The theory was at one time put forward that a staircase afforded communication between the western cella and the higher eastern cella, but several considerations establish the fact that they had a common level. The conclusive argument is that there are no cuttings in the rock for the cross-wall between the two cellae, although that rock lay only 1-1.50 m. below the base of the wall. In its rough and sloping surfaces (Fig. 9) there is not a single trace of a bed for a foundation which the supposed heavy cross-wall would demand. The rock betrays no evidences whatever of preparation to receive a foundation. The contention that points of rock were broken off is absurd. The foundations for the outside walls go down to and rest in such beds, that of the west wall being an ill.u.s.tration. Those who believe that the heavy cross-wall supported roof beams besides serving as a terrace wall for the western cella 3 m. lower than the eastern, seem not to have thought that such a wall would need a well cut bed in the rock. Now the east wall, the thinnest in the building, has a foundation which, though it consists of eight courses of heavy poros blocks, rests in deep cuttings in the rock.

Under one block of the lowest course, lies a smaller block of poros which also rests in deep cuttings in the rock. Why did not the eastern interior cross-wall likewise have a bed for it cut in the rock, especially since its foundation was so shallow, only two or three courses of poros, and not eight as in the case of the eastern wall? The only bit of outside wall which does not rest in cuttings in the rock is that at the southwest corner, but there the few courses below the lintel of the door rested on an object of cult of some sort which made impossible the normal foundation, while the weight above the lintel rested on the heavy block in the west wall and the firmly founded wall just east of the door.

The champions of the accepted plan of the Erechtheum must explain a striking inconsistency in construction presented by the two interior cross-walls. The western, a screen-wall (D'Ooge, _The Acropolis of Athens_, p. 202) which reached only five courses above the orthostates and supported no other weight whatever, had a foundation which rests partly in cuttings in the rock, while the eastern interior wall which reached quite to the ceiling, supported the weight of it, besides being of the nature of a terrace wall, had a foundation which rested only on the rough and sloping rock. How is this inconsistency to be explained?

The inconsistency cannot be avoided. The logical inference from the facts is one which makes Pausanias intelligible. The eastern cross-wall could not have reached to the ceiling except at the ends where the blocks keyed into the side-walls and shared their foundations. The inference that this wall for its entire length must have been as high as the traces on the side walls is altogether unnecessary. Except at the ends this wall was as high as the other part.i.tion-wall, and like it supported no weight. The pilasters lessened a span of thirty feet by perhaps two feet and with the outside walls served to support a heavy cross-beam. Wall-pilasters are not unknown in Greek architecture as the temples of Apollo at Ba.s.sae and the Heraeum at Olympia prove (Frazer, _op. cit._, III, p. 589).

Pausanias walked into the cella of Athena from that of Erechtheus without ascending a step. Since all the interior chambers of the Erechtheum had the same level as the north portal it is unnecessary to maintain that he should have entered the Athena cella first on coming from the east. In perfect keeping with the new plan of the interior is the simple sequence of the topographical indications in his description: (1) p?? t?? ?s?d??, (2) ?se????s??, (3) ??d?? (d?p???? ??? ?st?? t?

????a), (4) ????tat?? ??a?a (cf. ? ?e?? ?? ? t? ???a??? ??a?a), (5) t?

?a? d? t?? ?????? ?a?d??s?? ?a?? s??e???.

But what of the protruding poros foundations of the east and south walls and of the unfinished surface of the north wall which have always readily confirmed the theory of a higher level for the cella of Athena?

Certainly these were not visible. They must have been concealed behind marble shelves on north and south and marble shelves and steps on the east (Fig. 7). The builders of the Erechtheum were economical, using the foundations of the peristyle of the Hekatompedon as far as possible and then adding blocks of poros to complete a foundation for the south wall of their temple. There was no more need for a wall of marble behind the south shelf than there was for a marble floor beneath the pedestal of the statue in the Parthenon. These shelves were convenient for the exhibition of the many objects deposited in the cella which was a religious museum. The surface of the marble walls is not preserved to a sufficient height to show whether there was any trace of contact with the top of the shelf, just as they can give no positive evidence of a floor at the higher level.

