A history of the Dead Sea Scrolls

Somebody once wrote a line of Isaiah above another line and on down the margin. This page follows that leather from a hole in a cliff through a shut door, a war and a ruled box under Miscellaneous for Sale, then lays the Isaiah scrolls beside the medieval text word by word, and says what the count shows and what it cannot.

A history in 17 parts · about 1 hr 23 min · c. 125 BC – 2026

Isaiah 40:7–8 · By a later hand

The Grass Withers

In the thirty-third column of a leather roll, somebody has written a sentence where there was no room for it. A thread of smaller writing runs above a line of older writing, carries on past the end of the line, turns the corner and goes down the left margin for about ten lines, close beside the linen stitches that join this sheet of skin to the next. The two inks differ in tone. Down the margin the letters lie on their sides, with their tops toward the seam. Near the start of the thread, where a name belongs, stand four dots. The roll is the Great Isaiah Scroll, the one complete copy of Isaiah to have come out of the caves by the Dead Sea, and the line under the thread is Isaiah 40:7–8, the verses in which the grass withers.

The first hand has the grass wither once: “The grass withers, the flower fades, and the word of our God stands forever.” A later hand wrote in the longer form above the line and on down the margin, its words turned a quarter turn, and wrote the divine name as four dots. One study puts that hand, by its script, at about 100 to 75 BC. In most English Bibles Isaiah 40 has the long form, in which the grass withers and the flower fades twice, and between the two the breath of Yahweh blows on it and the people are like grass. The Greek has the short form, and so does 1 Peter. The Greek is the old translation of the Jewish scriptures, whose story is told on the Septuagint page. 1 Peter 1 quotes the verse with one withering, and with “the Lord’s word” for “the word of our God.”

The four dots stand where the long form has the divine name, the four letters that Bibles render as Yahweh or as the LORD. On the line, all through the roll, the first hand writes those letters out in full, in the same square script as the words on either side. Above the line the Name is four dots. “Withers, fades” comes twice in the long form, and an eye can pass from the first pair to the second and go on from there, leaving out whatever lies between. That is one explanation. The other is that the source the first hand copied was shorter, and the longer form grew up in other copies afterward. The first hand wrote straight on here, with no space left, and no other scroll survives at these verses, so the only Hebrew copy of the short form is this one, and it has been corrected.

The Greek does not decide it. A translator’s eye could pass over a repeated phrase as a copyist’s could, or the two could share a shorter Hebrew. The scroll’s modern editors count the long form among the additions that the Hebrew behind a modern Bible took on over time, and another specialist sees an eye that passed on, in the scroll and in the Greek alike. Whether the first copyist’s eye slipped or his source was shorter, nobody can say. Either way, the words written above the line are the form that the printed Hebrew Bible has at these verses.

Since 2021 one English Bible has marked the place in its text. The New Revised Standard Version, in its updated edition, prints the long form’s extra words inside double brackets, and does the same for five other passages that the scroll’s first hand lacks. Its preface keeps those brackets for the few passages that are generally held to be later additions to the text, kept in it for their age and their weight in the tradition. The words are still printed, and still read, and a note at the foot of the page says that a manuscript from the caves at Qumran and the Greek lack them. That is how the first hand’s short form reaches a reader’s page.

This page works from photographs and a modern transcription. It has laid the whole roll beside the medieval Hebrew text that modern Bibles translate, word by word, and the count comes after the story. The story left a mark here too. On the back of the skin, behind the next column and near that seam, sits a patch of modern paper, stuck on to hold a weak place together by one of the men who handled the scroll on its way to a buyer.

1948 to 1961 · Who said it first?

Who Must Be a Goatherd

In the story as it was later told, a young Ta’amireh goatherd known as Muhammad edh-Dhib, the Wolf, threw a stone into a hole after a strayed animal and heard something break. The Ta’amireh are a Bedouin people whose country, an American traveler wrote after camping with them in 1838, lies between Bethlehem, Tekoa and the Dead Sea. In the fuller tellings the animal has gone up into the limestone cliffs above the sea, and he has climbed after it, calling. The hole is high in the rock face and not much wider than a man’s head. The stone goes in, and the sound that comes back is pottery breaking. He runs. The next day he returns with an older companion, and the two squeeze through and drop into a narrow cave where jars stand along the walls, some with lids like bowls, among broken pieces and rock fallen from the roof. In one jar they find bundles of rag, under the rag something black, and under the black a roll of smooth brown leather that reached, as they told it afterward, from one end of a tent to the other.

Who said it first, and when? Put to this story, the question sorts it into layers, and the first layer has no goatherd in it. The first full account in English, in September 1948, has Bedouin carrying goods from the Jordan valley, a cave partly fallen in, and jars crushed by the fall. It has no goat, no stone and no name. John Trever, the young American scholar who photographed the scroll in 1948, printed it as he had it from a monk of St. Mark’s, a monastery in Jerusalem, and the monk’s brother, a customs officer at the bridge over the Jordan. They had it from the Bedouin through others, so the account was third-hand, and it came from the people who were holding the scrolls.

The word goatherd reached print in 1949 on a guess: the original finder, wrote Lankester Harding, Jordan’s director of antiquities, “must be a goatherd.” He had not met him. His own footnote says that Joseph Saad of the Palestine Museum had met him since. That July Harding spoke in London, and an American journal reported him as saying that the cave was found by a Bedouin searching for a lost goat, with an exclamation mark after the goat. It gave the year as 1937, though Harding’s own report of that season said 1947. By 1955 the stone was in a book printed in New York, which set the versions side by side. He was herding goats, or looking for a lost sheep, or he and one or two others were taking goods across from the Jordan, perhaps smuggled, or they had sheltered in the cave from a thunderstorm. In another a runaway goat jumped into the cave, he threw a stone after it, and the sound of breaking pottery drew him in.

In October 1956 the finder himself was heard. A man of Bethlehem took down his statement on 23 October, and William Brownlee, Trever’s colleague in the winter of 1948 at the American School in Jerusalem, printed it the next year. It gives 1945 for the year, an opening at the top of the cave like the mouth of a cistern, ten jars broken, a skin bag of scrolls hung up for more than two years, and pieces of the leather shared out to make sandal straps. Trever answered in print that no Bedouin would keep an antiquity for two years, and that the opening described was not this cave’s. Neither Trever nor Brownlee ever met him. When Trever’s researcher went looking for the finders in 1961, the men he interviewed were sent by the mayor of Bethlehem, and they asked to be paid. The researcher noted in his papers that they were cross with the dealers of Bethlehem for having exploited them. The finders who spoke in 1949 had been liberally tipped as well, and the speaker of 1956 was paid.

Set in the order they were printed, the tellings have a steady core and a moving surface. Bedouin, a hole, jars, cloth and a sale at Bethlehem are there from the first and stay. The name, the goat, the stone, the number of jars and the year come late or shift from one teller to the next. It was probably the winter of 1946–47. The men who could have said gave different years, and no one wrote it down at the time. Between them the tellings offer 1937, 1938, 1942, 1945 and the spring, the summer and the autumn of 1947, and the winter is itself a reckoning, made partly from the interviews of 1961 and partly by counting back from the sale. The year usually printed, 1947, is the year of the market.

The goat reached print later than the goods, and it came from nearer the man. Saad spoke to the goatherd in 1949, and a photograph of that year in the archive of the Ecole Biblique, the French Dominican school in Jerusalem, shows him at the cave with a companion. Roland de Vaux of the Ecole, who dug the cave with Harding, had him on the workforce in 1952 and made him tell his story in front of the other men. In May 1958 de Vaux told Trever that he had sat with him on a rock by the cave’s mouth and heard the story there. The goods and the goat need not exclude each other, since men driving goats across the Jordan to sell them are carrying goods.

He was somewhere between fifteen and eighteen when he went in, by the ages that others gave him, and by the word of a kinsman an orphan, a poor goatherd with no flock of his own. After the season at Qumran he worked as a laborer on another excavation, near Nablus in the north. In 1949 the Ta’amireh threatened legal action over the scrolls against the head of St. Mark’s and his diocese. A study of the cave’s records published in 2017 calls the Bedouin reports reasonably consistent, while other scholars set the men of 1961 aside as witnesses altogether. What survives of the finders’ own words was taken down and printed by other men.

About AD 800 Timothy, head of the Church of the East, the church of the Christians of Iraq and Persia, wrote to a metropolitan of his church about something he had heard from Jews lately converted, men he trusted. Ten years before, they said, an Arab hunting near Jericho lost his dog into a cleft in the rock, went in after it, and found a chamber inside the mountain with many books in it. He told some Jews in Jerusalem, and a crowd went out and found books of the Old Testament and others in Hebrew. Timothy wrote asking for the books to be searched for verses that the New Testament quotes and that he could not find in the Old, and he guessed that Jeremiah or Baruch had hidden them. Among the books, the man who told him said, was a Psalter with more than two hundred psalms.

July 1947 · By Samuel’s account

The Door at St. Mark’s

St. Mark’s stands in the heart of the Old City of Jerusalem, just south of David Street. On a Saturday in July 1947 a priest of the house opened its door to men carrying very dirty rolls of leather, several of them wrapped in dirty cloth with something black on it. He looked at the writing and saw that it was not Syriac. He told them his superior was not interested. They said they had been sent for, and he sent them away. His superior heard of it over lunch. He was Mar Athanasius Samuel, metropolitan of the Syriac Orthodox Church in Jerusalem, whose books and liturgy are in Syriac, and he had sent for the men and been waiting for them all morning. The scene is his own account of 1949, phrased by Trever from repeated interviews and checked by Samuel.

The men had come through Khalil Eskander Shahin, called Kando, a merchant of Bethlehem, the market town where the Ta’amireh came in on Saturdays. By the earliest accounts they came to him because the writing was taken for Syriac, and he was a Syriac Orthodox merchant who also mended shoes. In some tellings a Muslim sheikh of the town had looked at the script first and sent them on. By one later report a Muslim dealer was offered the great scroll of Isaiah for twenty pounds and would not pay that much for a thing he did not believe was old. Kando passed word to George Isha’ya, a merchant of the same church in Jerusalem who worked with him, and Isha’ya to the metropolitan. Every step of this rests on the memoirs of men who later fought one another over the scrolls, and Kando, who died in 1994, never wrote his down.

The Syriac Orthodox had been in Jerusalem since at least the fifth century, and St. Mark’s itself goes back to the Crusaders, but the community of 1947 was largely new. The massacres of Christians in the Ottoman Empire during the First World War, which they call the Seyfo, had driven refugees south, and the refugees swelled the church in Palestine from a few hundred to a few thousand. Samuel was one of them, a peasant’s son from a village now on the border of Turkey and Syria. In April, in Holy Week, Isha’ya had told him of Syriac manuscripts found by Bedouin in the wilderness by the Dead Sea, wrapped like mummies, and a few days later Kando and Isha’ya brought one roll to the monastery. He saw that the writing was Hebrew, broke a small piece from the margin and burned it, and knew by the smell that it was leather or parchment.

He bought later that summer, when two of the men came back. He got five rolls, two of which turned out to be pieces of one, and the great scroll was among the other three. By Kando’s later account he left the scrolls with Samuel on a security of twenty-four pounds. The American tax case of later years put their cost at fifty. Samuel held in 1949, and again in his memoir of 1966, that but for the priest at the door more scrolls would have come to him, since the men had at least seven that Saturday by Kando’s report. A third man had gone off with his share. If the five were among the rolls at the door, as Samuel believed, the door had turned away Isaiah.

