Distribution Maps

Specimens on loan from 11 natural history collections (see Acknowledgments) were examined and databased using the online PBI Plant Bug database ( https://research.amnh.org/pbi/locality/). The databased samples comprised 1,613 specimens of Z. tetracanthus and 1,745 specimens of Z. renardii. Specimen data is publicly available through Discover Life ( http://www.discoverlife.org/). Due to the geographic spread and number of records, these data provide the most comprehensive documentation of the distribution ranges of both species published to date. Many original locality labels lacked latitude and longitude information. These localities were georeferenced (i.e., latitude and longitude coordinates determined) using Google Earth, the USGS Geographic Names Information System (GNIS), or an online georeferencing gazetteer ( http://www.fallingrain.com/world). The specimen data were mapped using the mapping feature of the AMNH PBI database ( http://research.amnh.org/pbi/maps/). Distributions in North and Central America for both species are shown in Fig. 1. We used a map based on the Köppen climate classification to document climate zones for the plotted native range of Z. renardii and for the recorded extralimital distributions (Supplementary Fig. 1).

Fig. 1.

Distribution of Zelus renardii and Z. tetracanthus in mainland North and Central America. The range of Z. tetracanthus extends to the east coast of North America, whereas Z. renardii is restricted to roughly the western half of the continent.

Documentation of Habitats for Z. renardii and Z. tetracanthus in Riverside County (California, USA) and Baja California (Mexico)

Collecting dates and elevations of collecting sites were summarized based on specimen records in the database. Habitats were described and photo-documented for several collection events of Z. renardii and Z. tetracanthus in Riverside County, California (2007–2010) and Baja California, Mexico (May 2008).

Behavioral Studies

Laboratory cultures were established from adult males and females collected in Riverside County (California) for Z. renardii and Madera Canyon (Pima County, Arizona; N 31.731 W 110.879) for Z. tetracanthus. Specimens were kept individually in petri dishes at constant room temperature (25 °C, 28% RH) and 12:12 h L:D (8 am-8 pm) under fluorescent lights and fed 10 adult D. melanogaster Meigen every 2 d. All petri dishes used in this study were Ø 90mm. To document predatory behaviors, an unfed assassin bug that had been starved for 3 d was placed under a Canon XL2 3CCD Mini DV video camcorder (Canon USA) in a petri dish, 10 adult Drosophila melanogaster were introduced simultaneously into the dish, and behaviors of the Zelus specimens were recorded on digital videocassettes (Mini DV; DVM60 ME). The cassettes were transferred to Windows Movie Maker (version 5.1; Microsoft) for analysis of behavioral sequences and to quantify the 1) time elapsed between introduction of the Drosophila into the petri dish and start of feeding (defined by the insertion of stylets into the prey item) and 2) feeding duration (time between insertion of stylets and dropping of prey item). Seven predation events for 7 different specimens were recorded for both species (Table 1).

To record mating behavior, one virgin male and one virgin female (aged 3–7 d after the final molt to assure sexual maturity) were placed together in a new petri dish and behaviors were recorded using the Canon XL2 system. Similar to the approach above, the video tapes were transferred to Windows Movie Maker for analysis of behavioral sequences and to quantify the 1) time elapsed between the introduction of pairs in the petri dish and start of copulation and 2) time elapsed from start (defined by aedeagus inserted) to end of copulation (male genitalia retracted). Twelve matings were recorded for Z. renardii and 15 for Z. tetracanthus (Table 2).

TABLE 1.

DURATION OF PREDATORY BEHAVIORS (TO THE NEAREST MINUTE) FOR 2 SPECIES OF ZELUS (N = 14).

A Student's t-test (performed at  http://vassarstats.net/) was used, assuming both equal and unequal variances, to determine if differences in feeding and mating times between the 2 species were significant. Statistical significance was always obtained whether assuming equal or unequal variances. We only report p-values assuming equal variances.

Each female was mated only once, the male removed after copulation, and the female was subsequently kept in a separate petri dish under the same environmental conditions as described above. The first deposited egg batch of each female was kept with the female and observed for emergence for 15 d or until immatures hatched (usually 8–12 d after deposition; see also Swadner & Yonke [1973]). We recorded the number of immatures that hatched from these egg batches.

Distribution Ranges of Z. renardii and Z. tetracanthus and Climatic Conditions in Areas of Introduction

Figure 1 shows what we consider to be the original distribution range of Z. renardii and Z. tetracanthus in North and Central America based on 3,358 specimen records. The 2 species are sympatric in, among other areas, the Western USA, including California and Baja California, Mexico, but also throughout part of Central America. The range of Z. tetracanthus extends to the east coast of North America, whereas Z. renardii is restricted to roughly the western half of the continent. In addition, Z. tetracanthus extends further north into Canada than Z. renardii. The ranges of both species extend over a number of different climatic zones, indicating that both species are capable of dealing with diverse abiotic environmental conditions. The simplified climatic world map combined with the known distribution of Z. renardii demonstrates that 3 of the areas where this species has been introduced are Mediterranean type climates (Chile, Greece, Spain), whereas Hawaii, Johnston Island and Samoa are wet tropical (Supplementary Fig. 1).

