Hostname: page-component-7b7567c8c-wxvwp Total loading time: 0 Render date: 2026-11-14T03:10:23.223Z Has data issue: false hasContentIssue false
  • English
  • Français

Columbus's Method of Determining Longitude: An Analytical View

Published online by Cambridge University Press:  21 October 2009

Keith Pickering
Get accessRights & Permissions [Opens in a new window]

Extract

On 14 September 1494, Christopher Columbus observed a lunar eclipse while at the island of Saona near the eastern tip of Hispaniola. He later recorded in his Libro de las Profecias that, from his timing of the eclipse, he determined his longitude to be ‘five hours and more than one half’west of Cape San Vicente. His actual longitude was three hours and 59 minutes west of Cape San Vicente, so Columbus was off by over an hour and a half, some 23 degrees of longitude.

Keywords

1. History.2. Astronomical navigation.

Information

Type
Research Article
Information
The Journal of Navigation , Volume 49 , Issue 1 , January 1996, pp. 95 - 111
Copyright
Copyright © The Royal Institute of Navigation 1996

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

References

REFERENCES

1West, D. C. and Kling, A. (1991). The Libro de las profecias of Christopher Columbus. University of Florida Press, Gainsville, pp. 226–227.Google Scholar
2West, D. C. and Kling, A. (1991). The Libro de las profecias of Christopher Columbus. University of Florida Press, Gainsville.Google Scholar
3Olson, D. W. (1992). Columbus and an eclipse of the Moon. Sky & Telescope (October), 437–44O.Google Scholar
4Ashley, R. E. (1991). The search for longitude. The American Neptune 51: 4, 261–262.Google Scholar
5Molander, A. B. (1992). Columbus and the method of lunar distances. Terrae lncognitae 24, 65–78.CrossRefGoogle Scholar
6Dunn, O. and Kelley, J. E. Jr (1989). The Diario of Christopher Columbus's First Voyage to America, 1492–1493. Norman and London, University of Oklahoma Press, p. 329.Google Scholar
7Morison, S. E. (1942). Admiral of the Ocean Sea. Little, Brown, & Co., Boston, p. 65.Google Scholar
8Jane, C. (1988). The Four Voyages of Columbus, 2 vols. Dover Publications, New York, II 34,.Google Scholar
9Kelley, J. E. Jr. (1983). In the wake of Columbus on a Portolan chart. Terrae lncognitae 15, 102–107.Google Scholar
10Marden, L. (1986). The first landfall of Columbus. National Geographic 170, 573 – 575.Google Scholar
11McElroy, J. W. (1941). The ocean navigation of Columbus. The American Neptune 1, 209–240.Google Scholar
12Richardson, P. L. and Goldsmith, R. A. (1987). The Columbus landfall: voyage track corrected for winds and currents. Oceanus 30: 3, 3–10.Google Scholar
13Molander, (1992), 77–78.CrossRefGoogle Scholar
14Molander, (1992), 75.Google Scholar
15Dunn, & Kelley, (1989), p. 341.Google Scholar
16Jane, (1988), II, 26.Google Scholar
17Dunn, & Kelley, (1989), p. 373.Google Scholar
18Molander, (1992), 74, fn 19.CrossRefGoogle Scholar
19Davies neglected to calculate parallax when determining the Moon's apparent position. Rawlins' corrected analysis shows that Vespucci must have been off the coast of Guinea (not Brazil as Vespucci claimed) to have made the observation described. Both Davies' paper and Rawlins' rebuttal are unpublished; but see Dennis Rawlins, ‘Peary, verifiability, and altered data’, DIO 1: 1 (1991/1/14), 29.Google Scholar
20DrCharles, Kowal as quoted in The Washington Post, 11 Dec. 1990: A5.Google Scholar
21Molander, (1992), 68.Google Scholar
22Molander, (1992), 68–69.Google Scholar
23I have followed Molander's convention and used Columbus's daily positions as given by John W. McElroy in ‘The ocean navigation of Columbus’, The American Neptune 1 (1941), 219–239. The position on 27 February is not given by McElroy in this article, but it can be measured from the map in Morison (1942) (opposite p. 228) at 37°.7 N, 17°.6 W.Google Scholar
24 The conjunction occurred at 22:16:10 Universal Time on 24 September, which is 191 2 hours Local Mean Solar Time for an observer at 45°. 58.4 W, McElroy's position at dawn on the 26th-which would have been dawn on the 26th after applying Molander's 24-hour shift. Molander's given time of 1907 hours LMST is therefore off by at least 5 minutes. Eliminating the 24-hour shift puts Columbus at 45° 9'.2 at dawn on the 25th, and the conjunction at 0715 hours. But we should really interpolate McElroy's dawn positions for 1915 hours the previous night, which would put Columbus at 29° 8'.4 N, 44° 47'.8 W, and the time of conjunction at 1917 hours LMST. We should also note that in this era before accurate clocks, the concept of mean time was unknown; the time used in this period was local apparent solar time, or ‘sundial time’.Google Scholar
25 Lunar positions have been calculated from Michelle Chapront-Touze and Jean Chapront, Lunar Tables and Programs from 4000 B.C. to A.D. 8000 (Richmond: Willmann-Bell, 1991). Accuracy at this epoch is approximately 37 arcseconds. Planetary positions have been calculated from Pierre Bretagnon and Jean-Louis Simon, Planetary Programs and Tables from –4000 to +2800 (Richmond: Willmann-Bell, 1986). Accuracy at this epoch is approximately 7 arcseconds for Jupiter and Saturn, 23 arcseconds for Mars, 10 arcseconds for Venus and Mercury, and 2 arcseconds for the Sun. I have optimistically assumed that Columbus was aware of lunar parallax and corrected for it at all times; this correction is absolutely necessary for any longitude determination, since it can alter the moon's apparent location by over a degree and the resulting position fix by as much as 30° in longitude. I must point out, however, that there is no historical evidence that Columbus was aware of lunar parallax, nor of how to correct for it.Google Scholar
26 Dr Schaefer has contributed a letter on this subject to the ongoing Columbus Landfall Round Robin, of which both Molander and myself are contributing members. These correspondences have a wide distribution, and can be considered public, albeit unpublished.Google Scholar
27Molander, (1992), 75, fn 20.CrossRefGoogle Scholar
28 By coincidence, no Mercury conjunctions are correlated to a position fix.Google Scholar
29 By coincidence, a probably invisible daytime conjunction with Venus is correlated to a position fix.Google Scholar
30 Including daytime conjunctions would have given Molander one more correlation success, on 15 February, but the large increase in P caused by including all the uncorrelated daytime conjunctions actually would have been less favourable for the alleged lottery-odds coincidence, even with the 15 February success. Thus, Molander excludes the daytime conjunctions. By coincidence, Molander also excludes the position of 15 February and so avoids a correlation failure.Google Scholar
31 By coincidence, two of Molander's correlation successes occur below the horizon.Google Scholar
32 By coincidence, all of these errors favour a correlation between conjunctions and position fixes.Google Scholar
33Molander, (1992), 75, 78.CrossRefGoogle Scholar
34Molander, (1992), passim.Google Scholar
35Molander, (1992), 77.Google Scholar
36Jane, (1988), II, 12.Google Scholar
37Molander, (1992), 77.Google Scholar
38Molander, (1992).Google Scholar
39Morison, (1942), p. 68.Google Scholar
40Jane, (1988), II, 34, 84.CrossRefGoogle Scholar
41McElroy, (1941), 219–239.Google Scholar
42Chapront-Touzé, & Chapront, (1991).Google Scholar
43William, H. Beyer, CRC Standard Mathematical Tables and Formulae, 29th edition (CRC Press, Boca Raton, 1991), p. 492.Google Scholar
44William, H. Beyer (ed), Handbook of Tables for Probability and Statistics, 2nd Edition (Chemical Rubber Co., 1968), p. 283.Google Scholar