Paratorneuma, ROUDIER, 1956 RESYN.

Stüben, Peter E. & Astrin, Jonas J., 2010, Molecular phylogeny in endemic weevils: revision of the genera of Macaronesian Cryptorhynchinae (Coleoptera: Curculionidae), Zoological Journal of the Linnean Society 160 (1), pp. 40-87 : 79-83

PARATORNEUMA ROUDIER, 1956 RESYN.

( FIG. 21A–E )

(lastly: Paratyphloporus Solari, 1937 , (Stüben, 2007) – Mediterranean)

Type species of the genus Paratorneuma : Torneuma orbatum Wollaston, 1865 ( Roudier, 1956) ;

Wollaston, 1865; F. Solari, 1937; Roudier, 1956; González, 1971; Franz, 1981; Osella & Osella, 1984; Osella, 1986; Behne, 2000; Stüben, 2000g, 2007a, 2008b; Germann & Stüben, 2006.

Compilation of species of the genus Paratorneuma

Paratorneuma orbatum ( Wollaston, 1865) – La Gomera

Paratorneuma aphroditae Germann & Stüben, 2006 – La Gomera

Paratorneuma feloi (Stüben, 2007) – Tenerife

Paratorneuma franzi González, 1971 – Tenerife

Paratorneuma lindrothi Franz, 1981 – La Palma

Discussion ( Torneumatini ): The supraspecific classification of Torneumatini has always been a problem for taxonomists working exclusively morphologically ( Stüben, 2007a). The taxa of this tribe all have a very similar habitus. Furthermore, biological and ecological arguments are hard to apply because the animals live subterraneously. In the past the only decisive character was the pectoral canal, which is either present and fully developed ( Torneuma ; Fig. 20C ), only present as a shallow depression in front of the praecoxae ( Paratyphloporus ; Fig. 21C ) or totally absent ( Pseudotorneuma ). In a recent study, Torneumatini was classified using this character as well as the structure of the internal sac of the aedeagus ( Stüben, 2007a).

The simple classification suggested in this study bore in mind the practice of science and was an attempt to settle the taxonomic confusion that had resulted from almost 150 years of research on Torneumatini . It was always clear that the supposition of a continuous evolutionary transformation process (fully developed pectoral canal reaching midcoxae → pectoral canal only as shallow depression in front of the praecoxae → pectoral canal absent) would only deliver a heuristic/pragmatic, typological classification ( Stüben, 2007a: 95). We cannot even be certain about the polarity of this transformation series (although it is more likely that the fully developed pectoral canal constitutes the ancestral character state for Torneumatini ; see above).

The genus Paratorneuma Roudier (1956) (pectoral canal developed as a shallow depression in front of the praecoxae; Fig. 21C ) was at first interpreted as a younger synonym of Paratyphloporus Solari, 1937 ( Stüben, 2007a) . According to this definition, the species belonging to this genus included not only Mediterranean taxa (e.g. the type species of Paratyphloporus : Torneuma karamani Formánek, 1912 ), but also some Canarian species like Paratyphloporus franzi ( González, 1971) (originally Paratorneuma ) and the new species Paratyphloporus feloi Stüben, 2007 . However, the option was kept open: ‘If future molecular results should indicate the plausibility of parallel evolution between Canarian and South European Torneumatini , thus rendering the groups paraphyletic, a separation and thereby a re- synonymization of the genus Paratorneuma Roudier (1956) with the type species Torneuma orbatum Woll. (La Gomera) will be required’ (translated from Stüben, 2007a: 39).

According to our molecular dendrogram (from which many southern European species of Torneuma are absent), the pectoral canal of Torneumatini has been reduced twice ( Fig. 1A ): on the Canary Islands and in the Mediterranean independently. Now we have to consider four genera of Torneumatini and await further molecular results (sufficient material fit for sequencing has yet to be collected) in order to determine whether the species-rich Mediterranean group around the type species Torneuma deplanatum deplanatum (Hampe, 1864) requires one more taxon: namely Typhloporus Hampe, 1864 .

Differential diagnosis: See also the ‘Key to the genera and subgenera of Macaronesian Cryptorhynchinae’.

Distribution: Canary Islands, Madeira, Porto Santo, Mediterranean.

Figures

Figure 1. A, Bayesian 50% majority rule consensus for the two analysed mitochondrial genes. Numbers indicate nodal posterior probabilities. The scale shows the expected nucleotide substitutions per site. Abbreviations in parentheses denote subgenus (see text). Genus names in bold indicate new nomenclature. Circles mark colonization events (in black: invasive species). For many Macaronesian species, the respective host plants are depicted. The second page of the tree shows the taxa corresponding to the ‘Atlantic clade’. B, chronogram based on relaxed phylogenetic analysis of the two mitochondrial genes, using the uncorrelated lognormal model of substitution rate variation. Numbers next to the nodes indicate mean divergence times (in Myr), with 95% confidence intervals of divergence times depicted as bars at the corresponding nodes. N.B. Although the underlying topology is that delivered by MrBayes, BEAST does not allow polytomies, which it resolves arbitrarily.

Figures 2–22. Morphological, biological, and ecological comparison of the genera and subgenera of Macaronesian Cryptorhynchinae along with the respective sections of the tree; (TS) = type species. Figures 4E and 5E show a section of the tree that results when 16S sequences for Aeoniacalles aeonisimilis* and Dendroacalles euphorbiacus* are excluded.