From the wikipedia article on rotation matrices I gathered that rotations about the z-axis have the form:

```
cos(theta) -sin(theta) 0
sin(theta) cos(theta) 0
0 0 1
```

Is taking a matrix for a rotation in all three dimensions and setting the 0s and 1s an appropriate way to lock the rotation to the z-axis? It basically works, but seems inelegant.