Homework Assignment 5

Pattern 1: How to form the next row in the fraction tree

I have listed down the steps of how to construct the next row in the tree

Imagine forming a rectangle that has roughly the same area as the kth row. This rectangle stands for
the next row in the tree. Divide this box into two equal rectangles vertically. There will be n number
of fractions in each rectangle.

Step 1: The numerators of the fractions in the left rectangle will be the same as the numerators of
the fractions in the kth row

Step 2: The sum of the numerator and the denominator of all the fractions the kth row becomes the
numerators of the fractions in the right rectangle of the k + 1 row.

Step 3: The series of the denominators in the whole K + 1th row is the equal to the series of
numerators of the K + 1th row in reverse. The numerator of the last fractions becomes the
denominator of the first fractions. The numerators of the second to last fraction becomes the
denominator of the second fraction.

Example: Calculating the 6th row from the 5th row

5th row

6th row

The numerators of the left rectangle will be the numerators of the 5th row

1 2 3 3 4 5 5 4
|

The sum of the numerator and the denominator of all the fractions the 5th row becomes the
numerators of the fractions in the right rectangle of the 6th row.
1 2 3 3 4 5 5 4 5 7 8 7 7 8 7 5
|

Now filling out the denominators

1233455457877875
5787787545543321
 Pattern 2: Linking Fractions

The aim was to basically track all the fractions that have the same denominators and link them up.
There are many fractions like this, so I came up with a specific procedure: Start with any number n.
You have to locate the fractions that is equal to n (this will be n divided by 1). After doing so, locate a
fraction that has n in the numerator. There will be many fractions like this, but you have to locate
the fraction that is to the left of the previous fraction and that is the closest (in terms of distance) to
the previous fraction. You keep on repeating this.

An example: Let n be one. The fractions that we start with will be 1/1 as 1/1 equals 1. Now we have
to locate a fraction that has one in the numerator in the 3rd row that is left of the 1/1 fraction. This
fraction turns out to be ½. I continued to repeat this for different values of n. Here is a table of my
results:

n Fractions
1 1 1 1 1 1 1 1
, , , , , ,
1 2 3 4 5 6 7
2 2 2 2 2 2 2
, , , , ,
1 3 5 7 9 11
3 3 3 3 3 3
, , , ,
1 4 7 10 13
4 4 4 4 4
, , ,
1 5 9 13
5 5 5 5
, ,
1 6 11
6 6 6
,
1 7

I drew line cutting though all the fractions in the different rows. Then I marked the fractions that
were part of rule with blue dots. When I got rid of the fractional tree I ended up with this image:

Observation 1: If the gap between the first dot of a row and the edge is m, the distance between the
next pairs of dots are the powers of two. For example, in row 3, the distance between the first dot
and the second dot is twice the distance between the first dot and the left end. In the fourth this
goes one step further – the distance between the second and third dot is four times the distance
between the left end and the first dot. In the fifth row the distance between the fifth and fourth dot
is eight times the distance between the left side and the first dot.
Observation 2: When you pick fractions in this it leads to cool patterns. One of them is that the
fractions that have the same numerators has an interesting pattern in the denominator. If you list
out the fractions part of the procedure for a number n, you will see that the denominators form an
arithmetic sequence starting with one where the common difference is n. For example, if the value
of n is 2 the denominators form the odd number sequence.

Observation 3: I thought about adding up all the fractions part of the above procedure that are part
of the same row. One thing that I noticed was that the numerators are basically the natural
numbers, and the more rows you have the greater the numerator of the last fraction is. I have listed
down the fractions (by rows) below:

Row 1: 1
1
Row 2: 1 2
,
2 1
Row 3: 1 2 3
, ,
3 3 1
Row 4: 1 2 3 4
, , ,
4 5 4 1
Row 5: 1 2 3 4 5
, , , ,
5 7 7 5 1
Row 6: 1 2 3 4 5 6
, , , , ,
6 9 10 9 6 1
Row 7: 1 2 3 4 5 6 7
, , , , , ,
7 11 13 13 11 7 1

I wanted to find a summation notation that would give you the fractions in the same order. I applied
my knowledge from the Maths Study Circle for doing so. I paid attention to the denominators of the
fractions. I drew up a table with my results:

- 1 2 3 4 5 6 7

1 1 0 0 0 0 0 0

2 2 1 0 0 0 0 0

3 3 3 1 0 0 0 0

4 4 5 4 1 0 0 0

5 5 7 7 5 1 0 0

6 6 9 10 9 6 1 0

7 7 11 13 13 11 7 1

I could see a pattern in columns that I have mentioned earlier – the differences in the sequences
increase by one if you move to the right of the table. Because of this there would be some kind of
pattern in the formulas and there was:

1 2 3 4 5 6 7
n 2n – 1 3n – 2 4n – 3 5n - 4 6n - 5 7n - 6
 The general formula would be: 𝑐𝑛 − (𝑐 − 1). Where c is the column number.

