When we say "string theory" these days, we mean the magnificent theory that can be connected by physical transitions to the well-known supersymmetric vacua in 10 or 11 spacetime dimensions: "string theory" really means "superstring theory".
But the whole first volume of Joe Polchinski's textbook, among others, is about bosonic string theory in 26 dimensions. Is it just an awkward and mostly inconsistent toy model to learn some techniques that reappear in superstring theory or is it a part of "the" theory that we study?
In 1993, Berkovits and Vafa argued that the N=0 bosonic string can be viewed as a special case of the N=1 superstring, much like the N=1 string is a special case of the N=2 string, by reinterpreting some matter degrees of freedom as the new required (super)conformal ghosts etc.
Tonight, Hellerman and Swanson claim that they can construct a cosmological solution that interpolates between the bosonic string in the past and the superstring (although it's just type 0 string) in the far future. Light-like tachyon condensates and the identification of the RG flow with a time evolution play a crucial role here.