Search

The position of scaphopods in molluscan phylogeny remains singularly contentious, with several sister relationships supported by morphological and phylogenomic data: Scaphopoda + Bivalvia (Diasoma), Scaphopoda + Cephalopoda (Variopoda), and Scaphopoda + Gastropoda. Nervous system architecture has contributed significant insights to reconstructing phylogeny in the Mollusca and other invertebrate groups, but a modern neurophylogenetic approach has not been applied to molluscs, hampered by a lack of clearly defined homologous characters that can be unequivocally compared across the radical body plan disparity among the living clades. We present the first three-dimensional reconstruction of the anterior nervous system of a scaphopod, Rhabdus rectius, using histological tomography. We also describe a new putative sensory organ, a paired and pigmented sensory mantle slit. This structure is restricted to our study species and not a general feature of scaphopods, but it forms an integral part of the description of the nervous system in R. rectius. It also highlights the potential utility of neuro-anatomical characters for multiple levels of phylogenetic inference beyond this study. This potential has not previously been exploited for the thorny problem of molluscan phylogeny. The neuroanatomy of scaphopods demonstrates a highly derived architecture that shares a number of key characters with the cephalopod nervous system, and supports a Scaphopoda + Cephalopoda grouping.

Anna Meyer published her first (and apparently last) scientific paper in 1913, in which she reported, in part, on the organization of the renogenital systems in the “Diotocardia” and “Monotocardia,” and the evolutionary signal present in this character suite. Meyer’s phy- logeny broke with conventional wisdom of her time and she reconstructed early gastropod evolution as a sequence of branching events rather than a continuous grade. Her phylogeny was ignored because it did not correspond to the well-ordered grades that came to dominate gastropod systematics for over 75 years, but her schematics of renogenital configurations have been featured in the literature for over 55 years, sometimes without citation and redrawn to show a different set of relationships. We hope that the translation of her paper from German into English (http://www. ucmp. berkeley. edu/collections/archdut. html) will make her contribu- tion more accessible to workers. Additional

Cell lineage studies in the clade Eutrochozoa, and especially the Spiralia, remains a rich and relatively untapped source for understanding broad evolutionary developmental problems; including (1) the utility of cell timing formation for phylogenetic hypotheses; (2) the evolution of cell timing changes and its relation to heterochronic patterns; (3) stereotypy or lack thereof in rates of change of cell growth during evolution and its relation to both evolutionary history and current usage; and (4) how mosaic cleavage timing variation may be expected to differ from other groups. A compilation of available cell timing information was made from previous studies where each division was explicitly followed and the total number of cells followed was greater than 24. From that compilation, we performed a series of heuristic and quantitative analyses, including a phylogenetic analysis using cell timing data as characters and analyses of timing variation across all taxa. Our results show that: (1) cell lineage data reconstructs a phylogenetic hypothesis that has similarities, especially among the Mollusca. to the patterns found in morphological and molecular analyses; (2) the mesentoblast (4d) is a unique cell compared to other cell in that it speeds up and slows down relative to other cells in taxa with both unequal and equal cell sizes; (3) some cells that form in the same quartet at the same point in the cell lineage hierarchy have much lower variations than analogous other cells, arguing for architectural constraint or stabilizing selection acting on those cells; and (4) although variation in cell timing generally increases during development, timing of formation of progeny cells in the first quartet has lower variation than the parent cells, arguing that some regulation-like behavior might be present.

A three-trophic-level interaction among American Black Oystercatchers (Haematopus bachmani), limpets (Lottia spp.), and erect fleshy algae in rocky intertidal communities of central and southern California was documented via manipulative and “natural” experiments. Removal of the territorial limpet (Lottia gigantea) initially caused large increases in the percent cover of erect fleshy algae, followed by a more gradual increase in density of small limpets (Lottia spp.) and a decline in algal cover. Algal cover increased following the removal of small limpets at the sites from which L. gigantea had been removed earlier, thus demonstrating that the large and small limpets had similar inhibitory effects on plant populations. A comparison of sites with and without oystercatchers showed that L. gigantea occupied substrate inclinations in proportion to their availability at sites where oystercatchers were rare, whereas the distribution of L. gigantea was skewed toward vertically inclined substrates where oystercatchers were common. Survival rates of limpets translocated to horizontal and vertical substrates were similar in sites lacking oystercatcher predation, but were much lower on horizontal substrates where oystercatchers were common. Our results are consistent with those from several prior studies in demonstrating that shorelines frequented by humans typically lack oystercatchers. Humans also exploit L. gigantea and reduce populations to low densities of small individuals. These findings may explain why the midlittoral zone of rocky intertidal communities in western North America are so often dominated by high population densities of small limpets.

