The hypothesis that animals navigate magnetically is attractive because the
earth's magnetic field provides consistent information about position and
direction in all environments through which animals travel. However, the hy
pothesis has been difficult to test because (i) the structure and function
of the sense could not readily be analyzed in the laboratory and (ii) the e
ffects of experimental treatments on behavior in the laboratory and field c
ould not be reliably predicted. Our research is focused on the structure an
d function of the sense in rainbow trout and on applying the understanding
gained in the laboratory to studies of navigation by homing pigeons. We hav
e found iron-rich crystals (most likely single-domain magnetite) in candida
te magnetoreceptor cells located within a discrete layer of sensory tissue
in the nose of rainbow trout. The candidate receptor cells are closely asso
ciated with a branch of the trigeminal nerve that responds to changes in in
tensity but not direction of magnetic fields. In parallel work, we have dev
eloped a model of magnetic position determination in which pigeons derive m
agnetic analogues of geographic latitude and longitude from (i) the total i
ntensity and (ii) the direction of the intensity slope of the earth's magne
tic field. Taken together with our other results, the model gives us confid
ence that a coherent understanding of the structure, function, and use of t
he magnetic sense in animals is now developing. (C) 2000 American Institute
of Physics. [S0021-8979(00)93108-6].