Small examples of cube diagrams of knots

Abstract

In this short note we highlight some of the differences between cube diagrams and grid diagrams. We also list examples of small cube diagrams for all knots up to 7 crossings and give some examples of links. Comment: 15 pages

Full text

SMALL EXAMPLES OF CUBE DIAGRAMS OF KNOTS
SCOTT BALDRIDGE AND BEN MCCARTY
Abstract. In this short note we highlight some of the differences between cube diagrams and grid
diagrams. We also list examples of small cube diagrams for all knots up to 7 crossings and give
some examples of links.
1. Introduction
Cube diagrams, introduced in Baldridge and Lowrance [2], are 3-dimensional representations of
knots or links. We define cube diagrams carefully in Section 3. The easiest way to imagine a cube
diagram is to think of an embedding of a knot or link in a [0, n]×[0, n]×[0, n] cube (using xyz
coordinates) for some positive integer nsuch that the knot projection of the cube to each axis plane
(x= 0, y= 0, and z= 0) is a grid diagram.
Figure 1: A cube diagram for the Trefoil.
The integer nis called the size of the cube diagram and we will refer to a cube diagram as
“small” if its size is close to the crossing number of the knot. Small cube diagrams are very useful
for computing and testing for knot invariants using computers—calculating certain invariants of
cube diagrams with size n > 20 can be computationally intractable for most computers. In this
paper we discuss why small cube diagrams can be elusive to find for a given knot or link and we give
examples of small cube diagrams (n < 20) for knots with seven crossings or less. In particular, the
Appendix gives pictures of the cube diagrams we have generated. A Mathematica program [3] can
be downloaded from http://cubeknots.googlecode.com that can be used to calculate invariants
of cube diagrams and also to rotate in 3-dimensions the examples in this paper.
Grid diagrams and cube diagrams are useful representations in knot theory. In 1996 Cromwell
used grid diagrams as ways to represent embeddings of knots in open books (c.f. [5]). He described
Date: July 30, 2009.
S. Baldridge was partially supported by NSF Grant DMS-0748636.
1
arXiv:0907.5401v1 [math.GT] 30 Jul 2009

2 S. BALDRIDGE AND B. MCCARTY
a set of elementary grid moves that preserve topological knot type that can be used to check for knot
invariants. More recently grid diagrams were used in constructing a combinatorial version of knot
Floer homology (c.f. [9] and [10]) and they provide a natural presentation of the front projection of
a Legendrian knot (c.f. [11] and [12]). Cube diagrams share many similarities with grid diagrams.
Like grid diagrams, there is set of elementary cube moves that preserve the knot type and there
is a combinatorial knot Floer homology that can be computed from a cube diagram [2]. However,
cube diagrams have also appeared independent of grid diagrams in the study of unstable Vassiliev
theory [8].
2. Finding small cube diagrams: the lifting problem
In order to find small cube diagrams, it is important to understand how they are different from
grid diagrams. In this section we discuss the problem of lifting a grid diagram to a cube diagram.
A grid diagram Gis an n×nsquare grid decorated with Xand Omarkings in such a way that
every row (resp. column) contains exactly one Xand one Omarking. To get an oriented knot or
link projection from a grid diagram one draws edges from Xto Oin each column and from Oto
Xin each row, taking the vertical segment as the over crossing at any intersection (cf. [2, 5]).
Figure 2: Grid diagram: X={2,3,4,5,1}and O={5,1,2,3,4}.
Since each of the three projections of a cube diagram is a grid diagram, it is natural to think
of a cube diagram as a lift of a grid diagram of one of the three projections. Such a lift is clearly
not unique. What is not obvious is that many times such lifts do not exist: a given grid diagram
does not necessarily lift to an embedding of a knot or link in a lattice in an n×n×ncube such
that the projection to each plane is a well-defined knot projection (it can lift to an embedding, but
the intersections in the other two projections are not necessarily isolated double points). For an
example of a grid diagram that can not be the projection of such a lattice knot see [2].
The lifting problem to a true cube diagram is harder than just finding a lift of a grid diagram in
3-space such that the other two projections are valid knot projections. Even when a grid diagram
does lift to such an embedding, that embedding is rarely a cube diagram and the chances that a
grid diagram lifts to a cube diagram appears to decrease as the grid size increases. To support
this assertion we wrote a brute-force program to get statistics on grid diagrams that lift to cube
diagrams. The program looked at all size n= 5,6,7,8 grid diagrams and tested whether each grid
diagram was a nontrivial knot (we did not consider links) and whether each grid diagram lifted to
a cube diagram. The data for grid diagrams for nontrivial knots are presented below.