A peculiar cutting in the orthostate at the south-east corner of the temple should be noted in this connection. The cutting is in the interior angle and is so made that the orthostate could be set at this place on a horizontal surface which ran inward. Was this horizontal surface the floor level? Was the floor of the eastern cella raised one step above the threshold as D'Ooge says (_op. cit._, p. 207)? This is unlikely because the floor level would then have been above the base of the orthostates. The horizontal surface was the top of the shelf, for its vertical plane would have courses of the same height as ordinary wall-blocks. There is a Roman block 10 feet long and 1-1/2 feet high which the Christians reused as the base stone of the iconostasis when they converted the Erechtheum into a church. It had a base moulding of some sort which the Christians chiselled off. This long block probably formed part of the lowest course of the facing of the shelf. The fact that its dimensions are those of the ???????? ????? ??et??, ??t?????

ta?? ?p???a??t?s?? e???? de??p?? ?fs?? t????? ??p?d??? (_I. G._, I, 322, col. 1) causes a suspicion that the Roman block simply replaced a Greek one, which in its position at the base of the wall "corresponded to" the ?p???a??t?de? at the top of it.

An examination of the foundation for the east wall reveals an interesting condition which is unintelligible if the cella of Athena had a higher floor-level than the western cella. In the north-east corner, a marble block of the north wall is cut back to the line of the west face of the poros foundation (Fig. 10). If the marble block lay buried beneath the floor, why was it so carefully trimmed? The explanation may be offered that the cutting was done when the temple was made over into a church. But the chiseling is more careful than the chiseling done at that time in the Erechtheum. When the eastern part.i.tion-wall was removed, rough traces of it were left on the side-walls. The treatment of the block in question is Greek in its carefulness and the cutting was probably made to receive a slab of the marble facing which concealed the foundation-blocks of the east wall.

[Ill.u.s.tration: FIGURE 10

THE N.E. CORNER OF THE CELLA OF ATHENA]

There is another serious difficulty in the way of those who believe that the eastern cella had a higher level than the western. The south wall of the temple had orthostates on the outside but none on the inside where wall-blocks of the usual height took their place. These wall-blocks were easily torn out and have since completely disappeared.

In the western chamber orthostates would have been illogical because they would have been high above the level of the floor, but in the eastern cella, if it had the level of the eastern porch orthostates would have been used. Since there were wall-blocks behind the orthostates of the south wall in the western cella, one would reasonably expect orthostates behind wall-blocks in the north wall of the eastern cella, provided that cella was at the level of the eastern porch. But it is absolutely certain that such was not the case. The notched form of the orthostate at the north-east corner of the temple shows that it was in contact with two courses of wall-blocks of regular height in the north wall. Thus the eastern cella, if it lay at the level of its porch strangely lacked interior orthostates in its north and south walls. But if this cella lay at the level of the western cella, the lack becomes at once intelligible. The absence of orthostates at the supposed higher floor-level of the eastern cella combines with the absence of any cutting for a foundation for the wall between the cellae to prove the theory which is in perfect harmony with the simple sequence in the description by Pausanias.

The theory of one level within the Erechtheum seems to contradict and to be contradicted by the evidence which Stevens has found of a door in the east wall (_A. J. A._, 1906, p. 58 ff.). The contradiction is not necessary, for a flight of steps at the east end of the cella of Athena is perfectly possible. The construction of an apse for the church at the east end of the temple necessitated the removal of a number of foundation-blocks which might have given evidence of steps. However it is quite possible that the foundations for the steps which had no need to rest in rock cuttings were simply laid against, not keyed into the foundations of the east wall. The stairs are drawn in the plan (Fig. 7).

