Urodacidae Pocock, 1893

Soleglad, Michael E. & Fet, Victor, 2003, High-level systematics and phylogeny of the extant scorpions (Scorpiones: Orthosterni), Euscorpius 2003 (11), pp. 1-175 : 115-116

Family Urodacidae Pocock, 1893

Type Genus. Urodacus Peters, 1861 .

Synonyms.

Heteroscorpionidae Kraepelin, 1905 , new synonymy (valid as subfamily name).

Composition. The content of Urodacidae is changed here compared to that of Prendini (2000), as we include the subfamily Heteroscorpioninae , downgraded from family rank. We also reestablish the nominotypic subfamily Urodacinae . Thus, the family includes two subfamilies—one from Australia, another from Madagascar —with one genus each.

Distribution. Australia and Madagascar.

Taxonomic history. This taxon, long treated as a subfamily of Scorpionidae , was first elevated to the family level by Lourenço (1996a) who followed the unpublished classification of Stockwell (1989). Fet (2000g) still listed it as a subfamily of Scorpionidae , noting that the change in rank requires further justification. Such justification was provided by Prendini (2000) who upheld Urodacidae as a family and a sister group to family Heteroscorpionidae .

Biogeographic history. Subfamilies Urodacinae and Heteroscorpioninae can be interpreted as remnants of Gondwanaland scorpionoid fauna—two relict genera, one surviving in Madagascar, another in Australia. Koch (1977) discussed biogeographic distribution and desert adaptations of the diverse Australian genus Urodacus . Lourenço (1996a, 1996d) discussed Heteroscorpion in the context of endemic fauna of Madagascar, which generally is related to the African fauna, diverging with the split of Gondwanaland.

Diagnosis. (This diagnosis is based on derived characters for clade ( Heteroscorpionidae + Urodacidae ) as presented in Prendini (2000: Fig. 7)). Synapomorphies. Chelal fingers with multiple (>2) median denticle (MD) rows; ventral surface of pedipalp patella with single row of 4–20 trichobothria; external surface of pedipalp patellae with 14 or more trichobothria; ventral surface of pedipalp chela with six or more trichobothria; metasomal segments I–IV with single ventral median carina.

Discussion. Based on the many important characters, which genus Heteroscorpion uniquely shares with the family Liochelidae , and likewise, does not share with the genus Urodacus , we decided to investigate Prendini’s (2000) original cladistic analysis which combined these two genera as sister elements. This questioning of the clade “ Urodacus + Heteroscorpion ” was precipitated, in part, by the somewhat “high-level” approach to neobothriotaxy taken by Prendini (2000), which was discussed in detail in Soleglad & Sissom (2001: 71–73). They pointed out that Prendini considered almost all neobothriotaxic conditions found within the superfamily as single derivations within the pedipalp segment surfaces. This approach, in the opinion of Soleglad & Sissom (2001), predictively created severe homoplasy (i.e., the simplistic model did not convey true evolutionary lines for this complicated set of derivations). As stated in the discussion of Soleglad & Sissom (2001), three of these characters, those involving the chelal ventral surface (Prendini’s character 49), and patellar ventral and external surfaces (characters 43 and 45), exhibited the lowest overall character support in his entire analysis. Notwithstanding Prendini’s recent retort (2003b: 155) concerning the existence of “unambiguous homoplasious synapomorphies” —a fact Soleglad & Sissom (2001) never questioned —Soleglad & Sissom’s comment was aimed directly at Prendini’s superficial modeling of neobothriotaxy, and in particular, questioned the clade “ Urodacus + Heteroscorpion ” which was based on five synapomorphies, three of which involved his neobothriotaxy model. By doing a cursory study of the neobothriotaxy exhibited in Heteroscorpion and Urodacus , it is clear that the two events are separate

Diplocentridae Scorpionidae Urodacidae Hemiscorpiidae Liochelidae Heteroscorpionidae

