Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The response of invertebrate assemblies to grazing

The response of invertebrate assemblies to grazing Gibson, C W D . Brown, V K , Losito, L and McGavin, G C 1992 The response of invertebrate assemblies to grazing - Ecography 15 166-176 Invertebrate communities were monitored in a controlled sheep grazing experiment on ex-arable land on limestone and in nearby older calcicolous grasslands Results for vacuum sampling of Heteroptera and herbivorous Coleoptera are presented and compared with the responses of spiders, leafhoppers and leaf-miners Invertebrate groups differed in their responses to grazing and grassland age in terms of species diversity, abundance, variability over time and the development of a specific fauna on short turf Leaf-miner assemblies were most strongly determined by the effects of grazing on plant species composition, while spiders' responses could largely be explained by the effects of grazing on plant architecture Spider species simply accumulated over time, whilst leaf-miners were the most labile group Spiders, leaf-miners and leafhoppers all contained some common species restricted to old grasslands, whilst herbivorous Coleoptera and Heteroptera did not No single group demonstrated the full range of invertebrate responses Spiders, leafhoppers and leaf-miners each showed useful features as "indicator groups'", but more than one group should be chosen for study in the context of particular research projects C W D Gibson and V K Brown, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY, U K -L Losuo, 30 Cope Close, Botley. Oxford OX2 9AJ, U K G C McGavin, Umv Museum, Parks Road, Oxford OXl 3PW, U K Introduction The change in invertebrate communities dunng succession rarely follows changes in vegetation composition alone (but see Brown and Hyman 1986) Reality is not as simple as this Invertebrates often show complex resfKjnses because plant architecture may be as important as plant species composition (Southwood et al 1979) In addition, their resjKwises may be related to the sequencing of immigration and to hfe-histones (Brown 1986, Southwood et al. 1988), and to the effects of interactions within and between trophic levels (Lawton 1988). Invertebrate assemblies associated with vegetation influenced by grazing or cutting are known to show a vanety of responses compared with unmanaged areas In grassland, invertebrate speaes nchness and abundance may be dimmtshed in grazed or cut areas due to the lack of architectural diversity (Morns 1973a, 1990a). Accepted 13 February 1991 © ECOGRAPHY However, some species may require the particular features of microchmate and/or structure which only exist m grazed grasslands (e g Thomas 1983), or even need grasslands with a mosaic of structure to provide conditions for different phases of the life cycle (e g Chernll and Brown 1990a, b, Weidemann et al 1990) This picture is complicated by the apparent trade-off between plant species diversity and structural diversity in grasslands. Generally, the more plant species there are in a grassland, the more invertebrate herbivore species. However, management which maximises plant species diversity tends to decrease architectural diversity and thus precludes many species from establishing One compromise which has been suggested to maximise invertebrate diversity is rotational management (Bacon 1990, Morris 1990a), where grasslands are grazed or cut at intervals, not too short to decrease plant species diversity but long enough to allow architectural diversity to develop This recommendation can be regarded ECOGRAPHY 15 2 (1992) as an applied version of hypotheses that intermediate levels or rates of disturbance tend to maximise diversity (Connell 1979, Huston 1979) The different responses of invertebrate groups to plant species diversity and structure have produced severe problems in understanding invertebrate community development and in designing management to maximise diversity Among studies which have addressed more than one invertebrate group. Moms and colleagues found that many Heteroptera were simply diminished by increased cutting or grazing (Morns 1973b), while herbivorous Coleoptera appeared to be less affected (Morns and Rispin 1987) Since the amount of work involved m sampling all invertebrate groups IS usually prohibitive, objective methods are needed for choosing groups which are representative either of the response of the whole community, or the wildlife "value" to be expected under different management regimes Here, we report on the first six years of an experimental study where the changes in a number of invertebrate groups are examined dunng early secondary succession under different grazing regimes. These changes are viewed against the background of the fauna of older grazed grasslands Particular attention is paid to the way in which different taxa respond to grazing management and to establishing the nature of the communities that develop, especially the assemblies which are associated with the short turf produced by heavy grazing We report detailed results for two components of the fauna, the Heteroptera and phytophagous Coleoptera, but compare their responses with other taxa studied, mainly the auchenorrhynchan Homoptera, the leaf-miners and spiders (Brown et al unpubl data, Gibson et al m press ) studied The companson allows a judgement of whether there is an "ideal" invertebrate group which reflects widespread community changes, or whether there is no alternative to studying a wide vanety of taxa if invertebrate community development is to be understood quadrat recording was started in 1984, areas were not fenced until early 1985 Deer rarely enter the experimental paddocks The field lies within the University of Oxford's Wytham estate (1240 ha of which c 410 ha contains semt-natural vegetation) It is surrounded by woodland, but also abuts onto three patches of old-established hmestone grassland with varying degrees of scrub encroachment Ten further patches of ancient grassland are scattered elsewhere withm the woodlands (Gibson 1986) and these thirteen patches form the local potential source for calcicolous grassland flora and fauna (Gibson and Brown m press a) All the grassland patches and the experimental field he on Jurassic coralhan limestone (Arkell 1947) there are no other old hmestone grasslands within 5 km of the site and the nearest substantial areas are >15 km away Sampling methods D-vac samples were taken on five occasions (May, June, July, Aug , Sep in 1985 and 1986, two occasions (early Aug , early Oct ) m 1988 and three occasions (May, July and Aug ) in 1989 Ungrazed controls were sampled from May 1985, short-period spnng and autumn grazed from July 1985 onwards, although the first grazing penod m the latter was not until Sep. 1985 Heavily grazed treatments in Upper Seeds were sampled from 1988 onwards, and the two older grassland patches with spring-and-autumn grazing in 1989 only In paddocks, a sample unit of three sucks (total 0 274 m-) was taken from random positions m each of two 3 x 3 m plots where vegetation was assessed regularly by point quadrats (Gibson et al 1987b) In other treatments, similar samples were taken from notional 3 x 3 m squares surrounding each of twelve 1 x 1 m permanent quadrats in each treatment or area, where vegetation was also assessed Heteroptera and herbivorous Coleoptera were sorted from other invertebrates, counted and identified "bhnd" (unaware of which treatment which sample belonged to), by G McG and L. L respectively, to species or nearest taxon possible Other Coleoptera were Ignored as D-vac suction is known to be lneffiaent for their extraction (Duffey 1980) and may vary considerably according to vegetation structure (Morns and Rispin 1987). The minimum number of species in a sample was estimated by counting lmmatures as an extra species only if none of the species they could have belonged to was present as an adult The phytophagous Coleoptera were mainly herb-feeders Heteroptera were assigned to one of SIX feeding types" "moss-feeders", "grass-feeders", "other herbivores" (herb-feeders and generalist herbivores), "mixed feeders" (partial predators), "carnivores" and "tounsts" (sensu Southwood et al. 1982). Nomenclature follows Clapham et al (1987) for Methods Study area Grazing experiments took place m a 10 ha field abandoned from agriculture m 1982 (Upper Seeds) Details of the expenmental design and grazing regimes were given by Gibson et al. (1987). Briefly, eighteen 30 x 30 m paddocks have been arranged in two Latin squares (grids) where individual paddocks have been grazed by sheep each autumn, grazed each spring or left ungrazed, since 1985. Outside the paddocks, the field has been divided into areas either grazed by sheep from Apni to November ("SA"), grazed from late August to November ("AA") or left ungrazsd, but with free access to deer (Dama dama and Mimttacus reevesi). Although ECOORAPHY 15 2 (1»2) Results Heteroptera Forty-one species of Heteroptera were recorded in the D-vac samples, of which 37 were found m the experimental treatments on Upper Seeds However, no single species was regularly common enough to test the effects of grazing on individual species Grazing did, never- b) c) Fig 1 The composition of Heteroptera by feeding type m different grazing treatments and areas, using D-vac data from 1988-1989 The area of each circle is directly proportional to Heteropteran density Each circle represents a minimum of 103 individuals Treatments are (a) ungrazed controls on Upper Seeds, (b) short-penod spnng grazed on Upper Seeds, (c) short-period autumn grazed on Upper Seeds, (d) heavy autumn grazed on Upper Seeds, (e) spnng-and-autumn grazed on Upper Seeds, (f) spnng-and-autumn grazed on the old tussock grasslands The feeding types are indicated by solid shading (moss feeders), blank (grass feeders), horizontal lines (other pure herbivores), vertical lines (partial predators or unknown), and dot shading (carnivores). Fig 2 The relative abundance of heteropteran species in different treatments on Upper Seeds in 1988-^9 Species are scaled to log(N + l) and ordered by their abundance m ungrazed controls Species with a tenfold or greater difference between treatments are shaded. Treatments are (a) spring- and autumn grazed, (b) heavy autumn grazed, (c) short-penod autumn grazed, (d) short-penod spnng grazed and (e) ungrazed controls Speaes ape:— 1, Nabis flavomargmatus, 2, Dolichonabts limbatus, 3; Stalia major, 4, Stygnocorts pedestris, 5, Plagiognathus chrysanthemi, 6, Notostira elongata, 7, Amblytylus nasutus, 8, Leptoptema dolabrata, 9, Phytocoris varipes, 10; Drymus sylvaticus, 11, Capsus ater, 12; Orthops campestns, 13; Pentrechus gemcuiatus, 14, Lopus decolor, 15, Dictyonota tricomis, 16, Podops inuncta, 17; Berytmus monttvagus, 18, Stenodema calcaratum, 19, Pithanus maerkeli, 20; Tngonotytus ruficornis, 21, Plagiogruithus arbustorum, 22, Anthocons nemorum, 23; Megahcerexa recticomis, 24, Lygus rugidtpenms, 25, Stdduta salUUona, 26; Calocons norvegicus, 27, Tmgis cardm, 28, Onus btevtgatus. ECOGRAPHY 15 2 (1992) plants, Kloet and Hmcks (1977) for beetles, Merrett et al 1985 for spiders, and Southwood and Leston (1959) for Heteroptera. 