# Mashers, ricers, presses and mortars: research notes

Date: 2026-10-25. Scope: hand tools that crush food against or through a surface: potato and bean mashers (wire and plate), potato ricers and food mills, garlic presses and garlic rockers, and mortars and pestles (granite, molcajete, Thai khrok, suribachi, marble, porcelain, steel). In the Zamana taxonomy this is `supplies > kitchen > mashers`.

Overlaps with siblings, left to them: nutcrackers, ice crushers and can crushers are `openers` ("Openers, Crackers & Crushers"). A food mill is sometimes sold as a "sieve" and could be claimed by `strainers`; it stays here because experts compare it directly with the ricer for the same job. Tortilla presses and burger presses carry the word "press" but are shaping tools with no shared criteria; they are not faceted here and are flagged as an unclaimed tail.

Citrus hand presses: `cocktail-mixing-tools/notes.md` defers them here ("to whichever kitchen leaf holds presses (`mashers`…)"), and this register had deferred them back to `cocktail-mixing-tools` / `bottle-wine-serving-tools`, which never mentions them — a cycle. This register now claims no ownership: citrus presses share no facet with the five tools here (no mash texture, no mortar size, no press_style value fits a citrus reamer), so they belong with `cocktail-mixing-tools`. That register's notes.md needs a matching fix to close the cycle; not done here, since this task only touches files under `mashers/`.

Push-down garlic choppers: the boundary is blade versus hole. A push-down cube press that forces cloves through holes with no cutting edge is the `press_style` value "push-down chopper" here. A push-down chopper with a cutting blade or spokes (Chef'n GarlicZoom, "Pressed Garlic Chopper Manual Garlic Press for Potato, Peppers and Onions") cuts rather than presses and sits with `knives` or `graters`.

Bladed meat choppers (19 of 98 masher search titles: "Meat & Potato Masher", "Ground Beef Smasher", "5 Curve Blades") cut ground meat with blades rather than pressing it through holes or wires. They are excluded from `tool_family`'s masher value and flagged as an unclaimed tail, the same treatment as tortilla presses.

Pill crushers and lab mortars (16 of 77 mortar search titles: "Pill Crusher", "United Scientific … 70ml") crush non-food items and are excluded from `tool_family`'s mortar and pestle value.

Electric tomato strainers, mixer strainer attachments and electric grain mills are motorized and excluded from `tool_family`'s food mill value, which covers hand-crank mills only.

Quote labels: **verbatim** = the sentence was read in raw text I downloaded (curl plus tag stripping, pdftotext for PDFs, or the Wikipedia plain-text API); **fetched excerpt** = from the WebFetch summary tool; **search excerpt** = from a search-engine summary only. No search-excerpt number is used in the JSON.

## Seed

Seed output (verbatim from `seed.py --category mashers`):

```
=== Example entry (from an unrelated category) ===
category: kitchen-knives
what this shows: a numeric spec that only makes sense as one side of a trade-off, and a visual grammar borrowed from the hobbyist community that puts the trade-off on one axis.

<example>
{
  "id": "steel_hardness",
  "label": "Steel hardness",
  "definition": "How hard the blade steel is, measured on the Rockwell C scale, which sets how long an edge stays sharp and how easily it chips.",
  "why_experts_care": "Harder steel holds a thinner, sharper edge for longer but is brittle, so it chips on bones and hard squash and is harder to sharpen at home. Softer steel dulls faster but bends rather than breaks and comes back with a few strokes on a honing rod. Japanese knives cluster at the hard end and German knives at the soft end, which is most of what the Japanese-versus-German debate is actually about.",
  "kind": "numeric",
  "unit": "HRC",
  "range": { "min": 52, "max": 66, "typical": "German-style knives sit around 56 to 58. Japanese-style knives sit around 60 to 63." },
  "roles": ["annotation", "sort_key"],
  "novice_answerable": false,
  "source": "listing",
  "listing_field_hint": "Often in the specs table as HRC or Rockwell. Missing on cheaper knives, which usually means soft.",
  "backfill_cost": "low",
  "backfill_note": "Manufacturer pages carry it for anything mid-range and up. Below that, infer soft from the steel name if given.",
  "visual": { "chartable": true, "grammar": "edge retention versus toughness scale", "encoding": "position on a single horizontal scale, hard on the right, with chip risk shaded in" },
  "priority": { "narrows_set": 3, "cost_to_answer": 4 },
  "evidence": ["knifesteelnerds_hardness", "seriouseats_knives"]
}
</example>

Trade-off entry it belongs to:
{
  "optimise": "steel_hardness",
  "give_up": "toughness and ease of sharpening",
  "plain": "Push hardness up and you get a knife that stays sharp for months but needs a whetstone and care around bones. Push it down and you get a knife you can abuse and hone in the kitchen but will sharpen more often.",
  "see_more_branch": "See more low-maintenance knives"
}

=== Persona ===
A car mechanic, who thinks in wear, serviceability and what fails first

Ask yourself two things and write the answers under 'Seed' in notes.md:
  1. What is the mashers equivalent of the example above?
  2. What would this persona notice first about mashers that a typical buyer would not?
Keep what survives the data check. Drop the rest.
```

1. **The mashers equivalent of steel hardness.** Two specs that only make sense as one side of a trade-off. (a) The ricer's hole count: more holes of the right size means less force for the same mash, but holes on the hopper sides spray potato everywhere (ATK says it now accepts the mess for the lower force). Survived as `perforation_count`, but only test reviews count holes, so it is researched and high cost. (b) The mortar's interior roughness: rough unpolished stone grinds fastest but sheds grit until seasoned and traps food; smooth is easy to wash and cannot grind spices. Survived as `mortar_interior` and a trade-off. The hardness scale became the mash texture scale (chunky to silky, one mark per tool).
2. **What the car mechanic notices first.** Wear, parts count and what fails first. Survived: (a) soft metals and coatings: cast-aluminium and die-cast zinc heads, a scratched non-stick coating on the Zyliss Susi 3 in ATK's test, aluminium darkening in a dishwasher, and Zyliss's own page contradicting itself on dishwasher safety; kept as `head_material` and a judgement. (b) Parts count: extra ricer disks and a six-part garlic cube; kept in `interchangeable_disks` and `cleaning_access`. (c) What falls off first: the built-in cleaning pin on the Susi 3 in two reviews; kept as a medium judgement. (d) Plastic heads melting on a hot pot base; kept in `head_material`. Dropped: hinge and arm looseness on presses (ATK on the two OXO presses, qualitative, never on a listing); marble chipping as a facet (TechGearLab only, kept as a note in `mortar_material`); plastic suribachi wearing out sooner (Wikipedia only); a squeeze-force metric (no published measurements).

