Proceed to Safety

Notable Properties of Specific Numbers    

First page . . . Back to page 13 . . . Forward to page 15 . . . Last page (page 25)

2431 = 11×13×17

A product of 3 consecutive primes (see also 1001).

A permutation of the digits 1,2,3,4 that is also a multiple of 11 (see 132 for more about this; see also 163 and 101.0979×1019).


At one point in history, as the story goes (see this Numberphile video featuring N.J.A. Sloane) the mathematicians Armand Borel and Freeman Dyson were having a discussion of large prime numbers over lunch, during which Dyson points out that 1111111111111111111 is prime. Borel wants something "more intersting". It turns out that the rather "memorable" number 12345678910987654321 is prime.

This leads to speculation about what other numbers of this form are prime. The next one turns out to be 123456...24442445244624452444...654321, a 17350-digit number with "2446" in the middle. (It is not known what the next-larger one is, though one almost certainly exists.)

2520 = 23×32×5×7 = 5 × 7 × 8 × 9

2520 is the smallest number that is divisible by all the numbers from 1 to 10.

How do you find such a number? It doesn't need to be 1×2×3×4×5×6×7×8×9×10, there are many smaller answers. To consider a simpler example, 12 is the smallest number divisible by 1, 2, 3 and 4 but it is smaller than 1×2×3×4. There are different ways to find the answer, all of which amount essentially to looking at the prime factorisations of all the numbers and keeping the highest exponents of each prime that occur. The simplest way to describe the answer is as follows: Keep the highest power of each prime, and throw away all the rest. For example, in the numbers 1 through 10, the highest prime powers are 8=23, 9=32, 5=51 and 7=71. All the other numbers are composite (e.g. 6=2×3 and 10=2×5) or smaller powers of the primes already mentioned (2=21, 4=22 and 3=31).

This number is also related to 10 in another way: it is 10 times the central number (252) on row 10 of Pascal's Triangle. This is sort of related, because 252 = 10!/(5!×5!); see 5040 and 3628800.


This is 72×1000/25.4, the number of pixels per meter corresponding to 72 pixels per inch. A PNG image (such as a screen shot) created on a computer that uses the "72 pixels per inch" standard (e.g. many Apple systems going back to the original Macintosh) will often have 2835 in the "Pixels per Meter" attrbutes.


See 2834.64566929....


2592 = 25×92, the only 4-digit number of the form ABCD = AB×CD66 (from Jim Cook). The only other (known) similar number is 24547284284866560000000000.

See also 25920 and 13532385396179.


3003 occurs in Pascal's Triangle in rows 14, 15 and 78, because 3003 = 14!/(6!×8!) = 15!/(5!×10!) = 78×77/2. This is related [144] to the fact that 6=F3F4 and 15=F4F5 are two consecutive golden rectangle numbers. The part of rows 14 and 15 where this happens looks like this:

1001 2002 3003
3003 5005

and is the only place where three numbers appear side-by-side in the ratio 1:2:3.

In addition to rows 14, 15, and 78, 3003 also occurs in row 3003, making for a total of 8 appearances. It is the only number yet known to occur so many times. Singmaster's conjecture states that there is a finite upper bound to the number of times any single number appears in Pascal's triangle. Numbers that occur at least 5 times in Pascal's Triangle include: 120, 210, 1540, 3003, 7140, 11628, 24310, 61218182743304701891431482520, ... (Sloane's A3015).

3003 is also a palindrome, and the 77th triangular number. It shares this property with 1, 3, 6, 55, 66, 171, 595, 666, and many others (Sloane's A3098). Note that the index (77) is itself a palindrome. Possibly because of the simple formula Tn = n(n+1)/2, there are also at least two cases with a repunit index: T1111=617716 and T111111=6172882716.


This is 6×7×8×9, which also happens to be equal to the sum 5×6×7×8 + 4×5×6×7 + 3×4×5×6 + 2×3×4×5 + 1×2×3×4. A little investigation shows that this is one of a family of similar patterns: 1+2=3, 1×2+2×3+3×4=4×5, 1×2×3+2×3×4+3×4×5+4×5×6=5×6×7, and so on. The sums/products for the sequence: 1, 3, 20, 210, 3024, 55440, 1235520, 32432400, ... (Sloane's A6963) which counts "planar embedded trees of N nodes".


Magic constant of the famed Parker Square, an arrangement of perfect squares (such as 292, 372 and the ever-popular 12) into a 3×3 grid such that all rows and columns sum to the same thing. It's a bit like a magic square, only a little less magical. They made a website and Twitter account for it, so I suppose now it's a thing.

