Systematics of testudacarine torrent mites (Acari, Hydrachnidia, Torrenticolidae) with descriptions of 13 new species from North America

Abstract Thirteen new species of North American Testudacarus (Torrenticolidae: Testudacarinae) are described: Testudacarus deceptivus O’Neill & Dowling, sp. n., Testudacarus hitchensi O’Neill & Dowling, sp. n., Testudacarus harrisi O’Neill & Dowling, sp. n., Testudacarus dennetti O’Neill & Dowling, sp. n., Testudacarus dawkinsi O’Neill & Dowling, sp. n., Testudacarus radwellae O’Neill & Dowling, sp. n., Testudacarus kirkwoodae O’Neill & Dowling, sp. n., Testudacarus hyporhynchus O’Neill & Dowling, sp. n., Testudacarus smithi O’Neill & Dowling, sp. n., Testudacarus rollerae O’Neill & Dowling, sp. n., Testudacarus elongatus O’Neill & Dowling, sp. n., Testudacarus rectangulatus O’Neill & Dowling, sp. n., and Testudacarus oblongatus O’Neill & Dowling, sp. n. Testudacarus vulgaris Habeeb, 1954 is resurrected from synonymy with Testudacarus minimus and redescribed. Debsacarus (Habeeb, 1961), Testudacarus americanus Marshall, 1943, and Testudacarus minimus Marshall, 1943 are redescribed. All redescriptions are from original types. Species delimination was accomplished through examination of morphology, biogeography, and molecular phylogenetics of the barcoding region of COI. Other species are addressed and a key to world species is presented. For Testudacarinae, this represents the first published: 1) descriptions from multiple specimens (i.e. intraspecific variation); 2) colored photographs; 3) explicit illustrations and discussion of sexual dimorphism within the subfamily; 4) genetic data. A comprehensive testudacarine reference list is also included.


Introduction
Torrenticolidae Piersig, 1902 are ubiquitous and diverse in North America, but the majority of species remain undescribed. This study is the second in a series of descriptions of North American torrenticolids. The goal of this ongoing taxonomic project is to explore the family and make these mites amenable to other researchers.
Testudacarinae Cook, 1974 are found abundantly in riffles of fast flowing streams throughout most of North America and sporadically in Asia. Typical of lotic-dwelling water mites, testudacarines are dorso-ventrally flattened, heavily sclerotized, and possess robust legs with large tarsal claws used for crawling. Most testudacarines are less than 1 mm in size and can exhibit striking coloration. Larvae are reported to be ectoparasites of chironomid adults (Smith 1982).
Despite their abundance, few testudacarines are described worldwide and in North America the most recent description is over fifty years old. Limited morphological and distributional data have been presented, and no genetic data has ever been published on Testudacarinae. Minimalistic and incomplete descriptions have led to considerable confusion throughout testudacarine taxonomic history. There is a need describe new species with modern methods and to redescribe older species with the same thoroughness.
Thirteen descriptions and four redescriptions of North American Testudacarus Walter, 1928 are included within. Following Fisher et al. (2015), species were delimited using a combination of morphology, biogeography, and molecular data (i.e. "barcoding" region of COI). In addition to descriptions and redescriptions, sexual dimorphism within the subfamily is explicitly addressed, a comprehensive testudacarine reference list is included, and a key to world species is presented.
Testudacarus americanus and T. minimus were described by Marshall (1943) from one "small" male and one "large" female from the same creek in California. Habeeb (1954) described T. vulgaris from New Brunswick. Later, Habeeb (1967) synonymized T. minimus with T. americanus after noticing sexual dimorphism within Testudacarus (specifically, females are larger than males). Habeeb (1969) then synonymized T. vulgaris with T. americanus and established Testudacarus americanus galloi, from "two female mites rather like [T. americanus], yet atypical." He stated that T. americanus vulgaris was a blue form found from New Brunswick to as far west as Arizona, and T. americanus americanus and T. americanus minimus were "red to golden" forms found from California. Habeeb (1974a) then resurrected T. minimus and changed T. americanus vulgaris to T. minimus vulgaris, after realizing he had misread Marshall (1943). Habeeb (1961) described T. oribatoides from a male and female from California. This species has a "protrusable maxillary tube…reminiscent of Pseudotorrenticola," and is in other respects atypical for Testudacarus (Habeeb 1961). Habeeb (1974b) erected Debsacarus and designated Testudacarus oribatoides as the type specimen "due to the fact that many recent authors have no respect for subgeneric names." Only two authors, Viets (1987) and Smith (1982), address the hypotheses proposed by Habeeb (1969Habeeb ( , 1974aHabeeb ( , 1974b. Viets (1987) did not take a stance on the validity of any species, instead he catalogued all the names presented in the literature and asked the reader to "vergl." (short for the German vergleichen, or "compare"). However, concerning Debsacarus, Viets (1987) did state: "Diagnose und abbildungen dürftig; Genus-und Artberechtigung unklar," ("Diagnosis and illustrations poor; genus and art authority unclear."). Smith (1982) acknowledged that Habeeb (1969) "proposed a second subspecies from California," but otherwise took no stance on its validity.