SMALL CUBE DIAGRAMS 3
Grid Total grids Number that Percent
Size of nontrivial lift to cube
knots diagrams
5 10 3 30%
6 972 261 27%
7 85,022 19,722 23%
8 8,077,072 1,589,447 19.7%
Note that creating grid diagrams with sizes above 8 is time intensive and the computation to rule
out unknot grid diagrams grows quickly beyond the capabilities of most computers. For example,
the total number of grid diagrams of knots of size 8 including unknots is 101,606,400. It is interesting
to note that unknot grid diagrams do lift to cube diagrams more often: of those 101 million grid
diagrams, 72,109,568 of them or 71 percent lift to cube diagrams. But that fact is because many
unknot grid diagrams do not have any crossings and such grid diagrams often lift to cube diagrams.
Furthermore, the percentage of unknot grid diagrams that lift to cube diagrams also decreases as
the grid size increases in the examples we have calculated. To get the data for size 8 grid diagrams,
we ran 5 computers simultaneously night and day for 1 week. See the program [3] for other methods
of measuring the probability of finding a cube diagram when the size is greater than 8.
Therefore building a same-size cube diagram from a given grid diagram cannot be done in general.
However, a grid diagram can always be used to build a larger-sized cube diagram. In fact, given a
grid diagram of size nof a knot, a cube diagram of the same knot always exists of size at most
(2.1) n+ 2(# of bad crossings) + (# of twisted bends).
The bad crossings mentioned above are those crossings in the (y, z)−and (z, x)−projections that
do not follow the convention given in the definition of a cube diagram. For many grid diagrams,
this leads to large cube diagrams 3 or 4 times the size of the original grid. For example, a grid
diagram for a 7 crossing knot of size 9 may produce a cube diagram of size 35 or more. In this
paper we find cube diagram representations for many 7 crossing knots of size 9.
In order to derive the formula above and show how to improve upon it, we need a precise
definition of cube diagrams, which is the content of the next section.
3. Cube Diagrams of Knots
Let nbe a positive integer and let Γ be the cube [0, n]×[0, n]×[0, n]⊂R3thought of as
3-dimensional Cartesian grid, i.e., a grid with integer valued vertices. A flat of Γ is any cuboid
(a right rectangular prism) with integer vertices in Γ such that there are two orthogonal edges of
length nwith the remaining orthogonal edge of length 1. A flat with an edge of length 1 that is
parallel to the x-axis, y-axis, or z-axis is called an x-flat,y-flat, or z-flat respectively.
The embedding of a link in the cube Γ can be described as follows. A marking is a labeled point
in R3with half-integer coordinates. Mark unit cubes of Γ with either an X,Y, or Zsuch that the
following marking conditions hold:
•each flat has exactly one X, one Y, and one Zmarking;
•the markings in each flat forms a right angle such that each ray is parallel to a coordinate
axis;

4 S. BALDRIDGE AND B. MCCARTY
•for each x-flat, y-flat, or z-flat, the marking that is the vertex of the right angle is an X, Y,
or Zmarking respectively.
An oriented link can be embedded into Γ by connecting pairs of markings with a line segment
whenever two of their corresponding coordinates are the same. Each line segment is oriented to go
from an Xto a Y, from a Yto a Z, or from a Zto an X(note that the cube itself is canonically
oriented by the standard right hand orientation of R3). The markings in each flat define two
perpendicular segments of the link Ljoined at a vertex, call the union of these segments a cube
bend. If a cube bend is contained in an x-flat, we call it an x-cube bend. Similarly, define y-cube
bends and z-cube bends.
Arrange the markings in Γ so that the following crossing conditions hold:
•At every intersection point of the (x, y)-projection, the segment parallel to the x-axis has
smaller z-coordinate than the segment parallel to the y-axis.
•At every intersection point of the (y, z)-projection, the segment parallel to the y-axis has
smaller x-coordinate than the segment parallel to the z-axis.
•At every intersection point of the (z, x)-projection, the segment parallel to the z-axis has
smaller y-coordinate than the segment parallel to the x-axis.
If Γ satisfies these conditions, then it is called a cube diagram. We say that Γ is a cube diagram
representing the (oriented) link L.
The knot projections of a cube diagram to the three coordinate axis planes are grid diagrams.
What is not obvious, and the definition above specifies, is how the three projected grid diagrams
are oriented with respect to the cube diagram. In all three cases, the orientation is specified by
the order of the axes using the standard orientation of R3. For example, in the (x, y)-projection,
the x-axis specifies the ‘row’ and the y-axis specifies the ‘column’ of the grid diagram but in the
(y, z)-projection the y-axis specifies the ‘row’ and the z-axis specifies the ‘column’ of that grid
diagram (see Figure 3).
Figure 3: Crossing conditions for the projections
These orientations matter: if in the definition above, the first crossing condition was modified
so that the segment parallel to the x-axis has greater z-coordinate than the segment parallel to the
y-axis, then the set of knots types that are represented by modified ‘cube diagrams’ of a given size
is different than the set of knot types using the actual definition.

