Perisphaerinae Brunner von Wattenwyl and Hyposphaeria Lucas are valid names concealed by the unavailable names Perisphaeriinae and Perisphaeria Burmeister (Blattodea, Blaberidae)

Abstract The incorrect generic name Perisphaeria Burmeister, 1838 results from a mistake by Princis (1947). Perisphaeria is an incorrect spelling by Burmeister referring to Perisphaerus Serville, 1831. Princis (1947) incorrectly regarded Perisphaeria as proposed by Burmeister and as the senior synonym of Hyposphaeria Lucas, 1863, which is valid. The subfamilial name Perisphaeriinae based on Perisphaeria should be corrected to Perisphaerinae Brunner von Wattenwyl, 1865 based on Perisphaerus; but note Princis’ Perisphaeriinae is based on his Perisphaeria Burmeister. The taxa involved are listed.


Introduction
Perisphaeriinae and its type genus Perisphaeria are unavailable. Perisphaeria was never proposed as a new genus, but is an incorrect spelling of Perisphaerus Serville, 1831 by Burmeister (1838: 483). It is thus unavailable (ICZN 1999: Article 33.3), although it may be a Latinized spelling of Périsphère Serville, 1831, i.e. an unjustified emendation of Perisphaerus, by Burmeister (ICZN 1999: Article 33.2 and 33.5); the family name is therefore unavailable and should be corrected (ICZN 1999: Article 35.4.1). The genus Perisphaeria revised by Princis (1947: 51) refers to Hyposphaeria Lucas, 1863, which is valid but Princis regarded it as a junior synonym of Perisphaeria Burmeister of his own concept.
Perisphaeria was attributed to Burmeister by Princis (1947: 51). Here arises the question: did Burmeister establish such a genus or is it simply an unjustified emendation? Burmeister (1838: 483) added an asterisk to Perisphaeria with which he usually indicated a new taxon. However, it was normal practice for Burmeister to then provide an etymology for the proposed new name, but this was not the case for Perisphaeria, suggesting that this genus was not described as new. Secondly, Burmeister (1838: 484) included four subgenera in Perisphaeria, of which the first one (nominotypical subgenus) is Perisphaeria Serv., and doubted their ability to roll into a ball as described by Serville, whilst Serville (1831: 44) used this as a diagnostic character and as the etymology for his Perisphaerus. Thus Burmeister must have his Perisphaeria attributed to Serville was in fact Perisphaerus. Thirdly, Perisphaeria is not an emendation of Perisphaerus according to the rules of the Code (ICZN 1999: Article 33.2.1), i.e., there is no explicit statement of intention, Perisphaerus was not cited and replaced, and no similar case of treatment is found in the same publication. Therefore, Perisphaeria is deemed to be an incorrect subsequent spelling and thus unavailable (ICZN 1999: Article 33.3). It can be speculated that the asterisk annotation to Perisphaeria (Burmeister 1838: 483) was not only used by Burmeister to denote a new name, but also to denote names with new scope or definition, i.e. extending the scope of Perisphaerus to include species from both Asia and Africa. The species Burmeister (1838) described were from Africa and not Asia, from which Serville (1831: 44) described Perisphaerus armadillo; that is why he did not accept Serville's words. It is not impossible, however, that this asterisk is simply a mistake, especially when such a mark is absent next to the Perisphaeria in the key (Burmeister 1838: 481).
We speculate that Perisphaeria may be, in a practical but not a nomenclatural sense, a Latinized version of Périsphère Serville, 1831. In other words, an emendation of Perisphaerus rather than an incorrect spelling of Perisphaerus, because French names seemed to be more formal at that time. However, Serville (1831: 44) himself gave a Latinized (i.e. valid) version which was exactly Perisphaerus, just following PÉRISPHÈRE and followed by the Greek etymology; but such a change in spelling by Burmeister is not demonstrably intentional, thus not an emendation in nomenclatural sense (ICZN 1999: Article 33.2.1). Similarly, Perisphaera is also a Latinized version of Périsphère, being used by Serville (1839: 132) himself for the first time. The word ball is a feminine noun in both French (sphère) and Latin (sphaera); therefore Perisphaerus is a result of incorrect Latinization by Serville (1831) when he Latinized Périsphère the moment he proposed it. Subsequently he realized that and re-Latinized the French name in a later paper (Serville 1839). However, Perisphaerus should remain in use rather than Perisphaera since incorrect Latinization is not to be corrected (ICZN 1999: Article 32.5.1).
