The Southernmost region of Australia, the island of Tasmania, is also the most mountainous, with large areas of rugged alpine environments. This entomological frontier offers a distinct suite of insects for study including many endemic taxa. However, harsh weather, remote locations and rough terrain represent an environment too extreme for many existing insect trap designs. We report here on the design and efficacy of a new Alpine Malaise Trap (AMT), which can be readily hybridised with several other common insect trapping techniques. Advantages of the design include its light weight and portability, low cost, robustness, rapid deployment and long autonomous sampling period. The AMT was field tested in the Tasmanian highlands (AUST) in 2017. A total of 16 orders were collected. As expected, samples are dominated by Diptera. However, the trap also collected a range of flightless taxa including endemic and apterous species, Apteropanorpa tasmanica – closest relative of the boreal, snow scorpionflies (Boreidae). Combined and compared with other trap types the Alpine Malaise Traps captured less specimens but of a greater diversity than passive sticky traps, while drop traps captured less specimens but a greater diversity than AMT. The statistical potential of the catch is discussed.

Long term, flight intercept, sampling, invertebrate, Tasmania, Apteropanorpa

Most work on insect biodiversity ultimately relies on sampling populations in nature. The nature of the Tasmanian alpine environments is harsh. The Southernmost region of Australia, the island of Tasmania, is also the most mountainous with large areas of rugged alpine environments (Fig. 1). Low, wind pruned vegetation densely covers rocky, saturated soils. The winter season may see intermittent snow cover for several months, exposing plants to intense UV-B radiation in winter as well as summer. Among the multitude of different insect-trapping methods, few are well suited to the Tasmanian alpine environment.

Figure 1. 

Tasmania, Australia. Elevation Above Sea Level by https://maps.thelist.tas.gov.au/

A century ago René Malaise (1937) observed how efficiently his tent walls intercepted insects and funnelled them to the high points of the roof. His pioneering eponymous design for 'a new insect trap' was based on this observation. Malaise traps are still widely employed today. Malaise (1937) initially suggested the long-term, unmonitored operation of his trap tailored it for difficult to reach sites, like “high mountains”. The trap does boast the advantages of continuous autonomous operation; averaging out changing daily conditions and requiring no operator effort. However, early traps were large, gauzy constructions (Malaise 1937, Townes 1962); the adaptation published by Marston (1965) has more than 11 m of collecting face. As a result, classic malaise traps are actually ill suited to alpine sampling.

With such large collection faces, the collection chambers fill quickly. Rather than long term autonomous deployment, traps usually have to be emptied daily (Malaise 1937, Gressitt and Gressitt 1962, van Achterberg 2009, Russo et al. 2011, Diserud et al. 2013) though weekly, fortnightly (Clapperton 1999), and monthly (Doran 2003) are also reported. The fragile nature of the gauzy panes also makes them mismatched to the rough vegetation and wind exposure of Antarctic (Farrow and Greenslade 2013) and Tasmanian-highland sites (Hansen 1988, Doran 2003) although see Solem and Mendl (1989) and Finn and Poff (2008) for successful highland sampling elsewhere. In the advent of smaller traps based on the Malaise model, such as the SLAM and composite insect trap (Russo et al. 2011), we see devices which may be robust enough for alpine deployment. However other limitations persist.

Ethanol is a widely available and relatively harmless preservative now favoured for Malaise traps. However, evaporation puts a limitation on deployment time and any liquid component adds up to an intolerable weight when replication of samples is desired from a remote location (Russo et al. 2011). Dry killing agents, including cyanide-infused plaster, were used in the earliest traps (Malaise 1937), however, dry traps also require daily emptying to prevent dead and brittle specimens from being damaged by live ones (van Achterberg 2009). Propanol is substituted in remote traps for its slower evaporation rate (Farrow and Greenslade 2013), however, it does not address the issue of weight.

