pitfall trapping as a method for studying populations of carabidae

Pitfall Trapping as a Method for Studying Populations of Carabidae (Coleoptera)Author(s): P. J. M. GreensladeSource: Journal of Animal Ecology, Vol. 33, No. 2 (Jun., 1964), pp. 301-310Published by: British Ecological SocietyStable URL: http://www.jstor.org/stable/2632 .

Accessed: 09/03/2014 15:50

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

British Ecological Society is collaborating with JSTOR to digitize, preserve and extend access to Journal ofAnimal Ecology.

http://www.jstor.org

This content downloaded from 209.129.47.117 on Sun, 9 Mar 2014 15:50:25 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=briteco

http://www.jstor.org/stable/2632?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


301

PITFALL TRAPPING AS A METHOD FOR STUDYING POPULATIONS OF CARABIDAE (COLEOPTERA)

BY P. J. M. GREENSLADE

Department of Zoology, Imperial College, London*

Pitfall traps provide a convenient method of investigating the ecology of adult Carabidae and have been used in studies on such topics as the seasonal incidence of adults, the spatial pattern of distribution in populations, the relative numbers of a species in different vegetation types, in describing the Carabid fauna of habitats, and in work on daily rhythms of activity (Hikimiuk 1948; van der Drift 1951; Scherney 1955, 1960; Gilbert 1956; 1958; Dawson 1957; Kabacik 1957; Skuhravy 1956; Williams 1959b; Kirchner 1960; Thiele 1960; and others). These traps suffer from the disadvantage that catches depend both on the density of the population being sampled and the activity of individuals in it. Briggs (1961) concluded that in the Carabid Harpalus rufipes (Deg.) the size of the population played a minor part in determining numbers trapped, and Morris (1960) has stressed the desirability of obtaining estimates of absolute population in sampling pro- grammes, that is the number of individuals per unit of habitat.

As pitfall catches vary with the amount of locomotor activity shown by Carabidae they are influenced by weather, and Briggs (1961) and Greenslade (1961) found a cor- relation between catches and temperature. The problem of the relation between catch, population size and activity is not dealt with here, but Taylor's (1963) recent discussion of temperature and insect flight and suction trap catches is relevant. There are, however, other factors which may result in variation in pitfall catches of Carabidae, the two main ones, which have not been considered by some workers, are the effect of ground vegetation in impeding Carabid movement and reducing catches, and the differential susceptibility of species to trapping according to their behaviour. Taking these factors into account the cases in which it is legitimate to draw conclusions from pitfall trapping are discussed. This is necessary for despite their drawbacks pitfall traps are often the only method available in studies on Carabid populations. Although direct quadrat counts are to be preferred they may be precluded by low density or other factors.

On the Imperial College Field Station, Silwood Park, Berkshire, where the work described in this paper was carried out, quadrat counts of Carabidae were made in three habitats (Greenslade 1961), and the densities of adults were: beechwood, 1-33 per m2 (6 Carabidae in 4-46 m2 sampled); grass heath, 28 per m2 (52 in 1 86 m2); arable land, 80 per m2 (82 in 1 02 m2). With the sparse woodland population direct counts are not practicable on account of the time and labour they involve, while in other habitats quadrat sampling may result in excessive destruction of the site being studied, especially as Carabid distri- butions are often highly aggregated and centred on restricted favourable areas (Green- slade 1963a, 1964a, e).

TRAPPING METHODS

The traps used here simply consisted of 1 lb jam jars (11-5 cm high with a diameter at the mouth of 6 5 cm), sunk in the ground with their rims level with the soil surface. More

* Now at: Department of Agriculture, Honiara, British Solomon Islands Protectorate.



302 Pitfall trapping

elaborate traps have been used by Williams (1958) incorporating a device to segregate the catch according to time of capture, and by Fichter (1941) whose trap contained pre- servative and excluded rain water. Here such modifications were unnecessary as it was convenient to visit traps daily or even more frequently.

