Introduction

Invasive earthworms are one of the most important ecological drivers currently changing forest ecosystems in North America (Bohlen et al. 2004; Dávalos et al. 2015; Craven et al. 2017). Of particular concern are three Asian species in the genera Amynthas and Metaphire due to their impacts on forest floor, soil structure, carbon dynamics, nutrient concentration, and other soil organisms (Zhang et al. 2010; Snyder et al. 2011, 2013; Greiner et al. 2012; Qiu and Turner 2017). In the scientific and popular literature these species are frequently referred to as “Amynthas spp.” (e.g. Dávalos et al. 2015) and nicknamed “crazy worm” or “jumping worm”. They invade remote forests in the United States (Callaham et al. 2003), and one species in particular, Amynthas agrestis (Goto & Hatai, 1899), was recently discovered spreading north and west from the Mid-Atlantic into New England and the Midwest (Görres and Melnichuk 2012; Reynolds et al. 2015; Schult et al. 2016).

So far most of the Amynthas and Metaphire sightings in the northeastern United States involved only A. agrestis (Callaham et al. 2003; Snyder et al. 2011; Görres and Melnichuk 2012; Görres et al. 2014; Reynolds et al. 2015; Schult et al. 2016), with occasional reports of Metaphire hilgendorfi (Michaelsen, 1892), often as Amynthas hilgendorfi (e.g. Chang et al. 2016a; see Chang et al. 2016b for details), a species closely related to A. agrestis. The two species have never been reported at the same locality. However, recent genetic analysis indicated sympatric occurrence of three species-level cryptic lineages within invading Amynthas in upstate New York (Schult et al. 2016). Meanwhile, at several locations where A. agrestis or M. hilgendorfi was reported, we also found a previously overlooked species, Amynthas tokioensis (Beddard, 1892) (Chang et al. 2016b). These findings led us to re-examine archived and new specimens to understand the extent of co-occurrence of the three species.

We examined 1224 specimens previously reported from 15 sites as Amynthas spp. (Dávalos et al. 2015; Schult et al. 2016; Table 1) and 536 newly collected specimens from another 15 sites where Asian earthworms have never been reported (Wilmette, IL) or only one of A. agrestis or M. hilgendorfi was previously reported or observed (Szlavecz et al. 2006, 2014; Görres and Melnichuk 2012; Chang et al. 2016a). All samples were collected from 2002 to 2016 with a variety of techniques, including hand sorting, digging, and extraction by formalin or mustard solutions. Species identification followed Chang et al. (2016b).

Table 1 Sampling locations in northeastern USA and numbers of Metaphire hilgendorfi, Amynthas agrestis and Amynthas tokioensis recorded
Full size table

Results

A total of 1760 specimens (1124 identifiable) belonging to three species, M. hilgendorfi, A. agrestis and A. tokioensis, were collected from 30 sites across eight states in the Midwest and northeastern USA (Fig. 1; Table 1). Amynthas tokioensis, M. hilgendorfi and A. agrestis were found in 25, 23 and 22 sites, respectively. Of the 30 sites, 27 had at least two species, and 13 had all three. In addition to being the most common species, A. tokioensis was also dominant in 11 of the 25 sites based on relative abundance, whereas A. agrestis was only dominant in nine sites. Among 51 specimens that were analyzed genetically in Schult et al. (2016) and re-examined morphologically per Chang et al. (2016b), 22 were unidentifiable, but 29 were identified as A. tokioensis (8), A. agrestis (12), or M. hilgendorfi (9), and all species identifications corresponded unequivocally to the three species-level genetic lineages, A, B and C reported in Schult et al. (2016), respectively.