Combining trapping methods expands sampling parameters and improves catch (Moir et al. 2005). A single apparatus which combines sampling techniques takes less time to deploy and operate than multiple individual traps (Russo et al. 2011). These are desirable traits when sampling time is constricted by access time. However, the additional liquid preservative needed to operate pan traps or drop traps adds intolerable weight to sampling systems which must be carried any distance (Hansen 1988). A persistent conflict exists between accessibility and service requirements for traps in remote locations (Farrow and Greenslade 2013, Price and Baker 2016).

Intercept devices for sampling airborne insects in alpine habitats need to be: light weight, for on foot transportation to remote sites; robust against extremes of weather, especially high winds, ice and strong UV-B radiation; have long term capture capacity, while maintaining specimen quality at 'identification' standard; and collect effectively enough to generate at least semi-quantitative data useful for comparative purposes across a range of invertebrate orders. In the present paper, we assess the effectiveness in the alpine environment of a novel intercept trap that has these attributes, the Alpine Malaise Trap (AMT). We compare the catch of the AMT to those of both sticky traps and drop traps.

The AMT offers a number of advantages over existing designs, especially in relation to sampling in extreme habitats. Its lightness, inexpensiveness and the lack of a need to clear the trap on a daily or weekly basis, make it particularly suited to remote sampling sites (Table 3). The employment of a sticky plastic film in place of a liquid preservative considerably reduced the weight per trap, increasing portability, and eliminated the limitation of evaporation on operating time. Its small size meant a slower capture rate. The traps can be operated in an alpine environment for six weeks of summer without reaching capacity. They were also robust to several meters and months of snow and winds exceeding 100 kph (Bureau of Meteorology 2017). In spite of these extremes, the AMT preserved specimens representing a comprehensive cross section of the airborne alpine insect fauna including some unexpected apterous and brachypterous taxa. A compromise of this method is the quality of specimens recovered. In situ, the sticky gel can obscure and distort characters necessary for species level identification. In this project, specimens were easily assigned to family without further treatment; genus or species in the case of remarkable specimens. Tanglefoot can be dissolved and the specimens recovered into ethanol for taxonomic resolution where necessary. Miller et al. (1993) describe a citrus-oil solvent preferable to the traditional petrochemicals, though dry trapping methods are still recommended if specimen quality is critical.

Flies and wasps are attracted to white or yellow colours (classical Malaise traps are usually white). The transparency of our trap should partly eliminate this bias making for more representative samples. The use of transparent surfaces also allows our trap, when fitted with a drop capture array, to function like a classic window trap, capturing fliers strong enough to become unconscious upon impact with the panes (Hines and Heikkenen 1977, Wilkening et al. 1981). The bias in catches towards flies was expected, as Diptera are often dominant elements of the fauna in highland areas (Levesque and Burger 1982, Shaw and Taylor 1986, Konno 2006).

Our catch is largely congruent with that of Doran (2003) who extracted 85% Diptera and 6% Hymenoptera with classic malaise traps from the Warra LTER research site in alpine Tasmania. However, the inclusion of Mecoptera and Orthoptera in our AMT samples diverges from the Warra malaise samples and is more similar to the Warra pitfall samples (Bashford et al. 2001, Doran 2003). The apterous element in the AMT samples suggests that wind is augmenting the trap's flight intercept catch with otherwise sedentary or ambulatory taxa, resulting in samples that fall in-between classic malaise and pitfall traps. If this means that the AMT samples are more location specific, the trap may be more sensitive to differences in environment and treatment. Large mobile insects have already proven a poor indicator of environmental differences at Tasmanian (Driessen and Kirkpatrick 2017) and other sites (Polchaninova et al. 2016, Lazarina et al. 2017).

