TAS tas v.2010/02/18 Prn:2010/04/15; 9:36 F:tas09154.tex; (Milda) p. 1 1 59 2 60 3 61 4 62 5 63 6 64 7 65 8 66 9 67 10 68 11 69 12 70 13 71 14 72 15 73 16 74 17 75 18 76 19 77 20 78 21 79 22 80 23 81 24 82 25 83 26 84 27 85 28 86 29 87 30 88 31 89 32 90 33 91 34 92 35 93 36 94 37 95 38 96 39 97 40 98 41 99 42 100 43 101 44 102 45 103 46 104 47 105 48 106 49 107 50 108 51 109 52 110 53 111 54 112 55 113 56 114 57 115 58 116 Please see the online version of this article for supplementary materials. The First (Known) Statistical Graph: Michael Florent van Langren and the “Secret” of Longitude Michael F RIENDLY, Pedro V ALERO-MORA, and Joaquín I BÁÑEZ ULARGUI A 1644 diagram by Michael Florent van Langren, showing estimates of the difference in longitude between Toledo and Rome, is sometimes considered to be the first known instance of a graph of statistical data. Some recently discovered documents help to date the genesis of this graphic to before March 1628, and shed some light on why van Langren chose to display this information in this form. In the process, we discovered three earlier versions of the 1644 graph and one slightly later repro- duction. This article describes these early attempts on the so- lution of “the problem of longitude” from the perspective of a history of data visualization. KEY WORDS: Data visualization; History of statistics; Lon- gitude; Michael Florent van Langren; Selenography; Statistical historiography; Thematic cartography; Uncertainty. 1. INTRODUCTION Every picture tells a story (Rod Stewart) In the history of statistical graphics (Friendly 2008), as in other artful sciences, there are a number of inventions and de- velopments that can be considered “firsts” in these fields. The catalog of the Milestones Project in the history of data visual- ization (Friendly and Denis 2001) lists 70 events that can be considered to be the initial use or statement of an idea, method, or technique that is now commonplace. Some early examples related to the theme of this article are: • map projections of a spherical earth and use of latitude and longitude to characterize position (Claudius Ptolemy, c. 150 AD) • the first modern atlas, Teatrum Orbis Terrarum (Abraham Ortelius 1570) Michael Friendly is Professor of Psychology, ???, York University, Toronto, Canada (E-mail: friendly@yorku.ca). Pedro Valero-Mora is Professor, Fac- ultad de Psicologia, Universitat de Valencia, Spain (E-mail: valerop@uv.es). Joaquín Ibáñez Ulargui is Cientifico Titular del CSIC, Centro Nacional de In- vestigaciones Metalúrgicas, Madrid, Spain (E-mail: jiu@cenim.csic.es). This research was aided by the National Sciences and Engineering Research Coun- cil of Canada, grant OGP0138748 to the first author. We are grateful to Robert Karrow, Peter van der Krogt, Doug Sims, and other correspondents on the Map History discussion list for information related to van Langren and the topics discussed in this article. • first world map showing lines of geomagnetism (isogons), used in work on finding longitude by means of magnetic vari- ation (Guillaume le Nautonier 1604). Among these, there is also: • first visual representation of statistical data: variations in de- termination of longitude between Toledo and Rome (Michael Florent van Langren 1644). (Like others in his family and in this time, his name was written in various forms and in dif- ferent languages: Miguel Florencio, Michale Florent, etc. We use the English version here, abbreviated as MFvL where it serves in the following.) If this is truly the first exemplar of a graph of statistical data, van Langren should be canonized in this history, along with the contributions of William Playfair (1786, 1801) (invention of the line graph, bar chart, and pie chart), Charles Joseph Minard (“the best statistical graphic ever drawn”), Florence Nightingale (1857) (the use of the “coxcomb,” or rose diagram for social and political advocacy), and others. Yet, van Langren and the history of his graph remain little known. However, it is important to make clear (a) that many such “firsts” are relative rather than absolute and (b) that “priority chasing” for its own sake is often an unprofitable goal in his- toriography (May 1975). On the first point, many “first” devel- opments in the history of data visualization were preceded by other contributions that count as a “first” under different specifi- cations. For example, the idea of a system of latitude and longi- tude for a map of the world was first proposed by Eratosthenes in the third century BC; the idea of the “coxcomb”—a polar area chart—goes back at least to André-Michel Guerry (1829), and Nightingale almost certainly was introduced to this graphic form by William Farr (1852). On the second point, it is more useful to understand the context in which a significant histori- cal event occurred. This article describes the background and questions that led to van Langren’s graph (Figure 1), conventionally dated to a 1644 publication, La Verdadera Longitud por Mar y Tierra (The true longitude for sea and land). We show why using a graph was effective for van Langren to achieve his communica- tion goals. More importantly, some recently discovered letters by van Langren and others help tell the story behind this early graph and serve to date its genesis to early 1628. 1.1 Other Early Graphs We claim that this one-dimensional line graph by van Lan- gren is the first known visual representation of statistical data (i.e., empirical data where uncertainty of the observations is © 2010 American Statistical Association DOI: 10.1198/tast.2010.09154 The American Statistician, Accepted for publication 1