by Brian Tomasik
First written: 24 June 2013; last update: 15 Sep. 2017


Applied welfare biology involves assessing how environmental policies affect net wild-animal suffering. This research is important and should be pursued, and I outline some interesting questions in this field for people to study. That said, applied welfare biology has some risks, including encouraging policies that are unpopular and whose implementation could slightly increase international conflict. Therefore, I think the first priority of animal advocates should be to argue against future projects to spread wilderness into space or computational systems. Not expanding wild-animal suffering is an easier case to make and is ultimately more important in expectation.

See also the list of research topics for the "Animal Ethics Essay Prize".



Yew-Kwang Ng established the field of "welfare biology" in his classic 1995 paper: "Towards Welfare Biology: Evolutionary Economics of Animal Consciousness and Suffering." He proposed to systematically research how we can improve the welfare of wild animals, which is a topic that has so far been largely neglected in the fields of ecology and animal welfare.

Applied welfare biology involves analyzing the sign of net impact for various environmental policies on wild-animal suffering. Small, short-lived species that have many offspring dominate numerically

among wild animals, and given that almost all of their babies die shortly after birth, they produce far more suffering than happiness. Thus, the question of reducing wild-animal suffering will often boil down to a question about reducing populations of these r-selected organisms over the long run.a The best way to do this can be to examine the growth-limiting factors for an ecosystem (plant energy, dissolved oxygen, moisture, temperature, etc.) and then evaluate how a given environmental change impacts those factors.

Why animal advocates should focus on the future

Applied welfare biology typically targets short-term wild-animal suffering on Earth. Present-day suffering in nature is tragic, and our heart strings are tugged to try to do something about it now. And indeed we should explore to some extent what we could do to reduce the atrocities in the wild in the near term, even if only by modest amounts. However, I think that overall, animal advocates should focus on averting future wild-animal suffering, which could vastly dominate that in the present. I explain the main reasons below.

Global stability

In most cases, the conclusion of applied welfare ecology is that more habitat destruction is better, to avert future births of animals that will almost all die painfully shortly after coming into existence. Unfortunately, in many cases, habitat destruction runs contrary to another important variable: Global peace and stability.

For instance, in "Crop Cultivation and Wild Animals," I discuss how topsoil loss seems likely to decrease primary productivity in the long run but may also exacerbate future wars. The same could be said for other resource shortages -- in water, biodiversity, and so on -- that may also be accentuated by habitat degradation. Human conflicts matter a lot because they could lead to wars, arms races, and failures of compromise that in expectation plausibly increase future suffering. Even very small changes in prospects for international cooperation could translate into vast numbers of additional suffering organisms in the far future.

This doesn't mean we should abandon applied welfare ecology entirely. There may in fact be cases where wild-animal welfare and future stability align. For instance, I'm very uncertain about the net impact of climate change on wild animals in the short and medium terms, but if climate change proved to be net harmful for wild animals, then we would have two reasons -- wild-animal welfare and global stability -- for opposing it. However, it's too early to reach such a conclusion now, and it may very well turn out that climate change is net positive for wild animals, in which case we would have a similar tradeoff as appears for topsoil erosion or biodiversity loss.

There's also a risk that others who take up the mantle of applied welfare biology will ignore considerations of global stability and focus exclusively on short-term benefit to wild animals, which could cause net harm depending on the magnitudes of the different effects at play.

Habitat destruction is unpopular

Concerns about global stability are reasons that even pure suffering reducers acknowledge as directly relevant. However, many other people object to environmental degradation on a more visceral and simple basis: They like ecosystems and don't want them destroyed. Justifications for this view include

  • Aesthetic appreciation of nature's beauty and balance
  • Feeling that nature is special because it's millions of years old
  • Avoiding hazards to human health caused by pesticides and herbicides
  • Desire to preserve biodiversity for discovering new medicines, finding useful genes, and other scientific discoveries
  • Keeping the option open to preserve nature in case we later decide wild-animal lives are net positiveb
  • Wanting to avoid unforeseen ecological or economic consequences.

I personally suspect that raw beauty may be a strong driver behind environmentalism, but presumably this intuition is encouraged and fortified by the more "solid" economic and scientific reasons.

Regardless of motivations, many people care about ecological conservation and are upset when it's suggested that eliminating ecosystems would reduce wild-animal suffering. There's a risk that by promoting the habitat-destruction agenda too aggressively, wild-animal advocates may turn people away from caring about wild animals, possibly permanently. We wouldn't want our movement to do for the cause of helping wild animals what animal-liberation terrorists have done for the cause of anti-vivisection, namely, arousing public opposition that delays acceptance of the problem.

This is especially important given that most of the expected wild-animal suffering in humanity's future lies ahead of us, even though nature on Earth continues to shrink. We don't want to kill the goose laying golden eggs by prematurely turning people against us and tarnishing our movement for when it matters most down the road.

Arguing against spreading wildlife is easier

Future wild-animal suffering might take a few forms:

  1. Spreading plants and animals to Mars or other planets for terraforming
  2. Disbursing seeds of life into the wider galaxy in the project of directed panspermia
  3. Simulations of evolution or virtual-reality wilderness that are sophisticated enough to contain conscious suffering.

Unlike wildlife on Earth in the near term, these forms of wilderness are all basically optional. Perhaps terraforming will one day be widely seen as important for beginner-level space colonization, but at least presently, few people would complain if these projects were not pursued, and some object to them on grounds other than wild-animal suffering. So it should be relatively easier to convince people that spreading the cruelties of nature in these contexts would be morally objectionable, since most people are already not that supportive of them.

Other reasons it's easier to argue against spreading wildlife:

  1. Status-quo bias is on our side. People want to preserve existing nature but don't feel an obligation to create it de novo. Indeed, some deep ecologists feel that we should keep planets unpopulated because this is their natural state.
  2. We can appeal to "problem of evil" intuitions. Since we would be actively bringing about new wildlife, we have to ask why supposedly good creators are giving rise to so much suffering by innocent animals. (Thanks to a friend for originally proposing this angle.)

Some principal drawbacks of focusing on not spreading wildlife:

  1. The scenarios sound like science fiction to most people.
  2. Some might say it's too early to talk about this because nothing can be done now. (Of course, this isn't true. What we can do now is make the argument more widely recognized so that the ground is laid before the technological capabilities arrive.)
  3. Applied welfare biology has concrete action recommendations like vegetarianism does, while for averting future wild-animal suffering, the actions are more fuzzy, like research and outreach.

The bias for short-term action is a main reason that applied welfare biology feels so pressing. If we care about wild-animal suffering, the natural assumption is that we should focus on it in the present. It's possible that even if we focus on future wild-animal suffering, others will apply those principles to the present day. Of course, this can be a good thing, but it just requires caution about not stepping on toes, as discussed previously.

Some arguments for applied welfare biology

With the aim of presenting both sides of the debate, here are some reasons we might think applied welfare biology is important in the short term.

  • It has potential direct implications for our daily behavior (e.g., should we recycle?).
  • It defuses the humans are helpless to do anything objection to reducing wild-animal suffering. In fact, each human affects thousands of wild animals every day by his or her environmental choices. Governments spend major resources on environmental policies that affect trillions of animals.
  • Like vegetarianism, it can give people something concrete they can do about wild-animal suffering, rather than feeling hopelessly depressed and giving up.
  • The practice of doing this analysis may one day spill over into the actual academic discipline of welfare biology, and we can help set the foundation for that.
  • The analytical framework of reducing r-selected populations as much as possible can shape people's thinking about the problem in general. For instance, in macroeconomics, the task is to evaluate the net impact of various factors on GDP, and this may cause people unconsciously to come to believe that GDP is an important thing to optimize. Sometimes the analysis shapes people's goals, rather than the other way around.
  • The topic is really interesting, so it may organically bring in new people to the issue for whom mere speculative philosophy isn't so compelling. It also involves more hard science and demonstrates that advocates for wild animals aren't hopelessly naive about how nature works.

