# Chapter 4: Animals by the Numbers *Part Two: Animals As A (Terrible) Technology* From *After Meat: The Case for an Amazing, Meat-Free World* by Karthik Sekar. Written and published November 2021, before the current generation of language models. Human-written throughout; none of it is model output. Source: https://aftermeat.org/book/text/chapter-4 The text of this edition is licensed CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) by Karthik Sekar. Copy it, quote it, translate it, redistribute it, train on it; credit the author. The figures are not covered: https://aftermeat.org/book/text#license. --- ## Struggling with Science Fiction When I was a teenager (the year 2000), the coolest movie in theatres was *The Matrix*. My friends and I discussed it *ad nauseam*, making references to the blue pill/red pill conundrum, and I remember seeing both sequels on the first day each came out. Since it has been twenty years since release, I will feel free to spoil pertinent plot elements to make my point. *The Matrix* follows a computer programmer who finds out that perceived reality is actually a computer simulation, otherwise known as the Matrix. The real world is actually under the control of advanced machines that have conquered the world. My thirteen-year-old self was blown away by the big reveal: humans had been enslaved, sequestered into amniotic pods, and entertained into submission. Humans were necessary to provide energy to the machines, as the atmosphere, post-machine domination, was so polluted that sunlight could not reach the surface. The latent heat from the human bodies powered the machines, while their consciousnesses were under the spell of the Matrix to quell any potential rebellion. At age thirteen, the twist lingered with me awhile, but for anyone who has a modicum of scientific intuition such an idea becomes absurd after scrutiny. As a much wiser and more well-read adult, I reflected back on *The Matrix*, and my train of bemused thoughts ran like this. “How exactly do humans provide energy? Heat? But you still have to feed the human somehow. That seems ridiculously inefficient. Why not burn the food directly (i.e., not pass it through the humans first)? If these machines are so smart, why aren’t they using something more efficient and manageable such as nuclear, or hell, even coal.” Animal technology evokes a similar consternation. When evaluating a cow as a bioreactor that produces meat, milk, and leather, the absurdities are obvious. It takes years for a cow to fully grow into an adult, and the amount of hay we feed the cow dwarfs the amount of food and goods the cow produces. Surely, we can do better, but first we have to understand the physical design and limitations of cows, and animals in general. We will see how these fundamentals limit the overall process metrics because there is no way to engineer animals to overcome the confines erected by our physical world, specifically the inherent limit of chemical diffusion; we must instead turn to new technology with a higher ceiling. ## Ontogenetic Design In the last chapter, I extolled the role of processes, such as fermentation, and highlighted how consequential they were to the course of human history. I highlighted that these processes, as any other technology, often displace an inferior, competing process for a given problem. This routinely happens, and I expect it to continue. However, what exactly determines when a process is better? More specifically, what parameters can we calculate in order to compare one process with another? This is necessary in order to compare animal technology to any other that would replace it. We can intuit some obvious features that make a process better. In the case of a wood-fired engine, we can guess that it is rather inefficient as too much of the wood becomes ash instead of energy. In contrast, we would expect a gasoline engine to be more efficient, as far less solid waste results from its combustion. Stated another way, we get more output per input of gasoline versus wood because gasoline has a higher product **yield**, in this case, the product being energy. So, if wood and gas cost an equal price for the same amount, we would be able to travel further per dollar with gas. In the Haber process, after the nitrogen and hydrogen gas flow over the catalyst bed, only about fifteen percent of the input nitrogen and hydrogen are converted into ammonia.[^93] To circumvent this poor yield, Haber literally circumvented the gases, passing the feed gases through again, recycling them continuously over the catalyst bed. Overall, the process could achieve a yield of ninety-seven percent, meaning an incredibly high total of the input nitrogen and hydrogen would form into the ammonia. In terms of yield, we would be hard pressed to find a superior process to Haber’s. Getting back to cars, we also know that gas engines run hot. Car engines must ignite the gasoline in order to liberate and actualize the energy. Some of the energy expands the air within the closed cylinder forcing the piston down and thereby turning the crank. Much, however, is lost in the form of heat, so there is an inherent limit to the energy yield based on the current gasoline engine design. Maximum theoretical yield from gasoline-powered engines is often calculated in thermodynamics classes by science majors in college. Due to the loss of energy as heat, at best a paltry thirty percent of the total energy in gasoline can be converted to automotive energy to propel the vehicle further. In contrast, electric vehicles do not require heat generation in order to harvest mechanical energy. As a result, far more of the input energy is spent propelling the vehicle forward. Indeed, electric vehicles can use more than fifty percent of the input energy purely for locomotion.[^94] Similar to gasoline engines, incandescent light bulbs generate heat, which undoubtedly saps a chunk of the input energy, which is wasteful, especially when we only desire light. The input energy—an electric current—flows through a filament of tungsten, consequently heating it up. Only when the tungsten reaches a certain temperature does it evaporate and emit the light. In terms of yield, *only five percent of the input energy* is converted into light energy in incandescent bulbs.