Homepage    Articles    Locating Bethel    Battle for Ai    Benjamin Survey Chart    Exodus-Conquest Dating Fiasco

Locating Biblical Ai Correctly

by Dr. David Livingston

One of the greatest problems in identifying Beitin as Bethel is the uncertainty surrounding the location of Ai. Because of this, a number of theories have been advanced concerning its location:

  • Albright suggested that Ai was mistaken for Bethel by Bible writers (1934: 11; 1939: 16-17).
  • It was proposed that Ai was only an insignificant outpost of wooden structures, none of which remain (Owen 1961: 321).
  • Others have said that Bethel is near Shechem.
  • As a result of his excavations, Joseph Callaway’s conclusion is that et-Tell was Ai, but that the Conquest was in Iron I (12th century BC, Callaway 1968: 316) -- later than most scholars accept.
  • Grintz has suggested that et-Tell is Beth-Aven (Grintz 1961: 201).
  • Noth and others dismiss the biblical account as unhistorical and therefore unreliable, little more than etiologic (Albright 1939: 16). Thus it is understandable why Callaway concluded regarding the et-Tell=Ai equation: “Ai is simply an embarrassment to every view of the Conquest that takes the biblical and archaeological evidence seriously” (1968: 312).
Abram camped on a mountain with Bethel on the west and Ai on the east. Genesis 12:8

Criteria for Identifying Ai

Found in Genesis 12:8;  Joshua 7-8;  Ezra 2: 28; Nehemiah 7:32, 11:31;  I Samuel 13:5:

First, there must be a mountain (Heb. har) between Bethel and Ai.
Second, there must be a valley (Heb. gay) north of Ai.
Third, there must be a hiding place for the ambush.
Fourth, there must be a descent (Heb. morad) which leads to Jericho.
Fifth, there must be a geological feature fitting the description of the shevarim.
Sixth, there must have been a wall around the site with an entrance gate.
Seventh, it must be smaller than Gibeon (which is about 11 acres).
Eighth, it must be near Bethel (as seen below).
Ninth, it should show occupation in MB II, LB I, IA II, and Persian periods.

The location for Ai given by Eusebius and Jerome is indefinite. As one can see from these citations, and as Conder points out, Ai was unknown in the fourth century (Conder 1879: 108). A study of the accounts of the Palestinian surveyors of the last century, Robinson, Conder, Thomsen, McGarvey, Ritter, etc., indicates that Ritter summed up the situation well for his day, “There is now no uniformity among travelers in their judgments respecting the location of Ai” (Ritter 1866: 223). But, since their time, even with archaeological research, there is little more certainty today. One of the few people who seemed to be certain that et-Tell is Ai was Albright, as a result of Mme. Marquette-Krause’s early excavations (1939: 15).

Eusebius in Greek, and Jerome in Latin, give the location of Ai in their day.

'Aggai (Genesis 12:8). The sun goes down (over) Bethel, not far away. Bethel is situated going up to Jerusalem from Neapolis (Nablus) on the left at the 12th marker, and it still remains. 'Aggai is deserted and can only be pointed out. This is the record of Gai.

*Aggai ad occidentalem plagem uergit Bethelis, non multum ab ea distans. sita est autem Bethel euntibus Aeliam de Neapoli in laeua parte uiae duodecimo circiter miliario ab Aelia, et usque hodie, paruus licet uicus, ostenditur, sed et ecclesia aedificata est ubi dormiuil Iacob pergens Mesopotamiam, unde et ipsi loco Bethel, id est domus dei, nomen imposuit. Aggai uero uix paruae ruinae resident, et locus tantummodo demonstratur, et sciendum quod in Hebraeo G litturam non habet, sed uocatur Ai, scribiturque per elementum quod apud eos dicitur Ain (5, 26).

*Aggai goes west toward the region of Bethel, not much from it in distance. Now Bethel is situated on the left side of the road nearly 12 milestones from Aelia going toward Aelia from Neapoli(s). And to this day only a small village can be shown. But a church is built where Jacob slept on his way to Mesopotamia. The place itself he named "Bethel," it is "the house of God." Actually, there are scarcely any remains of Aggai, and the place is barely discernable. It must be known that Hebrew does not have the letter G, but it is called Ai, and written through the alphabet which among us is called Ain.

Ai Cannot Be Both Near Beth Aven and Far From Bethel

The Masoritic text of Joshua 7:2 is translated, “Now Joshua sent men from Jericho to Ai, which is near (, ‘im) Beth Aven, east (, miqqedem) of Bethel." In this description it seems as though Ai is near Beth Aven.

However, the Greek Old Testament (Septuagint, or LXX) leaves out the phrase, “. . . which is near Beth Aven . . .” Thus it reads, “Now Joshua sent men from Jericho to Ai, which is near (,‘im) Bethel.”

Not only this reference, but all other references clearly indicate Ai is near to and associated with Bethel, and Beth Aven is not.

