The Ark, the Table, the Lampstand, and the Court

• Anchor Bible Dictionary — “Ark of the Covenant,” “Mercy Seat,” “Cherubim,” “Lampstand,” “Tabernacle,” “Weights and Measures.”
• Dictionary of the OT: Pentateuch — “Tabernacle,” “Sacred Space,” “Craftsmanship,” “Furniture of the Tabernacle.”
• NIDOTTE / TDOT — כַּפֹּרֶת (kappōreṯ, mercy seat), מְנוֹרָה (mĕnôrāh, lampstand), חָצֵר (ḥāṣēr, court), כִּכָּר (kikkār, talent), פְּקוּדִים (pĕqûḏîm, records / musterings), צבא (ṣāḇā’, to serve, to muster).
• TDNT / EDNT — ἱλαστήριον, σκηνή, λυχνία, καταπέτασμα — because the Greek column of this passage carries the tabernacle vocabulary the Greek New Testament also uses.

• Childs (OTL), Durham (WBC), Propp (Anchor Yale, Exodus 19–40), Sarna (JPS), Stuart (NAC) on Exodus.
• Menahem Haran, Temples and Temple-Service in Ancient Israel; Carol Meyers, The Tabernacle Menorah; G. K. Beale, The Temple and the Church’s Mission.
• On the Greek: John William Wevers, Notes on the Greek Text of Exodus and Text History of the Greek Exodus — the standard treatment of the rearranged chapters 35–40. Named, not consulted for this page; see Stage 5.

A portable, gilded, carried shrine. Nothing in Exodus 37 would have puzzled an ancient craftsman. Gilded wooden chests with carrying poles and outstretched winged figures are known from Egypt — the nested gilded shrines and the poled chest from Tutankhamun’s tomb are the standing comparison — and thrones flanked by winged composite creatures are found across Syria and Canaan. The craft is shared; that is not in dispute and it is not a problem. What is unshared is the top of the box. The כַּפֹּרֶת carries two cherubim and nothing between them. In every neighbouring culture the winged guardians flank an image; here they flank an empty space, and God says he will speak from it (Ex 25:22; Num 7:89). The page does not argue from this that Israel borrowed or did not borrow. It notes what the object has and what it conspicuously lacks.

Hammered from one piece. The lampstand is מִקְשָׁה, hammered work — base, stem, branches, cups, calyxes and flowers “of one piece with it” (37:17), out of a single talent of pure gold (37:24). A talent is the heaviest weight in the Israelite system, worked by hand into an almond-blossom tree. Whatever else the menorah is, it is a tree of light standing in a tent, and it is the one tabernacle object of which we have an ancient picture: the Arch of Titus in Rome, erected soon after AD 81, shows a seven-branched lampstand carried in triumph out of Jerusalem. That is the Second Temple’s menorah, not Bezalel’s; but it is the object’s afterlife in stone.

Bronze mirrors. A polished bronze hand-mirror was ordinary, valuable and personal — Egyptian tombs are full of them. Exodus 38:8 says the priests’ washing basin was cast from a collection of them, given by women who served at the door of the tent. The Greek column of the same verse has τῶν νηστευσασῶν, αἳ ἐνήστευσαν, “of those who fasted, who fasted,” and adds “on the day he pitched it.” The two columns carry the same three consonants, צבא, and give them different senses. The two readings are printed side by side above. This page does not adjudicate between them and does not describe either as a mistake.

How the golden lid was read between the testaments. The Greek name of the mercy seat, ἱλαστήριον, stands 28 times across the 30,348 Greek Old Testament rows screened here — thirteen in Exodus, seven in Leviticus 16, five in Ezekiel 43, once in Numbers 7:89, once at Amos 9:1, and once outside the canon, at 4 Maccabees 17:22, where the deaths of the martyrs are called τοῦ ἱλαστηρίου τοῦ θανάτου αὐτῶν, “their atoning death,” through which “divine Providence preserved Israel.” 4 Maccabees is a witness to how Greek-speaking Jews used the word in the first century, not Scripture and not evidence of what Paul meant. It is worth knowing that the word was live and that it could be used of a death.

