Ontogeny and systematics of Toernquistiidae (Trilobita, Proetida) from the Ordovician of the Argentine Precordillera

Journal of Paleontology, Mar 1998 by Chatterton, Brian D E, Edgecombe, Gregory D, Waisfeld, Beatriz G, Vaccari, Norberto E

Most Hystricuridae show state 0.

49) Interpleural furrows behind first pygidial segment: 0) absent; 1) distinct.

Hystricuridae show state 0 or more commonly 1.

50) Width of interpleural furrows on pygidium: 0) of even width or disappear towards margin; 1) distinctly broader (triangular) distally (often terminated by a sharp border rim).

mall cranidia and Hystricuridae show state 0.

51) Prosopon near pygidial margin (on border where present): 0) smooth; 1) granulose; 2) tuberculose.

This character is variable in Hystricuridae.

52) Pseudoarticulating half ring(s) in pygidium: 0) absent; 1) present.

Hystricuridae appear to show both states. 53) Number of pygidial pleural ribs: 0) 1-2; 1) 3-5; 2) more than 5. Most Hystricuridae show state 1.

54) Marginal rim on pygidium (sharply interrupts interpleural furrows): 0) absent; 1) present.

This is commonly present in Hystricuridae (state 1).

55) Terminal piece of pygidium divided into two lobes/nodes: 0) absent; 1) nodes; 2) knobs.

This character is absent in Hystricuridae. Knobs appear during the meraspid period in Ischyrotoma, as soon as the posterior rings of the pygidial axis develop in the transitory pygidia.

CLADISTIC METHODS AND RESULTS

The 55 characters detailed above were analyzed for 38 taxa using PAUP 3.1.1 (Swofford, 1993). Trees were rooted by specif ying the hystricurids (all species of Hystricurus as well as Psalikilus) as outgroups. Neither the ingroup nor the outgroup were constrained as monophyletic (Nixon and Carpenter, 1993). Heuristic procedures were employed in all analyses. Two hundred random stepwise addition sequences were undertaken, saving no more than 20 trees in each replicate. We then applied TBR (treebisection-reconnection) branch-swapping on these trees with closest addition, holding ten trees in each addition sequence. In all analyses discussed below, polymorphisms were optimized as either of the observed states, rather than as missing, and all multistate characters were treated as non additive.

With these parameters, the total data set results in 138 minimal length cladograms of 305 steps (Consistency Index 0.403; Retention Index 0.642). The strict consensus of these cladograms is shown in Figure 4.1 (a semistrict consensus has the same topology). The separation of toernquistd and dimeropygid clades is a feature of all shortest trees. Some hystricurids (H. globosus and H. scrofulosus) are invariably more closely allied to the dimeropygids, and a majority of trees (67 percent) resolve another group of hystricurids (a clade composed of H. millardensis, H. genalatus, H. paragenalatus, H. penchiensis, H. Nectimembrus, and Psalikilus typicus) as more closely related to the toernquistds than are the dimeropygids. This leads us to favour the classification of Dimeropygidae and Toernquistdae as separate families. Uniting them in a single family to the exclusion of the hystricurids would recognize a polyphyletic group.

Dimeropygid ingroup relationships are well resolved in all shortest trees (with Ischyrophyma within Ischyrotoma; cf., Fortey and Owens, 1975, fig. 2), whereas considerable ambiguity confounds the interrelationships of the toernquistds. Only core groups for Chomatopyge and Mesotaphraspis, most of the species of Paratoernquistia, and a Chinese clade (minuta anxiensis) are unambiguous. New genera proposed in this work, Lasarchopyge and Paratoernquistia, are resolved as clades in 86 and 99 percent of the shortest trees, respectively.