Beth Aven Is Listed as an Independent Place Only Once

Although Bethel and Ai are early mentioned as twin cities twice (i.e., close together -- Genesis 12:8, 13:3), there is no reference to Beth Aven in the Pentateuch. The first mention of “Beth Aven” is in Joshua 7:2. Following that is an oblique reference in Joshua 18:12,13 where the “wilderness of Beth Aven” is used as a reference point. But the “wilderness of Beth Aven” can only be located after Beth Aven is positively identified.

In Joshua 16:1 and 18:21-28,  Beth Aven is not found in the list of cities belonging to Benjamin (nor in Ephraim, just north of Benjamin).

The only other reference is 1 Samuel 13:5. It must be ruled out as being at Beitin, since Beitin is mostly north of Michmash instead of west of it.

All Other References Place Ai Close to Bethel

There is no reference to Beth Aven in the Pentateuch.

“(Abram) . . . removed unto a mountain on the east ( miqqedem) of Bethel . . . with Bethel on the west ( miyyam), and Hai on the east ( miqqedem).” There is no mention of Beth Aven in Genesis 12:8.

“Ai which is beside (, mitsad) Bethel.” Joshua 12:9 (no mention of Beth Aven). Note also: means “side,” in Gen. 6:16, the side of the ark of Noah. In I Samuel 6:8, the jewels of gold are to be put on the same cart, next to, or immediately adjacent to () the Ark of the Covenant. II Samuel 2:16 tells of each warrior thrusting his sword into the side () of his opponent. Illustrations can be multiplied, but these make it clear that Ai and Bethel were near each other.

Later in Scripture, Bethel and Ai are listed together in:

  • Ezra 2:28 “The men of Bethel and Ai, 223.” (No mention of Beth Aven anywhere in the list.)
  • Nehemiah 7:32 “The men of Bethel and Ai, 123.” (No mention of Beth Aven.)
  • Nehemiah 11:31 “The children also of Benjamin . . . (dwelt) at . . . Aiya, and Bethel and in their villages.” (No mention of Beth Aven.)

Beth Aven is not listed at all, and apparently was not reinhabited after the Return from Babylon.

Ai Apparently Was Not Strategic Militarily

Jericho was strategic. It was the seat of  moon worship (in light of its name), just as Beit Yerach on Lake Kinneret had been the seat of moon worship around the time of Abraham. For a throng like the Israelites, most of whom had not experienced the miraculous deeds of their God, they needed an early victory to reassure them of Yahweh’s power over all the gods of Canaan.

However, Ai was small and, no matter where it will eventually be found, it could never be militarily strategic. It had some other function of importance to Israel. On the other hand, Bethel was strategic because it controlled the north-south road through the hill country.

We are not aware of anyone doing a study on the choice of Ai as the place to destroy in the hill country. Only one reason comes to mind: the battle at Ai was staged as much to destroy the army of Bethel, which came to help, as to destroy Ai itself.

The Meaning of the Name “Ai”

Before considering the topography, it will be of interest to examine the meaning of the name “Ai” (). Some maintain that it means “ruins,” or “heap of ruins.” However, that is the only name the city ever had. It is as doubtful that later Israelite settlers would rebuild it and call it “the ruins” (ha‘Ai ) as it is that the Canaanites would originally name their city “ruins.” The name must have an alternative meaning. Here are the results of research on this problem by four leading scholars:

  • J. Simons (1959: 270, ¶ 465): “The word means no more than ‘a heap of stones.’”
  • Y . Kaufman (1953: 77, fn. 46): “Ai does not mean ‘ruin’, but heap, a pile or piles of stones. , , , are always and only, names of inhabited places and never of ruins. Numbers 21:11, 33:44-45; Joshua 15:29;  Isaiah 10:28;  Jeremiah 49:3.”
  • J. Grintz (1961: 209-211):
    As for 'Ai, Biblical evidence is clear on this point that it was not a ruin but an inhabited place . . . Actually , not only is there no connection between 'Ai and Et-Tell but the word ‘Ai (and the more so ‘Aiyah) does not have the meaning of “ruin” (either in Arabic or Hebrew). . . . Et-Tell is not an unusual name. It is interesting that the same also applies to the name ‘Ai. Similar names exist in many places and they always, as far as they are known to us, indicate a settled town and never designated an actual ruin . . .”
  • Ziony Zevit (1983: 26; 1985: 62):
    Ostensibly the names Et-tell , "the tell," and ha’ay , "the Ai," always with the definite article in Hebrew , should support the identification of the site. [But] a common explanation, that both names refer to "the ruin," and hence the Arabic is a translation of the Hebrew , does not bear up under scrutiny. . . . the etymology of the word "Ai" negates any connection with a word meaning "ruin." Etymologically , Ai does not refer to a ruin. (Also see Zevit 1983: 27, 32.)

