s by
a low passage running eastwards from the central apse, perhaps not at
all.

The plan carried out by Matheus took the form of a rectangular hall
enclosing the central apse and the two smaller apses to the north and
south, but leaving--now at any rate--a space between it and the side
apses. Possibly the original intention may have been to pull down the
two side apses, and so to form a square ambulatory behind the high altar
leading to the great octagon beyond; but if that were the intention it
was never carried out, and now the only entrance is through an
insignificant pointed door on the north side.

The walls of the Pateo with their buttresses, string courses and parapet
are so exactly like the older work as to suggest that they may really
date from the time of Dom Duarte, and that all that Matheus Fernandes
did was to build the vault, insert the windows, and form the splendid
entrance to the octagon; but in any case the building was well advanced
if not finished in 1509, when over the small entrance door was written,
'Perfectum fuit anno Domini 1509.'

Two windows light the Pateo, one looking north and one south. They are
both alike, and both are thoroughly Manoelino in style. They are of two
lights, with well-moulded jambs, and half-octagonal heads. The
drip-mould, instead or merely surrounding the half octagon, is so broken
and bent as to project across it at four points, being indeed shaped
like half a square with a semicircle on the one complete side, and two
quarter circles on the half sides, all enriched by many a small cusp and
leaf. The mullion is made of two branches twisting upwards, and the
whole window head is filled with curving boughs and leaves forming a
most curious piece of naturalistic tracery, to be compared with the
tracery of some of the openings in the Claustro Real. (Fig. 58.)

No doubt, while the Pateo was being built, the great entrance to the
Imperfect chapels, one of the richest as well as one of the largest
doorways in the world, was begun, and it must have taken a long time to
build and to carve, for the lower part, on the chapel side especially,
seems to be rather earlier in style than the upper. The actual opening
to the springing of the arch measures some 17 feet wide by 28 feet high,
while including the jambs the whole is about 24 feet wide on the chapel,
and considerably more on the Pateo side,--since there the splay is much
deeper--by 40 feet high. To take the chapel side fir