In August a priest of St. Mark’s went out to the cave with Isha’ya and slept in it. The floor was covered with cloth and bits of manuscript among the broken jars, and one whole jar stood there, too heavy to carry in the heat. They brought back nothing. Then the experts turned Samuel away. A man of his own community who worked at the Palestine Archaeological Museum, Jerusalem’s museum of antiquities, and who read Arabic and Ottoman papers more than Hebrew, said the scrolls were worthless and told him not to pay a shilling for them. A friend at the Ecole Biblique brought a Dutch Dominican who knew Hebrew, and the two thought them fairly modern. When the Dutchman told his colleagues that evening, they reminded him of Moses Shapira, the Jerusalem dealer whose leather strips were called forgeries in 1883, and whose story is on the page on biblical archaeology. By other accounts the same visitor named the largest roll as Isaiah at once and saw that the scrolls were old, and the tellings conflict.

Late that summer or early in the autumn the museum man brought a Jewish scholar of his acquaintance, Toviah Wechsler, who agreed that the scrolls were not ancient and thought them perhaps taken from a synagogue. By Samuel’s report he pointed at a table and said that a box of that size full of pound notes would not pay for them, if they were really two thousand years old. Two years later, when he had come to think them genuine, he explained in print what had misled him. It was the corrections in the margin of one of the scrolls, written in ink so black, he said, that it could not be ancient.

1760 to 1939 · By the highest authorities

No Probability

The experts of 1947 had a sentence behind them, printed in 1895. “There is, indeed, no probability that we shall ever find manuscripts of the Hebrew text going back to a period before the formation of the text which we know as Massoretic.” Frederic Kenyon, a former fellow of Magdalen College, Oxford, wrote that in a history of the text for general readers, and it summed up more than a hundred and thirty years of looking. The looking began at Oxford. Benjamin Kennicott, a scholar there, set out to compare every Hebrew manuscript of the Bible he could reach, reading by reading, with the printed text. A subscription of nine thousand pounds paid for the work, and he heard, he wrote, “the Whisper of Detraction” about what a comfortable thing such a trust must be in one man’s hands. A woman of Wokingham whom he had never seen left him two hundred pounds in her will, to be paid when he finished. He had begun, by his own account, “strongly prejudiced in favour of the Integrity of our Hebrew Text,” as nearly all learned Europe was. In 1748 he became convinced that it “had suffered from transcribers, at least as much as the copies of other antient writings.” Robert Lowth, later bishop of London and the translator of Isaiah, had pointed him to the chapter that did it, 2 Samuel 23.

In 1760, in the first of his yearly reports to the subscribers, Kennicott set down what he expected the oldest copies to show: “the older the MSS are, the more they differ from the printed Text; for they generally read more agreeably to the Context, and also to the antient Versions.” The ancient versions were the old translations, Greek, Syriac, Latin and the rest, made from Hebrew copies long lost. His own oldest manuscripts were medieval. Lowth went further with Isaiah. At 21:8, where the Hebrew has a watchman cry out like a lion, he made “no doubt” that the true reading was the seer, the one who looks out, as the Syriac translator had found it in his copy. Kennicott’s Bible came out between 1776 and 1780, with readings from more than six hundred manuscripts, and another scholar added hundreds more within ten years. The older text that Kennicott had hoped for did not appear.

By 1895 Kenyon could give the result of all that labor in a line, that every Hebrew manuscript represented the same text, “and that without substantial variation.” The medieval copies could not reach behind themselves. They all came down from the work of the Masoretes, the scholars of Tiberias, by the Sea of Galilee, who from about the seventh century gave the bare consonants of Hebrew their vowel signs and filled the margins with notes for keeping them. The text they handed on is the Masoretic text, and the Masoretic Text page tells its story. Its best-known witness is a codex, a book of leaves bound at one edge rather than a roll, copied in 1008 or 1009 and known as the Leningrad Codex, a complete Hebrew Bible and the manuscript that most modern Hebrew Bibles print. The oldest Isaiah with a date that holds is the St. Petersburg Codex of the Prophets, written in 916.

Scholars on both sides of the quarrels over the Bible accepted that agreement and read it in opposite ways. S. R. Driver, Regius Professor of Hebrew at Oxford, held in 1890 that the Jews had come to the work late, when the damage was done, and then “proceeded to guard their sacred books with extraordinary care, with the result that corrupt readings were simply perpetuated.” W. H. Green, who had then taught at Princeton Theological Seminary for more than half a century, ended his book on the text in 1899 by saying that “no other work of antiquity has been so accurately transmitted.” Why so few old copies were left had more than one answer. The Talmud has a worn-out scroll of the Law buried beside a scholar, set first in an earthenware jar. Skin lasts only in dry ground. In that part of the world, dry ground means the desert and Egypt.

Kenyon’s next chapter opened with August 1883 and Shapira’s strips of Deuteronomy, in an ancient script, with an eleventh commandment added to the ten. He wrote that they seemed enough to cast doubt on the whole received text of the Old Testament. The sensation lasted a few days. Then they were shown to have been cut from the margins of an ordinary synagogue roll. His sentence about older manuscripts came straight after. The older text, he went on, could be reached only through the earliest translations made from it. The book stayed in print, and writers on the scrolls in the 1950s cited the sentence from its edition of 1939.

November 1947 · By Sukenik’s diary

Through the Wire

On 24 November 1947, by his diary, Eleazar Sukenik, professor of archaeology at the Hebrew University, stood at the gateway to Zone B in Jerusalem with a fence of barbed wire in front of him. On the far side an Armenian friend of his, a dealer in antiquities, held up a scrap of leather with writing on it. Even through the wire Sukenik could see that the letters were like those cut on the ossuaries of first-century Jerusalem, the stone chests in which the bones of the dead were gathered. The fence was British. Britain had governed Palestine under the League of Nations’ mandate of 1922, and that month it announced that it would leave by August, a date it soon brought forward to 15 May. Its officials had already divided the city into security zones, Jewish and Arab, with wire between them and a pass needed to cross. The Armenian was acting for an Arab dealer in Bethlehem, Faidi Salahi, who had bought leather scrolls from one of the Bedouin finders, and the Armenian asked whether Sukenik wanted to buy them. Sukenik suspected a forgery and said yes. By the diary he meant to meet the Armenian again on the 28th and go with him to Bethlehem.

He had been in America while the scrolls went the rounds of Jerusalem that summer, and he had heard nothing of them before the scrap at the wire. He weighed the risk, since Bethlehem lay in Arab territory, and went. In the last days of November he made the journey by bus and came home with two scrolls on approval, wrapped like an ordinary parcel, which the dealer let him study before any price was named. When he unrolled the first of them, he wrote in his diary, it came over him suddenly that destiny had chosen him to look at a Hebrew scroll that nobody had read for more than two thousand years. He set about finding the money to buy them for the Hebrew University, and the university’s president found him some.

On 29 November the General Assembly of the United Nations voted to divide Palestine into a Jewish state and an Arab one, and by the morning after the vote Sukenik had decided to buy. Jews celebrated in the streets that night. Arabs rejected the vote, and the violence between Arabs and Jews, which had been rising for months, grew worse within days. Sukenik’s son, Yigael Yadin, was then chief of operations of the Haganah, the Jewish underground defense force. The father’s diary is known through the book the son wrote about the scrolls ten years later. Early in December a man from the university library told Sukenik that St. Mark’s had scrolls too, which he had been shown at the monastery that autumn, but by then Sukenik could no longer get into the Old City to see them. In December he bought from the dealer in Bethlehem a crushed second roll of Isaiah.

Nothing more came of it until the second half of January, when a letter reached him from Anton Kiraz, a Syriac Orthodox owner of a taxi firm and an associate of Samuel’s, on whose land south of the city Sukenik had once excavated a tomb. In a divided city Kiraz was a useful man, with cars, drivers and friends among the British officials. He said later that he had lent the metropolitan part of the money for the scrolls and was his partner in selling them, and Samuel told it differently. By now there was no going back and forth between the Arab and Jewish quarters, so Kiraz found neutral ground. He arranged an office at the YMCA through its librarian, another member of his church, and the YMCA stood in Zone B, where a pass would let both men in.

There Sukenik was shown the scrolls of St. Mark’s, and he saw that they belonged with his own. Kiraz offered to sell them to the Hebrew University. He let Sukenik borrow the great scroll of Isaiah to study it and to raise the money, and while he had it Sukenik copied several of its columns. It seems that he photographed it as well, without the knowledge of Kiraz or the metropolitan. The crushed roll had been his since December. Now he had the great scroll too.

He meant to buy all four of Samuel’s scrolls, and a meeting was agreed at which the university’s president and the metropolitan would settle the terms. The meeting never took place. The talks stalled while Sukenik raised the sum, which came slowly, and when he had it and asked again, Kiraz proposed the Yugoslav consulate, where a member of the Syriac Orthodox community was employed. By then the metropolitan had turned to the Americans, and the war had closed the road. Kiraz assured Sukenik that the Hebrew University would have first refusal whenever the scrolls were offered for sale. Samuel later denied that any such talks had taken place, and said that Kiraz had acted without his knowledge. Sukenik was the first to print, all the same. Late in the summer of 1948, in the first volume of a book he called Hidden Scrolls, printed in Jerusalem, he set chapters 42 and 43 in Hebrew type as he had copied them, the first of the great scroll’s text to be printed anywhere. The scroll had left his hands months before. On 6 February 1948, at a second meeting in the office at the YMCA, he gave it back.

19 February 1948 · By Trever’s first article

Two Lines

Thirteen days later, on the afternoon of 19 February 1948, Father Butros Sowmy, a monk of St. Mark’s, and his brother brought scrolls wrapped in newspaper to the American School of Oriental Research in Jerusalem. Millar Burrows of Yale, the School’s director that year, had left for Iraq four days before, and the man in charge was John Trever, a young scholar with a new doctorate. He took the visitors to his own room, the only warm one in the building, since there had been no oil for the stoves that winter. The brothers opened one of the smaller rolls a little way, and it was too brittle to open further. Then the monk took out the largest. Trever found it supple, laid it on his bed and unrolled about a dozen columns, looking for any sign that it was forged. The script was plainly Hebrew. To his eye it looked strange.

In a drawer of his desk was a box of slides for a lecture called “What Lies Back of Our English Bible?” and he thumbed through it to the early Hebrew manuscripts. One slide showed a codex of the Law in the British Museum, then put in the ninth century, and a glance told him that the roll on his bed was far older. The next slide was the Nash Papyrus, a scrap with the Ten Commandments and the Shema on it, published in 1903, and its letters were like the scroll’s. The papyrus was undated too. Scholars had put it anywhere from the second century BC to the third century AD, because a hand is dated by setting it beside hands whose dates are known, and what comes out is a span of years, never a year. His cameras were away at the antiquities museum, so he could photograph nothing that afternoon.

Sowmy let him copy two lines instead. When the brothers had gone, Trever took a phrase that came twice in the first of them and followed it to Isaiah 65:1. “What I had copied was there word for word!” he wrote. The next morning he got into the Old City on a pass stamped at the Jaffa Gate, met the metropolitan, copied a few lines from what looked like the head of the scroll, and found after lunch that they were Isaiah 1:1. The power failed that afternoon, and in the evening he and Brownlee sat at a small table with two kerosene lamps beside them and an article of 1937 on the Nash Papyrus. In it William Albright of Johns Hopkins University had placed the papyrus between about 165 and 37 BC, most probably in the second century. By midnight the two young men were sure that the scroll was as old as the papyrus or older.

On 21 February the metropolitan and the monk brought the scrolls to the School to be photographed, and as Trever mended the great scroll for the camera he found crumbs of its lower edge in the satchel it had traveled in. Albright’s answer to the first prints came by air letter on 15 March. He congratulated Trever on the greatest discovery of manuscripts in modern times, judged the script older than the Nash Papyrus, preferred a date about 100 BC, and had no doubt that it was genuine. By its handwriting, the great scroll is more than a thousand years older than the oldest dated Hebrew manuscript of Isaiah.

A crumb of leather three millimeters by eight, carrying the first letter of the word for fire in Isaiah 65:5, came loose in 1948 and was gone by April 1949. The letter survives on Trever’s film. In the 1970s he found the crumb wandering from picture to picture, folded over a crack on his first negative and loose on other columns later, and by then the color in his transparencies was fading too.

By the middle of March Burrows wanted to see the cave, since pieces of the jars would help to date the scrolls, and a man from the shrine of Nebi Musa offered to arrange it. The trip was called off. The way was too dangerous, Burrows was told, because Jewish troops were training on the plain north and west of the Dead Sea. Who really stopped it, he wrote, would probably never be known. He had already told the metropolitan that the scrolls were not safe in the Old City, and Trever had suggested another monastery of the church down by the Jordan. On or about 25 March Sowmy left the country with them, for a bank vault in Beirut, without any license. The Mandate’s law was plain: “No antiquity may leave the country without an export licence from the said Department.” Burrows wrote later that taking them out without a license was illegal, and that the Americans had tried to tell the metropolitan so. The Israel Antiquities Authority calls it smuggling, and the Jordanians came to regard Samuel as a smuggler, while his own reasons, as he later gave them, were the war and the needs of his people.

Burrows’s diary for those weeks is mostly a record of shooting and explosions. A bomb in the city killed more than fifty people while he was away, and on 11 March explosives damaged the building of the Jewish Agency. Station wagons were being taken at gunpoint in broad daylight, “though most politely,” and the Americans kept theirs in. On Easter Sunday an attempt at a truce for the day broke down completely. The British left in the middle of May, the State of Israel was proclaimed, the armies of neighboring Arab states came in, and the war went on into the next year. Israelis call it their War of Independence. For Palestinians it is the Nakba, the catastrophe, in which the establishment of the state displaced about 750,000 of them from their homes. Kiraz’s house and his taxi firm ended up on the Israeli side of the line, and by early 1949 he, his parents and his sisters were living in a single room near the Jaffa Gate. Sarah Irving, the historian of this small community, argues that the first journeys of the scrolls were bound up with the catastrophe that overtook most ordinary Palestinians in 1948.

The first public word of the find came from Yale, in a statement of 10 April 1948 which said that the scrolls had “been preserved for many centuries in the library” of a monastery. Burrows had written that the monastery acquired them, and he did not know who changed his sentence on its way to the press. The library story was older than the release. When Sowmy first telephoned the School, by Trever’s first article, he said that he had come upon the scrolls while cataloging the monastery’s books. Irving points out that a find in a monastery library would have made the monastery’s ownership lawful, while a find in a cave that nobody had reported could be seized under the law of the Mandate. Samuel’s later book gives the idea of the library to Butros Sowmy, who could not answer for it. He had been killed that May, in what Irving calls Israeli shelling of the monastery.

January to March 1949 · By the spade

A Cigarette Roller

On 28 January 1949 Akkash el-Zebn, an officer of the Arab Legion, Jordan’s army, sent out to find the cave, saw fresh earth and broken pottery thrown out in front of a hole in the cliffs above the Dead Sea. By then the cliffs, like Bethlehem, were under Jordanian control, which is why the cave was dug from Amman. Harding had heard of the cave only in November, when the American journal that announced the find reached Jerusalem. Early in February he went out to the hole, found fragments of manuscript and sherds of pottery, and came back to dig.

He and de Vaux dug from 15 February to 5 March, with a few workmen and a guard from the Legion. Two men at a time could work. They dug by hand, mostly with knives. The cave is a little over two meters across at its widest, and the old way in, high in the rock, is so narrow that a jar would pass through it only on its side. They found linen, hundreds of fragments, part of one scroll still rolled, and broken jars, cylinders about sixty centimeters high, with the bowls that seem to have served as their lids. Some ladies who saw the linen said it was modern, but bits of manuscript were stuck to it. By the finders’ report some scrolls lay in the jars. No archaeologist ever saw one there. Joan Taylor, Dennis Mizzi and Marcello Fidanzio, in their study of the dig’s records, count the sherds of more than fifty jars in a cave that size.

What looked like pitch was the skin of the scrolls, decayed; the British Museum showed that in 1949. The pitch had been in print since April 1948, in an American journal’s note passing on news from Sukenik, long before anyone trained had seen the cave. The Bedouin had carried off two whole jars to hold water, and those are lost, so the pair in Jerusalem today are others from the cave. Set beside the tellings, the dig bore out a cave, jars, linen, a fall of rock from the roof, and fragments that belonged to scrolls already sold. It bore out nothing about who went in, or when, and nothing of a goat or a stone. Taylor, Mizzi and Fidanzio take the fragments as enough to place in this cave all seven scrolls sold from Bethlehem, the four of St. Mark’s and Sukenik’s three, and Brent Nongbri, a historian of ancient manuscripts, asks how much a few small pieces from a disturbed cave can show.

Mixed with the ancient debris were cigarette stubs, scraps of modern cloth, and newspaper that had been used to wrap food. Men had dug in the cave in secret the November before, going down several inches in the hope that the Bedouin had missed a scroll, and by a later account Kando and Isha’ya were among them. They found no scroll. The bits of manuscript that the Bedouin had let fall lay churned into the earth the secret diggers had turned, and the excavators of 1949 gathered them out of it a fragment at a time. Among the leavings of the secret dig lay a cigarette roller with its owner’s name on it, and Harding gave it back.

1950 to 1967 · By the museum’s ledger

A Pound a Square Centimeter

From 1950 the Palestine Archaeological Museum, now called the Rockefeller Museum, kept its purchases in a handwritten ledger headed Purchase of Antiquities. It was ruled into four columns, for who sold, what was paid, where the seller said the piece had been found, and what reward was given for information. Early that year Harding and de Vaux had decided that the museum would pay for scrolls and for news of where they lay, though under Jordan’s law every antiquity in the country already belonged to the state. Digging for one without leave was a crime. Kando gathered what the Bedouin brought him and offered it to the museum, and de Vaux paid him. The rate paid in the early years was a pound a square centimeter, a pound sterling being then worth one Jordanian dinar, and at that rate nobody gained anything by tearing a piece smaller.

The Ta’amireh went on finding. They found the caves of Wadi Murabba’at, a valley a few miles to the south, with documents of the second century AD, and of the eleven caves near Qumran that held scrolls they found five. In September 1952 they found Cave 4 almost under the feet of the archaeologists who had dug the ruin nearby. Jozef Milik, one of the scrolls’ editors, told how it happened. Some of the Bedouin sat one evening in a tent, talking over the prices their fragments had fetched, when an old man who had been dozing woke and said that in his youth he had followed a wounded partridge into a hole in the rock and found an old lamp and some broken pottery. The young men took flour, ropes and lights and went to find the place. When Harding arrived with two soldiers he found them still digging, in relays.

Cave 4 held thousands of fragments from about 500 scrolls. In all, by the count of a later chief editor of the scrolls, the caves near Qumran gave up some 930 manuscripts, about 200 of them books of the Bible and the rest rules, hymns, commentaries and other writings of their time. Some of those writings are the rule books of one strict Jewish community, and Sukenik was among the first to take it for the Essenes, an order that the Jewish historian Josephus describes. The ruin of Qumran stands near the northwest shore of the Dead Sea, and Harding and de Vaux dug it from late 1951 to 1956. No scroll was found in it. Eighteen of the twenty-two copies of Isaiah known from the caves came out of Cave 4, most of them in scraps. A later part of this history lays those scraps beside the great scroll, and they change what the great scroll means. By the autumn of 1952 the museum had nothing like the money for Cave 4. Jordan found fifteen thousand dinars, from funds that Burrows says were badly needed for health and farming. De Vaux asked foreign institutions to buy the rest, promising that each could take its lots home once the fragments had been studied, in two or three years. McGill, the Canadian university, bought nearly five hundred fragments sight unseen with a Montreal family’s gift, and the local press called Cave 4 McGill’s Cave, while Manchester, Heidelberg, the Vatican Library and a seminary in Chicago bought lots as well.

The promise was a government’s, and a later government withdrew it. In 1961 Jordan wrote to the buyers that the fragments would stay in the country, as “an indivisible part of the history of Jordan” and of the spiritual legacy of all mankind, and that their money would be repaid. McGill had its money back two years later. The historians of McGill’s purchase find the tourist trade in the decision too, since by then the scrolls were drawing visitors and offers of fees for loans abroad. Whatever the motive, the refunds kept the collection whole. The fragments McGill paid for are in Jerusalem still. The study took decades, and the photographs of the fragments were opened to all only in 1991.

Jordan had nationalized the museum the year before the war of June 1967, which Israelis call the Six-Day War and Arabs the Naksa, the setback. On the second day of the fighting three Israeli scholars crossed the lines with an escort and got into the Rockefeller by a back entrance, while shots were still being exchanged with Jordanian soldiers on the city walls. Most of the fragments, mounted between plates of glass, stood packed in wooden crates, apparently for Amman. Two days later Israeli officers questioned Kando and his son, hard, and that evening they took him home and lifted from under the tiles of his bedroom floor the long scroll now called the Temple Scroll, which he had kept back, wrapped in cellophane, in a shoebox. He was paid for it afterward, more than $100,000 by the reports, which differ on the sum.

The Rockefeller’s scrolls have been in Israel’s keeping since that week. When some were shown in Toronto, Jordan asked Canada to hold them under the Hague Convention on cultural property in war, and Canada declined to act as a third party. Germany would not promise a museum in Frankfurt that they would be safe from seizure, and Israel withdrew the loan. Israel says that it holds them as custodian, with the right to show and conserve them, counts them part of Jewish heritage, and argues that a loan abroad is not an export. Jordan says that Israel took them unlawfully from its museum in East Jerusalem. The Palestinian Authority claims them by the land they were found in, the West Bank, and Palestinian scholars, granting that the scrolls are Jewish, count them part of Palestinian heritage as the Roman ruins of that land are. No court has decided who owns the Rockefeller’s scrolls, and in practice they travel only where the host promises not to seize them.

1 June 1954 · By the Tax Court

Miscellaneous for Sale

“The Four Dead Sea Scrolls” stood in a ruled box under Miscellaneous for Sale in the Wall Street Journal of 1 June 1954. The title stood in its own quotation marks. Below it ran the sentence: “Biblical Manuscripts dating back to at least 200 BC, are for sale.” The box suggested them as a gift for an educational or religious institution, from a person or a group, and gave a reply address at the newspaper, Box F 206. Its neighbors in the column offered a stove works in Kalamazoo and steel tanks, and nothing in the box said who was selling. The advertisement, by his family’s account, was the idea of a man of Worcester, Massachusetts, who since 1951 had been a trustee, with Samuel himself, of the Archbishop Samuel Trust, into which Samuel had put the four scrolls.

Samuel had brought the four scrolls into the United States at the end of January 1949, by the date the tax case gave, and five years of showing had found no buyer. The Library of Congress put them on view in the autumn of 1949, and they went on to Duke, the University of Chicago and the art museum in Worcester. In 1948 Trever had won leave to photograph the scrolls by telling the men of St. Mark’s that publication would raise their value, as it had raised the value of the Codex Sinaiticus. One explanation given for the empty years is that the American School’s editions of 1950 and 1951 had already put the text into print. The other is that Jordan claimed the scrolls, so that no buyer could be sure of his title to them. Those who have written on the sale give the two different weights. No document settles which counted more.