For those who insist that “Ai” means “ruins,” the lexicons say, “No.” Brown, Driver and Briggs Hebrew and English Lexicon of the Old Testament, says nothing about Ai meaning “ruins.” It says only that Ai was an “old Canaanite city near Bethel. . . .” (1962: 743). In Holladay’s A Concise Hebrew and Aramaic Lexicon . . . nothing is found about “Ai” meaning “ruins” (1971: 271).

And in Koehler and Baumgartner’s Lexikon the meaning of ha’ai is that Ai, or ‘ay, is a corruption of ‘iy. In Gesenius’ Hebrew Grammar, as well as in the lexicons mentioned, this word, 'iy, coincidentally similar to 'ay, does mean “heaps of ruins” (Kautzch 1963: 82) in Psalm 79:1; Jeremiah 26:18; and Micah 1:6, 3:12.

But ‘iy is not ‘ay. So possibly this word is causing confusion as to the actual meaning of ‘ay. At any rate, one can hardly appeal to lexicons for the meaning of “the Ai” as “the ruins.”

Considering genuine ruins, the word horbah (from which we get modern Khirbet) meaning “ruin,” would have been a much better choice than the enigmatic name Ha’Ai, or Ha’ Gai.

Another Problem with Calling Ai “Ruins”

But grant for a moment that the ruins of et-Tell (traditional Ai) were “the ruins” of Abraham’s time (Genesis 12:8). Then, after Joshua’s Conquest, a new “Ai” settlement was established close to et-Tell. Would the Israelites name their newly built settlement “The Ruins”?  It was a new town -- not a ruin. It is difficult to understand why so many scholars insist that “Ha ‘Ai” means “The Ruins.”

Kh. Maqatir (near et-Tell) is a candidate for Ai as we will see later. However, a situation making it problematical for Kh. Maqatir to be Ai is that nearby et-Tell (the traditional site for Ai) was occupied in Iron Age I. Maqatir was also occupied at the same time. So, -- if et-Tell was known as “Ai” at that time, how could it be that Maqatir was also called “Ai” during Iron Age I? It is doubtful they would BOTH name their town, “Ai,” and especially if it means “The Ruins.”

If “Ai” Does Not Mean “Ruins,” What Is the Point?

“The Ruins” cannot be the meaning of “Ha ‘Ai,” even though some continue to insist on that. The Ai of Abraham must have been an inhabited community (sometime between EB IV-MB II). Thus Et-Tell could not have been the Ai of Abraham’s time because it was in ruins when he passed by. Therefore, we should not consider Ai at et-Tell as “landmark ruins” in Abraham’s time. The point is -- if “Ai” does not mean “ruins,” then et-Tell drops out of the picture as “The Ruins” and we must look elsewhere for Ai.

A third candidate for Ai, Kh. Nisya, was a village in Abraham’s day. This was especially true if ‘Ai was an outpost for Bethel. The tall mountain east of Bethel (at el-Bireh) blocked the view eastward. Therefore, Ha‘Ai () could be, literally, the “eye” () eastward of Bethel.

Dates for Abraham’s Coming to Bethel and Ai

Some evangelicals hold to a date of ca. 2165 BC/BCE for Abraham’s birth, during EB IV or MB I. However, there is a wide difference of dates among them. Young’s Concordance and Bishop Ussher place his birth at 1996 BC/BCE; Anstey (1913:130) has his entry into Canaan in 1875 BC/BCE (provided the Exodus took place ca. 1445 BC/BCE). The LXX puts his entry into Canaan as late as 1700 BC/BCE (Finegan 1964:193). A recent account by Al Hoerth (1998: 60) dates Abraham’s birth to “shortly after 2000 BC.”

Thus we conclude that the dating of Abraham, as well as EB IV and MB I and IIA, is still somewhat fluid with a variety of opinions when these periods and subperiods begin and end.

The consensus puts Abraham in the Middle Bronze II era beginning ca. 2000 BC/BCE, rather than in Early Bronze. Thus, it is probable that Abraham entered the land and camped at Bethel and Ai (Genesis 12:8) during MB II. If that is the case, there was occupation at both Kh. Nisya and Kh. Maqatir for that period with many types of ceramics as evidence.

The only other time et-Tell was occupied was Iron Age I. This means that if et-Tell was at that time called “Ai,” then we can assume that no other site nearby would also be named “Ai” at the same time.

Three Possibilities for Ai:  (et-Tell, Khirbet Maqatir, Khirbet Nisya)

First Possibility: et-Tell

What can we conclude about et-Tell’s being identified with Ai? J. Simons gives four reasons why he believes et-Tell cannot be Ai:

  1. It is not near Bethel (at Beitin); it is approximately two miles away.
  2. It was not inhabited during the Conquest (by any reckoning).
  3. It is too large (larger than Gibeon).
  4. There is no broad valley to the north (Simons 1959: 270. See also Simons 1947: 311).

Simon’s reasons numbers two and three are valid. On the other hand, number four is not really correct. There is a valley to the north. The first reason may or may not be correct, depending on the meaning of Hebrew ‘im (“with”). As one walks over the area between Beitin and et-Tell, he hardly has the impression that they were “twin cities,” because of the considerable distance and open space between. But, in addition to Simon’s valid reasons, et-Tell is wrongly located in relation to Michmash. It is north of Michmash when it should be west of it. Furthermore, there is only a small hill, instead of a mountain, west of et-Tell, which hardly fits the biblical description. Probably the most important of these reasons is the fact that et-Tell could not have been inhabited during the Conquest by whatever date that event is calculated.

Finally, it does not have the proper archaeological periods present to match the Biblical requirements for Ai. (See chart of occupation for all three sites at the end of this article.)

Fig. 1.   Author at massive Early Bronze fortification wall around et-Tell.

Fig. 2.   ABR volunteers on tour of et-Tell examining Iron Age four-room house.

Second Possibility: Khirbet Maqatir

Excavations began at Maqatir under Bryant Wood’s direction in 1996. Considering Kh. Maqatir as a candidate for Joshua’s Ai, it does have walls and pottery for the period of the Conquest (ca. 1400 BC/BCE).

To begin, let’s assume that Kh. Maqatir was Ai at the time of Joshua’s Conquest. Sometime later it was resettled by Israelites -- possibly as early as Iron Age I (the pottery of which is present at the site). If so, the settlement should have continued through Iron Age II until the Babylonian Captivity, at which time all the families of Ai were carried away to Babylon. Then, at the end of the captivity, during the Persian period, some of the former inhabitants and their families resettled Bethel and Ai. However, Kh. Maqatir has neither Iron Age II remains, nor Persian.

In Ezra 2:1, 28;  Nehemiah 7:6,32; 11:31, the men of Bethel and Ai returned from the Babylonian Captivity “to their own cities.” So both Iron Age II and Persian periods must be present for a site to meet the conditions for Ai. If they resettled at a location other than Maqatir during the Persian period, there are no other sites with Persian remains in the area except at Beitin and Kh. Nisya. Excavator Wood accounts for this lacuna in his article in Bible and Spade (Winter 1999, p. 29) by suggesting, “The men of Ai could have settled anywhere in the vicinity of et-Tell. Since they were few in number, they would be difficult to locate archaeologically.” (However, see the Benjamin Survey Chart.) This might have been the situation after the Babylonian Captivity (during the Persian period). But it does not explain the absence of Iron Age II before the captivity when the population was denser.

Does Kh. Maqatir Meet the Requirements for Ai?

  • First, there must be a mountain (Heb. har) between Bethel (at el-Bireh) and Ai. There are some hills between the two.
  • Second, there must be a valley north of Ai. There is a valley.
  • Third, there must be a hiding place for the ambush. There is, but in full view of the army of Bethel as they run over to aid Ai.
  • Fourth, there must be a descent (morad) down to Jericho. There is.
  • Fifth, there must be a geological feature fitting the description of the shevarim. Maybe a rock outcrop below the site could be this feature.
  • Sixth, there must be a wall around the site with a gate. There is.
  • Seventh, it must be smaller than Gibeon (11 acres). It is much smaller.
  • Eighth, it must be near Bethel. It is 3.5 km or 2 miles in a straight line with several valleys between, which hardly describes “near .”
  • Ninth, in light of the references to Ai in Ezra and Nehemiah there should be occupation during the Persian period, after the Babylonian Captivity. And since this was a Reoccupation, there should be Iron II pre-captivity occupation present also. However, there is so far not a trace of Iron II nor Persian. Furthermore, Maqatir was occupied, following the Conquest, only during the Hellenistic period), therefore it did not suffer the destruction down to bedrock as did so many other nearby sites.
    See also the Benjamin Survey Chart. As we pointed out above, no other site in the region has Iron Age II and Persian, meaning that there is no alternative to Maqatir for those periods. Peter Briggs, in his dissertation, overlooked this fact when describing Maqatir (Briggs 2001: 108).

A New Location for Beth Aven

Bethel was equated with Beth Aven by Hosea (4:15; 5:8; 10:5,8,15). Nearly all scholars take Beth Aven here to be a pejorative name for Bethel. This probably means Beth Aven was no longer occupied in Hosea’s time and we note that Maqatir was not occupied then either.

A recent suggestion for the solution of the location of Beth Aven is, “. . . not only have Bethel and Joshua’s Ai been misidentified, but also Beth Aven. . . . The three sites that meet all of the Biblical requirements are: Bethel at el-Bireh (instead of Beitin), Joshua’s Ai at Kh. el-Maqatir(instead of et-Tell) and Beth Aven at Beitin” (Wood 1999: 106).

