Fractions and Decimals

This page is about converting a fraction (i.e. a ratio of two numbers, also called a rational number) into a decimal fraction and the patterns that occur in such a decimal fraction. It is interactive and you can use the calculators on this page to investigate fractions for yourself to many decimal places. No special knowledge beyond decimals and division is required.

A quick look at what's on this page ...

Abbreviations used on this page:

The calculators on this page require JavaScript but you appear to have switched JavaScript off (it is disabled). Please go to the Preferences for this browser and enable it if you want to use the calculators, then Reload this page.
Contents of this page
The You do the maths... icon means there is a You do the maths... section of questions to start your own investigations. The calculator calculator icon indicates that there is a live interactive calculator in that section.

Changing a Fraction into a Decimal number

Converting a fraction to a decimal is just a division operation. So the fraction 1/2 means 1÷2.
When we do the long division (or use a calculator!) we find 1÷2=0·5. This fraction was easy - just one digit and we are done.

Some examples and patterns

People have always been fascinated by the patterns in decimal fractions, trying to find the order in the seeming chaos and intrigued by the sequences.

When we make a table of the first few reciprocals of the numbers 2,3,..., that is when we turn the whole numbers upside down: from 2=2/1, 3=3/1, 4=4/1,... to 1/2, 1/3, 1/4, ... we get the following:

StoppingEndless
1/2 = 0.5
1/3 = 0.3 3...
1/4 = 0.25
1/5 = 0.2
1/6 = 0.16 6 6...
1/7 = 0.142857 142857...
1/8 = 0.125
1/9 = 0.1 1...
1/10 = 0.1
1/11 = 0.09 09 ...
1/12 = 0.08 3 3...
1/13 = 0.0 769230 769230...
StoppingEndless
1/14 = 0.0 714285 714285...
1/15 = 0.06 6 6...
1/16 = 0.0625
1/17 = 0.0588235294117647 0588....
1/18 = 0.05 5 5...
1/19 = 0.052631578947368421 0526...
1/20 = 0.05
1/21 = 0.047619 047619...
1/22 = 0.045 45 45...
1/23 = 0.0434782608695652173913...
1/24 = 0.0416 6 6...
1/25 = 0.04

Three types of decimal fractions

  1. In the table we can see that some decimal fractions stop after a few decimal places - those in the left-hand columns - such as 1/2, 1/4, 1/5, 1/8.
    These are called terminating decimals.

    Their denominators are 2, 4, 5, 8, 10, 16, 20, 25, 32, 40, 50, ... A003592

  2. Others become an endlessly repeating cycle of the same digits - those in the right-hand columns - such as 1/3, 1/7, 1/9.
    Their denominators are called recurring (or repeating) decimal fractions. Decimal fractions that are purely a collection of digits that repeat from the beginning, such as 0.3 which is just 3 repeating for ever and 1/7 which is 142857 endlessly repeated.

    These are called purely repeating decimal fractions.
    Their denominators are 3,7,9,11,13,19,21, ... A045572 and are all the numbers ending in 1, 3, 7 or 9.

  3. The rest start off with a fixed part, a finite series of digits before they too eventually settle down to an endless repetition of the same digits, the recurring part for example:
    1/6 = 0.1666666...
    which begins 0.1 and then cycles 6 indefinitely
    1/12 = 0.0833333...
    which starts 0.08 before it too starts to repeat 3 for ever.
    These are also called mixed recurring decimal fractions and are the rest of the numbers that are neither denominators of purely repeating fractions nor denominators of terminating fractions:
    Their denominators are 6, 12, 14, 15, 18, 22, 24, 26, 28, 30, 34, ... A105115

At first it is surprising that every fraction fits into one of these three categories:

The decimal fraction of every proper fraction is either terminating or else it becomes recurring.

To see why we have just two types of decimal fraction: terminating or repeating, think about what happens when you try to compute n/d as a decimal fraction: d ÷ n.
Here is the division process for 1/4 and 1/6:
1/4 = 1 ÷ 4
 .  2 5
4 )1 . 0 0
 8
2 0
2 0
0
 
1/6 = 1 ÷ 6
 .  1 6
6 )1 . 0 0 0
  6
4 0
3 6
40
...
Two things can happen in this process:
either at some stage d divides exactly into a number in the division process
and so the division stops because the remainder is 0.
We have a terminating decimal fraction as in 1/4 above.
1/4 = 0.25
or we find a remainder which is the same as an earlier remainder.
so the division process would continue with the same divisors and remainders as when we first found that remainder and then this cycle would repeat endlessly.
We have found a repeating cycle as in 1/6 above.
1/6 = 0.16666666....
Note that we must stop OR else get into a cycle because when we divide by d there are only d different remainders: 0, 1, 2, ..., d-1 so after at most d-1 places, we will find an earlier remainder is repeated and then we have a cycle.
When we find a repeating cycle, it depends if the repeating remainder was the same as the first remainder in which case we have a purely repeating cycle, or if it was a later remainder that was repeated, after some fixed number of digits before the repeating part.

We could simplify this even further by saying that all terminating decimals end with the infinite cycle of 000000... so that every proper fraction is a recurring decimal !

Patterns in recurring decimals

If we take all the fractions with the same denominator, that is, the lower number in a fraction, we can find some amazing patterns too. The first and simplest are the sevenths, the ninths and the elevenths:
1/7 = 0.142857 142857 142857 ...
2/7 = 0.285714 285714 285714 ...
3/7 = 0.428571 428571 428571 ...
4/7 = 0.571428 571428 571428 ...
5/7 = 0.714285 714285 714285 ...
6/7 = 0.857142 857142 857142 ...
    1/9 = 0.1111111...
2/9 = 0.2222222...
3/9 = 0.3333333...
4/9 = 0.4444444...
5/9 = 0.5555555...
6/9 = 0.6666666...
7/9 = 0.7777777...
8/9 = 0.8888888...
    1/11 = 0.0909090909...
2/11 = 0.1818181818...
3/11 = 0.2727272727...
4/11 = 0.3636363636...
5/11 = 0.4545454545...
6/11 = 0.5454545454...
7/11 = 0.6363636363...
8/11 = 0.7272727272...
9/11 = 0.8181818181...
10/11= 0.9090909090...

Notation for the recurring part of a decimal fraction

Mathematicians use several notations to indicate which of the digits in a decimal fraction are in the repeating part (the period or recurring part or cycle):
  1. a dot is put over the first and last digits in the recurring sequence (or sometimes over each of the digits in the period)
    This notation goes back at least to Robertson (1768)
    ..
    1/7 = 0.142857 142... = 0.142857
    ..
    3/44 = 0.06818181... = 0.0681
    .
    2/3 = 0.66666...= 0.6
  2. a line is drawn over the repeating part
    1/7 = 0.142857
    3/44 = 0.0616
    2/3 = 0.6
Both of these are a little awkward on web pages and in the output from computer programs and calculators, so an alternative is also used:
  1. bracket the recurring part with square brackets [ and ]
    1/7 = 0.[142857]
    3/44 = 0.06[16]
    2/3 = 0.[6]
On this page and in the Calculators we will use the [ ] notation .

Here is a question to test your understanding of the bracket notation:

/ You do the maths...

  1. Which of these decimal fractions is not the same as 0· 123 123 123 ...? [Press the button to check your answer.]:
    1. 0.[123] Wrong! This is 0. 123 123 123 123 ...
    2. 0.123 [123] Wrong! This is another way of writing 0. 123 123 123 123 ...
    3. 0.1[231]> Wrong! This is 0. 1 231 231 231 231 ... which is the same
    4. 0.1[23123] Right! Correct! This is 0. 1 23123 23123 ... and is not the same as 0. 1231231231231...
    5. 0.12[312] Wrong! This is 0. 12 312 312 312 ... which is the same

People have suggested that all fractions are recurring ones because they all end with 000000... or they can end with 99999999... . And anyway, is 0.499999... the same as 0.5 = 0.5000000... or not?
The answer is "Yes, they are the same!" but here is a longer explanation if you need more convincing...

Let's examine these two special periods, [0] and [9]:-
Aren't all fractions recurring? It's really a matter of taste as both arguments are correct.
Such decisions are made, choosing one as the preferred method, so that we can all conveniently talk the same mathematical language. These choices are called conventions.
The same is true when deciding on which side of the road to drive. It is a convention in the UK that we drive on the left, but the convention in France is to drive on the right. So long as you go with the convention when driving in Britain and go with the other convention when in France, then there is no problem. But make sure you know which convention is being used in any other country!

/ You do the maths...

Use the interactive calculator following these questions to help you answer them:
[This calculator can give as many decimal places as you like, unlike an ordinary hand-held calculator which often only gives you 8 or perhaps 12 decimal places.]
  1. Convert the following fractions to decimals: 1/7, 2/7, 3/7, 4/7, 5/7, 6/7.
    What do all the 6-digit cycles of these 7th fractions have in common ?
    They all have the same six digits in their repeating part, starting at different points in the cycle:
    1/7 = 0.[142857]
    2/7 = 0.[285714]
    3/7 = 0.[428157]
    4/7 = 0.[571428]
    5/7 = 0.[714285]
    6/7 = 0.[857142]
  2. Is this true of the eigths? Try all the fractions from 1/8 to 7/8. No!
  3. Find another number, N, all of whose decimal fractions 1/N, 2/N, 2/N, ... are made from the same cycle of digits as we found for 1/7.
    Hint: there are two more with N < 20
    n/17 has a period of 16 digits all of which are in the same cycle as
    1/17 = 0.[0588235294117647]
    Similarly n/19 with the cycle being
    1/19 = 0.[052631578947368421]

Fraction to Decimal Calculator

Fraction to Decimal C A L C U L A T O R



 decimal places
in base
R E S U L T S



Which decimal fractions terminate and which recur?

