Listen: https://serve.buildtall.systems/72b81dc9d09df7662a5dfeb73b54a28204168a9deda54ff31f938bd258ed493a.mp3
Source: https://serve.buildtall.systems/85258127800abbed3476e10ee9bba0ce8a259c1f3530d741fc6e331c9721c2a3.pdf
The Sand Reckoner is a short mathematical treatise that the text presents as written by Archimedes and addressed to King Gelon.
OpenAlex has no record of this document: no work in the index matches its title and authors, as of 2026-10-16, so this episode makes no claim about how often it has been cited. The document itself states no venue of publication, so none is given here. The work opens by taking up a claim that some people hold, that the number of the sand is infinite in multitude, and it argues against that view as well as against the view that no named number could exceed the grains of sand. Its method is a chain of geometrical proofs rather than an experiment or a field study. It sets out assumptions about the sizes of the earth, the moon and the sun, bounds them with inequalities, and then builds a system of named numbers large enough to count the grains of sand in spheres of growing size. The data are assumed measurements, not collected observations, although the author describes a simple rod-and-disc method for estimating the angle the sun subtends at the eye. The evidence is strongest where the steps are explicit proofs and weakest where it rests on chosen figures, which the author states as upper limits, such as an earth whose perimeter is taken to be no more than about 3,000,000 stadia. The text was extracted from a PDF, so some exponents in its number scheme are garbled and the exact powers should be checked against the original.
The work opens by addressing King Gelon and setting out a claim about sand. Some people, the writer says, hold that the number of grains of sand is infinite, and by sand the writer means not only the sand around Syracuse and the rest of Sicily but all the sand found in every region, whether inhabited or uninhabited. The writer puts the first view as "the number of the sand is infinite in multitude." A second group does not regard the number as infinite, but still thinks that no number has yet been named that is large enough to exceed the multitude of sand.
The writer argues that the second group is further from the truth than it realises. Imagine, the writer says, a mass of sand equal in size to the earth, with all the seas and the hollows of the earth filled up to the height of the highest mountains. Those who hold the second view would, in the writer's words, be "many times further still from recognising that any number could be expressed which exceeded the multitude of the sand so taken."
The writer promises to show, by means of "geometrical proofs, which you will be able to follow," that some of the numbers named in an earlier work, sent to Zeuxippus, exceed not only the number of grains in that earth-sized mass of sand but also the number of grains in a mass of sand equal in size to the universe. The rest of the work sets out to prove that claim.
The author begins by defining a word. Most astronomers, the text says, use 'universe' as the name for a particular sphere: its centre is the centre of the earth, and its radius is "equal to the straight line between the centre of the sun and the centre of the earth." On this view, the universe reaches as far as the sun is from the earth. The author presents this as the common account, the view astronomers have handed down, and not as a claim of the author's own.
The text then turns to Aristarchus of Samos, who, the author reports, wrote a book of hypotheses. A hypothesis here is an assumed starting proposition, and the book's premisses, the statements it builds from, lead to a striking result: "the universe is many times greater than that now so called."
The text sets out the hypotheses one by one. First, the fixed stars and the sun stay unmoved. Second, the earth travels round the sun along the circumference of a circle, with the sun lying in the middle of that orbit. Third, the sphere of the fixed stars shares its centre with the sun and is so large that the circle of the earth's orbit stands in a fixed ratio to the distance of the stars. The effect of that ratio is to place the stars at a distance so vast, compared with the earth's orbit, that the orbit becomes a very small thing by comparison. The author is using these assumptions to show how a much larger universe could follow from ideas that were already in circulation.
Now it is easy to see that this is impossible, and the author explains why. He writes that "since the centre of the sphere has no magnitude, we cannot conceive it to bear any ratio whatever to the surface of the sphere." A point has no size, so it cannot stand in a measurable relation to a surface.
