[CW: discussion of creepy bugs and animal suffering, some links show those things. IANAB]

What's the difference between parasites and predators? I'll re-ask this question in different forms as we go. I invite you to figure out the answers for yourself. I'll give an answer, but it may be wrong and it's definitely not the whole story.

1. Natural enemies

The Lion

It's after dark. Drawn by sound or scent or habit to a herd of buffalo, or a loner, he approaches his prey, from their back or from behind vegetation. When he's close enough, or the herd is spooked, he sprints out towards them in a burst, visually homing in on the smallest and slowest. The lion leaps up to grab the dorsal skin or ridge of hair between his jaws, or grapples the buffalo's legs. He holds on to his prey, or backs off to try another tackle or grab, and if he can, he overpowers the buffalo and drags it to the ground. He closes his jaws around his prey's neck, suffocating it. When the buffalo stops struggling, the lion has his meal of flesh.

The Worm

From A Functional Biology of Parasitism (Libgen):

Eggs of [the trematode worm, Alaria canis] are released from the parasite and shed in the faeces. After a period of incubation, eggs hatch and release motile miracidia that seek out and penetrate individuals of any of several species of the pulmonate snail, Helisoma. After two generations of sporocysts, the daughter sporocysts begin production of furcocercous, or fork-tailed, cercariae. These free-swimming cercariae, when released from their snail hosts, are particularly sensitive to water currents. When tadpoles of any of several species of ranid frogs swim in the vicinity of the cercariae, the larval trematodes will attach themselves to the skin of the tadpole, drop their tails and penetrate the tadpole's surface. Once inside, the cercaria develops into an unencysted mesocercaria, a larval stage that is intermediate between a cercaria and a metacercaria. The life cycle may then go in one of two directions. If an infected tadpole or a metamorphosed adult frog is eaten by a canid, the mesocercaria will be freed from the ranid's flesh by digestion and penetrate the gut wall of the new host. It will migrate through the diaphragm and into the lung where the mesocercaria transforms into a metacercaria. Eventually, the parasite then migrates up the trachea and is swallowed. In the small intestine of the canid definitive host, it matures sexually (note that the presence of metacercariae and sexually mature adults in the dog make it both an intermediate and definitive host, a most unusual arrangement). This unusual migratory odyssey is made even more remarkable by the fact that canids normally do not prey on either tadpoles or adult ranid frogs. In order to bridge this ecological barrier, a paratenic host is employed. In this particular case, water snakes eat infected tadpoles and adult frogs. Mesocercariae are freed from the flesh of the second intermediate host and penetrate the gut wall of the snakes where they enter the coelom and remain as mesocercariae. Over a period of time, the numbers of mesocercariae increase rather dramatically in the snake. When canids eat water snakes, they acquire mesocercariae which migrate as if the parasite had been directly transmitted to the canid by an infected tadpole or an adult frog.

Other natural enemies

A natural enemy is a living thing that feeds on another living thing. Parasites and predators are natural enemies.

Predators: Lions, tigers, bears. Foxes, dogs, hyenas, killer ants. Spiders. Sharks, pirahnas, octopi. Eagles. Anteaters. Praying mantises. Snakes, frogs. Dolphins, whales. Storks. Crocodiles. Owls. Ladybugs. Penguins. Scorpions, crabs, dragonflies, giant centipedes.

Parasites: Eukaryotes like Plasmodium (which cause malaria). Bacteria like Mycobacterium tuberculosis and Clostridium tetani. Worms like Schistosomes. Viruses. Ticks, mites, leeches, lice, fleas, mosquitos. Ten meter long intestinal tapeworms.

2. Clusters

Reference clusters

Adapted from A Functional Biology of Parasitism (Libgen):


____________________________________________________________________________________________
|                                       | Parasites                 | Predators            |
| Feeding strategy                      | Gradual feeding           | Total consumption    |
| Size relative to prey                 | Smaller than prey         | Larger than prey     |
| Reproductive potential                | Greater than prey         | Less than prey       |
| Population size relative to prey      | More numerous             | Less numerous        |
| Specificity of prey species           | Usually specific          | Usually non-specific |
| Effect on prey                        | No reduction in viability | Prey killed          |
| Superinfection by a single species    | May reduce prey viability | Does not apply       |
| Density-dependent effect on prey pop. | No density effect         | Yes, Lotka–Volterra  |
--------------------------------------------------------------------------------------------

For some purposes of asking the question "Parasites and predators, what's the difference?", having these two lists of features is already enough. To guess whether there are more Schistosoma than humans, we can just note that Schistosoma are tiny, and predict that they're also numerous, and also guess that probably a single Schistosoma is likely to have many more offspring than a human could. We use a set of related tools and methods to defend against parasites as a class, and a different mostly non-overlapping set of tools and methods to defend against predators.

These clusters of features are also sufficient to reference what we mean by "parasite" and "predator" (though just giving a list of examples is as good or better). If you gave someone these lists of features, I think they could correctly classify natural enemies that are centrally predators or centrally parasites, and then they could investigate the empirical clusters thereby identified.

Mere clusters

Have we answered the question? In asking what the difference is between predators and parasites, are we satisfied with this list of features? Can we expect to get any more of an answer? Is there some essence of parasite-ness and essence of predator-ness we can discover?

Just because we have these two words, "parasite" and "predator", doesn't mean there are corresponding essences. Once we know all the answers to questions of fact about the lion, there's nothing else to ask about the lion itself. Just because it feels like there's a central node Blegg vs. Rube, doesn't mean there's something in the word that "really a Blegg" stands for. (How An Algorithm Feels From Inside (LessWrong))

Though, we might still want to know good ways to think about the lion and the worm as one among many other animals. Even if we reached a point where there really weren't any further questions of empirically answerable fact about the lion or the worm, we might still want to understand other species using the lion and the worm as analogs. In that case, these underlying nodes like "Blegg" or "predator" become relevant and truth-value-bearing as compressions of predictions of empirically answerable facts about analogous species.

So what's the essence of parasite-ness?

Dredging essences

Maybe we can get there by showing which traits are the way they are due to other traits; when we get to the bottom of things, that's the essence.

(Throughout, I'll take dimensions to be relative, so "the predator is large" means "the predator is large, relative to its prey"; a fox is large, i.e. it's large relative to mice, and a lion is large, i.e. large relative to antelopes, while a worm is small, i.e. small relative to pigs. And I'm going to sometimes call the host of a parasite its "prey".)

Why aren't parasites lethal? Because they feed gradually, only eating a tiny portion of their prey, which doesn't kill it.
Why do predators eat their entire prey? Because as long as they've killed their prey, they may as well consume the whole thing.
Why are parasites small? So that they require less food and can therefore have many offspring.
Why don't parasites have a density-dependent effect on prey population, but predators do? Because they aren't lethal, whereas predators are lethal, so more predators causes fewer prey.
Why are parasites more numerous than prey? Because they don't have density-dependent effects on the prey population; predators can't be more numerous than their prey.
Why do parasites feed gradually instead of eating their whole prey? Because they're tiny. If you're tiny, you only eat a tiny bit of your prey at a time.
Why are predators lethal? Because they're large, so they have to eat a lot, which kills their prey.
Why are parasites numerous (relative to prey)? Because they're tiny, so one prey can feed many parasites, whereas many prey are needed to feed one predator. There can't be more large predators than prey, at least not for very long.

A lot of the features in the table are explainable as consequences of size. Predators are lethal because they're big, and therefore are able to kill their prey. They eat their whole prey because they're big enough to eat that much. They're less numerous and have less reproductive potential than their prey because they're big, so it takes more prey to feed them. They have density-dependent effects on prey population because they're lethal, which is because they're big.