The idea of a stair-case at the east end of a cella is ill.u.s.trated by the temple at Didyma. The eastern door of the Erechtheum was not the normal, not the intended entrance to the cella of Athena, but served as the traditional eastern entrance toward which the xoanon faced.

Pausanias like other visitors entered by the p??stas?? ? p??? t??

????at??, the main entrance to the temple.

It is interesting to note some evidence which shows that in the period before the Erechtheum was converted into a Christian church there was no difference of level within the building, namely, the ma.s.ses of rubble masonry which were placed close to the north wall at approximately equal distances from the eastern cross-wall. They are firmly founded on the rock and reach up nearly to the base of the orthostates. They have no counterparts along the south wall. The screen-wall of the north aisle of the church stood directly over one of the ma.s.ses. The threshold of it is still in place. These heavy foundations and the interior longitudinal walls of the church cannot be contemporary. The latter were sufficient to carry the weight of the roof of the church; and the screen-wall in the aisle, since it rests partly on a filling of earth, shows that the heavy foundation of rubble masonry underneath had ceased to serve any purpose after the church was built. It was there before that time and therefore must have been laid in a Roman period when the level within the temple was the same.

Any discussion of the workmanship of this ma.s.s of stones and mortar has no bearing on the question of its date and that of the threshold above.

The point is, the masonry is earlier than the Christian church, and quite embarra.s.ses the advocates of a higher level for the eastern cella in the period before the conversion of the temple into a Christian church. This foundation then is perfectly intelligible in the light of the theory that in Greek times there was but one level within the temple. What the purpose of this rubble masonry was is uncertain. The substantial and solid character of the ma.s.ses leads one to believe that they were foundations for piers or pillars which reached to the top of the adjacent wall and together with it supported heavy cross-beams which spanned the cella from north to south. The idea may have come to the Romans from the Greek pilaster which as noted above lay approximately midway between the ma.s.ses of rubble masonry. This was, then, apparently a device for reducing the span from the north to the south wall. The fact that this masonry was laid before the period of the church is of far greater importance than its purpose.

The new plan of the Erechtheum is interesting in the light of the Chandler inscription. If one feels that the magnificent north porch determines the front of the building, then the first room is a satisfactory p??st??a??? and lies in front of (p??) the * st??a??? in which was the important object of cult, the f??a? (st????). The following proportion may be set down: p???a??: ?a??:: p??st??a???: *

st??a???. ???st??a??? and * st??a??? are conjectured to have been the official names in the fifth century for the two chambers of the d?p????

????a of Pausanias.

The order followed by the commissioners in their report upon unfinished interior walls was as follows: In the first room entered from the ????a, the p??st??a???, 12 tetrapodies were ??at??sesta. The phrase ?? t? p??st??a?? favors the theory that more walls than one are meant.

Then in the inner chamber 3 tetrapodies of the pa?ast??,[7] i.e., that part of the part.i.tion-wall east of the door in the west cella. Then in the third room 6 (?) tetrapodies of the wall p??? t????at??. The order in which the chambers were examined for unfinished walls was that of Pausanias in describing their contents.

Again the new plan fits the treasure list of 306/5 B.C. (I.G., II,^2 733). The remarkable feature of the inscription is that it mentions three pa?ast?de?, first an isolated one, and then a pair of them, one on either side of a door. The single pa?ast??, the first to be mentioned is again that part of the part.i.tion-wall east of the door in the west cella. This door was near the west end of the wall, so that the s.p.a.ce between it and the west wall of the temple was negligible. Thus for one entering by that door there was a pa?ast?? on the left, but none on the right. When however he pa.s.sed into the ?a?? t?? ?????? through a door which stood a little south of the middle of the wall (and opposite the door in the west wall of the temple) he had a pa?ast?? upon his left and also upon his right. The pa?ast?de? are interior walls on either side of a door which in the Erechtheum reached up only five courses above the orthostates. The paintings which Pausanias found in the first room favor the opinion that the treasures which hung on the parastas were on the south side of that wall--i.e., in the second room of the d?p???? ????a.

