Techniques for sampling Auchenorrhyncha in grasslands

22 Pages·2002·2.3 MB·English

by  Stewart, Alan J A

© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
Techniques for sampling Auchenorrhyncha
in grasslands
ALAN J.A. STEWART
Abstract
The relative merits of different techni- They can be used to give estimates of
ques for sampling the Auchenorrhyncha absolute population density in grasslands
community in grasslands are reviewed. As and tend to sample the epigeal species bet-
is the case when studying many other ter than sweep nets. However, a true pic-
invertebrate groups, no single technique ture of the species living in the lowest
can be relied upon to reveal the full range vegetation stratum or close to the soil sur-
of species at a site or provide unbiased face can best be obtained by using pitfall
estimates of population density for all spe- traps. A comprehensive inventory of spe-
cies. Nevertheless, with moderate effort cies would therefore need to combine pit-
and inexpensive equipment and if due fall trapping with either sweep netting or
attention is paid to the importance of stan- suction sampling. Brief discussions are
dardising sampling procedures to allow also presented of techniques for sampling
comparisons across both space and time, the aerial fauna and for estimating disper-
reliable estimates of both relative and sal and movement between populations.
absolute population density can be achie-
ved. Sweep-netting is an inexpensive and
Key words: Auchenorrhyncha, sam-
simple method for providing relative esti-
pling, grasslands, sweep-netting, suction
mates of population density but it is hard
samplers, pitfall traps, population estima-
to standardise and it under-samples the
tes, standardization.
epigeal species. The conventional D-Vac
suction sampler has now been largely
superseded by a variety of smaller and
cheaper hand-held suction devices that
have been developed by modifying devices Denisia 04,
that are sold for collecting garden refuse. zugleich Kataloge des OÖ. Landesmuseums,
Neue Folge Nr. 176 (2002), 491-512
491
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
Introduction red to closely grazed or regularly mown gras-
slands where the vegetation remains generally
The Auchenorrhyncha form an important short (MORRIS 2000). For this reason, the
component of the invertebrate fauna of most Auchenorrhyncha community is a good
temperate grasslands. The group has a number reflection of the intensity of grassland mana-
of properties which make it very suitable for gement, responding both rapidly and precisely
monitoring the biotic conditions and assessing to any changes in the management regime
the conservation status of a range of grassland such as the imposition or cessation of grazing
types (HILDEBRANDT & NICKEL in press): (MORRIS 1981a, 1981b; MORRIS &. PLANT
(i) Population densities in grasslands often 1983).
exceed those of other key invertebrate taxo- (v) At a more practical level, there are
nomic groups such as the Heteroptera and well tested and widely accepted techniques for
Coleoptera and can reach remarkably high sampling the Auchenorrhyncha community
levels (in excess of 1,000 individuals per m2; in grasslands, although there is no technique
WALOFF 1980). Potential species richness is that will suit all circumstances, nor one that is
high enough to be a useful ecological indica- devoid of sampling bias. Nevertheless, reliable
tor, with individual grassland sites often sup- population estimates can be generated using
porting in excess of 40 species (MORRIS 1971). simple, inexpensive and portable equipment.
(ii) The autecology of many grassland In general, at least in Europe, the taxonomy of
Auchenorrhyncha is well studied in terms of the group is well documented, stable and sup-
host plants, habitat associations and responses ported by high quality identification keys.
to management. It is therefore possible to pro- When embarking upon a programme of
vide a detailed ecological interpretation of the research on the ecology of grassland Auche-
community from data on species occurrence or norrhyncha, one of the first questions to be
relative abundance. answered will be: what sampling technique
(iii) The Auchenorrhyncha perform an should be used? This deceptively simple que-
important functional role in the grassland stion turns out to be remarkably complex to
community as herbivores, by tapping into the answer, as the rest of this paper will show.
