play and relax: games for kids games
  Cut the knot: learn to enjoy mathematics
A math books store at a unique math study site. Learn to enjoy mathematics.
Google
Web CTK
Try our no ads browsing

Sites for teachers
Sites for parents
Terms of use
Awards

Interactive Activities
CTK Exchange
CTK Insights - a blog

Games & Puzzles
What Is What
Arithmetic/Algebra
Geometry
Probability
Outline Mathematics
Make an Identity
Book Reviews
Stories for Young
Eye Opener
Analog Gadgets
Inventor's Paradox
Did you know?...
Proofs
Math as Language
Things Impossible
Visual Illusions
My Logo
Math Poll
Cut The Knot!
MSET99 Talk
Other Math sites
Front Page
Movie shortcuts
Personal info
Privacy Policy

Guest book
News sites

Recommend this site

Sites for teachers
Sites for parents

Education & Parenting

Manifesto: what CTK is about Buying a book is a commitment to learning Table of content Try our no ads browsing Things you can find on CTK Chronology of updates Email to Cut The Knot Recommend this page

 

Stern-Brocot Tree

The Stern-Brocot tree is built with the mediant fractions. Place two starting terms 0/1 and 1/0 some way apart and their mediant fraction 1/1 in between and a little down. This creates two gaps: one between 0/1 and 1/1 and another between 1/1 and 1/0. Compute two corresponding mediants and place them below 1/1. Continue in this manner. The next step will add a row of four fractions, then 8, and so on.

It's a remarkable property of the Stern-Brocot tree that the mediant fractions obtained in that manner always come out in lowest terms. The fact is proven by induction based on the following property of the tree: any two fractions m1/n1 and m2/n2 whose mediants may be computed at any stage of the construction, satisfy

(1) m2n1 - m1n2 = 1

A mediant fraction generated by two terms that satisfy (1) stands in the same relation with both of its progenitors. From here we observe that the rows of numerators and denominators of the terms in the Stern-Brocot tree are computed independently of each other. They may be dealt with separately. The row of numerators starts with the pair of integers 0,1. The row of denominators starts with the pair of integers 1,0. They are just symmetric reflections of each other. Let denote the first tree as [0, 1] and the second [1, 0]. We may generalize and consider trees that grow from an arbitrary pair of integers [x,y].

 

Since, at any stage of construction, we only carry out linear operations (actually only addition), we get a whole vector space of trees. The space is 2-dimensional as we can write [x,y] = x[1,0] + [0,1]. Each tree [x,y] consists of two parts: the left tree [x,x+y] =x[1,1] + y[0,1] and the right tree [x+y,y] = x[1,0] + y[1,1]. The tree [1,1] is a mirror image of itself. In particular, all its rows are palindromic.

Let's record several observations:

  1. The tree of the denominators [1,0] has [1,1] as its left part.
  2. The left part of the tree of the numerators [0,1] is still [0,1] while [1,1] is now the right part.
  3. Fractions on the left from 1/1 are less than 1; fractions on the right from 1/1 are greater than 1

Combining all together we see that

  1. The left part of the tree of the denominators is palindromic.
  2. The right part of any row of numerators coincides with the left part of the corresponding row of denominators.
  3. The left part of the row of numerators coincides with the previous row of numerators.

Surprisingly, the Stern-Brocot tree contains all non-negative fractions. Therefore, its left subtree contains all fractions between 0 and 1. Somehow it must be possible to pluck off the tree the Farey series of any order. Whatever the exact process, since the only criteria for selecting a fraction into the series is the magnitude of its denominator, and, as we know, the right part of any row of denominators is palindromic, the same will always be true for the denominators in the Farey series.

Copyright © 1996-2008 Alexander Bogomolny

30240992Page copy protected against web site content infringement by Copyscape


Search:
Keywords:



Latest on CTK Exchange
try this puzzle ?/?? + ?/?? + ?/? ...
Posted by albert1950
3 messages
10:42 AM, Sep-25-08

Help me find Hisashi ABE, Pythago ...
Posted by likesmath
2 messages
11:11 AM, Oct-06-08

triangle construction
Posted by Elianto84
11 messages
12:20 PM, Oct-09-08

Interesting facts about hyperbolas
Posted by qwendy
1 messages
08:34 AM, Oct-08-08

Circles through the Orthocenter
Posted by Bui Quang Tuan
3 messages
09:08 AM, Sep-09-08

Vincenty Formulae & Common Tangen ...
Posted by Cathy Gordon
1 messages
12:26 PM, Oct-09-08

Possible mistake in "What Is Geom ...
Posted by David
1 messages
09:05 AM, Sep-25-08