How to find the length of a terminating decimal fraction

If the proper fraction n/d is in its lowest terms then it terminates if and only if d is divisible by 2 or by 5 or by both and is not divisible by any other prime number.
This is because in the process of dividing d into n by long division, some power of 10 will then be an exact multiple of the denominator and so will leave no remainder and our process of long-division will stop.
We can write this as d = 2α × 5β where α and β are whole numbers but may be 0.
The decimal fraction for n/d has MAX(α, β) decimal places where the MAX(α, β) function means the largest of α and β.

The list of denominators with terminating decimals begins:

2, 4, 5, 8, 10, 16, 20, 25, 32, 40, 50, ... A003592

How to tell if a fraction is purely recurring as a decimal fraction

The rule for a fraction which is purely recurring, that is, its period starts immediately after the decimal point, is just the opposite of the rule for detecting a terminating fraction above:
The fraction n/d, in its lowest terms, is purely recurring if and only if neither 2 nor 5 are factors of d
n1/nPeriod
length
n is
Prime?
30.31Prime
70.1428576Prime
90.11not Prime
110.092Prime
130.0769236Prime
170.058823529411764716Prime
190.05263157894736842118Prime
210.0476196not Prime
230.043478260869565217391322Prime
270.0373not Prime
290.034482758620689655172413793128Prime
The first column is 3, 7, 9, 11, 13, 17, ... A045572 which can be described in several equivalent ways as: The third column is the lengths of all the purely periodic decimal fractions: 1, 6, 1, 2, 6, 16, 18, 6, ... A002329. There is no explicit formula for this series but we can describe it with more detail as we see in the next secion.

Did you notice that all the primes numbers are in this list, except of course 2 and 5?
Also, the length of the period of 1/n when n is such a prime is sometimes n–1 and sometimes not! Can you spot any patterns in the period length for a prime n?

The length of the period of a purely periodic decimal fraction

When we divide n into 1 and we eventually reach a remainder of 1 again, then the decimal fraction will repeat. At this point, for example, we have the following if we are using long division to find the decimal fraction for 1/21:
. 0 4 7 6 1 9 21 ) 1 . 0 0 0 0 0 0 8 4 1 6 0 1 4 7 1 3 0 1 2 6 4 0 2 1 1 9 0 1 8 9 1
The division of course is equally correct if we ignored the decimal point so that we would be calculating just with whole numbers. In that case, we have found that 1 followed by six zeroes, 106, when divided by 21, leaves a remainder of 1. That is:
106 = 47619 × 21 + 1 or:
106 - 1 is divisible by 21
The length of the recurring part of a purely periodic decimal fraction
is the smallest power of 10 that leaves a remainder of 1 when divided by the denominator.
The decimal form of a fraction is purely recurring if and only if the denominator does not have either 2 or 5 as a factor.
In the same way that we write n / d to mean n divided by d we can also write n mod d to mean just the whole number remainder when we divide the integer n by the integer d.
13 / 5 = 2.6 = 23/5
13 mod 5 = 3
mod is used only between whole numbers so the possible remainders "mod n" are 0, 1, 2, 3, ..., n-1 but also sometimes it is useful to use negative remainders -n+1, -n+2, ... -1, 0 or some other set on n-1 integers.
Another definition of a mod b = r is b divides into (is a factor of) a – r.
mod looks like the other arithmetic operations such as +, –, ×, / in that it goes between the two numbers it "operates" on.

You will often see an alternative notation in maths books, where two numbers a and b are congruent or equivalent (≡) if, when divided by a given number, the modulus n , they have the same remainder:
a ≡ b (mod n):

13 ≡ 3 (mod 5)
This is a little more general than the mod operator notation whose result is the remainder whereas this notation means that the two numbers have the same remainder when divided by the modulus, the number in the brackets.
13 mod 5 = 2 or 13 ≡ 2 (mod 5)
27 mod 5 = 2 or 27 ≡ 2 (mod 5)
⇒ 13 ≡ 27 (mod 5)
The equivalence is also called a congruence and the "mathematics of remainders" is called the Theory of Congruences or Modular Arithmetic. For instance, we can add and subtract congruences and multiply them too:
If a ≡ A (mod n) and b ≡ B (mod n) then:
a + k ≡ A + k (mod n);
a + b ≡ A + B (mod n);
a – b ≡ A – B (mod n);
a × b ≡ A × B (mod n);
k a ≡ k b (mod n)
but we must be careful about division since although
6 ≡ 12 (mod 6) but, dividing by 2:
3 ≡ 6 (mod 6) is wrong!

In terms of lengths of recurring parts of decimal fractions:
"The order of 10 mod 7" gives the length of the recurring part of decimal (base 10) fractions with denominator 7. This means the smallest power of 10 which, when divided by 7, leaves a remainder of 1.
101 ÷ 7 = 1× 7 + 3 so has remainder 3: 101 ≡ 3 (mod 7)
102 ÷ 7 = 14×7 + 2 so has remainder 2: 102 ≡ 2 (mod 7)
103 ≡ 6 (mod 7)
104 ≡ 4 (mod 7)
105 ≡ 5 (mod 7)
106 ≡ 1 (mod 7)
So fractions with denominator 7 are periodic with 6 digits in the period.
Note:


Modular arithmetic has many important applications in modern Number Theory. The great mathematician Carl Friedrich Gauss (1777 - 1855) was the first to fully develop this topic and show its power in his book Disquisitiones Arithmeticae written in Latin. See the References at the foot of this page for the English translation in paperback.

Here we have found that

106 mod 21 = 1
and that 6 is the smallest power of 10 that leaves a remainder of 1 when divided by 21.

The same is true for all fractions 1/n which have purely periodic decimal fractions.
Often in textbooks you will see the Greek letter λ (lambda) used for this length:

If λ is the smallest power of 10 that leaves a remainder of 1 when divided by n, then there are λ digits in the period of 1/n:
λ is the smallest number for which 10λ mod n = 1
The list of denominators of purely periodic fractions begins:
3, 7, 9, 11, 13, 17, 19, 21, 23, 27, 29, 31, ... A045572

Mixed Recurring Fractions

All fractions as a decimal will either terminate or recurr. Some recurr from the first decimal place and are the purely recurring decimal fractions of the previous section. Others have some digits at the start and then they recurr. These are the mixed recurring decimal fractions that we look at in this section.
For the fraction n/d, in its lowest terms, to terminate have seen that the prime factors of d must also be prime factors of 10, namely 2 or 5 or both.
Examples are: 2, 8 = 2×2×2, 40 = 2×2×2×5, 100 = 2×2×5×5

For it to be purely recurring, none of the prime factors of d must be a prime factor of 10.
Examples are: 3, 7, 9 = 3×3, 11 and all the other prime numbers bigger than 5.

So that leaves the mixed recurring decimal fractions which have at least one prime factor in common with those of 10 and at least one not in common with those of 10
Examples are 6 = 2×3, 15 = 5×3, 30 = 2×5×3
These denominators are of the form

d = 2a1 5a2 p1b1 p2b2 ...
where the pi are prime numbers other than 2 and 5 and a1 or a2 can be 0 (which effectively excludes that prime as a factor).

How long are the fixed and recurring parts of a mixed recurring fraction?

Using the above description of the denominator d, we separate its factorization into two parts, Fixed and Recurring where
d = F × R where
F includes only the prime factors 2 and/or 5 or else is 1 and
R is the rest of the prime factors or else is 1.
The size of the fixed part is determined solely by the factor F
The size of the recurring part is determined by R.
1/R will be purely recurring and the length of its period is the same as the length of the recurring part of 1/d.

Let's look at some examples:

Below there is a calculator which will find denominators or reduced fractions for given lengths of fixed and/or recurring parts.

The easiest way to find F and R for a given denominator d is to repeatedly find g, the GCD of 10 and d, then divide d by g and repeat until the GCD is 1. We then multiply together all the g's which is F. What is left in d is the R part.

The denominators of fractions with purely recurring decimals are:

3, 7, 9, 11, 13, 17, 19, 21, 23, 27, 29, 31 ... A045572
The denominators that have decimals that eventually recur, either as purely recurring decimals or else as mixed recurring decimals are those which do not have a terminating decimal expansion:
3, 6, 7, 9, 11, 12, 13, 14, 15, 17, 18, 19, 21, 22, 23, 24, 26, 27,... A085837

Fractions with the same denominator

We have mainly looked at 1/n up to now and how many decimal places it has and if it is periodic or not.
What can we say about k/n?
Some examples are:
nn/5Decimal
11/50.2
22/50.4
33/50.6
44/50.8
 
nn/6Decimal
11/60.1[6]
21/30.[3]
31/20.5
42/30.[6]
55/60.8[3]
 
nn/8Decimal
11/80.125
21/40.25
33/80.325
41/20.5
55/80.625
63/40.75
77/80.875
 
nn/12Decimal
11/120.08[3]
21/60.1[6]
31/40.25
41/30.[3]
55/120.41[6]
61/20.5
 
nn/12Decimal
77/120.58[3]
82/30.[6]
93/40.75
105/60.8[3]
1111/120.91[6]
It looks as if the numerator does not matter but that it is only the denominator of the fraction in its lowest terms that matters - and indeed this proves correct!
The character of the decimal fraction - whether it terminates or not and the lengths of the fixed and recurring parts - is solely determined by the denominator provided the fraction is in its lowest terms.

The rules for decimal (base 10) fraction for 1/d in its lowest form

The decimal fraction terminates if and only if
the only prime factors of d are 2 or 5 or both.
The decimal fraction is purely periodic (its period begins immediately after the decimal point) if and only if
neither 2 nor 5 are factors of d.
The decimal fraction has an initial number of non-repeating decimal places and then is periodic if and only if
d has both a prime factor which is a prime factor of 10 and a prime factor which is not a prime factor of 10

Here are the mathematical details (optional) to show why the numerator does not matter:


For n/d we look at
n
d
,
10 n
d
,
100 n
d
, ...
If n/d terminates, then eventually 10k n will be divisible by d.
This only happens if the prime factors of d are also prime factors of 10.
If n/d is in its lowest terms then, since d has no factor in common with n, d will divide only into a power of 10, no matter what the value of n is.

