JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. B5, PAGES 9045-9050, MAY 10, 1994 Toward the discrimination of fine-grained ferrimagnets by magnetic measurements in lake and near-shore marine sediments F. Oldfield Department of Geography, Universityof Liverpool,Liverpool,England Abstract. The magnetic properties of low field susceptibility (Z), frequency dependent susceptibility (Zfd), and susceptibility of anhysteretic remanent magnetisation (Z^RM) are used to define sets of natural soilandsediment samples withinwhich, by thecriteria of King et al. (1982) andMaher (1988), the modal ferrimagnetic grain diameter is less than-0.07 -0.1 gm. Within this sample set,two distinct groups with low andhigh values, respectively, for the quotients Z•M/Z and Za,•/zfd are defined. Thefirst group includes sediment samples from sites where published studies propose a detrital origin for thefine-grained ferrimagnetic content. Where catchment samples are available for comparison, theyfall withinthe same envelope, asdo clays from palaeosol samples within Chinese loess. This envelope of low quotient values also overlaps with the values for the fine grained synthetic magnetite samples within Maher's(1988) New MT series.The highquotient envelope of values includes sediments from theAdriaticSeaandclays from saltmarshand shallow water marine sediments in NW Britain. In these cases, no catchment source is postulated for the fine-grained magnetite. The size range of magnetite in thissetappears to be almost exclusively stable single domain (SSD), anda bacterial originseems likely. Sample sets from estuarine environments between river inflow sites andthe open sea show intermediate values which, in the case of the Potomac, are ordered by distance downriver. Althoughat this stage, magnetic measurements alone cannot discriminate between bacterial andfine-grained detrital ferrimagnets in sediments, they holdout some promise of doingsoprovided the distinctions proposed here canbe substantiated by transmission electron microscopy. Introduction Following on from the initial susceptibility, based on studies of Lough Neaghsediments by Thompson et al. [1975], many papers have used the "rock" magnetic record in Quaternary sediments primarily as an indication of the nature and rateof erosion fromthecatchment. Theyinclude attempts at bothqualitative [Oldfield et al. 1985] andquantitive [Yu and Oldfield, 1989, 1993] sediment source ascription. By now, it is clear that the general proposition upon which suchstudies rest (that magnetic minerals in sediment fully and accurately record a terrigenous, detrital signal) must often be qualified,since the "detrital" signal may be modified by oneor more of the following: (1) reductivediagenesis leading to 'magnetite' dissolution[e.g., Anderson and Rippey, 1988; Oldfield et al., 1992a], (2) authigenicgreigite formation [Snowball and Thompson, 1988;Snowball,1991], and(3) the in situ production of bacterial magnetite, commonly though not exclusively, as intercellular magnetosomes [e.g., Chang et al., 1989; Petersen et al., 1986]. A full range of susceptibility and remanence measurements can usually identify horizons where diagenesis strongly affects the sediment record. The distinctivegrain size and mineral selectivity of the process[Karlin and Levi, 1983; Copyright 1994 by the American Geophysical Union. Paper number93JB03137. 0148-0227/94/93JB-03137 $05.00 Bloemendal et al., 1992] can be used to detect it from its effects and to indicate those sites and horizons to be avoided if the aim of the research is to establish sediment-source linkages from the magnetic properties. Snowball [1991] has recently confirmed that sediment samples strongly influenced by authigenic greigiteformation can be recognised from their distinctivelyhigh isothermal remanence/susceptibility quotients and remanence hysteresis characteristics. They often occur in discrete, well-defined zones [cf. Reynolds et al., 1990]. Experimental results [Oldfield et al., 1992] showthat the more general lossof g and isothermal remanent magnetization (IRM) in some sediments, which was initially ascribed to the oxidation of greigite [Snowball and Thompson, 1988], arisesinsteadfrom the effects of slow drying out of palaeomagnetic subsamples stored in the moist condition. In the many suites of sediment samples which have escaped the influence of both diagenesis and authigenic greigite formation, there is little doubt that the ferrimagnetic grains, with diameters from - 0.2 gm diameter upward,are of detrital and overwhelmingly terrigenous origin. The evidence for source-sedimentlinkages and the visual characteristics from electron micrography concur to reinforce this conclusion. Much more controversy surrounds the origin of the fine (<0.1 gm diameter) stable single domain (SSD) and superparamagnetic (SP) grains in sediments. It is clear that in some deep-sea sediments,bacterial magnetite in the form of SSD size magnetosomes predominates [Kirschvinck et al 1984; Petersen et al., 1986; Vali et al., 1987; Bloemendal et 9045