Luke Muehlhauser gives similar reasons for why he thinks his organization, MIRI, should work on object-level research in addition to just public outreach.

Topics in applied welfare biology

I think it's good to work on applied welfare biology as long as we keep in mind the caveats discussed previously. If we can find policies that both help wild animals and align with global stability (and as a bonus, people's ecological sentiments), these could be win-win-win solutions and should be promoted.

In the remainder of this piece, I discuss some topics in applied welfare biology and sketch some of my speculations about them. Note that I wrote this section earlier and thus have not in this section given much consideration to effects on international peace or avoiding clashes with other value systems, and these factors should be introduced into the analysis.

Types of interventions to explore

The basic argument why we should address wild-animal suffering can be made at a philosophical level: Wild-animal suffering matters, there's an extraordinary amount of it, and research on how to reduce it has high expected value. Even failing that, we should work to make post-humans less inclined to spread wildlife into space and sentient simulations.

This point is solid, but many people seem to find it too abstract and still get hung up by the idea that "We can't do anything to help wild animals, so we can ignore the issue." Largely I think this sentiment reflects either (a) an attempt to resolve cognitive dissonance between caring about animals and not wanting to worry about suffering in nature or (b) lack of interest in high-risk high-potential-reward strategies like long-term research and moral edification. A friend of mine suggested that in order to "break the ice" for people who feel this way, it would help to have a few clear examples of "bulletproof" interventions to help wild animals that (1) clearly reduce wild-animal suffering (2) at low cost (3) without risk of significantly impinging upon human interests. These interventions needn't be the most cost-effective; rather, the point is to provide an existence proof for our ability to do something to help wild animals.

Trivial cases

There are some trivial examples of clearly good ways to help wild animals, like when people at the Kapani Safari Lodge in Zambia saved a mother and baby elephant from sinking in deep mud in 2011. The CNN article on the rescue actually states explicitly the point that we want to make: "Most conservationists would agree that you should not interfere with mother nature. But there are exceptions to every rule." Of course, we hope that people will accept more and more exceptions to this unfortunate principle of not interfering with nature as time goes on.

A similar rescue happened in Kenya in 2012. Saving elephants seems good naively, but it's probably also good from a more sophisticated perspective that considers smaller organisms like insects, because elephants consume lots of plant biomass that would otherwise support many more little animals.

Systematic safe bets

The trivial examples can help in philosophy arguments, but they still don't illustrate the idea that larger-scale interventions are possible. It would help to explore more efforts of this latter type. A few ideas:

  • Humane insecticides: See this section.
  • Eliminating predators of large herbivores: See below for details. This intervention has the benefit of seeming good on the surface to naive animal advocates (since you're saving herbivores from predation) as well as being good from a more sophisticated viewpoint that is concerned with small organisms (because big herbivores eat food that would have been consumed by many more smaller, more r-selected animals). The main drawback here is that humans may not be happy with this intervention -- e.g., people complain about deer overpopulation -- but at least we could defend killing predators in cases where farmers favor it to protect their livestock or where humans drive out predators by development.
  • Parking lots, buildings, and sprawl: Parking lots and building construction seem like some of the most clear ways to reduce wild-animal suffering insofar as they prevent plant life on a region of land for decades into the future without many obvious drawbacks for animals elsewhere. Problems with this approach are that (a) it's opposed by conventional animal-rights people because you're just preventing animal life with no other justification for doing so and (b) it's opposed by ecology supporters. Even when this example doesn't work from a public-relations standpoint, we can know privately that it's one of the "safe bets." And there are arguments to be made for how it benefits humans -- the same arguments that building developers make when justifying their projects.

Messier but most important cases

In addition to trivial cases of intervention and more systematic "safe bet" cases, we should also explore areas that we think likely have the highest leverage for actually making a difference to wild animals, even if the details are more messy and speculative. These are what we would focus on if our goal was to prevent as much wild-animal suffering as possible on Earth in the near term, rather than proffer examples for making a philosophical point. Below in the "Research topics" section I enumerate some top priorities for areas that makes a large-scale difference to wild animals.

Eco-friendly interventions?

People cite various reasons against changing human environmental policies to reduce wild-animal suffering, but fundamentally, one of the biggest motivations for opposition is that people find nature beautiful and don't want to mess with it. Therefore, if we could identify environmentalist policies that also happen to benefit wild animals, these would be an easier sell and would help to break the idea that we shouldn't try to consider wild-animal suffering in our policy choices.

The danger here is that we might in the process reinforce the assumption that "the best way to help wild animals is to reduce human interference in ecosystems." Indeed, many environmentalists believe that wild animals would suffer less if humans backed off from nature, which is in fact probably wrong: It's likely that humans have, on the whole, prevented more suffering in nature than they have caused. This makes me worry that, unless our rhetoric was very clear to defuse the environmentalist assumption, we might primarily reinforce ecological tendencies if we promote an intervention to help wild animals that's consistent with nature preservation.

Research topics

What's the net impact on wild-animal suffering of various human activities?

Net impact of crop cultivation [high priority]

See "Crop Cultivation and Wild Animals" for details.

Net impact of pasture grazing [high priority]

Here are some articles on this topic.

Even more land is used for grazing (3.2-3.6 billion hectares) than for growing crops.

Which plant foods are best to eat? [high priority]

The previous section discussed the question of determining whether crop cultivation is net good or bad. In addition, we can further ask which types of foods are better than others, where better mostly reduces to meaning prevents the most insect deaths per calorie of human-consumed food.c I drew a rough sketch of what this kind of analysis could look like in a section titled "Ranking foods".

Global warming [high priority]

See "Climate Change and Wild Animals" for details. The topic has broad implications, because almost every environmental choice involves a difference in energy consumption.


Biofuel releases stored plant energy without powering creatures that suffer. From this perspective, it seems promising. Of course, its production entails similar consequences as crop cultivation in general, so the net balance of those effects would need to be worked out further. My main concern is that fertilizer and irrigation might increase crop yields so much (at least for crop cultivation in wealthier countries) that even after accounting for lost biomass for biofuel, there might still be an increase in net primary productivity. However, assuming that a large portion of the plant is used for biofuel production, maybe this worry is implausible.

In addition, biofuel is sometimes a political ploy to subsidize farmers in the Midwestern US and may actually require more energy to produce than it yields. If so, it contributes to climate change, and those effects need to be considered.

The impact of biofuel might be substantial. Human appropriation of net primary production is predicted to reach 27–29% of potential plant biomass by 2050 but could reach 44% with greater use of bioenergy. This study says:

According to our results, humans today already harvest over 8 Pg C/yr. This biomass amounts to an approximate gross calorific value of ≈300 exajoules (EJ) per year, of which some 35–55 EJ/yr are used for the provision of energy services (35). Prominent studies suggest that the use of biomass for energy generation could grow to 200–300 EJ/yr in the next decades (32, 35). The additional harvest of 4–7 Pg C/yr needed to achieve this level of bioenergy use would almost double the present biomass harvest and generate substantial additional pressure on ecosystems.

In cases where biofuel does actually reduce climate change, perhaps pushing for it could even be seen as socially/environmentally positive.

Some have proposed biofuel from algae. Tradeoffs here are similar as for land-based cultivation of biofuel. I'm not sure if harvesting algae would kill more or fewer small animals than harvesting corn.

