The addition of nimbers is the exclusive or of non-negative integers in binary representation.
It is closed under finite sets $~\{0,...,2^n-1\}~$ and the infinite set $~\{0,1,2,3,...\}~$.
The finite groups are isomorphic to elementary abelian groups Z2n, and that's how they are called in the following.

## Z23

 With the Cayley table these are the 16 subgroups of Z23:

The number of elements by rank in the lattice   (graphically: by layer of the Hasse diagram)   is 1, 7, 7, 1.
This is row 3 in the triangle of 2-binomial coefficients. (Gaussian binomial coefficients)

The rows of these matrices can be read as 3-ary Boolean functions.
(Compare e.g. matrix number 3 and Boolean functions like 1001 0000.)
Only the matrices' top rows are used to identify the subgroups,
so each subgroup corresponds to one 3-ary Boolean function, and thus can be represented by a binarily colored cube:

 (The subgroups also correspond to elements of the Fano plane: )

The cubes in black ovals are essentially the same. The corresponding Boolean functions belong to the same big equivalence class.
There are 8 of 22 essentially different cubes. Their number by layer of the Hasse diagram is 1, 3, 3, 1.
This is row 3 in .

## Z24

With the Cayley table these are the 67 subgroups of Z24: File:Z2^4; subgroups list.svg
In this file the subgroups are ordered lexicographically, without respect to their rank.
The corresponding integer sequence is .

The lattice of subgroups can be represented in a Hasse diagram:

The number of elements by rank in the lattice is 1, 15, 35, 15, 1.
This is row 4 in the triangle of 2-binomial coefficients.

Each subgroup of Z24 corresponds to a 4-ary Boolean function, and thus could be represented by a binarily colored tesseract.
One may ask, whether two such tesseracts are essentially the same or not,
i.e. if one can be turned into the other by any rotation of the tesseract.
But the answer is less easy than for the 3-dimensional case.

All binarily colored tesseracts that can be turned into each other form a big equivalence class.
The files linked in the following table list the corresponding 4-ary Boolean functions in the big equivalence classes.

16 N(4,1)1 N(3,1)4 N(3,2)6 N(3,3)4 N(3,4)1 N(2,1)6 N(2,2)12 N(2,3)4 N(2,4)4 N(2,5)3 N(2,6)6 N(1,1)4 N(1,2)6 N(1,3)4 N(1,4)1 N(0)1

(The table is organized upside down, so the rows are arranged like the layers in an Hasse diagram.)

There are 16 of 402 equivalence classes. Their number by rank in the lattice is 1, 4, 6, 4, 1.
This is row 4 in .

 All (4) = 307 functions in these equivalence classes (in a matrix like this one) The respective matrices for smaller sona are the 16x16, 4x4, 2x2 and 1x2 submatrices. All (4) = 67 functions that correspond to entries of These are the entries in the odd columns of the matrix on the left.

## Z25

Z25 has 1 + 31 + 155 + 155 + 31 + 1 = 374 subgroups of order 1, 2, 4, 8, 16, 32.

The number of equivalence classes is 25 =32.

Equivalence classes belonging together as counterparts:

 N(0) N(5,01)  N(1,01) N(4,01) N(1,02) N(4,02) N(1,03) N(4,03) N(1,04) N(4,04) N(1,05) N(4,05)  N(2,01) N(3,01) N(2,02) N(3,02) N(2,03) N(3,05) N(2,04) N(3,03) N(2,05) N(3,06) N(2,06) N(3,07) N(2,07) N(3,04) N(2,08) N(3,08) N(2,09) N(3,09) N(2,10) N(3,10) 

## Z26

This text refers to the table of all 2825 subgroups: Subgroups of Z2^6     (very large)

Z26 has 1 + 63 + 651 + 1395 + 651 + 63 + 1 = 2825 subgroups of order 1, 2, 4, 8, 16, 32, 64.

These are the (6) = 68 equivalence classes:

 N(6,01)1 N(5,01)6 N(5,02)15 N(5,03)20 N(5,04)15 N(5,05)6 N(5,06)1 N(4,01)15 N(4,02)60 N(4,03)60 N(4,04)30 N(4,05)6 N(4,06)20 N(4,07)45 N(4,08)90 N(4,09)60 N(4,10)60 N(4,11)15 N(4,12)15 N(4,13)90 N(4,14)10 N(4,15)60 N(4,16)15 N(3,01)20 N(3,02)90 N(3,03)60 N(3,04)15 N(3,05)60 N(3,06)90 N(3,07)180 N(3,08)60 N(3,09)60 N(3,10)90 N(3,11)15 N(3,12)60 N(3,13)60 N(3,14)20 N(3,15)20 N(3,16)90 N(3,17)15 N(3,18)90 N(3,19)45 N(3,20)45 N(3,21)180 N(3,22)30 N(2,01)15 N(2,02)60 N(2,03)20 N(2,04)60 N(2,05)45 N(2,06)90 N(2,07)30 N(2,08)60 N(2,09)60 N(2,10)90 N(2,11)6 N(2,12)15 N(2,13)15 N(2,14)10 N(2,15)60 N(2,16)15 N(1,01)6 N(1,02)15 N(1,03)20 N(1,04)15 N(1,05)6 N(1,06)1 N(0)1

(Subscripts show the number of subgroups in the equivalence class.)

Array with only the numbers of subgroups:

6 1 1 5 63 6 4 651 16 3 1395 22 2 651 16 1 63 6 0 1 1 1 6 15 20 15 6 1 15 60 60 30 6 20 45 90 60 60 15 15 90 10 60 15 20 90 60 15 60 90 180 60 60 90 15 60 60 20 20 90 15 90 45 45 180 30 15 60 20 60 45 90 30 60 60 90 6 15 15 10 60 15 6 15 20 15 6 1 1

In the table it can be seen that the binary string's Walsh spectra share the pattern of another binary string.
So each subgroup has a counterpart with the index number shown in column .
These pairs of counterpart subgroups usually belong to pairs of counterpart equivalence classes.
10 equivalence classes are their own counterparts. Among them is N(3,17), containing the only 15 subgroups that are their own counterparts.

 N(0) N(6,01)  N(1,01) N(5,01) N(1,02) N(5,02) N(1,03) N(5,03) N(1,04) N(5,04) N(1,05) N(5,05) N(1,06) N(5,06)  N(2,01) N(4,01) N(2,02) N(4,02) N(2,03) N(4,06) N(2,04) N(4,03) N(2,05) N(4,07) N(2,06) N(4,08) N(2,07) N(4,04) N(2,08) N(4,09) N(2,09) N(4,10) N(2,10) N(4,13) N(2,11) N(4,05) N(2,12) N(4,11) N(2,13) N(4,12) N(2,14) N(4,14) N(2,15) N(4,15) N(2,16) N(4,16)  N(3,01) N(3,02) N(3,03) N(3,05) N(3,04) N(3,11) N(3,06) N(3,07) N(3,08) N(3,12) N(3,09) N(3,13) N(3,10) N(3,16) N(3,14) N(3,15) N(3,17) N(3,18) N(3,19) N(3,20) N(3,21) N(3,22)