The western cella of the Erechtheum was in all probability divided into two chambers by a wall running east and west (Fig. 7). The chief evidence in the building for this is that the west door of the Erechtheum does not stand in the middle of the wall, a peculiarity often remarked (Penrose, The Principles of Athenian Architecture, p. 88). The unusual position of a door under a column is structurally objectionable (Michaelis, Jb. Arch. Inst., 1902, p. 18). Had the door been placed in the middle, it would have stood directly under the central intercolumination of the west colonnade. The latest theory (D'Ooge, The Acropolis of Athens, p. 201) is that the position of the door was determined by the structure which abutted against the west wall just south of the door. The presence of an adjoining structure is then to be credited with some magic power of attraction which drew the door from its normal position into one structurally objectionable. The unsymmetrical position of the door was doubtless determined by the interior cross-wall which stood just north of the door and divided the west cella into a north and south chamber of approximately the same size. The door connecting the two very probably lay in the axis of the north and south doors of the temple (Fig. 7), thus very near to the west wall. The distance of the top course which could not have reached above the lintel of the west door was 8¼ feet above the bottom of the orthostates of the west wall. The height of the doors mentioned in the Chandler inscription is 8¼ feet. Of this cross-wall there are no traces of contact with the west wall. It must be noted, however, that the surface of the west wall is at that place badly broken away (Fig. 8). The surface of the orthostate is in part well preserved but orthostates at the place of contact with interior walls have nowhere left any indication of such contact—no anathyrosis. This is especially peculiar in the case of the eastern cross-wall where the supposed higher level on the east side would lead one to expect a careful joining with anathyrosis (Fig. 9). Had the north wall been destroyed beyond recovery down to this orthostate, there would have been no evidence now to show that a cross-wall ever was in contact with it. The orthostate next the door in the west wall cannot be cited as evidence against the existence of an interior cross-wall running east and west. The blocks above this orthostate are badly broken away except one just below the lintel which has some original surface preserved. The lintel like the orthostate is a block two courses high and may have the same exemption from any signs of contact, as far as the surface is concerned, with the interior of the wall. It is possible that not a single course of the cross-wall keyed into the west wall because the former was merely a low partition-wall. The top of the lintel in the line of the wall is broken away so that there, as in the case of the blocks below, no evidence of clamps can be expected. Neglecting for a moment the remarkable position of the door, it may be said that the interior surface of the west wall just north of the door is in no condition to give definite evidence pro or con of the existence of this interior cross-wall. The conclusive answer must be found in the simple description of Pausanias to whose text one may now turn (I, 26, 5). The new plan fits perfectly.
In the first room (ἐσελθοῦσι) Pausanias found the altars of Hephaestus, Poseidon-Erechtheus and Butes, and the paintings of the Butadae. The wall space lighted directly from the west windows was finely adapted for the paintings. There were only two doors and those at the west ends of the long walls. There could have been an uninterrupted series of paintings, whereas the προστομιαῖον of the other plan had five doors, and therefore offered less desirable space. With the words διπλοῦν γάρ ἐστιν τὸ οἴκημα, Pausanias passes to the next room (ἔνδον) where he found the well of sea-water. Now the name with which Pausanias introduces his description is significant: ἔστι δὲ καὶ οἴκημα Ἐρέχθειον καλούμενον. He named the temple from the part which he entered first and then he says a moment later that this οἴκημα is double, i.e. the part which he has just entered. Up to this point there is no suggestion of Athena. The διπλοῦν οἴκημα of Erechtheus consisted of two chambers one behind the other with reference to the porch.
The φρέαρ in the new plan is in the inner (ἔνδον) room of the οἴκημα near the west wall of the temple, where water was accumulated in later times and probably therefore in Greek and Roman times, while there is no indication whatever of a well of any sort in the inner chamber according to the old plan. At present the cistern in the western part of the temple reaches from north door to south door, but there is evidence to show that originally in Greek times it did not extend so far north. Just inside the north door, the pavement consisted of thin slabs, 0.13 m. thick, which ran in under the heavy blocks below the orthostates of the west wall and fitted into a cutting in the topmost course of the poros foundation. The thinness of the pavement is inconsistent with the theory of a hollow vault of any sort beneath the floor. There must have been a filling of earth for the pavement to rest on. This confirms the theory that the originally smaller place for the accumulation of water within the building was the south-west corner. The drain at the south-west corner of the North Porch which brought water from the direction of the Caryatid Porch both before and after the present Erechtheum was built may have carried excess water from the φρέαρ. It is possible that the absence of a proper foundation beneath the threshold of the door in the Caryatid Porch was due to the presence there of a course or courses of stone which surrounded the well and trident-mark. The architect, unable to secure consent to their removal, was compelled to build upon them and to raise the door. He placed the threshold above the bottom of the orthostates, and the position of this threshold may have determined the high position of the orthostates of the western wall. Both are placed at the same level.
In the inner room Pausanias saw the trident-mark, naturally near the φρέαρ. The first produced the second, according to Apollodorus, III, 14, 2. Pausanias did not see them πρὸ τῆς ἐσόδου but ἔνδον. There is no authority whatever for identifying the marks in the rock beneath the north porch with those made by the trident of Poseidon, except common consent in recent times. If the trident-mark lay within the Erechtheum what deity made that outside, and beneath the porch, a mark which was beyond question an object of cult? "Die Stelle welche Zeus mit seinem Blitze getroffen hatte, wurde mit einem Puteal umgeben und blieb unter freiem Himmel" (Dörpfeld, Ath. Mitt., 1903, p. 467). An altar of Zeus Hypatus stood before the entrance. The coincidence of place πρὸ τῆς ἐσόδου and ἐν τῇ προστάσει τῇ πρὸς τοῦ θυρώματος where, according to the official inscription the altar of the Thyechous stood, outweighs any objection to the identification of the two altars based on difference of name in the two records, ὁ βωμὸς τοῦ θυηχοῦ and Διὸς βωμὸς Ὑπάτου. Pausanias departs from the official terminology of building inscriptions. The rotunda at Epidaurus was called in the building inscription θυμέλη (cf. Cavvadias, Τὸ Ἱερὸν τοῦ Ἀσκληπιοῦ ἐν Ἐπιδαύρῳ, p. 50). Pausanias called it θόλος. The official name for the Erechtheum does not occur in literature nor in inscriptions except in the report of the commissioners. It is not surprising then if Pausanias failed to call the altar ὁ βωμὸς τοῦ θυηχοῦ. This name gives not the slightest clue to the god to whom it was erected. The suggestion of Michaelis (Jb. Arch. Inst., 1902, p. 17) that the altar may have been one to Poseidon proceeds from the logical idea to make it that of the god who is thought to have made the marks in the rock beneath the porch.