Four Farm Observations and the Dung Pat Clock

#The pat that would not go away

In late February, on a camp that has had good rain, you find a dung pat from before Christmas still sitting there like a scone. The crust is intact, the grass under it dead. Kick it over: no tunnels, no beetles, no smell of anything working.

That pat is not a soil problem. It is a receipt. Months earlier at the crush, a decision was made about what went into those animals, and the pat is the paperwork coming back. The South African review of record on macrocyclic lactones found around 75% of doramectin residue still present at 180 days (Jacobs & Scholtz 2015, OJVR). Six months. Dung stays a chemical product long after the animal has stopped being one.

The most decision-changing information on a South African livestock farm is lying on the ground, and it is free.

#The four observations

Mark three points with a GPS pin and something physical — a steel dropper, a painted rock. Not "somewhere in the top camp". The same three points, forever, so that in five years you compare the same soil to itself.

Choose them using what this module has already established about water points: one near a trough, one at mid-distance, one as far from water as the camp allows. Kotzé found soil damage concentrating in the piosphere in both grassland and savanna systems, with even well-run rotational camps showing early deterioration of aggregate structure near the water point (Kotzé 2015, UFS). A point that avoids the trough tells you all is well right up until it isn't.

Do not attach an absolute infiltration figure to observation 2. Most of the "South African" grazing and infiltration numbers in circulation are North American, New Zealand or Irish, and one widely quoted pair does not appear in its cited source at all. There is no verified South African rangeland measurement of infiltration by stocking density — which is why your own paired comparison on your own soil is the better instrument. Capping itself is documented across all nine provinces, driven by low organic matter, high silt and fine sand, and adverse surface chemistry (review of soil crusting in SA).

#If you are going to measure carbon, measure it properly

Testing badly and then believing the result costs you twice: once for the analysis, once for the decision made on a meaningless number. Five rules; break one and the five-year comparison is void.

• Sample by soil form and terrain unit, not by camp. Mixing a shallow crest with a deep footslope in one composite destroys the signal before it reaches the lab.
• Always measure bulk density on the same cores. Without it and a fixed depth, a carbon percentage cannot become a stock in tonnes per hectare, and compaction alone will manufacture an apparent gain.
• Fix the depth increments — for example 0–5, 5–15 and 15–30 cm — and repeat them exactly.
• Fix the GPS points and go back to them. Not the same camp: the same spot.
• Never compare across methods. Walkley-Black measures oxidisable carbon, dry combustion total carbon; two results three years apart from two laboratories prove nothing (SGS South Africa).

Detecting real change against South African background variability takes five years or more and a lot of cores, so set the baseline up properly on day one — a bad baseline cannot be repaired later. We deliberately quote no per-sample price, because SA laboratory pricing could not be confirmed as at July 2026: phone Labserve, SGS or the ARC and get it in writing.

#The arithmetic that saves you from a false result

Your laboratory reports 0.8% organic carbon at 0–30 cm, bulk density 1.40 g/cm³. That is 10 000 m² × 0.3 m × 1.40 t/m³ = 4 200 t of soil per hectare, so 0.8% of it is 33.6 t C/ha. Three years later, same depth and density, the report says 0.9% — 37.8 t C/ha, an apparent gain of 4.2 t C/ha in three years, or 1.4 t C/ha/yr.

Now test it. The international literature credits improved grazing-land management with roughly 0.05–0.5 t C/ha/yr (Lal, Carbon Management) — an international figure, and SA's arid and semi-arid systems sit at the bottom of that band. Your 1.4 t C/ha/yr is nearly three times the top of that international band, and about fourteen times a realistic SA rate of 0.1 t C/ha/yr.

The honest reading is not "the grazing plan is working" but "something in the measurement changed". At 0.1 t C/ha/yr, a 100-hectare Highveld holding gains about 10 t C a year, roughly 37 t CO₂e: slow, unglamorous, and nothing like a 0.1 percentage-point jump between two lab reports. And treat the quantity itself as argued over. The South African evidence for improved grazing is reasonable on cover, rest-driven recovery, forage supply, drought resilience and stocking capacity, and weak, context-dependent and hard to measure on soil carbon. Anyone selling you the second as though it were the first is selling.