So is size the answer to our question? Is large size relative to prey the essence of predator-ness, and small size the essence of parasite-ness? Are the other features explainable as consequences of size? What does this explanation predict about other features?

"Because"

There was something very fishy about the reasoning in the previous section. It's easy to come up with circular chains of "because". Are predators lethal because they're big and powerful and can kill their prey, or are they big in order to be lethal, or do they have to be lethal in order to support their big bodies? We mixed up different kinds of "because", and besides our reasoning was just not reliable, and so we got circularities and contradictions. In one spot we said predators eat their entire prey because the prey is dead and they may as well, but in another spot we said predators kill their prey because they have to eat a lot of the prey's body, which kills it.

So this plan of getting to the essence of predator-ness is dubious, if we're not more careful about what it means for one feature to "explain" another feature. Maybe there's not an answer to be found here. It's useful to find constraints between features with these "because" arguments, e.g. because contraints explain correlations, but maybe there's nothing further to say; various constraints describe natural enemy species, and that's it.

On the contrary, correctly described constraints do not have to be silent about "because" and "explains". I elaborate on this here: Expanding the domain of discourse reveals structure already there but hidden. Causal relations, directed logical implications, and the strength of relations can be revealed by considering more causal elements, fewer assumptions, and more cases.

Besides that, our reasoning just wasn't very solid. Some of the questions we asked assumed a premise that isn't universally true; some parasites (e.g. malaria) have density-dependent effects on population, some parasites are lethal (e.g. Toxoplasma gondii sometimes causes rat hosts to be eaten by cats), and so on. We have to look for reasons that make sense of noisy data where true universals are relatively rare.

We could try to be more careful about deductions and get to the bottom of things. The approach of this essay is instead to make hypotheses about the essence of parasiteness and predatorness, and see if those hypotheses make sense of the data. Our first hypothesis, in the previous section, was size: size seems to be naturally correlated with, if not explanatory for, many features. What else is left to explain?

3. Solid dichotomy

Gap

Whether or not size is the essence of parasite/predator-ness, there's still a phenomenon begging to be explained: the dichotomy of size.

Stereotypically, predators are larger than their prey: the lion eats the lamb, the cat eats the mouse, the eagle eats the squirrel. This is mostly right; really though, predators often eat animals that are bigger than they are: the lion eats the zebra, the eagle eats the raccoon, the wolf eats the bison. Still, there is a size limit. Lions rarely hunt hippos or elephants (especially adult elephants, which are ~20 times the weight of a single lion; a pride of lions might take down an adult), and humans have no squirrel-sized predators. This paper gives these graphs ("Body sizes of animal predators and animal prey in food webs" (PDF)):

With few exceptions, both by weight and length, predators are at least 1/10th the size of their prey, and in the great majority of predator-prey pairs the predator is larger. (Except for vampire bats. Vampire bats are weird.) (Is this right? What are some other counterexamples?)

Parasites, on the other hand, are much smaller than their prey, I'm guessing almost always at least 10000 times (and very nearly always >500 times) smaller than prey by weight at the time of encountering the prey (so, not counting parasites that grow large when feeding). (Except for vampire bats. Vampire bats are weird.) [[Citation needed! Not sure this is right. Especially, there may not be a dichotomy if the prey is small in absolute terms, e.g. if the prey is an insect (mosquitos for example supposedly sometimes feed on cicadas (Wiki)); or if the prey is fish. Leeches may be a counterexample, I didn't immediately find info on what mammals leeches suck blood from. My reasoning: first, I just can't think of parasites that are that big. Second, ticks are among the biggest things centrally called parasites. Mice are among their smallest prey. (I think? Their usual prey is deer.) If I'm skimming these random things rightly (Link 1) (Link 2), pre-engorgement adult ticks are <10mg, and so if white-footed mice are ~20g, that's a factor of 2000. (Though post-engorgement adults are only 100x smaller.) Most parasites are smaller, e.g. lice, filariform worms, bacteria, viruses. Even if the dichotomy is not total, it still seems real.]]

Assuming these numbers are right, why is there this dichotomy? Size determines lots of features, such as population size, so if there are two discrete clusters of animals with either one whole set of features or the other whole set of features--parasite features or predator features--that's explained by a dichotomy of size: the big ones form the big-predator cluster, and the small ones the small-predator cluster. But why are there "tiny predators" (parasites) and "big predators" (predators) in the first place, with nothing in between? To rephrase, taking the standard term "natural enemy" (which excludes other heterotrophs such as scavengers), why aren't there mid-sized natural enemies?

Appeal to authority

Let's see what is standardly said about parasites vs. predators. This blog (parasiteecology) gives what's claimed is a standard definition: a predator feeds off multiple prey in its lifetime; a parasite feeds off a single prey (per stage of its lifecycle).

Combes (The Art of Being a Parasite (Libgen)) states that there's "a major difference between the two types of systems", explaining:

In the predator–prey system, there is no post-encounter interaction. The mouse is eaten as soon as it is captured. [...] By contrast, we can say that in the host–parasite system, the real action of the interaction actually begins after the encounter: either the mouse is able to destroy the infective stage of the parasite or the parasite is able to survive in the mouse. If the parasite survives, it manages to install itself in the right microhabitat and reproduces. This state can last for weeks or months, even for years. This second phase in the host–parasite relationship, after the encounter, is that of compatibility. We say there is a lasting, intimate interaction between the two partners.

In other words, predators don't interact with prey after their initial encounter; parasites do interact after the encounter, ongoingly.

These are two different definitions, but maybe they boil down to the same thing? A natural enemy that is large relative to its prey must eat many prey; if it's small, there's no need to spend energy and risk the outside world in search of new prey, so it feeds on one. A large natural enemy has no reason to interact with their prey after the initial encounter, since they've already eaten it; a small natural enemy will stick around to opportunistically continue feeding on the same prey.

The mosquito in the ointment

Hold on... What about mosquitoes? Mosquitos are small. But they have many prey, and a mix of other features. Here's the table, expanded with features discussed above and with a * marking features of mosquitos.


_____________________________________________________________________________________________
|                                       | Parasites                  | Predators            |
| Feeding strategy                      | *Gradual feeding           | Total consumption    |
| Size relative to prey                 | *<1/1000 as big as prey    | >1/10 as big as prey |
| Reproductive potential                | *Greater than prey         | Less than prey       |
| Population size relative to prey      | *More numerous             | Less numerous        |
| Specificity of prey species           | Specific prey species      | *Non-specific prey   |
| Effect on prey                        | *No reduction in viability | Prey killed          |
| Superinfection by a single species    | *Reduces prey viability    | Does not apply       |
| Density-dependent effect on prey pop. | *No density effect         | Yes, Lotka–Volterra  |
| Number of prey per natural enemy      | One prey per life stage    | *Multiple prey       |
| Interaction after first encounter     | Lengthy, ongoing           | *[fairly] Brief      |
---------------------------------------------------------------------------------------------

Odd. Both our authoritative definitions call mosquitos predators, but most of our features have it that mosquitos are parasites. What should we predict about mosquitos? On further investigation, will they prove to have more predator-like features or more parasite-like features? Here's the table, extended again with more features.