If n/d eventually becomes periodic, then we find a remainder when dividing by n by d is equal to one found earlier. This is because in this case we do not find an exact division and the remainders can only be 1, 2, 3, ... , d-1 and therefore after at most d decimal places, we must find a remainder being repeated.
Suppose the remainder first repeated is 10sn and is again found at 10s+tn.
We can write
10sn ≡ 10s+tn (mod d)
But since n and the modulus d have no factor in common (the fraction n/d is in its lowest terms), then we can divide this equivalence by n:
10s ≡ 10s+t (mod d)
The start of the period - after s initial decimal places - and the length of the period - t decimal places - are therefore independent of n.

For example: 13/108:

.1 2 0 3 7 ... 1 0 8 ) 1 3.0 0 0 0 0 ... 1 0 8 2 2 0 2 1 6 4 0 0 4 0 0 3 2 4 7 6 0 7 5 6 4
So we see that 1300 ÷ 108 has the same remainder 4 as 1300000 ÷ 108 . Checking this:
1300 = 12×108 + 4 : check!
1300000 = 12037×108 + 4 : check!
Using congruences, we can write this as:
13 102 ≡ 13 105 (mod 108) and, since GCD(108,13)=1,
102 ≡ 105 (mod 108) :checking:
   100 = 0×108 + 100
10000 = 925×108 + 100 : check!
For example:
13/108 has an initial fixed part of 2 decimal places and a periodic repeating part of 3 dps and so will all decimal fractions for n/108 provided that n has no factor in common with 108:
n/108lowest
terms
Decimal
1/1081/1080.00[925]
2/1081/540.0[185]
3/1081/360.02[7]
4/1081/270.[037]
5/1085/1080.04[629]
6/1081/180.0[5]
7/1087/1080.06[481]
8/1082/270.[074]
9/1081/120.08[3]
10/1085/540.0[925]
n/108lowest
terms
Decimal
11/10811/1080.10[185]
12/1081/90.[1]
13/10813/1080.12[037]
14/1087/540.1[296]
15/1085/360.13[8]
16/1084/270.[148]
17/10817/1080.15[740]
18/1081/60.1[6]
19/10819/1080.17[592]
20/1085/270.[185]
n/108lowest
terms
Decimal
21/1087/360.19[4]
22/10811/540.2[037]
23/10823/1080.21[296]
24/1082/90.[2]
25/10825/1080.23[148]
26/10813/540.2[407]
27/1081/40.25
28/1087/270.[259]
29/10829/1080.26[851]
30/1085/180.2[7]

Same denominator Decimal Fraction Calculator

This calculator finds all the denominators of fractions with a given fixed part length and period length.
Note that any numerator with these denominators will also have the same fixed and period lengths.
The denominators are restricted to a given maximum which can be a number up to 16 digits long.
Same denominator Decimal Fraction C A L C U L A T O R
Same denominator Decimal Fractions




with a denominator of
in base

R E S U L T S




The number of fractions with the same denominator and Euler's Totient function φ(n)

The Calculator above computes the number of fractions with the same denominator by using a relatively simple formula which is why it is so fast. The number of fractions with the same denominator d is an important function in mathematics and in Number Theory in particular. Euler used it in his research and called it φ(d), the (Euler) Totient function. It has nothing to do with the golden mean numbers φ as that is a completely different use for the Greek letter phi!
Euler's phi function φ(d) is
the number of fractions 0 < n/d < 1 which in their lowest form have d as their denominator.
Here is a table of the number of proper (reduced) fractions with a given denominator d or φ(d):
d234567891011121314151617181920
φ(d)12242646410412688166188
After φ(2)=1, all the values are even.
This is the series A000010, the totient numbers and it plays an important part in many results in Number Theory.
Another mathematical description of φ is
Euler's phi function φ(d)> is
the number of integers between 0 and d that have no factor in common with d
φ(d) = # { 0<n<d | GCD(n,d)=1 }
Sometimes it is called Euler's totient function and written as φ.

Euler found some very important properties of φ(d) that lead to a quick and efficient method of computing it for large d:

φ(d) = d – 1 when d is prime.
If d is prime then all numbers lower than it, from 1 to d – 1, have no factors in common with it.
φ(d) is multiplicative.
e.g. φ(3)=2, φ(5)=4 φ(15)=8
If we take 2 primes such as 3 which by the result above has 2 numbers with no factor in common with it, and similarly, if there are 4 numbers with no factor in common with 5 then there will be 2×4 numbers with no factor in common with 3×5:
because the multiples of 3 will not overlap with the multiples of 5 for numerators less than 3×5=15.
This to any pair of numbers with no common factor:
If gcd(a,b)=1 then φ(a b) = φ(a) φ(b)
φ(pa) = pa – pa − 1 = pa − 1(p − 1) = pa − 1 φ(p) if p is prime.
When the denominator is a power of a prime then the only numerators that will have a factor in common with it are just the multiples of the prime: p, 2p, 3p, .., p2, ... pa-p, pa and there are pa–1 of these. Hence
φ(pa) = pa – pa–1
φ(na) = na–1 φ(n) for all n
This follows because φ(n) is multiplicative and from the previous property.
φ(p1a1 p2a2 ... pkak) = φ(p1a1) φ(p2a2) ... φ(pkak) for primes pi
This looks daunting but it merely means that we split a number into its separate prime factors pi with their powers ai and find the φ count for each of these piai, using the previous property, then, since all the primes have no factors in common, we can multiply them all using the multiplicative property!

Nontotient Numbers

All values of φ(n) are even for n>2. But not all even numbers are the result of φ(n) for some n.
In other words when we list the counts of proper fractions with a given denominator, there are some counts that never appear in the list and these are the nontotient numbers. For instance,
we can find n for φ(n) = 12, since φ(13) = 12 and there are 12 proper fractions less than 1 with a denominator of 13.
Similarly, φ(n) = 16 since φ(17) = 16 and there are 16 proper fractions less than 1 with a denominator of 17.
But there is no solution to φ(n) = 14: there is no denominator that has exactly 14 reduced fractions!
14 is the smallest nontotient number.
The list begins:
14, 26, 34, 38, 50, 62, 68, 74, 76, 86, 90, 94, 98, ... A005277.

Fractions with the same lengths of period, fixed part or terminating part

We have seen how to compute the length of the terminating part of of the fixed and periodic parts of a given fraction. Now let's ask the question in reverse:
Given the length of terminating part or the lengths of fixed and periodic parts, how do we find such fractions?

Terminating Decimals of a given length

A terminating decimal fraction n/d must have d = 2a × 5b ( a and b may be 0) and the length of its (terminating) decimal is LCM( a, b ) which we saw above.
So, given a length t of a terminating decimal fraction we need to find a and b with LCM( a, b ) = t. This means that both a and b are among the factors of t or else one of them is 1 and the other is t.
For example, if t = 2, we have
a2ab5bd = 2a×5b1/d
01225250.04
12225500.02
240140.25
2415200.05
242251000.01
As a check, there are 99 decimals with just 2 decimal places, from 0.01 to 0.99, so the number of fractions in their lowest terms with the above denominators must sum to 99, you might think... but this is wrong! Why? Because 0.10 is not a terminating fraction of 2 digits but just 1.
We must exclude those 2-digit numbers that end in 0, leaving just 90 terminating fractions of 2 digits (decimal places).

Purely periodic decimals of a given length

A common symbol used for the length of the period of the decimal expansion of 1/d is the Greek lowercase letter lambda for length: λ.
λ will be the smallest power of 10 that has a remainder of 1 when divided by d. If we are given a value for λ we want to find the d for which 10λ = 1 (mod d) which means that d will be a factor of 10λ - 1.
Here are the factors of 10λ - 1 for various values of λ = length of period:
λ10λ–1prime
factorisation
factors>1
found ealier
new factors>1
1932 3, 9
29932×113, 911, 33, 99
399933×373, 927, 37, 111, 333, 999
4999932×11×1013, 9, 11, 33, 99101, 303, 909,
1111, 3333, 9999
59999932×41×273, 941, 123, 271, 369, 813, 2439,
11111, 33333, 99999
699999933×7×11×13×373, 9, 11, 27, 33, 37, 99,
111, 333, 999
7, 13, 21, 39, 63, 77, 91, 117,
143, 189, 231, 259, 273, 297,
351, 407, 429, 481, 693, 777,
819, 1001, 1221, 11287, 1443,
2079, 10101, 10989, 12987,
15873, 25641, 27027, 30303,
37037, 47619, 76923, 90909,
111111, 142857, 333333, 999999
7999999932×239×46493, 9239, 717, 2151, 4649, 13947, 41841,
1111111, 3333333, 9999999
Some interesting facts arise from this table:

Decimals of given fixed part length and period length

Let's call the fixed part length t and the period length and either may be 0.
Combining the results of the above two sections, we find all numbers T with terminating decimal of length t and all numbers R which are purely recurring with a period of length λ, as described above.
The mixed fraction denominators which have both fixed part length t and a period of length λ are those numbers made from a product of one of the T with one of the R.
From the sections above we see that

Similar sized Decimals Calculator

This Calculator finds the number of distinct denominators that can occur for decimal numbers with a given number of fixed-part decimal places and a given number of periodic-part decimal places.
For example, there are 9 distinct decimals of the form 0.d with 1 fixed decimal digit d>0 only each having a denominator of 10. But when reduced to their lowest terms, there are only 3 possible denominators for these fractions: 2, 5 and 10.
Similar sized Decimals C A L C U L A T O R


denominators up to
for decimals with fixed part of length
and period of length
in base

R E S U L T S




Converting a decimal fraction into a proper fraction

Converting a terminating decimal to a fraction

First let's look at a method of converting a decimal fraction to a proper fraction by hand. This illustrates the maths behind the process. Then we can explore with an online calculator to do it for us!
For example, 0.25:
0.25 =
2
10
+
5
100
=
25
100
=
1
4

The process is to write the fraction as a whole number divided by 10the number of decimal places and then simplify this fraction until it is in its lowest form.

Converting a periodic fraction to a fraction

If the decimal fraction is periodic then it never ends and we need a different approach.

