Replicating biological information is usually stored only within nucle
ic acid. The existence of 'strains' of agent in prion disease (scrapie
, BSE, CTD has been taken to indicate an independent genome within the
transmissible agent. Other replicable information exists, however, bo
th in biology and elsewhere, including, for example, the 'meme' (the n
eural correlate of ideas which replicate in human brains by communicat
ion) and the computer virus. From this broader viewpoint, we explore t
he possibility that 'strain' differences in prion disease reflect biol
ogical information stored within the prion protein rather than in nucl
eic acid. Much of the disease variation in mice (used as evidence for
strain differences) can be accounted for by the primary structures of
the prion protein of the host (the experimentally infected mouse) and
the donor mouse (from which infectious tissue is taken). Information d
etermining residual disease variation (when these factors have been ex
cluded) map reside in different conformational states of host prion pr
otein, Prion protein can adopt different conformational and glycosylat
ion states. The information which these states contain is only partial
ly conserved on transmission between animals, permitting the appearanc
e of both 'strain stability' and 'strain mutation'. Different sources
of replicating, biological information including information in the 'a
gent' (the abnormal form of prion protein) and in the host prion gene
(PRNP) are in evolutionary competition, We argue that, in the prion di
seases, replicating information is not carried in nucleic acid in eith
er the host or the 'agent' but is carried within the conformational st
ate of the abnormal form of prion protein.