Uniform partitions of 3-space, their relatives and embedding Thanks: This work was supported by the Volkswagen-Stiftung (RiP-program at Oberwolfach) and Russian fund of fundamental research (grant 98-01-00251).

Abstract

We review 28 uniform partitions of 3-space in order to find out which of them have graphs (skeletons) embeddable isometrically (or with scale 2) into some cubic lattice ${\bf Z}_{n}$. We also consider some relatives of those 28 partitions, including Achimedean 4-polytopes of Conway-Guy, non-compact uniform partitions, Kelvin partitions and those with unique vertex figure (i.e. Delaunay star). Among last ones we indicate two continuums of aperiodic tilings by semi-regular 3-prisms with cubes or with regular tetrahedra and regular octahedra. On the way many new partitions are added to incomplete cases considered here.

1 Introduction

A polyhedron is called uniform if all its faces are regular polygons and its group of symmetry is vertex-transitive. A normal partition of 3-space is called uniform if all its facets (cells) are uniform polyhedra and group of symmetry is vertex-transitive. There are exactly 28 uniform partitions of 3-space. A short history of this result follows. Andreini in 1905 proposed, as the complete list, 25 such partitions. But one of them ($13^{\prime}$, in his notation) turns out to be not uniform; it seems, that Coxeter [Cox35], page 334 was the first to realize it. Also Andreini missed partitions 25-28 (in our numeration given below). Till recent years, mathematical literature was abundant with incomplete lists of those partitions. See, for example, [Cri70], [Wil72] and [Pea78] (all of them does not contain 24-28) and [Gal89]. The first to publish the complete list was Grünbaum in [Grü94]. But he wrote there that, after obtaining the list, he realized that the manuscript [Joh91] already contained all 28 partitions. We also obtained all 28 partitions independently, but only in 1996.

We say that given partition $P$ has an $l_{1}$-graph and embeds up to scale $\lambda$ into the cubic lattice ${\bf Z}_{m}$, if there exists a mapping $f$ of the vertex-set of the skeleton of $P$ into ${\bf Z}_{m}$ such that

$$\lambda d_{P}(v_{i},v_{j})=||f(v_{i}),f(v_{j})||_{l_{1}}=\sum_{1\leq k\leq m}|f_{k}(v_{i})-f_{k}(v_{j})|\mbox{ for all vertices $v_{i},v_{j}$}.$$

The smallest such number $\lambda$ is called minimal scale; all embeddings below are given with their minimal scale.

Call an $l_{1}$-partition $l_{1}$-rigid, if all its embeddings (as above) into cubic lattices are pairwise equivalent, i.e. unique up to a symmetry of the cubic lattice. All embeddable partitions (and all embeddable tiles, except of Tetrahedron, having two non-equivalent embeddings: into $\frac{1}{2}H_{3}$ and $\frac{1}{2}H_{4}$) in this paper turn out to be $l_{1}$-rigid and so, by a result from [Shp93], having minimal scale 1 or 2. Those embeddings were obtained by constructing a complete system of alternated zones; see [CDG97], [DSt96], [DSt97], [DSt98].

The following 5-gonal inequality ([Dez60]):

$$d_{xy}+(d_{ab}+d_{ac}+d_{bc})\leq(d_{xa}+d_{xb}+d_{xc})+(d_{ya}+d_{yb}+d_{yc})$$

for distances between any five vertices $a,b,c,x,y$, is an important necessary condition for embedding of graphs. It turns out that all non-embeddable partitions and tiles, considered in this paper, are, moreover, not 5-gonal.

Denote by $De(T)$ and $Vo(T)$ the Delaunay and Voronoi partitions of 3-space associated with given set of points $T$. By an abuse of language, we will use the same notation for the graph, i.e. the skeleton of a partition. The Voronoi and Delaunay partitions are dual one to each other (not only combinatorially, but metrically). Denote by $P^{*}$ the partition dual to partition $P$; it should not be confounded with the same notation for dual lattice.