Sampling and curation
Mites were collected and preserved using protocols detailed in Fisher et al. (2015).

Morphological terminology
Terminology used in this study is detailed in Figs 1-5 and follows Goldschmidt (2007) as modified by Fisher et al. (2015). Hyphens are used for directional or numbered morphological features: for example, dorsoglandularia 1 will be expressed as dorsoglandularia-1. This is to prevent confusion when terms are followed by numbers and to make longer, more complicated terminology more accessible to unfamiliar readers. "Colorless" refers to a lack of pigmentation in the cuticle; as the cuticle itself is typically yellowish, "colorless" species are thus yellowish.

Material deposition of Nearctic types
All holotypes, allotypes, and some paratypes have been deposited in the Canadian National Collection of Insects, Arachnids, and Nematodes (CNC), Ottawa, Canada. Additional paratypes have been deposited in the Acari Collection of the University of Arkansas (ACUA), Fayetteville, Arkansas. Specific numbers of slides deposited at the CNC and ACUA are noted within each species description. Collection abbreviations are used throughout.

Morphological and distributional examinations
Material from the CNC and additional collections provided tens of thousands of testudacarines for morphological examination from across North America; a portion of these were examined closely for morphological variation. Previous torrenticolid studies suggested color and size were not necessarily important characters in distinguishing species (e.g., Fisher et al. 2015). Therefore, testudacarine "morphotypes" were chosen conservatively, giving more weight to drastic character differences, such as the presence of four instead of five pedipalp segments, over potentially more ambiguous characters, such as color and size variation. Many morphological characters were examined including general characteristics (e.g., color, size, body shape) and specific morphological features such as the following: shape of the dorsal plate, platelets, coxal field, and genital field; positioning of glandularia and lyrifissures; setae on the dorsum, venter, and gnathosoma; and structure of the gnathosoma and ejaculatory complex. Over 100 measurements per specimen were taken and compared and proportions between many of these measurements were analyzed. Finally, distributional Left -coxo-glandularia (cg); latero-glandularia (lg); ventro-glandularia (vg); Middle -coxae (c). Right -gnathosomal bay (gb); coxae-II+III midline (ml); genital field (gf); acetabula (a); line of secondary sclerotization (ss); excretory pore (ep). data was considered for each "morphotype" and probable ranges were hypothesized. Differences and similarities in ranges were considered as further supporting evidence of putative species. ; latero-glandularia (lg); ventro-glandularia (vg); coxae (c); coxae-II+III midline (ml); genital field (gf); acetabula (a); line of secondary sclerotization (ss); excretory pore (ep). Scale: 100 µm. Photo Michelle Hoppner and Ian Smith (used with permission).