But there was a catch there were zeros in before the actual start of the series. For example, the
normal odd number sequence is 1, 3, 5, 7, 9, 11… But the sequences now is 0, 1, 3, 5, 7, 9, 11. The
formula for the normal odd umber is 2n – 1 but the formula for the tweaked series would be 2(n-1) –
1. The number of zeros is equal to the columns number minus one.

The general formula would be

(𝑐)(𝑛 − (𝑐 − 1)) − (𝑐 − 1)

When I simplified the formula I got

−𝑐 2 + 𝑐𝑛 + 1 = 𝑐(𝑛 − 𝑐) + 1
This becomes the denominator of the fraction. The value of c will start at one and go on till the row
number r. The numerators just becomes column number. The user just has to input the row number
they want the fractions. The whole thing complied becomes:
𝑟
𝑐
∑
𝑐(𝑟 − 𝑐) + 1
𝑐=1

Pattern 3: Reciprocals

If you draw a line through the centre of the tree – you will have a line of symmetry. The only catch is
that the fractions on one side are the reciprocals of the other side.

This leads to the fact that the product of all the fractions in the row is always one. The sum of the
product of the reciprocals would be exactly half the number of fractions in a row.

Pattern 4: Mediants

If there are two fractions b and c, below a fraction, a, then the mediant of the fractions b and c wll
equal the fraction a. Let me take an example:

Here the mediant of the fractions 8 by 13 and 7 by 11 will equal 5 by 8.

8 7 8+7 15 5
, = = =
13 11 13 + 11 24 8

This is true not only fractions present in the row right above it. In fact, IF any fraction a, that is
directly above the centre of the of the two fractions b and c, a is equal to the mediant of b and c. Lat
me take another example:
The centre of the two fractions 5 by 6 and 6 by 5 is directly below the fraction 1/1. The mediant of
the fractions 5 by 6 and 5 by 6 will equal 1 by 1
5 6 5 + 6 11 1
, = = =
6 5 6 + 5 11 1
This is also true for the other fractions such as 6 by 7 and 7 by 6. This is why when you make a series
of mediants for each row (the mediant of Fraction 1 and Fraction 2, The mediant of Fraction 2 and 3
etc) you get a combination of all the fractions in the rows above it. Let me take an example.
1 2 3 3
I am going to try thing out with the 4th row: 3 , 3 , 2 , 1. Now I find the mediant sequence:

1+2 2+3 3+3 1 1 2

, , = , ,
3+3 3+2 2+1 2 1 1
1 2
The rows above the first rows are: 2 , 1 and 1 by 1. The mediant sequence is a combination of both of
the rows! I made the diagram below in an attempt to explain why its happens using the previous
conclusion:

Doe this have anything to do with that the mediant is exactly the mid mark between two fractions ?
Because in all the above examples with the tree, the mediant of the two the fractions, b and c,
always lies in between both the b and c, spatially?
 Row 4 (Mediant Series): 1 1 2
, ,
2 1 1
Row 5: 1 1 2 1 3 2 3
, , , , , ,
3 2 3 1 2 1 1
Row 6: 1 1 2 1 3 2 3 1 4 3 5 2 5 3 4
, , , , , , , , , , , , , ,
4 3 5 2 5 3 4 1 3 2 3 1 2 1 1

Now I replaced these fraction sequences with numbers – each fraction would be replaced by a
number that was its row number. The table then becomes:

Row 4: 323
Row 5: 4342434
Row 6: 545354525453545
Row 7 (predicted): 65646563656465626564656365
646562

I could find a patten in this type of arrangement. Starting from sequence that stand for row k, the
sequence for row k + 1 would be to add the number k between the number of the previous
sequence so that there is that number k to the left are right of the previous sequence.

