be so clearly indicated, but there will be a more marked
shrinkage in the width of the cloth as well as in the pattern, and
this is what has taken place in the girdle giving us the diminishing
taper.

[Illustration: "THE LINEN GIRDLE OF RAMESES III."

_Reproduced by kind permission of Dr. Clubb, Director, The Museums,
Liverpool._]

If this diminishing taper were indicated by a decrease in the width of
the pattern commencing at the selvedges, then it might be presumed
that a reed had been used for the central portion only--a very clumsy
even if feasible arrangement, but the pattern begins to decrease along
the middle and hence no reed could have been used.

It does not follow that because a loom was not provided with a reed it
was without heddles. Anyone who will examine the large series of
primitive looms at Bankfield Museum, will observe that heddles
preceded reeds; this must necessarily be so as the making of the shed
is the first step in weaving, while the reed's work is more that of a
finisher. But the heddles are all extremely primitive, and in my
experience do not exceed four in number where there is no reed. Such a
quantity of heddles with its complicated harness as Mr. Lee considers
necessary is quite out of the question with a loom so undeveloped as
not to be provided with a reed. Hence the indication is that the
girdle was woven on a loom of a primitive character.

In carrying out the work the weaver has made many mistakes. On the
left hand side of the right hand row of red crosses (they come out
black in the photograph) there is an "end down" for a considerable
distance--that is a thread has been missed.

On the same row of crosses three white threads show above and below,
while on the left hand row of crosses there are five white threads
above and below. The crosses are neither the same size nor shape in
the two columns and curiously their white hafts in both columns point
to the left instead of one row pointing to the left and the other to
the right. Then again the white point at the right apex of the zigzag
on the left corresponds to a red point at the left apex of the right
hand zigzag, but if the girdle had been woven on an advanced loom with
dobby and harness these points would have been red in both places.

As regards the large number of warp threads to the inch which Mr. Lee
puts down as 272-340 (107-134 per cm.), this does not by any means
indicate a complicated piece of machinery for the we