장음표시 사용
1쪽
AN INTERNATIONAL JOURNAL OFANIMAL TAXONOMY AND ECOLOGY
Acta Mologica Academiae Scientiarum Hungaricae is published quarterly DomFebruary l 994 sother issues in May, August and November) by the Hungarian Natural History Museum and the Biological Section of the Hungari an Academyos Sciences with the financiat support of the Hungarian Academy of Sciences.
2쪽
Actia Mologica Academiae Scientiarum Hungaricae 46 I). ρρ. J-II, 2000
SPATIAL DISTRIBUTION OF CARABID SALONG GRASS FOREST TRANSECTS
sive groups by indicator species analysis: si) habitat generali sis 2) forest generali sis 3)Species of the grass, sq) forest edge species, and ) forest speciali sis. Distributions of the
eighteen most frequent carabids were generalty aggregated. There were significant correlations belween the carabid abundance and the illo ing abiotic factorso relative cover of thel eas litter, the herbs, the shrubs. and the canopy layer. Biotic factors, like the abundance of the carabids' prey, and the Occurrence of other carabids were also correlated significantly with the distribution os particular species at the studi ed spatiat scale. For the eighteen most frequent species we mund P significant positive and 4 significant negative correlations of the abundance patierns. For twO species sMoloPS Piceus sPANZER. l 93) and Pterostichus hiar-meisteri HEER, i 84 l), whicli are of similar siZe. spatiat patieria and seasonat acti Vity, wefound significant negative interactiora suggesting interspeci sic competition belween them. The resulis stress the importance of an integration os biotic and abiotic factors in carahidecology, and at So provide an empirical approach sor understanding spatiat distribution os ca- rabidS. Key word S forest edge, aggregation indices, indicator Species, communi ty organi sation
There is a current debate whether the spatiat distribution os a species in acommuni ty is determined by their autecological characteristi cs or by communi ty
Quantitative knowledge about habitat associations and spatiat patieria os ca- rabids in Hungary is relatively poor. Further studies, haSed On comparis Onsamong habitat type S are necessary to demonstrate more precisely the habitat requirements of carabid species. This knowledge is essentiat to asses s the relative importance of en viron mental factors and species interactions in structuring communities in different habitats. Knowledge of exact habitat requirements is also
3쪽
needed for the purpo ses of nature conservation and conservation biology, in order
SPENCE l 994. NIEM EI A et ah l 990). This non-rando ira carabid occurrence is determined by the heterogeneotis distribution os ab iotic conditions and refources NIEMELA et ah l 992a). However. there are severat scales of en Viron mental heterogenei ty, and species responses may be determined by different factors at variotis leve is of en viro iamental patiern NIEM EI A et a l. l 992b). For example, on theregional scale the distribution os a particular carabid species may be determinedmain ly by geographic, climatic and historical factors HENGEVELD l98 P. PENEV& TURIN l 994. PENEV l 996). But on a locat scate, that is, in an area whicli is welt with in the dispersat radius of the species, variation in distribution acrosshabitat types may be determined by en viron mental conditions and interspecificinieractions NIEM EI A et ah l 985. l 996. LUFF et ah l 992. HALME & NIEM ELA
In this paper we present an analysis of ground - beet te distribution on a localscule, in grass- forest transecis at the ecological scale of interacting populations NIEM EI A L SPENCE l 994). Our objectives were to assess the extent of Variation in the distribution os carabid species, and to relate this to habitat characteristicsand to spatiat distribution os co- occurring carabid species.