Whether or not there is any order in the enumeration of the treasures is a question. If there is, then it naturally begins with treasures first seen after entering from the p??stas?? ? p??? t?? ????at??, just as the record of the commissioners in the case of interior walls begins with walls in the first room, just as the description of Pausanias begins with the contents of the first room. This coincidence is remarkable, and is true of no other theory about the temple.

It is a necessary consequence of this interpretation that some treasures were in the west part of the Erechtheum. Perhaps then something may be said for the scholiast on Aristophanes, _Plutus_, 1183 (reading ?????

for t????? and keeping in mind the d?p???? ????a of Pausanias's description): ?p?s? t?? ?e? t?? ?a??????? ?????d?? ?????? d?p???? ?????

(t?????) ???? ???a?, ?p?? ?? ??sa???f???????. The words ???? ???a?

suggest that the scholiast wished to distinguish between a d?p???? ?????

the two parts of which were connected by a door and another type the two parts of which were not so connected but separately entered from without. Pausanias seems to give an instance of the latter in II, 25, 1.

White (_Harvard Studies_, Vol. VI, p. 39) refers the scholium to the restored west part of the Hekatompedon but does not discuss the meaning of ???? ???a?, which Michaelis was unable to explain. In White's so-called opisthodomus, to which door of three possible ones does the scholiast refer? The three chambers of his opisthodomus do not satisfy the requirements of a d?p???? ?????, the reading which he accepts (_op.

cit._, p. 4, note 3). More reasonable is the interpretation that the scholiast had in mind the west cella of the Erechtheum in which some treasures seemed to have been placed, and that he used the words ?e??

?a??????? ?????d?? ?????? in the stricter sense, just as Pausanias called the east cella ?a?? t?? ?????d?? (I. 27. 1), and regarded the d?p???? ????? as lying behind it. The ?e?? t?? ?????? was oriented east, and what was immediately west was behind it. But it is not to be supposed that the west cella of the Erechtheum was ever called an opisthodomus. The scholiast seems however to have the oldest Athena temple in mind.

There is a point perhaps of slight moment which deserves a word. One of the paintings, that of Erechtheus driving a chariot, was painted, according to the scholiast on Aristides, I, 107, 5, behind the G.o.ddess.

A possible interpretation is that the painting was in the cella of Athena on the wall behind the xoanon, but the paintings of the Butadae were in the first room which Pausanias entered. Unless the painting of Erechtheus was separate from those of the Butadae, then the new arrangement of the interior permits a satisfactory solution of the difficulty. For the east wall of the room in which were the paintings of the Butadae was behind the G.o.ddess. According to the old plan, Pausanias found the paintings in the western chamber of the d?p????

????a, that is, between them and the wall against which stood the xoanon, was a chamber. The pa.s.sage may mean that in a painting Erechtheus appeared behind Athena driving a chariot (Petersen, _Jb.

Arch. Inst._, 1902, p. 64; _Burgtempel_, p. 110). In the sequence of words in the sentence, ?? t? ????p??e? ?p?s? t?? ?e??, the second phrase seems to be a closer definition of the place than is given in the first.

Furthermore, position was determined by reference to the xoanon. An interior wall was located with reference to it, t? p??? t????at??. The scholiast on Aristophanes, _Equites_, 1169, is interesting in this connection because he shows what part a statue might play in the designation of a temple: d?? e?s?? ?p? t?? ????p??e?? ?????? ?a??, ? t??

?????d?? ?a? ? ???se?efa?t???.