In Tables 1, 2 we remind the results (from [DSt96]) on embedding of uniform polyhedra and plane partitions. In the Table 1, ${d(P)}$ denotes the diameter of polyhedron $P$.

Table 1: Embedding of uniform polyhedra and their duals

Polyhedron $P$; face-vector	$\ell_{1}$-status of $P$	$d(P)$	$\ell_{1}$-status of $P^{*}$	$d(P^{*})$
Tetrahedron; (3.3.3)	$\to\frac{1}{2}H_{3}$	1	$\to\frac{1}{2}H_{4}$	1
Cube; (4.4.4)	$\to H_{3}$	3	$\to\frac{1}{2}H_{4}$	2
Dodecahedron; (5.5.5)	$\to\frac{1}{2}H_{10}$	5	$\to\frac{1}{2}H_{6}$	3
Cuboctahedron; (3.4.3.4)	non 5-gonal	3	$\to H_{4}$	4
Icosidodecahedron; (3.5.3.5)	non 5-gonal	5	$\to H_{6}$	6
Truncated tetrahedron; (3.6.6)	non 5-gonal	3	$\to\frac{1}{2}H_{7}$	2
Truncated octahedron; (4.6.6)	$\to H_{6}$	6	non 5-gonal	3
Truncated cube; (3.8.8)	non 5-gonal	6	$\to\frac{1}{2}H_{12}$	3
Truncated icosahedron; (5.6.6)	non 5-gonal	9	$\to\frac{1}{2}H_{10}$	5
Truncated dodecahedron; (3.10.10)	non 5-gonal	10	$\to\frac{1}{2}H_{26}$	5
Rhombicuboctahedron; (3.4.4.4)	$\to\frac{1}{2}H_{10}$	5	non 5-gonal	5
Rhombicosidodecahedron; (3.4.5.4)	$\to\frac{1}{2}H_{16}$	8	non 5-gonal	8
Truncated cuboctahedron; (4.6.8)	$\to H_{9}$	9	non 5-gonal	4
Truncated icosidodecahedron; (4.6.10)	$\to H_{15}$	15	non 5-gonal	6
Snub cube; (3.3.3.3.4)	$\to\frac{1}{2}H_{9}$	4	non 5-gonal	7
Snub dodecahedron; (3.3.3.3.5)	$\to\frac{1}{2}H_{15}$	7	non 5-gonal	15
$Prism_{3}$; (4.3.4)	$\to\frac{1}{2}H_{5}$	2	$\to\frac{1}{2}H_{4}$	2
$Prism_{n}$ (odd $n\geq 5$); (4.n.4)	$\to\frac{1}{2}H_{n+2}$	$\lfloor\frac{n+2}{2}\rfloor$	non 5-gonal	2
$Prism_{n}$ (even $n\geq 6$); (4.n.4)	$\to H_{\frac{n+2}{2}}$	$\frac{n+2}{2}$	non 5-gonal	2
$Antiprism_{n}$ ($n\geq 4$); (3.3.n.3)	$\to\frac{1}{2}H_{n+1}$	$\lfloor\frac{n+1}{2}\rfloor$	non 5-gonal	3

Table 2: Embedding of uniform plane tilings and their duals

Tiling $T$; face-vector	$\ell_{1}$-status of $T$	$\ell_{1}$-status of $T^{*}$
$Vo(Z_{2})=De(Z_{2})$; (4.4.4.4)	$\to Z_{2}$	$\to Z_{2}$
$De(A_{2})$; (3.3.3.3.3.3)	$\to\frac{1}{2}Z_{3}$	$\to Z_{3}$
$Vo(A_{2})$; (6.6.6)	$\to Z_{3}$	$\to\frac{1}{2}Z_{3}$
Kagome net; (3.6.3.6)	non 5-gonal	$\to Z_{3}$
(3.4.6.4)	$\to\frac{1}{2}Z_{3}$	non 5-gonal
truncated (4.4.4.4); (4.8.8)	$\to Z_{4}$	non 5-gonal
(4.6.12)	$\to Z_{6}$	non 5-gonal
truncated (6.6.6); (3.12.12)	non 5-gonal	$\to\frac{1}{2}Z_{\infty}$
chiral net; (3.3.3.3.6)	$\to\frac{1}{2}Z_{6}$	non 5-gonal
3,3,3,4,4	$\to\frac{1}{2}Z_{3}$	non 5-gonal
dual Cairo net; (3.3.4.3.4)	$\to\frac{1}{2}Z_{4}$	non 5-gonal