Samuel’s own account, in his book of 1966, is of a churchman who believed that Providence had sent the scrolls to St. Mark’s so that their price could repair a monastery damaged in the war and help the destitute families of Jerusalem and Bethlehem. By the same book, the first money spent after the sale went to churches in Mosul and in Tur Abdin, in Turkey. In 1949 Kiraz and his mother had written against his appeals in America, calling his words “fake and hypocrisy” and saying that all the help they had seen was some secondhand clothes from England. His parish in New Jersey remembers him otherwise. One of its members, whose uncle he met off the boat from Turkey at a Manhattan pier and whose parents he married, calls him the protector of the scrolls, and the church there grew in part on the money. He died in 1995 and was buried at a monastery of his church in the Netherlands. His will left his estate for education and the training of clergy.

Yadin, an archaeologist by then and lately the chief of staff of Israel’s army, was on a speaking tour of the United States when a correspondent of a London Jewish weekly, told of the box by a colleague, telephoned him. The rest is told by Harry Orlinsky, a Jewish scholar of the Bible, in an account he published twenty years afterward. About noon on Thursday 1 July his car was packed for two weeks in Toronto, and his wife was already sitting in it, when the telephone rang as he locked the door. He let it ring three or four times. Then he went back in. It was Yadin, calling from the office of Israel’s consul general in New York, and he needed an expert to look at the scrolls under another name. Orlinsky became Mr. Green. By his account he went to a bank’s branch at the Waldorf-Astoria, met the banker who was acting for the buyer, looked at the scrolls, and telephoned the agreed word, a Hebrew toast to life.

Samuel’s trust sold them on 1 July 1954 for $250,000. The money came, in all or in most, from a New York donor, Samuel Gottesman, and the metropolitan, by his side’s account, learned only afterward whom the buyer had acted for. The four, the great Isaiah, a community’s rule book, a commentary on Habakkuk and a retelling of Genesis, went to Jerusalem to join Sukenik’s three, the crushed Isaiah, a book of hymns and a scroll on a last war. Then came the American tax case. The tax authorities treated the gain as Samuel’s own income, because the trust let him draw ten thousand dollars a year, which made it his and not a charity, and the courts agreed with them. The deficiency for 1954 came to $60,813.33, about a quarter of the price, before any interest or additions. In 1954 Isha’ya had asked for between twenty-five and fifty thousand dollars for his help and Kando’s in getting the scrolls to Samuel. Three years later the trustees paid Kando $4,000, “as a gift and not as compensation.” Isha’ya, whom the court calls Shamoun, received the same, and Kando’s money went through his hands. Each signed away his claims for a dollar.

The purchase was announced in February 1955, and that September TIME told its readers that so far “no important variations from accepted Bible texts” had come to light. In 1957 the seven scrolls of the first cave were shown together for the first time, in an air-conditioned vault in the basement of a Hebrew University building, under guard day and night. Their own building, the Shrine of the Book, opened in Jerusalem in 1965, most of it underground beneath a white dome. Among its architects was Gottesman’s son-in-law. Trever, who visited it the next year, thought the dome had been shaped like the lid of one of the jars from the cave. The great scroll ran around a drum beneath the dome, a cylinder more than twenty-five feet round, and it was failing on display. Trever found the leather darker and the writing far harder to read than he remembered, and he blamed damp where the authorities in Israel blamed light and air. The Shrine’s drum was built so that the scroll could be lowered into a rock-cut cave in an emergency. By 1968 the scroll had come off it, and a facsimile has stood there since.

Later second century BC · By its handwriting

The Hand and the Seam

The great scroll, which all those men had carried, hidden, photographed and sold, is seventeen sheets of skin sewn into a roll a little over seven meters long and about a quarter of a meter high, holding fifty-four columns, copied in the later second century BC by its handwriting. Most of the sheets carry three or four columns, of about thirty lines each. The back of the scroll is dark, and Trever put that down to the hands that held it. He set the darkening beside the many ancient repairs as marks of long use, older than the cave. The leather also changes along its length. It is very dark at the beginning, which by his account formed the outside layers of the roll through the centuries in the cave, and it lightens toward the middle until at the far end it is cream or light tan.

As it stands it holds Isaiah from the first verse to the last, every verse in the order printed today but one: it lacks 2:10. It has no chapter numbers and no verse numbers, which came much later, and its own paragraphs fall only sometimes where the chapters do, so the order has to be tested by laying the modern numbers against the leather. Laid out that way, from the first column to the last, its 1,290 verses run forward one after another, and none is out of place.

The skin is sheepskin by the standard description, though the kind of animal is still being studied, and the sheets are sewn with linen thread. The writing was probably done with a reed, though no pen has survived, in a black carbon ink, to judge from the fragments of the cave that the British Museum tested in 1949. Tears were mended in antiquity, some by careful sewing and some by strips of skin stuck to the back. A date here can mean three things, and they have three separate answers. The hand dates the copying, and set beside other hands it gives the later second century BC. Carbon dates the skin, the death of the animal and not the day anyone wrote on it, and at column 39, where laboratories at Zurich and Tucson each took a sample, the likelier of two ranges runs from about 250 to about 100 BC. The going into the cave has no date at all, though Trever, judging from the wear, thought it more than likely that the scroll went in long before about AD 70.

Someone marked it. The skin carries more than a hundred signs, by the list made for the exhibition of 2026, most of them in the margins and far more than on any other scroll from Qumran. Twelve times an X stands in the margin, once beside “Behold, my servant” and once beside the Sabbath verse in chapter 58. What the X meant, and whether copyist or reader drew it, nobody knows. The commonest sign is a short stroke beside a break in the text, nearly all sloping down to the left, a slope found on no other scroll. Chapter 40 begins on the last line of a column, as a new paragraph, with a mark in the margin that also stands at the start of chapter 36. There is no blank line.

The first hand can be watched at work, and the scenes are small. At 4:5–6 the word for “by day” comes twice, and the eye went from the first to the second, so that the cloud by day runs straight on into the heat and the sense breaks. Fourteen words of the medieval text are gone from that place, the fire by night and the pavilion among them. At 1:15 the hands full of blood are followed by “your fingers with iniquity,” words from 59:3, and at 51:3 the sorrow and sighing of 35:10 flee a second time. Where the medieval text has a rare verb at 13:10 for the shining of the stars, the scroll has the common one. Scholars have called this copyist careless in part, free, a simplifier for later readers, an emender of hard passages, and a man who left gaps sooner than invent, and more than one of the labels may be true. All of these are labels for the first hand, a name that still covers one scribe or two.

After chapter 33 the writing stops three lines short of the foot of the column, a new sheet is sewn on, and chapter 34 opens at the top. Nothing is missing there. The text runs on unbroken from 33:24 to 34:1. From column 28 the spelling grows fuller, and words that the medieval text lacks become about four times as many. Eugene Ulrich and Peter Flint, who edited the scroll, hold that one scribe copied it all, and a colleague in their own edition finds a break in the language of the two halves. Mladen Popovic and his colleagues, reading the strokes by computer in 2021, argue for two, trained so alike that the eye cannot part them. Marcello Fidanzio, of the study of the dig’s records, argued in 2026 that the two parts were made separately and joined, pointing to folds that run both ways on the first eight sheets and one way on the last nine, and to repairs that crowd the first part and hardly touch the second. He does not say how many scribes there were, and two scribes could have worked on two pieces. Carbon tests on the two parts, reported in 2026, showed no significant difference, which leaves room for a gap of some years. The seam is on the leather, and its cause is argued.

At four places in the second part the first hand appears to have left room, and others filled it. At one of them two lines of very small script are crowded into the space, and Burrows had already noticed in 1955 that the copyist sometimes left a gap for something missing or unclear in the copy before him. Drew Longacre, a scholar of the scrolls’ text, argued in 2013 that the first hand’s source was damaged. Rearrange the columns as a lost source would have held them, and all four places fall at about the same height, as if at the ruined foot of each column. Ulrich and Flint read the same places as words that the first hand’s source, older and shorter, had not yet taken on. Neither reading has displaced the other. The first of the four falls just after a line the scroll does have: “Search in the book of Yahweh, and read.”

All of this is one scroll. There was another. Sukenik’s crushed roll of Isaiah turned out, when it was opened, to be one large piece and several small ones, apparently broken already when it went into the cave, with the writing in many places legible only by infra-red photography. It had been so tightly compressed that the attempt to unroll it was put off until the other scrolls were open and the people doing the work had learned how, and so it was left until last.

2026 · By this page’s count

Word by Word

“What I had copied was there word for word!” Trever’s article does not reproduce his two lines. The verse they led him to, the first of Isaiah 65, can now be set out and tested. It has fourteen words in the scroll and fourteen in Leningrad, the same fourteen in the same order. He was right about that. Six of the fourteen differ in their letters. Five of the six are spelling, one of them on a damaged letter, and the sixth has one ending more: where Leningrad has those who did not ask, the scroll has those who did not ask me. The King James already has “asked not for me,” and the American Standard Version sets those two words apart as supplied by its translators, while the scroll has the “me” on the leather. Who said it first, and what did he mean by a word? He had just found two lines from a leather roll in a printed Bible, and for that purpose a word was a word however it was spelled.

This page has put the same test to the whole scroll. Each word was set beside its word in Leningrad, and the two counted as one only if every letter matched, so that a single extra letter fails a whole word and “and” weighs as much as the name of God. That is a harsh count, and it is the one printed here. Only letters on the leather were counted. Where an editor has restored a letter that the skin has lost the word was left out, so that no guess could count for the match or against it, and the scroll was taken as it now stands, with the work of its later hands. The letters come from one modern transcription, typed from the official editions. Nobody working on this page has seen the leather or read the official edition of the two Isaiah scrolls, and nobody outside has tested the count. Leningrad is the yardstick because it is what readers have, and the note on sources at the foot of the page sets out how the work was done.

Set beside the Leningrad Codex, counting only letters on the leather, the Great Isaiah Scroll matches letter for letter in 69 words of every hundred by this page’s count. Of about 16,700 words compared in the great scroll, about 5,100 differ. Two thirds of those are spelling. Take a row of a hundred words, then, and walk down it. About 69 are the same. That is lower than the 95 percent a reader may have heard, claimed of the two Isaiah scrolls together, a figure this part comes back to. The next 20 or so are the same word spelled otherwise, because a few Hebrew letters could stand for a vowel, a scribe might write one or leave it out, and the word stays the word. The words for “not,” “for” and “all” make about a fifth of all the spelling differences between the scroll and Leningrad. Edward Young of Westminster Seminary, writing in 1958, compared differences of this kind to “honor” and “honour.”

That leaves about 11 words in the row. About 4 of them differ in a way an English translation need not show, as the “me” of 65:1 need not, and about 4 in one small word or an ending, a “his” for a “my” or an “and” that English would print. One or two differ in a word with content, as at 45:7, where a Bible has “I make peace” and the scroll has “I make good,” and the last is a slip of the pen or a piece of a clause. That sorting comes from a second reading, since every difference beyond spelling was read twice, by AI readers working apart, to one rule. In about 230 places a content word differs. In 28 or more others what the verse says differs, and in about 8 more what looks like a copying slip loses or repeats a clause. In about 1,060 of Isaiah’s 1,291 Hebrew verses the scroll differs in no content word, plain slips of the pen aside.

Some of those places are verses a reader would look for. At 63:3 the scroll lacks the eight Hebrew words that run from “I trod them in my anger” to the lifeblood sprinkled on the garments, and the crushed roll has them. At 2:9 it lacks “therefore don’t forgive them,” and after that the whole of 2:10, and two copies from Cave 4 have the place in broken pieces. Anthony Ferguson, a specialist in these scrolls, has sorted the same differences his own way, into those that alter the sense and have a plain scribal explanation and those that have none. His lists, made apart from this count, come out about the size of its own. That is a comparison and not a validation.

