A In the early fifteenth century, a European who wanted a copy of a book had essentially two options: borrow one and copy it out by hand, or pay a professional scribe to do so. Monasteries had long maintained scriptoria for this purpose, and commercial workshops also served students and lawyers. Even so, a substantial volume might occupy a copyist for several months, and its price placed it beyond the reach of all but institutions and the wealthy. Europeans were already familiar with a cruder form of printing, in which an entire page was carved into a block of wood, inked and pressed onto paper; this was adequate for playing cards and simple devotional images, but hopelessly laborious for long texts. It was into this world, around 1400, that Johannes Gutenberg was born, in the German city of Mainz. Little is known of his youth, although his later achievements suggest that he had been trained in metalwork, possibly in connection with the goldsmiths' trade.

B Gutenberg was not, however, the first to print with movable type. In China, an artisan named Bi Sheng had produced individual characters from baked clay in the 1040s, and Korean printers were casting metal type by the thirteenth century; the oldest surviving book printed in this way, a Buddhist text known as the Jikji, dates from 1377. Why, then, did the technique not transform East Asian societies in the way it would later transform Europe? The explanation most commonly offered concerns writing systems. Chinese requires thousands of distinct characters, so a printer needed an enormous stock of type, and carving whole pages in wood often remained the more economical choice. A European printer, by contrast, could set almost any text from a few dozen letters and symbols. Historians warn, though, against concluding that East Asia was somehow backward: woodblock printing there sustained a vast and sophisticated book culture for centuries.

C What Gutenberg seems to have achieved was not a single invention but the combination of several into a workable system. Fragments of evidence about his early experiments survive only because of a lawsuit brought against him in Strasbourg in 1439 by the relatives of a deceased business partner; the testimony refers to a press, to purchases of lead and to an undisclosed 'art'. The centrepiece of the eventual system was an adjustable hand mould, which allowed a worker to cast large numbers of identical letters quickly and accurately. The type itself was made from an alloy consisting mainly of lead, with smaller amounts of tin and antimony, a mixture that melted at a relatively low temperature and hardened rapidly without shrinking. Because the water-based inks used by scribes simply ran off metal surfaces, he developed a thicker ink based on oil, closer to the varnishes then used by painters. Finally, he adapted the screw press, a machine long familiar to Rhineland farmers, who used it to squeeze grapes for wine.

D The most celebrated product of this system, a Latin Bible laid out in two columns of 42 lines, was completed in Mainz by about 1455. Roughly 180 copies appear to have been printed. Most were on paper, while the remainder used vellum, a costly material prepared from calfskin, and some 49 survive, in whole or in substantial part. The project demanded heavy investment, and Gutenberg had borrowed large sums from a wealthy citizen, Johann Fust. In 1455 Fust sued for repayment and, having won the case, took control of much of the equipment, continuing the business with Peter Schöffer, Gutenberg's former assistant. When Gutenberg died in 1468, his name did not appear on a single book he had produced.

E The technology might have remained concentrated in Mainz had it not been for politics. In 1462 the city was sacked during a struggle between two rival candidates for its archbishopric, and many printers, their workshops disrupted, carried their skills elsewhere. Within a few years presses were operating in Italy, where two German printers set up near Rome in 1465, and Venice soon emerged as the leading centre of the European book trade. Printing reached Paris in 1470, and William Caxton, an English merchant who had learned the craft on the Continent, established a press at Westminster in 1476. By 1500 printing had been carried out in more than 200 towns, and some historians have estimated that Europe's presses had by then turned out around twenty million volumes.

F The consequences were profound. Books became steadily cheaper, and works in everyday languages such as German, French and English multiplied alongside the traditional Latin texts. When Martin Luther's criticisms of the Church appeared in 1517, printed pamphlets carried them across the German-speaking lands within weeks, something no army of copyists could have managed. In an influential study published in 1979, the historian Elizabeth Eisenstein argued that print also changed the nature of knowledge itself: because identical copies could circulate widely, scholars in different cities could compare the same tables, maps and texts, and errors could be identified and corrected in later editions. Yet uniformity worked in both directions. In 1631, a London edition of the Bible omitted the word 'not' from the commandment against adultery, and the slip was faithfully reproduced in every copy of the print run. Nor did everyone welcome the change. In 1492 the German abbot Johannes Trithemius wrote a treatise praising the labour of scribes and warning that printed books would not last — a treatise that, ironically, was itself soon printed.