This paper sadly proposes to generate biofuel from "zooplankton in aerated wastewater treatment lagoons". This would involve painfully killing huge numbers of zooplankton. It's much better to produce biofuel from plants before that plant food is eaten by animals. While one could argue that burning zooplankton lipids is better than letting those zooplankton be eaten by predators, since doing so would bring into existence additional predator animals, the problem is that biofuel production creates pressures to increase yields, i.e., to create more zooplankton than would have existed naturally, such as by increasing algae production in a given water body. In other words, biofuel production that extracts lipids from zooplankton may increase rather than decrease the total number of animals created.


Some wood -- such as that used in "doors, window frames, crates, coffins, furniture, plywood sheets, chopsticks, household utensils and other items" -- is harvested in the rainforest, which degrades some of the most animal-rich ecosystems on the planet and thereby prevents a lot of suffering in the long run. However, other wood comes from sustainable harvesting of forests or from tree farms, and the net impact of these on wild-animal suffering is not clear.

Logging on US federal lands

This page discusses the debate in the US over whether to increase logging of US federal lands. On balance, I suspect that more logging would be beneficial, since forests tend to have higher net primary productivity than other land types. However, there are some potential counterarguments:

  1. Less imported timber: Increasing the supply of US timber might slightly reduce pressure for imports, including some imports from tropical countries where one hectare of deforestation probably reduces more total wild-animal suffering. However, increasing the supply of US timber would probably expand the total amount of timber used, and the reduction in imports might not be that huge? One would have to study the details of the timber markets to evaluate this better.
  2. Uncertain impact of forest fires: It's claimed that logging forests reduces risk of forest fires. As is discussed later, forest fires have both positive and negative consequences for wild-animal suffering and might be beneficial overall.

Recycling paper

In 2005, before I thought about animal suffering, I lobbied my school board to switch to using recycled paper for its copy machines. At the time I approached the question from the standpoint of benefiting future humans by reducing negative externalities. Now I can consider the question from the standpoint of impact on wild-animal suffering.

On balance, I tentatively still support recycling paper (and do so myself), though I haven't explored the topic in much depth from the perspective of wild-animal suffering.

The net impact of recycling paper on the amount of suffering in nature involves many considerations, although probably just a few of them dominate:

  • If non-recycled fiber is sourced from tree farms, then using virgin paper instead of recycled paper means more tree farms need to displace native vegetation. So we need to ask whether the annual net primary production (NPP) that will eventually be made available to animals to eat is higher on the native vegetation or on tree farms. This paper says regarding "Forest management and wood harvest" that "juvenilization can increase forest NPP, in particular NPP allocated to stemwood growth". And even though some of this NPP is harvested for wood and fiber, that wood and fiber will eventually decompose, sometimes in ways that will feed animals.
  • Non-recycled paper uses more energy, contributing to global warming.
  • Non-recycled paper adds more chemicals to water. The sign isn't clear, because some of them (e.g., dioxins from bleaching) may reduce fertility (which can be good or bad depending on the species affected), but others may merely cause acute harm without reducing populations.

Contrary to popular perception, it seems most virgin fiber in the US these days doesn't come from old-growth forests:

Much paper now comes from sustainable forests. These sustainable forests are essentially "tree farms," where trees are grown as a crop, just like broccoli or wheat. [...]

Overall, one-third of the fiber used to make paper comes from wood chips and sawmill scraps; another third comes from recycled paper.

Recycling cans and bottles

I haven't looked into the details of the environmental impacts of non-paper forms of recycling, but my impression is that one of the main reasons to recycle is to save energy. If climate change is at least as likely to increase wild-animal suffering as to decrease it, then saving energy is at least neutral in expectation and maybe positive. Combined with a default stance in favor of socially acceptable behaviors, this consideration leads me to recycle cans, bottles, and plastic containers. But as I mentioned, I haven't really researched this topic.

Humans have begun finding organisms that can eat plastic, and it's plausible this trend will continue. Perhaps within a century, many types of plastic waste that humans produce will be eaten by bacteria or other life forms? If so, then creating more plastic from oil / natural gas might increase the total amount of life that's brought into existence, thereby increasing suffering. In contrast, recycling plastic would prevent new plastic from being created.

Power plants killing fish

A power plant kills possibly millions of fish and many more fish eggs per year. Does this reduce net fish populations in the long run? A Sierra Club report seems to suggest that the answer is generally "yes". If so, this could be good, both because it prevents births of future fish as well as because fish are generally predators, which plausibly tend to cause more deaths of their prey per unit time than if the prey didn't have predators. See below for more on eliminating predators. On the other hand, ocean food chains have many levels of predation, so eliminating some predators might increase those at a level below. The net impacts of power plants on fish might be comparable to those of fishing.

Some argue that power-plant killings lead to more pressure for r-selective reproduction, which would increase suffering if the fish population stays roughly constant.

Urbanization and sprawl

According to Vaclav Smil, major cities cover 500 million hectares, an area 50% bigger than India. Earth's total land area is ~15 billion hectares. So major cities cover ~3% of Earth's land area. (In comparison, agricultural crops cover "10.9% of global land area".) Of course, cities are not devoid of primary productivity (they contain grass and trees), so the total reduction in Earth's primary productivity may be less than 3%. (On the other hand, cities tend not to be located in deserts or tundra, so the potential primary productivity of the land that cities occupy might be higher than the global average?)

Of course, sprawl might cover even more land per person than cities do. Geoffrey West says: "You need less roads per capita in a bigger city."

Here's a quick ballpark check on the figure of 500 million hectares. In the US, "In 2013, the average size of new houses built increased to an all-time high of 2,679 square feet [...]. Over the last 40 years, the average home has increased in size by more than 1,000 square feet, from an average size of 1,660 square feet in 1973". Suppose that a typical house is 2000 square feet, which is 0.02 hectares. There were 133.9 million houses in the US in 2014. So US houses cover ~2.7 million hectares. Of course, we also have to include driveways, lawn ornaments that cover grass, sidewalks, roads, parking lots, shopping malls, office buildings, etc.

This page says: "The US road network exceeds 6.58 million kilometres in total length, making it the world's longest and biggest road network. It comprises approximately 4.3 million kilometres of paved roads including 76,334km of expressways and 2.28 million kilometres of unpaved roads." Road widths can vary from 20 feet to 60 feet. Let's assume an average width of 25 feet = ~8 meters. Then US roads cover (6.58 million km) * (0.008 km) = 53,000 km2 = 5.3 million hectares. This is about double the area covered by houses.

I don't have data on office buildings, shopping malls, etc., but it's plausible that these should in aggregate cover an amount of land in a similar ballpark as the area covered by US homes, given that people spend about half their waking hours in these places (with the other ~half spent at home). Office buildings might occupy less land than homes because they tend to have higher densities of people per unit area (especially if they're tall buildings). All together, a reasonable estimate for the area covered by buildings and gravel/pavement in the US might be 2 * (2.7 million hectares) + 5.3 million hectares = 10.7 million hectares.

The US comprises 4.4% of the world's human population, so a naive estimate of global land covered by all humans is (10.7 million hectares)/0.044 = 243 million hectares. This is just half of Smil's estimate of land area covered, and Smil's estimate was only for major cities. Part of the difference might be due to Smil's figure including parks, lawns, etc., and part of the difference might come down to errors in my crude estimate. Still, I would think that, if anything, my estimate is too high, because land cover per person is probably lower in Europe (which is more densely populated than the US), and the same is plausibly also true in the developing world, where houses and roads are presumably on average smaller.

A nice way to sanity-check these numbers is to look at random locations on Earth on Google Maps with the "Earth" setting turned on so that you can see satellite images of green vegetation and human structures. An estimate of ~1% for the fraction of all land that's covered by roads and buildings seems about right to me based on Google Maps.


See "How Wild-Caught Fishing Affects Wild-Animal Suffering".