2 28 uniform partitions

In Table 3 of 28 partitions, the meaning of the column is:

1. the number which we give to the partition;

2. its number in [And05] if any;

3. its number in [Grü94];

4. a characterization (if any) of the partition;

5. tiles of partition and respective number of them in Delaunay star;

5^∗. tiles of its dual;

6. embeddability (if any) of partition;

6^∗. embeddability (if any) of its dual.

Notation $\frac{1}{2}{\bf Z}_{m}$ in columns 6, 6^∗ means that the embedding is isometric up to scale 2.

The notations for the tiles given in the Table 3 are: tr$P$ for truncated polyhedron $P$; $Prism_{n}$ for semi-regular n-prism; $\alpha_{3}$, $\beta_{3}$ and $\gamma_{3}$ for the Platonic tetrahedron, octahedron and cube; $Cbt$ and $Rcbt$ for Archimedean Cuboctahedron and Rhombicuboctahedron; $RoDo$, tw$RoDo$ and $RoDo-v$ for Catalan Rhombic Dodecahedron, for its twist and for $RoDo$ with deleted vertex of valency 3; $Pyr_{4}$ and $BPyr_{3}$ for corresponding pyramid and bi-pyramid; $BDS^{*}$ for dual bidisphenoid.

Remark that [Cox35] considered 12 of all 28 partitions; namely, No’s 8, 7, 18, 2, 16, 23, 9 denoted there as $t_{A}\delta_{4}$ for $A=\{1\}$, $\{0,1\}$, $\{0,2\}$, $\{1,2\}$, $\{0,1,2\}$, $\{0,1,3\}$, $\{0,1,2,3\}$, respectively, and No’s 6, 5, 20, 19, 17 denoted as $q\delta_{4}$, $h\delta_{4}$, $h_{2}\delta_{4}$, $h_{3}\delta_{4}$, $h_{2,3}\delta_{4}$.

Table 3. Embedding of uniform partitions and their duals.

