Alteration of ultramafic rocks plays a major role in the production of hydrocarbons and organic compounds via abiotic processes on Earth and beyond and contributes to the redistribution of C between solid and fluid reservoirs over geological cycles. Abiotic methanogenesis in ultramafic rocks is well documented at shallow conditions, whereas natural evidence at greater depths is scarce. Here we provide evidence for intense high-pressure abiotic methanogenesis by reduction of subducted ophicarbonates. Protracted (≥0.5–1 Ma), probably episodic infiltration of reduced fluids in the ophicarbonates and methanogenesis occurred from at least ∼40 km depth to ∼15–20 km depth. Textural, petrological and isotopic data indicate that methane reached saturation triggering the precipitation of graphitic C accompanied by dissolution of the precursor antigorite. Continuous infiltration of external reducing fluids caused additional methane production by interaction with the newly formed graphite. Alteration of high-pressure carbonate-bearing ultramafic rocks may represent an important source of abiotic methane, with strong implications for the mobility of deep C reservoirs.

SUMMARY We present results from a study of a partial magnetic polarity transition record from a 9.7 Ma 40Ar/39Ar dated volcanic section on the island of Gomera (Canary Islands, Spain). The record provides results that bear on the debate over whether long‐term persistent features of field behaviour exist during transitions. The sampled section, containing more than 20 flows, begins after the onset of the transition. The nine lowermost flows in the section, which are all transitional, yield virtually identical directions and similar low paleointensities. Following the recovery of virtual geomagnetic pole positions (VGPs) to near the spin axis, paleointensities progress to very high values. Transitional directions lay due west, reflecting the presence of a non‐zonal intermediate field. VGPs for these flows cluster in the west Atlantic, which according to Hoffman would correspond to a zone of fastest P‐wave propagation in the lower mantle and to a radial flux centre of the present‐day non‐dipole field. Hoffman had noted a similar occurrence at two other locations. He found that in a select set of volcanic records VGPs tended to group over two areas (Australia and South America), which also corresponded to fast P‐wave and non‐dipole field anomalies. Given the discontinuous and irregular character typical of volcanic sequences, it is natural to question whether these clusters are merely artefacts of episodic volcanism or actually reflect standstills of the reversing field. We address this question for our record by comparing chemical and magnetomineralogic characteristics of the flows. We find that all of the transitional flows have very similar chemical and rock magnetic properties. Furthermore, these flows differ significantly from the overlying reverse flows, which give distinctly different values and much greater flow‐to‐flow variation. From this we conclude that the packet of transitional flows was derived from the same source and had erupted over a relatively short period of time, as reflected by a magma source that presumably did not have time to evolve significantly. While the VGP cluster in the Gomera record is consistent with long‐term mantle control on transitional fields, we cannot conclude that the cluster represents a standstill of the field. These results highlight the difficulties of interpreting temporal changes in field behaviour from volcanic sequences. In addition, they provide a clear incentive for investigating physical characteristics of lavas to assess relative eruptive rates.