Ocean litter

Marine debris, such as from thrown-away plastic and lost fishing nets, painfully kills animals that eat it or get trapped by it. This might have roughly comparable effects as fishing, except that fishing removes biomass from the oceans, while merely killing fish in the oceans leaves the fish biomass around to support populations of other marine life, which may increase suffering.

Marine debris might prevent a small amount of photosynthesis by blocking light, but the effect is probably not huge; harm to the animals killed directly and side effects on the marine food chain seem more important.


Blue whales may eat as many as 40 million krill per day. Even though krill matter vastly less than bigger animals, the collective moral weight of this harm is substantial. On the other hand, if whales weren't eating the krill, presumably fish would? Fish have shorter lives than whales, and fish ecosystems have many more levels of predation. So if the krill must be eaten by someone, maybe it's better for it to be whales. On the other hand, is it certain that krill will be eaten by someone if not by whales?

The same argument doesn't necessarily apply to whales that predate upon larger fish, although as we saw in the previous section, the analysis is very complicated depending on how one layer of predation affects those below it.

Whales may cause significant net harm via nutrient cycling. The "whale pump" or "whale poop hypothesis" refers to the process of whales bringing nutrients, especially iron and nitrogen, from the deep ocean back up to the surface, significantly increasing primary productivity and therefore suffering.

Blue whales have decreased 100-fold in the past few centuries. For whales in general, Joe Roman reports: "Conservative estimates are that large whales have been cut to 25 percent, though the work done on whale genetics shows that we're probably closer to 10 percent [of previous population sizes]."


Cats eat meat, cost money, and take time to care for. On the other hand, they may increase people's concern for animals in general.

Cats also affect wildlife. They kill billions of rodents and birds in the US per year. It's probably bad for cats to kill herbivorous rodents. The sign of the impact on birds is less obvious, since birds themselves kill hundreds of bugs per day.

HSUS has written about the effects of cat predation on wildlife. It's interesting how in this context, an animal-welfare organization recognizes that predation can be a bad thing. Unfortunately, this logic is not extended beyond human-caused predation.

Rat sterilization

Replacing more painful with less painful forms of population control is generally good. If population numbers remain roughly the same, then there are unlikely to be big ecological side-effects, in which case more humane control methods are clearly an improvement. Thus, insofar as efforts to sterilize rats -- instead of poisoning or otherwise killing them -- reduce population numbers equally effectively as current pest-control methods, sterilization seems unequivocally beneficial.

The analysis becomes more complex if sterilization reduces rat populations much more than other control methods would. We then have to ask who will eat the food that the rats had been eating, and whether this will cause more or less suffering than the rats themselves experienced from their short lives and painful deaths when their populations were supported by this food. In the case of city rats infesting garbage cans, if there are fewer of these rats, there may be a few more cockroaches, flies, and other insects, and more food will end up in landfills. I'm uncertain what fraction of landfill food waste is consumed by invertebrate animals, but my current impression is that most landfill organic matter is eaten by bacteria and other non-animal organisms? If we care less about bacteria than about rats per unit of food consumed, landfilling food waste seems better than letting rats eat it, assuming insects don't eat a significant fraction of the landfilled waste.

The issue is more difficult in the case of rats on crop fields or otherwise outdoors, who consume (mostly but not exclusively vegetarian) food that wouldn't otherwise be put in the trash. Assuming a significant fraction of this food, if not eaten by rats, would be eaten by invertebrates, is it bad to reduce rat populations? I don't know. In this section, I present an inverted-U-shaped curve for "moral importance per gram" for different kinds of organisms. (For my purposes here, it should be "moral importance per Joule of food consumed", since we're considering different animals eating the same amount of food. Rats have faster metabolisms per unit of body weight than humans do, so "importance per gram of body weight" isn't quite the right metric.) Relative to my current moral intuitions, I think rats and insects might both be roughly near the top of this curve, i.e., might have roughly equal moral importance per unit of food consumed. I think both larger animals (e.g., elephants) and smaller animals (e.g., bacteria) matter less per unit of food consumed. If we take this view, then sterilizing even wild rats probably doesn't cause net harm, even if it remains unclear whether it's net good.


Elephants, like other big grazing animals, plausibly reduce net insect populations by eating vegetation that could otherwise be eaten by smaller animals. In addition, elephants may help prevent savanna from turning into forest, which seems good because grassland generally has lower net primary productivity compared with forest. This page explains:

In Africa, a heavy concentration of elephants in protected parkland have created a savanna by eating leaves and twigs and breaking off the branches, smashing the trunks and stripping the bark of trees. Elephants can convert a dense woodland into an open grassland in a short period of time. Annual fires then maintain the area as a savanna.

This thesis offers some further details on how elephants affect savanna vegetation:

Pronounced reductions in trees and other woody plants have been experienced across the continent, including Cameroon, Tanzania, and South Africa (Barnes 1983a, Pamo and Tchamba 2001, Jacobs and Biggs 2002a). Conservationists and reserve managers have expressed concern about loss of rare or vulnerable trees and a possible concomitant loss of biodiversity. This has led to the paradoxical situation whereby managers of reserves with high elephant densities develop plans to limit or reduce population numbers of an endangered species (Barnes 1983b, Caughley et al. 1990). [...]

In concert with environmental factors, elephants can nonetheless precipitate declines in tree populations or marked changes in community composition.

In his early years, Allan Savory recommended culling elephants to prevent land deterioration, though he later came to believe that grazing improves rather than degrades grasslands. However, Savory's views are rejected by many ecologists.

While the presence of elephants in native grassland probably prevents net suffering relative to the absence of elephants, it's obviously probably bad to conserve land against human development in order to allow elephants to live on that land.


Unfortunately, earthworms tend to increase soil fertility. This book explains:

The rank growth of grass around earthworm casts suggests an increased availability of plant nutrients therein. Earthworms are noted for their favorable effect on soil productivity.

Earthworms are important in other ways. The holes left in the soil serve to increase aeration and drainage, an important consideration in soil development. Moreover, the worms bring about a notable transportation of the lower soil to the surface. They also mix and granulate the soil by dragging into their burrows quantities of undecomposed organic matter such as leaves and grass which they use as food.

The "Nutrient cycle" Wikipedia page explains that "Ecosystem engineers can influence nutrient cycling efficiency rates through their actions." Then it cites earthworms as prototypical ecosystem engineers. But do worms accelerate nutrient cycling more than whatever organisms would take their place? Plausibly yes? Worms mix soil and release nitrogen for plants.

This chapter explains:

According to Anderson (2000) the smaller soil animals contribute disproportionately to the metabolic activity of the edaphon, and there are many interactions between them and the microflora, e.g. by grazing. Conversely the larger soil animals are more important in the modification of the physical properties of soils, e.g. by geophagy, comminution of dead plant biomass and by promoting both vertical and horizontal translocation by the process of bioturbation or biomixis. Therefore the larger soil animals are often called ‘soil engineers’ because they have a strong input to such overall soil parameters as infiltration capacity and texture.

On the plus side, though, earthworm invasions of forests can reduce understory growth.

How to dispose of food waste: compost, sink, or trash bin?

This topic is interesting in part because disposing of food scraps is something everyone does, so a ranking of methods of food-scrap disposal would cut against the claim that "we can't do anything to affect invertebrate suffering." The content of this section has moved to an upcoming article, which should be online by Nov./Dec. 2017.

Mowing the lawn

Mowing grass hurts a lot of insects. You can often see a stream of butterflies and beetles fleeing as a mower comes through, and you're less likely to see those that get crushed or torn up. In some cases, a mower may even run into a frog or mouse.

One might think that mowing could have a benefit of reducing primary production. This may be true in some cases, like if you mow the grass too short. But often, mowing actually stimulates growth of the blades and runners (although at the expense of roots). So it's doubtful that ordinary mowing appreciably reduces primary production. Thus, in general, mowing seems probably net harmful.