#The dung pat as a diagnosis of your own programme

Southern Africa has an exceptional dung beetle fauna, and more than 100 species can colonise a single pat in savanna conditions, working in guilds: rollers cart dung off in balls, tunnellers bury it beneath the pat, dwellers breed inside it, kleptocoprids steal from the others (Wildlife ACT). Between them they bury dung, aerate soil, cycle nutrients into the root zone, cut fouling, suppress pest flies and reduce nematode larval survival in the pat. That last one matters: the beetles are part of your worm control. (The national species total often quoted traces back to a user-generated list, so this course leaves it out.)

Macrocyclic lactones act on invertebrate chloride channels, are poorly metabolised, and pass largely unchanged into dung. Jacobs & Scholtz rank them, most to least harmful to dung fauna: doramectin > ivermectin > eprinomectin > moxidectin, moxidectin being least likely to disturb the natural insect assemblage. Persistence differs enormously — ivermectin residues become undetectable within about a fortnight, moxidectin by around three weeks, eprinomectin at roughly a month, and doramectin still sits near 75% of residue at 180 days. Route matters as much as active: subcutaneous ivermectin caused longer delays in dung degradation than the pour-on, despite pour-ons being dosed higher, while sustained-release boluses are the worst option of all, excreting toxic concentrations persistently over prolonged periods.

Intact pats persisting three weeks or more in the wet season point to a dung-fauna problem, and that is usually a drenching problem. Before concluding your soil biology is weak, check what went through the crush and when. And note what targeted treatment rests on: in a 2025 Eastern Cape study of communal sheep, resistance turned up on eight of eight farms, every one to at least two drug classes (Mavundela et al. 2025). Blanket dosing is not cheap insurance — it costs you the beetles and the drenches at once.

#Kraaling: fertility you place on purpose

If dung is a fertility input, you can decide where it lands. A 2023 meta-analysis found kraals and bomas approximately double soil concentrations of carbon, nitrogen, phosphorus and potassium and slightly raise pH relative to adjacent non-kraaled soil, with kraal age not predicting the effect (Momberg et al. 2023, AJRFS). Short-duration overnight kraaling — a few nights per site, then move — is used on ranches to build nutrient-enriched patches (PMC9049567).

#Manure you did not produce

Importing organic matter — stable bedding, horse manure, a neighbour's straw — is one of the fastest ways to lift a degraded patch, and on a small holding beside a stable yard it can be free for the fetching. Take it. But you are importing three things you did not choose, and only one of them is fertility.

You are importing somebody else's drenching programme. Everything this lesson says about macrocyclic lactones applies to the animal that produced the manure, not to yours. Horses are dewormed routinely, frequently with ivermectin or moxidectin, and often on a calendar nobody wrote down. Spreading that manure onto ground you are trying to build dung-beetle populations on can work directly against the thing you are doing it for — and doramectin's roughly 75% residue at 180 days shows how long "old manure" can stay a chemical product. Composting helps and does not fully solve it.

You are importing weed seed. Straw and manure carry viable seed, and under the Conservation of Agricultural Resources Act 43 of 1983 and the NEMBA Alien and Invasive Species Regulations the duty to control declared invaders is yours the moment they are on your land — regardless of which bakkie they arrived on.

You are importing biosecurity risk, low between horses and ruminants but not zero, and higher if the source yard also runs stock.

Fertility placed on ground that carries no animal is a soil amendment, not a grazing system. It changes nutrient status; it does not defoliate a single plant. Module 4 deals with what happens to a grassland that is fed but never grazed.

#What to do on Monday

Buy a 20 cm ring, put three steel droppers where they mean something — one at the trough, one at mid-distance, one far out — and write the four quarterly dates on the wall calendar. Then, before you order another drench, turn over a summer pat and count what is living in it. If nothing is, the first thing to change is not your soil. It is your programme.