• Josephus, Antiquities 3.134–150 — the ark, the table and the lampstand described for a Roman readership; Philo, Life of Moses 2.71–108 — the same furniture allegorised (witnesses to interpretation, not Scripture).
• 4 Maccabees 17:20–22 — ἱλαστήριον of a martyr’s death; measured from the CATSS file on this shelf, quoted here as a Second-Temple usage.
• The Arch of Titus, Rome (AD 81) — the seven-branched lampstand of the Second Temple carried in triumph; the object’s afterlife, not a picture of Bezalel’s work.
• 1 Samuel 2:22 — the only other passage that places women serving at the entrance of the tent of meeting; it shares three consecutive words with Exodus 38:8.

• Romans 3:25; Hebrews 9:5 — the two occurrences of ἱλαστήριον in the Greek New Testament: Christ publicly set forth, and the golden lid itself (see Just and the Justifier).
• Hebrews 9:1–5 — the New Testament’s own inventory of this furniture: the lampstand, the table, the bread of the Presence, the golden altar, the ark, the jar of manna, Aaron’s rod, the tablets, and the cherubim overshadowing the mercy seat. Measured against Greek Exodus 37–38 it reaches two words — the page’s own counter-example, printed in Stage 4.
• Acts 7:44; Revelation 15:5 — the only two places the Greek New Testament uses the Septuagint’s name for the tabernacle, ἡ σκηνὴ τοῦ μαρτυρίου (see The Same Twelve Words).
• Numbers 7:89 — the voice speaking to Moses “from above the mercy seat that was on the ark of the testimony, from between the two cherubim” (see the voice from the mercy seat).
• Leviticus 16:14–15 — the blood sprinkled on and before the כַּפֹּרֶת on the Day of Atonement (see the Day of Atonement).

Exodus 35–40 narrates the building of what Exodus 25–31 commanded. The obvious question is how closely. Each of the four ‘making’ chapters was screened, verse by verse, against every other verse of the Hebrew Bible. The result is not uniform, and the sharpest break in the whole account falls exactly at the seam this page sits on.

Chapter 37 has the highest ceiling of the four and chapter 38 has the lowest — and they are adjacent. Exodus 38 is the only one of the four whose ceiling never reaches eight: not one of its thirty-one verses shares eight consecutive words with anything in the Hebrew Bible. Chapter 37, next door, does it twelve times.

Here is chapter 37 in detail. Every deep row is band 1 — the words stand nowhere else in the Hebrew Bible in that order — and 26 of the 29 rows are band 1 overall.

Why eighteen, and why there? The run engine matches word forms. The command chapters say תַּעֲשֶׂה, you shall make; the making chapters say וַיַּעַשׂ, and he made — so the transposition from command to report cuts almost every run at the verb. The eighteen-word run at 37:21 is the longest precisely because the stretch it copies contains no verb at all: it is the triple repetition of ‘and a calyx under two branches of it.’ It runs from the first word of the verse to the eighteenth — one word short of the end of Exodus 37:21, which has nineteen words, and two short of the end of Exodus 25:35, which has twenty — because Exodus 25:35 finishes ‘going out of the lampstand’ where Exodus 37:21 finishes ‘going out of it.’ The ceiling is set by grammar as much as by fidelity, and a page that reported the 18 without that sentence would be overselling it.

The low ceiling of chapter 38 is real, but it must not be read as unfaithfulness, and the bands say why.

Three honest qualifications, printed here rather than buried. (a) Chapter 38’s material is different in kind: the court command in Exodus 27:9–19 is terse and largely numerical to begin with, and the metal accounts of 38:21–31 have no command behind them at all — there is nothing for them to repeat. (b) Its deepest run is not even a command match: 38:6 at seven words matches itself, the poles of the table and of the incense altar one chapter back. (c) Its widest bands are stock phrases. ‘Blue and purple and scarlet yarns and fine twined linen’ reaches six words at a band of ten; ‘the entrance of the tent of meeting’ reaches three at a band of forty. A run of six with a band of ten is a formula, not a quotation, and this page counts it as one.

