Longtime BAR readers know that two theories vie with each other regarding where the Temple once stood on Jerusalem’s Temple Mount. The first was advanced by Asher Kaufman, a Hebrew University physicist with a longstanding interest in the Temple Mount, the second by Leen Ritmeyer, an architectural draughtsman who worked on the Temple Mount excavations under Benjamin Mazar and who is himself a Temple Mount specialist. Hoping to advance the issue, BAR editor Hershel Shanks asked me to weigh the merits of each theory. I, too, am a specialist on the Temple Mount and have published on the subject. I replied that I agreed with neither. This naturally led to an invitation to state my own view; this article is the result.
Kaufman places the Temple about 280 feet northwest of the seventh-century Dome of the Rock, which now graces the Temple Mount.1 Ritmeyer places the Temple’s Holy of Holies, the innermost sanctuary, at the spot where the rock mass called al-Sakhra (Arabic, “the Rock”)2 sits under the golden dome of the mosque.3
My own location is closer to Ritmeyer’s than to Kaufman’s: I place the center of the main hall of the Temple (called the Hekhal), rather than the Holy of Holies, over al-Sakhra. This on its own is an original suggestion: Most scholars identify al-Sakhra with either the Holy of Holies or the Altar of Sacrifice.4 However, what I believe is especially important about my proposal is the reasoning I use to reach this result.
Nothing can be said about the location of Solomon’s Temple (often called the First Temple) based on existing features. The same is true of the Temple known as the Second Temple, which was rebuilt by the exiles upon their return from Babylon. Herod the Great totally reconstructed this Temple (also known in Jewish tradition as the Second Temple). Some of the large-scale features of Herod’s 5Temple are preserved in the existing Temple Mount (some of the underground cisterns, however, are believed to predate Herod).
The site, as we know it today, is mostly the product of two distinct building phases: Herodian and early Islamic. In 20/19 B.C.E., Herod initiated a vast project to greatly expand the Temple enclosure by creating a huge level platform bordered by a massive retaining wall. This retaining wall was built of large rectangular blocks of limestone with recessed margins and a large, slightly raised central panel, or boss.
Herod more than doubled the size of the previous Temple Mount. His enlarged enclosure constituted the largest sacred precinct in the whole of classical antiquity—approximately 36 acres. By comparison, the precinct to the goddess Athena on the Acropolis of Athens—including the famous Parthenon—occupies barely a fifth of the area of its Jerusalem counterpart.
Herod’s glorious sanctuary was destroyed in 70 C.E. when the Romans conquered and burned Jerusalem, effectively ending the First Jewish Revolt against Rome. (Only a few holdouts in the Judean Desert, notably Masada, managed to withstand the Roman assault for another three years.)
After the destruction of 70 C.E., the Temple Mount remained largely unoccupied until the late seventh century C.E., when the Umayyad caliph ‘Abd al-Malik and his successors reclaimed the site and established what we see today: the splendid Dome of the Rock, on a raised platform in the middle of the esplanade, and the al-Aqsa Mosque, at the esplanade’s southern end. The Muslims renamed the sacred enclosure al-Haram al-Sharif, which means “the Noble Sanctuary.”
The Haram probably does not have precisely the same measurements as Herod’s Temple Mount. It will be important to keep the likely differences in mind.When we refer to the Temple Mount, we will be referring to Herod’s esplanade. When we refer to the Haram, or al-Haram al-Sharif, we will be referring to the platform as it exists today.
The location of Herod’s Temple has mostly been considered in the context of the existing topography of the Temple Mount, particularly in relation to al-Sakhra. Later on, I shall also relate al-Sakhra to the ancient Temple. But I would like to approach the question first by considering what we can learn from Herod’s enclosure wall. In fact, the only surviving part of the Temple complex from Herod’s period that we arecertain of is the incomplete circuit of the enclosure wall.
The Herodian walls of the Temple enclosure are preserved on three sides—south, west and east. Fortunately, the Herodian masonry is distinctive, so there is no dispute as to whether it is Herodian. And it is by this means that we can identify the remains of the Herodian perimeter wall.
