Development Is Not Necessarily the History of the Individual

At the beginning of our exploration of developmental phenomena, it seems reasonable to address a semantic question: what do we mean by development? Let us focus on the development of living organisms, without worrying about what development may mean, for example, to an economist or an educator.

What can be considered as development is a controversial issue. A few years ago, a group of biologists and philosophers of biology thought it necessary to consider this question seriously. Overall, the debate involved 24 scholars. Two important things emerged from their responses. First, only half of those concerned said that a definition of development was necessary; the others argued that they could safely do without, and one even added that a definition of development is impossible. Second, the proposed definitions were very different from one another, to the point that several important biological phenomena would fall within the sphere of developmental biology for some scholars but not for others.

A look at the list of proposed definitions is useful. It will serve as a guide for our itinerary, not so much to seek answers to our questions as to widen horizons as much as possible and to try to formulate sensible questions. Here are the definitions as proposed. Development is…

• the process by which a single cell gives rise to a complex multicellular organism;

• the generation of a new individual form;

• the change of biological form over time;

• the temporal change of organization along the life cycle;

• the biological reading of DNA-encoded gene networks that determine the structure of the organism;

• the irreversible increase in the complexity of a biological system over time.

In different form, all of these definitions capture important aspects of development and many of the problems on which most research focuses. However, it is sensible that from the very first pages of this book, the reader should assume a critical attitude towards a number of positions that it is all too easy to take for granted, e.g. that

• development is a series of structural changes affecting multicellular organisms; single cells per se, including unicellular organisms, do not develop;

• development is about the biological individual – indeed, it is the process that shapes the individual;

• development necessarily entails an increase in complexity;

• the body contains (multiple copies of) a programme according to which it is formed;

• development is irreversible.

These widespread beliefs are based, at best, on a generalization of conditions that apply only to some living organisms, not to all. They overemphasize aspects that cover only a part of the phenomena that deserve to be called developmental processes.

An excellent starting point to clear the path through these problems is a few sentences by the great French physiologist Claude Bernard:

This text is from the Leçons sur les phénomènes de la vie communs aux animaux et aux végétaux (Lectures on the Phenomena of Life Common to Animals and Plants) published in 1878, immediately after Bernard’s death.

Thus, in Bernard’s vision, development is not the history of the individual from zygote to adult, but a series of changes in which each step depends strongly on the conditions in which the living organism was until then. It is like a chess game, where different choices (some more advantageous for the player, some less so, but this is not important) are generally possible with each move; these choices depend on the current arrangement of the pieces on the board and this, in turn, depends on the previous moves. The comparison between the succession of changes in development and the configurations of the pieces on a chessboard, however, is only valid as long as the game is in full swing. When the game is over, the board is emptied: there is no continuity between one game and the next, whereas there is between one biological generation and the next. It is precisely here that it is advisable to take a closer look, to make clear what should be considered as development.

In fact, we are faced with two possibilities. If we listen to Bernard, development is a process (better, perhaps, a set of processes) that continues through the generations. If instead we follow the popular notion (shared by many professionals), development is the individual story of the changes that transform an egg first into an embryo, then into a juvenile and finally into an adult.

However, there is a way to overcome this dichotomy. If we do not want to leave a good number of important topics outside the scope of developmental biology, we should accept a notion of development consistent with Bernard’s observation. The chapters of developmental biology, therefore, will not be (only) those that correspond to moments in the history of an individual, such as cleavage (the division of the egg into a progressively increasing number of cells, the blastomeres) or gastrulation (the embryonic developmental phase shared by almost all animals, during which germ layers are formed; more on germ layers on p. 78). Instead, the mechanisms of regeneration will fall within the scope of this discipline, even if most of what we know about them are the responses to challenges an animal would never face in nature; or the structural changes resulting from a pathological situation, primarily cancer, or those induced by the presence of a parasite.

In this broader perspective, the history of the individual does not become less interesting or less central in developmental biology, except that the succession of stages from egg to adult animal (or from seed to mature tree) is no longer the development, but a particular history of development. In the pages of this book, there will be space both for individual stories (often very different from what we might describe as the normal development of an individual of our own species) and for developmental processes as such, over a time span that may be shorter or longer than one generation.

