapid erosion, and a relatively gentle outboard zone in the rain shadow east of the topographic divide. Detritus, including gold, shed from the rising mountains are transported eastward from the outboard zone into a foreland basin, and westward from the inboard zone across the plate boundary to the Australian plate. Rivers in the outboard zone are widely spaced (>20 km) and structurally controlled oblique to the trend of the orogen. These rivers are long (>150 km) with relatively low gradients, and form braid plains showing net aggradation over the past million years. Gold placer formation in outboard rivers draining the central Southern Alps requires concentration factors of ca. 10⁵ to 10⁶, which has occurred only during localized postdepositional uplift and recycling of sediments. Tectonic shortening coupled with low outboard erosion rates results in capture of some outboard rivers by inboard rivers, with consequent redirection of gold into the inboard region. Inboard rivers are generally short (20 km) and closely spaced (10 km), and oriented perpendicular to the plate boundary. Aggradation of sediments is temporary and periodic scouring of the valleys to bedrock by floods and glaciers results in ephemeral placer deposits in river bed lags and flood gravels on the Pacific plate. Gold is ultimately transported across the plate boundary on to the indentor (Australian) plate, in till or river gravels, where it is concentrated in degradational lags in fluvial sediments and in marine beach placers. Dextral tectonic motion at the plate boundary has transported detrital gold up to about 180 km laterally from Southern Alps sources during the past 5 m.y.
 
 

Gold eroded from quartz veins on the slopes of actively uplifting and physically weathering tributary catchments is transported to the fluvial system with little modification by mass movement mechanisms. Gold concentrates initially in the bed-load component of spatially limited, high-grade, primitive placers in moderate- to high-gradient, incised, juvenile valleys in which the rivers flow on or near bedrock. Interglacial and/or uplift-induced downcutting events and major floods periodically reconcentrate gold into lags on bedrock. Maximum gold particle size increases for the first few kilometers of primitive placers then decreases progressively downstream. Coarse gold (>3 mm, >0.5 g) and gold with low flatness index (<7) becomes lodged in bedrock crevices and is only remobilized by bedrock erosion. Gold undergoes minor to moderate rounding in primitive placers but little or no flattening or folding.

Gold transported through primitive placer valleys enters the bed load of a short, lower-gradient, transition zone and then the trunk river system. Spatially extensive, moderate- to low-grade trunk placer deposits form in basins where trunk rivers emerge from gorges. Basin uplift, interglacial downcutting, and major floods periodically reconcentrate bed-load gold into intraformational or basal lags in the trunk rivers. Gold flattening and folding commence in the transition zone, and flatness index maxima, roundness, and the proportion of folded particles increase progressively down the trunk rivers. Distal trunk placer gold is typically up to 2 mm in size, has a relatively high flatness index (>10), and is well rounded and commonly folded.

Flattening progressively increases both the surface area-to-volume ratio and the entrainment potential of gold particles. A reliable relationship between fluvial transport distance and flatness index maxima throughout the river system defines a critical flatness index above which gold is entrained and below which gold is not transported. At given values of river gradient, velocity, and bed roughness, both entrainment and retention of gold in primitive placer valleys are controlled by shape and mass characteristics that are largely inherited from the primary sources. In transition and trunk placer valleys, gold entrainment is controlled by flattening of particles to a critical flatness index state, whereas gold retention is controlled by flatness index reduction to subcritical state by particle folding, as well as input of relatively proximal, locally derived, primary or paleoplacer gold with subcritical flatness index or mass.
 

A semiquantitative study of various physical features of approximately 2,700 gold particles from 21 lode deposits and 36 placer deposits in the Klondike district, Yukon Territory, Canada, was undertaken. The shape parameters of roundness and flatness were determined using a classification method. There appears to be a systematic variation between these parameters and between the parameters and the distance of transport of the particle. These relationships were confirmed by quantitative measurements on critical samples. Rim characteristics were also estimated. They show a systematic relationship between one another, with the particle shape, and less clearly, with the distance of transport.

Gold particle shape data shows a smooth, well-defined relationship to distance of transport from the lode source of the gold. Both roundness and flatness show a rapid increase within the first 5 km from the lode source and a slower more linear increase beyond. This relationship seems to apply worldwide. The relationship between the shape of a gold particle and the rim characteristic supports the conclusion that rims are formed by the removal of Ag, Hg, and Cu, and not by the precipitation of Au. Thus, rim thickness and the associated high fineness is probably the result of the dynamic process of Ag removal in stagnant sediments (to create the rim) and abrasion in active sediments (to remove the rim). The composition of the rims represents the stable composition of gold in normal surficial environments. Together these observations place limits on the occurrence of new gold.