3121 = 55-4

The solution to the "Monkey and Coconuts problem" (1926 Saturday Evening Post version), which according to Martin Gardner62,63 is "probably the most worked on and least often solved of all" Diophantine equations. Here is the problem (in my words):

Five sailors, stranded on a desert island, spend the day gathering coconuts and then go to sleep, agreeing to divide them up in the morning.
     After a while, one sailor wakes up, concerned that he might not get his fair share. He gives one coconut to the monkey to keep him quiet, then divides the pile into five parts, and finds that it divides evenly. He hides one fifth and puts the other 4/5 back into one pile (the monkey keeps his).
     One after another, each of the other four sailors does the same thing — wakes up, gives a coconut to the monkey, divides the rest into five parts, which comes out evenly, hides one-fifth and puts the remaining 4/5 back into one pile.
     In the morning, the men together divide the remaining pile and find it divides equally into five parts. How many coconuts were there in the beginning?

A little trial-and-error reveals that the answer is probably somewhat large. One can work it out methodically by working backwards from the end: The final pile F must be a multiple of both 4 and 5, so it can be any number in the sequence {20, 40, 60, 80, ...}, the pile prior to that has to have been 5/4F+1, which admits the numbers in the sequence {76, 176, 276, ...}, etc. There are also more elaborate solution techniques, some of which are outlined in an excellent Numberphile video. There is more than one answer, but the smallest is 3121. The general solution for n sailors is nn-n+1 for odd n, and (2n-1)nn-n+1 for even n.64

See also 7.76×10202544.


3151 is a "pentagonal" number by both definitions: the "square plus triangle" definition (OEIS sequence A0326: 0, 1, 5, 12, 22, 35, 51, 70, 92, ..., (3n2-n)/2) and the symmetrical "centred" definition (A5891: 1, 6, 16, 31, 51, 76, 106, ..., (5n2+5n+2)/2). Illustrations of both types are in the 51 entry. There are infinitely many such numbers, increasing geometrically (A128917 : 1, 51, 3151, 195301, 12105501, 750345751, 46509331051, 2882828179401, ...)


This number is prime, and all its digits are individually prime. Such numbers make up OEIS sequence A19546. See also 2357.


3435 = 33+44+33+55, and is the only number in base 10 (apart from the trivial 11=1) formed from the sum of each of its digits raised to its own power. (From van Berkel [206])

This property is similar to properties of 135, 153, 48625, and some of the Friedman numbers.


3456 = 1!×2!×3!×4!×3!×2!×1! which is somewhat related to 34560. It is also the determinant of this matrix of factorials: [ 1! 2! 3! 4! ] [ ] [ 2! 3! 4! 5! ] [ ] [ 3! 4! 5! 6! ] [ ] [ 4! 5! 6! 7! ]

3456 is twice 1728, and has a notable pattern of ascending digits. The digits 3,4,5,6 are reminiscent of the sums-of-cubes identity 33+43+53=63 (see 216), and it is a multiple of this sum, so it can be related to these cubes in other ways, such as 13×23×(33+43+53+63)=3456.

It is 8 times 432 which has three digits in descending sequence, and that is 8 times 54 which has two digits descending.

It also has the following cute relationship to 27: 27 × 27 = 3456, and 2+7+2+7=3+4+5+6.

After this tweet by a couple of the creative folks at Numberphile, I have jokingly called it the "Grime-Parker Carriage Number", in allusion to the Hardy-Ramanujan Taxicab number.

See also 4567.


The only other known Wieferich prime (see 1093).

3581.875516... = π(e-1))

This is πe, where is the lower-valued form of the hyper4 operator. See also πe, 4979.003621... and 11058015.34616.


The number of seconds in an hour (60×60), the number of fingers in a cord (units of length), the number of shekels in a talent (units of weight), the number of years in the (Babylonian) long "saros", and the product of squares of the simplest Pythagorean triangle (32×42×52). All of these were significant to the Babylonians.

Although it is not a divisibility record-setter itself, 3600 is the square of the popular record-setter (and base of the Sumerian/Babylonian number system) 60.