Development Is Not Necessarily the Sequence of Changes from Egg to Adult

When we describe an animal or a plant as a ‘monster’, this is because it departs significantly – although sometimes in one feature only – from the morphology of a typical individual of the species. A calf should have one head rather than two; a fruit fly should have two wings rather than four, as found in a well-known mutant. However, it is not always easy to say what the typical structure of an animal should be.

First, an animal may undergo metamorphosis over the course of its life. A newt, for example, spends many weeks as a tadpole, and the differences between tadpole and post-metamorphic newt are important: the tadpole is an aquatic animal that breathes through gills, while the adult can move on land and breathes through lungs (and skin).

A tadpole and a newt are the same animal; nevertheless, when we identify in the adult morphology the form typical for the species, we give the adult form an absolute value. The egg and the embryonic stages, to continue with larva and juveniles, are thus downgraded to mere preparatory stages. The ‘true form’ of the newt is the form of the adult. This is acceptable in the everyday use of the term ‘newt’, but not if we want to understand developmental biology.

As in other situations (I will give more examples in this book), useful suggestions come from the study of individuals that have undergone a less than normal development. Some newts, for example, become sexually mature without having undergone metamorphosis: they retain their larval shape and continue to increase in size, while their gonads mature as in a normal adult. If we follow the standard terminology, these newts, although able to reproduce, are not ‘real’ adults. The ‘true form’ of the animal is another.

Second, there are organisms in which it is impossible to recognize a standard form. The tiny fungus Candida albicans, for example, can easily switch between a single-cell form, comparable to a yeast, and a filamentous, multicellular one (see p. 27).

When discussing development, it is critically important, even if difficult, to move away from the traditional attitude that deserves the name of adultocentrism, according to which all the embryonic, larval and juvenile stages – and the developmental processes in which they are involved – are only steps or means required to become an adult. This old attitude has not changed much with the modern concept of development as the deployment of a genetic programme, because the latter is intended as a programme for the production of an adult.

In the traditional adultocentric view of development:

• The adult condition is the goal to be reached. However, we will see that this is not always true; moreover, the very notion of adult is sometimes problematic.

• Once the adult condition is reached, development is stopped. If life extends beyond the reproductive stage, the adult faces ageing – a phenomenon that traditionally belongs to the discipline of pathology rather than to the biology of development. But we will see that changes in the organism in post-reproductive age occur according to processes of the same nature as those that characterize previous stages.

• Developmental mechanisms have been consolidated through natural selection, therefore they are adaptive. But we will see that from the point of view of the cellular or molecular mechanisms involved, ‘normal’ developmental processes and phenomena such as the production of tumours are not necessarily very different.

• The sequence of events that characterize an individual’s development is irreversible. On this topic too we will have something to say.

It is difficult to deny that the adultocentric vision contains a good deal of finalism. From this point of view, a comparison between developmental biology and evolutionary biology can be interesting. In the latter, finalism survives only in rather superficial popular versions of the theory, in which evolution is considered synonymous with progress, rather than a continuous and always imperfect adjustment to the changing conditions faced by a population. In developmental biology, however, sentences with a finalistic flavour often come from the pen of authoritative scientists. For example, Eric Davidson, a scholar to whom we owe major achievements in the molecular genetics of development, wrote that “development is the execution of the genetic programme for the construction of a given species of organism,” and that “a particular function of embryonic cells is to interact in specific ways, in order to generate morphological structure.”

Also adultocentric is the term ‘set-aside cells’. This designates groups of cells found in the larvae of many invertebrates, which are not parts of the larva’s organs but remain dormant until metamorphosis. Only at this point, while the larval structures are reabsorbed or lost, does the adult take shape precisely from those cells that, until then, had been ‘set-aside’, almost with the intention – one might say – of using them later in adult morphogenesis. It would be preferable to say that those cells, rather than set-aside, were temporarily marginalized from active life.

