We consider the hypothesis that galactic magnetic fields are primordia
l. We also discuss the various objections to this hypothesis. To test
this hypothesis properly we assume that there was a magnetic field pre
sent in the galactic plasma before the galaxy formed and explore how s
uch a field would evolve assuming a specific model for the interstella
r medium in the galactic disk. After the galactic disk formed, the lin
es of force thread through it and remain connected to the external cos
mic medium. They enter through one side of the disk, proceed horizonta
lly a distance l in the disk, and then leave through the other side. W
e find that the lines of force are stretched by the differential rotat
ion of the galactic disk, amplifying the toroidal component of the fie
ld and increasing l. When the magnetic field is strong enough, it prod
uces ambipolar velocities that try to lift the line out of the galacti
c disk but in opposite directions on different parts of the line. The
result is that, instead of the line being expelled from the disk, its
horizontal length I is shortened, both in the radial and in the toroid
al direction. This leads to a reduction of the rate of horizontal stre
tching and finally a reduction in the magnetic field strength. After a
sufficient time, the magnetic field at all points goes through this r
eduction and the field strength approaches a universal function of tim
e. This function is slowly decreasing and only depends on the ambipola
r properties of the interstellar medium. At any given time the magneti
c field is toroidal and has the same strength everywhere. On the other
hand, it turns out that its direction varies rapidly with radius, cha
nging sign every 100 parsecs or so. However, if the initial cosmic mag
netic held is not uniform, the areas of one sign of the toroidal field
dominate over the other. The resulting field has a net Faraday rotati
on. If such a held were observed with low resolution in an external ga
laxy, then the field would appear toroidal in between the spiral arms.
The spiral density wave would turn it so that the lines appear to tra
ce out the spiral arm, although the apparent lines really are the sum
of pieces of magnetic lines as they cross the disk. They do not necess
arily extend very far along the arms. We contend that this model of th
e magnetic held, which arises naturally from a primordial origin, can
fit the observations as well as other models for the magnetic held, su
ch as those arising from the mean field dynamo theory. Finally, becaus
e the field lines are topologically threaded through the disk, they ca
nnot be expelled from the disk. This counters the objection against th
e primordial origin, namely that such a field could not survive very l
ong in the galaxy.