phloem or xylem sap or extracting the con- Field entomologists have shown considerable
tents of the mesophyll cells of their host ingenuity in developing a wide variety of
plants. When population densities are high, collection and sampling methods, responding
this action induces a substantial photosynthe- to the considerable range of habitat associa-
tic drain on the plants and may influence the tions and behaviour patterns exhibited by this
outcome of competition between plant species group of insects. Most field research program-
and hence the course of succession (BROWN et mes will initially be concerned with tackling
al. 1988). The transmission of plant pathogens two primary questions: what species occur in a
by many Auchenorrhyncha may compound habitat and at what densities? Faced with this
such effects. Removal of plant sap followed by challenge, selection of the most appropriate
excretion of soluble waste material by these sampling technique is not always straightfor-
insects will undoubtedly have significant ward. There is no universally applicable sam-
effects on nutrient cycling within the gras- pling technique that will suit all purposes. Fur-
sland ecosystem, although this effect has thermore, all techniques are selective to some
received rather less attention than it deserves. degree, the extent of the bias being dependent
on a number of factors. Reliable estimation of
(iv) The structure and composition of
population density may require employing
grassland Auchenorrhyncha communities
several techniques in combination.
generally reflect a combination of the species
composition and physical structure of the The researcher who is faced with deciding
vegetation (BROWN et al. 1992). Auchenorr- which sampling techniques to employ has to
hynchan species richness tends to be much take a number of considerations into account.
greater in undisturbed and lightly grazed gras- The most important of these is sampling effi-
slands where the vegetation is taller, compa- ciency: the effectiveness with which the tech-
492
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
nique reveals all the individuals in the target ring on which sampling techniques are most
area. This will vary between species, habitats, appropriate in different circumstances. These
seasons, environmental conditions and, to a include:
lesser extent, between field workers. Cost is i) Vertical stratification of species:
always a major consideration. It includes not Many species select particular strata wit-
just the cost of the materials and the time nee- hin the vertical structure of grassland vegeta-
ded to sample, but also the time required to tion (DENNO 1980; DENNO et al. 1980). This
process any material collected. The latter is demonstrated particularly clearly by the use
includes separation of the insect specimens of different techniques (e.g. pitfall trapping,
from plant material and other debris inadver- suction sampling and sweep netting) that sam-
tently collected during the sampling process, ple different subsets of the total fauna (AND-
identification of specimens and curation of RZEJEWSKA 1965; PAYNE 1981; PETER 1981;
the samples or selected voucher material for TÖRMÄLÄ 1982; NOVOTNY 1992; CHERRILL &
future reference. The time required to do this SANDERSON 1994). Furthermore, there is evi-
post-sampling work is often under-estimated. dence that this stratification changes seaso-
For quantitative studies, it must be possible to nally (ANDRZEJEWSKA 1965) and diurnally
standardise the technique, so that it can be (ROMNEY 1945).
replicated with confidence over both space
ii) Sexual differences in activity patterns:
and time. Not all techniques lend themselves These may result in biased sex ratios in the
to this. Other considerations such as the amo- catches produced by particular techniques. For
unt of training required and the extent to example, pitfall trap catches of certain species
which sampling is weather-dependent may tend to be dominated by males (LEQUESNE &
also have to be built into the selection pro-
MORRIS 1971; PAYNE 1981; TÖRMÄLÄ 1982),
cess. Generally, the final choice is based on a possibly because the females are more seden-
balance of the appropriateness of the techni- tary.
que measured against time and cost.
iii) Differences in sampling efficiency bet-
Programmes for sampling Auchenorrhyn- ween life history stages:
cha in grassland must be guided by these gene- The same technique may not be appro-
ral principles. Here, I review the techniques priate for all life history stages, even within
that have been developed or adapted to sam- the same species. This may be because, for
ple these insects in the grassland habitat and example, nymphs and adults inhabit different
comment on their effectiveness. The review is parts of the host plant (overwintering nymphs
concerned solely with the efficacy of the tech- of certain species typically reside close to the
niques themselves; issues relating to the stati- soil surface) or have different susceptibilities
stical design of field sampling programmes to being caught. SlMONET et al. (1979) con-
(HURLBERT 1984; EBERHARDT & THOMAS cluded that the D-Vac was the most appro-
1991; DUTILLEUL 1993) and subsequent analy- priate technique for sampling adults of the
sis of the data (see summary of key references potato leafhopper, Empoasca fabae; however,
in POTVIN & TRAVIS 1993) have been covered nymphs were more efficiently extracted by
in general terms elsewhere. SOUTHWOOD placing excised branches of the plant for 24 hr
(1978), DENT (1991) and KUNO (1991) consi- in containers with small Dichlorvos squares
der the application of general principles speci- (SlMONET et al. 1978).
fically to sampling insect populations. Techni-
iv) Diurnal changes in behaviour:
ques for sampling the predators and parasitoids
Whilst diurnal periodicities in leafhopper
of Auchenorrhyncha (many of which are the
flight activity have been known for some time
same as for their hosts) are not dealt with
(LEWIS & TAYLOR 1965), rather little is known
here, but are covered fully by POWELL et al.
about whether analogous changes occur in
(1996).
other behavioural traits. DONDALE et al.