SMALL CUBE DIAGRAMS 5
4. Building a cube diagram from a given grid diagram
Section 2 described the problems with lifting a grid diagram to a cube diagram. However, a
grid diagram can always be used to create a cube diagram. There are two issues to overcome: (1)
changing the grid diagram so that it can be the knot-projection of a lattice knot such that the
other two projections are well-defined knot projections and (2) fixing the crossings of that lattice
knot in the other projections so that each crossing satisfies the crossing conditions.
The first issue can be fixed by removing all of the twisted bends from a grid diagram. A bend in
a grid diagram for a knot Kis a pair of segments in Kthat meet at a common Xor Omarking. If
a bend passes over some other segment of Kand passes under some other segment of K, then call
it twisted (cf. [2]). There are two ways to partition a grid diagram of a knot into non-overlapping
bends, depending on whether the two segments in each bend intersect in an Xor Omarking. A
grid diagram of size ncan be used to construct a lattice knot embedded into an n×n×ncube
such that the other projections are knot projections if a partial order can be put on either of the
two partitions of bends following the convention that if two bends cross the bend crossing over the
other is greater (cf. [2]). If a grid diagram has a partition with no twisted bends, then a partial
order always exists for that partition. If a partition has one twisted bend, then stabilizing at the
vertex of the twisted bend produces a new grid diagram with no twisted bends. The new grid is
one size larger, which explains the third term in the formula above when there are multiple twisted
bends.
If a grid diagram has no twisted bends, the bends can be stacked to form a lattice knot that
projects to valid knot projections in all three planes. While these knots do indeed satisfy the
crossing conditions for the (x, y)-projection they may not satisfy the crossing conditions in one
or both of the other projections. As observed in [2] the invalid crossings may be repaired by the
insertion of a rotated crossing as shown in figure 4 (c.f. [2]).
Figure 4: Insertion of a rotated crossing
The rotated crossing corrects the crossing condition for that projection while leaving the crossing
data of the other two projections unaffected. This procedure has the effect of rotating the crossing
so that the overcrossing is correct, but at the cost of increasing the size of the cube diagram by 2,
which explains the second term in the formula above.
5. Small cube diagrams for knots with small crossing knots
The algorithm in Section 4 tends to produces large cube diagram representations for a given knot
that are computationally intractable for computers to calculate invariants like knot Floer homology.
To find small cube diagrams for small crossing knots, we wrote a computer program that searches

6 S. BALDRIDGE AND B. MCCARTY
all size ≤9 grid diagrams looking for grid diagrams that lift to cube diagrams. If a grid diagram
does, we check to see what knot type it is and record the cube diagram. The idea is simple enough,
but virtually all grid diagrams that lift to cube diagrams are unknots, and checking whether each
grid diagram is possibly an unknot involves an O(n3) number of calculations per grid diagram of
size n. The total number of size ngrids (for knots and links) is
(n!)2
41 + 2(1 + n)Γ(1 + n, −1)
eΓ(2 + n)−6Γ(3,−1)
eΓ(4) ,
where Γ(s) is the gamma function and Γ(s, x) is the incomplete gamma function. Clearly, checking
whether each is an unknot is too time intensive for size 9 grids. We describe next how we reduced
the number of calculations to a routine that runs in days rather than years.
Generating grid diagrams of size ninvolves the choice of two size npermutations σ, τ ∈Sn. The
coordinates of the X-markings of the grid diagram are given by (i, σ(i)) and the O-markings are
given by (i, τ(i)). The program generates every possible grid diagram by an outer loop/inner loop
structure. The outer loop generates a new grid diagram by always starting with σ= (1,2,3, . . . , n)
(X’s along the diagonal) and cycling through the choices of τthat generate a grid diagram. The
inner loop then runs through all permutations of the columns using a routine that picks the ‘next
largest’ permutation in lexicographic order. For example, if n= 3, then the order of permutations
from ‘smallest’ to ‘largest’ is σ1= (1,2,3), σ2= (1,3,2), σ3= (2,1,3), σ4= (2,3,1), σ5= (3,1,2),
and σ6= (3,2,1). By setting up the inner and outer loop in this way we reduce the number of times
we need to check for the unknot significantly because the difference between the two diagrams from
σiand σi+1 is often either a (1) column commutation move or (2) a column cyclic permutation
move. Therefore if the grid diagram associated to σiis the unknot and σi+1 is a commutation or
cyclic permutation of σi, then we know that the grid diagram associated to σi+1 is also an unknot
(no time intensive calculations are necessary). Specifically, if the current grid diagram is merely a
commutation of the previous diagram and the knot determinant of the previous diagram is 1, then
the current diagram is also considered a potential unknot. It is reasonable to throw out determinant
1 knots—according to [4] there are just two nontrivial knots, 10124 and 12242, that have arc index
≤9 and determinant 1.
Once the program finds a grid diagram that potentially represents a nontrivial knot, the program
looks for several X−Oconfigurations known not to lift to a cube diagram including eliminating
links. These configurations are relatively easy to check for and also significantly reduce the number
of diagrams the program attempts to lift to cube diagrams. Next, the knot determinant is computed.
As before, if the determinant is 1, the diagram is discarded as the unknot, 10124 knot, or 12242
knot. The program then attempts to lift the grid to a valid cube by choosing a third permutation
ζ∈Snthat determines the order in which the z-cube bends will be stacked in order from smallest
z-coordinate to greatest. For each permutation the program checks to ensure that the stack is
compatible with the crossing conditions determined by the grid. If a valid stack permutation is
found, the crossing conditions are checked in the (y, z)−and (z, x)-projections. If the crossing
conditions are satisfied, the tuple of permutations (σ, τ, ζ ) is a cube diagram and the program
computes a variant of the Jones polynomial to determine exactly what knot the cube diagram
represents (c.f. [13]). This final and most time intensive calculation involves O(c22c) number of
calculations where cis the crossing number of the diagram. Fortunately, this calculation is not
necessary very often.
Using the routine sketched out above and beginning with grid diagrams of size ≤9 we obtained
most of the list in Appendix A. For the 6 crossing knots, however, a slightly different strategy