The division of the hour into 60 minutes or 3600 seconds is the most familiar relic of the old Sumerian base-60 numbering system. It also survived in the divisions of angles into degrees, minutes (short for "minute divisions"), seconds, thirds14, and so on. The division of an hour into 3600 parts also happens to be convenient and useful. The hour had been long established as 1/12 part of the daylight period (to such an extent that, in many cultures, the length of the hour increased and decreased with the seasons!). The pace represented by 3600 beats per hour arises naturally because it is close to the frequency of a human heartbeat; most people feel their heartbeat when resting quietly. Since hours and heartbeats were already pretty well established, it was useful to choose a number that was pretty close to the right ratio but also was arithmetically convenient to work with. 3600 was by far the best choice.


A triangular Mersenne number, and the largest. See 91.


The number of ones left on the tape by the current best-known Turing busy beaver with 5 states; see 47176870 for a full description.


The first Fibonacci number with a prime subscript that is not itself prime: F19=4181=37×113.


The side-length of the first-published example of a squared square, a square divided up into smaller squares, all of different sizes and each with integer-length sides. It was assembled from squared rectangles by Roland Sprague in 1939. Numberphile has a video on the topic, Squared Squares. See also 112.


See 6.397665...×1043.


A prime number with consecutive increasing digits. See also 2357, 3257, 3456, and 23456789.


4879 is a Kaprekar number under the most relaxed Kaprekar rules, in which the point of division can be in a place other than the number of digits in the original number. 4879^2 is 23804641, but instead of dividing it into 2380+4641, it is divided into 238+04641 = 4879. Even though the original number 4879 has four digits and its square has eight, the square is divided unevenly giving three digits to the "left half" and five digits to the "right half" 04681. Under the more strict Kaprekar rules, the division always happens in the place corresponding to the number of digits in the original number.


In 1875 Edouard Lucas considered the problem: "A certain number of cannonballs can be arranged in a square on the ground, and can also be stacked into a square pyramid. How many are there?"

The number of cannonballs in the pyramid is the sum of consecutive squares 12+22+32+...+x2. This problem is formally stated as the diophantine equation x(x+1)(2x+1)=6y2. The only known answers were the trivial case x=y=1 and the intended answer to the problem, x=24, y=70, with y2=4900 cannonballs.

Lucas conjectured that there were no other solutions; it was proven by Watson in 1918, but using a rather complicated method. Interest in the original problem continued, and in 1985 D.G. Ma published the first proof using entirely elementary means.

Amazingly, the fact that 24(24+1)(2×24+1)=702 is connected to 26-dimensional Lorentzian space-time, the Leech lattice, and the Monster group. The connection appears96 to be evident in Conway and Sloane [150] and is made more explicit by Borcherds in his 1984 Ph.D. thesis [170].

Mike Hill, instructor of an abstract algebra course at University of Virginia, describes it this way97 in course notes on the Monster group:

A path on which the time-distance is always zero in a higher dimensional (> 4) space-time (Lorentzian space) yields a perpendicular Euclidean space of two dimensions lower.
   ex. 26-dimensional Lorentzian space yields the 24-dimensional Euclidean space which contains the Leech lattice.

The Leech lattice contains a point (0,1,2,3,4,...,24,70)
Time distance from origin in Lorentzian space
      0 = 02 + 12 + 22 + ... + 232 + 242 - 702
This point lies on a light ray through the origin

Borcherd said a string moving in space-time is only nonzero if space-time is 26-dimensional.

A similar description is given by Ronan [198], page 224.

See also 3.377×1038.

4979.003621... = e(e(π-1))

This is eπ, where is the lower-valued form of the hyper4 operator. See also eπ 3581.875516... and 4341201053.37.


5000 ("five thousand") is the largest number whose name has no repeated letters, and the longest such number-name (with 12 letters). See also 34, 84, 1005, 1025, 1084, and 1000000000008020.


5040 = 7! = 7×6×5×4×3×2×1 = 10×9×8×7. There are other numbers with a similar property (see 720 and 3628800)44.

5040 is also divisible by all the numbers from 1 to 10, and by 144 and several other useful numbers. Plato44 cited this as a reason for 5040 being the ideal number of citizens in a state.

5040 is also a divisibility record-setter (also called "highly composite" or "antiprime"). Numberphile used it as the subject of their video 5040 and other Anti-Prime Numbers which explains how the number of factors of a number is computed directly from that number's prime factorisation.

It is a member of a "chain" starting with 293318625600 and iterating the "number of divisors" function A0005: 293318625600 has 5040 divisors; 5040 has 60 divisors, 60 has 12 divisors, etc.; each of these are highly composite.


5041 = 7!+1 = 712. This is the highest known case of a square which is one more than a factorial. The other cases are 25 = 4!+1 = 52 and 121 = 5!+1 = 112. 5041 is my favourite, because it's the largest one and because of the recurrence of the digits 7 and 1 letting you write: √7!+1 = 71.

