Introduction

The worldwide decline in taxonomic expertise (Godfray 2002, Walter & Winterton 2007, Wheeler 2004) is inhibiting prevention and management of biological invasions (Lodge et al. 2006). Protocols for diagnosing pests and diseases underpin essentially all quarantine activities, and low-quality, inaccessible or absent taxonomic keys are major impediments to effective border biosecurity and pest management (IPPC 2006, Lodge et al. 2006). For a taxonomic specialist, identifying a specimen may only require a brief check of critical characters (Walter & Winterton 2007). However, less specialised diagnosticians dealing with unfamiliar species in quarantine situations require extremely well-designed taxonomic tools to obtain correct identifications. Therein lies the problem, as traditional dichotomous keys are “compiled by those who do not need them for those who cannot use them” (Lobanov 2003), resulting in tools that can be inadequate for quarantine requirements (Lodge et al. 2006, Walter & Winterton 2007).

One way to alleviate the shortage of taxonomic expertise is to capitalise on new, rapidly evolving technologies, such as polytomous (matrix based) electronic interactive keys (Agnarsson & Kuntner 2007, Chesmore 2002, Lodge et al. 2006, Norton 2002, 2005, Walter & Winterton 2007). While paper-based dichotomous keys are simple, portable, and do not require associated technologies such as software and computers to be used, they have particularly limited utility, both when specimens are damaged and when diagnosticians have difficulty recognising particular characters. Reaching a conclusion with a dichotomous key depends on the user being able to progress stepwise through couplets or less commonly a greater number, e.g. triplets, of questions in a manner predetermined by the key's author. However, damaged structures and incorrect interpretations of characters can render couplets unanswerable. Polytomous keys allow users to bypass characters that are unrecognisable, hidden, damaged or missing and choose those that are most easily observed or recognised on the specimen. Although paper-based polytomous keys can be difficult to use (Walter & Winterton 2007), this can be overcome by presenting them in electronic format using custom, e.g. Walker et al. 2005, or commercial software such as Delta (Dallwitz 2007, 2009,  http://delta-intakey.com) and Lucid® (Norton 2002, 2005,  http://Lucidcentral.com). Further advantages of electronic keys over paper-based keys are that they can be easily corrected or updated and made readily available via the World Wide Web, and that genetic data and media files can be easily incorporated. Disadvantages associated with the use of electronic keys are their cost, reliance on the user having access to a computer on which the appropriate software is loaded, and the ability to utilise electronic resources in field situations. Whilst the ongoing and rapid development of software and handheld computing devices is likely to overcome the aforementioned problems, it is likely that there will be a need for both paper and electronic versions of keys for the foreseeable future.

Here, I describe the application of computer software to develop a key to hard tick species (Acari: Ixodidae) that are significant to New Zealand quarantine activities. The advantages offered by polytomous over dichotomous keys are particularly relevant to ticks because most specimens submitted for identification are collected by nonspecialists and are often crushed, have missing structures, or are contaminated with debris.

Ticks are blood-feeding external parasites of birds, mammals, reptiles and amphibians that can act both as reservoirs and as vectors for a range of diseases caused by bacteria, viruses and rickettsiae (Varma 1993). New Zealand's tick fauna consists of eight hard tick (Acari: Ixodidae) species, all of which are formally described, and two soft tick (Acari: Argasidae) species, one of which is undescribed (Dumbleton 1953, 1963, Heath 1977). The hard tick Haemaphysalis longicornis Neumann is the only non-indigenous tick species in New Zealand and was likely introduced from Japan via Australia (Hoogstraal, et al. 1968). It feeds on a range of mammalian and avian species and is of economic importance, both in New Zealand (McKenna 1996) and elsewhere (Hoogstraalet al. 1968). It is also the only tick that is commonly encountered by humans in New Zealand (McKenna 1996). Four hard ticks (Ixodes auritulus Neumann, I. eudyptidis Maskell, I. kerguelenensis André & Colas-Belcour, and I. uriae White), and one soft tick (Ornithodoros capensis Neumann) are primarily parasites of seabirds (Dumbleton 1953, 1961, Roberts 1970). Ixodes eudyptidis is limited to coastal regions of New Zealand and southern Australia (Heath 1977, Roberts 1970), while the others have widespread distributions in the Northern and Southern Hemispheres (Amerson 1968, Arthur 1960a, Cooley & Kohls 1945, Dumbleton 1973, Heath 1977, Hoogstraal 1985, Roberts 1970). Of the remaining four species, I. anatis Chilton, I. jacksoni Hoogstraal and Amblyomma (formerly Aponomma) sphenodonti (Dumbleton) are endemic to New Zealand and are associated with kiwi, Apteryx spp. (Struthioniformes: Apterygidae), cormorant, Stictocarbo punctatus (Sparrman) (Pelecaniformes: Phalacrocoracidae) and tuatara, Sphenodon punctatus Gray (Rhynchocephalia: Sphenodontidae), respectively (Heath 1977). The fourth is an undescribed soft tick collected from the native bat Mystacina tuberculata Gray (Chiroptera: Mystacinidae) (Heath 1977).