A purely periodic decimal fraction

First, let's take a purely periodic fraction such as 0.[037] = 0. 037 037 037 .... Let's call our decimal fraction d.
First we multiply d by 10the length of the PERIOD to make the fractional part of the decimal (to the right of the decimal point) the same as the original number. For this example, we would use 103 since the period is of length 3:
103 d = 37.037 037 037 ...
Now, if we subtract the original number form this the part to the right of the decimal point will disappear:
103 d = 37.037 037 037 ...
d = 0.037 037 037 ...
103d – d=37.000 000 000 ...
(103 – 1) d =37
d = 37/999
d=1/27
The process is to multiply the purely periodic decimal by 10length of the period, subtract the original from this and then divide to form a fraction.

A mixed periodic decimal fraction

If the decimal fraction is mixed, we use a combination of the terminating and purely periodic methods.
For example, 0.12[037] = 0.12 037 037 037 037.... Let's call this m.
First, multiply m by 10length of the FIXED part to get a purely periodic fraction. The power to use here is 2:
100 m = 12.037 037 037 ... and so:
100 m = 12 + 0.[037]
Then use the purely periodic fraction method above on this value to find 100 m as a fraction:
100 m = 12 + 1/27
Now find a proper fraction that is the value of m:
100 m = 12 + 1/27
100 m = (12×27 + 1)/27
100 m = (324 + 1)/27
100 m = 325/27
m = 325/2700
m = 13/108
The calculator in the next section does all the arithmetic using the same methods as above.

Alternative forms for some recurring fractions

There are different ways to write the same recurring decimal fraction,
for instance 0.00010101010... could be written 0.000[10]
or, equally correctly, as
0.00[01].
This does not make much of a difference for this number but for other numbers such as 1/11 it does! Mathamaticians choose one of two conventions: use the longest or the shortest fixed part.
In the software package Mathematicain new window the results from the RealDigits function) uses the longest fixed part and so starts a recurring part with a non-zero digit:
1/11 = 0.0[90]. This makes a difference because 1/11 = 0.0[90] is a mixed recurring decimal fraction with a fixed part of 1 digit and a period of two digits whereas 1/11 = 0.[90] is a purely recurring decimal fraction with no fixed part.

On this page, we use the shortest fixed part for all recurring decimals in any base.

Decimal to Fraction Calculator

You can input a decimal fraction and the calculator will convert it to a proper fraction. Give the fixed part (if it has one) followed by the recurring part (if your decimal fraction is not a terminating one).
Decimal to Fraction C A L C U L A T O R

[ ]

in base

R E S U L T S



/ You do the maths...

  1. Can you find a fraction which is all 1's: 0.[1]=0.111111...?
    1. Can you find a fraction which is all 2's: 0.[2]=0.222222...?
      Looking at the result from the Calculator, how could you have got this without using the Calculator>
    2. Without using the calculator, what fraction is 0.[3]=0.3333...?
    3. ... and what is 0.[4]=0.44444...?
  2. Can you find a fraction which repeats your age (e.g. 0.[14]=0.14 14 14 14 14... )?
    1. What fraction will it be on your next birthday?
  3. Can you find a fraction which is your birthday? e.g. for 5 January 1985 you might try 0.05 01 1985 or 01 05 1985 if you prefer the American system of writing dates
  4. Can you... find a fraction which is 0.123456789?
  5. By stopping any long fraction after a few decimal places, we can find a fraction that approximates the original one.
    For instance π is 3.1415926535....
    1. What is 3.1 as a fraction?
    2. What is 3.14 as a fraction?
    3. What about 3.141 and 3.142 if we round to 3 dps?
    In fact, 22/7 = 3.1428.. is better than all those above so using just the first few decimal places to make a fraction is not always the best way to get a good fraction as an approximation.
    For a much better way see my page on Continued Fractions in new window

Fractions with maximum length periods

We know that all denominators D that recur that have a period of at most D-1 digits.
Those with the maximum length: D-1 digits (regardless of the base) are always primes and are called maximal length primes or long primes or full repetend primes. For instance, in base 10, we have
1/7 = 0.[142857]
1/17 = 0.[0588235294117647]
1/19 = 0.[052631578947368421]
...
Such maximal D are always prime and the series of denominators begins 7, 17, 19, 23, 29, 47, 59, 61, 97, 109, … A001913
The cyclic numbers themselves are 142857, 588235294117647, 52631578947368421, ... A180340
but this does not show the initial zeroes which you must remember to include when rotating them!

Hardy and Wright's "Introduction to the Theory of Numbers" (see references) on page 148 says that "very little is known about these".

However we do know all of these periods will have complementary halves meaning that the second half of the period (which will be of even length if it is maximal) are the 9-complements of the first half, meaning that corresponding digits in each half will always sum to 9.
But even more than halves we have:

If the period length of 1/p, p a prime, has a factor A then
the total of all blocks of A digits of the period will always be
a multiple of 99..9: a number with A 9s.
For example, 1/19 = 0.[ 052631578947368421 ] has a period of 18 digits. We can sum these in block of size A:
A=2: 05 + 26 + 31 + 57 + 89 + 47 + 36 + 84 + 21 = 396 = 4×99
A=3: 052 631 + 578 + 947 + 368 + 421 = 2997 = 3×999
A=6: 052631 + 578947 + 368421 = 999999 = 1×999999
A=9: 052631578 + 947368421 = 999999999 = 1×999999999
For maximal length primes, an odd prime will have an even length period and so the complmentary halves result will follow.

Reference:

Cyclic Numbers

If we take the digits of a maximal period prime, then all of its digits are just rotated when we multiply it by 2,3,4,... up to p-1.
1/7 = 0.[142857] so 142857 is a cyclic number:
142857 × 1 = 142857
142857 × 2 = 285714
142857 × 3 = 428571
142857 × 4 = 571428
142857 × 5 = 714285
142857 × 6 = 857142
Try it with one of the other maximal prime's periods : 1/17 = 0.[0588235294117647] or 1/19 = 0.[052631578947368421]. Remember that the inital zeroes are also needed! The series of prime denominators is given in A001913 and their corresponding cyclic numbers are A004042

How common are maximal period fractions?

It seems they are quite common among the primes.
In 1927 the mathematician EilArtin conjectured that the proportion of primes with maximal period is about 0.374 for all bases that are not a power of a number. The excluded bases are 4,8,9,16,25,27,32,36,49… A001597. The precise value is as follows, where p runs through all the prime numbers:
2×1 - 1 3×2 - 1 5×4 - 1 7×6 - 1 ... p(p-1) - 1 ... = 0.3739558136192...
2×1 3×2 5×3 7×6 p(p-1)
This is now called Artin's Constant and its decimal places are given in A005596.

Maximal Period Fractions Calculator

Find denominators D of fractions with the maximal length of D-1 digits.
Maximal Period Fractions C A L C U L A T O R


maximal length periods
from
up to
in base
R E S U L T S



Plots of Period lengths

Here are plots of the lengths of the periods of the purely periodic fractions 1/N for different ranges of N:

Almost-maximal periods

If you look closely at the plots above, you will see the clear line of primes p that have a maximal period p-1. The next line below them is of those (odd) numbers n whose period is (n-1)/2. All the other points are below this line.

But....there are a few odd stragglers that lie in between the top two lines!

The smallest (in base 10) is 49 with a period of 42 and the next are 289, 343, 361, 529,... A158248. These are odd composite numbers with 10 as a primitive root:

They are all powers of a maximal period prime, but not all powers of maximal prime are in this list, for example 4872 = 237169 has a period of 486 (below the half-way line) which is the same as the period for 487, making 487 a maximal period prime.

Decimal fractions that look like special sequences

You may have noticed two things in the Table of the first few fractions 1/n above: It is amazing to find that there are even more fractions whose periods contain a special pattern ...

You Do The Maths...

  1. Can you guess which simple fraction begins with the powers of 3: 0.01 03 09 27 ... ?
    Answer 1/97=0.[01 03 09 27 83 50 51 54 63 91 75 25 77 31 95 ...
  2. And be prepared to be surprised at what the powers of 3 are with 1 dp each when summed as a decimal series!
    Answer
    .1 .03 .009 .0027 .00081 .000243 .0000729 .00002187 .000006581 .0000019683 .00000059049 .000000177147 .0000000531441 ... ...       0.142857.. = 1/7
  3. How about the powers of 2 but with 3 digits each: 0.001 002 004 008 016 032 ...?
    Answer 1/98

However after a few more decimal places the series often get "muddled" (as we saw with 1/7 and 1/19 above) and the pattern seems to disappear. There must be some nice mathematical reason behind these patterns....
Which series appear in the period of a a fraction in this way?
For those, how can we find a fraction for a particular series?
This section unravels some of the simple maths behind these questions.

Powers of a number

If we look more closely at some simple fractions near 1/100 we see their decimal fractions are special:
1/99 = 0. 01 01 01 01 01 01 01 01 01 1 ...
1/98 = 0. 01 02 04 08 16 32 65 30 61 2 ...
1/97 = 0. 01 03 09 27 83 50 51 54 63 9 ...
But 1/98 looks like 01 02 04 08 16 32 ... the powers of two and 98 = 100 - 2.
and 1/97 looks like 01 03 09 27 8(1)... the powers of three and 97 = 100 - 3.
In fact, even 1/99 fits this pattern because 99 = 100 - 1 and it would then be the powers of 1, that is 01 01 01 01 01 ... which it is!
But soon the pattern of powers disappears. Why?
Because it is really there, just masked as the powers get larger.....

Shift-and-add applied to the a series makes the fraction!

We take the powers of 2 but moving each power exactly 2 places to the right each time and then we add them:
0.01+
02+
04+
08+
16+
32+
64+
128+
256+
512+
1024+
...
0.010204081632653061......
and we see this is indeed the start of 1/98 as a decimal.

Did you notice that

Any guesses as to a fraction whose decimals are the powers of 2 using three digits for each?

/ You do the maths...