Molecular examination
The "barcoding" region of COI was used as an independent test of morphological species hypotheses. COI was used to determine if any morphological characters, conservative or ambiguous, indicated species boundaries by sorting into distinct genetic lineages. COI was also used in the same way to test distributional hypotheses. Taxon sampling included roughly 300 specimens spanning "morphotypes" from across North America. Unfortunately, ethanol collections were limited from Mexico, northern Canada, and the eastern United States and therefore do not fully represent the ranges of species from these regions. Later, twenty specimens were included for phylogenetic analysis of 28S (D1-3) to investigate interspecific relationships. Genbank accession numbers of specimens for which sequences were obtained and used in this study are located in Table 1. Based upon recommendations by Chakrabarty et al. (2013), GenSeq nomenclature is used in the table to indicate the status of types and non-types sequenced.   Genomic DNA extraction was completed with Qiagen DNeasy Tissue Kits (Qiagen Inc.,Valencia, California). Amplifications of the target region of COI were performed with LCO1490 and HCO2198 (Folmer et al. 1994). Amplifications of the target region of 28S were performed with D23F and D6R (Park and Ó Foighill 2000). PCR was performed in a DNA Engine Peltier thermal cycler. COI samples were denatured for two minutes at 94 °C, followed by forty cycles of fifty seconds at 94 °C, thirty seconds at 48 °C, and one minute at 72 °C, with a final ten minute extension on the last cycle. 28S samples were denatured for two minutes and thirty seconds at 94 °C, followed by forty cycles of thirty seconds at 94 °C, twenty seconds at 53 °C, and one minute at 72 °C, with a final ten minute extension on the last cycle. Purification was done with Qiagen QIAquick PCR Purification Kits and test gels of 1.5% agarose were used to confirm PCR product quality. The purified product was then sequenced by Macrogen USA, based in Rockville, Maryland (http://macrogenusa.com/). DNASTAR© Lasergene SeqMan (Madison, Wisconsin) was used to reconcile forward and reverse sequences. The contigs that resulted were examined for contamination with GenBank BLAST searches. Clustal X (Thompson et al. 1997) was used to align sequences, and then BioEdit (Hall 1999) was used to conservatively edit the resulting sequences. COI sequences were around 650bp and 28S sequences were around 800bp. MrBayes (3.2.2) was used to perform Bayesian analyses over 5 million generations with Lebertia Neuman, 1880 as an outgroup. Monophyly was tested across Torrenticolidae as part of a forthcoming study. Molecular analysis was performed with the Extreme Science and Engineering Discovery Environment infrastructure available through the Cipres Portal (Miller et al. 2010).