CATSS/Rahlfs Exodus does not merely number chapters 35–40 differently. It arranges them differently. The correspondence below was established row by row from the files on this shelf.

One block has nowhere to go. Hebrew Exodus 37:25–28 — the making of the golden altar of incense, 58 Hebrew word tokens — has no counterpart anywhere in CATSS Exodus 35–40. The Greek book has the object: at 40:5 Moses is told to place τὸ θυσιαστήριον τὸ χρυσοῦν εἰς τὸ θυμιᾶν before the ark, and at 40:26 he does it. What the Greek construction account never narrates is anyone making it. This page states the fact and stops there. It does not say the Greek is shorter because something dropped out, and it does not say the Hebrew is longer because something was added; both explanations are current in the literature and neither is measurable from these files.

Length, measured within each language rather than across them. Comparing Hebrew word tokens with Greek word tokens tells you about tokenisation, not about text, so the useful comparison is each column against itself. In the Masoretic column the construction account (35–40) runs at 88% of the length of the instruction account (25–31) — 2,772 tokens against 3,156. In the Greek column it runs at 67% — 3,204 against 4,806. Verse units: 214 Masoretic against 173 CATSS.

Two Greek sentences with no Hebrew counterpart, printed above in the codex. At LXX 38:22 the bronze altar is made ἐκ τῶν πυρείων τῶν χαλκῶν, ἃ ἦσαν τοῖς ἀνδράσιν τοῖς καταστασιάσασι μετὰ τῆς Κορε συναγωγῆς — from the bronze censers of the men who rebelled with Korah’s company. The name Korah stands in Greek Exodus at exactly three places: 6:21, 6:24 and 38:22. Its subject matter is that of Numbers 16–17, but the longest run LXX Exodus 38:22 shares with LXX Numbers 17:4, the censer verse, is zero — and on the Greek side of this shelf a zero means the two verses share no word form at all. The link is topical, not verbatim, and the page will not call it a quotation. And at LXX 38:26 the basin is cast from the mirrors τῶν νηστευσασῶν, αἳ ἐνήστευσαν — of the women who fasted — where the Hebrew has הַצֹּבְאֹת אֲשֶׁר צָבְאוּ, the women who served, plus a Greek plus, “on the day he pitched it.”

And the accounts do not agree.

Note which two agree. The silver figure is the one total in the passage that can be derived: 603,550 men at half a shekel each is 301,775 shekels, which at 3,000 shekels to the talent is exactly 100 talents and 1,775 shekels — and the text says so itself at 38:27–28, casting one talent per base for a hundred bases. Both columns carry that figure and both carry the head-count. The gold and the bronze cannot be computed from anything, and those are the two that differ. That is an observation about the numbers, not a theory about how either column came to read as it does, and this page offers no such theory.

All 48 CATSS rows of Greek Exodus 37–38 were screened against all 7,927 verses of the Greek New Testament. Across the 48 rows: all 48 reach two, 23 reach three, 5 reach four, 1 reaches five, and nothing reaches six. For scale, the whole of Greek Exodus reaches thirteen — Exodus 20:11 against Acts 14:15 and Acts 4:24, band 2.

Mind the chapter labels. Because the Greek rearranges these chapters, ‘Greek Exodus 37–38’ is not the same 60 verses as ‘Hebrew Exodus 37–38.’ The 48 CATSS rows screened here carry all of Hebrew 37:1–24 and 37:29, all of Hebrew 38:1–23, and also six rows belonging to Hebrew 36:8–38 — while Hebrew 38:24–31, the metal totals, sit in Greek chapter 39, which was screened separately and is reported below. The Hebrew verse is given beside every Greek label.

