Further observations on scorpion genera Hadrurus and Hoffmannihadrurus (Scorpiones, Caraboctonidae)

Abstract Multiple populations of Hadrurus pinteri from Baja California Sur, Mexico have been examined. It is demonstrated that the southern populations of this species have a larger number of accessory trichobothria (neobothriotaxy) than the northern populations, numbers exceeding the maximum currently recorded for the genus. Examination of carapace and chela coloration and its patterns show a close affinity between Hadrurus pinteri and the dark phase of Hadrurus concolorous. A new morphometric ratio of the carapace is defined that distinguishes Hadrurus from Hoffmannihadrurus, further supporting the monophyly of the latter genus.


Introduction
presented a cladistic analysis of scorpion superfamily Iuroidea. In this analysis they demonstrated the monophyly of genus Hoff mannihadrurus, a taxon which had been recently synonymized by Francke and Prendini (2008). In this present contribution, new information is presented from the evaluation of additional specimens of Hadrurus pinteri Stahnke, 1967 andH. concolorous Stahnke, 1967, all from Baja California Sur, Mexico. Th is information is relevant to the cladistic analysis of Fet and Soleglad (2008) as follows: 1) it supports their hypothesis that the northern populations of H. pinteri are losing accessory trichobothria as demonstrated by a larger number being found in the southern populations; 2) analysis of coloration and its patterns of the carapace and chela of H. pinteri show close affi nities to the dark phase of H. concolorous, not Hoff mannihadrurus gertschi, as proposed by Francke and Prendini (2008); and 3) a new morphometric ratio involving the carapace is defi ned which can be used to diagnostically separate genera Hadrurus and Hoff mannihadrurus. Items 2 and 3 further support the result of Fet and Soleglad (2008) that confi rms the validity of genus Hoff mannihadrurus.

Material
Th e following material was examined for analysis and/or illustrations provided in this paper. It must also be noted that many observations and statistics provided in this paper are augmented, in part, from other data previously collected and discussed in Soleglad (1976), Fet et al. (2001Fet et al. ( , 2004, Soleglad and Fet (2004), and Fet and Soleglad (2008 Stahnke, 1940: Hawthorne, Mineral Co., Nevada, USA (VF).

Neobothriotaxy in Hadrurus pinteri
Neobothriotaxy in Hadrurus was fi rst reported by Gertsch and Soleglad (1972: fi gs 96-107) when the fi rst trichobothrial pattern for this genus was illustrated for H. arizonensis. Th e unusual and complicated pattern exhibited in this genus was later investigated by Soleglad (1976) where he illustrated and provided trichobothria statistics of the chela for eight species (now divided into two genera, Hadrurus and Hoff mannihadrurus). Th is analysis of Soleglad (1976) involved the study of over 200 specimens and the description of new species Hoff mannihadrurus gertschi (assigned to Hadrurus at the time). Th e most important result of this study was that the major species groups could be diagnosed using chelal trichobothrial patterns alone. Th ese diagnoses were based in most part on the presence/absence and numbers of accessory trichobothria on three surfaces of the chela, ventral, internal, and external. In other words, neobothriotaxy could be used in large part to distinguish the species. Years later, Fet et al. (2001) provided the fi rst systematic analysis of the six species of Hadrurus using DNA. Supporting this analysis was morphology analyzed from a cladistic perspective; again, neobothriotaxy contributed several characters. Th e original data set of Soleglad (1976) was expanded by over 50 % with many additional specimens being added, primarily in the "hirsutus" group. At this time the sample set was over 600 (i.e., both chelae are considered). Based on the large collection of neobothriotaxy data from these specimens as well as from chactoid scorpion genera (Anuroctonus and Euscorpius), data that were closely correlated with the specimen's geographical locality, Soleglad and Fet (2004: 102-106) presented an "accessory trichobothria loss" hypothesis. We will not present this hypothesis here, it is discussed in detail in the aforementioned reference as well as in the most recent work of Fet and Soleglad (2008: 273-277). In the latter reference, germane to this study, it was suggested that the accessory trichobothria were being lost in the northern populations of H. pinteri, presumably the direction of the radiation of this species. I.e., the assumption being the species origin is in Baja California Sur, the primary location of the "hirsutus" group, which also includes species H. concolorous and H. hirsutus as well as H. pinteri. However, Fet and Soleglad (2008) had very little H. pinteri material from the south-ern area of Baja California, so this assumption of accessory trichobothria loss in the northern half was not based on substantial data.