Species delimitation results
Phylogenetic analysis of COI and 28S resulted in five well-supported (posterior probability greater than 95%) clades; however, analyses did not produce resolution at the base of Testudacarinae, resulting in a five-branched polytomy (Fig. 6). Each of the five lineages show at least 15% COI divergence from another. Within these five lineages are 16 distinct and well-supported species. With few exceptions, these species exhibited relatively high COI divergence (greater than 5%) between clades and relatively low divergence within a given clade (less than 1.5%). Genetic extractions were unsuccessful for a 17 th species, T. radwellae.
Three morphotypes (Testudacarus minimus, T. hitchensi, and T. elongatus) exhibited more intraspecific variation than expected, suggesting potential cryptic species. Further investigation of specimens identified morphological and biogeographic differences suggesting three Testudacarus minimus-like species, four T. hitchensi-like species, and three T. elongatus-like species. However, some of these "species" exhibit high intraspecies COI divergence with restricted geographic ranges and no diagnosable morphological variability, and should be the target of further research.
In summary, we find strong support through a combination of morphology, biogeography, and phylogenetic analysis of COI and 28S for 17 species sorted into four robustly supported species complexes. The following species complexes are proposed to better organize the subfamily: Testudacarus minimus complex, T. hitchensi complex, T. americanus complex, and T. elongatus complex. Each complex is treated below within the taxonomic descriptions. Subfamilial diagnosis. For larval diagnosis see Smith (1982). Adults differ from torrenticolines in having three pairs of acetabula (six in Torrenticolinae); condyles present over the insertions of leg-IV; long posterio-dorsal subcapitular apodemes (also long in Monatractides); a ridge extending anteriorly from the leg-IV socket; and a ring of platelets closely affiliated with the central dorsal plate, i.e., not hidden within a dorsal furrow as in torrenticolines (Fig. 1). They are further characterized by having a single anterio-medial dorsal platelet and pedipalps without ventral projections, although some torrenticolines also have these characters. Testudacarinae can be further diagnosed by the following combination of characters. Medial dorsal plate exhibiting secondary and occasionally tertiary sclerotization. Dorsal platelets variable in size, shape, and coloration. Anterio-medial platelet smaller than anterio-lateral platelets and trapeziform (rounded to rectangular). Anterio-lateral platelets long with anterior bulge and posterior tapering. Seven pairs of lateral platelets present. Lateral-platelet-2, -4, and -6 large and elongate and -1, -3, -5, and -7 smaller and rounded. Lateral-platelet-3 highly variable and positioned either anterior or lateral to lateral-platelet-4. Lateral-platelet-4 highly variable in shape mostly depending on lateral-platelet-3 position. Dorso-glandularia-2 and post-ocularial setae located together on anterio-lateral platelet. Dorsoglandularia-3, -5, and -6 located on lateral-platelet-1, -5, and -7 respectively. Dorsoglandularia-4 located on the large medial dorsal plate. Latero-glandularia-4 located on lateral-platelet-3. Ventro-glandularia-3 posterior to coxae-IV (on coxae-IV in other torrenticolids). Coxo-glandularia-4 located at tip of coxae-I (as in Monatractides and many Torrenticola). Pedipalp, femur, and genu with plumose setae ventrally. Also similar to Monatractides, posterio-dorsal subcapitular apodemes are long. Rostrum short.
Distribution. Testudacarines have been reported on many occasions outside of their original descriptions. Furthermore, the Canadian National Collection in Ottawa, Canada includes thousands of testudacarines collected from across most of North America . In Asia there have only been a handful of additional reports (Walter 1929, Pešić and Smit 2007, Jin et al. 2010, Morimoto 2012). This is not completely due to a lack of torrenticolid work in Asia, for an extensive list see Walter et al. (2009, pg. 256) andFisher et al. (2015). Extensive work has also been done on water mites in Europe, Africa, and Australia without any reports of testudacarines. Therefore, testudacarines are currently thought to be widely distributed throughout most of North America (with southern limits in Mexico and northern limits around the 60 th parallel), and sparsely distributed in parts of Asia.
Remarks. The three pairs of acetabula, coxae-IV condyles, and "generalized" pedipalps are plesiomorhphic states that clearly show testudacarines as retaining ancestral torrenticolid characteristics (Wiles 1997a). Wiles (1997a) and other authors suggest latero-glandularia-3 is present on the dorsum of testudacarines. However, we suggest that this is latero-glandularia-4 due to its posterior-most positioning. We also detail sexually dimorphic characters (Fig. 7). Although Habeeb (1954) first noted differences between the sexes of T. vulgaris, he did not present these distinctions in their wider context as overall conditions of Testudacarinae. Sexual dimorphism present in Testudacarinae include: 1) female dorso-glandularia-4 positioned closer to the muscle Figure 7. Testudacarine sexual dimorphism: female dorsal shield (A) and ventral shield (C) differing from male (B, D) by the following characters: 1) dorso-glandularia-4 positioned far closer to muscle scares; 2) area of secondary sclerotization always present (males rarely present; very small if present); 3) with shorter coxae-II+III midline; 4) genital field enveloped by coxal field; 5) larger and rounder body (males around 80% of female size); 6) excretory pore well separated from ventral line of secondary sclerotization. scars; 2) dorsal secondary sclerotization always present in females and usually absent in males (very small if present in males); 3) female coxae-II+III midline short; 4) genital field almost entirely enveloped by coxal field in females but only around half of male genital field within coxal field; 5) females idiosoma larger and rounder (males around 80% of female size) with less of the ventral shield composed of coxal field; and 6) excretory pore well separated from ventral line of secondary sclerotization in females, and is either in direct contact with or nearly so in males. Habeeb, 1974 http://zoobank.org/9C344329-32F6- 4C4B-8167-196E030B2ED8 Habeeb 1974b: 1;Viets 1987: 222, 724;Zhang and Guo 2010: 117. Type species. Debsacarus oribatoides (Habeeb, 1961).

Debsacarus
Generic diagnosis. Debsacarus differ from all other Testudacarinae in having foursegmented pedipalps (instead of five) and projections on the anterio-tips of coxae-I. With the exception of Testudacarus hyporhynchus, Debsacarus differ from all other Testudacarinae in having an elongate gnathosoma and an extremely narrow gnathosomal bay that is covered dorsally and ends anterior to the leg-I insertion ventrally.

Remarks.
Debsacarus oribatoides show at least 15% COI divergence from all other Testudacarinae and less than .6% divergence from one another (Fig. 8). Additionally, Habeeb (1961) describes a protrusible maxillary tube, however, we find no evidence in the additional specimens examined that the maxillary tube or subcapitulum is any more protrusible than what is commonly found in other Testudacarus, and certainly is not protrusible like in Pseudotorrenticola. Habeeb (1961) did not designate types, however, he described the species from the only two specimens available. From those two specimens, we have designated a lectotype (♀) and paralectotype (♂).