SMALL CUBE DIAGRAMS 7
was used. Beginning with a valid grid diagram for each knot type the cube stacking algorithm
described above was used to produce lattice knots that projected to valid knot projections in all
three planes. While these knots did indeed satisfy the crossing conditions for the (x, y)-projection
they all had invalid crossings in at least one of the other two projections. As observed above, the
invalid crossings may be repaired by rotating the crossing as shown in Figure 4. It is not known if
these examples are the smallest.
Appendix A. Knot Examples
The following table lists examples of cube knots up to 7 crossings as well as several 8, 9, 10
and 12 crossing knots. Each picture displays the cube knot from the point of view of the (x, y)-
projection. The code presented to the right of each diagram is designed to work with a Mathematica
notebook found at [3]. Since projections of cube diagrams are grid diagrams, the arc index α(K)
of a knot gives a lower bound for the size of a cube diagram. It has been shown in [1] that for
alternating knots α(K) = c(K) + 2 where c(K) is the crossing number of the knot. Therefore in
searching for cube diagrams up to size 9 the program could only be expected to find alternating
knots up to 7 crossings (one exception below, 815, was found by running a partial search for size 10
diagrams). However, for non-alternating knots the arc index may be much smaller. This fact helps
to explain why some of the 9, 10 and 12 crossing knots show up in the table below. The examples
that were found were of knots with relatively low arc index. For example 12591 listed below is a
non-alternating knot that has arc index equal to 9 (see [4]).
It is interesting to note that for the 6-crossing knots, 76, and 77the cube diagrams were obtained
using the second method described above. This method was necessary because the program found
no valid cube diagrams of size ≤9.
Diagram Mathematica Code
K31={X[{1, 5, 4},{4, 3, 2},{5, 4, 3},{2, 1, 5},{3, 2, 1}],
Y[{1, 5, 1},{2, 1, 2},{3, 2, 3},{4, 3, 4},{5, 4, 5}],
Z[{1, 2, 1},{2, 3, 2},{3, 4, 3},{4, 5, 4},{5, 1, 5}]}
31
K41={X[{1, 1, 2},{2, 2, 4},{3, 3, 3},{4, 7, 5},{5, 8, 6},
{6, 4, 8},{7, 5, 7},{8, 6, 1}],
Y[{1, 1, 8},{2, 2, 1},{3, 3, 6},{4, 7, 4},{5, 8, 7},
{6, 4, 5},{7, 5, 2},{8, 6, 3}],
Z[{1, 5, 2},{2, 7, 4},{3, 6, 3},{4, 4, 5},{5, 3, 6},
{6, 1, 8},{7, 8, 7},{8, 2, 1}]}
41

8 S. BALDRIDGE AND B. MCCARTY
Diagram Mathematica Code
K51={X[{1, 7, 1},{2, 1, 2},{3, 2, 3},{4, 3, 4},{5, 4, 5},
{6, 5, 6},{7, 6, 7}],
Y[{1, 7, 6},{2, 1, 7},{3, 2, 1},{4, 3, 2},{5, 4, 3},
{6, 5, 4},{7, 6, 5}],
Z[{1, 2, 1},{2, 3, 2},{3, 4, 3},{4, 5, 4},{5, 6, 5},
{6, 7, 6},{7, 1, 7}]}
51
K52={X[{2, 1, 6},{1, 7, 7},{3, 2, 1},{4, 4, 3},{5, 5, 4},
{6, 3, 2},{7, 6, 5}],
Y[{2, 1, 2},{1, 7, 3},{3, 2, 4},{4, 4, 5},{5, 5, 6},
{6, 3, 7},{7, 6, 1}],
Z[{2, 3, 2},{1, 4, 3},{3, 5, 4},{4, 6, 5},{5, 1, 6},
{6, 7, 7},{7, 2, 1}]}
52
K61={X[{1, 16, 7},{2, 1, 8},{3, 13, 11},{4, 5, 14},{5, 15, 13},
{6, 14, 10},{7, 7, 9},{8, 12, 12},{9, 9, 3},{10, 2, 4},
{11, 3, 2},{12, 4, 5},{13, 6, 6},{14, 8, 1},{15, 10, 15},
{16, 11, 16}],
Y[{1, 16, 14},{2, 1, 2},{3, 13, 13},{4, 5, 12},{5, 15, 15},
{6, 14, 8},{7, 7, 11},{8, 12, 10},{9, 9, 6},{10, 2, 7},
{11, 3, 3},{12, 4, 4},{13, 6, 16},{14, 8, 5},{15, 10, 1},
{16, 11, 9}],
Z[{1, 2, 7},{2, 14, 8},{3, 7, 11},{4, 16, 14},{5, 13, 13},
{6, 12, 10},{7, 11, 9},{8, 5, 12},{9, 3, 3},{10, 4, 4},
{11, 1, 2},{12, 8, 5},{13, 9, 6},{14, 10, 1},{15, 15, 15},
{16, 6, 16}]}
61
K62={X[{1, 9, 5},{2, 12, 8},{3, 1, 9},{4, 7, 3},{5, 4, 2},
{6, 3, 6},{7, 2, 1},{8, 5, 10},{9, 8, 11},{10, 10, 12},
{11, 6, 7},{12, 11, 4}],
Y[{1, 9, 7},{2, 12, 12},{3, 1, 5},{4, 7, 4},{5, 4, 1},
{6, 3, 3},{7, 2, 8},{8, 5, 6},{9, 8, 9},{10, 10, 10},
{11, 6, 2},{12, 11, 11}],
Z[{1, 1, 5},{2, 2, 8},{3, 8, 9},{4, 3, 3},{5, 6, 2},
{6, 5, 6},{7, 4, 1},{8, 10, 10},{9, 11, 11},{10, 12, 12},
{11, 9, 7},{12, 7, 4}]}
62

