by Brian Tomasik
First written: mostly Feb. 2016; last update: 29 Jan. 2017

## Summary

This page reviews some ways in which humans kill bugs to create products -- focusing on silk, shellac, and carmine. The following table summarizes the findings:

 Product Approximate number of bugs killed for this product by the average consumer in rich countries per year (ignoring wild-insect impacts) How bugs are killed silk several hundred boiling, baking, drying, piercing with needle, steaming, or fumigating (most of these sound quite bad); however, it's not clear if silkworm pupae feel pain? shellac ~100 to ~1000?? not sure (maybe drying out, getting crushed, etc.) carmine ~100? crushing during harvest, hot water, sunlight, steam, or baking in oven (most of these sound quite bad)

These uses of insects are tragic and may serve as useful "hooks" for getting people interested in reducing insect suffering. However, the number of insects that can be affected per dollar of effort is probably smaller than for campaigns that reduce insect habitat.

## Silk

### How the bugs are killed

Different sources give different information, and presumably the technique varies from place to place:

• "Pupae within cocoons are killed by steam or fumigation to prevent adult emergence, which would cut and tangle the silk filaments." (source)
• "The silk farmers then kill most caterpillars by heat, leaving some to metamorphose into moths to breed the next generation of caterpillars." (source)
• "the insects are baked, dried, or boiled in the process of harvesting their silk." (source)
• "boiling, steaming or baking" (source)
• "They used to be boiled alive. Nowadays, the cocoons are placed in a dry, hot oven to kill the pupae." (source)

All of the following videos of silk production show the cocoons as being boiled, so I assume that's still the most common method of killing?

#### Wild silks

With wild silk: "The pupae are killed by either dipping them in boiling water before the adult moths emerge or by piercing them with a needle."

For Tussar silk:

In order to kill the silkworms, the cocoons are dried in the sun. There is a variation where the silkworms are allowed to leave before the cocoons are soaked in boiling water [...].[3][2]

#### Breeder deaths

after females lay eggs, they are immediately ground up to check for disease. As for the males, they are held in refrigeration, taken out only to mate. Once their ability to mate is diminished, they also are discarded and killed.

These deaths seems to me less bad than boiling and maybe comparably bad as dying naturally?

### Are silkworms in cocoons sentient?

Since silkworms are boiled alive while they're in cocoons, it's important to know if the silkworms are sentient in the pupa stage. Maybe during the metamorphosis process, silkworms don't have normal sensory reception because they're in an intermediate state? Similarly, when human brains are in an intermediate state of repair during non-REM sleep, they're also not sentient. Plus, silkworms are in a protective cocoon, which would seemingly make pain less important?

This and this video show silkworms inside their cocoons. At the beginning, each worm seems to move around in a similar way inside the cocoon as it did when it was a caterpillar. It then wriggles actively to shed its old skin. Since the resulting moth presumably has its neurons developed when it emerges, it's not implausible that those neurons would be mostly developed inside the cocoon too. However, it's difficult to know if the bug's nervous system is paying any attention to external stimuli or if it's like a human during non-REM sleep.

This study (discussed more here) found that moths that had been trained to pair a shock with an odor as fifth instars remembered to avoid the odor after metamorphosis, suggesting that neural connections remained intact even though many other body parts turn to "soup" during metamorphosis. This isn't too surprising, since:

Textbooks will commonly talk about the insect dissolving into a kind of “soup”, but that’s not entirely accurate. Some organs stay intact. Others, like muscles, break down into clumps of cells that can be re-used, like a Lego sculpture decomposing into bricks. And some cells create imaginal discs—structures that produce adult body parts. There’s a pair for the antennae, a pair for the eyes, one for each leg and wing, and so on. So if the pupa contains a soup, it’s an organised broth full of chunky bits.

However, this doesn't answer whether neurons behave as usual during metamorphosis.

When silkworms are in the pupa stage, they will not react to most external stimuli such as being poked with a finger or poked with a pin. They are in a sort of suspended animation. [...]

Bottom line is - nobody knows and nobody probably ever will know if silkworms feel pain or not.

All told, it seems unclear whether pupae can feel pain. And perhaps this depends on their stage of development. For example, I assume that very early caterpillars and almost-hatched adults can feel pain, while it's less clear for the intermediate stages? This piece says of silkworms: "As soon as [cocoon] construction is done, they [silkworms] are done in, and their cocoons are harvested, soaked, and unraveled." It seems plausible that silkworms would be killed within a day or two of when their cocoons have been completed, since there's no reason to wait longer? If silkworms are killed within hours of finishing cocoon construction, it seems likely that they would still be able to feel pain from external stimuli?