In the light of the new arrangement within the Erechtheum, the reference of Vitruvius (IV, 8, 4) to the temple becomes clearer. Speaking of it and other temples he says: "cellae enim longitudinibus duplices sunt ad lat.i.tudines uti reliquae, sed is omnia quae solent esse in frontibus ad latera sunt translata" (Petersen, _Burgtempel_, p. 144). If the cella of Athena was completely separate from that of Erechtheus and at a higher level, he could not have said reasonably of the cella of the temple that it was twice as long as wide like other temples. For the cellae of Athena and Erechtheus ought then to have been considered separately. In the new plan such a statement applies with greater force because the low part.i.tions might be readily disregarded. The second statement shows that Vitruvius regarded the east facade of a temple as the front, and normal place of entrance, but that this and the more elaborate porch were transferred in the case of the Erechtheum to what would be the side of other temples. As Petersen, (_op. cit._, p. 143) says, the words "columnis adjectis dextra ac sinistra ad umeros p.r.o.nai" are a clear reference to the north porch. This too seems to be the p???a?? which Lucian refers to in Piscator, 21: ??ta??? p?? ?? t? p????? t?? p????d??

d???s?e?. ? ???e?a d???e? ??? t? ???a, ?e?? d? ?? t?s??t?

p??s????s?e? t? ?e?. This interpretation is perfectly consistent with the fundamental contention that the p??stas?? ? p??? t?? ????at??

determines the front of the building.

The theory set forth in the above pages is in perfect accord with the description in Pausanias. It is confirmed by the evidence of the inscriptions and of the building itself so far as that evidence goes.

The serious criticism of the accepted plan of the Erechtheum is that all theories based upon it disagree with the written evidences, not with one written record of a later period like the simple account of Pausanias, but with another record centuries earlier, namely the contemporary official inscription. Investigators attempt the solution of the problem after accepting the restored interior as certain. The keynote of the present theory is that the interior of the temple has been too far destroyed to make any one restoration absolutely certain on the basis of the evidence of the building alone, and that all available evidence must be used simultaneously to determine the correct restoration.

IV

THE ERECHTHEUM AS PLANNED

The question as to the original plan of the Erechtheum follows naturally the interpretation of the building as built. That the west wall was planned for its present place seems improbable for a number of reasons.

The north porch is out of proportion to the room into which it opens, and by reaching beyond the west wall of the temple becomes in part porch to an open precinct. The west front has columns and Caryatids at different levels (Dorpfeld, _Ath. Mitt._, 1904, p. 101). The displeasing effect of this difference could not have been concealed by the walls of the Pandroseum, the south one of which reached as high as the parapet of the porch of the maidens. The latter porch ill.u.s.trates the skill of the architect in concealing differences of level. The unique closed wall on which the maidens stand was his device for concealing from view from without, a door which was below the level of the porch and which belonged to the interior whereas the porch belonged to the exterior. The architect, by placing the entrance to the porch at the north east corner close to the wall, completely concealed the presence of the low door.

With this care to conceal a difference of level, the west side of the temple is in marked contrast.

The north-west corner of the western cella is peculiar in two ways. The western jamb of the door cuts 3-1/2 cm. into the west wall of the temple. This suggests crowding and is satisfactorily explained by the condition of the foundations below. The foundation of the west wall does not key into that of the north wall (Fig. 11), a fact seeming to prove that when the latter foundation was laid, it was not the intention of the architect to place a foundation in the line of the present west wall, and to crowd the door jamb into that wall.

Of the symmetrical exterior proposed by Prof. Dorpfeld there lies a suggestion in the fact that the north and south doors have the same axis, although the Caryatid porch has not. The porch seems to have been moved a little to the east of its intended place that it might not project beyond the west wall, but not far enough to prevent the cornice of the porch from so projecting.

