Paleomagnetic data frequently display bipolar data that are not exactl
y antipodal. One cause of this biasing is the addition of another late
r (secondary) magnetization that cannot be removed by laboratory demag
netization experiments. The magnitude of this hidden secondary magneti
zation can be calculated when only two values are known: the non-antip
odal angle (measured), and the angle between the Normal direction (mea
sured) and secondary direction (measured or assumed). Furthermore, a m
inimum magnitude of the secondary magnetization can be calculated with
knowledge only of the nonantipodal angle. Applied to three published
data sets (sandstone, limestone and diorite intrusion) with non-antipo
dal angles of 11 degrees, 28 degrees and 11 degrees respectively; a se
condary component is calculated at 25%, 28% and 16% of the characteris
tic magnetization remaining after laboratory cleaning. Near surface al
teration (weathering) appears to be the source of secondary magnetizat
ion in two of these cases. A directionally consistent, biasing effect
of a few degrees is made on the mean direction as compared to the conv
entional calculation of averaging the nonantipodal directions.