Skill: course-build → write-mcqs → write-explanation, design-check-sequence @ b0fa825

Topic 18 — How igneous and sedimentary rock form — the PP100's ten question types, with their exemplars (v3, 2 October 2026: the blind read, the guidance review and the figure review applied; then the v2 re-read's polish — four type 6 scenes and the type 4 limestone scene made fresh, two type 6 moves varied)

What this page is. James reads a bank as its question TYPES first (vault: A bank is reviewed as its question types first, one exemplar each). Under the Grade 4 standing brief the bank is built straight away (53 items in content/G4T18_pp100_bank.py); this sheet is the way in. Each type line says what the stem asks, what the four options (or the accept list) are, and prints one exemplar in full. Tags follow depth of reasoning — easy = a one-step read of a described rock or one fact; medium = the rule applied to a fresh case, the reason picked, or a step placed; hard = two ideas chained, a claim judged by its correct reason, a set answer, or a coined term recalled — and rise with the atom's order: EAR4-007 (what igneous rock looks like) carries easy items, one medium and one hard term item; EAR4-010, the capstone, carries medium and hard only. Every hard item carries a one-line "hard because" (hb) in the bank file.

Rulings already applied. Four options on every MCQ (five fixed sets: the family set "igneous rock / sedimentary rock / a mineral / soil"; the grains set "tiny grains, or glass / big grains / flat layers / pieces of shell"; the cooling set "It cooled fast. / It cooled slowly. / It never cooled. / It was pressed by layers."; the steps set "The pieces settle in a layer. / More layers press down on them. / The pieces stick together. / The pieces melt."; the rocks set "sandstone / shale / limestone / conglomerate"); no option bundles a verdict with its reason (every claim is recast as "X is wrong … What is the correct reason?" with four single-statement reasons); every stem is a positive question (the two "will NOT have" stems are single negatives over positive options, the form the Topic 11 bank keeps as a legitimate DOK 1 / DOK 3 read); an item figure is a family card or a close-up box and carries no verdict chip, tint or family name; "igneous rock", "sedimentary rock" and "sediment" are recalled by MCQ among their sisters (RECALL_BY_MCQ: the first two are eleven and sixteen letters, the third is on the checker's hard-to-spell list), so the one typed key is "layers"; two multi-selects, both where the answer is genuinely a set; every item names ONE atom; US spelling; units as words; no number of years anywhere (the simplification); a stem of three or more sentences is tagged hard.

The bar list (real things a generated item must never use): any rock whose family a child cannot read from the topic's clues — marble, slate, gneiss, quartzite (metamorphic: Topic 19's); chalk and coal (sedimentary rocks the topic never shows); flint and chert; any gem or mineral name offered as a rock (quartz, mica, feldspar are Topic 17's minerals); "magma" (not said in this topic — "melted rock deep underground"); "lava rock" as a rock name (say basalt, or "the dark rock of a lava field"); a time in years; any lesson scene (the tray of rocks, Jace's black rock from Hawaii, Ian's limestone from a Texas hill, Miss Lopez's jar, the Grand Canyon rim, Yellowstone's obsidian); a rock that floats (pumice floats, and Topic 13 bars it from any water scene — here pumice is a dry rock in a hand); a claim-maker from another topic's row (only Ian, Isaac and Jace).

The ten types

Hard because — type 6: each claim is judged by its correct reason, which needs the forming story held against the belief (hot rock against cooled long ago; always-there against formed, asked as what the rock or the layer was before; growing grains against slow cooling then no change; painted stripes against settled layers; one settled layer against many layers and a very long time). Type 7: the formation story names the family, then the family's clue is predicted or denied. Type 8 and 9: a term recalled from its meaning among its sisters, or produced. Type 10: a set held together. Type 4's Indiana item and type 5's conglomerate item: a two-step read (which pieces settled sets which rock; the step before the finished rock).

