NEW RECORDS OF THE TRIBE BRYOBIINI BERLSESE ( ACARI : TETRANYCHIDAE : BRYOBIINAE ) FROM SERBIA , WITH NOTES ABOUT ASSOCIATED PREDATORS ( ACARI : PHYTOSEIIDAE )

This paper gives an overview of the present knowledge and some new faunistic and zoogeographic data of the insufficiently researched tribe Bryobiini in Serbia. In Serbia, this group of mites is represented by eight species, including four species new to Serbian fauna: Bryobia angustisetis Jakobashvili, B. lagodechiana Reck, B. ulmophila Reck and B. vasiljevi Reck. New data on host plant species and families have also been obtained – two new host plant species for B. angustisetis, two host plant species and two host plant families for B. graminum, one host plant species for B. lagodechiana, four host plant species and one host plant family for B. rubrioculus, two host plant species and one host plant family for B. ulmophila, and one host plant species for B. vasiljevi. This overview was supplemented with data on the other groups of leaf-inhabiting mites that coexist with Bryobiini species. Statistical analysis of interspecific association was done to determine the role of 15 associated predatory mite species belonging to the families Phytoseiidae, Anystidae and Trombidiidae.

Mites of the genus Bryobia are rather conspicuous, darkly colored and larger than most of the other leaf-inhabiting mites.These organisms are mainly dis-tributed on the upper leaf cuticle and are easily visible with the naked eye.Nevertheless, based on zoogeographical data, in most of the European countries there is a small number of recorded species.More than half of registered European species (36) was found only in a single country.Besides the fact that in several European countries mites of the tribe Bryobiini have not yet been discovered, in 20 European countries only 1-5 species have been recorded, in five countries 6-10 species; a greater number of species have been found only in France (12) , Ukraine (19), Italy (20) and Greece (29), (Migeon andDorkeld, 2006-2011;Bolland, 2011).For the completeness of the allegations, the significant data of 13 recorded species from Moldova should also be taken into consideration (Kulikova, 2008(Kulikova, , 2011)).Based on the above, it can be assumed that, with the exception of a few widely distributed species, the majority of Bryobia species are not frequently found in Europe, and that they probably exist in lower population densities.
Over the past three decades in Serbia a relatively extensive and simultaneous sampling of the most important leaf-inhabiting mites has been conducted, primarily from the groups Tetranychoidea, Eriophyoidea and Phytoseiidae (Stojnić and Petanović, 1994;Petanović and Stojnić, 1995;Petanović and Stanković, 1999;Glavendekić et al., 2002;Stojnić et al., 2002;Petanović et al., 2003;Stojnić et al., 2007;Mladenović et al., 2010).Mites were collected mostly from cultivated plants and weeds in agrobiocoenoses, but also from forest and ornamental woody plants, as well as ruderal herbaceous plants.Despite the effort expended in the period reviewed, only five species of subfamily Bryobiinae have been found (Tomašević 1965;Stojnić, 1993;Glavendekić et al., 2002;Stojnić et al., 2007): Bryobia rubrioculus (Scheuten), B. graminum (Schrank), B. kissophila van Eyndhoven and Pseudobryobia longisetis (Reck), (Bryobiini), as well as Tetranycopsis horridus (Canestrini and Fanzago), (Hystrichonychini).In comparison with other spider mites, a small number of Bryobia individuals was collected in samples, and from rather few sites.This research aimed to investigate the presence of Bryobia mites in natural forest habitats in Serbia, and to summarize all Bryobiini data collected so far.Statistical analysis of interspecific association was performed in order to assess the possibility of possible predator-prey interactions between predatory mites and Bryobia sp.

MATERIALS AND METHODS
Leaf samples were collected from deciduous forests in different parts of Serbia.In the laboratory, the mites were collected from standard and nonstandard leaf samples, consisting of 100 and 300-500 leaves respectively.Prior to mite extracting, both types of sample were exposed to ethyl acetate (a mite-killing agent) for 15 min.From the standard leaf sample, mites were collected by direct examination under a stereomicroscope.Mite extraction from the nonstandard leaf samples was performed by a simple beating technique over white paper.Prior to microscopy, the mites were cleaned in ethyl alcohol and lactic acid solution for several days (Evans and Browing, 1955) and then mounted in Hoyer's mounting medium (Boczek, 1980).Identification was performed using suitable taxonomic keys for the families Tetranychidae (Reck, 1959;Mitrofanov et al., 1987;Hatzinikolis and Emmanouel, 1991) and Phytoseiidae (Begljarov, 1981;Karg 1993).All examined material is kept in the Acari collection, Department of Entomology and Agricultural Zoology, Faculty of Agriculture, University of Belgrade, Serbia.
Statistical analysis of interspecific association (two species case) was used (Dajoz, 1977;Ludwig and Reynolds, 1988) in order to assess the possibility of possible predator-prey interactions between predatory mites and Bryobia sp.The Jaccard Index (JI) presents the proportion of sampling units where both species occur to the total number of sampling units where at least one of the species is found: where a is the number of sampling units where species A occurs; b is the number of sampling units where species B occurs; c is the number of sampling units where species both species occur; d -represents the number of sampling units where neither A nor B are found; N is the total number of sampling units: a+b+c+d).
The probability or expected frequency of occurrence of both species in sample unit is given as: P = ab/N P>c, negatively associated pair of species; P=c, species are not associated; P<c positively associated pair of species.
The significance of obtained values was tested by the chi-square test: X 2 = N 3 (c-P) 2 / ab(N-a)(N-b); if value X 2 >3.84, the null hypothesis of independent co-occurrence of species A and B is rejected, and it can be concluded that species A and B are significantly associated (*); if value X 2 >6.64, it can be concluded that species A and B are very significantly associated (**).