In this study, we have analyzed a collection of southern populations of H. pinteri from the California Academy of Sciences. We tabulated the number of chelal accessory trichobothria of the ventral, internal, and external surfaces and compared it to the data of the northern populations. Figure 1 shows the geographic localities of H. pinteri from which the accessory trichobothria data are derived. Th e northern samples, totaling 13 specimens, are from the northern half of their known range, the majority of specimens from Oakies Landing. Th e most southern of these specimens is from the Arroyo Calamajué. Th e southern samples, comprised of 28 specimens, span the entire southern range originally outlined by Williams (1970). Th e most northern specimens, from Punta Trinidad, are farther north than originally reported by Williams (1970), these specimens not listed in his material examined. San Miguel de Comondú provided the most specimens, ten in number. Interestingly, the range of H. pinteri, as stated by Williams (1970) and shown in our map ( Fig. 1) is disjunct, roughly 150 km separating the northern and southern ranges. H. pinteri, as reported by Williams (1970: 18): "… was never found in predominantly sandy habitats or away from habitats of volcanic origin …". As can be seen in Fig. 1, the terrain is certainly mountainous, the eastern portion of the peninsula volcanic from Puertocitos to well south of Loreto. What is interesting in this map is that the area of disjunction between the two ranges of H. pinteri is also volcanic leading to the conclusion that H. pinteri distribution is probably not disjunct. Th e gap in reported localities is probably due to the lack of collecting in this area. Williams and his associates conducted their monumental collecting expeditions in Baja California during the late 1960s, Williams (1980: 1), in his monograph on the scorpions of Baja California, Mexico, reported that 60,000 specimens were examined from Baja California! However, during this time, access to this particular eastern area of Baja California was diffi cult if not impossible, as the primary road (unpaved during the time) was on the western side of the peninsula. Table 1 shows the statistical breakdown of accessory trichobothria for the entire "hirsutus" group, involving more than 260 samples. Of importance to this discussion is the breakdown of H. pinteri into its northern and southern populations. As predicted by Fet and Soleglad (2008), we see that the southern populations of H. pinteri exhibited larger numbers of accessory trichobothria than the northern samples in two of three chelal surfaces, the internal accessory trichobothria showing slightly higher numbers in the northern populations. For the ventral series, the mean value of the southern populations was 8 % larger than that seen in the northern populations, roughly two additional accessory trichobothria on an average. Similarly, the external surface of the southern populations exhibited over a 7 % mean value diff erence. For the internal series, the northern populations had 1.4 % more accessory trichobothria on an average, though the standard error range maximum value was slightly larger in the southern population.
Of particular interest, we found that the southern populations of H. pinteri exhibited the largest number of accessory trichobothria found in the three individual chelal surfaces for the entire genus Hadrurus (as well as for its sister genus Hoff mannihadrurus). Previously, based on data presented in Fet et al. (2004), the largest number of internal, ventral, and external accessory trichobothria were seven (H. arizonensis), 27 (23 accessory, H. pinteri), and four (H. pinteri and Hoff mannihadrurus gertschi), respectively. In these newly examined populations of H. pinteri we found seven internal accessory trichobothria in two specimens, from Loreto and Agua Verde Bay, thus matching the maximum counts found in H. arizonensis. For the latter species, four specimens exhibited seven internal accessory trichobothria, one from ABDSP in California, and three from Sonora, Mexico, all belonging to the pale form of this species (previously referred to as H. a. pallidus; synonymized by Fet et al., 2001). Now, comparing H. pinteri to H. arizonensis, which includes subspecies H. a. austrinus, we see that the former species mean value is slightly larger, 5.637 vs. 5.490 (note that 253 samples represent H. arizonensis). For the southern populations of H. pinteri, we found a specimen with 32 ventral trichobothria (28 accessory), exceeding the previous largest number by fi ve. In addition to the largest ventral number, one specimen (from Punta Trinidad) exhibited 30 ventral trichobothria, fi ve with 29, and four with 28. Th e largest ventral trichobothria count from the northern half of Baja California is 26. Th e mean value diff erence between the two localities in Baja California is 8 %. We encountered no less than fi ve instances of fi ve external accessory trichobothria Table 1. Statistics showing neobothriotaxy of the pedipalp chela of the Hadrurus "hirsutus" group. In particular, two disjunct populations of H. pinteri are contrasted showing that for two of the three trichobothrial series, the southern population (i.e., Baja California Sur) has the largest number of accessory trichobothria, roughly a 7.5 % increase. MVD = mean value diff erence; p-value = Anova output. Statistical data group: minimum-maximum (mean) (±SDEV) [N] {standard error range} (coeffi cient of variability) (SDEV/mean). * includes orthobothriotaxic trichobothria V 1 -V 4 . Many of the statistics are from previous studies, as well as new material examined in this project. See Soleglad (1976), Soleglad and Fet (2004), Fet et al. (2004), and Fet and Soleglad (2008). (Punta Trinidad, San Miguel de Comondú, and Isla Danzante). Th is count includes the diagnostic and unique accessory trichobothrium found on the base of the fi xed fi nger. Figures 2-4 illustrate examples of these large accessory trichobothria numbers for each chelal surface.

