The debate over the role of fluid flow in the precipitation of diagenetic cements is a longstanding one that arose because it is often difficult to find a sufficient local source of cement to account for observed cement volumes, and it is equally difficult to justify the large volume of pore waters required to transport the necessary chemical components from distant sources (e.g. McBride, 1989). The debate has been particularly heated in cases where the cement sources and sinks are not readily apparent. For example, dolomitization of limestone requires a hydrologic mechanism that provides magnesium and removes excess calcium. A recent paper by Kaufman (1994) addressed the issue of dolomitization and suggested that downward moving fluids associated with sea-level falls may serve as one possible mechanism. The role of fluid flow processes are also of concern with respect to Al-bearing species (e.g. Hayes and Boles, 1992). However, the problem is perhaps most acute in the case of quartz cement, which is the principal cause of porosity loss in many clastic reservoirs. A number of recent papers emphasize this subject, and I will discuss these papers to illustrate the nature of the debate.
Major factors that control quartz precipitation have been extensively studied, and include such variables as temperature, framework grain composition, quartz grain surface area, solubility, and availability of fluids. However, an accurate, process-driven predictive method remains elusive. Until the late 1970's, local dissolution at grain contacts, known as intergranular pressure solution and recognized by ``sutured'' grain contacts, was thought to be the predominant source of diagenetic quartz. However, that notion has now been largely discarded (e.g. Paxton et al., 1990; Houseknecht, 1987; Sibley and Blatt, 1976). More recently, the debate has revolved around the role of large-scale fluid flow.
A good example of the tenor this debate has taken over the past four years can be found in a number of papers dealing with homogenization temperatures of aqueous fluid inclusions in quartz cement from North Sea Mesozoic sandstones (Walderhaug, 1994a; Osborne and Haszeldine, 1993; Robinson et al., 1992; and Gluyas and Coleman, 1992; among others). In each of the studies, fluid inclusion homogenization temperatures are observed to increase with present-day bottom-hole temperature of cored wells. If the inclusion temperatures are valid, then these observations imply that quartz precipitated within the last few million years of burial, during a rapid, relatively short-lived series of cementation episodes (e.g. Robinson and Gluyas, 1992). Was large-scale fluid flow required, or was temperature the dominant factor, with silica being locally redistributed from sources not immediately obvious from petrographic examination?
Newly acquired supporting data have been used to support both arguments. For example, Coleman and Eggenkamp (1994) have made innovative use of new chlorine isotope analyses to argue that fluids were clearly moving around in the basin, and are therefore likely to have been involved in silica cementation. Conversely, Aplin et al. (1993) concluded that fluid transport mechanisms must have been restricted to reservoir-scale convection and/or diffusion based on the compositional heterogeneity of the precipitating fluids, as measured by oxygen isotopes and fluid inclusion salinities. Bjø rlykke and Gran (1994) argue that distinct stratification of salinity in porewaters from the North Sea precludes large-scale fluid flow, as such flow would have homogenized the pore water salinities. Finally, calculations of likely fluxes of aqueous silica in sedimentary basins suggest that quartz cement sources must be predominantly local (<10 m), assuming equilibrium processes predominate, as there is not enough meteoric or compactional water to supply the needed components (for a recent calculation, see Bjø rlykke and Egeberg, 1993).
Even if large-scale flow is not the primary mechanism by which
the components in cements are transported, there is good evidence
that large-scale fluid movements do occur in some sedimentary
basins, as shown by extensive analyses of Gulf Coast porewaters
(e.g. Hanor, 1994; Land and Macpherson, 1992a; Land, 1991). Some
of these fluids are clearly derived from shale dewatering.
However, new evidence, including minor anomalies in exotic metals
and 
He, also points to a contribution from so-called
thermobaric fluids derived from deep crustal fluid circulation
and/or low-grade metamorphic reactions (e.g. Elliot et al., 1993;
Land, 1991). These observations have led to considerable
speculation regarding the role of deep basinal fluids as agents of
diagenesis, but the importance of such fluids to quartz cement
precipitation remains unclear at present.
Ultimately, quartz precipitation is likely a complex affair involving a variety of mechanisms whose importance varies in different settings, and it may be many years before we can adequately resolve the relative contributions of the various mechanisms to create an adquate predictive model. In the meantime, promising empirical approaches continue to develop. For example, Walderhaug (1994b) combined fluid inclusion homogenization temperatures with burial history information to develop an empirical kinetic model for quartz precipitation. Also, McKeever et al., (1994) recently presented some of the first oxygen isotopic analyses of quartz cement collected using the ion microprobe. The reproducibility of their analyses was equivalent to conventional isotopic analyses, and they are now analyzing samples as small as 30-40 microns, close to the dimensions of an individual overgrowth. Other researchers such as Elias et al. (1993) are continuing to perform experiments to test the circumstances under which pressure solution is likely to be an important source of quartz cement. As these types of developments continue to emerge, the odds of developing a comprehensive predictive model continue to improve.