You are on page 1of 27

ELSEVIER

Sedimentary Geology 124 (1999) 329

Processes controlling the composition of heavy mineral assemblages in sandstones


Andrew C. Morton a,b, , Claire R. Hallsworth a
b

British Geological Survey, Keyworth, Nottingham NG12 5GG, UK Department of Geology and Petroleum Geology, University of Aberdeen, Kings Buildings, Aberdeen AB9 2UE, UK Received 1 October 1997; accepted 14 July 1998

Abstract Sandstone compositions result from a complex interplay between provenance and factors that operate during the sedimentation cycle. Accurate identication and discrimination of provenance depends on isolating provenance-sensitive features, and avoiding parameters that are inuenced by other factors. Heavy mineral analysis offers a high-resolution approach to determination of sandstone provenance, because of the diversity of mineral species found in sandstones and because the factors affecting assemblages have been comprehensively evaluated. This paper presents the current understanding of the effects of processes operative during the sedimentation cycle. The original provenance signal may be overprinted by weathering at source prior to incorporation in the transport system; by mechanical breakdown during transport; by weathering during periods of alluvial storage on the oodplain; by hydraulic processes during transport and nal deposition; by diagenesis during deep burial; and by weathering at outcrop. The most inuential of these processes are hydraulics, which fractionates the relative abundance of minerals with different hydraulic behaviour, and burial diagenesis, which reduces mineral diversity through progressive dissolution of unstable mineral species. There is also evidence that weathering during alluvial storage plays a signicant role. Two alternative, complementary approaches are recommended to identify provenance from heavy mineral data. The relative abundances of minerals with similar hydraulic and diagenetic behaviour are largely unaffected by processes operative during the sedimentation cycle, and utilize information gained from the entire heavy mineral suite. Determination of such ratios can be augmented by acquisition of varietal data, concentrating on the varieties shown by mineral types within the assemblage. A number of different varietal techniques are recommended, including optical differentiation of types based on colour, habit and internal structure, single-grain geochemical analysis, and single-grain geochronology. 1999 Elsevier Science B.V. All rights reserved. Keywords: heavy minerals; provenance; weathering; transport; diagenesis

1. Introduction Heavy mineral analysis (HMA) is one of the most sensitive and widely-used techniques in the determination of sandstone provenance. In common with in Corresponding

author.

formation such as whole-rock petrography, feldspar compositions and quartz types, heavy mineral data provide constraints on the mineralogical nature of the source terrains. One of the main strengths of HMA is that a wide variety of detrital heavy minerals has been found in sandstones. For example, over 50 translucent detrital minerals were described

0037-0738/99/$ see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 7 - 0 7 3 8 ( 9 8 ) 0 0 1 1 8 - 3

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

by Mange and Maurer (1992), and in addition there are several opaque species of common occurrence. Many of these minerals have very specic and restricted parageneses, and therefore provide crucial provenance information that cannot be acquired by any other means. Further renement to provenance evaluation can be provided by detailed analysis of individual mineral species (varietal studies). This can be achieved by acquiring geochemical data, for example using the electron microprobe (Morton, 1991), because mineral compositions are commonly diagnostic of the physical and chemical conditions at the time of crystallization. Alternatively, crystal morphology and internal structure of mineral grains (such as zircon) are indicators of petrogenesis, and can therefore be used in provenance studies (Lihou and Mange-Rajetzky, 1996). However, the composition of heavy mineral assemblages is not controlled only by the mineralogical composition of the source region. The original provenance signal is modied by several other processes that operate during the sedimentation cycle (Fig. 1). As a result, heavy mineral data from sandstones do not solely reect the composition of the parent rocks. Indeed, in some cases they bear very

little resemblance to the source rock. The overprinting processes can cause an originally homogeneous sandstone body, derived from a single source, to now contain markedly heterogeneous heavy mineral assemblages. Alternatively, sandstones from different sources, originally having had different suites, may now possess supercially similar assemblages. Therefore, the uncritical use of heavy mineral assemblages to characterize and differentiate provenance may yield erroneous conclusions. For the purposes of this paper, the term provenance is used in its most restricted sense, to represent the mineralogical composition of the sediment source region, and is not intended to encompass other source area attributes such as climate or relief. Heavy mineral assemblages are affected by three processes, physical sorting, mechanical abrasion, and dissolution. Physical sorting takes place as a result of the hydrodynamic conditions operative during the transport and depositional stages, and controls both absolute and relative abundances of heavy minerals. Mechanical abrasion takes place during transport, and causes grains to diminish in size, by a combination of fracturing and rounding. Dissolution causes partial or complete loss of heavy minerals in a va-

Fig. 1. Schematic diagram showing processes controlling heavy mineral assemblages in sandstones, modied from Morton and Hallsworth (1994).

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

riety of geochemical conditions at several stages in the sedimentation cycle, from weathering at source, subaerial exposure in non-marine depositional settings, during burial, and nally during weathering at outcrop. The clearest indication of the action of dissolution processes is the presence of corrosion textures on heavy mineral grain surfaces, as rst discussed by Edelman and Doeglas (1932, 1934). Corrosion textures include etch pits, surface mamillae, facets, ragged edges, hacksaw terminations and skeletal structures. Variations in corrosion textures are related to differences in mineral composition, crystallography and possibly the nature of the corrosive uid. Caution should be taken in interpreting all such textures as indicative of ongoing dissolution, as they can be inherited (Hubert, 1971). For example, etched staurolite and garnet co-exist with unetched epidote in Palaeozoic sandstones of the Clair Field, west of Shetland, UK (Allen and Mange-Rajetzky, 1992). Such instances provide evidence for sediment recycling. The corrosion textures observed on natural grain surfaces have been reproduced by acid leaching in laboratory experiments. Natural hacksaw terminations on clinopyroxene and amphibole have been simulated by Berner et al. (1980), as have faceted surfaces on garnet (Hansley, 1987). Some authors have argued that features generally attributed to corrosion are in fact overgrowths, the facets on garnet grains being the subject of greatest controversy. This question was considered in detail by Morton et al. (1989), who concluded, on the basis of thermodynamic considerations, experimental studies, distribution of faceted garnets in the subsurface, textural relationships and compositional variations within faceted grains, that the overgrowth hypothesis cannot be sustained. The processes of physical sorting, mechanical abrasion, and dissolution potentially affect not only the heavy mineral assemblages, but all aspects of the sediment composition, including whole-rock mineralogical and geochemical attributes, feldspar compositions and isotopic data. It is therefore important that their effects are evaluated when using any such techniques to determine sediment provenance. Because HMA is a long-established technique, there has been extensive evaluation of the effects of these processes (Hubert, 1971; Mange and Maurer, 1992;

Morton, 1985), but this cannot be said of several of the alternative analytical approaches. The purpose of this paper is to review the present understanding of the effects of processes operative during the sedimentation cycle, and to highlight areas that require further research. Physical sorting, mechanical abrasion, and dissolution operate at several stages during the sedimentary cycle. Weathering of the parent rocks causes modication of source rock mineralogy prior to incorporation into the transport system through selective dissolution. During transport, minerals may be selectively lost through mechanical abrasion and through weathering during storage periods prior to, or at, the ultimate site of deposition. Hydraulic processes during transport and at the time of deposition strongly affect relative abundances of minerals with different hydraulic behaviour. Burial diagenesis selectively removes unstable minerals as a result of changing physico-chemical conditions. Finally, weathering at outcrop, following uplift and exposure to present-day or ancient subaerial processes, may cause further depletion of unstable minerals.

2. Weathering of the parent rocks Although weathering in the source area alters the provenance information prior to sediment entering the transport system, this process is commonly considered to be an inherent feature of the provenance, rather than as an overprinting factor. Despite this, it is important that weathering effects are evaluated in order to relate sediment composition to source. The degree to which heavy mineral assemblages are modied by weathering in the source area depends on three main factors: the composition of the heavy mineral suite (controlled by the parent rock lithology), the physiographic setting, and the climate. The latter two control the extent to which soil formation occurs. If the transport processes removing weathered material from an area are potentially more rapid than the weathering processes generating the material, erosion is said to be weathering-limited (Johnsson et al., 1991). Under such conditions, detritus is quickly removed without signicant modication by chemical weathering, and thus the provenance signature of the parent rocks is likely to be

Table 1 Relative stability of translucent heavy minerals in weathering proles Gneiss, Dolerite, Amphibolite Northern USA Goldich (1938) Crystalline schist Granite Tertiary kaolinitic sand Bavarian molasse Calcareous sandstones Germany Lemcke et al. (1953) Zircon, Rutile, Tourmaline, Staurolite Aeolian coversands

Northern USA Dryden and Dryden (1946) Zircon

Germany Piller (1951) Zircon

Germany Germany Weyl and Werner (1951) Grimm (1973) Zircon, Rutile

England Bateman and Catt (1985) Zircon, Rutile, Tourmaline, Andalusite, Kyanite, Staurolite, Titanite

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

Tourmaline Sillimanite Monazite Kyanite Calcic amphibole Staurolite Staurolite Epidote Kyanite

Tourmaline, Andalusite, Kyanite

Kyanite, Epidote Calcic amphibole Staurolite Epidote Apatite

Garnet Calcic amphibole

Garnet

Garnet

Garnet Calcic amphibole

Garnet, Apatite Calcic amphibole

Garnet

Garnet

Epidote Clinopyroxene Orthopyroxene Clinopyroxene, Olivine Apatite Olivine Most stable minerals are at the top of the table, least stable at the foot. Clinopyroxene, Orthopyroxene, Calcic amphibole Apatite

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

largely transferred into the transport system. If, in contrast, the maximum weathering rate exceeds the ability of transport processes to remove material, then erosion is transport-limited (Johnsson et al., 1991). Under transport-limited conditions, weathering products have a longer time to react with soil and groundwaters, strongly affecting the composition of detritus. In this situation, extensive modication of the heavy mineral suite could occur before the sediment reaches the transport system. There have been a number of published studies on the effects of present-day soil-forming processes on heavy mineral suites. These deal with a variety of weathered materials, but the coverage of climatic regime is relatively restricted, with virtually all of the case studies coming from acidic podzol-type soils from northern Europe and the northern part of the USA. In general, the diversity of detrital heavy minerals in weathering proles decreases from bottom (least weathered) to top (most weathered), due to dissolution or transformation of the less stable minerals. A comparison of the relative order of stability of heavy minerals in these proles (Table 1) shows a reasonable degree of consistency, with zircon, rutile, tourmaline, kyanite, andalusite and sillimanite invariably showing high stability, and apatite, olivine and the pyroxenes all showing very low stability.