Watering lawns

According to "Soil ecology": "Moisture is a major limiting factor on land." Presumably watering the soil, by bringing up groundwater to the surface, increases flora and fauna productivity and hence suffering.

Using water in the home

Other things being equal, I encourage reducing home water use because water treatment often kills zooplankton.

Leaving standing water

Leaving pools of water around the yard -- such as in buckets, other containers, or plastic tarps -- seems quite bad. For one thing, it may cause some bugs to drown in the water if they fall in and can't get out. Maybe more importantly, standing water breeds mosquito larvae. I've seen ~100 larvae in a single bucket of water. It's bad to breed mosquitoes because this increases insect populations, and it doesn't seem like mosquitoes are displacing other insects that would have existed (or at least not very much). The blood that mosquitoes eat probably would otherwise be decomposed by predators or microbes. Mosquitoes also eat sugar from plants, which might prevent other bugs from eating it, but this consideration is probably not worth the cost of increased mosquito populations. Fortunately, reducing mosquito breeding grounds is a cause that everyone can get behind whether they care about insect suffering or not.


Respiration converts stored plant energy to CO2 and H2O but powers sentient creatures in the process. Burning wood also converts stored plant energy into CO2 and H2O but does so without supporting appreciable amounts of sentience, which is great.d Wood biomass has accumulated over centuries and would have otherwise fed termites, decomposers, etc.

The main possible downside of burning wood is that it contributes to climate change by releasing stored carbon, so if climate change is net bad for wild animals in the long run (which isn't clear but is plausible) then this would be unfortunate. That said, insofar as firewood replaces other fuels, this possible cost would be somewhat offset.

Another small downside is that if firewood still contains bugs when it's burned, then some bugs will die in one of the worst ways imaginable.

Forest fires

This section has moved here.

Mountaintop removal mining

Mountaintop removal mining reduces plant cover on mountain tops, which seems good. The rubble that gets moved elsewhere may cover even more plants. Reclamation that restores plant cover is probably bad.

Mountaintop removal mining increases the amount of fossil fuels available to burn, contributing to climate change, as well as air pollution. Mining itself creates significant water pollution. How these impacts affect wild-animal suffering could be explored in greater depth.


Some vegans eschew honey, and one motivation is that honey production can harm bees. For example: "Some beekeepers kill off their hives before winter. This practice can make economic sense."

Bees may be accidentally crushed by beekeepers. For example, on this thread, when one person asks, "How do you best avoid crushing bees when you put your inner cover and top cover back on?", some replies include "Its gonna happen" and "Its one thing I don't enjoy. I just hate the crunching sound."

Even the most humane bee farmers still inflict great suffering on their bees by causing them to be born into inherently short lives that typically end with painful deaths.

For these reasons, I generally try to avoid honey.

It's possible that the biggest animal impacts of honey occur via effects on plant pollination.

"Managed honey bees" contribute $14.6 billion to the US economy via improved crop yields. According to USDA, "In the United States, pollination contributes to crop production worth $20-30 billion in agricultural production annually." They add: "An estimated one-third of all food and beverages are made possible by pollination, mainly by honey bees."

Hughes Honey Apiary reports that "Honeybees are responsible for pollinating approx 80% of all fruit, vegetable and seed crops in the U.S". Furthermore, wild bees may not be good enough:

In most areas there are not enough feral honeybee colonies to provide adequate pollination. The feral bees that are available are often some distance from the farm field. In inclement weather they only fly short distances and therefore would not reach the crops.

Farms wanting maximum pollination require honeybee colonies from apiarists to help guarantee adequate bee populations in their fields.

In the wake of colony collapse disorder for farmed bees, there's debate about whether wild bees can pick up the slack. Gretchen LeBeun says "There have been some studies that show that where you don’t have honey bees, if you have enough habitat for regular bees, they are able to provide full pollinator service for some crops." On the other hand, Rusty Burlew explains:

perhaps one day native pollinators will shoulder the bulk of our pollination needs -- but it won’t happen within our current system of agriculture. It can’t. Successful transition to native pollinators will require nothing short of a complete overall of our current farming system. [...]

In the “old days,” let’s say before the end of WWII, people who kept honey bees kept them for honey. And if you didn’t keep bees, you didn’t worry about pollination. [...]

But the Green Revolution changed how we farm and, before long, there weren’t enough native pollinators to do the job. The fields were too big, the habitat was too scarce, and pesticides were everywhere. As farms got bigger and more mechanized, honey bees had to be trucked in along with other forms of migrant labor.

Even the people who are currently studying native pollinators concede that without significant changes, native bees might supplement -- but not supplant -- honey bees. Some experts estimate that up to 30% of the farmland would have to be converted to bee habitat. Hedgerows, borders, and habitat strips would have to be interspersed with crops. This reserved land would need to remain un-tilled and be planted with large numbers of flowering plants so that something was always in bloom.

Given the current state of affairs, it seems that buying honey may increase the efficiency of crop cultivation (which is plausibly bad) by increasing farmed-bee populations. In general, increasing plant productivity per hectare is bad. This would further vindicate the vegan stance on honey in a roundabout way. Of course, there are many additional variables at play, including what kinds of crops would be grown without as many bees, the effects of food instability on digital suffering in humanity's long-term future, and so on. I have only begun to explore this question and so don't have solid opinions here.

Note that native plants aren't as much affected by bee populations in the Americas, since honey bees are not indigenous to the Americas.

Dioxins and other pollution

Dioxins reduce animal fertility. This can be good, but it may be bad in the case of, e.g., large mostly-herbivores (including deer, pandas, and humans) that dispose of lots of plant energy with relatively little suffering. Dioxins also cause other painful defects, cancers, etc.

In general, it would be interesting to explore the net impact of various forms of air and water pollution, especially those forms of pollution that result from industries where consumer demand can make a difference to the quantity of product supplied.

NOx and NOy

This paper says:

Vegetation exposure to ozone reduces photosynthesis, growth, and other plant functions. Ozone formation in the atmosphere is a product of NOx, which are also a source of nitrogen deposition. Reduced carbon sequestration of temperate forests resulting from ozone is likely offset by increased carbon sequestration from nitrogen fertilization. [...]

While a primary effect of NOx is to increase soil fertility, it also has negative effects on vegetation due to acid rain and loss of stomatal control by direct absorption into leaves [...]

The results of this analysis show that, in all cases, the benefits of NOy deposition on carbon sequestration outweigh the negative effects of ozone on carbon sequestration in temperate forests.

What this paper calls "benefits", I would probably call "costs", since more forest biomass generally means more food to create larger invertebrate populations.

However, the paper adds (p. 792) that on crop fields that are already fertilized with nitrogen, NOy reductions won't reduce productivity (since there will still be enough artificial nitrogen fertilization), but they will reduce ozone damage to the plants. Hence NOy reductions may increase crop yields, which may increase insect suffering.

Acid rain

Acid rain can somewhat reduce plant growth, depending on its severity.

This page claims:

acid rain can cause phytoplankton in lakes to die. Insects, which rely on phytoplankton for food, now have less food to eat, and they begin to die as a result. [...] although acid rain may not directly affect a certain species of plant or animal, it can affect the entire food web by limiting the amount of food available.

This page reports the opposite:

An important experiment was performed in a remote lake in Ontario, in which sulfuric acid was added to slowly acidify the entire lake, ultimately to about pH 5.0 from the original pH of 6.5. During this whole-lake acidification, [...] there was a small increase in algal biomass after acidification because of an increased clarity of the water. [...]

the abundance of zooplankton increased by 66-93% after acidification, a change attributed to an increase in algal biomass. [...]

