We discuss self-similar solutions to $O(4)$ textures in Minkowski space and in flat Friedmann-Robertson-Walker backgrounds. We show that in the Minkowski case there exist no solutions with winding number greater than unity. However, we find besides the known solution with unit winding number also previously unknown solutions corresponding to winding number less than one. The validity of the non-linear sigma model approximation is discussed. We point out that no spherically symmetric exactly self-similar solutions exist for radiation or matter dominated FRW cosmologies, but we find a way to relax the assumptions of self-similarity that give us approximative solutions valid on intermediate scales.

Global textures are interesting because they are promising candidates for seeds in the structure formation of the universe. The important configurations are those who will be able to collapse. The type of configurations that I will consider are half knots and true knots. I define a half knot as one that covers more than a half-$S^3$, and a true knot as a configuration that covers the whole of $S^3$. Configurations that are half knots will be able to collapse, a new argument for this collapse criterion is given in this paper. I also give some critical remarks on the method of using the total winding number as a criterion for identifying a knot. I propose another more direct method and make use of it when I estimate the probability for formation of both true and half knots by numerical simulations.Taking into account the chance of annihilation the probability for collapse is found to be in the interval $p_{coll} \in [0.01,0.04]$. In the skyrmion picture of hadrons the investigations of formation of true knots are important, and some implications of my result on the predictions of a "topological" theory of baryon production in jets are discussed.