On the south, the ancient masonry has now been laid bare along the entire length of the wall, which is also the southern wall of al-Haram al-Sharif, although the upper portion of the stonework is more recent.
On the west, too, we can now trace the entire length of the Herodian wall, although until recently, the northern part of the western wall was inaccessible. Near the southern part of the western wall is the Jewish devotional section known as the Western Wall, or, formerly, the Wailing Wall. Like the southern wall, the western wall of the Temple Mount is also the western wall of al-Haram al-Sharif.
The eastern wall is somewhat more complicated. At a point 105 feet north of the southern end of the eastern wall is the “Straight Joint,” a straight seam in the masonry that tells the careful observer that the wall south of that point was an addition. It is clear from the masonry that Herod added this section to the previously existing eastern wall of the Temple Mount. North of the Straight Joint, the ashlars have a distinctly different appearance. The older blocks are generally shorter and, while they also possess drafted margins, are not as finely dressed. On the other hand, these ashlars are arranged more systematically in alternate rows of headers and stretchers (narrow and long sides facing out). Also, some of the lateral joints on the ashlars are diagonal () rather than atright angles to the horizontal line of the stones, as is generally the case in Herodian construction ().
It is difficult to say who was responsible for this construction, although we do know that it was earlier than the Herodian extension to the south. Based on stylistic comparisons with fortifications in Greece that are firmly placed in the Hellenistic period, this earlier masonry dates somewhere between the fourth and the first century B.C.E.6
The pre-Herodian masonry on the eastern wall appears to continue northwards, below present ground level, possibly as far as the Golden Gate, where the same distinctive masonry is again visible. A little beyond the Golden Gate, there is a further stretch of Herodian ashlars, which continues beyond the northeast corner of al-Haram al-Sharif.
From this we conclude that Herod used the east wall of the existing Temple Mount, but extended it both toward the south and the north. Note, however, that on the north the Herodian wall extends beyond the wall of today’s al-Haram al-Sharif. This is important to remember.
Moreover, the eastern wall is not perfectly straight. The original, pre-Herodian builders tried with some difficulty to construct a wall that does not follow the contours of the hill but rather cuts across the hillside, with the terrain descending quite steeply both to the south and to the north. In addition, the angle formed by the southern wall and the eastern wall is slightly more than 90 degrees (actually 92 degrees), as compared to the 90 degree angle formed by the southern wall and the western wall.
The north wall of Herod’s Temple Mount no longer exists—or at least remnants of it can no longer be seen. Where we place the northern wall will determine the precise shape and size of Herod’s Temple Mount.
But before we try to fix the line of the northern wall of Herod’s Temple Mount, let’s look at the shape of the present enclosure, al-Haram al-Sharif. It is an irregular quadrilateral. Each of its four sides is a different length, yet it follows the same line as the enclosure of Herod’s Temple Mount on three sides (all except the north). On the south and west, the al-Haram al-Sharif is the same length as Herod’s Temple Mount—920 feet and 1,590 feet, respectively (all these lengths are rounded).7 On the east, al-Haram al-Sharif is 1,540 feet long; on the north, 1,030 feet.
Now let’s look at the angles formed by this quadrilateral. There are two right angles (90 degrees)—on the southwest and on the northeast. The angle on the southeast, as previously noted, is 92 degrees, a little more than a right angle. On the northwest, it is 85 degrees.8
The right angle at the southwest corner is exactly what we would expect of the Herodian Temple Mount: the intersection of two lines established by Herod withoutearlier lines to influence his architects. If the right angle at the northeast corner of al-Haram al-Sharif is also the corner of the Herodian Temple Mount, this would enable us to fix all four walls of the Temple Mount (assuming, of course, that they were straight lines).