Samplisag area was located at the Northeria Motintains in the Aggtelek Nationat Park. nearthe Mogyoros Peak Haragi stya). In this region oak-horiabeam forests fouerco- Pinetum) and agrasS association Polygialo vici jori-BrachyPodietum Pinniali) are the most extensive. There are three habitat types On the research area si) Grass Polygalo majori-Brachyρodietum Pinnati), with dense herbaceous vegetation dominated by Polygiala major, Brachyosilietum Pinnatum. Fili eridum vulgaris, Salvici Priateri yiS. Dum hir Ici. Gerianium Sanguineum. The litter layer. the shrubs and the canopy layer are missing in iliis habitat. Its area was approx. 40 lin. 2) Forest edge. with dense herbaceous vegetation originating srom the adjacent grasS. Theshrub layer is also dense in this habitat . consisting main ly of shmbs and saplings of the canopytreeS PinuS heiulus, Corylus avellana and Prunus minosia). The litter layer is thick and the canopy layer is tess closed than in the forest interior. 3) Forest interior: Oak-horiabeam forest. with thick litter layer, moderate herbaceous and shrub layer and with 85-95 R canopy cover. The si Ze of the forest stand was greater than l00 ha. Beelles were collected using unbai ted pittali traps, consisting os plastic cups diameter l00 min. volume 500 mi) containing ethylene-glycol as a kill ing-preserving solution SPENCE & NIEME LA l 994). Three replicated parallel transects of pittali traps were set across the three studi ed habitats. about 50-J0 meters hom each Other. The transects were perpendicular to the forest edge.
4쪽
SPATIAL DISTRIBUTION OF CARABIDS
To study associations botween the distribution Os carabids and the environ mentat variabies, we studi ed sive environ mental sactors. We estimated the percentage cover of the leas litter layer, theherbs. the shrubs and the canopy layer within a radius of 0. 5 meter around each trap. we also stud ted the abundance of the potentiat 1 ood refources of the carabilis sabundance of other animais that seli in the traps). These invertebrates se. g. Lumbricidae. Mollusca. I Sopoda, Chilopoda, Aranea. and Coleoptera larvae) are sursace dwellings, theres ore they cara be regarded as 1 ood sotirce sor the carabitis. It was proved by serological method SERGEEVA l 994) that these invertebrates are prey sorcarabitis.
Principat coordinates analysis PCOA) using the Bray-Curtis index of similari ty was used forcara id abundaraces to asses similarities in carabid assemblages of the traps GAUCH l 986). He
The IndVal Indicator Value) approach was applied to si nil indicator species and species assem blages characteri sing the grass, the forest edge and the forest interior DUERENE & LEGENDREl99J). Iis goat is to finit indicator species and or species assemblages characteri Sing grou pS Of Sam-ples. The novel ty of this approach lies in the way that this method combines a species' relative abundance to iis relative frequency of occurre iace in the variotis groups of samples. Statistical signisi cance of the species indicator values is evaluated using a rando mi sation procedure. The Stari os this approach consi sis os obtaining a classis cation Os sample uniis using one of the classical methods of data analysis. We obtained a typology si ona the principat coordinates analysis using Bray- Curtis index os percentage similari ty sor carabid abundances. Based on this si te typology Ind Validentis es the indicator species corresponcling to the variotis groups. Indicator species are defined asthe most characteristic species of each grou p. soland mostly in a single group of the typology and present in the majori ty of the sites belonging to that groti p. This duali ty. whicli is of ecological in terest. is rarely exploi ted completely in such analyses: osten only the distribution os abundances in the Variotis groups is VSed. In these cases, species occupying only one or two sites in orae habitat group and present only in that group rare species) receive the fame indicator value as species occupying ali sites of that habitat group and mund only in that groti p. There is an important dissere iacebetween these two species. The sirst One is an asymmetrical indicator. according to the IndVal terminology, whicli means that iis presence cannot be predicted in ali sites of one habitat. but contributes to the habitat specii city. The seconii type of species, Ora the contrary, i S a true symmetrical indicator: iis presence contributes to the habitat specifici ty and one casa predici iis presen ce in alisites os the groti p. With the IndVal it is possibie to distinguisti the two types of indicator species: the species that have an indicator value greater than ββR are regarded as symmetrical indicator species sDUERENE & LEGENDRE l99 ). To characteri se the spatiat distribution of carabills. we calculated the index os dispersion. Iδ whicli is desinod as the variance-to- mean ratio sDIGGLE l 983) lδ α S lx. where 1' the Vari ance, an lx the average number of individuat s. An la value close to one indicates a rando in distribution. H l indicates aggregate i. while I3 Q l indicates regular distribution. The departure si om rando ira nesscara be tested by the test statistic ID n- l)Xylx. It has approximate4 χ distribution with sn- l) de-grees of 1 reedom. This approximation is reasonable provided n 6 and x l. Multiple regression was used to study whether any of the enViron mental measu rements andos abundance of the other carabids could he used to predici distribution os a particular carabidspecies. Catches of the common carabid species were compared among murteen traps in habitatswith Kruskal-wallis nonparametric A NOVA . A Tukey-type multiple comparison was then used tocompare catches frona the habitat types SOKAL Ω ROHLF l98 l). For the eighteen most frequent species the correlation het ween the number of trapped individua is and the degree of their aggregation was calculated by the Pearson 's product- moment correlation. Beire the calculations the num-
5쪽
her of individuals was transformed by a togarithmic transformation to provide normali ty. The correlation belween the abundance of a species and number of traps Dorn whicli it was recordeclwas calculated by the Spearinan rank correlation. The analyses were done by the SPSS-PC pro-
Tahle 1. Two-way indicator table showing the species indicator power for the habitats. In the columns for each species the Lirst number indicates the number of specimens, the second the numberos traps wherein the species was captured, in this Sample group. The IndVal R) columia indicates the species indicator value for the corresponding clustering leve l. '' indicates a significant Pς0. Ol), while nS means a not significant IndVal value.
