Invertebrate Biology x(x): 1–14. © 2016 The Authors. Invertebrate Biology published by Wiley Periodicals, Inc. on behalf of American Microscopical Society. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. DOI: 10.1111/ivb.12145

Phylogeographic analysis of invasive Asian () in the northeast United States

Nancy Schult, Kelly Pittenger, Sam Davalos, and Damhnait McHugha

Department of Biology, Colgate University, Hamilton, New York 13346, USA

Abstract. Phylogeographic studies are useful in reconstructing the history of species inva- sions, and in some instances can elucidate cryptic diversity of invading taxa. This can help in predicting or managing the spread of invasive species. Among terrestrial invasive species in North America, earthworms can have profound ecological effects. We are familiar with the centuries-old invasions of European earthworms (Lumbricidae) and their impacts on nutrient cycling in soils. More recent invasions by Asian earthworms of the family are less fully understood. We used data for two mitochondrial gene frag- ments, cytochrome oxidase I (COI) and 16S rRNA, to examine the relationships among populations of Asian earthworms in the megascolecid genus Amynthas in the northeast Uni- ted States. Recent reports have indicated that one species in particular, , is having detrimental effects in mixed forest ecosystems, and we were interested in under- standing the invasion history for this species. We were surprised to discover three divergent mitochondrial lineages of Amynthas occurring sympatrically in upstate New York. Given the gap between intra- and inter-lineage sequence divergences, we propose that these three lineages represent cryptic species of Amynthas, one of which is A. agrestis. For each of the three lineages of Amynthas, we observed shared haplotypes across broad geographic dis- tances. This may reflect common origins for populations in each lineage, either by direct routes from native ranges or through post-introduction spread by natural dispersal or human-mediated transport within North America. Management efforts focused on horticul- tural imports from Asia, commercial nurseries within the USA, and on prohibition of bait disposal may help to reduce the further invasion success of Amynthas. Additional key words: 16S rRNA, cytochrome oxidase I, haplotype frequency, barcode, Megascolecidae

With enhanced globalization, the spread of inva- the ecosystem in ways that are detrimental to native sive species has increased dramatically over the past species. With the ability to disrupt natural ecosys- several decades as non-native species are being tems, invasive species are regarded as a leading transported across the world either unintentionally threat to biodiversity, the environment, and the or deliberately. International trade provides human- economy (Tsutsui et al. 2000; Pejchar & Mooney mediated pathways for the transmission of organ- 2009; Vila et al. 2011; Qiu 2015). If we are to isms across bodies of water and great distances over address the effects of invasive species, we need to land, with the consequence that many species understand their establishment, dynamics, and rela- become established beyond their natural ranges tionship with invaded ecosystems. However, we are (Hulme 2009). Once introduced, invasive species can often poorly informed about the history of a species act as agents of change that threaten biological invasion, and know even less about the ecology, life diversity by outcompeting native species or altering history, or physiology of the species itself. In some instances we do not even understand the systematics aAuthor for correspondence. of the invading taxa (e.g., Yassin et al. 2008; E-mail: [email protected] Folino-Rorem et al. 2009; Bastos et al. 2011). This 2 Schult, Pittenger, Davalos, & McHugh makes it very difficult to predict or manage the Metaphire, and the fact that many recognized spe- spread of invasive species. cies in the Amynthas-Metaphire complex are Among invasive species, those that engineer parthenogenetic lineages. Furthermore, Blakemore ecosystems to alter the physical environment and (2013) suggests that a proliferation of taxonomic limit the quantity and quality of resources available descriptions of new species without due attention to to native species have the biggest impacts. Invasive ICZN rules has created a situation where molecular earthworms are among such species in North Amer- analyses are needed to help bring clarity to the ica, where most species are non-native. In validity of Japanese and Korean species. For this North America, recession of the Wisconsin glacial reason, many ecological studies of Amynthas do not shield ~10–12,000 bp exposed defaunated landscapes identify which species is under investigation, (Tiunov et al. 2006; Archer 2012), and subsequent although it is assumed that populations being stud- evolution of northern forests proceeded with few ied include just a single species. native earthworm species. The dominant earthworm Field studies of invasive Amynthas spp. have indi- species currently inhabiting northern forests and agri- cated that they alter the forest floor habitat by reduc- cultural lands were introduced by settlers from Eur- ing organic matter within the soil horizon and ope over the past few hundred years (Hendrix et al. increasing soil aggregation and nitrification, thus lim- 2006). The ecological impacts of these invasions iting available food resources for other fauna (Stein- include the physical alteration of soil layers, the inter- berg et al. 1997; Zhang et al. 2010). The worms can ruption of nutrient cycling, and competition with cause destruction of soil surface layers (O-horizon native invertebrates and plants (Bohlen et al. 2004a, and A-horizon) until only the castings of the worm b). Over the past few decades, with the increase in remain, and by creating significantly larger soil aggre- trade across the Pacific, the number of Asian invasive gates that are stable even over months, Amynthas earthworm species has been increasing in North worms can have a lasting impact on soil dynamics America and among these, species of the genus that may reduce biodiversity by minimizing soil Amynthas Kinberg 1867 (Megascolecidae) have been heterogeneity (Gorres€ & Melnichuk 2012; Greiner of particular concern (Hendrix & Bohlen 2002). et al. 2012). The invasion of Amynthas into forested These epi-endogeic worms live beneath leaf litter or habitats has also been reported to negatively impact mulch or in topsoil layers, where they alter the soil native millipede species richness and density, likely structure and remove organic materials from the soil due to the rapid removal of the top-most organic soil surface (Archer 2012; Gorres€ & Melnichuk 2012; horizons (Snyder 2008; Snyder et al. 2009, 2013), and Greiner et al. 2012; Ikeda et al. 2015). The relatively the worms detrimentally affect the behavior of native large size of some Amynthas worms (up to 25 cm), salamanders (Ziemba et al. 2015). their violent defensive behavior, and their serpentine To reconstruct the invasion of Amynthas in North locomotion have earned them the common names America, we can draw on published survey records “Crazy Worm,” “Jumping Worm,” and “Snake and online reports by environmental agencies to Worm.” The reproductive mode in species of Amyn- provide estimates of earliest introduction to particu- thas, as inferred from the degree of functionality or lar areas. The presence of Amynthas was recorded in degradation of testes, spermathecae, etc., can be sex- Mississippi as early as ~1900 (Gates 1958), and sub- ual or parthenogenetic; in either case, the worms have sequent reports of worms in this genus in northeast- an annual life cycle in temperate regions, with fast- ern US forests suggest that it has been spreading growing juveniles hatching from over-wintering egg into this region over the past several decades. Most cases in late spring (Reynolds 2010; Gorres€ et al. of these reports have been attributed to Amynthas 2014). The worms are found in natural forests and agrestis (Goto & Hatai 1899), which according to are often locally patchy, reaching densities up to 50 Gates (1982) was first recorded in Maryland, USA individuals m2 (Gorres€ & Melnichuk 2012). in 1939, with subsequent records in arboreta, nurs- Members of the genus Amynthas are easily recog- eries, and green houses in Massachusetts in 1953, nizable by a light-colored, smooth clitellum that New York in 1954, Louisiana and Connecticut in fully encircles the body. However, distinctive mor- 1958, Missouri in 1962, and Maine in 1966. Later phological characteristics are not apparent at the reports indicate A. agrestis invading natural habitats species level. In his review of Japanese earthworms, ranging from Alabama, Connecticut, Georgia, Blakemore (2003) reports that the of Louisiana, Maryland, Massachusetts, Missouri, New Amynthas is complicated by the lack of diagnostic Jersey, New York, North Carolina, Oklahoma, characters, the interrelationships among species South Carolina, and Tennessee (Reynolds 1978; attributed to Amynthas and another putative genus, Gates 1982; Callaham et al. 2003; Reynolds &

