Where Did That Number Come From?

#The number that was never measured

The first draft of this course's own soil and water research carried a pair of infiltration figures: 22 mm per hour under high stock density against 41 mm per hour under low. It arrived with a South African university thesis attached. It read like a measurement taken on veld near you.

Open the thesis. The numbers are not in it. Not on another page, not under another heading, not in an appendix. They do not exist in the document they are attributed to.

That error was not caught by anyone reading more carefully, thinking harder, or knowing more about soil. It was caught by a click.

This lesson is built entirely out of errors found inside this course's own research, not out of somebody else's bad advice. If the people writing your course made these mistakes, so will the consultant, the feed rep, the podcast and the neighbour who is very sure.

#Four figures, one table, three countries

The draft pack presented four figures as measured South African numbers for what grazing does to infiltration and soil structure. Checked one by one against the source:

• "High stock density 22 mm/h against low density 41 mm/h." Not present in the cited thesis. Cut entirely, along with the teaching example that rested on it.
• "Light-to-moderate grazing retains about 75% of ungrazed infiltration; heavy grazing about 50%." Gifford & Hawkins (1978) — a North American review, quoted inside the thesis's literature chapter.
• "Trampling raises bulk density 8–17% and cuts macroporosity 57–83%." Singleton et al. (2000), reached via Kurz et al. (2006). New Zealand and Irish work, again from the literature chapter.
• "Karoo crusted soils: 40 mm/h on shale, 83 mm/h on dolerite." Genuinely South African, but paywalled and never read at source. It stays flagged, and should not be quoted until someone has the paper open.

Nobody lied. A South African researcher wrote a proper literature review summarising overseas work, as every thesis does. Somebody downstream lifted numbers out of that review, kept the South African citation, and dropped the fact that the measurements were taken on another continent. Two steps later it is "a South African study found".

The bitter part is what got skipped. That thesis's own South African results are more useful than any number taken from it: soil degradation was less pronounced under rotational than under continuous grazing in a clayey grassland, and the damage concentrated around the water point rather than across the paddock. Real, local, peer-reviewed, and buried under borrowed figures.

#Seventy-five percent of what?

Page 2 of the Department's ostrich value chain profile opens with the line the whole industry repeats: South Africa dominates the ostrich industry worldwide with 75% of the global market share.

Twenty-two pages later, same document, on international trade data: in 2022 South Africa's exports were 1% of world exports of ostrich meat, ranked 19th, at 80 tonnes and about USD 642 000.

The first reaction is that somebody is lying. Nobody is. The two statements measure different things and the document never reconciles them: "market share" refers to birds slaughtered and to the hide trade, the meat tariff line captures a sliver of a trade that mostly moves as leather, and 2022 was a collapsed year. Agribook states the defensible version — South Africa accounts for around 75% of the ostriches slaughtered in the world.

#Minus fifteen point three percent of somebody else's cattle

Here is a number now quoted across South African beef discussion: weaning weight carries a relative economic value of −15.3% for feedlot profit, against carcass weight at +34.1% and deaths, disease and injury at −31.6%. It is usually introduced as "the ARC found", and it is the backbone of the right-size-your-cow argument, which this course accepts.

Read the paper's own second paragraph. The study was done in the United States on Charolais-sired feedlot cattle. South African authors adapted it by removing bovine respiratory disease and marbling score, then re-estimated the values. So what you have is a US decomposition, minus two variables, re-run for local use. No South African feedlot dataset producing these values was located.

That does not make it useless. It makes it a structural argument, and structure travels where magnitudes do not. But it changes what you may say. You may say a feedlot's economics penalise weight that arrives before the feedlot and reward weight added inside it. You may not say South African data proves it.

#Units: where the money actually leaks

Bad provenance costs you a decision. Bad units cost you a budget, quietly, in your own spreadsheet.

Greasy against clean fibre. Mohair is quoted on two bases: the clean indicator is what the fibre is worth after yield loss, and the greasy indicator is what actually leaves your farm on a bale. At the sale of 2 June 2026, kid hair traded at R699.60/kg greasy and R797.26/kg clean; strong adult hair at R283.26 greasy and R321.59 clean.

Work it through on a small flock. Two hundred Angoras, shorn twice a year. Farm-level clip is reported at roughly 1–1.5 kg per animal per shearing, so call it 1.25 kg — 2.5 kg per animal per year, 500 kg of greasy hair leaving the farm. Priced against the kid clean indicator you write down 500 × R797.26 = R398 630. Priced against the greasy indicator, which is the basis it will actually be sold on, 500 × R699.60 = R349 800.

You have just overstated the enterprise's gross income by R48 830, about 14%, without a wrong keystroke. A real flock's class mix changes the rands but not the error — the 14% is the basis, not the class.

Carcass against live weight. A South African pastured-pig operation publishes on its own website that its pigs "slaughter out at ±40 kg after twelve months". That is a carcass figure — the farm does not say whether warm or cold. The benchmark people set beside it, a commercial pig reaching about 100 kg in six months, is a live weight. At the dressing percentages measured on hardy South African pig types, a 40 kg carcass is near 50 kg live. The honest comparison is that the outdoor system takes twelve months to reach roughly half the live weight a commercial pig reaches in six. That is still a heavy penalty and still the whole economics of pastured pork here — but state it in matched units.

As-fed against dry matter. The Department's rule of thumb for a growing ostrich is about 2.5 kg of feed per bird per day. The measured trial figure from Oudtshoorn is about 2.02 kg of dry matter per bird per day overall, running from 1.34 kg in the starter phase to 2.65 kg in the finisher. The first is a rule of thumb and very likely as-fed; the second is measured dry matter. Put them in one feed budget and the answer is wrong by more than the margin you were arguing about.

Depth, the same problem wearing a different hat. Long-term South African burning plots show organic carbon 70% lower in the top 0–10 mm — an alarming figure that travels well. The same source reports the loss over 0–100 mm at 12%, nitrogen down 13%. Same trial, same fire. Sampling depth is doing the rhetorical work. Quote both numbers or neither.

#The four questions

Everything above collapses into four questions you can ask out loud in ninety seconds.

Run those four questions on the next three claims you meet about regenerative grazing, and expect one to fail. That is not cynicism; it is the ordinary condition of applied agricultural information. Every stocking rate, price and threshold in these lessons carries its source. Click one this week and see whether the number is where it says it is.