Navigation

Defining a Cairo-type Tiling

Under construction

How to exactly define a 'Cairo-type' tiling is not as straightforward a task as might otherwise be imagined. Although the basic framework is established by convention, with pentagons either (a) equilateral, or (b) the dual of the semi-regular 32. 4. 3. 4 tiling, this has no actuality in fact! (This situation is made all the more galling in that Macmillan’s 1978 article in the Mathematical Gazette, with first-hand experience, set out the conditions of the in situ tiling, from which everyone else has ignored, and chosen instead more ‘aesthetic’ pentagons, likely from Gardner’s 1975 article in Scientific American!)
Such matters thus require addressing. Therefore, I begin on firm foundations, with the premise of the in situ pentagon, which I describe as the ‘serving model’, and then I broaden the scope of the investigation. Simply stated, there is no one single Cairo pentagon tiling that can serve as the defining model. However, by general consensus, it can be described as of one symmetrical pentagon, with two opposite sides forming a 90° angle, of which when formed as a tiling when four pentagons are suitably combined form a subsidiary (or can be interpreted as) par hexagon. Furthermore, when shown as a tiling, the ‘subsidiary hexagons’ intersect at right angles. However, this feature is applied to a variety of pentagon tilings, such as an (a) equilateral, (b) the dual of the semi regular tiling, (c) and of a stick cross reflected in a square matrix. All these examples give a somewhat alike pentagon, at least at first glance (hence the general uncertainty). Somewhat ironically, the two that are most often given, (a) and (b) are not actually found in the in situ tilings!
Although the term ‘Cairo-type’ can be applied to all these, it is possible to vary the pentagon even more, and yet retain the attributes as described above. For example, even with just the one pentagon, one can have variations, with sides of different lengths, of four the same, and change the length of the base from ‘long’ to ‘short’ (but retaining the 90° condition. One can increase the core number of pentagons, with two, three, of four. One can relax the side length rations. One can relax the sides. One can relax the 90° angles. One can relax the symmetry condition, with asymmetric pentagons. One can relax the convexity of the par hexagon, with asymmetric par hexagons. One can even relax the definition somewhat, and include a type of pentagon that although strictly lacks the intersecting aspect, still retains much of a Cairo-type tiling.
An open question is how best to present the material. One could have different formats, as arising from the varied conditions. However, what seems the most obvious, and indeed adopted, is that of ordering as to the most basic condition, namely that of the number of pentagons involved. The following studies, grouped under 1, 2, 3, or 4 pentagons, are broadly some examples that arguably qualify as ‘Cairo-type’ tiling.


Serving Model


As such, given that the in situ tiling by its very nature must serve as the model, this is thus examined and described. The most obvious observation is that it consists of one symmetrical pentagon, with two opposite angles of 90°, with four sides the same length, the fifth side (base) being longer. Further to this, when viewed as a tiling, subsidiary convex hexagons intersect at right angles. A further feature of this is that of collinearity when the subsidiary hexagon is extended. However, the exact condition as regards to angles has still to be established; I have two possibilities, from Macmillan or myself, either of which would appear likely
As such, the studies are based on the premises above, albeit with relaxed conditions; the number of pentagons, symmetry, angles, the side lengths, subsidiary concave hexagons, all being unrestricted.



1 Pentagon


Given the in situ serving model, as detailed, I now analyse how the features of this can be ‘relaxed’, beginning with a single pentagon (larger number of pentagons are considered in succeeding chapters). Just for general reference, the properties of the serving model are:

One pentagon

Symmetrical

4, 1 sides (four short, one long)

Two 90° opposite angles

As a tiling, subsidiary convex par hexagons at right angles

As a tiling, lines of ‘nearest neighbour collinearity’

Special Cases

1. Equilateral pentagon (114° 18’, 131° 24’)


This possesses most of the properties of the serving model namely, symmetrical, two 90° opposite angles, subsidiary convex par hexagons at right angles. Lacking are the 4: 1 ratios and a long base and collinearity. However, although in normal circumstances such ‘omissions’ could be considered a drawback, in this instance it is not. Arguably, with the greater symmetry, this ‘improves’ upon the serving model, at least in an abstract sense.