Following observations by Walsh (1931, 1933), Kaufmann (1937) and others, experiments were made to determine whether baiting traps with carrion resulted in a marked increase in catch of Carabidae, in addition the effect on catch of the density of traps in a plot was investigated. Observations were also made on camouflaging traps in case diurnally active species might see and avoid them; white-painted traps were compared with black, and unpainted controls. In preliminary experiments it was found that neither baiting nor camouflaging traps caused variation in catches, and no trap-density effects were recorded. Therefore these topics were not pursued and the simple type of trap was retained, although they were rarely spaced with a distance between them of less than approximately 2 m, except in experiment II described here.

GRASSES

TRAP SOI L

. =7

A B C

FIG. 1. Three types of simple pitfall trap.

Other work was carried out to determine the optimum position of the mouths of traps in different types of ground cover, and for different species of Carabidae. The simplest approach to this problem is comparison catches when a known number of Carabidae are confined within an area containing one or more traps of given types. However, in such experiments it was found that there was a decline in the incidence of marked Carabidae which had been introduced into galvanized metal strip enclosures, and with recaptures outside them, this showed that there was continual escape. Van der Drift (1951) also reported difficulty in enclosing Carabid populations. Instead information on the relation between the position of mouths of traps, and their catches of different Carabidae, and ground vegetation types, was obtained from two other experiments and some supple- mentary observations.

Experiment I

A plot of nine traps was laid out in a uniform area of grass heath whose composition is described in the second paper in this series (Greenslade 1964d). The traps were of three types as shown in Fig. 1:

A. The mouth of the jar level with the upper surface of the vegetation. B. The jar sunk in the ground with the lip level with the ground surface, within a

cleared area of 60 cm diameter from which all vegetation had been removed. C. The jar sunk in the ground but without a cleared area.

The traps were arranged in a Latin square of total size 5 5 x 5 5 m, and were operated for two periods in 1959: in the spring from 18 March until 17 May, and in the summer



P. J. M. GREENSLADE 303

from 14 July until 25 August. In the spring trapping the ground cover consisted of dead vegetation from the previous growing season, mainly grasses, forming a mat 10-13 cm deep, with occasional tussocks. By the summer this mass of dead grass had been reduced to a depth of 6-9 cm and was bound together by fresh growth of grasses and ruderals, many of the latter having a rosette growth form.

Experiment II

A further four traps were situated near the grass heath plot, in woodland under beach, where the ground was covered with loose leaf litter and a sparse recent growth of Holcus mollis and Poa pratensis. Traps of types B and C were compared from 18 March until 23 May 1959, and were arranged in a square of side I1 m, in which similar traps were diagonally opposite each other.

Other observations

A transect of five traps of type B at intervals of 3-7 m was laid out in leaf litter under beech, crossing a sandy track. The traps were operated from 13 May until 27 July 1959, and litter depths were measured at five points within a radius of 1 -8 m of each trap.

Finally quadrat samples were taken in arable land in August 1960. A 30 cm square quadrat was used and counts of Carabidae from a total of 1 m2 of soil were obtained. At the same time, during August, three pitfall traps were also in use in the area of these quadrat samples and their catches were recorded.

The results from the two experiments and other observations are shown here with catches expressed for different categories of Carabid distinguished according to size and habit. Detailed catches of the species in each trap are given by Greenslade (1961).

RESULTS

The distribution of the catch in the two experiments can be considered in terms of the four variables: (1) Position of trap; (2) Type of trap; (3) Type of Carabid; (4) Vegetation.

Position of trap

In experiment I the plot was designed to avoid effects due to spatial trends in the density of Carabidae. As studies of Nebria brevicollis (F.) and Pterostichus madidus (F.) described in the following paper (Greenslade 1964d) show that these species commonly travel 10 m in a 24 h period, the possibility that there is systematic variation in catch in terms of the rows and columns of the plot need not be considered. There is, however, likely to be variation in catches due to the position of traps in relation to near-by grass tussocks and other minor irregularities of the ground surface, and it was these local differences that traps of type B were intended to overcome.

Type of trap

It is convenient to add the catches in a plot for each type of trap, and to test the signi- ficance of the differences between them by calculating x2. The total number of Carabidae taken in the three types of trap in the two experiments are compared in Table 1.

