We do not know for certain when the compass saw the light of day. The phenomenon of magnetism was well known in ancient times, in China as well as the West, but the first reports of a magnetic needle which could swing freely are little more than 900 years old. As often, it seems, the Chinese were the first to document the compass around 1090 AD. European reports come from a century later, but we don't know whether the discovery was independent or whether it came from the East. Moreover, it is interesting to note that, while we automatically think of the compass needle as pointing toward north, the Chinese thought of it pointing south.

The compass became known in Europe during the true renaissance of natural history. During the 12^th century a flood of translations from ancient Greek and Arabic texts into Latin appeared supplemented by later Arabian works. The thoughts of Archimedes, Aristotle, Euclid etc., inspired the learned circles and transformed that and later centuries to an exciting period. One of the most remarkable scientific papers from this epoch is in fact concerned with magnetism; it takes the form of a long letter (therefore often referred to as Epistola de magnete - letter on magnetism) written in 1269 by the army engineer Petrus Peregrinus to a colleague during maneuvers in Italy. Herein theory and experiment concerning magnetism, including concepts such as magnetic poles, attraction and repulsion, were described. This work was studied for centuries subsequently. It is interesting to note, furthermore, that Petrus was of the opinion that the compass needle pointed to the Pole Star, while the general opinion of that time held that a mountain of magnetite must be located at the north pole.

The compass presumably came into common use as a navigation instrument in the 14^th century and it was commonly held that it indeed pointed true north. However, in the following century it became clear that the situation was not quite so perfect and there was a slight deviation from true north. In Europe, the compass pointed a few degrees to the east, the first certain indication coming from portable solar clocks from the 15^th century. Such clocks were built as combinations of sundials and compasses because they had to be oriented correctly when set up. The construction of such clocks in Germany in the middle of the century indicate that such magnetic deviation was a known phenomenon. Indeed we see the deviation noted on German maritime charts from the same period. It is a common belief that Christopher Columbus discovered magnetic deviation during his first trip to America in 1492, but the phenomenon was in fact known in central Europe earlier among mariners.

The compass needle therefore points a little to the side of the geographic north pole and this varies depending on where one is on the Earth. Mapping of the deviation became important for mariners and in the 1530s the Portuguese developed a method for this based on observations of the sun. The pioneer of this technique was the marine officer João de Castro (1500-1548) and a series of high-quality measurements of deviation may be found in his remarkably accurate and informative logbooks for the years 1538 to 1541. The method was subsequently employed during a multitude of travels on all oceans, and, thanks to this mapping, supplemented by an number of land-based measurements, we are able to form a picture of the Earth's magnetic field as long ago as the 1500s, and indirectly ascertain the location of the magnetic poles as they then were.

Towards the end of the 16^th century it became apparent that the compass needle possessed another characteristic: if it were allowed to rotate about a horizontal rather than vertical axis, it would point obliquely into the ground - something a similar but unmagnetitised needle would not do. Sufficient information lay on the table: deviation, declination and the knowledge of corresponding forces which act between magnets. In a book published in London in 1600, the conclusion was dawn that the Earth is itself a huge magnet! The book is known under the name De magnete (the actual title being much longer) - About the magnet - and was written by the Englishman William Gilbert (1540-1603), a renowned doctor in Elizabeth I's England, and appointed as her personal physician until the plague took his life two years later.

In De magnete, Gilbert follows the thread from Petrus Perengrinius Epistola de magnete going through magnets and their properties, magnetic poles, describing deviation and declination and arriving at the result that the Earth has the same properties as a sphere of magnetite. The book aroused considerable attention, and was particularly praised by Galilei and Kepler, although the Church was less impressed because parts of the theory supported the heliocentric theory of astronomy. Today, De magnete is recognized as the first modern work of natural history and represents the birth of geomagnetism as a systematic discipline. In De magnete, the Earth was given a new property - previously only her form and size had been known: now she had two new poles, the magnetic ones. These poles had to be located somewhere in the arctic and Antarctic regions but exactly where, no-one could say.