99-1 Ninber of mdividuals 9- 099- •4 - i J] 99-1 affected in a similar manner by heavy grazing, but not by short-period grazing (Fig lb, c) This change in feeding type composition was produced in an apparently simple manner; as grazing pressure increases more and more species became rare or disappeared Only one species {Capsus ater) was more abundant in a heavily grazed treatment (heavy autumn grazing), and only two (Lopus decolor, Dictyonota tricornis) were most abundant in a short-period grazed treatment (Fig 2) By contrast, eight common species (> 10 individuals m any single treatment in 1988-89) were both commonest in ungrazed controls and had a tenfold difference or more m abundance between grazing treatments The extreme dominance of spnng-andautumn grazing by herbivores (Fig le) is shown to be the result of two common and ubiquitous species {Plagiognathus chrysanthemi and Amblytylus nasutus) being unaffected by any grazing treatment applied (Fig 2a) Coleoptera As with the Heteroptera, individual species were too sparse and/or patchily distributed for individual analysis between treatments and time Forty-four species were recorded, of which only two were confined to the older grasslands The effect of grazing on the fauna (Fig 3) was more complicated than m the Heteroptera As with that group, there were no common species character- 20 Species Number of mdividuQls 99-1 Fig 3 The relative abundance of plant-associated coleopteran species in different treatments on Upper Seeds in 1988-89 Conventions follow Fig 2, but species are - l,Sitorm lineatus, 2, Longitarsus spp , 3, Meligethes aeneus, 4, Longitarsus jacobaeae, 5, Meligethes spp , 6, Apion dichroum, 7, A loti, 8, A seniculus, 9, A trifolu, 10, Oedomera lunda, 11, Rhagonycha femoralis, 12, Rhagonycha fulva, 13, Trachys scrobicutatus, 14, Miccotrogus picirostris, 15, 5jto/jfl humeralis, 16, Apion fihrostre, 17, Anthonomus rubi, 18, Hypera postica, 19, /Ipwn vfreni, 20, 5«ona hispidulus, 21, Orthochaetes setiger, 22, y4pion mmiatum, 23, 4 ptsi, 24, 5«ofia sulcifrons, 25, Sphaeroderma rubidum, 26, 5«o/ifl lepidus, 27, Phaedon tumidulus,28, Apion vorax,29, A carduorum.30,A onopordi,3l, A rufirostre theless, have considerable effects on the total abundance and faunal composition (x^ = 196, df = 12, p<O.(X)l). Carnivorous species were reduced more by grazing withm Upper Seeds than were mixed feeders (Fig 1). This was not so m the older grasslands (Fig. If), where the fauna was dominated by Nabidae in spite of heavy grazing. However, in 1989, these grasslands still retained much of the tussock structure and the deep litter of Brachypodium pinnatum from their previous ungrazed state. By contrast, the heavier grazed areas on Upper Seeds (Fig. le) were a more uniform short turf without tussock or deep htter. The mixed feeders were Species Fig 4 The relative abundance of plant-associated coleopteran species in different groups of five sampling occasions over 12 months in 1985(a), 1986(b) and 1988-89(c) Conventions follow Fig 2, but species 1 to 22 follow Fig 3 and the remaining species are — 23, Ceutorhynchus punctiger, 24, Apion onopordi, 25, Chaetocnema confusa, 26, Apion put, 27, Suona puncticoUis, 28, Rhmoncus pencarpius, 29; Suona sulcifrons, 30; Ceutorhynchus trimaculatus 169 ECOORAPHY 15 2 (1992) lstic of spnng-and-autumn grazing Short-penod spnng grazing and heavy autumn grazing did not dimmish species richness or abundance relative to controls indeed short-penod spnng grazing held marginally more individuals and species than any other treatment. Among common species, the four which showed a tenTreatment AA SA fold or more difference between treatments were all commonest in either short-penod spring grazing or con22 19 17 11 Heteroptera 18 trols 24 14 19 16 Herbivorous Coleoptera 22 The above pattern may be an effect of the large 26 20 23 17 18 Auchenorrhyncha 36 34 27 22 16 numbers of legume feeders m the fauna Both common Leaf miners 40 27 21 17 Spiders 25 legumes in the experimental treatments {Trifolium repens and Medicago lupulma) were strongly affected by b) Numbers of species in groups sampled by D-vac in ungrazed grazing and by 1988-89 largely restncted to grazed controls from 1985 to 1989 treatments (Gibson and Brown in press b) Further interpretation is rendered difficult by the lack of comYear 1985 1986 1988-89 prehensive information on the foodplants of individual species For instance, Apion loti, only known to feed on Herbivorous Coleoptera 15 15 22 Lotus spp (Joy 1932). was common and widespread by Auchenorrhyncha 24 20 26 Spiders 27 26 40 1988-89 This may reflect the use of additional foodplants, but the species is highly mobile (Morns pers c) Numbers of species of leaf miners ir1 ungrazed controls and comm ) Despite the scarcity of L corniculaius on Upshort-period grazing treatments Data are from four matching per Seeds, it is thus impossible to distinguish long disdates for each year tance dispersal from the possible use of additional foodplant sjjecies Year 1986 1987 1989 Changes through the study in ungrazed controls (Fig Ungrazed controls 25 29 30 4) revealed a changeover between four species very Spring grazed 23 28 26 common in 1985 (Fig. 4a) and probably associated with Autumn grazed 22 29 22 Medicago lupulma {Sitona humeralis, Hypera postica, Apion ptsi and 5 puncticollis) and five taxa with a greater variety of feeding habits, including only one {Sitona lineatus) associated with legumes. Of the four Table 2 Comparison of the effect of grazing on the absolute abundance of different invertebrate groups in D-vac samples, using two dates, when all treatments were sampled (August 1988, August 1989) Excluding Diptera and Hymenoptera as "tourists", responses differed between groups (p < 0 001 under f in both 1988 and 1989) 1989 data are included for leaf miners to show the absolute density of this group compared to taxa m the D-vacs Treatment codes follow Table 1 Columns D = density in animals m"-. R = ratio of density to that m ungrazed controls N = total number of animals a) August 1988 D Diptera Hymenoptera Auchenorrhyncha Coleoptera (herb) Heteroptera Spiders b) August 1989 D Diptera Hymenoptera Auchenorrhyncha Coleoptera (herb) Heteroptera Spiders Leaf miners 24 9 21 3 115 32 2 244 C Table 1 Responses of species richness to treatment and successional age m different invertebrate groups a) Total species richness between treatments in 1988-89 Treatments are C = ungrazed controls. S = short-period spnng grazing. A = short-penod autumn grazing, AA = heavy autumn grazing. SA = spnng-and-autumn grazing A R AA R D 209 SA R 0 86 0 82 0 24 2 50 105 060 D 249 127 319 70 15 15 2 SA R 0.62 046 0.51 0.40 0 23 028 034 D 416 20 7 135 118 27 30.4 14 R 102 159 0 24 088 009 0 10 D 268 94 5 513 26 4 25 8 117 S 17 3 160 C 104 A 65 6 32 5 19 7 17 9 %0 AA 3894 1434 1003 245 271 1622 N R 10 D 17.3 10 0 94.5 32 5 12 541 22 R 070 0 47 0 82 101 0 31 D 14 0 15 2 101 18 5 R 0 56 0 71 0.88 0 58 0 62 0 41 0 62 D 15 5 97 58 0 12.8 09 51.9 20 LO ]0 ]0 ]0 ECOORAPHY 15 2 (1992) Table 3 Changes m species m different invertebrate groups as a function of grazing treatment and successional age Species are regarded as changing if there was a tenfold difference or more between pooled data for the five different treatments in 1988-89 or between the groupings of sample dates for ungrazed controls in 1985, 1986 and 1988-89 (D-vacs) or 1986, 1987 and 1989 (leaf miner study) a) Response to grazing treatment Changing Not changing % Age changing Group Herbivorous Coleoptera Spiders Auchenorrhyncha Heteroptera Leaf miners b) Response to time (year) in ungrazed controls Spiders Auchenorrhyncha Leaf miners Herbivorous Coleoptera 5 10 11 9 41 26 27 21 10 9 27 8 28 9 30 0 Data for Heteroptera were not comparable with the remainder species initially common in controls, one (S puncticollis) was not seen at all m 1988-89 (Fig 4c), while the other three persisted in at least small numbers in grazed treatments (Fig. 3) Comparative responses of invertebrate groups Species richness and abundance The manner m which speaes richness declined as grazing intensity increased was similar for the five groups for which comparable data were available (Table la, X = 7 31, df = 16, p > 0 95) In general, short-period ^ grazing treatments contained slightly fewer species than controls. Heavy autumn grazing produced a further small decline, with a larger decline m the spring-andautumn grazed treatments These pattems, on the scale of whole data sets, are similar to those seen in local (quadrat scale) richness in those groups abundant enough to test (Brown et al. unpubl. data, Gibson et al in press) Comparable data were available for four groups (Table lb,c) to examine the change m total nchness over time In all these, there was a