Invertebrate Biology vol. x, no. x, xxx 2016 Phylogeographic analysis of invasive earthworms 3

Wetzel 2004; Wetzel 2005). The most recent records also sampled worms from a mixed forest site near of the species in natural habitats have been in Ver- Auburn, Alabama and from the arboretum at the mont (2012), Wisconsin (2014), and just last year in University of Madison-Wisconsin (Table 1). Individ- New Hampshire, Maine, Illinois, and Ontario, ual worms were collected from beneath leaf litter Canada (Gorres€ & Melnichuk 2012; Reynolds et al. and under logs. Voucher specimens were preserved 2015; B. Herrick [unpubl. data], chicagobotanic.org, in 70% isopropanol after a tail snip (~2mm3) was sleloinvasives.org). New reports of A. agrestis in preserved in RNALater (Ambion) and stored at 4°C mixed deciduous forests of upstate New York over until DNA extraction. External morphological char- the past several years and concerns about the spread acters were used to confirm identification to the of the species into the Adirondack Mountains and level of genus at the time of collection, and a sub- across the Great Lakes region (Bernard et al. 2009) sample of worms was later dissected and identified prompted our study of populations sampled in and by S. James (University of Iowa). around Hamilton, New York. We discovered that the assemblages of earthworms we sampled included DNA extraction, amplification, and sequencing three sympatric, cryptic lineages of Amynthas, which has implications for understanding this invasion, its Total genomic DNA was extracted from tail snips further spread, and its mitigation. For example, the of 130 individuals of Amynthas using the DNeasy existence of separate lineages of Amynthas may be Blood & Tissue Kit and manufacturer’s protocol problematic for ecological studies, which could be (Qiagen). A ~650-bp fragment of the mitochondrial hampered by misidentifications, and also for puta- cytochrome c oxidase I gene was amplified using tive autecological studies that may include multiple primers modified from those of Folmer et al. (1994) cryptic species. Thus, we redirected our study to by aligning available sequences for megascolecid document the genetic diversity of Amynthas in earthworm species (MegaCOI-F [50-TAYTCWAC upstate New York and to include worms from a WAAYCAYAAAGAYATTGG-30] and MegaCOI- southern area in which A. agrestis has been report- R[50-TAKACTTCTGGRTGMCCAAARAATCA-30]). edly established for a few decades (Alabama) (Rey- M13 tails were added to the 50 end of each COI pri- nolds & Wetzel 2004; Reynolds 2011), and also mer to facilitate direct sequencing of PCR products from one of the more recently reported invasions of (Regier & Shi 2005). Each PCR reaction mix con- Amynthas, in Madison, Wisconsin (B. Herrick, tained the following: 4 mM MgCl2, 0.2 mM dNTPs, unpubl. data). 1x PCR buffer, 0.5 U Taq DNA polymerase (Invit- Here, we report our results based on analyses of rogen), 0.125 lg/lL BSA, 1 lM MEGA-COI pri- the mitochondrial genes cytochrome oxidase I (COI) mers (F and R), 2 lL of genomic DNA, and H2O and 16S rRNA. The COI barcode sequence has to reach a final reaction volume of 10 lL. For been used extensively in phylogeographic analyses amplification, an annealing temperature of 40°C was of earthworms, including invasive species (e.g., used for 40 PCR cycles. For a subset of 27 speci- Cameron et al. 2008; Novo et al. 2009; Porco et al. mens of Amynthas, a 360-bp fragment of the 16S 2013) and has also been applied in the detection of rRNA gene was amplified using the Ann16S primers cryptic diversity of earthworm species (Novo et al. of Sjolin€ et al. (2005). For these amplifications, we 2009, 2010; James et al. 2010; Deca€ens et al. 2013). used a similar reaction mix (without BSA) and a We used COI and 16S rRNA sequences in phyloge- touchdown PCR with annealing temperatures from netic analysis of lineages of Amynthas from upstate 60°Cto40°C decreasing 0.5°C per cycle followed New York, and COI sequences to provide prelimi- by 30 cycles at 40°C. nary assessment of the levels of genetic diversity All PCR products were visualized on a 1% agar- within and among populations of each lineage, with ose gel using ethidium bromide fluorescence under the goal of understanding the invasion history of UV light and cleaned using Exonuclease I and rSAP Amynthas in the northeast United States. (NEB) before sequencing. We used either M13Rev/ M13(-21) primers or the Ann16S primers listed Methods above and the BigDye Terminator v3.1 Cycle Sequencing Kit to sequence each PCR product in both directions. Sequencing reactions were cleaned Sample collection and preservation using Ethanol/EDTA precipitation and re-suspended Earthworms were sampled at four mixed decidu- in 12 lL HiDi formamide (Life Technologies) ous wooded sites in upstate New York separated by before running on an ABI 3130A Genetic Analyzer 2–3 km, and a fifth site in southern New York; we (Thermo Fisher Scientific). Overlapping sequence

Invertebrate Biology vol. x, no. x, xxx 2016 4 Schult, Pittenger, Davalos, & McHugh

Table 1. List of specimens of Amynthas in the study including label; the location, GPS coordinates, and date of collec- tion; and the number of specimens from each location sequenced for cytochrome oxidase I (COI, by lineage A, B, and C, as in Fig. 1) and for 16S rRNA (16S).