## Sources and what each supports

### Specialist reviewers who test (America's Test Kitchen pages read as raw HTML)

**ATK, The Best Potato Ricers, Miye Bromberg, 11 June 2025** https://www.americastestkitchen.com/equipment_reviews/1389-potato-ricers
- verbatim: "All the models came with a hopper or disk featuring small perforations that measured about 1.5 mm in diameter—the best size for ricing potatoes to a fine puree."
- verbatim: "The more perforations of that size that a hopper has within a given surface area, the more exit points there are for the potatoes—and the faster and easier it is to evenly extrude all the potatoes. Our top model has 459 perforations—the second-most of any we tested"
- verbatim: "Adding perforations to the sides of the hopper can reduce the force required to rice the potatoes."
- verbatim: "when we surveyed all the cooks in the test kitchen, only one said he'd ever used a disk other than the fine one. We think a single disk or setting for “fine” ricing is all you're likely to need."
- verbatim: "These mashed potatoes rival those made in a food mill , but the potato ricer is less expensive than the food mill and takes up a lot less space in your kitchen."
- verbatim (Norpro 162): "because none of its parts are removable, it can be a bit tricky to clean." Hole counts across models: 169, 169, 201, 347, 459, 569.
- verbatim (OXO 26981): "It also acquired stains from sitting against the perforations after washing". This is a stainless model (Amazon Material row: Stainless Steel), so it is not used for the aluminium claim.
- Supports: hole size, hole count versus force, disks rarely used, ricer versus food mill on cost and space, cleaning. Confidence high.

**ATK, The Best Potato Mashers, published 17 October 2026** https://www.americastestkitchen.com/equipment_reviews/1713-potato-mashers
- verbatim: "We vastly preferred potato mashers with heads made from perforated disks to those made from wavy wires. Perforated disks mashed our potatoes, beans, and ground beef to a smoother, finer consistency than wavy wires and did so more quickly too."
- verbatim: "Wavy Heads: This style of masher never got our potatoes even close to a smooth texture. The gaps between the waves of most mashers left large chunks of untouched potato"
- verbatim: "Plastic Heads: Some mashers had heads made from nylon or other plastics, which can melt if exposed to high heat (say, the bottom of a pot left over a burner.)"
- verbatim thresholds (single source, kept out of the JSON): "at least 50 holes measuring about 0.25 inches in diameter"; "an area of at least 9 square inches"; "at least 10 inches tall"; handles "at least 5 inches long". Note the winner itself has 45 holes and 8 sq in, below the page's own thresholds.
- Supports: plate versus wire, plastic heads melt, hole counts per masher (37 to 73). Confidence high for the direction, medium as a cross-reviewer finding because Serious Eats could not be read (see "Could not reach").

**ATK, The Best Food Mills, 6 November 2024** https://www.americastestkitchen.com/equipment_reviews/1705-food-mills
- verbatim: "a spring-loaded blade presses the fruit or vegetable through a perforated disk, extruding a smooth, dense puree and leaving most skins and seeds behind in the hopper. Because you often don't need to peel or seed produce before milling, the food mill can save a lot of time and effort."
- verbatim: "Each mill we tested has at least three interchangeable disks"; "While the perforation sizes and patterns varied from model to model, puree quality was not an issue."; "The main factor in determining speed was the force exerted by a small spring at the center of the mill."
- verbatim: Cuisipro Deluxe "Disks : 3 (2 mm, 3 mm, 4 mm) Weight : 2.6 lb Capacity : 12 cups"; Küchenprofi "Disks : 4 (1 mm, 2 mm, 3 mm, 5 mm) ... Capacity : 8 cups".
- verbatim: "Models with legs that extended out from the bottom of the hopper were the most stable ... Models that hooked onto the sides of the cooking vessel tended to wobble" (single source; trimmed, see below).
- Supports: food mill strains skins and seeds, disks, spring not holes sets speed (so hole count is irrelevant on the mill branch), measured capacity. Confidence high.

**Cook's Illustrated (ATK), Which Mashing Method is Best?** https://www.americastestkitchen.com/cooksillustrated/how_tos/5808-which-mashing-method-is-best
- verbatim: "For the fluffiest mashed potatoes, the goal is to use the gentlest touch possible to avoid bursting the potatoes' swollen starch granules. Once released, the sticky gel inside will turn the mash gluey. We've found that the vigorous action of a food processor guarantees glueyness"
- verbatim: "While tasters found the riced potatoes a tad fluffier, they were also a bit grainy. The potatoes that passed under the food mill's sweeping blade were almost as fluffy and boasted a far smoother texture."
- verbatim: "Electric Mixer: A real no-no, unless you like gluey potatoes Masher: Even with a good workout, some lumps are inevitable."
- Supports: the potato half of the mechanism and the texture order. Confidence high.

**ATK, The Best Mortars and Pestles and Molcajetes, 24 June 2025** https://www.americastestkitchen.com/equipment_reviews/1348-mortars-and-pestles
- verbatim: "we recommend looking for medium-size models with a capacity of 1.5 to 4 cups"; "avoid mortars that hold less than 1.5 cups".
- verbatim: "We liked mortars that weighed at least 5 pounds"; "Mortars weighing less than 5 pounds were too light; we had to hold them down with our nondominant hand while working to keep them from tipping over"; "Pestles weighing about 1 pound did more of the work".
- verbatim: "Models made from coarse-textured materials, including granite, volcanic rock, and marble, are more abrasive than mortars made from metal or stoneware"; "Smooth metal and stoneware mortars provide no traction, so whole spices skittered around their interiors"; "Coarser mortars do have to be seasoned prior to use to rid them of any gritty particles".
- verbatim: "Instead of simply cutting ingredients into tiny pieces as a food processor might do, the pestle crushes them, extracting more of the aromatic oils and flavor compounds in the process."
- verbatim: "we washed every model with soap and water between each test and didn't notice any off flavors".
- verbatim: Cilio Goliath "It held 3 cups of food" in the intro and "Capacity: 4 cups ... Mortar Weight: 10 lb, 12⅝ oz" in the spec line (the page disagrees with itself).
- Note: ATK groups marble with the coarse materials; Reviewed and TechGearLab call marble smoother than granite. The register puts marble in the middle "textured" value.
- Supports: capacity bands, weight floor, interior roughness, seasoning, soap, molcajete form. Confidence high.

