The Siwaliks are continental molasse deposits that extend from the North of Pakistan to Assam in Northeast India along the Himalayan Range. The Siwalik Series have been known worldwide since the 1830s for their Neogene and Quaternary fossil vertebrates, especially in the Upper Indus Basin, with special attention paid to the human origins in this area of the Indian sub-continent (e.g., Dennell, 2010, see review of the history in Dambricourt Malassé et al., 2016a). The 6000 meter thick Siwalik Series is divided into three subgroups: Lower, Middle and Upper Siwalik (see Patnaik, 2013 for a review; Stidham et al., 2014) and further into zones or formations based on Mammalian fauna, called faunal zones (Pilgrim, 1913). These biostratigraphic subdivisions are named Kamlial and Chinji for the Lower Siwalik, Nagri and Dhok Pathan for the Middle Siwalik, Tatrot, Pinjor and Boulder Conglomerate for the Upper Siwalik. Unfortunately, the paleontological record is spatially unequally distributed, and the deposits are characterized by lateral facies variations complicating the regional correlations (Nanda, 2002). To ease this difficulty, a well-developed chronostratigraphic framework was established based on studies of magnetic polarity (Barry et al., 1982, Barry et al., 2012, Keller et al., 1977 and Opdyke et al., 1979), coupled with fission track dating of volcanic tuffs (Johnson et al., 1982).

The type locality of the Tatrot Formation is described in the Potwar Plateau, Pakistan, and the type locality of the Pinjor Formation in a northeastern area of the Chandigarh anticline, near the Pinjaur Township (Fig. 1). Numerous fossil species were exhumed from several localities of the Tatrot and Pinjor Formations (Badam, 1973, Barry et al., 1982, Barry et al., 2012, Gaur and Chopra, 1984, Nanda, 1973, Patnaik, 2003, Patnaik, 2013, Raghavan, 1990, Sahni and Khan, 1964, Sahni and Khan, 1968, Sahni and Mitra, 1980 and Stidham et al., 2014). Four tuffaceous mudstones have been discovered in the Pinjor Formation near the Ghaggar River (Tandon and Kumar, 1984), but only one was dated by fission tracks, providing an age of 2.14 ± 0.5 Ma (Mehta et al., 1993). Thus, this tuffaceous layer became a chronological benchmark to correlate the magnetostratigraphy of the Pinjor Formation with the geomagnetic polarity time scale (GPTS) (Gradstein et al., 2004, Kumaravel et al., 2005 and Tandon et al., 1984).

In other Upper Siwalik localities in Pakistan and India, the chronostratigraphic and paleontological results showed that the uppermost part of the Tatrot Formation provides a paleontological assemblage with new species, especially Equus sivalensis, emerging during the very Late Pliocene (Upper Tatrot Formation) and developing over the Pleistocene (Pinjor Formation) (Nanda, 1994 and Nanda, 2002). This paleontological assemblage is called “transitional fauna” (Sahni and Khan, 1968 and Sahni and Mitra, 1980). At Masol, the top of this transition coincides with the Gauss-Matuyama reversal, dated to 2.58 Ma (Cande and Kent, 1995 and Ranga Rao, 1993).

The studied area is located in the Siwalik Frontal Range (SFR) near Masol village about ten kilometers north of Chandigarh. The structure of the SFR is an anticline parallel to the sub-Himalayan foothills (Fig. 1). The Patiali Rao, a seasonal river, cuts them perpendicularly from north to south, and this fluvial incision exposes a transect showing inclined strata of the Upper Siwalik Subgroup from the source upstream to the Punjab Plain downstream (for more details, see Gargani et al., 2016).

The Oil Natural Gas Commission (ONGC) began a geological survey in 1956; it divided the Upper Siwalik Subgroup of the Chandigarh anticline into five units based on the lithology, which are, from the oldest to youngest: Masol Formation, Rupar Formation I, II, III and IV (Ranga Rao, 1993 and Ranga Rao et al., 1995). Referring to its paleontological assemblage, the Masol Formation was correlated with the Quranwala faunal zone defined by Pilgrim (1913) into the Tatrot Formation, while Rupar I, II and III have been linked with the Pinjor Formation. The conglomeratic unit Rupar IV was correlated with the Boulder Conglomerate. Later, Sahni and Khan, 1964 and Sahni and Khan, 1968 discovered numerous faunal remains in the Masol Formation and confirmed their attribution to the Tatrot fauna. This collection was next enriched by the paleontological findings of Badam, 1973 and Badam, 1979, Gaur (1987) and Nanda (1994) (see a review in Moigne et al., 2016).