Hammering is the main cause of shape change in fluvially transported gold particles. Its principal effect is to flatten particles. Abrasion has its most marked effect on surface texture. Hammering and abrasion both decrease the mass (size) and increase the roundness of the gold particles. It is therefore possible to make a connection between hydraulic conditions during gold particle transport and its shape change.
 

We have determined the major and trace element composition of approximately 2,700 gold particles from 21 gold-bearing mesothermal quartz veins and 35 placer gold samples from the Klondike district in western Yukon Territory. Measured Au, Ag, Cu, and Hg contents were used to define a characteristic geochemical signature (or fingerprint) for each of the vein samples. These signatures were then compared with the various compositional populations that we have distinguished within each of the placer samples. Preliminary conclusions derived from the study include: (1) placer gold in both recent stream deposits and in the Pliocene to Pleistocene White Channel Gravels is detrital in origin; (2) the placer gold is mainly, if not entirely, derived from mesothermal quartz veins; (3) all lode sources for the placer gold have not yet been located; and (4) gold composition data can be used to identify the lode signature for placer gold and, if it has not been removed by erosion, help locate and link placer gold to its specific lode source.
 

Toroidal-shaped gold particles, extracted from a sample of the Basal reef orebody in the Welkom gold field that contained crossbedded gold, have been recognized in thin section, demonstrating that the form in question is not an artifact of extraction but is a genuine detrital form. Similar forms have been recovered in great abundance from an eolian paleosurface in Yakutia and from a recent storm-beach placer in New Zealand, indicating the likelihood that eolian processes produced these forms. This has been confirmed by means of wind-tunnel experiments. Toroids in various states of preservation have been extracted from a number of other Witwatersrand paleoplacers where there is an association with ventifacts, indicating that wind deflation might have played a role in concentrating gold from placer-sand deposits.

The concentration of heavy minerals in a sedimentary deposit depends, inter alia, on the composition of the source material, local hydraulic conditions during deposition, sizes and densities of the sediment grains, and sizes of the grains relative to the bed roughness. Effects of these factors have been investigated in a laboratory flume by depositing pure heavy mineral and mixed heavy and light mineral sediments on a stationary granular bed. Experiments with pure heavy minerals indicate that once a minimum mobility condition is exceeded, deposits will be fairly extensive and uniform, and only slightly affected by the flow condition. In all experiments, deposits were generally finer grained than the supply sediments and, in some cases, were bimodal. The concentrations and total amounts of heavy mineral in the deposits were influenced by the concentrations in the supply mixtures and were strongly dependent on the sizes of the light mineral grains. The results show that settling velocity is a poor indicator of hydraulic equivalence, while critical shear velocity appears to be a good indicator.
 
 

The Ventersdorp contact reef is an auriferous conglomerate horizon and economic orebody on top of the Late Archean Witwatersrand Supergroup in South Africa. It differs from the other conglomerate beds (reefs) in the Witwatersrand basin in having a massive metabasalt sequence as its hanging wall and by its intense postdepositional alteration. A geochemical, petrologic, and fluid inclusion study of the Ventersdorp contact reef reveals a polyphase hydrothermal alteration history. Regional metamorphism under acidic conditions was followed by potassic alteration along the reef. The potassic alteration, which also extended into the hanging-wall and footwall rocks, was subsequently followed by a chloritization episode that was restricted to the reef and its immediate contacts. Mass-balance calculations suggest that the latter alteration required only minor fluid influx from an external source, whereas the potassic alteration reflects the progressive dehydration of the siliciclastic footwall rocks and associated hydration of the basaltic hanging wall. Based on microtextural observations on quartz using SEM-cathodoluminescence imaging, the chloritization event is ascribed to the 2020-Ma Vredefort impact event that created the necessary secondary permeability in the form of microfractures particularly within the conglomerate bed. Considering that the Ventersdorp contact reef is generally regarded as the most intensely altered reef in the Witwatersrand basin, the relatively minor fluid infiltration established in this study suggests that, contrary to prevailing opinion, the other auriferous conglomerate beds might have experienced even less influx of externally derived hydrothermal fluids.
 

The eastern Arctic shelf is a typical passive continental margin and is characterized by complex, branched, fluvial channel networks of various ages. At least four main generations of buried valleys can be distinguished: Eocene, Oligocene, early Miocene, and Pliocene. Younger Quaternary buried valleys typically occur in both inshore and offshore zones covered by thick middle to late Pleistocene ice-loess sediments. These drainage networks reflect gradual restructuring and subsidence of the Mesozoic platform, as well as repeated sea-level changes. The accumulation of ice-loess sediments on the shelf and coastal plains during the late Pleistocene epoch was also a significant factor in the restructuring and burial of the fluvial channels.

