The NRM of all the samples was studied by stepwise (AF) demagnetization of variable peak field, according to the coercivity of the sediments. The intensity of magnetization is different from site to site with the highest value observed at Site 974 and lowest value measured at Site 976. This reflects the different source of the magnetic minerals between the sites. As shown in Figure 2, the magnetic inclinations are almost exclusively positive, which indicates that all the samples belong to the Brunhes Chron. A couple of exceptions have been observed, showing negative inclinations: a small interval between 7.9 and 8.4 mcd at Site 975, and a couple of samples at ~14 m depth at Site 976. The former is clear after the AF cleaning. This observation gives credit to the event as representative of a primary magnetization. The stratigraphic position, in particular, strongly supports a correlation with the Blake Event. This event is documented by full-reverse field directions at several sites around the world. By correlation with the oxygen-isotope stratigraphy the event has been dated in the Mediterranean between 117 and 123 ka by Tucholka et al. (1987). This age is well in agreement with the Chinese loess record that shows the Blake Event just above the Eemian interglacial (Fang et al., 1997). The stratigraphic position of the observed reverse interval is coincident with the beginning of the reversal at about 123 ka while the length appears a little bit shorter. On the basis of the sedimentation rate, the event appears to have ended at about 119 ka. It occurs in the Core 161-975B-2H, just one meter above the sapropel S5, which is known to have been deposited during the Eemian at about 124 ka (astronomical age; Lourens et al., 1996). Moreover, the detailed oxygen-isotope stratigraphy available for this interval (Pierre et al., Chap. 38, this volume), assigns this event midway between the substages 5d and 5e, exactly in the same position recognized by Tucholka et al. (1987) in eastern Mediterranean cores.
In spite of the high sedimentation rate, this event has not been observed at Site 976. Heavy remagnetization induced by the drilling was documented for the sediments recovered during Leg 161 (Comas, Zahn, Klaus, et al., 1996), so it is possible that the strong overprint masks the event. The down-core plot of the magnetic inclinations shows, after cleaning, two samples with negative inclinations at about 14 m of depth (Fig. 2). At this stage they cannot be considered unequivocally representative of a geomagnetic reversal; however, if they do represent a true event, the stratigraphic position within oxygen-isotope stage 3 may be compatible with the Laschamp excursion that has been reported in literature with ages scattered between 32 and 50 ka (Nowaczyk et al., 1994).
Site 974 is located in the central Tyrrhenian Sea on the western side of the De Marchi Seamount at 3668 m water depth, ~300 m from ODP Site 652. Twenty-six samples were studied from the uppermost 25 m of sediments, which consist of hemipelagic deposits, mainly nannofossil-rich clay. Volcanoclastic layers occur throughout the cores and several samples exhibit a size fraction of only <40-63 µm, attesting that fine-grained turbidites are frequent in these sediments (Fig. 3). The presence of fine-grained turbidites, together with a low sedimentation rate (about 5.5 cm/k.y. for the Pleistocene-Holocene; Comas, Zahn, Klaus, et al., 1996) and a possible hiatus in the interval studied, accounts for the poor resolution obtained at this site. Nevertheless, a comparison between the paleoclimatic trend and the oxygen-isotope record obtained at Hole 653A, also in the Tyrrhenian Sea (Vergnaud-Grazzini et al., 1990), allows us to attribute our last sample (161-974B-3H-7, 10-12 cm; 25.12 mbsf) to isotope stage 15 (about 0.65 Ma). The warm-water microfaunal assemblage found at 1 mbsf is a record of the oxygen-isotope substage 5e. The sapropel layer recognized in Sample 161-974B-3H-3, 19-21 cm, could be related to the oxygen-isotope stage 13 (0.478-0.524 Ma, according to Imbrie et al., 1984) and probably corresponds to the sapropel S11 or S13.