SMALL CUBE DIAGRAMS 9
Diagram Mathematica Code
K63={X[{1, 1, 4},{2, 3, 1},{3, 13, 9},{4, 2, 14},{5, 15, 13},
{6, 14, 5},{7, 8, 6},{8, 9, 7},{9, 12, 8},{10, 7, 3},
{11, 4, 2},{12, 6, 10},{13, 10, 11},{14, 11, 12},{15, 5, 15}],
Y[{1, 1, 14},{2, 3, 9},{3, 13, 13},{4, 2, 12},{5, 15, 15},
{6, 14, 8},{7, 8, 4},{8, 9, 5},{9, 12, 11},{10, 7, 1},
{11, 4, 7},{12, 6, 3},{13, 10, 6},{14, 11, 10},{15, 5, 2}],
Z[{1, 8, 4},{2, 7, 1},{3, 3, 9},{4, 1, 14},{5, 13, 13},
{6, 9, 5},{7, 10, 6},{8, 4, 7},{9, 14, 8},{10, 6, 3},
{11, 5, 2},{12, 11, 10},{13, 12, 11},{14, 2, 12},{15, 15, 15}]}
63
K71={X[{1, 9, 1},{2, 1, 2},{3, 2, 3},{4, 3, 4},{5, 4, 5},
{6, 5, 6},{7, 6, 7},{8, 7, 8},{9, 8, 9}],
Y[{1, 9, 8},{2, 1, 9},{3, 2, 1},{4, 3, 2},{5, 4, 3},
{6, 5, 4},{7, 6, 5},{8, 7, 6},{9, 8, 7}],
Z[{1, 2, 1},{2, 3, 2},{3, 4, 3},{4, 5, 4},{5, 6, 5},
{6, 7, 6},{7, 8, 7},{8, 9, 8},{9, 1, 9}]}
71
K72={X[{2, 1, 2},{1, 9, 3},{3, 2, 4},{4, 4, 5},{5, 5, 6},
{6, 6, 7},{7, 7, 8},{8, 3, 9},{9, 8, 1}],
Y[{2, 1, 8},{1, 9, 9},{3, 2, 1},{4, 4, 3},{5, 5, 4},
{6, 6, 5},{7, 7, 6},{8, 3, 2},{9, 8, 7}],
Z[{2, 3, 2},{1, 4, 3},{3, 5, 4},{4, 6, 5},{5, 7, 6},
{6, 8, 7},{7, 1, 8},{8, 9, 9},{9, 2, 1}]}
72
K73={X[{1, 1, 2},{2, 2, 3},{3, 3, 4},{4, 8, 5},{5, 4, 6},
{6, 5, 8},{7, 6, 7},{8, 7, 9},{9, 9, 1}],
Y[{1, 1, 9},{2, 2, 1},{3, 3, 2},{4, 8, 7},{5, 4, 3},
{6, 5, 5},{7, 6, 4},{8, 7, 6},{9, 9, 8}],
Z[{1, 3, 2},{2, 4, 3},{3, 6, 4},{4, 5, 5},{5, 7, 6},
{6, 9, 8},{7, 8, 7},{8, 1, 9},{9, 2, 1}]}
73