Label Collection location GPS coordinates Collection date No. specimens sequenced COI No. specimens 16S for each COI lineage NW ABC BCA7 Hamilton, NY #1 42°48.8630 75°32.0920 9/2014 29 7 4 18 4 BCA22 Hamilton, NY #2 42°47.8170 75°30.3330 10/2015 15 7 8 3 BCA8, 24 Clinton, NY 43°02.9540 75°24.0110 9/2014, 10/2015 34 25 9 7 MCHS Cazenovia, NY 42°56.1900 75°50.9930 9/2015 9 3 4 2 4 BCA21 Millbrook, NY 41°52.5670 73°40.5500 11/2015 9 9 2 BCA10 Auburn, AL 32°32.1480 85°26.2720 4/2015 10 10 4 BCA2 UW Arboretum, WI 43°04.1500 89°42.7750 9/2014 24 24 3 Total 130 66 35 29 27 fragments were assembled into consensus sequences topologies was based on 1000 pseudoreplicates. using Sequencher v.5.2.4 (Gene Codes), aligned Results were visualized in FigTree v.1.4.2 (http:// using Mesquite v.3.03 (Maddison & Maddison tree.bio.ed.ac.uk/software/figtree/). 2015), and formatted using SequenceMatrix (Vaidya We identified unique haplotypes using the FASTA et al. 2011). Each sequence was compared with Tools Unique Sequences program, and haplotype known earthworm sequences using the BLAST diversity was calculated using Arlequin v.3.5.1.2 search algorithm in GenBank (Altschul et al. 1997), (Excoffier & Lischer 2010); haplotype relationships and similar sequences were included in subsequent were visualized in HapStar v.0.7 (Teacher & Grif- phylogenetic analyses. fiths 2011) based on a minimum spanning tree from Arlequin. Uncorrected pairwise p-distances among lineages were calculated using MEGA v.6.06 Data analysis (Tamura et al. 2013), and within-lineage pairwise Molecular phylogenetic analyses were conducted distances were calculated using Arlequin. for single-gene datasets (COI or 16S rRNA) and a + combined COI 16S rRNA dataset. The COI data Results matrix consisted of 546 bp for 130 sequences gener- ated for this study; sequences chosen for their close A 546-bp fragment of COI was analyzed for 130 similarity based on BLAST results (Amynthas worms and 24 haplotypes were detected (GenBank tokioensis [GenBank AB542559], Metaphire hilgen- Accession No.s KX454181-KX454310). The ML dorfi [AB543234], and six Amynthas agrestis tree for COI sequences shows a monophyletic group sequences [AB542599, AB542600, AB542601, (bootstrap support=89%) of three well-supported AB542602, AB542604, AB542609]); and four addi- lineages labeled A, B, and C (100% for each lin- tional sequences (Amynthas diffringens [EF077550], eage), with a sister relationship between lineages B Metaphire californica [AY960810], Amynthas gracilis and C strongly supported (91%) (Fig. 1). Thus, [KP688582], and Amynthas robustus [AB542538]). worms originally attributed to Amynthas agrestis For 16S rRNA, we analyzed a 320-bp matrix of 27 comprise three reciprocally monophyletic groups. new sequences aligned with 16S rRNA sequences The three lineages are highly divergent, with mean from GenBank for the same taxa used in the COI pairwise distances for COI ranging from 15.84% analysis (DQ257309, AB474335, NC027258, between B and C, to 24.03% between A and B; the AB474301). The GTR+I+G evolutionary model was mean pairwise distance for A and C is 22.15% chosen, and Maximum Likelihood (ML) analyses (Table 2). Within-lineage distances are very low and were run using the RAxML program (Stamatakis range from 0.01% to 0.4% (Table 2), which leaves a 2006; Stamatakis et al. 2008) on the Cipres Portal gap of over 15% between intra- and inter-lineage (www.phylo.org); bootstrap support for the resulting distances for COI.

Invertebrate Biology vol. x, no. x, xxx 2016 Phylogeographic analysis of invasive earthworms 5

Amynthas_BCA2.13_WI Amynthas_BCA2.17_WI Amynthas_BCA8.24_NY Amynthas_BCA2.20_WI Amynthas_BCA2.12_WI A Amynthas_BCA8.2_NY Amynthas_BCA8.4_NY [NY, WI] Amynthas_BCA2.16_WI Amynthas_BCA2.1_WI Amynthas_BCA8.17_NY Amynthas_BCA7.21_NY Amynthas_BCA2.14_WI Amynthas_BCA2.22_WI Amynthas_BCA8.3_NY Amynthas_BCA8.22_NY Amynthas_BCA7.8_NY 100 Amynthas_BCA8.9_NY Amynthas_BCA8.5_NY Amynthas_BCA8.13_NY Amynthas_BCA8.1_NY Amynthas_BCA7.14_NY Amynthas_BCA8.7_NY Amynthas_BCA7.12_NY Amynthas_BCA8.15_NY Amynthas_BCA8.21_NY Amynthas_BCA8.19_NY Amynthas_BCA2.24_WI Amynthas_BCA24.2_NY Amynthas_BCA24.6_NY Amynthas_BCA8.11_NY Amynthas_BCA8.8_NY Amynthas_BCA8.23_NY Amynthas_BCA2.21_WI Amynthas_BCA2.2_WI Amynthas_BCA2.23_WI Amynthas_BCA8.12_NY Amynthas_BCA7.27_NY Amynthas_BCA7.16_NY Amynthas_BCA8.10_NY Amynthas_BCA8.6_NY Amynthas tokioensis_AB542559 Amynthas_BCA22.14_NY Amynthas_MCHS7_NY Amynthas_MCHS9_NY Amynthas_BCA22.17_NY Amynthas_BCA7.10_NY Amynthas_BCA24.10_NY Amynthas_BCA2.4_WI Amynthas_BCA8.20_NY Amynthas_BCA22.15_NY Amynthas_MCHS10_NY Amynthas_BCA22.11_NY Amynthas_BCA2.8_WI Amynthas_BCA22.18_NY Amynthas_BCA22.12_NY Metaphire hilgendorfi_AB543234 Amynthas_BCA22.19_NY Amynthas_BCA2.5_WI Amynthas_BCA8.14_NY Amynthas_BCA2.25_WI Amynthas_BCA2.9_WI Amynthas_BCA2.7_WI Amynthas_BCA2.3_WI Amynthas_BCA2.10_WI Amynthas_BCA2.6_WI 89 Amynthas_BCA2.11_WI Amynthas_BCA2.18_WI Amynthas_BCA2.15_WI Amynthas_BCA24.7_NY Amynthas_BCA10.4_AL Amynthas_BCA10.7_AL Amynthas_BCA10.5_AL B Amynthas_BCA22.9_NY Amynthas_BCA24.4_NY [NY, AL] Amynthas_MCHS6_NY Amynthas_BCA10.2_AL Amynthas_BCA8.16_NY Amynthas_BCA22.13_NY Amynthas_BCA7.15_NY Amynthas_BCA10.9_AL Amynthas_MCHS2_NY 1 Amynthas_BCA10.6_AL Amynthas_BCA24.12_NY Amynthas_BCA22.10_NY Amynthas_BCA10.10_AL Amynthas_BCA8.18_NY Amynthas_BCA22.21_NY Amynthas_BCA24.3_NY Amynthas_BCA22.7_NY Amynthas_BCA22.8_NY Amynthas_BCA22.20_NY Amynthas_BCA10.3_AL Amynthas_BCA22.16_NY Amynthas_BCA10.8_AL Amynthas_BCA24.11_NY Amynthas_BCA24.8_NY Amynthas_BCA7.11_NY Metaphire agrestis_AB542600 Metaphire agrestis_AB542604 Metaphire agrestis_AB542599 Metaphire agrestis_AB542602 Metaphire agrestis_AB542601 Amynthas_BCA10.1_AL Amynthas_MCHS8_NY Amynthas_MCHS3_NY 100 Amynthas_BCA7.23_NY 100 Amynthas_BCA7.2_NY Amynthas_BCA24.5_NY Metaphire agrestis_AB542609 Amynthas_BCA7.13_NY Amynthas_BCA7.20_NY Amynthas_BCA7.6_NY Amynthas_BCA7.1_NY C 91 Amynthas_BCA7.7_NY Amynthas_BCA7.3_NY [NY] Amynthas_BCA7.9_NY Amynthas_BCA7.4_NY Amynthas_BCA7.26_NY Amynthas_BCA7.22_NY Amynthas_MCHS4_NY Amynthas_BCA21.10_NY 100 Amynthas_BCA7.19_NY Amynthas_BCA7.5_NY Amynthas_BCA7.29_NY Amynthas_BCA7.24_NY Amynthas_BCA21.8_NY Amynthas_BCA7.30_NY Amynthas_BCA21.5_NY Amynthas_MCHS5_NY Amynthas_BCA7.17_NY Amynthas_BCA7.28_NY Amynthas_BCA21.7_NY Amynthas_BCA21.2_NY Amynthas_BCA21.1_NY Amynthas_BCA21.4_NY Amynthas_BCA21.6_NY Amynthas_BCA7.25_NY Amynthas_BCA21.9_NY Amynthas diffringens_EF077550 89 100 Amynthas robustus_AB542538 100 Metaphire californica_AY960810 Amynthas gracilis_KP688582

0.06 Fig. 1. Maximum Likelihood (ML) tree based on the analysis of a 546-bp fragment of the COI gene for 130 specimens of Amynthas sequenced for this study and additional Amynthas sequences from GenBank. Amynthas diffringens was used to root the tree. Bootstrap values >85 are shown on branches. Lineages A, B, and C are shown in blue, red, and green, respectively, along with the states in which they have been found. Location labels for each specimen sequenced match those in Table 1. AL, Alabama; NY, New York; WI, Wisconsin.