Aesthetically this is very pleasing indeed, and arguably this is the ‘best’ pentagon in the series, not only here of the 1 pentagon category but the series as a whole, in that it is the most ‘basic’ of all the pentagon types, in that it possesses the most symmetry, namely of all five sides the same length, of which I consider the overriding issue as regards aesthetics. (The other examples in the series here can be described, at least in comparison, as weaker, with different side lengths). Indeed, this is my own personal favourite. Such aesthetics probably account for its popularity as the frequently stated serving model, likely from Gardner’s account. However, it does posses one relative shortcoming, in that it does not possess collinearity.

As can be seen, it is not necessarily a requirement that the sides be of different proportions, as in the 4, 1 of the serving model. However, this is very much the exception to the rule; this is the only 5, 0 pentagon possible.

As can be seen, aside from this, the subsidiary features remains the same, with two opposite 90° angles and convex par hexagons intersecting at right angles.

Also, of very much of minor importance, is that upon a ‘broad, casual glance’, it has the pleasing appearance of a ‘faux’ regular pentagon tiling, despite this being known not to tile.

2. Dual of the 32. 4. 3. 4 (120°)


This possesses most of the properties of the serving model, namely symmetrical, two 90° opposite angles, 4, 1 ratios, subsidiary convex par hexagons at right angles. Lacking is a long base and collinearity. Again, as above with the equilateral, likely on account of its aesthetics, many authors cite this as the defining model.

Aesthetically this is very pleasing indeed, and arguably, the second ‘best’ pentagon in the series, due to its underlying background, namely that of the dual of the semi-regular 32. 4. 3. 4 tiling. However, I consider that it pales aesthetically, in relative terms, as when compared with the equilateral, in that the major shortcoming is the sides lack those all of the same length. But that said, due to its underlying source it still has much aesthetic value. Furthermore, it has ‘pleasing’, round figure, ‘aesthetic’ interior angles, two of 90°, and three of 120°, something which the other ‘core value’ examples lack. In contrast to the equilateral pentagon, the sides not of the same length, with four of these being the same, with one decidedly ‘short’ in relation to the other four, described as of the ‘4, 1 short’ type (in contrast to ‘4, 1 long’, to others which is also possible).


3. Cordovan, by Buitrago and Iglesias (112.5°, 135°)


This possesses most of the properties of the serving model, namely symmetrical, two 90° opposite angles, 4, 1 sides, subsidiary convex par hexagons at right angles and a long base (the latter in relative terms). Lacking is collinearity of sides. However, it does indeed possess collinearity aspects, but of intersection rather than sides. When the sides are so produced it results in the ‘nearest neighbour’ pentagon in a like orientation intersecting at a nearest vertex.

Aesthetically, this is pleasing in regard as to the general resemblance to the serving mode, and is one of the strongest ‘round figure’ pentagons, and indeed in particular on account of its collinearity intersection. However, in matters of collinearity I still consider the side collinearity as a superior aspect.

Surprisingly, this particular pentagon is very little known, with apparently only Buitrago and Iglesias having written about this, in relation to their studies of the Cordovan proportion, hence the title.


4. Bailey (105°, 15°)


This possesses most of the properties of the serving model, namely symmetrical, two 90° opposite angles, 4, 1 sides, subsidiary convex par hexagons at right angles and a long base (the latter in relative terms). Lacking is collinearity of sides. However, it does indeed possess collinearity aspects, but of intersection rather than sides.

Aesthetically, this is pleasing in regard as to the general resemblance to the serving model, and is one of the strongest ‘round figure’ pentagons, and indeed in particular on account of its collinearity intersection. However, in matters of collinearity I still consider the side collinearity as a superior aspect.

This is somewhat alike in premise to the Cordovan example above, and essentially echoes its premise, in that it possesses the property not of collinearity of sides but of intersection, as when the sides are produced it results in the ‘nearest neighbour’ pentagon in a like orientation intersecting at a far vertex.