Using the sequence of row 5 let’s try to crate the sequence for to 6. We have to add the number 5 in
between the digits for row 5 (and at the ends):

Row 5: 4 3 4 2 4 3 4

Row 6: 5 4 5 3 5 4 5 2 5 4 5 3 5 4 5

Pattern 5: Sum of the rows

I tried summing up all the fractions in each row to see weather there would be some pattern:

Row 2: 1/1
Row 3: 5/2
Row 4: 11/2
Row 5: 23/2
Row 6: 47/2

I could see a pattern – the denominators were always two throughout the rows. The next numerator
would always be the current one times two plus one. This was true for all the other sums except for
the first one 1 by 1. The recursive formula for the next fraction is this sequence would be

𝑅𝑜𝑤𝑛 = 𝑅𝑜𝑤𝑛−1 ∗ 2 + 1
Then I tried to compute a closed formula for the fractions- but I hit a problem – this sequence wasn’t
a k-constant sequence just like the powers of 2:

Sequence: 1, 5, 11, 23, 47, 95, 191

Difference 1: 4, 6, 12, 24, 48, 96
Difference 2: 2, 6, 12, 24, 48
Difference 3: 4, 6, 12, 24
Difference 3 is the same thing as difference 1! The difference will keep on oscillating between 4, 6,
12, 24, 48, 96 and 2, 6, 12, 24, 48. There is no closed formula for this sequence!

Pattern 6: Fibonacci Numbers

Fibonacci fractions:
1 1 2 3 5 8 13
, , , , , ,
1 2 3 5 8 13 21
Plotting them on the fraction tree and connected them up

The line moves left then right then left and then right. However, the lines length becomes smaller
and smaller so over time the spatial distance between consecutive Fibonacci fractions decreases.

I tried making another Fibonacci fraction sequence – this time picking alternate Fibonacci numbers:
1 1 2 3 5
, , , ,
2 3 5 8 13

Now the same thing occurs but the whole thing moves down to the right. I juxtaposed both the
sequences to get this figure. If you remove the first fraction and line you will get a symmetrical
shape:
 Then I skipped two Fibonacci numbers to make another sequence:
1 1 2 3 5 8
, , , , ,
3 5 8 13 21 34
I tried to map these fractions to tree but I could find the fraction 2 by 8!

Proving that the mediant lies in the middle of the two fractions.

I started with drawing the two sides and the diagonal of the parallelogram. In the diagram below the
point B is (c, d) and the point C is (a, b). The line “k” is the diagonal of the parallelogram.

Next I started to calculate the slopes of the lines. The slope of the red line can a over b or b over a
depending on what cartesian system you follow. If a is equal to the x coordinate, then the slope
would be b over a and if a is equal to the y coordinate then the slope would be a over b.
0 − 𝑎 −𝑎 𝑎
𝑆𝑙𝑜𝑝𝑒𝑟𝑒𝑑 = = =
0 − 𝑏 −𝑏 𝑏
The slope of the blue line would be d over c if c is the x coordinate and c over d if c is the y
coordinate.
0−𝑐 −𝑐 𝑐
𝑆𝑙𝑜𝑝𝑒𝑏𝑙𝑢𝑒 = = =
0 − 𝑑 −𝑑 𝑑
 The dashed lines indicate the y coordinates or the x coordinates of the points c and b.

Then I started to figure out the slope of the diagonal of the parallelogram. The slope of the diagonal
would be its y coordinate divided by its x coordinate. The y coordinate would be a + line q (see the
diagram).

The line p stands for the y coordinate of point D. Line p is equal to a which is the y coordinate of
point C (which is a) plus line q.

When I saw that line q was connected to line c (y coordinate of point B) via line g (red line) I was
tempted to connect points h and f. This line looked (and was – I measured the angle) parallel to the
red line and line q was parallel to line “c”. This set up formed a parallelogram. An important property
is that in a parallelogram opposite sides are equal! This meant that q was equal to c! Hence:

𝑦 =𝑎+𝑐
I did a similar thing to find the x coordinate of point D. The line “p” signifies the x cordite of the point
D. This is equal to the x coordinate of point B which is equal to d plus the line q.

Again, I did a similar thing – I connected points H and E to form the parallelogram ABHE. The side BH
or q will equal the x coordinate of the point ad which is b. Hence:

𝑥 =𝑏+𝑑
𝑎
• Slope red: 𝑏
𝑎+𝑐
• Slope diagonal: 𝑏+𝑑
𝑐
• Slope blue: 𝑑

I guessed that the slope of the red line is equal to fraction 1 and the slope of the blue line is equal to
fraction 2 and the diagonal is mediant of the two fractions (by definition). The diagonal lies right in
the middle of the two sides (a diagonal bisects the parallelogram). This is supposed to signify hat the
mediant is in the middle of the two fractions.

Using mediants to figure out which fraction is greater

I tried my best to solve this sum, But I couldn’t reach a reasonable conclusion. You can just find the
mediant of the two fractions and do the cross multiplication of the mediant and the two fractions
but you can do that with the original fractions too.