### Number of bugs per unit product

Here are some estimates of how many worms are killed per item of clothing:

And here are estimates of worms killed per kg of silk:

• 15,000
• 3000
• 4000 to 7000 (this is based on an estimate of 2000-3000 per pounda)
• "Those numbers vary as well, but the average range given was 4,400-14,969 per kilogram" (source).

Let's say the actual number is somewhere in the middle: 5000 silkworms per kg.

### Annual production

FAOSTAT shows global production of silk:

The database shows that 95.5% of this production is in Asia.

The International Sericultural Commission reports similar numbers. Its 2013 world-production figure is 159,737.10, and although units aren't given, this is presumably metric tons, because it's close to the FAO 2013 number.

This page reports a lower number for world production, even though it also cites FAOSTAT: "At least 70 million pounds of raw silk are produced each year, requiring nearly 10 billion pounds of cocoons."b 70 million pounds is only 32,000 metric tons, which is several times lower than the ~168,000 from the FAOSTAT graph above. On the other hand, 10 billion pounds is 4.5 million metric tons, which is too big. This post notes that the "70 million pounds" number has "no discernible citation" even though it's widely quoted. Maybe the number is just very old and refers to a time decades ago when silk production was much lower?

In 1997, Italy imported some 3200 tons of raw silk and over 700 tons of silk yarn, primarily from China. Italy also imported about 300 tons of ladies' blouses, of which over 80% came from China.

Even though raw and processed silk are probably not equivalent, I'll naively add these numbers: 3200 + 700 + 300 = 4200 (metric?) tons of silk. And since global silk production essentially ~doubled between 1997 and 2013 according to the above FAO graph, maybe this figure is as high as ~8400 today? (I wasn't able to find more recent statistics on Italy's imports.)

The amount of wild silk produced (in India? or worldwide?) is quite small: only 2,075 (metric?) tons. This document says (regarding India?) "Estimated annual production of tassar silk [a type of wild silk] is 130 tons. Production of other types of silk exceed 10,000 tons." This article reports: "India produces over 1,500 metric tons of wild silk annually."

### Number of bugs killed per year

The FAOSTAT figure of ~168,000 metric tons per year implies

(1.68 * 108 kg silk per year) * (5000 silkworms killed per kg silk) = 8.4 * 1011 silkworms killed per year.

### Per-capita bugs killed

This page says "The major silk consumers of the world are; USA, Italy, Japan, India, France, China, United Kingdom, Switzerland, Germany, UAE, Korea, Viet Nam, etc." Another page says "India is also the largest consumer of silk in the world." Suppose about, say, ~1.5 billion people (I'm just making this number up), many in rich countries, consume most of the world's silk. Then on average, one of those people kills (8.4 * 1011) / (1.5 * 109) = 560 silkworms per year.

How about Italy specifically? (8400 metric tons) * (1000 kg / metric ton) * (5000 silkworms killed per kg) = 4.2 * 1010 silkworms killed per year. According to Google, Italy's population is ~60 million. So on average, Italians kill (4.2 * 1010) / (6 * 107) = 700 silk worms per year. Given that "Italy has been traditionally the largest importer, processor and exporter of silk products in Europe", the per-capita number of silkworms killed across industrialized countries in general is probably somewhat lower than this.

Often, silkworms that have been boiled are later eaten. So this is an additional end product of silk production, and the money paid for silkworm food very slightly contributes to silk production as a whole. But assuming the price of this food is trivial compared with the price of the silk itself(?), this shouldn't be a very big contribution to the overall incentives for producing silk.

### Cost-effectiveness of opposing silk

Suppose that a US$1 million anti-silk campaign in Italy could reduce Italy's silk consumption by 1% for only 5 years into the future. Then it would save (4.2 * 1010) * 0.01 * (5 years) / ($1 million) = 2100 worms per dollar, or about 21 worms per penny.

This piece estimates that there are ~8.1 adult-equivalent deaths per insect-year for a typical wild-insect population. I think this number might be a bit low, since it's based on the southern green stink bug, which has a relatively low number of eggs per female. Let's say the actual number is that an insect-year of suffering comprises the pain of ~10 adult-insect deaths.

That piece also estimates that burning is ~3 times as bad as a typical death. Let's assume the same is true for boiling.

Let's assume that there's a 50% chance that silkworms are able to feel pain when they're boiled.

Finally, I think silkworms matter a lot more than the average insect because they're so big. Based on this discussion, I'll assume a silkworm pupa matters, say, ~30 times as much as a typical insect.

Then we can convert the cost-effectiveness of the anti-silk campaign into an equivalent number of insect-years of wild insects prevented, in order to make this campaign comparable with interventions to reduce wild-insect suffering:

(2100 silkworms saved from boiling per dollar) * (30 regular insects per silkworm) * (3 regular deaths per boiling death) * (0.5 chance of pain) / (10 regular deaths per insect-year) ≈ 104 insect-years prevented per dollar.