Now the counter-example this page is obliged to print. Three of the five deepest runs land on the same verse, Revelation 15:5 — ‘the sanctuary of the tent of witness in heaven was opened.’ It is tempting to read that as Revelation reaching back to the building of the tabernacle. It is not. ἡ σκηνὴ τοῦ μαρτυρίου is simply the Septuagint’s ordinary name for the tabernacle: the genitive string occurs 91 times across the 30,348 Greek Old Testament rows screened here — 35 in Numbers, 31 in Leviticus, 19 in Exodus, 3 in Deuteronomy, 2 in 1 Chronicles, 1 in Joshua. A run of four or five on those words is a title, not a citation, and it would score the same against a hundred other verses in the Pentateuch. The measurement finds verbatim contact; it does not find allusion, and here the difference is the whole story.

Meanwhile the New Testament passage that really does describe this furniture measures at the floor. Hebrews 9:1–5 walks through the lampstand, the table, the bread of the Presence, the golden altar, the ark, the cherubim and the mercy seat — the exact contents of Exodus 37 — and against every verse of Greek Exodus 37–38 it reaches two words.

Hebrews 9:5 and LXX Exodus 38:5 both contain τὸ ἱλαστήριον, and that two-word run is the whole of their measured contact. Nobody would conclude from this that Hebrews is not thinking about Exodus 37. Obviously it is. What the number shows is that Hebrews is describing the tabernacle in its own words rather than copying a Greek column — which is precisely the distinction a run screen can see and a thematic reading cannot. It is also the reason this shelf reports the band beside every run and refuses to convert depth into dependence.

And there is a coda that sharpens the point rather than softening it. Greek Exodus 39 — the chapter that carries the Hebrew metal totals of 38:24–31 and then the long list of finished furniture handed over to Moses — was screened too, and it does reach Hebrews. Across its 23 CATSS rows: all 23 reach two, 16 reach three, 6 reach four, 2 reach five, and nothing reaches six. Both fives are Hebrews, both band 1: LXX Exodus 39:14 against Hebrews 9:4 at καὶ τὴν κιβωτὸν τῆς διαθήκης, and LXX Exodus 39:19 against Hebrews 9:21 at καὶ πάντα τὰ σκεύη τῆς. Both Greek rows are lists — inventories of finished objects being carried in — and so is Hebrews 9. The Greek verses that report the making reach two; the Greek verses that report the inventory reach five. That is not a claim that Hebrews used those verses; it is a demonstration that a run screen measures genre and phrasing as much as it measures dependence, which is why this shelf never reads a run as a citation on its own.

• Exodus 25:10–40 — the command that chapter 37 reproduces, at up to eighteen consecutive words (see the ark, table and lampstand).
• Exodus 27:1–19; 30:1–10, 17–21 — the commands behind the bronze altar, the court, the golden altar and the basin (see the bronze altar and the court and the altar of incense).
• Exodus 32:1–6 — the same gold, the same eager giving, the wrong pattern (see the golden calf).
• Exodus 39:32–43; 40:34–38 — the work inspected and blessed, and the glory that fills it (see the tabernacle built and the glory fills the tabernacle).
• Numbers 7:1–89 — the tabernacle anointed, the leaders’ offerings weighed in the same sanctuary shekel, and the voice from the mercy seat (see the offerings of the leaders).
• Numbers 16:1–17:13 — the censers of Korah’s company, which the Greek column of Exodus 38:22 names and the Hebrew does not (see Korah’s rebellion).

• Brevard Childs (OTL), John Durham (WBC), William H. C. Propp (Exodus 19–40, Anchor Yale), Nahum Sarna (JPS), Douglas Stuart (NAC) on Exodus.
• John William Wevers, Text History of the Greek Exodus and Notes on the Greek Text of Exodus; Anneli Aejmelaeus and the Finnish school on the translation technique of Greek Exodus — the standard literature on the rearranged chapters 35–40.
• Menahem Haran, Temples and Temple-Service in Ancient Israel; Carol Meyers, The Tabernacle Menorah; G. K. Beale, The Temple and the Church’s Mission; Beale & Carson, Commentary on the NT Use of the OT (Rom 3:25; Heb 9).