_____________________________________________________________________________________________
|                                       | Parasites                  | Predators            |
| Feeding strategy                      | *Gradual feeding           | Total consumption    |
| Size relative to prey                 | *<1/1000 as big as prey    | >1/10 as big as prey |
| Reproductive potential                | *Greater than prey         | Less than prey       |
| Population size relative to prey      | *More numerous             | Less numerous        |
| Specificity of prey species           | Specific prey species      | *Non-specific prey   |
| Effect on prey                        | *No reduction in viability | Prey killed          |
| Superinfection by a single species    | *Reduces prey viability    | Does not apply       |
| Density-dependent effect on prey pop. | *No density effect         | Yes, Lotka–Volterra  |
| Number of prey per natural enemy      | One prey per life stage    | *Multiple prey       |
| Interaction after first encounter     | Lengthy, ongoing           | *Brief, punctual     |
| Vision                                | No eyes                    | *Yes eyes            |
| Way to encounter prey                 | Vector; diet; habitat      | *Sensory homing      |
| Speed relative to prey                | Slow moving                | *Fast moving         |
| Somatic traits to evade prey defenses | *Anti-immune/pain/clotting | Visual camouflage    |
| What's consumed                       | *Specific tissue/substance | Non-specific flesh   |
| Location of living and mating         | In or on prey              | *Away from prey      |
| Lifespan relative to prey's           | *Much shorter              | Comparable           |
---------------------------------------------------------------------------------------------

(These features are not at all universal among even unambiguous parasites or predators; for example, while lice IIUC live and mate on hosts (occasionally hopping between hosts), ticks and worms have transitional periods off hosts, and ticks mate off-host.)

Note that some of these features are in some cases ambiguous / continuous; lions have no interaction past their punctual encounter with prey, endoparasites have a lengthy interaction, but is a mosquito's interaction ongoing or punctual? What about a vampire bat who remembers and returns to the same sleeping goat on multiple consecutive nights? (Vampire bats are weird.) Also, adult female mosquitos usually stop feeding after a single blood meal (or sometimes two) to rest and develop and lay eggs, and have only a few blood meals in their short life: "...we determined that the overall frequency of multiple blood meals was 18.9% [in Anopheles arabiensis]." (Frequency of Multiple Blood Meals). Typical predators have many meals per reproductive period.

Mosquitos break our supposed correlation / clustering. They share many features with parasites and many with predators, and on further investigation they keep sharing more features with both parasites and predators. They also invalidate some of our reasoning: we argued that large natural enemies have multiple prey because they couldn't survive off one small prey, and small natural enemies have single prey because there's no need to find multiple prey. But mosquitos are small and have multiple prey, and it's not because one prey isn't big enough to feed them.

This somewhat randomly chosen paper ("Parasites as predators" (Sci-hub)) lists a few parasite vs. predator dichotomies, argues that each has issues, and then says, "We therefore prefer to recognize predators and parasites as overlapping classes of natural enemies, distinguished by consumption of prey tissue or symbiosis with a host organism, respectively.". Maybe in light of the table, we should follow suit and give up looking for any sort of "essence of predatorness" or "essence of parasiteness"; we can be sure that the picture, whatever it is, won't be simple and will involve overlapping crisscrossing features and classifications, so why think there's any further "overarching" features to find, rather than just more detail?

Contra Wittgenstein on family resemblances

Wittgenstein (Philosophical Investigations, (Libgen)) argues that it's a mistake to think that clusters imply some feature shared by every member of the cluster. In the context of critiquing generalizations about language, he writes:

For someone might object against me: 'You take the easy way out! You talk about all sorts of language-games, but have nowhere said what the essence of a language-game, and hence of language, is: what is common to all these activities, and what makes them into language or parts of language. So you let yourself off the very part of the investigation that once gave you yourself most headache, the part about the general form of propositions and of language.'.

He agrees to this description of what he's doing, and explains: if for example we notice "the various resemblances between members of a family: build, facial features, colour of eyes, gait, temperament, etc. etc.", then "if you look at them you will not see something that is common to all, but similarities, relationships, and a whole series of them at that". [The second quote is about "games", but IIUC the point is supposed to also apply to what he calls a family resemblance.] Rather, in e.g. calling things "numbers": "we extend our concept of number as in spinning a thread we twist fibre on fibre. And the strength of the thread does not reside in the fact that some one fibre runs through its whole length, but in the overlapping of many fibres.".

My counter-claim: many-featured clusters and correlations are very unlikely to show up for no reason. There's very likely some underlying reason/s for the clustering; families show a family resemblance because they have some relations of descent, games are forms of play that use self-imposed constraints on players's behavior to put the players in another (more fun) situation. When we use a word for something (nouns and verbs), we are, so to speak, applying a theory that there are clusters or correlations and that there are probably compact underlying causes. That's why once we recognize something as a table, we believe that we can fix its wobble by putting something under the leg in the air, even if we'd only just discovered that method yesterday; it's a table, so we believe new features are likely shared with other tables, even if we've never observed that correlation. Yes our concepts are multiplex, extensible, ambiguous, incomplete, indefinite, and so on, and even more so our language; but still, our concepts have boundaries and contours, and by virtue of labeling a family resemblance our concepts will generally also point to some compact underlying generators of that family resemblance, which give rise to further clustering / correlation that we'll find if we look. Since clusteredness tend to point to more clusteredness, that we use a word to refer to some things is evidence that there is more clusteredness among those things to be made explicit.

All this is to say: Where are all the Reggs and Blubes? Something got rid of them, or something is making Bleggs specifically and Rubes specifically, or there's some incompatibility in mixing features too evenly between those of typical Bleggs and typical Rubes, or all the features of Bleggs tend to gradually causally reinforce each other, or something else, but there pretty much has to be a reason. "So why might someone feel an impulse to go on arguing whether the object is really a blegg?" (Quoting: How An Algorithm Feels From Inside (LessWrong)) It might be a confusion; or, they might be hypothesizing that there's a reason that there are Bleggs and Rubes, and they're wondering which side of the dichotomy this object falls into.

Etymons

If we're going to stubbornly hold on to our intuitions that there's "something there", we may as well go full naive-meaning-ist for a moment: what do the words "parasite" and "predator" mean, literally, etymonically?

(Wiktionary: parasite)

From Middle French parasite, from Latin parasitus, from Ancient Greek παράσιτος (parásitos, "person who eats at the table of another"), from noun use of adjective meaning "feeding beside", from παρά (pará, "beside") + σῖτος (sîtos, "food").

(Wiktionary: predator) (Wiktionary: praeda) (Wiktionary: gʰed-)

Borrowed from Latin praedātor, from praedor ("loot, pillage"), from praeda ("booty, spoils, prey"). Likely from the o-grade Proto-Italic *praiɣodā, from (with the prefix *prai-) Proto-Indo-European *gʰed- ("find, hold, seize, take"), whence also the second element in prehendō and probably also hedera.

Ok, so a parasite is a dinner guest, and a predator is a seizer / apprehender. Makes sense.