6쪽
SPATIAL DISTRIBUTION OF CARABIDS
7쪽
than i individuals per trap. can he referred to as dominant species. Ele Ven other species, which were mund in 20-40 traps and were represented by a mean calchos more than O .lf individuals per trap. can be designated as Subdominant species. The other 22 species that were captured in less than 20 traps with less than O .l findividuals per trap, ure rure SpecieS. The result of the ordination PCOA) shows that there was an arcti effect suggesting a gradient in the data GAUCH l 986), namely, the carabid assem h-lages change gradually Dom the grass towards the forest interior a long the tran- secis Fig. l). Carabid samples of traps Dona the grass, the forest edge, and the forest interior were separated frona each other. The composition of the carabidsamples of the forest edge was more similar to the forest interior than to the grass. It is also evident, that there is a gradient in the species composition, be-
cause the traps are arran ged a long an arch.
Fig. 1. Ordination Os pittali catches by principat coordinates analysis using the Bray-Curtis index of dissimilari ty. grass. - forest edge. A forest interior
8쪽
SPATIAL DISTRIBUTION OF CARABIDS
Tahle 2. The index os dispersion to characteri se the spatiat distribution os frequent carabid species in the habitats. A value close to one indicates a random distribution. A significantly greater valuethan orae indicates aggregated spatiat distribution, while a value smaller than one indicates a regulardispersion. The sigia shows that the statistical test is not applicabie. and indicate significant spς0. 0 Pς0. 0l and PQ0. 00 l. respecti Vely). whil e ns not signi cant departure froin the ran-
9쪽
The dominant and subdominant species generalty Showed aggregated distribution Table 2). There was significant positive correlation belween the num-her of individuals and the degree of aggregation for a Species sm0. 63J5,ρα0.0033, n l9) indicating that more abundant Species were more aggregated. Sometimes, Variation in carabid catches among the murteen traps in thesame habitat was greater than the variation among habitats Table 3) indicatingilie importance of strong microhabitat variations controlling carabid distribution. Differences among the habitais in the number caught individuals, on the other hand. were significant for ali the dominant and subdominant species, excepi -rahuS convexuS FABRICIUS, i J5, Carabus nemorialis O. F. MULLER. lJ64 and Carabu1 intricatus I INNAEUS. t 6 l. This result suggesis the clear habitat preseerence of carabid species. Multiple regression analyses between the distribution os carabids and theen Viron mentat Vari abies and occurrence of other carabids were significant
10쪽
SPATIAL DISTRIBUTION OF CARABIDS
Table 3. The value of Kruskal-Wallis non-parametric A NOVA and iis statistical significancestis: not significant. PQ0. 05. PQ0. 0l and PQ0. 00 l) for the variation among the habitat types and for the variation among transects with in habitat types. The result of Tukey-type multiplecomparison among habitat types is gi ven by indicating significant or not significant ) differences: the significance levet is PQ0. 05. Legends: F - forest interior, E - forest edge and G - grass. Species
Variation Result os a Variation Variation Variation