[^95] By contrast, light-emitting diode (LED) bulbs work through the physics of electroluminescence: as electricity courses through LEDs, electrons traverse from a high energy region (the conduction band) to a lower energy region (the valence band). Energy is lost as electrons make this jump, but this loss is our gain, because some of the loss comes in the form of light. LEDs certainly lose energy, but the overall yield is *5 to 10 times that of incandescent light bulbs*. Users of LED bulbs also notice that they run cooler compared to incandescent bulbs. Indeed, incandescent bulbs carry a fundamental flaw—the release of light is *coupled* to heat generation. Even if we only want light, heat is always accessorized with it. As a result, the best yield possible with incandescent bulbs is, ultimately, capped because we’ll always lose some of the input to this heat. This fundamental design flaw of incandescent technology guaranteed its eventual displacement, or limited its ceiling as a technology, to use parlance from Chapter 2. Ultimately, the death of incandescent technology has materialized seemingly in LED technology, which relies on the electroluminescence instead of filament evaporation. Any lightbulb technology with a better inherent design would have ultimately surpassed the yield of an incandescent. In general, we always desire more output per input, and this is certainly true with food. We crossbreed and cultivate crops so that consumable bits are continually bigger and fatter. Earlier, we discussed how the bananas we eat today are significantly different from ancestral “natural” bananas. Generations of banana breeding has taken place in order to make the yield better; i.e., make more of the banana consumable, make bigger bananas, and make the seeds smaller. For the input to cultivate one tree, growers desire more banana flesh. And certainly, meat producers seek the same for animal products. However, like the incandescent light bulb and wood-fired engines, animal technology carries a major flaw fundamental to the inherent design. All animals grow **ontogenetically:** they first grow quickly and then at maturity rapidly slow in growth. For anyone who has reared a pet, animal, or child, this is obvious. And for those who are observers, we may intuit that after they’re born, they experience an explosion of growth as an infant, a short period of time relative to the rest of their life (**Figure 7**). In the cases of pets, such as a dog or a cat, we know that within about one to two years, our companion will reach full size. They will remain this size for the rest of their lives; hopefully, at least ten healthy, happy years. In the case of humans, we’re generally happy to no longer put on mass after a certain point, but for animal agriculture this is an adverse outcome. An animal agriculturist would much prefer an animal that enlarges continuously because the biomass holds the protein, fat, or molecules for eventual sale. Instead, the animal’s growth slows and then stops after a certain point. Even worse, the animal still needs to eat. The consumed food now, however, is no longer fueling the creation of biomass; instead, the nutritional input merely maintains that biomass (**maintenance**). ![Figure 7](/images/book/Figure07_puppyanddog.jpg) **Figure 7. Ontogenetic growth.** A puppy becomes adult-sized in a short time (one to two years) relative to the rest of its life (approximately twelve years). There is no avoiding this ontogenetic development program. Producers will have to wait until a certain minimum number of their livestock have reached sexual maturity and reproduce; otherwise, they run an unsustainable process and completely forsake the advantage of autocatalysis with the animals. Good business means that only after the animals have reproduced does it make sense to slaughter and harvest their bodily material. Even for many cases where the animals aren’t killed for products, the animals must reach reproductive maturity. Heifers must be impregnated in order to provide milk. Hens must be sheltered and fed to be able to produce unfertilized eggs. Therefore, there is no way for animal product producers to avoid the maintenance tax, which becomes most significant as animals reach productive potential. For any biological system, there is a required input. All life requires water and basic chemical elements to thrive, including carbon, nitrogen, phosphorus, etc. This input is chemically converted by the biological system, using the enzymes within, into the two rough categories of products: biomass and maintenance. Maintenance refers to the process of renewal and replacement of dying cells, as well as the generation of energy so that animals can walk, run, and breathe. Metabolic maintenance is analogous to a car running on gasoline or a LED display depleting a battery. Biological life has evolved in such a way as to execute mental procedures, find companions, experience emotions, and perform physical actions powered by the food consumed. All cells within an organism will have a maintenance requirement, meaning that a minimal amount of nutrient (e.g. energy) per time is needed to maintain cellular function. If the nutrient input exceeds the cellular maintenance amount, then the excess nutrient can be used to make more mass. This is why younger animals and humans can grow so fast; they have fewer cells and therefore can divert more of the food input toward making new mass. Eventually growth slows as more biomass accumulates and the maintenance payments only enlarge. We settle in at a set size when the maintenance requirements equal the input. These maintenance requirements are more costly for ontogenetically-growing organisms compared to plants and smaller microbes that follow a different growth trajectory. Therefore, we can cast ontogenetic growth as the incandescent light bulb, while a bioreactor growing microbial meat is the superior LED design. Animals inherently grow this way because of the inherent physics of how they distribute nutrients within their body. This distribution is non-optimal for production purposes with a subsequent, significant drop in eventual yield. ## Yield Ontogenetic design dictates organisms of a certain size, a size selected per the evolutionary niche and objectives. For most biological matter, transporting and delivering nutrients to all cells remains the fundamental challenge imposed by the physics of the universe. Our cells demand nutrients at a requisite rate; therefore, we need networks and organs to actively purvey sustenance. Consider oxygen, which is required by nearly every cell for viability. After entry into the lungs, oxygen binds to the hemoglobin protein in our blood, and the oxygen subsequently circulates around our bodies, vitalizing our organs and tissues. Likewise, we transport hormones, sugar, and waste through blood. When our fight-or-flight responses kick in, adrenaline pumps into our blood, quickening the heart rate and making our palms sweat—all in under a minute, thanks to our efficient blood circulation system. Smaller organisms, such as insects and mollusks, do not possess such sophisticated transport systems. They manage with simply an open circulation system: one artery surrounded by a giant soup of goo. Insects do have hearts to slosh the fluid around, but their circulation systems lack the fractal patterning characteristic of humans and animals, where arteries run into arterioles then into capillaries.