The author then says how the astronomer Aristarchus of Samos must be read. Aristarchus described the earth as moving in a circle, and the author takes him to mean that, since the earth is treated as the centre of the universe, "the ratio which the earth bears to what we describe as the universe is the same as the ratio which the sphere containing the circle in which he supposes the earth to revolve bears to the sphere of the fixed stars." The fixed stars were, in the ancient picture of the cosmos, the outermost sphere carrying the stars. The author says Aristarchus framed his proofs to fit this hypothesis, and in particular treated the sphere carrying the earth's orbit as equal in size to what we call the universe.
The author then sets his goal. He says that even if a sphere were filled with sand, as large as Aristarchus supposes the sphere of the fixed stars to be, he will still prove that some of the numbers named in the Principles exceed in multitude the grains of sand that would fill it. He adds that this holds provided certain assumptions are granted.
The first assumption concerns the earth. He writes:
The perimeter of the earth is about 3,000,000 stadia and not greater.
A stadion was an ancient Greek unit of length, roughly the length of a sports stadium's running track. The author notes that some earlier writers had tried to show the perimeter was about 300,000 stadia. He goes further, putting the earth at ten times the size his predecessors had thought.
The second assumption orders the bodies by size. The diameter of the earth is greater than the diameter of the moon, and the diameter of the sun is greater than the diameter of the earth. The author says that in this assumption he follows most of the earlier astronomers.
The third assumption concerns the sun. The author writes:
The diameter of the sun is about 30 times the diameter of the moon and not greater.
He sets this against earlier estimates. Eudoxus had declared the sun about nine times the moon's diameter, and the author's father, Pheidias, twelve times. Aristarchus had tried to prove that the sun's diameter is greater than 18 times but less than 20 times the moon's. The author says he goes even further than Aristarchus.
Archimedes opens this stretch by saying what he means to establish. His claim, he writes, may be established "beyond dispute", using the working figure that the sun's diameter is "about 30 times that of the moon and not greater." A translator's note explains that this is not strictly an assumption but a proposition the author proves later, by means of an experiment he is about to describe.
The main claim is that the diameter of the sun is greater than the side of a chiliagon, a figure with a thousand equal sides, inscribed in the greatest circle of the sphere of the universe. To prove it, the author sets up a cross-section. He pictures a plane through the centre of the sun, the centre of the earth and the eye of the observer, taken at the moment the sun has just risen above the horizon. That plane cuts the earth in one circle, the sun in another, and the sphere of the universe in a great circle. The argument that follows is a comparison of lengths and angles drawn from this diagram.
Before the geometry, the author describes how he arrived at the angle. He credits Aristarchus with an observation about the sun's apparent size against the circle of the zodiac, and says he tried a method of his own to measure the angle the sun subtends at the eye. The method is simple. Take a long rod, fasten a small cylinder or disc to one end, and point the rod at the sun just after it rises, when it is still possible to look at it directly. Then move the cylinder to the distance at which it just hides the sun and just fails to hide it, and measure the angle the cylinder subtends. He also allows for the fact that the eye does not see from a single point: "the eye does not see from one point but from a certain area."
The experiment gives two bounds on the angle the sun subtends, an upper and a lower limit, each stated as a fraction of a right angle. The translator notes that the treatise has been rendered literally up to this point, because of the historical interest of Archimedes' own words, and that the remainder can be reproduced more freely. The comparison with the thousand-sided figure rests on those two bounds.
The argument now turns to the sizes of the sun and the earth. Two tangents are drawn from the point E to a circle, touching it at P and Q, and two more are drawn from C, touching it at F and G. Lines are then joined from O to E, F, G, P and Q, and a chord AB is drawn across the great circle, meeting CO at M.
The first step compares angles. The work notes, "Now CO > EO, since the sun is just above the horizon," and from that concludes that the angle PEQ is greater than the angle FCG. It then argues that the angle FOG is less than a right angle, and that the stronger claim follows more easily from this than the weaker one would.
Next, the chord AB is shown to subtend an arc of the great circle that is smaller than a small fraction of its circumference. From this, AB is shown to be less than the side of a 656-sided polygon inscribed in the circle. The work then uses a known bound on the perimeter of any polygon inscribed in a circle, drawn from an earlier proposition on measuring a circle, to bring AB below a fixed fraction of CO. The text labels this inequality (a).
