Abstract
Bryozoans of the class Stenolaemata have been diverse and abundant members of the marine epibenthos since the Ordovician and have a rich fossil record. Only one order in this class, the Cyclostomatida, survives in modern seas. Like other marine bryozoans, cyclostomes are small, colonial coelomates equipped with crowns of ciliated tentacles used to gather suspended food particles. There is strong evidence that all stenolaemates shared these basic traits, making extant cyclostomes a potential resource for understanding the biology of long-extinct palaeostomates. However, our knowledge of the cyclostomes is far from being complete. For instance, morphometry of the feeding apparatus and its correspondence with skeletal parameters has not been comprehensively studied. Although some predictive models exist for palaeostomate soft parts, little is known about the extent of intraspecific variability and thus the limitations inherent in such modelling for palaeobiological and palaeoecological reconstructions. Similarly, at the anatomical and ultrastructural level, variation within Cyclostomatida soft parts remains largely unknown. Existing studies made with modern imaging techniques have barely scratched the surface, with only a few genera investigated. Indeed, most studies focus on just one family, the Crisiidae. Thus, it is difficult to assess the degree of morphological consistency within the group at the scale of cells, tissues and organs. Since cyclostomes vary greatly in their life styles and colony composition, the assumption of fine-scale anatomical uniformity needs to be checked.
In this thesis I address both the issues outlined above. In pursuit of these goals I use a variety of methods, ranging from simple light microscopy to TEM and serial block-face SEM, and a number of statistical tools. In Chapter 2 I investigate the relationships between feeding apparatus parameters and skeletal traits within and across 13 species from 8 cyclostome families. I use these data to develop and test predictive models, and outline their limitations if applied to reconstructing extinct stenolaemates. The strongest of these models is subsequently used to assess trophic partitioning of several palaeostome orders (Appendix II). In Chapters 3 and 4 I examine polypide and body wall ultrastructure in four species in two genera from the family Horneridae. Hornerids are phylogenetically distant from the Crisiidae and have different colony organisation and life history. These data enable nested comparisons: (1) between three species within one genus, (2) between two genera in one family, and (3) between disparate cyclostome families. The analysis of morphometry within and across cyclostome species revealed great intraspecific variability, almost without exception disconnected from skeletal traits. At a higher, interspecific level, however, the relationships were positive, linear, and moderately strong, in line with previous reports. Still, single-predictor models of soft-body characters have limited predictive power and applicability. Inferences derived from them are best cross-correlated with reconstructions based on other evidence. Ultrastructural study of the polypides of four hornerids, compared with findings on other genera, support a stablebauplan for Cyclostomatida, with a single exception of the funiculus. The composition of its inner core differs in Hornera, Crisia and Cinctipora, while in the rest of the cyclostome
families this organ remains unexplored. Such disparity hints at either a widespread variability of this trait, or at a more conservative structure with unusual outliers and clearly deserves further study. Organisation of the hornerid body wall differs from that in crisiids, revealing deviations from typical epithelial composition, including missing extracellular matrix (ECM) and disrupted cell continuity across orificial wall and portions of cystid lining. The classic coelomate body wall composition (epidermis — ECM — coelothelium) is only present in an unmodified form in the tentacle sheath and in skeletal attachments of the membranous sac. In addition, the architecture of the atrial polypide attachment(s) in particular emerges as a potentially useful taxonomic trait. Present findings, taken together and applied to the task of reconstructing fossil stenolaemates, do, paradoxically, delineate an area of uncertainty that is unlikely to be clarified, but also help narrow down the range of plausible and biologically meaningful interpretations. This study represents a step toward a better understanding of both living cyclostomes and palaeostomates.