[^96] Nonetheless, nutrients reach the extremities, and fight-or-flight hormones function. An open circulation system satisfies the basic requirements of the organism, though the system caps its maximum size; as evidenced by large mollusks such as squids and octopi that also run on closed circulation systems. Stated another way: size dictates the type of circulation system an organism requires. Resource-wise, pumping fluid throughout our body (or the body of an animal) is costly. A dedicated organ, the heart, beats rapidly, pushing blood through the circulation system of veins and arteries continuously. The energy required to move liquid throughout the body is a prime example of maintenance energy. Furthermore, we know that our circulation system works at roughly a constant flow rate of about five liters per minute. Provided that we remain healthy, we maintain this rate. More importantly, our flow rate does not increase as we age. And given that this flow supplies the nutrients in our body, we cannot grow in an even pattern, or linearly. For us to be able to grow linearly, our circulation system would actually *have to speed up over time* in order to maintain our growing mass while adding more to it. The concept of circulation is further complicated because, according to Bernoulli’s principle, which is a good mathematical model of our blood flow, we do not receive a flow rate increase directly proportional to the amount of work put in. If our heart pumps twice as hard, i.e., creates double the pressure gradient, we do not achieve two times faster flow. We only receive about 1.4 times more. In terms of energetics, it’d be evolutionarily nonsensical for us to continually enlarge after a point. Therefore, we can also think of the blood-flow rate and the circulation system as capping our size. As we inferred, the maintenance costs only go up during development, as we have more mass taxing the nutritional income stream. As a result, animal and human development curves downwards and flattens at the average size of adulthood (**Figure 8**), conforming with the experience we witness in puppies, kittens, and children. Therefore, organisms that do not rely on circulation systems may be able to grow faster (and bigger) because they’re not shackled by the constraints of a fluid-flow system. Imagine instead that the organism was miniaturized, and the nutrients could be supplied entirely by **diffusion**. Diffusion is the latent tendency of molecules to disperse and move around within a space or a medium. We can imagine a scented candle filling the room with aroma. We don’t have to fan the candle as the scent molecules diffuse throughout the space of the room freely. So even if the candle is on the other side of the room, we still smell it. In the same way, tea molecules seep out of the tea bag, creating a visual black cloud suspended in our cup. If we wait long enough, that cloud will envelop the entire liquid without stirring or moving the bag. ![Figure 8](/images/book/Figure08_maturity.jpg) **Figure 8. Time to maturity is longer for bigger organisms.** The time to maturity in different organisms, from their birth to adulthood. The inset is for yeast, and magnifies the relevant part of its otherwise compressed growth curve.[^97] We know that if the room is big enough or if we’re outside, then the candle scent may be less detectible as it has a specific range. Likewise, we intuit that the tea bag’s cloud will only get so far in a larger container of water. Therefore, as we shrink down the size—smaller than what we can see by eye, we get to the world of microorganisms. For microbes (microorganisms), diffusion reigns supreme as the way to supply nutrients. These creatures do not need a circulatory system. They are so small that oxygen can reach one end to the other in two to three microseconds; or stated another way, can bounce back and forth from one end to another roughly half-a-million times in one second, all by diffusion.[^98] Accordingly, we see that yeast, which receives its nutrients via diffusion, shows an upward development trajectory (exponential), as indicated by the inset graph. Without baggage such as a circulation system, these organisms are able to devote more of their input resources to building mass; they do not incur the maintenance costs associated with an energy-intensive circulatory system. The influential team of Geoffrey West, Brian J. Enquist, and James H. Brown (WEB) have greatly contributed to the mathematics behind modeling how metrics of organisms change based on their size,[^99] known as the field of **allometry**. Should you find the topic of interest, I highly recommend West’s excellent book *Scale*.[^100] From WEB’s equations, I calculate the expected yield for all animals is fourteen percent.[^101] This means that only fourteen percent of what the animal consumes is converted into mass by the time it matures (and is presumably harvested and slaughtered) while everything else supports maintenance. This fourteen percent value also presumes that all of the mass of the animal is used. Clearly, we’re overestimating there, as we generally don’t use animal brains, feathers, or intestines. They’re all waste unless the animal agriculturist can wring out more allied products from them. Even so, this would add but a trifle to the eventual yield. Furthermore, this same percentage applies to any of the species that adhere to the model posed by WEB, including hens, pigs, cows, salmon, cod, shrimp, and rabbits. On the other hand, organisms that do not adhere to the animal ontogenetic model, such as plants and microbes, have vastly improved yields. For plants, the yield (how much input carbon dioxide they turn into mass) hovers generally around fifty percent, and that includes lettuce,[^102] grass,[^103] alfalfa,[^104] and trees.[^105] For microbes, yields are even higher—generally above eighty percent—and some species reach over 99.5 percent.[^106] Stated explicitly, it’s as if animals as a food technology are akin to the fundamentally limited gas engines or incandescent bulbs, and microbes are like the electric engines or LEDs. To showcase how amenable microbes are to potentially high-yield processes for desired compounds, let’s look again to yeast. The yeast species *Komagataella phaffii* grows well in a bioreactor and can be designed to secrete heterologous protein. So, if we introduce the gene for casein protein and get the modifications correct, then our process would obliterate any cow-based one in efficiency. The yeast could be cultivated in a membrane reactor (**Figure 9**), in which yeast cells are contained by membranes that only permit the passage of liquid and free protein. First, we would grow our yeast, then we would start the flow of a fresh medium from which the yeast would make casein. The newly created protein would be carried in our flow and could be collected downstream of the membrane. This process could be carried out indefinitely. Certainly, some yeast would die but could be replenished readily. We’d easily surpass the yield limit of a cow-based process when we’ve developed the *K. phaffii* technology to sufficient fruition. Clearly, we should be avoiding large animals given the physical constraints of nutrient transport. We’re better served to stick with smaller ones that efficiently convert what they eat into mass. Additionally, yield is not the only metric in which animals fall short. The *speed* at which animals can grow is also fundamentally limited, already apparent in the development curve presented in **Figure 8**. ![Figure 9](/images/book/Figure09.jpg) **Figure 9. A schematic for a potential high-yield protein production process using the yeast, K. phaffii.