As is the case in other invertebrate groups, (1972) and SCHAEFER (1973) report diurnal
certain features of the ecology and behaviour changes in pitfall trap catches, whilst PAYNE
of Auchenorrhyncha have an important bea- (1981) has suggested that these changes differ
493
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
between the sexes. ROMNEY (1945) reports BUNTIN 1988). Where the use of more than
diurnal changes in vertical positioning of one sampling technique is unavoidable, an
Eutetrix teneüus on beet plants, but this effect attempt should be made to calibrate between
has not been widely investigated in other spe- the results (e.g. CHERRY et al. 1977; SlMONET
cies. In the light of these possible effects, sam- et al. 1978, 1979; TöRMÄLÄ 1982; BUNTIN
pling programmes that involve any sort of eva- 1988).
luation of treatment effects (e.g. different gras- It is important to distinguish at the outset
sland management regimes) should stipulate between different types of field work. 'Collec-
that samples are taken at similar times of day ting', for species inventory work or to obtain
to ensure comparability of catches, material, perhaps for experimentation or a
v) Spatial distribution patterns: taxonomic investigation, is an essentially
The number and location of replicate sam- non-quantitative exercise; there is no particu-
ples within a site or experimental plot need to lar interest in determining population size. On
be chosen bearing in mind the importance of the other hand, true 'sampling' has the speci-
spatial factors in determining abundance. In fic objective of providing an unbiased estima-
addition to the possibility of direct spatial te of population density and is by definition
autocorrelation (the phenomenon where the quantitative. This paper is concerned primari-
similarity between samples is related, positive- ly with techniques to achieve the latter objec-
ly or negatively, to their physical distance tive. Of course, most of the techniques used
apart (LEGENDRE 1993)), insect species occur- for quantitative sampling can also be used for
rence and/or abundance may be highly corre- general collecting.
lated with one or more key environmental fac-
tors (such as soil conditions, aspect or vegeta-
Absolute versus relative populati-
tion composition) which are themselves spati-
on size estimates
ally correlated. Rather few studies on Auche-
norrhyncha have tested for such autocorrelati- In quantitative sampling, there is an
on (SANDERSON et al. 1995) or attempted to important distinction between estimates of
quantify spatial variation in general terms and absolute as opposed to relative population size
how this changes temporally (GYÖRFFY & or density (SOUTHWOOD 1978; DENT 1991).
KARSAI 1991). An estimate of absolute population density is
a count of the numbers of individuals within a
A comparison of the relative efficiencies
specified area. As it is an estimate of the actual
of the different available sampling techniques
density, it should be comparable both spatial-
should be done as part of the preparatory work
ly and temporally (i.e. with estimates derived
for any field study. Failure to do this may result
from other sites or on other dates). One
in erroneous conclusions based on inappro-
should realise however that techniques desi-
priate comparisons, for example by comparing
gned to estimate absolute population density,
results from the same technique in different
whether by visual searching or some sort of
habitats or in different environmental conditi-
extraction technique, rarely detect 100% of
ons. Such considerations are particularly rele-
the insects actually present; in fact, extraction
vant in community studies, where apparent
efficiencies are frequently much less than this.
differences in the relative abundance of spe-
As the resultant count will therefore be an
cies may simply reflect differences in sampling
under-estimate of the true population density,
efficiency. Few community studies address this
caution should be exercised when extrapola-
problem, despite TöRMÄLÄ 's (1982) warning
ting from small samples to produce population
that different techniques for sampling grass-
estimates for large areas, as the under-estima-
land faunas produce very different results.
tes then become greatly magnified.