Ranga Rao (Ranga Rao, 1993 and Ranga Rao et al., 1995) investigated the lithostratigraphy and magnetostratigraphy of the deposits all along the Patiali Rao and made a synthesis of his results with those of the studies initiated in 1956 in the Chandigarh region (Azzaroli and Napoleone, 1982, Tandon et al., 1984 and Yokoyama, 1981). After all these geological and paleontological investigations, the first terminology of the subdivisions was then replaced by that of Tatrot (Masol Formation), Pinjor (Rupar I, II, III Formations) and Boulder Conglomerate (Rupar IV). Table 1 displays the stratigraphic correlations between these different terminologies used for the Chandigarh anticline.

The 1296-m-thick sequence observed along the Patiali Rao shows alternations of sandstones and claystones easily attributed to Tatrot, then thick benches of sandstone belonging to Pinjor, and finally Boulder Conglomerate on the southern edge of the anticline (Ranga Rao, 1993, Sahni and Khan, 1964 and Sahni and Khan, 1968). Structurally, the Masol 1 locality is located at the top and in the center of the anticline where the oldest layers of the Tatrot Formation outcrop are unearthed. For the magnetostratigraphy study, Ranga Rao (1993) selected fifty-two sites along the Patiali Rao transect. In the laboratory, he cut several 2.5 cm cubes extracted from 3–4 oriented blocks collected from each site. These samples were demagnetized by alternating field and measured using an astatic magnetometer. Ranga Rao et al. (1995), hence, revealed three normal and four reversed magnetozones in the sequence (Fig. 2).

Referring to the biostratigraphic data, the lowest magnetic reversal was assigned to the Gauss-Matuyama reversal, and the highest to the Matuyama-Brunhes. With a calculated sedimentation rate of 0.63 m/1000 years between both magnetic limits (based on the ages of LaBrecque et al., 1977), Ranga Rao (Ranga Rao, 1993 and Ranga Rao et al., 1995) identified the Reunion, Olduvai and Jaramillo magnetozones and, thus, highlighted a continuous sequence ranging from 2.7 Ma to 630 ka.

The Masol 1 stratigraphic sequence, 30 m thick, was completed with the sequence of the neighboring Masol 2 locality. The composite stratigraphy correlates with the uppermost part of the Tatrot Formation, and begins before the Quranwala faunal zone. It is characterized by the alternation of claystone, siltstone and sandstone (Fig. 3) (Chapon Sao et al., 2016). The lithostratigraphic sequence from unit F (lower) to unit A (upper) is composed as follows: unit F: sandy silts, grey in color, 2.40 m thick; unit E: red wine silts, 2 m thick; unit D: orange silts, 1 m thick; unit C: sandstone with white mica, light-grey in color, 1.20 m thick; unit B: coarse sandstone with gravels and clay lenses, cross-bedded, 2.40 m thick; unit A: brown silts, 0.80 m thick (Fig. 3). Unit B sandstone corresponds to the “Elephant sand” (Chapon Sao et al., 2016) and the Masol 1 sequence is equivalent to units c3 through s4 defined by Tudryn et al. (2016).

The bovid tibia shaft with cut marks was discovered on the slopes of a cliff formed by the outcrop of units B to E, less than 10 m below the top of the cliff. The yellow color of the cortical bone, similar to other fossils collected on the same surface of unit D and the fine micaceous sandy crust that covers it, allows researchers to conclude that its lithostratigraphic origin is very likely the yellow claystone D, near the micaceous sandstone C (Dambricourt Malassé et al., 2016b).

