Monday, May 16, 2016

some tilings on regular grids

The only regular polygons that can tile the plane by themselves are equilateral triangles, squares, and regular hexagons (these are the regular tilings).


In the previous post, we noticed that the regular hexagon tiling can be used as a basis for creating decorative tilings of other polygons, such as nonagons, decagons, 15-gons, and so on, that include increasingly curved (but still polygonal) six-pointed stars to fill in the gaps, observing with the help of a formula form another post, that polygons that can be arranged around a regular hexagon like this have a number of sides equal to a multiple of 3.

15-gons on a regular hexagonal grid

We can do the same sort of thing with the regular triangle tiling, arranging hexagons, dodecagons, 18-gons, or any n-gon where n is a multiple of 6, and the gaps become increasingly-circular triangles.

And the same with the square tiling, arranging squares, octagons, dodecagons, 16-gons, or any n-gon where n is a multiple of 4, where the gaps becoming somewhat astroidal.

The polygons that can be arranged on the triangular, square and hexagonal grids can all be found by looking at the intersection of the green horizontal lines with the family of curves below, which as explained in an earlier post, describe a "regular polygons-in-rings formula" k = 2n/(n -(2+2m)).


The  regular polygon with the fewest number of sides that can be arranged this way on all of these grids is the dodecagon (some other decorative tilings from dodecagons here)


The image below shows where some of these tilings sit on the k = 2n/(n -(2+2m)) curve family.



Friday, May 13, 2016

a tiling of regular nonagons and stars


When I found this nice nonagon and star tiling, I couldn't get They Might Be Giants out of my head. The underlying symmetry comes from hexagons though:


The relationship described in the previous post tells us that the next regular polygon that can be put around a hexagon like this is the dodecagon (see the line k = 6):


And sure enough, these dodecagons can also form a nice hexagon-based tiling, with somewhat pointier stars.



The k = 6 line above gives us a whole family of polygon and star hexagonal tilings that follow the relationship n = 3m + 3 - the stars don't start to appear until m = 2



Friday, April 22, 2016

regular polygons in rings, part two

In the last two posts, I was playing around with rings of polygons, relying on a formula developed in this post from last summer. I realize that the explanation offered in that original post was not so great, so I am going to try and better explain what the "regular polygon ring formula" k = 2n/(n -(2+2m)) means, and mention a few more interesting patterns.

The first thing to notice, is that some regular polygons can be fitted together nicely in a ring where there is one edge (along the interior of the ring) between adjacent polygons, like these:

And some regular polygons can be fitted together skipping over two edges along the interior, like these:


However, some regular polygons cannot be put into such arrangements for a given skip count.



So, a reasonable question to ask is which regular polygons can be arranged in a ring by skipping over a fixed number of edges. More precisely, can k regular n-gons be place in a ring formed by skipping over m edges (along the inside of the ring) between adjacent n-gons?

Before answering that, it is good to know a couple of things about angles in regular n-gons. The angles at the vertex of a regular n-gon is equal to (n-2)pi/n. A nice proof of this involves cutting the n-gon into triangles that all share one vertex of the original n-gon: (2-n) triangles each contribute a total angle measure of pi, which is then divided evenly among the n vertices of the original n-gon. Also, if you form a wedge in the n-gon, from its center to two adjacent vertices, the angle formed will be 2pi/n. These two angle facts will help us come up with formulas for which n-gons can form the rings described above.


Let's start with just skipping one edge (= 1) when forming the ring. Like for the pentagons shown below.


Consider k n-gons arranged in a ring so that there 1 side is skipped between adjacent polygons.

If you connect the centers of the k n-gons, you obtain a bigger interior polygon, a regular k-gon.

On the one hand, since it is a regular k-gon, its interior angles are (k-2)pi/k (the first fact about angles in regular polygons mentioned above). On the other hand, if you look at the angle as it sits within each n-gon, you can see that the angle is formed by connecting the midpoints of two sides, one side apart, to the center. From this, you can see that this angle is also equal to 4pi/n (from the second fact about angles in regular polygons mentioned above).

Equating these two gives you the rule k = 2n/(n-4).



We want k and n that are integers, so we can graph the function  k = 2n/(n-4) and see what works. If we choose n = 4, things are undefined, which corresponds in our geometric model to an infinite ring of squares:


The only actual solutions we get are for n equal to 5, 6, 8, and 12. After this k is approaching the limit of 2, which geometrically corresponds to the fact that you can't make a ring with just two polygons.