Many studies have been made of the physiological effects of acidification on fish. Younger life-history stages are generally more sensitive than adults, and most losses of fish populations can be attributed to reproductive failure, rather than mortality of adults (although adults have sometimes been killed by acid-shock episodes in the springtime).

Population reductions due to reproductive failure seem relatively good from the perspective of wild-animal suffering. However, the increases in zooplankton numbers are troubling from the perspective of zooplankton suffering.

This paper reports mixed effects of acid rain on productivity of fresh-water bodies:

In spite of geomorphological differences between acidifying systems and discrepancies in methods of plankton analysis, this review allows some general conclusions. [...] There are no clear indications for changes in the biomass and primary productivity of phytoplankton, however, the biomass of attached algae tends to increase. The scarce data on zooplankton indicate a decrease in standing stock.

Rule of law

In general, better law enforcement may improve conservation? For example:

On paper, environmental laws in the Brazilian Amazon are among the world’s most stringent. Landowners are required to keep 80 percent of their land forested, but lack of law enforcement has undermined this regulation, while economics and politics have conspired to thwart efforts to slow deforestation on the Amazon frontier. For environmental groups used to working under the rule of law, success on the frontier is an enigma.

Of course, rule of law also helps economic growth, which can lead to more resource use down the road.

Property rights may increase incentives for habitat conservation, which could be bad for wild animals. Property rights help curb tragedies of the commons. Stronger property rights would probably help preserve rainforest. In Brazil:

[in the 1990s], eastern Mato Grosso was a frontier in the truest sense of the word—a region where armed land invasion was rife, conflict between Indian tribes and outsiders raged, disputes were settled in blood, and law enforcement was only an abstract concept. In other words, a land without governance. The circumstances perpetuated a forest-clearing bonanza[.]

Rancher John Carter said of the Brazilian rainforest: "One of the main drivers to deforest is to maintain property rights" against illegal land-grabbing.

More stable countries allow people to have lower future discount rates, which increases incentives for sustainability.

Family planning

Population trends have significant impacts on far-future trajectories that should be evaluated. But we can also ask whether human environmental impacts, in aggregate, are good or bad for wild animals in the short run.

On the face of things, the effect of humans seems to have been positive on balance. If climate change is net bad and is sufficiently bad, there's a chance this assessment could flip around.

Presumably increasing access to birth control slows human population growth, which may increase wild-animal suffering. Family planning also tends to increase per-capita income, but my guess is that even if increased per-capita income reduces wild-animal suffering, this effect doesn't offset the wild-animal effect of slowing human population growth. Here's one illustration of this intuition. The next section of this piece ("Reducing poverty") contains a figure (called "Table S1") showing that per-capita biomass consumption per year is 582 kg in Africa and 1070 kg in Western Industrial countries. Meanwhile, nominal GDP per capita according to a 2016 estimate was US$37,477 in North America but only US$1,809 in Africa, a difference of 21 times. A naive reading of these numbers is that you have to increase per-capita income by 21 times to less than double biomass consumption. And family planning that halves the future human population size probably doesn't by itself increase per-capita income by 21 times.

Another reason to think that the environmental effects of increases in per-capita income won't offset the environmental effects of population reductions is that most experts believe that family planning reduces humanity's environmental footprint, and this wouldn't be true if the income-increasing effects were stronger than the human-population-reduction effects. (This argument was inspired by a comment from Jiwoon Hwang.)

Reducing poverty

If, as seems plausible, the net impact of human activity tends to reduce wild-animal suffering (see previous section), then it could be that alleviating poverty also alleviates wild-animal suffering. That said, there are some additional factors to worry about:

  • Anti-poverty advocates sometimes argue that poverty reduction won't hurt the environment on balance because richer families have fewer children. Could a reduction in birth rates more than offset the increased per-capita environmental impact that poverty reduction entails?
  • The hypothesis of an environmental Kuznets curve postulates that as wealth continues to increase, environmental impacts actually decline. If this is true, then even if poverty reduction increases per-capita environmental impact now, it might also more quickly bring about a situation where per-capita environmental impact starts to decline.

Some examples where poverty seems to increase environmental impact are

  1. Burning wood for fuel, especially in Africa. While this carries an immense human toll, it accelerates deforestation and eliminates a lot of stored plant energy in a non-sentient way.e This paper reports the (perhaps surprising) finding that global per-capita consumption of biomass products actually declined from 1961 to the early 2000s. "This is related to a shift in the patterns of the use of biomass products, above all reductions in the consumption of fuel wood."
  2. Slashing and burning the rainforest.f This also eliminates a lot of stored plant energy and significantly reduces future animal populations on the affected land.
  3. Desertification (in some cases).
  4. Stronger rule of law may reduce environmental destruction, and poverty reduction tends to improve rule of law.
  5. Crop yields per hectare are often much lower in poor countries, which might mean that fewer total insects can be supported by the food produced on a given area of cropland?

Moreover, climate change is an environmental impact that primarily results from industrialization, and the sign of its net impact on wild animals is unclear (and plausibly negative). So reducing poverty might exacerbate this potentially bad anthropogenic impact.

Here are some of many examples where long-term increases in a society's wealth lead to reduced per-capita environmental impact:

  • in-vitro meat
  • indoor and underground farms
  • water- and air-quality regulations.

This piece says "There are some indications that we humans are starting to [... use] technology and ingenuity to shrivel our environmental impact even as living standards keep rising."

This paper includes the following table showing biomass consumption per capita for different world regions in 2005.

You can see that richer Western countries have somewhat higher per-capita biomass consumption, even after adjusting for net exports. That said, the difference isn't huge and is much less than differences in wealth.

On the other hand, the same paper also suggests an inverse correlation between wealth and biomass consumption over time:

This paper explains:

During the last century, total human appropriation of plant growth [HANPP] has almost doubled [...].

Although HANPP grew, it grew at a far slower rate than population and the economy. During the same period, population grew by 274% and gross domestic product (GDP) (in constant 1990 dollars) grew by 1,655% (22). HANPP per capita has therefore declined from a world average of roughly 3.9 tC/cap per y to 2.3 tons of carbon per capita and y.

Shade balls

Shade balls inhibit sunlight penetration into reservoirs and thereby limit growth of algae and bacteria. This seems plausibly good insofar as reducing primary productivity generally reduces animal populations, but the net impact of eutrophication in general is unclear. The environmental effects of producing the balls would also be worth studying.

Road salt

Salt inhibits plant growth, so runoff of salt applied to icy roads seems good. Salt certainly causes short-term suffering to wildlife, but this is probably outweighed by long-term population reductions.

Satellite monitoring

"Big data" has allowed for better tracking of illegal/unreported fishing and Brazilian rainforest destruction. Insofar as this helps with habitat conservation, it's plausibly bad. But this trend doesn't seem like an effective leverage point for people to push on; this technology will be increasingly used for lots of other reasons as well.

Invasive species

It's very unclear in the general case whether invasive species increase or decrease total wild-animal suffering. Evaluation should be made on a case-by-case basis.

With regard to plant invaders, maybe one could make a case that the invading plants are on average more productive than the native plants(?) (as evidenced by the fact that the invaders outcompete the native plants?), which would increase total primary productivity and hence total animal populations. But this can easily be untrue in many cases -- e.g., invasive ivy may contribute to killing trees, and trees generally have higher productivity than other plants.

Also, maybe one can make the argument that disrupting ecosystem equilibria generally reduces total ecosystem productivity?

How about invasive animals? Even if one animal species is more productive than another, that may just mean that it eats more food that would have fed other animals, leaving the impact on total animal populations ambiguous. Maybe it's generally good to have invasions by large herbivores if they displace greater numbers of smaller herbivores.