An adultocentric view of development requires that each phase be compatible with the next. In my view, the opposite perspective is much more reasonable: that is, development can proceed so far as each phase is compatible with the previous one. In this perspective, there is no difficulty in including in developmental biology the individual stories that stop before the adult condition is reached. I am referring not just to the philosophically uninteresting case of a developmental history truncated by accident, but to stories in which, through intrinsic causes, development is arrested in a condition other than the normal: the so-called ‘monsters’.

Disregarding those created in the lab (often invaluable for the progress of developmental biology), monsters sometimes show up in nature, even in our species. Their study is the subject of a specific scientific discipline, teratology. To approach this field, I suggest we turn the pages of the first treatise on comparative teratology (three volumes of text plus one of plates), published in the years 1832–37 by Isidore Geoffroy Saint-Hilaire. In this work, monsters are arranged according to a classification similar to Linnaeus’ distribution of animal and plant species. This exercise is very important: if monsters can be classified, this means that their deviations from the normal condition are not arbitrary, but fall within a finite, perhaps small, number of kinds. And even monsters usually obey the laws of biological form, including two-headed calves or fruit flies with the antennae replaced by two legs, at least in so far as they do not depart from bilateral symmetry.

There Is Not Always a Species-Specific Limit to Individual Growth

In 1864, a year before succeeding his father William as the director of Kew Gardens – one of the most prestigious botanical institutions in the world – Joseph Dalton Hooker, one of Charles Darwin’s closest friends, described under the name of Welwitschia mirabilis a truly unusual plant discovered 5 years previously by the Austrian botanist Friedrich Welwitsch. The homeland of this unique plant is the desert that extends along the border between what are today Namibia and Angola. Its massive woody trunk, which has no branches, resembles a low stump a few tens of centimetres high. From its upper margin sprout two broad ribbon-shaped leaves, each of which can be up to 4 metres long. The tip, which is the oldest part of the leaf, is dry and frayed, especially in older plants. But the two leaves continue to grow, thanks to the proliferative activity of basal cells, throughout the life of the plant. Specimens a thousand years old are not uncommon, and some are believed to be twice as old.

Welwitschia mirabilis is the only living species of a lineage of gymnosperms – a plant with some affinity with conifers, but not very close to them. In the other major group of seed plants, the angiosperms (flowering plants), there are also a few species with continuously growing leaves: in this case, however, growth takes place from the distal tip, and the whole leaf will wither within a few years. These plants are tropical trees of the mahogany family (Meliaceae), more precisely those classified in the genera Guarea and Chisocheton.

Indeterminate growth, however, seems to be a widespread feature in trees even if, sooner or later, the process will necessarily come to an end. We will discuss in Chapter 8 whether ageing affects all living beings, or whether some organisms do not experience it. But we do not need to invoke ageing here: even the most robust tree ends up succumbing to attacks by fungi or insects, helped perhaps by severe atmospheric events.

We might think that things are different in animals. In humans, growth in height eventually slows down, then ceases altogether. Other familiar vertebrates follow the same trend. But it would not be safe to generalize. Even among mammals there are species in which growth continues throughout life, even if this slows with the onset of maturity. Examples are bison, giraffes and elephants. There are many more examples of animals with indeterminate growth among the amphibians and, more conspicuously, among fishes such as the grouper. There is also no shortage of examples among invertebrates, for example the giant clams of the genus Tridacna, which can live over a century, reaching enormous size. The largest known shell of a giant clam weighs 330 kilograms; the mollusc that produced it weighed perhaps another 20 kilograms.

In other cases, the arrival of reproductive maturity puts a neat end to growth, even if this was previously very rapid. Among the plants, bamboos reach the most extraordinary growth rates, up to 90 centimetres in a single day, but the plants die after their only flowering season. Animal embryos often elongate particularly fast, especially those supplied with a large amount of yolk. The increase in size of a tumour is also often very fast. In the context of normal post-embryonic development, extraordinarily rapid growth is exhibited by many tapeworms. Within 2 weeks after infection, Hymenolepis diminuta, a tapeworm 20 to 60 centimetres in length that lives in the rat intestine, increases 3400 times in length and 1.8 million times in weight, producing the fantastic figure of 2200 proglottids (the technical term for the ‘segments’ of tapeworms).