The intermediate group (garnet, calcic amphibole, staurolite and epidote) tend to vary in their relative stability between the various proles. One possible explanation is that the stability variations result from compositional variations of amphibole, epidote and garnet between the different proles. Another possibility is that there were variations in the composition and concentration of the groundwaters responsible for the leaching (Mitchell, 1975). There is, therefore, no correct absolute order of stability in weathering processes, as variations in the macro- and microenvironments have a signicant control on relative mineral stability. Nevertheless, an approximate order of stability can be formulated, giving an overall framework within which minor relative stability variations can be expected (Table 2). Nickel (1973), in his experimental work on the stability of heavy minerals, showed that minerals have markedly different relative stabilities in different pH conditions. The order of stability determined from observations of soil proles bears a strong resemblance to those seen at pH values of 3.6 and 5.6, but strongly contrasts with that at pH of 10.6 (Table 2). This is predictable given the acidic conditions associated with podzol formation. For example, pH in the proles described by Bateman and Catt (1985) varies between 3.3 and 4.7. There are some minor

Table 2 Relative stability of heavy minerals in podzol-type soil proles, adapted from Bateman and Catt (1985) compared with experimental stability patterns determined by Nickel (1973) Podzolic soil proles Rutile, Anatase, Brookite Zircon Tourmaline Sillimanite Andalusite Kyanite Staurolite Topaz Titanite Monazite Garnet Epidote Calcic amphibole Orthopyroxene Clinopyroxene Olivine Apatite pH 3.6 Zircon, Rutile Calcic amphibole Kyanite Staurolite Tourmaline Epidote Garnet Kyanite Staurolite, Tourmaline Epidote Calcic amphibole Garnet Epidote pH 5.6 Zircon, Rutile pH 10.6 Zircon, Rutile Garnet, Staurolite Calcic amphibole Apatite Kyanite Tourmaline

Apatite

Apatite

Most stable minerals are at the top of the table, least stable at the foot.

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

differences, most notable being the high stability of calcic amphibole in the experiments run at a pH of 3.6. This may result from a difference in the composition of the amphibole concerned, but is more likely to indicate that natural weathering processes cannot be precisely simulated by the use of HCl in the laboratory. Despite this, the observed order of stability is compatible with the experimental work, and can be used to predict which minerals are likely to disappear, or at least become relatively depleted, in the period leading up to incorporation in the transport system. Despite the clear evidence for loss of heavy minerals within soil proles, there is virtually no documented evidence that shows signicant reduction in overall heavy mineral diversity between source rock and transport system. Recent sediments derived from areas subjected to intense tropical weathering, such as Puerto Rico and the Orinoco basin, contain abundant unstable minerals such as pyroxene and amphibole (Walker, 1974). The fact that apatite, considered by Bateman and Catt (1985) to be the most unstable mineral in soil proles, is virtually ubiquitous in sands and sandstones, provides further support for this contention. Nevertheless, there is evidence that the relative proportions of stable to unstable minerals are affected by the intensity of source-area weathering. Allen (1948) compared Eocene and Cretaceous sandstones in the Great Valley of California (USA), both derived from basement rocks of the Sierra Nevada complex that include granitoid bodies rich in calcic amphibole. The Cretaceous sandstones contain abundant amphibole, together with garnet, titanite and other minerals, whereas the Eocene largely contains zircon, tourmaline, epidote, andalusite, sillimanite and rutile, with only minor amphibole and garnet. This was attributed by Allen (1948) to a change in climate between the Cretaceous and the Eocene, with a tropical regime operating in the Eocene. Thus, although a number of heavy minerals are unstable during soil-forming processes, studies of both present-day and ancient examples indicate that source-area weathering does not signicantly affect the diversity of heavy mineral suites prior to incorporation of sediment into the transport system. Despite the general lack of change in mineral diversity, however, it is likely that the proportion of

stable to unstable minerals in the transport system will be higher than in the parent rock, and that the degree of enrichment of stable minerals will be greatest in transport-limited erosional regimes, notably in deeply-weathered areas with low relief. In circumstances such as this, the relative importance of parent rocks largely composed of unstable minerals (basalt, for example, which would yield mainly olivine, pyroxene and opaque phases) could be signicantly under-represented in the heavy mineral suite.

3. Mechanical abrasion during transport The effects of abrasion on grains during transport has often been considered as a possible mechanism to account for decreases in mineral diversity (Mackie, 1923; van Andel, 1959). There have been a number of experimental studies which have determined the relative mechanical stability of heavy minerals, the most comprehensive being those of Freise (1931), Thiel (1940, 1945) and Dietz (1973). The mechanical stability patterns of Freise (1931) and Thiel (1940, 1945) are in excellent agreement, with only rutile showing a discrepancy in position on the two stability sequences (Table 3). The pattern determined by Dietz (1973) is markedly different, with zircon and kyanite being much more resistant and staurolite and tourmaline being much less resistant compared with the earlier studies. This is probably because Dietz determined shape and roundness as a guide to resistance to abrasion, whereas Freise and Thiel also considered weight loss during the experiments. The anomalous position of kyanite in Dietzs stability pattern, for example, probably arises from its well-developed cleavage, which causes it to break rather than become rounded during transport. Although there has been a considerable effort into determining the mechanical stability of heavy mineral grains, there is no evidence from natural studies that heavy minerals disappear from assemblages through this process. Studies of major river systems with extremely long transport distances, such as the Mississippi (Russell, 1937), the Nile (Shukri, 1949) and the Rhine (van Andel, 1950) show no evidence for any decrease in mineral diversity downstream. A recent study of the Cascade River and associated beach sediments in southern Westland

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329 Table 3 Mechanical stability of heavy minerals, from Freise (1931), Thiel (1945) and Dietz (1973) Freise (1931) Tourmaline Corundum Chrome spinel Spinel Rutile Staurolite Augite Thiel (1945) Tourmaline Dietz (1973)

4. Weathering in non-marine depositional environments Heavy mineral assemblages can be fundamentally altered by weathering in non-marine depositional environments, due to dissolution by circulating groundwaters. The degree and pattern of dissolution depends on the nature of the groundwater at the depositional site, and is thus facies-dependent. Broadly speaking, there are two different sets of conditions, one that operates in well-vegetated humid settings, and another related to conditions conducive to the formation of red beds. 4.1. Well-vegetated humid environments These conditions, whether under tropical or temperate climates, lead to the generation of acidic groundwaters through bacterial degradation of organic matter (Burley et al., 1985). These acidic groundwaters cause extensive dissolution of unstable heavy minerals, as they do during soil formation in the source terrain. The reaction of heavy mineral suites during such circumstances has been extensively studied in late Tertiary uvial sandstones of Denmark (Friis and Johannesen, 1974; Friis, 1976; Friis et al., 1980). In these sandstones, extensive mineral dissolution has occurred at the top of sedimentary units relating to downward percolation of acidic groundwaters generated during periods of exposure and plant colonization. Minerals gradually appear with depth below the exposure level, the order in which they appear being governed by their stability (Fig. 2). The minerals also show strong evidence of corrosion, with etched grain surfaces. The stability patterns observed in the Danish sandstones are closely comparable with the generalized stability pattern found in acid soil proles, although the one of the key minerals, apatite, was not recorded in the Danish sediments. Studies of heavy mineral distribution in Jurassic sandstones of the North Sea (Morton, 1986) showed that apatite is abundant in marine sandstones, but tends to be depleted in sandstones deposited in marginal marine environments and is commonly absent entirely from uvial and delta-top sandstones. Thus, apatite is a useful key mineral to detect the inuence of acidic groundwaters, especially as it is stable in deep burial conditions. Morton and Humphreys (1983) noted etch-

Staurolite Zircon Kyanite Titanite

Topaz Garnet Epidote Apatite Zircon

Garnet Epidote Zircon Calcic amphibole Rutile

Garnet

Rutile Staurolite Tourmaline

Kyanite Olivine Andalusite Diopside Monazite

Orthopyroxene Apatite Augite Kyanite Olivine Andalusite

Highly resistant minerals are at the top of the table, least resistant at the foot.

(New Zealand) showed that there was no loss of heavy mineral diversity through abrasion in transport (Morton and Smale, 1991), even though one of the major minerals in the river is olivine, experimentally determined as one of the least mechanically-stable minerals. Furthermore, monazite, determined experimentally as the least mechanically-stable mineral of all, is commonly found as a placer mineral in beach deposits, even though beach sediments probably suffer the greatest actual degree of mechanical abrasion of all environments, with the possible exception of aeolian dune sands (van Andel, 1959). The effects of abrasion on heavy mineral proportions during transport therefore appear to be be generally negligible, but it is possible that abrasion may have a signicant effect in exceptional circumstances.

10

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

Fig. 2. Variation in abundance of heavy minerals below a Miocene subaerial exposure surface, marked by a brown coal horizon, in the Lavsberg borehole, Denmark (from Friis et al., 1980). Ky D kyanite, Gt D garnet, Ep D epidote, Am D Amphibole, ZRT D zircon C rutile C tourmaline.

ing on the generally stable phases chloritoid, chrome spinel and gahnite (zinc spinel) in zones of intense apatite and garnet leaching in Middle Jurassic sandstones of the North Sea. It appears, therefore, that these minerals can also be dissolved under such circumstances. In association with the intense leaching, some zircons develop outgrowths, believed to consist of xenotime (Mange and Maurer, 1992). These have also been noted in uvial sandstones from the Carboniferous Millstone Grit and Jurassic sandstones of Yorkshire (England) by Buttereld (1936) and Smithson (1941), and may be further indicators of intense acidic groundwater leaching, although the possibility that they may occur during deep burial should also be considered. 4.2. Red-bed forming environments Red beds, in which iron is in the ferric state, may form in either arid, desert conditions or in seasonal

climates. The stability of heavy minerals in desert environments has been studied by Walker (1967) and Walker et al. (1967, 1978) in both modern and late Cenozoic sequences. These studies showed that extensive dissolution of ferromagnesian minerals, notably orthopyroxene, clinopyroxene and calcic amphibole, has taken place. The iron taken into solution through decomposition of these minerals has been oxidized to Fe3C , imparting the typical reddish coloration to the sediments. Thus, heavy mineral dissolution is an important process in the formation of red-bed sequences. However, there have not been any detailed studies of the relative stability of the other heavy minerals in red-bed forming conditions, either under arid or seasonal conditions, and the behaviour of minerals such as apatite, garnet, staurolite and zircon is therefore not known. Apatite, rutile, tourmaline and zircon are generally abundant in the ancient red-bed sediments of NW Europe, and therefore appear to be mainly stable under such circumstances.