Amazingly, I believe this pentagons intersection properties is completely unknown, it arising as a result of my (hence my attribution) systematic search of pentagons with 5° increments. Certainly, I have not seen this quoted elsewhere.


5. Squared Intersections (108° 43’, 143° 13’)


This possesses most of the properties of the serving model, namely symmetrical, two 90° opposite angles, 4: 1 ratios, subsidiary convex par hexagons at right angles, collinearity, and a long base (the latter in relative terms). Indeed, it has all the same attributes!

Aesthetically, this is pleasing in regard as to the general resemblance to the serving model, and indeed is the strongest of its type; at a casual glance, they are indistinguishable. Indeed, as discussed elsewhere, this could possibly serve as the in situ model, with Macmillan having been mistaken. This example, based on the reflected stick figure premise, is remarkably like the in situ example, with convex pentagons, with near like angles (contrast 108° 26’, 143° 8’), same side proportions (4, 1), with collinearity, with convex subsidiary hexagons and a long base. Note that further examples of a similar stick cross premise/construction lack such close angles and this collinearity aspect.


6. Squared Intersections (135°)


Aesthetically, this is pleasing, but not in regard as to the general resemblance to the serving model; indeed, the pentagon is fundamentally different. This example, based on the reflected stick figure premise, with convex pentagons can be described as a special case, in that it possesses not two 90° angles, but three, with side ratios of 3, 2, and with convex subsidiary hexagons intersecting at right angles. Furthermore, and most interestingly, it is yet another instance of a one pentagon type possessing ‘nearest neighbour’ collinearity (and, I might just add, newly discovered at the amazingly late date of 19 October 2011! Are there yet others?). Another, pleasing aesthetic feature of this is that a Greek cross tiling can be discerned upon judicious selection, this further adding to its aesthetics.

This possesses some of the properties of the serving model, namely symmetrical, subsidiary convex par hexagons at right angles, collinearity. Lacking are the 4, 1 ratio, in that here the sides are 3, 2, and the two 90° angles, where here there are three.

The examples above can all be described as of angle considerations per se. Aside from such examples, one can also have examples based on unit intersections, which can have interesting connections, of which although these are also described as by angle, are, in general, more of a generic type.

Firstly, I show some examples of what can be determined as ‘special cases’, followed by ‘general cases’.



No.1
As can be seen, this is of pentagons, of a square matrix, of just one type, with a line of mirror symmetry, of which when four of them are suitably combined form a subsidiary (or can be interpreted as) par hexagon. Furthermore, when shown as a tiling, the hexagons intersect at right angles. Indeed, this intersecting aspect is very much at the heart (and accounts for its appeal, although this is not unique) of this type of tiling. This can serve as the basic model; however, it is possible to vary this theme (below, as well as examples that although is very much a Cairo-esque nature, is arguably not a true member).



2 Pentagons

Pentagons proportions: 2, 1, 1, 1
Subsidiary hexagons: 1 (one side par)
Of note is the one subsidiary pentagon, which has only one side parallel. Also of note is that for each subsidiary hexagon these are (likely by their nature) of an asymmetric distribution.

No.4
Pentagons proportions: 2, 1, 1, 1
Subsidiary hexagons: 2 (both par hexagons)
Here, the pentagons in each subsidiary hexagon are distributed in a symmetrical way.
Of note also of the subsidiary hexagons is that these differ in their ‘type’, with one consisting of two different (with four lengths of sides the same), whilst the other has three different sides.


3 Pentagons

No.1


4 Pentagons



Near Misses

All of the above so far can be considered as bona fide members of the Cairo pentagon type. However, there is another example whose inclusion/exclusion is somewhat more problematical. The example above displays all the intersecting features as above. However, the ‘subsidiary’ par hexagon aspect raises considerable queries. Here, we have a decagon, although in layman’s terms it is certainly very 'close' to a hexagon. Strictly, then, this is not a Cairo-type if a condition of par hexagons is defined. However, the premise remains of a ‘wobbly sided’, loose, par hexagon here. Consequently, this is decidedly different from the above. Should this be included as a Cairo-type tiling or not?

Subsidiary Decagons



Subsidiary Octagons

Created 24 December 2010. Last Updated: