Chemotaxonomical characterisation of males of Bombus lucorum ( Hymenoptera : Apidae ) collected in the Czech Republic

Labial gland secretions of 26 males of the bumblebee Bombus lucorum (L.), collected in the Czech Republic, were ana­ lysed. The secretions consisted of 60 compounds; ethyl (Z)-9-tetradecenoate was the main component (average 53%). Although the males varied in colour, their labial gland secretions were similar in composition, which indicated they belonged to one species. Chemically the B. lucorum occurring in the Czech Republic correspond to the earlier described “blonde form” of this species.


INTRODUCTION
The species of bumblebees (Bombus Latreille, 1802) belonging to the subgenus Bombus s.str.are often difficult to determine.Colouration is very variable and other morphological differences are indistinct.Individuals of the "lucorum species group" are especially difficult to distinguish morphologically (Krüger, 1954).In West and Central Europe this group consists of Bombus (s.str.)lucorum (Linnaeus, 1761) and two other closely related species -B.(s.str.)magnus Vogt, 1911 and B. (s.str.)cryptarum (Fabricius, 1775).
Bombus lucorum in the broadest sense (the taxonomy and nomenclature is very confused -see Williams, 1991, p. 81-85) is a widely distributed bumblebee species, which occurs throughout the Palaearctic.It is one of the commonest bum blebee species in Central and Northern Europe (Rasmont, 1984).The males are very variable in colour throughout their distribu tion.Due to their morphological similarity B. cryptarum (Ras mont, 1984) and B. magnus (Loken, 1973) have either been regarded as distinct species or have been synonymised with B. lucorum, sometimes even with B. terrestris (Warncke, 1981(Warncke, , 1986)).This has led several authors to use characteristics other than morphological and morphometrical ones to separate these taxa.Species can be distinguished using enzyme electrophore sis (Scholl et al., 1983(Scholl et al., , 1992) ) or the chemical composition of the marking pheromone of the males (Pamilo et al., 1997;Bertsch, 1997a, b).The second characteristic is known to be speciesspecific.In Scandinavia, the labial gland secretions of males were used to identify two forms of Bombus lapponicus (Berg strom & Svensson, 1973), which were later classified as distinct species, Bombus lapponicus (blonde form) and Bombus scandinavicus (dark form) (Svensson & Bergstrom, 1977).Svensson (1979Svensson ( , 1980) ) later synonymised Bombus scandinavicus with Bombus monticola.
The first analysis of the chemical composition of the labial gland secretion of B. lucorum males was reported by Calam (1969).He identified the main component (ethyl 9-tetradecenoate) and several of the minor components.B. lucorum occurring in Sweden was studied extensively by Kullenberg et al (1970) and Bergstrom et al. (1973).The labial gland secretions of individuals of a wide variety of colour, ranging from light yellow to dark, from the island Öland were analysed.While that of the blonde form contained mainly ethyl (Z)-9-tetradecenoate, the dark form produced mainly ethyl dode-canoate (Kullenberg et al., 1970).The "dark" form was later transferred to B. cryptarum (Rasmont et al., 1986;Bertsch, 1997a, b).
Bumblebees belonging to the "B.lucorum species group" occur in Czech Republic (Tkalcu, 1974(Tkalcu, , 1999)), but the chemical composition of the males' labial gland secretion is unknown.The purpose of this study was to make a detailed analysis of the labial secretions of specimens collected in different localities in the Czech Republic, to determine the species on the basis of both morphological and chemical traits and to compare our results with the data in the literature.

MATERIAL AND METHODS
Insects.Males (26 individuals) of the bumblebee species Bombus lucorum were collected in the summers of 1994-1999 at seven different localities in the Czech Republic (both in Bohemia and Moravia, Table 1).Some of the males came from colonies that were established artificially using mated females from the previousyear.The insect material is deposited in the collection of one of the authors (O.H.).
For the chemical analyses, living insects were transported to the laboratory and then kept in a freezer until the labial glands were dissected.The glands were extracted with hexane (50 pl per gland).After 15 minutes of shaking and 2 h standing in the refrigerator, the hexane extract was filtered and stored in a freezer before analysis.Each sample was analysed separately.
Gas chromatography.The extracts were analysed using a gas chromatograph with a splitless injector (200°C) and a mass detector (Fisons MD 800).A BPX5 column (5% phenyl methyl silicone; 30 m x 0.22 mm, film thickness 0.25 pm) and helium (flow 0.55 ml/min at 50°C) were used for the separations.The temperature program started at 70°C (2 minutes delay) after which the temperature was increased to 140°C at a rate of 40°C/min, then to 240°C at a rate of 2°C/min, and finally to 300°C at a rate of 5°C/min.Compounds were identified by comparing their mass spectra with those in the National Institute of Standards and Technology Library (NIST, U.S.A.) and by co-chromatography with synthetic or commercially available standards.
The double bond positions were determined from mass spectra of dimethyl disulphide (DMDS) adducts of unsaturated components.The configurations of the double bonds in the fol lowing compounds were determined (after their chroma tographic separation) from the retention times and co chromatography with the corresponding E-and Z-standards (iso thermally 160°C for ethyl 9-hexadecenoate, 200°C for ethyl 9-octadecenoate, 170°C for 9-hexadecenol and 11-octadecenol on a DB-1 column ( methyl silicone -30 m x 0.25 mm, film thickness 0.25 pm); hydrogen as carrier gas, flow 1.2 ml/min; Hewlett-Packard 5890A).The double bond configurations in unsaturated hydrocarbons were determined from infrared spectra measured on a Bruker IFS-88 instrument in a KBr micropellet (1.5 mm diameter).The elution order of the isomers of unsaturated ethyl esters was determined from the analogy with chromatographic properties (equivalent chain length values) of fatty acids methyl esters according to the data in the literature (Christie, 1988;Stránský et al., 1997).The identity of (Z,Z)-9,12-octadecadienol and (Z,Z,Z)-9,12,15-octadecatrienol was verified by co-chromatography with commercial standards, on two different columns, and by their mass spectra.

Preparative column chromatography and derivatisation.
A hexane extract of one selected gland (50 pl) was chromatog raphed on silica gel Merck 60 (0.040-0.063 mm; 290 mg) in a Pasteur pipette.The elution of the sample started with pentane, followed by hexane/ether mixtures (1%-40% of ether).Frac tions were checked by GC-MS and those containing unsaturated compounds were derivatised with DMDS according to the pro cedure published by Attygalle et al. (1993).The products were analysed by GC-MS using the same temperature program as for the original extracts.
Evaluation of GC data.The analytical results were evalu ated by multivariate data analyses PCA (principal components analysis) and PLS-DA (projections to latent structuresdiscriminant analysis) (Wold et al., 1989), which are included in CODEX® (SumIT System AB, Sweden).The GC-data were pre treated as in earlier investigations (Sjodin et al., 1989;Valterová et al., 1995).The normalised data were logarithmically trans formed and represented as a matrix X.Each variable was scaled to unit variance (autoscaling).
In PLS-DA, the Y matrix was made up as a single dummy variable, each male being given a value -1or1 depending on the group of males to which it belonged.The significance, i.e. the number of significant components, was determined by cross validation.The importance of the constituents in differentiating between the two groups ofbumblebees was determined from the loading plots (Wold et al., 1989).

RESULTS
All the males belonged to the "blonde" form, in particular the medium blonde form as figured in Bergstrom et al. (1973, Plate I, Figs 2-4).The material was slightly variable.The variability in the colouration of the Czech males corresponds to that cited for B. lucorum males by Rasmont et al. (1986).
The composition of the labial gland secretion of B. lucorum males is summarised in Table 2.The positions of the double bonds are specified in Table 2 except for minor components (content < 0.1%) and polyunsaturated compounds where the DMDS adducts were not found.The configurations of the double bonds, as determined either from the co-chromatography with standards or from infrared spectra (absence of the band 965 cm-1, which is intensive and typical for E-isomers), were Z in all cases.
Traces of the sesquiterpenic alcohols farnesol and 2,3-dihydrofarnesol (below 0.1%) were detected in Czech bum blebees.The Scandinavian blonde form did not contain any isoprenoids (Bergstrom et al., 1973).We also detected several compounds with larger molecules than pentacosane.The majority were present in trace amounts, but substantial quanti ties of 9-heptacosene and wax-type esters were present, espe cially in some of the samples.It is possible that these com pounds were not analysable by the older techniques, which may account for why they were not recorded previously.
The principal components analysis (PCA) showed a slight tendency for the data to separate into two groups, one con taining the males collected in 1997 and one containing all the other males.However, cross-validation revealed no significant components (CSV/SD(1) = 0.97).A PCA, in which hydrocar- "Males from a colony artificially established using a mated queen from the previous year bMales from one nest bons and fatty acids were excluded, resulted in a model with one component on the limit of significance (CSV/SD = 0.94, vari ance explained (1) = 22%) that showed the same pattern of groups.Hydrocarbons present in the gland extracts are not con sidered to be the active components (Bergman, 1997).Free fatty acids are most probably precursors of the components in the secretion (Bergman, personal communication).Therefore, hydrocarbons and free fatty acids were excluded from the fol lowing analysis.

DISCUSSION
Although the males collected in 1997 differed from the rest of the males, the order of difference is very small compared to that between the Scandinavian dark and blonde forms described by Bergstrom et al. (1973).Therefore, B. lucorum in the Czech Republic are all chemically similar and more similar to the Scandinavian blonde form than to the dark form (Bergstrom et al., 1973).All the major components we found were also found by Bergstrom in the blonde form.Compounds reported only from the dark form (ethyl decanoate, ethyl octadecadienoate, geranylgeraniol, and geranylgeranyl acetate) were not found in the Czech males.The literature on "lucorum species group" and our results point to the conclusion that the Czech specimens belonged to B. (s.str.)lucorum and not to B. cryptarum or B. magnus (Bertsch, 1997a).Differences between individuals were most probably due to differences in age or physiological condi tions rather than their origin.
The composition of the labial gland secretion of B. cryptarum (Bertsch, 1997a) corresponds to that of the "dark form" of B.  Bergstrom et al. (1973) in the blonde form (semiquantitative data), n.d.= not determined because of the low content of the component lucorum (Bergstrom et al., 1973).The components of the marking pheromone of B. magnus have not been described in detail.Bertsch (1997a) mentions the two main components (9,12-octadecadienol and 9,12,15-octadecatrienol) of the secre tion collected from artificially reared B. magnus, that originated from Scotland.Czech B. lucorum males produced these two compounds, too, but both in small amounts (0.8% and 1.3%, respectively).
B. lucorum is the most common species of the "lucorum spe cies group" in the localities where we collected our specimens.This agrees with the data in the literature.B. cryptarum and B. magnus are rare in the Czech Republic as we have not found any individuals that belong to these species.Although the males we collected were slightly variable in colour, they all belonged to one "chemical form", similar to the Scandinavian blonde form of this species or more simply B. lucorum.It is apparent from our results and the literature that the composition of the labial gland secretion can be used to determine species in cases where morphological traits fail or are difficult to use.

Table 1 .
Data on where and when the Bombus lucorum males were collected.

Table 2 .
Composition ofthe labial gland secretion of B. lucorum males.-configuration in all cases as determined either by co-chromatography with standards or from infrared spectra b reported by