4. Back to the data

Generalizations

Do all predators struggle with their prey? No, a bird eating an insect might meet with no resistance from its prey.
Do parasites ever struggle with their prey? Depends what we mean by struggle. Many parasites have chemical adaptations that suppress their host's immune system and resist its effects, which is a sort of struggle. Some ectoparasites latch on, making them difficult to dislodge. I don't think any parasites struggle against their prey muscle push against muscle push.
If a heron rides a hippo to conveniently scan the water for prey, is it being parasitic on the energy the hippo puts into moving around? I don't know, sure? (YouTube)
Do all parasites eat some of their host's tissue? No, tapeworms stay in their host's digestive tract and eat the host's food.
Do all predators have absolutely small brood / litter sizes? Spiders (>100). Do all absolutely large predators have absolutely small brood / litter sizes? Ostriches (up to 40). Do all absolutely large predatory carnivores have absolutely small brood / litter sizes? I don't know, maybe. Do any predators have significantly larger brood / litter sizes than their prey? Spider broods might be 4x the size of moth broods.
Do all predators consume their entire prey? No, pirahnas; pack hunters share, especially large kills.
Do all predators have non-specific prey? Apparently "specialization in army ants is unexpectedly high" (abstract) and probably there are lots of other examples.
Are all parasites non-lethal? No, many parasites are lethal, such as many viruses, and bacterial infections like malaria.
Do all parasites have single prey? No, some leeches have multiple prey (Pond life).
Do all predators see? No, deep sea predators and some cave predators are blind (blog).
Do all predators use sensory homing to encounter prey? No, a web-spinning spider has often already caught its prey by the time it's aware of its prey.
Are all predators fast moving relative to the size of their prey? Not if you believe these researchers claiming based on its metabolism that the colossal squid likely drifts and entangles its prey (abstract), though it doesn't look too slow in this video (YouTube), and not if you count carnivorous plants (Wiki). (Other examples? Maybe lazy spiders? Other "entanglement" predators or ambush predators? Some insectivores like toads (YouTube) and iguanas (YouTube) don't move their bodies fast to catch prey, just their tongues.)
Do any predators use chemical warfare? Yes, e.g. venomous predators such as snakes.

Weird cases

There are many:

Hyperparasites: parasites of parasites. (Yes, there are also parasites of parasites of parasites. No futher though of course.) (Wiki)
Parasitoids: parasites that kill or disable their prey, e.g. wasps that paralyze their prey and lay eggs in them, which eat the prey alive as they grow. (Wiki)
Behavior-altering parasites, e.g. Toxoplasma gondii. (Wiki)
Things that eat plants. Fungi, many arthropods, many vertebrates.
Pack hunters, pirahnas, ant colonies. An individual red-bellied pirahna is ~30x smaller than a capybara, and swarms of pirahna can consume a capybara. Pack hunters can take down animals very roughly 20x larger than their individuals. Ant swarms sometimes eat animals (alive!) that are 1000x or more larger than an ant; slug (YouTube), crab (YouTube).
Enemies of eusocial colonies. Many animals eat ants, termites, bees, or wasps. If an animal eats some members of a eusocial colony, was that a predatory event? Only a small part of the colony was eaten, and the colony went on living. What if the animal keeps returning to the same anthill? Is this beetle an endoparasite (YouTube) of the ant colony?
Viruses. Unlike all other natural enemies I'm aware of, viruses are distinctly non-alive; they don't produce anything at all, and what they steal from their prey is the very means of production. ...Oh well also there's prions (Wiki). Some viruses have viruses (Structural Studies of the Sputnik Virophage).
Brood parasitism. (Wiki)
Large hematophages. (Wiki) Vampire bats (obligate) and some birds (facultative), perhaps contiguously with eating ectoparasites off another animal, drink the blood of wounds created by pecking. Lampreys are another exceptional case; apparently they feed on the blood of e.g. paddlefish (YouTube), which are roughly 150x their weight. Some leeches may also be exceptional in terms of weight. Vampire bats are exceptional in being (if accepted as a parasite) the only mammalian parasite, and the only parasite that is also social [citation needed].
Farming. E.g. humans farming animals, ants farming aphids.
Mutualism. E.g. gut bacteria that digest parts of the host's food that it can't digest by itself and produce something the host can use.
Cancer.
Carnivorous plants. (Wiki)
Parasitic males. Some gonochoristic (sexually dimorphic) species have males that are symbiotically parasitic on females, e.g. anglerfish (YouTube), or are not even really organisms (see Combes ch. 2 on Entoconcha, Thyonicola, and Enteroxenos).
Very selfish genes (Wiki). Transposons (Wiki) are segments of DNA that move around or are copied to other places in the genome, sometimes harmful and sometimes helpfully. B chromosomes are extra chromosomes that piggyback in an organism; in the wasp Nasonia vitripennis, haploid-male transmitted B chromosomes can cause themselves to be transmitted by throwing out the rest of the paternal genome before fertilization (Combes, ch. 2). Also see Combes for a summary of Alain Pagano's research on the genome of Rana ridibunda in some sense parasitizing Rana lessonae to survive in intermediate habitats by producing hybrids that only pass on R. ridibunda genes, never R. lessonae genes.

What do parasites and predators have in common?

Roughly all natural enemies:

have some method, whether passive or active, for encountering their prey;
have methods to prevent the prey from separating itself from the enemy (whether by harming (e.g. by punching or by immune cells), killing, or fleeing from the enemy); and
consume something inside the prey's body, whether flesh, specific tissue, or the prey's own food.

Intuitions

Intuitions are data too, though of a different order. Intuitively, are these parasites? My answers:

Herbivorous animals: no.
Fungus on a tree: yes.
Mosquitos: yes.
Vampire bats: yes?
Pirahnas: no.
Wasps laying eggs in a paralyzed insect: no? But the eggs are?
Brood parasites: only analogically?
Hyperparasites: yes.
Viruses: no, because they aren't alive?

5. The S/D hypothesis

Keeping in mind the exceptions and weird cases, there's nevertheless a big size dichotomy among natural enemies. There are very many members of very many species of tiny parasites of large land animals, and there are many members of many species of large predators of large and small land animals, and there are very few species of medium-sized natural enemies of large land animals.

Squirrel!

What would it be like, if humans had squirrel-sized predators? It wouldn't make sense, we wouldn't just let a squirrel-sized animal bite into our flesh, we'd attack it, remove it, run away from it.

The Sneak/Disable hypothesis: there's a size dichotomy because there are basically two strategies for feeding on live prey. The Sneak strategy tries to avoid being noticed by the prey, and feed unmolested because the prey isn't aware and so doesn't respond by fleeing or fighting. The Disable strategy tries to disable the prey, rendering it unable to flee or fight while the enemy feeds. If a natural enemy is noticed by its prey and tries to feed, the prey will flee or fight. If a natural enemy is 1/100th the size of its prey, it will be noticed by the prey, and it will be basically powerless to disable the prey, so it will be unsuccessful. That's why there are very few natural enemies 1/100th the size of their prey.

If the chance of being noticed by prey is continuous in the size of the natural enemy, and the chances of successfully disabling the prey is also continuous in size, how do we get the quasi-discreteness of a dichotomy? The situation is like this (oversimplified here):

To successfully feed, a natural enemy can't be both noticed, and also lose the fight once engaged in a struggle with the prey. The chance of that happening is, very roughly, the product of the chance of being noticed and the chance of losing the fight:

Note that even with these oversimplified and impossibly linear curves (e.g. a prey has way less than 5% chance of noticing an enemy 1/10^6th its size), we still get a dichotomy: the probability of feeding being prevented is low on both extremes of size, and at least significant in the middle. In this way, dichotomies can emerge even from two independent, monotone, smooth (say, linear) factors, if the factors are 1. monotone in opposite directions, and 2. combined by multiplication (or by logical AND, or by taking the minimum, or anything like that).

That said, in the case of natural enemy size, the dichotomy is stronger because noticing is somewhat discrete (i.e. has a regime of especially high slope, i.e. its derivative looks like a bump). A slightly more realistic graph (though still very imprecise, and leaving the animals involved unspecified):

So the Sneak/Disable hypothesis says that the reason there's a dichotomy between natural enemies that are relatively way smaller than their prey and natural enemies that are at least relatively not much smaller than their prey, is that there's a region of disallowed sizes demarcated on one side fairly sharply by the enemy being noticeable by the prey, and demarcated fairly softly on the other side by the enemy losing in a fight against the prey, where natural enemies can't exist because they would be noticed and then lose the ensuing fight.