Unlike many other countries, New Zealand is currently free of most ticks and tick-borne diseases that impact animal production and human health (Heath 2002a, b). The sole exception is Theileria orientalis, which is present in the North Island and is vectored by H. longicornis (Heath 2002b, James, et al. 1984). However, should additional tick species or tick-borne diseases become established in New Zealand, this situation could change. Between 1955 and 2009, the total border interceptions and post-border detections (hereafter referred to as “quarantine detections”) of exotic hard ticks made by New Zealand quarantine authorities was 122 (Heath 2001, pers. comm. 2009, Loth 2005), including 17 species from the genera Amblyomma, Bothriocroton, Dermacentor, Haemaphysalis, Ixodes and Rhipicephalus (Heath 2001, 2009, Loth 2005). To date, there have been no quarantine detections of exotic soft ticks in New Zealand (Heath 2001, pers. comm. 2009, Loth 2005). Approximately 40% of tick quarantine detections are from animals, typically dogs, imported into New Zealand (Loth 2005). A similar proportion is from humans and their clothing, whilst the remainder arrive in luggage, containers or via other routes not involving mammals (Loth 2005). Quarantine detections from humans, luggage and other objects are of particular concern as these all occur post-border and rely on the public bringing specimens to the attention of the appropriate authorities (Heath pers. comm. 2009).

In New Zealand, quarantine authorities, health professionals, members of the agricultural industries and ecologists currently have very limited tools for accurately identifying tick specimens and may resort to the dubious assumption that any specimen they encounter is H. longicornis. This situation has prompted the Ministry of Agriculture and Forestry Biosecurity New Zealand (MAFBNZ) to call for better tick taxonomic tools (MAFBNZ pers. comm. 2008). Developing electronic keys to ticks is also a top priority of the Quadrilateral Scientific Collaboration in Plant Biosecurity ( http://www.quadscoop.org/), involving quarantine agencies from Australia, Canada, New Zealand, and the United States.

This contribution describes two keys that enable New Zealand quarantine authorities, health professionals, members of the agricultural industries and ecologists to determine whether a tick specimen submitted for identification is a species already established in New Zealand or a potential exotic. I provide electronic and paper-based dichotomous and polytomous keys that enable non-experts to identify: i) nymphs and adults of the three genera and eight species of Ixodidae present in New Zealand, ii) adults of all 12 extant genera of Ixodidae in the world, and iii) nymphs of eight of the 12 extant genera of Ixodidae in the world. Genera excluded from the nymphal key were those containing species with very restricted geographical distributions and host ranges that are unlikely to be encountered during New Zealand quarantine activities.

Methods

The taxa included in my keys are listed in Table 1. Dichotomous (Table 2) and polytomous keys (Tables 3 & 4) to identify adults and nymphs of Ixodidae to genus were constructed while examining specimens of Amblyomma, Amblyomma (formerly Aponomma), Rhipicephalus (formerly Boophilus), Bothriocroton, Dermacentor, Ixodes, Hyalomma, and Rhipicephalus, and with reference to the following literature: Arthur 1960a, b, Arthur & Chaudhuri 1965, Barker & Murrell 2004, Beati et al. 2008, Belozerov et al. 2001, Guglielmone et al. 2009, 2010, Hoogstraal et al. 1970, Horak et al. 2002, Kaufman 1972, Keirans et al. 1994, Klompen et al. 2002, Matthysse & Colbo 1987, Nuttall & Warburton 1911, 1915, Roberts 1970, Sonenshine 1991, Varma 1993, Volzit 2002, Volzit & Keirans 2003, 2007, Walker et al. 2000 and Walker et al. 2003.