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

11

However, zones of apatite depletion are found locally within such deposits (Jones et al., in press), suggesting occasional periods of acidic groundwater percolation. The presence of such zones within red bed sequences may indicate a seasonal climate rather than arid conditions. 4.3. Alluvial storage Heavy mineral assemblages may be affected by non-marine weathering not only at the ultimate depositional site, but also during temporary alluvial storage, for example on river oodplains or in coastal dunes. Friis (1978) identied this process as a probable cause of variations in heavy mineral assemblages in Miocene marine sandstones of Denmark. Weathering during alluvial storage is considered to have a strong inuence on the gross mineralogical composition of sands, particularly in tropical areas such as the Orinoco basin (Johnsson et al., 1988; Savage and Potter, 1991), where tropical weathering in the low-lying Llanos area alters rst-cycle feldspathic-lithic arenites to quartz arenites, by destroying the unstable lithic and feldspar grains. Morton and Johnsson (1993) evaluated the effects of this process on heavy mineral suites in the Apure River system, which ows across the Llanos (Table 4), concluding

that the ratio of apatite to tourmaline is signicantly reduced during alluvial storage. The effect on other parameters is less well-dened, although abundances of garnet and clinopyroxene both tend to decline with increased alluvial storage. The Llanos case study indicates that heavy mineral suites are affected by weathering during short term alluvial storage in uvial systems. The study has shortcomings, mainly because of the heterogeneity in composition of assemblages entering the transport system. This heterogeneity is the principal cause of the high standard deviations associated with the mineralogical parameters in Table 4. The ndings need to be substantiated by further similar studies, ideally designed to minimalize compositional variations of the starting material. Nevertheless, the ratio of apatite to tourmaline appears to be a useful guide to the extent of weathering during alluvial storage, but in consequence may not always be an accurate guide to provenance. In many circumstances, notably in marine sediments and in uvial and aeolian sediments deposited in arid and semi-arid conditions, the apatitetourmaline ratio is likely to be effective in discriminating sediments of differing provenance. However, in uvio-deltaic sandstones deposited in humid tropical settings, variations in this ratio are likely to be at least in part a function of weather-

Table 4 Mean heavy mineral abundances in sands from different geographic and physiographic regions of the Apure River basin, from Morton and Johnsson (1993) Within orogenic belt (region 1) Calcic amphibole Clinopyroxene Chloritoid Epidote Garnet Rutile Sillimanite Staurolite Titanite Zircon Number of samples ZTR index Apatitetourmaline index (ATi) 19 9 11 34 32 10 11 5 34 22 35 11 17 11 25 26 14 57 13 Proximal Llanos (region 2) 25 7 53 45 28 8 73 53 42 22 21 20 10 25 24 12 51 11 Middle Llanos (region 3) 23 6 11 12 32 6 73 31 41 11 21 17 9 10 23 11 43 17 Distal Llanos (region 4) 19 8 0.1 0.3 12 30 22 64 44 31 21 33 28 29 5 35 32 20 17

ZTR (zircontourmalinerutile) index calculated as dened by Hubert (1962). ATi (apatitetourmaline index) is dened as 100 apatite=(apatite plus tourmaline). Uncertainties represent 95% condence limits as determined by Students t distribution. Uncertainties of ratio data are not statistically rigorous.

12

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

ing during alluvial storage, and should be used to indicate changes in provenance only if supported by other data. Because the apatitetourmaline ratio is affected by weathering on the oodplain, it may be possible to relate variations in this parameter to the extent of oodplain storage. The Upper Jurassic Piper Formation in the Tartan Field (Outer Moray Firth, North Sea) contains packages of sandstones with differing apatitetourmaline ratios (Fig. 3). Units with consistently high apatitetourmaline ratios are interbedded with units characterized by low or variable ratios. Other mineralogical parameters, including the monazitezircon ratio, remain relatively unchanged throughout (Fig. 3), suggesting no signicant change in provenance. Consequently, an alternative explanation for the variations in apatitetourmaline ratio is required. In light of the relationship between apatite tourmaline ratio and extent of oodplain storage, one possible explanation is that sandstones with high ratios were transported to the depositional site without signicant oodplain modication, whereas those with low or variable values were stored for relatively long periods. This is given some credence by sedimentological studies (Kadolsky, pers. comm.), which indicate that the units with high apatite tourmaline ratios represent either highstand or transgressive systems tracts, whereas units with low or variable apatitetourmaline ratios represent aggradational systems tracts (Fig. 3). It appears that high sea level during the highstand tends to reduce the potential area for sediment storage on the oodplain, thus helping to maintain high apatitetourmaline ratios. During the subsequent aggradational phase, sediment is interpreted to spend longer periods on the oodplain, promoting apatite dissolution and thereby reducing the apatitetourmaline ratio. Friis (1978), Larsen (1966) and Carver and Kaplan (1976) have also noted that sandstones deposited during regressive phases have different heavy mineral assemblages from those deposited during transgression. The possible relationship between mineralogical parameters (such as the ATi value) and extent of alluvial storage, which itself is linked to the type of depositional sequence, is clearly an area that requires further research.

5. Hydraulic processes 5.1. The hydraulic equivalence concept Because heavy minerals are denser than quartz and feldspar grains, they behave differently during transport and deposition. Thus, the size distribution of heavy minerals in sandstones differs from that of the light minerals, with the denser heavy minerals having ner mean grain sizes than the associated quartz and feldspar. Detailed attention to this phenomenon was rst drawn by Rubey (1933), who introduced the concept of hydraulic equivalence by showing that grains with the same settling velocity are deposited together. Rubey proposed that settling velocity is a function of grain size and grain density, so that smaller, denser minerals are deposited with larger, less dense ones, and introduced the term hydraulic equivalent size, dened as the difference in size between a given heavy mineral and the size of a quartz sphere with the same settling velocity in water. A number of detailed studies have shown that there is often a good relationship between density and hydraulic equivalent size, with a strong correlation between the two parameters (Fig. 4). Although the concept of hydraulic equivalence is useful, it has been subsequently shown that it does not always adequately explain the behaviour of heavy minerals during transport and deposition. Occurrences where heavy minerals are not in precise hydraulic equilibrium with the associated light minerals have been described frequently in the literature. For example, van Andel (1950) showed that garnet in Rhine sediments is anomalously coarse, being only slightly ner-grained than quartz. The presence of heavy mineral laminae in sandstones cannot be explained in terms of the hydraulic equivalence concept (Reid and Frostick, 1985), because in many such laminae, the heavy mineral grains are too large to be in settling equivalence with the neighbouring quartz grains. Since many factors control the selective entrainment and deposition of grains with different densities, Reid and Frostick (1994) suggested that the term hydraulic equivalence should be replaced by settling equivalence. Mechanisms in a given transporting or depositional regime determine which heavy minerals are carried along and which are deposited. During low-

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

13

Fig. 3. Variations in apatitetourmaline index (ATi) in Upper Jurassic sandstones of well 15=6-9 and -10 (Tartan Field, Outer Moray Firth, UK). Note the relationship between ATi value and sequence stratigraphy (from Kadolsky, pers. comm.). UJ6, UJ7 etc. are Texaco in-house sequence stratigraphic units. Main Sand, Six Sand etc. are lithostratigraphic units identied in the Tartan Field. HST D highstand systems tract, TST D transgressive systems tract, AST D aggradational systems tract.

energy transport stages, grains tend to move by sliding or rolling, but saltation and suspension dominate in high-energy stages (Allen, 1994). Heavy minerals respond sensitively to changes in these conditions

because of their different densities. Consequently, the heavy mineral composition of sand in an active channel may differ considerably from that deposited on the oodplain of the same river.

14

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

Fig. 4. Relationship between hydraulic equivalent size and mineral density in two case studies. Rio Grande river data from Rittenhouse (1943), New Jersey beach sand data from McMaster (1954). Correlation coefcient between hydraulic equivalent size and density is 0.91 for the Rio Grande and 0.87 for the New Jersey beaches. Hydraulic equivalent size is dened as the difference in size between a given heavy mineral and the size of a quartz sphere with the same settling velocity in water (Rubey, 1933).

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

15

The hydraulic behaviour of heavy minerals is therefore a complex topic, with many factors involved. Settling is governed principally by grain size and density, although grain shape is another potentially signicant factor. The transport and deposition of heavy minerals also depends on how grains are entrained in the ow, on the nature of the transporting medium (water-borne or air-borne), and on the mode of dispersal within the transporting medium. Ultimately, the grain size distribution in the parent rocks controls the grain size availability in the sediment. The behaviour of heavy mineral grains during transport is therefore affected by a combination of processes. Fletcher et al. (1992) suggested that the term transport equivalence be used to describe the behaviour of particles during the interplay of entrainment, motion and deposition. 5.2. Grain shape Density is not the only control on settling velocity and thus on hydraulic=settling equivalence. The original denition of hydraulic equivalence by Rubey (1933) assumes that the heavy mineral grains are spherical. One possible explanation of the deviation between theoretical and observed mean grain size is that many mineral grains are not spherical. The deviation away from the sphere is extreme in some cases, notably amphibole and kyanite. Flores and Shideler (1978) showed that deposition of heavy minerals on the outer continental shelf of south Texas was in part controlled by winnowing of bladed species such as amphibole, while the more equant grains such as epidote remained at the site of original deposition. Experimental work by Briggs et al. (1962) showed that grain shape can be as important as density in controlling the hydraulic behaviour of heavy minerals. Mica shows the most extreme manifestation of the effect of the shape factor on hydraulic equivalence. Although mica is a heavy mineral, with a density around 2.8 to 3.0 gm cm 3 , sand-size mica particles are hydraulically equivalent to silt-sized light minerals (Doyle et al., 1983). This highly anomalous hydraulic behaviour results from the platy habit of detrital mica. 5.3. Entrainment equivalence In many cases, the relative availability of heavy minerals is dependent on transport distance (van An-

del, 1950; Lowright et al., 1972). This is because heavy minerals, once deposited during transport, are more difcult to re-entrain in the transporting medium than quartz (Hand, 1967; Slingerland, 1977, 1984: Komar and Wang, 1984). Thus, heavy minerals tend to remain at more proximal depositional sites compared with the light minerals. Entrainment equivalence is the expression describing the relative ease of a heavy mineral grain to become entrained in the transporting medium relative to quartz. The difculty of entraining heavy minerals is further exaggerated because the heavy minerals, being relatively small through their hydraulic behaviour, lie in interstices between larger light mineral particles (Reid and Frostick, 1985). The entrainment potential of heavy mineral grains is considered to be the principal factor in generating placer deposits, both in uvial and shallow marine settings, and the understanding of this process is crucial in exploration of economic placer deposits. In order to determine the limits of downstream heavy mineral distribution trends, Hattingh (1992) conducted an experiment in an ephemeral river in South Africa by placing high concentration of heavy minerals on the dry river bed. The investigation revealed that after a ood event, dense plant growth in swampy area ltered out heavy minerals which were still in suspension. This demonstrates that vegetation is important in removing some of the heavy minerals from the transporting system. Hattingh (1992) also noted that heavy mineral concentrations occur on point bar faces, formed by heavy minerals inltrating open-work gravel pores up to 5 cm deep, thereby forming miniature placers. 5.4. Dispersive equivalence In grain ow deposits, in which deposition takes place from a highly concentrated ow of sand-sized and larger particles, grains are supported by dispersive pressures resulting from grain collision (Bagnold, 1954). This causes larger (or heavier) particles to rise, generating an inversely-graded deposit. Sallenger (1979) observed dispersive equivalence in subaerial grain ow deposits and in inversely-graded beach foreshore sediments, and argued that under such circumstances, heavy mineral grains are dispersively-equivalent to light minerals, rather than

