of
prayer.

After Jocelin's death the works came to a standstill, for the
sufficient reason that the chapter was "overburdened with an
intolerable debt," owing to the enormous expense of the litigation
with Bath Abbey over Bishop Roger's election (p. 153). This, however,
was the last attempt of the rival cathedral of St. Peter; and the
debt, which was at its worst in 1248 (the year after Roger's death),
was bravely met by a contribution of a fifth of the income of each
prebend, as well as by gifts and obits; so that towards the end of
William Bytton's episcopate the debt was nearly cleared, and in 1263
Bytton made over the sequestrations of vacant benefices to the fabric
fund.

In 1248 an earthquake had done much damage, shaking down the _tholus_
(either the vault, or the stone capping) of the central tower, as we
learn from Matthew Paris _(Hist. Angl._ iii. 42). Accordingly, in
1263, preparations were made for further building; and in 1286 we hear
of a chapter meeting, summoned by Dean Thomas Bytton, whereat the
canons bind themselves to give one-tenth of their prebends for five
years, "to the finishing of the works now a long time begun (_jam diu
incepta_), and to repair what needed reparation in the old works."

The reparation here mentioned refers in all probability to the roof
and piers of the transepts and eastern part of nave, damaged by the
fall of the _tholus_. The famous western capitals of the transepts,
with their frequent representations of the miseries of toothache, must
refer to the second William Bytton, who had died in 1274, and whose
tomb became famous for its dental cures (p. 125). No doubt, the
offerings at the shrine of this local saint helped considerably to
swell the funds for the building operations.

The works "now a long time begun" can hardly be anything else than the
chapter-house undercroft, the outer walls of which may have been built
some forty years before. Professor Willis, who had access to the
document, decided, on architectural evidence, that the undercroft must
have been already completed at this time, and his view may be safely
accepted (_Arch. Inst._, "Bristol" vol., p. 28). The passage to the
undercroft would seem to be the first result of the chapter's
undertaking; its ornament is of a more advanced type than that of the
undercroft itself, and one of its carved heads is swollen as by the
toothache, and tied in a handkerchief. There can be little or no doubt
that the "finis