Invertebrate Biology vol. x, no. x, xxx 2016 6 Schult, Pittenger, Davalos, & McHugh

Table 2. Mean pairwise uncorrected p-distances for COI five populations, and two haplotypes are shared within (italics, below diagonal) and between (bold, above between the Madison, Wisconsin population and diagonal) lineages A, B, and C of Amynthas from Fig. 1. populations in upstate New York, which are ~1500 km apart (Figs. 3, 6). Lineage B shows a Lineage A Lineage B Lineage C range of haplotype diversities from 0 to 0.8333 Lineage A 0.0048 0.2403 0.2215 across the five populations sampled (Table 3). In Lineage B 0.0000 0.1584 this lineage, the most common of the six haplotypes Lineage C 0.0003 is shared among all populations in upstate New York and in Auburn, Alabama (~1600 km apart), while four of the haplotypes are unique to one pop- Lineage A includes 66 specimens and 11 haplo- ulation sample (Figs. 4, 6). Lineage C population types (18 polymorphic sites) from four locations in samples had a total of seven haplotypes, five of upstate New York and all 24 worms in our Madi- them unique to a single site, and haplotype diversi- son, Wisconsin sample. This lineage also includes ties ranged 0–0.7222 (Table 3, Fig. 6). We found unpublished GenBank sequences recorded as Amyn- worms of lineage C at two upstate New York sites thas tokioensis and Metaphire (Amynthas) hilgen- and Millbrook, New York, separated by ~275 km, dorfi (Fig. 1). Lineage B includes 35 specimens and and the most common haplotype is shared among six haplotypes (five polymorphic sites) for the same the three sites (Fig. 5). We have no record of lineage four locations in upstate New York and all 10 C in either Wisconsin or Alabama. worms in our Auburn, Alabama sample. Several Meta- unpublished GenBank sequences attributed to Discussion phire (Amynthas) agrestis fall within this lineage (Fig. 1). Lineage C is comprised of 29 specimens Our study has extended the available COI and with seven haplotypes (five polymorphic sites) 16S rRNA sequences for Amynthas by 130 and 27, recorded for four sites separated by between 50 km respectively, and represents the first phylogenetic and 200 km in New York; no sequences similar to analysis of Amynthas relationships in North Amer- lineage C haplotypes were found in GenBank. ica. The results presented here suggest the presence A fragment of the 16S rRNA gene was sequenced of three cryptic lineages of Amynthas in upstate for a subsample of 27 specimens spanning the sam- New York, and we propose that these lineages rep- pled sites and representing the three COI lineages resent sympatric species occurring in the newly (GenBank Accession No.s KX469294-KX469320). invaded range of these Asian earthworms in the Alignment of 16S rRNA sequences was straightfor- northeast USA. There is an increasing number of ward with a single indel noted; lineages A and B reports of cryptic earthworm divergence identified worms shared a 320-bp fragment of the 16S rRNA using sequence data, with many of these studies gene, while a 319-bp fragment of the gene was found drawing on divergent COI barcode sequences for for all worms from lineage C. The ML tree for 16S species designation in cases where morphological rRNA shows strong support (100%) for each of the characters distinguishing species are lacking (e.g., three lineages identified in the COI tree (Fig. 2), with Klarica et al. 2012; Deca€ens et al. 2013). For earth- a sister relationship between lineages A and C worm species, low levels of intraspecific COI bar- (100%). ML analysis of a concatenated 16S rRNA code sequence divergences contrast with high levels and COI dataset for the 27 specimens also supports of interspecific COI barcode divergences, ranging monophyly of the three lineages (100%); as with from ~13% upwards (Hebert et al. 2003; Chang COI, lineages B and C are supported as sister taxa in et al. 2009; James et al. 2010; Chang & James 2011; this combined analysis (97%, not shown). Deca€ens et al. 2013; Nygren 2013). For example, COI haplotype diversity and nucleotide diversity Novo et al. (2009, 2010) reported five cryptic species for the three lineages are given in Table 3; haplo- of hormogastrid earthworms with allopatric distri- type frequencies for each lineage at each site are butions on the central Iberian Peninsula based on shown in Figs. 3–5; and Figure 6 shows a minimum intraspecific divergences of up to 4.27% for COI spanning tree representing the relationships among compared with divergences up to 20.2% between identical COI sequences in our samples. COI haplo- cryptic species. In a study of sympatric cryptic type diversities within population samples for lin- diversity, King et al. (2008) identified multiple spe- eage A ranged 0.2857–1 (Table 3). Of the 11 cies of Allolobophora occurring in Great Britain; in haplotypes in lineage A, seven are found in only one this case, the barcode gap between intraspecific and population sample, one haplotype is common to all interspecific divergences based on the Kimura 2-

Invertebrate Biology vol. x, no. x, xxx 2016 Phylogeographic analysis of invasive earthworms 7

Amynthas_BCA8.16_NY

Amynthas_BCA24.5_NY B [NY, AL] Amynthas_BCA22.16_NY 100 Amynthas_BCA24.7_NY

Amynthas_MCHS2_NY

Amynthas_BCA10.3_AL

Amynthas_BCA24.3_NY

Metaphire agrestis_AB474331

Amynthas_BCA7.2_NY

Amynthas_BCA10.1_AL

Amynthas_BCA10.2_AL

Amynthas_MCHS6_NY

Amynthas_BCA8.18_NY 100 Amynthas_MCHS8_NY

Amynthas_BCA10.4_AL Metaphire hilgendorfi_AB474297 A Amynthas_BCA8.15_NY [NY, WI] 90 Amynthas_BCA8.21_NY

Amynthas_BCA2.2_WI 100 Amynthas_BCA22.14_NY 90 Amynthas_BCA2.8_WI

Amynthas_MCHS7_NY

89 Amynthas_BCA22.17_NY

94 Amynthas_BCA2.9_WI Amynthas_BCA21.1_NY C Amynthas_BCA21.6_NY [NY] Amynthas_BCA7.1_NY

Amynthas_BCA7.30_NY 100 Amynthas_BCA7.24_NY

Metaphire hilgendorfi_AB474340

Amynthas gracilis_NC027258

99 Metaphire californica_AB474335 100 Amynthas robustus_AB474301

Amynthas diffringens_DQ257309

0.03 Fig. 2. Maximum Likelihood (ML) tree based on the analysis of a 320-bp fragment of the 16S rRNA gene for 27 spec- imens of Amynthas sequenced for this study and additional Amynthas sequences from GenBank. Amynthas diffringens was used to root the tree. Bootstrap values >85 are shown on branches. Lineages A, B, and C are shown in blue, red, and green, respectively, along with the states in which they have been found. Location labels for each specimen sequenced match those in Table 1. AL, Alabama; NY, New York; WI, Wisconsin.

parameter (K2P) model ranged 8.4–12.5%. For 23 reported interspecific mean divergences for COI of species of Amynthas and Metaphire from China greater than 15% in all cases, with intraspecific (Sichuan, Hebei and Beijing), Huang et al. (2007) divergences ranging 0–2.3%. Although sample sizes

Invertebrate Biology vol. x, no. x, xxx 2016 8 Schult, Pittenger, Davalos, & McHugh

Table 3. Numbers of individuals (N) for which COI was sequenced from each collection location, haplotypes found, and diversity estimates for lineages A, B, and C of Amynthas in Fig. 1.

Lineage Collection location N Number Polymorphic Haplotype Nucleotide haplotypes sites diversity diversity A Hamilton, NY #1 7 2 4 0.2857 0.0021 Hamilton, NY #2 7 3 5 0.5238 0.0029 Clinton, NY 25 3 5 0.3533 0.0019 Cazenovia, NY 3 3 9 1.0000 0.1098 Madison, WI 24 6 8 0.7246 0.0041 B Hamilton, NY #1 4 3 2 0.8333 0.0024 Hamilton, NY #2 8 1 0 0.0000 0.0000 Clinton, NY 9 2 1 0.2222 0.0004 Cazenovia, NY 4 3 2 0.8333 0.0018 Auburn, AL 10 2 1 0.2000 0.0003 C Hamilton, NY #1 18 3 2 0.5425 0.0011 Cazenovia, NY 2 1 0 0.0000 0.0000 Millbrook, NY 9 5 3 0.7222 0.0019

A

Clinton, NY (n=25)

Cazenovia, NY (n=3)

Madison, WI (n=24)

Hamilton, NY #1 (n=7) Hamilton, NY #2 (n=7)

NAD 1983 Projection 0 50 100 200 300 Miles Scale 1:8,342,920 Fig. 3. Pie charts showing frequencies of the 11 COI haplotypes recorded for lineage A of Amynthas by location. Each color represents a different haplotype. n, sample size at each location. were small (n=3 for 22 species, 6 for one species), between 10.2% and 25.5% for six species of Meta- the authors proposed that the high contrast between phire. Novo et al. (2015) studied the phylogeography intra- and interspecific divergence values supports of two invasive species of Amynthas on Sao~ Miguel COI barcoding as a reliable method for distinguish- Island in the Azores and found intraspecific pairwise ing among Amynthas species and other earthworms distances for COI up to 6.9% for Amynthas corticis, (Huang et al. 2007). In a study focused on assessing and up to 4.3% for Amynthas gracilis. In our study, the synonymization of two Amynthas species in Tai- the COI divergences between lineages is between wan, Chang & Chen (2005) reported K2P distances 15.8% and 24.0% and, with very low within-lineage of 0.2% up to 10.3% within two allopatric lineages divergences compared to previous studies of Amyn- of Amynthas formosae, and interspecific distances thas species (this study: 0.0–0.4%), the barcode gap

Invertebrate Biology vol. x, no. x, xxx 2016 Phylogeographic analysis of invasive earthworms 9

CLADE B

Clinton, NY (n=9)

Cazenovia, NY (n=4)

Hamilton, NY #1 (n=4) Hamilton, NY #2 (n=8)

Auburn, AL (n=10)

NAD 1983 Projection 0 50 100 200 300 Miles Scale 1:8,342,920 Fig. 4. Pie charts showing frequencies of the six COI haplotypes recorded for lineage B of Amynthas by location. Each color represents a different haplotype. n, sample size at each location.

CLADE C

Cazenovia, NY (n=2)

Millbrook, NY (n=9)

Hamilton, NY #1 (n=18)

NAD 1983 Projection 0 50 100 200 300 Miles Scale 1:8,342,920 Fig. 5. Pie charts showing frequencies of the seven COI haplotypes recorded for lineage C of Amynthas by location. Each color represents a different haplotype. n, sample size at each location. for all the three lineages is greater than 15% and is Two recent papers call the COI barcode gap into well above the suggested 10x rule of thumb differ- question as a suitable means of earthworm species ence between intra- and inter-lineage divergences for discovery when used alone. In Martinsson et al. species delineation (Hebert et al. 2004; Carr et al. (2015), the maximum pairwise distance between COI 2011). lineages of Apporectodea longa was 7.8%, and

Invertebrate Biology vol. x, no. x, xxx 2016 10 Schult, Pittenger, Davalos, & McHugh

A lineages of Amynthas represent cryptic species is clear. In the absence of nuclear data thus far, we can draw on the pattern of internal reproductive struc- tures as corroboration in support of our hypothesis that multiple species exist among invasive popula- tions of Amynthas in the northeast USA. We sent 61 unlabeled specimens to S. James (University of Iowa) for his independent assessment of species identification based on any inconspicuous genital markings, internal glands, and spermathecal struc- tures; in combination, these characters provide a good guide to species designation for Amynthas. 39 After independently examining the worms without C knowledge of our molecular results, S. James reported that: specimens we later identified as belonging to lineage A lack all sexual organs except B ovaries, and are likely in the Amynthas hilgen- dorfi/tokioensis complex; worms from lineage B have Fig. 6. Haplotype network for 24 haplotypes of the cyto- three pairs of spermathecae and are members of chrome oxidase I gene fragment in 130 samples of Amyn- Amynthas agrestis; lineage C worms have two pairs thas from New York, Wisconsin, and Alabama. Each of spermathecae and would conventionally be identi- circle represents a unique haplotype and each black dot fied as Amynthas hilgendorfi. From the internal mor- along branches in the network represents a mutational phology, all of the worms seemed to lack any male step between haplotypes. The areas of circles on the net- work are proportional to the number of individuals shar- function (S. James, unpubl. data), and thus the three ing each haplotype (range 1–36). Lineages A (blue), B lineages are presumably parthenogenetic. This (red), and C (green) from Fig. 1 are represented in the accounts for the very low within-lineage divergences network with the proportion of each haplotype accounted we observed (0–0.4%). Interestingly, the identifica- for by worms from New York, Wisconsin, or Alabama tions provided by S. James are reflected in the represented by diagonal lines, diamonds, or hatching, grouping of GenBank sequences with two of the lin- respectively. eages. COI lineage A includes sequences from Gen- Bank attributed to A. tokioensis and Metaphire hilgendorfi (=Amynthas hilgendorfi), and lineage B distances within COI lineages ranged between 0.0 includes several sequences labeled as Metaphire and 2.0%, with a barcode gap of up to 6.4%; how- agrestis (=Amynthas agrestis) (Fig. 1). We found no ever, the COI lineages were not recovered in analy- COI sequences in GenBank similar to those of lin- sis of ITS2 sequences for the same specimens. For eage C. populations of Lumbricus rubellus in central Europe, Our 16S rRNA ML tree provides strong support Giska et al. (2015) observed pairwise distances from for each of the three lineages of Amynthas evident 1.3% up to 16% between co-occurring COI lineages; in the COI topology. The two datasets do conflict in no within-lineage divergences were given, but from the relationship among the three lineages, although supplementary data provided they are estimated to the combined analyses support a sister relationship be between 0.04% and 2.3%. By contrast, the analy- between lineages A and B, and it is noteworthy that sis of genome-wide RADseq data (over 5000 SNPs) there is a shared indel in the 16S rRNA sequences by Giska et al. (2015) grouped specimens according for worms in these two lineages. We do not make to population of origin and did not support the conclusions about the relationships among the three reproductive isolation indicated by COI analysis. lineages here. It is beyond the scope of this study Neither of these studies meet the suggested 10x rule because our sampling across the diversity of Amyn- of thumb difference for barcode gaps. As mentioned thas taxa is limited thus far. This taxon sampling above, the COI barcode gaps we observed among limitation might explain the difference we see in the Amynthas did, and they were greater than those for relationships among the three lineages for the two mitochondrial markers in Martinsson et al. (2015) genes, or perhaps different rates of evolution or Giska et al. (2015). Nonetheless, the need for between and across COI or 16S rRNA are not well analysis of a nuclear marker to assess whether the accounted for in the chosen models. While the