Gilbert's sketch showing how magnetic needles orient themselves on the surface of a magnetic sphere. At the magnetic equator, the needles lie on the surface, but away from the equator they become steadliy more tilted until, at the poles, they stand perpendicular.	A modern version of Gilbert's model but with the source of magnetism well inside the earth. The magnetic needles are replaced by magnetic field lines extending from pole to pole.

Gilbert had laid the cornerstone and the building work could now begin; slowly but surely, the Earth's magnetic field became mapped out. The first map of deviation was published in 1701 following an exhaustive voyage by the English astronomer Sir Edmond Halley (1656-1742). It was no coincidence that this map covered the Atlantic ocean, good navigation being of great importance for England at that time, for both economic and military reasons. Furthermore, in the course of the 17^th century, it became apparent that the magnetic field was not constant and that the deviation changed slightly from year to year, and therefore it became important, not only to map the magnetic field, but also to keep the maps up to date. Knowledge of the polar regions also grew considerably around this time and consequently measurements of the magnetic field nearer the poles also became available. These measurements gave seed to ideas on the location of the magnetic north pole, Halley being of the opinion that it lay somewhere north of Spitzbergen.

In 1811, Den Kongelige Danske Videnskapers Selskab (The Royal Danish Society of Scientists) advertised an award for answering the question "Can the Earth's magnetic field be described by only one magnetic axis, or are several necessary?" One of those who delivered an answering treatise was the young Norwegian Christopher Hansteen (b. 1784) who's answer drew considerable attention and contributed to his professorship of the University of Christiania (later Oslo) in 1816. Hansteen became well known in international as well as Norwegian research circles and made many contributions to the Norwegian community right up to his death in 1873. The treatise answering the Danish challenge was published in 1819 in an expanded form under the title Untersuchungen über den Magnetismus der Erde (Investigations of the Earth's magnetism). Assembled here are almost all observations of the magnetic field up to that time along with maps and the attempt to build a mathematical model in which the observations might be reproduced by a system of magnets within the Earth. The conclusion was that one magnet was insufficient to describe the Earth's field, two being required, giving four magnetic poles in all, the two new poles being located north of Siberia and in the south-easterly Pacific. The original strong poles were located in the American far north and in eastern Antarctica. The magnetic conditions in Siberia interested Hansteen greatly and he undertook an long trip there in the years 1826-8; regrettably, his measurements from this trip were never analysed.

An important aspect of Hansteens book was the attempt to make a mathematical model of the magnetic field. Given such a model, one could, in principle, calculate the deviation, declination and strength of the field for any point on the Earth's surface. Hansteen's description, however, did not survive long: the great German mathematician, Carl Friedrich Gauss (1777-1855), attacked the problem with characteristic elegance, and, in 1838, presented a mathematical description which is still in use today. Gauss did not speculate about what might have to exist in the Earth's interior, rather, he developed a purely empirical model which merely described the observations best. Scientists did not start to solve the problem of how the magnetic field might be generated until a century later. In Gauss model, the two additional poles of Hansteen's model were dispensed with and are replaced by significant irregularities in the bipolar field.

The sea passage to the East along the northern extremity of the Americas had remained a dream of merchants and geographers, but slowly these northern coasts were indeed mapped. One of the many workers was the Englishman John Ross (1777-1856), a seasoned explorer of polar regions who, in 1829, set out to search for this Northwest passage. Like many before him, his expedition became trapped in the numerous inlets of the difficult waters around northern Canada and four years passed before he was able to make his way home. The expedition's second in command was James Clark Ross (1800-1862), a nephew of John Ross, and when they lay trapped in the pack-ice, he used the opportunity to localize the magnetic north pole which he felt had to be in the vicinity. It was normal to carry instruments for magnetic observations on such expeditions at that time, and so Ross succeeded and at 70° 5´ N, 96° 47´ W the magnetic field was determined to be quite vertical as far as the instruments accuracy could show. The location of the magnetic north pole had been determined for the first time. The real objective of the expedition, the Northwest passage, remained illusive, but the discovery of the magnetic north pole was a respectable prize to take home and considerable kudos followed: John Ross received a knighthood, and the island on which the pole was located was named Boothia Felix after the expedition's sponsor, the gin distiller Felix Booth. Later, James Clark Ross travelled on new expeditions, among them one attempting to locate the southern magnetic pole, a venture which failed, although the Ross Sea could then be added to the maps, and he too could add "Sir" to his name. Excursions into the Northwest passage region continued with increasing drama culminating in the Franklin catastrophe in the 1840s in which 129 men lost their lives on the tundra of the magnetic pole.