Is this really the location of these three towns? Bethel is unquestionably at modern el-Bireh (instead of at Beitin). But is Beitin to be equated with Beth Aven? I think not, even though I published that suggestion in the Palestine Exploration Quarterly (1994: 158). New evidence from Maqatir has caused a change of mind on the location of Beth Aven. Since that time a more satisfactory site for Beth Aven closer to Michmash has been discovered that fits well with the Biblical situation.

In view of the finds at Khirbet Maqatir, this new possibility for Beth Aven is Khirbet Maqatir itself.

Does Kh. Maqatir fulfill the requirements for Beth Aven? The archaeological profile there matches the Biblical profile for Beth Aven. Middle and Late Bronze pottery has been found, fitting the references in Joshua. There was a destruction during the Late Bronze period. Apparently, Maqatir was resettled during Iron Age I. Of interest is an Iron Age pagan cultic stand which has come to light in the Maqatir excavations. This might fit with the meaning of Beth Aven -- “house of wickedness.”

As in Joshua 7:2, Kh. Maqatir is found almost straight east of Bethel (now located at el-Bireh). Next, Maqatir faces the Judean wilderness (the “wilderness of Beth Aven”) as it should according to Joshua 18:12-13.

Mukhmas is southeast of Maqatir. With Bethel at el-Bireh and Michmash at Mukhmas, Kh. Maqatir is closer to Michmash than is Beitin, thus a better location for Beth Aven. Assuming the Philistines camped north of Michmash they would have been east and a little south of Maqatir. (See I Samuel 13:5.) In their flight the Philistines had to run right over Beth Aven (at Kh. Maqatir).

It appears to fit the situation well. (See also I Samuel 14:23,31.)

Third Possibility for Ai: Khirbet Nisya
Size of Kh. Nisya Compared with Gibeon (el-Jib)

The size of Gibeon is usually considered to be about 10-11 acres. But the Benjamin Survey (Finkelstein 1993: 46*, 235) says it is 60 dunams, or about 15 acres. Joshua 10:2 says that Gibeon was “greater than Ai.” This is interpreted as a reference to the comparative sizes of Gibeon and Ai. Ai should be smaller than Gibeon. Kh. Nisya is about 4-5 acres in size depending with which period one is concerned. In some periods it may have been smaller, in others larger than 5-6 acres.

The Benjamin Survey gives the size of Nisya as 15 dunams, or almost 4 acres (Finkelstein 1993: 32*,166), but this may be a subjective figure since few of the sites were actually measured by the Benjamin Survey team. Kh. Nisya meets the condition of being smaller than Gibeon.

* The asterisk is actually a part of the numbering system in the book. It signifies page numbers of the English section of the report.

An Excursis on The Difficulty with Discovering Canaanite Architecture in the Hill Country

Kh. Nisya is not unique, no early architecture (Early Bronze, Middle Bronze, or Late Bronze) was found at other nearby sites either, although they were occupied in those periods. At least three close to Kh. Nisya are Tel el-Ful, Tel el-Jib, and Tel en-Nasbeh.

The topography around Khirbet Nisya matches every detail given in the account of the destruction of Ai in Joshua 7-8. Ai should be east of Bethel (at el-Bireh) on the other side of Jebel et-Tawil (Genesis 12.8). Excavations at Khirbet Nisya have shown that the site was occupied during the biblical periods when Ai was in existence. Periods of significant occupation, determined by ceramics, artifacts and architectural evidence are: Middle Bronze II, Late Bronze I, some Late Bronze II B, Iron Age I and II, Persian, Hellenistic, Early Roman, Byzantine and Early Islamic (Livingston 1987, 1989).

At this time no architecture earlier than Iron Age I has been uncovered at Khirbet Nisya. In considering this problem, it should be recalled that no early architecture (Early Bronze, Middle Bronze, or Late Bronze) was found at other nearby sites either, although they were occupied in those periods. At least three close to Kh. Nisya are Tel el-Ful, Tel el-Jib,and Tel en-Nasbeh. Ceramics were found for earlier periods at each of these sites, but later building and agricultural activities obliterated earlier structures. Of Tel el-Ful, Paul Lapp said (Lapp 1981: 6):

From a very few maceheads and potsherds, Sinclair judged that the earliest occupation at Tell el-Ful was Middle Bronze; there were no building remains. . . . The building of the fortress and the housing on the eastern edge of the mound destroyed most of the pre-fortress evidence, except what could be found in pockets in the bedrock. It is possible that some of the silos were cut in this period, but they were reused in later periods, rendering a date for most of them impossible.

