letters to nature NATURE | VOL 409 | 18 JANUARY 2001 | www.nature.com 331 Acknowledgements We thank P. Jenden and M. Laroche for discussions, and L. Price and B.Marty for comments and suggestions. This work was supported by the US Department of Energy, the Gas Research Institute project ‘Advanced stable isotope techniques’ and the ETH, Zu ¨rich. Correspondence and requests for materials should be addressed to C.J.B. (e-mail: ballentine@erdw.ethz.ch). ................................................................. The electric Moho Alan G. Jones* & Ian J. Ferguson² * Geological Survey of Canada, Ottawa, Ontario, Canada, K1A 0E9 ² Department of Geological Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, R3T 2N2 .............................................................................................................................................. Since Mohorovic ˇic ´ 1 discovered a dramatic increase in compres- sional seismic velocity at a depth of 54 km beneath the Kulpa Valley in Croatia, the ‘Moho’ has become arguably the most important seismological horizon in Earth owing to its role in defining the crust–mantle boundary. It is now known to be a ubiquitous feature of the Earth, being found beneath both the continents and the oceans, and is commonly assumed to separate lower-crustal mafic rocks from upper-mantle ultramafic rocks. Electromagnetic experiments conducted to date, however, have failed to detect a corresponding change in electrical conductivity at the base of the crust, although one might be expected on the basis of laboratory measurements 2 . Here we report electromag- netic data from the Slave craton, northern Canada, which show a step-change in conductivity at Moho depths. Such resolution is possible because the Slave craton is highly anomalous, exhibiting a total crustal conductance of less than 1 Siemens—more than an order of magnitude smaller than other Archaean cratons. We also found that the conductivity of the uppermost continental mantle directly beneath the Moho is two orders of magnitude more conducting than laboratory studies on olivine would suggest, inferring that there must be a connected conducting phase. Earth materials conduct electricity predominantly by the flow of ions in fluids and the flow of electrons in solids. Dry silicate rocks are highly resistive, so a region of enhanced electrical conductivity represents an interconnected network of a fluid and/or mineral conducting phase. Laboratory studies suggest that for dry rock assemblages there may be an observable difference in conductivity between deep crustal mafic rocks and upper-mantle ultramafic rocks 2 . However, a convincing step-change in conductivity at the base of the crust has not previously been reported because of one still inadequately explained feature of the Earth; the enhanced electrical conductivity of much of the continental lower crust 3 . This characteristic of the deep crust has been observed globally over the past 30 years, principally using the natural-source magneto- telluric (MT) method, and explanations for its existence remain controversial 3 . Suggestions of an interconnected brine below the brittle–ductile transition 4,5 have been met with scepticism on petrological grounds 6 . A counter suggestion of an interconnected, thin, grain-boundary carbon film 7 also has its critics 8 . Notwith- standing an explanation of its cause, a direct consequence of the existence of this lower-crustal conducting layer is that it is virtually impossible to determine its thickness, and hence derive total crustal thickness using electromagnetic (EM) methods. Where the Earth consists of horizontal layers the external sources induce only horizontal electric currents in these layers, and the MT method is incapable of resolving independently the conductivity and thickness of a conductive layer sandwiched between two Slave craton C a n a d i a n s h i e l d Wopmay fault zone Central site Yellowknife Great Slave lake P a l a e o z o i c p l a t f o r m G r e a t S la v e l a ke s h ea r z o n e T h e l o n f r o n t T a l t s o n M a g m a tic z o n e 0 100 km 61° 62° 117° 1,000 km Figure 1 Tectonic map of the Slave craton in the northwestern part of the Canadian Shield. Also shown is the location of the profile and the central site (star). –3 –2 –1 0 1 2 3 4 103 104 106 107 108 109 105 Degrees –3 –2 –1 0 1 2 3 4 log [frequency (Hz)] W E Kilometres 0 10 20 30 40 78 24 42 60 YK Rae log [frequency (Hz)] Figure 2 Contoured magnetotelluric phase responses for the seven sites from Yellowknife to the surface expression of the western boundary of the Slave craton. Distance is along the abscissa, and log(frequency) is along the ordinate. Top, phases with the electric component directed N41W. Bottom, phases with the electric component directed N49E. © 2001 Macmillan Magazines Ltd