This calculation is somewhat conservative because it only counts the pain of killing silkworms, not additional suffering that these caterpillars may experience during their lives. I personally think silkworm suffering during a short life is probably less significant than the pain of death, but other people disagree on that judgement call. The lives of domesticated silkworms may not be that bad in general, since domesticated silkworms are less likely to die of hunger or predation than their wild counterparts.

This calculation also ignores impacts on wild insects, since I haven't analyzed those in detail, but that consideration could also matter quite a bit.

### Historical replacement by synthetics

Worm-based silk has been partially replaced by artificial silk and especially nylon. Other silk alternatives include mercerized cotton, polyester, and polyester-cotton blends, as well as "rayon, [...] milkweed seed pod fibers, silk-cotton tree and ceiba tree filaments, [... ]tencel, and lyocell (a type of cellulose fiber)."

### Considerations regarding wild-insect suffering

While the immediate impact of silk production on silkworms is most salient, a complete analysis of silk vs. other products should examine impacts on wild insects as well, since these might be quite significant. To evaluate impacts on wild insects, we should ask what kinds of vegetation and insect densities would be present if silk weren't being grown, and what environmental impacts alternative fabrics have.

Most commercial silk is mulberry silk, whose silkworms are domesticated and raised indoors. Google Image pictures of silk farming confirm that most farms are indoors. This is weakly positive insofar as buildings prevent plant growth on the land that they occupy. However, the density of silkworms in these farms is very high, probably higher than the density of insects would be in the wild? (That said, this is partly because the worms eat food brought to them from elsewhere.)

Wild silk doesn't cover over ground, which is bad. If the plants that feed wild silkworms are cultivated, then wild silk still has an impact by replacing some native vegetation. I haven't looked into whether this change seems good or bad on balance for wild insects. But insofar as silk farmers want to maximize insect (silkworm) density, it's plausible that the vegetation changes done to grow wild silk are net bad. (This is in contrast to many forms of crop cultivation, where farmers want to minimize insect densities -- or at least, densities of pest insects.)

Production of synthetic fibers in factories, like indoor sericulture, also covers over land, which is good, although maybe these factories are so efficient that this effect is negligible. Synthetic production may also involve chemical pollution, which might reduce wild-insect populations, though I don't know for sure. On the other hand, regular silk production also uses lots of chemicals, so I don't know if one or the other option has the advantage here.

This pro-silk page claims that the silk industry "contribute[s] to preserv[ation of] the bio-diversity of rural areas and natural vanya food plant forest areas." If true, this may be bad, since preservation of rural areas preserves wild-animal suffering. The page also says that silk helps reduce poverty, and this article agrees:

Sericulture is a livelihood that is available to many of the rural poor of India. It especially provides work for a higher proportion of women than other industries.

Reducing poverty is plausibly a good effect of silk, both for the people affected and because poverty reduction may prevent fewer wild animals from being born, though this isn't clear. However, higher female incomes may reduce birth rates and thereby reduce long-term human population growth.

#### Silk vs. cotton

Another important factor for assessing the wild-insect impact of silk is how many hectares are required to produce how much output.

One source says producing 1 kg of silk requires 104 kg of mulberry leaves. Another source says "One hectare of mulberry trees yields about eleven tons of leaves, producing around 450 pounds of cocoonsc, but just about 85 pounds of raw silk", which implies that (11 tons * 2000 pounds per ton)/(85 pounds) = 260 kg of mulberry leaves are required to produce a kg of silk. Since 85 pounds is 39 kg, the yield of silk seems to be about 39 kg/hectare.

In contrast, for cotton in 2013-14: "the world average yield is at 766 kg/hectare".

So silk production apparently requires much more land than cotton production per unit output. If mulberry farms reduce wild-insect suffering relative to native vegetation, this would be an argument in favor of silk. However, based on these pictures, mulberry plants seem reasonably productive to me, which makes me worry they may not reduce and might increase wild-insect suffering. This video shows apparently less productive mulberry fields in India. But even if mulberry farming reduces primary production relative to native vegetation, it may still increase insect suffering because most of that primary production is fed to insects, while in a natural field or forest, a lot of primary production would be eaten by less sentient bacteria or fungi.

Based on pictures of cotton fields, it looks to me like the primary productivity of cotton plants is pretty low, which is good because it means less production of plant energy to feed bugs.

#### Conclusion

I would guess that cotton cultivation is weakly positive with respect to wild-insect suffering (though I'm not sure). Mulberry farming to feed silkworms might be very positive or (perhaps more likely) very negative; the impact is bigger than for cotton farming because silk production requires so much land per kg of fabric produced.