A Camouflage is among the most familiar of natural phenomena and, until surprisingly recently, among the least rigorously studied. For much of the twentieth century, claims about how a particular colour pattern concealed its owner rested largely on the impressions of human observers, who would look at an animal against its background and judge whether it was hard to make out. The weakness was obvious: the viewers who matter are predators, and predators do not necessarily perceive the world as we do. Since the 1990s, however, the field has been transformed by field experiments and computer models of animal vision, which together allow researchers to ask not whether an animal looks hidden, but whether it actually survives longer because of its appearance.

B The first systematic theory of camouflage came not from a zoologist but from an artist. In the 1890s the American painter Abbott Handerson Thayer pointed out that an animal lit from above should, in principle, appear lighter on its back and darker on its belly, and that this gradient of shading is one of the main cues by which the eye recognises a solid, rounded object. Many animals, from deer to fish, show precisely the reverse pattern — dark above and pale below — and Thayer proposed that this 'countershading' cancels out the effect of the light, making the body appear flat and therefore difficult to detect. Thayer's law, as it became known, was widely accepted, but Thayer damaged his own reputation by insisting that virtually every colour pattern served to conceal. He maintained, for example, that the vivid pink of flamingos allowed them to vanish against a sunset sky, a claim that the former American president Theodore Roosevelt publicly ridiculed in 1911.

C A more disciplined account was provided by the British zoologist Hugh Cott, whose book Adaptive Coloration in Animals appeared in 1940. Cott distinguished between background matching, in which an animal resembles the colour and texture of its surroundings, and what he called disruptive colouration. In the latter, bold patches of contrasting colour create false edges that break up the animal's outline, so that a predator searching for a familiar shape fails to find one. Cott's ideas had immediate practical applications: during the Second World War he served as an adviser on camouflage to the British army. Within biology, however, disruptive colouration remained largely a hypothesis, supported by striking examples rather than by data.

D That changed in 2005, when a team led by Innes Cuthill at the University of Bristol published the results of an elegant field experiment. The researchers made artificial 'moths' from triangles of waterproof paper, printed with patterns derived from photographs of oak bark, and pinned them to trees in a woodland, each with a dead mealworm attached as bait. Some targets carried markings that touched their edges; others carried identical markings positioned well inside the outline. Wild birds were then left to forage. Targets whose markings broke the edge were attacked significantly less often than those with the same patches confined to the interior, and they also outperformed targets that simply matched the colour of the bark. The result was the first rigorous evidence that disruption works independently of background matching.

E Countershading proved harder to test. In experiments carried out in the late 2000s, Hannah Rowland and colleagues made caterpillar-shaped prey from pastry, dyed green, and placed them on the upper surfaces of branches in woodland, where wild birds could find them. Prey that were darker on the side facing the sky and paler underneath survived longer than uniformly coloured prey, a finding consistent with Thayer's law. Rowland has nevertheless been careful to point out that such results do not prove that flattening the appearance of the body is the only mechanism at work; countershading may also, for instance, help an animal to match a background that is itself unevenly lit. In this sense the experiments vindicated Thayer's observation while leaving open the question of exactly why it works.

F Fixed patterns aside, the cuttlefish poses a stranger challenge. These relatives of the octopus can alter their appearance in less than a second, using millions of pigment-filled cells in their skin that are expanded or contracted by muscles under the control of the nervous system. Roger Hanlon, of the Marine Biological Laboratory in Massachusetts, has argued that, despite this enormous flexibility, cuttlefish rely on just three broad categories of body pattern — uniform, mottled and disruptive — selected according to the size and contrast of objects in their surroundings. What makes their performance especially puzzling is that cuttlefish appear to be colour-blind, possessing only one type of light-sensitive pigment in their eyes, yet they match the colours of their backgrounds with remarkable accuracy. How they do so remains unresolved.

G Not all camouflage involves being invisible. Some caterpillars resemble twigs so closely that a predator may see them perfectly well yet fail to recognise them as food, a strategy biologists call masquerade. In a 2010 study, John Skelhorn and colleagues found that domestic chicks which had previously encountered real twigs took longer to attack twig-like caterpillars than chicks with no such experience — evidence that the deception works by exploiting what the predator has learned rather than by preventing detection. Such work underlines the central lesson of modern camouflage research: concealment exists only in relation to a particular observer. Many birds, for example, can detect ultraviolet light, which is invisible to humans, so a pattern that looks perfectly cryptic to a human eye may be conspicuous to the very predator it is supposed to deceive.

