began to employ, in Doric buildings, the device of triglyphs and the metopes between the beams.
3. Later, others in other buildings allowed the projecting principal rafters to run out till they were flush with the triglyphs, and then formed their projections into simae. From that practice, like the triglyphs from the arrangement of the tie-beams, the system of mutules under the coronae was devised from the projections of the principal rafters. Hence generally, in buildings of stone and marble, the mutules are carved with a downward slant, in imitation of the principal rafters. For these necessarily have a slanting and projecting position to let the water drip down. The scheme of triglyphs and mutules in Doric buildings was, therefore, the imitative device that I have described.
4. It cannot be that the triglyphs represent windows, as some have erroneously said, since the triglyphs are placed at the corners and over the middle of columns — places where, from the nature of the case, there can be no windows at all. For buildings are wholly disconnected at the corners if openings for windows are left at those points. Again, if we are to suppose that there were open windows where the triglyphs now stand, it will follow, on the same principle, that the dentils of the Ionic order have likewise taken the places of windows. For the term "metope" is used of the intervals between dentils as well as of those between triglyphs. The Greeks call the seats of tie-beams and rafters ὀπαί, while our people call these cavities columbaria (dovecotes). Hence, the space between the tie-beams, being the space between two "opae," was named by them μετόπη.
5. The system of triglyphs and mutules was invented for the Doric order, and similarly the scheme of dentils belongs to the Ionic, in which there are proper grounds for its use in buildings. Just as mutules represent the projection of the principal rafters, so dentils in the Ionic are an imitation of the projections of the common rafters. And so in Greek works nobody ever put dentils under mutules, as it is impossible that common rafters should be underneath principal rafters. Therefore, if that which