Nielsen and Meier (2002) are skeptical of our claim of phyletic information being present in the cell division patterns of spiralian taxa. They conclude that cell division patterns must be known in greater detail, and that coding methods for cell lineage data should be better refined before phyletic information can be identified. As noted by Nielsen and Meier (2002), earlier studies had focused almost exclusively on molluscan taxa. In addition, only one or two cell lineages had been typically considered, and none of these previous studies had used an explicit cladistic methodology to investigate potential phyletic information. Our analyses (as well as Nielsen and Meier's reanalysis of our data [2002, Fig. 1]) produce several groupings that were congruent with both molecular and morphological phylogenies. For example, we recovered an Oligochaete-leech clade, the sister relationship of the two Echiura species, and the overall structure of the gastropod mollusc clade. We also noted that the bivalve mollusc and annelid groupings in our tree were not monophyletic. However, the vast majority of the annelid taxa (Echiura + Polycheata + Pogonophora + Oligochaete + Leech) formed a grade outside the Polyplacophora + Gastropoda. Members of these groups were not randomly interdigitated within other taxa across the tree as would be expected if no phylogenetic information were present (as in the case of some polychaete worms and bivalve molluscs). Polyphyletic polychaetes, bivalves, and other worm groups have also been included in analyses of some molecular datasets so we were not that surprised by this result (e.g., Winnepenninckx et al. 1995). We reported the polyphyly in these and other groups (relative to traditional spiralian classifications) as part of our assessment of the degree of phylogenetic information remaining in cell lineage data. Like all characters, the cell lineage characters provided better resolution and more congruent information in some clades than they did in others. Nielsen and Meier (2002) suggest that our "puzzling phylogenetic results" might be partially explained by the coding of the cell division data into character states for cladistic analysis. Determining a metric by which to score cell lineage data is problematic. Absolute time from fertilization can vary greatly depending on temperature, and cleavage cycle is invariant with cell lineage and therefore noninformative. For our analysis we chose embryonic cell number, a standard comparative metric for spiralian embryos that dates from the late 1890s. More recently, cell number has been used for almost 20 years to argue acceleration and retardation of molluscan cell lineages during early development (e.g., Verdonk and Van den Biggelaar 1983; Freeman and Lundelius 1992; Haszprunar 1993; Van den Biggelaar 1993, 1996; Van den Biggelaar and Haszprunar 1996). For our cladistic analysis these same data were transformed into discrete character states by dividing the cell number by three. This was done to transform the cell number into a character state within PAUP's allowable range for ordered characters (31 steps; Swofford 1998). At that time we considered this to be a relatively assumption-free transformation as well as conservative because small variation in cell numbers would be grouped into a single state by rounding error (e.g., 23, 24, 25 cells would all be scored as state 8) rather than each given a unique state (e.g., Nielsen and Meier 2002, Fig. 2b). We do not doubt that homoplasy was present in the data, but this can be discovered only after the analysis, not a priori. The scenario constructed by Nielsen and Meier (2002) in their Figure 2 presents a similar problem for all studies using embryonic cell number as a metric, not just those that use a cladistic methodology. For example, contrary to Nielsen and Meier's (2002) interpretation, Van den Biggelaar and Haszprunar (1996) did not map developmental characters onto existing phylogenies. Instead, these latter two authors actually constructed alternative trees based on two characters presence or absence of the polar lobe and cell number at time of 4d formation, and we would suggest that Nielsen and Meier's (2002) scenario is much more problematic for these limited datasets than it is for our larger dataset in which all available data were included in the analysis. And, as with molecular data, it was anticipated that any robust phylogenetic signal would have to emerge from the background noise of homoplasy. Although we appreciate Nielsen and Meier's (2002) suggestion to instead map this data onto an existing tree, we wonder why they would accept our coding for mapping the characters but not for using them in the analysis itself. We consider Nielsen and Meier's suggested coding prohibitively restrictive in that differences early in cell lineages are propagated to later lineages. Their argument is that any upstream change in other connected lineages prohibits the downstream coding of similarity. For Nielsen and Meier (2002), although two cells may form at 32 cells, they should only be coded as similar (e.g., putatively homologous) if the