frcw does not follow the published ReCom algorithm (nor the GerryChain implementation when it a spanning tree cannot be split within the population bounds. The ReCom algorithm simply generates new spanning trees until it succeeds. frcw, however, repeats the previous plan (a la RevReCom) when it cannot split a tree.