The Lagoon

by Armand Marie Leroi

  In the seventeenth and eighteenth centuries Tournefort at Paris, Pontedera at Padua, Cavolini at Naples – even Linnaeus at Uppsala – all tackled the mystery of the fig. They made some progress but left it unsolved. In 1864 Guglielmo Gasparrini, Professor of Botany at Naples, reviewing the fig from Aristotle to the results of his own extensive experiments, came to exactly the wrong conclusion. The ‘wild’ and ‘cultivated’ figs, he said, were quite different species, indeed belonged to different genera, and the need for the proximity of wild figs to fruiting trees was mere peasant superstition. Unlucky Gasparrini: he, too, studied an asexual strain and, like Theophrastus, generalized too far.*

  But that’s the scientist’s perpetual dilemma: how far to push his data. Aristotle’s generalizations tend to the sweeping; Theophrastus is more cautious. It makes him a duller read. Both occasionally question the quality of their evidence, but neither ever expressed the doubts that Gasparrini did at the end of his fig monograph, and piteous they are:

  Having now reached the term of my labors I cannot conceal a certain anxiety which has secretly grown up in my mind. I fancy I hear from all quarters that the custom of tying wild to cultivated figs being of such ancient date, and having been upheld by so many distinguished men of science, both ancient and modern, cannot but be founded on experience, against which no theories, no subtleties of science, are of any avail. Verily does the rise of such ideas in my breast so agitate me, that many times in the midst of my labors my breath has been stopped by the fear that some fact illy understood has drawn a veil over my mind.

  Unfortunate, but that’s how it goes. Much of science is about navigating between the general and the particular, between unifying the phenomena and dividing them, and sometimes you just get it wrong.

  LXXXII

  BEFORE THE RISING of the Pleiades, there is no honey to be had. Honey comes with the morning risings of the stars – when alkyōn (Eta 25 Tauri) and sirios (Sirius A) first appear in the pre-dawn sky – and when the rainbow descends to the earth. Honey can be harvested when the wild fig fruits. This is all in late May and June.

  Honeybees delight Aristotle. More pages in Historia animalium are given to them than to any other animal, bar man himself. He describes their foods, predators, diseases, the various products they collect or make, their uncanny industry and the complexity of their social lives. He says they are divine.

  You wonder how he knows so much about them. The Arab encyclopaedist al-Damîrîb al Din (d. 1405) claimed that Aristotle had a glass hive made so that he could see the bees going about their work. The bees, resenting his curiosity, smeared the inside of the glass with clay. The last detail alone makes the story unlikely, and the Arab is vague about his source. I am not sure that Aristotle saw the inside of even an ordinary hive. Beekeeping was, however, a major industry in fourth-century Greece and he certainly interrogated beekeepers at length.

  Among the bee problems that Aristotle discusses is the origin of honey. You wouldn’t think it a problem at all. Every child knows that bees make honey from nectar that they collect from flowers. As expected, Aristotle describes how bees enter calyced flowers and gather up the sweet juices with an organ that resembles a tongue.* He also says that honey from white thyme is better than red; Theophrastus speaks of white and black thyme.* It all seems very clear. But in other passages, some directly adjacent, he contradicts himself. Honey doesn’t come from flowers since, if beekeepers take honey from the hive in autumn, when there are plenty of flowers about, it is not replaced. Instead it falls from the sky. His tone is that of a man correcting popular error.

  That’s what the business of honey and the stars is all about. Honey production is tied to the astronomical calendar in some quite direct way; it seems to be a phenomenon rather like dew. That seems absurd. It isn’t – or isn’t very. His sky-honey is ‘honeydew’, the droplets of sweet fluid that suddenly appear in the spring on branches and leaves in the woods. They are, though he does not know it, the excretory product of aphids and other sap-feeding bugs.* These days, about 65 per cent of Greek honey comes from honeydew. (Greeks, who can connoisseur you to death on such matters, disagree as to whether flower or pine honey is the tastier.) As for why Aristotle disagrees with himself, someone has been meddling with the text. I suspect Theophrastus, who wrote a book, On Honey, now lost, which, as far as can be made out from second-hand reports, said that it comes from three sources: honeydew, flowers and ‘reeds’, the last of which may be Indian sugarcane.

