# String Theory: Why should(n't) the string snap?

My question is related to a comment in the margin of the popularising site The Official String Theory Web Site.

There are two basic types of string theories: those with closed string loops that can break into open strings, shown above, and those with closed string loops that can't break into open strings, shown below

Is this a choice in a specific theory or is a consequence of an underlying principle inside the theory? How can a closed string break inside one theory and not in another, what prevents it or induces the breaking?

• This just depends on weather or not your theory includes D-branes. Since an open string has to end o such a brane, there are only open strings in a theory with branes. Jan 8, 2015 at 11:32

Towards the end of the 1980s people realised that open strings could also end on dynamical $D$-branes, thus preserving Lorentz invariance. This meant that you could have open strings in theories where previously they didn't crop up. In particular Type II string theories contain open strings which end on types of $D$-brane. For technical reasons to do with the field content of the theory, only certain dimensions of $D$-brane are allowed!
But even with $D$-branes there are still some string theories in which you can't have open strings at all! These are the heterotic theories. In such string theories, the left and right moving oscillations along the string have different amounts of supersymmetry. It turns out this is fine if the string is closed, but not allowed for open strings! So there are no open string states in a heterotic theory.