**ATK, The Best Garlic Presses, 24 June 2025** https://www.americastestkitchen.com/equipment_reviews/1857-garlic-presses
- verbatim (Leifheit rocker): "we often had to use our full body weight to get a good mince, which strained our shoulders and wrists."
- verbatim (Zyliss Susi 3): "the built-in cleaning tool—which latched onto the handle—was easy to knock out during use, so we kept having to retrieve and reinsert it. This model's coating also showed scratches from the paring knife".
- verbatim: "Swing-Out Hopper: Some presses had a hopper that swung out from the hinge, making it easy to remove clove remnants"; "Plungers that were covered in nubs trapped garlic and were difficult to clean."
- verbatim (garlic cube): "it had six parts"; "A cube-shaped model had a push-down lid that put users' fingers at risk; it once snapped shut on a tester's finger and drew blood."
- verbatim (ORBLUE, sold as "Mincer, Crusher & Peeler"): "with unpeeled cloves, a lot of garlic was left behind in the hopper".
- Supports: press style, cleaning, unpeeled claims versus performance. Confidence high.

**Reviewed (USA Today network), The Best Mortars and Pestles, Lindsay D. Mattison** https://www.reviewed.com/cooking/best-right-now/the-best-mortars-and-pestles
- verbatim: Cilio Goliath "weighing in at a massive 14.5 pounds"; "it can hold about three cups of liquid".
- verbatim: ChefSofi "a still-impressive 7 pounds"; "the two-cup capacity of the bowl is more than sufficient to make a two-avocado guacamole or a four-serving batch of pesto."
- verbatim: "If you cook for a crowd and you want plenty of extra space, go for a mortar with a four-cup capacity."
- verbatim: "the granite models seemed to do the best job at gripping and ripping"; "the marble and stone mortars and pestles didn't have a rough enough interior to fully pulverize fresh and dried chilies into a paste."
- verbatim: seasoning is "grinding up white rice to remove any granite bits".
- Supports: the 2-cup and 4-cup bands, weights, granite best. Confidence high.

**Reviewed, The Best Garlic Presses, updated 25 July 2025** https://www.reviewed.com/cooking/best-right-now/best-garlic-presses
- verbatim (Kebley rocker): "Cons Hard to clean Can require significant force to operate".
- verbatim (Susi 3): "Cons Cleaning tool may fall out during use".
- verbatim: "Stainless steel is by far the most common material for garlic presses: It's durable, it's long-lasting, and, most importantly, it doesn't retain that distinctive garlic odor"; plastic presses "are prone to hold onto persistent odors".
- verbatim (OXO Soft-Handled): "thanks to its sturdy die-cast zinc build, this garlic press can go right in the dishwasher".
- Inconsistency kept: the page calls the Kuhn Rikon Epicurean "made of cast aluminum" in one card and "made of stainless steel" in the next. Kuhn Rikon's retailer copy, Amazon and ATK all say stainless.
- Supports: rocker force and cleaning, cleaner falls out (second source), material odour. Confidence medium-high.

**TechGearLab, The Best Mortar and Pestle Set** https://www.techgearlab.com/topics/kitchen/best-mortar-and-pestle
- verbatim: "The textured granite makes for great friction when it comes to grinding"; Tera stainless "Too smooth for grinding"; Sagler marble 1/3 cup "Not efficient".
- verbatim: Cooler Kitchen porcelain "offering only about a cup of capacity, making it less of a candidate for something like guacamole or pesto" and "added some texture to the interior, which made this surprisingly effective at crushing and grinding for a porcelain model."
- verbatim: "Granite sets generally require a bit more maintenance — they need to be pre-seasoned and washed without soap"; Laevo marble "we did experience some small chipping".
- verbatim: Vasconia 4-cup molcajete "weighs almost eight pounds ... it's more of a pain to wash".
- Supports: rough beats smooth, small mortars, soap disagreement, weight cost. Confidence high.

### Primers (Wikipedia plain-text API)

**Potato ricer** https://en.wikipedia.org/wiki/Potato_ricer
- verbatim: "forcing it through a sheet of small holes, which are typically about the diameter of a grain of rice"; "the potatoes are smashed only once. With this method, the cell walls are much less likely to break open."; potatoes "become glutinous ... when over-processed."
- Second source for the potato mechanism. Confidence high.

**Garlic press** https://en.wikipedia.org/wiki/Garlic_press
- verbatim: "a clove of garlic can be passed through a sturdy press without even removing its peel. The peel remains in the press while the garlic is extruded out."
- verbatim: "Garlic crushed by a press is generally believed to have a different flavor from minced garlic, because more of garlic's strong flavor compounds are liberated."; Cook's Illustrated: "a good garlic press can break down cloves more finely and evenly than an average cook using a knife"; Bourdain called presses "abominations"; Elizabeth David's essay "Garlic Presses are Utterly Useless".

**Molcajete** https://en.wikipedia.org/wiki/Molcajete
- verbatim: "Although true molcajetes are made of basalt, imitations are sometimes made of a mixture of pressed concrete and volcanic rock particles."
- verbatim: "A new basalt molcajete needs to be 'broken in' ... repeatedly grind uncooked white rice"; "As the porous basalt is impossible to fully clean and sanitize, molcajetes are known to 'season'".

**Dishwasher** https://en.wikipedia.org/wiki/Dishwasher
- verbatim: "Aluminium, brass, and copper items will discolor, and light aluminum containers will mark other items they knock into."

**Suribachi** https://en.wikipedia.org/wiki/Suribachi
- verbatim: "a pottery bowl, glazed on the outside and with a rough pattern called kushi-no-me on the unglazed inside"; "The surikogi pestle is made from wood to avoid excessive wear"; "plastic versions ... have a much shorter life-span."

**Food mill** https://en.wikipedia.org/wiki/Food_mill
- verbatim: "interchangeable plates with different hole sizes may be supplied"; "Uses of a food mill include removing the seeds from cooked tomatoes"; "usually made of stainless steel or aluminum".