net increase over the study period, with most species being taken in 1988-89 Only in leaf-miners were there sufficient data to examine the effect of grazing on this pattem (Table lc). Here, it appears that species nchness in short-penod grazed treatments, unlike controls, did not simply increase over time. ECOORAPHY 15 2 (1992) Changes in abundance between grazing treatments did not present such a simple picture Comparing August samples m the two years when all treatments were monitored (1988 and 1989, Table 2) highlights the unpredictable nature of the groups classed as "tourists" (Diptera and Hymenoptera) "The D-vac samples only the adults of these groups (which were not identified to species), and even though most may originate from Upper Seeds, they are likely to be highly mobile between treatments (cf Hendnx et al 1988) Heavy grazing produced the greatest numbers, especially in the dry summer of 1989 However, this is more likely to be due to the presence of adult feeding attractants (e g fresh sheep dung) than to preference for this type of habitat throughout the life cycle Adults of leaf-minmg Diptera were too sparse in the D-vac samples for their distributions to be compared with those of the larval mines Among other groups, spiders and Heteroptera (rare in 1989 in the drought) were severely affected by grazing, whilst herbivorous Coleoptera and (most of all) Auchenorrhyncha were less affected As with herbivorous Coleoptera, the leafhoppers in spring-and-autumn grazed treatments were numerically dominated by very few species (Brown et al. unpubl data) Leaf-miner densities were also strongly affected by grazing Table 2 also demonstrates the typical abundances of these invertebrate groups The low numbers in some groups restnct the analyses that can be applied Both Auchenorrhyncha and spiders were regularly found at densities of 1(X) m"- or more, allowing analysis by multivariate methods and by the examination of common species alone (Brown et al unpubl data, Gibson et al in press) In this study, Heteroptera and herbivorous Coleoptera never reached such high densities Leafminers, however, could be studied in a much more precise way (Gibson et al unpubl data), so their generally lower densities mattered less Indeed, leafminer densities were unusually low in August 1989, with summer peaks otherwise ranging from 23 6-55.8 mines m"^ in controls Lability of different taxa There was considerable difference between taxa m the extent to which the abundance of species changed with grazing treatment (Table 3a, x^ = 9 98, df = 4, p < 0 05), but less m the proportion of species which changed with Ume (Table 3b). Table 3(a) shows that herbivorous Coleoptera were the group with the smallest proportion of species changing tenfold or more in abundance between grazing treatments, whilst leaf-miners had the greatest proportion changing and are therefore the group in which changes according to grazing management are most likely to be detected Spiders were the only group with a significantly distinct chance of species changing with time (Table 3b, x^ between spiders and other groups' data combmed = 4.80, df = 1, p<0.05). Spi171 Table 4 Fluctuations in leaf-miner species as a function of grazing Species are regarded as changing if there was a tenfold difference or more between pooled data for the four directly comparable dates in 1986. 1987 and 1989 Treatment codes follow Table 1 Grazing Changing species increase decrease fluctuate Other species Total seen in period C S A jjerse Table 5 shows that there are more spiders and Auchenorrhyncha which were slow to appear compared to other groups (/^ = 4.93, df = 1, p < 0 05 after pooling data from other groups to avoid low expected values) Some caution is needed m this interpretation, since plants such as Helianthemum nummulanum and Carex flacca, which were still rare in the Brachypodium pinnatum tussock grasslands studied, support a rich leafminer fauna elsewhere m Wytham (Sterling and Gibson unpubl data) The usefulness of spiders as a single group IS highlighted m their low lability with time compared with an apparently nch fauna in the old grasslands ders which changed through time were all species rare or absent at the beginning of the exf)enment, which subsequently became common (Gibson et al. in press) The other three groups suggested a constant chance of c 30% of a species changing tenfold or more in abundance over three years of sampling The greater precision of the leaf-miner data allowed further analysis, which revealed that an increase in the tendency for species to change as a result of grazing (Table 4) cannot be discounted (x" = 3 8 without Yates' correction, df = 1, p near 0 05) The apparent constancy in the number of species increasing or decreasing between treatments compared to the number fluctuating was notable but there were too few data to test Discussion Plant species composition, structure and prediction of invertebrates The biology of the invertebrate taxa in this study should predict some of their response to vegetation composition and architecture as it changes with grazing management and age of habitat Leaf miners are a purely herbivorous group which are known to contain a high proportion of plant species or genus specialists (Henng 1951, Strong et al. 1983, Godfray 1985) Indeed, many species are even more selective than expected (Brown et al 1990, Gibson et al unpubl data) They should thus be highly sensitive to plant species composition, with responses to plant archiRelationships with plant species and tecture only appeanng as modifications to this type of structure pattern. In this study, leaf-miner assemblies were best explained by axes derived from ordination of the vegeFauna of heavily-grazed grasslands tation, relating to plant species composition (Gibson et The commonest species in the short turf, produced by al unpubl data) TTiere were additional grazing respnng-and-autumn grazing, were those common or sponses needing architectural effects for their explanawidespread in early-successional or arable habitats with tion, such as a reduction in density of the dipterous much bare ground This fauna included a leafmmer miner Agromyza frontella in heavily grazed treatments {Cerodontha denticornis), two leafhoppers {Euscelis in- where leaves of its foodplant Medicago lupulma were cisus and Macrostelis laevis) and three spiders {Erigone small. Such effects via structure were, however, generatra, E dentipalpis and Oedothorax fusciis) No common Heteroptera or phytophagous Coleoptera were most abundant m the short turf In some groups (e g. Table 5 The number of speaes in different invertebrate groups spiders - Gibson et al in press), there were further only found in samples from older grassland patches (OG) sparse species charactenstic of heavy grazing On the Upper Seeds (US) was more intensively sampled for a longer experimental timescale, heavy grazing thus appears to time, so data on species restncted to this field are not comparable and not shown prevent fast successional change in the fauna, whilst allowing the possibility for other rarer species to estabOG US total % restncted lish A similar process appears to be operating in the flora (Gibson and Brown in press b). Group 49 Auchenorrhyncha 11 18 3 Fauna of older grasslamls Some of the rarer species might be expected to be those charactenstic of older grasslands which are slow to dis172 Spiders Heteroptera Leaf mmers' Herbivorous Qdeoptera 11 4 4 'Excluding spedes restricted to woody plants ECOGRAPHY 15 2 (1992) ally weaker than the effects of grazing on miner density via foodplant abundance alone (Gibson et al unpubi data) The feeding specialisations of many herbivorous Coleoptera are probably less well known than those of leaf miners, but they are a relatively specialised group (e g Brown and Hyman 1986, in press), with most of the species reported from this study being herb rather than grass feeders Many species specialise on parts of the plant which are selectively removed by grazing (e g flowers, seeds), so their response should be less simply to plant species composition than the miners In this study, many of the common phytophagous beetles were legume feeders, and the commonest legumes were only retained in grazed treatments, not surprisingly these beetles largely reflected foodplant abundance However, one ubiquitous legume feeder (Sttona Itneatus) was common in the short turf produced by heavy grazing, whilst two other species had disappeared by 1988-89, and two more had declined, in parallel with the decline in their foodplants in ungrazed controls Auchenorrhyncha, m contrast to the leaf-miners and Coleoptera, are dominated by grass- rather than herbfeeders They are well known to need particular features of sward structure (e g Andrzejewska 1965) or nutrient status (Prestidge and McNeiU 1983) which are to some extent independent of species composition It might be expected that their responses would reflect plant architecture in addition to species composition. This expectation was confirmed m the current study Auchenorrhyncha abundance and diversity were more closely related to structural measures, such as vegetation height and density, than to grass species composition Only a few species, such as Zygtntdta scutellaris on Agrostts stolonifera, were directly dependent on the distribution of a particular grass foodplant (Brown et al unpubi data) The Heteroptera is a heterogeneous group containing herbivores as well as mixed feeders and carnivores Many herbivorous species are known to need particular plant structures for feeding or shelter (Southwood and Leston 1959, Moms 1973a, b, Wetton and Gibson 1982), and further studies (Morns 1990b, c,d) have shown them to be particularly sensitive to sward structure In general carnivores, by definition lacking plant species requirements, should be relatively more sensitive than the others to plant architecture Apart from the two ubiquitous herbivores, Amblytylus nasutus and Plagiognathus chrysanthemi, all