Looking for larger numbers with the property N! = A2+1 for integer N and A is rather hard, see Brocard's problem. It is known that the next solution, if any, is for N>4×109. There are several similar special properties of numbers (for examples, see 39, 89, 91 and 51381) where the distribution falls off so quickly that seems likely that there are only a finite number of numbers with the property, but for which it's difficult to prove. In this case for example, the odds of N! + 1 being a square are about 1 part in 2×√N!, assuming there is no special relationship between the distribution of squares and factorials. Since the factorials grow very quickly, the infinite sum

SUM [ 1 / 2×√N! ]

converges very quickly, and in fact it's a bit of a surprise that there are as many as three solutions for N! + 1 = M2. The fact that there are three suggests that the distribution of factorials and squares might have a relationship — but one should be wary of the Strong Law of Small Numbers (again see 91). See also 40401.


An approximation of the number of years in the "Mayan long count" calendar, according to some theories. This calendar began at a date that is equivalent to 3114 BC in the Gregorian system; it counts days and has an integer value expressed in a 5-digit mixed-base system using bases 13, 20, 20, 18 and 20 in each of the five places. The rightmost two places, counting in cycles of 18×20, correspond to a 360-day year, and the other three places count "years" (periods of 360 solar days). The total number of days is 1872000. Expressed more accurately, the "5126 years" is 13×202×18×20/365.242189670 = 5125.3662719. See 2012.

The use of 13 as the base of the highest place in this calendar's counting system is uncertain, and it could just as likely be 20, which would lead to a "long count" of 18×204/365.242189670 or about 7885 years.

See also 18980 and 550420.

5280 = 25×3×5×11

Number of feet in a mile. 5280 is close to the Roman mile defined as 1000 paces, or 5000 feet, but was adapted to accommodate other units of length including the furlong (660 feet) and chain (66 feet). These numbers (66, 660 and 5280) are all multiples of 11 because of their relation to the old unit of length called a rod, also called a perch or pole, which is the same length as an old farmer's tool called an "ox goad"92, used in plughing. However, the origin appears93 to go back to Denmark and Prussia sometime before the 11th century; evidence for this fact lies in the fact that the English rod agrees with the old Danish fod and the Prussian Rheinfuss to within less than half an inch, or 0.19% of the total length.

See also 1852 and 63360.


The length (counted in moves for White) of the longest possible chess game, assuming that each player would exercise their option to declare a draw under the 50-move draw rule. Such a game involves stalling for as many turns as possible, then moving a pawn or capturing a piece (actions that reset the 50-move counter). The number is calculated as "2 players × (6 pawn-moves per pawn to reach promotion × 8 pawns + 8 captures of opposing promoted pawns + 3 captures of other opposing pieces excluding the 4 that were captured by pawns on their way to promotion) × 50 moves between events - 4 times that the side executing "events" switches × (1/2 move) = 2×(6×8+8+3)×50 - 4/2 = 5898. It was described by Matthew van Eerde on Math Forum in 1997.

See also 8848, 10120.


This is the ending value you get using the "Kaprekar transformation" on 4-digit numbers. See 495 for details.


The number of synapses (junctions between neurons) in the nematode worm C. elegans. See 959.


This is 34 = 812, and is the (semi-)magic constant of this semi-magic square, an arrangement of perfect squares into a 3×3 grid such that all rows and columns sum to the same thing (but not the diagonals), and each number appears exactly once:


Finding one like this but for which the diagonals also work is an open problem. See also 3051.

6585.3213142 ≈ 223 × 29.530588853

(the saros cycle in days)

Number of mean solar days in the saros.

See also 19756 and 161178.


6720 = 4×5×6×7×8, and is the magic constant for this "multiplicative" complete magic square:

  1  6 20 56
40 28   2   3
14  5 24   4
12  8   7 10

which is built on the principle of doing an elementwise multiplication (Hadamard product) on the following components:

1 2 4 8   1 3 1 1   1 1 5 1   1 1 1 7
8 4 2 1 1 1 1 3 5 1 1 1 1 7 1 1
2 1 8 4 o 1 1 3 1 o 1 5 1 1 o 7 1 1 1
4 8 1 2    3 1 1 1    1 1 1 5    1 1 7 1

each of which satisfy the row, column, diagonal, and 2×2 subsquare requirements, but with repeated numbers. See also 216, 1514, and 6227020800.


