a
much-disputed question, some, as for instance Artis, the excavator of
the Castor potteries in Northamptonshire, claiming that it was a
natural result obtained in the baking, after polishing of the surface,
by means of specially contrived kilns. But it is now generally agreed
that it was artificial. It is true that the Roman lamps and many of
the commoner wares have a gloss produced by polishing only, varying in
colour and brightness with the proportion of iron oxide in the clay
and the degree of heat at which the pieces were fired. But the surface
finish of the finer or _terra sigillata_ wares is something quite
distinct, and reaches a high and wonderfully uniform perfection.

It is possible that the technical secret of the potters of the Roman
world was only a development from the practice of the Greeks, but it
does seem as if the finer Roman wares were coated with a brilliant
glossy coating so thin as to defy analysis, yet so persistent as to
leave no doubt of its existence as a definite glossy coat. Repeated
attempts have been made to determine its nature by analysis, but
chemists ought to have known better, for the coating is so thin that
it is impossible to remove it without detaching much more body than
glaze. Examination shows it to be much more than a surface polish or
than the gloss of the finest Greek vases, and we shall have to wait
for a final determination of its nature until some one who is at once
a chemist and a potter can reconstruct it synthetically. Whatever its
nature and method of production, it is certain that the glaze itself
was a transparent film which heightened the natural red colour of the
clay, until in the finest specimens it has something of the quality of
red coral.

In the manufacture of vases the Romans used the same processes as the
Greeks. They were all made on the wheel, except those of abnormal
size, such as the large casks (_dolia_), which were built up on a
frame. Specimens of potters' wheels have been found at Arezzo and
Nancy, made of terra-cotta, with a pierced centre for the pivot, and
bearing small cylinders of lead round the circumference to give a
purchase for the hand and to aid the momentum of the wheel. For the
ornamental vases with reliefs an additional process was necessary, and
the decoration was in nearly all cases produced from moulds. The
process in this case was a threefold one: first the