It has been conjectured that for knots $K$ and $K'$ in $S^3$, $w(K#K')=
w(K)+w(K')-2$. Scharlemann and Thompson have proposed potential counterexamples
to this conjecture. For every $n$, they proposed a family of knots ${K^n_i}$
for which they conjectured that $w(B^n#K^n_i)=w(K^n_i)$ where $B^n$ is a bridge
number $n$ knot. We show that for $n>2$ none of the knots in ${K^n_i}$ produces
such counterexamples.