Pentagon Tilings

In association with Vincent Kreugel, a Dutch architect (I detail at the end our respective contributions), we have been expanding on a pentagon tiling as apparently first devised by Marjorie Rice. Rice is famous for pentagon tilings, having found three new types of convex pentagons in the mid to late 1970s. An excellent article on the pentagon problem is Doris Schattschneider’s ‘Tiling the Plane with Convex pentagons’. Much of Rice’s investigations remain unpublished, in that only the product of her investigations are shown. How she devised these is not generally shown. However, some of her investigations are indeed shown in The Mathematical Gardner, a compilation of articles in hour of the late Martin Gardner, with Doris Schattschneider’s article ‘In Praise of Amateurs’ (mostly concerning background detail on Rice’s pentagon tiling findings), pages 140-166. Pages 154-155 contain numerous convex pentagon tilings, of which one of the tilings, or more precisely the tile, is the subject of this page. Unfortunately, the reproduction is at such a small scale that the annotations are barely discernible and so referring to these directly is difficult. Also, the drawings are hand drawn, and so matters of accuracy arise. To ease matters, for the sake of general convenience I show the Rice tilings below, of a type 1 convex pentagon (complete with shading for mirror image tiles), Fig. 1a and Fig. 1b. 1a correlates to Rice 4c, whilst 1b is the diagram to the extreme right of 4c (the text is illegible). Both these diagrams appear to be of the same pentagon (again, it’s difficult to be certain, for reasons as detailed above). Simply stated, two different stackings of the same pentagon by accident or design are shown (although no reference is made to this feature in the article).

Note that no claim is made as to completeness; indeed, I would not be surprised if there are other ‘configurations’ (made of larger blocks of greater unit thickness) below. The nature of the study militates against a systematic analysis of all possibilities. However, a start at least has been made in determining the basics, of which here I am on firmer ground; I believe that the basics are indeed in place.

Figure 3a--c: Strip tilings

Figure 3d-f Strip tilings

Type 2. An ‘extendible’, 180° block, in which a core 2-unit block can be added to extend the block. 

Figure 4a-c

Figure 5a-c

Somewhat curiously, it can be seen that Type 5a is the same as Type 7a. From this, one might expect that there would be an echo here of the respective types i.e. Type 5b is the same as 7b; 5c is the same as 7c, but this is not the case. Quite what explains this feature is unknown.

Figure 6a-c (Kreugel 6a)

Figure 7a-c

Figure 8a-c. See [2] diagram 21, which has similarities, albeit with reservations, due to the nature of the hand-drawn tiling which makes it difficult to be sure of a direct correlation

Figure 9: Non Extendible

Varied premise

A feature of the Rice type 4a is that the tiling can be varied, and not just ‘extended’ as above. A sloping column of the underlying 4-unit block is given, at left. Any of the three blocks in pink can be seen to tile, leading to infinity of tilings of any combination desired of this procedure.

Figure 10: Variation of Rice block placements

Extending block premise

Figure 11: Extending blocks

Many of the above can be seen to possess the 2-unit block in green, from which this can be inserted time and again, and so extend the block to infinity.

(6) Central block 'wraparound'

Figure 12a, b: 'Wraparounds' (Kreugel)

Note the curious spiral form that wraps around the inner block, the only instance of this type found. Although the spiral is admittedly 'weak', in that this is only apparent in a vertex to vertex arrangement (as against ‘strong’ in a tiling of contiguous tiles). Furthermore, as the 'inner core' tile can be extended, so can the spiral.

Figure 13: A 'novelty', of a combination of chevrons and spiral. (Kreugel)

Central tilings

Aside from the single pentagon, other tilings are possible with related polygons, with what I term as ‘standard’ size pentagon and ‘small’ size pentagons. These are based upon a central pentagon

Figure 14 a, b: Central tiling with inner pentagons 'tiny' and 'standard'

Possible Tilings

An open question is to the number of tilings possible; as such, I have not discussed how the above were formed. The most obvious process is to take a pentagon and find out all the possible ways of joining, as a 2-unit block, and then seeing which, if any of these tile.

In principle, this can be extended to, 3, 4… unit blocks. However, the number of tiles increases exponentially with each successive increase; even calculating by hand for 3- unit is only just practical but even so, I have not done so, as a rough calculation shows that the numbers are too large (never mind the actual drawing), possibly approaching 1000 blocks. I could do this, but the time involved is judged too excessive. Simply stated, such combinatory matters are a task for a computer. However, as alluded to above, I have indeed computed for the 2-unit instances, of which there are a set of 34 ‘all possibilities’, of which not all distinct; omitting the repeat instances gives a set of 21 distinct.

Figure 15: Set of 34, all possibilities

Figure 16: Set of 22, distinct

One could then continue with this core set, forming 3-unit blocks. However, as alluded to above, this is right at the limits of practicality, and of which I refrained from doing by hand; the time involved is simply too great. An alternative, quicker, more practical continuation is that of forming blocks with a likely tiling possibility that is forming tiles with parallel sides. To do this, all one has to do is take a member of set of 22, and ‘pair’ it in turn with the set. This is more ‘intuitive’, and can be done with speed and ease. Many blocks are formed that obviously will not tile, and so can in effect be discarded, but some do indeed form blocks of parallel sides that do at least offer the possibility of forming a tiling. Indeed, this is how the Type 2 tilings above were formed, of which it was later observed that they could be ‘extended’.