$$\begin{array}[]{||l|l|l||l||lr|l||l|l||}\hline\cr 1&2&3&4&5&&5^{*}&6&6^{*}\\ \hline\cr No&&&&\mbox{tiles}&&\mbox{of dual}&\mbox{emb.}&\mbox{dual}\\ \hline\cr 1&1&22&De({\bf Z}_{3})=Vo({\bf Z}_{3})&\gamma_{3}&8&\gamma_{3}&{\bf Z}_{3}&{\bf Z}_{3}\\ 2&3&28&Vo(A_{3}^{*}=\mbox{bcc})&tr\beta_{3}&4&\sim\alpha_{3}&{\bf Z}_{6}&-\\ 3&4&11&De(A_{2}\times{\bf Z}_{1})&Prism_{3}&12&Prism_{6}&\frac{1}{2}{\bf Z}_{4}&{\bf Z}_{4}\\ 4&5&26&Vo(A_{2}\times{\bf Z}_{1})&Prism_{6}&6&Prism_{3}&{\bf Z}_{4}&\frac{1}{2}{\bf Z}_{4}\\ 5&2&1&De(A_{3}=\mbox{fcc})&\alpha_{3},\beta_{3}&8,6&RoDo&\frac{1}{2}{\bf Z}_{4}&{\bf Z}_{4}\\ 6&13&6&\mbox{F\"{o}ppl partition}&\alpha_{3},tr\alpha_{3}&2,6&\sim\gamma_{3}&-&{\bf Z}_{4}\\ 7&14&8&\mbox{boron CaB}_{6}&\beta_{3},tr\gamma_{3}&1,4&Pyr_{4}&-&-\\ 8&15&7&De(J\mbox{-complex})&\beta_{3},Cbt&2,4&\sim\beta_{3}&-&-\\ 9&22&27&&Prism_{8},&&\sim\alpha_{3}&{\bf Z}_{9}&-\\ &&&&trCbt&2,2&&&\\ 10&6&24&De((4.8^{2}\times{\bf Z}_{1})&Prism_{8},\gamma_{3}&4,2&\sim Prism_{3}&{\bf Z}_{5}&-\\ 11&8&18&De(3.6.3.6\times{\bf Z}_{1})&Prism_{3},&&\sim\gamma_{3}&-&{\bf Z}_{4}\\ &&&&Prism_{6}&4,4&&&\\ 12&12&17&De(3^{4}.6\times{\bf Z}_{1})&Prism_{3},&&\sim Prism_{5}&\frac{1}{2}{\bf Z}_{7}&-\\ &&&&Prism_{6}&8,2&&&\\ 13&11&13&De(3^{3}.4^{2}\times{\bf Z}_{1})&Prism_{3},\gamma_{3}&6,4&\sim Prism_{5}&\frac{1}{2}{\bf Z}_{4}&-\\ 14&11^{\prime}&14&De(3^{2}.4.3.4\times{\bf Z}_{1})&Prism_{3},\gamma_{3}&6,4&\sim Prism_{5}&\frac{1}{2}{\bf Z}_{5}&-\\ 15&7&19&De(3.12^{2}\times{\bf Z}_{1})&Prism_{3},&&\sim Prism_{3}&-&\frac{1}{2}{\bf Z}_{\infty}\\ &&&&Prism_{12}&2,4&&&\\ 16&18&25&\mbox{a zeolit}&\gamma_{3},tr\beta_{3},trCbt&1,1,2&\sim\alpha_{3}&{\bf Z}_{9}&-\\ 17&20&21&&tr\alpha_{3},tr\gamma_{3},&&\sim\alpha_{3}&-&-\\ &&&&trCbt&1,1,2&&&\\ 18&17&9&&\gamma_{3},Cbt,Rcbt&2,12&\sim BPyr_{3}&-&-\\ 19&16&5&&\alpha_{3},\gamma_{3},Rcbt&1,3,1&\sim BPyr_{3}&\frac{1}{2}{\bf Z}_{7}&-\\ 20&21&10&\mbox{boron UB}_{20}&tr\alpha_{3},tr\beta_{3},Cbt&2,1,2&\sim Pyr_{4}&-&-\\ 21&9&16&De(3.4.6.4\times{\bf Z}_{1})&\gamma_{3},Prism_{3},&&\sim\gamma_{3}&\frac{1}{2}{\bf Z}_{4}&-\\ &&&&Prism_{6}&4,2,2&&&\\ 22&10&23&De(4.6.12\times{\bf Z}_{1})&\gamma_{3},Prism_{3},&&\sim Prism_{3}&{\bf Z}_{7}&-\\ &&&&Prism_{12}&2,2,2&&&\\ 23&19&20&&\gamma_{3},Prism_{8},&&\sim Pyr_{4}&-&-\\ &&&&tr\gamma_{3},Rcbt&1,2,1,1&&&\\ 24&2^{\prime}&2&De(\mbox{hcp})&\alpha_{3},\beta_{3}&8,6&twRoDo&-&-\\ 25&-&3&De(\mbox{elong. }A_{3})&Prism_{3},\beta_{3},\alpha_{3}&6,4,3&RoDo-v&\frac{1}{2}{\bf Z}_{4}&{\bf Z}_{4}\\ 26&-&4&De(\mbox{elong. hcp})&Prism_{3},\beta_{3},\alpha_{3}&6,4,3&RoDo-v&-&-\\ 27&-&12&&Prism_{3}&12&BDS^{*}&\frac{1}{2}{\bf Z}_{5}&-\\ 28&-&15&De(\mbox{elong. 27})&Prism_{3},\gamma_{3}&6,4&\sim Prism_{5}&\frac{1}{2}{\bf Z}_{5}&-\\ \hline\cr\end{array}$$