Several investigators have concluded that this northeast corner of al-Haram al-Sharif, which now forms a kind of tower, is also the corner of the Herodian Temple Mount because on the north and east sides it contains ashlars that are unmistakably Herodian. However, this fact can be misleading. The Herodian ashlars are mixed together with other types of masonry, indicating that the ancient stonework, seen aboveground, is in reuse. An additional reason for doubting that this corner is an original Herodian feature is the lack of pilasters, or square half columns, high up on the northern and eastern faces of this corner, whereas remains of a decorative pilaster course existed at this level on the west wall. One fragment of a wall pilaster survives in place towards the northern end of the western wall,9 and other fragments have been found among the debris at the foot of the southwestern corner by Israeli archaeologists since 1967. Such pilasters still grace the upper part of the wall of the beautifully preserved enclosure at Hebron that Herod built around the Tomb of the Patriarchs.10 Owing to its excellent state of preservation, the Tomb of the Patriarchs has much to teach us about the Temple Mount enclosure wall that Herod built at about the same time in Jerusalem. In many ways, it is a smaller version of the Jerusalem enclosure. It is built of ashlars stylistically identical to those in Jerusalem and has a similar pilaster course.
The reuse of the beautiful Herodian ashlars thrown down in the Roman destruction of the Temple Mount was common in Jerusalem. The most prominent examples are walls of early Islamic palaces outside the southwest corner of the Temple Mount, built almost entirely of Herodian ashlars. So the presence of this type of masonry in the northeast corner of the enclosure wall does not necessarily mean that it is Herodian.
Indeed, it would be surprising if Herod’s Temple Mount enclosure wall did not follow a more regular plan like the Hebron enclosure, which is a perfect rectangle—not only because both were built by the same ruler at the same time and in the same style, but also for more general reasons. The then-current precepts of Roman imperial architecture usually called for a symmetrical ground plan unless there was an obstruction,such as another important building that confined the available area, or if there was an older wall that the builders wished to incorporate.
Another factor militates against accepting the northeast corner of al-Haram al-Sharif as the northeast corner of Herod’s Temple Mount: The eastern wall of the Temple Mount extends northwards, without a break, for about 65 feet beyond the northeast corner of al-Haram al-Sharif. Although we cannot see this aboveground, we know it from the work of Charles Warren in Jerusalem between 1867 and 1870. At that time, Warren, who was conducting a survey of western Palestine for the London-based Palestine Exploration Fund, managed—at great personal risk and despite opposition from the Ottoman authorities—to dig deep shafts down to the foundations of the enclosure wall and then drive galleries along the base of the walls. It is the record of his observations and measurements that tells us what the walls look like underground.11
The combination of one or more peristyle courts with a temple aligned with one of the main axes in a symmetrical arrangement became a hallmark of the great temple complexes of the Roman Empire. Some of the most famous examples are in the Near East, in particular, the temple precinct of Jupiter Heliopolitanus at Baalbek, the temple of Bel at Palmyra and the temple precinct of Artemis at Gerasa. Other examples are found in Asia Minor and around the Mediterranean.12 According to the first-century C.E. Jewish historian Josephus, Herod’s Temple complex in Jerusalem followed this same arrangement. We are told that the Temple platform was bordered by a circuit of porticoes, which included an immense triple-aisled basilica on the south side,13 and that the sanctuary occupied the central position in the scheme.14 Incidentally, the incorporation of a triple-aisled basilica at one side of a peristyle court is encountered at various Roman sites across the eastern Mediterranean, including Cyrene in North Africa and Smyrna (Izmir), Ephesus and Kremna in Asia Minor.15
Archaeological evidence for the portico on the southern side of Herod’s Temple Mount has been obtained in Professor Benjamin Mazar’s excavations on the southern edge of the Jerusalem esplanade.16 Although this says nothing about the Temple itself, excavations at Samaria (Sebaste) have revealed that there Herod’s architects worked according to a similar formula (although without including a triple-aisled basilica in the temple enclosure). Herod refounded Samaria as Sebaste in honor of his Roman patron Augustus (Sebastos in Greek) in about 25 B.C.E.17 Herod’s Sebasteion, one of three temples he dedicated to Augustus (the others were at Caesarea and Paneion), was symmetrically oriented in relation to a peristyle court and was probably axially aligned with its monumental entrance.18 This temple was begun in the year that Herod ordered the rebuilding of Samaria, that is, 25 B.C.E., approximately five years before ordering the reconstruction of the Temple in Jerusalem.