As can be seen, the construction of Rice’s pentagon tile above is most simple and straightforward, and surprisingly for such a simple, core value tile, not apparently been discussed before in the literature (has anyone seen this?). By any stretch of the imagination, this is an ‘interesting’ pentagon, given the numerous ways it tiles. An open question is to whether there are any other pentagons of a like ‘simple’ nature that may also tile in such ‘interesting’ ways, and if so, how to devise these? As alluded above, Rice’s pentagon is derived from a regular pentagon, and can be described of as a side produced and a side of a parallel line. As a variation of the theme, more pentagons can be devised by this ‘produce and parallel’ procedure. Upon taking a regular pentagon, producing and extending all the sides as according to these two rules, from an irregular star decagon as in the diagram as below. Other pentagons, with a little imagination, can be seen to be embedded here. Furthermore, in the search for even more pentagons, one can join the outer vertices, thus forming an irregular decagon diagram. Again, even more pentagons can be seen embedded here, of which I now investigate. As with Rice’s pentagon, I have not seen these discussed in the literature, has anyone seen these anywhere?

Figure 17: Pentagon forming from an irregular star decagon

From this, one can then ‘see’ a variety of pentagons, of which I show below

Figure 18: Set of 19 Pentagons with underlying construction.

Determining the exact number of possibilities, although it might be thought simple, is not a straightforward task; adopting a systematic process seems beyond me. Those I have found were largely by trial and error. However, I believe that this is the complete enumeration. Rice’s pentagon can be seen as Tile 1. As such, to the best of my recollections, I believe these others to be ‘new’. For interest’s sake I include the side ratios and angles below:

Figure 19: Set of 19 Pentagons with side ratios and angles

Now, do any of these pentagons tile, and if so are there any ‘interesting’ types? As such, the same procedure for finding tiles as above can be adopted, with 2-unit blocks etc. To a degree I have studied this, but as can be seen by my struggles above, unifying this satisfactorily is a Herculean task. This being so, I show a arbitrary selection of tilings.

Tile 2 (Type 1)

Figure 20: Three tilings

Tile 3 (Type 1)

Figure 21, a-b: Two tilings

At first glance these may appear identical, but they are indeed distinct

Tile 4 (Type 1)

Figure 22: One tiling

Tile 5 (Type *)

Figure 23: One tiling

Tile 6 (Type 1)

Figure 24: One tiling

Tile 7 Non Tiler

Tile 8 (Type 1)

Among the many of the tilings that have more than one placement here that can be described as ‘interesting’, tile 8 is perhaps of even more interest than most. Of note is that a core unit can tile in strips, b-d, and of which when rotated 180°, due to the asymmetric tile, can tile in an uncountable infinity of ways. Putting this into words in not a straightforward task, due to the intricacies involved. Much better is a picture, from which one can see the premise.

Figure 25, a-d: Four tilings and an infinity of others are possible 

Figure 25 e: Showing the strip premise

Tiles 9 and 10 Non tilers

Tile 11 (Type 1)

Figure 26: One tiling

Tile 12 (Type 1)

Figure 27: Two tilings

Tile 13 (Type 1)

Figure 28: Three tilings

Tile 14 (Type 1)

Figure 29: Two tilings

Tile 15 (Type 1)

Figure 29: Four tilings

Tile 16 (Type 1)

Figure 30: Two tilings

Tile 17 Non Tiler

Tile 18 (Type 1)

Tile 18 can be said to tile in essentially two different ways, Type 1, with a single instance, and Type 2, a-c, with ‘uncountable infinity’. The more ‘complex’ Type 2 is an example of the ‘uncountable infinity’ types of tilings, of which in this instance it can be seen that the tiles tile by means of a par hexagon, with special features. Simply stated, the tiles can be arranged ‘differently’ in the par hexagon, and so thus giving the possibility of uncountable infinity of tilings. Of this type, I show three examples, arguably with the greatest degree of structure. As alluded to above numerous trivial combinations are also possible, but are not shown.

Figure 31: Uncountable infinity of tilings; four 'core' examples shown

Tile 19 (Type 1)

Figure 32: Two tilings

Combining

As the sides have lengths in common, it is possible to combine certain of the tiles, Below I show an example, with tiles 1 and 2

Figure 33 Two tilings of Tiles 1 and 2 in combination

Matters of attribution

Seemingly as ever with matters concerning this tile, this is not a straightforward task! Although I have done much original and innovative work with this tile, and its spin–offs as above, the idea was not mine. The seed for my investigations arose as a result of correspondence with Vincent Kreugel, who sent me a pentagon tiling believing that he had discovered it, being unaware of Rice’s previous work. This took my interest, and upon researching it I found Rice’s own study. Kreugel’s findings include figures 6a, 11, 12, 13. Everything else is my own.

References

[1] Klarner, David A. editor. The Mathematical Gardner. Wadsworth Inc. 1981

See: ‘In Praise of Amateurs’, by Doris Schattschneider, pages 140-166 re Marjorie Rice and pentagons

[2] Schattschneider, Doris. ‘Tiling the Plane with Congruent Pentagons’. Mathematics Magazine Vol.1, 51, No.1 January 1978. 29-44.See page 32, Figure 3, diagram 20. Also see page 37 Figure 8, which repeats Rice’s tile, although the diagram is poorly drawn.

Created: 23 December 2013. Revised 28 January 2014. Revised 5 March 2014, with tiles (and tilings therof) 14-19 added