Invertebrate Biology vol. x, no. x, xxx 2016 Phylogeographic analysis of invasive earthworms 11 placement of the lineages within a phylogeny of earthworms (Marinissen & van den Bosch 1992), Amynthas will require further work, both the identi- perhaps because of the unusually rapid body move- fication of the worms based on internal morphology ments of Amynthas worms, but it is still not likely and the phylogenetic analyses of COI and 16S to account for the apparent recent spread of Amyn- rRNA point to the presence of three sympatric spe- thas worms across the northeast USA. According to cies of Amynthas in upstate New York, with one of various online blogs and news reports (D. McHugh, them being A. agrestis. unpubl. data), bait worms in the genus Amynthas Because we sampled populations of Amynthas are growing in popularity in the region, and as with without a priori knowledge of the cryptic lineage other invasive earthworms, the abandonment of live diversity, population sample sizes for each lineage bait may also facilitate dispersal to new locations are relatively small and uneven for COI and even (Bohlen et al. 2004a,b; Hendrix et al. 2006; smaller for 16S rRNA. Unfortunately, this limits Cameron et al. 2008). any in-depth analysis of haplotype diversity and dis- tribution within each lineage; nonetheless, some observations are worth noting regarding COI haplo- Conclusion types. More than one COI haplotype was recorded Our results indicate that multiple divergent mito- for all samples except the Hamilton, NY #2 sample chondrial lineages of Amynthas appear to be invad- (n=8) for lineage B and the Cazenovia, NY sample ing the northeast USA. Three well-supported, (n=2) for lineage C; and wide-ranging levels of hap- reciprocally monophyletic lineages of Amynthas lotype diversity occur in each of the three lineages show levels of COI sequence divergence that are of Amynthas in our study. Similar observations have consistent with species-level differentiation in other been made in studies of other parthenogenetic earth- earthworms, and this is corroborated by internal worms. In Dendrobaena octaedra, for example, hap- morphology. We are complementing mitochondrial lotype diversity ranges 0–0.6444 in invasive data with nuclear markers, and we continue to sam- populations in Alberta, Canada (Cameron et al. ple additional populations of Amynthas across the 2008) and from 0.242 to 0.917 in populations of the northeast USA as well as from native ranges to ver- same species in its European native range (Knott & ify the proposed cryptic species. While we do not Haimi 2010). Although low, the levels of haplotype have sufficient data to reconstruct the invasion his- diversity in many of the invasive parthenogenetic tory of Amynthas, the diversity and distribution of populations of D. octaedra are greater than 0 and COI haplotypes we observed across broad geo- thus suggest multiple independent introductions or graphic distances for each lineage suggest that multi- the single introduction of a large number of worms ple introductions likely occur, perhaps from from a diverse source population (Cameron et al. common origins, either by direct routes from native 2008). Either of these scenarios may explain the ranges or through post-introduction spread by natu- haplotype diversities for the three parthenogenetic ral dispersal or human-mediated transport within lineages of Amynthas also, with the transport of North America. The extensive movement of horti- worms in horticultural materials (mulch, potted cultural materials for trade and the growing use of plants) the likely mode of introduction in either case “Crazy Worms” as bait in the northeast USA likely (Richardson et al. 2009; Belliturk et al. 2015). The facilitates widespread dispersal of the worms and general suitability of horticultural materials for spe- may allow simultaneous transportation of multiple cies of Amynthas, as well as the extensive trade in lineages to the wide range of mixed forest habitats those materials both internationally and nationally, suitable for Amynthas. Focusing management efforts might also explain the sympatric occurrence of the on control measures in commercial nurseries and on three lineages in upstate New York. It is possible prohibition of bait disposal could help to reduce the that worms are transported as an assemblage, as has further invasion success of Amynthas. been documented for some marine species (e.g., Per- eyra et al. 2015). Alternatively, the three lineages might have invaded upstate New York simultane- Acknowledgments. We are grateful to the Society for Integrative & Comparative Biology for hosting the special ously or sequentially along leading edges from the session in honor of Kristian Fauchald at the annual meet- south and east. Shifting of population boundaries ing in Portland, OR in January 2016, where this work has been reported to be just 12 m per year for was first presented. We thank the following colleagues for worms identified as Amynthas agrestis in Tennessee providing specimens for this analysis: Ken Halanych (Snyder et al. 2011). This is substantially farther (Auburn University), Brad Herrick (University of Wis- than reports of 2–4 m per year for other consin-Madison Arboretum), and Tim McCay (Colgate

Invertebrate Biology vol. x, no. x, xxx 2016 12 Schult, Pittenger, Davalos, & McHugh

