There are lots of well known wormhole solutions, which are constructed to be stable. But can a "normal" system evolve to form such wormholes? In thin shell formalism it is possible to glue Schwarzschild spacetimes together. If we glue the an inner Schwarzschild to an outer one, we get a standard shell moving in standard Schwarzschild vacuum. Above the horizon the gravitational mass has to be positive, I mean M_out > M_in, where, M is the Schwarzschild mass parameter. However it is also possible to glue outer Schwarzschild to another outer Schwarzschild, in this case we get a wormhole solution, and gravitational mass can be negative (the rest mass is always positive). But this is a prepared situation, I mean we get a wormhole because we set the initial parameters such a way. However it is known that during shell crossing, mass inflation occurs near the center singularity. If we consider for example two shells, and they collide transparently under the horizon, the gravitational mass can be negative (M1<M2<M3 is not valid), this means that between the two shells we get a region, which is part of a wormhole solution (not the whole spacetime is a wormhole because we have the inner Schwarzschild below the two shells) This is strange if I am right. What about the radius in this region? Between the two shells the radius is not increasing outwards? Let us suppese that after the collision the inner shell collapse to the centrel singularity without shell crossing, in this case we get a global wormhole which has a central singularity?