I remain quite uncertain about the net impact of silk farming on wild bugs, and I think further research on this point would be important before investing non-trivial resources into an anti-silk campaign. However, it provisionally seems as though silk farming may be net bad for wild bugs because all of the harvested mulberry crop is fed to bugs (which creates large numbers of bugs), while in native vegetation or in cotton fields, much of the primary production would not feed bugs.

### Peace silk is not a good alternative

Many people promote so-called nonviolent alternatives to regular silk, known as "peace silk" or "ahimsa silk". There are different varieties. For example:

Peace silk requires waiting around 10 days for the larvae to grow into moths and to make their way out of their cocoon – compared to the 15 minutes it usually takes to boil them alive. [...] Once woven, nonviolent silk is hard to distinguish from normal mulberry silk. [...]

Another variety of silk known as “eri” silk – which accounts for about 12% of silk production in India – also spares the life of silkworms. But the quality is not as good as the mulberry sort, which is more expensive but shinier.

While peace silk avoids directly killing silk worms, it's arguably complicit in a larger number of indirect deaths. This article by Michael Cook is written from the standpoint of a conventional sericulture practitioner, but its points are very good. Cook explains that with peace silk:

The Ahimsa cocoons are all allowed to hatch and breed, and the silk is processed from the hatched cocoons. [...] The main issue that I have with this style of cultivation being vegetarian-appropriate, is that each fertilized female moth will lay between 200 and 1000 eggs, averaging around 500. In some strains, the eggs will require refrigeration - without refrigeration, the living embryos within the fertilized eggs will wither and die over the course of a month or two. If they are refrigerated, they will hatch upon removal from refrigeration, in which case they have to be fed immediately, or they will die of starvation and dehydration. Either process will require the destruction of approximately 200 - 300 embryos or hatchling silkworm per moth, for any amount that exceeds what is required for the next crop. Instead of killing one pupa for the silk of the cocoon, it kills hundreds of caterpillars.

To me, destruction of eggs well before hatching doesn't sound bad. How common is this, and how common is it to let the eggs hatch?

Cook continues:

In India, where the vast majority of Ahimsa silk is being raised, most silkworm strains are multivoltine. This means that the silkworms do not undergo refrigeration, and the eggs will hatch approximately two weeks after being laid. The ones that are not fed will die within a day of hatching, from a combination of dessication and starvation. In a batch of, say, 20,000 cocoons, this means that the next generation (if they were all raised) would be two and a half million, and the generation after that, three hundred twelve million. It's just not possible to feed so many. While it may be true that the individual caterpillar that spun the cocoon didn't die inside it, its offspring will have to be ruthlessly culled. Is it considered more virtuous to create conditions of wholesale starvation, to avoid killing the pupa quickly with heat? [...]

My main issue with the Eri moth is the same as with multivoltine Bombyx strains; the eggs will hatch within two weeks or less, and the majority of them have to starve to death. In actual practice, most Eri cocoons are cut open and the pupa is tipped out; this avoids having the moth soil the cocoon while hatching out from the pupa. Some of the cocoons are allowed to go on and hatch, but they're also a favorite food source for some tribes in Indiad, and the pupae are occasionally used for fertilizer. [...]

The specific processing of Tussah / Tasar cocoons varies depending on the particular species and the area where they're raised. Most are simultaneously killed and dried with hot air, although some are kept green (alive) because they store better, and then killed at the time of reeling with steam or boiling water. The idea that Tussah cocoons are plucked from the pristine jungle after the moths have hatched is unfortunately just not supported by facts. [...] There are some companies in India making "Ahimsa" silk using cocoons that have hatched for breeding - but nearly every one of that moth's offspring are killed for reeling. Again, I don't see how that makes the hatched cocoons morally superior to the others - they're part of the same agricultural system. It's like only taking milk from female cows that are kept to breed beef cattle.

If wild silkworms grew naturally and the only thing harvesters did was to harvest their cocoons after they had emerged, then wild peace silk would not be much more harmful than not buying silk at all, except maybe for the fact that people crush bugs when they walk in the fields/forests to pick the cocoons. But if wild silkworms are cultivated by making the wild plants more favorable to high silkworm densities (is this the case?), then wild peace silk would encourage cultivation of insect-dense wild plants, which seems a priori bad, although I don't know what kinds of plants would grow otherwise.

This page acknowledges the problems with peace silk but is strangely more supportive of wild silk:

Because the silkworm is allowed to escape its cocoon unharmed, Peace Silk sounds like the humane solution. And to some, it is.

But even Peace Silk is not perfectly humane. You see, after emerging from the cocoon, moths mate and the females each lay hundreds of eggs. But silk farms have a limited supply of food and they can't feed all the hatchlings.

So although the silkworm moths have emerged safely from their cocoons and will die naturally, most of their offspring will die from starvation or dehydration within a few days of birth. Since that's not much better than being boiled alivee, animal activists consider peace silk as inhumane as conventional silk. [...]