  If the origin of honey is problematic, the origin of the bees themselves is even more so. It’s not that Aristotle doesn’t know about the development of bees, for he describes it in some detail with moderate accuracy. A bee deposits the brood in the cell and then incubates it like a bird as it develops into a larva. While the larva is still small it lies obliquely in its cell; later it sits upright, eats, excretes and clings to the comb. It changes into a pupa, gets sealed up in its cell, grows feet and wings and then breaks out and flies away. No, the problem is: which bee deposits the brood?

  As with the ‘wild’ and ‘cultivated’ fig, there are too many actors on the stage. If there were just two kinds of bees in the hive, they would be like any other animal and it would be a fairly simple matter to work the sexes out. But there are three kinds of bees – workers, drones and leaders – and no one has ever seen any of them copulate.* ‘The generation of bees is a great puzzle,’ he says, but it’s just the sort of puzzle he loves.

  In The Generation of Animals he sets out the hypotheses in a logical sequence. For the sake of economy, I tabulate them. Bees might be:

  1. spontaneously generated;

  2. the progeny of some other kind of animal;

  3. the progeny of bees.

  If 3, then they might be produced:

  3.1. without copulation;

  3.2. with copulation.

  If 3.2, then the following mating, offspring combinations might be possible:

  3.2.1. w × w → w (and so for the other kinds);

  3.2.2. q × q → w + d + q (or some other homotypic mating);

  3.2.3. w × d → w + d + q (or some other heterotypic mating).

  Where w, d and q stand for worker, drone and queen, × stands for a cross and → for the offspring.* Pages of analysis follow that draws on general zoological principles and bits of bee lore to eliminate the possibilities in succession.

  Quickly disposing of the idea that bees are spontaneously generated, or produced by some altogether different kind of animal that doesn’t live in the hive, he turns to bee genders. He begins with some sexual stereotypes or, to put it more kindly, empirical generalizations. Males have offensive weapons (horns, tusks), females don’t; females look after the young, males don’t. Neither workers nor drones, however, fit these moulds since workers have stings, but look after the brood, while drones are stingless and do nothing. So workers and drones must be neither male nor female, but a bit of both. They’re like plants or those sexually ambivalent fishes that putatively reproduce without copulation. (He knows that drones sometimes swarm out of the hive, but not that they are in pursuit of a virgin queen with the intent of having her at altitude.)

  He then investigates which bee generates which. Aristotle reports that drones can appear in hives that have neither drones nor a queen but only workers. (Remarkably, he’s right – they can.) So workers must generate drones. Workers never appear in hives without queens. So queens generate workers. Neither workers nor drones generate queens (he asserts), so queens must generate themselves. We have the order of generation: q → q + w → d. Finally, he considers whether any of them copulate. Workers are sexually ambivalent so can reproduce without copulation. Besides, if they did copulate, someone would certainly have seen it, and no one has. Since workers don’t copulate, we can assume that queens don’t either. Only one possibility remains: a multi-generational sequence of asexual reproduction in which queens generate queens and workers, and workers generate drones, and drones generate nothing.

 
; It’s a strange system, though not nearly as strange as reality.* Nor is it as strange as it is often supposed to be. Should you read Aristotle on bee reproduction, you will find that I have edited out one of its seemingly outré aspects: I have called his ‘leader’ bee, the bee that generates all the others, the ‘queen’, since that’s what it’s called today. He, however, often calls it the basileus – ‘king’.

  Some scholars have accused Aristotle of gender-bias. (She’s female, after all, so why not basileia?) But Aristotle is innocent. He cannot possibly believe that the leader bee is male. In his biology males are no more capable of reproducing by themselves than they are in ours. Driven by his data, he take the leader bee to be neither male nor female but a mix of both, and simply calls it by its popular name. Social wasps, he says, have a leader wasp commonly called ‘the mother’ – and, for them, that’s the term he adopts.

