Authors' reply

The main criticism by Tucholke and Sibuet on our paper1 concerns the position, nature and age of the J anomaly that occurs on both sides of the Newfoundland–Gibraltar Fracture Zone. Despite the fact that Tucholke previously proposed a 100 km offset for the J anomaly across the Newfoundland–Gibraltar Fracture Zone2, Tucholke and Sibuet argue that this anomaly is “collinear” and thus formed at the same time (∼125 Myr ago; chron M0). However, at that time, the nature of the crust north and south of the Newfoundland–Gibraltar Fracture Zone was different. To the south excess magma resulted in a thickening of the oceanic magnetic layer2 and amplification of its associated magnetic anomaly (M0), whereas the northern domain was formed by exhumed subcontinental mantle and polyphase magmatic activity, as evidenced by rocks from Ocean Drilling Program (ODP) sites 1277 and 1070 (ref. 3). To account for the drilling and seismic results, we proposed that this excess magmatic event occurring south of Newfoundland–Gibraltar Fracture Zone at around M0 time2 propagated into the Newfoundland–Iberia rift and finally resulted in onset of seafloor spreading in the northern part (ODP sites 1277 and 1070) at about 112 Myr ago. Such a mechanism for break-up has been previously proposed4, and we suggested that the J anomaly is the signature of this diachronous propagating event1.

The consequence of our finding is that the J anomaly does not result from a sharp boundary as expected from a reversal of Earth's magnetic field, but rather corresponds to a ∼20- to 40-km-wide zone of magma that overprints previously exhumed mantle rocks of unknown age. In contrast to true oceanic crust, the basement underneath the J anomaly is polyphased. This implies that it can neither be viewed as a conventional seafloor-spreading anomaly nor used to constrain plate reconstructions of Iberia and North America. Therefore, our model invalidates the assumptions used to construct Tucholke and Sibuet's Fig. 1a,b and their related ensuing arguments.

Tucholke and Sibuet's Fig. 1c could potentially provide a first approach to alternative palaeogeographic restorations before about 84 Myr ago (chron C34). Nevertheless, this interpretation is also problematic because it requires a 112 Myr age for the crust at ODP holes 1277 and 1070. Yet, the oldest rocks drilled at those locations are 128 Myr old. Moreover, the offset of the Newfoundland–Gibraltar Fracture Zone used to construct this figure is the same as the one observed at present, but, in the absence of robust magnetic anomalies, this point remains unconstrained for times older than chron C34.

To conclude, we note that the classical interpretation using the J anomaly as a seafloor-spreading anomaly of M0 age, as suggested by Tucholke and Sibuet, requires an up-to-600-km-wide domain between Iberia and Europe to have been subducted during the subsequent convergence that led to the present-day Pyrenees5. However, hints for the existence of such a subduction zone can neither be observed in the field6 nor in the most recent tomographic images7. Thus, we do believe that our model is a first step to elucidating inconsistencies in the conventional plate kinematic reconstructions within the southern North Atlantic, as well as in the Pyrenean domain.