The differences do not fall evenly along the roll. From column 28 on, the text stands further from Leningrad, in spelling above all: by the count of letters on the leather, 77 percent of the words before that point match it letter for letter, and 63 percent after. The step falls at the three blank lines. The count takes in no letter shapes and no folds, and it chooses no camp among those who argue over the seam.

The scraps of Cave 4 go part of the way to explaining that distance, since laid beside Leningrad in the same way, the eight Isaiah scrolls with more than a hundred words left to compare fall into two groups. Six match at about 90 words of a hundred. Two come near 70, the great scroll and one copy from Cave 4, and what those two have in common is a way of spelling, largely a house style, with no sign of decay. Where the scroll differs from Leningrad in more than spelling and another Isaiah scroll has that word, the other scroll sides with Leningrad four times in five, which is partly what a copy’s own slips would produce. Leningrad does not descend from this scroll. The copies that led to Leningrad are what this history will call the line, and the great scroll is a cousin of that line and no ancestor of it. The distance between two relations comes from what happened on both sides, so it does not measure copying down one line.

One of the six is the ruin that was left until last. The second Isaiah scroll is the crushed roll Sukenik bought, sold as from the same cave as the great scroll and copied about the turn of the era by its hand, and for a few days in the winter of 1948 he held both. About a fifth of the book’s words survive in it, whole or in part, and in what survives the lower third of each column is gone, on the estimate of the Shrine’s curator. It is no clean copy. It has real differences of its own, and at 60:19–20 the eye went from one “everlasting light” to the next. Counted the same way, it matches Leningrad letter for letter in 92 words of every hundred. Most of its words lie in the later chapters, where the great scroll stands further off, so the 69 and the 92 do not rest on the same stretch of the book.

On the words the two scrolls share, about 2,700 of them, they match each other in 64 of every hundred. Two scrolls sold as from one cave stand further apart, mostly in spelling, than the second stands from a codex written about a thousand years later. Spelling aside, they differ from each other in about ten words of a hundred, the great scroll differs from the codex in about ten as well, and the second differs from the codex in about three. That holds in both halves of the book and on words with no damaged letter. Kennicott expected the oldest copies to stand furthest from the printed text, and one of these two does, and the other does not.

The great scroll is the weaker witness for the letter and a strong one for the book. Nothing in the count touches its order. Its 1,290 verses still run forward in order, with one verse lacking, and no other copy from the caves has the book from end to end. For the letter a reader has to go to a ruin that has kept about a fifth of the words, and to the scraps of Cave 4 that stand beside it.

The claim that the text had not moved is far older than the find. Josephus wrote about AD 95, in an argument with the Greeks, that “no one has been so bold as either to add any thing to them, to take any thing from them, or to make any change in them,” which is a boast of his people’s loyalty and says nothing of what scribes did. In 1948 the first fear ran the other way, and John Allegro, a member of the scrolls’ team and a partial witness, remembered “news-hungry journalists” talking of changing the whole Bible.

Burrows’s first paper that October set spelling apart from wording and found the scroll’s chief importance in “supporting the fidelity” of the Masoretic tradition. In 1955 he wrote that it was a matter for wonder how little the text had altered in something like a thousand years. He was writing of one scroll of one book, and a paragraph earlier he had said that this did not prove it the original text. The next lines of his book report Paul Kahle, who in 1937 had made the Leningrad Codex the basic text of the standard printed Hebrew Bible. For Kahle the variants were the discovery. Burrows, holding his ground, gave them second place.

Then came the sentence most readers have met. In a later printing of Gleason Archer’s textbook, first published in 1964, the two Isaiah scrolls “proved to be word for word identical with our standard Hebrew Bible” in more than 95 percent of the text, the rest chiefly slips of the pen and spelling. No count for it has been found. A sentence of that kind can now be tested, and it has to be tested on both of the bases a reader might give it, since one reader takes a word to mean every letter and another, like Trever, lets spelling go. The two scrolls together match Leningrad letter for letter in about 73 words of a hundred by the count, and in about 91 with spelling set aside. Only the second scroll, in what survives of it and with spelling set aside, reaches the figure, at about 97.

1948 to 2021 · By a committee’s vote

Thirteen Readings

In the summer of 1948 the Old Testament section of the committee making the Revised Standard Version was in session at Northfield, Massachusetts, and a man staying at the same hotel remarked to them one day, “You will have to revise your translation of Isaiah all over again now, won’t you?” Burrows was just back from Palestine, and the list of the scroll’s variant readings that he had made on the way home lay before the committee while it went through Isaiah. He says nothing of the scroll in that room or of any photograph, and his printed edition was two years off. In 1952 the Revised Standard Version marked thirteen readings in Isaiah as “One ancient Ms,” meaning the great scroll, and a reader of that page could not have told that a cave was involved. Burrows, who had voted on them, wrote in 1955 that he now thought some had been mistakes.

Each reading was argued on its merits and settled by vote, rarely a unanimous one. Thirteen counts the footnotes, and more of the English moved than that. At 7:1 and 49:17 the new Bible agrees with the scroll and carries no note. One of its members, meanwhile, was publishing against the scroll paper by paper. Orlinsky, the only Jew on the committee, told the American Oriental Society at New Haven on 6 April 1949 that the copyist, or the copyist of the scroll’s source, “was not a particularly careful one,” and by his own footnote Burrows stood up afterward to say that he agreed with the paper as a whole. Six years later, at Dibon in 15:9, Burrows gave up one of the thirteen by name. He gave it up to Orlinsky.

The questions were older than the committee. In 1611 the King James translators had printed them down the outer edge of the page. Where their text has “the golden city” at 14:4, the margin offers “exactresse of gold.” At 19:18 “The city of destruction” has “of Heres: or of the Sunne” beside it, and at 21:8, where the watchman cries “A lion,” the margin tries “cryed as a lyon.” Their preface defends the habit: “doth not a margin do well to admonish the Reader to seek further, and not to conclude or dogmatize upon this or that peremptorily?” They had judgment, and no Hebrew older than the Middle Ages to try it on. The great scroll has the sun at 19:18, the second half of their note, and at 21:8 it has a word their margin never thought of.

Lowth printed the seer at 21:8 in 1778, reasoning from the Syriac, 170 years before it turned up on the leather. The same book kept “the exactress of gold” at 14:4 and “the offered cities” at 33:8, where the scroll has other words. At 53:11 it reads “Of the travail of his soul he shall see [the fruit,] and be satisfied,” the brackets his, and the Revised Standard Version of 1952 supplies the same words with no note. A reading that stands on old leather is an old reading, which is a smaller thing than the original one. Lowth’s reason at 21:8 was sense. In the World English Bible the watchman still cries like a lion, as he does in the Jewish Publication Society’s Tanakh.

A difference does not say who changed. All through the count Leningrad stood still and the scrolls were measured from it, because a measure needs one end fixed. The updated New Revised Standard Version and the New International Version follow the scroll at 14:4, 33:1, 45:2 and 49:24, though the United Bible Societies’ committee, which weighed these readings for translators, rates three of those only C, which means considerable doubt. Leningrad has at 14:4 a word found nowhere else, and the scroll differs from it by one letter, an r for a d. Leningrad’s word at 33:1 is unknown, and the scroll has a common word for finishing, close to what the King James had already printed. At 49:24 Leningrad has the captive of a righteous man, “lawful captives” is the translators’ rescue of that, and the scroll has “tyrant,” the word of the next verse. A yardstick can be the bent one. At 19:18 the Bible Societies’ committee goes further and reads Leningrad as the altered text, saying that its “destruction” seems to be a slighting name for Heliopolis in Egypt, the city of the sun, where a Jewish temple stood.

The scholars who had guessed at the Hebrew fared worse than the old translations. R. Laird Harris, a critic of the Revised Standard Version, writing in 1968, went through about thirty places in its Isaiah where the committee had mended the Hebrew by conjecture, and wrote that “not one of these thirty scholastic conjectures is supported by the Isaiah scroll.” Read against the transcription, one of the thirty is clearly borne out, Syene at 49:12, and two in part. So where modern scholars had only guessed, the scroll almost never agreed, and where an old translation had differed from the Hebrew, it sometimes did. A modern critical Bible follows the scroll in a few dozen places in Isaiah, out of nearly thirteen hundred verses, and most of those change a word or a letter. The scroll had no part in the change readers may have been told of at 7:14, since the word that the 1952 committee rendered as a young woman is in the great scroll as it is in the medieval text.

One chapter carries a figure of its own. The claim that Isaiah 53 has 166 words and only seventeen letters in question was in print by 1968, in A General Introduction to the Bible by Norman Geisler and William Nix. As it is quoted, ten of the letters are spelling, four are style, and three are the word for light. Leningrad’s chapter does have 166 words, counting linked words separately. By this page’s count the Great Isaiah Scroll matches 95 of them letter for letter, and it differs from the chapter in about 94 letters. The figure of seventeen comes close only for the second scroll, at about 18, and about a third of that chapter there is lost or damaged. Whichever scroll is meant, the word for light is not the only word in question.

At 53:11 Leningrad has a verb of seeing with no object written after it, and other copies have the word for light. Three scrolls have it: the great scroll, the second Isaiah scroll and a copy from Cave 4, where its last letter is restored. So does the Greek, in a sentence built differently. Aquila, Symmachus and Theodotion, the later translators into Greek, lack it. The next verse has two more places where the same three scrolls and the Greek stand together against Leningrad, with the plural “the sins of many” where Leningrad has the singular, and “their transgressions” for “the transgressors,” and the letters that make the difference are on the leather in all three. The Bible Societies’ committee rates all three readings B, which means probable with some doubt, and what it grades is the earliest text it could reach. All five modern Bibles checked keep Leningrad’s “transgressors” or a word like it at 53:12, and none has a note on it.

At 52:14, where the Bibles have “marred,” one copy, the great scroll, has one letter more. A scholar read it as “I have anointed,” and others read it as a spelling or another verb. The second scroll has Leningrad’s letters there, and no Bible checked prints the longer form. “Light” at 53:11 was not one of the thirteen. It entered the New English Bible, the New International Version and the New Revised Standard Version, went into the Christian Standard Bible in 2017, and came out of the NRSV’s 2021 update. The World English Bible prints it.

c. 50 BC to AD 135 · As it now stands

Nine Hundred and Eighty Letters

Two scrolls lay under the floor of the synagogue at Masada, the last fortress of the first war with Rome, each at the bottom of its own pit, and the pits had been filled in with earth and stones. Yadin, who dug the rock, found a Deuteronomy in one and an Ezekiel in the other. The copies from Masada were written about 50 BC to AD 30 by their hands and left there by AD 73 or 74, which makes the rock a fair place to ask how old Leningrad’s letters are. Another scroll from that rock, a Leviticus that did not come from the pits, has 278 words left to compare, holding 980 letters, and as the scroll now stands none differs. Three small places in it stand above the line, a word at 10:17 and single letters at 11:32 and 11:35, and each gives Leningrad’s reading.

Outside Qumran, in scrolls from Masada and the caves of the second revolt, the war of AD 132 to 135, 98 of every hundred words match Leningrad letter for letter by this page’s count of about 4,300 words, restored letters left out. Only seventeen of them are Isaiah. The scrolls were chosen by where they were found, those from the caves were copied by their hands as late as about AD 115 and left by about 135, and 58 in every hundred of the words are one scroll, a copy of the Twelve Prophets, Hosea to Malachi. These closest scrolls are also the youngest. Closeness and the years down to Leningrad move together, so the figures alone cannot say whether the text held still or the span was only shorter. To call them copies of a line is partly circular as well. One modern scheme of sorting defines such a copy by a limit of two percent of its words, so that a scroll that fell outside would not have been counted in the line at all.