Rock magnetic parameters measured on the 26 samples collected from Hole 974B do not show any particular trend as can be observed in the susceptibility profile shown in Figure 3. Two peaks in concentration parameters 
, Karm, and SIRM occur at 6.5 and 14.5 mbsf and represent volcanoclastic layers. Interparametric ratios exhibit a narrow range of values, which indicates only minor changes in magnetic grain-size (Fig. 3). The S-ratio is above 0.9 in all the samples (Fig. 3) and the coercivity of the remanence (B0cr) is in the range 30-40 mT, indicating that ferrimagnetic minerals (magnetite-type) dominate the magnetic properties.
Site 975 was drilled in the Balearic Basin in 2415 m water depth. The interval studied (uppermost 22 m) consists of nannofossil ooze with five well-identified organic-rich layers that correspond to the sapropel S5-S9. The total organic carbon content ranges from 0.1% to 1.5% (Table 1), but, in general, is quite low, especially in the uppermost 9 m, which is below 0.3%, the common value for typical deep-sea sediments (McIver, 1975). Only in some of the well-identified sapropel layers do the TOC values exceed 1% (Fig. 4).
The paleoclimatic trend recognized at Site 975 identifies into eight intervals, which correspond to oxygen-isotope stages 1-8 (Last 260 ka, according to Martinson et al., 1987). The age of stage and substage boundaries is listed in Table 3 and shown in Figure 5.
The bottom of the paleoclimatic curve shows low values associated with a cold assemblage (dominance of G. quinqueloba, G. bulloides, G. glutinata, N. pachyderma, and G. scitula), which is related to oxygen-isotope stage 8.
Between 21 and 17.42 mcd, the ameliorating climatic conditions are attested to by abundant frequencies in G. ruber and G. inflata. This interval corresponds to oxygen-isotope stage 7. Positive fluctuations in the paleoclimatic curve are documented at 20.46 and 17.84 m and are characterized by the presence of a warm planktonic foraminiferal assemblage similar to the microfauna recognized in the sapropels S9 and S7 found in the Eastern Mediterranean Sea (Cita et al., 1982). The negative interval around 19.4 m contains a higher percentage of N. dutertrei and corresponds to the stagnant phase related to the deposition of sapropel S8.
Following this interval, the paleoclimatic values vary between -60 and -100 in correspondence to oxygen-isotope stage 6. At 15.81 and 16.37 m, a peak in the frequencies of N. dutertrei and G. bulloides allow us to recognize the two phases that characterize sapropel S6.
The rapid transition between cold and warm conditions at 9.89 m corresponds to the beginning of the isotope stage 5. A significant and rapid cooling is documented by three samples between 8.63 and 8.19 m. This event abruptly breaks the gradual decrease of temperature between substage 5e1 (8 m) and substage 5d (7.05 m). This interpretation is supported by the Blake Event recorded at 8.19 m (Fig. 2).
Above 6 m of depth, a drastic decrease in the paleoclimatic values suggests a rapid transition to oxygen-isotope stage 4. A positive peak observed at 4.44 m contains a sapropel faunal assemblage that we interpret as time-equivalent to the S2 deposition in the eastern Mediterranean, if this sapropel existed (Vergnaud-Grazzini et al., 1977; Muerdter et al., 1984). A rapid shift occurs at the transition from the Last Glacial Maximum (1.32 m) to the Holocene.
The magnetic parameters measured on the 63 samples collected from the cores recovered in the Menorca Rise yield well-defined trends that fit very well with the paleoclimatic record observed at Site 975 (Fig. 6). A very good correlation exists between the KARM profile and the paleoclimatic curve, which suggests that stable single domain (SSD) grains are somehow related to climate (Fig. 7). Glacial stages are also characterized by larger grain size as well as by higher coercivity and lower S-ratio (Fig. 6). The contribution of superparamagnetic (SP) grains, expressed by the fd parameter, is higher during the interglacials 1, 5, and 7 and reaches the maximum value in the uppermost three samples, which belong to the Holocene (Fig. 6). Magnetic mineralogy is dominated by ferrimagnetic minerals; however, as shown by the HIRM profile, the contribution of antiferromagnetic minerals (hematite-type) increases during the glacial stages and seems quite significant during isotope stage 2 (Fig. 6).