Pritchard said likewise of Gibeon (speaking of Early Bronze, but also implying Middle Bronze). It is highly probable that Gibeon, like its neighbors in this period, was surrounded by a strong city wall, although no trace of it has been found:

They (the Romans) had dug their foundation trenches down to bedrock, completely demolishing the earlier buildings as they salvaged all usable building stones (81). . . . They scraped bare the bedrock for secure foundations and thus destroyed whatever evidence there may have been of previous occupation (103). . . . Perhaps in an area not yet excavated -- to date we have dug into but a fraction of the total area -- the remains of the “great city” of Joshua’s day are to be found (158). (See also page 136.)

[The evidence, or lack of it, Pritchard speaks of in relation to Gibeon fits the situation at Kh. Nisya.]

Magen Broshi writes of Tel en-Nasbeh (Avi-Yonah 1976: 914):

(Of the Calcolithic-Early Bronze Age) No architectural remains whatsoever were discovered on the mound from these periods. The only evidence of occupation were the sherds found scattered over the surface of the site and in the caves.

C.C. McCown (one of the excavators at Tel en-Nasbeh and editor of the excavation report, 1947: 68) said this about the situation there:

When the entire lack of houses of EB date is considered, the total amount of EB Age pottery found at TN is surprising. [Obviously there was occupation during this Caananite period, but all architectural remains for EB have been destroyed.]

A little farther away, Israel Finkelstein discovered phenomena similar to the above at the hill-country site of Shiloh (Finkelstein 1988: 211).

Even before we began to excavate Shiloh, we were aware of two problems typical of hilltop sites. The summit had been badly eroded and bedrock was exposed in many places. Moreover, since the masons of every period attempted to lay the foundations of their buildings directly on bedrock, the activities of later periods -- notably Roman, Byzantine, and Medieval -- severely disturbed, or even razed the early strata on the southern slope of the tell and on the summit. (Our emphasis.)

Other sites parallel the destruction of early architecture at Kh. Nisya.

Fig. 3. Narrow place widening out below shevarim.

Fig. 4. Shevarim now much wider than it was 3400 years ago. Boulders continue to fall away, ever increasing the width.
The shevarim () pictured here is not far below Kh. Nisya in Wadi Sheiban which becomes Wadi Qelt going down to Jericho. The left photo shows volunteers going through the defile; the right photo is up close. On the right large boulders are shevaring off the outcrop on that side and falling into the stream. With only a little imagination, one can visualize that this was a very narrow place 3400 years ago. Since then, large boulders splitting and falling off have filled the stream while increasing the width of this originally very narrow defile.

Ubiquitous Walls Around Canaanite Cities and Towns

The biblical narrative is quite clear about walled cities in the land before the Israelites arrived. Although most scholars do not believe that there ever was a “Conquest,” most of those who do believe have opted for a late date at which time (LB II) there were very few walled cities.

Conversely, the biblical narrative clearly says that when Moses sent spies into the land, ca. 1440-50 BC/BCE, (about the time of the new date for the end of MB II), to bring back a report, their chief complaint was that “the cities are walled, and very great . . .” (Numbers 13:28). In Deuteronomy 1:28 the spies cried, “The people are greater and taller than we are; the cities are great and walled up to heaven . . .” In spite of their lack of faith that the Lord would make them victorious, Moses told them they would possess “. . . cities great and fenced up to heaven” (Deuteronomy 9:1). Later (ca. 1400 BC/BCE), when Joshua led Israel into the land, mention is made of the conquest of walled cities: Jericho was walled and well fortified (Joshua 2:15; 6:5); the army of Ai chased Israel “from before the gate (you don’t have a gate without walls) even unto the shevarim” (7:5); “The rest which remained of them (of the five kings) entered into fenced cities” (10:20; 11:13; 14:12; 19:35).

Considering only archaeological finds, and opting for a late “Conquest” date, one might conclude that the Bible narratives are describing a nonexistent situation concerning “walled” cities at the time of the Conquest. However, this is not the situation when considering a biblical date of ca. 1400 BC/BCE for the Conquest.

Looking at the archaeological evidence for walled cities in the land at an early Conquest date (ca 1400 BC/BCE), Benjamin Mazar, in 1968, was already noting the “. . . great upsurge in the construction of large fortresses in the hill country and the Shephelah” during the Middle Bronze Age (1968: 92). And since there is no cultural break between Middle Bronze II (and III) and Late Bronze I, there is every reason to conclude the situation Mazar mentions above and Dever refers to below, continued through LB I. An abundance of Middle and Late Bronze ceramics and artifacts found at Kh. Nisya make these remarks (from his article on Middle Bronze II) by William Dever cogent:

Beginning with Middle Bronze II, and continuing until the end of Middle Bronze III, the archaeological record at nearly every site shows a continual process of defensive construction. . . . Not only are all the larger sites fortified, as might be expected, but even towns and villages as small as 2 to 4 acres are surrounded by city-walls. . . . Indeed, scarcely a single excavated Middle Bronze Age site in Palestine has failed to yield formidable fortifications (1987:154, our emphasis).