Testudacarus minimus complex
Species complex diagnosis.These species can be distinguished from most other testudacarines by their small size (female and male dorsal length less than 700 and 600 µm, respectively), highly variable coloration (red, orange, blue, violet, and rarely colorless), and small (<140 µm), rounded anterio-medial platelet (differing from Testudacarus rollerae, which has a large (>140 µm) anterio-medial platelet more than or nearly twice as wide as long). Additionally, only this complex and the T. hitchensi complex are pre- Figure 12. Testudacarus minimus complex molecular phylogeny: 28S and COI Bayesian analysis showing strong support for a soft polytomy with three distinct clades (•: >95% posterior probability); colored clades exhibit <2.5% divergence in COI within and >6.5% divergence between; continuation of (A) lineage from Fig. 6. sent east of the Great Plains. These two complexes resemble each other morphologically in many respects, but can be easily distinguished because members of this complex have uniform coloration across all three anterior platelets while T. hitchensi-like mites have a colorless anterio-medial platelet and colored anterio-lateral platelets. With the exception of T. radwellae, males of this complex differ from T. hitchensi-like mites in having dorso-glandularia-4 positioned less anterior to and more lateral to the muscle scars. This complex is abundant and present across most of North America and comprises the following species: T. deceptivus, T. minimus, T. radwellae, and T. vulgaris. Remarks. Molecular data show strong support for three distinct clades (Fig. 12). All three clades exhibit less than 2.5% COI divergence within the clade and greater than 6.5% divergence between clades. In California there is currently no reliable way to diagnose these three clades morphologically as they are all roughly the same size and color (colorless to orange). However, outside of California it is possible to diagnose clades based on color, size, and geographic distribution. Members of this complex exhibit the broadest geographic ranges and thus exhibit the highest and not unexpected intraspecies divergence of the four complexes. Given the broad geographic sampling conducted in this complex, we feel comfortable designating the three main clades, exhibiting intra-clade divergence of more than 6.5%, as multiple species: T. minimus, T. vulgaris, and T. deceptivus. A fourth species, T. radwellae, belongs to this complex based on morphology, but genetic extractions were unsuccessful. Testudacarus radwellae males also share morphological similarities with T. hitchensi-like mites (the positioning of dorso-glandularia-4). Therefore T. radwellae is potentially important in discovering the relationship between these two species complexes and deserves further investigation. Marshall, 1943 http://zoobank.org/CD1D1B50-6A37-4099-86D7-A6DAB8A17CA6

Testudacarus minimus
Testudacarus minimus : Marshall 1943: 322;Bergstrom 1953: 159;Mitchell 1954: 40;Imamura 1955: 182, 188;Viets 1956: 255;Habeeb 1959a: 21;Crowell 1961: 329;Mitchell 1962: 42;Lundblad 1967: 418;Conroy 1968: 29;Habeeb 1974a: 1;Conroy and Scudder 1975: 307;Imamura 1976  Type deposition. Holotype (1♂) deposited at the CNC. Diagnosis. Testudacarus minimus most resemble T. vulgaris and T. deceptivus. Throughout the majority of their shared range in the west, T. minimus are orange to red and T. vulgaris are violet to blue. While these two species have overlapping size ranges, T. minimus are generally larger. Testudacarus vulgaris females rarely exhibit a dorsal length over 600 µm and males rarely exceed 500 µm while T. minimus females and males are usually larger than 600 and 500 µm, respectively. Testudacarus deceptivus have only been found in two counties in California and cannot be distinguished from either T. minimus or T. vulgaris using morphology. Testudacarus minimus are the only members of their complex that have been found in Washington and northern Oregon.
Redescription. Female (n=14) with characteristics of the genus with following specifications.
Gnathosoma Distribution. Abundant throughout North America, ranging from the Pacific Northwest to the southwestern United States (and potentially into northern Mexico), and east into the western Great Plains.
Remarks. Commonly colorless or orange in the southwestern United States; red, pink, or orange-red in the northwest, Rocky Mountains, and western Great Plains; and uncommonly red-violent in the northwest, Rocky Mountains, and western Great Plains. Habeeb, 1954 http://zoobank.org/AD09023D-849F-4F13-BD0C-1CF7B6623748

Remarks. Commonly orange and uncommonly violet in the southwestern
Description. Female (n=2) with characteristics of the genus with following specifications.
Gnathosoma  Venter (Fig. 21)   Male (n=7) similar to female except for sexually dimorphic characters previously discussed and with following specifications. Etymology. Specific epithet radwellae after the late Dr Andrea J. Radwell, the American water mite researcher who collected the specimens needed for this description. Dr Radwell collaborated with us on the larger torrenticolid project as a whole, giving us invaluable advice and mentorship. Without her, large portions of this project would not have been possible. She is dearly missed.
Distribution. Reported from only two counties (Polk and Montgomery) in Arkansas.