Coloration and Patterns of Hadrurus and Hoffmannihadrurus
While examining the southern populations of H. pinteri, we discovered that two of the specimens were in fact not H. pinteri, but H. concolorous. Based on coloration and patterns these two specimens certainly looked like H. pinteri, only after detailed trichobothrial analysis could we isolate the two specimens from H. pinteri. One specimen, from Tambobiche, had a somewhat small number of accessory trichobothria, only 14-15 ventral and three internal. Th e other specimen, from San Miguel de Comondú, exhibited 18-19 ventral, fi ve internal, and one external accessory trichobothria. In both specimens, the diagnostic accessory trichobothrium on the fi xed fi nger was absent. Of special interest, the specimen from San Miguel de Comondú was contained in a vial with a large H. pinteri male, thus they were collected together. Figures 5-9 show the carapaces of H. pinteri, the two H. concolorous misidentifi ed for H. pinteri, and two additional color phases of H. concolorous. Th e carapaces of H. pinteri and the H. concolorous from Tambobiche are indistinguishable, both uniformly dark in color. Th e carapace of the specimen from San Miguel de Comondú is lighter in color, more close to the reddish specimen from Santa Rosalia (Fig. 8). Th e specimen from the sand dune area in Las Bombas (Fig. 9) is typical of H. concolorous, as indicated by its name "concolorous".
Th e chela of H. pinteri (Fig. 13) and the H. concolorous from San Miguel de Comondú (Fig. 14) are indistinguishable. Again, the chela from the Las Bombas specimen (Fig. 15) is typical of the "concolorous" phase of H. concolorous. Figures 11-12 show the carapacial coloration and patterns of Hoff mannihadrurus gertschi and H. aztecus. Although the carapace of H. gertschi is considerably darker, its interocular area is lighter in color, exhibiting similar light/dark patterns as seen in its sister species H. aztecus. It is clear that the carapace of H. gertschi is not patterned as in Hadrurus pinteri (Fig. 5). Similarly the chela of H. gertschi (Fig. 16) is darker than that seen in Hadrurus pinteri (Fig. 13). Francke and Prendini (2008) attempted to associate the carapace and chelal coloration/patterns of H. gertschi to that of Hadrurus pinteri (their "coloration" characters 2-3, and 5). However, the dark color phase of H. concolorous is much closer to H. pinteri, a more reasonable association since the two species are closely related and share the same geographical area! Th e above observations of essentially identical coloration in H. pinteri and the dark phase of H. concolorous in conjunction with the lack of close compliance to Hoff mannihadrurus gertschi further endorses the observations of Fet and Soleglad (2008: 272-273).

Carapace Morphometrics of Hadrurus and Hoffmannihadrurus
While studying the carapacial coloration and its patterns in Hadrurus and Hoff mannihadrurus we observed that the convexed anterior edge exhibited in both genera was considerably more exaggerated in Hoff mannihadrurus. Th is is quite visible in the photographs presented in Figs 5-12 as well as in Fet et al. (2004: fi gs 57-58) where Hadrurus pinteri and Hoff mannihadrurus aztecus are shown. We analyzed this diff erence in  (Figs 6-9,  14-15) exhibiting dark coloration patterns essentially identical to H. pinteri (Figs 5, 13) to pale yellow with little or no patterns. Also of interest is the darken posterior half of the carapace in Hoff mannihadrurus gertschi, matching the same area that is also darkened in H. aztecus 5, 13  the anterior edge of the carapace in both genera, including all species and subspecies of Hadrurus. Several morphometrics were obtained in an attempt to quantify this visible diff erence between the two genera. From these morphometrics we discovered that the anterior edge (measured from the lateral eyes to the distal aspect of the carapace, see Figure 17) was longer in Hoff mannihadrurus. We concluded that the more overt convexed aspect seen in the Hoff mannihadrurus carapace contributes directly to its elongation. Th is was evident from two ratios we constructed: the anterior edge length as compared to the carapace length and the anterior edge length as compared to the position of the median eyes. Th e results derived from these two ratios are nearly identical, so we present here only the latter of the two ratios. Figure 17 illustrates exactly how these two measurements are taken and Table 2 presents the results involving 30 samples spanning all species of Hadrurus (20 samples), and Hoff mannihadrurus (ten samples). Th e sampling included four Hoff mannihadrurus gertschi, six H. aztecus, and ten samples each from the