The argument then uses right-angled triangles. Because OA equals CO and AM is perpendicular to CO, while OF is perpendicular to CA, the work concludes that AM equals OF. Doubling AM gives AB, and AB is therefore the diameter of the sun. Applying the bound labelled (a), the diameter of the sun is shown to be less than that fixed fraction of CO. The same reasoning, the work says, shows a fortiori that the diameter of the earth is smaller still.
The proof ends by comparing two lengths. From the same steps, the ratio of CO to HK is shown to be less than 100 to 99.
The argument now turns to the angles. The work compares two right-angled triangles, CFO and EQO, and sets the angle at E against the angle at C. It notes that "EQ < CF (since EO < CO)", so the side in the second triangle is shorter than the side in the first, and it draws on the fact that CO is greater than CF. From these comparisons the proof reaches an inequality between the two angles, carries a ratio of 100 to 99 over from a step set out earlier, and doubles the angles to finish the comparison.
The author then uses a hypothesis about the angle PEQ, which is taken to be larger than a stated fraction of a right angle. Combined with the previous step, this gives the angle ACB a value above the same fraction. The work draws the consequence that the arc AB is greater than a fixed fraction of the circumference of the great circle through A, O and B. A great circle is a circle on a sphere whose plane passes through the sphere's centre.
A chiliagon is a polygon with a thousand sides. The argument uses one inscribed in that great circle. Since the arc AB is larger than the fraction just established, the work concludes, "Hence, a fortiori," that AB is greater than the side of that chiliagon. The phrase a fortiori means that the conclusion follows all the more strongly from what was shown. The work then recalls that "AB is equal to the diameter of the sun, as proved above", so the sun's diameter is larger than the side of the inscribed chiliagon.
The proof depends on an inequality that the work itself flags as a step needing justification. In its words, "The proposition here assumed is of course equivalent to the trigonometrical formula" for two acute angles, one larger than the other. Stated in modern terms, the larger angle's tangent divided by the smaller angle's tangent is greater than the larger angle's sine divided by the smaller angle's sine. The reader can take this as a known relation between the tangent and the sine of two angles.
With the sun's diameter established against the chiliagon, the work sets out the quantities it will compare. "Suppose, for brevity," the author writes, that one symbol stands for the diameter of the universe, another for the sun, a third for the earth, and a fourth for the moon. The text then gives three assumptions: the sun's diameter is more than thirty times the moon's, the earth's diameter is greater than the moon's, and the sun's diameter is less than thirty times the earth's. The text labels these as assumptions and uses them as the basis for the comparisons that follow.
The work states the results it can now prove. The diameter of the universe is less than 10,000 times the diameter of the earth, and it is also less than 10,000,000,000 stadia, an ancient unit of length. The sum of these steps is that the universe, as the author measures it, is far smaller than the earlier estimates. The passage ends at the point where the sun's diameter is shown to exceed the side of the chiliagon, which the next step of the proof builds on.
The author now fixes the sizes that will bound the argument, beginning with a polygon of a thousand sides, called a chiliagon, drawn inside a circle. The key fact is a basic property of polygons:
But the perimeter of any regular polygon with more sides than 6 inscribed in a circle is greater than that of the inscribed regular hexagon, and therefore greater than three times the diameter.