There is something else that suggests that the shape of Herod’s Temple Mount enclosure was a trapezium, almost a rectangle, rather than an irregular quadrilateral, like al-Haram al-Sharif. Three of the streets of Jerusalem north of al-Haram al-Sharif, including the eastern section of the Via Dolorosa, are aligned broadly parallel to one another and are oriented at right angles to the western wall of the Temple Mount enclosure. In other words, they are parallel not only to themselves, but to the southern wall of the enclosure.19 On the other hand, the north wall of al-Haram al-Sharif has an orientation without any parallels elsewhere in the city.
All of this suggests that the northern wall of the Temple Mount enclosure was on a different alignment from the northern wall of al-Haram al-Sharif and was parallel to the southern wall of the Temple Mount. But where was it?
Assuming that the Herodian western wall ended where it does today (at the northwest corner of the al-Ghawanima Minaret), I looked first at the northern end of the western wall.20 This seems especially reasonable because a cornerstone appears to remain in situ at the northern end of the Herodian wall, alongside the northwest corner of the minaret.21 The ashlar in question bears distinctive marginal drafting on both its north and west faces, whereas blocks belonging to a straight section of the enclosure walls only have the drafting on the single exposed face.
If we proceeded eastward after making a 90 degree turn, we would hit the eastern Herodian wall close to the point where Warren found it ended underground.22 Warren labeled the vicinity of this spot on his drawing of the elevation of the east wall as the “possible termination of Old Wall with marginal drafts.”23
Now comes the surprising part. I created a right-angled triangle by drawing a line from the northwest corner of the Herodian Temple Mount to the southeast corner (the right angle exists between the west and south walls). The south wall is 920 feet; the west wall is 1,590 feet, as we have already said. The ratio is 1:1.73. The result is a 60-degree angle between the south wall and the diagonal (hypotenuse) of the triangle. This 60-degree angle has been totally overlooked by all previous investigators. As we shall see, this vital measurement not only confirms the location of the “missing” northern wall of the Herodian Temple Mount, but enables us to deduce the position of the Temple itself.
This ratio of the south wall to the west wall, which creates a right-angled triangle with a 60-degree angle at the southeast corner, I discovered to my profound surprise, was precisely the same as the width-to-length ratio of the rectangular enclosure that Herod constructed at Hebron around the Tomb of the Patriarchs.24 The walls of this enclosure form an almost perfect rectangle, the proportions of which average 193.7 feet to 111.5 feet, a ratio of 1:1.737, the same ratio as in the Temple Mount. And triangles drawn from the adjacent sides again produces angles of 60 degrees.
But this is not all. A 60-degree angle has special significance in classical architecture. The same proportions are to be found in other architectural schemes 25from the early Roman Empire, among them the enclosures of the temple of Jupiter Damascenus and the Delphineion at Miletus. Also, these proportions were used in the Teatro Maritimo in Hadrian’s villa at Tivoli and other structures from the classical period.
Recently, while reviewing the final reports of Yigael Yadin’s excavations of Masada,26 I noticed that a similar geometrical scheme, involving an equilateral (equal-sided) triangle with 60 degree angles, governs the proportions of the banqueting hall on the lowest terrace of the Northern Palace at Herod’s desert fortress.27
This type of geometrical plan therefore seems to be akin to a signature of the architects employed by Herod on his most prestigious building projects. A supreme example of geometrical planning in Herodian architecture is the perfectly circular hilltop château at Herodium, near Jerusalem.28
Abstract geometrical designs tended to dominate the artistic repertoire of Judea in the period of the Second Temple because the Second Commandment was interpreted at the time as proscribing representational images. (A similar emphasis on geometrical patterns is observed in later Islamic art.) Especially favored in Herodian decoration are compass-drawn rosette patterns, which occur with great regularity in mosaics and stone reliefs of this period.29 Most common of all are three- and six-petal rosettes, which are based on a 60-degree symmetry. One of these rosettes, with six black petals on a white background, graces a mosaic on the floor of a small bathroom in a luxurious mansion Nahman Avigad excavated in Jerusalem’s Upper City, which directly overlooks the Temple Mount.30 The craftsmen who were responsible for this mosaic may have received their geometrical training from the same school as the architects who designed the ground plan of Herod’s Temple Mount.