Remarks to Table 3:

1. The partition 15^∗ is only one embeddable into ${\bf Z}_{\infty}$ (in fact, with scale 2).

2. All partitions embeddable with scale 1 are, except 25^∗, zonohedral. The Voronoi tile of 25^∗ is not centrally-symmetric. It will be interesting to find a normal tiling of 3-space embeddable with scale 1 such that the tile is centrally-symmetric; such non-normal tiling is given in [Sht80]: see item 35 in Table 4 below.

3. An embedding of tiles of tiling is necessary, but not sufficient, for embedding of whole tiling; for example, $26^{*}$ and $27^{*}$ are non-embeddable while their tiles are embeddable into $H_{4}$ and $\frac{1}{2}H_{8}$, respectively. In fact, all dual uniform partitions $P^{*}$ in Table 3, except of 24^∗, have no non-embeddable tiles. Among all 11 non-embeddable uniform partitions only items 11, 15, 24 and 26 have only embeddable tiles. The same is true for tilings 30, 33 and 32^∗, 33^∗, 34^∗, 46^∗ of Table 4.

4. Among all 28 partitions only No’s 1, 2, 5, 6, 8 have same surrounding of edges: polygons (4.4.4.4), (4.6.6), (3.3.3.3), (3.3.6.6), (3.3.4).

5. Partitions 8 and 24 are Delaunay partitions of lattice complexes: namely, a 3-lattice called J-complex and a bi-lattice hcp; the tile of $Vo$(J-complex) has form of jackstone (it explains the term ”J-complex”)) and it is combinatorially equivalent to $\beta_{3}$.

6. Partitions 1, 3, 5 are Delaunay partitions of lattices ${\bf Z}_{3}$, $A_{2}\times{\bf Z}_{1}$, $A_{3}$=fcc. Partitions 2 and 4 are Voronoi partitions of lattices $A_{3}^{*}$ =bcc and $A_{2}\times{\bf Z}_{1}$. No’s 10, 11, 12, 13, 14, 15, 21, 22 are Delaunay prismatic partitions over 8 Archimedean partitions of the plane; the embeddability of them and their duals is the same as in Table 2, but the dimension increases by 1.

7. Partitions 7 and 20 occur in Chemistry as borons CaB₆ and UB₂₀, respectively. Partition 20 occurs in zeolites.

8. The ratio of tiles in partition is 1:1 for 6, 7, 8, 10; 2:1 for 5, 11, 13, 14, 24, 28; 3:1 for 9; 8:1 for 12; 2:1:1 for 17, 19, 20; 3:1:1 for 16, 18; 3:2:1 for 21, 22, 25, 26; 3:3:1:1 for 23.

3 The table of other partitions

Table 4. Embedding of some other partitions.

