I am doing a report on the Ising model, without going into the complex details of the model, in its two-dimensional form it involves a lattice of points.

Lattice of Points

Which, when periodic boundary conditions are imposed can be though of as on a torus.

For clarification, what I mean by "periodic boundary conditions" is that the right neighbour to a point in the far right column is a point in the far left column in the same row, and the bottom neighbour to a point in the bottom row is a point in the top row in the same column.


Therefore, what I want to do is plot a grid on a torus, and place dots on that grid.

I also eventually want to place arrows going from one point to the next on the torus, as I did here for the lattice:

Sweep through lattice

I'm just having a lot of trouble trying to put it all together. How can this be achieved with Tikz? If not with Tikz, then how can be achieved at all?

Thank you in advance for any answers or comments.

I've done a fair bit of research on drawing a torus:

And a few other things:

And if it helps, here is the source code to the drawings I made in this post:

    \clip (-1,-1) rectangle (6cm,6cm); 
    \draw[style=help lines,thick] (0,0) grid[step=.5cm] (5,5);

    \foreach \x in {0,1,...,10}
        \foreach \y in {0,1,...,10}
            \node[draw,circle,inner sep=2pt,fill] at (.5*\x,.5*\y) {};


    \clip (-1,-1) rectangle (6cm,6cm); 

    \foreach \x in {0,1,...,10}
        \foreach \y in {10,9,...,0}
            \node[draw,circle,inner sep=1pt,fill] at (.5*\x,.5*\y) {};
            \draw[thick,->] (.5*\x,.5*\y) -- (.5*\x+.4,.5*\y);
            \draw[thick,->] (.5*\x,.5*\y) -- (.5*\x,.5*\y-.4);

My only request is that graphics created in non-LaTeX applications like InkScape, Blender, etc not be suggested.

  • 6
    I think you really want to look into Asymptote for these kinds of 3D diagrams.
    – Jake
    Commented Dec 5, 2013 at 14:04
  • Mathematica is the best choice. Commented Dec 5, 2013 at 14:37
  • 2
    @DonutE.Knot I don't have access to Mathematica, nor the money to purchase it. Commented Dec 5, 2013 at 14:42
  • 3
    @DonutE.Knot Student + Thesis + Broken Laptop + New Apartment = not happening. Anyway, this is off topic. If you'd like to continue, we should do so in chat. Commented Dec 5, 2013 at 14:49
  • 2
    @DonutE.Knot I've been using Mathematica since version 3 and used to be a fan of it. However, for serious computations it is to slow and memory hungry. Graphics look nice on screen, but if you export they are huge or crap.
    – Alex
    Commented Dec 5, 2013 at 17:43

2 Answers 2


enter image description here

Edit A bug fixed (the outer equator midpoints was not calculated correctly, as pointed out by @Dror).

MWE with Asymptote, file lattice.asy:

import graph3;

pen surfPen=rgb(1,0.7,0);
pen xarcPen=deepblue+0.7bp;
pen yarcPen=deepred+0.7bp;


real R=2;
real a=1;

triple fs(pair t) {
  return ((R+a*Cos(t.y))*Cos(t.x),(R+a*Cos(t.y))*Sin(t.x),a*Sin(t.y));

surface s=surface(fs,(0,0),(360,360),8,8,Spline);

int m=20;
int n=10;
real arcFactor=0.85;

pair p,q,v;

for(int i=1;i<=n;++i){
  for(int j=0;j<m;++j){

Compile with asy -f pdf -noprc -render=4 lattice.asy to get a standalone lattice.pdf.

  • It looks great, the only reason I choose the other answer is because it looks more "scientific". Commented Dec 5, 2013 at 19:36
  • I get this mess when I render yours: s30.postimg.org/rijii354x/Page_1_torus_latice.jpg Commented Dec 6, 2013 at 2:46
  • @NictraSavios: The command should be asy -f pdf -noprc -render=4 lattice.asy as noted.
    – g.kov
    Commented Dec 6, 2013 at 5:52
  • Got it, I wasn't using the render command because it errored on my machine. Turns out Ubuntu's in-repository package doesn't support it, so I compiled from source and it worked. See here: tex.stackexchange.com/questions/148622/… Commented Dec 7, 2013 at 13:42
  • Very nice solution! Is there some artifact along the (outer) equator of the torus?
    – Dror
    Commented Dec 11, 2013 at 7:22

Are you thinking about something like this?



import graph3;
import three;


int nb = 20, ns = 10;
real rb = 5.0, rs = 2.0;

triple torus(pair z) {

  return ((rb + rs*cos(2*pi*z.x/ns))*cos(2*pi*z.y/nb),
      (rb + rs*cos(2*pi*z.x/ns))*sin(2*pi*z.y/nb),


surface site = scale3(0.1)*unitsphere;

for(int k1=0; k1<ns; ++k1) {
  for(int k2=0; k2<nb; ++k2) {


enter image description here

It's made with Asymptote.

  • 5
    What sorcery is this! It's... it's beautiful... if it weren't free, I'd force the developers to take my money. Heck, I'll do it anyway! "Give me a donate button for I'll show up at your house" I say! Thank you! Commented Dec 5, 2013 at 15:37
  • Really nice! Would it be very complicated to use round surfaces instead of planar ones (i.e. circular arcs for the arrows instead of straight lines)?
    – Jake
    Commented Dec 5, 2013 at 15:39
  • Can it do 2D graphics as well? Commented Dec 5, 2013 at 15:41
  • 1
    @Jake No, it's not complicated, have a look at the gallery.
    – Alex
    Commented Dec 5, 2013 at 15:49
  • 2
    I know, but I can't figure out how. The documentation isn't that good and not well indexed. Commented Dec 5, 2013 at 19:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .