PALEOMAGNETIC INVESTIGATION OF MIDDLE DEVONIAN LIMESTONES OF ALGERIA AND THE GONDWANA RECONSTRUCTION

We have carried out a palaeomagnetic investigation of the upper middle Devonian marly limestone of the Hazzel Matti Formation outcropping at Meredoua in the northern Ahaggar in the Sahara Desert. The Givetian a ge of the formation is well constrained by palaeontological arguments. The formation has been deformed later than the Moscovian (upper middl e Carboniferous). In the altered yellowish samples and in the heteroge neous sample made of mixed fresh and altered zones, several obviously secondary magnetization components are present. In the freshest blue-g rey samples a SE down-dipping component can be isolated, particularly when a combined thermal and AF treatment is applied. This component, o bserved in 53 samples from three sites, passes the fold test of McFadd en & Lowes (1981) and McFadden & Jones (1981) at the 95 per cent proba bility level. The mean direction for the three sites after tilt correc tion is: D(S) = 118.9-degrees, I(s), = 34.2-degrees, k = 378, alpha95 = 6.3-degrees, which corresponds to a south pole (MER) situated at phi (p) = 61.7-degrees, lambda(p) = -16.2-degrees, A95 = 4.2-degrees and p laces the site area under a palaeolatitude of - 18.8-degrees. However, the best grouping of the directions is achieved after 85 per cent of unfolding using a statistical method based on a bivariate extension of the Fisher's statistics (Legoff 1990; Legoff, Henry & Daly 1992). Suc h a tilt correction changes the pole position by 1.5-degrees. As the a lpha95 and K curves obtained from stepwise unfolding display a broad m inimum and maximum value respectively, and because there was probably slight original syn-sedimentary dips, it is likely that 85 per cent of tilt correction is not significantly different from complete unfoldin g. The magnetic behaviour of the samples upon thermal and AF demagneti zation, the evolution of the initial susceptibility upon thermal treat ment, hysteresis cycles and investigation of the magnetic mineralogy t hrough microscope examination and electron microprobe analyses lead to the following conclusion: greigite was probably a primary magnetic mi neral formed during the diagenesis of the rock. However, its primary m agnetization has been totally replaced by a VRM, except perhaps in rar e samples in which the SE component can be found in the narrow range o f 300-350-degrees-C. Magnetite is thought to carry the SE magnetizatio n component and may have formed early during the diagenesis or soon af ter greigite. To consider the fold test negative would assign the MER pole a post-Moscovian age and place it in a situation inconsistent wit h the other poles of the West Gondwanian APWP, unless all the middle C arboniferous and probably lower Permian African poles are rejuvenated. We rather favour the interpretation of a positive fold test, which ho wever does not constrain the magnetization to be of primary Givetian a ge; the SE component may as well have been acquired any time between t he Givetian and at least the late Moscovian. Assuming that the MER pol e has a Givetian age implies that the Rheic Ocean, which separated Lau russia from Gondwana in the early Palaeozoic, was almost closed in the middle Devonian, in agreement with palaeobiogeographic arguments; but it increases the discrepancy between the APWP of respectively West an d East Gondwana. An intermediate lower Carboniferous age might be in b etter agreement with both African and Australian data. Clearly, howeve r, more Gondwanian poles are needed to test the reliability of both We st and East Gondwanian APWP, and check simultaneously the possibility of relative displacements between these two blocks during the Devono-C arboniferous time period.