To develop the key to New Zealand species, I referred to previously published keys and taxonomic descriptions contained in Arthur (1963), Chilton (1904), Dumbleton (1943, 1953, 1958, 1961, 1963, 1973), Hoogstraal (1967), Hoogstraal et al. (1968), McKenna (1996) and Roberts (1970), and I examined representative male, female and nymphal specimens of all species in New Zealand except I. jacksoni. This allowed a range of discriminating features to be identified and included. The keys were digitised using Phoenix® and Lucid® software ( http://Lucidcentral.com).

The polytomous key was constructed using LucidBuilder software. LucidBuilder enables the user to develop a data matrix based on character states scored as being common, rare, uncertain (not known), common and misinterpreted, rare and misinterpreted, not scoped or absent. Character states are scored by placing a tick mark in the appropriate cell of the data matrix as illustrated in Table 5. It is also possible for the software to assign more than one state to an entity. For example, males of the genus Ixodes almost always have 7 plates on the ventral surface. However, males of I. jacksoniare an exception to this in that their ventral plates are obsolete (Hoogstraal et al. 1967). In this situation the character state can be scored as both commonly having 7 ventral plates and rarely with the ventral plates being obsolete (Table 5).

TABLE 1.

List of ixodid taxa included in the dichotomous and polytomous keys in Tables 2–4.

TABLE 2.

Dichotomous key to the ixodid genera and species known to occur in New Zealand.

continued

continued

continued

continued

Dependent and not scoped scoring is included in the data matrix to streamline or create a folding key. When a character state that has been scored using a positive dependency is selected, further questions are unfolded and presented to the user. In the polytomous key the following dependencies are used: character states scutum (present) and capitulum visible dorsally (yes) unfolds the character “number of legs”; character state number of legs (8) unfolds the character “size of scutum”; character state size of scutum (scutum covers the whole body) unfolds the eight characters (T-AA) associated with determining which genus a male ixodid belongs to; character state size of scutum (scutum covers 1/2–1/3 of body) unfolds the characters “porose areas” and “genital aperture”; character states porose areas (absent) and genital aperture (absent) unfold the five characters (G-K) associated with determining which genus an ixodid nymph belongs to; and the character states porose areas (present) and genital aperture (present) unfold the six characters (AL-AR) associated with determining which genus an ixodid female belongs to (Table 3). Not scoped scores were also used in the construction of the polytomous key. This type of scoring enables the coding of characters that are useful and applicable only for a subset of the entities in the key. Entities contained in the subset are coded for these characters in a normal fashion, but all the other entities are assigned the not scoped score for these characters. The characters associated with a subset of entities will only be presented to the user or unfolded when all other entities in the key have been eliminated. In the polytomous key outlined in this contribution, not scoped scoring is used to unfold the key and present the user with characters that discriminate between: the nymphs of the Ixodes species present in New Zealand (subset 1, characters L-S); the males of eyeless Amblyomma and Bothriocroton (subset 2, character AB); males of the Ixodes species present in New Zealand (subset 3, characters AC-AK); the females of eyeless Amblyomma and Bothriocroton (subset 4, character AR); the females of Rhipicentor andRhipicephalus (subset 5, character AS); and the females of the Ixodes species present in New Zealand (subset 6, characters AT-BA) (Table 3).

I conducted initial testing of the keys on adult and nymphal specimens of Amblyomma (eyed and eyeless), Bothriocroton, Dermacentor, Haemaphysalis, Hyalomma, Ixodes and Rhipicephalus using a blind testing protocol. The keys were then reviewed by a tick taxonomist, a quarantine diagnostician who had limited experience of tick taxonomy, and a laboratory technician who had no previous experience at identifying ticks.

TABLE 3.