No matter which name, or to be more accurate, which spelling-Perisphaerus, Perisphaera or Perisphaeria-the family name is based on, the name established by Brunner von Wattenwyl (1865: 303) should be corrected to Perisphaeridae (ICZN 1999: Article 32.5.3.3). Note also that Brunner v. W. (1865) proposed it at the rank of tribe, which should be corrected to Perisphaerini under current nomenclature. The situation is further complicated by Princis (1947: 51) who attributed Perisphaeria to Burmeister and regarded it as restricted to African species, whilst the Asian species were re-combined with Perisphaerus Serville. Princis (1960: 438) subsequently included Perisphaeria Burmeister in Perisphaeriidae and established a separate family, Derocalymmidae, to include Perisphaerus. As a result, the true type genus of Perisphaeriidae was transferred to another family, but one of its incorrect spellings and wrong authorship remained; Perisphaeriidae is thus no longer correct. Perisphaeriidae and Derocalymmidae are now considered synonyms , whilst Princis' erroneous use of the genus-group names has become widespread since his catalogue (Princis 1964). It is noteworthy that Princis' Perisphaeriidae has the same spelling as Perisphaeriidae Brunner v. W. that was based on the incorrect spelling Perisphaeria Serville. These two names, identical in spelling and similar in scope, are inherently different.
Before Princis (1947), most authors were well aware of the concept of these generic names (spellings), and cited correct authorship (i.e. Serville) of Perisphaerus when referring to it under its incorrect spelling (Perisphaeria), though an exception is Fischer (1853: 94) who attributed Perisphaeria to Burmeister. However, he gave Perisphaera Serv. as a synonym, which suggests that Fischer understood well the concept of them but cited an incorrect spelling with its author instead of the original spelling with its author. Lucas (1863: 408) recognized Burmeister's Perisphaeria as the same name as Perisphaera Serville, but suggested that they cannot refer to each other, i.e., he disputed the extension of its scope. Lucas stated that Perisphaera should be restricted to the Asian species which can roll into a ball. Lucas proposed a new genus Hyposphaeria Lucas, 1863 to include African species that do not roll into a ball and included species described by Burmeister. Brunner v. W. (1865), Walker (1868), Saussure and Zehntner (1895) and Kirby (1904) all treated Perisphaeria Serville as valid and provided synonyms including Perisphaera and/ or Perisphaerus; it is inferred from this that they accepted Burmeister's spelling to be the valid name of the genus proposed by Serville, and that Perisphaeria and Perisphaeriidae were once in prevailing usage. Brunner v. W. (1865) and Walker (1868) did not mention Hyposphaeria, while Saussure and Zehntner (1895) agreed with Lucas and included only Asian 'rollable' species in Perisphaeria Serville. However, they did not adopt Hyposphaeria since they believed species in this genus should be placed in various genera. Kirby (1904) in his catalogue treated Hyposphaeria as a valid name and included four species in it, with H. stylifera (Burmeister, 1838) designated as type species.
In summary, Perisphaerus Serville, 1831 is a valid name and includes the 'rollable' species, mostly from the Oriental Region, with Perisphaera and Perisphaeria as unavailable spellings, although the latter spelling was once enlarged in its scope and widely used. Hyposphaeria Lucas, 1863 is a valid name and includes the African 'unrollable' species, of which the early ones were described under Perisphaeria by Burmeister (1838). This name was once incorrectly synonymized with the unavailable name Perisphaeria Burmeister by Princis (1947). Perisphaeridae Brunner v. W., 1865 is a valid family name based on Perisphaerus Serville. Perisphaeriidae Brunner v. W. is unavailable since it was based on Perisphaeria Serville. The Perisphaeriidae used by Princis has the same spelling but is not identical since this name was based on Perisphaeria Burmeister which is a consequence of his misunderstanding.
Princis' erroneous revision of the names mentioned above should be no longer adopted.
Serville signed his name as J. G. Audinet-Serville in his published works and some authors attributed the scientific names established by him to Audinet-Serville; however, cockroach taxonomists (e.g. those mentioned above) prefer Serville in citation and authorship of names, and as such this convention is followed to avoid ambiguity. The abbreviation Brunner v. W. for Brunner von Wattenwyl is used for brevity.