10 S. BALDRIDGE AND B. MCCARTY
Diagram Mathematica Code
K74={X[{1, 1, 4},{2, 5, 5},{3, 2, 2},{4, 3, 6},{5, 6, 7},
{6, 4, 8},{7, 7, 9},{8, 8, 3},{9, 9, 1}],
Y[{1, 1, 9},{2, 5, 3},{3, 2, 7},{4, 3, 1},{5, 6, 4},
{6, 4, 2},{7, 7, 6},{8, 8, 8},{9, 9, 5}],
Z[{1, 6, 4},{2, 9, 5},{3, 4, 2},{4, 7, 6},{5, 2, 7},
{6, 8, 8},{7, 1, 9},{8, 5, 3},{9, 3, 1}]}
74
K75={X[{1, 1, 2},{2, 2, 3},{3, 3, 4},{4, 9, 5},{5, 4, 6},
{6, 5, 9},{7, 6, 7},{8, 7, 8},{9, 8, 1}],
Y[{1, 1, 9},{2, 2, 1},{3, 3, 2},{4, 9, 8},{5, 4, 3},
{6, 5, 5},{7, 6, 4},{8, 7, 6},{9, 8, 7}],
Z[{1, 3, 2},{2, 4, 3},{3, 6, 4},{4, 5, 5},{5, 7, 6},
{6, 1, 9},{7, 8, 7},{8, 9, 8},{9, 2, 1}]}
75
K76={X[{1, 1, 2},{2, 14, 5},{3, 2, 6},{4, 3, 11},{5, 6, 7},
{6, 5, 12},{7, 4, 10},{8, 10, 3},{9, 11, 4},{10, 12, 9},
{11, 7, 8},{12, 9, 13},{13, 8, 14},{14, 13, 1}],
Y[{1, 1, 14},{2, 14, 7},{3, 2, 1},{4, 3, 5},{5, 6, 10},
{6, 5, 11},{7, 4, 13},{8, 10, 6},{9, 11, 2},{10, 12, 3},
{11, 7, 12},{12, 9, 9},{13, 8, 8},{14, 13, 4}],
Z[{1, 11, 2},{2, 3, 5},{3, 10, 6},{4, 5, 11},{5, 14, 7},
{6, 7, 12},{7, 6, 10},{8, 12, 3},{9, 13, 4},{10, 9, 9},
{11, 8, 8},{12, 4, 13},{13, 1, 14},{14, 2, 1}]}
76
K77={X[{1, 1, 7},{2, 2, 13},{3, 6, 8},{4, 5, 12},{5, 4, 9},
{6, 14, 6},{7, 8, 5},{8, 13, 11},{9, 7, 10},{10, 16, 14},
{11, 12, 15},{12, 15, 16},{13, 17, 17},{14, 10, 3},{15, 9, 4},
{16, 3, 2},{17, 11, 1}],
Y[{1, 1, 16},{2, 2, 1},{3, 6, 9},{4, 5, 13},{5, 4, 14},
{6, 14, 3},{7, 8, 7},{8, 13, 6},{9, 7, 12},{10, 16, 11},
{11, 12, 10},{12, 15, 8},{13, 17, 15},{14, 10, 2},{15, 9, 5},
{16, 3, 17},{17, 11, 4}],
Z[{1, 8, 7},{2, 5, 13},{3, 15, 8},{4, 7, 12},{5, 6, 9},
{6, 13, 6},{7, 9, 5},{8, 16, 11},{9, 12, 10},{10, 4, 14},
{11, 17, 15},{12, 1, 16},{13, 3, 17},{14, 14, 3},{15, 11, 4},
{16, 10, 2},{17, 2, 1}]}
77

SMALL CUBE DIAGRAMS 11
Diagram Mathematica Code
K815 ={X[{1, 1, 5},{2, 2, 2},{3, 3, 4},{4, 5, 6},{5, 7, 7},
{6, 8, 8},{7, 4, 9},{8, 6, 10},{9, 9, 3},{10, 10, 1}],
Y[{1, 1, 10},{2, 2, 8},{3, 3, 1},{4, 5, 3},{5, 7, 5},
{6, 8, 6},{7, 4, 2},{8, 6, 4},{9, 9, 7},{10, 10, 9}],
Z[{1, 7, 5},{2, 4, 2},{3, 6, 4},{4, 8, 6},{5, 9, 7},
{6, 2, 8},{7, 10, 9},{8, 1, 10},{9, 5, 3},{10, 3, 1}]}
815
K819 ={X[{1, 7, 2},{2, 1, 3},{3, 2, 4},{4, 3, 5},{5, 4, 6},
{6, 5, 7},{7, 6, 1}],
Y[{1, 7, 6},{2, 1, 7},{3, 2, 1},{4, 3, 2},{5, 4, 3},
{6, 5, 4},{7, 6, 5}],
Z[{1, 3, 2},{2, 4, 3},{3, 5, 4},{4, 6, 5},{5, 7, 6},
{6, 1, 7},{7, 2, 1}]}
819
K821 ={X[{1, 1, 3},{2, 2, 2},{3, 3, 4},{4, 5, 5},{5, 7, 6},
{6, 6, 8},{7, 10, 7},{8, 4, 9},{9, 8, 10},{10, 9, 1}],
Y[{1, 1, 8},{2, 2, 6},{3, 3, 1},{4, 5, 3},{5, 7, 7},
{6, 6, 4},{7, 10, 10},{8, 4, 2},{9, 8, 5},{10, 9, 9}],
Z[{1, 5, 3},{2, 4, 2},{3, 6, 4},{4, 8, 5},{5, 2, 6},
{6, 1, 8},{7, 7, 7},{8, 9, 9},{9, 10, 10},{10, 3, 1}]}
821
K949 ={X[{1, 9, 2},{2, 1, 3},{3, 8, 4},{4, 2, 5},{5, 3, 8},
{6, 4, 6},{7, 5, 9},{8, 6, 7},{9, 7, 1}],
Y[{1, 9, 8},{2, 1, 9},{3, 8, 7},{4, 2, 1},{5, 3, 4},
{6, 4, 2},{7, 5, 5},{8, 6, 3},{9, 7, 6}],
Z[{1, 4, 2},{2, 6, 3},{3, 3, 4},{4, 5, 5},{5, 9, 8},
{6, 7, 6},{7, 1, 9},{8, 8, 7},{9, 2, 1}]}
949