The smallest number with a persistence of 6: 6×7×8×8 = 2688; 2×6×8×8 = 768; 7×6×8 = 336; 3×3×6 = 54; 5×4 = 20; 2×0 = 0.

6939.60160373 = 19 × 365.242189670

(the metonic cycle)

Number of mean solar days in the metonic cycle of 19 tropical years or 235 synodic months. Its integer approximation (6940) can be used as the basis of a lunisolar calendar that repeats every 19 years, and with some effort you can also make the months fairly regular. For example, every year can have 12 months each of which has a specific name, 6 with 29 days and 6 with 30 days; 7 out of 19 years have a 13th month of 30 days, and 4 of those 7 years also add an extra day to one of the normal 29-day months. Other similar systems are possible but all solutions have equal complexity; actual lunisolar calendars (like the Hebrew calendar) are more complex but achieve greater accuracy.

Although it is complex, a lunisolar calendar has very strong practical motivations. The importance of the solar day is obvious; the tropical year is important to anyone living in a climate with seasons. And the synodic month tells us when it is possible to see at night by the light of the moon, and what time of day the high and low tides will take place. These things are important even in modern urban society. Under a lunisolar calendar you can agree to meet outdoors at 8PM every month on the 15th of the month and know that there will be moonlight (weather permitting), or you can agree to go fishing every 10th of the month at 5AM and know there will be a low tide (assuming that's how the tide lines up at your location).


The integer approximation of 6939.60160373. There would be this many days in 19 years, if the 19-year metonic cycle were exact. See also 20819 and 161178.


See 74994632695013731827926060475067499.

7776 = 65

One of the more memorable small powers. While I was 10 and 11 years old I memorised integer exponents of integers just for fun. I still know all of these by heart:

2 4 8 16 32 64 128 256 512 1024 all the way up to 131072
3 9 27 81 243 729 2187
4 see 2
5 25 125 625 3125 15625
6 36 216 1296 7776 46656
7 49 343 2401
8 see 2
9 see 3
10 trivial
11 121 1331 14641
12 144 1728

7776 is also a Kaprekar number for 5th powers: 77765=28430288029929701376, and 2843+0288+0299+2970+1376=7776.

(And better still, 7+7+7+6 = 27.)


The Boolean Pythagorean triples problem asks if all integers that belong to a Pythagorean triple can be assigned a colour (either red or blue) in such a way that all Pythagorean triples have at least one red and one blue. The answer is no, as stated succinctly by the following (true) statement:

The set {1, 2, 3, ..., 7824} can be divided into two subsets, such that neither subset contains three numbers that are a Pythagorean triple; but this cannot be done for the set {1, 2, 3, ..., 7825}.

The proof of this was developed by Marijn Heule, Oliver Kullmann and Victor Marek. It involves a massive amount of computation which eventually produces a "proof" about 2×1014 characters long.


This is the 1000th prime number, and has all odd digits making it somewhat more psychologically random. It is perhaps for this reason that Veritasium worked it into his great video on psychologically random numbers like 17 and 37.

7920 = 8×9×10×11

7920 is the ratio between 11!=39916800 and 7!=5040, and the ratio 7920/5040 = 11/7 = 1.571428571428... is kind of an approximation to π/2.


7980 = 15×19×28, the product of three numbers that have an important relation to calendars in the Roman, Byzantine and Christian worlds. 28 years is the "solar cycle", 4×7, the number of years it takes before any date falls on the same day of the week again (that's 4 years per leap-year cycle, and 7 days per week). 19 is the length of the metonic cycle, see 19 for more. 15 years is the "indiction cycle", a period related to certain phenomena such as taxation. Since all three numbers are relatively prime to each other, the least common multiple is 7980. The "Julian day number", used in astronomy, is based on a system proposed in 1583 just after the adoption of the Gregorian calendar. Julian day 1 is January 1st 4713 BC. The year 4713 BC happened to be the most recent time that all three cycles (the 15- 18- and 28-year cycles) were aligned. It is convenient primarily because it pre-dates all recorded history, even in China, Egypt, Greece and Mesopotamia.


8000 = 203 = 113+123+133+143, the smallest cube that can be expressed as the sum of 4 consecutive cubes. See also 216. See also 246924.


This number has a property related to vampire numbers: if you take the first two digits and multiply by the other two, the product has the same four digits (but in a different order): 81×27=2187.

See also 8712.

(Personal: For a while during my childhood the numbers 7, 27 and 127 were my favourite 1- 2- and 3-digit numbers. I have since forgotten why the properties of 127 appealed to me, but I suspect it was partly because it ends in "27". If I had continued the series, I probably would have picked 8127, because it is divisible by 7 and by 27, and consists of the cubes 8, 1 and 27 strung together; also the "81" is 3×27.)