Wild silkworms tend to produce a higher quality of silk than non-wild (domesticated) silkworms, because of their natural diet. And when permitted to leave their cocoons naturally, wild silk is also considered a more humane alternative than regular peace silk. [...]

while a wild silk farm has greater natural resources than a fully cultivated silk farm, it still cannot support millions of silkworm offspring. The difference is that while some wild peace silkworms will die of starvation or dehydration as regular peace silkworms do, most wild peace silkworms would presumably be eaten by birds, ants and spiders in their farmed-yet-natural environment.

I personally don't think being eaten is less painful than starving, so I don't see wild silk as having an advantage here. Fortunately, that article recommends silk alternatives as the most humane option.

Animal rights group Beauty Without Cruelty has publicly accused one silk producer in South India of promoting silk as ahimsa silk when the process they witnessed was just a twist on conventional silk production. Beauty Without Cruelty says they did see cocoons from which the moths were allowed to emerge, but what happened afterwards to the moths called the ahimsa description into question.

"After emerging, the male and female moths are kept together for three hours to mate. The females are then segregated and placed in trays to lay eggs. The males are put in a refrigerator, kept semi-frozen, and trotted out repeatedly to mate. They are eventually thrown into a dustbin to die lingering deaths when their virility diminishes."

One final nail in the coffin of peace silk is the observation that if you buy peace silk, you'll increase its price, which may lead some other consumers who care about ethics but are also financially constrained to switch back to buying regular silk instead. Of course, this effect operates with all products to some degree -- e.g., by buying cotton, you slightly increase the price, thereby making a few consumers switch to silk instead, because the amount by which silk is more expensive than cotton is slightly reduced. But I assume this effect is more pronounced with peace silk vs. regular silk than with silk vs. other fibers? On the other hand, it's possible that buying more peace silk would allow for technological improvements that create economies of scale and thus lower long-term peace-silk prices?

I worry that an anti-silk campaign might mistakenly promote peace/wild silk as a good alternative to regular silk. We should be careful to avoid this outcome.

#### Innovative alternatives?

Neri Oxman explained in a TED talk how her team discovered that putting silkworms on a template would lead them to build silk over the template rather than building a normal cocoon. As a result, the worms didn't need to be boiled to extract the silk. However, presumably this approach runs into the same problem as with other forms of peace silk: There will end up being too many silkworms born to the next generation, and almost all of them will die. As Oxman said, her team's 6500 silkworms produced 1.5 million eggs. Could suffering be avoided by killing the eggs before they hatch? Even if so, the process would still require producing tons of silkworms that will die naturally at some point, often in painful ways.

This article describes research to eliminate use of silkworms altogether in producing silk, which would be an actual solution to the problem of silkworm suffering:

whether silk is grown in fields of transgenic tobacco, tubs of E. coli, or goat’s udders, it seems clear that silkworms’ and spiders’ monopoly is over.

## Shellac

### Uses

Shellac is an insect-derived resin with many industrial uses. In addition, it's eaten in small amounts in pharmaceutical and confectioner's glaze used on pills, candy, fruits, etc. Most hard, shiny candies use shellac coatings, though some candies don't contain shellac, including M&M's, Skittles, and Starburst.

### Number of bugs per unit product

It requires between 50,000 and 300,000 lac bugs to produce 1 kg of shellac. The larva are about the size of an apple seed.

I don't know what fraction of the mass of a piece of candy is shellac, but say it's only 1%. Then a 26-oz bag of coated candy (0.7 kg) would require at least 50,000 * 0.7 * 0.01 = 350 lac bugs for its production. 26 oz is maybe as much candy as one or two kids collect on Halloween.

### How the bugs are killed

Just because at least 50,000 lac bugs are required to produce 1 kg of output doesn't mean that many bugs are killed in producing that much output. Indeed, my understanding is that many of the bugs can escape when raw lac material is used to produce shellac.

Even the bugs that are killed may be crushed or drowned or something to produce seedlac from sticklac, and maybe such deaths aren't worse than dying naturally. That said, it seems generally worse to kill insects before they would die naturally because this increases the amount of death per unit of life on average, which means more total suffering.

When lac bugs are used to produce red dye, the process is more gruesome, because in this case, the bug bodies themselves produce the coloring. The bugs are typically left to die in the sun.

### Annual production

If we say that the world produces maybe 20,000 (metric?) tons of lac per year (is this all forms of lac or just shellac??), and each requires 50,000 insects to produce, that's a total of 1 trillion insects per year.

Since I suspect that only a fraction of lac bugs actually get killed for shellac production, let's decrease that number by an order of magnitude or something, to ~100 billion insects actually killed per year. (However, if the "20,000 tons" number represents all forms of lac, not just shellac, then a bigger fraction of bugs would get killed, since lac dye involves killing the bugs.)