12 S. BALDRIDGE AND B. MCCARTY
Diagram Mathematica Code
K10124 ={X[{1, 1, 3},{2, 2, 4},{3, 3, 5},{4, 9, 6},{5, 4, 7},
{6, 5, 8},{7, 6, 9},{8, 7, 2},{9, 8, 1}],
Y[{1, 1, 9},{2, 2, 1},{3, 3, 2},{4, 9, 8},{5, 4, 3},
{6, 5, 4},{7, 6, 5},{8, 7, 6},{9, 8, 7}],
Z[{1, 4, 3},{2, 5, 4},{3, 6, 5},{4, 7, 6},{5, 8, 7},
{6, 9, 8},{7, 1, 9},{8, 3, 2},{9, 2, 1}]}
10124
K10128 ={X[{1, 9, 2},{2, 8, 3},{3, 1, 4},{4, 2, 5},{5, 3, 6},
{6, 4, 8},{7, 5, 9},{8, 6, 7},{9, 7, 1}],
Y[{1, 9, 8},{2, 8, 7},{3, 1, 9},{4, 2, 1},{5, 3, 3},
{6, 4, 4},{7, 5, 5},{8, 6, 2},{9, 7, 6}],
Z[{1, 6, 2},{2, 3, 3},{3, 4, 4},{4, 5, 5},{5, 7, 6},
{6, 9, 8},{7, 1, 9},{8, 8, 7},{9, 2, 1}]}
10128
K10139 ={X[{1, 1, 2},{2, 2, 4},{3, 9, 5},{4, 3, 6},{5, 4, 7},
{6, 5, 8},{7, 6, 9},{8, 7, 3},{9, 8, 1}],
Y[{1, 1, 9},{2, 2, 1},{3, 9, 8},{4, 3, 2},{5, 4, 3},
{6, 5, 4},{7, 6, 5},{8, 7, 6},{9, 8, 7}],
Z[{1, 3, 2},{2, 5, 4},{3, 6, 5},{4, 7, 6},{5, 8, 7},
{6, 9, 8},{7, 1, 9},{8, 4, 3},{9, 2, 1}]}
10139
K10145 ={X[{1, 1, 2},{2, 2, 4},{3, 9, 5},{4, 4, 6},{5, 3, 7},
{6, 5, 9},{7, 6, 8},{8, 7, 3},{9, 8, 1}],
Y[{1, 1, 9},{2, 2, 1},{3, 9, 8},{4, 4, 3},{5, 3, 2},
{6, 5, 5},{7, 6, 4},{8, 7, 7},{9, 8, 6}],
Z[{1, 3, 2},{2, 6, 4},{3, 5, 5},{4, 8, 6},{5, 7, 7},
{6, 1, 9},{7, 9, 8},{8, 4, 3},{9, 2, 1}]}
10145

SMALL CUBE DIAGRAMS 13
Diagram Mathematica Code
K10161 ={X[{1, 1, 2},{2, 2, 4},{3, 9, 5},{4, 4, 6},{5, 3, 7},
{6, 5, 8},{7, 6, 9},{8, 7, 3},{9, 8, 1}],
Y[{1, 1, 9},{2, 2, 1},{3, 9, 8},{4, 4, 3},{5, 3, 2},
{6, 5, 4},{7, 6, 5},{8, 7, 7},{9, 8, 6}],
Z[{1, 3, 2},{2, 5, 4},{3, 6, 5},{4, 8, 6},{5, 7, 7},
{6, 9, 8},{7, 1, 9},{8, 4, 3},{9, 2, 1}]}
10161
K12591 ={X[{1, 9, 2},{2, 1, 4},{3, 2, 3},{4, 3, 5},{5, 4, 6},
{6, 6, 7},{7, 7, 8},{8, 5, 9},{9, 8, 1}],
Y[{1, 9, 7},{2, 1, 9},{3, 2, 8},{4, 3, 1},{5, 4, 2},
{6, 6, 4},{7, 7, 5},{8, 5, 3},{9, 8, 6}],
Z[{1, 4, 2},{2, 6, 4},{3, 5, 3},{4, 7, 5},{5, 8, 6},
{6, 9, 7},{7, 2, 8},{8, 1, 9},{9, 3, 1}]}
12591

