Metal
A curated, evidence-aware list from the frozen product graph.
- gold (hiraṇya)metalGold, both as an object of wealth and as the colour of anything radiant.84linked passages
- metal (ayas)metalMetal, and specifically the dark metal the translators render "iron": the substance of Indra's bolt, of a fort, of a hook and of a blade.11linked passages
- lead (sīsa)metalLead, used in the Atharvaveda as the material with which a demon is pierced and in the Yajurveda as one of the metals of the sacrifice.7linked passages
- silver (rajata)metalSilver, named beside gold and iron in the short metal lists and in the amulet charms.2linked passages
- copper (loha)metalCopper or bronze, attested once: VS 18.13 asks that "my copper and my lead and my tin" prosper by sacrifice, in the corpus's one enumeration of the metals.1linked passages
- dark metal (śyāma)metalThe dark metal named at VSM 18.13, standing inside that verse's metal enumeration between ayas and loha. The registry records the attested token and its position in that list; it does not assert a modern metallurgical identification, for which the traditional readings (iron, or a dark copper alloy) disagree. Only the exact inflected forms are registered: the bare stem token-matches the unrelated verb form śyāma at RV 6.5.7 and substring-matches 22 times inside aśyāma. `śyāmaṃ` is the Yajurvedic witness at VSM 18.13 and `śyāmam` the Atharvavedic one at AVS 11.3.7, one verse before the trapu witness in the same hymn.2linked passages
- tin (trapu)metalTin, named in the Atharvaveda's material description at AVS 11.3.8. The exact Sanskrit token is registered because the stored text already attests it; this entry does not infer metal absence from Vedas where that token is not matched.2linked passages
How this list was scoped1
`passage_count` counts DISTINCT passages reaching the object through the lexical, conceptual and ritual predicates this surface counts, and is the same figure the detail view reports. It is NULL, never 0, where no passage reaches it: the mention layer's measured recall is 0.8857, so a null is a lower bound and not an assertion that the corpus is silent.
measured