Traps of type B consistently took larger numbers of Carabidae than the others although in experiment II the difference between B and C was not significant. It was noticed that the




cleared areas around these, if kept clean, were not affected by seasonal changes in vegetation, or its wearing when the traps were visited frequently over a long period. In experiment I, in the spring observations, traps of type C were more efficient than A (X2 - 13'33, P<0-001), while in the summer the situation was reversed, A showing significantly higher catches than C (X2 = 16-0, P<0 001). This suggests that in the spring some Cara- bid movement occurred within the tangle of vegetation, while in the summer the more compact mass of dead grasses permitted greater activity on its surface, and less between it and the soil.

Table 1. The catches of Carabidae in three types of pitfall trap

Type of trap r A _

X22 P

A B C

Expt I f Spring 5 57 25 47-4 <0 001 Expt. I XSummer 219 300 173 35-4 <0-001

-~---~ x22 p B C

Expt. II 71 53 1P7 >010

For explanation see text and Fig. 1.

Type of Carabid

The Carabidae whose habits are known and which occurred in numbers in experiment I in the spring, can be divided into two groups: the larger ground-living, mainly predacious species (Davies 1959), Pterostichus madidus and P. caerulescens (L.), and those species which are commonly associated with plants and are often collected by sweeping, Amara species (A. lunicollis Schdt., A. communis (Pz.), A. aenea (Deg.), A. familiaris (Duft.) and A. tibialis (Payk.) occurred), and Bradycellus harpalinus (Serv.).

Table 2. The catches of two types of Carabidae in experiment I in the spring (excluding traps of type A)

Type of trap Type of Carabid Xi2 P

B C Larger ground-living species 31 5 18 4 <0 001 Species associated with plants 17 16 01 >0-7

As very small numbers were taken in traps A in this period comparison is restricted to B and C in Table 2 which shows that type B traps took significantly more of the Pterostichus species, while those associated with plants occurred equally in both kinds of trap. This indicates that the larger species were moving within and under the vegetation, and the cleared patches in B traps may have presented a larger catching area than the simple sunken traps of type C. Ibbotson (1958) recorded high catches of the Frit fly (Oscinella frit L.) on sticky traps in cleared areas of 1-5 m in diameter in oat fields which he attributed to klinokinetic movements due to abrupt temperature gradients at the periphery of the cleared circles. Here the higher catches of the Pterostichus species in B traps may result from rapid turning movements after entry into the unfavourable, or at least strange, environment of the cleared areas. In contrast the Bradycellus and




Amara species move on or in the vegetation to a greater extent, and may tend not to leave it when encountering a cleared area.

The same comparison cannot be made for experiment I in the summer or for experiment II, as Amara and Bradycellus species were infrequent. But in the former other species occurred which are known to be phytophagous (Davies 1959), and their distribution can be compared with that of Pterostichus madidus, the only abundant large predatory species which was active in the plot then. The phytophagous Carabidae fall into two groups, Harpalus rufipes, and H. affinis (Schrank), and Pterostichus melanarius (Ill.), all larger species which seem to feed to some extent on fallen plant seed (Briggs 1957; Davies 1959), and the smaller species, Bradycellus harpalinus (a single specimen) and the more agile Calathusfuscipes (Gz.) and C. melanocephalus (L.). Catches of these groups are shown in Table 3.

Table 3. Catches of different types of Carabidae in experiment I in the summer

Type of trap Type of Carabid _ P- X22 p

A B C Pterostichus madidus 155 263 136 64-66 <0 001

flarger 8 15 14 2.33 >0 3 Phytophagous species ><smaller 19 12 29 7.3 <0 05

Here again for the non-phytophagous Pterostichus madidus type B traps are the most efficient, while for the larger phytophagous species the distribution of the catch can be attributed to chance, suggesting that they move more within the vegetation. For the smaller species traps of type C are the most efficient and these can be regarded as moving mainly on the surface of the vegetation.

Here these trapping differences between types of Carabidae are attributed to the strata in which they are active within the ground vegetation, but whether or not this is correct, the point demonstrated is that, within one type of ground cover, different types and species of Carabidae are trapped at different rates in any given trap.

Vegetation

P. madidus formed a high proportion of the total catch throughout these experiments and the relative numbers in different traps are compared in Table 4.

