Thermal noise holds the pigeon's proposed ear compass to 0.15 bits per second

Induction in a semicircular canal is ordinary physics. A conducting loop turning in a magnetic field develops a voltage, and the canals are rings of salty fluid that conducts electricity, sitting inside a head that turns ^[2]. The proposal was that this voltage can be read.

Reading it is where the trouble starts. Charges in the fluid jitter with heat, and that jitter shows up as voltage fluctuations on the same terminals that would have to carry the signal ^[5]. A compass has to match a change in signal to a particular head direction ^[7]. The usual answer to the noise is a narrow filter, which does cut noise. It also cuts the rate at which information can arrive, and the bird's head keeps turning while it does ^[6].

So Kattnig priced the sensor in bits per second. His example task, a rapid scan able to tell apart directions 5 degrees apart, needs more than 560 bits per second ^[9]. Divide that by the 0.15 bits per second the idealised canal managed and the sensor comes out about 3,700 times too slow ^[10]. Allowing charge to leak across the cupula weakened the signal further ^[11]. The shortfall is in information, so more sensitive cells and extra brain processing cannot recover a direction the sensor never encoded in time ^[12].

None of this touches the biology. In 2019, researchers publishing in Current Biology generated electrical signals in an enlarged laboratory model of a canal and identified molecular components associated with electrical sensing in pigeon inner ear tissue ^[13]. The enlargement matters: what produced the signal was a scaled bench version of a canal.

In the experiments that strengthened the case, the birds' heads were held still and the magnetic field was rotated around them, changing its direction relative to the inner ear without the bird turning its head ^[15]. That design isolates the field change from self-motion, which is what an electrophysiologist wants. It also leaves out the scanning behaviour the compass account depends on.

Kattnig's study is theoretical, and it examined one arrangement ^[1][18]. Its conclusion is that another sensing arrangement or a different physical process would be needed, and that whatever replaces induction has to account for the brain activation as well as supply enough directional information to navigate ^[17]. The analysis challenges the induction mechanism while leaving unexplained why magnetic fields activate parts of the pigeon brain connected to the inner ear ^[16].