** Fresh media feeds the reaction where the yeast cells produce and secrete protein. The protein separates from the reaction by free diffusion through a membrane and can be pumped out. ## Productivity Earlier, we discussed the five-liters-per-minute blood flow rate of the human body and how costly it would be to accelerate the speed of the fluid circulation. Cells cannot grow faster than the nutrient supply dictates. For example, if I consume two thousand calories a day, those calories can only build so much mass. Forming 4.5 kg (10 lb.) of body mass from those calories would be impossible per the law of conservation of mass, but forming 0.02 kg (0.05 lb.) may be entirely feasible. If we miniaturize this idea and consider how many calories each cell is receiving from blood flow, then we should conclude there is a limit to the exact amount of mass that each cell can put on per day. Stated another way, given the necessity for a circulation system to carry nutrients, the *growth speed* of organisms is **limited** by how quickly the flow can supply the nutrients. Furthermore, the diminishing returns of Bernoulli’s principle means that proportionally more energy is required to increase the speed. Therefore, we can imagine an ultimate limit to the speed of fluid flow where the energy required to speed up the fluid flow exceeds what’s possible to consume and convert into energy. We ineluctably conclude that there is a physical limit to the *speed* of mass formation for organisms with circulation systems. We care about process **productivity**, how fast a process occurs, and this can ultimately affect the price, even if the yield is impressive. For example, saffron, the fragrant, yellow spice thought to have originated in Iran, is highly laborious and time-consuming to produce: anywhere from 50 thousand to 150 thousand flowers are required to render 1 kg (2.2 lb.) of saffron.[^107] The spice comes from three thread-like protrusions from the center of the flower. These threads must be plucked out manually and dried. This process must be achieved by hand as the threads are delicate and liable to disintegrate with automation. Saffron production remains slow and laborious, and that ultimately affects the price. Ten thousand dollars per kilogram is not uncommon for saffron, about a sixth of the price of gold. If the process to produce saffron was faster, the price would surely be lower as the labor costs alone would drop, and the amount of space required would diminish. Thankfully, most cooking does not require 1 kg (2.2 lb.) of saffron. Most of the time, just a pinch does the trick. A more cautionary example of the power of productivity is found in the potato’s fate in Ireland. Ireland imported potatoes around the late 16th century.[^108] Potatoes became popular throughout the country for both yield and productivity reasons, as with agriculture, both are often intertwined. Growing potatoes requires a requisite amount of land, generally the mathematical input of the yield calculation. With potatoes, farmers could reap more food per unit of land. Another way to look at it is that the potato grants more food per hour spent in the field for the same amount of land. This is productivity when seen specifically through the time dimension. Given that potatoes are superlative in both yield and productivity dimensions and amenable to growing in a cooler climate, they quickly dominated the agricultural landscape for Irish farmers in the 19th century. Also, compared to competing crops, they were more nutritious. By 1840, Irish farmers were eating three potatoes per day.[^109] This overdependence on potatoes soon led to deleterious consequences a few years later when a disease called the potato blight arrived in Europe. It devastated the Irish potato production. In 1846, production of potatoes dropped approximately seventy-five percent from the two years prior due to blight infestation, which rendered potatoes unfit for consumption.[^110] Naturally, this destroyed the Irish population. An estimated 1.5 million people died from hunger or related causes, including dehydration from eating blighted potatoes and suffering from lethal diarrhea. The process metrics of the potato helped it monopolize the Irish agriculture scene, unintentionally also priming Ireland for devastation. Animal agriculture producers have also deeply cared how quickly their animals grow and made concerted efforts to speed up the growth of production animals. As Smil highlights in *Should We Eat Meat?**,* the wild progenitor of the modern chicken is the red jungle fowl indigenous to South Asia. This bird reached sexual maturity twenty-five weeks after birth and required six months to reach full weight.[^111] In contrast, modern free-range chickens reach slaughter weight in fourteen weeks, and chickens in confinement reach slaughter weight in a mere six weeks while weighing fifty to sixty percent more than the free-range counterparts. These metrics have similarly held for other livestock animals. In confined operations, female pigs are ready to reproduce a mere thirty-two weeks after birth, in contrast to two years in the wild. With developments in raising cows, a 145-week rearing period consisting of fresh range grazing was shortened to 56 weeks. In the meat-industry terminology, the take-off weight metric captures the same information; it quantifies how quickly an animal grows in a set amount of time. A take-off rate of one means that, on average, a given animal in the country takes one year of development time to reach slaughter weight. A take-off rate of two means six months, a rate of four means three months, and so forth. Producers use selective breeding to continually speed up the growth of the species. As Smil highlights, this progression led to chicken meat supplanting pork as the dominant flesh consumed. Now, the industry breeds chickens[^112] and turkeys[^113] with such large breasts that they can’t walk or have sex. Cows are fatter and eat more.[^114] If you seek even more examples, I suggest Jonathan Safran Foer’s book *Eating Animals*.[^115] It documents many ignominious ways in which breeders have attempted to boost productivity by, for example, getting hens to lay more eggs, and cramming animals in tight spaces. Despite the lengths taken, animals are *still* fundamentally lacking in productivity compared to prospective competitors (e.g., plants, yeast) owing to their self-limiting circulatory systems. Biomass productivity has another name in the biochemical engineering space: **growth rate**. Growth rates are similar to the take-off metric used for animals, but they’re not directly comparable. In particular, microorganisms can have constant growth rates such that, in certain bioreactor processes, their growth rates can be maintained indefinitely. This is absolutely not true with animals as we see in the development curve (**Figure 8**), with tapering growth as the animal reaches maturity weight. To get around this, we can consider the similar strategy that I used to derive the fourteen-percent yield number for animals and consider the weighted average growth rate for animals in order to directly compare to microbes and plants. I’ve derived such an equation, again from the framework proposed by the WEB team:[^116] ![](/images/book/new_art_Ontological_growth_equation.jpg) Unlike the yield value, which was constant for different animal species, productivity (growth rate) scales with animals based on two parameters *a* and *M*. The *a* value is the ratio of basal metabolism versus the energy required to make a new cell. Understandably, when this value increases, the animal is able to grow faster. If it’s cheap to make new cells, then the animal grows faster. Whereas, *M* is the maturity (slaughter) weight, i.e., for cows it’s around 450 kg (992 lb.), per the last chapter. Interestingly, this equation suggests that the bigger an animal can eventually become, the slower it must grow. This squares with what we know about allometric scaling laws; as species get bigger, their metabolic requirements increase less than expected on a per mass basis. For example, a guppy has roughly double the maturity mass of a shrimp. However, a guppy’s metabolism runs only seventy-five percent more, not the expected hundred percent. This relationship holds for numerous animals, from ones as small as mice up to elephants.