Having identified the plan of the Herodian Temple Mount, it now remains to see if we can determine the location of the Temple itself. The first step is to identify the principal axes of the Temple enclosure. I took the west-east axis to be parallel to the south wall and midway between this wall and the proposed north wall. The longitudinal north-south axis presented a less obvious choice because the west wall is not closely parallel to the east wall and the east wall does not follow an exactly straight course. Here I was helped by a key surviving feature of the Herodian Temple enclosure, a pair of passageways that led up to the Temple platform from the southern entrances known as the Huldah Gates.31 The western one is known as the Double Gate, and the eastern one as the Triple Gate. The Triple Gate is so named because it is fronted by a now-blocked triple archway, behind which is a lobby of three bays, as compared with the double entry and lobby of the Double Gate. However, as Warren observed, the Triple Gate “was a gateway of about the same style as the Double Gate, and is very likely at that time to have exactly corresponded to it in only having two passages.”32
Claude Conder, who worked on the survey of western Palestine after Warren, concluded that “these two double tunnels in the south wall, at the Double and Triple Gates, thus correspond to each other in their length, width and slope of ramp; the sill of each is on the same level.”33
These two entries are oriented approximately perpendicular to the south wall, although they are not situated symmetrically with respect to the center of the south wall. However, since they were part of the original Herodian scheme, and were the principal gateways to the Temple enclosure for the general public, I took the north-south axis to be on a line drawn through the midpoints between these two passages. I found additional support for this decision in parallels found in slightly older temple complexes with twin ramped entrances—the Hellenistic temple of Apollo at Didyma, in Ionia, begun in the third century B.C.E., and the Temple of Fortuna Primigenia at Praeneste, in Latium, Italy, dating to about 80 B.C.E. These two temple complexes, which were very influential in the development of classical architecture, both have ramped passages that are symmetrically situated on either side of one of their main axes.
The two Temple Mount axes thus drawn are perpendicular to one another to within half a degree. They intersect at a point now occupied by the Dome of the Chain (Qubbat al-Silsila), in front of the eastern entrance to the Dome of the Rock.
The focal point of this intersection must have coincided with an important feature of the Temple. It could not have been the Holy of Holies, the inner sanctum of the Temple, if only because it would have left insufficient space on the Temple Mount for the Court of the Women (‘Azarat ha-Nashim), to the east.
The better alternative is that this spot marked the location of the great sacrificial altar, the focus of public worship. Incidentally, this view is found in Jewish tradition. We may find it in the description of the Temple Mount by Rabbi Jechiel of Paris in the 13th century:
Round the Even Shetiyah [foundation stone]34 the Ishmaelite [Arab] kings have built a very beautiful building for a house of prayer and erected on the top a very fine cupola [the Dome of the Rock]. The building is on the site of the Holy of Holies and the Sanctuary, and in front of the mosque towards the altar [on its eastern side] is a structure of pillars [the Dome of the Chain] and the cupola is at the top of these pillars and it would seem that this was the place of the outer altar which was in the Court of Israel.35
In more recent times, Claude Conder also deduced that the Dome of the Chain coincided with this altar.36 Frederick Hollis and Miriam Rosen-Ayalon, an expert on al-Haram al-Sharif, have independently reached the same verdict.37 As further support for this idea, I found evidence that the altar in front of the temple of Jupiter Heliopolitanus at Baalbek stood at the center of its courtyard during the lifetime of Herod.38
Based on the location of the altar, we can locate the Temple—provided we can determine the modern equivalent of the cubit, the standard unit of length used in the ancient descriptions of the Temple. The cubit used in Herod’s Temple is quoted in the scholarly literature as anywhere between 17 inches (.43 meters) and 20.5 inches (.52 meters). From a study of artifacts and structures from the Herodian period, the scholar Arye Ben-David has deduced that the short cubit of that time had a length of about 18 inches (.465 meters).39 This is a fairly average value for the cubit and appears to be reliable. It is the one I will use (although a small percentage difference in the value of the cubit would not fundamentally affect my conclusions; it would merely incur a small shift of the Temple sanctuary relative to the altar).