$$\begin{array}[]{||l||l||l|l||l|l||}\hline\cr 29&De(L_{5})&\alpha_{3},Pyr_{4}&ElDo&\frac{1}{2}{\bf Z}_{4}&{\bf Z}_{5}\\ 30&De(D\mbox{-complex})&\alpha_{3},\sim\beta_{3}&\mbox{triakis tr}\alpha_{3}&\frac{1}{2}{\bf Z}_{5}&-\\ 31&De(\mbox{Kelvin})&\alpha_{3},\beta_{3}&RoDo,twRoDo&-&-\\ 32&De(\mbox{Gr\"{u}nbaum})&Prism_{3}&Prism_{6},BDS^{*}&\frac{1}{2}{\bf Z}_{5}&-\\ 33&De(\mbox{elong. Kelvin})&\alpha_{3},\beta_{3},Prism_{3}&RoDo-v&-&-\\ 34&De(\mbox{elong. Gr\"{u}nbaum})&Prism_{3},\gamma_{3}&\sim Prism_{5}&\frac{1}{2}{\bf Z}_{5}&-\\ 35&P(S_{1})&S_{1}&&{\bf Z}_{3}&-\\ 36&P(S_{2})&S_{2}&&{\bf Z}_{4}&-\\ 37&P(S_{3})&S_{3}&&{\bf Z}_{5}&-\\ 38&A\mbox{-19}&Prism_{\infty}&&{\bf Z}_{2}&\\ 39&&Prism_{\infty}&&{\bf Z}_{2}&\\ 40&A\mbox{-20, $n$ even}&C_{n}\times P_{\bf Z}&&{\bf Z}_{\infty}&\\ &A\mbox{-20, $n$ odd}&C_{n}\times P_{\bf Z}&&\frac{1}{2}{\bf Z}_{\infty}&\\ 41&A\mbox{-22}&Aprism_{\infty}&&\frac{1}{2}{\bf Z}_{3}&\\ 42&A\mbox{-23}&Prism_{\infty},Aprism_{\infty}&&\frac{1}{2}{\bf Z}_{3}&\\ 43&\parallel-type&\gamma_{3},C_{4}\times P_{\bf Z}&&{\bf Z}_{3}&\\ 44&\perp-type&\gamma_{3},C_{4}\times P_{\bf Z}&&{\bf Z}_{3}&\\ 45&chess-type&\gamma_{3},C_{4}\times P_{\bf Z}&&{\bf Z}_{3}&\\ 46&A\mbox{-13}^{\prime}&\alpha_{3},\mbox{tr}\alpha_{3}&R,\mbox{ twisted }R&-&-\\ \hline\cr\end{array}$$

In Table 4 we group some other relevant partitions. Here $L_{5}$ denotes a representative of the 5-th Fedorov’s type (i.e. by the Voronoi polyhedron) of lattice in 3-space and $ElDo$ denotes its Voronoi polyhedron, called elongated dodecahedron. Remaining four lattices appeared in Table 3 as No 1 =$De({\bf Z}_{3})=Vo({\bf Z}_{3})$, No 5 =$De(A_{3})$, No 2 =$Vo(A^{*}_{3})$, No 3 =$De(A_{2}\times{\bf Z}_{1})$, No 4 =$Vo(A_{2}\times{\bf Z}_{1})$. Remark that $De(L_{5})$ and $De(A_{2}\times{\bf Z}_{1})$ coincide as graphs, but differ as partitions.

In the notation $De$(Kelvin) below we consider any Kelvin packing by $\alpha_{3}$ and $\beta_{3}$ (in proportion 2:1) which is proper, i.e. different from the lattice $A_{3}$=fcc (face-centered lattice) and the bi-lattice hcp (hexagonal closed packing). Any proper Kelvin partition, as well as partition $13^{\prime}$ in [And05] (given as 46 in Table 4 and which Andreini wrongly gave as uniform one), have exactly two vertex figures. (The Voronoi tiles of tiling 46 are two rhombohedra: a rhombohedron, say, $R$, i.e. the cube contracted along a diagonal, and twisted $R$; both are equivalent to $\gamma_{3}$.) The same is true for Grünbaum partitions; see Section 5 below for those notions and items 32–34 of Table 4.

See Section 4 below for items 38–45 of Table 4. $D$-complex is the diamond bi-lattice; triakis tr$\alpha_{3}$ denotes truncated $\alpha_{3}$ with $Pyr_{3}$ on each its triangular faces. Partitions 29 and 30 from Table 4 are both vertex-transitive, but they have some non-Archimedean tiles: $Pyr_{4}$ for 29 and non-regular octahedron for 30.