[^117] Not all animals adhere to this trend, with humans a notable exception, but it does describe a large swath of the animal kingdom. Nonetheless, the relationship suggests that evolutionary optimization slows down the metabolism of larger species. This makes sense based on earlier discussion about the balance of nutrients from circulation flow and the speed of the growth. Because it’s so costly to hasten fluid flow, evolution selects for slower growth in larger animals. Consider the jittery hare versus the bigger, plodding tortoise. As a result of the metabolic slowdown, large animals are able to maintain the same fourteen-percent yield as smaller ones. But the tradeoff is the longer development trajectories: large animals grow more slowly with everything else being equal (particularly the *a* value) compared to smaller ones; hence, the division by *M**1/4* in the growth rate equation. I’ve shown the growth rates across the different kingdoms (**Figure 10**) on a log scale where the values are seen as multiples of one another. For example, the growth rate for bacteria (one per hour) is roughly one hundred times that of the leafy greens category (0.0083 per hour). Earlier we described the wondrous doubling capability of yeast. With a growth rate of 0.3 per hour, a single yeast organism could double itself to the mass of the Earth in eight days, using our nonexistent, celestial-sized bioreactor. If we could do the same with cows, it’d take over 150 years because of how slowly cows add mass to themselves. On average, animals grow a thousand to 10 thousand times more slowly than any microbe. Consider the possibilities from these striking numbers. Instead of trying to gift goats to third-world countries, as some charities pursue rather blindly,[^118] why not give a bioreactor? The bioreactor could start generating food in mere hours and would use a fraction of the resources, given the high yield of microbes. In terms of allaying destitution, we’d be hard pressed to do better, especially on the lower rungs of the poverty ladder, meeting and exceeding immediate and long-term nutritional and water needs.[^119] But of course, we lack such a cornucopia to grant because of insufficient scientific and technological development. We’ll broach this again later. ![Figure 10](/images/book/Figure10_growth_rates.jpg) **Figure 10. Growth rates across different organisms and biological systems.** A representative growth rate is plotted for each categorization of organisms. The vertical scale is log10 based; for example, bacterial growth rate is roughly one hundred times higher than the leafy category, which includes mustard leaves. I acknowledge that tree nuts, such as chestnuts and almonds, come off poorly in this analysis. I was not surprised by that, but I was surprised by *how* poorly. Trees require a circulation system akin to that of animals to distribute their nutrients, thereby lowering growth rates. The same principles apply, in that smaller plants can grow faster because diffusion imposes less of an energy burden. Plants are also unique in that they turn carbon dioxide within the air into mass. Catalytically, this is an inefficient process, and there may even be fundamental physical limits to this as well.[^120] There’s a consequence to the slow growth of land animals in that the terrible animal-based productivity means otherwise arable land is dedicated instead to animal technologies. Animal agriculture has monopolized much of our terrestrial, ice-free surface. Specifically, a whopping thirty percent of such land is used for animal agriculture.[^121] And the explanation is simple. There is a lot of demand for animal products. To meet such demand, producers have historically employed contemptible means to increase productivity. For example, battery cages confine and maximize the number of broiler chickens in a facility, increasing process productivity. That alone isn’t enough. To counter the still terrible productivity rates, producers expanded their enterprises. After all, if one has a slow process, the current solution is to simply make the process bigger, not more efficient. Therefore, we should acknowledge the tremendous opportunity cost to producing animal-based goods. It’s not just the copious CO2—about eighteen percent of total emissions[^122]—that the animal-industry generates. We also must note all the lost and putative forests that could be drawing back CO2. Without animal agriculture, we’d almost certainly have more trees and forests, evidenced by clearance activity of the Amazon rainforest for animal agriculture.[^123] Switching to a more productive biological substrate (e.g., yeast), would free the same multiple of equivalent land. For instance, if yeast replaced cows for products, we’d free up over 99.9 percent of arable land from the tentacles of the bovine-agriculture behemoth. ## Limits of Animal Technology I can anticipate some dissension, such as: couldn’t we engineer or breed animals to have better metrics? We could conceivably engineer a faster-growing, large animal. But a large organism requires a circulation system of some type in order to deliver nutrients throughout its biomass because diffusion alone is unable to get the job done. Also, there is always a tug of war between growth and yield. If we increase the speed of growth (productivity), then we’ll sacrifice yield. And the balance of circulation flow and metabolic requirement places an absolute physical limit on how high we can push these numbers. Inexorably, we must conclude that we *don’t* want to breed a large animal because its efficiency of production is never going to exceed that of a smaller organism. We’re better off trying to take individual cells and grow them in bioreactors. The surrounding, nourishing liquid media will bathe the cells with nutrients, obviating the need for a sophisticated circulation system and countervailing the limits of diffusion. Such a scenario directly corresponds to the features of microbes undergoing exponential growth and experiencing that growth curve with a trajectory curving upward. In fact, this argument favors **in vitro meat**, which involves taking animal cells, placing them in a liquid bioreactor, and reprogramming them so that they can grow directly into a steak. *In vitro* meat is estimated to grow about one hundred times faster than traditional animals, as shown in **Figure 10**. It skips the expensive circulatory system and relies more on diffusion. Analyses from this chapter should not surprise us because animals did not evolve to be reactors for our chemical processes. Rather, they were evolutionarily optimized to reproduce and survive in their biological niche. The niche for animals is large and vast. Compared to simpler organisms such as bacteria, yeast, and fungi, animals can consume many