common Heteroptera in this study were highly sensitive to the structural changes induced by grazing The carnivorous species were more strongly affected by grazing than other feeding groups, the apparent exception m the older grasslands was due to their retention of a tussock structure in 1989, despite one year's grazing Spiders are the only purely carnivorous group studied here. Grassland spiders are well known to contain a range of specialists on different parts and structures of ECOGRAPHV 15 2 (1992) the sward (eg Duffey 1962, 1978, Robinson 1981, Greenstone 1984, de Keer et al 1989) and should be the group most independent of species composition and most dependent on structure Here the responses of spider assemblages were not so simple Although the commonest members of the Upper Seeds spider communities were simply reduced by grazing, with the reduction being directly related to the intensity of grazing, the larger web-spinners were highly dependent on the occurrence of one or two plant species (Gibson et al in press) This was because architecture and species vanables are necessarily confounded' on Upper Seeds Pasttnaca sattva and Senecio jacobaeae were the mam species which provided prominent and complex stems, which often persisted over winter to provide web sites for spiders Duffey (1962), m a study of the old grassland patches reported here but some thirty years previously, noted the same phenomenon, with P sattva and Cynoglossum offtctnale being the main species involved In addition, there was a number of species, sparse in D-vac samples, which appeared to be restricted to older grasslands and/or specialised on short turf (Gibson et al in press) The lability of invertebrate groups Even the short period of this study has suggested that some invertebrate taxa have high proportions of species which are liable to fluctuate Here, herbivorous Coleoptera and spiders were the least likely to fluctuate, whilst leaf miners were the most likely to do so. Auchenorrhyncha, despite Prestidge and McNeill's (1983) emphasis on the number of species fluctuating, were intermediate Although other authors have examined the degree of fluctuation in long-term studies of all invertebrates (reviewed m Strong et al 1983), there appear to be no other comparative data between groups under controlled conditions The lack of fluctuation in spiders and the large amount in leaf miners are likely to be robust results since the amount of data on these groups was considerable There is also a suggestion that grazing increased the likelihood of species changing among leaf-miners, the only group with enough data to test m this way If this is borne out as the study continues, the simplest explanation would be that, as with the plants, grazing increases the chance of colonisation (here because of increased foodplant diversity), but also imposes a "disturbance" which increases the chance of extinction (Gibson and Brown in press a). Turnover is thus generally increased and the number of species under grazing is determined by a balance between colonisation and extinction rates, m a phase of succession when ungrazed treatments are undergoing simple accumulation of species Such dynamics could produce ultimately either situations where "undisturbed" communities have the greatest diversity, or where intermediate levels of disturbance favour di173 Table 6 Possible criteria for the choice of "indicator" groups from those studied Herbivorous Coleoptera Accessible to sampling Typical abundance in usual sample size and method Identification' Availability to ldentifyResponse to management Extent of fluctuation Characteristic common species m all treatments Species restricted to old grassland Fair Poor Poor Poor Low Medium No Few Spiders Good Good Good Fair Fair Leafminers Very good Fair Fair Poor High Medium Yes (Many'') Heteroptera Good Poor Good Good Fair Unknown No Few Leafhoppers Good Good Fair Fair Fair Low Yes Many Medium Yes Many '' Identification" is scored as "good" if nearly all species can be easily identified as adults, "fair" if there are a few groups of species which cannot be separated without extensive work and "poor" if there are many such critical species, including very common ones -"Availability to identify" is a judgement based on the accessibility and existence of up-to-date literature and the number of experts in the group versity (cf Tansley 1939, Connell 1979, Huston 1979), as seen m the plants (Gibson and Brown in press a) "Indicator" taxa - do they exist? The choice of invertebrate taxa in studies of management, succession or other ecological investigations is rarely arbitrary groups are chosen on some sort of compromise between theoretical and convenience reasons (e g Brown and Southwood 1983) It is generally agreed that a chosen group must be large enough for responses to treatment to depend on at least several species Despite this, groups such as British butterflies are often used because of their ease of identification (Nature Conservancy Council 1986) Here, data from quantitative (leaf-miner quadrats) and semi-quantitative (vacuum sample) studies allow the companson of five groups of invertebrates Practical reasons for choice would suggest taxa with an efficient sampling method, whose typical sample size is large compared to the animals' usual density, ease of identification, both with resjject of availability of literature and experts and to the amount of work (e.g dissection) that has to be done for reliable identification Theoretical reasons for choice would suggest taxa which show clear responses to management compared with temporal fluctuations, contain common species characteristic of all treatments applied and, where studies cover a successional senes, contain species charactenstic of all stages. Table 6 shows that, not surprisingly, no single taxon meets all these cntena. The single group which meets most of them is the spiders, their only drawback is that vacuum sampling only addresses a part of the spider fauna. The remaining fauna, which is too sparse or has individuals too heavy or concealed for D-vacs, contains many species with strong responses to habitat change (Duffey 1962) Sampling methods for this remaining fauna are time-consuming or non-quantitative Herbivorous Coleoptera emerge f)oorly from this study, their single most important drawback is their low typical density (see also Brown and Hyman 1986, in press. Morns and Rispin 1987) Plant-based sampling methods may show them to be a more generally useful group Heteroptera in this study were also generally too sparse to be useful Two groups stand out both for practical and theoretical reasons Leaf-hoppers have already been widely used in community studies (e g. Waloff 1980, Moms 1981a, b, 1990b, c,d, Prestidge and McNeill 1983). their responses m this study have been shown to be much more sophisticated than the simple response to grass architecture than has sometimes been suggested before (e g Andrzejewska 1965) Leaf-miners have been neglected because of feared difficulties in identification, with notable exceptions (Godfray 1982, 1985, Brown et al 1990) This study suggests that they should be in the first rank of consideration for community studies or for the choice of "indicator groups". ImpUcations for management of grassland Each group studied has produced quantitatively and quahtatively different information about the way in which grazing affects the invertebrate fauna Only among leaf-miners, and to a lesser extent Auchenorrhyncha, could examples of each different type of grazing response be found. Each grazing treatment also produced a distinct fauna, with its own balance between different groups There are two direct consequences for management and monitoring of grasslands. First, there is a severe hmit to the amount and type of infonnation that a single ECOGRAPHY 15 2 (1992) invertebrate group can give about the whole fauna Secondly, management goals must be extremely clear about their desired effects on particular taxa Rotational management (Morns 1990a) is not necessarily "good" or "bad", nor is grazing "too heavy", "too light", "too early in the season" or "too late" the value of management is only relevant with reference to clearly defined goals A particular example concerns the use of intensive management to produce a close-cropped turf, often recommended on traditional grounds, to maximise plant diversity (Smith 1980) or needed for the conservation of particular rarities m Britain (e g Thomas 1983, Thomas et al 1986, Cherrill and Brown 1990a, b) The young grasslands on Upper Seeds suggest that the predominant fauna associated with this management is a numerically nch but species-poor one dominated by common animals of open ground, early successional or arable land The httle evidence from study of closely grazed chalk grasslands tends to support this With the caution that more studies of the fauna of closelycropped downlands are needed, this suggests that management successfully designed to promote a number of sparse rarities may inevitably lmpovensh the more abundant compionents of the fauna Acknowledgements - This study was made possible by generous support from the Nuffield Foundation 1985-1987 and continued with support from the Nature Conservancy Council 1988-90 Experiments on the older grasslands were set up with the aid of a grant from the Nature Conservancy Council's Southem Region We also thank BIOSCAN (UK) Ltd for support from 1988 Grazing management has been enabled by the staff of Oxford Umv Expenmental Farm at Wytham, and Oxford Umv Wytham Management Committee gave permission for the land to be used Data collection would not have been possible without the cheerful assistance of a growing band of field recorders We thank M G Morns for helpful comments on a previous version of the manuscript http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Ecography Wiley

The response of invertebrate assemblies to grazing

Ecography , Volume 15 (2) – Apr 1, 1992

Loading next page...