Table 4. Catches of Pterostichus madidus in experiments I and II

Type of trap A____ _

I_ _

X22 P

A B C

Expt f I Spring 3 26 5 28 66 <0 001 Expt. I Xj,Summer 155 263 136 64-66 <0.001

X2 p

B C

Expt. II 47 27 5 4 <0 05

In each case B type traps were most effective for this species. Their superiority was least marked, but still significant, in experiment II and this can be attributed to the sparser vegetation allowing more freedom of movement over the surface of the ground, so reducing the difference between the traps. In experiment I B traps took a lower proportion




of the total catch in the summer observations, 470% as opposed to 760% in the spring. As the differences between A and C are slight in both seasons, the data for this species in experiment I in Table 4, can be treated as a contingency table, and x2 calculated applying Yate's correction: X22 = 11X17, P<0001. Traps of type B were therefore, compared with A and C, significantly more efficient at trapping P. madidus in the spring than in the summer.

80-

60- w

40-

020-

O 1 2 3 LITTER DEPTH (cm)

FIG. 2. Pitfall catches of Nebria brevicollis, and litter depth in a woodland transect.

As observation showed no evidence for major changes in behaviour in P. madidus in the two seasons, the difference in the effectiveness of B type traps can be related to changes in the ground vegetation. It is suggested that in the summer compaction of the dead plant material and the effect of new growth reduced the amount of activity which was possible within and beneath it, and so also reduced the differences between the three trap types.

Other observations

Catches of Nebria brevicollis which comprised 72% of the Carabidae taken in the woodland transect traps, are shown in Fig. 2, where they are plotted against litter depth. There is an inverse relation between the depth of the litter around a trap and its catch of N. brevicollis.

Table 5. Carabidae of three types of daily activity cycle taken in soil quadrats and pitfall traps in an arablefield, August 1960

Activity rhythm -` Total

Diurnal Plastic Nocturnal Soil quadrats 36 13 33 82 Pitfall traps 26 16 70 112

Carabidae taken in quadrat samples in an arable field are compared with those occur- ring in pitfalls in the same area at the same time in Table 5. In the table the Carabidae are grouped according to their daily periodicity of locomotor activity, where the plastic category distinguishes those which may be active in the day or night in response to variation in their environment, especially climatic factors (Greenslade 1963b). The direct




population counts in the quadrats show that diurnal and nocturnal species are present in similar numbers while in the pitfall traps the catch of nocturnal species is significantly higher than that of diurnal. If the plastic species are omitted the data can be treated as a 2 x2 contingency table: X12 = 1078, P<0-01.

DISCUSSION

The type of trap with a cleared area, B, was selected for further studies of Carabidae, as it consistently made higher catches, the only exception being Calathus species in experiment 1, and can be used in most sorts of ground cover; it was also noticed that in comparison with C, B traps were less frequently fouled by dead leaves and other debris. B traps also eliminate variation in catch due to grass tussocks and similar obstacles close to traps.

Other variation in pitfall catches, apart from that which the traps are commonly used to investigate, comes from two main sources, the type of Carabid, and the nature of the ground cover. Experiment I showed that the traps are selective in that different Cara- bidae are taken at different rates. It was suggested that this was due to various species and life-forms being active in different strata in thick grass heath. Kabacik (1957) found that Carabus arcensis Herbst moved only on or near the surface of leaf litter, while Pterostichus niger (Schall.) may be active both on and within it. If large Carabus and small Bembidion are at risk among traps for a given time, the larger species, travelling further, will tend to be trapped in higher numbers. While this is clear for an extreme example it also holds when species are compared which are apparently similar in size and activity as in Pterostichus madidus, P. melanarius and Calathusfuscipes in experiment I. In addition to this the daily time of activity of species affects trap catches. Van der Drift (1951) commented on the ability of diurnal Notiophilus species to evade capture by traps; they are small and often recover balance and escape when on the lip of a pitfall, but may also see and avoid traps. This is supported by the comparison of pitfall catches and direct counts of population densities in Table 5, which shows that the diurnal Cara- bidae are under-represented in the former.