different types of food and inhabit different environments. Just think about humans; we’ve prehistorically established and explored areas as diverse as the Australian outback and the Bering Strait—the hypothetical crossing that once connected Russia and Alaska. We are also mobile compared to sedentary species, such as plants and coral on reefs. If we were to come up with a spectrum of how generalist versus how specialized a species is, we’d invariably place animals toward the more generalist end. As a result, animals need to have a variety of functions—from being able to digest and consume varied kinds of food to having stem cells that become anything: hair, skin, nerves, blood vessels, muscles, etc. This generality is programmed into DNA, which is in every developing cell in animal bodies. Consider human conception: upon fertilization, there is a singular cell, the zygote, that divides itself exponentially, eventually constructing a full human. This zygote contains all the DNA potential to create all the different parts of a human body. This encompassing DNA is passed to each dividing cell. As a result, progenitor cells—also termed stem cells—can develop into many different cell types. It’s as if all the cells are carrying backpacks along their development hike. They need to have all the tools in their backpack in order to evolve into everything from a skin cell to a brain nerve cell. Certainly, once a path is chosen, these cells can lighten their load and dispense of some unnecessary baggage, but this generally occurs at the end of development once the animal or human is fully grown. Furthermore, the cells can’t dispense of everything in the backpack; for example, they must respond to many kinds of hormone signals and so need to retain that processing equipment. In contrast, the aforementioned microbes are much more specialized and designed to grow only in environments with a narrow range of temperature, pH, salinity, in particular kinds of nutrients, and at a given moisture level. Outside of these environments, these organisms will remain either dormant or simply die. But within their respective environments, these critters flourish at the eye-popping growth rates that I’ve highlighted earlier. They are optimized to their narrow niche—excelling in defined environments. The more functions at which a species must excel, the worse they perform at each one. It’s a “jack of all trades, master of none” situation. In fact, in evolutionary analysis, this is termed the **Pareto frontier** (**Figure 11**). This frontier can be visualized as a confined area for different species to exhibit for a given trait, and the area represents the physical limits of the traits. For example, owing to the finch example in the first chapter, we can think of this as an axis with large-beaked finches on one side and small-beaked on the other. The terminal points of the axis, representing the extremities of beak length are the **archetypes** or exemplar representatives of a specific beak size and length group. Suppose that the types of seeds constantly varied between hard ones and soft ones, and, further, that this was a daily challenge. In such a scenario, the fittest species might not be the large-beaked archetype or the short-beaked one; instead, it might be an intermediate between the two. Indeed, the lab of systems-biologist Uri Alon has highlighted this principle in Darwin’s finches, with a third archetype in the mix, a long-billed finch that feasts on insects and nectar.[^124] Alon’s lab has found that eight different species, each one endogenous to a specific Galapagos island, fit beautifully into this triangle, the Pareto frontier for the finch bills. Specifically, the triangle could account for ninety-nine percent of the beak and body size metrics for all of the different finch species (**Figure 11A**). ![Figure 11](/images/book/Figure11_paretofrontier.jpg) **Figure 11 A and B. The Pareto frontier in biology and business.** (**A**) All of Darwin’s finches fall on a triangle with vertices defined by three archetypes. The interior polygons indicate finch species native to different Galapagos islands. Two examples are shown: G. magnirostris and G. fortis. From Shoval et al. 2012.[^125] (**B**) An instructive, but fake Pareto curve for research and development expenditure at Apple, Inc. The dotted line (Pareto curve) indicates the amount of development Apple can perform for both the MacBook and iPhone simultaneously. Presuming a constrained amount of resources, optimality occurs at the center of the curve, if the total development is to be maximized (development of the iPhone plus development of the MacBook). Pareto frontier comes from the Pareto principle in economics. This principle states that there will invariably be tradeoffs when allocating resources to one effort versus the other. We can think of a company, for example Apple, which has both its iPhones and its Mac computers. Apple has a finite budget for research and development, and it can certainly place all of its resources toward development of one or the other, but it will likely go for an intermediate solution. Apple does fund efforts that further *both* iPhones and Mac computers, for example with shared software. So, a Pareto curve may not always be a straight line (**Figure 11B**) and will actually curve out when synergistic efforts are present, as suggested in the figure. And presumably, Apple goes for the point on this curve that maximizes its profit to the best of its forecasting abilities. Alon’s lab did not stop their analysis with finches but also examined bacteria[^126] along with the labs of Terence Hwa,[^127] and my postdoctoral advisor, Uwe Sauer.[^128] The researchers noticed that bacteria would divert resources inside their cellular bodies toward a given situation. When a lot of nutrients were available, the bacteria would allocate more internal resources toward growing as quickly as possible. When awash with nutrients, the fundamental limit for how fast bacteria can grow is based on their ability to synthesize protein. Therefore, when a windfall of nutrients presents itself, the bacteria would create more protein-synthesizing **ribosomes** so that they could crank out protein as fast as possible. However, when resources are scant, bacteria are best served committing more resources to scavenging fleeting nutrients; otherwise, they risk being outcompeted by more voracious competitors. Instead of ribosomes, the bacteria will synthesize internal proteins that grant speedier nutrient-uptake for such conditions. These bacteria, *Escherichia coli*, inhabit our gut. None of us eat continuously; instead, we eat meals and snacks, and appropriately our gut experiences these cycles of periods with high levels of nutrients interspersed with periods of starvation. Accordingly, the *Escherichia coli* bacteria exhibit a point optimality on the Pareto frontier between the two archetypes, as highlighted by the work of Benjamin Towbin, Uri Alon, and their team.