Take a look at Wikipedia's List of environmental issues to see all sorts of possibilities:

We should triage this list based on which environmental impacts affect the most animals and how much humans can do to decrease or increase these environmental impacts.

How can we reduce plant populations?

Changes that primarily affect animals run the risk that by decreasing one species, you allow more of another species to take its place. If we instead favor policies that reduce plant populations in a long-term way, it seems like we can address the root of the problem. How can this best be done?

  • Road and building construction might help, though these have many additional effects as well (including roadkill).
  • Land salination due to deforestation, irrigation, etc. also reduces plant growth.
  • Elimination of predators that eat big herbivores? See the next section.

Should we eliminate predators?

Short answer: I think predators of large herbivores are generally bad. It's less clear how to evaluate a situation where predators eat lower-level predators.

In a talk on re-wilding, George Monbiot explains how the reintroduction of wolves to Yellowstone National Park allowed for an explosion of wildlife because plants that had been eaten by the elk were able to come back. Some claim that the ecological impact of wolf reintroduction may be exaggerated. But reintroduced wolves do eat something, often large herbivores.

The "green world hypothesis" notes that less than 17% of the Earth's net primary production is eaten by herbivores. It postulates that this is due to a variety of limiting factors, including predation.

In general, large herbivores are an efficient way to eliminate plant energy without too much suffering in the process, both because the herbivores are long-lived and more K-selected and so experience less suffering per unit time but also because they're just big, so there are many fewer total animals to suffer.g So it seems plausible that eliminating wolves, bears, and other predators of herbivores may indeed be one of the best things we can do to reduce wild-animal suffering. To skeptics who ask us, "What are you going to do? Stop lions from eating gazelles?" we can say "Yes." :)

This book reports that

In Canada at Jasper and Banff National Parks as well as in Yellowstone National Park in the U.S.A., concerned range managers frequently demonstrate wild ungulate overgrazing of the natural grasslands. Overgrazing has predisposed these temperate natural grasslands and shrublands to reduced biomass productivity, [and among other things ...] increased soil erosion.

Soil erosion and reduced productivity both mean less plant food for small animals like insects and hence probably less future small-animal suffering.

This page adds: "Studies have shown that carnivore density is correlated with ecosystem productivity and services like the health of soil, water, and vegetation."

The usual concern raised with eliminating predators is: What about prey overpopulation? Won't that lead to a population boom and then a population bust, resulting in more total suffering through starvation? First, remember that more prey is a good outcome if it means fewer smaller animals eating the same food. But yes, we might not want prey to grow exponentially. In this case, we could introduce contraception or sterilization in place of predation. After all, contraception is the method humans have chosen for themselves to control their numbers. (Sadly, some nature-loving animal advocates prefer for herbivores to be eaten alive rather than having populations controlled by contraception: "contraception and neutering in nature is a treacherous bargain, for it puts coyotes and bobcats out of their natural evolutionary work. Do we really want a patented, FDA-approved pharmaceutical plan to control the destinies of untamed animals?")

Jennifer Everett claims:

Predation intervention would surely cause frustration and suffering for predators in the short term and in the long-run would very likely require human hunting of prey species whose populations grew beyond the carrying capacity of their habitat (both of which are of obvious concern to animal welfarists). It is reasonable to conclude that the state of affairs in which predation occurs in nature without regular interference from humans is better than that in which we attempt to prevent it.

But this is a false dilemma. A better option than either allowing predation to continue as is or painfully killing predators would be to sterilize predators. And rather than let prey overpopulate, one could sterilize prey as well, although reducing herbivore overpopulation would help preserve primary productivity, which I'm nervous about.

There are potential downsides to eliminating predators:

  1. Sometimes eliminating top predators can increase populations of lower-level predators, and this may actually reduce prey populations. Watching out for this possibility would be important.
  2. We should also study more whether it's actually the case that large herbivores reduce insect populations on balance. For instance, Allan Savory has claimed that proper livestock grazing in Africa can actually improve soil fertility and reduce desertification, but his thesis has been widely rejected by many scientists. Moreover, even if Savory is right, it seems clear that overgrazing by a population boom of herbivores in the absence of predators should reduce primary production and hence food for other animals.
  3. In general, herbivory may either decrease or increase nutrient cycling,h although it's not obvious that big herbivores do this more or less than small insect herbivores. In the discussion of whales above, it seemed that big whales cycled more nutrients than small ocean critters could, but intuitively there seems to be a substantive difference between oceans -- where it may take a big animal to transport nutrients over large distances -- versus land, where flying insects can also spread nutrients pretty well. Another consideration is that large land herbivores feed lots of insects via their poop and dead corpses, whereas the grass that those herbivores eat may be less juicy to insects. This is worrisome, but it's worth remembering that even if insects don't eat grass much when it's alive, detritivores like earthworms and millipedes may eat lots of decaying grass (or the bacteria that are decomposing said grass). It's plausible that the total amount of food that bugs get from herbivore poop based on an herbivore eating X kilograms of grass is less than the food that bugs get from eating X kilograms of rotting grass itself (since the herbivore should have dissipated lots of the food energy as heat), but this seems important to investigate further. For instance, if non-bacteria bugs eat ~20% of all poop but only ~1% of all rotting grass (just making up numbers for illustration), then the total amount of non-bacteria fed by poop might be higher than from the equivalent amount of grass.
  4. Loss of apex predators may contribute to climate change. Vegetation can sequester carbon, so less vegetation means less carbon sequestration. That said, a lot of sequestration occurs through a plant's roots, so as long as plants aren't completely killed by being grazed, they may still sequester carbon. Plus, animal hooves may stomp old grass into the soil to increase sequestration.


Hunting of herbivores is probably very bad for the same reasons (listed above) that predators of herbivores are probably very bad. Indeed, hunting amounts to predation by humans. Hunting of predators, while tragic for the injured animals, may have good consequences overall due to reduced predation, though this depends a lot on the details.

I find it interesting that many animal supporters oppose hunting but support predation in the same breath. One example appears in this podcast episode (beginning at 54:13), in which vegans discuss how hunting of wolves increases wolves' stress and disrupts their social structures. That's certainly tragic for the wolves. But remember that wolf predation increases the stress of prey and systematically changes prey behavior. And many prey animals suffer these effects for every one predator. Mainstream attitudes on this topic involve an amazing feat of doublethink, in which "humans killing animals" is bad, but "animals killing animals" is good.

Hunting can also be an incentive for habitat preservation, which is bad. For example, an important source of funding for prairie preservation in one area came from a hunting-conservation organization, Pheasants Forever. "Wildlife stamp" programs require hunters to pay a small fee that goes toward habitat conservation.

Humane insecticides

  • Which ways of killing crop pests cause the least suffering per death?
  • Are chemical insecticides net beneficial in the sense that they prevent more painful deaths than they cause in the long run by reducing insect populations?
  • Are there practical ways to lobby for more humane pest-control methods without decreasing pesticide use?

What is it like to die as a 2-day-old insect?

Dying hurts (usually). And I think one of the biggest reasons why there's so much suffering in nature is that most species give birth to dozens, hundreds, or tens of thousands of offspring, most of which die a few days or weeks after birth. Even if the quality of life for an insect that reached maturity were positive, the large numbers of offspring that die young would more than offset the balance.

It would help to dig deeper into the details here. For example:

  1. How many offspring do different species give birth to, on average? (Yew-Kwang Ng has a few figures in "Towards Welfare Biology.")
  2. How many of those offspring are sentient before they die? (I assume that many of them get eaten or fail to thrive while still eggs, which may or may not suffer.)
  3. How do the babies die? What fraction are eaten? Killed by parasites? Overcome by bacteria? Dehydrate? Starve? Freeze? What else?
  4. How painful are each of those ways of dying?

Simon Knutsson has explored some of these questions for several kinds of insects in this piece.