The most notable captures of flightless taxa were the apterous alpine scorpionfly, genus Apteropanorpa, an endemic family of four species similar in appearance to Northern Hemisphere snow scorpionflies (Boreidae). It was first identified by Carpenter in 1940 and formerly presumed rare. Recent reviews identified the new species A. evansi, A. warra and A. hartzi; highlighting the potential for more discoveries (Byers and Yeates 1999, Palmer et al. 2007). Our survey on Mount Field in late summer captured only 39 specimens, however a review of the unprocessed autumn-winter samples reveal 335 more, predominantly (17:1) captured by the AMT. Trial traps on kunanyi/Mount Wellington captured almost exclusively A. tasmanica. Thousands of Apteropanorpidae were captured in pitfall traps at the Warra Long-Term Ecological Research site (Doran 2003). Pitfall traps were decided against for our study due to dense vegetation on thin, rocky, saturated soil. As the AMT is sampling a cross section of flying and pit-fall taxa it may be a useful alternative to pitfalls at other difficult sites.

Despite being considerably shorter than classic malaise traps, 30 cm high intercept faces fit precisely within the 'boundary layer – allowing independent insect flight' as hypothesised and tested by Taylor (1974). While this height was established over grass, and the true boundary layer at our site may be impacted by the height of vegetation, the addition of passive sticky traps projecting above 30 cm, into the 'free air' (Taylor 1974) helps address this potential short fall. Conversely, the extreme wind conditions at our site will have increased turbulence and decreased the boundary layer at times. The same conditions that dictated the small size of the traps compensates for the potential loss of invertebrate catch. Indeed, while flightless species may climb into the traps from contact points with the ground, their even positioning on the sticky sample surface (Fig. 5) suggests a passive carriage to the traps on strong winds. In the case of A. tasmanica, such transport constitutes a significant contribution to sampling.

Combining trapping methods is a proven way to counter the limitations of particular trap types and improve sample yield (e.g. Querner and Bruckner 2010). Adding compatible devices to create a single trapping station has been found to reduce the cost and time of using multiple individual traps (Campos et al. 2000, Russo et al. 2011). Similar trap designs to ours include the Composite insect trap (Russo et al. 2011), and the Sea Land and Air Malaise (SLAM) trap.

As the sticky CD traps alone would constitute a robust, cheaper and lighter alpine sampling technique, we were interested to compare the sampling strengths of each. While CD traps captured significantly more specimens overall this is obviously tied to the capture of nearly 2,000 more Diptera. Otherwise the traps are either comparably effective or the AMT captured significantly more specimens of a given order. The AMT did not catch as many taxa or specimen as CD traps, however it does deliver a more taxonomically balanced sample. The CD trap catches were dominated by Diptera and Hymenoptera. While still the top two orders sampled by AMT, dominance of Diptera and Hymenoptera was balanced by higher counts of other taxa. Similarly, samples from the small drop trap trial were both dominated by Diptera followed by Hemiptera. While drop traps captured significantly less specimens overall, a greater diversity of orders contributed to the total catch. As predicted by the literature, the drop capture array significantly increased the capture of beetles (Russo et al. 2011) as well as other apterous or cursorial species like spiders and flightless Tasmanian alpine grasshoppers (Russalpia spp.), particularly larger bodied specimens.

The success of the Alpine Malaise Trap is illustrated by our ability to deploy 34 replicates in rough terrain, 1.5 hours hike from vehicle access, with three people in 9 hours. The traps were able to operate continuously and autonomously for 6 weeks in summer, collecting 8,029 readily identifiable invertebrates (Basic AMT) and a further 7,155 from the hybrid CD attachment and 1,229 from hybrid drop capture array. Further, the traps were robust to extremes of wind, rain, snow and UVB. The invertebrate profile of samples is an intermediate of classic malaise and pitfall traps. The environmental sensitivity this conveys over standard malaise traps is being investigated further.

We thank Larissa Giddings, Julie Gryphon and Emalisa White for outstanding assistance in the field, the Tasmanian Parks and Wildlife Service for permit FA17335 to conduct research in Tasmanian Reserves and the cooperation of the Mount Field National Park staff.