Fiddly clarifications to make the claim more precise

By "noticed", I mean noticed by whatever controls the gross movements of the prey (so basically, the nervous system controlling muscle movements, or for tiny creatures without muscles, whatever sense apparatus makes it wiggle around or squirt chemicals or whatever). A Sneak may in some sense be "noticed" by the prey's immune system, but that's a type of defense separate from muscular force (/gross movement). Muscular force is important because it scales with the size of the organism more than immune strength does [citation needed]. I'm guessing that immune responses are stronger in larger animals, since they have more resources, more ability to withstand attrition, and longer lifespans to train recognizition; but only sublinearly stronger, because there's a defeat-in-detail (Wiki) dynamic, where the strength of an immune response is limited by e.g. the surface area of the enemy, but the strength of a muscular response depends mainly on the size of the wielder. A monkey can easily remove a tick from its (or its friend's) skin, but it will have trouble dealing with tiny endoparasitic worms with biochemical adaptations for fooling, disabling, or weathering its immune responses. Likewise, a endobiotic Sneak may cause illness that its host notices, but the host won't know what's happening and won't be able to do anything about it by that point (except for humans).
By "noticed", I mean noticed at the time of first encounter, when the prey's muscular movements could affect the enemy (including the enemy's position relative to the prey, by fleeing from the enemy). So if an endoparasite is at first invisible to its prey, e.g. by being tiny and hidden in the prey's food, but then later causes disease that the prey notices, that still counts as a Sneak strategy because by the time it's noticed, the endoparasite is out of reach of the prey's muscular defenses and can no longer be muscularly fled from.
The S/D dichotomy only applies to natural enemies of a given prey to the extent that the prey flees or fights its natural enemies using its muscular force when it notices them.
Disablers also sneak, i.e. also use stealth. To avoid giving their prey more chance to flee, many predators are preferentially nocturnal (e.g. lions and owls), have a camouflaged appearance (e.g. leopards), and approach quietly, under cover, and/or with great speed. The difference is that Disablers can't hope to feed on prey without being noticed. Likewise, some Sneaks disable parts of their prey, e.g. the immune system or sensory apparatuses (e.g. anesthetics in ectoparasite saliva), but most Sneaks don't disable their prey's muscular movements.
Note that a Disabler may not ever be actually noticed by its prey. E.g. an insectivorous bird might take its prey completely by surprise, disabling it before it has a chance to notice the bird. But the bird would have been noticed if it tried to somehow feed on the insect without disabling it.
I'm glossing over a potential prey noticing a natural enemy vs. it being worth it for the prey to spend energy fleeing or fighting the enemy. I think that hypothetical meso-sized natural enemies are basically always worth fighting, but that might not be true in all cases.
I'm considering pack hunters like wild dogs to be "one natural enemy". A pack of wild dogs might take down a buffalo that's ~25x the weight of an individual dog, but a lone dog couldn't take down the same prey; so the size dichotomy applies more clearly to "the whole entity that makes the kill".
Disablers might disable by killing or paralyzing their prey. They might disable not by penetrating the prey's body, but by immobilizing the pray with claws, jaws, limbs, webs, etc.; e.g. these lions eat a baby elephant alive after grounding it, since they can't kill it (YouTube). They might also not even permanently disable their prey at all. For example, this bug is eating this other bug alive and might've let it go half-eaten (YouTube); the disabling in this case is simply holding the prey immobile. Another kind of disabling is persistence hunting (Wiki), i.e. chasing prey until it's too tired to flee or fight, practiced by dogs and humans (also, this video claims the mongoose is trying to exhaust the snake (YouTube)). Combes: "bivalve molluscs living in deep hydrothermal vents, such as species of Calyptogena and Bathymodiolus, get their energy from symbiotic bacteria."; humans farm many kinds of animals; ants farm aphids. Thus enclosing the prey is another type of Disabler strategy. Anglerfish shoot prey insects with water, stranding them in a body of water (YouTube); what are other "stranding" Disablers?
The "Disabler" moniker isn't meant to strictly imply decreased evolutionary fitness for the prey (which effect is sometimes used to classify natural enemies), though decreased fitness is by far the usual outcome, since killing is the central Disabler strategy. I don't think it makes much sense to say that a born farm animal's fitness is decreased by being a farm animal, since it wouldn't exist without the farm, and a wild animal caught and bred is in some sense also fit. If two populations are strongly separated, they form (at least) two separate evolutions; the abstractions of "species", "evolution of a species", "gene pool", and "fitness" all break down. (All this is more or less irrelevant to suffering, as opposed to evolutionary fitness; an adaptation-executing rabbit could still suffer, even if it's not counterfactually close to spreading its genes.)
I'm glossing over some factors that change with enemy size, in addition to noticeability and ability to disable the prey. For instance, a larger natural enemy is more worth fleeing from or fighting off, not just more noticeable. Also, the S/D hypothesis focuses on the chance of successful feeding, but there's also the factor of danger to the natural enemy. A larger natural enemy has to feed more frequently and therefore has more interactions with prey, which greatly increases the cost to the natural enemy of being prone to be harmed in the course of a struggle with a prey. A mosquito can risk a few blood meals, but a vampire bat, or for that matter a lion, has to be very wary of harm from their prey, who are after all struggling for their lives (against Disablers) or at least for some of their body. Maybe vampire bats are basically as big as you can get just feeding on blood (which is nutrient poor), pushing the limits by feeding constantly, gorging, and relying on (sometimes non-kin) social bonds of feeding when in need, and any larger creature would have to eat the prey's flesh, which would be much more painful and harmful, hence noticeable and worth defending. I think these factors can be summed up rightly enough as: if a natural enemy is too big to be unnoticed by, or unconcerning to, its prey, but it's too small to chase and fight the prey effectively, then it's going to be overall unsuccessful. In the graphs, "lose a fight" means "can't catch the prey, or can't disable the prey long enough to eat some of it, or can't avoid being too harmed by the prey".
The "Sneak" moniker isn't meant to imply behavioral sneaking. Many parasites encounter their host passively and go unnoticed. We do see sneaking behavior at the upper end of the size range of Sneaks. Lice avoid light, and vampire bats are extremely sneaky: they're active at night, they have exceptionally sharp teeth to make thin painless cuts, they approach prey cautiously, they perhaps have anesthetic in their saliva (at a glance I find some sources claiming this but I'm not sure), and in fact their prey seem to not notice them or be only dimly aware that something's around. (Though, I haven't ruled out that the prey, e.g. a goat or a pig, are sometimes aware of the vampire bat but can't do anything; I think they could do something, e.g. run away or roll on the ground, and would want to, but I'm not totally sure.) What's a better moniker? "Stealth"? "Covert"? "Unseen"?
There's no guarantee that tiny Sneaks won't be relevantly noticed by prey just because they're tiny. For example, some animals may avoid rotten-smelling food, feces, old carcasses, etc., I assume usually due to a disgust instinct evolved under pressure to avoid parasites, though in some cases learned from previously getting sick. In this way, the evolution of humans "notices" the evolution of microscopic worms before the humans notice the worms. But generally both evolved and learned disgust is much less reliable and comprehensive than vision (in their respective realms), and meso- or macro-sized natural enemies have no chance of going unnoticed if they have an extended interaction with their prey. (Except for vampire bats.)
Sneaks may be lethal for a purpose. Toxoplasma gondii makes its rat hosts unafraid of cats, causing them to be eaten, transfering T. gondii to the cat, its next host. That's not an initial interaction with the rat, and it's a strategy for sneaking into the cat.