• No critical apparatus was consulted, on either side. The Hebrew is the Masoretic text as printed in Sefaria’s Miqra according to the Masorah. The Greek is CATSS’s digitisation of Rahlfs. Neither the Samaritan Pentateuch nor the Qumran Exodus manuscripts (4QpaleoExod^m and its relatives) were examined — and for chapters 35–40 those are precisely the witnesses a specialist would reach for first. Every statement here about “the Greek” means these files, not the whole Greek tradition.
• The missing golden altar is reported, not explained. Whether CATSS Exodus 35–40 is short because a Hebrew Vorlage was short, or because a translator abridged, or because the Hebrew grew, is the central disputed question in the literature on these chapters (Wevers, Aejmelaeus, Gooding). This page takes no position and could not, from local files. It measures absence; it cannot measure why.
• The ten-shekel and nine-hundred-shekel differences in the metal totals are reported without a cause. The observation that the two divergent totals are the two that cannot be derived arithmetically is an observation, not a demonstration. Numerals are notoriously unstable in transmission in every direction, and nothing on this shelf can tell you which column stands nearer an original.
• The run screen finds verbatim contact and nothing else. It cannot see paraphrase, allusion, shared imagery or common structure. Hebrews 9:1–5 measuring two is the demonstration, printed in Stage 4 for exactly that reason. A low number here is not evidence of independence, and a high number is not evidence of borrowing.
• The Hebrew screen’s zeros mean “no two consecutive words” (single words are not counted), while the Greek screen’s convention would report 1 for a row sharing only isolated words. No row in this passage came out at zero on either convention, so nothing on this page turns on the difference — but the two conventions are not interchangeable and the distinction is stated here rather than assumed.
• Numbers 7 was screened against, not for. Several Exodus 38 rows find their nearest partner in Numbers 7’s offering lists rather than in Exodus 27. This page reports the fact and does not build on it; the direction of that relationship is exactly the kind of claim the house rules forbid.
• No acrostic or letter screen was run, because the passage is prose inventory and there is nothing for such a screen to find. No Second-Temple Greek corpus outside the CATSS files was searched; the 4 Maccabees datum is from the file on this shelf, and Josephus and Philo are cited from the standard editions and were not measured.
• The ESV is the English column. No Crossway API key has ever been issued to this site; the ESV here is quoted under Crossway’s standing permission, and the shelf tracks its footprint per book. This page adds sixteen verses of Exodus.

• Obedience is measurable, and here it was measured. Exodus 37 repeats the command of Exodus 25 at up to eighteen consecutive words, and the words stand nowhere else in the Hebrew Bible in that order. The chapter is not filler. It is a receipt.
• God keeps the accounts in public. Twenty-nine talents and 730 shekels of gold; 100 talents and 1,775 of silver; 70 talents and 2,400 of bronze — audited, assigned, and put under a named officer (38:21). Church money and church work have been open books since Sinai.
• Somebody gave what they could not replace. The priests washed in bronze that had been women’s mirrors (38:8). The participle Scripture uses of those women stands in only one other verse in the whole Hebrew Bible — 1 Samuel 2:22, of women at the same door. Their names are gone; their gift is still in the text.
• At the centre there is a space. The mercy seat had two cherubim and nothing between them — and God said he would speak from there. Paul calls Christ the ἱλαστήριον (Rom 3:25); Hebrews uses the same word of that lid (Heb 9:5). The vacancy was a promise.

Step in response: this week, do one unglamorous, exactly-as-asked piece of work for God’s people and let nobody know you did it — the kind of thing that would end up in a chapter nobody memorises. Then give something you would rather keep, as those women gave their mirrors. And take five quiet minutes at the gold lid: two cherubim, an empty space, a voice. That is where the blood went; that is where the voice came from; and that is the word the New Testament reaches for when it wants to say what the cross was.