RESULTS AND DISCUSSION
A total of 85 samples of mites of the tribe Bryobiini were collected at 56 sites from the following plant species: Allium cepa L., Cornus mas L., Corylus colurna L., Crataegus monogyna Jacq., Daucus carota L., Hedera helix L., Malus domestica Borkh., M. pumila Mill., M. silvestris Mill., Prunus avium (L.), P. cerasifera Ehrh., P. cerasus L., P. domestica L., P. serotina Ehrh., P. spinosa L., Pirus communis L., Pseudotsuga menziesii (Mirb.),Salvia pratensis L., S. nemorosa L., and from mixed samples of Poaceae and Lamiaceae.This study revealed the presence of eight species of the tribe Bryobiini in Serbia, of which four species are new to our fauna.The finding places of species are shown in Fig. 1.
Collection data (4 samples: 31 females, 18 nymphs, 16 larvae ): Zemun, Belgrade, 12.11.1990, in fallen cones of Douglas-fir, Pseudotsuga menziesii (Mirb.)Franco -5 females, 2 nymphs, 1 larva (leg.B. Stojnić); Stublina, Aleksinac, 31.08.1991, on wild carrot, Dau- Remarks: Taking into account the worldwide distribution of B. graminum, the small number of samples does not provide sufficient data of its distribution in Serbia.The presence of this mite in the cones of Douglas fir certainly does not refer to a host plant, but to a hibernaculum.On the other hand, A. cepa (Alliaceae) and D. carota (Apiaceae) are new records, both for host plant species and families.
Remarks: In Serbia, among the Bryobiini species, B. rubrioculus has been most frequent, found at most of the sites and on the greatest number of host species (10).Three new hosts of B. rubrioculus have been recorded from Rosaceae (Malus pumila, M. silvestris and Prunus serotina).The presence of this mite spe-cies on Corylus colurna is the first information about the one more host family, Betulaceae.
Remarks: This is the first report of B. ulmophila from Serbia.New records of host plant species (M.sylvestris, C. mas) as well as a new host family (Cornaceae), indicate that B. ulmophila prefers woody hosts.
Hosts: mints, Lamiaceae (Mentha sp., Phlomis sp., Salvia nemorosa, S. officinalis, S. sibthorpii, Salvia sp.), also Asteraceae (Cynara scolymus), Campanulaceae (Asyneuma argutum) and Convolvulaceae (Convolvulus sp.) (Migeon andDorkeld, 2006-2011).Remarks: In both samples, P. longisetis has been found in sages on sandy soil, in moderate population density, associated with A. salviae and T. turkestani.Different stages of P. longisetis were present on both sides of sage leaves, as well as yellow spots.The silk webs of the other tetranychid species, T. turkestani, have been poorly spun.This was the first report of P. longisetis from Serbia and second report from Europe.Meadow sage, S. pratensis, was new host data (Stojnić et al., 2007).

Notes on the associated predatorz mites
Mites of the tribe Bryobiini have been found associated with other mite groups in the majority of the examined samples.predatory mites were present in 43.5% of samples with Bryobia sp.(Table1).Twelve species of the family Phytoseiidae were found: Anthoseius rhenanus (Ouds.),Dubininellus corniger (Wainstein), D. echinus (Wainstein and Arutunjan), D. juvenis (Wainstein and Arutunjan), D. macropilis (Banks), Euseius finlandicus (Ouds.),Kampimodromus aberrans (Ouds.),Paraseiulus talbii (Athias-Henriot), P. triporus (Chant and Shaul), Seiulus aceri (Collyer), Typhlodromus pyri Scheuten and T. tiliae Ouds., as well as unidentified juvenile stages of phytoseiid mites in a number samples.In addition, few samples contained mites of families Anystidae (Anystis sp.), Stigmaeidae and Trombidiidae (Allothrombium sp.) At the outset, it should be clearly recognized that interspecific association between pairs of species, as well as the Jaccard similarity coefficient drawn from these analyses, are limited by the relatively small number of samples, and also by the relatively small number of collected Bryobia individuals, commonly accompanied with the other more numerous phytophagous and mycophagous mites.However, it is believed that these data could provide a useful base for future sampling.
The main result of this analysis is a conclusion that the pair of two most frequent species, E. finlandicus and B. rubrioculus, is not significantly associated in a broad spectrum of host plant species and habitats.Contrary to these findings, the results of some earlier studies obtained mainly in neglected or-chards, showed that the same pair of species was very significantly associated (JI%=18.3**)(Stojnić, 1993).Similar to the above, a very significant association (JI%=13.9**) between K. aberrans and B. rubrioculus was inevitably determined by the origin of samples -also from neglected fruit trees.The same goes for all the other significant and very significant values (Table 1), obtained on the basis of too few simultaneous findings of pairs of species.Further studies of interspecific association between these pairs of species are necessary to obtain representative results and to draw valid conclusions.

Table . 1
. Proportion of predatory species in samples containing Bryobia sp.
In Serbia, B. kissophila is noticeably present on the leaves of ivy only early in the season, but later population densities are lower.
This is the first report of B. vasiljevi from Serbia.Crab apple, M. sylvestris, is new host record.