Recently, we described the first known larvae of the enigmatic aquatic beetle family Meruidae (Alarie et al., 2011) and coded 28 larval features into a limited character matrix of 12 taxa to provide a phylogenetic context to the morphology that we described and discussed. In our analysis, Meru was placed as sister to the remaining Dytiscoidea – an identical placement to a preceding morphological analysis by Dressler & Beutel (2010). Subsequently, incorporating the newly illuminated Meru larval characters, Dressler et al. (2011) provided a re-analysis in the context of an expanded dataset. Based on this new analysis, Dressler et al. (2011), refuted the prior topology and made surprising claims about the morphology and ecology of Meru and Notomicrus. Further, despite basing their analysis and findings on the assumption that Meru adults and larvae were associated properly, they concluded that this association should be considered unconfirmed. Here, we defend the justification for our larval association, present molecular data further confirming its identity, and challenge their conclusions about the ecology of Meruidae and Notomicrus.
Standards of larval association
Our association of the larvae of Meru was based on several independent factors, including their abundant co-occurrence with adults in a very unusual habitat and locality (i.e. algal seepages on granite outcrops in the northwestern Guiana Shield), their tiny size and their unusual (and apparently phylogenetically distinct) morphology – including the fact that the larvae share with adult Meru a pectinate or ‘combed’ claw, unique among all known aquatic adephagan adults or larvae. Although thousands of adult and larval beetle specimens were found on the seepages that contained Meru, the only other adephegans (adult or larvae) found on the seep could be assigned unambiguously to the family Dytiscidae. Nevertheless, Dressler et al. (2011, 705) question the veracity of our association, suggesting that ‘it is questionable whether this evidence alone is sufficient for a holometabolan species' and that ‘the association… of this taxon cannot be considered as completely confirmed as long as the identicalness on the molecular level is not published’. Just why would the authors express doubts about our larval association while simultaneously publishing a comprehensive re-analysis and discussion of adephagan phylogeny contingent upon it being correct? Further, we had communicated to the corresponding author that a positive molecular association had been confirmed well before our manuscript was published.
While we reject the implication that our enumerated observations – in this case ecological and temporal association with the adults, morphological autapomorphies and lack of evidence for any alternative candidates – is ‘questionable’, we are happy to publish a partial sequence of the mitochondrial gene COI for a larval Meru specimen (GenBank accession JQ684026; methods for DNA extraction, amplification and sequencing followed as in Short & Caterino, 2009; voucher in the collection of the Universidad Central de Venezuela, Maracay, Venezuela). The ‘identicalness’ of the larval sequence to the previously published adult DNA (GenBank accession FM163588; Balke et al., 2008) is more than 99%, with only 6 bp differences in more than 758 bp of overlap between the fragments. This is well within the intraspecific threshold for ‘DNA barcoding’, to say nothing of differences that would be expected between valid genera or families. While we agree that there are instances (perhaps even in the majority of cases) where confirmation using molecular methods is essential for associating adult and unreared subimaginal stages of insects, we cannot agree with the assertion that this has singular standing over all other morphological, ecological and phylogenetic inferences for all ‘holometabolan species'.
Phylogenetic placement and classification of Meruidae
Dressler et al. (2011, 709) suggest that, ‘In summary, it can be said that the dataset and applied analytical methods in Alarie et al. (2011) are not sufficient for obtaining robust phylogenetic results'. To be sure, the dataset of Dressler et al. (2011) is larger and more comprehensive than the one we published, and includes adult as well as larval characters. However, the stated purpose of our dataset was not to present a total-morphology phylogeny of Adephaga, but rather to focus on the larval characters of Meru in the context of a larval-based phylogeny. We reanalysed the dataset of Dressler et al. (2011) using only their larval data (characters 88–143) in TNT (Goloboff et al., 2008) using a traditional TBR search of 5000 replicates, keeping 20 trees per replicate. The strict consensus of the resulting 80 most parsimonious trees of 125 steps is a completely unresolved polytomy. We agree that we could have applied more rigorous search methods in our prior parsimony analysis – although we disagree with their implication that it would have yielded a more ‘robust’ result. Using TNT, we reanalysed our previous matrix using an exhaustive search resultng in a single most parsimonious tree with identical placement of Meru as found previously, as well as an otherwise identical branching topology for all terminals except for (Aspidytes+Amphizoidae), which is collapsed due to a zero length branch.
Dressler et al. (2011) pose the titular question, ‘Is Meru a specialized noterid?’ and conclude that ‘it appears disputable to maintain family rank for Meruidae’ and that ‘it would be justified to treat Meru as a noterid genus' (p. 709). These conclusions seem to suggest that despite the results of their current analysis as well as four previous (though not all completely independent) analyses that provide support for the reciprocal monophyly of Merudiae and Noteridae (Dressler & Beutel 2010; Dressler et al. 2011, Beutel et al. 2006, Balke et al. 2008, Alarie et al. 2011) suffer from various insufficiencies, or Meruidae is unworthy of family status despite their peculiar morphology, ecology, and distribution. They admit this is an ‘academic argument' and we agree: if Meru can be argued to be a ‘specialized noterid’, it could also be argued that noterids are merely ‘derived’ meruids. Of course, neither hypothesis is well supported and Dressler et al. (2011, 711) appear to hedge their bets by surmising their results might ‘suggest that life in more or less typical hygropetric habitats may be ancestral for the meruid-noterid clade, or possibly even the entire Dytiscoidea, as Aspidytes also lives in a similar habitat'. If this proposed ecological sequence is true, then following their logic, noterids would be considered ecologically ‘derived’, insofar as they have transitioned away from seeps and back into more traditional lotic and lentic systems. In reality, this is a futile discussion and the blanket use of the words ‘specialized’ and ‘derived’ seemingly imply that the alternate condition is not itself specialised. While it may be appropriate to refer to ‘derived’ morphologies in the context of polarised character state changes, the semantics often (and in this case) are not related to specific character states, but rather are conferred on clades or habitats in the context of their extant diversity. In the same way that ‘basal’ is often applied to the smaller of two asymmetrically diverse sister-groups, so too is ‘specialized’ used here to refer to taxa that occupy the ‘lesser’ habitat.