Paleochannels contain gold- and tin-bearing placer deposits. Four types of placer regions are distinguished on a mineralogical basis: (1) regions with gold- and rare earth element-bearing placers (area of subsidence of the Kharaulakh and Kular fold structures); (2) regions with tin-bearing placers derived mainly from cassiterite-silicate primary sources (Chokurdakh-Lyakhovskaya and similar transform zones); (3) regions with gold- and tin-bearing placers (main part of the coastal area east of the Kolyma River mouth); (4) regions with rare metal-(e.g., Ta, Nb, Sc, Be), gold-, and tin-bearing placers (in the area of median massifs). Placer potential generally increases with decreasing age of the channelsóthe Pliocene and early Pleistocene channels are the most productive. Fluvial placer formation, however, decreased drastically in the early Pleistocene when connections between sources and valleys were broken. Some of the eastern Arctic shelf placers are large or superlarge deposits and these contain more than half the placer gold and tin resources of the area. Small geochemical anomalies associated with sea-floor sediments are potentially important indicators of buried fluvial placers.

Tin placers of the east Arctic shelf area are Arctic analogs of the Tin Islands shelf placers of southeast Asia. There are also distinctive parallels between the Chukotka coastal plain gold placer and equivalent beach placers at Nome, Alaska. In contrast, gold placers of the western sector of the eastern Arctic shelf (Kular gold region) formed without marine influence. Technology and economics determine the spatial limits within which the search for placers is possible. In general, placer prospecting is limited to areas of moderate subsidence with less than 30 m of sedimentary cover. The width of this zone varies from 10 km near the Chukotka coast up to 100 km at the coastal and inshore plains of Yakutia. The zone can also extend up to some hundreds of kilometers seaward within the boundaries of the large transform structural and metallogenic zones such as at Chokurdakh-Lyakhovskaya.
   

The post-Gondwana history of the major rivers in the western part of South Africa is important because these rivers were instrumental in the development of diamond placers along the west coast of southern Africa. The evolution of the drainage systems that developed after breakup of west Gondwana can be viewed in three time-slots: the middle to Late Cretaceous, the early to middle Cenozoic, and the late Cenozoic periods.

During the middle to Late Cretaceous there were two main river systems draining the interior. The one in the south, also referred to as the Karoo River, had its source in the present upper Orange/Vaal drainage basin and its outlet was at the present Olifants River mouth. The second and more northerly system, also known as the Kalahari River, drained southern Botswana and Namibia and entered the Atlantic Ocean via the lower Orange River. Erosion dominated the period immediately after breakup of west Gondwana and most of the diamonds released during erosion of Cretaceous kimberlites in central South Africa were transported by the Karoo River to the coast.

By early Cenozoic times, the lower Kalahari River had captured the upper part of the Karoo River and the broad configuration of the present Orange River network was established. This capture and northerly shift of the Orange River, on the newly exhumed pre-Karoo surface, was the result of an accelerated uplift of the southern and eastern subcontinental margins ca. 100 to 80 Ma. During the early and middle Cenozoic, the climate was arid to semiarid. This resulted in a substantial reduction in erosion rates and hence few diamonds were released from the primary bodies during that time.

Late Cenozoic fluvial gravels, however, dated as either middle Miocene or Plio-Pleistocene, contain diamonds that were reworked out of older Tertiary fluvial deposits. Sediments at the base of the Koa Valley and in the upper terraces in the Sak Valley formed the Koa River, a major tributary of the Orange River during the Miocene, and drained most of the area previously occupied by the lower Karoo River. The Koa River thus reworked diamonds trapped in the Cretaceous Karoo River deposits or terraces.

Younger sediments of the Carnarvon Leegte were never part of the Koa system. In fact, the Sak River captured the upper Koa River by late Pliocene times and the Plio-Pleistocene lower terraces in the Sak Valley and the paleo-Carnarvon Leegte joined as the paleo-Hartbees Riveróanother major tributary of the Orange in the Plio-Pleistocene.

Although climatic changes were the major controls that initiated the alluvial pulses during the Cenozoic, asymmetric uplift of the subcontinent was ultimately responsible for the northwesterly shift of the Orange River.
 