The west wall itself offers evidence of a curtailment of the original plan. By way of introduction let us compare the east facade, which is Greek with the west facade, the part of which above the closed wall is Roman (_Arx Athenarum_, Pl. XXV, D, and _A. J. A._, 1906, Pl. VIII). The windows in the east wall which Stevens has determined with accuracy were placed at the height of four ordinary courses above the base moulding and two courses from the top of the wall, just as were the Roman windows in the west wall. The second course above the eastern windows was a moulding, the corresponding course above the western windows is plain probably because of the adjacent capitals. Below both sets of windows were three courses of blocks. In the east wall orthostates were justifiable, in the west wall they would have been illogical because on neither side was there a floor, but three courses equal in height to four ordinary courses were placed there. Stevens has shown that the eastern windows were seven courses high including the lintel. The western windows are five courses high. The explanation of the difference of height is simple. The eastern wall was thirteen courses high, the western eleven. The western windows were two courses shorter in order that they and their counterparts, the eastern windows, might be equidistant from the base of the wall, namely four ordinary courses, and from the top of the wall, namely two courses. The fact that the sills of the Greek windows were one meter lower than the Roman windows is of no consequence whatsoever. The fact of great importance is that the east and west windows occupied the same relative position in the facade. The stylobate of the western facade could not be placed so low as the eastern because of the door and the necessity of a heavy block three courses high at the south end of the wall. This block could not be placed lower because of the Cecropium (= temple of Pandrosus?) which crossed the line of the wall, to judge from the cuttings in it beneath the heavy block. Had the architect wished equality of height for the eastern and western colonnades he would have been compelled to place the stylobate of the western two courses lower. This would have made it impossible to place a door in that wall which was necessary probably for a reason of cult.

[Ill.u.s.tration: FIGURE 11

THE INTERIOR N.W. CORNER OF THE ERECHTHEUM. MODERN MASONRY UNDER N. END OF W. WALL]

In Roman times therefore the western windows were placed with careful reference to the eastern. Between the columns in each case appeared windows, two in the eastern wall with door between, three in the western where a door was impossible. Both facades were surmounted by epistyle, frieze and pediment. The wall below the western colonnade was a subst.i.tute for the higher ground level of the east side. The Romans who repaired the wall repaired it with reference to the east front. For them the west facade was simply a combination of wall with windows, and colonnade. Unless the Greeks had a western facade of columns and wall with windows essentially like the Roman restoration, we are forced to make a strange a.s.sumption. The Greek architect conceived the idea of combining wall with colonnade in one plane and then instead of carrying his idea to its conclusion put in a wooden grille in the intercoluminations above a low wall of three courses, a grille which answers to nothing in the east facade, and then left it to the Romans to exploit his idea by placing there three windows.

The only obstacle to the perfectly natural a.s.sumption that the Romans restored the essential features of the west wall as it was in Greek times is the testimony of a contemporary inscription (I. G., I, Suppl., 321. col. III, 18) that one Comon a carpenter was paid a sum of 40 dr.

for "fencing" (d?af???sa?t?) four intercolumniations on the wall toward the Pandroseum: d?af???sa?t? t? eta?????a t?tta?a ??ta t? p??? t??

?a?d??se??. The accepted interpretation of the pa.s.sage is that a wooden grille was the final form of the west wall and remained so until Roman times. The objection to this interpretation is that we must then believe that the Greek architect planned a wooden grille for a marble building in a wall exposed to the elements where repair would be necessary from time to time and that only in the Roman period did the change to more enduring marble take place. It is probable that the wooden grille was only temporary and was soon replaced by a wall with windows. Whatever the interpretation of the inscription, the fact remains that the present form of the west wall is a restoration made with deliberate reference to the east facade. It is a studied restoration which far from being an arbitrary creation of the 4th century A.D., as Penrose (_op. cit._, p.

93) regarded it, is too original for a Roman period. The imitation is Roman, the idea is Greek. The very same idea is expressed in the Sidon sarcophagus of the mourning women, an Attic work of about 350 B.C. The illusion produced by the sarcophagus is that of female figures standing between the columns of the peristyle of a temple (Hamdy Bey-Reinach, _Une Necropole Royale a Sidon_, p. 241). The west facade in Greek times as in Roman was simply a compression together in one plane of colonnade and wall--a combination to which the architect was forced by the curtailment of his plan.