  1. Check this out with 1/97 = 1/(100 - 3) and the powers of 3.
    Make a table, moving the unit digit two places to the right each time and adding. Check you have enough powers to verify the first 20 decimal places as we did for the powers of two above.
    30=0.01 +              
    31=   03 +            
    32=     09 +          
    33=       27 +        
    34=         81 +      
    35=          243 +    
    36=            729 +  
    37=             2187 +
    3...               ............
    TOTAL:                  
  2. To about 20 or 30 decimal places, what are 1/9980 and 1/9970?

So we observe (or guess!) that

the powers of p taken d digits at a time appear in the decimal fraction for 1/(10d–p)
A series of powers is called a geometric progression since each term is a constant multiple of the previous term.
In this example: a, a×b, a×b2, a×b3, ... we have a as the first term and b as the multiplier or common ratio.
We write the sum of the first n terms using the Greek capital letter sigma: Σ:


i = 0
a bi = a + a×b + a×b2 + a×b3 + ...
The sum of a geometric progression (G.P.) will exist provided -1 < b < 1 so that the terms converge to a limit, the sum.
The infinite sum - let's call it S - is easily calculated as follows:
Multiplying the sum by b gives:
b S = a×b + a×b2 + a×b3 + ... = S – a
a = S ( 1 – b)
S = a / ( 1 – b )
For our decimal sums, if we have powers of p and each power is shifted by k places, then we have
1 + p/10k + p2/102k + p3/103k + ...
where the common ratio is p/10k and the first term is 1. The sum has a limit provided that p/10k < 1, i.e. p < 10k.
Using the formula above for the sum of a G.P., the sum of all powers is
1 / (1 – p/10k) = 10k / (10k – p)
But notice that we began the sum at class=maths>1, so, to get class=maths>1 as the first term in our decimal fraction, we need to divide this fraction by 10k:
0.0001 000p 000p2 ... with k digits per power = 1/(10k – p)

The powers backwards!

At the top of this page, you might not have noticed among the examples of fractions converted into decimals this particular one:
1/19 = 0 . [052631578947368421]
It does not seem interesting except when you look at it backwards: ..., (1)6, 8, 4, 2, 1. Surely these cannot be the powers of 2 in reverse order and added, can they?
               1
              2 
             4  
            8   
          16    
         32     
        64      
      128       
     256        
    512         
  1024          
 ....           
 ....78947368421
Indeed they can!
Here is another, this time the powers-of-2 ..., 32, 64, 32, 16, 8, 4, 2, 1 have first 2 and then 3 digits each:
1/199 = 0.[00502512562814070351...14572864321608040201] with a period of 99 digits,
1/1999 = 0.[00050025012506253126...64032016008004002001] with a period of 999 digits, ...
Did you notice that the start of the period is the powers of 5? 1, 5, 25, 125, 625, ...
What about powers of 3 backwards? ..., 81, 27, 9, 3, 1
1/29 = 0.[0344827586206896551724137931] with a period length of 28,
1/299 = 0.[00334448160535117056...48829431438127090301] with a period length of 66,
1/2999 = 0.[00033344448149383127...29243081027009003001] with a period of 1499 digits,
and the powers of 4 backwards: ..., 256, 64, 16, 4, 1
1/39 = 0.[025641] with period length of 6
1/399 = 0.[002506265664160401] with a period length of 18
1/3999 = 0.[00025006251562890722...97024256064016004001] with a period length of 105
and they seem to begin with the powers of 25: 1, 25, 625, 15625, 390625,... but with 5 digits per number!

An interesting thing happens with the Triangular Numbers

1, 3, 6, 10, 15, 21, 28, ...
which we will meet later on this page. Here they are in a fraction with 2 digits per triangular number:
100/970299 = 0.[000103061015212836...5545362821151006030100], period length is 19602
The series appears both forwards from the left and backwards from the right!

The Fibonacci numbers backwards appear as the periods of:

10/109 = 0.[0917431192660550458...064220183486238532110] (108 digits)
100/10099 = 0.[00990197049212793345...34211308050302010100] (3366 digits)
1000/1000999 = 0.[0009990019970049920...008005003002001001000] (500499 digits)

Why? I will expand this section soon to include some of the theory so that we can directly find a fraction for a given series....watch this space!

See if you can spot some more rules for other series in this Calculator:

Series to Decimal Fraction Calculator

Use the Show button to display any number of decimal places, even thousands of dps if you want!
Select your series then press the button it to set the fraction in the input boxes
OR enter your own fraction
Then use the buttons below the fraction.
The and buttons will expand or shrink the fraction boxes.
Series to Decimal Fraction C A L C U L A T O R

Series numbers have dps
in base




 decimal places
in base
R E S U L T S



Summing powers in a decimal

/ You do the maths...

    1. What series do you notice in the decimal expansion of 1/9801?
    2. What fraction would give the same series but starting at 1 instead of 0? 100/9801
    3. Can you guess a fraction that has three places for each member of this series? 1/998001
    4. ... and what about a fraction with four places for each number? 1/99980001
    1. Look at your answer to the previous question and take a guess as to what 2/9801 would look like.
      Check your answer with the calculator above. 0, 2, 4, 6, 8, ... the evens, two digits at a time
    2. What series appears in the decimal for 3/9801? 0, 3, 6, 9, 12, ... the triples two digits at a time

Why do the powers patterns ultimately disappear?

We said earlier that all fractions when put into decimal form either terminate or recurr. This is true for our fractions whose decimal fractions correspond to a particular number series. They cannot go on for ever in the decimal digits! The reason lies in the fact that the numbers (the powers of 2, say) appear with two digits each. When we get to a power greater than 100, there will be an "overflow" into the 2-digit power before it. In fact what happens is that we do include every number from 0 upwards but the overflows eventually cause the decimal to get into a recurring sequence.

Number of Divisors

The fractions 1/10k – 1 have a fascinating application.
They are all periodic as none of their denominators has a factor of 2 or 5:
1/10k-1decimal
11/90.[1]0.111111111...
21/990.[01]0.010101010...
31/9990.[001]0.001001001...
41/99990.[00001]0.000100010...
51/999990.[000001]0.000010000...
60.000001000...
70.000000100...
80.000000010...
90.000000001...
...
Sum0.122324243...
The sum might not seem interesting until we realise that So in column C there is a 1 on row k whenever k is a divisor of C.
The sum of the rows is therefore the series of the number of divisors of C but as a decimal!
C123456789...
Divisors11,21,31,2,41,51,2,3,61,71,2,4,81,3,9...
#Divisors122324243...
Surprisingly this is correct for 46 decimal places:
0.1223242434262445262644283446282644492448282664
and then we find the number of divisors of 48 is 10 and its tens digit overflows into the 47th decimal place sadly!
We can increase the accuracy of the decimal by taking 2 digits per value, so we have the following as the first 100 dps:
0.0102020302040204030402060204040502060206040402080304040602080206040404090204040802080206060402100306...
We can see the last three values are now 10, 3 and 6 as the number of divisors of 48, 49 and 50.
This decimal is now correct up to 90717 decimal places because the first number with more than 99 divisors is 45360 which has 100 divisors.

The number of divisors of n is sometimes written as τ(n) using the Greek letter tau. For more values and more details see A000005.
The digits of the decimal sum are given in A073668.
Unfortunately, none of these decimals correspond to a proper fraction! We return to such non-fractional numbers later on this page.
For now, let's return to those series of number which do correspond to a proper fraction and we will discover some methods and techniques for finding a fraction which has the series as its decimal expansion.

How to turn a Series into a decimal fraction

There are lots of other series such as powers of 2: 0, 1, 2, 4, 8, 16, ... or the Fibonacci numbers 0,1,1,2,3,5,8,13,21,... which appear as the decimal form of some special fractions, but not all series! There is no fraction that gives the primes numbers for example. So which series can we find in our decimal fractions and which can we not find?
A list of some of those we can find is given in the selection menu in the above Decimal Series to Fraction Calculator. Pick one and the decimal fraction with this as an initial segment in its periodic decimal expansion will be filled in for you in the calculator's boxes so you can use the other buttons to explore it.

In this section we look at various methods of finding a fraction and what kinds of series they can generate as a decimal fraction.

Start with the series of 1's

Let's look at the series
S = 1 + x + x2 + x3 + x4 +...
If we multiply S by x we have
x S = x + x2 + x3 + x4 +...
But this is just S − 1 so
S − 1 = x S
Collecting the S terms on the left:
S − x S − 1 = 0
S(1 − x) − 1 = 0
S(1 − x) = 1
So S = 1/(1 − x):
1 = 1 + x + x2 + x3 + x4 +... starting with 1. (1)
1 − x
x = x + x2 + x3 + x4 + x5... starting with 0. (2)
1 − x

Thie first equation (1) is just the formula for the infinite sum of a geometric progression (with x<1) starting with 1,x,x²,... and the second is the formula for the series starting x,x²,... .
These are the basic formulas for our powers-in-a-decimal form
if we let x = p/10d for powers of p with d digits per power.

It is also convenient to use x = 1/D and then D is bigger than 1 and formula (2) becomes

1/D = 1 = 1 + 1 + 1 + ... (3)
1 - 1/D D − 1 DD2D3

If we look at powers of 10 in our decimal series so that D=100 in formula (3), we get
1
100
+
1
10000
+ ... = 0.01 01 01 01 ...
and this fraction is 1/D-1 = 1/99
Choosing different values for d, the number of places per item in the series, we have:
1
9
= 0. 1 1 1 1 ... (D=10)
1
99
= 0.01 01 01 01 01 ... (D=100)
1
999
= 0.001 001 001 001 ... (D=1000)
Now we can get any constant series k, k, k, k, ... with any number of digits per item by multiplying the above fractions by k
k
9
= 0. k k k k ...
k
99
= 0.0k 0k 0k 0k 0k ...
k
999
= 0.00k 00k 00k 00k ...
k
10d – 1
= 0. [k] if 0≤k≤10d-1

Add two series

If our series is a sum of two series, we can add their fractions.
For example the series 2n + 3n is 0, 1+1, 2+3, 4+9, 8+27, ... = 2, 5, 13, 35, 97, ... (A007698).
With 3 digits per term, 2n and 3n appear in the fractions
1
103-2
=
1
998
= 0. 001 002 004 008 016 ...
1
103-3
=
1
997
= 0. 001 003 009 027 081 ...
so the sum is the fraction
1
998
+
1
997
=
1995
995006
= 0. 002 005 013 035 097 ...