Testudacarus hitchensi complex
Species complex diagnosis. Only this complex and the T. minimus complex are present east of the Great Plains. These two complexes resemble each other morphologically in many respects, but can be easily distinguished because members of this complex have non-uniform coloration across all three anterior platelets (colorless anterio-medial platelet and colored anterio-lateral platelets) while T. minimus-like mites possess uniform coloration across all three platelets. Males of this complex differ from T. minimus-like mites in having dorso-glandularia-4 positioned more anterior to and less lateral to the muscle scars. These mites are common in eastern United States and rare in eastern Canada and Florida, small (female and male dorsal length less than 700 and 600 µm, respectively), and violet to blue in color. This complex comprises the following species: T. harrisi, T. dennetti, T. dawkinsi, and T. hitchensi. Remarks. Distinguishable morphological characters can be found for four lineages while genetic data indicates more diversity (Fig. 23), suggesting cryptic speciation within the clade. Three clades (violet and blue clades in Fig. 23) exhibit less than 1.5% COI divergence within the clade and greater than 6% divergence between clades. This relatively low divergence within clades over their large ranges compared to the high divergence exhibited between clades even in the same streams strongly supports multiple species. The fourth clade (green in Fig. 23) proves problematic as no morphological variability has been found within the clade, but COI divergence of up to 4.5% is present and within a small geographic area (North Carolina and Tennessee). Ethanol Figure 23. Testudacarus hitchensi complex molecular phylogeny: 28S and COI Bayesian analysis showing strong support for at least four distinct clades, but suggesting more (•: >95% posterior probability); excepting green clade, clades exhibit <1.5% divergence in COI within and >6% between; green clade exhibits <4.5% within and >9.5% between other clades; specimens in red constitute additional suspected species based on genetic data, but lack morphological or distributional variation from green clade; continuation of (B) lineage from Fig. 6. collections were limited from this region and more data is needed. Furthermore, examinations of GAW collections provided by the CNC suggest there are other potential "morphotypes" of this species complex unrepresented in the genetic data presented. More species almost certainly exist in this complex, and further research is needed. Type deposition. Holotype (1♀), allotype (1♂), and eight paratypes (4♀, 4♂) deposited at CNC; ten paratypes (5♀, 5♂) at ACUA.
Diagnosis. These mites differ from all others in the complex in having large medial pores on the dorsal plate surrounded by a ring of smaller pores (all pores uniform in other species). Males also have a "bleached" or colorless area posterior to the coxal plate that is colored in other members of the complex.
Description. Female (n=10) with characteristics of genus with following specifications.
Gnathosoma Etymology. Specific epithet hitchensi after the late Christopher Eric Hitchens, the English author, journalist, and literary critic. As Sam Harris' wife, Annaka, said: "Nothing Hitchens does is ever boring." Hitchens has inspired thousands of free-thinkers to remain clever and engaged in our attempts to understand the world around us.
Distribution. Eastern United States east of the Mississippi River with southern limits in Florida.
Remarks. As it is likely that this species represents a cryptic species complex, measurements were only included from specimens exhibiting less than 2% COI divergence within the clade (those highlighted in red in Fig. 23 were excluded). This was done so measurements would remain useful if more species were diagnosed in the future. Type deposition. Holotype (1♀), allotype (1♂) and ten paratypes (5♀, 5♂) deposited at Canadian National Collection; eleven paratypes (7♀, 4♂) at ACUA.
Diagnosis. These mites have violet to blue coloration over the majority of their anterio-lateral platelets while the rest of the complex have coloration restricted to the posterior half of the platelet.
Description. Female (n=13) with characteristics of the genus with following specifications.
Distribution. Eastern United States east of the Mississippi River, with southern limits in Florida. Male (n=9) similar to female except for sexually dimorphic characters previously discussed and with following specifications. Distribution. Eastern United States east of the Mississippi River, with southern limits in Florida.

Testudacarus dawkinsi
Etymology. Specific epithet dawkinsi after Clinton Richard Dawkins, the English evolutionary biologist and writer. Dawkins has proven repeatedly that one can change the world as a biologist by day and keep going as a free-thinker by night.

