P. Grandcolas, THE PHYLOGENY OF COCKROACH FAMILIES - A CLADISTIC APPRAISAL OF MORPHOANATOMICAL DATA, Canadian journal of zoology, 74(3), 1996, pp. 508-527
Seventy-two morpho-anatomical characters were examined in 221 genera b
elonging to the families Blattidae, Polyphagidae, Blattellidae, and Bl
aberidae. They were cladistically analyzed and polarized using two man
tids and two termites. As no autapomorphies of the family Blattellidae
were found, the constituent subfamilies were used as terminal taxa to
gether with other families. Three trees were found (CI = 0.81 and RI =
0.88, without autapomorphies) that differed only by the position of N
yctiborinae relative to Blattellinae and Ectobiinae. The strict consen
sus tree was [Blattidae [Polyphagidae [Anaplectinae [[Pseudophyllodrom
iidae, Blaberidae] [Nyctiborinae, Blattellinae, Ectobiinae]]]]]. The m
ain discrepancies with McKittrick's tree were the monophyly of Polypha
gidae (instead of paraphyly) and that the Blaberidae is the sister-gro
up of Pseudophyllodromiinae (instead of the sister-group of Blattellin
ae, Ectobiinae, and Nyctiborinae). These results made it necessary to
elevate the Anaplectinae and Pseudophyllodromiinae to familial status,
and to give a new sense to the family Blattellidae, which includes on
ly the subfamilies Blattellinae, Ectobiinae, and Nyctiborinae. This ph
ylogeny was used to test current evolutionary hypotheses concerning so
ciality and reproductive behaviour; many traits were assumed to be anc
estral to all cockroaches (protozoan symbionts and familial life habit
s) or preadaptive (progressing from advanced oviparity in Blattellidae
to ovoviviparity in Blaberidae) that must actually be considered homo
plasic.