Moving a series several places

To move a series, we merely multiply the fraction by a power of 10 to move it to the left in the decimal, or divide it by a power of 10 to move it to the right, introducing zeroes as the new decimal places after the decimal point. We can also just add an integer before we divide too.
The series 0, 2, 5, 13, 35, 97, ... which is the series 2n+3n but starting with 0 is
1995
995006
= 0. 002 005 013 035 097 ...
1995
995006×1000
= 0. 000 002 005 013 035 097 ...
and the series 5, 13, 35, 97, ... which omits the first term is
1995
995006
= 0. 002 005 013 035 097 ...
1995×1000
995006
= 2. 005 013 035 097 ...
1995×1000
995006
− 2 =
2494
497503
= 0.005 013 035 097 ...

Multiplying by a constant

To make our series more general, we can multiply a power series by a constant.
For example, 1/7 = 0.142857 and this looks suspiciously like the beginning of the series
14, 28, 56, 112, ... = 14×1, 14×2, 14×22, 14×23,...
which is the powers of two with 2 digits each, multiplied by 14.
The powers of 2, with 2 digits each is 1/98 = 0.0102040816326530612244897959183673469387755 and so multiplying by 14 gives
14/98 = 1/7 =0.142857
Our suspicion is proved correct!

The Accumulating Sum

Finding the successive sums of a series, or accumulating the sum as we progress down a series, is a useful technique:
P(x) = a + b x + c x2 + d x3 + ...
Accumulating the sums:
a + (a + b) x + (a + b + c) x2 + (a + b + c + d ) x3 + ...
If a series is the successive sums of another series, we call it an accumulating series. For instance, accumulating the sums of the constant series 1, 1, 1, 1, 1, ... gives the series 0, 1, 2, 3, 4, 5, ... and note that we begin with 0 as the first "sum". If our series is d digits per series item then we can multiply the fraction by
10d
10d–1
to accumulate the sums beginning with the first term or multiply by
1
10d–1
to accumulate the sums from 0:
0. 01 01 01 01 01 01 ... =
1
99
and accumulating its sums we have
0. 01 02 03 04 05 06 07 ... =
100
99×99

0. 00 01 02 03 04 05 06 ... =
1
99×99
=
1
9801
To remove the initial zero, multiply by 100:
0. 01 02 03 04 05 06 ... =
100
99×99
=
100
9801
Again, accumulating these sums we have
0. 00 01 03 06 10 15 21 =
100
9801×99
=
100
970299
which are called the Triangular Numbers
Starting at 1 rather than 0, we move the series to the left:
0. 01 03 06 10 15 21 =
10000
970299
If we accumulate the sums of the constant series 2, 2, 2, 2, 2, ... we get the even numbers:
0. 02 02 02 02 02 ... =
2
99

0. 00 02 04 06 08 10 ... =
2
99×99
=
2
9801
Earlier we saw that 1/9801 is the fraction for the natural numbers 2 digits at a time and this fraction shows us that doubling it will also give the even numbers.

To get the odd numbers, we can add 1 to each of the even numbers because the series 2n+1 is double (the natural number series) + (the constant 1 series):

0. 01 03 05 07 09 ...
= 2 × ( 0. 00 01 02 03 04 05 ... ) + 0. 01 01 01 01 01 ...
=
2
99×99
+
1
99

=
101
9801
The numerator 101 tell us that we can also take the series for 1/9801, which is the natural number series 0, 1, 2, 3, ..., and shift it two places to the left (multiply by 100) then add it to the natural numbers series again:
0. 01 03 05 07 09 = 100 (0. 00 01 02 03 04 05 ...) + (0.00 01 02 03 04 05 ...)
=
101
9801

=
10 a + b
89
for 1 digit per term
=
100 a + b
9899
for 2 digits per term
=
1000 a + b
998999
for 3 digits per term

Differences

Let's take a look at the series of square numbers 1, 22=4, 32=9, 42=16, 25, 36, ... . What proper fraction produces this series in its decimal fraction?
Here we can find successive differences:
Squares:0149162536...
Differences1357911
Now we can see that the squares are the accumulated sums of the odd numbers!
This is easily verified since we add an odd number to one square to get the next:
(n + 1)2 = n2 + 2n + 1
Let's then take the fraction for the odd numbers of two digits each: 101/9801 and apply the "accumulate the sums" operation of dividing by 99:
0. 01 03 05 07 09 ... =
101
9801

0. 00 01 04 09 16 25 ... =
101
9801×99
=
101
970299
So the squares, 3 digits at a time would be
0. 000 001 004 009 016 025 ... =
1001 = 1001
998001×999997002999
Also, we see that the the square numbers are the sum of two consecutive Triangular numbers since the numerator here is 1001 and the denominator is the one for the Triangular numbers that we found above:
0. 000 001 003 006 010 015 ...    1001/(998001×1000)
+ 0. 001 003 006 010 015 ...+1001/998001
= 0. 001 004 009 016 025 ...=1002001/998001000

If we take the cubes, we can find differences and then take their differences, the second differences, and when we take differences of the second differences (the third differences), we find a constant series:

Cubes:0182764125216...
1st Differences:1719376191...
2nd Differences:612182430...
3rd Differences:6666...
This shows us how to find a decimal fraction for the series of cubes:
First, the constant series of 6's with 3 digits each:
0. 006 006 006 006 ... =
6
999
=
2
333
Accumulating the sums:
0. 000 006 012 018 ... =
2
333×999
=
2
332 667
But we need to start at 1, so we add on 0.001 = 1/1000:
0. 001 006 012 018 ... =
2
332 667
+
1
1000
=
334 667
332 667 000
and these we accumulate without introducing an initial zero term:
334 667
332 667 000
×
1000
999
=
334 667
332 334 333
= 0. 001 007 019 037 ...
and finally one more accumulation, with initial zeroes, to give the cubes:
334 667
332 334 333 × 999
=
334 667
332 001 998 667
= 0. 000 001 008 027 064 125 ...

Why this works (optional)

Show why this works The accumulated sums of power series in x is found by dividing by 1 - x:
a + b x + c x2 + d x3 + ...
1 - x
= a + (a+b) x + (a+b+c) x2 + (a+b+c+d) x3 + ...
Another way of showing this is that in Summing a power series we see that
1
1 - x
= 1 + x + x2 + x3 + x4 +...
Using this and multiplying out the brackets to collect terms in x we have:
a + b x + c x2 + d x3 + ...
1 - x
= (1 + x + x2 + x3 + x4 +... ) (a + b x + c x2 + d x3 + ...)
= a + (a+b) x + (a+b+c) x2 + (a+b+c+d) x3 + ...
Note that in decimal, base 10, we use x = 1/10d since we want d decimal places at a time for each number.
This accumulating powers of x by multiplying by 1 / (1–x) means multiplying by
1
1 - x
=
1
1 - 1/10d
=
10d
10d – 1
Multiplying by the simpler fraction 1/(10d–1) means we have also divided by 10d which will have moved our series 1 number to the right or d decimal places, which is where the initial d 0's come from.
What is behind this method for differences is that if P(x) = a + b x + c x2 + d x3 + ... is some power series, for the series a, b, c, d, ... then if we multiply it by 1 – x we get the series of differences:
( 1 – x ) P(x)
= (1 – x) ( a + b x + c x2 + d x3 + ... )
= a + b x + c x2 + d x3 + ... – x (a + b x + c x2 + d x3 + ...)
= a + (b – a) x + (c – b) x2 + (d – c) x3 + ...
Taking differences and accumulating sums are opposite actions on a series (power series).
If we take the differences of the accumulated sums series, we find we arrive back at our original series:
Sums(x) = a + (a + b) x + (a + b + c) x2 + (a + b + c + d ) x3 + ...
the differences here are: a + (a + b – a) x + (a + b + c – (a + b)) x2 + (a+b+c+d–(a+b+c)) x3 + ...
= a + b x + c x2 + d x3 + ... = P(x)

Fibonacci Numbers

The Fibonacci Numbers occur in many parts of mathematics as well as in the patterns of seeds on flowerheads. They have a simple recursion rule:
Add the last two numbers to get the next, starting from 0 and 1 (or 1 and 1 or 1 and 2...):
0, 1,1, 2, 3, 5, 8, 13, 21, ... A000045
On my Fibonacci pages, in the section The Fibonacci Series as a Decimal Fraction we saw how to prove that 1/89 was the Fibonacci series-in-a-fraction by using a generating function, which is a polynomial which encoded the Fibonacci numbers as its coefficients:
P(x)=F(0) x + F(1) x2 + F(2) x3 + ... ...+F(n-1)x n + ...
which we showed was
P(x) = x2
1 - x - x2
If we divide top and bottom by x², we have
P(x) = 1
1/x² - 1/x - 1

Now let X = 1/x and we have
P(X) = 1
X² – X –1
where X>1.
If we let x = 1/10 or X = 10 then we have
0 + 1/10 + 1/100 + 2/1000 + 3/104 + ... = 0.011235... = 1/100-10-1 = 1/89
For two digits per Fibonacci number, we let x = 1/100 so that we have
0 + 1/102 + 1/104 + 2/106 + 3/108 + ... = 0.010102030508... =1/10000-100-1 = 1/9899

Fibonacci Numbers and Pascal's Triangle illustrated by decimal fractions

Another way of looking at this fraction, 1/89 is that it is 1/100-11 and so, from our results above on Summing Powers in a Decimal it is related to powers of 11, that is 1 + 11/10 + 112/100 + 113/1000 + ...:
1/89 = .01110 = 1
+.0011 111 = 11
+.000121 112 = 121
+.00001331 113 = 1331
+.0000014641 114 = 14641
+.0000001510105 115 = 15(10)(10)51 = 161051
+...
=.0112358
which, in turn, is related to Pascal's Triangle. Since we are looking at the decimal fraction for powers of 11, then each power will be shifted by 1 place and so we get out diagonal line in Pascal's Triangle.

