Nature 421, 495 (2003)

My Brief Communication about thermoelectricity in shark gels neglected a systematic effect of surface electrochemistry: electrode potentials vary with temperature in electrolyte solutions. However, silver leads in sea water1 and accepted values for likely electrode reactions2 show a sign opposing the gel signals, making it unlikely that an artefactual signal is the origin. Our subsequent work3 discussed artefacts and repeated the signal with platinum electrodes. Although another report4 finds a zero signal using salt bridges, it ignores thermopower in gel-filled leads, which risks building a ‘null thermocouple’ from two similar materials (see ref. 5, for example). A temperature function of the electrosensors is not known, but the thermoelectric transduction hypothesis still stands.