A multilocus molecular phylogeny of the endemic North American camel spider family Eremobatidae (Arachnida: Solifugae)

Highlights

• We provide the first family-level phylogenetic analysis of camel spider relationships.

• Several genera and a major group comprising most species of the subfamily Eremobatinae are monophyletic.

• The subfamily Therobatinae and its component genera are not monophyletic.

• Eremobatidae originated in the late Eocene to early Miocene as North American deserts began to form.

• We provide a framework for future taxonomic and biogeographical investigations.

Abstract

Camel spiders (Solifugae) are a diverse but poorly studied order of arachnids. No robust phylogenetic analysis has ever been carried out for the order or for any family within the Solifugae. We present a molecular phylogenetic analysis of the endemic North American family Eremobatidae Kraepelin, 1899, the first such analysis of a family of Solifugae. We use a multi-locus exemplar approach using DNA sequences from partial nuclear (28S rDNA and Histone H3) and mitochondrial (16S rRNA and Cytochrome c Oxidase I) gene loci for 81 ingroup exemplars representing all genera of Eremobatidae and most species groups within the genera Eremobates Banks, 1900, Eremochelis Roewer, 1934, and Hemerotrecha Banks, 1903. Maximum Likelihood and two Bayesian analyses consistently recovered the monophyly of Eremobatidae, Eremorhax Roewer, 1934 and Eremothera Muma, 1951 along with a group comprising all subfamily Eremobatinae Kraepelin, 1901 exemplars except Horribates bantai Muma, 1989 and a group comprising all Eremocosta Roewer, 1934 exemplars except Eremocosta acuitalpanensis (Vasquez and Gavin, 2000). The subfamily Therobatinae Muma, 1951 and the genera Chanbria Muma, 1951, Hemerotrecha, Eremochelis, and Eremobates were polyphyletic or paraphyletic. Only the banksi group of Hemerotrecha was monophyletic; the other species groups recognized within Eremobates, Eremochelis, and Hemerotrecha were paraphyletic or polyphyletic. We found no support for the monophyly of the subfamily Therobatinae. A time-calibrated phylogeny dated the most recent common ancestor of extant eremobatids to the late Eocene to early Miocene, with a mean estimate in the late Oligocene (32.2 Ma).

Introduction

Solifugae, known by various common names including camel spiders, sun spiders, and wind scorpions, is a poorly studied order of mostly nocturnal, cursorial, hunting arachnids noted for their powerful two-segmented chelicerae, a voracious appetite, and tremendous speed (Punzo, 1998, Beccaloni, 2009). Solifuges are dominant predatory arthropods in arid ecosystems (Banta and Marer, 1972, Cloudsley-Thompson, 1977, Wharton, 1987, Punzo, 1994, Polis and McCormick, 1986) and also serve as important prey for many other desert taxa (El-Hennawy, 1990, Henschel, 1994, Arlettaz et al., 1995, Anderson et al., 1999, Catenazzi et al., 2009).

Solifugae is the sixth most diverse order of arachnids with 12 families, 141 genera, and approximately 1100 described species (Harvey, 2002, Harvey, 2003, Prendini, 2011). In addition, five monotypic fossil genera have been described (Poinar and Santiago-Blay, 1989, Selden and Dunlop, 1998, Dunlop and Rössler, 2003, Dunlop et al., 2004, Dunlop et al., 2008, Dunlop et al., 2015, Dunlop and Klann, 2009). The oldest known solifuge, Protosolpuga carbonaria Petrunkevitch, 1913, is from the Upper Carboniferous (Pennsylvanian) of Mazon Creek, Illinois, USA and dates back to 313–304 Ma (Selden and Shear, 1996).

Based on morphological and molecular analyses, the order Solifugae is indisputably monophyletic (Wheeler et al., 1993, Wheeler and Hayashi, 1998, Giribet and Ribera, 2000, Giribet et al., 2002) although its position within the Arachnida remains uncertain (Weygoldt and Paulus, 1979a, Weygoldt and Paulus, 1979b, Shultz, 1990, Weygoldt, 1998, Wheeler and Hayashi, 1998, Dunlop, 2000, Giribet et al., 2002, Alberti and Peretti, 2002, Dabert et al., 2010, Pepato et al., 2010, Klann and Alberti, 2010, Regier et al., 2010, Dunlop et al., 2012).