The fourth perfect number, and of the form 2(p-1)(2p-1) where p=7.

This number appeared along with 8191 and 8208 in the 2006 Simpsons episode "Marge and Homer Turn a Couple Play" [230]. See also 6.397665...×1043.


A Mersenne prime, 213-1 = 8127.

This number appeared along with 8128 and 8208 in the 2006 Simpsons episode "Marge and Homer Turn a Couple Play" [230]. See also 6.397665...×1043.


An Armstrong or "narcissistic" number, 8208 has 4 digits and is equal to 84+24+04+84.

This number appeared along with 8128 and 8191 in the 2006 Simpsons episode "Marge and Homer Turn a Couple Play" [230]. See also 153, 6.397665...×1043.

8712 = 23 32 112

The smallest number that is not a palindrome, and is divisible by the number you get when you reverse its digits: 8712 = 4×2178. The next number like this is 9801=992. The numbers with this property comprise OEIS sequence A31877 and their factors are A8919. (Thanks to Tavi Laiu for this one). See also 1089. Another related property: 8712 × 2178 = 664.


The length (counted in moves for White) of the longest possible chess game, assuming that players cooperate in not exercising their option to declare a draw under the 50-move draw rule, but still avoid invoking the mandatory 75-move draw rule. The calculation is the same as for 5898, but substituting 75 for 50: 2×(6×8+8+3)×75 - 4/2 = 8848. See also 10120 and 1012500.


9000 was used as the starting value in an xkcd forum discussion/contest "My number is bigger!":

I'll start us off with nine thousand. Because we all know that anything over nine thousand is f***ing unimaginably huge.

This was a reference to the "over 9000" meme, see 9001.


An internet meme: "Anything over 9000 is L337". The origin seems to be a Dragonball Z line 八千以上だ! "it's higher than 8000!" mis-translated into English as "It's over 9000!!!" (see Dragon Ball wiki,  Over 9000), and then spread through videogame culture. There is no relation to HAL 9000 or ISO 9001. See also 9000 and 900901.

See also 573, 73735963, and 1597463007.


Grafting numbers

9797 has the curious property that its square root includes itself right after the decimal point: √9797 = 98.9797959181... Any number of the form (10n-3)×(10n+1) has this property: 77, 9797, 997997, ...

These are a subset of a more generally defined class of numbers with very similar properties, for example √98 = 9.89949493661...; √77327 = 278.077327374...; etc. Such numbers are called "grafting numbers", a name given by Matt Parker. Watch this Numberphile video for his account of how he discovered and investigated them.

The Grafting numbers are OEIS sequence A232087: 0, 1, 8, 77, 98, 99, 100, 764, 765, 5711, 5736, 9797, 9998, 9999, 10000, 76394, 77327, 997997, 999998, 999999, 1000000, 2798254, 7639321, 8053139, ...

See also 0.763932....


9801 = 992 and has a curious reciprocal: 1/9801 = 0.00010203040506070809101112131415161718192021222324252627... For more on this see my separate article Fractions with Special Digit Sequences; see also 89, 99.9998, 199, 998, 9899, and 997002999.

9801 = 9×1089, a multiple of its "reversal". See 1089 and 8712.


9899 = 1002-100-1, and has a reciprocal containing Fibonacci numbers: 1/9899 = 0.00 01 01 02 03 05 08 13 21 34 55... For more on this see my separate article Fractions with Special Digit Sequences; see also 89, 99.9998, 199, 998, 9801, and 997002999.

10000 = 104

(a myriad)

(basis of the classical Greek large number system)

From the 3rd century BCE to as late as the 12th century, in Greek, Coptic and Armenian texts, letters of the alphabet were sometimes used to represent numbers122:

Α=1 Β=2 Γ=3 Δ=4 Ε=5 ς=6 Ζ=7 Η=8 Θ=9 Ι=10 Κ=20 Λ=30 Μ=40 Ν=50 Ξ=60 Ο=70 Π=80 Ϙ=90 Ρ=100 Σ=200 Τ=300 Υ=400 Φ=500 Χ=600 Ψ=700 Ω=800 Π=900    ͵Α=1000    ͵Β=2000    ͵Γ=3000    ͵Δ=4000    ͵Ε=5000    ͵Δ=6000    ͵Ζ=7000    ͵Η=8000    ͵Θ=9000

The necessary letters are combined to make a quantity, so for example the number 8127 would be ͵ΗΡΚΖ. For numbers above this they used a capital letter mu (Μ) to represent 10000, whose name in Greek is myriad. To distinguish it from Μ=40, a small alpha α was written above the Μ. The precise way in which larger numbers were handled differed through the centuries in ancient Greece. In the most ambitious system, a number (like ͵ΗΡΚΖ) was written in small letters (in this case ͵ηρκζ) directly above a letter Μ to represent 10000 raised to that power; see 1040000. However, in much more common usage the small number placed above the letter Μ merely multiplies the value; see 108.

