Thorne's (Proc. R. Soc. Lond. B 241, 37 (1990)) studies of the laboratory behaviour of extant woodroaches (Cryptocercus) and termites (Zootermopsis) suggest that transfaunation of hindgut protozoans potentially could occur by aggression and consumption in the field. However, existing literature suggests no overlap in protozoan species composition for these two taxa. Furthermore, it is doubtful that transfaunation would occur in the solitary ancestral `termitoid' and `roachoid' lines proposed by Thorne: not only is it unlikely that such insects would encounter each other, but it is doubtful that they would show the degree of aggression exhibited by the termite soldiers in her study. Inheritance from an ancestor common to Cryptocercus and the lower termites remains the most logical explanation for the presence of the unique cellulolytic oxymonad, trichomonad and hypermastigote flagellates in these two groups.