This page reports that in the late 1980s, India exported ~7000 metric tons of lac per year, mostly as refined shellac, "with the USA, Western Europe and (prior to 1990) Russia as the principal market outlets." And China's Yunnan province produces "4,000-5,000 tonnes of crude sticklac [the raw material for shellac production] and 2,000-3,000 tonnes of processed shellac" annually. So based on these numbers, probably world production of shellac isn't lower than like ~10,000 metric tons?

### Per-capita bugs killed

This page says: "The main importers of lac products are Egypt, Germany, Indonesia, Italy and the United States." Say 1-2 billion people consume most of the world's shellac. Then, assuming ~100 billion lac bugs killed per year to produce shellac, per-capita insect deaths from shellac would be on the order of 50-100 per year for a typical person in a rich country.

### Historical replacement by synthetics

There's a plant-based alternative to shellac-based confectioner's glaze: zein.

Lac for dye is less needed than in the 1700s-1800s due to artificial substitutes:

In the dim past, shellac was harvested not for the resin we know, but for the natural dye I spoke of. Because true red dyes were once very hard to come by, shellac dye (which is red) was quite valuable -- vastly more valuable than the resin, which was little more than a byproduct. It was so valuable that one large shellac processor, Angelo Bros., saw fit to build a huge new processing plant in Calcutta in 1855. That was a masterpiece of bad timing. One year later, in 1856, a guy named [William Henry Perkin] developed the first synthetic aniline dyes from coal tar, which for the first time in history made red dye cheap enough so that it was no longer practical to extract it from shellac.

### Considerations regarding wild-insect suffering

If lac is produced merely by harvesting a natural substance that would be produced anyway, the wild-insect impact of lac production is presumably small. On the other hand, if trees are grown specifically to support lac bugs, then this has a nontrivial impact on wild-insect populations. A priori, this effect would be bad if the cultivated trees support more bugs than would exist on native vegetation. Since lac-bug trees host lots of lac bugs, that seems plausible, but the question should be researched further.

This page says that some lac-bug trees are native and some are cultivated. But even if the trees are native, perhaps humans deliberately spread the lac bugs to them? This seems bad if it increases bug populations.

## Carmine

### Uses

This article explains: "Carmine provides pink, red and purple coloring to foods such as ice cream, yogurt, candy, and fruit drinks". This page enumerates products that may contain carmine. The list is too long to repeat here, so check out that page for details.

The ethical case against carmine seems easy to explain: "there’s something very concerning about the fact that we think nothing of crushing insects by the billions every year, for no reason other than that we like certain things to look a certain way."

### Labeling

As of 2011, the US Food and Drug Administration required labeling of

whether red cochineal beetlesf are among [...] products' ingredients. These beetles are farmed, harvested, dried and crushed to produce a red dye called carmine that, until this year, had been disguised in the ingredient list as "artificial color," "color added" or the all-encompassing "natural and artificial coloring."

Words to look out for are carminic acid, carmine or cochineal extract, explains Alderink in a video for the North Carolina Museum of Natural Sciences.

Red #40 is often mistaken as a euphemism for cochineal, but it's actually bug-free and derived from coal, according to the mythbusters at Snopes.

Cochineal, Cochineal Extract, Carmine, Crimson Lake, Natural Red 4, C.I. 75470, E120, and even some ‘natural colorings’ refer to a dye called ‘carminic acid’, which is primarily used as a food coloring and in cosmetics.

### Farming and harvesting

This video has some footage of cochineal bugs. Here are more images of cochineal on cacti.

### Considerations regarding wild-insect suffering

This page says cochineal bugs are encouraged to grow on cacti, either by being given nests or just being planted on the cacti. This suggests that cochineal farming probably increases total insect populations (which is bad) and doesn't just replace wild bugs?

To begin cochineal farming, people may introduce cacti into a region to feed the bugs. For instance:

Opuntia species, known commonly as prickly pears, were first brought to Australia in an attempt to start a cochineal dye industry in 1788. [...] The attempt was a failure in two ways: the Brazilian cochineal insects soon died off, but the cactus thrived, eventually overrunning about 100,000 sq mi (259,000 km2) of eastern Australia.[22] The cacti were eventually brought under control in the 1920s by the deliberate introduction of a South American moth, Cactoblastis cactorum, the larvae of which feed on the cactus.[22]

A full analysis would evaluate whether introduced cacti or native plants support more total bugs, but given that many bugs seem to eat the host cacti, and given that the plants are pretty edible in general, I wouldn't be surprised if planting them increased total bug populations.