Bedouin found the Twelve Prophets scroll about 1955, badly damaged, in a cave of Wadi Murabba’at, the valley the Ta’amireh had opened. Of its 2,456 words compared, six differ in anything beyond spelling or a tiny point of grammar, three of them on damaged letters. Read through for sense, the words from those sites hold about six places that differ in a content word. One lies at Psalm 22:16, the verse of the hands and feet, in a copy from Nahal Hever, and the Masoretic Text page takes up that verse. Emanuel Tov, who took over the scrolls’ publication and wrote the handbook on their scribes, had put the difference at “up to 2%” of the words, with no count shown. He also notes how far down the likeness goes in the Leviticus. Leningrad’s Torah often writes the feminine “she” with the letters of the masculine “he.” By his report every copy from Qumran spells that word the ordinary way, and the Leviticus alone has Leningrad’s form, at 11:6 and at 10:17, the first of the three small places.

For Isaiah itself the evidence is the second scroll and its Cave 4 kin, at about 92, then Murabba’at’s seventeen words, and then no Hebrew Isaiah that this page has found until the ninth or tenth century. So the book’s evidence has two ends. It begins with the great scroll, about 125 BC by its hand, the oldest copy, and after the seventeen words it breaks off until the medieval codices, of which the Prophets codex of 916 is the first with a date that holds.

How one form of the text came to win is the question under all of this, and the desert puts it sharply. By their hands some of Masada’s copies and some of Qumran’s were written in the same decades, in the same desert, and they do not look alike. One answer, Ian Young’s, is the date: the text had settled before Masada, so the scrolls at Qumran must be older as a deposit. Tov answers that the copying overlaps and the difference lies in who owned them. Masada’s people followed the leadership in Jerusalem, on his reading, and the other forms of the text lost their readers. Peter Gentry, a specialist in the text, holds that a text like Leningrad’s was dominant all along, near the Temple and after its fall. Ulrich objects that the question is asked from the winner’s side. All of them accept what came after. From about AD 100 every Hebrew biblical manuscript that survives in Jewish hands is of one type.

Burrows had moved by 1958. Writing after Murabba’at, he called the text fluid at Qumran and fixed by the second revolt, with the form behind Leningrad as the trunk and the others as branches, and he quoted the Bible scholar Moshe Greenberg, writing in 1956: “The prevalence of the standard, not its creation, came after 70 a.d.” He added that the Masoretes sometimes chose wrongly, and that a verse’s first wording may already have been lost from every copy they had. Not everyone has followed. Tov rejects the word stabilization, and Gentry the consensus that the text was standardized in the first century.

The line of copies that led to the Masoretic text changed very little in the centuries between the desert and Tiberias. It was not the only line. That steadiness belongs to a text, and the scroll usually cited for it is not where it shows. The span is about nine centuries from the Twelve Prophets scroll and about a thousand years from Masada, and for Isaiah it rests on the second scroll and the scraps of Cave 4 beside it. None of this speaks for the time before Isaiah’s oldest copy.

The Masoretes themselves kept a record of other readings, as the Masoretic Text page tells. Where they thought a word should be read otherwise, they left the letters standing and put the other word in the margin. At 49:5 the letters have Israel “not” gathered and the margin has “to him,” so that the King James prints the one and the World English Bible the other. Where the Masoretes marked two readings in Isaiah, the great scroll goes with one about as often as the other, and at three places, 49:5, 54:16 and 37:30, another scroll has the other reading, so both were in circulation long before anybody marked them.

Dates argued · By later hands

Four Dots

Three letters stand above the line in the first verse of the great scroll, and one of them is the ayin of Isaiah’s own name, set small and high over a word that had gone on without it. The modern transcription records where such a letter stands and not whose it is, so the scenes that follow can name no writer. A hand has canceled “far” at 21:1, the usual word for a land at a distance, and written above it the word that the King James renders “a terrible land,” which is Leningrad’s. “In Jerusalem” was written after “this altar” at 36:7 and then canceled, words that belong to the same speech as 2 Kings 18:22 gives it. Hezekiah’s name is written above the line at 37:14, in the verse where he spreads the letter before Yahweh. At 3:17 a hand canceled “Lord” for the Name, where Leningrad has “Lord.”

The leather carries hundreds of corrections, and more than one hand made them. The introduction to Trever’s color plates of 1972 calls the first hand’s own corrections “not infrequent,” and it speaks of correctors of various dates, writing a less formal script. Fidanzio said in 2026 that at least seven hands had added to the scroll. Some of the ink is blacker than the rest, the black that misled Wechsler in 1947. Tov identifies one of these people, reporting that the copyist of the Rule, the rule book of the people of the caves, found in Cave 1, put corrections into this scroll, though no single correction can be given to him. That copy of the Rule is usually put by its script about 100 to 75 BC, and one study sets the hand at 40:7 in that copyist’s time, which does not make it his. No date has been fixed for the other hands, and a letter above the line need not be a later letter.

Twice in the scroll the Name is a row of dots, four at 40:7 and four or five at 42:6, and both rows are above the line. The Rule writes the Name the same way where it quotes Isaiah, which is a habit shared and gives no reason, since the first hand writes the Name out freely all through the roll, and at 3:17 it is written above the line in full letters. The additions have been read against the scroll as well as for it. In September 1950 Kahle argued that one insertion, agreeing with the Masoretic text, could not have been written before the third century AD, so that the scrolls must have gone into the cave late. Burrows called this a “strange and quite unwarranted assumption,” since agreement with the standard text does not rule out an early date. He thought the difference between such words and the writing around them suggested a scroll already old when they were copied in.

The count can follow the pen word by word. By its reckoning, of the corrections that make or break agreement with Leningrad, seven in eight make it, leaving the large repairs aside. That shows less than it seems, since mending a slip usually restores what nearly every copy read. Whether from a second copy or the first hand’s own source, something close to that text lay behind them. The one fairer test, where another Isaiah scroll differs from Leningrad, has only seven corrected words to go on. In spelling the corrections lean away from Leningrad.

The great scroll is a side branch. No known copy descends from it. These corrections did not keep the line, and they hardly moved the scroll itself. Take away every letter written in over the text, put back what was canceled, leave the large repairs where they are, and the 69 in every hundred falls by about half a word. The scrolls from outside Qumran carry about a fifth as many corrected letters, about 3 in every thousand against about 14 here, with none at all in the Psalms from Nahal Hever or from Masada, and the second scroll has about 1. The pattern is not clean. Masada’s Ezekiel has about 11 and a Cave 4 Isaiah about 12. A clean scroll may also have had a good scribe or a lazy checker.

Later tradition remembered more than the leather shows. The Talmud has the proofreaders of scrolls in Jerusalem drawing their wages from the Temple funds, on a saying handed down about two centuries after the Temple fell, and a tractate later still tells of three scrolls found in the Temple court, where the reading of two was followed against one. Tov joins such memories to the desert copies and proposes that the finest of them were corrected against a central text in that court, which he himself calls a hypothesis. Many readers have heard that one slip condemned a sheet, and the Talmud says otherwise, allowing a Torah scroll with some three errors to the column to be mended. Rules written down centuries later forbid much of what this scroll does, from its linen thread to its canceled Name, and they do not describe its scribes, since no such rule is known to have bound them and the Isaiah scroll is no Torah scroll. Care has a limit of its own, as Driver complained in 1890: it keeps an error as faithfully as a true reading.

The witnesses come to one sentence. The famous scroll shows the whole book, the forgotten scroll shows the wording already there, and the margin shows people who cared about it. At some date unknown, someone with a pen thought one letter of a prophet’s name worth going back for. That is a scene and not an explanation, and the copies that stand nearest Leningrad, where an explanation might show, carry fewer marks to read it from. How the line was kept is not on this leather.

Not every gap found a pen. At 16:8–9 the vine of Sibmah comes twice within two verses, and the eye crossed from one to the other, so that twenty words of Moab’s vine are gone from the great scroll. The second scroll, crushed as it is, keeps “the lords of the nations” and “Therefore I will weep” from inside that gap on its own leather. No hand on the great scroll ever put them back.

1 Samuel 10:27 · Above the line

Four Cubits and a Span

“But he held his peace.” So 1 Samuel 10 ends in the World English Bible and the King James, with Saul newly chosen, despised by worthless men and saying nothing. In the Samuel scroll from Cave 4, copied about the middle of the first century BC by its hand, a run of writing stands above the line at that place, and it reads “and it was about a month,” with Nahash the Ammonite going up after it. The Greek has the month, and so does Josephus, who writes “After one month.” Leningrad has the silence. The two phrases look alike in Hebrew, down to a resh in one and the very similar dalet in the other.

The words above the line close a paragraph that survives on the leather with gaps and in no Masoretic manuscript. Nahash king of the Ammonites pressed the people of Gad and Reuben hard, it says, and put out every man’s right eye, and left Israel no deliverer, until no man beyond the Jordan had kept that eye, but seven thousand men reached Jabesh Gilead. Josephus knew the story, and the reason he gives for the cruelty, that a shield covered the left eye and left a man useless in war, is his alone. Frank Moore Cross, one of the scroll’s editors, thought the paragraph lost from the Hebrew, and Alexander Rofé, who answered him, thought it added, as the Masoretic Text page tells. The New Revised Standard Version has printed it since 1989, the New International Version gives it in a footnote, and the English Standard Version says nothing. On the page it is the latest arrival, and on the leather it is nearly a thousand years older than any Hebrew copy of the place without it.

Seven chapters on, a numeral differs. Little of 1 Samuel 17:4 survives on the scroll’s leather: the going out, the word for a man, the numeral “four” with its first letter damaged, then “cubits” and “and a span.” The name of Goliath in that verse is an editor’s restoration. Leningrad has six, and the World English Bible with it, while Josephus has “four cubits and a span,” as the two great Greek codices do. At a cubit of eighteen inches the four comes to about six feet nine and the six to about nine feet nine, the height that Weights and measures gives for Goliath. In one journal a scholar argued in 2005 that four was the first reading and six a scribal rise, and another answered in 2007 that six came first and four was a conversion into the cubits of Egypt.

Samuel is where the wording moves. Set beside Leningrad in 1 Samuel, this scroll has another word or no word in 14.6 percent of the words compared, about 520 by this page’s count, against 5.3 in the great scroll and 1.5 in the second. Before about AD 100 the Hebrew Bible existed in more than one form. In most books the copies differ the way copies do, and Isaiah, with all its spelling, is one of those books. In a few the difference is of another size, and the plainest of them is a prophet.

The Greek of Jeremiah is about a seventh shorter than the Hebrew that Bibles translate, and the Septuagint page tells that argument. Hebrew fragments from Cave 4 share the shorter form in places. At 43:6 one of them runs on without a gap from Nebuzaradan to Gedaliah son of Ahikam, with no “captain of the guard” and no “son of Shaphan,” as in the Greek. Another copy from the same cave has the long form, in the passage on the Branch at 33:14–26: “Yahweh” and the first letters of “our righteousness,” the rest restored, and a man sitting on the throne, with David’s name restored. That passage is the largest thing the Greek lacks, though its substance appears at 23:5–6 in both. Tov and others read the two as editions of one book. The two lay in one cave, with nothing there to say how their owners ranked them. In Jeremiah, Samuel, Exodus and Numbers the caves show different editions.

The Psalter of more than two hundred psalms has not come back. The Psalms scroll from Cave 11 runs in another order, with Psalm 118 after 109 and Psalm 150 well short of the end, and it holds psalms that no Hebrew Bible keeps, and a piece of Sirach among them. In the Psalms the caves show a different order, which Flint reads as another Psalter and Tov as a collection for worship. Masada’s Psalms, on the other side, reach 150 in the usual order and stop. The Cave 11 scroll also holds a verse a reader can look up, since Psalm 145 is an alphabet, a verse to each Hebrew letter, and Leningrad has no verse for the letter nun. The scroll has one, as does the Greek. The verse was in English long before the cave was opened, since the Douay Bible had it from the Latin in 1610, and the Revised Standard Version of 1952 printed it from one medieval Hebrew manuscript, the Greek and the Syriac. What the scroll gave it was a Hebrew witness far older than that manuscript, and whether the verse was lost or supplied is still open.