Although Dever leaves a little room for exceptions to his rule, the author’s examination of excavation reports on 30 MB II/LB I sites (by 1984) showed almost no exceptions to walled cities. One exception is Gibeon, but only because the wall has been obliterated in all excavated areas. The 30 examples include: (central hill country) Beitin, Beth-Shemesh, Tell el-Farah (north), (Jordan Valley) Jericho, (northern region) Tel Acco, Tel Dan, Hazor, Megiddo, Ta’anach; (southern region) Beth-Zur, Tel Beit Mirsim, Lachish, Tel Malhata, Tel Masos; (coastal region) Achzib, Tel Ajjul, Aphek, Tel Ashdod, Tel Burga, Tel el-Farah (south), Tel Gerisa, Gezer, Tel el-Hesi, Tel Mevorakh, Tel Mor (1 1/2 acre). Tel Nagila, Tel Poleg, Yavne Yam, Tel Zeror, Kh. Zukeriyeh (1 acre). Dozens of others could be added to this list. We have noted two sites of one acre each (or slightly more) to point out that even the smallest sites were walled. An exception to this are the small clusters of houses in the countryside where farmers lived outside the walls (as in medieval times when serfs lived outside castle walls but went inside as part of the fighting force during invasions). This is, no doubt, the type of “unwalled settlement” Gophna and Beck (1981) have in mind below. Gophna described one such “village” in the Jordan Valley (1979: 32). Actually, it was only a cluster of five or six houses on less than one acre.

Admittedly, some scholars maintain that there were many villages (by their calculations, as many as one-third in MB II) which had no protective walls during MB II and LB I  (Gophna and Beck 1981). However, Gophna and Beck’s chief interest was in population studies and their information regarding unexcavated sites was only from surface surveys. We agree that surface surveys are helpful in identifying MB/LB sites. However, it is a mistake to conclude that there were no fortifications based on surface surveys. Later Broshi and Gophna acknowledge:

Complete accuracy is impossible here because we do not know whether any number of sites were surrounded by ramparts, and this fact often cannot be established without an excavation (Gophna and Broshi 1986: 88).

We may conclude then that if remains from Middle Bronze II and III (and LB I) are present at all, a city, town, or village would be protected by a wall. Since Khirbet Nisya has considerable ceramic and artifactual remains from these periods, and is larger than the small villages mentioned above, we should expect it to have had a wall and gate even though we have not yet found evidence for these at the site.

Finally, Ze’ev Herzog noted:

The most striking change in the LB cities is the total disregard for any form of general fortification. In most cases even the walls of the MB were dismantled and domestic units were constructed over them. (Herzog 1997: 272)

Khirbet Nisya Was Occupied in MB II/LB I

Sherds and artifacts indicative of MB II and LB I, are included on several plates of drawings (See Plates 3.1-3.5, pages 34-43 in  Khirbet Nisya: the Search for Biblical Ai, 1979-2002). These are only examples from among many. But there is sufficient quantity of each type, and variety of styles of every kind of water and storage jar, juglet, cooking pot, bowl, chalice, krater, etc. to indicate that Kh. Nisya was inhabited during MB II and LB I.

These are not simply chance finds dropped by wandering Canaanite shepherds. They indicate habitation during these periods. For example,60-plus sherds (not all are displayed on Plate 5.1 of  Khirbet Nisya: the Search for Biblical Ai, 1979-2002) of handmade, flat-bottomed, pie-crust-rimmed cooking pots are considerable when compared with those found by Kenyon in her much larger excavations at Jericho.

Kh. Nisya Should Have Had Walls During MBII/LBI

Evidence of walls and a gate have not yet been found at Khirbet Nisya. The reason is not difficult to understand. First, it is important to note how city walls were built in late Middle Bronze II and into Late Bronze I. Almost without exception at MB sites, city walls were built on a fieldstone base about one m high (for drainage?) topped with mudbricks to the desired height. Examples of this type of wall were found at Beth-Shemesh (Avi-Yonah 1975 1:250), Shechem (Seger 1975:35), Beitin (Avi-Yonah 1975 1:192), Jericho (Ibid. 2:562), Dan (Biran 1984:8), Hazor (Avi-Yonah 1975 2:481), Beth-Zur (Ibid. 1:265), Aphek (Kochavi 1975:30), Tel Nagila (Avi-Yonah 1975 3:896), Kh. Zurekiyeh (1 acre in size; Gophna and Ayalon 1981: 69), to cite a partial list.

Only when these mudbrick walls were covered rather quickly with debris have they been preserved. If not, the bricks have eroded due to winter rains. An example of the ephemeral character of mudbricks are the MB mudbrick walls and gate at Tel Dan. The mudbricks were deteriorating so rapidly that there was concern whether they can be preserved at all. On the other hand, the fieldstone bases remain, unless removed or reused in terrace walls or houses by later occupants.

