344 
RELATIVE ANTIQUITY 
[Ch. XXIV. 
notwithstanding the horizontality of the tertiary formations 
of that age. 
In order to illustrate the grave objections above advanced^ 
which are aimed at the validity of the whole of de Beaumont's 
reasoning, let the reader suppose, that in some country three 
styles of architecture had prevailed in succession, each for a 
period of 1000 years ; first the Greek, then the Roman, and 
then the Gothic ; and that a tremendous earthquake was known 
to have occurred in the same district during some part of the 
three periods, — a shock of such violence as to have levelled to 
the ground every building. If an antiquary, desirous of dis- 
covering the date of the catastrophe, should first arrive at a 
city where several Greek temples were lying in ruins and half 
engulphed in the earth, while many Gothic edifices were 
standing uninjured, could he determine on these data the era 
of the shock ? Certainly not. He could merely affirm that 
it happened at some period after the introduction of the Greek 
style, and before the Gothic had fallen into disuse. Should he 
pretend to define the date of the convulsion with greater pre- 
cision, and decide that the earthquake must have occurred in 
the interval between the Greek and Gothic periods, that is to 
say, when the Roman style was in use, the fallacy in his reason- 
ing would be too palpable to escape detection for a moment. 
Yet such is the nature of the erroneous induction which we 
are now exposing. For, in the example above proposed, the 
erection of a particular edifice is not more distinct from the 
period of architecture in which it may have been raised, than is 
the deposition of chalk, or any other set of strata, from the geolo- 
gical epoch to which they may belong. Yet, if on these grounds 
we are compelled to include in the interval in which the ele- 
vation of each chain may have happened, the periods of those 
two classes of formations before alluded to, the deranged and 
the horizontal, it follows that, even if all the facts appealed to 
by de Beaumont are correct, his intervals are of indefinite 
extent. He is not even warranted in asserting that the chain 