16

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

hydraulically-equivalent to them. Reid and Frostick (1985) consider that although the process could generate layers rich in heavy minerals near the top of grain ows, it is unlikely that placer deposits could form directly in this way. However, they comment that these layers could subsequently be enriched in heavy minerals by entrainment processes. 5.5. Grain size distribution in the parent rocks The grain size distribution of heavy minerals inherited from the parent rocks by the transported sediment clearly has an important role. If a certain heavy mineral is available only in the very ne grain size fraction, it will be absent from the coarser fraction even if the hydraulic equivalence concept requires it to be there. The effect of initial grain size on the spatial distribution and preferential concentration of heavy minerals is exhibited by the dominantly ophiolite-derived coastal and shelf sediments of the northeastern Mediterranean, off southern Turkey (Mange-Rajetzky, 1983). Due to their coarse grain size in the parent rocks, pyroxenes and amphiboles dominate the coarse .0:6 1:8 / frac-

tion of the coastal sediments, and these minerals also show a peak in the coarser .1:4 2:6 / fraction of the ner grained and better sorted shelf sediments. In longshore transport the shape-controlled concentration of particular minerals, especially amphiboles, was also detected in this study. The role of shape and size on the offshore dispersal of the principal heavy mineral components was revealed by mapping their abundance in the surcial sediments: coarser and denser pyroxenes are most abundant at sediment entry points such as river mouths. Diallage (a pyroxene characterized by excellent parting) is susceptible to mechanical breakdown, and its ne lamellae are carried far offshore by currents, resulting in a concentration of diallage close to the shelf edge. The importance of inherited grain size distributions is demonstrated in a study of river and beach sediments in the Baixada de Jacarepagua, Rio de Janeiro (Brazil). The river sands represent rst-cycle detritus derived from Precambrian metamorphic and igneous basement (Savage et al., 1988). There is no correlation between mean grain sizes and densities of the heavy minerals in the river sands (Fig. 5), which therefore have highly anomalous hydraulic equiva-

Fig. 5. Mean heavy mineral grain sizes in river and beach sediments from the Baixada de Jacarepagua (Rio de Janeiro, Brazil), showing that mean heavy mineral size in the river sands is not controlled by density, in contrast to that in the beach.

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

17

lence relationships. In contrast, the heavy minerals in the associated beach sediments display almost ideal hydraulic equivalence behaviour, with mean grain sizes arranged in order of their density (Fig. 5). This is believed to be the result of the long transport distances that the beach sands have travelled, which have allowed the heavy minerals to reach hydraulic equilibrium with the associated light phases.

6.3. Poreuid circulation The recognition that zones protected from poreuid circulation (through cementation, original low porosity, restricted uid connection and inltration of hydrocarbons) contain more diverse assemblages than the surrounding friable sandstones, indicates that dissolution rate is governed by the rate of poreuid movement through the rock. 6.4. Geological time The length of time that a sandstone has been subjected to poreuid movement at elevated temperatures clearly governs the degree of mineral depletion. For example, in the North Sea, it is generally true that Jurassic sandstones at similar depths to Paleocene sandstones are more depleted in heavy minerals. As sandstones are buried, poreuid temperature rises. In addition, other reactions take place to alter the composition of the porewaters, such as the transformation of smectite to illite (Hower et al., 1976) and the decarboxylation of organic matter to generate hydrocarbons (Tissot et al., 1974). The effects of the burial diagenesis on heavy mineral assemblages can be determined by studying sandstones across a range of burial depths. Ideally, the analysis should be conned to a unit that shows considerable variation in burial depth (thereby involving a range of poreuid temperatures), has a comparatively simple burial history (not involving any inversion phases or thrust faulting), has a uniform provenance (so that any variations in assemblage cannot be ascribed to differences in source), and was deposited under relatively uniform hydraulic conditions (precluding a marked hydraulic control on mineralogy). The best-constrained study of the effects of burial diagenesis on heavy mineral assemblages deals with the Late Paleocene to Early Eocene sandstones of the Dornoch and Forties Formations in the central North Sea (Morton, 1984). The Dornoch Formation represents a prograding coastal-deltaic complex which seismic data show to be derived from the northwest (Rochow, 1981). The Forties sandstones formed in a submarine fan complex at the foot of the delta slope, basinward of the area where the Dornoch sandstones were deposited. The top of the sequence shows a variation in burial depth from less

6. Burial diagenesis The reaction of heavy mineral suites to deep burial has been the focus of much research in recent years, although there have not yet been enough case studies to condently predict mineral dissolution patterns under all circumstances. Essentially, heavy mineral suites react to burial diagenesis by dissolution of unstable phases (known as intrastratal solution) and growth of secondary minerals. Although these secondary minerals have no bearing on provenance, they provide indications of the diagenetic regime. Consequently, the composition and abundance of secondary phases should be recorded during analysis. Several factors govern heavy mineral dissolution during deep burial. These are the stability of the mineral, poreuid temperature and composition, poreuid circulation, and time. 6.1. Stability Some minerals are more stable than others. A mineral can be recognized as unstable on the basis of its surface textures and on its distribution in the subsurface. 6.2. Poreuid temperature and composition The relationship between burial depth and dissolution is a result of increasing poreuid temperature as a result of geothermal gradient, probably in combination with changes in poreuid chemistry caused by reactions such as the smectiteillite transformation and decarboxylation of organic matter.

18

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

Table 5 Order of persistence of heavy minerals in deeply-burial sandstones from the North Sea (Morton, 1984, 1986), US Gulf Coast (Milliken, 1988; Milliken and Mack, 1990), New Zealand (Smale and Morton, 1987) and Yugoslavia (Scavnicar, 1979) North Sea Apatite, Monazite, Spinel, TiO2 minerals, Tourmaline, Zircon Chloritoid Garnet Staurolite Kyanite Titanite Epidote Calcic amphibole Andalusite, Sillimanite Pyroxene Olivine US Gulf Coast Apatite, Chloritoid, Monazite, Spinel, TiO2 minerals, Tourmaline, Zircon Garnet New Zealand Apatite, Chloritoid, Spinel, TiO2 minerals, Tourmaline, Zircon Garnet Staurolite Titanite Epidote Kyanite Yugoslavia Apatite, Chloritoid, Garnet, Spinel, TiO2 minerals, Tourmaline, Zircon

Titanite Epidote Kyanite Calcic amphibole Pyroxene

Most stable minerals are at the top of the table.

than 200 m at the basin margin to nearly 3 km at the basin centre, and the thermal history is one of continued subsidence throughout the Tertiary. It thus forms an ideal case study of the effects of increased poreuid temperature on heavy mineral suites. The pattern of heavy mineral stability deduced primarily from the North Sea Paleocene, with additional data from younger Tertiary shallow marine boreholes and from the deeper Jurassic, is shown in Table 5. Suites become steadily less diverse as burial increases due to elimination of heavy mineral species through dissolution (Fig. 6), beginning with calcic amphibole, followed by epidote, titanite, kyanite and staurolite. Each phase becomes more corroded as burial depth increases. Garnet does not disappear entirely, but declines markedly in abundance relative to the stable mineral zircon. Concurrently, garnet compositions become less diverse due to preferential dissolution of the more calcium-rich garnets (Morton, 1987). In the underlying Jurassic, garnet disappears virtually completely below about 3500 m (Fig. 7). The remaining minerals (apatite, rutile and the associated TiO2 minerals, tourmaline and zircon) rarely display signs of etching, and are apparently stable (Morton, 1979). Indeed, as burial approaches 4 km, apatite tends to develop overgrowths. The provenance-sensitive minerals monazite and chrome spinel also appear to be generally stable, because they lack evidence for surface corrosion even at con-

siderable burial depths (Fig. 8). Chloritoid, another provenance-sensitive mineral, occurs frequently in garnet-depleted sandstones, and is therefore considered to be more stable than garnet. However, chloritoid frequently shows signs of corrosion in sandstones buried to between 3 and 4 km (Fig. 8), and its absence in very deeply-buried sandstones may therefore be due to intrastratal solution. There are some cases where chloritoid appears to be less stable than garnet. For example, chloritoid is present in calcitecemented sandstones (doggers) in the Lower Cretaceous sequence of the Britannia Field (North Sea), but is absent in the encasing uncemented sandstones (Mange, pers. comm.). The reasons for the variation in relative stability of garnet and chloritoid are unclear, and further documentation is needed before conclusions can be made. However, both minerals show wide compositional variations, and this could be one of the important factors inuencing their behaviour during burial. In rare instances, North Sea sandstones contain etched tourmaline grains (Fig. 8), but this is not a ubiquitous feature in deep burial. The process causing tourmaline etching has not been investigated in detail, but it appears to have taken place in sandstones that contain hydrocarbons with high H2 S contents. The possible relationship between H2 S and tourmaline etching is worthy of further study. Another fairly well-constrained study is that of Milliken (1988), who examined dissolution of heavy

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

19

Fig. 6. Burial depth distribution of heavy minerals in Upper Palaeocene sandstones of the central North Sea, showing the decrease in mineral diversity with increasing burial caused by dissolution of unstable minerals (Morton, 1984).