But there are also animals that go through periods of negative growth. This is not simply a matter of weight loss due to lack of nourishment for a prolonged period, but of a somewhat ‘regulated’, although regressive, developmental process that allows the animal to resume positive growth when environmental conditions or availability of food are back to normal. Cases of negative growth have been observed in many invertebrate groups, but we will take a look at just three examples.

Under fasting conditions, 1-centimetre-long planarians (a group of free-living flatworm, the most popular of which live in freshwater) can be reduced to a tiny worm less than a millimetre long, but their complex anatomical structure remains substantially preserved, through a proportional reduction of the various organs.

Even more intense is the effect of negative growth in some nemertines, a group of worm-like animals, almost all marine, also known as the ribbon worms. Some nemertines, for example some species of the genus Lineus, can endure fasting for more than a year, reducing their size from a few tens of centimetres to a microscopic mass of a few hundred cells: in the process, the gonads, digestive tract and other organs are resorbed. The outcome of this negative growth can be described as a return to an embryonic level of morphological complexity.

The final example is from insects. In conditions of prolonged fasting, the larva of the small beetle Trogoderma glabrum progressively decreases in weight and also in length, but it cannot be said that its development is suspended. On the contrary, it goes through a higher number of moults than normal. Whereas under normal feeding conditions the larva of this insect goes through five or six stages, a fasting larva continues to moult an indeterminate number of times, even for years: after each moult, its size is a little smaller.

Developmental Change Is Not Necessarily Different from Metabolism

In a part of the tree of life where the divide between unicellular and multicellular conditions is particularly thin, we find a species of tiny marine organisms known as Salpingoeca rosetta. This is the best studied species among the choanoflagellates, a small group of microscopic eukaryotes with some characteristics that suggest affinity with the evolutionary lineage of the animals. The trait that gives them their name is the presence of a whip-like structure called a flagellum, surrounded by a showy collar. What is remarkable, of course, is not the flagellum: flagellated cells are present in many groups of organisms, such as the spermatozoa of most animals, including humans. It is the collar that is not as common. This is a sort of circular palisade formed by microvilli, those slender rod-shaped extensions found, for example, on the side of the cells of our digestive tract that faces the lumen of the intestine, greatly increasing their surface area and therefore the efficiency of the absorption of digested substances. The cells of the gut mucosa, however, possess microvilli but not flagella. Cells similar to those of choanoflagellates are characteristic of sponges, where they are called choanocytes.

In both cases (choanoflagellates and choanocytes), the continuous movement of the flagellum and the presence of the surrounding palisade of microvilli allow for efficient circulation of water. This forwards tiny nutrient particles to the cells, which then engulf them. What makes the similarity between choanoflagellates and sponge choanocytes more interesting is that some choanoflagellates, including Salpingoeca, often live in groups: that is, they exhibit a rudimentary form of multicellularity. In recent years, DNA comparisons have confirmed that choanoflagellates are indeed among the closest relatives of animals. This has motivated a search in choanoflagellates for clues about the transition from the uni- to multicellular condition: for example, the presence in choanoflagellates of those molecules that allow cells resulting from a mitosis (the normal cell division that produces two daughter cells, both of them with a complete copy of the parent cell’s genome) to remain together. A number of these molecules have actually been found in choanoflagellates. In the latter, however, the transition between the single-cell and the multicell organization is not an obligate developmental step, as in animals, but depends on environmental conditions. What triggers this transition? In the case of Salpingoeca rosetta, the stimulus is the presence of certain species of bacteria, in particular Algoriphagus machipongonensis. But the bacteria do not only act as the trigger of this important morphogenetic event: Algoriphagus is also a prey species for the choanoflagellates.

We are thus faced with a situation where one external agent serves both as food and as the trigger of a structural change – that is, it is involved both in metabolism and development. However, this is not an isolated case. Perhaps this distinction was not yet clear before the chain of transformations that we call development took on an autonomous and precise form in the evolutionary lineage of animals. Problematic situations are found almost everywhere.

