> In that case, isn't it peculiar that Russian, Czech and Old
> Polish converged, completely independently, on a gerund in
> -a?
Not really. The search went out to collect those that had -a, so the
resulting collection all have -a. A random search would have contained
many other forms.