In what follows, I use “veg*n” to mean vegan or vegetarian where it doesn’t make any difference.

Mike Archer claimed in an article just before Christmas 2011 that:

” … if you want to minimise animal suffering and promote more sustainable agriculture, adopting a vegetarian diet might be the worst possible thing you could do.”
His evidence related to the number of mice killed during cereal production.

Even if we was right about the number of mice killed during cereal production, his claim is still obviously false because the average Australian consumes far more grain, embodied in the bodies of their meat, than any veg*n could consume. So clearly being veg*n isn’t the worstyou can do.

But Archer didn’t get the number right. He’s out by about a factor of about 400. Read on for the details.

Mice do die during grain production … but how many?

Here’s how Archer does the calculation. Firstly he calculates that you need to kill 2.2 cattle to get 100 kg of protein. That’s his first inaccuracy. People need to eat an amount of food each day in accordance with their weight and activity level. The amount is measured in Calories (or kiloJoules) and if you eat the calculated amount, then it is almost impossible not to get adequate amount of protein. The amount of protein you need is a function of your weight and has little to do with your activity level.

Hence the proper measure of the value of a kilogram of food for most purposes is Calories, not grams of protein.

That introduces an error factor of about 2 to Archer’s calculations because beef has roughly twice as much protein as grain per kilogram.

Archer proceeds:

“Each area of grain production in Australia has a mouse plague on average every four years, with 500-1000 mice per hectare. Poisoning kills at least 80% of the mice. At least 100 mice are killed per hectare per year (500/4 × 0.8) to grow grain. Average yields are about 1.4 tonnes of wheat/hectare; 13% of the wheat is useable protein. Therefore, at least 55 sentient animals die to produce 100kg of useable plant protein: 25 times more than for the same amount of rangelands beef.”
When I first read this little description, a few alarm bells rang, but not loudly and I didn’t really think it possible that Archer could be out by a big enough factor to matter and I knew enough about mouse plagues to know they were a significant issue. Animal Liberation SA was a co-sponsor of a Conference on Fertility Control in Wildlife back in 1990. But you can’t deal with everything and we have rather lost track of the issue in the intervening years.

But, happily various people (Brett, Syd, Jane, Shirley) commenting on Archer’s article raised questions about the area of mouse plagues and eventually curiousity prompted a serious search for some data.

Archer really didn’t seem to understand the question, here’s part of one of his responses:

“No-one has asserted that mouse plagues cover ‘all of Australia’s grain growing regions on average every four years’. The assertion is that on average, each part of the eastern Australian wheat belt is subject to a plague every 4 years–on average. Plagues are numerous, generally localised and common. For example, Brown and Singleton (2002) write “Mouse plagues have been a feature in grain growing regions of Australia since the first plague in 1904. They occur somewhere in Australia once every four years on average, however since 1980 the frequency of mouse plagues in some regions appears to have increased to once every three years.”
The sentence quoted by Archer as evidence … (“They occur somewhere in Australia once every 4 years …”) means what it says and not what Archer wants it to say. Consider the following 3 analogous statements:
  1. During each four year period, there is a single day on which I break every bone in my body,
  2. During each four year period, each bone in my body is broken exactly once
  3. During each four year period, one of my bones is broken.
The number of bones broken over a 4 year period is the same in both A and B. Archer seems to think they are different. The qualitative experience is definitely different but the number of bones broken isn’t. The quote Archer gives in support of (B) clearly means (C). This was pointed out to him by Brett and Syd, but Archer simply didn’t get it.

Regardless of Archer’s logical confusion, what does the data say.

And what was the result …?

There are indeed frequent mice plagues in Australia with densities in the 500-1000 mice per hectare range, or even higher. But that’s about the only thing Archer gets right.

His thumbnail calculation implies A or B. This is wrong. It’s not even with a factor of 10. His statment that poisoning kills 80 percent of mice is also wrong because most of the time, people don’t bother to poison during plagues. It’s time consuming and expensive.

And the proof?

Australia has a Co-operative Research Centre specially set up to examing feral animal problems and it coordinates research on all manner of things, including mice plagues.

Its 2004 report, “Counting the Cost: Impact of Invasive Animals in Australia”, has a section on mouse plagues which gives annualised costs and annualised areas impacted (page 9).

  1. Plagues affect 100,000 to 500,000 hectares of grain crops every year.This confirms that we are dealing with situation C.Given that we generally plant in excess of 20 million hectares of grain every year, then plagues affect between 1 in 200 and 1 in 40 hectares each year.
  2. The report gives a break down of costs. They include grain losses, managment and research. There is simply no section called baiting or poisoning. However, a figure of $24.4 million is quoted as the annual cost of plagues with $22.79 million being production losses. This leaves at most $1.61 million for other expenses, including baiting. At $15/ha (a 1996 figure), this is enough to bait at most 100,000 hectares.So the area poisoned is at most 1/200 of the total area planted and in that area, 80 percent of the mice may well be killed.
  3. Lastly, I won’t try to quantify it, but most mouse plagues occur in the eastern wheat belt, where yields are lowest and not in the WA wheat belt where yields are highest.
  4. Reading the literature, it seems that most of the deliberate killing of mice associated with grain production is done to prevent plagues, not during plagues. The poisoning during the 1993 plague is mentioned frequently and seems to have been exceptional. If people were poisoning half a million hectares annually, you can bet the costs would appear in that CRC report.
Summary and conclusion

In summary, Archer’s estimate of the poisoning deaths associated with mice plagues is out by at least a factor of about 400. 200 for the number killed multiplied by another 2 for considering grams of protein rather than Calories.

While individual death free lunches are certainly possible, the provision of any kind of food for 22 million people (or globally for 7 billion) will impact other species adversely. But overall, Archer hasn’t given any convincing evidence that a vegan diet isn’t the best way to reduce both total suffering and environmental impact.

Sometimes things that look obviously true, remain true, even after the best efforts of sloppy thinkers to muddy the water with a mish mash of dodgy statistics.

Postscript … food for thought

I thought I might add a couple of points people might like to think about.

  • Archer seems to think that since mouse are killed to protect grain output, their deaths can be attributed to people who eat grain. Fair enough. Does this imply that since rabbits are frequently killed to protect grazing areas, those deliberate deaths are attributable to people who eat sheep and cattle? If so then the number of deaths to produce a kilo of meat will be rather more than in Archer’s calculations.
  • At one point (Jan 4, 11:19), Archer waxes lyrical about the number of sentient lives never allowed to be lived because of grain production. Is he serious? If so, then this allows vegans to feel immense pride for all those mice who would otherwise have never been born.