The less marked superiority of B traps over C in experiment II as compared with I implies that a sparse ground cover allows more movement over the surface. This is also noticed in the spring and summer trapping in experiment I when surface movement was easier on the compressed mat of vegetation in the summer. Although this variation in the efficiency of traps caused by differences in the physical characters of various types of grassland is slight, its presence implies that there will be a greater discrepancy between the proportions of populations of the same species trapped when more diverse types of ground cover are compared, for example a thin layer of leaf litter, and rank pasture. This is supported by the woodland transect which shows that catch varies inversely with the depth of the ground cover. In the small area of the transect the high catches near the track, where low litter depths were measured, cannot be attributed to a locally dense population of Nebria brevicollis, for hand collections did not show any concentration of individuals there; but the distribution of catches can be explained by the absence of litter allowing in places a greater freedom of movement, and hence a greater likelihood of any individual Carabid being trapped.

These observations demonstrate that pitfall trapping cannot properly be used for the quantitative assessment of the Carabid fauna of any habitat; nor should it be employed to compare the numbers of one species in different habitats. There are also indications that seasonal changes within a type of vegetation cover may affect the trapping of Carabidae.




However, Fichter (1941), discussing pitfall traps, states that they may be used to compare the 'composition of soil surface arthropod populations in two or more habitats'. But for quantitative work this is evidently untrue, especially if it embraces groups more

diverse than a single family of Coleoptera. Boyd (1960) used pitfall traps to compare the surface active invertebrate fauna of

grazed and ungrazed maritime grassland on the premise that 'pitfall catches probably give a fairly accurate indication of the surface activity of the soil fauna'. The comparison was between closely cropped grassland and an ungrazed sward which reached ankle depth, and while the results showed distinct qualitative differences in the fauna of the two areas, it is questionable whether any quantitative comparison is permissible.

Williams (1959a) used pitfall traps to measure the activity of ground-living inverte- brates in different habitats interpreting the catches of traps as measures of the level of activity in their respective vegetation types. However, catch represents locomotor activity expressed as movement in a horizontal plane, and for any given energy outlay the amount of such movement will be greater in an essentially two-dimensional habitat, for example a thin layer of leaf litter, than in a three-dimensional one such as grassland.

After using pitfalls to investigate the Carabid fauna of arable land, Scherney (1960) concluded that Carabidae were associated with cultivation; the numbers taken in pitfall traps in different habitats were in the order wheatfields>barley>potatoes>clover>grass meadow>densely weed-covered waste ground. His conclusion may be valid, for the quadrat counts mentioned in the present work showed the highest Carabid density on arable land, but it is noticeable that the sequence of increasing Carabid abundance in traps, from a dense weed cover to cereal fields, is also a sequence of decreasing resistance to the

passage of Carabidae over the ground surface. These examples illustrate the fact that pitfall traps must be used with discretion,

especially for comparative purposes. This does not affect their value as a sampling method in marking and recapture work, or for studies on life histories. In the latter case the chance of variation in catches due to seasonal changes in vegetation must be noted, and results from pitfalls are best combined with observations on larvae, and gonad dissections, as in studies on Calathus species (Gilbert 1956), and Pterostichus madidus (Greenslade 1964b). They can be used also in comparing the fauna of different habitats in conjunction with marking and recapture or quadrat counts, and for investigating the distribution of one species in a single vegetation type.

Apart from these cases pitfall results are of limited value but three exceptions may be noted. In conjunction with hand collections they can be employed for the qualitative assessment of different Carabid faunas, especially when many species are restricted to one habitat or another, and they yield some information on the frequency of species; but it is desirable that observations should be made also on the distribution of larvae. If two vegetation types are sampled with pitfalls and a species is trapped in higher numbers in the one which presents the most resistance to Carabidae movement, for example in long grass when it is compared with a thin litter layer, it can be concluded that traps indicate the species' true distribution. Examples of pitfalls used in these ways are given by Greenslade (1964c).

Finally in the second paper in this series (Greenslade 1964d) it was found that where a plot of traps extended over more than one type of ground cover, two species might occur each associated with one type, usually being absent when the other was present. It was concluded that in inverse patterns of this kind, pitfall catches did show the true distribu- tions of the species involved.