Could we reduce r-selection?

Dying (through predation, parasitism, disease, cold, etc.) presumably constitutes a large fraction of the pain of a short life, because the life itself is not long enough to have large amounts of suffering of other types. Hence, to a decent approximation, our goal may be to reduce death rates rather than sheer populations of small animals. Death rates are higher per capita in more r-selected species that have more offspring per parent pair. In general, greater ecological instability favors r-selected organisms, suggesting that, ceteris paribus, if an ecosystem is going to exist, it's best for it to be stable.

We should explore further: What sorts of environmental policies would reduce r-selection? This would be a way to reduce suffering even without eliminating a habitat entirely.

Interestingly, temperate regions tend to have higher r-selection than tropical ones because the winter period incentivizes organisms to die quickly in the fall and leave eggs to survive until spring. In the tropics where winters are not harsh, organisms can live through the whole year without disruption. Of course, tropical regions also have more total organisms (including r-selected insects) and hence presumably more total suffering than temperate regions, but insofar as tropical regions are less r-selected, the comparison may be less dramatic than it seems at first glance.

Reducing the length of trophic chains

Consider these simplified trophic chains, with "->" meaning "is eaten by":

  1. Grass -> deer -> bacteria
  2. Grass -> deer -> wolf -> bacteria
  3. Phytoplankton -> zooplankton -> minnows -> medium-sized fish -> big fish -> shark -> ocean bacteria.

For a given amount of plant energy at the beginning, longer trophic chains often mean more animals get eaten before the energy fully dissipates or feeds decomposing bacteria. That means more suffering from painful deaths in the jaws of predators. (Thanks to a friend for this point.) For instance, if deer are eaten by wolves rather than left to decompose, their energy powers more total non-bacterial organisms, which increases total suffering (assuming that even large masses of bacteria aren't disvalued as much as big animals).

Of course, the food chains are more complex than what's depicted above. Decomposing deer are also eaten by scavengers and bugs. So another way to draw the food chain might be

Grass -> deer -> bugs -> small birds -> big birds -> bugs -> ...(repeat until energy fully dissipates).

On its face, this looks potentially worse than when the deer was eaten by the wolf, since the deer's corpse powers many small bugs compared against a few weeks of life for a single wolf. But I think this conclusion is shortsighted, because more wolves also means fewer total deer, which means more bugs eating the food the deer would have eaten, which is probably worse than having some extra deer corpses feeding some extra bugs.

This page explains that

Habitat fragmentation is known to also affect specialized populations of higher trophic levels, for example, in a plant-herbivore-parasitoid food chain. [...] Theoretical models and empirical evidence show that specialized parasitoids (and predators) suffer even morei, so that food chain length tends to be shortened and herbivores tend to become released from possible control of their natural enemies.

So habitat fragmentation may be one way to reduce the length of trophic chains. However, while I probably support shortening trophic chains with regard to large herbivores, I'm more uncertain about it in the case of insect predators, because reducing insect predators increases populations of insect herbivores. While I probably support having more big mammalian herbivores insofar as they crowd out insects, I generally oppose having more small insect herbivores.


  1. Actually, the total amount of suffering may be most related to the number of deaths per unit time, assuming that the painful process of dying is one of the worst experiences an organism goes through. In general, number of deaths per unit time is directly proportional to the population size, though not always. For example, an environmental change that doesn't affect populations in the long run but leads to greater r-selection owing to ecological instability might imply more total suffering due to more deaths per unit time.  (back)
  2. This argument isn't convincing for several reasons:
    1. Most opportunities to cheaply reduce wild-animal suffering are "now or never" decisions. For instance, either we build this parking lot now or we don't. Of course, in theory, if society doesn't build the parking lot now, it could revisit the question in another century and decide that it should have built the parking lot after all. But by that point, building the parking lot might be an expensive undertaking.
    2. Society in the future will probably not care much about wild-animal suffering, so it's misleading to ask what we should do if society did care about wild-animal suffering enough to choose the optimal time to intervene to help wild animals. I will not be in charge in the future, and even if reducing wilderness proves to be the better choice upon further research, future humans probably won't do what I would have wanted. Hence, my actions are essentially "now or never", and I can't actually "keep the option open" to decide later, since I won't be around beyond a few decades. As an analogy, suppose the Democratic Party controls the US Congress but has lost the latest elections and will soon be replaced by a Republican Party majority. The Democrats are debating a bill that would be better informed by an ongoing research program. The Democrats think the bill is probably good by their lights, but they'd be more certain of that once the research is completed. However, by the time the research is completed, the Republicans will be in control and will kill the legislation. It seems the Democrats should pass the bill now rather than waiting to "keep their options open".
    3. I care primarily about reducing suffering rather than creating happiness, so from my perspective, it's overwhelmingly clear that wilderness contains net disvalue.
    4. It should be possible within a few centuries to create more artificial, digital happiness using the matter on Earth than whatever modest amounts of happiness the Earth's wilderness contains at present.

    Still, spreading concern for wild-animal suffering so that future generations can better act on this concern is a valuable form of keeping options open, since future people will have better insight into the relevant issues than we do, and to the extent they share our values, we should generally approve of what they decide.  (back)

  3. This is per calorie because you only eat so many calories. Amusingly, if crop cultivation is net good, then ignoring grocery expenses and potential global-warming effects, it's better for you to burn more calories so that you can eat more and take away the energy from insects. Of course, literally burning the food in a fire would be even easier.  (back)
  4. Ultraviolet oxidation is another form of non-sentient elimination of stored plant energy.  (back)
  5. This page reports:

    The number of small trees and shrubs that are either cut down or uprooted for fuel is enormous. In Jordan, it was estimated that Bedouins annually uproot 182 million fodder shrubs just for cooking and in the Sudan, 548 million acacia shrubs are annually destroyed for the same purpose (FAO, 1976). [...]

    Data regarding the extent of fuel gathering in Latin America and the Caribbean are limited. Stuart (1992) reported that each family in Guatemala's Petén region consumes 10 to 20 cords of wood a year, requiring hundreds of trees. Similar consumption rates are likely applicable in many other countries where fossil fuels are either not available or too expensive. Fuel gathering is an important factor contributing to Haiti's serious desertification problem.


  6. One can debate whether poverty is the main culprit here. This article explains that poor, moving farmers ("shifted cultivators") account for 60% of trophic-rainforest loss (according to this book). However: "Robert Repetto of the World Resources Institute ranks commercial logging as the biggest agent of tropical deforestation" because it not only causes direct destruction but also creates roads through the forest that are used by the shifted cultivators. "Most of the rainforest timber on the international market is exported to rich countries." So the rich help facilitate rainforest destruction by the poor. Since both poverty and affluence play key roles in this process, it's not obvious what the net impact would be of reducing poverty on the whole.  (back)
  7. One might also wonder whether the bodies of herbivorous mammals, being endothermic, eliminate more plant energy per unit of brain activity than other animals, thereby preventing more other animals from suffering using that energy. If brain activity is proportional to brain energy consumption, then this seems not to be the case, because: "There is [...] no significant difference in the percentage of energy supplied by the body to the brain between endotherms and ectotherms [24],[25]." If brain energy consumption is less efficient in endotherms, or if we care more about the total number of brains powered than about brain activity, then large mammalian herbivores would have less morally relevant mental activity per calorie than ectotherms. On the other hand, if we think sentience is directly proportional to brain energy consumption, then ectotherms would generally have less moral importance per unit of brain size.  (back)
  8. "Insect herbivory accelerates nutrient cycling and increases plant production"  (back)
  9. By "suffer", the text means that the populations decrease. In fact, the parasitoids and predators themselves suffer less in the sense of having fewer bad experiences, because there are fewer of them.  (back)