The hexagon's perimeter is three times the diameter of the circle, so any regular polygon with more sides has a longer perimeter, and the chiliagon's perimeter therefore exceeds three diameters. This gives a lower limit that the author uses to keep the circle's size in check. Next the author takes the earth as a sphere. It is assumed that the earth's perimeter is no more than 3,000,000 stadia, a stadion being an ancient Greek unit of length of roughly 180 metres. From that the earth's diameter is taken to be less than 1,000,000 stadia. The universe's diameter is then taken to be less than ten thousand times the earth's, which with the earth bound gives less than 10,000,000,000 stadia. The author then makes a physical assumption that is needed to turn these sizes into a count. The argument supposes "a quantity of sand taken not greater than a poppyseed, and suppose that it contains not more than 10,000 grains." A poppy-seed's diameter is likewise taken to be no less than a small fraction of a finger-breadth, so that a sphere filled with sand can be bounded from both sides. With the sizes set, the author turns to the problem of naming numbers. Greek counting words reach only as far as a myriad, or 10,000, and so a myriad myriads, or 100,000,000, is the largest number the usual names express. The author calls these "numbers of the first order." To go further, the author treats 100,000,000 as a new unit, the unit of a second order of numbers running from that unit up to its own square. The same device is repeated: "Let this again be the unit of the third order of numbers," and the process continues through successive orders until it reaches a very large final order, which the author labels P. The purpose is to have names for numbers large enough to count the grains of sand needed for the claim that follows.
The argument now moves from naming numbers to building a scheme that can name very large ones. The ordinary Greek words stop at the myriad, which is ten thousand, so Archimedes sets out a system of periods, each divided into orders, that can climb far beyond it.
The first period counts from one up to a hundred million, which the text calls a myriad-myriad. Each order then runs on from the one before it. The unit that begins the second period is the number that closes the first, and the text says the second period consists "of the numbers from P up to 100,000,000 P", where P names the end of the first period. The unit of the first order of the second period is therefore P itself, and each later order multiplies the one before by a further hundred million. The third period takes its starting unit from the end of the second in the same way, and the fourth takes its start from a yet higher power. The extracted text garbles some of the exponents at this stage, so the exact powers should be checked against the original.
The passage describes the largest number in the scheme in the author's own terms, as "a myriad-myriad units of the myriad-myriad-th order of the myriad-myriad-th period". The text then states that this number is a hundred million times a product built from a hundred million raised to a power, with the exponents again hard to read in the extraction. The author adds that "The scheme of numbers thus described can be exhibited more clearly by means of indices as follows", and sets out a table in which each order and period is written as a power of ten. That table is the clearest guide to how the orders stack, and it is the basis for the counting that follows.
The work now turns to a system for naming numbers far larger than ordinary counting reaches. Its scheme sorts numbers into periods and orders, and the author is careful to show how quickly the scale grows. The text says that "the last number in the first period would be represented now by 1 followed by 800,000,000 ciphers". A cipher here is a written zero. The author then notes that the last number of the later period "would require 100,000,000 times as many ciphers, i.e. 80,000 million millions of ciphers". The point is that the notation still works after the numbers have become impossible to write out in full.
To place long runs of numbers in this scheme, the author uses a series in continued proportion, which is a geometric progression: each term is found by multiplying the one before it by a fixed ratio. The series starts with 1 and then 10, so its terms are successive powers of ten. The author writes:
Consider the series of terms in continued proportion of which the first is 1 and the second 10.
The terms are grouped in blocks of eight, which the work calls octads. The first octad, running from 1 up to ten million, falls under the first order of the first period, and the second octad falls under the second order of that period. The author adds that
We can, in the same way, place any number of octads.
The work then states a theorem about this series. In modern terms, it says that multiplying two powers of ten adds their positions in the series. The text puts it this way: "If there be any number of terms of a series in continued proportion", and if any two terms are taken and multiplied, "the product will be a term in the same series". The product also sits at a distance from the first term that is one less than the sum of the distances of the two factors, counted with their end terms included. To prove this, the author picks a term that lies the same number of steps from the second factor as the first factor lies from the start. That count is given as m, so the term to be taken is the one reached by counting m steps from the second factor, which is the term that completes the sum.
The author begins by stating what must be shown: "We have therefore to prove that" the grains of sand can be counted within a stated bound. The argument first needs a property of continued proportion, a chain of terms in which each is the same multiple of the one before. The author states the rule: "Now terms equally distant from other terms in the continued proportion are proportional." Written in powers of a single base term, the product of the term m places from the start and the term n places from the start is the term m plus n places from the start. The author records this as a first result, numbered (1), and it is the tool the rest of the argument depends on.