**Mortar and pestle** (read, not cited) https://en.wikipedia.org/wiki/Mortar_and_pestle: pharmacy mortars are glazed porcelain ("Wedgwood"); granite in Southeast Asia and India. Background only.

### Standards and regulators (food contact)

**Council of Europe / EDQM, Metals and alloys used in food contact materials and articles, 1st edition (2013), PDF read with pdftotext** https://www.icoval.org/files/documentacion/documentos/metalls_alloys.pdf
- verbatim, Table 1: "Al Aluminium 5" and "Zn Zinc 5" (SRL, mg/kg food).
- verbatim: "Storage of acidic (e.g. fruit juices), alkaline (e.g. lye dough products) or salty, liquid foodstuffs in uncoated aluminium utensils should be limited in order to minimise release." and "The producer should provide specific labelling of uncoated aluminium for users."
- verbatim: "Zinc is easily dissolved in dilute acids and by bases".
- The Euro Inox summary of the same guide (https://www.edelstahl-rostfrei.de/fileadmin/user_upload/ISER/downloads/CoE-Guidelines_EN.pdf), verbatim: "The Council of Europe resolution is not legally binding for member states, but serves as a reference for the implementation of Article 3 paragraph 1 of Regulation (EC) No. 1935/2004." and, on stainless, "No particular health concerns have been raised".
- The 2nd edition (2024) draft on edqm.eu returned an HTML bot check, not the PDF. EUR-Lex (Regulation 1935/2004 itself) returned an empty response, as on earlier runs.

**Sander, Kappenstein, Ebner et al. (BfR), PLoS ONE 2018, PMC6056035** https://pmc.ncbi.nlm.nih.gov/articles/PMC6056035/
- verbatim: "The SRL is set at 5 mg aluminium /kg food"; "Heating up to 72°C core temperature of the food for two minutes caused significant releases of aluminium up to 5 mg/kg for tomato puree, 10 mg/kg for sauerkraut juice"; "The frequent use of such kitchenware for the preparation of acidic food should be avoided."
- Second source for the 5 mg/kg aluminium figure. The zinc 5 mg/kg figure has only the CoE source, so it stays here and not in the JSON.

**US FDA, 21 CFR 174.5 (eCFR renderer API)** https://www.ecfr.gov/current/title-21/chapter-I/subchapter-B/part-174/section-174.5
- verbatim: "Any substance used as a component of articles that contact food shall be of a purity suitable for its intended use."
- Supports: the US general rule. There is no US release limit for metal utensils comparable to the CoE table. Confidence high on what it says, low on its usefulness for picking a tool.

### Brand spec sheets and seller pages

**Zyliss, Susi 3 Garlic Press** https://us.zyliss.com/products/zyliss-susi-3-garlic-press
- verbatim: "Material(s): Aluminium, Polypropylene (PP), Ilaflon Coating"; "Sturdy, cast aluminum body"; care instructions "Dishwasher Safe"; FAQ: "Is the garlic press dishwasher safe? No, hand washing is recommended to maintain the longevity of the aluminium construction."; "No need to peel cloves beforehand".
- The page contradicts itself; the kitchen-standard `dishwasher_safe` trap comment does not cover this pattern. Confidence high.

**Kuhn Rikon Epicurean, via Kitchen Conservatory** https://www.kitchenconservatory.com/Tools/Garlic-Tools/Epicurean-Garlic-Press-in-Stainless (kuhnrikon.com returned a bot check)
- verbatim: "Made by Kuhn Rikon of stainless steel. Dishwasher safe. Air drying after washing is recommended."; "Stainless steel sieve swings out for easy cleanup."

**ImportFood, Solid Thai Granite Mortar and Pestle** https://importfood.com/products/thai-cookware/item/mortar-and-pestle-solid-thai-granite
- verbatim: "6\" has 1.5 cup capacity, weighs 7 lbs."; "7\" has 2+ cup capacity, weighs 12 lbs."; "8\" has 3+ cup capacity, weighs 15 lbs."; "9\" ... has 5-6 cup capacity, weighs 24 lbs"; the 6-inch "should meet the needs of most home users".
- verbatim: "They cut, while it bruises and mashes, releasing the oils and flavours"; "The ones designed to make som tam ... are made from unglazed, fired clay ... The pestle is always made of wood".
- verbatim: "The pestle and inside surface of the mortar are polished smooth" (and ATK still rated it a Best Buy, so "polished" alone does not mean it fails; the JSON no longer lists polished stone as a failing material).

**Masienda, Molcajete** https://masienda.com/products/molcajete
- verbatim: "Weight: 20 lbs" (original) and "Weight: 8 lbs" (Chico); "grind the rice into a rough powder ... repeat the process until the resulting rice powder maintains its original color with no tint of gray." ATK weighed the same molcajete at 15 lb 11½ oz.

### Retail listings (what data is already present)

Amazon pages were read as raw HTML with curl (ASINs taken from ATK's buy links). Product-information rows and bullets, verbatim:

| Listing | Rows and bullets that matter |
|---|---|
| Cilio Goliath mortar B00133Q3S6 | Material Granite; Finish Type Glossy; Capacity 8 ounces; Item Weight 11 pounds; Item Diameter 5.25 inches |
| Cole & Mason 5.5" granite B00EAU5VQE | Material Stone; Finish Type Unfinished; Capacity 2 cups; bullet "holds a 3 cup (20oz.) capacity"; "season mortar with white rice" |
| Cuitáxi molcajete B09MTC2LQ2 | Material Lava Stone; Item Weight 4.53 kg; title "Made of Authentic Lava Rock in Mexico | Not Concrete" |
| Cuisipro Deluxe Food Mill B000G18A3W | Material Stainless Steel; Capacity 6 quarts; "3 interchangeable disks in 2mm, 3mm, and 4mm sizes" |
| Norpro 162 ricer B0000VLUM6 | Material Tin; Handle Material Tin; bullets "Capacity: 2.5C / 20OZ", "Shiny triple chromed finish", "Hand washing recommended" |
| OXO Good Grips ricer B00004OCJQ | Material Stainless Steel; Handle Material Silicone; "Dishwasher safe" |
| OXO Good Grips garlic press B00HEZ888K | Material Zinc; "Die-Cast Zinc"; "Flip the top handle over to use the built-in cleaner" |
| Kuhn Rikon Epicurean B0000CD0HX | Material Stainless Steel; "Press Peeled or Unpeeled"; "pull-out sieve" |
| OXO Good Grips smooth masher B00004OCL9 | Material Stainless Steel; "Grid Pattern for Smooth Mashed Potatoes" |
| Chef Craft masher B002JG8ETC | Material Stainless Steel; title "10.25 inch"; "Great to use on stainless steel and cast iron pans" |

