There follows on this what is perhaps the earliest extant description of a glass furnace. ‘The furnace of the glass-makers should have six compartments, of which three are disposed in stories one above the other.... The lower compartment should be deep, in it is the fire; that of the middle story has an opening in front of the central chambers—these last should be equal, disposed on the sides and not in the centre (?), so that the fire from below may rise towards the central region where the glass is and heat and melt the materials. The upper compartment, which is vaulted, is arranged so as uniformly to roof over the middle story; it is used to cool the vessels after their manufacture.’[85] We have also the description of a smaller furnace, which is perhaps that in which the more fusible glass for enamels and minor objects of _verroterie_ was melted. Finally, an oven with a floor of brick-earth is mentioned, for fritting the sand and alkali. In spite of much that is obscure in this description, we can trace in it the general type of furnace which, doubtless handed down from Roman times, has survived in places with few important changes to the present day. [Illustration: PLATE XIX MEDIÆVAL GLASS FURNACE FROM AN ILLUMINATED MANUSCRIPT OF RABANUS MAURUS] And here I may call attention to a contemporary drawing of a mediæval glass furnace—a source of information as unique as it is unexpected. This is to be found in a manuscript of an encyclopædic work, _De Originibus Rerum_, compiled by Rabanus Maurus, one of the earliest of the schoolmen. Rabanus lived in the Benedictine monastery of Fulda, in the first half of the ninth century. The manuscript in question, which is attributed to the year 1023, has been carefully reproduced by the monks of Monte Cassino where it is preserved. The full-paged miniature is to be found in a chapter headed _De Vitro_; I can, however, discover nothing in the text that throws any light on our subject. In the illustration we see to the left a nearly naked workman who holds a mass of some green material, perhaps the frit; another man is blowing through a tube what is probably meant for the unfinished cup; to the left a chalice-like vessel, perhaps the model, is depicted. Notice, too, in the pediment of the roof (probably to be regarded as the annealing oven) a cup with a knob for stem, and hemispherical foot. Cups of a similar form, apparently in these cases of metal, are found in other illustrations of the manuscript (Plate XIX.).