thirty-five years without the interruption of a month."--_Hist. of the
Inductive Sciences_, vol. ii. book vii.]

One would have liked to see this elaboration more clearly, to have
been allowed a glimpse into his workshop while he was so engaged.
Unfortunately the editor of his journals has selected the relatively
unimportant records of his earlier studies, and left us in the dark as
regards this far more interesting period. He was such an indefatigable
diarist that it is unlikely that he neglected to keep a journal in
this crisis of his studies. But it has not been published, and it may
have been destroyed. All that we have is this short paragraph in his
Memoirs:--

"The classics, as low as Tacitus and the younger Pliny and
Juvenal, were my old and familiar companions. I insensibly
plunged into the ocean of the Augustan history, and in the
descending series I investigated, with my pen almost always
in my hand, the original records, both Greek and Latin, from
Dion Cassius to Ammianus Marcellinus, from the reign of
Trajan to the last age of the Western Caesars. The subsidiary
rays of medals and inscriptions of geography and chronology,
were thrown on their proper objects, and I applied the
collections of Tillemont to fix and arrange within my reach
the loose and scattered atoms of historical information.
Through the darkness of the middle ages I explored my way in
the _Annals and Antiquities of Italy_ of the learned
Muratori, and diligently compared them with the parallel or
transverse lines of Sigonius and Maffei, Baronius and Pagi,
till I almost grasped the ruins of Rome in the fourteenth
century, without suspecting that this final chapter must be
attained by the labour of six quartos and twenty years."

When the time for composition arrived, he showed a fastidiousness
which was full of good augury. "Three times did I compose the first
chapter, and twice the second and third, before I was tolerably
satisfied with their effect." His hand grew firmer as he advanced. But
the two final chapters interposed a long delay, and needed "three
successive revisals to reduce them from a volume to their present
size." Gibbon spent more time over his first volume than over any one
of the five which followed it. To these he devoted almost regularly
two years apiece, more or less, whereas the first cost him three
years--so disproportionately