Consequences

§ Tests

A consequence, and test, of the Sneak/Disable hypothesis, is that the less a prey can defend against enemies, the less its natural enemies will show a size dichotomy. If the prey doesn't notice enemies and flee or fight enemies that it notices, then there's no S/D dichotomy. If the obvious size dichotomy of natural enemies for large animals is really because of the S/D dichotomy and not because of something else, then when the prey doesn't notice enemies or doesn't flee/fight enemies, there's no S/D dichotomy and nothing else to cause / enforce a size dichotomy, so there shouldn't be a size dichotomy.

Plants are a big class of example. Plants have no fast macro movements, don't flee, don't fight, and don't notice. And indeed, as far as I know there's no size dichotomy for things that eat plants. Among animals that feed on (parts of) trees, the full range of sizes, from fungus to nematodes to beetles to caterpillars to mice to squirrels to giraffes, is occupied.

Correspondingly, the stronger an animals behavioral defenses against natural enemies, the greater the divide between large Disablers and small Sneaks. Animals that groom each other, e.g. social primates, should tend to have a greater gap between their largest Sneak and their smallest Disabler. (This could be confounded by there being pressure to groom due to a greater load of Sneaks preexisting for other reasons.) Apex predators have few or no Disabler natural enemies (especially vulnerable lions may be eaten by hyenas or other lions; humans can disable anything), but still have plenty of Sneak natural enemies. Humans have the strongest defenses of all, and some are even planning to extinct entire species of larger Sneaks. (Aside: On the other extreme, viruses are nearly "nadir parasites" (unless there are DNA/RNA sequences that in some sense parasitize viruses), and are sometimes predated on by bacteria (Single Cell Genomics Reveals Viruses Consumed by Marine Protists).) Here's a monkey trying to remove a big leech (YouTube). Itch probably evolved in part to get rid of parasites (as well as other stuff like toxins or dead skin).

(If there are size dichtomies for behaviorally defenceless prey, that weakly argues against the S/D hypothesis, because it implies some other source of size dichotomies, which could also explain the dichotomy for defenseful prey; that counterevidence is removed if the reason for the other dichotomy is known and can be shown to not apply to some large class of defenseful prey. If there's not a size dichotomy for defenseful prey, that strongly argues against the S/D hypothesis, unless there's some surprising reason the dichotomizing force is overcome.)

§ Near the boundaries

The larger a Sneak is, the more likely it is to be noticed by prey, so the more pressure there is for it to be behaviorally stealthy (rather than just small):

Maybe that's why ticks have worse vision than mosquitos, and mosquitos have worse vision than vampire bats (and I'd guess lice have worse vision than ticks, not sure, see Wigglesworth 1941 (nominative determinism!) (Libgen)). (I don't think this is a historico-phylogenetic phenomenon, since ticks are arachnids, and predator spiders have good vision.)
Endoparasites aren't nocturnal any more than their prey, mosquitos are crepuscular, and vampire bats are nocturnal.
Endoparasites don't do any stealth behavior; ticks and lice might flee light; mosquitos avoid body parts in motion (which maybe is also viewable as being engaged in a fight); and vampire bats do sneaky stealthy stuff (YouTube).
Endoparasites take their time eating, whereas ticks, mosquitos, some leeches (Pond life), and vampire bats gorge, drinking multiples of their body weight in one meal, or in the case of vampires, urinating while feeding to make room (someone get Dracula a change of clothes...), presumably to minimize the need to hunt more and be exposed to more danger from prey; then they go somewhere away from the prey to digest and maybe reproduce.

(I think pirahnas seek mainly wounded or dead animals, e.g. by testing prey to see if it responds denfensively, putting them mainly outside the S/D dichotomy; not sure.)

One way this story seems plausibly wrong, is, does a cow really not notice a vampire bat on its back? Maybe there's nothing it can do? (For large things, a bull bucks.) Or maybe it doesn't actually mind enough? This sea lion certainly seems to mind (YouTube) (and according to the filmers, the bats are targeting spots difficult for the sea lion to defend, which is in a grey zone between Sneak and Disabler). These penguins seem to mind, and the babies seem to not notice the sneaking bats (YouTube).

The smaller a Disabler is, the more it's a struggle to catch and disable the prey. Smaller Disablers hunt in packs (or to say it another way, lone Disablers hunt smaller prey). A pride of lions trying to take down a giraffe, a hippo, or an elephant is in for a dangerous struggle with no guarantee of success. Disablers target small, lone, wounded prey when given the chance.

§ Classifications

Normal parasites: All Sneaks.
Normal predators: All Disablers.
Mosquitos: Sneaks.
Leeches, lampreys: Not sure. Either Sneaks, or S/D doesn't apply because their prey are defenceless.
Vampire bats: Common vampire bats are Sneaks. I'm not sure about the other two species; if the hairy-legged vampire bat, which feeds on birds, grapples with the birds, then I'd call this a non-lethal Disabler. Though, the first segment here (YouTube) is maybe a form of exhaustion hunting. (Vampire bats are weird.)
Hematophagous birds: Sneaks if anything, but this is a weird one; a hypothesis about why their prey don't fight back, is that the prey are adapted to let the hematophages pick harmful parasites off their skin, and they don't notice when the hematophages slip into hematophaging. See (YouTube)
Pirahnas: not sure; some pirahna feeding interactions might witness the S/D hypothesis not applying to fish.
Wasps that paralyze their prey and lay eggs in them: Disablers if anything, though this doesn't quite fit in the S/D rubric because it's the wasp's offspring that feed; the larva could be said to be Disablers with the strategy "get Mom to do it".
Fungi: S/D doesn't apply to fungi on trees, which are defenceless (although there's something intuitively very parasite-y about them, so being a Sneak is not the one true essence of parasitism). Fungi growing in an animal is a Sneak. Sneak / Disabler is thus not a fact about an organism, but about a natural enemy feeding on a prey. That said, a main reason to talk about Sneak / Disabler is to understand niches and evolution, so we could say an organism is a Sneak if a major part of its niche is sneaking; this would distinguish rare "random accidental" fungal infections that don't affect the evolution of anything from e.g. fungal STDs.
Farmers: Disablers.
Cancer: S/D doesn't apply because cancer doesn't appreciably evolve (not beyond the lifespan of its host) and so doesn't have a niche. Likewise prions.
Viruses: Sneaks (contra Rick and/or Morty (YouTube) ).
Ants eating a crab: One Disabler, the colony.
Very selfish genes: Sneaks. (This is a weird case because very selfish genes never had behavioral encounters with their prey, and so are in some sense further in the space of counterfactual worlds from behaviorally encountering their prey.)

(Aside: Viruses, STDs, and other contact infections like lice have a cool property: they live almost entirely in "host spacetime", the spacetime induced by the movement and contact patterns of the host, or of parts of the host like the head or genitals. Perhaps there are parasites that live entirely in host spacetime by being transmitted between parent and child when the parent regurgitates food to feed the child, or when the parent breastfeeds the child, or when the parent lays an egg or gestates the fetus.)

6. Trailheads

Other dichotomies, correlations, and questions

Are there other major dichotomies?

I'm curious what other sources of dichotomies there are with natural enemies, as well as features that correlate with each other. As a conjecture meant more to provoke the counterevidence than as an actual guess: Maybe all features of natural enemies are continuous except for those made dichotomous by the S/D dichotomy.