SUMMARY

Pitfall trapping is reviewed as a sampling method for Carabidae. Catches are determined primarily by the size of the population at risk and the level of locomotor activity but species may show differential susceptibility to trapping according to size, behaviour, and the strata in which they are active in the ground vegetation. Catches of a single species may vary in different types of ground cover depending on the resistance they present to horizontal movement. These observations severely restrict the number of cases in which it is permissible to make quantitative comparisons between the catches of different pitfall traps.

REFERENCES

Boyd, J. Morton (1960). Studies of the differences between the fauna of grazed and ungrazed grassland in Tiree, Argyll. Proc. Zool. Soc. Lond., 135, 33-54.

Briggs, J. B. (1957). Some experiments on control of ground beetle damage to strawberry. Rep. E. Mall. Res. Sta. 1956, pp. 142-5.

Briggs, J. B. (1961). A comparison of pitfall trapping and soil sampling in assessing populations of two species of ground beetles (Col.: Carabidae). Rep. E. Mall. Res. Sta. 1960, pp. 108-12.

Davies, M. J. (1959). A contribution to the ecology of species of Notiophilus and allied genera (Col., Carabidae). Ent. Mon. Mag. 95, 25-8.

Dawson, N. (1957). Ecology of Fenland Carabidae. Ph.D. thesis, Cambridge. Drift, J. van der (1951). Analysis of the animal community in a beech forest floor. Tijdschr. Ent. 94,

1-168. Fichter, E. (1941). Apparatus for the comparison of soil surface Arthropod populations. Ecology, 22,

338-9. Gilbert, 0. (1956). The natural histories of four species of Calathus (Coleoptera, Carabidae) living on

sand-dunes in Anglesey, North Wales. Oikos, 7, 22-47. Gilbert, 0. (1958). The life history patterns of Nebria degenerata Schaufuss and N. brevicollis (Fabricius)

(Coleoptera, Carabidae). J. Soc. Brit. Ent. 6, 11-14. Greenslade, P. J. M. (1961). Studies in the ecology of Carabidae (Coleoptera). Ph.D. thesis, University

of London. Greenslade, P. J. M. (1963a). A concentration of Carabidae (Coleoptera) at Fleet in Dorset. Ent. Mon.

Mag. 99, 46-8. Greenslade, P. J. M. (1963b). Daily rhythms of locomotor activity in some Carabidae (Coleoptera).

Ent. Exp. Appl. 6, 171-80. Greenslade, P. J. M. (1964a). Further notes on aggregation in Carabidae (Coleoptera) with especial

reference to Nebria brevicollis (F.). Ent. Mon. Mag. (In press). Greenslade, P. J. M. (1964b). On the ecology of some British Carabidae (Coleoptera) with especial

reference to life histories. Trans. Soc. Brit. Ent. (In press). Greenslade, P. J. M. (1964c). The habitats of some Carabidae (Coleoptera). Ent. Mon. Mag. (In press). Greenslade, P. J. M. (1964d). The distribution, dispersal and size of a population of Nebria brevicollis

(F.) with comparative studies on three other Carabidae. J. Anim. Ecol. 33, 311-33. Greenslade, P. J. M. (1964e). Aggregation in British Carabidae (Coleoptera). Ent. Mon. Mag. (In press). Hikimiuk, A. I. (1948). (A quantitative study of seasonal phenomena in the life of Carabus species).

[In Russian.] Bull. Sov. Nat. Moscow. 1153, 4-13. Ibbotson, A. (1958). The behaviour of frit fly in Northumberland. Ann. Appl. Biol. 46, 474-9. Kabacik, D. (1957). Z badan nad biegaczowatymi (Carabidae) lasu sosnowego. Ecol. Polska. B, 3,

239-42. Kaufmann, R. R. U. (1937). Beetles associated with carrion in Pannal Ash, near Harrogate. Ent. Mon.

Mag. 73,78-81, 227-33. Kirchner, H. (1960). Untersuchungen zur Okologie feld bewohnender Carabiden. Dissertation, University

of Cologne. Moore, B. P. (1957). The British Carabidae (Coleoptera). Part I. A check list of the species. Ent. Gaz.

8, 129-37. Morris, R. F. (1960). Sampling insect populations. Annu. Rev. Ent. 5, 243-64. Scherney, F. (1955). Untersuchungen uber Vorkommen und wirtschaftliche Bedeutung rauberisch

lebender Kafer in Feldkulturen. Z. PflBau, 6, 49-73. Scherney, F. (1960). Beitrage zur Biologie und okonomische Bedeutung rauberisch lebender Kaferarten.