Characters and states used in the polytomous key to ixodid genera and species known to occur in New Zealand.

continued

continued

continued

continued

Results and Discussion

I rectified an error in an earlier key whereby the second dichotomous pairing of Dumbleton's 1963 key incorrectly stated that female I. anatis has spurs on the first three coxae. In contrast, the specimens of I. anatis that I examined were consistent with earlier descriptions of female I. anatisbecause a single spur was present on the first coxa, while coxae II-III were without spurs (Chilton 1904, Dumbleton 1953).

Tables 2–4 present the dichotomous key (Table 2), and the features, states (Table 3) and data matrix (Table 4) used in the polytomous key (Table 4) to the ixodid taxa listed in Table 1. On paper, the dichotomous key (Table 2) is easier to use than the polytomous key (Tables 3 and 4), but the reverse is true when it is presented electronically. The electronic key to New Zealand Ixodidae is available directly from the author on a compact disk or from  http://keys.lucidcentral.org/keys/v3/hard_ticks/Ixodidaegenera.html. It contains the dichotomous and polytomous keys and has been designed to enable a nonexpert to identify New Zealand Ixodidae. Upon starting, users select either the dichotomous or the polytomous key. If the dichotomous option is selected, then Phoenix software (Norton 2003, CBIT) and the associated dichotomous key to New Zealand Ixodidae are initiated (Table 2). If the polytomous key is selected, characters and their associated states (Table 3) are presented to enable the specimen to be identified to family, life stage, genus and finally species. Users are not required to follow any particular sequence of questions within families, life stages, genera and species. In both the dichotomous and polytomous keys, all features referred to either in couplet questions or as character states are illustrated with line drawings, micrographs, or both. Once identification has been reached, users can access embedded fact sheets that contain detailed descriptions, images, or both for the male, female and nymph. Notes on the distribution and ecology of each species, and how to distinguish it from similar species are also provided. This additional information assists in confirming the identification.

During initial testing of both keys using species exotic to New Zealand, an experienced tick taxonomist correctly identified all 15 adult and 10 nymphal specimens to genus level. When presented with adults and nymphs of H. longicornis and four of the Ixodes species that are known to occur in New Zealand, all specimens again were correctly identified. A quarantine diagnostician correctly identified 10 adults and 10 nymphs to genus. An inexperienced technician was able to identify 15 of the 10 adults and 10 nymphs to genus. Both the quarantine diagnostician and inexperienced technician were able identify the nymphs and adults of H. longicornis. The latter result is important as this is the species that professionals in the health and agricultural industries in New Zealand are most likely to encounter. All the testers and in particular the quarantine diagnostician and inexperienced technician commented that the images and fact sheets contained within the key helped them to confidently identify the specimen they were examining. The inexperienced technician also commented that the key was interesting because until they were exposed to it they had always assumed that New Zealand had only one species of tick. However, all reviewers commented that they would like to see a glossary and front page included in the key. These are currently being developed and will be added to later versions.

TABLE 4A:

Polytomous key to nymphs of all Ixodidae genera and to the Ixodidae species known to occur in New Zealand.

TABLE 4B:

Polytomous key to males of all Ixodidae genera and to the Ixodidae species known to occur in New Zealand.

TABLE 4C:

Polytomous key to females of all Ixodidae genera and to the Ixodidae species known to occur in New Zealand.

TABLE 5.

LucidBuilder data matrix for polytomous key to genera of male Ixodidae.

QUADS ( http://www.quadscoop.org/) has indicated that, for optimal utility, electronic identification tools need to be extended to cover more than one faunal region. I am currently developing another electronic key that includes Australasian Ixodidae, Ixodidae that have previously been intercepted at New Zealand's border, those ixodid genera that contain a small number of species (eyeless Amblyomma (former Aponomma), Anomalohimalaya, Bothriocroton, Dermacentor, Margaropus, Rhipicephalus (former Boophilus), and Rhipicentor), the monotypic genera Cosmiomma and Nosomma, and the subgenera of Ixodes. In the future, there is scope to expand the key to include all known ixodid taxa plus genetic data such as nucleotide sequences and high resolution melt curves (Winder et al. in press).