Let us take snakes as an example. For a large python, coping with food needs is no small problem. Like almost all snakes, pythons are predators, and their appetite can be satisfied only by rather large prey. A python has remarkable strength, and the tightness of its turns allows it to kill a variety of prey, including monkeys and pigs. However, by the time the snake opens its jaws and starts swallowing the prey, its problems have only just begun. The victim is not chewed, cut into small pieces and suitably sprayed with saliva before being pushed down through the oesophagus; instead, it ends up in the stomach still almost untouched. Digestion, which is long and difficult, requires support from all the organs of the snake. In 2 days from the time the python swallowed the prey, the length of the intestinal villi increases fivefold and the muscle mass of the heart ventricle by 40%. This latter increase is due not to cell proliferation but to growth in the amount of contractile proteins in the fibres that make up the heart mass. But this is only the tip of the iceberg. In both pythons and rattlesnakes, in conditions of prolonged fasting, the intestine undergoes morphological and functional regression; after a meal, the gut epithelium resumes its organization and functionality, not only through an increase in cell volume, but also thanks to reactivated cell proliferation. Therefore, the entire cycle of structural and physiological changes that accompany a period of fasting and the subsequent feeding phase translates into alternation between a phase of reduction and dedifferentiation and the following regenerative phase.

The next story confirms that the usual divide between development and metabolism as distinct chapters of biology is subjective. The green sea slug Elysia chlorotica, 2 to 3 centimetres long, is common in salt marshes and coastal pools along the eastern coasts of the United States and Canada. The species owes its specific name (chlorotica) to its lively green colour. This is due to the filamentous alga Vaucheria litorea which it feeds on without digesting it completely. After piercing the cell wall of the alga with the teeth of the scraper (technically, the radula) in its mouth, a sea slug sucks up the contents. In the mollusc’s digestive tract, which has extensive ramifications throughout the animal’s body, the chloroplasts of the alga remain intact for months, engulfed within the cells of the intestinal wall, and continue to perform photosynthesis. The sugars thus produced contribute to meeting Elysia’s food needs: we are therefore in the sphere of metabolism. But this story also affects development closely. The mollusc does not only change from a heterotrophic to an optional autotrophic condition; when it begins feeding on Vaucheria, its larva also finds it easier to complete metamorphosis.

Developmental Biology Is Not the Same as Embryology

In 1673, the Italian physician, physiologist and anatomist Marcello Malpighi published a small work entitled De formatione pulli in ovo (On the Chick’s Formation in the Egg), a true milestone in the history of embryology. The theme was not new: since Aristotle’s time, the chicken had been the most accessible animal in which to study embryonic development. Even though chick embryos develop inside an opaque shell, their study has many advantages, such as availability in large numbers and, more important, the possibility of creating a rigorous time series, by incubating eggs in uniform environmental conditions and measuring the time elapsed between an egg’s deposition and the moment in which its shell is broken by a scientist to observe the embryo.

In times closer to Malpighi’s, other scholars such as the Italian Girolamo Fabrizi d’Acquapendente and the Englishman William Harvey had also dealt with the chick embryo. However, these authors had made their observations with the naked eye: Malpighi was the first to use a microscope. Without this tool, he would not have been able to make the observations he describes and illustrates in the 25 figures of the booklet. The chick remained the main study object of some of the great researchers of the following century, such as Albrecht von Haller, Lazzaro Spallanzani and Caspar Friedrich Wolff, and of the first half of the following century, such as Christian Pander. Karl Ernst von Baer tackled a more difficult topic, the study of the embryonic development of mammals, culminating in 1827 with his discovery of the elusive egg (or ovum) of these animals. I will look later (p. 70) into von Baer’s great work of comparative embryology (two volumes, published respectively in 1828 and 1837), which represents the culmination of these studies.

In the following decades, studies extended to the embryonic development of many different animals, including a number of marine invertebrates, whose transparent eggs and embryos allow much easier observation.

Towards the end of the nineteenth century, purely descriptive study started to be complemented by experimental embryology, which we will deal with later (p. 60). The effects of manipulations (at first exclusively mechanical, later by means of chemical treatments as well), carried out at different developmental times, allowed the identification of critical developmental steps and turning points. Further advances would have to await developments in molecular biology. Scientists would eventually learn to interfere with the expression of those genes for which a morphogenetic effect is known or suspected.