  HONEYBEE – APIS MELLIFERA LEFT TO RIGHT: MELISSA – WORKER; KĒPHĒN – DRONE; BASILEUS – ‘KING’, OR HĒGEMŌN – ‘LEADER’ (OUR QUEEN)

  Aristotle likes his scheme. There’s a three-generation sequence of bees that ends with the sterile drones. It has, he says, a kind of order in that each member of the sequence differs from the others in only one way. (Queens are big and have stings, workers are small and have stings, drones are big but don’t have stings.) ‘Nature has arranged this so well that the three kinds will exist for ever, even though they do not all generate.’ He has a life cycle for bees.

  Aristotle’s analysis of the mystery of bee reproduction is a model of how he often does science. It resembles the procedure described in the Nicomachean Ethics. He starts with the ‘appearances’, collects the best explanations for them and deductively eliminates them according to the evidence. By the time he’s done he seems to have demonstrated an essential property of bees.

  Has he? Aristotle thinks that demonstrations deliver truth. Yet, as he well knows, any demonstration is only as good as its premises and his – though he does not admit it – are weak. They rely on generalizations that he must know are at best ‘for the most part’ true. (Do males never take care of young? Then what about paternal care in the glanis, his catfish? Do females never have offensive weapons? Then what about cows?) Then, too, his data come from beekeepers, no more reliable than fishermen. His text is littered with ‘they says’.

  So a doubt niggles, and his discussion ends on a tentative note. I can’t pretend it’s typical. He’s usually so confident that no one will ever surpass his work, so final. But here, for once, he looks to the future and tells us that he hasn’t worked everything out; more than that, he tells us how future discoveries will be, or should be, made. And though we may respect Aristotle in his more Olympian moods, the moods to which great scientists are prone, here we have to love him:

  So this, at least as far as theory goes, seems to be the situation on the generation of bees – in conjunction, that is, with what people believe to be the facts about their behaviour. Not that there is, currently, any proper understanding of what those facts are. If in the future they are understood, it will be when the evidence of the senses is relied on more than theories, though theories have a part to play so long as what they indicate agrees with what is seen . . .

  – since that is exactly what happened.

  LXXXIII

  IN MARCH THE SWALLOWS arrive in Lesbos. They come from Africa on the khelidonias or swallow wind – Theophrastus uses the phrase, presumably a synonym for the ornithiai anemoi. Aristotle, who lists swallows among his migratory birds, also speaks of the intelligent way in which a pair will build a nest from mud and straw, rear their chicks and keep their house neat and clean.* He seems to admire the little birds. But, dispassionate as always, he also says that if you poke out the eye of a nestling swallow it will regenerate. He really believes this. He repeats it three times. And, though it seems like a bizarre thing to say, I’m not sure that he’s wrong. He may actually have done it.*

  Behind this claim lies a study in comparative embryology. Surveying the ontogenies of the various creatures that he’s studied, he rates their progeny on a scale of ‘perfection’ that tries to captures how much they change between emerging from their mothers and adulthood. Holometabolous insects such as butterflies have very imperfect progeny. (He thinks that the growth of a caterpillar is equivalent to the formation of an egg inside the reproductive tract of a chicken, and the chrysalis is the equivalent of an egg.) Cephalopod, crustacean and fish eggs are soft and ‘grow’ a little after being laid, so they’re also low on the scale of perfection. The progeny of birds, snakes, turtles and lizards are more perfect since they have hard-shelled eggs that don’t grow, and the young of cartilaginous fishes are more perfect yet since they start out as hard-shelled eggs in the womb but hatch out internally and are born alive. The young of live-bearing tetrapods (mammals) are the most perfect of all.