A Few institutions attract such consistent hostility as the standardised test. Pupils dread it, and commentators denounce it as an instrument of privilege. Yet the idea behind it was once regarded as radical, even liberating. When the rulers of China's Sui dynasty began selecting officials by written examination around the year 605, they were attempting to weaken the hold of aristocratic families over government posts, which had traditionally been filled largely through recommendation and family connection. The system, which survived until 1905, was far from perfect: it came to reward the mastery of rigid literary formulas, most notoriously the 'eight-legged essay', rather than practical judgment. But its founding principle — that public office should go to those who can demonstrate competence rather than to those who happen to be well born — seems to me entirely sound, and one that even its fiercest modern critics would hesitate to reject.

B The modern history of testing contains a darker lesson. In 1905 the French psychologist Alfred Binet devised a series of tasks intended to identify children who needed additional help at school. Binet was explicit that his scale did not measure a fixed or inborn quantity, and he worried that it might be used to label children permanently. Within two decades his fears had been realised. In the United States, group tests inspired by his work were administered to more than 1.7 million army recruits during the First World War. Carl Brigham, who had used the army data to rank national groups by intelligence, went on to design the Scholastic Aptitude Test, first administered in 1926 — though he later publicly retracted his earlier conclusions. The lesson, I think, is not that testing is inherently sinister, but that an instrument designed for one purpose becomes dangerous when it is quietly put to another.

C Defenders of testing make a case that deserves more attention than it usually receives. Their central claim is that a common examination, marked in the same way for every candidate, is less vulnerable to bias than the alternatives — teachers' grades, personal statements, interviews and letters of reference — all of which tend to favour applicants whose families know how to present them well. Recent research lends this argument some support. In a study of admissions to highly selective American universities, the economists Raj Chetty, David Deming and John Friedman found that test scores predicted students' later academic performance considerably better than their high-school grades, which have become inflated and therefore less informative. Several universities that had made tests optional, including MIT and Dartmouth College, reinstated them, Dartmouth noting that some talented applicants from poorer backgrounds had withheld scores that would in fact have strengthened their applications. I find this evidence harder to dismiss than many critics seem willing to allow.

D The most familiar objection is that scores rise steadily with family income. Wealthier families can pay for private tutoring and practice materials, and their children typically attend better-resourced schools. This pattern is not in dispute; its interpretation is. Critics take it as proof that the tests themselves are biased. Defenders reply that a test which reveals inequalities in schooling is no more to blame for them than a thermometer is to blame for a fever: abolish the measurement, and the underlying problem merely becomes harder to see. There is force in this reply, but the analogy is imperfect. A thermometer does not alter the temperature it records, whereas a high-stakes test changes the behaviour of the very people it measures, and those with the most money are best placed to change their behaviour in ways that raise their scores.

E It is this capacity to alter behaviour that underlies the second major objection. In the 1970s the social scientist Donald Campbell observed that the more heavily a numerical indicator is used to make social decisions, the more it becomes subject to pressures that corrupt it — and the more it distorts the very activity it was meant to monitor. Daniel Koretz has documented this effect in American schools, where gains on high-stakes state tests have often failed to appear on lower-stakes national assessments, suggesting that pupils had been drilled in the test rather than taught the subject. Some researchers have also pointed to 'stereotype threat', the finding, first reported by Claude Steele and Joshua Aronson in 1995, that students reminded of a negative stereotype about their group may perform below their potential. Later attempts to reproduce the effect have often found it smaller than originally reported, however, and I would not rest much of the argument upon it.

F Where, then, does this leave the question posed in the title? Part of the difficulty, it seems to me, is that the question is wrongly framed. A standardised test does not measure some innate quantity called 'ability'; it measures what a candidate has learned and can do on a given day, which reflects talent, effort, teaching and circumstance in proportions that no test can separate. Judged as a gauge of innate potential, every test is unfair. Judged as a reasonably consistent record of current attainment, a well-designed test compares favourably with the alternatives, provided it is used with restraint. In practice, that means treating a score as one piece of evidence to be weighed alongside others rather than as a verdict, and allowing candidates more than one opportunity to sit the examination, so that a single bad morning does not decide a future. The aim should be not to abandon measurement but to stop asking it to do more than it can.