minerals in Plio-Pleistocene sediments of the US Gulf Coast. These have a single source, the Mississippi River, and because of extremely high sedimentation rates are buried between about 450 m and 5 km. This study was carried out on thin sections coupled with SEM examination to identify grain surface corrosion, and it is therefore possible that the distribution of low-abundance phases may not have been fully appreciated. Milliken (1988) found a pattern of heavy mineral loss similar to that observed in the North Sea, the only major difference being that kyanite is apparently less stable than epidote and titanite. The burial depths at which the heavy minerals dis-

appear in the Gulf Coast sequence are considerably greater than in the North Sea. Titanite, for example, persists to 34 km in the Gulf Coast, compared with just over 1 km in the North Sea. Oligocene sediments in the US Gulf Coast show a similar pattern (Milliken and Mack, 1990). Stable heavy mineral phases in the US Gulf Coast are considered to be apatite, chrome spinel, TiO2 minerals, monazite, tourmaline and zircon. The stability pattern inferred from the US Gulf Coast is shown in Table 5. The study of heavy mineral dissolution in the EoceneOligocene McKee sandstones of New Zealand (Smale and Morton, 1987) is a well-con-

20

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

Fig. 7. Relationship between garnet=zircon ratio and depth in Upper Jurassic sandstones of the central North Sea, showing that complete garnet dissolution is commonplace in sandstones buried in excess of 3.5 km. The garnet=zircon ratio is expressed as GZi (see Table 7 for denition).

strained study in the sense that the sequence had a uniform provenance: however, the burial depth range is fairly limited (from about 2100 m to 4100 m), and some of the area has been subject to uplift, so that the maximum burial depth is not known for all cases. Because the depth range is limited and shallow samples were not available, all the suites are relatively limited in diversity. Nevertheless, there is a clear decrease in diversity with depth, staurolite disappearing at about 3100 m and garnet at about 3600 m. Both species show evidence of surface corrosion as they

approach the levels at which they disappear. Zircon, rutile, tourmaline, apatite, chloritoid and spinel are apparently stable. Smale and van der Lingen (1989) showed that calcium content governs the stability of the garnets, as in the North Sea. Miocene and Pliocene sandstones of the Sava River basin of Yugoslavia also show a decline in heavy mineral diversity with depth (Scavnicar, 1979), interpreted as resulting from intrastratal solution. Kyanite disappears rst, followed by epidote group minerals, titanite and nally staurolite. Garnet,

Fig. 8. Scanning electron micrographs of selected mineral species. (a) Monazite from Triassic Skagerrak Fm sandstone, central North Sea (4403 m, 14444 ft), showing no evidence for dissolution. (b) Chrome spinel from Westphalian C =D sandstone, southern North Sea (3818 m, 12553 ft), showing no evidence for dissolution. (c) Gahnite (zinc spinel) from Middle Jurassic Brent Group sandstone, northern North Sea (3016 m, 9895 ft), showing no evidence for dissolution. (d) Chloritoid from Middle Jurassic Brent Group sandstone, northern North Sea (2728 m, 8950 ft) showing incipient corrosion manifested by triangular-shaped etch pits. Garnet in the same sample shows extensive corrosion. (e and f) Tourmaline from Upper Jurassic Claymore Fm sandstone, Outer Moray Firth (4135 m, 13567 ft), showing evidence for corrosion in the form of hacksaw terminations (e) and etch pits (f).

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

21

22

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

chloritoid, rutile, tourmaline, zircon and apatite are not lost from the assemblages. The pattern matches those from the North Sea, the US Gulf Coast and New Zealand in most respects, the only difference being the relatively shallow disappearance of kyanite (Table 5). Cavazza and Gandol (1992) demonstrated that a sandstone unit with an essentially uniform provenance (the Contessa megabed in the northern Apennines of Italy) shows mineralogical variations related to burial depth. Of the heavy minerals, zircon, monazite, rutile, spinel, garnet and chloritoid are ubiquitous over the area, and are considered to be stable under the conditions experienced by the megabed, but staurolite and titanite show variable distribution attributed to dissolution during burial. This study is particularly signicant because it demonstrates that other mineralogical parameters also vary because of burial depth, notably the abundance of smectite and interstratied illite=smectite and the abundance of intrabasinal carbonate clasts. There are a number of other basins where burialrelated dissolution is interpreted to have taken place, such as the Vienna Basin in Austria (Wieseneder and Maurer, 1958) and the North Sumatra Basin of Indonesia (Morton et al., 1994). Another observation worthy of note is the distribution of etched staurolite in Karroo (Permian) sandstones in Zimbabwe. Bond (1943) noted that staurolite, present throughout the Karroo in the area of study, was etched only in a sandstone immediately below a basaltic lava ow, and he therefore considered that the etching was related to the basalt. In view of the relationship between dissolution and elevated poreuid temperatures, it seems likely that the staurolite in this example was dissolved by hot uids introduced at the time of the basalt extrusion. The observation of grain surface corrosion coupled with patterns of subsurface disappearance is strong evidence for the widespread operation of dissolution processes, but is only circumstantial. Consequently, it has been argued that intrastratal solution is not widespread, but is a local phenomenon (van Andel, 1959; Weissbrod and Nachmias, 1986). However, further evidence for the extensive operation of subsurface dissolution comes from the identication of zones in which heavy minerals have been preserved within sequences where they are other-

wise absent. For example, preservation of heavy minerals may occur in early carbonate concretions (Bramlette, 1941), in mudstones (Blatt and Sutherland, 1969), and in oil-bearing sandstones (Yurkova, 1970). In North Sea Paleocene sandstones, preservation of heavy minerals has been observed in early calcite concretions, in analcite-cemented zones, in mudstones, in sandstones with limited poreuid connections, and in oil-bearing intervals. Identication of such horizons is crucial to the recognition of the operation of intrastratal solution processes, and the preserved assemblages give important information about the composition of the detrital assemblage prior to dissolution. There has been relatively little work on calibrating the precise conditions under which mineral dissolution occurs. Etch facets have been generated on garnets by subjecting them to warm dicarboxylic acid by Hansley (1987), who concluded that garnet dissolution in the subsurface is caused by circulation of organic acids at temperatures exceeding 80C, generated through alteration of organic matter. The presence of etched garnets may therefore be a useful indicator that sediments have reached the zone of oil generation. The work of Nickel (1973) on relative stability of minerals in different pH uids failed to simulate the pattern of stability identied during burial. This is presumably because the pH of the uids (3.6, 5.6 and 10.6) involved do not simulate the natural situation in the subsurface. It is worthy of note, however, that a limited series of experiments using a near-neutral uid (pH of 8) simulated the relative stability of apatite, kyanite and garnet as seen in deep burial, with kyanite < garnet < apatite (Nickel, 1973). This suggests that near-neutral uids best describe the conditions responsible for the subsurface dissolution patterns. A nal comment here is that apatite, known to be unstable under acidic conditions, is stable in deep burial, even though organic and carbonic acids are thought to be generated through decarboxylation of organic matter. Further work is clearly required on the stability of apatite in such conditions, as this may throw light on the nature of poreuids in deep burial. The pattern of mineral disappearance related to increasing burial depth can be used to identify areas of basin inversion and possibly give an estimate of maximum burial. The Middle Jurassic sequence

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329 Table 6 Comparison of order of persistence of heavy minerals, adapted from Pettijohn (1941) with observed stability patterns in acidic weathering and deep burial conditions Acidic weathering stability pattern TiO2 minerals Deep burial stability pattern Apatite, Monazite, Spinel, TiO2 minerals, Tourmaline, Zircon Chloritoid Garnet Staurolite Kyanite Titanite Epidote Calcic amphibole Andalusite, Sillimanite Pyroxene Olivine Order of persistence

23

7. Weathering at outcrop Following burial, sediments may be uplifted and exposed at the surface before subsequent phases of sedimentation, leading to the formation of subaerial unconformity surfaces. Under these circumstances, sandstones are likely to be invaded by acidic groundwaters generated in the same way as those involved in the surface weathering. Thus, subaerial unconformities can be recognized by zones of mineral depletion below the unconformity level, the pattern of depletion being comparable to those found in soil proles and in sands subjected to circulation of acidic groundwaters. The most critical mineral in this stability pattern is apatite, which is stable in deep burial but is exceptionally unstable under weathering. There are several examples where heavy minerals have helped to identify subaerial unconformity surfaces. Morton (1982) identied a dissolution prole at the top of the shallow marine Paleocene Thanet sandstones of southeast England that indicates exposure and subaerial weathering prior to the transgression that laid down the shallow marine Woolwich Formation. This weathering prole showed amphibole and titanite to be least stable of the minerals in the sandstones, followed by apatite, epidote and garnet, with staurolite, kyanite, zircon, tourmaline and the TiO2 minerals being stable. Another example is given by Weissbrod and Nachmias (1986) in their review of the Late Precambrian to Cretaceous sequences of Israel and adjacent areas, in which a major sub-Carboniferous unconformity caused extensive dissolution of detrital apatite. This example is of particular interest because the zone of detrital apatite depletion is characterized by abundant secondary apatite, considered to result from reprecipitation of the phosphate previously taken into solution. Sandstones presently exposed at outcrop are also susceptible to this process. For example, Hester (1974) showed that garnet and epidote are present in unweathered Upper Cretaceous sandstones of Alabama and Georgia (USA), but are absent in weathered samples. In cases such as these, where friable sandstones have been exposed at the surface for long periods, extensive alteration of the assemblages may have occurred, and it may be difcult to determine

Zircon Tourmaline Sillimanite Andalusite Kyanite Staurolite Topaz Titanite Monazite Garnet Epidote Calcic amphibole Orthopyroxene Clinopyroxene Olivine Apatite

TiO2 minerals Zircon Tourmaline Monazite Garnet Apatite Staurolite Kyanite Epidote Calcic amphibole Andalusite Topaz Titanite Clinopyroxene Sillimanite Orthopyroxene Olivine

of Yorkshire was deposited in the Cleveland Basin, which was subsequently inverted during the Tertiary, so that the basin axis now approximates to the axis of the Cleveland Anticline. There is a pattern of decreasing mineral diversity approaching the centre of the domal structure (Smithson, 1941), with assemblages rst losing staurolite and subsequently garnet. Minerals that remain in the assemblages at the basin centre are apatite, zircon, TiO2 minerals, tourmaline, spinel, chloritoid and monazite. The pattern of mineral disappearance is identical to the stability pattern deduced from deep burial (Tables 5 and 6), and therefore the distribution of minerals at the presentday surface reects the original burial pattern. Since extensive garnet dissolution starts to occur at about 80C (Hansley, 1987), temperatures at the basin centre must have exceeded this gure. This is in close accord with the evidence from uid inclusions in secondary sphalerite, which indicate temperatures of 7982C (Shepherd, quoted in Hemingway and Riddler, 1982), and from vitrinite reectance and spore coloration, which indicate a maximum temperature of 95C (Cooper and Barnard, quoted in Hemingway and Riddler, 1982).