By "continuous feature", what I mean is, there's a fairly homogeneous distribution of natural enemies falling all along the spectrum of that feature. For instance, this conjecture claims that number of prey is continuous, in the sense that "if there were no S/D dichotomy", then there'd be some natural enemies averaging .01 prey in a lifetime, some averaging .1 prey, some 1 prey, some 3 prey, some 10 prey, some 30 prey, some 100 prey. We might see a dichotomy on average prey size, where some tiny Sneaks average .001 prey, ticks average .2 prey or something, lice average 1, female mosquitos average 1.5, and then the natural enemy with the next highest prey count is a Disabler averaging 100 prey per lifespan. But if so, that's just because of the size dichotomy.

This is a somewhat silly conjecture in the sense that, really, every species boundary constitutes a dichotomy (i.e. a discreteness), and every "strategy" (how to disable a prey's immune system, which tissue to eat, how to encounter the prey, how to exit the prey, where and how to reproduce, etc.) potentially gives rise to a cross-species dichotomy of specialization. So really the conjecture / question is asking, are there other "broad, overarching" dichotomies of niche that divide natural enemies into clusters? (This is more or less equivalent to asking about the broad structure of [the space of strategies, along with a function on that space representing fitness].)

The ecto/endobiosis dichotomy

I thought there'd be a dichotomy between endobiotic and ectobiotic natural enemies. But I don't actually see it clearly, at a second glance.

Endoparasites typically have highly specific prey, have exactly one prey per life stage, reside in specific areas in their prey which they reach by specific routes, have chemical anti-immune properties, and mate inside their prey. Macropredators, on the other end of the spectrum, typically have non-specific prey, have many prey, don't reside in their prey, don't have anti-immune properties, and don't mate inside their prey. Small ectobiotic natural enemies, a little ways along the endo/ectobiotic spectrum from endoparasite to macropredator, are partway on these features, I think:

Ticks, lice, and mosquitos feed in somewhat specific areas of their prey (e.g. where the blood is most easily reached), which they find by sensory homing.
Adult ticks take a single blood meal, lice reside on 1 or 2 hosts, while mosquitos have 1 or a few blood meals. Ticks, leeches, and mosquitos have anti-immune substances in their saliva, as well as other chemical warfare (anti-coagulants and possibly anesthetics).
Specificity isn't known very well AFAIK, but I think some species of ticks, lice, and mosquitos have pretty specific prey while others have fairly non-specific prey (Host specificity in a diverse Neotropical tick community) (The population genetics of host specificity: genetic differentiation in dove lice).
Lice mate on hosts, some ticks mate on hosts and some don't, and mosquitos don't mate on hosts.

This apparently continuous pattern might be understandable as following from non-dichotomous correlations; e.g. there's a continuous spectrum of how long a natural enemy is in contact with its prey, and this gives rise to a continuous spectrum of evolutionary pressure to deal with the prey's chemical defenses, which gives rise to a continuous spectrum of prey specificity (assuming that chemical adaptations tend to be fairly prey-specific because tuned to affect chemicals that differ between prey species, as opposed to e.g. adaptations for powerful jaws, which can pretty much bite whatever). (Not that any of these features have simple explanations or explanations in terms of the natural enemy itself, e.g. prey specificity might also be modulated down/up by scarcity/abundance of prey.) See Combes chapter 2 describing a series of parasites along a supposed path of evolution of endoparasitism, giving examples along a spectrum from a commensalist just along for the ride, to an ectoparasite, to a parasite anchored / embedded in the outer layers of the host, to an embedded parasite with more deeply penetrating proboscises, through more deeply embedded parasites and finally to a parasite that releases its eggs into the host's digestive tract; other parasites themselves move around in the internals of the host.

The existence of intermediate cases doesn't demonstrate that there's no dichotomy though; maybe it's rare for parasites to be embedded but only a little deeply in the surface of their prey. Is that so? Note that if there's an endo/ectobiotic dichotomy, it's nested in the size dichotomy, and hence in the S/D dichotomy; a natural enemy can't make it in inside its prey if it's large when it first encounters its prey, so all endobiotic natural enemies are Sneaks. We could try describing natural enemies by these two nested dichotomies as follows (with added features, some of which are guesses):


____________________________________________________________________________________________
|                                       | Sneak                     | Disable              |
| Strategy at initial prey encounter    | Be unnoticed/unconcerning | Disable prey         |
| Feeding strategy                      | Gradual feeding           | Total consumption    |
| Size relative to prey                 | <1/1000 as big as prey    | >1/10 as big as prey |
| Reproductive potential                | Greater than prey         | Less than prey       |
| Population size relative to prey      | More numerous             | Less numerous        |
| Effect on prey                        | No reduction in viability | Prey killed          |
| Superinfection by a single species    | Reduces prey viability    | Does not apply       |
| Density-dependent effect on prey pop. | No density effect         | Yes, Lotka–Volterra  |
| Number of prey per natural enemy      | Few prey                  | Many prey            |
| Interaction after first encounter     | Some ongoing interaction  | Brief, punctual      |
| Way to encounter prey                 | Vector; diet; habitat;    | Sensory homing,      |
|                                       | sensory homing            | ambush, entanglement |
| Speed relative to prey                | Slow moving               | Fast moving          |
| Somatic traits to evade prey defenses | Anti-immune/pain/clotting | Visual camouflage    |
| What's consumed                       | Specific tissue/substance | Non-specific flesh   |
| Lifespan relative to prey's           | Much shorter              | Comparable           |
| Development                           | Multi-staged              | Single-staged        |
| Approach to prey                      | Little or no              | Initially stealthy,  |
|                                       | stealth behavior          | then violent         |
| Nervous system relative to size       | Small                     | Big                  |
| Sociality                             | Never social              | Sometimes social     |
--------------------------------------------------------------------------------------------

I think there are many more species of Sneaks than Disablers, even excluding viruses. See (Wiki).

This table has some guesses, partly (probably over-)generalized from Combes ch. 2:


______________________________________________________________________________________________________
|                                   | Endobiotic Sneaks                | Ectobiotic Sneaks           |
| Specificity of prey species       | Very specific prey               | Somewhat specific prey      |
| Number of prey per natural enemy  | One prey per life stage          | One or a few prey           |
| Interaction after first encounter | Lengthy, ongoing                 | Brief or lengthy            |
| Vision, other senses              | No / reduced sense organs        | Yes eyes and other senses   |
| Way to encounter prey             | Vector; diet; habitat            | Habitat, sensory, host-host |
| What's consumed                   | Blood, other tissue, host's food | Blood, skin                 |
| Location of living                | In prey                          | On or away from prey        |
| Location of mating                | In prey                          | On or away from prey        |
| Digestion                         | Osmotrophic, no digestive tract  | Normal digestive tract      |
| Egg dispersal                     | Through host's excretions        | Away from host, or through  |
|                                   |                                  | a siphon if embedded        |
| Shell                             | No                               | If ancestral                |
| Adult shape                       | Worm-like                        | Whatever was ancestral      |
| Approach to prey                  | No stealth behavior              | Some stealth behavior       |
|                                   |                                  | through whole interaction   |
| Nervous system relative to size   | Very small                       | Somewhat small              |
------------------------------------------------------------------------------------------------------

Mosquitos are almost but not quite perfectly described as Ectobiotic Sneaks. What are some natural enemies that don't fit well with these tables? (Other than vampire bats.) Do the tables fit better if they're used not as classes, but as a bunch of imperfectly correlated spectra with one or two dichotomies?