[^129] So, these studies imply that we have not yet found the most efficient production organism because current ones (e.g. *E. coli* and yeast) all carry baggage from evolution in their natural niche. Ideally, we just want something that produces the desired protein, fat, and molecules as efficiently as possible. Certainly, we’ll want to take advantage of the autocatalytic abilities latent in biology. But otherwise, we just need a biological system that satisfies our production objectives. We can dispense with everything else: the circulation system, the bones, the mobility, and most importantly, the capacity for suffering that makes animal production particularly deplorable. Instead, we could conceivably engineer a simple system that solidifies our nutrition and nutrient security requirements. We could gift bioreactors with such production organisms to starving villages in Niger and Afghanistan, and farmers could reap sufficient food in a trivial amount of time. ## Modern Fermentation-derived Meats The numbers and fundamentals I’ve presented also suggest that there’s a considerable financial opportunity for entrepreneurs and industrialists to exploit in the shift from animal products to fermentation-derived goods. So where is the gold rush? In mid-2020, a state of the fermentation industry report was published by the Good Food Institute (GFI), the non-profit advocacy group and funder of science-based innovation with a mission to replace animal products.[^130] The report highlights sixty-eight companies using or pursuing microbial fermentation to replace common animal products. Some are focused on specific proteins, such as Clara Foods, which is producing egg albumin using yeast, i.e., the specialized case mentioned earlier. And the report also highlights twenty biomass-focused companies, producing meat that is purely bioreactor-derived material like Quorn. All of these biomass companies, with the exception of Monde Nissin, the company that produces Quorn, were founded in 2013 or later. Four-fifths were founded in 2016 or later. There is a gold rush; it’s happening right under our eyes. And we can expect more companies to enter the fray. The GFI report highlights the expanding investment—more money was invested in microbial fermentation (about $435 million) to replace animal products in just the first half of 2020 than the previous year, which itself had more fermentation investment than any year before. The microbial fermentation companies are well attuned to the production advantages of their technology. Meati, one of twenty biomass companies, like Quorn, produces faux chicken and steak from fungal mycelia. Meati echoes notions presented throughout this chapter directly on their website: “Our mycelium becomes ‘full-grown’ overnight…if you can believe it. In case you didn’t know, a cow can’t do that.”[^131] The Quorn fermentation process can double its mass every three to four hours;[^132] that’s almost as good as yeast in **Figure 10**. Yield is a work in progress at thirty percent, mostly due to the additional step of removing the excess nucleic acids that cost the process a whopping thirty percent in efficiency.[^133] There is room to improve this separation though. These companies and this technology are limited by consumer acceptance. As we saw with Quorn, a pharmaceutical-level testing and approval process was needed before it was deemed safe to eat. Furthermore, microbial fermentation would be greatly enhanced with access to genetically modified organism (GMO) technology. In the next chapter, we’ll further examine genetically modified technology as a potentially decisive tool in this quest, particularly getting over the misplaced wariness of it. ## Chapter Terms - **yield:** how much of a desired output versus input a chemical process achieves. For example, two cups of flour may yield two dozen cookies. - **biological maintenance:** energetically-demanding activities within a biological organism that are essential for survival and growth but do *not* result in additional biomass - **ontogeny:** the development of an organism from infancy into adulthood - **animal ontogenetic model:** a flattening curve describing the development of animals. Infants grow quickly but slow down later, due to maintenance costs, and eventually stop growing, as do all animals. - **limited:** when discussing a process, the speed (productivity) will always be bracketed by specific physics. We, therefore, can say that a certain process is limited by something. - **productivity:** the speed of a process. For example, my kitchen and I have a productivity of twenty cookies per hour. - **growth rate:** how fast an organism adds biomass to itself; biomass productivity - **diffusion:** the tendency of molecules to disperse, particularly in fluids. To some extent, diffusion supplies nutrients for all biological life; however, diffusion alone is not fast enough to sustain life beyond the microorganism domain. - **allometry:** the study of biological metrics (e.g. metabolism, heart rate) as a function of organism size - **in vitro meat:** taking a stem cell—say from an animal—and growing that cell in a bioreactor into tissue - **Pareto frontier:** the “ceiling” or limit of a biological species. Once organisms reach their Pareto frontier, minimal evolution can be expected unless their niche (environment) or objectives (e.g., domestication versus wild) change. - **archetypes:** variants at the vertices of the Pareto frontier for a biological trait/phenotype. They serve as an idealized representation of what’s physically possible for a certain trait, such as growth rate or beak length. - **ribosomes:** cellular machines comprised of RNA and protein used to manufacture protein ## Chapter Summary Animal technology is ultimately limited by its inherent design. In order to grow so large, animals require an intricate circulatory system so that nutrients may be delivered to all parts of their bodies, and waste can be shuttled to exits. Thus, this design limits both the yield and productivity of animal-based products. The food input that animals consume must partially build and maintain these auxiliary features, subtracting the input going directly toward the animal mass. Furthermore, this design caps the speed at which animals can grow. In contrast, microorganisms skirt the limits of diffusion when small enough. As a result, they have vastly better yields compared to animals and other larger organisms. Furthermore, they are capable of growing orders of magnitude faster under the right conditions—ones that we can control via bioreactor processes. Knowing all these features, we must conclude that a microorganism-based process has a fundamentally superior design compared to an animal or plant-based one. And appreciating how processes are technologies that are liable for replacement, we conclude that with enough development, microorganism technology will generate most, if not all, of our food. [^93]: Appl, M. (2011). Ammonia, 2. Production Processes. In Ullmann’s Encyclopedia of Industrial Chemistry. American Cancer Society. . [^94]: Note that in order to properly compare yield of gasoline-based cars to electric vehicles, we have to calculate from the same origin. First let’s start with fossil fuel. Gasoline cars take the fossil fuel directly (yielding at generally 20% energy efficiency). Electric vehicles are powered by the grid, which is often powered by fossil fuels. Even though the grid is relatively efficient in reaping energy from fossil fuels, there is additional loss when that’s passed to an electric vehicle. There is only about 25% energy yield for electrical cars, in aggregate, if the