On the purported niche overlap between Meruidae and Noteridae
An auxiliary concern raised by Dressler et al. (2011, 711) relates to the reliability of our larval association, asserting that ‘Notomicrus [a genus of Noteridae] co-occurs with Meru on the natural water slide in Venezuela (Beutel et al., 2010)’. This statement errs in that (1) Meru has never been collected with any frequency – if ever – on the ‘water slide’, (2) Notomicrus has never been collected on any seepage habitats at the Tobogan de la Selva site, and (3) Notomicrus has never been taken together with Meru.
As emphasised previously, the aquatic habitats at Tobogan de la Selva are extremely diverse and complex. Furthermore, we refer to the fact that most (if not all) collections of Meru are not from the water slide, but rather from peripheral seeps, root mats or detritus (Alarie et al., 2011). Thus, special care must be made when making ecological statements about locally occurring species, as locality co-occurrence does not necessarily imply ecological or temporal co-occurrence. The list of taxa cited as ecologically co-occurring with Meru– including Notomicrus– in Beutel et al. 2010 (which is extracted from Spangler & Steiner, 2005) is correct; but only in the context of a diverse suite of habitats where Meru has been found, including stream margins and sifted wet litter, as well as hygropetric situations. Thus, the assertion that Notomicrus occurs on the seepages at Tobogan is unsupported by any data.
We reviewed: (i) several hundred pages of Paul Spangler's original, hand-written field notes from the Venezuela expeditions where he collected Meru; (ii) the Notomicrus specimens from his fieldwork at the Smithsonian Institution; and (iii) our own collections, notes and observations from Tobogan de la Selva. No field notes, observations or specimens – historical or recent – support the assertion that Notomicrus occur in seeps at this locality, nor have Meru and Notomicrus ever been collected together. Thus, Dressler et al.'s (2011, 711) assertion that ‘The larvae of Notomicrus, which are still unknown, must occur in the same habitat as the larvae described by Alarie et al. (2011)’ is speculation without factual basis, and stands opposed to the data from many thousands of specimens and a half-dozen expeditions conducted over various seasons.
Notomicrus aside, there is indeed a remarkable seep-inhabiting species of Noteridae at the same locality as Meruidae, namely Tonerus wheeleriMiller, 2009. This unusual taxon, supported as a monogeneric tribe by cladistic analysis based on adult morphology (Miller, 2009), occurs exclusively (so far as is known) in seepage habitats around Tobogan de la Selva. Surprisingly, given its ecological, biogeographic and apparent phylogenetic significance to Meruidae, Dressler et al. (2011) did not discuss Tonerus or the phylogeny of the Noteridae presented alongside its description. Nevertheless, this taxon also has never been taken together with Meruidae.
Are pectinate claws an adaptation to hygropetric habitats?
Concerning Meruidae, Dressler et al. (2011, 711) claim that ‘The specialized [pectinate] claws are an obvious adaptation to the hygropetric habitat' and thus ‘may also occur in the larva of Notomicrus’. Although we have dispensed already with the idea that the larvae are associated incorrectly, or that Notomicrus are demonstrably hygropetric at Tobogan, no further explanation or citations are provided as to why the adaptation of combed claws would be ‘obvious’ even for Meruidae, let alone other hygropetric taxa. To our knowledge, no other aquatic beetles have pectinate claws. Furthermore, there are hundreds of other species known to occupy comparable hygropetric habitats (including, as mentioned above, the hygropetric noterid Tonerus) and none have pectinate claws (e.g. Spangler, 1972; Reichardt, 1973; Miller, 2009; Short & García, 2010). Thus, while the claws of Meru may be linked directly to their occurrence in seeps, it is in no way ‘obvious’; rather, it is an ad hoc hypothesis without basis in comparative morphology or ecology.
An abundance of ecological, morphological and molecular data provides unanimous support for the positive association of adult and larval life stages of Meruidae. The specific ecology of Meruidae has become confused in the still-nascent literature on the family, but most published accounts either are incomplete or unintentionally misleading (as in the case of Dressler et al., 2011). Based on known collections, Meru phyllisae occurs primarily on seasonal seepages over granite, and although a handful of adult specimens have been collected in true ‘cascades’, root mats and sifted riparian detritus, these represent a fraction of known specimens and may represent opportunistic, seasonal or accidental records. While various phylogenies based on adult, larval and molecular characters have been published – all of which place Meruidae and Noteridae in close proximity – the precise relationship between (or within) the two families remains to be settled.