At each transition from descriptive embryology to experimental embryology to developmental genetics, the understanding of developmental processes has risen to levels unimaginable in the previous stage. This progress, however, has been achieved at a price we are only just starting to realize. For obvious practical reasons, the number and diversity of species studied has tended to shrink. This opens an entire chapter of the life sciences – the biology of model organisms – that will be the subject of the next section.

Up to this point, we have focused mostly on animals. Until the eighteenth century, there were no important contributions to the knowledge of plant developmental biology. Even the existence of sexuality in plants was not accepted before the last years of the seventeenth century. In the middle of the following century, the eminent figure of the German scholar Caspar Friedrich Wolff emerged, author of a theory of generation based on a comparative study of embryonic development in both plant and animal species.

In plant science, the use of the term embryo was occasional and lacked a clear circumscription until 1788, when the German botanist Joseph Gaertner successfully used it in the first volume of his large treatise De fructibus et seminibus plantarum (On Plant Fruits and Seeds). Gaertner defined the embryo as “the most noble and essential part of the seed, the only part that provides the new plant and to which all the other parts of the seed are added for at least temporary use.” Until then, there had been no name for the future seedling, in the phase in which it is still enclosed within the seed casings (integuments). It is understandable that Gaertner, faced with the uncertain nomenclature of the few accounts on the seeds of plants published before then, none of which were accurate and comprehensive, turned to animal embryology, which was undoubtedly more advanced. From this literature Gaertner borrowed many terms. ‘Embryo’ had long been in use to indicate an early stage of development, both of viviparous animals, including our species, and of oviparous ones, such as the chick. Of the many other terms of zoological origin used by Gaertner, some (such as placenta and cotyledon) come from the embryology of viviparous animals, others (such as egg white and yolk) from the embryology of oviparous ones. More than two centuries later, some of these terms have remained in use for both plants and animals, but nobody today would venture to say, for example, that the placenta or cotyledons of plants are the same thing (‘homologous structures’, to use the technical term) as the animal parts known under the same term. Unfortunately, the idea of an equivalence between what is called an embryo in either kingdom is widespread even among professionals. To realize this, it is often necessary to lift the veil of apparent modernity provided by questions formulated in molecular terms. For example, some scientific papers address the question of whether there are similarities between the trend of gene expression along the different embryonic stages of animals and plants, but fail to explain how plant and animal developmental phases can actually be compared.

The Success of Modern Developmental Biology Has Not All Been Based on Model Species

What are the main model organisms used in laboratories all over the world for experimental research – the species on which is based much of what we know about development biology, and biology in general?

Among the first entries in the list, we find a few familiar animals, such as mice (Mus musculus) and the fruit fly (Drosophila melanogaster). We may not pay much attention to fruit flies as they wander among glasses of new wine or overripe fruits that are beginning to rot. However, the role they have played in genetics is well known, and this insect has been increasingly popular since 1908, when the American geneticist Thomas Hunt Morgan and his collaborators began experimenting in what has remained famous as the Fly Room at Columbia University, New York. Other model animals are less popular, but in elementary biology courses it is likely that one will at least encounter photos of zebrafish (Danio rerio) and of the tiny nematode worm whose scientific name is Caenorhabditis elegans. Among plants, the most fashionable model species is the humble thale cress (Arabidopsis thaliana), but also important are tomato (Solanum lycopersicum), snapdragon (Antirrhinum majus) and rice (Oryza sativa), plus the moss Aphanorrhegma patens (usually known as Physcomitrella patens). Let’s add to this list two representatives of the fungi (baker’s yeast Saccharomyces cerevisiae and the red bread mould Neurospora crassa), and the cellular slime mould Dictyostelium discoideum, an odd kind of organism whose life cycle will be the subject of a later section (p. 36). Discoveries based on most of these species will be mentioned repeatedly in later chapters. The problem with model species is that the results of observations and experiments limited to them cannot be extrapolated to other species so broadly and uncritically as we might hope.