One exemplar per type, in full

Type 1 (easy, EAR4-007). A hard rock from a Colorado mountain has pink, white and black grains locked together with no gaps, and no layers. Which family is it in? — igneous rock / sedimentary rock / a mineral / soil. Why: Igneous rock is hard, with grains locked together or smooth like glass, and it never has layers. This rock's grains are locked together, and it has no layers. So it is igneous rock. (nearest lesson-side instance: L01-q4, a hard gray rock with tiny locked grains, yes / no — here the form is the four-way family set and the scene is a named mountain)

Type 2 (easy, EAR4-007). Which of these clues is never found on igneous rock? — a hard surface / grains locked together / flat layers / a smooth glassy surface. Why: Igneous rock is hard, with grains locked together or smooth like glass, and it never has layers. Flat layers are the one thing igneous rock never has. (nearest: L01-q8 "Which of these does igneous rock never have?", three options — same move by design, four options here, the hard clue added)

Type 3 (easy, EAR4-008, drawn). Here is a close-up of a rock that formed from melted rock. How did the melted rock cool? — It cooled slowly. / It never cooled. / It cooled fast. / It was pressed by layers. (figure pp_grains_glass, an even dark red-brown fill with a shine streak, no grains — not obsidian's black, so the colour does not cue the key; nearest: L02-q7, a glassy rock in words, fast / slowly — here the glass is drawn and the set is four)

Type 4 (medium, EAR4-008). A rock from a Georgia mountain has grains locked together that are big enough to see. What is the correct reason its grains are big? — The melted rock cooled slowly, deep underground. / The melted rock cooled fast, in the open air. / The rock is very old. / Many layers pressed down on it. (nearest: L02-q4, the correct reason for tiny grains — the mirror case)

Type 5 (medium, EAR4-010). Mud settles on a lake bottom in thin layers, and the layers are pressed for a very long time. Which rock forms? — limestone / conglomerate / shale / sandstone. (nearest: L04-q8, pebbles and sand → conglomerate among granite and obsidian — here the four sedimentary rocks are the sisters)

Type 6 (hard, EAR4-010). A flood covered an Iowa field with a flat layer of mud, and a year later Jace is wrong to say the mud has turned into rock. What is the correct reason the mud is still loose? — Mud can never turn into rock at all. / The field was too cold for rock to form. / Nothing has pressed it for a very long time. / Field mud is the wrong kind of mud for making rock. Hard because: the forming story is held against one settled layer — settling is only the first step; the pressing by more layers and the very long time are missing. Every hard key is one clause and no key is the longest option (the guidance review's Essential). (nearest: L04-q7, Ian's TRUE claim about a bucket, yes / no — a different child, a different scene kind (a field, not a vessel), and here the verdict is given and the reason is the target; the v2 glass bowl was the jar with a new vessel, the re-read's flag)

Type 7 (hard, EAR4-008). A rock formed when melted rock cooled and hardened deep underground. Which of these will the rock NOT have? — big grains / grains locked together / flat layers / a hard surface. Hard because: the forming story names the family (igneous), then the family's one never-clue is picked. (nearest: none on the lesson side — the lesson quizzes go looks → family, never formation → clue)

Type 8 (hard, EAR4-009, typed). In a sedimentary rock, small pieces settled time after time, each lot on top of the last. The rock shows them as flat parts that run right through it, one on top of another. Type the one word for those flat parts. — accept: layers, layer, a layer, the layers, flat layers and ten child spellings (lairs, layrs, layres, lyers, laiers, layors, leyers, layerz, layes, lares); "stripes", "bands", "sheets" and "lines" each get a wrong-term line (stripes and bands are what the layers look like from outside; shale splits into sheets along its layers). Hard because: the visible clue is produced as a word from its description. (nearest: L03's article line "The layers go right through the rock"; no lesson item types it)

Type 9 (hard, EAR4-009). A rock is made of grains or pieces pressed together, often in flat layers, sometimes with shells inside. What is a rock like that called? — igneous rock / sedimentary rock / sediment / a mineral. (nearest: L03-q10, the family from "pieces pressed together in flat layers" among three — here the definition is whole and the sisters are four)

Type 10 (hard, EAR4-008, multi). Select the TWO rocks that formed when melted rock cooled and hardened. — granite, with its locked grains / sandstone, with its pressed sand grains / obsidian, smooth like glass / shale, with its thin flat layers. Hard because: each rock's clue must be read to its family, and the two igneous rocks held together. (nearest: none — the lesson side never asks a set)

Design-move inventory (move · scene kind · nearest earlier item), one line per type with its scenes

The bank's 53 moves against the lesson side (L01 9 quiz items, L02 11, L03 11, L04 12, 4 attached, 3 summary, 13 practice; the lessons' own cases: granite, obsidian, basalt, pumice, Jace's striped rock (L01); Jace's Hawaii rock, lava in Hawaii, the cross-section, the fast / slow table (L02); sandstone, limestone, shale, conglomerate, granite (L03); the Texas hill, Miss Lopez's jar, the step cards, the pieces-to-rock table, the two-family table (L04)):