Fig. 1
figure 1

Presence of Metaphire hilgendorfi (H, yellow), Amynthas agrestis (A, orange) and A. tokioensis (T, green) at sampling locations in northeastern United States (the Mid-Atlantic, New England, and the Midwest) based upon morphological examination of 1760 specimens; absence of a species at a sampling location is labeled white. Labels for each sampling location match those in Table 1

Full size image

Discussion

Our data highlight widespread species co-occurrence in the current range of Amynthas and Metaphire in northeastern USA, and provide the first confirmed reports of A. tokioensis in Maryland, Delaware, Vermont, Illinois, and Wisconsin and M. hilgendorfi in New Hampshire. We herein emphasize that co-occurrence of two or three species is what should be expected, whereas the situation with a single species presence is rare. Sympatric occurrence between A. agrestis and A. tokioensis has been confirmed in compost (Gates 1958), but never before from natural habitats such as forest soils. Moreover, the universal presence or even dominance of A. tokioensis in forested habitats has been overlooked for a long time. This lack of reporting might be due to difficulties with specimen identification, rather than absence of the species from a locality. Recent scientific literature documented the invasion as only by A. agrestis. In some cases, species identification of A. agrestis was achieved by examining only a small number of individuals, or was based on the color and shape of clitellum and the jumping and snake-like behavior. Unfortunately, all of these characteristics are also shared by A. tokioensis and M. hilgendorfi and thus should not be used as a basis for species diagnosis. While we recognize the difficulty of identifying pheretimoid earthworms, we urge researchers and managers to re-examine their collections and confirm species identities by taxonomically reliable morphological traits (e.g., number and location of genital pores). A key to identify Amynthas and Metaphire species in North America is now available for this purpose (Chang et al. 2016b). Identification should preferably include at least 20 adult specimens to adequately address potential co-occurrence of more than one species.

Amynthas tokioensis has been previously reported living in natural habitats in only six states, all of which are in the Southeast or Mid-Atlantic (Reynolds and Wetzel 2008). Confirmed records of A. tokioensis in Vermont, Illinois and Wisconsin clearly show that the known distribution of this species has expanded further north and northwest into New England and the Midwest and the species is able to survive cold winters (presumably as cocoons; Görres et al. 2016). Similarly, M. hilgendorfi had not been recorded outside of greenhouse or compost habitats in any of the New England or Midwest states (Reynolds and Wetzel 2008) until it was found recently in Vermont (Reynolds et al. 2015) and now in New Hampshire and Illinois. Altogether, our findings of M. hilgendorfi and A. tokioensis and recent reports of A. agrestis from Wisconsin, Illinois, Vermont and New Hampshire (Görres and Melnichuk 2012; Reynolds et al. 2015; Schult et al. 2016; Qiu and Turner 2017) clearly show a northward and westward range expansion of the three species into colder regions, presumably in the last 10–20 years. The expansion of A. tokioensis may have taken place alongside A. agrestis and/or M. hilgendorfi into the four states, but was overlooked in previous surveys (Chang personal observation).

Correct species identification and assessment of each species’ abundance reaches beyond the simple descriptive characterization of earthworm assemblages. While functional groups help describe how one earthworm versus another may profoundly alter soil characteristics via litter consumption and soil vertical mixing (Bohlen et al. 2004), species identity more fully explains how non-native earthworms affect soil organic matter translocation and aggregate formation in interaction with native earthworms, other soil fauna, and the microbial community (e.g. Chang et al. 2016a, c, 2017). To date, A. tokioensis, A. agrestis, and M. hilgendorfi are only loosely categorized as epi-endogeic, and data are scarce with regard to their life history, diet, and activity in different soils and to which depths. More importantly, we have essentially no knowledge on how these species differ ecologically. The widespread species co-occurrence implies niche differentiation and possible co-invasion. Autecological studies targeted toward each species are needed to understand how co-occurrence affects invasion success and competition with resident species, as documented between Lumbricus rubellus and the ecologically similar A. agrestis (Zhang et al. 2010) or M. hilgendorfi (Chang et al. 2016a). Pheretimoids are frequently the sole species in invaded forests in Vermont and at West Point New York (Görres personal observation; Dávalos et al. 2015). The recognition of an expanding and multi-species system reported herein provides unique opportunities to study earthworm invasions and their impacts on forest soil properties. This assemblage information, while raising serious concerns, also allows a new set of questions to be asked regarding the complex interactions among co-invading species and the potential non-additive impacts they may have on ecological processes and ecosystem functions.