The small nematode worm Caenorhabditis elegans, about a millimetre long, has been reared for research since the middle of the last century, but it was only in 1963 that an important research effort based on this species started, involving a large number of developmental biology labs. Several reasons motivated this choice. In part, these are the usual reasons that make an animal a good candidate for a model species: small size, short life cycle, easy and cheap breeding in the lab. To this, we can add the remarkable transparency of the body of this worm; and, finally, a trait that made C. elegans particularly promising for development biologists: in small nematodes and some other invertebrates of similar size, the number of cells is the same in all individuals, both the whole body count and the number of cells in each organ. This characteristic is known as eutely. To be more precise, numbers in the case of C. elegans are given for nuclei rather than cells, because some sets of cells are fused to form a multinucleate unit with common cytoplasm – a syncytium (more on this subject on p. 36). In adult hermaphrodite individuals there are 959 nuclei; of the total, 302 belong to neurons, 95 to muscle cells in the body wall and so on. Another 131 cells were formed in the embryo, but underwent programmed cell death (apoptosis; p. 24).

All these cells stem from the fertilized egg, through a series of divisions whose course is rigidly fixed within the species. The highly conserved sequence of mitoses was expected to be associated with a progressive restriction of the final fate of the cell, eventually allocating it to specific parts of the body. The surprise came when it was discovered that the cell’s final fate is not fully dictated by its position in the cell lineage tree but also by its interactions with other cells (p. 63).

Drosophila melanogaster is far from being representative of insects in general, precisely because of a quality that is otherwise appreciated in a model species: the short duration of its life cycle. In the lab, at 25 °C, the whole embryonic development of a fruit fly takes just one day. But this speed is the effect of dramatic deviations from the usual insect developmental pathways. In Drosophila, the first 13 cycles of mitosis occur in syncytial rather than cellular conditions; a bit later, all body segments will be formed synchronously rather than by adding new segments at the posterior end. Correspondingly, the expression of genes with major effects on establishing the general structure of the embryo is quite different from the spatial and temporal patterns recorded in most other insects. Many ground-breaking discoveries in developmental genetics were made while experimenting with Drosophila: a lot of adjustments became necessary in order to apply these results to other insects and non-insect animals.

Problems are no less important in the case of Arabidopsis thaliana. It would be unfair to describe this as ‘typical’, not only of the whole world of flowering plants – about 300 000 species, including herbs, grasses, bushes and trees – but even of the 3700 species in the cabbage family (Brassicaceae) with which it is classified.

The genome of A. thaliana is much smaller than most other plant genomes sequenced to date, probably resulting from a drastic reduction suffered by this species after it split from its closest relatives. It may be useful to make a comparison with another species of the same genus, Arabidopsis lyrata. On the geological scale, the divergence from their last common ancestor is recent, at about 10 million years ago. During this time, the differences between the two species have become very considerable. The genome of A. lyrata is larger than that of A. thaliana: 32 700 genes are estimated in A. lyrata as against 27 000 in A. thaliana, but a quarter of the latter are not present in A. lyrata, half of the genome of which seems to have no equivalent in A. thaliana. We are a long way from the 1.2% gene differences estimated to separate humans from chimpanzees. The uniqueness of A. thaliana is also found in morphology, starting with the haploid number of chromosomes (see box on p. 12), which is just five, while it is eight in other Arabidopsis species and in Brassicaceae in general.

More seriously, given the role that A. thaliana has had and still has in developmental biology, it is poorly representative of plants in some aspects of genetic control of developmental processes. Genes involved in a combinatorial way in controlling the specification of the parts of the flower (sepals, petals, stamens and carpels) were first identified in A. thaliana. In its original formulation, the so-called ABC model of genetic specification of flower part identity envisaged that the flower organs where a gene function A is expressed develop into sepals, those where A and B expressions overlap develop instead into petals, those with B and C into stamens, those with C only into carpels. Important differences in the genes actually involved were soon found in the snapdragon; and in fact, the pattern of gene expression found in the latter applies to a much larger set of plant species than the pattern found in the first plant model.