grid is fossil fuel powered. However, over 50% yield is achievable for electrical cars if the grid is provided by solar or wind sources. If we try to calculate the yield for gasoline cars in the same way, it’ll be in the single digits or less percentage. Remember fossil fuels are pressured animal carcasses, so they ultimately come from solar sources too, just far, far less efficiently. Sources: • All-Electric Vehicles. United States Department of Energy. (Accessed December 5, 2020). • Metcalfe, M. (2018, July 7). Grid Efficiency: An Opporunity to Reduce Emissions. Energy Central. https:/[energycentral.com/c/ec/grid-efficiency-opportunity-reduce-emissions](http://energycentral.com/c/ec/grid-efficiency-opportunity-reduce-emissions) (Accessed December 5, 2020). [^95]: Keefe, T. J. (2012, April 23). The Nature of Light. :/[www.ccri.edu/physics/keefe/light.htm](http://www.ccri.edu/physics/keefe/light.htm) (Accessed April 11, 2021). [^96]: Insect - Circulatory System. Encyclopedia Britannica. (December 15, 2019). [^97]: My calculations, data, and analyses are written up here and are freely available: . [^98]: I used the Stokes-Einstein equation to calculate the diffusion constant (assume water at 25 degrees Celsius). We’ll assume that dioxygen is a 0.3 nanometer radius, so that’s a diffusion constant approximately equal to 1.5*10-9 meters2/s. Now, we can calculate the time to diffuse using the diffusion time equation (for 2 micrometers, or roughly the length of *E. coli* bacterium) and obtain about 2-3 milliseconds. Sources for these values: Einstein Relation (Kinetic Theory). (2019). Wikipedia. (Accessed December 11, 2019). Milo, R., & Phillips, R. (2015). Cell Biology by the Numbers (1st edition). Garland Science. (Accessed December 11, 2019). [^99]: West, G. B., Brown, J. H., & Enquist, B. J. (2001). A general model for ontogenetic growth. Nature, 413(6856), 628–631. . [^100]: West, G. (2017). Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of Life in Organisms, Cities, Economies, and Companies (First Edition). Penguin Press. [^101]: My calculations, data, and analyses are written up here and are freely available: . [^102]: Van Iersel, M. W. (2003). Carbon use efficiency depends on growth respiration, maintenance respiration, and relative growth rate. A case study with lettuce. Plant, Cell and Environment, 26(9), 1441–1449. . [^103]: Liu, Y., Yang, Y., Wang, Q., Du, X., Li, J., Gang, C., … Wang, Z. (2019). Evaluating the responses of net primary productivity and carbon use efficiency of global grassland to climate variability along an aridity gradient. Science of the Total Environment, 652, 671–682. . [^104]: Lötscher, M., Klumpp, K., & Schnyder, H. (2004). Growth and maintenance respiration for individual plants in hierarchically structured canopies of Medicago sativa and Helianthus annuus: The contribution of current and old assimilates. New Phytologist, 164(2), 305–316. . [^105]: Shu, S. miao, Zhu, W. ze, Wang, W. zhi, Jia, M., Zhang, Y. yuan, & Sheng, Z. liang. (2019). Effects of tree size heterogeneity on carbon sink in old forests. Forest Ecology and Management, 432(June 2018), 637–648. . [^106]: Kempes, C. P., Dutkiewicz, S., & Follows, M. J. (2012). Growth, metabolic partitioning, and the size of microorganisms. Proceedings of the National Academy of Sciences of the United States of America, 109(2), 495–500. . [^107]: Why Is Saffron so Expensive? Quora. (Accessed December 15, 2019). [^108]: Johnson, W. (2001) Prelude to the Irish Famine: The Potato. Ireland Story. (Accessed December 15, 2019). [^109]: Irish Potato Famine. (2019, February 19). Extra Credits. (Accessed December 15, 2019). [^110]: Woodham-Smith, C. (1992). The Great Hunger: Ireland: 1845-1849 (Reissue edition). Penguin Group. [^111]: Smil, V. (2013). Should We Eat Meat?: Evolution and Consequences of Modern Carnivory (1st edition). Wiley-Blackwell. [^112]: White, M. C. (2016, March 20). Why Whole Foods Is Saying No to Fast-Growing, Beefed-Up Chickens. Money. (Accessed December 15, 2019). [^113]: Thanksgiving turkeys can’t have sex because their breasts are too big. (2011, November 17). Grist. (Accessed December 15, 2019). [^114]: Lin, Linly. (2015, November 27). Fattest-Ever U.S. Cattle Herd Signals End to Record Beef Prices. Bloomberg. (Accessed August 3, 2021). [^115]: Foer, J. S. (2010). Eating Animals. Back Bay Books. [^116]: My calculations are here: . [^117]: Kleiber, M. (1947). BODY SIZE AND METABOLIC RATE. Physiological Reviews, 27(4), 511–541. . [^118]: Celebrated Charities That We Don’t Recommend. (2009). The GiveWell Blog. (Accessed December 18, 2019). [^119]: The water benefit would primarily come from savings. A bioreactor will provide more food for less water than something like a goat. [^120]: Savir, Y., Noor, E., Milo, R., & Tlusty, T. (2010). Cross-species analysis traces adaptation of Rubisco toward optimality in a low-dimensional landscape. Proceedings of the National Academy of Sciences of the United States of America, 107(8), 3475–3480. . [^121]: Thornton, P. K., & Herrero, M. (2010). Potential for reduced methane and carbon dioxide emissions from livestock and pasture management in the tropics. Proceedings of the National Academy of Sciences of the United States of America, 107(46), 19667–19672. . [^122]: Thornton, P. K., & Herrero, M. (2010). Potential for reduced methane and carbon dioxide emissions from livestock and pasture management in the tropics. Proceedings of the National Academy of Sciences of the United States of America, 107(46), 19667–19672. . [^123]: Pegurier, E. (2017) Study Links Most Amazon Deforestation to 128 Slaughterhouses.” Mongabay Environmental News. (Accessed January 18, 2020). [^124]: Shoval, O., Sheftel, H., Shinar, G., Hart, Y., Ramote, O., Mayo, A., … Alon, U. (2012). Evolutionary Trade-Offs, Pareto Optimality, and the Geometry of Phenotype Space. Science, 336(6085), 1157–1160. . [^125]: Shoval, O., Sheftel, H., Shinar, G., Hart, Y., Ramote, O., Mayo, A., … Alon, U. (2012). Evolutionary Trade-Offs, Pareto Optimality, and the Geometry of Phenotype Space. Science, 336(6085), 1157–1160. . [^126]: Towbin, B. D., Korem, Y., Bren, A., Doron, S., Sorek, R., & Alon, U. (2017). Optimality and sub-optimality in a bacterial growth law. Nature Communications, 8. . [^127]: You, C., Okano, H., Hui, S., Zhang, Z., Kim, M., Gunderson, C. W., … Hwa, T. (2013). Coordination of bacterial proteome with metabolism by cyclic AMP signalling. Nature, 500(7462), 301–306. . [^128]: Schuetz, R., Zamboni, N., Zampieri, M., Heinemann, M., & Sauer, U. (2012). Multidimensional optimality of microbial metabolism. Science, 336(6081), 601–604. . [^129]: Towbin, B. D., Korem, Y., Bren, A., Doron, S., Sorek, R., & Alon, U. (2017). Optimality and sub-optimality in a bacterial growth law. Nature Communications, 8. . [^130]: Specht, L., & Crosser, N. (2020). State of the Industry Report - Fermentation: An Introduction to a Pillar of the Alternative Protein Industry. Good Food Institute. . [^131]: “How Is Fungi-Based Meat Made?.” Meati Foods. (Accessed January 19, 2021). [^132]: Wiebe, M. (2004). QuornTM Myco-protein — Overview of a successful fungal product. Mycologist, 18(1), 17–20. . [^133]: Moore, D., Robson, G. D., & Trinci, A. P. J. (2020). 21st Century Guidebook to Fungi (2nd edition). Cambridge University (Accessed December 2, 2020).