Predators and What Lethal Control Actually Does

#The jackal you shoot is replaced by one you did not raise

The number the whole South African predation argument runs on comes from one telephonic survey of farmers, done in 2010. It has never been repeated. Sixteen years later it is still quoted as the size of the problem, and farms are still spending against it.

What that spending buys is genuinely disputed, and the dispute is not ideological. Serious people disagree about scale and coordination — and about one South African result that is species-specific in a way almost nobody repeats correctly.

#Two compensatory mechanisms

The core South African work is Minnie's doctoral research at Nelson Mandela University and the associated paper in the Journal of Applied Ecology (thesis; paper). It describes a landscape in which hunting pressure is heterogeneous — heavy on farms, light or absent on reserves — and shows what that does to a black-backed jackal population.

Compensatory immigration. Unhunted reserves act as a source. Hunted farms act as a sink. Every animal removed from a farm opens a territory that a dispersing subordinate from the source population moves into. The killing does not lower the number of jackals on the farm so much as it raises the turnover of jackals on the farm.

Compensatory reproduction. The farm subpopulation, being continually cropped, ends up with a younger age structure and increased reproductive output — earlier first reproduction and more animals breeding. A heavily hunted population is a young, fast-breeding population.

Put the two together and you get the mechanism that matters: the farm becomes an attractive sink, or ecological trap. It looks good to a dispersing young jackal — food, territory, no resident competitor — and the animal cannot perceive the mortality risk before it commits. Minnie's own conclusion is blunt: "These compensatory processes will continue to counter population management actions as long as recruitment from unmanaged areas persists."

Read that last clause carefully, because it is the operational content. The finding is not "control never works". It is that single-farm, year-round, uncoordinated control next to an unhunted refuge is close to self-defeating — you are effectively farming jackals for the district.

#The finding everyone gets backwards

Now the part that is routinely repeated wrong, including in an earlier draft of this course's own research.

The UCT Karoo Predator Project reports that "killing caracals is probably counter-productive for farmers as it is associated with an increase in lambs lost the following year" — and states explicitly that "this is not true for jackals" (UCT CSSR).

There is a further limit worth stating. The Karoo Predator Project result is observational and correlational. Causality could run the other way — farms with rising losses are precisely the farms that cull more, so the correlation between culling and subsequent losses can be produced by the losses rather than by the culling. The project's own sample description could not be confirmed for this course, so no sample size is quoted here.

Before you spend money on predator control, read that Scientific Assessment. It is the national synthesis — nine chapters plus a Summary for Policymakers, edited by Kerley, Wilson and Balfour at NMU's Centre for African Conservation Ecology, covering the history, the socio-economics, the ethics, the legal framework, current practice, jackal and caracal ecology, and the ecosystem roles of mesopredators. Nothing else in South Africa covers the ground.

#The R1.39 billion number, and its three weaknesses

The most-cited national figure comes from Van Niekerk's 2010 MSc at the University of the Free State: R1 390 453 062 per year in small-stock predation cost, split Northern Cape R540.8 m, Eastern Cape R412.8 m, Free State R247.1 m, Western Cape R105.0 m and Mpumalanga R84.7 m. It rests on telephonic interviews with 1 424 farmers across five provinces.

Three weaknesses must travel with that number wherever it goes. It is self-reported — farmers estimating their own losses by telephone, without carcass verification, in a policy debate where a big number is useful to the respondent. It is an extrapolation from five provinces to the whole country. And it is sixteen years old, in 2010 rands, and has never been repeated.

It is simultaneously the best number available and a weak number. Cite it with all three caveats or do not cite it.

#Budget it, do not fear it

The practical move is to stop treating predation as a tail risk that occasionally ruins a year, and start treating it as an annual line item with a number next to it — the same way you treat lick.

Work an example. A 300-ewe flock. The Laingsburg enterprise study found total losses of 12 lambs killed or lost per 100 ewes, with predation accounting for roughly three-quarters of recorded losses (Conradie & Landman 2015) — and read that denominator carefully, because it is 75% of losses, not 75% of the flock.

• 300 ewes × 12 per 100 = 36 lambs lost
• 75% of those to predation = 27 lambs
• At the published Eastern Cape stock-theft valuation for sheep, R1 675.10 per head (recorded and claimed values across 2018–2024), that is roughly R45 000 a year

Now the decision rule falls out of it. Spend up to your expected annual loss on prevention, and measure whether the loss moves. R45 000 a year buys a great deal of lambing-camp fencing, a night kraal, a herder's wages for part of the season, or a guardian animal programme — and each of those is testable. R45 000 a year spent on ammunition and diesel with no before-and-after record is not a strategy; it is a habit.

Two honest corrections to that arithmetic. The R1 675.10 is a 2018–2024 recorded value, well below what a replacement animal costs today, so the number understates the loss. And the Laingsburg figures are Karoo figures — run your own.

#Density concentrates both problems at once

One regenerative practice deserves a specific warning here. Ultra-high stocking density puts a large number of animals in a small area at a predictable place and predictable time. That is a gift to a jackal and a bigger gift to a stock thief.

The theft side is not hypothetical. 190 000 head of livestock were reported stolen nationally in the year to March 2025, including 80 600 sheep, at a total value of about R1.4 billion — sheep are the majority of animals stolen but cattle are the majority of the money (African Farming, 14 July 2025). Recovery is poor: an Eastern Cape study puts it at 17.2% of animals and 6.9% of value (Ndzungu, Ngoshe & Jaja 2025). And reporting rates are somewhere between roughly 30% and 18% depending on which study you believe, so the official figures understate the problem.

In South Africa, then, the stocking-density decision is a security decision as much as a plant decision. Mark and register your animals under the Animal Identification Act 6 of 2002 — that mark is the precondition for any recovery at all — and think about where the mob will be standing at 2 a.m. before you design the paddock layout.

#What to do on Monday

Open a page in the farm book with three columns — predation, theft, other mortality — and start filling it in, with causes and photographs, from today. In two lambing seasons you will know something about your own farm that no national study can tell you, and you will be able to answer the only question that settles this argument: did the money I spent change the number? Until then, every predator decision you make is a guess wearing the clothes of a policy.

The next lesson takes the non-lethal side seriously, with the South African guardian-animal evidence and its failure rates.