One scroll reads “sons of God” at Deuteronomy 32:8 where the Masoretic text reads “sons of Israel.” Some scholars think the Masoretic reading was altered on purpose, and others doubt the motive. Little of the verse beyond those two words survives on the leather, in a scroll of excerpts that does not run on, and a standard edition of the Greek has “angels of God.” Tov holds that the scroll keeps the first wording and that the change avoided a sense of many gods, and Ferguson answers that “sons of God” is untouched at Genesis 6 and in Job. The World English Bible, the King James and the American Standard all print “children of Israel.”

Kenyon had an answer ready for readers alarmed by such things, in the same book that ruled out older manuscripts. “No fundamental doctrine of the Christian faith rests on a disputed reading,” he wrote in its edition of 1896, of the whole Bible and chiefly of the New Testament, and the word “fundamental” carries the sentence.

Can a reader trust the Bible in their hands? On everything the scrolls can test, the answer is yes, and the test has been a hard one. By this page’s count the second Isaiah scroll, about a thousand years older than Leningrad, differs from it beyond spelling in about three words of a hundred of what survives. The great scroll, the furthest from it, still holds the whole book, and in about 1,060 of its 1,291 verses it differs in no content word, plain slips of the pen aside. Where the differences are large, as in Samuel and Jeremiah, they are known and argued in the open, and some stand in the notes of the Bible a reader already owns. No article of the creeds turns on a reading in the scrolls. All of this rests on what survives, since about one verse in three of the Hebrew Bible has even a letter on the skin. The scrolls can show that these words were handed down with care. Whether the words are true was never a question for the leather. Believers rest that on other grounds, and nothing in the caves has taken those grounds away.

No list of books turned up in the caves. The fixed lists are later and come from elsewhere, and what the caves hold is a record of what was copied, with no Esther and a good deal of Enoch and Jubilees. By Tov’s count of the copies from Qumran there were about 36 of the Psalms, about 30 of Deuteronomy and about 21 of Isaiah.

About 100 BC · By their own Rule

Prepare the Way

Where ten men were, the Rule said, one must always be studying the Law, day and night, in relays, one man after another. All of them together were to keep watch for a third of the nights of the year, to read in the Book, to study the ruling and to bless together. The words are in the Cave 1 copy, written about 100 BC by its hand, and the Hebrew leaves open whether a third of all the nights means a third of each night or every third night. The Book has no name there.

Isaiah was among the books they kept. Its copies at Qumran are more than three times the number of Jeremiah’s or Ezekiel’s, and five manuscripts from Cave 4 hold commentary on the book, a verse quoted and then explained. The rule for such reading is set out in the commentary on Habakkuk from Cave 1, and Isaiah’s commentators work under it. God told the prophet what was coming on the last generation, it says, and did not make known to him when the age would end. He made known all the mysteries of the prophets’ words to the Teacher of Righteousness, as the writings of this group call the man they followed. On those terms a prophet did not know what his own words meant, and the Teacher did. Read that way, the book described their own day. A commentary from Cave 4, on the feasting of chapter 5, says that the feasters are the men of scoffing in Jerusalem, the words of 28:14, and it has living opponents in view.

All four Gospels set Isaiah 40:3 over John the Baptist, and the Rule from Cave 1 sets the same verse over a community told to go to the wilderness. No text joins the two. That copy of the Rule tells its people to separate from the men of injustice and go to the wilderness to prepare the way there. Then it quotes the verse, starting at “in the wilderness” and leaving out the voice that the Gospels begin with. The Name in the quotation is four dots, and the next line says what the way is, the study of the Law. A copy of the Rule from Cave 4 seems to have had the sentence without the verse, and that copy is the younger, so whether the verse was added or cut is argued. Where the wilderness goes in the verse changes with the version a reader has open, and each placing makes sense.

The American public found the scrolls in May 1955, when The New Yorker gave most of an issue to the critic Edmund Wilson. He suggested, as Burrows reported it, that Qumran might have been “the real cradle of Christianity,” the issue sold out in a few days, and the book made from it became a best seller. Burrows spent the next years answering questions after lectures, and he decided that the interest was chiefly religious. People wanted to know, he wrote, “what these documents would mean for traditional beliefs,” and some were afraid for their faith. Taller claims followed for decades, of a Teacher crucified before Jesus, of a Gospel on a Greek scrap and of Christians hidden in the commentaries under other names. No scroll from Qumran names Jesus, John the Baptist, Paul or any Christian.

The question comes closest with John. Some scholars hold that he had once been among these people, and others that the likeness belongs to the times. The Gospels put him in the wilderness and at the Jordan, and his washing was done once, for repentance, on all comers. The washings of the Rule were repeated, for purity, and it told its members to hide their counsel from outsiders. Josephus, who knew the Essenes, calls John “a good man” and describes his baptism with no word of them. The ruin at Qumran, which de Vaux read as the home of the people of the scrolls, had pottery that matched the first cave’s, and later excavators read it as a fort and then a pottery works, with the scrolls carried to the caves by refugees of the war with Rome. Whether the people of the caves were those Essenes, as Sukenik proposed in 1948 before the ruin was dug, now rests in part on that link, which is an inference.

The closest parallel to a saying of Jesus is small. A Hebrew fragment from Cave 4, by its script about a century before the events of the Gospels, says that God will heal the wounded, raise the dead and bring good news to the poor. That is the order of the answer Jesus sends to John in Luke 7:22, where “the dead are raised up, and the poor have good news preached to them.” It shows that the list was already being said, and it shows nothing of any Gospel writer knowing the scroll. The one carbon date on the fragment came out later than its script and is doubted.

No New Testament quotation of Isaiah turns on a reading found only in a scroll, by a check of the forty-one quotations that lean to the Greek or the Hebrew, made for this page by one AI reader and published nowhere else. Those that follow the Greek against the Hebrew still do, and those that follow the Hebrew fall where the scrolls and Leningrad agree. The word “light” at 53:11 is never quoted, and 1 Peter, which has the grass wither once as the first hand does, takes its words from the Greek.

At Nazareth, in Luke 4, the book of the prophet Isaiah was handed to Jesus, and he found the place and read. In the great scroll that place lies near the foot of column 49 of 54, five columns from the end, and there “He has sent me” stands above the line. Luke does not say what copy it was. “He closed the book, gave it back to the attendant, and sat down.”

February to October 2026 · By the museum’s page

The Empty Case

The whole scroll went on show on 23 February 2026, in the museum’s phrase for the first time since 1968. The Israel Museum in Jerusalem, which holds the Shrine of the Book, gave it a white gallery at the heart of the building and a glass case eight meters long, made in Belgium, for a show called “A Voice from the Desert,” which marked the museum’s sixtieth year. At the opening Israel’s president said that a scroll tells two stories, the one written in it and the story of its own journey. Visitors reserved a time in advance. By the Thursday of that week guards were counting them in at the door, no more than twenty-five at once. For the sake of the leather each party had ten minutes. The last hall held a reproduction, marked to show the seam and an ancient patch, for anyone who wanted to look longer.

The walk took in the whole book. Column 1 was dark, its skin worn by long lying on the outside of the roll, with a square-cornered loss near its foot and the ayin of a name set in over its first words, and the vision of Isaiah the son of Amoz opens there. The count begins with the oldest copy. What happened to these words before it, no scroll can say. Repairs crowded the early sheets, signs stood in the margins, and further along the leather paled toward cream. At column 27 the writing stopped with three lines never written, beside the stitches of the next sheet, and at column 33 stood the corner, the thread of small writing turned on its side down the margin, and the four dots. Column 44 held the passage of the Servant, from “Behold, my servant” to the last verse of the chapter after it, in twenty-two lines, and column 54 had eighteen lines, ending on “loathsome to all mankind,” then bare skin for nearly half its height, and no more. A visitor with ten minutes could see all of it.

Not in the case were the first hand and the later hand, who left no names, Kando the cobbler, who passed the scroll on and had $4,000 for it ten years later as a gift, Butros Sowmy, who carried it across Jerusalem on a winter afternoon and was dead by May, Anton Kiraz’s family in one room near the Jaffa Gate, and the man who went into the hole and later dug for wages beside it. By one report he died a refugee in Jordan in the late 1990s. No one can say more than that with confidence. The film that visitors saw first told of Bedouin shepherds and a cave in 1947, of a dealer and an archbishop, and of an anonymous purchase in New York.

The curator called the show a chance to bring together people connected with the scroll’s history. Fidanzio was its scientific director. The Ecole Biblique took part, and pottery and cloths that its excavators had taken from the caves stood in the rooms before the case. The Syriac Orthodox Church lent a New Testament in Aramaic, copied in the fifteenth century, to stand with them.

Five days after the opening, in the war between Iran on one side and the United States and Israel on the other, as sirens sounded and Iranian missiles came in, staff moved the scroll to a secure place, under the emergency orders of the Home Front Command, Israel’s civil defense command. It was very early on Saturday 28 February, and the same orders closed every museum in the country. Pieces from the museum’s Judaica, archaeology and art galleries went with it, and across the street the National Library took its treasures down that morning to vaults deep below the ground. The drum under the dome had been built with such a day in mind, and the scroll had been off it for nearly sixty years. The eight meters of glass stood empty for eleven weeks.

It went back on show on 15 May, a Friday, and by the museum’s page the show closed on 3 October, four months past the date set at its opening. The leather is kept out of the light. It darkened on display in the 1960s, skin of that age can rarely bear being shown in full, and in the Shrine only a small part of the original is normally on view, in a side case. The scroll that most visitors there have seen is a facsimile, and the museum’s calendar for November 2026 offers guided tours of the Shrine. Under the dome, on the drum, the copy stands.