 
/lp/wiley/the-response-of-invertebrate-assemblies-to-grazing-JjOieBdPgq

References (54)

Publisher
Wiley
Copyright
Copyright © 1992 Wiley Subscription Services, Inc., A Wiley Company
ISSN
0906-7590
eISSN
1600-0587
DOI
10.1111/j.1600-0587.1992.tb00020.x
Publisher site
See Article on Publisher Site

Abstract

Gibson, C W D . Brown, V K , Losito, L and McGavin, G C 1992 The response of invertebrate assemblies to grazing - Ecography 15 166-176 Invertebrate communities were monitored in a controlled sheep grazing experiment on ex-arable land on limestone and in nearby older calcicolous grasslands Results for vacuum sampling of Heteroptera and herbivorous Coleoptera are presented and compared with the responses of spiders, leafhoppers and leaf-miners Invertebrate groups differed in their responses to grazing and grassland age in terms of species diversity, abundance, variability over time and the development of a specific fauna on short turf Leaf-miner assemblies were most strongly determined by the effects of grazing on plant species composition, while spiders' responses could largely be explained by the effects of grazing on plant architecture Spider species simply accumulated over time, whilst leaf-miners were the most labile group Spiders, leaf-miners and leafhoppers all contained some common species restricted to old grasslands, whilst herbivorous Coleoptera and Heteroptera did not No single group demonstrated the full range of invertebrate responses Spiders, leafhoppers and leaf-miners each showed useful features as "indicator groups'", but more than one group should be chosen for study in the context of particular research projects C W D Gibson and V K Brown, Imperial College at Silwood Park, Ascot, Berkshire, SL5 7PY, U K -L Losuo, 30 Cope Close, Botley. Oxford OX2 9AJ, U K G C McGavin, Umv Museum, Parks Road, Oxford OXl 3PW, U K Introduction The change in invertebrate communities dunng succession rarely follows changes in vegetation composition alone (but see Brown and Hyman 1986) Reality is not as simple as this Invertebrates often show complex resfKjnses because plant architecture may be as important as plant species composition (Southwood et al 1979) In addition, their resjKwises may be related to the sequencing of immigration and to hfe-histones (Brown 1986, Southwood et al. 1988), and to the effects of interactions within and between trophic levels (Lawton 1988). Invertebrate assemblies associated with vegetation influenced by grazing or cutting are known to show a vanety of responses compared with unmanaged areas In grassland, invertebrate speaes nchness and abundance may be dimmtshed in grazed or cut areas due to the lack of architectural diversity (Morns 1973a, 1990a). Accepted 13 February 1991 © ECOGRAPHY However, some species may require the particular features of microchmate and/or structure which only exist m grazed grasslands (e g Thomas 1983), or even need grasslands with a mosaic of structure to provide conditions for different phases of the life cycle (e g Chernll and Brown 1990a, b, Weidemann et al 1990) This picture is complicated by the apparent trade-off between plant species diversity and structural diversity in grasslands. Generally, the more plant species there are in a grassland, the more invertebrate herbivore species. However, management which maximises plant species diversity tends to decrease architectural diversity and thus precludes many species from establishing One compromise which has been suggested to maximise invertebrate diversity is rotational management (Bacon 1990, Morris 1990a), where grasslands are grazed or cut at intervals, not too short to decrease plant species diversity but long enough to allow architectural diversity to develop This recommendation can be regarded ECOGRAPHY 15 2 (1992) as an applied version of hypotheses that intermediate levels or rates of disturbance tend to maximise diversity (Connell 1979, Huston 1979) The different responses of invertebrate groups to plant species diversity and structure have produced severe problems in understanding invertebrate community development and in designing management to maximise diversity Among studies which have addressed more than one invertebrate group. Moms and colleagues found that many Heteroptera were simply diminished by increased cutting or grazing (Morns 1973b), while herbivorous Coleoptera appeared to be less affected (Morns and Rispin 1987) Since the amount of work involved m sampling all invertebrate groups IS usually prohibitive, objective methods are needed for choosing groups which are representative either of the response of the whole community, or the wildlife "value" to be expected under different management regimes Here, we report on the first six years of an experimental study where the changes in a number of invertebrate groups are examined dunng early secondary succession under different grazing regimes. These changes are viewed against the background of the fauna of older grazed grasslands Particular attention is paid to the way in which different taxa respond to grazing management and to establishing the nature of the communities that develop, especially the assemblies which are associated with the short turf produced by heavy grazing We report detailed results for two components of the fauna, the Heteroptera and phytophagous Coleoptera, but compare their responses with other taxa studied, mainly the auchenorrhynchan Homoptera, the leaf-miners and spiders (Brown et al unpubl data, Gibson et al m press ) studied The companson allows a judgement of whether there is an "ideal" invertebrate group which reflects widespread community changes, or whether there is no alternative to studying a wide vanety of taxa if invertebrate community development is to be understood quadrat recording was started in 1984, areas were not fenced until early 1985 Deer rarely enter the experimental paddocks The field lies within the University of Oxford's Wytham estate (1240 ha of which c 410 ha contains semt-natural vegetation) It is surrounded by woodland, but also abuts onto three patches of old-established hmestone grassland with varying degrees of scrub encroachment Ten further patches of ancient grassland are scattered elsewhere withm the woodlands (Gibson 1986) and these thirteen patches form the local potential source for calcicolous grassland flora and fauna (Gibson and Brown m press a) All the grassland patches and the experimental field he on Jurassic coralhan limestone (Arkell 1947) there are no other old hmestone grasslands within 5 km of the site and the nearest substantial areas are >15 km away Sampling methods D-vac samples were taken on five occasions (May, June, July, Aug , Sep in 1985 and 1986, two occasions (early Aug , early Oct ) m 1988 and three occasions (May, July and Aug ) in 1989 Ungrazed controls were sampled from May 1985, short-period spnng and autumn grazed from July 1985 onwards, although the first grazing penod m the latter was not until Sep. 1985 Heavily grazed treatments in Upper Seeds were sampled from 1988 onwards, and the two older grassland patches with spring-and-autumn grazing in 1989 only In paddocks, a sample unit of three sucks (total 0 274 m-) was taken from random positions m each of two 3 x 3 m plots where vegetation was assessed regularly by point quadrats (Gibson et al 1987b) In other treatments, similar samples were taken from notional 3 x 3 m squares surrounding each of twelve 1 x 1 m permanent quadrats in each treatment or area, where vegetation was also assessed Heteroptera and herbivorous Coleoptera were sorted from other invertebrates, counted and identified "bhnd" (unaware of which treatment which sample belonged to), by G McG and L. L respectively, to species or nearest taxon possible Other Coleoptera were Ignored as D-vac suction is known to be lneffiaent for their extraction (Duffey 1980) and may vary considerably according to vegetation structure (Morns and Rispin 1987). The minimum number of species in a sample was estimated by counting lmmatures as an extra species only if none of the species they could have belonged to was present as an adult The phytophagous Coleoptera were mainly herb-feeders Heteroptera were assigned to one of SIX feeding types" "moss-feeders", "grass-feeders", "other herbivores" (herb-feeders and generalist herbivores), "mixed feeders" (partial predators), "carnivores" and "tounsts" (sensu Southwood et al. 1982). Nomenclature follows Clapham et al (1987) for Methods Study area Grazing experiments took place m a 10 ha field abandoned from agriculture m 1982 (Upper Seeds) Details of the expenmental design and grazing regimes were given by Gibson et al. (1987). Briefly, eighteen 30 x 30 m paddocks have been arranged in two Latin squares (grids) where individual paddocks have been grazed by sheep each autumn, grazed each spring or left ungrazed, since 1985. Outside the paddocks, the field has been divided into areas either grazed by sheep from Apni to November ("SA"), grazed from late August to November ("AA") or left ungrazsd, but with free access to deer (Dama dama and Mimttacus reevesi). Although ECOORAPHY 15 2 (1»2) Results Heteroptera Forty-one species of Heteroptera were recorded in the D-vac samples, of which 37 were found m the experimental treatments on Upper Seeds However, no single species was regularly common enough to test the effects of grazing on individual species Grazing did, never- b) c) Fig 1 The composition of Heteroptera by feeding type m different grazing treatments and areas, using D-vac data from 1988-1989 The area of each circle is directly proportional to Heteropteran density Each circle represents a minimum of 103 individuals Treatments are (a) ungrazed controls on Upper Seeds, (b) short-penod spnng grazed on Upper Seeds, (c) short-period autumn grazed on Upper Seeds, (d) heavy autumn grazed on Upper Seeds, (e) spnng-and-autumn grazed on Upper Seeds, (f) spnng-and-autumn grazed on the old tussock grasslands The feeding types are indicated by solid shading (moss feeders), blank (grass feeders), horizontal lines (other pure herbivores), vertical lines (partial predators or unknown), and dot shading (carnivores). Fig 2 The relative abundance of heteropteran species in different treatments on Upper Seeds in 1988-^9 Species are scaled to log(N + l) and ordered by their abundance m ungrazed controls Species with a tenfold or greater difference between treatments are shaded. Treatments are (a) spring- and autumn grazed, (b) heavy autumn grazed, (c) short-penod autumn grazed, (d) short-penod spnng grazed and (e) ungrazed controls Speaes ape:— 1, Nabis flavomargmatus, 2, Dolichonabts limbatus, 3; Stalia major, 4, Stygnocorts pedestris, 5, Plagiognathus chrysanthemi, 6, Notostira elongata, 7, Amblytylus nasutus, 8, Leptoptema dolabrata, 9, Phytocoris varipes, 10; Drymus sylvaticus, 11, Capsus ater, 12; Orthops campestns, 13; Pentrechus gemcuiatus, 14, Lopus decolor, 15, Dictyonota tricomis, 16, Podops inuncta, 17; Berytmus monttvagus, 18, Stenodema calcaratum, 19, Pithanus maerkeli, 20; Tngonotytus ruficornis, 21, Plagiogruithus arbustorum, 22, Anthocons nemorum, 23; Megahcerexa recticomis, 24, Lygus rugidtpenms, 25, Stdduta salUUona, 26; Calocons norvegicus, 27, Tmgis cardm, 28, Onus btevtgatus. ECOGRAPHY 15 2 (1992) plants, Kloet and Hmcks (1977) for beetles, Merrett et al 1985 for spiders, and Southwood and Leston (1959) for Heteroptera. 