Mandibulate convergence in an armoured Cambrian stem chelicerate

Background Chelicerata represents a vast clade of mostly predatory arthropods united by a distinctive body plan throughout the Phanerozoic. Their origins, however, with respect to both their ancestral morphological features and their related ecologies, are still poorly understood. In particular, it remains unclear whether their major diagnostic characters were acquired early on, and their anatomical organization rapidly constrained, or if they emerged from a stem lineage encompassing an array of structural variations, based on a more labile “panchelicerate” body plan. Results In this study, we reinvestigated the problematic middle Cambrian arthropod Habelia optata Walcott from the Burgess Shale, and found that it was a close relative of Sanctacaris uncata Briggs and Collins (in Habeliida, ord. nov.), both retrieved in our Bayesian phylogeny as stem chelicerates. Habelia possesses an exoskeleton covered in numerous spines and a bipartite telson as long as the rest of the body. Segments are arranged into three tagmata. The prosoma includes a reduced appendage possibly precursor to the chelicera, raptorial endopods connected to five pairs of outstandingly large and overlapping gnathobasic basipods, antennule-like exopods seemingly dissociated from the main limb axis, and, posteriorly, a pair of appendages morphologically similar to thoracic ones. While the head configuration of habeliidans anchors a seven-segmented prosoma as the chelicerate ground pattern, the peculiar size and arrangement of gnathobases and the presence of sensory/tactile appendages also point to an early convergence with the masticatory head of mandibulates. Conclusions Although habeliidans illustrate the early appearance of some diagnostic chelicerate features in the evolution of euarthropods, the unique convergence of their cephalons with mandibulate anatomies suggests that these traits retained an unusual variability in these taxa. The common involvement of strong gnathal appendages across non-megacheirans Cambrian taxa also illustrates that the specialization of the head as the dedicated food-processing tagma was critical to the emergence of both lineages of extant euarthropods—Chelicerata and Mandibulata—and implies that this diversification was facilitated by the expansion of durophagous niches. Electronic supplementary material The online version of this article (10.1186/s12862-017-1088-7) contains supplementary material, which is available to authorized users.


Modifications of the morphological matrix
We made some small modifications to the data matrix compared to its initial use (1). We improved the chelicerate section of the dataset by adding one character emphasized on this study and two characters put forward as critical in the early evolution of the clade (see a review in ref. (2) Character numbers discussed in Table 1 are shifted accordingly with respect to the character list in ref. (1).
In addition, we modified three characters following new observations described in the present study (see Discussion). Character 32 (number of somites defining the head tagma) was set to 3 (=8 somites, or 7 pairs of appendages) for Sanctacaris, Habelia and Offacolus (the original state 3 tentatively coding 9 somites for the larva of Pycnogonum was removed). In character 38, a cephalothorax had been coded as present in those fossil and extant chelicerates integrating an additional pair of appendages into their prosoma, along with mandibulates also having trunk appendages integrated to their head shields. A "cephalothorax" with such definition was arguably plesiomorphic to both clades, allowing for the formation of their more inclusive cephalic tagmata compared to the likely initial euarthropod condition of five somites (1,3). To be more informative, coding a cephalothorax in their respective taxa should therefore be relative to the plesiomorphic condition of head anatomy in these clades. With our new coding of the number of somites in the head tagma in basal "panchelicerates" (char. 32), we restricted the cephalothoracic condition to mandibulates and therefore chelicerates were coded 0 for character 38. We also tentatively coded a cephalothorax as present for Euthycarcinoidea sensu lato (1) given the broad posterior extension of the head shield, encompassing up to five additional segments behind the "specialized postantennal appendages." Character 113 was modified to exclusively code for the reduction of exopods on the fourth pair of appendages across all taxa (as the coding was partly redundant with char. 84 already dealing with the reduction of endopods).
Interestingly, we also had to remove the (Scorpiones, Araneae) clade from the set of backbone constraints we used, based on ref. (4). This enforced grouping was causing a broad inversion of polarity for the chelicerate clade, with habeliids being derived within a monophyletic "Xiphosura" and Chasmataspis and Megalograptus retrieved as derived arachnids, sister taxa to scorpions. Although we rejected this topological arrangement as a result of a clear polarization bias, it is worth noting that the phylogenetic position of Scorpiones has a strong impact of the overall topology of Chelicerata, and in particular that morphology seems to support a sister-group relationship of scorpions to other arachnids (as in ref. (5); but see ref. (6)).

Sanctacaris uncata Briggs and Collins, 1988
Sanctacaris was previously regarded as at least a cheliceromorph (7,8), but such statements were based on different interpretations of the head tagmatization than those proposed in the present paper. An interpretation of differentiated limbs in Sanctacaris as the exopods of the cephalic appendages was also used by Legg (8) as an argument for a chelicerate relationship, based on the segmented condition of similar exopodial branches in the Silurian Offacolus (9) and Dibasterium (10). We follow such view, especially given the "detached" condition of cephalic exopods in these taxa (see main text), but we disagree with other details of the interpretation. The spinose cephalic endopods 2 to 5, and likely endopod 1 as well, are composed of seven visible podomeres, including the distal claw complex, and the bases are not visible (we construe that the gnathobases equivalent to those of Habelia are confounded within the thick layers of fossilized cuticle at the front; Additional file 8). More importantly, appendages of three different types were labeled as the same cephalic exopods on the holotype: a short, broad, and likely lobate appendage bearing lamellate setae; another short, but elongate appendage with a setal brush at its tip (considered to be the sixth appendage by Briggs and Collins (7)); and the long, segmented "antenna-like rami" overlapped by the endopods (Additional file 8). The short appendage bearing a setal brush is the only one that resembles the stenopodous exopods of Offacolus and Dibasterium, but its identity in Sanctacaris is not obvious. The lobate appendage, located at the back of the anterior "bundle," best corresponds to the similar rounded exopod associated with the seventh cephalic pair in Habelia (Additional file 8), and which would be large enough to jut out slightly anteriorly if compressed ventrally against the head shield. Accordingly, the elongate branch could be the endopod of the seventh appendage, but a setose termination would be odd in the context of arachnomorphs. This appendage could also be a differentiated fifth cephalic exopod, but its very posterior position makes this interpretation questionable. The identity of this appendage can therefore not be disambiguated at present.
Nonetheless the position of what we interpret as the cephalic exopods 4 and 5, well-behind their corresponding endopods, corroborates the interpretation of a morphological dissociation between the inner and outer rami of the biramous appendages. The holotype of Sanctacaris also displays a frontal protrusion identical to the labrum identified in Habelia (Additional file 8).
Anteriorly, the protrusion is split into left and right elements, each bearing a small central reflective spot. This morphology extends the comparison with protocaridids, other hymenocarines and possibly some artiopods discussed in the main text, whose frontalmost labral organs are bilobed and/or display paired or more reflective remains (11). There seems therefore to be cumulative evidence for a common ancestry of that structure (independently from the accompanying sclerites) at the point of diversification between artiopods, mandibulates and chelicerates.