Three organic-rich layers well identified in the lithology (9.55, 15.81, and 17.84 mcd) and corresponding to sapropels S5, S6, and S7, exhibit a significant decrease in magnetic concentration, whereas Sample 161-975C-3H-5, 56-58 cm (20.46 mcd), corresponding to sapropel S9, yields a significant content of magnetite. However, the sample is characterized by an increasing magnetic hardness, as suggested by higher value of B0cr.
Site 976 is located in 1107 m water depth in the Alboran Sea about 60 km south of the Spanish coast and only 110 km west of the Gibraltar Strait. From an oceanographic point of view, it is very sensitive to the Atlantic-Mediterranean water-mass exchange. The sedimentation rate is high, ranging between 30 and 50 cm/k.y. (Comas, Zahn, Klaus, et al., 1996). The 45 m of sediment studied at Hole 976C is represented by open-marine hemipelagic facies of nannofossil-rich clay without any significant color change. The TOC measured on the samples studied is relatively high with values ranging between 0.35%-1.15%, averaging around 0.6% (Fig. 4; Table 2).
The agreement between the paleoclimatic curve and oxygen-isotope record (Fig. 8) indicates that our record spans the last six isotope stages. The depth and age of stage and substage boundaries are listed in Table 4 and shown in Figure 8.
The lower part of the section investigated shows low values in the paleoclimatic curve, which are related to the oxygen-isotope stage 6, as indicated by the dominance of cold species such as: N. pachyderma, G. quinqueloba, and G. bulloides. The warming indicated by the occurrence of G. ex gr. ruber, O. universa, and presence of G. inflata (a low percentage) characterizes the glacial/interglacial transition equivalent to the oxygen-isotope stage 6/5 boundary.
Between 40.70 m and 27.16 m the paleoclimatic values are characterized by a predominance of warm assemblages with high-amplitude variations. The major positive climatic cycles are related to oxygen-isotope substages 5e, 5c, and 5a. During these oxygen-isotope substages at 39.1-38.65 m, 34.18-33.05 m and 29.42 m, we have recognized three "sapropel-layer assemblages," corresponding respectively to the time deposition in the Eastern Mediterranean Sea of sapropels S5, S4, and S3 (Cita et al., 1977; Vergnaud-Grazzini et al. 1977). A decrease in the percentage of warm-water species allow us to identify oxygen-isotope cold substages 5d at 36.10 mbsf and 5b at 31.17 mbsf.
The transition between oxygen-isotope stages 5/4 is not straightforward. The paleoclimatic record exhibits an oscillation that cannot be unequivocally associated to the warm or cold stage (Fig. 8). Upon consideration of the sedimentation rate, we regard this swing as a warm oscillation at the base of isotope stage 4. In a speculative way, we can correlate this interval to the interstadial 20 in the GRIP ice core, (Dansgaard, 1993), which corresponds also to the Ognon II warm event at the Gran Pile peat bog in France (Woillard, 1978). Oxygen-isotope stage 4 is characterized by high frequencies of polar species G. quinqueloba (peak values 50%-70%) and the dominance of cold indicators such as G. bulloides, N. pachyderma, and G. glutinata.