University). We also thank Ali Hogue, Gian Sepulveda, Cameron EK, Bayne EM, & Coltman DW 2008. Genetic Vanessa Cortes, Nikki Doroshenko, and the students of structure of invasive earthworms Dendrobaena octaedra the fall 2015 BIOL206 course at Colgate University for in the boreal forest of Alberta: insights into introduc- their help in collecting the sequence data for the analyses. tion mechanisms. Mol. Ecol. 17: 1189–1197. We are grateful to Samuel James (University of Iowa) Carr CM, Hardy SM, Brown TM, Macdonald TA, & who kindly provided his taxonomic expertise in examina- Hebert PDN 2011. A tri-oceanic perspective: DNA bar- tion of Amynthas specimens. None of the authors have coding reveals geographic structure and cryptic diver- any conflicts of interest to declare. This work was sup- sity in Canadian polychaetes. PLoS ONE 6: e22232. ported in part by NSF-DEB 1036530 to DMH. Chang C-H & Chen J-H 2005. Taxonomic status and intraspecific phylogeography of two sibling species of References Metaphire (Oligochaeta: Megascolecidae) in Taiwan. Pedobiologia 49: 591–600. Altschul SF, Madden TL, Schaffer€ AA, Zhang J, Zhang Chang C-H & James S 2011. A critique of earthworm Z, Miller W, & Lipman DJ 1997. Gapped BLAST and molecular phylogenetics. Pedobiologia 54S: S3–S9. PSI-BLAST: a new generation of protein database Chang C-H, Rougerie R, & Chen J-H 2009. Identifying search programs. Nucleic Acids Res. 25(17): 3389– earthworms through DNA barcodes: pitfalls and pro- 3402. mise. Pedobiologia 52: 171–180. Archer KL 2012. The introduction of exotic earthworms by Deca€ens T, Porco D, Rougerie R, Brown GG, & James European colonists; Impact on agriculture and forestry: SW 2013. Potential of DNA barcoding for earthworm the earthworm invasion. Megadrilogica 15: 212–214. research in taxonomy and ecology. Appl. Soil Ecol. 65: Bastos AD, Nair D, Taylor PJ, Brettschneider H, Kirsten 35–42. F, Mostert E, von Maltitz E, Lamb JM, van Hooft P, Excoffier L & Lischer HEL 2010. Arlequin suite ver 3.5: Belman SR, Contrafatto G, Downs S, & Chimimba CT a new series of programs to perform population genet- 2011. Genetic monitoring detects an overlooked cryptic ics analyses under Linux and Windows. Mol. Ecol. species and reveals the diversity and distribution of Res. 10: 564–567. three invasive Rattus congeners in South Africa. BMC Folino-Rorem NC, Darling JA, & D’Ausilio CA 2009. Genet. 2011: 12–26. Genetic analysis reveals multiple cryptic invasive spe- Belliturk K, Gorres€ JH, Kunkle J, & Melnichuk RDS cies of the hydrozoan genus Cordylophora. Biol. Inva- 2015. Can commercial mulches be reservoirs of invasive sions 11: 1869–1882. earthworms? Promotion of ligninolytic enzyme activity Folmer O, Black M, Hoeh W, Lutz R, & Vrijenhoek R and survival of Amynthas agrestis (Goto and Hatai, 1994. DNA primers for amplification of mitochondrial 1899). Appl. Soil Ecol. 87: 21–31. cytochrome c oxidase subunit I from diverse metazoan Bernard MJ, Neatrour MA, & McCay TS 2009. Influence invertebrates. Mol. Mar. Biol. Biotech. 3: 294–299. of soil buffering capacity on earthworm growth, sur- Gates GE 1958. On some species of the Oriental earth- vival, and community composition in the western worm genus Kinberg, 1867, with a key to Adirondacks and central New York. Northeast. Nat. species reported from the Americas. Am. Mus. Nat. 16: 269–284. Hist. Nov. 1888: 1–33. Blakemore RJ 2003. Japanese earthworms (Annelida: Oli- ———— 1982. Farewell to North American megadriles. gochaeta): a review and checklist of species. Org. Megadrilogica 4: 12–77. Divers. Evol. 3: 241–244. Giska I, Sechi P, & Babik W 2015. Deeply divergent sym- ———— 2013. Megascolex (Perichaeta) diffringens Baird, patric mitochondrial lineages of the earthworm Lumbri- 1869 and Pheretima pingi Stephenson, 1925 types com- cus rubellus are not reproductively isolated. BMC Evol. pared to the Amynthas corticis (Kinberg, 1867) and A. Biol. 15: 217. carnosus (Goto & Hatai, 1899) species-groups (Oligo- Gorres€ JH & Melnichuk RDS 2012. Asian invasive earth- chaeta: Megadrilacea: Megascolecidae). J. Species Res. worms of the genus Amynthas Kinberg in Vermont. 2: 99–126. Northeast. Nat. 19: 313–322. Bohlen PJ, Scheu S, Hale CM, McLean MA, Migge S, Gorres€ JH, Melnichuk RDS, & Belliturk€ K 2014. Mortal- Groffman PM, & Parkinson D 2004a. Non-native inva- ity pattern relative to size variation within Amynthas sive earthworms as agents of change in northern tem- agrestis (Goto & Hatai 1899) (Oligochaeta: Megascole- perate forests. Front. Ecol. Environ. 2: 427–435. cidae) populations in the Champlain Valley of Ver- Bohlen PJ, Groffman PM, Fahey TJ, Fisk M, Suarez E, mont, USA. Megadriogica 16: 9–14. Pelletier DM, & Fahey RT 2004b. Ecosystem conse- Goto S & Hatai S 1899. New or imperfectly known species quences of exotic earthworm invasion of north temper- of earthworms. No. 2. Annot. Zool. Japan 3: 13–24. ate forests. Ecosystems 7: 1–12. Greiner HG, Kashian DR, & Tiegs SD 2012. Impacts of Callaham MA Jr, Hendrix PF, & Phillips RJ 2003. invasive Asian (Amynthas hilgendorfi) and European Occurrence of an exotic earthworm (Amynthas agrestis) (Lumbricus rubellus) earthworms in a North American in undisturbed soils of the southern Appalachian temperate deciduous forest. Biol. Invasions 14: 2017– Mountains, USA. Pedobiologia 47: 466–470. 2027.

Invertebrate Biology vol. x, no. x, xxx 2016 Phylogeographic analysis of invasive earthworms 13

Hebert PDN, Ratnasingham S, & de Waard JR 2003. (Annelida, Oligochaeta) in the central Iberian Penin- Barcoding life: cytochrome c oxidase subunit 1 sula: evolutionary and demographic implications. Zool. divergences among closely related species. Proc. R. Soc. Scripta 38(5): 537–552. Lond. B 270: S96–S99. Novo M, Almodovar A, Fernandez R, Trigo D, & Dıaz Hebert PDN, Penton EH, Burns JM, Janzen DH, & Hall- Cosın DJ 2010. Cryptic speciation of hormogastrid wachs W 2004. Ten species in one: DNA barcoding earthworms revealed by mitochondrial and nuclear reveals cryptic species in the neotropical skipper butter- data. Mol. Phylogenet. Evol. 56: 507–512. fly Astraptes fulgerator. Proc. Natl Acad. Sci. USA Novo M, Cunha L, Maceda-Veiga A, Talavera JA, Hod- 101: 14812–14817. son ME, Spurgeon D, Bruford MW, Morgan AJ, & Hendrix PF & Bohlen PJ 2002. Exotic earthworm inva- Kille P 2015. Multiple introductions and environmental sions in North America: ecological and policy implica- factors affecting the establishment of invasive species tions. Bioscience 52: 801–811. on a volcanic island. Soil Biol. Biochem. 85: 89–100. Hendrix PF, Baker GH, Callaham MA Jr, Damoff GA, Nygren A 2013. Cryptic polychaete diversity: a review. Fragoso C, Gonzalez G, James SW, Lachnicht SL, Zool. Scripta 43: 1–12. Winsome T, & Zou X 2006. Invasion of exotic earth- Pejchar L & Mooney HA 2009. Invasive species, ecosys- worms into ecosystems inhabited by native earthworms. tem services and human well-being. TREE 24: 497–504. Biol. Invasions 8: 1287–1300. Pereyra PJ, Narvarte M, Tatian M, & Gonzalez R 2015. Huang J, Xu Q, Sun ZJ, Tang GL, & Su ZY 2007. Iden- The simultaneous introduction of the tunicate Styela tifying earthworms through DNA barcodes. Pedobiolo- clava (Herdman, 1881) and the macroalga Undaria pin- gia 51: 301–309. natifida (Harvey) Suringar, 1873, in northern Patago- Hulme PE 2009. Trade, transport, and trouble: managing nia. Biol. Invas. Rec. 4: 179–184. invasive species pathways in an era of globalization. J. Porco D, Deca€ens T, Deharveng L, James SW, Skarzyn- Appl. Ecol. 46: 10–18. ski D, Erseus C, Butt KR, Richard B, & Hebert PDN Ikeda H, Callaham MA Jr, O’Brien JO, Hornsby BS, & 2013. Biological invasions in soil: DNA barcoding as a Wenk ES 2015. Can the invasive earthworm, Amynthas monitoring tool in a multiple taxa survey targeting agrestis, be controlled with prescribed fire? Soil Biol. European earthworms and springtails in North Amer- Biochem. 82: 21–27. ica. Biol. Invasions 15: 899–910. James SW, Porco D, Deca€ens T, Richard B, Rougerie R, Qiu J 2015. A global synthesis of the effects of biological & Erseus C 2010. DNA barcoding reveals cryptic diver- invasions on greenhouse gas emissions. Global Ecol. sity in Lumbricus terrestris L., 1758 (): resur- Biogeogr. 24: 1351–1362. rection of L. herculeus (Savigny, 1826). PLoS ONE 5: Regier JC & Shi D 2005. Increased yield of PCR product e15629. from degenerate primers with nondegenerate, nonho- Kinberg JGH 1867. Annulata nova. Ofv.€ Ak. Forh.€ 23: mologous 5’ tails. Biotechniques 38: 34–38. 97–103. Reynolds JW 1978. The earthworms of Tennessee (Oligo- King RA, Tibble AL, & Symondson WO 2008. Opening chaeta). Part 4. Megascolecidae, with notes on distribu- a can of worms: unprecedented sympatric cryptic diver- tion, biology and a key to the species in the state. sity within British lumbricid earthworms. Mol. Ecol. Megadrilogica 3: 117–130. 17: 4684–4698. ———— 2010. The earthworms (Oligochaeta: Acanthodrili- Klarica J, Kloss-Brandstatter€ A, Traugott M, & Juen A dae, Lumbricidae, Megascolecidae and Sparganophili- 2012. Comparing four mitochondrial genes in earthworms dae) of northeastern United States, revisited. – implications for identification, phylogenetics, and dis- Megadrilogica 14: 101–157. covery of cryptic species. Soil Biol. Biochem. 45: 23–30. ———— 2011. The earthworms (Oligochaeta: Acanthodrili- Knott KE & Haimi J 2010. High mitochondrial DNA dae, Glossoscolecidae, Komarekionidae, Lumbricidae, sequence diversity in the parthenogenetic earthworm Megascolecidae and Sparganophilidae) of the midwest- Dendrobaena octaedra. Heredity 105: 341–347. ern United States. Megadrilogica 15: 69–139. Maddison WP & Maddison DR 2015. Mesquite: a modu- Reynolds JW & Wetzel MJ 2004. Terrestrial Oligochaeta lar system for evolutionary analysis. Version 3.03. in North America north of Mexico. Megadrilogica 9: Available online at: http://mesquiteproject.org. 71–98. Accessed October 1, 2015. Reynolds JW, Gorres€ JH, & Knowles ME 2015. A check- Marinissen JCY & van den Bosch F 1992. Colonization list by counties of earthworms (Oligochaeta: Acantho- of new habitats by earthworms. Oecologia 91: 371–376. drilidae, Lumbricidae and Megascolecidae) in the states Martinsson S, Rhoden C, & Erseus C 2015. Barcoding of Maine, New Hampshire and Vermont, USA. gap, but no support for cryptic speciation in the earth- Megadrilogica 17: 125–140. worm Aporrectodea longa (Clitellata: Lumbricidae). Richardson DR, Snyder BA, & Hendrix PF 2009. Soil Mitochondrial DNA 28: 1–9. doi: 10.3109/19401736. moisture and temperature: tolerances and optima for a 2015.1115487. non-native earthworm species, Amynthas agrestis (Oli- Novo M, Almodovar A, & Diaz-Cosin D 2009. High gochaeta: : Megascolecidae). Southeast. genetic divergence of hormogastrid earthworms Nat. 8: 325–334.