99-1 Ninber of mdividuals 9- 099- •4 - i J] 99-1 affected in a similar manner by heavy grazing, but not by short-period grazing (Fig lb, c) This change in feeding type composition was produced in an apparently simple manner; as grazing pressure increases more and more species became rare or disappeared Only one species {Capsus ater) was more abundant in a heavily grazed treatment (heavy autumn grazing), and only two (Lopus decolor, Dictyonota tricornis) were most abundant in a short-period grazed treatment (Fig 2) By contrast, eight common species (> 10 individuals m any single treatment in 1988-89) were both commonest in ungrazed controls and had a tenfold difference or more m abundance between grazing treatments The extreme dominance of spnng-andautumn grazing by herbivores (Fig le) is shown to be the result of two common and ubiquitous species {Plagiognathus chrysanthemi and Amblytylus nasutus) being unaffected by any grazing treatment applied (Fig 2a) Coleoptera As with the Heteroptera, individual species were too sparse and/or patchily distributed for individual analysis between treatments and time Forty-four species were recorded, of which only two were confined to the older grasslands The effect of grazing on the fauna (Fig 3) was more complicated than m the Heteroptera As with that group, there were no common species character- 20 Species Number of mdividuQls 99-1 Fig 3 The relative abundance of plant-associated coleopteran species in different treatments on Upper Seeds in 1988-89 Conventions follow Fig 2, but species are - l,Sitorm lineatus, 2, Longitarsus spp , 3, Meligethes aeneus, 4, Longitarsus jacobaeae, 5, Meligethes spp , 6, Apion dichroum, 7, A loti, 8, A seniculus, 9, A trifolu, 10, Oedomera lunda, 11, Rhagonycha femoralis, 12, Rhagonycha fulva, 13, Trachys scrobicutatus, 14, Miccotrogus picirostris, 15, 5jto/jfl humeralis, 16, Apion fihrostre, 17, Anthonomus rubi, 18, Hypera postica, 19, /Ipwn vfreni, 20, 5«ona hispidulus, 21, Orthochaetes setiger, 22, y4pion mmiatum, 23, 4 ptsi, 24, 5«ofia sulcifrons, 25, Sphaeroderma rubidum, 26, 5«o/ifl lepidus, 27, Phaedon tumidulus,28, Apion vorax,29, A carduorum.30,A onopordi,3l, A rufirostre theless, have considerable effects on the total abundance and faunal composition (x^ = 196, df = 12, p<O.(X)l). Carnivorous species were reduced more by grazing withm Upper Seeds than were mixed feeders (Fig 1). This was not so m the older grasslands (Fig. If), where the fauna was dominated by Nabidae in spite of heavy grazing. However, in 1989, these grasslands still retained much of the tussock structure and the deep litter of Brachypodium pinnatum from their previous ungrazed state. By contrast, the heavier grazed areas on Upper Seeds (Fig. le) were a more uniform short turf without tussock or deep htter. The mixed feeders were Species Fig 4 The relative abundance of plant-associated coleopteran species in different groups of five sampling occasions over 12 months in 1985(a), 1986(b) and 1988-89(c) Conventions follow Fig 2, but species 1 to 22 follow Fig 3 and the remaining species are — 23, Ceutorhynchus punctiger, 24, Apion onopordi, 25, Chaetocnema confusa, 26, Apion put, 27, Suona puncticoUis, 28, Rhmoncus pencarpius, 29; Suona sulcifrons, 30; Ceutorhynchus trimaculatus 169 ECOORAPHY 15 2 (1992) lstic of spnng-and-autumn grazing Short-penod spnng grazing and heavy autumn grazing did not dimmish species richness or abundance relative to controls indeed short-penod spnng grazing held marginally more individuals and species than any other treatment. Among common species, the four which showed a tenTreatment AA SA fold or more difference between treatments were all commonest in either short-penod spring grazing or con22 19 17 11 Heteroptera 18 trols 24 14 19 16 Herbivorous Coleoptera 22 The above pattern may be an effect of the large 26 20 23 17 18 Auchenorrhyncha 36 34 27 22 16 numbers of legume feeders m the fauna Both common Leaf miners 40 27 21 17 Spiders 25 legumes in the experimental treatments {Trifolium repens and Medicago lupulma) were strongly affected by b) Numbers of species in groups sampled by D-vac in ungrazed grazing and by 1988-89 largely restncted to grazed controls from 1985 to 1989 treatments (Gibson and Brown in press b) Further interpretation is rendered difficult by the lack of comYear 1985 1986 1988-89 prehensive information on the foodplants of individual species For instance, Apion loti, only known to feed on Herbivorous Coleoptera 15 15 22 Lotus spp (Joy 1932). was common and widespread by Auchenorrhyncha 24 20 26 Spiders 27 26 40 1988-89 This may reflect the use of additional foodplants, but the species is highly mobile (Morns pers c) Numbers of species of leaf miners ir1 ungrazed controls and comm ) Despite the scarcity of L corniculaius on Upshort-period grazing treatments Data are from four matching per Seeds, it is thus impossible to distinguish long disdates for each year tance dispersal from the possible use of additional foodplant sjjecies Year 1986 1987 1989 Changes through the study in ungrazed controls (Fig Ungrazed controls 25 29 30 4) revealed a changeover between four species very Spring grazed 23 28 26 common in 1985 (Fig. 4a) and probably associated with Autumn grazed 22 29 22 Medicago lupulma {Sitona humeralis, Hypera postica, Apion ptsi and 5 puncticollis) and five taxa with a greater variety of feeding habits, including only one {Sitona lineatus) associated with legumes. Of the four Table 2 Comparison of the effect of grazing on the absolute abundance of different invertebrate groups in D-vac samples, using two dates, when all treatments were sampled (August 1988, August 1989) Excluding Diptera and Hymenoptera as "tourists", responses differed between groups (p < 0 001 under f in both 1988 and 1989) 1989 data are included for leaf miners to show the absolute density of this group compared to taxa m the D-vacs Treatment codes follow Table 1 Columns D = density in animals m"-. R = ratio of density to that m ungrazed controls N = total number of animals a) August 1988 D Diptera Hymenoptera Auchenorrhyncha Coleoptera (herb) Heteroptera Spiders b) August 1989 D Diptera Hymenoptera Auchenorrhyncha Coleoptera (herb) Heteroptera Spiders Leaf miners 24 9 21 3 115 32 2 244 C Table 1 Responses of species richness to treatment and successional age m different invertebrate groups a) Total species richness between treatments in 1988-89 Treatments are C = ungrazed controls. S = short-period spnng grazing. A = short-penod autumn grazing, AA = heavy autumn grazing. SA = spnng-and-autumn grazing A R AA R D 209 SA R 0 86 0 82 0 24 2 50 105 060 D 249 127 319 70 15 15 2 SA R 0.62 046 0.51 0.40 0 23 028 034 D 416 20 7 135 118 27 30.4 14 R 102 159 0 24 088 009 0 10 D 268 94 5 513 26 4 25 8 117 S 17 3 160 C 104 A 65 6 32 5 19 7 17 9 %0 AA 3894 1434 1003 245 271 1622 N R 10 D 17.3 10 0 94.5 32 5 12 541 22 R 070 0 47 0 82 101 0 31 D 14 0 15 2 101 18 5 R 0 56 0 71 0.88 0 58 0 62 0 41 0 62 D 15 5 97 58 0 12.8 09 51.9 20 LO ]0 ]0 ]0 ECOORAPHY 15 2 (1992) Table 3 Changes m species m different invertebrate groups as a function of grazing treatment and successional age Species are regarded as changing if there was a tenfold difference or more between pooled data for the five different treatments in 1988-89 or between the groupings of sample dates for ungrazed controls in 1985, 1986 and 1988-89 (D-vacs) or 1986, 1987 and 1989 (leaf miner study) a) Response to grazing treatment Changing Not changing % Age changing Group Herbivorous Coleoptera Spiders Auchenorrhyncha Heteroptera Leaf miners b) Response to time (year) in ungrazed controls Spiders Auchenorrhyncha Leaf miners Herbivorous Coleoptera 5 10 11 9 41 26 27 21 10 9 27 8 28 9 30 0 Data for Heteroptera were not comparable with the remainder species initially common in controls, one (S puncticollis) was not seen at all m 1988-89 (Fig 4c), while the other three persisted in at least small numbers in grazed treatments (Fig. 3) Comparative responses of invertebrate groups Species richness and abundance The manner m which speaes richness declined as grazing intensity increased was similar for the five groups for which comparable data were available (Table la, X = 7 31, df = 16, p > 0 95) In general, short-period ^ grazing treatments contained slightly fewer species than controls. Heavy autumn grazing produced a further small decline, with a larger decline m the spring-andautumn grazed treatments These pattems, on the scale of whole data sets, are similar to those seen in local (quadrat scale) richness in those groups abundant enough to test (Brown et al. unpubl. data, Gibson et al in press) Comparable