The author then says "The second result is now obvious", extending the same counting to a term one place further along. The reasoning is that the term at the position of the sum, m plus n, sits at a fixed distance from the base, so the rule carries over without new work.
With these results in hand, the author turns to the number of the sand. The assumption made earlier in the work, the fifth, bounds the size of a poppy-seed against a finger-breadth, taking the seed's diameter as a small fraction of that unit. Because spheres relate to one another in the cube of their diameters, the author concludes "since spheres are to one another in the triplicate ratio of their diameters, it follows that" a sphere one finger-breadth across holds 64,000 poppy-seeds. The number 64,000 is forty cubed, which is why the fortieth-part assumption produces it.
The author then multiplies that count by 10,000, reaching 640,000,000. The number is expressed on the author's own scale of orders, in which a unit of the second order is a hundred million and a unit of the first order is one. On that scale 640,000,000 is six units of the second order together with forty million units of the first order. The steps here are chained: each stage multiplies the previous count, so the bound on the seeds in a finger-breadth sphere is carried forward into the larger estimate that follows.
The argument now works through a table of sizes. The author takes an imagined sphere and gradually increases its diameter, and the key sentence is this:
We now gradually increase the diameter of the supposed sphere, multiplying it by 100 each time.
Because a sphere's volume grows with the cube of its diameter, a hundredfold increase in diameter multiplies the volume by a million, and that single fact drives every row of the table.
The first row sets the diameter at 100 finger-breadths. The author gives the number of grains of sand such a sphere could hold as less than a figure built from the million-fold factor and the previous bound, then states the result in the units of his large-number system: "10,000,000 units of the second order." Every figure is an upper limit, marked with "less than" throughout, so the argument needs only that the true count never exceeds the bound it reaches.
Each row repeats the same move. A million times the previous bound is written as a product of a power of ten from the series and the last term reached, and the resulting term number becomes the next entry. The table records these term numbers rather than writing out every digit, and the sixteenth term of the series stands for the bound reached at the hundred-finger-breadth size. Each step is simple arithmetic. What the method achieves is a way of naming numbers far larger than the ordinary system could express.
Following from the earlier steps, the author draws a further conclusion about the sphere of the fixed stars. The argument rests on a hypothesis taken from Aristarchus: that the earth is to the universe as the universe is to the sphere of the fixed stars. The earlier steps have already shown that the diameter of the universe is less than ten thousand times the diameter of the earth. From the hypothesis it follows that the diameter of the sphere of the fixed stars is less than ten thousand times the diameter of the universe.
Because the volume of a sphere grows with the cube of its diameter, the author then works with the cube of ten thousand. The work multiplies that figure by a thousand units of the seventh order of numbers, and then matches the product against the terms of its own series of powers. Its notation is built so that each term is a far larger power of ten than the last. The chain of comparisons ends at the 64th term, which the text glosses as ten to the sixty-third. The author states the result as a bound: the number of grains of sand that would fill a sphere equal to the sphere of the fixed stars is less than ten million units of the eighth order of numbers.
The work then closes by addressing king Gelon directly. The author writes:
I conceive that these things, king Gelon, will appear incredible to the great majority of people who have not studied mathematics, but that to those who are conversant therewith and have given thought to the question of the distances and sizes of the earth the sun and moon and the whole universe the proof will carry conviction.
The author adds that for this reason the subject was thought not inappropriate for the king's consideration.
The main takeaway is that the number of grains of sand needed to fill the universe, as the author measures it, can be bounded by a finite number, and that ordinary counting words can be extended to name numbers far beyond a myriad. The author frames his key figures as assumptions to be granted, and he treats the sun's size as something he proves later by an observation rather than taking on trust. The open questions the text leaves follow from those assumptions: whether the figures for the earth, the moon and the sun hold, and how sensitive the final bound is to them, since each rests on a chosen limit rather than a measured value. The translation notes that its later steps have been rendered more freely, so readers who want the exact figures should consult the original.