Testudacarus americanus complex
Complex diagnosis. These mites lack the four-segmented pedipalp of the Debsacarus oribatoides-like mites, the elongate body of the Testudacarus elongatus-like mites, and with the exception of T. rollerae, are much larger (female and male dorsal length usually more than 700 and 600 µm, respectively) than mites of the T. minimus and T. hitchensi complexes. Testudacarus rollerae have a larger (>140 µm) anterio-medial platelet that is more than or nearly twice as wide as long, while T. minimus-like mites have a smaller (<140 µm), more rounded anterio-medial platelet. These mites are present in western North America within and west of the Rocky Mountains, have very light to no coloration, have a large rectangular anterio-medial platelet, and comprise the following species: T. kirkwoodae, T. americanus, T. hyporhynchus, T. smithi, and T. rollerae. Remarks. Molecular data show strong support for five distinct clades (Fig. 32). Four clades exhibit less than 1.3% COI intraclade divergence, and all five clades exhibit greater than 9% divergence between clades. The fifth clade (pink in Fig. 32) exhibits 4.5% divergence within. However, only two specimens of this clade are available. One is teneral and badly damaged and therefore provides no characters for morphological diagnoses. More specimens should be collected and analyzed. Otherwise, all five clades have diagnostic morphological features that further warrant species designations. showing strong support for five distinct clades (•: >95% posterior probability); excluding pink clade, colored clades exhibit <1.3% divergence in COI within and >9% divergence between; pink exhibits 4.5% variation within; red specimen is a suspected species based on genetic data, but specimen is teneral and too badly damaged to diagnose; continuation of (C) lineage from Fig. 6. Marshall, 1943 http://zoobank.org/FE0E6228-D8AA-4063-A139-1C4A963454EB
Remarks. Having examined the type material, we suggest that T. americanus galloi Habeeb, 1969 is simply a teneral T. americanus that Habeeb confused for an "atypical" T. americanus. Type deposition. Holotype (1♀) deposited at CNC. Diagnosis. These are the largest known testudacarines and also differ further from their complex in having a smaller, more rounded anterio-medial platelet.

Testudacarus kirkwoodae
Description. Female (n=1) with characteristics of genus with following specifications.
Gnathosoma Male (n=0) unknown. Etymology. Specific epithet kirkwoodae after my (JCO) mother's maiden name. It was her bringing me to the leafcutter ant exhibit at our local science center that helped me become interested in biology as a child, and her endless support and advice that helped me finish my education.
Remarks. As it is likely that this species represents a cryptic species complex (with 4.5% COI divergence between the two available specimens), measurements were in-cluded from only one specimen (the other highlighted in red in Fig. 32 was excluded). This was done so measurements would remain useful if more species were diagnosed in the future. Measurments were also not included from the other specimen because it is teneral and badly damaged and would therefore prove poor for any description. While more than a single specimen is certainly desired for new descriptions, the included specimen has unique morphological characters such as its large size (it is larger than any other testudacarine, including species from Asia), and has strong support as a unique clade using COI. Therefore, this single specimen is unique enough that we are comfortable describing it.
Distribution. Only one specimen know from Douglas County, Oregon. Diagnosis. These mites differ from the rest of the complex in having a dorsally "covered" gnathosomal bay (short doral gnathosomal bay length) and an elongate gnathosoma with a long rostrum that exteneds below the gnathosoma ventral surface.
Description. Female (n=3) with characteristics of genus with following specifications.  Distribution. Reported from only two counties (Mendocino and Nevada) in California.