data were available for four groups (Table lb,c) to examine the change m total nchness over time In all these, there was a net increase over the study period, with most species being taken in 1988-89 Only in leaf-miners were there sufficient data to examine the effect of grazing on this pattem (Table lc). Here, it appears that species nchness in short-penod grazed treatments, unlike controls, did not simply increase over time. ECOORAPHY 15 2 (1992) Changes in abundance between grazing treatments did not present such a simple picture Comparing August samples m the two years when all treatments were monitored (1988 and 1989, Table 2) highlights the unpredictable nature of the groups classed as "tourists" (Diptera and Hymenoptera) "The D-vac samples only the adults of these groups (which were not identified to species), and even though most may originate from Upper Seeds, they are likely to be highly mobile between treatments (cf Hendnx et al 1988) Heavy grazing produced the greatest numbers, especially in the dry summer of 1989 However, this is more likely to be due to the presence of adult feeding attractants (e g fresh sheep dung) than to preference for this type of habitat throughout the life cycle Adults of leaf-minmg Diptera were too sparse in the D-vac samples for their distributions to be compared with those of the larval mines Among other groups, spiders and Heteroptera (rare in 1989 in the drought) were severely affected by grazing, whilst herbivorous Coleoptera and (most of all) Auchenorrhyncha were less affected As with herbivorous Coleoptera, the leafhoppers in spring-and-autumn grazed treatments were numerically dominated by very few species (Brown et al. unpubl data) Leaf-miner densities were also strongly affected by grazing Table 2 also demonstrates the typical abundances of these invertebrate groups The low numbers in some groups restnct the analyses that can be applied Both Auchenorrhyncha and spiders were regularly found at densities of 1(X) m"- or more, allowing analysis by multivariate methods and by the examination of common species alone (Brown et al unpubl data, Gibson et al in press) In this study, Heteroptera and herbivorous Coleoptera never reached such high densities Leafminers, however, could be studied in a much more precise way (Gibson et al unpubl data), so their generally lower densities mattered less Indeed, leafminer densities were unusually low in August 1989, with summer peaks otherwise ranging from 23 6-55.8 mines m"^ in controls Lability of different taxa There was considerable difference between taxa m the extent to which the abundance of species changed with grazing treatment (Table 3a, x^ = 9 98, df = 4, p < 0 05), but less m the proportion of species which changed with Ume (Table 3b). Table 3(a) shows that herbivorous Coleoptera were the group with the smallest proportion of species changing tenfold or more in abundance between grazing treatments, whilst leaf-miners had the greatest proportion changing and are therefore the group in which changes according to grazing management are most likely to be detected Spiders were the only group with a significantly distinct chance of species changing with time (Table 3b, x^ between spiders and other groups' data combmed = 4.80, df = 1, p<0.05). Spi171 Table 4 Fluctuations in leaf-miner species as a function of grazing Species are regarded as changing if there was a tenfold difference or more between pooled data for the four directly comparable dates in 1986. 1987 and 1989 Treatment codes follow Table 1 Grazing Changing species increase decrease fluctuate Other species Total seen in period C S A jjerse Table 5 shows that there are more spiders and Auchenorrhyncha which were slow to appear compared to other groups (/^ = 4.93, df = 1, p < 0 05 after pooling data from other groups to avoid low expected values) Some caution is needed m this interpretation, since plants such as Helianthemum nummulanum and Carex flacca, which were still rare in the Brachypodium pinnatum tussock grasslands studied, support a rich leafminer fauna elsewhere m Wytham (Sterling and Gibson unpubl data) The usefulness of spiders as a single group IS highlighted m their low lability with time compared with an apparently nch fauna in the old grasslands ders which changed through time were all species rare or absent at the beginning of the exf)enment, which subsequently became common (Gibson et al. in press) The other three groups suggested a constant chance of c 30% of a species changing tenfold or more in abundance over three years of sampling The greater precision of the leaf-miner data allowed further analysis, which revealed that an increase in the tendency for species to change as a result of grazing (Table 4) cannot be discounted (x" = 3 8 without Yates' correction, df = 1, p near 0 05) The apparent constancy in the number of species increasing or decreasing between treatments compared to the number fluctuating was notable but there were too few data to test Discussion Plant species composition, structure and prediction of invertebrates The biology of the invertebrate taxa in this study should predict some of their response to vegetation composition and architecture as it changes with grazing management and age of habitat Leaf miners are a purely herbivorous group which are known to contain a high proportion of plant species or genus specialists (Henng 1951, Strong et al. 1983, Godfray 1985) Indeed, many species are even more selective than expected (Brown et al 1990, Gibson et al unpubl data) They should thus be highly sensitive to plant species composition, with responses to plant archiRelationships with plant species and tecture only appeanng as modifications to this type of structure pattern. In this study, leaf-miner assemblies were best explained by axes derived from ordination of the vegeFauna of heavily-grazed grasslands tation, relating to plant species composition (Gibson et The commonest species in the short turf, produced by al unpubl data) TTiere were additional grazing respnng-and-autumn grazing, were those common or sponses needing architectural effects for their explanawidespread in early-successional or arable habitats with tion, such as a reduction in density of the dipterous much bare ground This fauna included a leafmmer miner Agromyza frontella in heavily grazed treatments {Cerodontha denticornis), two leafhoppers {Euscelis in- where leaves of its foodplant Medicago lupulma were cisus and Macrostelis laevis) and three spiders {Erigone small. Such effects via structure were, however, generatra, E dentipalpis and Oedothorax fusciis) No common Heteroptera or phytophagous Coleoptera were most abundant m the short turf In some groups (e g. Table 5 The number of speaes in different invertebrate groups spiders - Gibson et al in press), there were further only found in samples from older grassland patches (OG) sparse species charactenstic of heavy grazing On the Upper Seeds (US) was more intensively sampled for a longer experimental timescale, heavy grazing thus appears to time, so data on species restncted to this field are not comparable and not shown prevent fast successional change in the fauna, whilst allowing the possibility for other rarer species to estabOG US total % restncted lish A similar process appears to be operating in the flora (Gibson and Brown in press b). Group 49 Auchenorrhyncha 11 18 3 Fauna of older grasslamls Some of the rarer species might be expected to be those charactenstic of older grasslands which are slow to dis172 Spiders Heteroptera Leaf mmers' Herbivorous Qdeoptera 11 4 4 'Excluding spedes restricted to woody plants ECOGRAPHY 15 2 (1992) ally weaker than the effects of grazing on miner density via foodplant abundance alone (Gibson et al unpubi data) The feeding specialisations of many herbivorous Coleoptera are probably less well known than those of leaf miners, but they are a relatively specialised group (e g Brown and Hyman 1986, in press), with most of the species reported from this study being herb rather than grass feeders Many species specialise on parts of the plant which are selectively removed by grazing (e g flowers, seeds), so their response should be less simply to plant species composition than the miners In this study, many of the common phytophagous beetles were legume feeders, and the commonest legumes were only retained in grazed treatments, not surprisingly these beetles largely reflected foodplant abundance However, one ubiquitous legume feeder (Sttona Itneatus) was common in the short turf produced by heavy grazing, whilst two other species had disappeared by 1988-89, and two more had declined, in parallel with the decline in their foodplants in ungrazed controls Auchenorrhyncha, m contrast to the leaf-miners and Coleoptera, are dominated by grass- rather than herbfeeders They are well known to need particular features of sward structure (e g Andrzejewska 1965) or nutrient status (Prestidge and McNeiU 1983) which are to some extent independent of species composition It might be expected that their responses would reflect plant architecture in addition to species composition. This expectation was confirmed m the current study Auchenorrhyncha abundance and diversity were more closely related to structural measures, such as vegetation height and density, than to grass species composition Only a few species, such as Zygtntdta scutellaris on Agrostts stolonifera, were directly dependent on the distribution of a particular grass foodplant (Brown et al unpubi data) The Heteroptera is a heterogeneous group containing herbivores as well as mixed feeders and carnivores Many herbivorous species are known to need particular plant structures for feeding or shelter (Southwood and Leston 1959, Moms 1973a, b, Wetton and Gibson 1982), and further studies (Morns 1990b, c,d) have shown them to be particularly sensitive to sward structure In general carnivores, by definition lacking plant species requirements, should be relatively more sensitive than the others to