The partitions 35, 36, 37 of Table 4 are all 3 non-normalizable tilings of 3-space by convex parallelohedron, which where found in [Sht80]. The polyhedra denoted by $S_{1}$, $S_{2}$ and $S_{3}$ are centrally symmetric 10-hedra obtained by a decoration of the paralelipiped. $S_{1}$ is equivalent to $\beta_{3}$ truncated on two opposite vertices. $P^{*}(S_{i})$ for $i=1,2,3$ are different partitions of 3-space by non-convex bodies, but they have the same skeleton, which is not 5-gonal.

4 Non-compact uniform partitions

The non-compact uniform partitions, introduced in subsections 19, 20, 22, 23 of [And05], will be denoted here A-19, A-20, A-22, A-23, respectively, and put in Table 4 as No’s 38, 40, 41, 42. Denote by $Prism_{\infty}$ ($Antiprism_{\infty}$) and $C_{n}\times P_{\bf Z}$ the $\infty$-sided prisms (antiprisms, respectively) and the cylinder on $C_{n}$.

A-19 is obtained by putting $Prism_{\infty}$ on ($4^{4}$) and so its skeleton is ${\bf Z}_{2}$. We add, as item 39, the partition which differs from 38 only by another disposition of infinite prisms under net ($4^{4}$), i.e. perpendicular to those above it.

A-20 is obtained by putting the cylinders on ($4^{4}$), A-22 by putting $Antiprism_{\infty}$ on ($3^{6}$) and A-23 by putting $Prism_{\infty}$ and $Antiprism_{\infty}$ on ($3^{3}$.$4^{2}$).

In subsection $20^{\prime}$ [And05] mentions also the partition into two half-spaces separated by some of 10 Archimedean (and one degenerated) nets, i.e. uniform plane partitions. We can also take two parallel nets (say, $T$) and fill the space between them by usual prisms (so, the skeleton will be direct product of the graph of $T$ and $K_{2}$) or, for $T=(4^{4})$ or $(3^{3}.4^{2})$, by a combination of usual and infinite prisms. Similar uniform partitions are obtained if we will take an infinite number of parallel nets $T$.

The partitions 43, 44, 45 of Table 4 differ only by the disposition of cubes and cylinders. In 43 the layers of cylinders stay parallel ($\parallel$-type); in 44 they are perpendicular to the cylinders of each previous layer. In 45 we see ($4^{4}$) as an infinite chess-board; cylinders stay on ”white” squares while piles of cubes stay on the ”black” ones.

By a decoration of $Prism_{\infty}$ in 38, 39, one can get other non-compact uniform partitions.

5 Almost-uniform partitions

Call a normal partition of the 3-space into Platonic and Archimedean polyhedra, almost-uniform if the group of symmetry is not vertex-transitive but all vertex figures are congruent. Grünbaum [Grü94] gave two infinite classes of such partitions and indicated that he do not know other examples. In our terms, they called elongated proper Kelvin and elongated proper Grünbaum partitions. Kelvin and Grünbaum partitions are defined uniquely by an infinite binary sequence characterizing the way how layers follow each other. In Kelvin partition, the layers of $\alpha_{3}$ and $\beta_{3}$ follow each other in two different ways (say, $a$ and $b$) while in Grünbaum partition the layers of $Prism_{3}$ follow each other in parallel or perpendicular mutual disposition of heights. Unproper Kelvin partitions give uniform partitions 5 and 24 for sequences $...aaa...$ (or $...bbb...$) and $...ababab...$, respectively. Proper Kelvin and Grünbaum partitions are not almost-uniform; there are even $\infty$-uniform ones (take a non-periodic sequence).

Consider now elongations of those partitions, i.e. we add alternatively the layers of $Prism_{3}$ for Kelvin and of cubes for Grünbaum partitions.

Remark that $RoDo-v$ (the Voronoi tile of partitions 25, 26, 33) can be seen as a half of $RoDo$ cut in two, and that tw$RoDo$ is obtained from $RoDo$ by a twist (a turn by $90^{o}$) of two halves.The Voronoi tiles for proper Kelvin partition 31 are both $RoDo$ and tw$RoDo$ while only one of them remains for two unproper cases 5, 24. Similarly, the Voronoi tile of 34 (a special 5-prism) can be seen as a half of $Prism_{6}$ cut in two, and $BDS^{*}$ can be seen as twisted $Prism_{6}$ in similar way. The Voronoi tiles for proper Grünbaum partition 32 are both $Prism_{6}$ and $BDS^{*}$ while only one of them remains for two unproper cases 3, 27.