The time interval attributed to oxygen-isotope stage 3 shows high frequency climate fluctuations. This time interval is characterized by the presence of G. inflata and by marked fluctuations in N. dutertrei distribution. An interval with warm species between 20.99 and 20.10 mbsf yields a "sapropel planktonic fauna" that corresponds to an insolation maximum (von Grafenstein et al., Chap. 37, this volume) that would be coeval with the sapropel S2 in the Eastern Mediterranean, if this sapropel existed. It shows also the highest peak in the TOC profile (Fig. 4). The transition between oxygen-isotope stages 3/2 corresponds to a rapid decrease of G. inflata and a general cooling trend. The Last Glacial Maximum (18 ka) is identified at 7.5 mbsf and is characterized by the exclusive presence of cold species N. pachyderma (left and right coiling), G. quinqueloba, G. bulloides, G. glutinata, and G. scitula. The high sampling resolution in the upper part of the core allows us to recognize several well-defined climatic steps discussed in Pujol and Vergnaud-Grazzini (1989) in the same area.
The gradual warming during the deglaciation is characterized by different climatic phases related to the changing hydrological setting, as suggested by diachronous peaks in frequency of G. bulloides, N. pachyderma, and G. quinqueloba. The cold, short interval between 4.3-3.8 mbsf corresponds to the Younger Dryas event. This correlation is supported by pollen data (Combourieu Nebout et al., Chap. 36, this volume) and also by the finding of a Malletia shell in Sample 161-976C-1H-3 80-82 cm (3.8 mbsf). This bivalve protobranch is typical of cold waters and lives today in arctic regions (M. Taviani, pers. comm.).
The warm climatic variation at 2.50-2.82 mbsf shows a sapropel planktonic assemblage, which correlates to sapropel S1. The planktonic microfaunal association recognized in the uppermost samples is very similar to the present-day assemblage.
Each of the parameters that reflects variations in the magnetic concentration (, SIRM, HIRM), shows the highest value in the uppermost 3.5 m of sediments that belong to the Holocene (Fig. 9). This interval is also characterized by peak values (6%-9%) in the frequency dependent susceptibility. A drastic decrease in magnetic concentration together with increasing grain size, peak values in coercivity of the remanence, and minimum in S-ratio (S-0.3T) distinguish the interval 3.5-5 mbsf (Fig. 9), which includes the Younger Dryas. Similar trends in the magnetic properties are observed throughout the sequence studied. In fact, we have some intervals in which rock-magnetic parameters drop to very low values, with the greatest reduction observed in the case of the KARM parameter (Fig. 7). These fluctuations coincides with similar trends in the concentration-independent parameters (Karm/; SIRM/, and ARM/SIRM) and in the S-ratio (Fig. 9). The data indicate that intervals corresponding to low-magnetic content are also characterized by a coarsening of the grain size and an increase in the ratio of the goethite/hematite to magnetite. The longest interval showing the above-mentioned characteristics spans most of the oxygen-isotope stage 3, while a primary contribution of ferrimagnetic minerals occurs in samples deposited during oxygen-isotope stage 5 (Fig. 7).
These characteristics are typical of sediments that have undergone a significant diagenetic loss of magnetite. Selective dissolution of magnetite has been recognized in several suboxic/anoxic environments all around the world (e.g., Karlin and Levi, 1983; Canfield and Berner, 1987; Karlin, 1990; Tarduno and Wilkison, 1996) and seems to have been active also in the Alboran Sea. The organic-matter content of these sediments is higher than in deep-sea sediments and pyrite has been observed in several samples, which implies that sulfate reduction is an active process. Organic-carbon-rich sediments can be obtained by either an increasing organic productivity or an increased preservation of the organic matter under oxygen-deficient deep-water conditions (Stein, 1990). In order to decide which of these mechanisms is responsible for the organic matter accumulation, the relationship between the organic carbon and the sulfur content may be helpful (Leventhal, 1983; Berner, 1984). According to these authors, in normal-marine, fine-grained, detrital sediments, the mean C/S ratio is 2.8. A plot of the TOC vs. total S (Fig. 10) shows that anoxic conditions prevailed during the deposition of several samples. In comparison, Site 975 exhibits anoxic conditions only during the deposition of sapropel layers.