Testudacarus elongatus complex
Complex diagnosis. Unlike all other Testudacarinae, members of this complex have an elongate idiosoma. In contrast to most T. minimus-and T. hitchensi -like mites, these mites are colorless and much larger (female and male dorsal length greater than 700 and 600 µm, respectively). These mites are found in western North America west of the Rocky Mountains. This complex comprises three species: T. elongatus, T. oblongatus, and T. rectangulatus. Remarks. Combined molecular, distributional, and morphological data support three distinct clades within the T. elongatus complex (Fig. 43). All three clades exhibit less than 2.4% COI intra-clade divergence and greater than 3.3% divergence between clades. Intra-clade divergence of 2.4%, as seen with T. oblongatus, is not unexpected for a species exhibiting a large geographic range (British Columbia to California); how- Figure 43. Testudacarus elongatus complex molecular phylogeny: 28S and COI Bayesian analysis showing strong support at least three distinct clades (•: >95% posterior probability); colored clades exhibit <2.4% divergence in COI within and >3.3% divergence between; divergence of the two basal clades >9%; continuation of (D) lineage from Fig. 43. ever, a percent difference as high as 3.3% between two close localities (Mason and Snohomish County, between T. rectangulatus and T. elongatus) suggests separate species. Interestingly, COI divergence of more than 9% between the two sister clades (T. oblongatus and T. elongatus/T. rectangulatus) does not seem to produce high amounts of morphological diversity within this complex. Therefore, the morphological varation and geograhic varation found between T. elongatus and T. rectangulatus provide enough evidence for us to hypothesize two species, even if one is from a single specimen. Potentially, there is a coastal species (T. oblongatus), a species within and east of the cascade (T. elongatus), and a species from the Olympic Mountains (T. rectangulatus).  Since morphological variation is limited, a combination of morphology and distribution is best used to diagnosed members of the complex. These mites occur in the Olympic Mountains, while T. elongatus occur in Washington within and east of the Cascade Mountains, and T. oblongatus occur along the western Coast of Washington, Oregon, California, and British Columbia. Additionally, both T. elongatus and these mites differ from T. oblongatus in having more robust lateral platelets; most notably, lateral-platelet-4 tends to be larger in these two species than T. oblongatus, and is in direct or near direct contact with lateral-platelet-2. Reversely, T. oblongatus generally have less robust platelets and a smaller lateral-platelet-4 that has a noticeable gap between it and lateral-platelet-2. Limited specimens were found of T. elongatus and T. rectangulatus, but T. rectangulatus appear to have leg and pedipalp measurements roughly 10% larger than T. elongatus even between individuals of similar idiosoma size. More data is needed to better diagnose these species.
Male ( Holotype loans are not available from Ibaraki Nature Museum. The museum provided a low-magnification photograph through e-mail, though permission to print the photograph was not obtained. Diagnosis. These mites differ from all other Testudacarinae by distribution (Japan), and from T. tripeltatus by small size (dorsal length <700 µm). Testudacarus japonicus may be conspecific with T. okadai. More research and updated descriptions are needed for a better diagnosis.
Remarks. It is reasonable to assume Imamura (1955) had no knowledge of Habeeb (1954) because he never mentions T. vulgaris and there are inaccuracies in his description that could have been prevented if he had. Firstly, his "female" type specimen is almost certainly a male as "the genital area [is] relatively more to the posterior than in [females] and the two [dorsal muscle scars]… are located posterior to the [glandularia]" (Habeeb 1954). Furthermore, in his remarks he states the "Japanese specimen resembles most the Indian species," which with more current information is unlikely. At the time, T. japanicus would have been most similar in size, color, and shape to either T. vulgaris or T. minimus, not T. tripeltatus. Most importantly, the T. japonicus type is almost certainly male and therefore shares little morphology with the female T. tripeltatus. Therefore, the distinctions Imamura (1955) offers that T. japonicus are "different from [T. tripeltatus] in the anterior tips of the first [coxae], [pedi]palps, situations of [coxae] and genital organ" are unhelpful (Imamura 1955). He is referring to sexual dimorphism and comparing only the two most disparate species available to him. Imamura, 1976 http://zoobank.org/B66EFC5E-2C6A-448A-91E8-C5B1D7922654
Distribution. Mt. Fanjing (Guo and Jin 2005) and Fujian, China (Jin et al. 2010). Remarks. Testudacarus binodipalpis was described from one female and one "male." The described "male" is almost certainly a female as it exhibits all female sexual characters and no ejaculatory complex is noted in the description. However, these two females differ in some noteworthy respects. From illustrations it appears that the smaller female seems to have undergone tertiary sclerotization, while the larger female seems to have only undergone primary and secondary sclerotization. The size and positioning of lateral platelets are also quite different in each specimen. For these reasons the specimens should be reexamined as they might represent two species diagnosable by size. Guo and Jin (2005) state that T. binodipalpis can be separated from other Testudacarus by "the possession of 2 tubercles on the ventral surface of the" pedipalp tibia and the genu and femur "both with a feathered seta on the ventral surface." These pedipalp characters do not work as they are plesiomorphic for all Testudacarus (Fig. 5). Guo and Jin (2005) also state that the "dorsal and ventral apodeme both [have] a round terminal tip; [coxae-IV] with a triangular base." These additional characters are unhelpful in separating any testudacarines.