TABLE 2.

DURATION OF MATING BEHAVIORS (TO THE NEAREST MINUTE) FOR ZELUS RENARDII (N = 12) AND Z. TETRACANTHUS (N = 15).

Habitats, Microhabitats, and Collecting Dates of Z. renardii and Z. tetracanthus.

The 2 species are sympatric over large parts of their ranges including the western USA and parts of Central America. Zelus renardii is found from slightly asl (above sea level; 8 m in Baja California, Mexico) to more than 2,000 m asl (Durango, Mexico). Adults have been collected every mo of the yr, but records between Nov and Feb are confined to Hawaii, Texas, and Mexico. Similarly, Z. tetracanthus has been recorded from localities at sea level (Baja California, Mexico) to 2500 m asl in the Chiricahua Mountains in Arizona. Collecting records indicate that the majority of North American specimens were collected between Apr and Oct and only 5 specimens in Dec (Arizona, California, Texas, and Baja California in Mexico).

The 2 species are syntopic at some localities. Supplementary Fig. 2 shows a habitat in Baja California where both species were very abundant on blooming Prosopis sp. (Fabales:Fabaceae)that attracted large numbers of insects, including plant bugs and bees. However, based on our collecting in Baja California and Riverside County, only 1 of the 2 species tended to be abundant in certain habitats. Supplementary Fig. 2B shows chaparral in the San Pedro Mártir mountains (Mexico, Baja California) at an elevation of ∼1,100 m. Zelus tetracanthus was relatively common in this habitat on various species of shrubs and trees, but no Z. renardii were collected during the 2 collecting d one of the authors spent in this area. In contrast, numerous specimens of Z. renardii were collected by a team of 5 insect collectors sweeping vegetation, including blooming Chilopsis sp. (Bignoniaceae) in White Water Canyon in Riverside County, California (Supplementary Fig. 2C, D), but no Z. tetracanthus were netted. Zelus renardii is also very common in suburban yards in the Riverside city area (Supplementary Fig. 2E), but Z. tetracanthus is rarely found in developed areas. In suburban environments, Z. renardii is frequently found on native and non-native herbaceous and woody plants, including Ceanothus sp. (Rosales: Rhamnaceae), Asclepias spp. (Gentianales: Apocynaceae), and vegetables, such as tomato (Solanum lycopersicum L.; Solanales: Solanaceae) and pepper (Capsicum spp.; Solanales: Solanaceae). In addition to the adults, egg masses of Z. renardii were also frequently found attached to garden plants.

Observations on Predatory and Reproductive Behaviors of Z. renardii and Z. tetracanthus

Distribution Range

We proposed in the introduction that due to the large distribution range of Z. renardii across North and Central America, different populations may be pre-adapted to various climates, thus facilitating establishment under different climatic conditions; this ability may be absent in other Zelus species. We found that the distribution ranges of Z. renardii and Z. tetracanthus are broadly overlapping (Fig. 1); both species have native populations in tropical climates (large parts of Central America), dry, semi-arid and arid climates (Arizona and New Mexico in the USA), and Mediterranean climates (California, USA, and Baja California, Mexico). Based on these distribution ranges, populations of both species would be expected to have adapted to diverse climatic conditions.

Fig. 2.

Predatory and reproductive behaviors of Zelus renardii (A, C, E) and Z. tetracanthus (B, D, F). (A, B) Prey handling by Zelus spp. (C, D) pre-copulatory position in both species with the male assuming the dorsal riding position on top of the female. (E, F) Copulation in the 2 Zelus spp.

The origins of the Hawaiian, Chilean, and European Z. renardii from populations in the native New World range are unknown; we also lack data that would indicate if the Greek and Spanish populations are independent invasions or based on a single introduction to Europe. We speculate that the populations that became established in tropical areas in the Pacific region may have originated from Central America, and the Chilean and European populations may be derived from Mediterranean climate adapted populations, e.g., in California. This hypothesis could be tested in the future using molecular markers and phylogeographic methods.

Habitats and Microhabitats

Both species of Zelus are usually caught while beating and sweeping flowering and non-flowering vegetation. A pattern of host-plant preference is absent and both assassin bug species often seem to target flowering vegetation that attracts a diversity of potential prey items. Both species occupy various habitat types between the sea level and an altitude of up to 2,000–2,500 m. Zelus tetracanthus is more frequently confined to natural habitats that are relatively undisturbed by human activity. Zelus renardii occurs on a regular basis in agricultural settings and even urban and suburban environments, such as parks or other public spaces (i.e., our University campus). We have not surveyed plant nurseries, but would expect that Z. renardii is also common in such environments. Zelus females readily glue their egg batches on plants (usually on the ventral surface of the leaves) or other objects (i.e., tissue paper in the lab). These batches are typically firmly attached to the plant surface, eggs take 8–12 d to hatch, and immatures usually do not tend to disperse far from their egg batch during the first and second instars (C. Weirauch and G. Zhang, unpublished data.). We therefore propose that egg batches of Z. renardii may be shipped with nursery stock, presenting a possible avenue for their introduction to South America and Europe. Our observations indicate further that Z. renardii is more likely to be dispersed by humans than Z. tetracanthus, given the abundance of the former species in human-disturbed environments.