• Type 1 (10): described rock → family — a Colorado mountain rock (locked, no layers), Rock S's card (drawn: a dark red, gray and black locked tiling — not Rock J's gray and white), an Oregon hillside glass rock, Rock V's card (drawn: dark gray, glassy with holes — not pumice's pale gray), a Washington road-cut rock (dark gray, tiny grains, a grain will not rub off — the second pumice-in-words item was replaced), a Utah cliff rock (pressed sand, layers), Rock R's card (drawn: gray pebbles of many sizes in pale sand — not conglomerate's or L03-q5's brown pebbles), a rock dug up in Florida (shells, layers), Rock T's card (drawn: pale yellow sand in rows — not Rock K's tan and brown), a Kentucky road-cut rock (shells among tiny grains, layers; medium because two clues are read together). Nearest: L01-q1–q7 and L03-q1–q7 (yes / no on one family), L04-q9–q10, PR-4, PR-5, SUMMARY-Q1–Q2 (two-way sort). Same move by design (the fluency form); the bank's form is the four-way family set, and every scene is a fresh named place or a new lettered card (J, K, L, M, N, P are the quizzes' letters; R, S, T, V the bank's) whose rows, colours and close-up differ from every quiz card's.
• Type 2 (6): the clue that settles the family — flat layers as the never-clue (007); no-layers as the second clue for a glassy rock; locked grains as the second clue for a hard rock with no layers; the layers clue for sedimentary (009); a Maine beach rock with grains too small to see (which second clue says igneous, not sedimentary — new form: the two families contrasted); an Ohio river-bank rock with layers (which second clue says sedimentary). No option in the bank bundles a verdict with a reason: the one draft that did (shells → family, reasoned options) was recast as a type 1 item on a Kentucky road cut. Nearest: L01-q8, L03-q8 (three-option never / often items). The second-clue form is new.
• Type 3 (8): cooling ↔ grains — lava on an Oregon hillside (fast → grains set), melted rock deep underground (slow → big), a glass close-up (drawn, dark red-brown → cooled fast), a big-grain close-up (drawn: 12 gray cells of about 30 px, not L02-q6's 30 cells of about 20 px → cooled slowly), melted rock → igneous rock (the what), pea-sized grains → where it cooled (four places: new), a thin lava flow over cold Alaskan ground, hard within an hour (→ tiny or glass; the v1 Idaho-lake item was L02-q11's lava-into-the-sea with a lake for the sea), two rocks from one melted rock → which has the bigger grains (new comparison form). Nearest: L02-q3 (Idaho ground, what is the lava after cooling), q5–q7 (drawn close-ups and a glassy rock, fast / slowly), q11 (lava into the sea), PR-6 (cooled fast → grains, three options), SUMMARY-Q3 (slowly → big).
• Type 4 (5): the correct reason — a Georgia mountain's big grains; an Arizona hillside's glass; a New York shale cliff's layers; a Wisconsin wall's shells inside; an Indiana courthouse's limestone blocks, not sandstone (the v2 Kansas hill was Ian's Texas hill with a new state). Nearest: L02-q4 (tiny grains, the reason), L04-q4 (a sandstone cliff's layers, the reason). The glass, shells and limestone-not-sandstone reasons are new.
• Type 5 (8): the steps and the pieces — first / next / last (three items on the steps set), sand → sandstone, mud → shale, shells → limestone (the rocks set), where the pieces settle (four places), what happened to a conglomerate's pebbles before the pressing (backwards). Nearest: L04-q3 (what happens first, three options), PR-7 (what happens last, three options), L04-q8 (pebbles and sand → conglomerate among igneous rocks). The next-step, the rocks-set and the backwards forms are new.
• Type 6 (6): claims by their correct reason, or by what the thing was before — Jace's hot rock on a Minnesota museum shelf (the v2 desert rock was Jace's lava rock with a new place); Isaac's always-there black Montana hillside, asked as what the rock was before it was hard rock (the v2 reason form was L02-q9 / L02-q4 near-copied); Jace's growing grains in a Wyoming mountain; Isaac's striped pebble on a Michigan lake shore (the v2 Nevada cliff was L03-q9's cliff renamed); Ian's always-there layers in a Tennessee road wall, asked as what each layer was before it was rock (the v2 reason form was L04-q6 near-copied); the mud a flood left on an Iowa field (the v2 glass bowl was the jar with a new vessel). Nearest: L02-q8 (Isaac, TRUE: an old lava flow's rock is as cool as the ground), q9 (Jace, rocks always here), q10 (Ian, granite's grains kept growing), L03-q9 (Ian, a sandstone cliff's painted stripes), L04-q5 (Isaac, the rock grew its shells), q6 (Jace, a canyon wall's layers always there), q7 (Ian, TRUE: the bucket's layers need more), PR-8 (Isaac, TRUE: the grains stopped getting bigger when the rock hardened). Same beliefs by design (the misconceptions are the point); the v1 bank re-cased each quiz claim with the same child — the blind read caught it, and every bank item now has a different child AND a different scene from the quiz item on the same belief. The claim string on the lesson side is 3 TRUE of 8 (L02-q8, L04-q7, PR-8).
• Type 7 (3): formation → clue — settled sea-floor pieces → flat layers; a river-bed conglomerate → will NOT have locked grains; melted rock deep underground → will NOT have flat layers. Nearest: L04-q11–q12 (formation → family, two options). The chain to the clue is new.
• Type 8 (3): sediment named — a pond bottom (MCQ), a slow river's bed (MCQ), the flat parts typed as "layers". Nearest: L03-q11 (a stream bed, sediment among soil / a mineral), PR-10 (the sea floor), PR-11 (a puddle's mud — the v1 bay floor was PR-10 with bay for sea). The scenes differ (pond, river bed); the typed "layers" is new.
• Type 9 (2): the family from its definition — the two whole definitions. Nearest: L01-q9 (igneous rock among a mineral / soil), L03-q10, PR-12, PR-13 (from the forming sentence). The definition-whole form with four sisters is the bank's.
• Type 10 (2): a set — the two sedimentary clues; the two igneous rocks. Nearest: none.

