William Gilbert Predicts the Northeast Passage
William Gilbert's De Magnete, Book Four
This post is part of a series of posts about William Gilbert’s De Magnete (On the Magnet1), which is composed of six books. This post is about Book Four. Here are the previous posts related to De Magnete.
De Magnete, Nothing Less than the First Ever Work of Experimental Physics
Book One, part 1: William Gilbert Writes about the Loadstone
Book One, part 2: William Gilbert Examines Iron, Calls Aristotle's Earth Element Dead
Book Two, part 1: William Gilbert Compares Electric Bodies to Magnetic Bodies
Book Two, part 2: William Gilbert Discusses Magnetic Bodies
Book Two, part 3: William Gilbert Considers the Internal Structure of the Terrella
Book Two, part 4: William Gilbert States that the Moon Causes the Tides
Book Two, part 5: William Gilbert Aligns Several Loadstones
Book Three, part 1: William Gilbert Explains How Magnetic Bodies Acquire Direction
Book Three, part 2: William Gilbert Shows the Direction of Compass Needles
In this post, I will explain how William Gilbert understood the variation — today called declination — of compass needles from the direction towards true geographic north at any given point on the earth, and how, based on this understanding, that he was able to predict that there is a Northeast Passage connecting the Atlantic and Pacific Oceans through the Arctic Ocean.
For Gilbert, a compass needle varies the most in the ocean, when it is close to a continent. He had previously claimed that it is the earth’s mass that creates its magnetic properties, so it follows naturally from that claim that the presence of a large mass rising out of the ocean, i.e., a continent, would attract the compass needle. He also claims that variation tends to be higher at higher altitudes.
He illustrates the idea of variation near continents with the diagram below of a terrella with poles P and M, along with perturbations B, F and H, which he calls eminences. Versorium A varies towards eminence B, versorium C towards F, and versorium N towards H. Versorium G does not vary, as it is between eminences B and F; and versorium L does not vary, as it is sufficiently far from eminences, and the little island next to it is too insignificant.
The figure shows a terrella with uneven surface. The demonstration is made with small bars or short needles placed on the terrella: they turn from the terrella toward the projecting mass and the great eminences. In this way is verticity disturbed on the earth by the great continents which mostly rise above the beds of the seas and which at times cause the needle to deviate from the straight track, i.e., from the true meridian. The tip of the versorium A does not point toward the pole P if there be a large projection B on the terrella; so, too, the point C varies from the pole because of the projection F. Midway between the two eminences, the needle G points to the true pole, because, being equidistant from both projections B and F, it deviates to neither but keeps the true meridian, particularly when the energy of the projections is equal. But elsewhere, at N, the needle varies from the pole M toward the eminence H, nor is hindered nor stayed nor checked by the small eminence D on the terrella, which is like some island of the earth in the ocean. But L unhindered tends poleward. [pp.238-240]
He continues with this second diagram, which shows a terrella with north pole A in the center of the diagram, the equator and latitude 30ºN respectively labeled B and C, and eminences D and E. The versorium F (in eminence D) does not vary, versorium G (west of F) varies significantly because its proximity to D, and versorium H (west of G) varies, but not as much. As for the versoria near E, versorium I does not vary, while versoria L and M do.
Let A be the earth’s pole; B its equator; C a parallel circle at latitude 30 degrees; D an eminence reaching poleward; E another eminence stretching from the pole equatorward. Evidently the versorium F in the middle line of D does not vary; but G deflects very much, C very little as being more remote from D. So, too, the needle I, placed directly toward E, does not deflect from the pole; but L and M turn from the pole toward the eminence E. [p.240]
Gilbert assumed that the variation is constant. However, this was later to be proven false, as is shown by the following footnote:
Henry Gellibrand, English mathematician, professor of geometry at Gresham College, discovered, 1633-1635, the secular variation of the declination. In the words of Dr. Whewell (‘‘Hist. of the Ind. Sciences,” 1859, Vol. II, page 219): “Gellibrand discovered that the variation is not constant, as Gilbert imagined, but that in London it had diminished from eleven degrees east in 1580 to four degrees in 1633. Since that time the variation has become more and more westerly; it is now about twenty-five degrees west, and the needle is supposed to have begun to travel eastward again.
It may be added that the diurnal variation was subsequently found by George Graham, during the year 1722, whilst the annual variation was made known by Jean Jacques Dominique Cassini between 1782 and 1791. [p.240]
Gilbert discusses at length the construction of instruments to measure variation, either on the land or at sea. He also explains that published variation numbers for different parts of the globe must be treated very carefully, since the compasses built in different parts of Europe were calibrated to take into account the local variation where they were built.
There are in general use in Europe four different constructions and forms of compass. First, the form adopted throughout the Mediterranean, and in Sicily, Genoa, and the Venetian republic. In all of these compasses the pieces of iron are so attached beneath to the rotating card that (where there is no variation) they turn to the true points of north and south. Hence the mark for north, designated by a lily, always indicates exactly the point of variation: for the point of the lily on the card, together with the ends of the pieces of magnetized iron beneath, come to a standstill at the point of variation. Another form of compass is that of Dantzic [Danzig, now Gdańsk], employed in the Baltic Sea and in the Netherlands. Here the magnetized iron underneath diverges three fourths of one point eastward from the lily; for a voyage to Russia the divergence (recognized difference) is two thirds. But the compasses made at Seville, Lisbon, Rochelle, Bordeaux, Rouen, as well as throughout all England, have an interval of one half of a point. [pp.249-250]
Thus, when reading and using a chart, a mariner must know under what assumptions both the chart and his compass were made:
For the mariner, who, using British compass, should follow the directions of the Mediterranean marine charts, must needs stray far from his true course; so, one who should use an Italian compass in the North Sea, the German Sea, or the Baltic, in connection with the marine charts commonly used in those parts, would oft stray from the right direction. [p.250]
After these discussions based on a combination of theorizing, experimenting with terrellae, and reading reports from mariners, Gilbert moves on to examining the variation in different parts of the globe.
I was stunned when I read Chapter XVI, entitled “Of the variation in Nova Zembla.” On today’s maps, we read Novaya Zemlya (Новая Земля), which means ‘New Land’ in Russian. The island separates the Barents Sea and the Kara Sea.
Here is the paragraph in question:
THE variations are greatest in regions nigh to the poles, as has been proved, and there, too, the changes of variation are sudden, as Dutch observers noted some years ago, though their observations were not exact; yet the inexactitude can be excused, for, with the ordinary instruments, it is hard to get at the truth in such high latitudes — about 80 degrees. But now the variation of the compass gives the clear evidence of the existence of an open passage eastward through the North Sea — Arctic Ocean (Mare Scythicum), for, since the compass has so great an arc of variation to the west, it is evident that no continent stretches for any great distance along that whole route eastward. Therefore we can strive and explore more hopefully for a passage to the Moluccas by the northeast than by the northwest. [p.269, emphasis mine]
So, in 1600, almost a century before tsar Peter I “The Great” (1672-1725) became sole ruler of Russia and initiated exploratory voyages in the Arctic Ocean, and almost three centuries before the Vega expedition (1878-79) led by explorer Adolf Erik Nordenskiöld (1832-1901), funded by Swedish king Oscar II (1829-1907), William Gilbert had predicted the existence of the Northeast Passage!
What an amazing prediction!
If you wish to donate to support my work, please use the Buy Me a Coffee app.
William Gilbert. De Magnete. Dover, New York, 1958. Translation by P. Fleury Mottelay of De Magnete, first published in 1600.