66 



M. RUTA 



anterior half than in its posterior half. Anteriorly, it appears to be very 

 slightly raised above the level of the internal surface of the zygal 

 process. Posteriorly, it becomes almost semicylindrical in shape. The 

 transition between the anterior half and the posterior half is nearly 

 abrupt. The anterior end of the ridge merges gradually into the 

 internal surface of the zygal process. 



Stereom. Unlike the marginalia, all integument plates are made 

 of a 2-dimensional stereomic meshwork, consisting of regularly 

 arranged, subcircular to subhexagonal pores separated by short 

 trabeculae. The stereom fabric of the marginalia differs slightly from 

 that of the dorsal and ventral integuments. In addition, the texture of 

 the peripheral margin is more irregular than that of the remaining 

 surface area of the marginalia. 



There are indications that the suranal plate likewise consisted of a 

 2-dimensional meshwork, at least peripherally. However, the central 

 part of this plate may have been formed by a 3-dimensional meshwork, 

 albeit considerably less thick than the marginalia (e.g. PI. 2, fig. 4). 

 The stereom of the suranal plate is most characteristic in that it 

 changes remarkably from the centre towards the periphery of the 

 plate. The stereom texture of a roughly bell-shaped area of the 

 external surface of the suranal plate comprised between its centre 

 and its anterior margin is of the reticulate type observed in the 

 supracentralia and infracentralia, although the size of the holes is on 

 average slightly larger than in the integument plates. The size of the 

 holes becomes progressively smaller towards the lateral margins and 

 the antero-lateral angles of the plate, where the trabeculae are thicker 

 and shorter than in its central part. The smallest size of the holes is 

 observed at the level of a broadly crescentic area lying immediately 

 anterior to the posterior margin of the plate, where the trabeculae 

 seem to be flat-topped. Trabeculae and holes become more irregular, 

 antero-posteriorly elongate and arranged according to a broadly 

 fringe-like pattern along the posterior margin of the suranal plate. 

 The fringe extends slightly posterior to the rearmost ends of the left 

 and right parts of the thecal frame and consists of narrow, finger- 

 shaped trabeculae of approximately equal length separated by 

 elongate holes of variable length and outline. Rarely are adjacent 

 trabeculae connected by thin, transverse rods or walls. The rearmost 

 ends of the trabeculae are extremely variable in shape (spatulate, 

 pointed or rounded), so that the fringe appears to have an irregular 

 oudine. The lateral ends of the fringe consist of remarkably short 

 trabeculae with rare intercalated holes, and pass abruptly into the 

 postero-lateral angles of the suranal plate. 



The texture of the internal and external surface of both the 

 supracentralia (PI. 1, figs 2-5; PL 2, figs 2^, 5; PI. 3, figs 2, 4; PI. 5, 

 figs 1-2; PL 7, fig. 2, 4-5; PI. 8, fig. 3) and the infracentralia (PL 2, 

 figs 1, 3; PL 3, figs 1, 3, 5; PL 4, figs 1-2, 4) does not change 

 remarkably, although in the case of the supracentralia, their internal 

 surfaces show smaller, more irregular and more widely spaced pores 

 than their external surface. 



Changes in the stereom texture of the marginalia are best de- 

 scribed by reference to the illustrations (e.g. PL 1 , figs 2-5; PL 2, figs 

 4-5; PL 3, fig. 3; PL 6, figs 1-3; PL 7, fig. 2; PL 8, figs 1-2). I only 

 point out that the most remarkable feature of the stereom of the inside 



of the marginalia is the transverse elongation of some pores in the 

 deepest portions of the trough-like internal excavations of the plates 

 (e.g. PL 4, fig. 3). 



