A unit fraction is a fraction of the form , where is a positive integer. It represents one of the equal parts into which a whole has been divided. Examples include , , and . Unit fractions are fundamental to elementary arithmetic and play a distinctive role in the history of mathematics and in number theory, particularly in representations of rational numbers as sums of reciprocals.
Definition and interpretation
In , the numerator 1 indicates that one part is taken, while the denominator specifies the number of equal parts making up the whole. For example, if a length is divided into five equal segments, each segment has of the original length.
Under the definition allowing every positive integer denominator, is a unit fraction. Treatments restricted to proper fractions require , excluding this boundary case. Zero is not a unit fraction, and is undefined.
A rational number is a unit fraction precisely when its expression in lowest terms has numerator 1 and positive denominator. Thus represents the unit fraction , although its displayed numerator is not 1.
Every positive fraction with integer numerator and denominator can be expressed as a whole-number multiple of a unit fraction:
This connects the interpretation of fractions as equal parts with their interpretation as quotients.
Order and arithmetic
For positive integers and ,
A larger denominator therefore gives a smaller unit fraction: dividing the same whole into more equal parts makes each part smaller.
The product of two unit fractions is always another unit fraction:
Their sum and difference satisfy
A sum need not be a unit fraction. For example,
whereas
More generally, is a unit fraction exactly when divides , since its reciprocal must be the positive integer .
The quotient of two unit fractions is
which is a unit fraction exactly when divides .
Egyptian fractions
An Egyptian fraction is a finite sum of distinct positive unit fractions. For example,
The requirement that denominators be distinct distinguishes Egyptian fractions from the immediate expression of as copies of .
Ancient Egyptian mathematical texts, including the Rhind Mathematical Papyrus, dating to approximately 1650 BCE, make extensive use of unit-fraction decompositions. Egyptian notation also included a special representation for , so its treatment of fractions was not exclusively limited to unit fractions.
Existence and the greedy algorithm
Every positive rational number less than 1 has an Egyptian-fraction representation. One construction is the greedy algorithm: repeatedly subtract the largest unit fraction no greater than the remaining number.
For a remainder , choose
where the ceiling denotes the least integer not smaller than . Then
If the remainder is nonzero, its numerator before reduction satisfies
Reduction to lowest terms cannot increase that numerator. Consequently, the positive numerators strictly decrease, so the algorithm must terminate. The selected denominators also strictly increase, giving distinct unit fractions.
For example,
The greedy algorithm guarantees a representation, but it need not minimize either the number of terms or the largest denominator.
Nonuniqueness
Unit-fraction decompositions are generally not unique. The identity
splits any unit fraction with into two distinct smaller unit fractions. Repeatedly applying this identity to the term with the largest denominator produces infinitely many distinct finite decompositions of the same number.
Restrictions on the number or size of denominators lead to problems involving Diophantine equations. For instance,
is equivalent to
Positive integer solutions can therefore be found from factor pairs of .
Infinite sums and decimal representations
The sequence of unit fractions tends to zero:
Nevertheless, their infinite series, the harmonic series, diverges:
This illustrates that terms approaching zero are necessary, but not sufficient, for convergence of a series. Some more sparsely selected unit fractions do have a finite sum; for example, the geometric series
In decimal positional notation, a unit fraction terminates exactly when the only prime factors of are 2 and 5, including as the case with no prime factors. Otherwise, its decimal expansion is eventually repeating. Thus , while . These different decimal forms do not affect their status as unit fractions.