24

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

the composition of the unweathered suites. For optimum assessment of provenance, it is crucial that every effort should be made to collect fresh material.

8. Heavy minerals and geological age The recognition that heavy minerals are unstable and subject to disappearance from sandstones has been recognized since the time of Thoulet (1913). In his inuential paper, Pettijohn (1941) showed that older sandstones are progressively less rich in heavy minerals, and considered that the principal cause of heavy mineral loss was geological age. He formulated an order of persistence of heavy minerals from his evaluation of the evolution of heavy mineral suites through time (Table 6). This table bears a close similarity to the order of mineral stability identied during conditions of elevated poreuid temperatures, although it differs in that apatite is not considered to have a high degree of persistence. The concept that geological age is the principal control on heavy mineral suites was challenged by Turnau-Morawska (1984) who argued that the literature on which Pettijohn based his order of persistence was biased, and showed an alternative relationship between heavy mineral diversity and age, in which only Quaternary sediments have signicantly more heavy minerals than the remainder of the geological column. There are many examples of ancient sandstones that contain rich and diverse heavy mineral assemblages, such as the Permo-Triassic of the preUrals foredeep (Ukraine), which contains abundant amphibole and pyroxene (Sarksiyan, 1958), the upper Devonian of the Pyrenees (France and Spain), which contains abundant amphibole and epidote (Stattegger, 1976), and the Ordovician of the Southern Uplands (Scotland), which contains abundant amphibole and pyroxene (Kelling, 1962; Styles et al., 1989). These examples all demonstrate that although older sandstones tend to be less rich in mineral diversity, the age of the formation cannot be regarded as the sole controlling factor behind intrastratal solution. The sequences characterized by this anomalous behaviour occurs were deposited rapidly, with sediment generally displaying very poor sorting and containing high proportions of labile lithic material. In consequence, they have poor initial permeabili-

ties which are rapidly reduced further by mechanical and chemical alteration of the lithic component. The preservation of unstable heavy minerals is regarded as a result of the low permeabilities, which sealed the sediment from circulating poreuids and thus protected the heavy mineral grains. The factors controlling mineral diversity are the composition and temperature of the poreuids, the permeability of the sandstone, and the length of time to which the sandstone has been subjected to such conditions. Geological time is therefore a consideration, which is presumably why the observed pattern of decreasing diversity with increasing age exists, but the diagenetic environment and the sediment properties are more important factors. Thus, the order of persistence of Pettijohn (1941) is essentially an empirical listing that results from an intermixing of the two stability patterns derived from case studies of sandstones subjected to weathering under acidic groundwater conditions and those subjected to diagenesis caused by circulation of high temperature poreuids in the subsurface. This explains why apatite occupies an intermediate position in Pettijohns list.

9. Provenance-sensitive parameters Despite the complex overprinting processes that affect heavy mineral distribution in sandstones, assemblages nevertheless retain fundamental provenance information. For correct interpretation of provenance, it is critical that the parameters used are inherited from the source area and are not modied to any great extent by processes operative during the sedimentation cycle. There are two alternative ways to generate such data, one that uses the entire heavy mineral suite (conventional HMA) and one that concentrates on the attributes of individual mineral species (varietal HMA). To maximize the provenance information a combined approach is recommended. 9.1. Conventional HMA The two most important processes that overprint the provenance signal are hydraulics and burial diagenesis. Therefore, ratios of minerals with similar hydraulic and diagenetic behaviour are most likely

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329 Table 7 Provenance-sensitive heavy mineral ratios, as proposed by Morton and Hallsworth (1994) Index ATi GZi RZi RuZi MZi CZi apatitetourmaline index garnetzircon index TiO2 groupzircon index rutilezircon index monazitezircon index chrome spinelzircon index Denition 100 apatite count=(total apatite plus tourmaline) 100 garnet count=(total garnet plus zircon) 100 TiO2 group count=(total TiO2 group plus zircon) 100 rutile count=(total rutile plus zircon) 100 monazite count=(total monazite plus zircon) 100 chrome spinel count=(total chrome spinel plus zircon)

25

to reect provenance characteristics. The minerals selected for ratio determination should meet two criteria: (i) that they are stable within the diagenetic context of the study, and (ii) that they have similar densities and grain sizes, which are the main controls on hydrodynamic behaviour. Morton and Hallsworth (1994) proposed a number of mineral ratios that largely reect provenance characteristics, as shown in Table 7. Of the criteria shown in Table 7, the garnetzircon index (GZi) should be used only in zones where garnet dissolution is not advanced. Although the apatitetourmaline index (ATi) is an important provenance-sensitive parameter, it may also be affected by weathering at several stages in the sedimentation cycle, and therefore may provide information on sediment transport history (Fig. 3). 9.2. Varietal HMA The overprinting effects of hydraulics and diagenesis can also be minimized by concentrating on properties displayed by a single mineral group. Traditionally, differentiation of provenance using varietal HMA concentrated on parameters acquired using the petrographic microscope, such as colour, habit or zoning pattern. For example, zircons often carry the signatures of a multi-phase petrological history, and their morphology has long been regarded as a petrogenetic indicator (Frasl, 1963; Pupin, 1976). Colour varieties such as those of zircon (Mackie, 1923) and tourmaline (Krynine, 1946; Mange-Rajetzky, 1995) can also be provenance-diagnostic. Focusing on such characteristics in heavy mineral studies augments information on parent rock lithologies and provides a valuable tool in provenance studies (e.g. Poldervaart, 1955; Mange-Rajetzky, 1995; Lihou and Mange-Rajetzky, 1996). This approach also

yields important information on sedimentation history. For example, changes in apatite morphology within the DevonianCarboniferous sequence of the Clair Field, west of Shetland, are attributable to variable interplay between uvial and aeolian deposition (Allen and Mange-Rajetzky, 1992). The advent of microbeam analytical techniques such as electron microprobe analysis has added a new dimension to varietal HMA, by allowing the geochemical characterization of individual mineral suites (Morton, 1991). This type of information adds sophistication to discrimination of provenance and enables direct mineralogical comparison between source and sediment. This approach can be applied to many mineral species, but is best used on minerals that are stable in the context of the study because differences in composition may control diagenetic behaviour. For example, garnet stability in deep burial is controlled by calcium content (Morton, 1987; Smale and van der Lingen, 1989). The most recent development is the application of single-grain isotopic dating methods to provenance studies, most notably on zircon and monazite, both of which are stable phases during the sedimentation cycle. Singlegrain dating can be undertaken using conventional isotopic methods (e.g. Cliff et al., 1991), but the large amount of data needed to characterize source terrains is more readily acquired using microbeam techniques such as the sensitive high-resolution ion microprobe (Morton et al., 1996; Sircombe, 1997). Single-grain dating adds a major new dimension to provenance analysis because it provides a detailed control on the timing of major rock-forming events in the source terrain. This can be combined with data on the metamorphic or cooling history of the source terrain, such as 40 Ar39 Ar laser probe dating of mica (e.g. von Eynatten et al., 1996; Stuart et al., 1997)

26

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329

and ssion-track dating of minerals such as apatite or zircon (Hurford and Carter, 1991; Carter, 1999), thereby providing constraints on both the nature and geodynamic history of the source area. In conclusion, a threefold approach to characterization of sediment provenance using HMA is recommended. The rst stage is to characterize, differentiate and map sand types using conventional HMA methods, which places important constraints on the nature and location of source areas. The second stage is to select one or more mineral type for varietal HMA to provide direct mineralogical comparison with potential source areas and further information on source rock lithology. The nal stage is to acquire isotopic data, which give direct age constraints on the source region. This combined conventional, varietal and isotopic approach to HMA provides a powerful tool for evaluation of sediment provenance. 9.3. Further research

stratigraphy, which appears to be linked with the effects of weathering during storage, is another topic worthy of study. Despite the need for further investigation on some of the factors controlling heavy mineral assemblages in sandstones, the long history of heavy mineral research has provided an excellent basis for us to interpret provenance. The same is not necessarily true of other techniques, such as whole-rock geochemistry, which is a comparatively new approach but relies at least in part on variations in heavy mineral abundance to generate changes in chemical signature. Linking chemostratigraphic variations with changes in heavy mineral data is a potentially fruitful area of research (e.g. Preston et al., 1998), allowing a better understanding of the controls on whole-rock composition and identication of which parameters are most suitable for provenance evaluation.

Acknowledgements Although the factors controlling heavy mineral assemblages in sandstones are well-known, several areas of research remain. The effects of the most signicant processes (hydraulics and diagenesis) are reasonably well understood, although further case studies will undoubtedly increase our knowledge of their effects. For example, the effects of burial diagenesis are appreciated in NW European basins, but there has been comparatively little similar work carried out elsewhere. Also, the stability of less common heavy mineral phases during diagenesis is not well constrained because of the relative lack of case studies. Of the other factors involved, the effects of weathering during alluvial storage is the least well understood. The case studies available to date do not provide adequate constraints on the effects of this process. Although the study of the Apure River sediments in the Llanos area of Venezuela (Morton and Johnsson, 1993) showed that heavy mineral assemblages have been modied during storage, the composition of sediment fed into the river system was too heterogeneous to isolate the weathering effect. Further studies of river systems in tropical environments are needed, ideally using rivers draining less diverse terrains that the northern Andes. The possible relationship between mineralogy and sequence We are grateful to Robert Knox, Maria Mange, Ian Jarvis and Hilmar von Eynatten for their comments on earlier drafts of this manuscript. This paper is published with the approval of the Director, British Geological Survey (N.E.R.C.).