Invertebrate Biology vol. x, no. x, xxx 2016 14 Schult, Pittenger, Davalos, & McHugh

Sjolin€ E, Erseus C, & Kallersj€ o€ M 2005. Phylogeny of Teacher AGF & Griffiths DJ 2011. HapStar: automated Tubificidae (Annelida, Clitellata) based on mitochon- haplotype network layout and visualisation. Mol. Ecol. drial and nuclear sequence data. Mol. Phyl. Evol. 35: Res. 11: 151–153. 431–441. Tiunov AV, Hale CM, Holdsworth AR, & Vsevolodova- Snyder BA 2008. Invasion by the non-native earthworm Perel TS 2006. Invasion patterns of Lumbricidae into Amynthas agrestis (Oligochaeta: Megascolecidae): the previously earthworm-free areas of northeastern dynamics, impacts, and competition with millipedes. Europe and the western Great Lakes region of North PhD dissertation, University of Georgia, Athens, Geor- America. Biol. Invasions 6: 1223–1234. gia. pp. 1–149. Tsutsui ND, Suarez AV, Holway DA, & Case TJ 2000. Snyder BA, Boots B, & Hendrix PF 2009. Competition Reduced genetic variation and the success of an inva- between invasive earthworms (Amynthas corticis, sive species. Proc. Natl Acad. Sci. USA 97: 5948–5953. Megascolecidae) and native North American millipedes Vaidya G, Lohman DJ, & Meier R 2011. SequenceMa- (Pseudopolydesmus erasus, Polydesmidae): effects on trix: concatenation software for the fast assembly of carbon cycling and soil structure. Soil Biol. Biochem. multi-gene datasets with character set and codon infor- 41: 1442–1449. mation. Cladistics 27: 171–180. Snyder BA, Callaham MA Jr, & Hendrix PF 2011. Spa- Vila M, Espinar JL, Hejda M, Hulme PE, Jarosık V, tial variability of an invasive earthworm (Amynthas Maron JL, Pergl J, Schaffner U, Sun Y, & Pysek P agrestis) population and potential impacts on soil char- 2011. Ecological impacts of invasive alien plants: a acteristics and millipedes in the Great Smoky Moun- meta-analysis of their effects on species, communities tains National Park, USA. Biol. Invasions 13: 349–358. and ecosystems. Ecol. Lett. 14: 702–708. Snyder BA, Callaham MA Jr, Lowe CN, & Hendrix PF Wetzel MJ 2005. Checklist of the aquatic and terrestrial 2013. Earthworm invasion in North America: food Oligochaeta occurring in North Carolina, South Caro- resource competition affects native millipede survival lina, and/or Tennessee, with notations for species now and invasive earthworm reproduction. Soil Biol. Bio- known to occur in the Great Smokey Mountains chem. 57: 212–216. National Park. Illinois Natural History Survey, Univer- Stamatakis A 2006. RAxML-VI-HPC: maximum likeli- sity of Illinois, Champaign, IL. hood-based phylogenetic analyses with thousands of Yassin A, Capy P, Madi-Ravazzi L, Ogereau D, & David taxa and mixed models. Bioinformatics 22: 2688–2690. JR 2008. DNA barcode discovers two cryptic species Stamatakis A, Hoover P, & Rougemont J 2008. A rapid and two geographical radiations in the invasive droso- bootstrap algorithm for the RAxML Web servers. Syst. philid Zaprionus indianus. Mol. Ecol. Res. 8: 491–501. Biol. 57: 758–771. Zhang W, Hendrix PF, Snyder BA, Molina M, Li J, Rao Steinberg DA, Pouyat RV, Parmelee RW, & Groffman X, Siemann E, & Fu S 2010. Dietary flexibility aids PM 1997. Earthworm abundance and nitrogen mineral- Asian earthworm invasion in North American forests. ization rates along an urban-rural land use gradient. Ecology 91: 2070–2079. Soil Biol. Biochem. 29: 427–430. Ziemba JL, Cameron AC, Peterson K, Hickerson CM, & Tamura K, Stecher G, Peterson D, Filipski A, & Kumar Anthony CD 2015. Invasive Asian earthworms of the S 2013. MEGA6: Molecular Evolutionary Genetics genus Amynthas alter microhabitat use by terrestrial Analysis version 6.0. Mol. Biol. Evol. 30: 2725–2729. salamanders. Can. J. Zool. 93: 805–811.

Invertebrate Biology vol. x, no. x, xxx 2016