CITATION — REFERENCE ENTRY

The Fractal Geometry of Nature — W. H. Freeman

Revision 7d6fd8b9-58ed-46b2-a4e2-a1b0471a0f1b · 11/14/2026, 8:57:51 PM UTC
Key
mandelbrot1983fractal
Authors
Mandelbrot, Benoit B.
Issued
1983
Type
book
Publisher
W. H. Freeman
Publisher place
New York
Edition
Updated and augmented
Raw CSL JSON
{
  "ISBN": "9780716711865",
  "note": "Chapter 5, \"How long is the coast of Britain?\"",
  "type": "book",
  "title": "The Fractal Geometry of Nature",
  "author": [
    {
      "given": "Benoit B.",
      "family": "Mandelbrot"
    }
  ],
  "issued": {
    "date-parts": [
      [
        1983
      ]
    ]
  },
  "edition": "Updated and augmented",
  "publisher": "W. H. Freeman",
  "publisher-place": "New York"
}

Claims

  1. On a coastline altered by people, such as the coast at Chelsea, measured length changes little for measuring steps between roughly 20 metres and 20 centimetres, then rises again once individual stones affect the measurement.
    "there is an intermediate zone of ε's in which L(ε) varies little. This zone may go from 20 meters down to 20 centimeters"
    Locator: page: 27 · Quote language: en
  2. Mandelbrot's first-approximation model of coastline formation gives D = 3/2, which he says is much too large to describe real coasts but enough to show that a coastline's dimension should be expected to exceed 1.
    "to a first approximation D=3/2. This value is much too large to describe the facts"
    Locator: page: 30-31 · Quote language: en
  3. The value of D depends on the coastline chosen, and different pieces of the same coastline may give different values.
    "different pieces of the same coastline, if considered separately, may produce different values of D."
    Locator: page: 29 · Quote language: en
  4. Citing Richardson (1961), Mandelbrot reports that the lengths of the shared borders of Spain and Portugal, and of Belgium and the Netherlands, given in the neighbouring countries' encyclopedias differ by 20%, and that the discrepancy must in part come from different measuring steps.
    "the lengths of the common frontiers between Spain and Portugal, or Belgium and Netherlands, as reported in these neighbors' encyclopedias, differ by 20%. The discrepancy must in part result from different choices of ε."
    Locator: page: 27 · Quote language: en
  5. Mandelbrot proposes calling curves whose fractal dimension exceeds 1 'fractal curves', and concludes that within the scales of interest to geographers, coastlines can be modelled by fractal curves.
    "within the scales of interest to the geographer, coastlines can be modeled by fractal curves."
    Locator: page: 31 · Quote language: en
  6. Mandelbrot states that no mathematical proof that Richardson's D is a dimension is possible in a natural science; his aim is to show that length poses a conceptual problem and that D is a workable answer.
    "No such proof is conceivable in any natural science."
    Locator: page: 30 · Quote language: en
  7. Mandelbrot argues that a single measuring step is unlikely to be adopted by all agencies of one government, let alone all countries, and that geographic length is therefore not entirely objective: the observer is part of its definition.
    "the adoption of the same ε by all the agencies of a government is hard to imagine, and its adoption by all countries is all but inconceivable. ... The observer inevitably intervenes in its definition."
    Locator: page: 27-28 · Quote language: en
  8. Mandelbrot suggests it is unsurprising that a small country such as Portugal measures its borders more finely than its larger neighbour.
    "one should not be surprised that a small country (Portugal) measures its borders more accurately than its big neighbor."
    Locator: page: 27 · Quote language: en
  9. For a straight line, approximate measurements with different steps are essentially identical; for a circle they increase but converge rapidly to a limit. Curves whose length is defined in this way are called rectifiable. For coastlines, the approximate length tends to increase steadily without bound as the step shrinks.
    "For a circle, the approximate measurements increase but converge rapidly to a limit. The curves for which a length is thus defined are called rectifiable."
    Locator: page: 27 · Quote language: en
  10. Mandelbrot describes Lewis Fry Richardson as a scientist known for durable ideas about turbulence who also studied the nature of armed conflict between states.
    "He also concerned himself with other difficult problems, such as the nature of armed conflict between states."
    Locator: page: 29 · Quote language: en
  11. Richardson's diagrams of measured length against step size were found among his papers after his death and published in a little-read yearbook.
    "found among his papers after he died, were published in a near confidential (and totally inappropriate) Yearbook."
    Locator: page: 29 · Quote language: en
  12. Mandelbrot quotes Hugo Steinhaus (1954), who observed that the left bank of the Vistula, measured with increasing precision, would give lengths ten, a hundred or a thousand times that read off a school map, and concluded that most curves in nature are not rectifiable.
    "the left bank of the Vistula, when measured with increasing precision, would furnish lengths ten, hundred or even thousand times as great as the length read off the school map...[A] statement nearly approaching reality would be to call most arcs encountered in nature nonrectifiable."
    Locator: page: 28 · Quote language: en
Available in