What if we skip over two edges?



Here again, you form a regular polygon by connecting the centers of the k n-gons.

Again, since it is a regular k-gon, it will have interior angles of (k-2)pi/k.

However, now looking at this angle within the n-gon you can see it is formed by connecting the midpoints of two sides, two sides apart, to the center. So the angle is also 6pi/n.

Equating (k-2)pi/k and 6pi/n  gives you the rule k = 2n/(n-6).





For the "skip 2 edges" case, if we choose n = 6, we get an infinite ring of hexagons, and our function  k = 2n/(n-6) is undefined.


We can find which polygons work by finding the integer solutions to our rule. There are six polygons that work following the skip 2 edges rule of ring construction, as shown on the graph below. We can stop at the 18-gon, with it we've hit the minimum value of k = 3.



In general, we can try a "skip m edges" rule, and following the same reasoning, and note that the angles of the central polygon will be (2 + 2m)pi/n (see diagram below). And, as before, equating this with (k-2)pi/k, obtain the relationship k = 2n/(n - (2 + 2m)).

This also covers the case m = 0, where we make a ring skipping zero edges fitting the regular n-gons snugly together, and find (as maybe you knew already) that this can only be done for the triangle, the square, and the hexagon.

Looking at the graphs and solutions for several values of m, it looks as if there are some linear relationships between n and k that hold true for all m.



Playing connect the dots, there about six relationships that jump out right away: three lines sloping upwards, and three horizontal constant relationships (a fixed value for k), that slice through each m curve and give integer solutions.


For example, consider the line k = 2n; when substituted into the general formula k = 2n/(n - (2 + 2m)), this gives us n = 2m + 3. What does this mean? One way of putting it is, for any number of skips (m) we can always find an n-gon that can form a ring that is twice the size of n. Or, put another way, for any odd n, you can place n-gons in a ring double the size of n, and to do this you will have to skip over m = (n-3)/2 sides. Six triangles can form a ring skipping no sides (m=0), 10 pentagons can form a ring skipping one side, and fourteen heptagons can form a ring skipping two sides.


Similarly, the line k = n connects with all the m curves when n = 2m + 4, so for any even n, you can place n n-gons in a ring (with m = (n - 4)/2 skips), as described in this post.

The other relationships mentioned above (and there are more that were not mentioned) also reveal interesting patterns in how rings of regular polygons can be formed. Some questions you can look at would be: What values n allow you to create a ring of 3, 4, or 6 n-gons? Which "skip values" allow you to create a ring of 10 polygons?

Monday, April 18, 2016

rings of regular polygons in rings

As mentioned here and here, you can arrange regular polygons in regular rings.

If n is even, then n regular n-gons can be placed in a ring with each n-gon centered at the vertex of a another regular n-gon, where there are m = (n - 4)/2 edges between adjacent edges of the adjacent n-gons.

For example, squares can form a ring around a square, skipping no edges, while hexagons can form a ring around a hexagon, skipping one edge along the interior of the ring between adjacent hexagons.


Things look a bit more interesting for octagons and decagons, which due to the skipping of edges form a star shape.


So if a regular n-gon can form a ring around another regular n-gon, what happens if you make rings out of those (i.e. make a ring out of the red rings above)?

For n = 4 or 6, you get a tessellation - the small squares or hexagons will neatly fit up against each other and tile the plane (if you keep going). For n = 8, the octagons form patterns like this.


If we only show the initial small octagons, we get the pattern below, where the small octagons do not overlap, but form a nice pattern with various star-shaped holes.


We can only do one ring of rings of decagons before they start to overlap. In the ring of rings of decagons below, some edges of the secondary decagons were left in, for aesthetic effect.


Thursday, April 14, 2016

tetradecagons and heptagons

Taking another look at regular polygons in rings, here are some regular heptagons (7 sides) centered at the vertices of a regular tetradecagon (14 sides), and some regular tetradecagons centered at the vertices of a regular heptagon. The smaller polygons form rings where there are a fixed number of edges skipped between each pair of adjacent polygons.





When placing the heptagons around the tetradecagon, there are two edges between adjacent heptagons, and when placing the tetradecagons around the heptagon, there are four edges between the adjacent tetradecagons. In general, when placing polygons in rings like these, if you can place the regular n-gons  at the verticies of a regular k-gon, having adjacent n-gons sharing an edge while skipping over m edges in between if k = 2n/(n - (2+ 2m)).