16, 17, 18. Nave. Torcello.
19, 20. Transepts. St. Mark's.
21. Nave. St. Mark's.
22. External pillars of northern portico. St. Mark's.
23, 24. Clustered pillars of northern portico. St. Mark's.
25, 26. Clustered pillars of southern portico. St. Mark's.

Now, observe, first, the enormous difference in style between the bases
1 to 5, and the rest in the upper row, that is to say, between the bases
of Murano and the twelfth and thirteenth century bases of Venice; and,
secondly, the difference between the bases 16 to 20 and the rest in the
lower row, that is to say, between the bases of Torcello (with those of
St. Mark's which belong to the nave, and which may therefore be supposed
to be part of the earlier church), and the later ones of the St. Mark's
Facade.

Secondly: Note the fellowship between 5 and 6, one of the evidences of
the early date of the Casa Falier.

Thirdly: Observe the slurring of the upper roll into the cavetto, in 13,
14, and 15, and the consequent relationship established between three
most important buildings, the Rio-Foscari House, Terraced House, and
Madonnetta House.

Fourthly: Byzantine bases, if they have an incision between the upper
roll and cavetto, are very apt to approach the form of fig. 23, in which
the upper roll is cut out of the flat block, and the ledge beneath it is
sloping. Compare Nos. 7, 8, 9, 21, 22, 23, 24, 25, 26. On the other
hand, the later Gothic base, 11, has always its upper roll well
developed, and, generally, the fillet between it and the cavetto
vertical. The sloping fillet is indeed found down to late periods; and
the vertical fillet, as in No. 12, in Byzantine ones; but still, when a
base has such a sloping fillet and peculiarly graceful sweeping cavetto,
as those of No. 10, looking as if they would run into one line with each
other, it is strong presumptive evidence of its belonging to an early,
rather than a late period.

The base 12 is the boldest example I could find of the exceptional form
in early times; but observe, in this, that the upper roll is larger than
the lower. This is _never_ the case in late Gothic, where the proportion
is always as in fig. 11. Observe that in Nos. 8 and 9 the upper rolls
are at least as large as the lower, an important evidence of the dates
of the Casa Farsetti and Fondaco de' Turchi.

Lastly: Note the peculiarly steep profile of No. 22, with reference to
what