Aulacophore 



Proximal part. The proximal part of the aulacophore is invari- 

 ably disrupted or missing in the material studied (PL 1, fig. 2; PL 2, 

 fig. 5: PL 3, figs 1, 3, 5; PL 5, figs 1-3; PL 7, figs 2, 6; PL 8, figs 1- 

 2). Tectals and inferolaterals are likely to have formed two dorsal and 

 two ventral longitudinal rows respectively, although their precise 

 arrangement cannot be reconstructed accurately. In few specimens 

 (e.g. PL 7, fig. 2), it is possible to observe three to four small 

 segments aligned longitudinally despite the extensive disrupdon and 

 partly overlapping each other, although the latter condition may 

 represent an artifact of preservation rather than a genuine feature. 



Some of the plates of the proximal aulacophore are strongly 

 arcuate in transverse section and approximately square in outline. 

 One of their two longest margins is gently convex whereas the other 

 margin is slightly concave. The two shortest margins of the plates in 

 question differ in length and shape; the shorter of these margins is 

 straight, whereas the longer one is broadly concave and projects into 

 a small conical process at one of its extremities. 



By comparison with the proximal part of the aulacophore of 

 Reticulocarpos hanusi, as reconstructed by Jefferies & Prokop ( 1 972), 

 I suggest that the arcuate plates in question may belong to the 

 longitudinal ventral series and, therefore, may correspond to 

 inferolaterals. If this interpretation is correct, the conical processes 

 would occupy the ventral, antero-lateral angles of the arcuate plates, 

 following the anatomical orientation adopted in this paper. 



Another kind of plate is represented by slightly arcuate elements 

 without conical projections and rectangular in outline (e.g. PL 1 , fig. 

 2). These may belong to the longitudinal dorsal series and may thus 

 correspond to tectals. As in Reticulocarpos, the plates of the proxi- 

 mal part of the aulacophore probably decreased in size to a small 

 extent anteriorly. This is suggested by the fact that the posterior part 

 of the styloid oiJuUaecarpus, around which the most anterior tectals 

 and inferolaterals were probably wrapped up in life, is just smaller 

 than the thecal excavation for the aulacophore insertion. 



It is impossible to say whether these two kinds of plates were 

 sutured along their lateral and medial margins to form complete 

 tetramerous rings or whether elements of the dorsal and/or the 

 ventral longitudinal series were separate. Likewise, it is impossible 

 to ascertain whether the plates formed opposite or alternate rows 

 dorsally and/or ventrally. The presence of small intercalary plates, 

 reported by Jefferies & Prokop (1972) in the proximal part of the 

 aulacophore of Reticulocarpos, is difficult to document in 

 Juliaecarpus. Small, subrhomboidal elements found in some speci- 

 mens in the region of the proximal aulacophore may represent 

 intercalary plates or fragments of tectals and inferolaterals (e.g. PL 5, 

 figs 1-2). 



Intermediate part. The styloid is a robust structure, slightly 

 longer than wide, consisting of a saddle-shaped posterior part and a 



PLATE 7 



Juliaecarpus milnerorum gen. et sp. nov. Late Ordovician, Upper AshgiH, Rawtheyan, Upper Ktaoua Formation, Morocco. All latex casts coated with 

 ammonium chloride, la, b, EE 3175, x 20; posterior aspect of isolated ossicle; note changes in stereom structure along the spike. 2, EE 3070, x 15; 

 complete but slightly flattened theca in dorsal aspect with poorly preserved aulacophore; note dorso-lateral projections of M',. and M^. 3, EE 3141, x 10; 

 incomplete and heavily disrupted theca in dorsal aspect; note massive aspect of posterior zygal process in the centre of the photograph. 4, EE 3184, x 5; 

 incomplete and slightly disrupted theca in dorsal aspect; note well-preserved and articulated supracentralia, and disrupted zygal bar. 5, EE 3152, x 6; 

 poorly preserved theca in ventral aspect with damaged and displaced infracentralia. 6, EE 3146, x 10; poorly preserved aulacophore in right lateral view, 

 with complete styloid and broken first ossicle articulated with it. 