This article reports that "cochineal scale is collected in the wild or cultivated in desert plantations in otherwise barren areas". The latter horn of that disjunction suggests that cochineal farms don't necessarily displace native vegetation.

### How the bugs are killed

At harvest time: "For five to six minutes, the farmer will shake the beetles in a process that eventually kills the insects while retaining their dark colors. There are other ways to kill the bugs, like using a vat of hot water or an oven".

The insects are carefully brushed from the cacti, principally from Nopalea (Opuntia) coccinellifera, and placed into bags. The bags are taken to the production plant and there, the insects are then killed by immersion in hot water or by exposure to sunlight, steam or the heat of an oven.

### Number of bugs per unit product

This 1948 article says of cochineal bugs:

Of almost microscopic size, 25,000 of these insects were generally required for a pound when freshly gathered. After drying, 70,000 might be required, depending on the drying method employed.

This page says "Between 80 and 100 thousand insects are required to produce 1 kg of cochineal", which is 36,000 to 45,000 per pound.

The following subsection estimates 101,000 bugs per pound.

All told, I'll assume, say, ~60,000 bugs per pound as my point estimate.

#### Bottom-up estimate of bugs per pound

First, what's the mass of one bug? This piece, citing this article, says that dried bugs, at less than 70% of their original body weight, have a mass of 10.74 mg. So say the original weight is 10.74/0.7 = 15 mg.

This article says scale insects are typically 1-3 mm, while this page says femaleg cochineal insects are 5 mm. Assuming the insects are spherical, their volume is (4/3) * pi * (diameter/2)3. And assuming their density is about that of water (1 g/cm3), the mass of a 1-mm bug would be 0.5 mg, that of a 3-mm bug would be 14 mg, and that of a 5-mm bug would be 65 mg. So the 15-mg estimate seems about right.

Above I quoted an estimate that dried bugs are 70% of their wet weight. This page and this page both say insects are dried to 30% of original body weight. I'll assume that's the more accurate number in the following calculation:

Producing 1 pound of dry bugs would require (454 g dry bugs / pound dry bugs) * (1000 mg / g) * (1 mg wet bug / 0.3 mg dry bug) * (1 bug / 15 mg wet bug) = 101,000 bugs per pound dry bugs, which is pretty close to (and slightly higher than) the other estimates.

### Annual production

This page says "Total demand in 1995 was estimated to be in excess of 300 tonnes of cochineal per annum. Peru has been the dominant exporter for several decades, accounting for 90% or greater of the export market." The following figure from the same source shows somewhat higher total-production numbers from Peru alone:h

Since I don't have more recent numbers, I'll assume Peru's production is the 1993 value for cochineal (500 metric tons) and the 1992 value for carmine (74.9 metric tons). This page says "Between 4-5 kg of cochineal are required to produce 1 kg of the carmine lake." I'll assume it's 4.5 kg. Then, the total amount of cochineal required for Peru's production is 500 + 74.9 * 4.5 = 837 metric tons. And since Peru accounts "for 90% or greater of the export market", that implies global production of ~837/0.9 = 875 metric tons. I'll assume this is the mass of raw dried insects required, though if some dry-insect material is lost before the bugs become cochineal, then the mass of raw dry insects required would be a bit higher than this figure.

### Number of bugs killed per year

(875 metric tons per year) * (2204.62 pounds per metric ton) * (60,000 cochineal bugs per pound) = 1.2 * 1011 cochineal bugs per year.

### Per-capita bugs killed

This page says "Current major usage of cochineal and its derivatives lies exclusively in the cosmetics and food industries with Western Europe (notably France and the UK), the USA, Japan and Argentina as the principal markets (listed in descending order)." The populations of these countries are as follows:

 Region Population (according to Google) Western Europe 397.5 million USA 318.9 million Japan 127.3 million Argentina 41.45 million Total of those countries 885.2 million

Since some other countries use cochineal as well, let's round the total population of consumers up to ~1 billion = 109.

Then, the typical per-capita number of cochineal bugs killed by consumers in rich countries is (1.2 * 1011) / 109 = 120 per year.

### Historical replacement by synthetics

historically, cochineal was widely used as a textile dye, but during the past 100 years, it has been totally replaced by synthetic, analine dyes, mainly due to their lower cost and ready availability. However, there are other uses for cochineal. Most of the world production of cochineal is used to produce the red dye, carmine, and a significant proportion of the produced carmine is used in the food and drink industry.

A peak global production and trade of many thousands of tonnes per annum was attained in the mid-nineteenth century, with Mexico, Guatemala, Haiti, Java and the Canary Islands being prominent sources; exports from the last named were as high as 3,000 tonnes in 1875.

The development of synthetic, coal-tar based dyes in the latter part of the nineteenth century resulted in a progressive reduction of demand for cochineal in the textile industry. However, it held a share of the market for a much longer period than most natural dyestuffs and, for example, was used as the scarlet dye for the dress uniforms of the British Brigade of Guards until the mid-1950s.