Comparative studies should consider not just
sampling efficiency (the number of insects If areal densities are either inappropriate
extracted per unit area) but also the relative or impracticable, the next best estimate of
precision of each technique, measured as the absolute density expresses the population
variability amongst replicated samples (e.g. count in units of habitat; for the Auchenorr-
494
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
hyncha, this is most appropriately some com- with total counts adjusted accordingly if they
ponent of the host plant (e.g. a count of num- differ consistently.
bers per leaf, unit leaf area, stem length or Trapping techniques exploit the fact that
whole plant). These 'habitat units' will chan- most insects move through their habitat. A
ge as the plant grows, so a measure of the num- simple distinction can be made between
ber of habitat units per unit area is also needed 'interception traps' that collect insects moving
before a true population density estimate can through the habitat as part of their normal
be derived. behaviour and 'attraction traps' that provide a
When it is impossible to estimate absolute stimulus which draws the insects towards the
densities, the field researcher must resort to sampling point. Such techniques are usually
relative population estimates. Here, the esti- highly cost-effective as they are generally less
mate is no longer a true count of numbers in a time-consuming and require less skill than
given area, as the unit of measurement is active sampling techniques (LOTT & EYRE
usually unknown. Data from traps generate 1996). They also have the advantage of sam-
this type of relative estimate, as it is impossible pling continuously over an extended period,
to be certain about the absolute area or volu- including night as well as daytime. Due to the
me over which the trap is operating. If exter- length of operating time, often several days,
nal conditions (weather, habitat structure the influence of short-term fluctuations in
etc.) are similar, estimates using the same sam- weather are evened out. 'Instantaneous' sam-
pling technique should be broadly comparable pling techniques, by comparison, are always
across space and time. Active sampling tech- subject to the influence of time of day, weat-
niques such as sweep-netting can be standar- her conditions and other short-term factors.
dised by expressing samples in numbers caught Measured against this, traps left untended are
per unit of effort (usually sampling time or the vulnerable to adverse weather, human vanda-
number of sweeps), but can not readily be lism and damage by other animals (e.g. grazing
expressed directly in terms of densities. stock). The single greatest disadvantage,
however, of most trapping techniques is that
The distinction between absolute and
the resultant catch is strongly influenced by
relative population sampling techniques is not
the activity of the insects themselves. Seden-
always clear cut. Relative estimates can some-
tary species will be caught less frequently than
times be converted to absolute densities, if a
highly active ones, even if their actual popula-
good correlation can be demonstrated bet-
tion densities are similar. Consequently, data
ween counts from the technique and those
from attraction or interception traps should
from another more accurate estimate of abso-
not be analysed quantitatively until the relati-
lute density. However, the calibration is likely
onship between catch size and population
to be both species- and site-specific and assu-
density has been checked. This relationship
mes that various extraneous environmental
will undoubtedly vary between species and, for
factors are kept constant. Wherever possible,
any one species, between sexes, seasons or dif-
attempts should be made to quantify the effi-
ferent habitats.
ciency of the technique being employed. This
can be done by comparing the sample count In summary, in order to generate the most
with the number of insects added after a com- accurate population estimate, a selection of
prehensive search of the target area, perhaps different techniques should be tested simulta-
by removal or fumigation of the whole plant neously and the results compared. The final
or grassland turve and careful examination for choice of technique(s) to adopt will be a tra-
any individuals missed by the initial sampling. de-off between accuracy and cost. Similarly,
Similarly, active sampling (such as direct before proceeding with detailed studies, rese-
counts or sweep netting) should be carried out archers should have a clear appreciation of the
wherever possible by the same person, to avoid absolute efficiency of their chosen sampling
introducing individual operator bias. If more technique(s) for the species under investiga-
than one worker is involved, their relative tion and within the context of the particular
'sampling efficiencies' should be compared, habitat.
495
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
Estimation of absolute population the light or by using transmitted light. A simi-
density lar technique can be used for eggs laid into
small leaf veins, but those placed in larger
i) Direct counts
veins, petioles, buds or stems can usually be
Counting individuals in situ is clearly the
detected only by careful dissection of the
most direct method for estimating population
plant. This is laborious but produces very
density on plants, but is not always possible or
detailed information on oviposition behaviour
practicable, particularly with very active spe-
(CLARIDGE & REYNOLDS 1972; THOMPSON
cies. This approach is best applied to large or
1978; STILING 1980).
conspicuous species, but may also be favoured
Attempts have been made to accelerate
when counting the earlier life-history stages or
the process of detecting eggs laid within plant
adults of the more sedentary species. WHITTA-
tissue using a variety of chemical techniques.
KER (1965) for example was able to measure
These generally involve clearing the plant
the density of spittle masses of the cercopids
tissue in boiling lactophenol, which also serves
Neophilaenus lineatus and N. exclamations wit-
to coagulate the egg proteins so that the out-
hin wire quadrats in grassland, whilst vertical
lines of the eggs become visible under magnifi-
stratification in the community of planthop-
cation (CARLSON & HiBBS 1962). Other tech-
pers on the salt marsh grass Spartina patens was
niques use hydrogen peroxide or glacial acetic
quantified by counting directly the number of
acid for clearing the leaf tissue, followed by
individuals in five vertical strata up the stem
staining with acid fuchsin (CHATTERJEE &
(DENNO 1980).
RAM 1970). Such techniques, or modifica-
Many continental European workers stu-
tions thereof (SIMMONS et al. 1984), have now
dying grassland communities have favoured a
been used to detect eggs in a wide variety of
direct counting technique called the 'bioceno-
plant species (SlMONET &. PIENKOWSKI 1977;
meter' (KONTKANEN 1950; ANDRZEJEWSKA
SIMMONS et al. 1985; HEADY et al. 1985).
1965; NOVOTNY 1992). In essence, this invol-
Major disadvantages are that they are time-
ves delimiting a unit area of ground (typically
consuming to perform and involve the use of
0.25 m2) by covering it with a cylinder or box
hazardous chemicals.
that has an open base and gauze-covered top,
from which all insects are extracted by hand-
held aspirator. Other authors refer to this ii) Suction samplers
device as a 'box quadrat' (CHERRILL & BROWN Various mechanical devices have been
1990). It is designed to provide a standardized developed for the physical extraction of in-
areal count, but its accuracy is reliant both on sects from vegetation, using a strong current of
the box being positioned rapidly before any air generated by a motorised fan. The first device
highly mobile individuals escape and on the to achieve widespread use was the DlETRlCK,
observer detecting all the trapped insects. Alt- or 'D-Vac', suction sampler (DlETRlCK 1961)
hough the equipment costs are negligible, (Fig. 1). Typically, it comprises a fan unit
sampling using this method is very time-con- powered by a 100 cm-' two-stroke engine,
suming and has largely been superseded by the connected via a flexible hose to a plastic or
more automated methods dealt with below. fibreglass cylindrical inlet tube housing a mesh
The egg stage within the life cycle pre- collection bag. The inlet cylinder has gauze-
sents special sampling problems. In most covered apertures around the rim to allow air
Auchenorrhyncha species, the eggs are too to enter near the soil when it is held over the
small to be detected easily and most are inser- vegetation. The result is a powerful updraught
ted directly into the plant tissue of either lea- of air through the vegetation, which sucks the
ves or stems. Eggs laid within the leaf lamina insects into the collection bag.
are generally placed just beneath the surface As the material does not pass through the
and are therefore detectable under relatively fan, the insects are generally retained in near-
low magnification as simple bulges in the leaf perfect condition. This means that Auche-
epidermis. Sometimes, detection can be norrhyncha collected alive with this apparatus
improved by varying the angle of incidence of can be used for subsequent experimentation.
496
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
It should be noted however that smaller and estimate of population density, samples from
more delicate species may receive some dam- contrasting grassland types should be directly
age. Also, parasitised Auchenorrhyncha may comparable. However, there is a limit to the
suffer increased mortality and rearing parasi- vegetation height beyond which the process of
Fig. 1.
toids from surviving individuals may be less positioning the inlet tube will flatten or com-
D-Vac suction sampler. The engine
successful. press the plant material so that air is sucked
and fan unit is mounted on a back-
The whole unit is mounted on a backpack over, rather than through, the vegetation. pack frame (left) and connected via a
long flexible hose (middle) to the
worn by the operator, whose hands are left free The D-Vac, including various minor
inlet cylinder which contains the
to place the no::le over the vegetation and modifications of the original design (e.g. sample bag (right).
empty the bag at the end of sampling without
the need to take off the equipment or stop the
engine. The collection no::le is placed verti-
cally over the vegetation for a standard time
period (at least 20 s.), after which any insects
sucked into the collection bag can be emptied
into a separate container, killed and stored.
The greatest advantage of this technique is
that it facilitates sampling of a standardized
area of ground. As the cross-sectional area
delimited by the inlet nozzle is generally
c. 0.1 m:, it is customary1 to take ten such
'sucks' to produce a sample from lm2 of ground.
D-Vac suction sampling is often the pre-
ferred method for sampling grassland and low
crops when compared with other techniques
such as sweep netting and various types of trap
or beating tray (e.g. SlMONET et al. 1979; BUN-
TIN 1988). This is because it often produces
the highest density estimates and the lowest
variation between samples. When measured,
D-Vac extraction efficiencies have been shown
Fig. 2.
to vary for different insect groups (HENDERSON
Two types of G-Vac
& WHITTAKER 1977) and to be sensitive to a suction sampler.
number of extraneous factors (HAND 1986). Right: the simplest
design has a net
Rather few of these studies provide data speci-
collection bag inser-
fically for Auchenorrhyncha. DUFFEY (1980)
ted into the inlet
reported Auchenorrhyncha extraction effi- tube and secured
around the nozzle.
ciencies in rough grassland that varied from
The inlet tube has a
23% in May to 62% in August (presumably cross-sectional area
coinciding with the peak nymphal and adult of 0.01m2. Left: alter-
native design where
stages respectively). Efficiency rose to 70% on
a custom-built inlet
grazed (i.e. short) grassland. HENDERSON & tube has a larger
WHITTAKER (1977) also reported a sward-length cross-sectional area
(0.025m2) and a flan-
effect, with extraction efficiency in-creasing
ge with gauze-
from 32% in long grassland (20-30 cm height) covered holes that is
to 76% in short grass (<5cm tall). mounted beyond the
collection bag to
Efficiency is severely compromised if the allow unimpeded
entry of air into the
vegetation is moist (although probably not as
inlet tube. Both
seriously as when using a sweep net) or 'lodged' samplers are powe-
(flattened by wind, rain or trampling). Theo- red by 30 cm3
engines.
retically, as this method provides an absolute
497
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
THORNHILL 1978), has remained the standard and suction power generated by these machi-
equipment for quantitative sampling from a nes compared to the original D-Vac. The por-
variety of crops and grassland habitats for tability of these new machines make them ide-
many years. It does, however, suffer from a al for more general collecting, especially of
number of severe disadvantages, not least of those species which reside close to the ground
which are cost, weight and a poor reputation or in otherwise inaccessible places (WILSON et
for mechanical reliability. In recent years, al. 1993). It is also a useful technique for
various workers have developed smaller hand- collecting from very tall plants that are not
held suction samplers, by modifying 'suck-or- easily swept (e.g. the reed Phragmites australis)
blow' machines sold for collecting garden refu- or in damp or flooded situations where a sweep
se (SUMMERS et al. 1984; HOLTKAMP & net bag would soon become saturated and
THOMPSON 1985; DE BARRO 1991; WILSON et al. unworkable.
1993; ARNOLD 1994; MCLEOD et al. 1994, The inlet nozzle of any suction sampler
1995; SAMU & SÄROSPATAKI 1995; STEWART has to be placed quickly onto the ground in
6k WRIGHT 1995). BELL & WHEATER (2001) order to avoid invertebrates from outside the
refer to these types of machine as 'G-Vacs1.
delineated area being sucked into the sample.
The various models available are generally
SAMU et al. (1997) compared samples of spi-
similar in design and operation, in that power
ders taken from an alfalfa crop by a hand-held
is provided by a 30 cm3 petrol-driven engine
suction apparatus from within enclosed areas
which sucks air through a smaller (usually c.12 (each approximately 0.5m2) with samples
cm. diameter) inlet tube. The only modificati-
representing the same area of ground but
on of the gardening equipment that is needed
taken from a series of unenclosed sampling
to convert it to an insect suction sampler is to
points. Although the species compositions
attach a fine net bag to the inside of the inlet
and abundance rankings were similar for the
nozzle to retain the insect material collected
two sampling methods, the catches based on
(Fig. 2) (STEWART & WRIGHT (1995) provide unenclosed sampling points were substantially
more detailed instructions).
larger than those where the sampling area was
The inlet nozzle cross-sectional area of enclosed. They therefore suggest that the
these more compact machines is rather too action of placing the inlet tube nozzle onto
small (-0.01 m2) for each 'suck' to be regarded the ground draws in extra individuals from
as a single sample. However, sampling from a outside the target area and that such an 'edge
larger area can be standardized by delimiting a effect' may produce inflated estimates of popu-
fixed area of ground with an open-ended cylin- lation density.
der (e.g. 36 cm diameter, to be comparable to
The suction power of most conventional
the D-Vac collection nozzle) placed over the
samplers is severely reduced when the air flow
vegetation. The nozzle of the G-Vac suction
is impeded; this may happen when the inlet
sampler can then be inserted into the cylinder
nozzle is placed over the ground surface or
and passed repeatedly across the vegetation for
when a large amount of debris builds up in the
a set time interval to collect any insects trap-
collection bag. The Vortis sampler is designed
ped inside.
to circumvent these problems, firstly by intro-
In the only detailed study of the efficacy of ducing air into the system from higher up the
this equipment for sampling Auchenorrhyn- inlet tube, and secondly by dispensing with
cha, STEWART & WRIGHT (1995) showed that any sort of collection bag (ARNOLD 1994).
catches of most species using a G-Vac sampler Instead, insects are sucked up the inlet tube
were comparable with those taken from an into an enlarged chamber designed to create a
equivalent area of ground by D-Vac. Some spe- vortex of circulating air, from which centrifu-
cies known to inhabit the layer closest to the gal forces propel the insects into a detachable
ground were sampled in greater numbers with collection vessel mounted on one side (Fig. 3).
the G-Vac, although this effect was better Whilst the mechanical principles behind this
demonstrated in certain epigeal species of device represent an improvement on the
Coleoptera and Araneae. This observation design of previous suction samplers, its use in
reflects the considerably greater air velocity practice is prone to a new set of problems. The
498
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
insects have much further to travel before rea- easily removed and killed (see MOORE et al.
ching the collection vessel, including passing (1993) for full details). Both adult and nym-
through a set of fixed metal vanes that induce phal Auchenorrhyncha self-sorted very rapid-
the vortex of air; risk of damage to specimens ly in this apparatus, nearly 100% of individu-
is therefore increased. Similarly, there is an als separating within 4-8 hours. BuNTIN
enhanced danger of specimens adhering to the (1988) achieved 94% recovery of leafhoppers
interior walls of the suction tube or chamber if using a laboratory-based device that was simi-
either become coated in moisture. The only lar in principle, funnelling the insects straight
published data on sampling efficiency suggest into an ethanol-filled vial.
that Homoptera are sampled substantially bet-
Fig. 3.
ter by the conventional D-Vac (ARNOLD
Vortis suction sampler. Suction is pro-
1994). The absolute efficiency of the appa- vided by a 30 cm3 engine (top). Air is
ratus for this or any other insect group remains sucked into the inlet tube through a
gauze cone (bottom), drawn through
to be tested.
a set of radiating internal vanes that
Any type of suction sampler will remove generate a vortex and then into an
some quantities of dead plant material, soil expansion chamber (middle). Insects
that are circulated by the vortex in
particles and other debris (more powerful
the expansion chamber are propelled
machines will collect more). Sorting dead in- by centrifugal forces into an escape
sects from this waste material is probably the tube mounted on the side (middle
left) and drop down into a detachable
most time-consuming part of this sampling
collection vessel below.
method. Consequently, several workers have
attempted to develop techniques whereby,
before being killed, the insects' phototactic
responses are exploited to segregate them from
the unwanted material. DlETRlCK et al.'s
(1959) original method was to transfer the
material collected by D-Vac into a Berlese
funnel to sort the animals into tubes of alco-
hol. WALOFF (1980) reported using a 'sorting
frame' to separate Auchenorrhyncha from
debris collected by D-Vac sampling. This com-
prised a wooden-framed muslin funnel with a
clear plastic window at one end. The frame
was placed in front of a light source and the
whole sample was emptied into the funnel.
