321 



that the absence of the division described in character 1 and the absence of 110 are ple- 

 siomorphic for Blattaria and Mantodea as a whole but apomorphic within subgroup 2.2.3.2. 

 (discussion in 7.5. (Q), (R)). The character states achieved by these reversals can then be 

 properly defined as highly apomorphic states (Table lb). 



As a consequence of the combined presence of the first and the second problem, in the 

 character state matrix in Table la the items relating to the L2-divisions would suggest that 

 these L2-divisions have originated independently in Nahublattella and in subgroup 

 2.2.3.2.2. and are non-homologous. In a computer-based cladistic analysis this would cause 

 a misleading trend away from a holophyly of Nahublattella + subgroup 2.2.3.2.2. In the 

 matrix in Table lb this misleading impression is eUminated. 



As a result, the assessment of homology relations, the definition of character states, the 

 assignment of morphological condifions to certain character states, and the polarity 

 assumptions, and hence also the respective entries of items into the matrix, can in some 

 cases only be proper in dependence on a previous hierarchical analysis with reciprocal 

 illumination and on the resulting preUminary assumptions in terms of phylogeny. It is, at 

 least in the frame of the analysis presented here, not suitable to give a character list and 

 a character state matrix with the characters considered independently of each other and of 

 preliminary assumptions on phylogeny. 



8. HOMOLOGY RELATIONS ACCORDING TO MIZUKUBO & HIRASHIMA (1987) 

 AND GENERAL REMARKS ON THE ANALYSIS OF HOMOLOGY RELATIONS 



The assumptions and procedures of Mizukubo & Hirashima 



Mizukubo & Hirashima (1987) investigate the sclerites and the muscles of the phallomeres 

 of Periplaneta fuliginosa (Blattidae / Blattinae), of 3 species of Blattella (Blattellidae / 

 Blattelhnae), and of Opisthoplatia orientalis (Blaberidae). Additionally, they use data from 

 other writers concerning various species of Blattinae. The phallomeres of Blattinae are 

 regarded as the most primitive. The results of the authors comprise: (1) Homologies of 

 the phallomere elements of the different species. (2) Side-homologies of the elements of 

 the left and the right halves of the phallomere complex. (3) A ground-plan for the 

 sclerotisations of the phallomere complex of Blattaria, which is mainly based on the 

 morphology of Blattinae. 



As regards (1), the supposed homology relations are fundamentally different from those 

 I assume for the respective close relatives Eurycotis (Blattidae), Parcoblatta (Blattellinae), 

 and Blaberus (Blaberidae). For example, Mizukubo & Hirashima suppose that the hooks 

 designated here as hla have developed from completely different elements in the three 

 groups. (In my view these hla are strictly homologous.) Their opinions concerning the 

 ground-plan of the Blattarian phallomeres are also completely contradictory of my results. 

 The paper of Mizukubo & Hirashima must therefore be discussed in detail. 

 Mizukubo & Hirashima procede as follows: 



- They divide both the left and the right side of the phallomere complex into 11 

 "subregions". The definition of "subregion" is: "The smallest and indivisible unit which 