Search-result titles (queries: nylon potato masher nonstick, potato ricer, garlic press, mortar and pestle, molcajete, potato masher; first 12 titles each) show: tool noun in every title; ricer hopper size in ounces in 5 of 10 ricer titles (15oz four times, 22 oz once); mortar capacity in cups in 6 of 12 mortar titles; "Cast Aluminum" or "Stainless Steel" in half of the 12 press titles; "Rocker" in rocker titles; "for Non-Stick Cookware" on nylon mashers; two basalt molcajete titles adding "Not Concrete"; and granite sets titled "Molcajete" (PriorityChef "Granite ... Molcajete Mexicano", B07DFG7BT3 "Granite Mortar and Pestle Set 8-Inch (3.5-Cup), Molcajete", B0DDWLYRN3 "Unpolished Granite ... Molcajete Mexicano").

What this tells the scraper: Material rows are present on every Amazon page read. Item Weight is not: the Cole & Mason page (B00EAU5VQE) has no Item Weight row in my fetch, only "8 pounds" in the title against ATK's measured 5 lb 11 oz, so treat weight as present on some detail pages and weak in titles (1 of 77 mortar titles). Capacity is present on mortars and food mills but wrong or self-contradictory on three of four checked (Cilio 8 oz versus 3 to 4 cups measured; Cole & Mason 2 versus 3 cups on one page; Cuisipro 6 quarts versus 12 cups measured). Hole counts, interior roughness and press style are not fields.

### Could not reach
- Serious Eats (People Inc): curl returned an access-denied page and the search tool refuses the domain. Its masher, ricer and mortar reviews are therefore not cited; the search summary said Serious Eats agrees that plate beats wire, but that is a search excerpt and was not used.
- Wirecutter (NYT): JavaScript wall. Food Network: Akamai 403. Hot Thai Kitchen: Cloudflare challenge. OXO and Kuhn Rikon brand sites: bot checks.
- EDQM 2nd edition PDF and EUR-Lex 1935/2004: blocked or empty.
- The WebSearch budget for the session ran out (200 of 200) partway through; the rest of the research used curl on known or linked URLs.

## Data points used for derived metrics

No derived metric survived. The obvious candidate, converting a mortar's advertised width in inches to capacity in cups, fails on the data: the same "about 6 inch" class ranges from 1.5 cups (ImportFood 6 in, measured 1.6 by ATK) to 2.6 cups (Cole & Mason 5.5 in, ATK), and Cilio's 5.25 in diameter holds 4 cups (ATK). The table below is kept because it backs the capacity judgement and the weight claim.

| Mortar | Listed or brand capacity | Measured capacity (ATK unless noted) | Weight (source) |
|---|---|---|---|
| Cilio Goliath granite | 8 oz (Amazon) | 4 cups spec, "held 3 cups" text; about 3 cups (Reviewed) | 10 lb 12⅝ oz (ATK); 14.5 lb (Reviewed); 11 lb (Amazon) |
| Cole & Mason 5.5 in granite | 2 cups row, 3 cups bullet (Amazon) | 2.6 cups | 5 lb 11 oz (ATK); "8 pounds" in Amazon title |
| ImportFood 6 in Thai granite | 1.5 cups (ImportFood) | 1.6 cups | 5 lb 6 oz (ATK); 7 lb (ImportFood) |
| Masienda molcajete basalt | not stated | 4 cups | 15 lb 11½ oz (ATK); 20 lb (Masienda) |
| IMUSA 8 in granite molcajete | not read | 3.5 cups | 7 lb 14 oz (ATK) |
| Vasconia 4-cup molcajete | 4 cups (TechGearLab) | not measured | almost 8 lb (TechGearLab) |
| ChefSofi granite | 2 cups (title) | 2 cups (Reviewed, TechGearLab) | 7 lb (Reviewed) |
| Juvale stainless | n/a | 0.9 cups | 1 lb (ATK) |
| RSVP marble | n/a | 0.9 cups | 1 lb 9 oz (ATK) |
| Stoneware (PG4050-12) | n/a | 1.25 cups | 1 lb (ATK) |

Weight claim in the JSON ("every stone mortar that America's Test Kitchen and Reviewed recommended weighed 5 lb or more, and the largest exceed 14 lb"): ATK's recommended stone mortars run 5 lb 6 oz to 15 lb 11½ oz; Reviewed's two picks are 7 and 14.5 lb.

Ricer hopper sizes for `hopper_capacity`: Norpro 2.5 cups / 20 oz (Amazon bullet), PriorityChef and Zulay 15oz, Sopito 22 oz (titles). Food mills: Cuisipro 12 cups and Küchenprofi 8 cups (ATK).

## Numbers in category.json and the sources behind each

| Number | Where | Sources |
|---|---|---|
| 1.5 cup everyday floor | mortar_size | ATK (1.5 to 4 cups; avoid under 1.5); ImportFood (6 in, 1.5 cups, meets most home users) |
| about 2 cups for guacamole or pesto | mortar_size | Reviewed (two-avocado guacamole, four-serving pesto); TechGearLab (two cups, pesto batches) |
| 4 cups for a crowd | mortar_size, split 3 values | Reviewed; ATK (1.5 to 4 range, bigger for a crowd); TechGearLab (4-cup "Great for Big Batches") |
| 5 lb or more; over 14 lb | mortar_weight | ATK specs; Reviewed weights |
| 15 to 22 oz ricer hoppers | hopper_capacity | Amazon titles (PriorityChef, Zulay 15oz; Sopito 22 oz) and Norpro bullet (20 oz) |
| 5 mg/kg aluminium | judgement 8 | CoE technical guide; BfR PLoS ONE 2018 |
| Cilio 8 oz / 4 cups / about 3; Cole & Mason 2 / 3 cups; Cuisipro 6 qt / 12 cups | judgement 1 | each figure read on its own page; they are the observations that make the judgement, not thresholds |

Numbers deliberately kept out of the JSON as single-source: ricer hole size 1.5 mm and counts 169 to 569 (ATK); masher thresholds (ATK); zinc 5 mg/kg (CoE only).