References
Allen, J.R.L., 1994. Fundamental properties of uids and their relation to sediment transport processes. In: Pye, K. (Ed.), Sediment Transport and Depositional Processes. Blackwell, Oxford, pp. 2560. Allen, P.A., Mange-Rajetzky, M.A., 1992. DevonianCarboniferous sedimentary evolution of the Clair Area, offshore northwestern UK: impact of changing provenance. Mar. Pet. Geol. 9, 2952. Allen, V.T., 1948. Weathering and heavy minerals. J. Sediment. Petrol. 18, 3842. Bagnold, R.A., 1954. Experiments on a gravity free dispersion of large solid spheres in a Newtonian uid under shear. Proc. R. Soc. London A 225, 4963. Bateman, R.M., Catt, J.A., 1985. Modication of heavy mineral assemblages in English coversands by acid pedochemical weathering. Catena 12, 121. Berner, R.A., Sjo berg, E.L., Velbel, M.A., Krom, M.D., 1980. Dissolution of pyroxenes and amphiboles during weathering. Science 207, 12051206. Blatt, H., Sutherland, B., 1969. Intrastratal solution and non-

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329 opaque heavy minerals in shales. J. Sediment. Petrol. 39, 591 600. Bond, G., 1943. Solution etching of detrital staurolite. Geol. Mag. 80, 155156. Bramlette, M.N., 1941. The stability of minerals in sandstone. J. Sediment. Petrol. 11, 3236. Briggs, L.I., McCulloch, D.S., Moser, F., 1962. The hydraulic shape of sand particles. J. Sediment. Petrol. 32, 645656. Burley, S.D., Kantorowicz, J.D., Waugh, B., 1985. Clastic diagenesis. In: Brenchley, P.J., Williams, B.P.J. (Eds.), Sedimentology: Recent Developments and Applied Aspects. Geol. Soc. London, Spec. Publ. 18, 189226. Buttereld, J.A., 1936. Outgrowths on zircon. Geol. Mag. 73, 511516. Carter, A., 1999. Present status and future avenues of source region discrimination and characterization using ssion track analysis. Sediment. Geol. 124, 3145 (this issue). Carver, R.E., Kaplan, D.M., 1976. Distribution of hornblende on the Atlantic continental shelf off Georgia, United States. Mar. Geol. 20, 335343. Cavazza, W., Gandol, G., 1992. Diagenetic processes along a basin-wide marker bed as a function of burial depth. J. Sediment. Petrol. 62, 261272. Cliff, R.A., Drewery, S.A., Leeder, M.R., 1991. Sourcelands for the Carboniferous Pennine river system: constraints from sedimentary evidence and UPb geochronology using zircon and monazite. In: Morton, A.C., Todd, S.P., Haughton, P.D.W. (Eds.), Developments in Sedimentary Provenance Studies. Geol. Soc. London, Spec. Publ. 57, 137159. Dietz, V., 1973. Experiments on the inuence of transport on shape and roundness of heavy minerals. Contrib. Sediment. 1, 103125. Doyle, L.J., Carder, K.L., Steward, R.G., 1983. The hydraulic equivalence of mica. J. Sediment. Petrol. 53, 643648. Dryden, A.L., Dryden, C., 1946. Comparative rates of weathering of some common heavy minerals. J. Sediment. Petrol. 16, 91 96. Edelman, C.H., Doeglas, D.J., 1932. Reliktstructuren detritischer pyroxene und amphibole. Tschermaks Mineral. Petrogr. Mitt. 42, 482490. ber umwandlungserscheiEdelman, C.H., Doeglas, D.J., 1934. U nungen an detritischem staurolith und anderen mineralien. Tschermaks Mineral. Petrogr. Mitt. 44, 225234. von Eynatten, H., Gaupp, R., Wijbrans, J.R., 1996. 40 Ar=39 Ar laser-probe dating of detrital white micas from Cretaceous sedimentary rocks of the eastern Alps: evidence for Variscan highpressure metamorphism and implications for Alpine orogeny. Geology 24, 691694. Fletcher, W.K., Church, M., Wolcott, J., 1992. Fluvial-transport equivalence of heavy minerals in the sand size range. Can. J. Earth Sci. 29, 20172021. Flores, R.M., Shideler, G.L., 1978. Factors controlling heavy mineral variations on the south Texas outer continental shelf. J. Sediment. Petrol. 48, 269280. Frasl, G., 1963. Die mikroskopische Untersuchung der akzessorischen Zirkone als eine Routinearbeit des Kristallingeologen. Jahrb. Geol. Bundesanst. 106, 405428.

27

Freise, F.W., 1931. Untersuchung von Mineralen auf Abnutzbarkeit bei Verfrachtung im Wasser. Tschermaks Mineral. Petrogr. Mitt. 41, 17. Friis, H., 1976. Weathering of a Neogene uviatile ning upwards sequence at Voervadsbro, Denmark. Bull. Geol. Soc. Denmark 25, 99105. Friis, H., 1978. Heavy-mineral variability in Miocene marine sediments in Denmark: a combined effect of weathering and reworking. Sediment. Geol. 21, 169188. Friis, H., Johannesen, F.B., 1974. Late Tertiary weathering of uvial deposits at La sby, Denmark. Bull. Geol. Soc. Denmark 23, 197202. Friis, H., Nielsen, O.B., Friis, E.M., Balme, B.E., 1980. Sedimentological and palaeontological investigations of a Miocene sequence at Lavsjberg, central Jutland. Arbok Danmarks Geol. Undersgelse 1979, 5167. Goldich, S.S., 1938. A study in rock weathering. J. Geol. 46, 1758. Grimm, W.D., 1973. Stepwise heavy mineral weathering in the residual quartz gravel, Bavarian Molasse (Germany). Contrib. Sediment. 1, 103125. Hand, B.M., 1967. Differentiation of beach and dune sands using settling velocities of light and heavy minerals. J. Sediment. Petrol. 37, 514520. Hansley, P.L., 1987. Petrologic and experimental evidence for the etching of garnets by organic acids in the Upper Jurassic Morrison Formation, northwestern New Mexico. J. Sediment. Petrol. 57, 666681. Hattingh, J., 1992. Dispersal of tracer heavy minerals and clasts in the gravel-bearing braided Swartkops River, eastern Cape Province, South Africa. Bull. Geol. Surv. S. Afr. 112. Hemingway, J.E., Riddler, G.P., 1982. Basin inversion in North Yorkshire. Trans. Inst. Mini. Metall. 91, 175186. Hester, N.C., 1974. Post-depositional subaerial weathering effects on the mineralogy of an Upper Cretaceous sand in southeastern United States. J. Sediment. Petrol. 44, 363373. Hower, J., Eslinger, E.V., Hower, M.E., Perry, E.A., 1976. Mechanisms of burial and metamorphism of argillaceous sediment. I. Mineralogical and chemical evidence. Bull. Geol. Soc. Am. 87, 725737. Hubert, J.F., 1962. A zircontourmalinerutile maturity index and the interdependance of the composition of heavy mineral assemblages with the gross composition and texture of sandstones. J. Sediment. Petrol. 32, 440450. Hubert, J.F., 1971. Analysis of heavy mineral assemblages. In: Carver, R.E. (Ed.), Procedures in Sedimentary Petrology. Wiley, New York, pp. 453478. Hurford, A.J., Carter, A., 1991. The role of ssion track dating in discrimination of provenance. In: Morton A.C., Todd, S.P., Haughton, P.D.W. (Eds.) Developments in Sedimentary Provenance Studies. Geol. Soc. London, Spec. Publ. 57, 67 78. Johnsson, M.J., Stallard, R.F., Meade, R.H., 1988. First-cycle quartz arenites in the Orinoco River basin, Venezuela and Colombia. J. Geol. 96, 103277. Johnsson, M.J., Stallard, R.F., Lundberg, N., 1991. Controls on the composition of uvial sands from a tropical weathering

28

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329 Morton, A.C., 1982. The provenance and diagenesis of Palaeogene sandstones of south-east England as indicated by heavy mineral analysis. Proc. Geol. Assoc. 93, 263274. Morton, A.C., 1984. Stability of detrital heavy minerals in Tertiary sandstones of the North Sea Basin. Clay Miner. 19, 287 308. Morton, A.C., 1985. Heavy minerals in provenance studies. In: Zuffa, G.G. (Ed.), Provenance of Arenites. Reidel, Dordrecht, pp. 249277. Morton, A.C., 1986. Dissolution of apatite in North Sea Jurassic sandstones: implications for the generation of secondary porosity. Clay Miner. 21, 711733. Morton, A.C., 1987. Inuences of provenance and diagenesis on detrital garnet suites in the Forties sandstone, Paleocene, central North Sea. J. Sediment. Petrol. 57, 10271032. Morton, A.C., 1991. Geochemical studies of detrital heavy minerals and their application to provenance studies. In: Morton A.C., Todd, S.P., Haughton, P.D.W. (Eds.), Developments in Sedimentary Provenance Studies. Geol. Soc. London, Spec. Publ. 57, 3145. Morton, A.C., Hallsworth, C.R., 1994. Identifying provenancespecic features of detrital heavy mineral assemblages in sandstones. Sediment. Geol. 90, 241256. Morton, A.C., Humphreys, B., 1983. The petrology of the Middle Jurassic sandstones from the Murchison Field, North Sea. J. Pet. Geol. 5, 245260. Morton, A.C., Johnsson, M.J., 1993. Factors inuencing the composition of detrital heavy mineral suites in Holocene sands of the Apure River drainage basin, Venezuela. In: Johnsson, M.J., Basu, A. (Eds.), Processes Controlling the Composition of Clastic Sediments. Geol. Soc. Am., Spec. Pap. 284, 171 185. Morton, A.C., Smale, D., 1991. The effects of transport and weathering on heavy minerals from the Cascade River, New Zealand. Sediment. Geol. 68, 117123. Morton, A.C., Borg, G., Hansley, P.L., Haughton, P.D.W., Krinsley, D.H., Trusty, P., 1989. The origin of faceted garnets in sandstones: dissolution or overgrowth. Sedimentology 36, 927942. Morton, A.C., Humphreys, B., Dharmayanti, D.A., Sundoro, A., 1994. Palaeogeographic implications of the heavy mineral distribution in Miocene sandstones of the North Sumatra Basin. J. Southeast Asian Earth Sci. 10, 177190. Morton, A.C., Claoue -Long, J., Berge, C., 1996. Factors inuencing heavy mineral suites in the Statfjord Formation, Brent Field, North Sea: constraints provided by SHRIMP UPb dating of detrital zircons. J. Geol. Soc. London 153, 911929. Nickel, E., 1973. Experimental dissolution of light and heavy minerals in comparison with weathering and intrastratal solution. Contrib. Sediment. 1, 168. Pettijohn, F.J., 1941. Persistence of heavy minerals and geologic age. J. Geol. 49, 610625. ber den Schwermineralgehalt von anstehenPiller, H., 1951. U dem und verwittertem Brockengranit no rdlich St. Andreasberg. Heidelb. Beitr. Mineral. Petrogr. 2, 523537. Poldervaart, A., 1955. Zircon in rocks, 1: sedimentary rocks. Am. J. Sci. 253, 433461.