By the beginning of the 20^th century most parts of the Northwest passage region had been mapped and Roald Amundsen, with his modest sloop "Gjøa" could reap the honour for the first through voyage between 1903 and 1905. The expedition had en extra motive, however, and Amundsen wrote in his book on the Northwest passage: "I would like to combine my childhood ream of the Northwest passage with the scientifically far more important goal: To determine the magnetic north pole's current location." Amundsen had guaranteed himself some form of success by including this second scientific target in his plans. He took this scientific goal seriously, however, and worked systematically and thoroughly, and together with the engineer Gustav Juel Wiik learned about geomagnetism from leading experts in Germany where he also had instruments specially built. In the summer of 1902, Amundsen and Axel Steen of the Norwegian Meteorological office travelled the length of Norway to practice using the instruments. In the Northwest passage itself, he chose to spend the winter on King William Land near the magnetic pole, where the declination was all of 89.4° and there established a complete geomagnetic observatory which yielded continuous observations over a 19 month period. This was truly an impressive work, considering that the measurements were made on photographic plates which had to be exchanged and developed each day. The spring of 1904 was devoted to observations in the field to determine the exact location of the pole as exactly as possible; this was far from easy and required patience because magnetic disturbances originating from the sun - magnetic storms - moved the pole around slightly, so the exercise became like hunting a ghost in the tundra. Despite these problems, he succeeded and found that the pole had moved northwards since Ross' observation. Amundsen's expedition might alone be renowned for carrying out its scientific objectives - geomagnetic research.

In the summer of 1905, they broke camp and sailed the final stretch into known waters, but nonetheless had to spend yet another winter in the Arctic, this time on the north coast of Alaska. That the scientific aim was taken seriously was demonstrated yet again as the magnetic observatory was re-erected and observations made throughout the winter, a winter that, unfortunately, Wiik never survived.

So finally, in the spring of 1906, the triumph could be celebrated. After 400 years strife, the Northwest passage had been navigated. The magnetic and other scientific observations were packed and sent to Norway. Amundsen was soon under way with new daring projects and left the demanding analysis of the data to others, and it was not until 1929 that the final scientific result became available that the pole had moved 50 km north between 1831 and 1904.

The next time the pole position was determined was in 1948. The era of the several-year long expeditions was over and the pole could be reached by air in a matter of hours. The pole's position, lying currently in Canada is determined at regular intervals a few years apart by Canadian scientists. Since the days of Ross and Amundsen, the movement of the pole has speeded up considerably and now changes at a rate of 11 km per year.

Where has the pole been and where is it going? If we use all the observations from the 1500s until today and employ Gauss' model, we obtain the map shown here. The pole has enjoyed a sortie into Canada and is now on its way out into the Arctic Ocean again. Geologists find traces of the magnetic field in many minerals (pæleogeomagnetism) and can tell us how the pole has wandered round in the polar region for millions of years, a wandering it will surely continue. Sometimes, however, something quite remarkable occurs. In fact, the current north pole is a south pole in magnetic parlance, this having historical reasons. This has not always been the case, however, and now and again the two poles change place! During the course of the last 5 million years this has happened around 25 times. Approximately 720,000 years have passed sine the poles last changed around and the phenomenon will certainly repeat itself, we just don't know when. The dynamo in the Earth's interior is unstable such that on occasion the field weakens, loses its bipolar character and regenerates with reversed polarity. It is believed that the reversal process takes several thousand years, but some modern opinions suggest a shorter time. We can only imagine the confusion when the compasses of the world suddenly turn a half circle!

Originally published in: "Ultima Thule", Ravnetrykk no. 7, University of Tromsø. Translated from Norwegian by Dr. Chris Hall, University of Tromsø, 1996