Finally, why are powers of 11 related to Pascal's Triangle? Because if E is a power of 11 then the next power is 11 E or (10+1) E so we have the previous power's neighbouring digits added to get the digits of the next power of 11, just as we found elements in Pascal's triangle by adding the two in the previous row to get the one in the next row under them.

Reference:

Lucas Numbers

The Lucas Numbers, L(n), are another Fibonacci-type series but begin with 2, 1, ... and use the Fibonacci Rule of add the lastest two numbers to get the next:
2, 1, 3, 4, 7, 11, 18, 29, ... A000032
They are also just the sum of the two Fibonacci numbers on either side of a Fibonacci number:
L(n) = F(n+1) + F(n-1)
Here are the steps for building a fraction for the Lucas Numbers with two digits per item:
  1. The fraction for F(n) taken 2 digits at a time and starting at 0 is
    1/9899 = 0. 00 01 01 02 03 05 08 13 21 ...
  2. F(n+1) is seen in the decimal fraction of
    1/9899 × 100 = 100/9899 = 0. 01 01 02 03 05 08 13 21 ...
  3. F(n-1) is seen in the decimal fraction for
    1/9899 ÷ 100 = 1/989900 = 0. 00 00 01 01 02 03 05 08 13 21 ...
    although we need to introduce 1 before the 0 term to make the Fibonacci rule apply:
    1, 0, 1, 1, 2, 3, 5, ...
    so our fraction for the F(n−1) series is
    (1 + 1/9899) ÷ 100 = 9900/989900 = 99/9899 = 0. 01 00 01 01 02 03 05 08 13 ...
  4. the Lucas numbers L(n)=F(n+1)+F(n-1) are seen in the decimal fraction of
    99/9899 + 100/9899 = 199/9899 = 0. 02 01 03 04 07 11 18 29 ...

General Fibonacci-type series

Did you notice that the Fibonacci rule describes how the digits of 1/19 grow? 1/19 is 0.[ 05 26 31 57 89 ...] the follwoing two-digt pairs being affected by carries from later numbers.

This section explores these General Fibonacci series-in-decimals.

There are many other pairs of starting numbers that, with the Fibonacci Rule, give interesting series, called General Fibonacci series, G(a,b,n), where a and b are the two starting values. So

/ You do the maths...

  1. The Pentagonal Numbers are 0, 1, 5, 12, 22, 35, 51, 70, ... A000326 and have constant second differences. Find a fraction which has this series in its decimal expansion with 3 digits per number.
    Pentagonals:01512223551...
    1st Differences:147101417...
    2nd Differences:33333...
    0. 000 001 005 012 022 035 051 ... = ((3/999+1/1000)/999 + 1/1000)/999 =
    334
    332 334 333
  2. Look at the fractions (2 digits per number) for the triangular numbers, the square numbers and, from the previous question, the pentagonal numbers. What do you think is the next fraction in the sequence (the hexagonal numbers)?
  3. For the Lucas Numbers, 2,1,3,4,... what proper fraction expands to 2.01 03 04 07 11 ... ?

What about bases other than base 10?

If we consider bases other than 10, all our results still hold with some slight modifications. All fractions are terminating or purely recurring or mixed recurring.
Here is a table of the fractions 1/n for n from 2 to 12 in bases from 2 to 12. For bases above 10 we use the letters A=10, B=11, ... , Z=35, either UPPERCASE or lower.

1/n terminating in bases 2 to 12
Base:1/21/31/41/51/61/71/81/91/101/111/12
20.10.[01]0.010.[0011]0.0[01]0.[001]0.0010.[000111]0.0[0011]0.[0001011101]0.00[01]
30.[1]0.10.[02]0.[0121]0.0[1]0.[010212]0.[01]0.010.[0022]0.[00211]0.0[02]
40.20.[1]0.10.[03]0.0[2]0.[021]0.020.[013]0.0[12]0.[01131]0.0[1]
50.[2]0.[13]0.[1]0.10.[04]0.[032412]0.[03]0.[023421]0.0[2]0.[02114]0.[02]
60.30.20.130.[1]0.10.[05]0.0430.040.0[3]0.[0313452421]0.03
70.[3]0.[2]0.[15]0.[1254]0.[1]0.10.[06]0.[053]0.[0462]0.[0431162355]0.[04]
80.40.[25]0.20.[1463]0.1[25]0.[1]0.10.[07]0.0[6314]0.[0564272135]0.0[52]
90.[4]0.30.[2]0.[17]0.1[4]0.[125]0.[1]0.10.[08]0.[07324]0.0[6]
100.50.[3]0.250.20.1[6]0.[142857]0.1250.[1]0.10.[09]0.08[3]
110.[5]0.[37]0.[28]0.[2]0.[19]0.[163]0.[14]0.[124986]0.[1]0.10.[0A]
120.60.40.30.[2497]0.20.[186A35]0.160.140.1[2497]0.[1]0.1
130.[6]0.[4]0.[3]0.[27A5]0.[2]0.[1B]0.[18]0.[15A]0.[13B9]0.[12495BA837]0.[1]
140.70.[49]0.370.[2B]0.2[49]0.20.1A70.[17AC63]0.1[58]0.[13B65]0.12[49]
150.[7]0.50.[3B]0.30.2[7]0.[2]0.[1D]0.1A0.1[7]0.[156C4]0.1[3B]
160.80.[5]0.40.[3]0.2[A]0.[249]0.20.[1C7]0.1[9]0.[1745D]0.1[5]
170.[8]0.[5B]0.[4]0.[36DA]0.[2E]0.[274E9C]0.[2]0.[1F]0.[1BF5]0.[194ADF7C63]0.[17]
180.90.60.490.[3AE7]0.30.[2A5]0.2490.20.1[E73A]0.[1B834G69ED]0.19
190.[9]0.[6]0.[4E]0.[3F]0.[3]0.[2DAG58]0.[27]0.[2]0.[1H]0.[1DFA6H538C]0.[1B]
200.A0.[6D]0.50.40.3[6D]0.[2H]0.2A0.[248HFB]0.20.[1G759]0.1[D6]
210.[A]0.70.[5]0.[4]0.3[A]0.30.[2D]0.270.[2]0.[1J]0.1[F]

Here is a table of the sizes of the fixed part and of the period of the decimals of 1/n for n from 2 to 21 and the type of the decimal fraction in each base from 2 to 21:
Fixed and Recurring lengths of 1/n in bases 2 to 21
Base:1/n1/2 1/3 1/4 1/5 1/6 1/7 1/8 1/9 1/101/111/121/131/141/151/161/171/181/191/201/211/221/231/241/251/261/271/281/291/30
21002200412033006140102201213044008160182406110011320201120182302814
3011002041106022004051203061404016210180416050111202003300602814
41001100211032003120511061302200413091203150112101016091301412
501020110020602061105020406120401606091106050220220040180601412
610102001100230201101020012121140016200921121100113005112302201411
70101020401100203040100201211040201603030411010022020401209120704
8100210041201100214010120411042008120614021100111202014061102814
90110010211030110020511030312020811090213050111101003200301412
1010012010110630011002210616114001611018200612022312016032602811
110102020102030206011002012030204016060302061102202050120180602802
121010100410062020140110021614200162006141611011200201230160414
130101010401020203040100110020404040301804020100110202011090201404
141002200212103006120522011002400161601822121502232010110182002812
15011002101101022011051201201100208210181211050221220012300202811
1610011001110310031105110313011002130911031501111051309130711
1701020104020601020401002060604011002090406010022020200606060404
1810102004100330101401020041314400110022413110011300414202302814
1901010202010602010201002012060204080110020601002202010012030602802
20100210101202200610051201212122001616011002150222220112018120712
210110010111100220010211041111040421018011002022120504301102811

We usually call the base-B digits in the expansion of a base-B fraction the decimal form of the fraction even if the base is not 10. We ought really to talk about the bimals for base 2 and the trimals for base 3, etc, but this sounds strange, so we opt for the easier base 2 decimal or base 3 decimal, etc.

Terminating Fractions in other bases

For prime bases such as 2,3,5,7 and 11, the only fractions which terminate must have a denominator which is a power of that prime.
The same is also true for bases which are themselves a power of a prime such as base 4 = 22, base 8 = 23 and base 9 = 32. maths fractions are those with a denominator which is a power of that same prime.
So what about bases whose factors involve more than one prime?
basen where m/n
terminates
OEISDescription
22,4,8,16,32,...A000079{2i | i≥0}
33,9,27,81,...A000244{3i | i≥0}
42,4,8,16,32,...A000079{2i | i≥0}
55,25,125,...A000351{5i | i≥0}
62,3,4,6,8,9,12,16,18,24,...A003586{2i3j | i,j≥0}
77,49,...A000420{7i | i≥0}
82,4,8,16,32,...A000079{2i | i≥0}
93,6,9,27,54,81,...A000244{3i | i≥0}
102,4,5,8,10,16,20,...A003592{2i5j | i,j≥0}
1111,121,...A001020{11i | i≥0}
122,3,4,6,8,9,12,16,18,24,...A003586{2i3j | i,j≥0}
{ expression | condition } is the set of all numbers described by the expression but subject to, or generated by, the given condition. For example:
{2i3j | i,j≥0} is the set of numbers which are a product of a power of 2 and a power of 3 where the power i of 2 and the power j of 3 are the (whole) numbers 0 or more.

