This might have been intended to modify expansion and contraction, as well as to assist the gradual distribution of the colouring vapour. The mouth of the furnace and top of the kiln were, no doubt, stopped; thus we find every part of the kiln, from the inside wall to the earth on the outside, and every part of the clay wrappers of the dome, penetrated with the colouring exhalation. As further proof that the colouring of the ware was imparted by firing, I collected the clays of the neighbourhood, including specimens from the immediate vicinity of the smother kilns. In colour some of these clays resembled the ware after firing, and some were darker. I submitted them to a process similar to that I have described. The clays dug near the kilns whitened in firing, probably from being bituminous. I also put some fragments of the blue pottery into the kiln; they came out precisely of the same colour as the clay fired with them, which had been taken from the side of the kilns. The experiment proved to me that the colour could not be attributed to any metallic oxide, either existing in the clay or applied externally; and this conclusion is confirmed by the appearance of the clay wrappers of the dome of the kiln. It should be remarked, that this colour is so volatile that it is expelled by a second firing in an open kiln.” Fortunately, some of the kilns remained almost entire, and many had been left with the pottery partly packed in them for firing, so that there was no difficulty in understanding the nature of the process here employed by the Roman potters. [Illustration: Fig. 213.] [Illustration: Fig. 214.] [Illustration: Fig. 215.]