Z. angew. Ent. 47, 231-55. Skuhravy, V. (1956). Fallenfang und Markierung zum Studium Laufkafer. Beitr. Ent. 6, 285-7.

J.A.E. G




Taylor, L. R. (1963). Analysis of the effect of temperature on insects in flight. J. Anim. Ecol. 32, 99-117. Thiele, H. U. (1960). Gibt es Beziehungen zwischen der Tierwelt von Hecken und angrenzenden Kultur-

feldern? Z. angew. Ent. 47, 122-7. Walsh, G. B. (1931, 1933). Studies in the British necrophagous Coleoptera. Ent. Mon. Mag. 67, 76-81;

69, 28-32. Williams, G. (1958). Mechanical time-sorting of pitfall catches. J. Anim. Ecol. 27, 27-35. Williams, G. (1959a). The seasonal and diurnal activity of the fauna sampled by pitfall traps in different

habitats. J. Anim. Ecol. 28, 1-13. Williams, G. (1959b). Seasonal and daily activity of Carabidae, with particular reference to Nebr-ia,

Notiophilus and Feronia. J. Anim. Ecol. 28, 309-30.

APPENDIX

A recent paper by Mitchell (1963: J. Anim. Ecol. 32, 377-92) raises two further points concerning pitfall trapping on which some observations were also made during the work at Silwood (Greenslade 1961 and unpublished results). They are predation in traps, and the use of pitfall catches in conjunction with population estimates to investigate variation in locomotor activity.

Mitchell found that small Carabidae were preyed upon in traps both by larger species and by birds; although looked for in the work described here neither type of predation occurred. Briggs (1961) en- countered bird predation and placed guards above traps but these affected catches, and Mitchell's solu- tion, crumpled wire netting within traps, appears the more satisfactory. When predation by the larger Carabidae occurs it is possible that it might be avoided by supporting a false floor of coarse wire gauze (mesh size 4-5 mm) in traps a few centimetres above the bottom; this would allow small Carabidae to pass through but retain the large species. In this case a guard above the trap or a funnel in the mouth might be necessary to prevent bird interference. At Silwood Trechus species, which are soft-bodied and presumably susceptible to predation, were infrequent in traps and were not studied, but other trapped arthropods, especially spiders and Collembola, were usually attacked by some Pterostichus species.

As in Mitchell's work attempts were made to use pitfall catches and population estimates to study activity, and in particular to segregate the effects of climatic and other factors, using the partial regression technique. It was concluded that the method was of limited practical value although Mitchell shows that it can be useful in detailed studies on one or two species in a uniform habitat. There are three major difficulties:

(a) The complexity of the factors determining insect activity. Mitchell deals with time-trends, the average level of activity varying with the age structure of the adult population, and simple climate effects; among other problems to be considered are conditioning of Carabidae to temperature, and the interaction of temperature and humidity. It should be noted also that there may be time-trends in the sex ratio with associated differences in the activity of the sexes.

(b) The problem of describing the microclimate. Although activity can be related at times to standard meteorological records it may be necessary to make precise observations in a species' habitat; here it has been shown that different Carabidae may move in different strata where steep temperature and humidity gradients occur. In addition some species are active in response to weather and show no consistent daily rhythm (Greenslade 1963b), so that it may not be possible to represent the microclimate by observations made at the same time each day.

(c) The inadequacy of population estimates. Any analysis of activity will depend on the accuracy of these estimates, and if derived from marking and recapture they will usually have high standard errors. At the beginning and end of a species' seasonal activity, when it may be most important to distinguish between variation in activity and population density, pitfall catches tend to be so low that capture and recapture is inapplicable. In direct quadrat sampling, on the other hand, it is often difficult to differentiate between members of the active population resting between bursts of activity, and individuals still quiescent after hibernation or aestivation or before emergence.

I am grateful to Dr B. Mitchell for kindly sending me a copy of his paper, and to the editors for allowing the inclusion of an appendix.



pitfall trapping as a method for studying populations of carabidae

Documents