10000 is 万 in China (pronounced wàn) and Japan (pronounced man); the traditional form is 萬. Myriads and powers of a myriad are used in several east Asian systems for naming large numbers; see 100000000 and 104096 and my article on large numbers in Japanese. See also the Knuth -yllion number names.

See also 100000, 500000 and 10000000.


Generalised Fermat numbers

10001 is a "generalised Fermat number", of the form a2n + 1, with a=10 and n=2. They include the normal Fermat numbers (when a=2 or 4), OEIS sequence A78303 when a=6, A152581 when a=8, A80176, A178426, A152587, etc. If a is odd they are even (and thus composite); when a is even they have some of the properties of the normal Fermat numbers, such as being prime a little more often than one might expect, or having reasonably large prime factors. For example, 10001 is 73×137, and 1016+1 = 353×449×641×1409×69857. The "generalised Fermat primes" would be all primes expressible as a2n + 1, which is the same as OEIS sequence A2496, primes of form n2 + 1.


This is 640320/64 and appears in the Chudnovsky series approximation of pi.


10080 = 60 × 24 × 7, the number of minutes in a week. 10080 is also a divisibility record-setter with 72 divisors. Thus, there are 72 different ways to divide a week into equal parts that are whole multiples of minutes. 10080 is also 2×7!, see 40320 and 604800.

In France just after the revolution they switched to the now-standard system of weights and measures based on powers of 10 (SI, systeme internationale, or the "metric system"). They also set up (for a while) a calendar involving a 10-day week, and dividing the day into powers of 10. It didn't last long, and France switched back to the Gregorian calendar a few years later.

But there have been other times since then that people have considered the idea of measuring time in powers of 10. Clearly this "metric time" idea has some appeal, because there are dozens of web sites about it (do a search for "metric time"). Perhaps best-known is Swatch's "Internet time", based on the division of the day into 1000 parts called "beats" (with 86.4 seconds per "beat"). For subdivisions of the day it is easy, the hours and minutes we have now are just arbitrary divisions and there is no conflict with the other important environmental cycles.

When considering a metric replacement for the synodic month and tropical year, clearly there is no good practical solution. These are all very important physical cycles and have strange ratios that cannot be changed (at least not yet :-) and cannot easily be adapted to work with powers of 10.

Personally, I think the existing standard time system is just as cool as any 10-based system would be, because of the utilitarian properties of the factor record-setters and cute things like the minutes in February and the 86400000 property. However, if you insist on setting up a "metric time" system, you could do worse than to take advantage of the fact that a minute and a week are almost at an exact ratio of 10000. Such a system would leave the definition of "week" unchanged and define a special "minute" which is exactly 7/10000 of a day (which comes out to exactly 60.48 "standard" seconds). Then all weekly events would always start at the same minute-mark, which would be a 4-digit number from 0000 to 9999. You'd still have to deal with the fact that the number of weeks per year is odd (the fact that it's close to 100/2 makes this a little more bearable), the number of minutes per day is odd, and there is no suitable "month".

See also 3628800 and 86400000.


Many folks know 1/81 for its repeating decimal. Subtracting each digit from 9, we get something similar but with descending digits and leaving out the 1 instead of the 8: 80/81=0.987643209876432098764320.... Since the numerator 80 is (2,5)-smooth, we can convert this to a reciprocal of an integer: repeatedly multiply by 2/2 or 5/5 as needed until the numerator is a power of 10, then just replace the numerator with a 1:

       = 400/405 = 2000/2025 = 10000/10125

(suggested indirectly by reader danny12)

10201 = 1012

The 2nd row of Pascal's Triangle shows clearly in the 2nd power of 101. The pattern continues in higher powers, and goes further than for the powers of 11. See 14641 for more.


10744 is 143+203, and 10745 is 173+183. Like 854, the two sums fit the form (2t2-t-1)3+(2t2+t-1)3 and (2t2-1)3+(2t2)3; in this case t=3. (from Vladimir Shevelev) See also 18426689288.

