CO2 and/or H2O were reduced to CO/H2 in micro-tubular solid oxide electrolysers with yttria-stabilized zirconia (YSZ) electrolyte, Ni-YSZ cermet cathode and strontium(II)-doped lanthanum manganite (LSM) oxygen-evolving anode. At 822°C, the kinetics of CO2 reduction were slower (ca. −0.49Acm−2 at 1.8V) than H2O reduction or co-reduction of CO2 and H2O, which were comparable (ca. −0.83 to −0.77Acm−2 at 1.8V). Performances were improved (−0.85 and −1.1Acm−2 for CO2 and H2O electrolysis, respectively) by substituting the silver current collector with nickel and avoiding blockage of entrances to pores on the inner lumen of micro-tubes induced by silver paste applied previously to decrease contact losses. The change in current collector materials increased ohmic potential losses due to substituting the lower resistance Ag with Ni wire, but decreased electrode polarization losses by 80–93%. For co-electrolysis of CO2 and H2O, isotopically-labelled C18O2 was used to try to distinguish between direct cathodic reduction of CO2 and its Ni-catalysed chemical reaction with hydrogen from reduction of steam. Unfortunately, oxygen was exchanged between C18O2 and H216O, enriching oxygen-18 in the steam and substituting oxygen-16 in the carbon dioxide, so the anode off-gas isotopic fractions were meaningless. This occurred even in alumina and YSZ tubes without the micro-tubular reactor, i.e. in the absence of Ni catalyst, though not in quartz tubes. Unfortunately, larger differences between the thermal expansion coefficients of quartz and YSZ precluded using a quartz tube to house the micro-tubular reactor. However, the kinetic results, CO/H2 yields from off-gas analysis, diffusional considerations and model predictions of reactant and product gas adsorption on Ni suggested that syngas should be produced by electrochemical reduction of steam to H2, followed by its Ni-catalysed chemical reaction with CO2.