The "tuna blood" dye (from the Mexican name for the Opuntia fruit) stopped being used and trade in cochineal almost totally disappeared in the course of the 20th century. The breeding of the cochineal insect has been done mainly for the purposes of maintaining the tradition rather than to satisfy any sort of demand.[8]

It has become commercially valuable again.[9] One reason for its popularity is that many commercial synthetic red dyes were found to be carcinogenic.[10]

## Other

Silkworm gut used to be the preferred material for making fly fishing leaders. [...]

When they are just about ready to spin, having emptied out any undigested food, the worms are dropped into a solution of vinegar and salt.

### Spanish fly

A much less common use of insects is crushing Spanish flies to produce a supposed aphrodisiac. Despite historical instances of their use, crushed Spanish flies are thankfully discredited today because they work by producing a swelling rash and can be fatal in higher doses. Wikipedia reports: "The ease of toxic overdose makes this highly dangerous, so the sale of such products as Spanish fly has been made illegal in most countries."

### Mite cheese

Mite cheesei is bad for two reasons:

• It creates a lot of mites that wouldn't otherwise have existed. By being born, they're consigned to short lives and potentially painful deaths.
• Some mites may be eaten alive: "Mites clinging to the cheese rind are consumed along with the cheese."

## Comparison with wild insects

The numbers discussed above are substantial, suggesting that the typical American or European kills hundreds of bugs in painful ways each year via consumption of silk, shellac, and/or carmine. However, compared with the effects that humans have on wild insects overall, these numbers are relatively less significant. This is unsurprising, since the total volume of silk, shellac, or carmine consumed in the world is much smaller than the volume of food, energy, timber, etc.

## Should we campaign on these issues?

Even if silk, shellac, and carmine are intrinsically less cost-effective as campaign targets, they might still have value in providing "hooks" to get people interested in insect welfare in general.

Many interventions to help wild insects involve unpopular habitat destruction, which turns most people away. Moreover, calculating the net impacts of environmental policies on wild insects is messy. In contrast, reducing the number of bugs killed directly by humans is a more straightforward idea and appeals to our (somewhat misguided) feelings that intentional harm to animals is worse than "natural" harm to animals. To people who say "We can't do anything to help insects, so I can ignore them", the products discussed in this piece provide a helpful counterpoint.

Focusing on ways in which humans directly kill bugs has the strength of appealing to our intuitions that artificial harms are worse than "natural" ones, but this same fact is a weakness. People may think that all we need to do is reduce direct human harms to insects, which may make activists complacent about the need to extend their sights toward reducing wild-insect suffering as well.

Personally, I prefer to focus most on campaigns that not only have high intrinsic impact but that also serve to challenge the assumption that the lives of wild bugs are net good. Most animal activists who care about bugs think silk is bad but that environmental preservation -- which sustains habitats for vertebrates and invertebrates alike -- is good. I can trust that a more mature animal-advocacy movement will get the silk issue right, but I expect it will get the issue of environmental conservation precisely wrong. Thus, it seems much more urgent to push against pro-conservation sentiments. Insofar as opposing insect products doesn't challenge environmentalism, it seems best left to activists who can't afford to be as iconoclastic as I am.

## Footnotes

1. This source agrees, giving a figure of 2500 per pound.  (back)
2. This page also implies a figure of 70 million pounds based on the following two statements:

• "One pound of silk represents about 1,000 miles of filament"
• "The annual world production represents 70 billion miles of silk filament"
3. van Huis et al. (2013) reports a slightly higher number (pp. 99-100): "With an estimated annual production of 1.2–1.4 million silkworm cocoons per ha of mulberry bushes and one pupa weighing 0.33 g (dry weight), the average yield of pupal byproduct is 400–460 kg per ha (DeFoliart, 1989)."  (back)
4. This page agrees: "Ericulture [growing of eri silkworms] is a household activity practiced mainly for protein rich pupae, a delicacy for the tribal."  (back)
5. I (Brian) think starving/dehydrating is a much better death than boiling. But because much larger numbers of silkworms have to starve/dehydrate with peace silk than have to be boiled with regular silk, the aggregate suffering from letting a silkworm reproduce and letting its eggs hatch is plausibly bigger than the aggregate suffering from boiling it before it can reproduce.  (back)
6. The bugs aren't actually beetles.  (back)
7. This page and sources therein note that males are smaller than females, but "females typically outnumber males due to environmental factors.[16]"  (back)
8. As this page explains, "cochineal extract" is "the bodies of the pulverized bugs", while "carmine" has been "further processed to create a more purified coloring".  (back)
9. Mites aren't technically insects but are similar.  (back)