A NEW STYLOPHORAN ECHINODERM 



51 



posterior zygal plate with the posterior part of the right thecal 

 margin. Numerous isolated marginal elements reveal details of their 

 stereom texture and the degree of curvature of their internal surfaces 

 (i.e. the surfaces oriented towards the internal side of the theca). 



Rarely is the aulacophore found in place. Its proximal part, when 

 preserved, is invariably incomplete or crushed. Its intermediate and 

 distal parts are sometimes found articulated with each other and 

 slightly disrupted. More often, only disarticulated styloids, cover 

 plates and ossicles are found. 



The fine matrix in which the skeletal moulds occur reproduces 

 changes in the stereom texture of various skeletal elements, both 

 externally and internally. In some cases, the external surface of the 

 specimens appears to be heavily weathered and covered with a 

 network of criss-crossed cracks, often obliterating almost com- 

 pletely both plate boundaries and skeletal fabric. 



The fossil material required little mechanical preparation. Some 

 specimens were dissected with a needle to partially expose the 

 appendage and the peripheral margins of the thecal frame. Almost all 

 specimens were cast with black-stained latex. Latex peels were 

 coated with ammonium chloride sublimate and photographed at a 

 low angle of illumination to emphasize details of the external skel- 

 etal surface (see also Ruta & Bartels, 1998). 



The reconstruction of the external aspect of the animal and of 

 several details of its external and internal anatomy (e.g. suranal plate, 

 apophyses, internal surface of M', , peripheral flange, styloid, ossicles, 

 etc.) are based on camera lucida sketches drawn to scale and on a 

 series of engineering projections in dorsal, ventral, lateral, anterior 

 and posterior views (Fig. 2A-E). The engineering drawings were 

 based on morphological information gleaned mainly from the fol- 

 lowing specimens (see also list of the material examined below): 

 BMNH EE 3069a, b, 3070a, b, 3072a, b, 3101a, b, 31 15a, b, 31 19a, 

 b, 3127a, b, 3144a, b, 3163, 3178a, b. The cross-section of the 

 marginalia was reconstructed from the latex peels by cutting these 

 with a razor blade at different levels perpendicular to the direction of 

 maximum elongation of each marginal plate. Observations were 

 carried out using a Nikon SMZ-10 binocular microscope. 



SYSTEMATIC PALAEONTOLOGY 



Phylum ECHINODERMATA Fleming, 1828 



Class STYLOPHORA Gill & Caster, 1960 



Order ANKYROIDA Parsley, 1997 



Family unassigned 



Remarks. As conceived by Parsley (1997, 1998), the order 

 Ankyroida includes the former order Mitrata Jaekel, 1918 and some 

 representatives of the former order Cornuta Jaekel, 1901, the latter 

 almost universally considered as a paraphyletic group by sty lophoran 

 workers (e.g. Jefferies, 1986; Cripps, 1991; Daley, 1992; Woods & 

 Jefferies, 1992; Cripps & Daley. 1994; Parsley, 1997, 1998;Lefebvre 

 et al.. 1998). Although I am confident about the monophyly of the 

 Ankyroida, I shall not elaborate upon the interrelationships between 

 representatives of this group. These ought to be reassessed after the 

 inclusion of several recently described taxa in a new, expanded 

 character matrix for all known stylophorans. 



I have not included a family in the systemafic palaeontology 

 section above. In this, I follow some systematists' view that tradi- 

 tional systematic categories ought to be abandoned (e.g. see 

 discussion in Craske & Jefferies, 1989) and that hierarchically 

 clustered clades within a monophyletic group do not need naming. 

 This approach avoids the introduction of additional names in the 



systematic literature and eliminates ambiguities resulting from con- 

 tinuous redefinitions of previously established taxonomic names. A 

 rankless phylogenetic scheme is especially desirable in the case of 

 extinct groups (such as stylophorans) with a very discontinuous 

 fossil record, because the discovery of additional taxa and/or charac- 

 ters is likely to affect taxon relationships and character distribution to 

 a considerable extent. 



Because very few subgroups within the stylophorans can be 

 shown to be monophyletic (see discussion in Ruta, 1 998, in press), a 

 subdivision of the whole class into taxonomic ranks seems to be 

 inadvisable at present (but see also Caster, 1952 and Ubaghs, 1968, 

 1969). Current taxonomic subdivisions, especially for the cornutes 

 (Cripps, 1991; Daley, 1992; Cripps & Daley, 1994), ought to be 

 reconsidered. The major problem with them is the fact that they have 

 been generated by character coding built upon assumptions of char- 

 acter transformation at the cornute-mitrate transition (Lefebvre et 

 al.. 1998). Assumpfion-laden choice of some characters, based on 

 theory and process, also characterized a recently published data set 

 of mine (Ruta, 1997). 



The results of other phylogenetic analyses of the stylophorans as 

 a group or of different subsets of them differ significantly. Thus, 

 there is need for an integration of the various data sets and for a 

 critical evaluation of the characters used. 



In their monograph on Middle Ordovician cornutes from Nor- 

 mandy, Cripps & Daley (1994: 107) erected the family 

 Reticulocarpidae, for which they provided the following diagnosis: 

 'The Reticulocarpidae are cornutes with optic embayments, slit-like 

 terminal mouths, no branchial skeleton and ventral spikes on the 

 mid- and hind-tail ossicles'. As I shall explain in detail below, the 

 stylophoran described in the present work is extremely similar to the 

 genus Reticulocaipos. 



However, in the light of Parsley's (1997, 1998) cladistic analysis 

 and of the various problems associated with Cripps' (1991) and 

 Cripps & Daley's ( 1 994) character coding, I question the sister group 

 relationship between Reticiilocarpos haniisi Jefferies & Prokop, 

 1972 and Beiyllia miranda Cripps & Daley, 1994, grouped together 

 in the Reticulocarpidae by Cripps & Daley (1994). 



I am more inclined to accept Parsley's ( 1997, 1998) conclusion 

 that Reticiilocarpos and Prokopicystis mergli Cripps, 1 989<7 are each 

 other's closest relatives and that, together, they form the sister group 

 of the genus Hanusia Cripps, \9%9b (see also section on morpho- 

 logical comparisons below). 



Parsley's (1997, 1998) data matrix likewise presents some prob- 

 lems (e.g. presence of some redundant characters, poorly defined 

 states, low number of characters, wrong coding for certain features 

 in some stylophoran taxa, etc.). Despite this, a preliminary parsi- 

 mony analysis of Parsley's (1997, 1998) data matrix using PAUP 

 version 3.1.1 (Swofford, 1993) (with inclusion of modified coding 

 for several of his chosen characters, exclusion of redundant charac- 

 ters and corrections of several wrongly coded characters) produced 

 results similar to those of his original run (the latter was based on 

 reweighting of characters) and preserves in part the original tree 

 structure. These results will be extensively discussed in a separate 

 publication in conjunction with a revision of stylophoran interrela- 

 tionships. 



Genus JULIAECARPUS nov. 



Type species. Juliaecarpus milnerontm sp. nov.; late Ashgill (late 

 Rawtheyan). Upper Ktaoua Formation. Morocco. The only known 



species. 



Derivation OF NAME. After Julia J. Day (Department of Biology, 