## Against the existing add-on schema

Add-on read: `supplies/kitchen/add-ons/masher.jsonc` is empty (`{}`). The leaf composes kitchen-standard (`materials` with roles including `working_head`, `mesh_screen`, `body`; `max_heat_temp`; `dishwasher_safe`) and supplies-standard (`high_signal_markdown`; `sizing` with `capacity` and `weight`). Proposals only; nothing in zamana-expertise was edited.

| Facet in this register | In schema today | Proposed kind under AGENTS.md | Note |
|---|---|---|---|
| tool_family | no | hard filter | Proposal: `masher_tool_type: enum [masher, ricer, food_mill, garlic_press, garlic_rocker, mortar_and_pestle, other]`. The leaf bundles five tools with no shared criteria, so every row needs it. `garlic_rocker` carries press_style. |
| mash_texture | no | derived at query time | From tool type plus masher head (a photo fact). Not a stored field. |
| mortar_size | partly, `sizing.capacity` | category quality currency | The capacity enum has no `cup`, the unit mortars are sold in; propose adding `cup` or normalising to ml. Commit to augmenting: listed values are wrong often enough that the row should hold a reviewed figure for the top N. |
| mortar_interior | no | high_signal_markdown | "Unpolished interior" is a text claim; a photo half-answers it. Not a column. |
| mortar_material | partly, `materials[role=body]` | vertical anchor (already) | The materials enum has `marble` and `stone_other` but no `granite` or `basalt`; granite and basalt are the two materials experts rank first. Propose adding both to kitchen-standard, or a mortar-only enum. |
| mortar_weight | yes, `sizing.weight` | category quality currency | Dense on Amazon (Item Weight row). Keep. |
| head_material | yes, `materials[role=working_head]` | vertical anchor (already) | Fine as is. Prefer title words over the Material row (Norpro chromed ricer listed as Tin). |
| dishwasher_safe (folded into head_material here) | yes | keep | Extend the trap comment: a brand page can say dishwasher safe in the spec block and "hand washing is recommended" in its FAQ (Zyliss Susi 3). For aluminium heads read the FAQ and care text. |
| max_heat_temp | yes | keep | Matters for nylon and silicone masher heads, which can melt on a hot pot base. |
| masher_head | no | none (visual) | A photo answers wire versus plate; AGENTS.md rule 1 keeps it out. |
| perforation_count | no | none | A count a photo or a review answers; rule 5 cuts it. Keep only as a research annotation for the top N. |
| hopper_capacity | partly, `sizing.capacity` | sizing | Ricer titles state ounces; food mills quarts. Same unit issue as mortars. |
| interchangeable_disks | partly, `sizing.pieces_included_text` | high_signal_markdown | Disk sizes in mm belong in the markdown line. |
| press_style | no | folded into the tool-type enum | See tool_family. |
| unpeeled_cloves | no | high_signal_markdown | A claim, often overstated (ATK, ORBLUE). Rule 5 cuts curated claim vocabularies. |
| cleaning_access | no | high_signal_markdown | Swing-out sieve versus built-in cleaner is construction detail. |

Fields this register considers noise for the leaf: `sizing.dimensions` length and width (eyeballable for all five tools; the one sourced height rule, a masher at least 10 in tall, is single-source); `sizing.pack_count` (two-masher sets exist but carry no expert signal); `max_heat_temp` for mortars and garlic presses (stone and metal; it only matters for plastic masher heads).

## Branches

Every `when` value is a value of the split facet; the question-to-value rule is in each split's rationale.

**Split 1, tool_family ("Potatoes, sauce, garlic, or spices and pastes?").** Rule: potatoes, vegetables or sauce → masher, ricer, food mill; garlic → garlic press; spices and pastes → mortar and pestle. Changed 2026-10-25: "sauce" added to the question, since food mills are bought mainly for tomato sauce, applesauce and canning (Wikipedia, ATK, Amazon titles), not potatoes.
- when masher, ricer, food mill: then_ask mash_texture, head_material.
  - mash_texture: "Texture is the only thing that separates the three potato tools" (split 2 rationale) and the facet's values map onto these three tools.
  - head_material: "Nylon heads are sold for nonstick pans, but testers warn they can melt on a hot pot base" and "Aluminium darkens in a dishwasher" (head_material, why experts care); mashers and ricers are sold in stainless, cast aluminium, tinned steel and nylon (listing table above).
  - makes_irrelevant mortar_size: defined on the mortar ("How much the mortar bowl holds"), so split 3 is skipped. The other mortar facets are only reached through the mortar branch, so they are not listed.
  - makes_irrelevant press_style, unpeeled_cloves: defined on garlic presses ("Whether a garlic press is...", "Whether the press handles cloves with the skin on").
- when garlic press: then_ask press_style, unpeeled_cloves, cleaning_access, head_material.
  - press_style: "Lever presses with long handles need the least force. Two test teams found rockers need hard pressure" (press_style).
  - unpeeled_cloves: "Skipping the peel saves most of a press's time" (unpeeled_cloves).
  - cleaning_access: "Cleanup is where presses and ricers fail in daily use" (cleaning_access).
  - head_material: the Susi 3 is cast aluminium with a coating that scratched (ATK) and a self-contradicting dishwasher claim (Zyliss); OXO's is die-cast zinc (Amazon).
  - makes_irrelevant mash_texture ("in cooked potato") and mortar_size ("the mortar bowl"), the two later split questions; and hopper_capacity ("a ricer's hopper or a food mill's bowl"), named because a press also has a "hopper" (vocabulary) and listings say "large chamber", which a scraper could misread as capacity.
- when mortar and pestle: then_ask mortar_size, mortar_interior, mortar_material, mortar_weight.
  - mortar_size: "Too small and food flies out before it is ground" (mortar_size); split 3 follows.
  - mortar_interior: "rough granite and volcanic rock grind spices and shred herbs fastest ... smooth steel and stoneware let seeds skitter" (mortar_interior).
  - mortar_material: "Material sets roughness, weight and care" (mortar_material), now including cast iron.
  - mortar_weight: changed 2026-10-25, moved here from split 3's party-sized leg. ATK's "mortars weighing less than 5 pounds were too light" is a general floor, not a party-only concern, so it is asked for every mortar and pestle, not gated on size.
  - makes_irrelevant mash_texture (split 2, "in cooked potato"); head_material, now resolved by the expanded `mortar_material` (its values include stainless steel and cast iron, so cast-iron mortars no longer fit neither facet); press_style and unpeeled_cloves, defined on garlic presses.

