OT_173: FAO Animal Production & Health Paper 69 (1988) — Village milk processing: butter costed at 82% fat, cheese…

Source

https://www.fao.org/4/t0045e/T0045E00.htm — original source (opens in a new tab; the file is not redistributed)

FAO Animal Production & Health Paper 69 (1988) — Village milk processing

The milkfat → butter conversion denominator — and a scope correction the upstream research report got wrong

FAO’s village-scale milk processing paper. It supplies the 82% butter fat content that converts milkfat into butter: verbatim, “the cost price of 1 kg of butter will be based on the cost of 820 g of fat content” — giving butter_kg = milkfat_kg × recovery ÷ 0.82. It also gives a village cheese yield of “approximately 11 kg of cheese per 100 litres of milk” (Gouda-type). 🔴 SCOPE CORRECTION. The upstream research report cited this as “FAO butter composition”. It is not a compositional standard — it is a village dairy cost-accounting manual, and the 82% appears as an assumption inside a worked costing example. The proper instrument for a butter composition standard is Codex STAN 279-1971 or the NZ Food Standards Code. 🔴 The 90% fat-recovery figure the upstream report applied is NOT from this source and is uncited anywhere in corpus — the recovery term in the conversion above is still empty (RT_388).

Summary

FAO Animal Production and Health Paper 69, Village milk processing (1988), is a practical manual for small-scale, village-level milk handling and processing in developing-country contexts, written around the economics of small dairy enterprises: what it costs to turn milk into butter, cheese and other products at village scale, and how to price the output. It is not a food-standards document and it is not New Zealand data.

Its value to Neobiome is narrow but real: it supplies the denominator of the milkfat-to-butter conversion (butter is ~82% fat, so a kilogram of butter embodies ~820 g of milkfat) and a village-scale cheese yield. Its greater value is corrective — it is the source the upstream research report mis-scoped as a composition standard, and it is the source that does not contain the 90% fat-recovery figure that report applied. Ingesting it fixes both errors and makes the remaining gap explicit.

Key claims

- claim: "BUTTER IS COSTED AT 82% FAT — the milkfat→butter conversion denominator. Verbatim: 'the cost price of 1 kg of butter will be based on the cost of 820 g of fat content.' This gives the conversion butter_kg = milkfat_kg × recovery ÷ 0.82. [verified verbatim against the source]"
  source_location: "FAO Animal Production and Health Paper 69, Village milk processing (1988) — butter costing worked example."
- claim: "VILLAGE-SCALE CHEESE YIELD: verbatim 'approximately 11 kg of cheese per 100 litres of milk' (Gouda-type). ⚠ Village-scale and non-NZ. [verified verbatim against the source]"
  source_location: "FAO Animal Production and Health Paper 69, Village milk processing (1988) — cheese section / costing example."
- claim: "⚠ SCOPE CORRECTION — this is NOT a compositional standard. The upstream research report cited this source as 'FAO butter composition'. It is a village dairy COST-ACCOUNTING manual, and the 82% figure appears as an assumption inside a worked costing example, not as a specification. The proper instrument for a butter composition standard is Codex STAN 279-1971 or the NZ Food Standards Code. [scope finding of the verification pass, recorded verbatim from the approved draft]"
  source_location: "FAO Animal Production and Health Paper 69 (1988) — document scope and purpose; the 82% appears within the butter costing example, not in a standards section."
- claim: "⚠ THE 90% FAT-RECOVERY FIGURE THE UPSTREAM REPORT APPLIED IS NOT FROM THIS SOURCE AND IS UNCITED. The churning/separation recovery fraction — how much of the milkfat in the vat actually ends up in the butter — is therefore still unsourced. [negative finding of the verification pass, recorded verbatim from the approved draft]"
  source_location: "FAO Animal Production and Health Paper 69 (1988) — searched; no 90% fat-recovery figure present."

Neobiome Intelligence relevance

Feeds d02_smart_food_systems as the processing leg of the community-dairy → edible-fat chain. The chain is:

cows × kg milkfat/cow ([[ot_172_dairynz-dairy-stats-2023-24|OT_172]])
  × churning/separation recovery fraction  ← recovery % STILL EXTERNAL (RT_388); process documented in [[ot_196_butter-ghee-brief|OT_196]]
  ÷ 0.82 butter fat content (OT_173)
  = kg butter/yr

What this source settles: the 0.82 denominator, and the fact that the chain’s remaining unknown is a recovery fraction, not a composition figure.

What it does NOT settle — and this is the point of ingesting it:

  • The recovery fraction is empty. The 90% the upstream report used is not in this source and is not cited anywhere. Until RT_388 is filled, any butter output figure must be reported as a band across a stated recovery assumption, with the assumption visible — never as a point estimate. The small-scale separator+churn process is now documented in OT_196 (Practical Action / Fellows 2008), but that brief states no recovery % either — the numeric fraction remains external (Guelph Dairy Science eBook ~95% mechanical separator+churn vs ~60–76% traditional no-separator; FAO APHP 85), still unretrieved under RT_388.
  • It is not a composition standard. If the model needs a defensible butter fat content for anything other than an indicative conversion (e.g. a nutrition or labelling claim), cite Codex STAN 279-1971 or the NZ Food Standards Code, not this manual.
  • The cheese yield is village-scale and non-NZ. Usable as an order-of-magnitude cross-check on a community cheese pathway; not as an NZ coefficient.

Notes

  • 1988 vintage, developing-country village context, non-NZ. data_quality: medium reflects both the vintage and the mis-scoping risk: the figures are internally consistent for their own purpose (costing) but were never intended as engineering or compositional constants.
  • Why it was ingested despite the scope correction. Two of its four claims are negative — the scope correction and the missing 90% — and those negatives are the batch’s protection against a wrong number entering the fat model. Recording them on a page is what stops the “FAO says butter is 82% fat, and FAO says recovery is 90%” conflation from recurring.
  • Retrieval: HTML capture of the FAO document. Not AI-prepared, so no retrieval-provenance block.

Research targets

Documents to retrieve

  • RT_388 — the churning / separation fat-recovery fraction: what proportion of the milkfat entering a small-scale churn or separator actually ends up in the butter. This is the one empty term in the milkfat→butter chain above, and the figure the upstream research report supplied without a citation.

Research gaps

  • Not raised as RTs (recorded here instead, per the batch’s minimise-new-RTs decision):
    • A proper butter composition standard — Codex STAN 279-1971 or the NZ Food Standards Code. The 0.82 from this manual is adequate as an indicative conversion denominator, so retrieving the standard is a precision upgrade rather than a blocker; it is named here so the correct instrument is on record.
    • An NZ small-scale cheese yield. The 11 kg/100 L here is village-scale and non-NZ. A community cheese pathway is not currently modelled, so this stays a note rather than a target.

Connections

Links to

Sources (2): OT_172 · OT_196

Referenced by

Sources (3): OT_172 · OT_196 · OT_214

EDT domains (1): D02: Smart Food Systems & Agriculture