10800 = 180×60

The number of arc-minutes in 180 degrees, and thus the idealised length of a meridian (see 20003931.4585) in nautical miles (see 1852.216). See also 108 and 10080.


Inder J. Taneja, describes [233] how to construct numbers from chains of the digits 1 through 9 with arithmetic operations, digit-concatenation (assuming base 10) and parentheses, for example:

7926 = (1 + 2(3+4)) × 56 + 78 × 9
7926 = 9 × 876 + 5 + 4 + 32 + 1

The digits may only be in ascending or descending order, as shown here, and all nine digits must be used. The first number he was not able to form this way was 10958. This is described by Numberphile in two videos: the problem statement, and Matt Parker's solution using a generalised concatenation operator.


The name of the worker who is befriended by Freder in the 1927 movie Metropolis. Also a palindrome, and strobogrammatic.


In 2011 November, this was identified as the "smallest uninteresting number" in an episode of QI, based on its being the smallest (positive) integer not appearing in the listed terms of any sequence in the OEIS. It is important to note that some sequences, such as A0027, contain every integer — so this property of 12407 was only possible because each entry in the OEIS gives only a few lines of initial terms for each sequence. There is a lot more about this in [218].

See also 39, 51, 74


A few numbers are perfect, and others form amicable pairs. 12496 has the property that when you add its factors (not including itself), then do the same thing to the result, and repeat, you get back to 12496 after 5 steps. (The sequence is 12496 → 14288 → 15472 → 14536 → 14264, then back to 12496.) See also 14316.


A few numbers are perfect, and others form amicable pairs. 14316 has the property that when you add its factors (not including itself), then do the same thing to the result, and repeat, you get back to 14316 after 28 steps. (The sequence starts 14316 → 19116 → 31704 → ... and gets as high as 629072 at one point: see OEIS sequence A72890.) See also 12496.

14641 = 114

(Pascal's triangle)

The smaller powers of 11 (121, 1331 and 14641) give rows of Pascal's Triangle. Pascal's Triangle is a rather useful table of numbers (called binomial coefficients) that is formed by continually adding numbers in a cascade starting from a single 1, as shown here:

row 0: 1
row 1: 1 1
row 2: 1 2 1
row 3: 1 3 3 1
row 4: 1 4 6 4 1
row 5: 1 5 10 10 5 1
row 6: 1 6 15 20 15 6 1
row 7: 1 7 21 35 35 21 7 1
row 8: 1 8 28 56 70 56 28 8 1
row 9: 1 9 36 84 126 126 84 36 9 1
etc etc etc

Each is the sum of the (one or) two numbers above it.

These numbers have many applications because they count "combinations", such as "how many combinations of 3 colours can you make if you have 5 colours to choose from?" (the answer is the 3rd number in row 5 of the triangle: 10). Another question with the same answer is, "If you flip a coin 5 times, how many ways are there to end up getting heads exactly 3 times?" (the answer is also 10).

A general formula that gives the value of item N in row M is:

NCM = M!/((M-N)!×N!)

For example, 3C5 = 5!/((5-3)!×3!) = 120/(2×6) = 10. However this formula gets rather impractical to use when M is large, even when the value being computed is pretty reasonable. For example, consider 3C143=143!/(140!×3!). 143! has 248 digits, and 140!×3! has 242 digits — so you probably couldn't even get the answer on a calculator, as the calculator would overflow. However, most of the 143! in the numerator cancels out with the 140! in the denominator, leaving 3C143 = 143×142×141/3! = 477191, which you could even calculate by hand if you needed to.


16763, 16787, and 16811 are consecutive primes with the same spacing (24) between each. This is the first evenly-spaced prime trio with a spacing greater than 12 (the record set by 199). See also 47 and 69593.

First page . . . Back to page 13 . . . Forward to page 15 . . . Last page (page 25)

Quick index: if you're looking for a specific number, start with whichever of these is closest:      0.065988...      1      1.618033...      3.141592...      4      12      16      21      24      29      39      46      52      64      68      89      107      137.03599...      158      231      256      365      616      714      1024      1729      4181      10080      45360      262144      1969920      73939133      4294967297      5×1011      1018      5.4×1027      1040      5.21...×1078      1.29...×10865      1040000      109152051      101036      101010100      — —      footnotes      Also, check out my large numbers and integer sequences pages.

Robert Munafo's home pages on AWS    © 1996-2024 Robert P. Munafo.    about    contact
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Details here.