**Split 2, mash_texture ("Chunky, mostly smooth, fluffy, or silky with skins strained out?").** Rule: chunky → rustic and chunky; mostly smooth → smooth with a few lumps; fluffy → fluffy and lump-free; silky → silky, skins and seeds strained out. Changed 2026-10-25: chunky no longer implies wire-only. ATK's plate-head winner "is perfect for both smooth and chunky mashes", so a chunky answer should not filter out plate mashers; mash_texture is the finest texture a tool *can reach*, and the filter should be "at least the wanted texture", not an equality match.
- when rustic and chunky, smooth with a few lumps (mashers): makes_irrelevant hopper_capacity, interchangeable_disks, cleaning_access only.
  - head_material removed from then_ask 2026-10-25: it was already asked on this path at split 1 ("tool_family = masher, ricer, food mill" → then_ask mash_texture, head_material), so asking it again here was a duplicate.
  - masher_head is still not asked: it is an annotation-only facet (not a filter), settled by a photo, and is not fixed by this answer alone any more, since both wire and plate heads satisfy "chunky".
- when fluffy and lump-free (ricer): then_ask perforation_count, hopper_capacity, cleaning_access; makes_irrelevant masher_head.
  - perforation_count: "More holes of the right size mean less squeezing for the same mash" (perforation_count).
  - hopper_capacity: "It sets how many refills a batch takes" (hopper_capacity).
  - cleaning_access: "fixed ricers with no removable parts are tricky to clean" (cleaning_access).
  - masher_head: defined on mashers.
- when silky, skins and seeds strained out (food mill): then_ask hopper_capacity, interchangeable_disks; makes_irrelevant masher_head, perforation_count.
  - hopper_capacity: food mills "hold several times that and suit big batches" (hopper_capacity).
  - interchangeable_disks: "On a food mill the disks are the point" (interchangeable_disks).
  - perforation_count irrelevant: ATK food mills, verbatim above: "While the perforation sizes and patterns varied from model to model, puree quality was not an issue" and speed is set by the spring; the facet is defined on ricers and mashers.

**Split 3, mortar_size ("A pinch of spices, guacamole for two, or party batches?").** Rule: pinch of spices and guacamole for two both → everyday, 1.5 to under 4 cups; party batches → party-sized, 4 cups or more. Changed 2026-10-25: the "pinch of spices" answer no longer maps to the smallest band, renamed "small, under 1.5 cups". The mortars under 1.5 cups in the data were mostly the smooth ones that failed spices (ATK Juvale steel 0.9, RSVP marble 0.9, stoneware 1.25; TechGearLab Tera steel 1 cup, Sagler marble 1/3 cup, Cooler Kitchen porcelain 1 cup), and Reviewed puts the real floor nearer 2 cups ("anything smaller than that can likely only handle crushing small amounts"), so a spice buyer is steered to the everyday band instead.
- Both branches removed 2026-10-25 (were WEAK in the blind check):
  - spice-sized → mortar_interior was a no-op, since split 1 already asks `mortar_interior` on every mortar path, and its chain ("every mortar under 1.5 cups was smooth") is contradicted by the register's own values (marble is classed textured, and a textured porcelain under 1.5 cups worked).
  - party-sized → mortar_material, mortar_weight was also a no-op for `mortar_material` (now asked on every mortar at split 1) and rested on a chain the data contradicts (4- to 6.5-cup granite bowls that are not molcajetes exist). `mortar_weight` is kept, but moved to split 1's mortar branch as a universal ask, since ATK's 5 lb floor is general, not party-only.
- No branch left for any value: all three downstream mortar facets (`mortar_interior`, `mortar_material`, `mortar_weight`) are now asked for every mortar and pestle at split 1, so split 3 adds nothing further.

Generic test on split 3: a size question exists in many categories, but the bands and their dish anchors (guacamole for two, a party) are mortar-specific and sourced. Kept.

## Trimmed detail

- **mechanism.plain**: "Whether that is the goal or the defect is what separates the families" was merged into one sentence; "again and again" became "repeatedly"; "release more metal into acidic food than stainless steel does" shortened. No fact cut.
- **families[ricer]**: "about 1.5 mm holes" replaced by "rice-grain-sized holes", because 1.5 mm is ATK only. The 1.5 mm figure lives above.
- **mortar_size.why_experts_care**: "Testers advise against mortars under about 1.5 cups except for small spice jobs; about 2 cups makes a two-avocado guacamole or four servings of pesto" reworded to the two-source version; the two-avocado detail is Reviewed's (above).
- **perforation_count**: "the second-most holes" (459, ATK) became "one of the highest hole counts". Hole counts 169 to 569 above.
- **interchangeable_disks**: "only one of ATK's cooks" became "almost none"; "at least three disks" became "several". Both exact figures are ATK only (above).
- **food mills, legs versus hooks**: ATK found legs more stable than pot hooks and medium spring tension best. Single source; not a facet.
- **masher thresholds**: at least 50 holes of about 0.25 in, a 9 sq in head, 10 in tall, 5 in handles (ATK only; the winner itself misses two).
- **pestle weight**: about 1 lb pestles do more of the work (ATK only).
- **marble chipping**: TechGearLab saw small chips on a marble set; ImportFood says its granite "does not chip". Kept only as a clause in mortar_material.

## Corrections

1. (Self-check, before any blind check) Removed "polished stone" from the failing interiors: ImportFood's Thai granite is polished smooth inside and ATK still rated it a Best Buy.
2. Dropped ATK's OXO ricer stain as support for the aluminium dishwasher judgement: that ricer is stainless.
3. Restated the Susi 3 cleaner claim: ATK saw the cleaner knocked out repeatedly; Reviewed lists "may fall out" as a con. The earlier wording ("kept falling out in two separate tests") overstated Reviewed.
4. Removed `masher_head` from the then_ask of the masher branch on split 2: the answer already fixes it.
5. `vertical` is set to `other`, not `supplies` as the brief asked: `schema.json` allows only formulations, machines, fashion, furnishing and other, and validate.py rejects `supplies` ("'supplies' is not one of [...]"). The skill's files were not edited. Change the field to `supplies` once the schema enum includes it; nothing else depends on it (the catalog takes the vertical from the taxonomy path).
6. Cut the makes_irrelevant lists on split 1 to the later split questions and the facets an assistant could plausibly still ask (from 6, 9 and 9 facets to 3, 3 and 4). The longer lists were true but crowded the decision tree; facets reached only through another branch need no skip.

