In the introduction to the previous part published in this series (Holloway, 2005), an account was provided of developments in the classification of the Noctuoidea, particularly the status of the Noctuidae, given that several molecular studies had indicated that the traditional concept of the family was not monophyletic. If monophyly was the most important criterion, then the family should be restricted to the ‘trifine’ groups covered by Holloway (1988), with the inclusion also of the Plusiinae, Pantheinae and most of the traditional Acontiinae / Eustrotiinae complex. However, this left out virtually all the quadrifine subfamilies of the traditional concept of the Noctuidae. Molecular results showed these intermingled with the Lymantriidae and Arctiidae, with some surprising associations, such as a sister-relationship between the Aganainae and the ‘noctuid’ subfamily Herminiinae (discussed further on p. 22).
Therefore, for the Catocalinae / Ophiderinae complex, I decided to follow the consensus of a workshop on the Noctuoidea at the Societas Europaea Lepidopterologica (SEL) Congress in 2002 (Holloway, 2005: 4). This recommended a focus on definition of monophyletic groups throughout the quadrifine sequences of traditional noctuids, leaving nomenclatural problems to one side until a clearer picture emerged of the relationships of these groups to one another. Once a clear classificatory structure was available, it would be possible to develop a satisfactory, stable nomenclature.
However, two other publications appeared in 2005 (the second initially on-line only), one a review of the higher classification of the Noctuoidea (Fibiger & Lafontaine, 2005), the other an expansion of previous molecular analyses of the superfamily (Mitchell et al., 2006). The former recognised ten families: the traditional ones, but with the quadrifines divided off as the Erebidae, together with two additional small families, the Strepsimanidae and the Micronoctuidae. The latter favoured a classification that promoted several of the old quadrifine subfamilies to full family status to reflect the molecular results in a manner that enabled the retention of the Arctiidae and Lymantriidae as full families also. This led to a classification that recognised thirteen families, placing the Strepsimaninae within the Catocalidae (= Erebidae on priority), omitting the Micronoctuidae, and adding through promotion the Stictopteridae, Euteliidae, Aganaidae, Herminiidae and Hypenidae.
These publications formed a natural focus for a similar workshop at the SEL Congress held in Rome in September, 2005. Some attempt was made to reconcile the two classifications, but the problem of nomenclature again arose, particularly whether to accept a significant increase in the number of noctuoid families recognised in order to reflect the structure emerging from the molecular results, or whether to go in the opposite direction and roll the whole problem up within a very much larger concept of the Noctuidae. The latter course was that advocated subsequently by Lafontaine & Fibiger (2006). But, as for the 2002 discussion, we still do not have an extensively tested structural hypothesis for the classification, and labelling in a categorical manner the structural information currently available is still as dangerous.
One danger, given that Fibiger & Lafontaine (2005) included ‘with special reference to the Holarctic fauna’ in their title, is that, for historical reasons, the temperate tail is still wagging the much more diverse tropical dog. The greatest diversity at all taxonomic levels in the Lepidoptera is in the tropics. For example, though the type species of Micronoctua Fibiger was first recorded in Europe (Fibiger, 1997), its author (Fibiger & Lafontaine, 2005: 33; Fibiger, 2007, pers. comm.) has subsequently discovered more than 300 species, virtually all in the Old World tropics and subtropics, that relate to the family Micronoctuidae, several occurring in Borneo (Holloway, 2005: 30), a few already described. Fibiger & Lafontaine (2005) indicated that the family might be arranged in tens of genera and several subfamilies.
How many more such are there out there? It is likely that several may lurk within the sixth miscellaneous sequence of genera treated in Holloway (2005). Several genera traditionally assigned to the Hypeninae or Herminiinae appear obviously misplaced, once a clearer perspective is gained of the defining features of the core group. Before too many families are proposed for the Noctuoidea, it must be recognised that the major part of the subfamily, sister-group to the Notodontidae, is likely to have an extensive paraphyletic tail, probably mostly of tropical taxa. With such a tail, the tropical dog has yet to have its day.
The proposal for an all-embracing concept of the Noctuidae has its attractions in this situation, but also has one major disadvantage: that of nomenclatural squeeze at lower taxonomic rank. This is practical rather than scientific: for example, the subtribes recognised by Fibiger & Hacker (2005) for European Noctuidae would automatically be demoted out of existence. For this reason, many taxonomists, being on the whole rather conservative, will continue to use the traditional families, perhaps with some concession to well-supported and major structural features arising from further morphological and molecular studies. There is no inherent prejudice against large numbers of higher taxonomic categories. There are about 221 species of wild land mammals in Borneo (Payne et al., 1985). These are divided amongst 11 orders and 32 families.
This volume covers five groups from the traditional Noctuidae. They all fall within the Erebidae of Fibiger & Lafontaine (2005), and are scattered amongst the basal branches of their cladogram for the Erebidae together with the Boletobiinae (Holloway, 2006: 373 et. seq.; as Boletobiini), Araeopteroninae and Eublemminae (traditional Acontiinae, to be treated in Part 13). They placed the Hypeninae and Phytometrinae in a sister-relationship. The analysis of Mitchell et al. (2006) placed them with most of the Erebidae in an LAQ (Lymantriidae, Arctiidae + quadrifine Noctuidae) lineage, but with even greater dispersion: the Hypeninae are the basal branch; the Herminiinae are in a sister-relationship to the Aganainae; Rivula Guenée is placed as sister-taxon to an Arctiidae / Lymantriidae pairing; Phytometrinae and Hypenodinae were not sampled.
The groups in this treatment will therefore be given subfamily status, leaving open the question of family placement; obviously they could all be subfamilies of the very large concept of the Noctuidae of Lafontaine & Fibiger (2006), but the Herminiinae at least have had full family status in the past (e.g. Common, 1990; Nielsen et al., 1996) as discussed on p. 17. For each, some attempt will be made to assess their validity as monophyletic entities, but, for the Hypeninae at least, several of the genera do not show particular affinity with the core grouping, and several are held over for consideration with the traditional Acontiinae in case they prove to have affinities with taxa in that similarly heterogeneous assemblage.