The Late Cretaceous to early Tertiary Droogeveldt gravels, one of the more famous alluvial diamond diggings in the Vaal basin of South Africa for both grade and large stone size, were totally exploited in the earlier part of this century. The best recoveries were obtained from high elevation (1,100- to 1,130-m elev) bimodal gravels situated in narrow (2- to 20-m wide), straight-sided, bedrock depressions <1 km long, some 5 to 10 km from the current Vaal River (1,000-m elev). These gravels, forming a fining-upward profile, comprise cobble to pebble, silica-rich clasts characteristic of the current Vaal drainage basin set among now-weathered bedrock boulders derived locally from the underlying Archean Ventersdorp volcanic rocks. The gravel-filled depressions, known by the early mining community as "sluits" coincide with major structural lineaments in the Ventersdorp bedrock. Although the now rare, remnant weathered profiles resemble colluvium or debris flow-type deposits, the presence of potholed and smooth, polished, but irregular bedrock footwall in the sluits as well as rounded and percussion-scarred silica clasts points to a more energetic fluvial influence. The well-jointed bedrock sidewalls of the sluits provided readily accessible boulders (oversize clasts) to these gully-like features. Thus, the exceptional diamond concentrations and large stone size in the Droogeveldt gravels are attributed to a combination of primary, fixed turbulent zones formed by the narrow bedrock gullies with attendant pothole and scour trapsites, and secondary, quasi-fixed turbulent zones associated with the locally derived boulders (oversize clasts) in a paleo-Vaal River drainage.

The sluits are not eroded kimberlite dikes and the bulk of their diamonds are likely to be sourced from the Cretaceous kimberlites in the greater Kimberley area.
 

The Orange River, the principal conduit for transportation of diamonds from the southern African interior to the Atlantic coast, has within its lower valley two recognized suites of gravel terraces in which part of the passing diamond population has been trapped. Both suites lie on eroded bedrock and both are downstream thickening-and-fining coarse clastic wedges. The older and higher terrace suite, comprising the Arries Drift Gravel Formation, is early mid-Miocene in age (19-17 Ma) and is referred to locally as the Proto-Orange terrace. The lower and younger terrace suite, locally known as the Meso-Orange terrace, has not yet been dated but is considered to be Plio-Pleistocene in age. Middle to Late Proterozoic bedrock, underlying both suites of terrace deposits, was deeply eroded during the incision of the Orange River which began in Late Cretaceous times. The Proto-Orange terraces, richer in diamonds than the younger Mesozoic deposits, form low-grade but large-average, diamond-size, gem quality placers which are mainly discussed here.

Diamonds are best concentrated in cobble-boulder basal gravels on or close to the rough bedrock contact in three main trapsite settings: scour pools hosting oversize (obstacle) clasts return grades of 10 to 40 carats per hundred ton (cpht); push bars, where the grades range from 10 to 55 cpht; and bedrock highs with their associated oversize clasts, yielding grades of 6 to 12 cpht, do not concentrate diamonds to the same extent as the two former trapsites, but occur over a wider area and thus potentially enrich a greater volume of gravel. The average diamond (stone) size in these three trapsites is also upgraded, ranging from 1.5 to 2.8 carats per stone (ct/stn). The key factors in concentrating diamonds at the bedrock-gravel interface are turbulence scale and intensity created by the rough boundary conditions and the presence of fixed bedrock sites of turbulence, which are stable enough to initiate and retain around them the slow and stable growth of gravel that hosts the concentrated diamonds.

The upper, more mobile gravels have far lower concentrations of diamond than the basal gravels, with grades dropping to an average of 0.3 cpht and stone size decreasing to 1 to 1.3 ct/stn. Sites of higher concentration in the upper gravels of the Proto-terraces are bar platforms/riffles (1 to 3 cpht and average stone sizes of 1.1 to 1.3 ct/stn) and bar heads (0.7 to 1.5 cpht and average stone sizes of 1 to 1.3 ct/stn) while bar tails are practically barren.

Although the vertical decline in diamond concentration may be attributed to a decrease in diamond supply during the aggradation of the Proto-Orange gravels, the younger, lower grade Mesozoic gravels also display a similar vertical profile. Hence, diamond concentration in the lower Orange River deposits is influenced positively by the occurrence and nature of the localized fixed bedrock trapsites. From an exploration point of view, these fixed trapsites are more difficult to locate, often requiring extensive drilling, whereas the areally extensive but low-grade upper gravels are relatively easy to delineate. During exploitation, the localized but higher grade fixed trapsites provide a reliable source of above average-sized diamonds in an otherwise low-grade, high-tonnage risky mining operation.