Hadrurus "arizonensis" group (i.e., both H. arizonensis subspecies, H. obscurus, and H. spadix), and the "hirsutus" group (i.e., H. pinteri, H. concolorous, and H. hirsutus). Th e mean value diff erences Table 2. Statistics showing diff erences in the length of the carapace anterior edge in subfamily Hadrurinae based on the following morphometric ratio: anterior_edge_length / median_tubercle_position. See Fig. 17 for methods of measurement. Data shows that the anterior edge of Hoff mannihadrurus is approximately 48 % longer than in Hadrurus. Large standard error range separation and a very small p-value from variance analysis further support the signifi cant statistical diff erence between the two genera. Statistical data group: minimum-maximum (mean) (±SDEV) [N] {standard error range} (coeffi cent of variability). * Mean value diff erence, standard error range separation, and analysis of variance. Statistical data derived from specimens examined and the following references: Williams (1970), Stahnke (1971), Soleglad (1976), and Fet et al. (2004). of this morphometric ratio between these genera is 47.7%, implying that in Hoff mannihadrurus, the anterior edge of the carapace is roughly 50 % longer than in Hadrurus.
[note that the carapace of H. concolorous from Tambobiche (Fig. 6) was not included in the morphometric sampling due to its obvious damaged anterior edge.] Other relevant statistical indicators are no overlap of the absolute range, over 200 % separation of the standard error range, and a very small anova p-value of 6.12E-16. Fet and Soleglad (2008) presented a detailed cladistic analysis of the superfamily Iuroidea demonstrating monophyly of the families Iuridae and Caraboctonidae, the subfamilies Caraboctoninae and Hadrurinae, and the genera Hadrurus and Hoff mannihadrurus. Th is resulted in the reestablishment of the genus Hoff mannihadrurus, which had been recently synonymized by Francke and Prendini (2008). Fet and Soleglad's (2008) approach was to present their cladistic analysis in three successive layers: fundamental characters, low-level characters, and characters based on the accessory trichobothria loss hypothesis. Th e fi rst layer presented characters that dealt with higher-level systematic aggregates, characters that were assumption-and hypothesis-free. Low-level characters dealt with coloration and its patterns, setation, and etc., which are generally species-level characters. Th e last layer of characters, based on the loss hypothesis, as its name implies, is a hypothesis, thus formed from a set of assumptions. Although the hypothesis has been studied across three separate scorpion groups involving thousands of scorpions, it is still a hypothesis. It must be noted here that the monophyly of Hoff mannihadrurus was demonstrated at the fi rst layer of cladistic analysis, using only fundamental characters. As the other two layers were added, successively, this monophyly was further demonstrated with larger support (i.e., more characters and greater bootstrap/jackknife support). Th e new character described above showing diff erences in the anterior edge length between Hadrurus and Hoff mannihadrurus represents the fi fth new character supporting monophyly of Hoff mannihadrurus (i.e., four were previously identifi ed by Fet and Soleglad (2008)); these fi ve characters were not included in the analysis by Figure 17. Method of measurement of carapace for genera Hadrurus and Hoff mannihadrurus for determining morphometric ratio anterior_edge_ length (x) / median_tubercle_position (y). Shaded area indicates anterior_edge_length. Diagrammatic drawings based on Hadrurus pinteri and Hoff mannihadrurus gertschi. Francke and Prendini (2008). In this study, we added this character to the original cladistic analysis presented by Fet and Soleglad (2008: 265) by adding a new state (= 3) to the fundamental character 23 (carapace anterior edge). We then reinitiated the fundamental and fi nal character cladistic sequences with the following results: for the fundamental sequence, instead of three MPT's (most parsimonious trees) we obtained two; the bootstrap/jackknife results for monophyly of Hoff mannihadrurus improved from 68/66 % to 88/83 % (that is, 88/83 percent of the 5000 pseudoreplicates supported this monophyly); for the fi nal sequence, bootstrap/jackknife results for monophyly improved from 99/97 % to 100/98 %; and fi nally, character 23: state = 3 distributed unambiguously on the Hoff mannihadrurus node in both sequences, which is clearly a demonstrated synapomorphy for genus Hoff mannihadrurus. Refer to Fet and Soleglad (2008) for details in this cladistic analysis and defi nitions of specialized terminology.