14 S. BALDRIDGE AND B. MCCARTY
Appendix B. Link Examples
Diagram Mathematica Code
HL = {X[{1, 4, 1},{2, 1, 2},{3, 2, 3},{4, 3, 4}],
Y[{1, 4, 3},{2, 1, 4},{3, 2, 1},{4, 3, 2}],
Z[{1, 2, 1},{2, 3, 2},{3, 4, 3},{4, 1, 4}]}
Hopf Link
D0(41) = {X[{1, 1, 16},{2, 2, 15},{3, 3, 1},{4, 4, 2},{5, 5, 12},
{6, 6, 11},{7, 14, 8},{8, 13, 7},{10, 15, 13},{9, 16, 14},
{11, 7, 10},{12, 8, 9},{14, 10, 4},{13, 9, 3},{15, 11, 5},
{16, 12, 6}],
Y[{1, 1, 4},{2, 2, 3},{3, 3, 7},{4, 4, 8},{5, 5, 6},
{6, 6, 5},{7, 14, 10},{8, 13, 9},{10, 15, 11},{9, 16, 12},
{11, 7, 15},{12, 8, 16},{14, 10, 14},{13, 9, 13},{15, 11, 2},
{16, 12, 1}],
Z[{1, 10, 4},{2, 9, 3},{3, 13, 7},{4, 14, 8},{5, 12, 6},
{6, 11, 5},{7, 7, 10},{8, 8, 9},{10, 6, 11},{9, 5, 12},
{11, 2, 15},{12, 1, 16},{14, 16, 14},{13, 15, 13},{15, 4, 2},
{16, 3, 1}]}
D0(41)
D−4(52) = {X[{1, 1, 4},{2, 2, 3},{3, 3, 7},{4, 4, 8},{5, 5, 5},
{6, 6, 6},{7, 8, 9},{8, 7, 10},{9, 11, 11},{10, 12, 12},
{11, 9, 13},{12, 10, 14},{13, 13, 2},{14, 14, 1}],
Y[{1, 1, 14},{2, 2, 13},{3, 3, 1},{4, 4, 2},{5, 5, 12},
{6, 6, 11},{7, 8, 4},{8, 7, 3},{9, 11, 8},{10, 12, 7},
{11, 9, 6},{12, 10, 5},{13, 13, 10},{14, 14, 9}],
Z[{1, 8, 4},{2, 7, 3},{3, 12, 7},{4, 11, 8},{5, 10, 5},
{6, 9, 6},{7, 14, 9},{8, 13, 10},{9, 6, 11},{10, 5, 12},
{11, 2, 13},{12, 1, 14},{13, 4, 2},{14, 3, 1}]}
D−4(52)
References
[1] Y. Bae, C. Park. An upper bound of arc index of links. Mathematical Proceedings of the Cambridge Philosophical
Society. Volume 129, Issue 3, Nov 2000, 491-500.
[2] S. Baldridge, A. Lowrance. Cube diagrams and a homology theory for knots. arXiv:0811.0225v1.
[3] S. Baldridge, A. Lowrance. Cube Knot Calculator, http://cubeknots.googlecode.com.
[4] J. C. Cha, C. Livingston. KnotInfo: Table of Knot Invariants, http://www.indiana.edu/∼knotinfo, July 10,
2009.
[5] P. Cromwell. Embedding knots and links in an open book. I. Basic properties. Topology Appl., 64 (1995), no. 1,
37–58.
[6] M. Culler. Gridlink. http://www.math.uic.edu/∼culler/gridlink/.
[7] W.B.R. Lickorish. An Introduction to Knot Theory. Springer, 1997.
[8] C. Giusti. Plumbers’ knots. arXiv:0811.2215v1.

SMALL CUBE DIAGRAMS 15
[9] C. Manolescu, P. Ozsvath, S. Sarkar. A combinatorial description of knot Floer homology. arXiv:math/0607691v2.
[10] C. Manolescu, P. Ozsvath, Z. Szabo, D. Thurston. On combinatorial link Floer homology. arXiv:math/0610559v2.
[11] P. Ozvath, Z. Szabo, D. Thurston. Legendrian knots, transverse knots and combinatorial Floer homology.
arXiv:math/0611841v2.
[12] L. Ng, D. Thurston. Grid Diagrams, Braids, and Contact Geometry. arXiv:0812.3665v2.
[13] L. Zulli. A matrix for computing the Jones polynomial of a knot. Topology, 34, 717-729.
Department of Mathematics, Louisiana State University
Baton Rouge, LA 70817, USA
E-mail address: sbaldrid@math.lsu.edu
Department of Mathematics, Louisiana State University
Baton Rouge, LA 70817, USA
E-mail address: benm@math.lsu.edu