The oldest currently known true cephalopods are those of the genus Plectronoceras from the Upper Cambrian, much younger than Nectocaris. They had millimetric shells with a siphuncle and a small number of septa. Such diaphragms exist in a number of groups (not only molluscan). What makes cephalopod a cephalopod is its siphuncle, which could have originated at the extension of the pelagic life style from larvae to adults (Dzik 1981). ‘Monoplacophorans’ such as Hypseloconus were once thought to be ancestral to Plectronoceras by many workers (Yochelson et al. 1973; Peel 1991), but these should probably be better classified as brachiopods (Dzik 2010), with real ancestors somewhere close to ‘hyoliths’ such as Turcutheca (Dzik 2010).

Being much older, the ‘Kraken’Nectocaris does not have either a phragmocone or a shell. Could it be, that the cephalopod shell is not homologous to other molluscan shells, or may be it is a case of secondary loss in Nectocaris– an early, and surprisingly complete? Highly unlikely, given the micro-structural, genetic and ontogenetic unity, as well as the known common ancestry of that element. It is well known that the internal shell in coleoid cephalopods is a derived feature and during their evolution, it became completely reduced, as in recent Octopoda. However, it must be taken into account that the shell reduction in a most advanced cephalopod group such as the octopods occurred late during its evolutionary history (e.g. Boyle & Rodhouse 2005), as easily fossilized octopod shells are known from the Cretaceous (Fuchs et al. 2009). What’s more, the radula, a synapomorphic trait of a really high value, is absent too. Although the radula is lacking in bivalves and some aplacophorans, it is present in the rest of the molluscs including the cephalopods, with only Spirula having its radula absent or vestigial. However, a vestigial or totally reduced radula is believed to result from a secondary reduction (Warnke & Keupp 2005). Therefore, it is highly unlikely that the radula would be totally reduced in ‘early cephalopods’ such as Nectocaris. There is also no reason why it should not fossilized or being unnoticed in the material comprising nearly a hundred of specimens investigated, especially while soft tissues have been preserved. Moreover, molluscan radula are known to have been preserved in older, Lower Cambrian deposits (Butterfield 2008), and the oldest of cephalopod radulas are known to be preserved in Upper Ordovician (Gabbott 1999) and Late Silurian (see Mehl 1984) orthoconic nautiloids.

It appears then that two most important molluscan features (present in all cephalopods) are absent. The same applies to jaws (beaks), widely known in both Palaeozoic ammonoids (e.g. Tanabe & Mapes 1995) and Mesozoic coleoids (e.g. Klug et al. 2009), but which are not present in the putative ‘cephalopod’Nectocaris at all. The lack of beak and the presence of only one pair of frontal tentacle-like appendages seem to be completely incompatible with respect to cephalopod feeding behaviour, which rely on capturing the prey with tentacles, delivering it to the mouth and biting it with the beak (e.g. Hanlon & Messenger 2003; Boyle & Rodhouse 2005). Thus, the prey capturing by Nectocaris using its ‘tentacles’ must have been very inefficient, if not to say impossible, and in spite of the lack of jaw apparatus, its feeding behaviour seemed to be completely different (see below).

Similarities in gill and fin location and structure could well be convergent with any group, and is not considered specific here. The funnel is common to all cephalopods, but the one in Nectocaris is truly evident only in the artistic reconstruction (see Bengtson 2010). To work as a funnel properly, it should not get wider distally, as it does in Nectocaris. Rather, the funnel should have a small cross sectional area, as in Recent cephalopods. Only then, the jet has a considerable force and propels the animal in opposite direction (e.g. Boyle & Rodhouse 2005). The distally widening ‘funnel’ of Nectocaris suggests that this structure evolved for completely different purpose than for jet-propelled movements. Because of the poor preservation of the putative mouthparts in Nectocaris (Smith & Caron 2010), the ‘funnel’ most likely represents a specialized feeding apparatus. Its morphology suggests that it served as ‘probing’ device with a true mouth located most distally, for catching food, probably in the form of small animals. Interesting is the fact that a quite similar structure – in a general outline and orientation – a proboscis feeding apparatus, occurs in pycnogonid marine arthropods, and in its advanced form is known already from the Silurian (see Siveter et al. 2004).

Nectocaris also lacks another characteristic feature of cephalopods, namely the ink sac. The ink sac is known to be preserved in fossil coleoids (e.g. Wilby et al. 2004; Klug et al. 2009) and occasionally in ammonoids (Doguzhaeva et al. 2007), and likely could have been preserved as organic film in Nectocaris as well – that is, if it represents a cephalopod. However, as the ink sac is probably an advanced feature and may not have been present in early Palaeozoic forms, it is not taken into account during the current consideration.

The eyes of Nectocaris, although not discussed extensively by Smith & Caron (2010), may be, just as authors state, truly cephalopod-like; but we have to remember about the extreme number of convergences in that organ, and its appearances multiple (dozens of) times within the Animalia.

The presence of lateral flap-like appendages in Nectocaris, two in front of the head, and its stalked eyes resemble those of the anomalocaridids, a group of pelagic predatory ecdysozoan animals that flourished in the seas from the Cambrian to at least the Devonian (Whittington & Briggs 1985; Kühl et al. 2009). The animal bears striking differences to an anomalocaridid as well, with a ‘funnel’ as a necessary derived equivalent of a feeding apparatus of other members of that group. What must be always kept in mind is that only true similarities (apomorphies), and not differences (autapomorphies and convergences) are features that can tell as something about affinities (Van Iten et al. 2006). The similarity of Nectocaris pteryx to cephalopods is only superficial, being restricted to such convergent features as a general body outline, the presence of lateral fins for undulatory swimming and eyes. However, many important differences (see Table 1) deviate from the molluscan lineage. In addition, the phylogenetic position of Nectocaris below the shelled cephalopods (Nautiloidea), as proposed by Smith & Caron (2010, fig. 3), is inconsistent with the known and logical sequence of events in the cephalopod evolution. Therefore, at the present time, it would be safer to consider Nectocaris pteryx as Dinocaridida incertae sedis (Collins 1996). We, however, refrain from doing so at this moment, waiting for more information on the evolution and diversity of anomalocaridids and their relatives, to be established.

The constructive comments and suggestions provided by Jerzy Dzik and an anonymous reviewer significantly improved our discussion and made our arguments stronger. This is greatly appreciated.