Emergent insects were attracted to the light
and moved towards the plastic window, from
which they could be removed by hand-aspira-
tor. MOORE et al. (1993) have attempted to
take this principle a stage further towards
automation, by developing a field-based
method that can be employed immediately
after collection of the sample. Using the prin-
ciple employed in the traditional capture of
lobsters (Crustacea: Nephropidae), they con-
structed a 'light-sorter' device from joined sec-
tions of plastic soft drink bottles that were
painted matt black or left clear. The suction
sample debris was placed in the dark section of
the sorter and the insects were left to move of Fig. 4.
their own accord away from the debris and Sorting the catch from a suction sampler. Most suction samples contain substantial
amounts of plant and soil debris. The debris (left) is examined carefully and all insects
into the light section, where they could be
are removed with fine forceps (top). The catch can then be sorted into species (right).
499
© Biologiezentrum Linz/Austria; download unter www.biologiezentrum.at
Where such automatic sorting devices are ning). They found that trap efficiency was
not available or the insect sample has already generally unaffected by wind, temperature or
been killed, catches have to be separated from crop height but was significantly reduced at
the associated plant and soil debris by hand. lower sunlight levels. However, SlMONET et al.
Finding small insects in larger amounts of (1979) found that this method was considera-
debris can be extremely time-consuming and bly less efficient for sampling E. fabae adults
the process is also subject to a number of bia-
than either suction or sweep net techniques.
ses connected with the skill of the sorter, the
BUNTIN (1988) came to a similar conclusion
amount, nature and condition of the dead
for extracting cicadellid adults from bermuda-
plant material and the relative crypsis of the
grass, Cynodon dactylon, and found that the
insect species. The most efficient technique is
technique did not recover any nymphs. In the
to spread the entire sample evenly onto a clean
cooler climate of central Finland, TÖRMÄLÄ
light background, such as a white plastic tray.
(1982) reported this method to be very ineffi-
Individual specimens can then be picked out
cient for sampling the grassland Auchenorr-
from the debris using fine forceps (Fig. 4).
hyncha community, comparing unfavourably
Sorting effort can be standardised by imposing
with sweep netting, suction sampling and
a fixed time limit for processing each sample.
even pitfall trapping in terms of the numbers
Where either the number of insects or the
of individuals and species caught.
amounts of debris are excessive, it may be
pragmatic to process only a sub-set of the
material in each sample. This can be done by iv) Marking techniques
evenly spreading each sample over the sorting
Mark-release-recapture techniques have
tray and separating off a fixed fraction for
been used to estimate population sizes of a
detailed sorting.
wide variety of mobile animals including in-
sects (SOUTHWOOD 1978). At their simplest,
iii) Emergence traps these involve catching, marking and releasing
a number of individuals within a population,
Emergence traps make use of the positive
followed by re-sampling after a period to allow
phototactic response of many mobile insects.
for re-mixing. The ratio of marked to unmar-
Individual trap designs vary but all consist of
ked individuals in the second sample should
an open-ended opaque box or cylinder, with
be the same as that in the population as a
an aperture at the top providing the only sour-
whole; this fact allows a simple estimation of
ce of light. Insects moving up the chamber
the total population size. The technique
towards the light are funnelled into a collec-
makes a number of critically important
tion vessel. Samples are standardized because
assumptions, including (i) random selection of
the trap base covers a fixed area of ground.
individuals for marking (e.g. across sexes, phe-
The apparatus is placed rapidly over the vege-
notypes and age classes), (ii) fully random
tation and sealed at the soil surface to prevent
mixing of marked individuals with unmarked
insects escaping. CHERRY et al. (1977) suspen-
ones after release, (iii) a marking technique
ded the trap from the end of a long pole which
that is persistent but does not affect survival or
was used to lower the trap over the vegetation
subsequent behaviour of individuals, and (iv)
at a distance from the operator; this was inten-
a population which is closed (i.e. no birth,
ded to reduce disturbance of the resident
immigration, death or emigration) within the
Auchenorrhyncha by the operator. The alter-
period of study. These assumptions are rarely
native is to position the trap early in the mor-
fully met and their violation can produce
ning when the insects are likely to be least
serious biases in the resultant population esti-
active and then leave it in place for several
mates. The comparative ease and accuracy of
hours (TÖRMÄLÄ 1982) or days (CLEMENTS
other methods of population estimation have
1979). CHERRY et al.(1977) found that this
meant that this technique has not been widely
method recovered more than 80% of potato
used by workers studying Auchenorrhyncha.
leafhoppers E. fabae from alfalfa (absolute
densities being calculated after fumigation of However, various marking techniques
the trap to retrieve any individuals remai- have been employed in studies of local disper-
500