Stockwell (1989) and Soleglad & Fet (hypothetical) Prendini (2000)

derivations: in the patella external surface, none of the series match either in number of accessory trichobothria or in gross positions. In particular, the eb series is orthobothriotaxic in Heteroscorpion (as well as in the hemiscorpiine genus Habibiella ), and contains two accessory trichobothria in Urodacus (note: this distinction is obscured in two species exhibiting massive neobothriotaxy, Heteroscorpion magnus Lourenço and Urodacus yaschenkoi (Birula)) . We can conclude that the three characters assigned to the same state by Prendini for these two genera are incorrect (they were assigned to other scorpionoid taxa as well). Consequently we digitized Prendini’s (2000: Table 3) original data matrix and made the following alterations: 1) assigned separate states to his neobothriotaxy characters 43, 45, and 49 for the two genera in question (but retained the mappings for the other genera with neobothriotaxy); 2) changed Heteroscorpion’s state for character 11 to indicate subequal distal denticles on dorsal/ventral edges of cheliceral movable finger. This state change is indicated by viewing Figs. 11, 12, and 14 of three species of Heteroscorpion illustrated by Lourenço (2002b), two of which clearly show the denticles are subequal. Although the distal denticles are not subequal in Fig. 11 ( H. opisthacanthoides ), they do not exhibit the significant difference commonly seen in other scorpionoids as illustrated in our Fig. 45 ( Scorpio ) and Fig. 46 ( Brachistosternus ); 3) character 9, “shape” and dimensions of the sternum, was modified to indicate the posterior tapering as exhibited in Heteroscorpion (as indicated in a specimen observed by us). In addition, we augmented this character to reflect the two sternum types recently defined by Soleglad & Fet (2003); 4) character 33, Prendini’s modeling of the chelal finger median denticle (MD) rows, was changed to two rows for Heteroscorpion since two rows are visible on the distal third of the finger, although fused into “many rows” basally. We also question this somewhat simplistic modeling of this complex structure, considering “one-row”, “two-rows”, and “multiple rows (>2)”, where we believe Heteroscorpion is intermediate. However, this character requires some serious reanalysis involving many species in several genera, something not possible with the token species set used in the “exemplar method”; 5) as with neobothriotaxy, we question Prendini’s character 95, where he assigns three disparate genera groups that exhibit a single ventral median carina on metasomal segments I–IV to the same state, Heteroscorpion , Urodacus , and Hemiscorpius + Habibella. We assign each group its own state thus removing this assumption of homologous derivation (which also uncouples Heterometrus from Hemiscorpiinae as well).

Figure 126 shows the result of these changes to Prendini’s data matrix. Interestingly, our result is the same as that originally proposed by Stockwell (1989: Figs. 251, 259), Heteroscorpion binding with the liochelines and hemiscorpiines, and Urodacus grouping with the scorpionines and diplocentrines. This grouping has an intuitive appeal because, if for no other reason, the very unique and unprecedented character of the rounded lateral distal lobes of the leg tarsus is now cleanly distributed within the superfamily (i.e., nonhomoplasious).

Since our analysis is cursory at best, especially in issues involving the chelal finger dentition, we have not made changes to Prendini’s result. Additional study is required in several areas, in particular of chelal finger dentition across the superfamily Scorpionoidea , and a serious analysis of neobothriotaxy within closely related groups involving statistical analysis of many specimens across the species set. In summary, here is a list of characters (some of which were discussed above) that Heteroscorpion uniquely shares with subfamilies Liochelinae and Hemiscorpiinae : leg tarsus not exhibiting rounded lateral distal lobes (a key character); simple venom glands (a key character?); dorsal and ventral distal denticles of the cheliceral movable finger approximately same length; sternum tapers posteriorly; sternum is longer than wide; patella internal surface highly vaulted into a “projection” (a key character); and the chelal palm is somewhat flat, not exhibiting the highly vaulted palm commonly found in Scorpioninae and Diplocentrinae .