The most detailed source for the Temple proper is the Mishnah, the first great rabbinic legal text, compiled in about 200 C.E. The altar was doubtless centrally placed in front of the Temple edifice, as it was in most temples of the classical period. This axial arrangement is consistent with the description and detailed measurements given in the Mishnah. According to the Mishnah, the distance between the altar and the Temple (that is, the sanctuary building) was 22 cubits.40 The Temple was 100 cubits long,41 and the length and breadth of the altar were 32 cubits at its base.42 Accordingly, the distance from the midpoint of the Temple to the midpoint of the altar was 50 + 22 + 16 cubits (that is, half the sum of the lengths of the sanctuary and altar plus the distance between them), which totals 88 cubits, or 134 feet.
Now, the distance between the center of the Dome of the Chain and the Dome of the Rock is 127 feet.43 This is not very different from the 134 feet between the centers of the altar and the Temple. So, just as the position of the altar coincides almost exactly with the Dome of the Chain, the Temple corresponds closely in position to the existing Dome of the Rock.
But this is not the only near coincidence. The diameter of the Dome of the Chain is 47 feet,44 close to the 32 cubits, or 49 feet, for the length and breadth of the altar. Likewise, the length and breadth of the Dome of the Rock, 162 feet, is not far removed from the 100 cubits, or 152 feet, for the length of the Temple.45
Thus, we see the following pattern emerging, with the Dome of the Rock and the Dome of the Chain, both of which date back almost to the very beginning of Islam in the seventh century C.E., reproducing the Temple and altar in size and respective locations.
This can hardly be mere coincidence. We know that after the Roman destruction of the Temple in 70 C.E., memories of this hallowed shrine remained very much alive. In the third century, the early church father Origen noted that a statue of the emperor Hadrian on horseback stood on the site of the Holy of Holies. By the third and fourth centuries, all that could be seen inside the area once occupied by the Temple proper was the exposed crag of al-Sakhra and two imperial statues, the one of Hadrian immediately to the west of al-Sakhra (below the spot where the Holy of Holies was located, according to my reconstruction), apparently overlooking the rock in a pointed gesture of triumphalism.46 The so-called Bordeaux Pilgrim in the fourth century C.E. was also able to discern the site of the altar, which was delineated in the marble paving.47 The location of the Jewish Temple was still recognized when the Arabs conquered Jerusalem: They respected the exposed rock, al-Sakhra, as the Mihrab Dawud (the Sanctuary of David) and the Qibla Bani Isra’il (the Focus of Prayer for the Children of Israel). The winged motifs portrayed on the late-seventh-century wall mosaics in the Dome of the Rock allude to the seraphim guarding the Ark of the Covenant, according to Julian Raby, a scholar of Islamic Art at Oxford University. Moreover, early Islamic traditions about the rituals held at al-Sakhra appear to reflect Jewish Temple ceremonies.48 Therefore, over the course of time, the site of the Temple and altar continued to be revered, although the original associations with the Temple Mount became blurred and were superseded by newer Islamic traditions.
Thus we believe that we have located both the site of the Temple and the altar in front of it. So far, we have relied entirely on an analysis of the surviving walls of the Temple enclosure. Can the argument be strengthened by other evidence? And how does al-Sakhra, the rocky crag that rises above the floor of the Dome of the Rock, fit into this grand scheme? What was its significance in the Temple? To these questions, we will turn in the second installment of this article, in the next issue of BAR (25:05).