plant architecture Apart from the two ubiquitous herbivores, Amblytylus nasutus and Plagiognathus chrysanthemi, all common Heteroptera in this study were highly sensitive to the structural changes induced by grazing The carnivorous species were more strongly affected by grazing than other feeding groups, the apparent exception m the older grasslands was due to their retention of a tussock structure in 1989, despite one year's grazing Spiders are the only purely carnivorous group studied here. Grassland spiders are well known to contain a range of specialists on different parts and structures of ECOGRAPHV 15 2 (1992) the sward (eg Duffey 1962, 1978, Robinson 1981, Greenstone 1984, de Keer et al 1989) and should be the group most independent of species composition and most dependent on structure Here the responses of spider assemblages were not so simple Although the commonest members of the Upper Seeds spider communities were simply reduced by grazing, with the reduction being directly related to the intensity of grazing, the larger web-spinners were highly dependent on the occurrence of one or two plant species (Gibson et al in press) This was because architecture and species vanables are necessarily confounded' on Upper Seeds Pasttnaca sattva and Senecio jacobaeae were the mam species which provided prominent and complex stems, which often persisted over winter to provide web sites for spiders Duffey (1962), m a study of the old grassland patches reported here but some thirty years previously, noted the same phenomenon, with P sattva and Cynoglossum offtctnale being the main species involved In addition, there was a number of species, sparse in D-vac samples, which appeared to be restricted to older grasslands and/or specialised on short turf (Gibson et al in press) The lability of invertebrate groups Even the short period of this study has suggested that some invertebrate taxa have high proportions of species which are liable to fluctuate Here, herbivorous Coleoptera and spiders were the least likely to fluctuate, whilst leaf miners were the most likely to do so. Auchenorrhyncha, despite Prestidge and McNeill's (1983) emphasis on the number of species fluctuating, were intermediate Although other authors have examined the degree of fluctuation in long-term studies of all invertebrates (reviewed m Strong et al 1983), there appear to be no other comparative data between groups under controlled conditions The lack of fluctuation in spiders and the large amount in leaf miners are likely to be robust results since the amount of data on these groups was considerable There is also a suggestion that grazing increased the likelihood of species changing among leaf-miners, the only group with enough data to test m this way If this is borne out as the study continues, the simplest explanation would be that, as with the plants, grazing increases the chance of colonisation (here because of increased foodplant diversity), but also imposes a "disturbance" which increases the chance of extinction (Gibson and Brown in press a). Turnover is thus generally increased and the number of species under grazing is determined by a balance between colonisation and extinction rates, m a phase of succession when ungrazed treatments are undergoing simple accumulation of species Such dynamics could produce ultimately either situations where "undisturbed" communities have the greatest diversity, or where intermediate levels of disturbance favour di173 Table 6 Possible criteria for the choice of "indicator" groups from those studied Herbivorous Coleoptera Accessible to sampling Typical abundance in usual sample size and method Identification' Availability to ldentifyResponse to management Extent of fluctuation Characteristic common species m all treatments Species restricted to old grassland Fair Poor Poor Poor Low Medium No Few Spiders Good Good Good Fair Fair Leafminers Very good Fair Fair Poor High Medium Yes (Many'') Heteroptera Good Poor Good Good Fair Unknown No Few Leafhoppers Good Good Fair Fair Fair Low Yes Many Medium Yes Many '' Identification" is scored as "good" if nearly all species can be easily identified as adults, "fair" if there are a few groups of species which cannot be separated without extensive work and "poor" if there are many such critical species, including very common ones -"Availability to identify" is a judgement based on the accessibility and existence of up-to-date literature and the number of experts in the group versity (cf Tansley 1939, Connell 1979, Huston 1979), as seen m the plants (Gibson and Brown in press a) "Indicator" taxa - do they exist? The choice of invertebrate taxa in studies of management, succession or other ecological investigations is rarely arbitrary groups are chosen on some sort of compromise between theoretical and convenience reasons (e g Brown and Southwood 1983) It is generally agreed that a chosen group must be large enough for responses to treatment to depend on at least several species Despite this, groups such as British butterflies are often used because of their ease of identification (Nature Conservancy Council 1986) Here, data from quantitative (leaf-miner quadrats) and semi-quantitative (vacuum sample) studies allow the companson of five groups of invertebrates Practical reasons for choice would suggest taxa with an efficient sampling method, whose typical sample size is large compared to the animals' usual density, ease of identification, both with resjject of availability of literature and experts and to the amount of work (e.g dissection) that has to be done for reliable identification Theoretical reasons for choice would suggest taxa which show clear responses to management compared with temporal fluctuations, contain common species characteristic of all treatments applied and, where studies cover a successional senes, contain species charactenstic of all stages. Table 6 shows that, not surprisingly, no single taxon meets all these cntena. The single group which meets most of them is the spiders, their only drawback is that vacuum sampling only addresses a part of the spider fauna. The remaining fauna, which is too sparse or has individuals too heavy or concealed for D-vacs, contains many species with strong responses to habitat change (Duffey 1962) Sampling methods for this remaining fauna are time-consuming or non-quantitative Herbivorous Coleoptera emerge f)oorly from this study, their single most important drawback is their low typical density (see also Brown and Hyman 1986, in press. Morns and Rispin 1987) Plant-based sampling methods may show them to be a more generally useful group Heteroptera in this study were also generally too sparse to be useful Two groups stand out both for practical and theoretical reasons Leaf-hoppers have already been widely used in community studies (e g. Waloff 1980, Moms 1981a, b, 1990b, c,d, Prestidge and McNeill 1983). their responses m this study have been shown to be much more sophisticated than the simple response to grass architecture than has sometimes been suggested before (e g Andrzejewska 1965) Leaf-miners have been neglected because of feared difficulties in identification, with notable exceptions (Godfray 1982, 1985, Brown et al 1990) This study suggests that they should be in the first rank of consideration for community studies or for the choice of "indicator groups". ImpUcations for management of grassland Each group studied has produced quantitatively and quahtatively different information about the way in which grazing affects the invertebrate fauna Only among leaf-miners, and to a lesser extent Auchenorrhyncha, could examples of each different type of grazing response be found. Each grazing treatment also produced a distinct fauna, with its own balance between different groups There are two direct consequences for management and monitoring of grasslands. First, there is a severe hmit to the amount and type of infonnation that a single ECOGRAPHY 15 2 (1992) invertebrate group can give about the whole fauna Secondly, management goals must be extremely clear about their desired effects on particular taxa Rotational management (Morns 1990a) is not necessarily "good" or "bad", nor is grazing "too heavy", "too light", "too early in the season" or "too late" the value of management is only relevant with reference to clearly defined goals A particular example concerns the use of intensive management to produce a close-cropped turf, often recommended on traditional grounds, to maximise plant diversity (Smith 1980) or needed for the conservation of particular rarities m Britain (e g Thomas 1983, Thomas et al 1986, Cherrill and Brown 1990a, b) The young grasslands on Upper Seeds suggest that the predominant fauna associated with this management is a numerically nch but species-poor one dominated by common animals of open ground, early successional or arable land The httle evidence from study of closely grazed chalk grasslands tends to support this With the caution that more studies of the fauna of closelycropped downlands are needed, this suggests that management successfully designed to promote a number of sparse rarities may inevitably lmpovensh the more abundant compionents of the fauna Acknowledgements - This study was made possible by generous support from the Nuffield Foundation 1985-1987 and continued with support from the Nature Conservancy Council 1988-90 Experiments on the older grasslands were set up with the aid of a grant from the Nature Conservancy Council's Southem Region We also thank BIOSCAN (UK) Ltd for support from 1988 Grazing management has been enabled by the staff of Oxford Umv Expenmental Farm at Wytham, and Oxford Umv Wytham Management Committee gave permission for the land to be used Data collection would not have been possible without the cheerful assistance of a growing band of field recorders We thank M G Morns for helpful comments on a previous version of the manuscript

Journal

EcographyWiley

Published: Apr 1, 1992

There are no references for this article.