There must have been a wall around Khirbet Nisya during MB II and LB I. It was probably constructed of mudbricks with a fieldstone base. Because the soil of all earlier strata on the site has been so disturbed and so little stratigraphy remains, there is no hope that any mudbricks of the wall will ever be found, as they have not been found at el-Ful, en-Nasbeh, Gibeon, and Shiloh. As for the original fieldstone base, it has likely been dismantled or incorporated into a terrace wall. This has been evident over the seasons of excavation at the site. We have consistently found that terraces where one would expect to find walls, were rebuilt from bedrock to the top as late as the Early Islamic period. But likely, many of them were built during the Iron Age.

Does Khirbet Nisya Meet the Requirements for Ai?

  • First, there must be a mountain (Heb. har) between Bethel and Ai. And there is: Jebel (Mount) et-Tawil (“Tall One,” or “Long One”).
  • Second, there must be a valley north of Ai. There is a deep valley just north of Kh. Nisya, with a hill beyond, where Joshua could have stood and been seen easily by the ambush west of Ai.
  • Third, there must be a hiding place for the ambush. In this case, the ridge west and south of the site is admirably suited to hiding an ambush which would be near Ai but out-of-sight from Bethel (el-Bireh) since it forms a semicircle behind Nisya. (See map.)
  • Fourth, there must be a descent (morad) which leads to Jericho. Wadi Sheiban becomes Wadi Qelt and descends all the way from just south of Nisya to Jericho.
  • Fifth, there must be a geological feature fitting the description of the . There is a shevarim - a very narrow place in the wadi about 2.5 kms southeast of the site which could be considered the shevarim (“breakers” or “broken”). Here huge boulders, from rock outcroppings above the wadi, are even now breaking off and falling down into the wadi. 3400 years ago this narrow defile could have been a trap where the men of Ai (in the first battle) cut down 36 Israelites. A second possibility is that the first contingent fled northward, then to the east on the ridge going to Jericho.
  • Sixth, there must be a wall around the site with an entrance gate. This has, as yet, not been discovered.
  • Seventh, it must be smaller than Gibeon (11 acres). Nisya is only half as big as Gibeon.
  • Eighth, it must be near Bethel (in el-Bireh). It was a twin city with Bethel, about one km from it.
  • Ninth, it was occupied in four necessary periods: MB II, LB I, IA II, Persian (all of which can be seen on the pottery plates of the excavation report in Khirbet Nisya: the Search for Biblical Ai, 1979-2002).

Conclusion in Light of the Evidence

We concluded that there was an MB/LB settlement on the site. We found evidence on several terraces on the south and east sides where the soil was mostly dark gray to black from habitation sites. In Areas 50-54 pottery in the fill was mostly MB II-LB I. It is clear that at a relatively late date the soil in the center of the site was transported to use as fill around the foundations of buildings and to enlarge areas under cultivation, thus destroying MB/LB remnants.

In Areas 100 and E/G most of the pottery was Persian. In Area 102 it was mostly Hellenistic. All these, however, also had ceramics as late as Byzantine, indicating that the terraces were rebuilt in that period.

As mentioned already, many of the building stones have been moved or removed (and possibly transported to Ramallah and el-Bireh) for building purposes. Thus, not only little remains of ancient buildings on the site, but any town walls from early times have also disappeared.

Where one would expect to find a town wall on one of the terraces, the terrace itself was torn up and entirely rebuilt from bedrock to the top-soil fill and retaining wall. Other sites in the surrounding countryside experienced the same phenomena -- at nearby Tel en-Nasbeh and at Tel el-Ful. At Gibeon, Canaanite walls have never been found, even though we know it was occupied at that time, evidenced by tombs of the periods and pottery. It is possible that Kh. Nisya may have been an open village (without walls) during the Canaanite periods. But it is more reasonable, in light of cities, even villages, being well fortified in MB II and on into LB I, that the walls were dismantled later.

Is Khirbet Nisya the site of biblical Ai? In light of its match with the biblical account for topography, geography, and archaeological periods present, it should be considered as a candidate. Even though no fortification walls have yet been found, the site must have been walled some time in the Canaanite periods.

Table Showing Occupation Periods at Three Sites.

Bible Periods
Khirbet Nisya Et-Tell Khirbet Maqatir
Early Bronze       Occupied?   X  
Middle Bronze 1 X   Genesis 12:8      
Middle Bronze 2 X   Genesis 12:8, 13:3 X   X
Late Bronze 1 X   Joshua 7-8 X   X
Late Bronze 2       Occupied?      
Iron Age 1       Occupied? X X X
Iron Age 2 X   Isaiah 10:28
Persian X   Ezra 2:28
      Nehemiah 7:32
Hellenistic   X   X
Early Roman   X   X
Late Roman X   Eusebius'

Homepage    Articles    Locating Bethel    Battle for Ai    Benjamin Survey Chart    Exodus-Conquest Dating Fiasco

© 2003 David Livingston