## Corrections after the blind check (2026-10-25)

1. `mash_texture` split rationale and facet `why_experts_care` rewritten: chunky no longer maps to wire masher only. ATK's plate-head winner "is perfect for both smooth and chunky mashes," so chunky now accepts any masher; the equality mapping was filtering chunky buyers away from the better tool.
2. Removed the duplicate `head_material` ask from split 2's "rustic and chunky, smooth with a few lumps" branch; it was already asked on this path at split 1.
3. `mortar_size` value "spice-sized, under 1.5 cups" renamed "small, under 1.5 cups," and the split's rationale and facet `why_experts_care` rewritten: dropped "testers agree" (ATK says 1.5 cups, Reviewed says 2, ImportFood is a seller not a tester). The "pinch of spices" cold-start answer now routes to the everyday band (1.5 to under 4 cups), since the tested mortars under 1.5 cups mostly failed spice jobs.
4. Resolved the citrus hand press ownership cycle on this side: this register no longer claims or defers them circularly; notes.md now states plainly that citrus presses belong to `cocktail-mixing-tools`, since they share no facet here. The matching fix in `cocktail-mixing-tools/notes.md` is not made, since this task is scoped to files under `mashers/` only.
5. Reconciled the push-down-chopper boundary in notes.md: hole-based push-down cube presses stay here as `press_style` = "push-down chopper"; bladed push-down choppers (Chef'n GarlicZoom and similar) go to `knives`/`graters`. No facet value changed.
6. Added explicit exclusion rules to `tool_family`'s `backfill_note` and to notes.md: bladed meat choppers (flagged as an unclaimed tail, like tortilla presses), pill crushers and lab mortars, and electric tomato strainers, mixer attachments and grain mills.
7. `tool_family`'s question and rationale changed to "Potatoes, sauce, garlic, or spices and pastes?" so a sauce buyer has an answer; `skip_if_known`'s dish-answer entry now notes that guacamole fits masher or mortar and garlic paste fits press or mortar, so an ambiguous dish alone does not decide.
8. Judgement 8 (safety) note rewritten: "caps" became "sets a non-binding guideline"; named the 2013 (1st) edition and noted an unread 2024 (2nd) edition; added BfR's actual test conditions (hot acidic trays, 2 hours, up to six times the limit); marked the zinc figure as CoE-only; added that lead in glazed ceramic mortars is not covered.
9. Judgement 1 (listed capacity often wrong) note rewritten to quote both of ATK's self-contradicting Goliath figures ("held 3 cups" and "Capacity: 4 cups") and to flag that Cole & Mason's own bullet is self-contradictory (20 oz is 2.5 cups, not 3; ATK measured 2.6).
10. `mortar_interior`'s `why_experts_care` rewritten and its "rough" value relabelled without "unpolished": only ATK calls stoneware smooth (Reviewed calls it "lightly rough"), and polished granite (ImportFood, Tera) still grinds well, so polish no longer excludes a stone from the rough band. Judgement 4 reworded to match and `sources_disagree` set to true.
11. `perforation_count`'s `why_experts_care` rewritten: ATK's easiest ricer owed its ease to 459 holes *and* geared hinges, and the most-holed model (569 holes) was not the easiest, so the facet no longer credits holes alone.
12. `unpeeled_cloves`'s `why_experts_care` rewritten: the press ATK faulted is not "sold as a peeler," it bundles a separate silicone peeler accessory; the failed claim is its own "leaves the peel inside" bullet. Added a backfill rule that "Peeler" in a title is not proof of unpeeled handling.
13. `mortar_weight`'s `why_experts_care` rewritten: the 5 lb floor is ATK's; Reviewed only states weight for 2 of its 5 picks, so it can confirm but not independently set the floor.
14. Moved `mortar_weight` onto `tool_family = mortar and pestle` as a universal ask (ATK's 5 lb floor is general, not party-only), and dropped both of split 3's WEAK branches (spice-sized → `mortar_interior`, party-sized → `mortar_material`/`mortar_weight`), since `mortar_interior`, `mortar_material` and `mortar_weight` are now all asked once at split 1 and the split-3 legs were no-ops. See "Branches" above for the full trace.
15. Added `tinned steel` to `head_material`'s values and narrowed the "Tin row is wrong" note to Norpro specifically; tinned steel is real on Louis Tellier mills and ricers.
16. Added `cast iron` to `mortar_material`'s values, and added `cast iron` (to the rough band) and `wood` (to the textured band) in `mortar_interior`'s values, so cast-iron mortars (Sungmor) now fit a facet.
17. Changed `hopper_capacity` and `interchangeable_disks` from `researched` to `listing`: ricer ounces (25 of 59 titles), food-mill quarts and disc counts (30 of 50 mill titles, 11 of 59 ricer titles) are title- and bullet-carried, not research-only.
18. Added a note to `mash_texture`'s `backfill_note` that it depends on the researched `masher_head` field, so its accuracy for mashers is bounded by that backfill.
19. Fixed `press_style`'s `cold_start_question` to name its third value: "Squeeze-handle, rocker, or push-down chopper?" (it previously omitted push-down chopper entirely).
20. Updated the redirected ATK mortars reference URL from `.../1348-mortars-and-pestles` to `.../2140-mortars-and-pestles`.
21. Corrected notes.md's scraper-data claim that "Item Weight rows are present on every Amazon page read": the Cole & Mason page has no Item Weight row; only 1 of 77 mortar titles carries weight.

Rejected: the checker's four boundary caveats on Part A finding 3 (4 cups sitting at the edge of both bands) and finding 6 (whether ricer ounces are weight or volume) were graded SUPPORTED, not PARTIAL or UNSUPPORTED, and are not in the checker's numbered "Required fixes" list, so no claim in the JSON needed correcting; left as is.