The general rule is

The Terminating Fraction Rule
A fraction N/D terminates in base B if all the prime factors of the denominator D are also prime factors of the base B.
The numerator does not affect the termination of the "decimal".
Why does a fraction N/D terminate?
Let's look at an example first:
1/500 in base 10 is a terminating decimal fraction because
1/500 = 0.002 = 2/1000, so the denominator, 500 divides exactly into 1000=103 and we have 3 decimal digits in the decimal fraction.
We have found a power of 10 which is an exact multiple of the denominator, 50.
In any base B, let's look at 1/D.
This terminates if when we can find some power of the base B which is an exact multiple of D.
Only if all of D's prime factors are also prime factors of B will we be able to find such a multiple.
In base 10, 1/3 will not terminate as a decimal because 3, a prime, is not a prime factor of 10. Similarly for 1/6 for though 6 has 2 as a prime factor which is also a prime factor of 10, the prime factor 3 is not. Any number of the form 2i×5j will always be a factor of 10i×j at the most.
The smallest power of the base B which is an exact multiple of the denominator is the number of base-B digits in the base B decimal fraction for N/D (in its lowest form).

This describes all the terminating fractions in base B and therefore:
A fraction N/D recurs in base B if D has a prime factor that is not a prime factor of the base B

Decimal Fractions Calculator for any Base

Digits in bases greater than 10 are shown separated by a space:
0.12 in a base greater than 12 means a fixed length decimal of a single digit "12"
whereas 0.1 2 means two digits.
Decimal Fractions in any Base C A L C U L A T O R




showing up to periodic dps
with denominators
from
to
in bases
from
to
R E S U L T S



/ You do the maths...

  1. What is 1/2 in bases 2 to 12?
    Write your result mathematically.
    An answer
    1/22b+1 = 0.[b]
    1/22b = 0.b
  2. Which denominators that are terminating in base 10 (decimal) are also terminating in base 2 (binary)?
    An answer
    All of them because they are just the powers of 2
    1. Find a formula for the fractions which, in binary, have a period of n 0s followed by n 1s
      [Hint: use the Decimal to Fraction Calculator above] :
      They start with 0.[01]=1/3, 0.[0011]=1/5
      Answer with n 0s and n 1s 0.[0..01..1]2 = 1 /( 2n + 1)
    2. What if we started with n 1s followed by n 0s?
      Answer with n 1s and n 0s 0.[1..10..0]2 = 2n /( 2n + 1)
  3. If all the recurring fraction's denominators in base B are also all the recurring fraction's denominators in base C, what can you say about B and C?
    An answer Since recurring fraction's denominators have no prime factor in common with the base, then both B and C must have the same prime factors.
  4. What do you notice about the digits of 1/(b-1) in base b?
    An answer
    They are all 0.[1]
  5. What do you notice about the digits of 1/(b-1)² in base b?
    An answer
    They are all 0.[012...(b-3)(b-1)], omitting digit b-2
    For instance, in base 10, 1/9² = 1/81 = 0.[012345679] omitting 8.
  6. What fractions in binary have a period of n 0s followed by n+1 1s?
    Hint: Use the Decimal to fraction Calculator above
    Answer
    2n+1-1 / 22n+1-1
  7. What are the maximal primes (those with the longest possible periods) less than 100 in base 2?
    [Hint: Use the Maximal Period Fractions Calculator above]
    Answer 3, 5, 11, 13, 19, 29, 37, 53, 59, 61, 67, 83 A001122 Also known as the primes with a primitive root of 2
  8. What operation on a fraction of a maximal prime will rotate its digits one place to the left cyclically?
    Hence cycle the binary fraction for 1/13 in base 2 to find all its 12 multiples without recalculating the periods each time.
    Multiple the fraction by the base
    In base 2, rotating the period of 1/13 cyclically one place left each time, we have:
    1/13 = 0.[000100111011]
    2/13 = 0.[001001110110]
    4/13 = 0.[010011101100]
    8/13 = 0.[100111011000]
    16/13 = 1+3/13, 3/13 = 0.[001110110001] : 2×8 = 16 ≡ 3 (mod 13)
    6/13 = 0.[011101100010]
    12/13 = 0.[111011000100]
    22/13 = 1+11/13, 11/13 = 0.[110110001001] : 2×12 = 24 ≡ 11 (mod 13)
    22/13 = 1 + 9/13, 9/13 = 0.[101100010011] : 2×11 = 22 ≡ 9 (mod 13)
    18/13 = 1 + 5/13, 5/13 = 0.[011000100111] : 2×9 = 18 ≡ 5 (mod 13)
    10/13 = 0.[110001001110]
    20/13 = 1 + 7/13, 7/13 = 0.[100010011101] 14/13 = 1 = 1/13, .... and we start this list again! Answer

Maximal primes in other bases

For a given prime p, the smallest base in which it is has the maximal period (of p-1 digits) is called the least primitive root of the prime. These are as follows:
primebase base decimal
230.[1]
320.[01]
520.[0011]
730.[010212]
1120.[0001011101]
1320.[000100111011]
1730.[0011202122110201]
1920.[000011010111100101]
2350.[0102041332143424031123]
Apart from the prime 2, these bases form the series 2,2,3,2,2,3,2,5,2,3,2,6,... A001918 of the least primitive root of each prime.
There are many bases in which a prime denominator p is maximal - but not all. There are no maximal period primes in bases that are a square number!

A Theorem by Fermat (published in 1640)

For a prime p which is not a factor of the base B, Bp-1 = 1 (mod p)
This means that 1/p for p a prime has a period of at most p-1 digits in a base where p is not a factor.
Reference:

Are there any other types of numbers that are not fractional?

Optional Are there numbers whose decimal fractions neither terminate nor recurr?
Numbers whose decimal fraction terminate or end up recurring are always proper fractions as we have seen on this page. No other numbers have these properties.

The irrational numbers

So what about the decimal number 0.01 001 0001 00001 ... which does not end but also never repeats as each set of 0's-and-1's that we write has an extra 0 in it compared with the previous set.
But clearly no pattern of a fixed length repeats here and so this decimal does not represent any proper fraction.
Many important numbers that we use often in mathematics are non-fractional or non-ratio-nal numbers; they are called irrational numbers.

Probably the earliest number found to be irrational was √2 = 1.414213562373095... most likely found by Theodorus, the tutor of Plato, around 400BC.

Some other examples are
If r is irrational then so are the following for any positive or negative rationals k, ℓ
  • k + ℓ r and k – ℓ r
  • k / r and k × r
The reason is that if k + ℓ r = X say, is rational, then we can solve for r and find r = (X - k)/ℓ which would also be rational.
So, conversely, if r is irrational then so is k + ℓ r . Similarly for
k – ℓ r
k / r
k × r
.

According to Hardy and Wright (see above), Theorem 45:

If x is a root of an equation
xm + c1 xm–1 + c2 xm–2 + ... + cm= 0
where the ci are integers of which the first is 1,
then x is either an integer or is irrational.
Since √2 is a root of x2 – 1 = 0 and is clearly not an integer, then this Theorem shows it must be irrational.
Also, kn is a root of xk – n = 0 so, if n is not obviously an integer (is not the k-th power of an integer), then it too is irrational.

/ You do the maths...

  1. Find some more English words that begin with ir- where the ir- means not and negates the meaning of the rest of the word:
    e.g. an irregular polygon is a polygon that is not regular (a regular polygon has all sides and all angles equal).
    This may be irrelevant or you may find it an irresistible challenge. :)]
  2. Find some other words which begin with ir- which do not have this meaning.
    I know an Irish man who likes to iron out such irritating questions. :)
The square root of 2 as a decimal fraction is 1.41421 35623 73095 .... .
No matter how far we go on expanding this number as an ever more precise decimal fraction, its decimal digits will never get into any repeating pattern.
Suppose we can write √2 as the fraction p/q, where p and q have no common factors, so that p/q is in its lowest terms.
√2 = p/q ............... (1)
Squaring we have: 2 = p2/q2
Multiplying by q2: 2 q2 = p2 ........... (2)
So we see that p2 is even and therefore p is even and we can write it as p = 2 k.
Putting this in (2) we have 2 q2 = (2k)2 = 4 k2
Dividing by 2: q2 = 2 k2
But now we have shown q too must be even.
This cannot be since we said p/q was in its lowest terms so there is no factor in common between them.
Since assuming √2 is a fraction which can be written in its lowest terms leads to a logical contradiction, then √2 cannot be written as a fraction.

OR, putting this argument in another way....

Let's assume √2 = p/q where we make no assumption about lowest terms.
The reasoning aboves shows that both p and q are both even, so we can write p=2k and q=2n This means that √2 = p/q = (2k)/(2n) = k/n
But now the same reasoning shows that k and n are also both even.
We can continue this for ever, in an infinite decent with the numbers getting smaller and smaller with no end.
This is impossible for positive integers whose lowest value is 1.
We have to again conclude that √2 cannot be written as p/q with two integers p and q.


There are lots of other irrational numbers too -- such as:

The Algebraic Numbers

One thing all the integers, the fractions and some irrational numbers have in common is that they are all the roots of a polynomial. This includes We usually put all the terms of the polynomial on one side and and the values that make the expression 0 are called its roots.
Any number that is a root of a polynomial with integer or fractional coefficients is called an algebraic number
The reason we insist on integer or fractional coefficients (of the powers of x) is that otherwise we could have the equation
x = π
or any value we like!
In finding methods of solving equations, we get a heirarchy from each type of polynomial with a different degree.
The degree of a polynomial in x is the highest power of x in the polynomial.
For instance, a degree 1 polynomial in its most general form is a x = b for rational numbers (or integers) a and b.
It can be solved to give x = b/a which is a rational number or integer.

For degree 2, we have quadratic equations of the general form a x2 + b x + c = 0.
Here solutions involve square-roots in general and sometimes square-roots of negative numbers, which leads us to complex numbers x + i y where i = -1.
If y is zero, the number is real and if not, it is complex.
Solving higher degree polynomial equations involvers no other types of number than complex numbers!

... and even more numbers!

There are other numbers that are not the roots of any finite polynomial with rational coefficients: they are called transcendental numbers.
Such numbers are π, e and many other trigonometric and logarithmic values.

It is generally quite hard to prove a number is transcendental since we have to show it is not the root of any polynomial with rational coefficients.

References


© 2000-2019 Dr Ron Knott
created: 14 August 2000,
Back to the Main (Fibonacci) page