Referring to the previous works (Ranga Rao, 1993 and Rendell et al., 1987), the sandy sediments of Tatrot are generally too weakly magnetized to yield reliable paleomagnetic data. Thus, only the finer sediment units were sampled (Fig. 3). Besides sedimentological loose sampling, rectangular prism-shaped blocks were taken as samples from the Masol 1 section with a hammer and chisel. Before separating the blocks from the outcrop, they were coated with plaster bands for consolidation and to avoid fracturing during the transport. Each block was meticulously oriented by marking on the plaster coating the north and vertical lines. As the sampling took place at the top of the anticline, the stratigraphic strata are horizontal. Unit A was discarded from the paleomagnetic study, for its stratigraphic position corresponded to a sub-structural surface exposed to hazards of climatic variations and biological factors (plants, animals, shepherds). The middle and the upper parts of unit B were not sampled because they contained medium or coarse-grained sandstone. Finally, the thick unit E composed of silts was sampled with a first block in the lower part (PPM2) and a second in the upper part (PPM3).

All the paleomagnetic samples were exported to France to be analyzed in the paleomagnetic laboratory of the UMR 7194 CNRS, Department of Prehistory, National Museum of Natural History, located in the Institut de Paléontologie Humaine, Paris. After drying in an oven at 30 °C for one month, the sediment blocks were carefully cut with an amagnetic saw and knife to obtain at least four 8 cm³ cubes per block. The fine sandstone lenses of unit B were too brittle to allow the sampling of an 8-cm³ cube. Twenty-six cubes were made, but 12 were selected: four samples in unit F, six in unit E, and two in unit C.

The magnetic measurements furnace (MMTD 80) were used for the thermal demagnetization at 13 temperature steps, 100, 200, 300, 360, 400, 440, 460, 500, 520, 540, 560, 600 and 660 °C. The Molspin shielded alternating field demagnetizer enabled the removal of the natural remanent magnetization (NRM) by applying 15 steps, 2.5, 5, 7.5, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90 and 100 mT.

Before beginning the demagnetizations, the NRM of each cubic sample was measured with a high-sensibility, low-speed spinner magnetometer JR6 Agico. To determine the stability of the magnetization recorded in the sediment blocks, selected specimens were subjected to alternating field (AF) and thermal demagnetizations. However, as the sediment from the upper unit E and unit C were characterized by a soft magnetization, only the thermal demagnetization that was more effective was used.

The demagnetized data were analyzed using the REMASOFT 3.0 software developed by Chadima and Hrouda (2006).

The intensity of the magnetization was weak but provided reliable data, further strengthened by Fisher statistics (Table 2). For samples from all units, the thermal demagnetization was more effective than the AF demagnetization. No unstable secondary component was highlighted in this study, whereas Ranga Rao (1993) noted the presence of a secondary component, which was removed after 20 mT. The measurements showed that the sediments from the three lithostratigraphic units F, E and C were deposited during a normal magnetozone and recorded a transitional polarity in the lower part of unit E (Fig. 10). This transitional polarity was between two samples recording a normal polarity. The magnetostratigraphy of Ranga Rao (1993) did not reveal this transitional polarity.

The identification of the collected fauna (Moigne et al., 2016) allowed us to correlate the studied sediments with the Quranwala fossiliferous zone (Sahni and Khan, 1964) and, thus, with the upper part of the Tatrot Formation. Any indication to explain the transitional polarity in the lower unit E was missing. By reference to the biostratigraphic data (Moigne et al., 2016) and to the stratigraphic studies (Chapon Sao et al., 2016 and Tudryn et al., 2016), the sediments of the three units presented in this study were deposited during the Gauss magnetozone. Thus, the sediments can be considered older than the Gauss-Matuyama reversal, dated to 2.58 Ma (Cande and Kent, 1995 and Gradstein et al., 2004).

Many paleomagnetic studies have been performed in the Upper Siwalik Subgroup in Pakistan and India aiming to build a chronological framework for the paleontological sites and, therefore, establish correlations between them. The latest prehistoric discoveries in the Masol inlier (district Mohali, Punjab) led to the launch of a multidisciplinary program of research. The new magnetic polarity studies of samples selected from the four stratigraphic units at the Masol 1 locality allowed us to precisely locate the stratigraphic position of these discoveries in relation to the Gauss-Matuyama reversal. During the sampling process, the thermal demagnetization was more effective than the AF demagnetization and provided reliable results. Our study demonstrated that the deposits mostly recorded a normal and a transitional polarity. In addition, with data from the paleontological assemblage, these results showed that the investigated sequence was, therefore, older than the Gauss-Matuyama reversal, dated to 2.58 Ma.