Koch, 1842, Simon, 1879, Kraepelin, 1899, Kraepelin, 1901 conducted cursory taxonomic assessments of solifuges. Roewer, 1932, Roewer, 1933, Roewer, 1934 presented the first comprehensive classification of the order, which was criticized by several subsequent researchers for relying too heavily on spinal and setal patterns; characters that vary widely within genera and sometimes within species (Fichter, 1940, Panouse, 1950, Turk, 1960, Della Cave, 1971, Delle Cave and Simonetta, 1971, Muma, 1951, Muma, 1989). Fichter (1940) dismissed most of Roewer’s characters and Muma, 1951, Muma, 1962, Muma, 1963, Muma, 1970 was unable to utilize Roewer’s spine-like setal patterns of legs to identify genera. Muma instead established genera based on the form of the male cheliceral fixed finger, type of modified setae in the male flagellar complex, and presence or absence and form of the male cheliceral fixed finger mesal groove. He recognized species groups within genera by minor differences in the above male characters, gross differences in the genital operculum of the female, and proportionate sizes of the cheliceral fondal teeth of both sexes. Della Cave (1971) criticized Roewer’s characters and pointed out that characters used by Roewer to establish new genera are often variable within species. He suggested the need to re-assess the taxonomy proposed by Roewer but did not offer standardized characters to use instead. Brookhart and Muma (1981) established the A/CP ratio (appendage lengths/combined cheliceral + propeltidial lengths) as an approximation of leg length in relation to body size. Despite such attempts to find homologous characters for family, genus, or species group classification, Harvey (2003) decried a taxonomy devoid of any phylogenetic interpretations. Recently, however, an attempt has been made to provide homology assessment for cheliceral characters throughout the order (Bird et al., 2015).

The solifuge family Eremobatidae Kraepelin, 1899 is endemic to North America and has been recorded in southern Canada (British Columbia, Alberta, Saskatchewan), western U.S.A. (west of the Mississippi River), and in arid habitats throughout Mexico. The family presently comprises 179 described species placed in two subfamilies, eight genera, and 18 species groups (Table 1). Roewer, 1933, Roewer, 1934 laid the foundation for the systematics of Eremobatidae, raising Kraepelin’s (1901) subfamily Eremobatinae to family status as the Eremobatidae and further dividing the family into two subfamilies, the Eremorhaxinae and the Eremobatinae, on the basis of fourth leg tarsal segmentation and the number of tarsal claws on leg I. The Eremorhaxinae was a monotypic subfamily, including only Eremorhax magnus (Hancock, 1888), and defined as having an unsegmented fourth tarsus and no claws on the first leg. Muma (1951) synonymized the Eremorhaxinae with Eremobatinae after re-examining Eremorhax magnus and added the subfamily Therobatinae. Included in the Eremobatinae were genera with one claw on the tarsus of leg I, chelicerae about twice as long as wide, a style-like fixed cheliceral finger on the males, and with a ventral or mesoventral groove on the male fixed cheliceral finger. The Therobatinae included genera with two claws on the tarsus of the first leg, chelicerae from two and one-half to three times as long as wide, and males with a style-like fixed cheliceral finger with or without a ventral or mesoventral groove and with or without modified teeth on the fixed finger. In an unpublished manuscript completed shortly before his death (in JOB’s possession, available upon request), Muma created an additional subfamily, the Hemerotrechinae that included genera with two tarsal claws on leg I, and males with an indistinct, very indistinct, or no mesal groove on the fixed cheliceral finger. He placed Chanbria spp. and most species of Hemerotrecha in this new subfamily. He also moved the genus Horribates from the Eremobatinae to the Therobatinae.

Muma (1951) revised the North American family Eremobatidae, separating genera based upon gross morphological differences including the shape of the male fixed finger, cheliceral dentition patterns, structure of the flagellar complex, and structure of the female genital operculum. The species groups were subsequently created based upon additional morphological characters (Brookhart, 1965, Brookhart, 1972, Muma, 1970, Muma, 1976, Muma, 1989, Brookhart and Muma, 1981, Brookhart and Muma, 1987, Muma and Brookhart, 1988, Brookhart and Cushing, 2002, Brookhart and Cushing, 2004, Brookhart and Cushing, 2005).

Except for a morphological phylogenetic analysis of the scaber group of Eremobates (Brookhart and Cushing, 2004), the monophyly of the subfamilies, genera, and species groups of Eremobatidae has never been tested phylogenetically. We present a molecular phylogenetic analysis of Eremobatidae, the first such analysis of a family of Solifugae, to test whether the family, subfamilies, genera, and species groups are monophyletic. We conducted Maximum Likelihood and Bayesian analyses on a multilocus dataset representing all genera of Eremobatidae and most species groups of Eremobates, Eremochelis, and Hemerotrecha (Table 1). In addition, we used the mulitilocus data to produce a time-calibrated phylogeny using a Bayesian approach, providing the first insight into camel spider diversification and establishing a framework for future biogeographical investigations within the family.