⟂ draw constraints (an item that prints another's answer; near-duplicate pairs)

• Type 2's never-clue item ("Which of these clues tells you a rock is NOT igneous rock?" → flat layers) ⟂ type 7's "deep underground → will NOT have flat layers" and type 1's any card with LAYERS none keyed igneous: the never-clue is the answer the others lean on.
• Type 9's two definition items ⟂ every type 1 item (the definition is the sort's rule) and ⟂ type 8's two sediment MCQs (sediment sits in their option sets as a sister).
• Type 3's "melted rock → igneous rock" (the what) ⟂ type 10's "two rocks that formed from melted rock" and type 6's two lava-field reason items (they state the forming story).
• Type 5's three rocks-set items (sand → sandstone, mud → shale, shells → limestone) ⟂ type 4's Kansas item (limestone because the pieces were shell) and type 5's conglomerate read backwards.
• Type 3's two drawn close-ups (glass; big grains) ⟂ each other on the drawing convention (a reader who sees both learns the pair), and ⟂ the type 4 Georgia / Washington reason items.
• Near-duplicate pairs, ruled ⟂ by the blind read (2 Oct 2026): PP-27 ⟂ PP-28 ⟂ PP-29 (the three steps items on one set — each prints the order the others ask; draw one per test); PP-38 ⟂ PP-39 (the two sediment MCQs — the same key word; draw one); PP-16 ⟂ PP-17 (the two drawn fast-or-slow close-ups — a reader who sees both learns the pair; draw one); the two always-there claims (type 6, PP-44 and PP-47, both now asking what the thing was before it was rock) — draw one.

Freshness claim, corrected (2 Oct 2026)

The v1 record's scene-noun grep searched "lava flow" and missed "lava field", which five items carried (L02-q4, PR-8, PP-25, PP-43, PP-44). The v2 grep runs over every content file for both phrases and for every quiz scene noun (lake, sea, bay, bucket, canyon, Hawaii, Texas, tray, jar, thumb): "lava field" 0, "lava flow" in L02's own headache and L02-q8 only, "bucket" in L04-q7 only, "lake" 0 in the bank, "bay" 0. The blind-text grep over the bundle's [Drawing: …] lines for igneous | sedimentary | cooled | settled | melted | sediment | tiny | big: 0 hits of 15 lines.