Questions

I'm sufficiently out of steam for this essay that I won't investigate the situation with fish and small invertebrates and microorganisms. They could provide counterevidence. I'm confused about fish; many fish seem fairly defenceless, and lampreys feed on relatively large-ish fish harmfully. If the S/D hypothesis is right, why aren't there more meso-enemies? Are there? Are fish less defenceless than they seem? But then why lampreys? For small invertebrates, I think I read somewhere that they have relatively larger ectoparasites, but a lot of e.g. insects don't seem defenceless, so I don't know what's going on there. How relatively big do insect parasites tend to be? Can the S/D dictotomy exist even for a particular defenceless species, if a member of such a species is difficult to distinguish from other non-defenceless prey? My guess is that the S/D dichotomy doesn't apply to single-celled organisms because they aren't flexible enough to get a defeat-in-detail advantage when kinetically defending themselves, but I don't know. Bdellovibrio is a bacterium that preys on something that looks to be maybe 100x its size (YouTube) (Wiki)

I'm curious what modulates prey specificity. Hypotheses include scarcity of prey and adaptations like anti-immune adaptations that respond to features not shared across potential prey. Another possibility: many parasites have specific definitive hosts, where they mate to sexually reproduce, and they have other non-definitive hosts, where intermediate hosts are more likely to be non-specific. Maybe parasites that mate in or on their prey underwent an assortative mating speciation, where the (evolutionary/genetic) choice of definitive host serves a sort of coordination point. This predicts that ectoparasites that mate on their hosts should have especially specific prey, comparably to endoparasites, and more than otherwise similar ectoparasites that mate away from their prey. This doesn't explain why non-definitive hosts are sometimes specific. Perhaps certain animals have particularly weak immune systems and so make good hosts for abundantly asexually reproducing life stages. Transitional hosts are then less specic, just whatever trophically links to the definitive host.

An alternative explanation of the prey-count dichotomy, is that there's a dichotomy between specializing in finding many prey vs. not, arising because once a natural enemy has multiple prey, it "rolls down a hill" in the fitness landscape where more skill at finding prey causes finding more prey, which adapts the rest of the organism to having multiple prey (e.g. by having multiple reproductive cycles in place of a single one), which increases the incentive for skill at finding prey. This doesn't explain why mosquitos have a few prey but not many prey; but, neither does the S/D hypothesis, unless most mosquitos die because a prey animal kills it. (Well, maybe S/D can claim that mosquitos are pushing at the edge of noticeability, and so "try" to make due with minimum exposure to prey, which works out to roughly one blood meal per reproductive cycle.) A quick google claims mosquitos die of cold or of being eaten by predators. Maybe there's an explanation in terms of many-prey-specialization as a slightly runaway process with an equilibrium set by factors like pressure from predators. A possible test would be to look at detritivores / scavengers / herbivores; anything with defenceless and sometimes scarce food. At a guess, herbivores don't have a prey-count dichotomy; a fungus lives on one tree, a cankerworm may visit a few trees, a caterpillar eats one or two hundred leaves, a leafcutter ant colony feeds on hundreds of trees, a giraffe feeds from thousands of trees in its life. If feeders with defenceless and scarce food have a size dichotomy, that's anti-predicted by the S/D hypothesis compared to the prey-finding dichotomy: feeders that aren't skilled at finding food might be tiny and go through many generations on a single carcass (which they reach by some unskillful process whose unreliability is compensated by sheer numbers), while feeders that are skilled at finding food might be larger and have better sense organs, brains, and muscles. I'm wary of relying on explanations in terms of runaway processes, not because the dynamic is implausible, but because I don't know how to make simple deductions from them; when, for instance, does an autocausation hypothesis / basin-finding hypothesis predict a dichotomy, vs. a spectrum produced by some controlling factor being distributed continuously on a spectrum (e.g. predator pressure)? Literally any feature could be autocausal if there's runaway recursive perceptual mate selection.

A mystery: Why aren't there 1/50th size Disablers who use venom? Are there? Maybe there's no evolutionary pathway to a venom that's toxic enough to take down a 50x sized prey (though this would be surprising; aren't some venoms that strong? E.g. brown recluse's venom?). Is there some reason that a 50x size prey is not worthwhile, e.g. because it'd be hard to defend such a big kill for long enough to eat much of it or because the kill would rot? That seems strange though. A natural enemy could feed its young, attack scavengers, take the kill back to its lair... For that matter, why don't lions have venom?

A mystery: if there can be vampire bats, why isn't that strategy more common? The answer could just be, by the S/D hypothesis, their niche is very difficult (also because blood is nutrient poor), so few species have filled it. But this is a bit of a weird theory; shouldn't any niche that can be occupied and can be reached by evolution, be occupied roughly to its capacity? One kind of explanantion would explain vampire bats as having reached their niche by first having a pre-adaption, and then expanding / changing their niche towards their current niche after having their old niche "abut" their current niche, in a rare combination (either the pre-adaption or the niche-abutment). For example, it could be that being social is rare enough among small animals (though this seems not right, e.g. rodents are social), and sociality serves as a pre-adaptation present in bats in general. Then some bats followed the path of other large hematophages (YouTube): first, they ate ectoparasites off of some animals; then they occasionally drank blood from ensuing cuts (this is the niche-abutment); then they specialized, and bats got further into the niche than hematophagous birds because they were social and so could handle the precarious obligate hematophage niche by feeding each other through variations in hunting success. (This predicts that if crows, already pre-adapted to be intelligent, social, and clad in black, ever get a taste for blood...) Also, why aren't vampire bats smaller? Are they still evolving to be smaller?

What are some clear examples of an animal noticing a parasite and being capable of getting rid of it, but not getting rid of it?

Why does behavioral defense enforce the S/D dichotomy, but the immune system doesn't enforce another dichotomy for small endoparasites? Or does it? One explanation is in terms of defeat-in-detail; a multicellular worm wins each battle with individual immune cells. (Are there any immune system agents that are multicellular?) But this would still predict a size dichotomy around the size of immune cells. And really, why wouldn't the immune cells be able to clean up anything not much bigger than them? Combes points out that parasites have much greater evolvability because they have large generations with lots of death (Combes: "The common tapeworm of humans (Taenia saginata) produces up to 10 billion eggs during its lifetime."; Genghis Khan eat your heart out), short and hence numerous generations, and a high (and sometimes evolvably variable) mutation rate; therefore cellular-scale offense might simply beat cellular-scale defense for tiny vs. large organisms.

What are some properly flesh-eating parasites? A worm that takes tiny bites of flesh would be an unambiguous example. The first two google results for "flesh eating parasites" are Leishmania and Staphylococcus aureus (of MRSA fame), but I didn't quickly see what their diet is; do they kill flesh incidentally on their way to eating something more specific, or are they actually eating whole cells / undiscriminated cell material? E.g. Wikipedia (Wiki) claims "In order to avoid destruction by the immune system and thrive, the Leishmania 'hides' inside its host's cells.".

Are there natural enemies that are simultaneously Sneaks for some prey and Disablers for other prey? Mosquito larva are Disablers, but that's not simulatneous with the adult Sneaking. Some vampire bats might sneak to prey on goats but disable bird prey by grappling, I'm not sure. (Wiki)

Why do Sneaks consume a specific part of their prey, while Disablers consume most of their prey's body, including many different tissues? For relatively large Sneaks, eating flesh might be too noticeable and concerning to their prey. For Disablers, there's no downside to eating flesh, and it'd be too much work to pick out certain tissues (beyond gross features e.g. avoiding bones), many going as far as eating their prey whole and then regurgitating undigestible parts. For relatively small Sneaks, maybe it's just easier to digest specific tissues like blood or the host's partially digested food.

Why don't adult ticks have multiple blood meals and multiple reproductive cycles?

Search This Blog

The Lion and the Worm