As indicated above, Z. renardii could be considered a pioneer species, and we suggest it might also be more tolerant to adverse conditions, such as high temperatures and scarce food supply, compared with its congeners. We do not have experimental data to support this argument, but our experience with collecting live specimens in the field and bringing them back to the lab to establish cultures provide some circumstantial evidence. Wild-caught Z. renardii specimens survive the transfer to the laboratory even under hot conditions, establish colonies easily, and even irregular food supply usually does not impact their reproductive success. This contrasts with Z. tetracanthus, which easily overheat and die during transport and appear to depend on more regular food supplies. The physiological properties of the 2 species warrant further study to evaluate how these factors may affect the survival rate of specimens during dispersal.

Predatory Behavior

Zelus renardii and Z. tetracanthus are of roughly the same total length (both of 12.5 mm, Hart 1986) and Drosophila flies appear to be suitably sized prey items for both species. Zelus renardii caught prey much faster upon release in the rearing container than Z. tetracanthus; Z. renardii also consumed drosophilid flies faster than its congener. Prey capture in the wild usually occurs on vegetation, and frequently, the assassin bugs remain rather exposed during capture and feeding (C. Weirauch, unpublished data). As a result, they may be more vulnerable to predation by predators such as birds. We hypothesize that under this scenario, the shorter predation and feeding time in Z. renardii might present an ecological advantage that reduces the risk of being preyed upon; and thus may contribute to the high abundance of this species in many areas.

Reproductive Behavior

The drastically different duration of pre-copulatory behaviors in the 2 Zelus species warrants further study. Zelus renardii engaged in copulation much more quickly than Z. tetracanthus. The male in the former species appeared rather “aggressive” in its approach to the female and copulation was achieved within 5 to 18 min, on average after 12 min, after the virgin pairs are assembled. Pre-copulatory behaviors in Z. tetracanthus ranged from 11 to 39 min and copulation began on average 29 min after introduction into the experimental chamber; the approach of the male is distinctly less aggressive than in Z. renardii. In this study, we did not evaluate details of the pre-copulatory behaviors beyond the general sequence of events that were similar in the 2 species (antennal movements of both sexes; approach of the male; male assumes dorsal position; tapping of female head and pronotum; genitalic extension). Given that prolonged pre-copulatory behaviors are often associated with chemical communication, additional studies should investigate a potential role of such compounds in the mating behavior of Z. tetracanthus. Edde & Phillips (2006) reported that males of Z. tetracanthus are attracted to bostrichid pheromone traps in the field. The bostrichid pheromone may act as a kairomone or aggregation pheromone, but it may also indicate that chemical communication could play a role in Z. tetracanthus mating behavior.

Another unexpected result was the dramatically different percentage of egg batches with viable first instars: 67% for Z. renardii and only 29% for Z. tetracanthus. These rates are based on egg batches derived from each of the females in this study, all of which were mated only once. We realize that our sample size is too small (n = 12 for Z. renardii and n = 15 for Z. tetracanthus) to draw far-reaching conclusions. We here propose 2 hypotheses as possible explanations for this greatly different number of first instars in the 2 species, which may deserve future examination. Zelus tetracanthus females, but to a lesser degree females of Z. renardii, may require multiple matings to ensure fertile egg batches. This hypothesis can be tested with additional mating experiments. Secondly, and assuming that the lab conditions under which the 2 species were kept were not optimal, we propose that egg batches of Z. renardii may have higher tolerance to adverse environmental conditions (temperature, humidity), thus resulting in higher hatching rates. Such tolerance could be another important factor in allowing Z. renardii to be successful in pioneer situations and to quickly establish relatively high population densities.

In summary, Z. renardii and Z. tetracanthus show overlap in their distribution ranges and are exposed to the same range of climatic conditions throughout the Western United States; however, Z. tetracanthus occurs under a broader range of climatic zones than Z. renardii. The distribution range alone can therefore not explain the differential dispersal observed for the 2 species. Zelus renardii appears to be more abundant in man-made environments including gardens and parks; this association may facilitate dispersal by humans, i.e. with nursery stock. A comparative study of predatory behaviors indicated that Z. renardii spent less time preying, thus potentially evading vertebrate predators more readily than Z. tetracanthus. Comparisons of reproductive behaviors of the 2 species revealed that Z. tetracanthus was involved in more time-consuming pre-copulatory behaviors, indicating the potential role of chemical communications. Zelus renardii may be able to establish large populations faster than Z. tetracanthus, given the higher rate of viable egg batches produced. From our observations and experiments, we conclude that the association with humans that is observed in Z. renardii alone, but not Z. tetracanthus, may be the reason for the dispersal of this species; biological parameters including predation efficiency and higher rate of offspring per mating may further facilitate the establishment of this species outside its native range.