In the great Temple of Jupiter Olympius (Woodcut No. 196), as described by Vitruvius,[144] the nave had two storeys of columns all round, and the middle was open to the sky. It is suggested, however, by Dr. Dorpfield that the temple in Vitruvius’s time was incomplete, and that subsequently when Hadrian erected the great chryselephantine statue in it the nave may have lost its hypæthral source of light. (In that case its light may have been introduced through the court or hypæthron in front of the cell, such as is shown on the plan in Woodcut No. 196.) The Ionic temples of Asia are all too much ruined to enable us to say exactly in what manner, and to what extent, this mode of lighting was applied to them, though there seems no doubt that the method there adopted was very similar in all its main features. [Illustration: 155. Plan of Erechtheium. (From Stuart.) Scale 100 ft. to 1 in.] [Illustration: 156. Elevation of West End of Erechtheium. Scale 50 ft. to 1 in.] The little Temple of Nikè Apteros and the temple on the Ilissus, were both too small to require any complicated arrangement of the sort, but the Ionic temple of Pandrosus was lighted by windows which still remain at the west end, so that it is possible the same expedient may have been adopted to at least some extent in the Asiatic examples. The latter, however, is, with one exception, the sole instance of windows in any European-Greek temple, the only other example being in the very exceptional temple at Agrigentum. It is valuable, besides, as showing how little the Greeks were bound by rules or by any fancied laws of symmetry.