Besides of two unproper cases 25, 26 (elongation of uniform 5, 24) which are uniform, we have a continuum of proper elongated Kelvin partitions (denoted 33 in Table 4) which are almost-uniform. Among them there is a countable number of periodic partitions corresponding to periodic $(a,b)$-sequences. Remaining continuum consists of aperiodic tilings of 3-space by $\alpha_{3}$, $\beta_{3}$, $Prism_{3}$ with very simple rule: each has unique Delaunay star consisting of 6 $Prism_{3}$ (put together in order to form a 6-prism), 3 $\alpha_{3}$ and 3 $\beta_{3}$ (put alternatively on 6 triangles subdividing the hexagon) and one $\alpha_{3}$ filling remaining space in the star. Each $b$ in the $(a,b)$-sequence, defining such tiling, corresponds to the twist interchanging 3 $\alpha_{3}$ and 3 $\beta_{3}$ above (i.e. to the turn of the configuration of 4 $\alpha_{3}$, 3 $\beta_{3}$ by $60^{o}$).

Similar situation occurs for elongated Grünbaum partitions. Besides two uniform unproper cases 13, 28 (elongation of uniform 3, 27), we have a continuum of proper elongated Grünbaum partitions (denoted 34 in Table 4) which are almost-uniform. The aperiodic $(a,b)$-sequences give a continuum of aperiodic tilings by $Prism_{3}$, $\gamma_{3}$ with similar simple rule: unique Delaunay star consisting of 4 $\gamma_{3}$ (put together in order to form a 4-prism), 4 $Prism_{3}$ put on them and 2 $Prism_{3}$ filling remaining space in the star. Each $b$ in $(a,b)$-sequence, defining the tiling, corresponds to a turn of all configuration of 6 $Prism_{3}$ by $90^{o}$.

6 Archimedean 4-polytopes

Finite relatives of uniform partitions of 3-space are 4-dimensional Archimedean polytopes, i.e. those having vertex-transitive group of symmetry and whose cells are Platonic or Archimedean polyhedra and prisms or antiprisms with regular faces. [Con65] enumerated all of them:

1) 44 polytopes (others than prism on $\gamma_{3}$) obtained by Wythoff’s kaleidoscope construction from 4-dimensional irreducible reflection (point) groups;

2) 17 prisms on Platonic (other than $\gamma_{3}$) and Archimedean solids;

3) Prism on $Antiprism_{n}$ with $n>3$;

4) A doubly infinity of polytopes which are direct products of two regular polygons (if one of polygons is a square, then we get a prisms on 3-dimensional prisms);

5) Gosset’s semi-regular polytope called snub 24-cell;

6) A new polytope, called Grand Antiprism, having 100 vertices (all from 600-cell), 300 cells $\alpha_{3}$ and 20 cells $Antiprism_{5}$ (those antiprisms form two interlocking tubes).

Using the fact that the direct product of two graphs is $l_{1}$-embeddable if and only if each of them is, and the characterization of embeddable Archimedean polyhedra in [DSt96] (see Table 1 above), we can decide on embeddability in cases 2)–4). In fact, the answer is ”yes” always in cases 2)–4), except prisms on tr$\alpha_{3}$, tr$\gamma_{3}$, Cuboctahedron, truncated Icosahedron, truncated Dodecahedron and Icosidodecahedron, which all are not 5-gonal.

Now, the snub 24-cell embeds into $\frac{1}{2}H_{12}$ and the Grand Antiprism (as well as 600-cell itself) violates 7-gonal inequality, which is also necessary for embedding (see [Dez60], [DSt96]).