  Having established this rather crude scale of embryonic perfection among his greatest kinds, he allows variation within them too. Perfection now depends on relative size at birth and readiness for independent life: imperfect animals are born blind. Within the live-bearing tetrapods, Aristotle says that solid-hooved (horses and asses) and cloven-hooved (cows, goats, sheep) animals have perfect young; the whelps, cubs and pups of multi-toed tetrapods (bear, lion fox, dog, hare, mouse, etc.)* are, by contrast, quite imperfect. Within the birds, jays, sparrows, woodpigeons, turtledoves and pigeons also have very imperfect hatchlings.* And so do swallows. Thus, when Aristotle speaks of poking swallow chicks in the eye, his point is that regeneration is more likely in embryos than in adults, and that since swallow nestlings can regenerate, they’re really very foetal when they hatch.

  Aristotle’s science is not quantitative. It’s not, to be sure, resolutely qualitative either, for he often uses terms such as ‘large and small’, ‘the more and the less’ and ‘for the most part’. He can also discuss quantitative relationships, such as body-size scaling, with subtlety. Yet he rarely gives what modern scientists love and need: numbers. When describing the life histories of various birds and mammals, however, he does.

  You can even put them into a table. That shows that his data, although spotty, are quite good.* As always, he’s interested in the associations. The web that he weaves is wide, but five features – adult body size, longevity, gestation time, embryonic perfection, litter (or clutch) size and neonate size – are central. Some of the associations that he detects among these features are fairly obvious: longer gestation times result in more perfect (precocial) neonates. Others are quite counter-intuitive. You might expect, he says, that large animals would have larger litters (or clutches) than smaller ones, but actually that’s not so – they have smaller litters. Horses and elephants bear only one infant at a time. He obviously believes that such associations have predictive power. ‘Stories are told’, he says, ‘of the [deer’s] longevity, but none of them has been established as true: besides, the period of gestation and the swift growth of the fawns do not suggest that it is a long-lived creature.’ Given a positive association between longevity and gestation time, if deer truly were very long lived (something that’s hard to observe) then they should also have a very long gestation period, but they don’t.*

  Aristotle doesn’t want just any associations: he wants causal associations. Any association may be, in his terms, ‘accidental’ rather than ‘essential’ and so not require explanation at all. He notices that the negative association between adult body size and litter size is confounded with foot morphology. Solid-hoofed animals tend to be large and have one offspring at a time; cloven-hoofed animals tend to be medium sized and have a few; multi-toed animals tend to be small and have many. Perhaps, then, the litter–body-size association is really all about feet. But no: it’s body size that matters. ‘The evidence is that the elephant is the largest animal but is multi-toed; the camel, the largest of the rest, is cloven-hoofed’ – the foot-type–body-size association is, in fact, poor. Moreover, ‘It’s not only on land that large animals
produce few offspring and small ones many, but also in animals that fly and swim, and the reason is the same. Similarly the biggest plants do not bear the most fruit.’ Thus, not only is he aware of the possibility of confounding variables, but he also has a solution: to search for the same association in quite different groups of creatures.* In the same way, he asserts that the positive association between gestation time and longevity in viviparous tetrapods (so informative about deer) is not causal.* Here, at least, he doesn’t jump from association to demonstration; here he considers causation v. correlation.

  To explain his web of life-history features Aristotle wields all his familiar devices. Considering the association between litter size and adult body size, he reaches for his bodily economics. It’s particularly effective since fecundity depends on seed production and seed is the most refined, and hence expensive, nutritional product of all. Expending seed drains the body. That’s why (he says) men are so exhausted after sex, fat people are infertile, castrated animals and mules are so big and fierce, big animals tend to have few offspring, and highly fecund animals tend to be small. (The Adrianic fowl* is said to be a super-fecund dwarf.) Sex may be fun, and reproduction may be necessary, but growth and vitality drain away through our genitals.

  So animals must choose between bearing offspring and doing something else. Aristotle the ornithologist knows that, each year, some birds (partridges) lay a single clutch with many eggs, some (pigeons) lay many clutches with a few eggs, while others (raptors) lay only a single clutch with one or two. To explain these differences he postulates a resource-allocation network that links wings, legs, body size and fecundity and a few other features besides. A bird may invest in some, but not all, of these features for each has a cost in terms of the others.