c. 125 BC – 2026

The whole thing, in order

1. About 125 BCThe Great Isaiah Scroll is copied, by its handwriting. Carbon on one column gives the skin a likelier range of about 250 to about 100 BC.
2. After thatLater hands write on it. One study puts the hand at 40:7, by its script, at about 100 to 75 BC. The others have no fixed date.
3. About the middle of the first century BCThe Samuel scroll from Cave 4 is copied, by its hand.
4. About 50 BC to AD 30Masada’s copies are written, by their hands, and left there by the fall of the fortress, AD 73 or 74.
5. About the turn of the eraThe second Isaiah scroll is copied, by its hand.
6. About AD 95, by the usual datingA boast, in an argument with the Greeks: Josephus writes that no one has added, taken away or changed anything.
7. About AD 115The Twelve Prophets scroll is copied, by its hand, and left in a cave by about 135.
8. About 800Timothy hears of a chamber of books near Jericho, by his letter.
9. 916The Prophets codex, the oldest Isaiah with a date that holds, by its colophon.
10. 1008 or 1009The Leningrad Codex, by its colophon.
11. 1760Kennicott expects the oldest copies to differ most. His own oldest are medieval.
12. 1776–80His Bible, from more than six hundred manuscripts. The older text does not appear.
13. 1778Lowth prints the seer at 21:8, reasoning from the Syriac.
14. 1895Kenyon sees no probability of an older Hebrew manuscript, on the copies then known.
15. Winter of 1946–47, probablyThe find, reckoned from interviews of 1961 and by counting back from the sale.
16. July 1947Men with rolls are sent from the door of St. Mark’s, by Samuel’s account of 1949.
17. Summer 1947Samuel gets five rolls: 24 pounds as a security by Kando’s later account, 50 by the American tax case.
18. 24 November 1947Sukenik sees a scrap through the wire, by his diary.
19. December 1947He buys the crushed second roll, by later histories.
20. 6 February 1948He returns the great scroll at the YMCA, by later histories.
21. 19 February 1948Trever lays the largest roll on his bed and copies two lines, by his first article.
22. About 25 March 1948Sowmy takes the scrolls to Beirut without any license, by Trever and Burrows.
23. 10 April 1948The first public word, by Yale’s statement: a monastery’s library.
24. May 1948The British leave and Israel is proclaimed. The establishment of the state displaces about 750,000 Palestinians, by Irving’s history. Butros Sowmy is killed in what she calls Israeli shelling of the monastery.
25. September 1948The first full account in English, third-hand: goods and a rock fall, no goat, no stone, no name.
26. 28 January 1949An officer of the Arab Legion sees the hole, by Trever’s outline of 1965, as Nongbri gives it.
27. 15 February to 5 March 1949By the spade: a cave, jars and linen, and no scroll in a jar.
28. 1949The word goatherd reaches print, on Harding’s guess. That September a journal has him, in a talk of July, on a lost goat, with the year 1937.
29. From 1950The museum’s ledger records what each seller said.
30. September 1952The Ta’amireh find Cave 4, as Milik told it.
31. 1 June 1954The ruled box in the Wall Street Journal, as the histories date it.
32. 1 July 1954Samuel’s trust sells the four for $250,000, by the tax case.
33. By 1955The stone is in print, in a book printed in New York.
34. 1955Burrows writes how little the text has altered, of one scroll of one book, and warns that this does not make it the original.
35. October 1956A man of Bethlehem takes down the finder: 1945 and ten jars. The speaker is paid.
36. 1961Men sent by the mayor of Bethlehem tell the story, and ask to be paid.
37. 1964Archer’s textbook first appears. A later printing has 95 percent, of the two scrolls together. No count for it has been found.
38. June 1967The Rockefeller’s scrolls pass into Israel’s keeping, as the histories tell it. No court has decided who owns them.
39. By 1968A facsimile stands on the drum, by the museum’s account.
40. By 1968Geisler and Nix, as quoted: 166 words in Isaiah 53, seventeen letters in question. The letters come close only for the second scroll.
41. 2021Popovic and his colleagues argue for two scribes, reading the strokes by computer.
42. 23 February 2026The whole scroll goes on show, the first time since 1968, by the museum’s own phrase.
43. 28 February 2026As Iranian missiles come in, staff move it to a secure place, by a museum spokesman.
44. 15 May 2026It goes back on show, by the museum’s announcement.
45. 3 October 2026The show closes, by its page.
46. 2026The count on this page. Nobody outside has checked it.
47. NextA headline about the scrolls. Who said it first?

Who says so

A note on sources

This is a secondhand history, built on the books, papers and saved pages named here and on what could be read of the older editions, and where they disagree the disagreement has been left in the prose. It stops being secondhand at the count, which this page made for itself. Scripture is quoted from the World English Bible unless the sentence names another version. A rendering of a scroll’s reading is this page’s own and keeps the World English Bible’s words wherever it can, and the Rule, the commentaries and the Cave 4 fragments are given in this page’s own renderings or in paraphrase.

The count rests on one transcription, by Martin Abegg, James Bowley and Edward Cook, dated 2015 and drawn mainly from the official editions, in the Text-Fabric conversion by Jarod Jacobs, Martijn Naaijer and Dirk Roorda (ETCBC/dss, version 2.0.2), under the license CC BY-NC 4.0. This page republishes none of its Hebrew, and prints counts made from it and English renderings of single readings. Against it stands the Westminster text of the Leningrad Codex in its written consonants, the word the Masoretes wrote and not the word their margin gives to be read, with the linking hyphen taken as a space and final letters folded. “On the leather” means in that transcription, which nobody here checked against the skin. A letter an editor restored is never counted either way, and a word partly restored is left out. The lack at 63:3 is as the transcription has it, since whether a later hand touched that place could not be read. Phylacteries and doorpost texts stay out of every total, and the one verse in three is 7,645 of 23,144.

The count can be attacked, and these are the places to start. It counts words, so one vowel letter fails a whole word. Spelling here means a waw or yod inside a word, the fuller endings of the period, or one letter exchanged among aleph, he, het and ayin, about 120 words. Per thousand letters the great scroll differs in about 111 and the second in about 27.5, the same story at another scale. Every pairing runs through Leningrad, which measures from the winner and drops about 140 words that the great scroll alone has. Leningrad is also one manuscript, and a reading the scroll shares with another medieval copy still counts as a difference here. How often that happens was not tested. The objection from the winner is fair, and the sentences say “differs” because of it. The controls were modest. Leningrad’s Isaiah set against itself matched in all 16,932 words, changing the measure that aligns the words moved neither scroll, and aligning by verse in place of chapter moved the great scroll by a tenth of a point at most. Taking out every word with a damaged letter moves the great scroll from about 69.3 to 69.5 and the second from about 91.8 to 91.5.

The rounded figures stand on 16,726 words of the great scroll, 2,802 of the second, 2,721 that the two share, and 4,267 from outside Qumran. The 69 and the 92 do not rest on one stretch of the book, since 89 percent of the second scroll’s whole words fall in chapters 34 to 66, where the great scroll stands further off. Like for like they are about 65.4 and 92.1, and the comparison of the two scrolls holds in both halves. The four times in five rests on 432 words, and 1,032 more could not be tested. Run the other way, where the second scroll differs from Leningrad beyond spelling, the great scroll goes with Leningrad 48 times and with the second 32. The seven in eight is 148 corrected words toward Leningrad and 20 away, or 224 and 21 with the large repairs, and it pools every hand. Where another Isaiah scroll differs from Leningrad, only seven corrected words survive, four toward Leningrad, one away and two neither, and in spelling the corrections go 44 away and 25 toward. Taking the corrections out moves the great scroll from about 69.3 to about 68.7. The three places above the line in the Masada Leviticus sit inside its 980 letters.

The sorting of those differences was done by AI, and it should be weighed as that. Ten AI agents, not scholars, coded the differences in three passes, working apart from one another, to one written rubric that asked what would change in an English translation. Their agreement measures steadiness and not two minds, since these were runs of one model on one rubric. The rubric told a torn reader to take the lower code, and its first version gave “light” added and “the seer” for “a lion” as examples of places where, by its terms, the statement of the verse stands. So about 28 is a floor. Where the readers split, the same AI session that built the count settled it. An audit against Leningrad’s vowel signs took 56 words out of the spelling class, and a dozen or so real differences may still be filed there. Published counts of like size, above all Ferguson’s, about 900 that alter the sense, to set against this count’s small words and content words together, and about 30 with no plain scribal explanation, to set against the 28, are a comparison and not a validation.

Nobody working on this page saw the leather. Five readings were checked on the Israel Museum’s digital photographs, by AI readers with no specialist beside them: the dots above the line at 40:7, the words turned a quarter turn beside the seam, “he will see light and be satisfied” on the line at 53:11 with no correction near it, the word above the line at 61:1, and the small ayin at 1:1. Trever’s color plates were read as printed in 1972, every other reading here is the transcription’s, and the check of the 41 New Testament quotations was one AI reader’s. That no article of the creeds turns on a reading in the scrolls is a judgment and not a count. “At least seven” hands rests on Fidanzio’s words in a newspaper interview, and the chapter behind them was not read. The updated New Revised Standard Version was read on Bible Gateway and not in print.

Some of the work this page most needed was not read, because it stands behind a paywall or on library shelves that nobody here reached, and the first of it is the official edition of the two Cave 1 Isaiah scrolls, Ulrich and Flint’s of 2010. Every reading here that carries weight should be tried against it. Also out of reach were Kutscher on the scroll’s language, the editions of Burrows and of Parry and Qimron, Fields’s Full History, Trever’s and Samuel’s own books, Longacre’s article, Fidanzio’s volume of 2026 beyond one chapter, Tigchelaar on the later hands, the articles of Greenberg and Ian Young, Stegemann’s paper behind the report of the finder’s death, Harding’s report in the Cave 1 edition, Brownlee’s article of 1957 with the finder’s statement and Trever’s reply of 1961, Anton Kiraz’s published archive, Wilson’s article, Cross and Rofé on the Samuel paragraph, and Justnes on the hand at 40:7. All are used only as others report them, and Yadin’s book was not read either, so that Sukenik’s diary comes here from excerpts. De Troyer and Herbison’s article in Textus 29 (2020) on Isaiah 40:7–8 was met only in a search summary of its abstract, and the date given for the hand rests on that. The pages of Archer and of Geisler and Nix were not seen, and those sentences rest on later quotation and their years on catalogs and searches. No Bedouin or Palestinian telling in its own right was reached, in English or Arabic, and what was not found here may still exist.

The engravings are modern and imagined. They show no letters and no person, and they record no object.

Further reading

• Frederic Kenyon, Our Bible and the Ancient Manuscripts (1895) — The sentence about older manuscripts, on p. 43, straight after Shapira’s strips. The edition of 1896 is free online.
• John C. Trever, The Discovery of the Scrolls (1948) — In the Biblical Archaeologist 11:3, open online. The first full account in English, with the bed and the slide box in his first words.
• Athanasius Y. Samuel, The Purchase of the Jerusalem Scrolls (1949) — In the Biblical Archaeologist 12:2, open online. The door at St. Mark’s as the metropolitan told it.
• Millar Burrows, The Dead Sea Scrolls (1955) — The sentence on how little the text had altered, with its neighbors on both sides. A scan of the London printing is free online.
• Millar Burrows, More Light on the Dead Sea Scrolls (1958) — Fluid and then fixed, trunk and branches, three years on. A scan is free online.
• Tax Court of the United States, Archbishop Samuel Trust v. Commissioner (1961) — 36 T.C. 641. The sale, in a court’s words, free online.
• Sarah Irving, Palestine’s Syriac Orthodox community and the Dead Sea scrolls (2021) — In Contemporary Levant, with an author’s copy free online. The story from the community that sold them.
• Joan E. Taylor, Dennis Mizzi and Marcello Fidanzio, Revisiting Qumran Cave 1Q and its Archaeological Assemblage (2017) — In the Palestine Exploration Quarterly 149. What the first cave held, read back from the records of the dig.
• Jonathan Poletti, Jinn and the Myth of the Shepherd (2022) — In Gods, Spirits, and Worship in the Greco-Roman World and Early Christianity. The case against the story. A preprint is online.
• Emanuel Tov, Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert (2004) — The handbook on these scribes, with a table of corrections.
• Emanuel Tov, Hebrew Bible, Greek Bible, and Qumran (2008) — His overview of the biblical scrolls and their count. Several chapters are on his website.
• Peter Gentry, The Text of the Old Testament (2009) — In the Journal of the Evangelical Theological Society 52, free online. The case against a text standardized in the first century.
• Drew Longacre, Developmental Stage, Scribal Lapse, or Physical Defect? (2013) — In Dead Sea Discoveries 20, behind a paywall, and unread here. The damaged source. His blog post of 2012 on the Old Testament Textual Criticism blog sets it out free.
• Anthony Ferguson, How Much Can the Most Famous Dead Sea Scroll Prove? (2022) — From the Text & Canon Institute, free online. His 900 and 30 are in a post of January 2025 on the Evangelical Textual Criticism blog, also free.
• Eugene Ulrich and Peter Flint, Qumran Cave 1.II: The Isaiah Scrolls (2010) — The official edition, not free, and the book this page most needed and could not read.
• The Israel Museum, The Digital Dead Sea Scrolls (2011) — The great scroll column by column, free at dss.collections.imj.org.il.

---