environment: Sands of the Orinoco drainage basin, Venezuela and Colombia. Bull. Geol. Soc. Am. 103, 16221647. Jones, D.G., Haslam, H.W., Kemp, S.J., Leng, M.J., Milodowski, A.E., Morton, A.C., Strong, G.E., in press. Provenance of the Cheshire Basin ll. Mem. Br. Geol. Surv. Kelling, G., 1962. The petrology and sedimentation of Upper Ordovician rocks in the Rhinns of Galloway, south-west Scotland. Trans. R. Soc. Edinburgh 65, 107137. Komar, P.D., Wang, C., 1984. Processes of selective grain transport and the formation of placers on beaches. J. Geol. 92, 637655. Krynine, P.D., 1946. The tourmaline group in sediments. J. Geol. 54, 6587. Larsen, G., 1966. Rhaetic-JurassicLower Cretaceous sediments in the Danish Embayment (a heavy mineral study). Danmarks Geol. Undersgelse, Raekke 2, 91, 128 pp. Lemcke, K., von Engelhardt, W., Fu chtbauer, H., 1953. Geologische und sedimentpetrographische Untersuchungen im Westteil der ungefalteten Molasse des su ddeutschen Alpen-vorlandes. Beih. Geol. Jahrb. 11. Lihou, J.C., Mange-Rajetzky, M.A., 1996. Provenance of the Sardona Flysch, eastern Swiss Alps: example of high resolution heavy mineral analysis applied to an ultrastable assemblage. Sediment. Geol. 105, 141157. Lowright, R., Williams, E.G., Dachille, F., 1972. An analysis of factors controlling deviations in hydraulic equivalence in some modern sands. J. Sediment. Petrol. 44, 635645. Mackie, W., 1923. The source of purple zircons in the sedimentary rocks of Scotland. Trans. Edinburgh Geol. Soc. 11, 200213. McMaster, R.L., 1954. Petrography and genesis of the New Jersey beach sands. Bull. N.J. Dep. Conserv. Econ. Devel., Geol. Ser. 63, 239 pp. Mange, M.A., Maurer, H.F.W., 1992. Heavy Minerals in Colour. Chapman and Hall, London. Mange-Rajetzky, M.A., 1983. Sediment dispersal from source to shelf on an active continental margin, S. Turkey. Mar. Geol. 52, 126. Mange-Rajetzky, M.A., 1995. Subdivision and correlation of monotonous sandstone sequences using high resolution heavy mineral analysis, a case study: the Triassic of the Central Graben. In: Dunay, R.E., Hailwood, E.A. (Eds.), Non-biostratigraphical Methods of Dating and Correlation. Geol. Soc. London, Spec. Publ. 89, 2330. Milliken, K.L., 1988. Loss of provenance information through subsurface diagenesis in Plio-Pleistocene sediments, northern Gulf of Mexico. J. Sediment. Petrol. 58, 9921002. Milliken, K.L., Mack, L.E., 1990. Subsurface dissolution of heavy minerals, Frio Formation sandstones of the ancestral Rio Grande Province, South Texas. Sediment. Geol. 68, 187 199. Mitchell, W.A., 1975. Heavy minerals. In: Gieseking, J.E. (Ed.), Soil Components II. Inorganic. Springer, New York, pp. 449 480. Morton, A.C., 1979. Surface textures of heavy mineral grains from the Palaeocene of the central North Sea. Scot. J. Geol. 15, 293300.

A.C. Morton, C.R. Hallsworth / Sedimentary Geology 124 (1999) 329 Preston, J., Hartley, A., Hole, M., Buck, S., Bond, J., Mange, M., Still, J., 1998. Integrated whole-rock trace element geochemistry and heavy mineral chemistry studies: aids to the correlation of continental red-bed reservoirs in the Beryl Field, UK North Sea. Pet. Geosci. 4, 716. Pupin, J.-P., 1976. Signication des caracteres morphologiques du zircon commun des roches en pe trologie. Base de la me thode typologique Applications. PhD thesis, University of Nice. Reid, I., Frostick, L.E., 1985. Role of settling, entrainment and dispersive equivalence and of interstice trapping in placer formation. J. Geol. Soc., London 142, 739746. Reid, I., Frostick, L.E., 1994. Fluvial sediment transport. In: Pye, K. (Ed.), Sediment Transport and Depositional Processes. Blackwell, Oxford, pp. 89155. Rittenhouse, G., 1943. The transportation and deposition of heavy minerals. Bull. Geol. Soc. Am. 54, 17251780. Rochow, K.A., 1981. Seismic stratigraphy of the North Sea Palaeocene deposits. In: Illing, L.V., Hobson, G.D. (Eds.), Petroleum Geology of the Continental Shelf of North-West Europe. Heyden, London, pp. 255266. Rubey, W.W., 1933. The size distribution of heavy minerals within a water-lain sandstone. J. Sediment. Petrol. 3, 329. Russell, R.D., 1937. Mineral composition of Mississippi River sands. Bull. Geol. Soc. Am. 48, 13071348. Sallenger Jr., A.H., 1979. Inverse grading and hydraulic equivalence in grain ow deposits. J. Sediment. Petrol. 49, 553 562. Sarksiyan, S.G., 1958. Upper Permian continental Molasses of the Pre-Urals. Eclogae Geol. Helv. 51, 10431051. Savage, K.M., de Cesaro, P., Potter, P.E., 1988. Mineralogic maturity of modern sand along a high-energy tropical coast: Baixada de Jacarepagua , Rio de Janeiro, Brazil. J. South Am. Earth Sci. 1, 317328. Savage, K.M., Potter, P.E., 1991. Petrology of modern sands of the Rios Guavare and In rada, southern Colombia. J. Geol. 99, 289298. Scavnicar, B., 1979. Pjescenjaci Pliocena i Miocena savske potoline. Zb. Rad. Sekcija Priml. Geol. Geoz., Geokem., Ser. A 6, 351382. Shukri, N.M., 1949. The mineralogy of Nile sediments. Q. J. Geol. Soc. London 105, 511529. Sircombe, K.N., 1997. Where does the sand come from? The provenance of heavy detrital minerals in coastal sands and sedimentary rocks of east Australia using the SHRIMP ion probe. Terra Nova 9 (Abstr. Suppl. 1), 593. Slingerland, R.L., 1977. The effect of entrainment on the hydraulic equivalence relationships of light and heavy minerals in sands. J. Sediment. Petrol. 47, 753770. Slingerland, R.L., 1984. The role of hydraulic sorting in the origin of uvial placers. J. Sediment. Petrol. 54, 137150. Smale, D., Morton, A.C., 1987. Heavy mineral suites of core samples from the McKee Formation (EoceneLower Oligocene), Taranaki: implications for provenance and diagenesis. N.Z. J. Geol. Geophys. 30, 299306. Smale, D., van der Lingen, G.J., 1989. Differential leaching of garnet grains at a depth of 3.5 km in Tane-1, Offshore

29

Taranaki, New Zealand. N.Z. Geol. Surv. Rec. 40, 5760. Smithson, F., 1941. The alteration of detrital minerals in the Mesozoic rocks of Yorkshire. Geol. Mag. 78, 97112. Stattegger, K., 1976. Schwermineraluntersuchungen in den klastischen Serien der Variszischen Geosynklinale der Ostund Zentral-pyrena en. Mitt. Osterr. Geogr. Ges. 69, 267290. Stuart, F., Bluck, B., Pringle, M.F., 1997. Provenance of British Carboniferous sandstones from laser 40 Ar=39 Ar ages of individual detrital muscovite. Terra Nova 9 (Abstr. Suppl. 1), 593. Styles, M.T., Stone, P., Floyd, J.D., 1989. Arc detritus in the Southern Uplands: mineralogical characterisation of a missing terrain. J. Geol. Soc., London 146, 397400. Thiel, G.A., 1940. The relative resistance to abrasion of mineral grains of sand size. J. Sediment. Petrol. 10, 103124. Thiel, G.A., 1945. Mechanical effects of stream transportation in mineral grains of sand size. Bull. Geol. Soc. Am. 56, 1207. Thoulet, J., 1913. Notes de lithologie sous-marine. Ann. Inst. Oceanogr. 5, 114. Tissot, B., Durand, B., Espitalic, J., Combaz, A., 1974. Inuence of nature and diagenesis of organic matter in formation of petroleum. Bull. Am. Assoc. Pet. Geol. 58, 499506. Turnau-Morawska, M., 1984. Importance of heavy mineral analysis in solving geological problems. In: Luepke, G. (Ed.), Stability of Heavy Minerals in Sediments. Van Nostrand Reinhold, New York, pp. 280287 (translated from the original 1955 article in Acta Geol. Pol. 5, 363388). van Andel, T.H., 1950. Provenance, Transport and Deposition of Rhine Sediments. Veenman, Wageningen. van Andel, T.H., 1959. Reection on the interpretation of heavy mineral analyses. J. Sediment. Petrol. 29, 153163. Walker, T.R., 1967. Formation of red-beds in modern and ancient deserts. Bull. Geol. Soc. Am. 78, 353368. Walker, T.R., 1974. Formation of red beds in moist tropical climates: a hypothesis. Bull. Geol. Soc. Am. 85, 633638. Walker, T.R., Ribbe, P.H., Honea, R.M., 1967. Geochemistry of hornblende alteration in Pliocene red beds, Baja California, Mexico. Bull. Geol. Soc. Am. 78, 10551060. Walker, T.R., Waugh, B., Crone, A.J., 1978. Diagenesis in rstcycle desert alluvium of Cenozoic age, southwestern United States and northwestern Mexico. Bull. Geol. Soc. Am. 89, 1932. Weissbrod, T., Nachmias, J., 1986. Stratigraphic signicance of heavy minerals in the late PrecambrianMesozoic clastic sequence (Nubian Sandstone) in the Near East. Sediment. Geol. 47, 263291. Weyl, R., Werner, H., 1951. Schwermineraluntersuchungen im Jungtert a r und Altquarta r Schleswig-Holsteins. In: Proceedings 3rd. International Sedimentology Congress, Groningen Wageningen, pp. 293303. Wieseneder, H., Maurer, J., 1958. Ursachen der ra umlichen und zeitlichen Aenderung des Mineralbestandes der Sedimente des Wiener Beckens. Eclogae Geol. Helv. 51, 11551172. Yurkova, R.M., 1970. Comparison of postsedimentary alteration of oil-, gas- and water-bearing rocks. Sedimentology 15, 53 68.

You might also like