Discrimination of the bumblebee species Bombus lucorum , B . cryptarum and B . magnus by morphological characters and male labial gland secretions ( Hymenoptera : Apidae )

Frühjahrsköniginnen von B. lucorum, B. cryptarum und B. magnus von jeweils 2 Fundorten in Brandenburg (Deutschland) und Schottland (Vereinigtes Königreich) wurden mittels morphologischer Merkmale bestimmt. Dabei erwies sich die laterale Begrenzung des Collare am Rand des Pronotallobus oder auf dem Episternum als besonders brauchbares Merkmal. Farbfrische Königinnen der drei untersuchten Arten lassen sich sicher bestimmen, es sind gute Morphospezies. An Hand von sicher bestimmtem Material () werden von B. cryptarum und B. magnus Verbreitungskarten für Berlin und Brandenburg erstellt, aus den Fangdaten wird eine Frühjahrsphänologie der Flugaktivität für Königinnen rekonstruiert. Königinnen von B. cryptarum kommen im frühen Frühjahr aus dem Winterschlaf, sie sind 2-3 Wochen vor den Königinnen von B. magnus aktiv. Von sicher bestimmten Königinnen wurden Kolonien gezüchtet und die Labialdrüsen von Männchen aus diesen Zuchten gaschromatographisch/ massenspektrometrisch untersucht. Etwa 50 Substanzen, eine Mischung geradkettiger Fettsäurederivate (Alkohole, Ester und Kohlenwasserstoff e), wurden identifi ziert. An Hand der Labialdrüsensekrete lassen sich drei unterschiedliche Taxa sicher trennen. Die Labialdrüsensekrete der Männchen von B. magnus aus Schottland und Brandenburg sind identisch, die Arterkennungssignale sind also großräumig stabil. Die Labialdrüsensekrete der Männchen von B. cryptarum aus Brandenburg und aus Schottland sind ebenfalls identisch, damit ist B. cryptarum erstmals als Bestandteil der Fauna der Britischen Inseln nachgewiesen. Die Unterschiede der als Arterkennungssignale genutzten Sekrete der Labialdrüsen bestätigen den morphologischen Befund, B. lucorum, B. cryptarum und B. magnus sind gute Arten.

Th e nearly 250 known species of bumblebees have been grouped in subgenera which are classifi ed on the basis of morphology, especially the structures in the male genitalia (ITO 1985;WILLIAMS 1985WILLIAMS , 1994)).In spite of their size and conspicuous coloration, identifi cation of species within a subgenus is often diffi cult because several species share a similar general appearance in colour and morphology.Small diff erences in morphology have often been used as diagnostic characteristics to distinguish the species.Th e subgenus Bombus s. str.(syn.Terrestribombus) is a group in which classifi cation of spe-Bombus s. str.(syn.Terrestribombus) is a group in which classifi cation of spe-Bombus cies is especially complicated, partly due to considerable intraspecifi c variation of coloration and morphological characteristics.In Europe, fi ve species in the subgenus Bombus s. str., Bombus (Bombus) (KRÜGER 1939(KRÜGER , 1951(KRÜGER , 1954(KRÜGER , 1956(KRÜGER , 1958;;GER 1939GER , 1951GER , 1954GER , 1956GER , 1958;;GER PEKKARINEN 1979;LØKEN 1973;RASMONT 1984;RASMONT et al. 1986), enzyme electrophoretic data (SCHOLL & OBRECHT 1983;SCHOLL et al. 1992;PAMILO et al. 1984) and analyses of the compounds of the male labial glands (PAMILO et al. 1997;BERTSCH 1997a RASMONT (1983RASMONT ( , 1984) ) treated both taxa as separate species, WILLIAMS (1991,1998) lumped them with B. lucorum.Recent publications (PAMILO et al.1997;WILLIAMS 2000;PEDERSEN 2002) could not really solve these problems and did not clarify the status of B. cryptarum and B. magnus.In the present study, we use morphological characteristics of the collare and the compounds of the male cephalic labial glands to elucidate the taxonomic status of B. cryptarum and B. magnus.For this purpose, the specimens examined were taken from localities where these taxa occur sympatrically with B. lucorum.

Bumblebee samples
Queens of B. lucorum, B. cryptarum and B. magnus were collected in Brandenburg, B. magnus were collected in Brandenburg, B. magnus Germany, about 5 km Northwest of Menz, Lkr.Oberhavel (13° 02' 59'' O, 53° 06' 08'' N) during spring.All three species were abundant in a pine forest with sandy soil and a closed layer of mosses (Leucobryo-Pinetum sylvestris MATUSZKIEWICZ).In the beginning of May, luxuriant vegetation of Vaccinium myrtillus is a good food source for Vaccinium myrtillus is a good food source for Vaccinium myrtillus spring queens, which show strong fl ight activity in the early morning.Specimens were also taken from Teupitz,52° 08' 08'' N) 1975, maps ITE 1980)).Specimens were taken from the North coast of Scotland in the Sand Dunes of Dunnet, Caithness (3° 20 ' 30'' W, 58° 36' 30'' N) feeding on Anthyllis vulneraria and Anthyllis vulneraria and Anthyllis vulneraria near the light house at Duncansby Head, Caithness (3°01'40''W, 58°38'40'' N) feeding on Cirsium palustre.At both sites queens determined morphologically as B. cryptarum where also taken for investigation.
Th e spring queens collected in the fi eld were used in morphological examinations.For GC/MS investigations, males obtained from colonies developed artifi cially in a greenhouse were used.After collection, bumblebees were kept alive in a cool-box.During transportation, the characteristics essential for identifi cation, such as the tufts of hair at the thorax and abdomen, were sometimes soaked and stuck together, especially during wet weather.In such cases, we put the bees in small fl ight cages with some honey-water.After feeding, they started to clean and brush their hair by themselves, which resulted in the restoration of all the essential morphological characteristics.Specimens were preserved in a deep-freezer at -30°C.Voucher specimens of the examined bees are stored at the Entomological Collection of the Zoological Museum, Humboldt-University, Berlin.

Morphology
Details of morphology were studied using a stereomicroscope (Wild M8, Planar 1.0, Okulars 10x/21), only fresh specimen with undisturbed colour patterns were used.As already described by E. KRÜGER, details in hair was best studied in diff use light (use of GER, details in hair was best studied in diff use light (use of GER diff use fi lter and indirect light combined with Novofl ex Macrolight Plus) at high magnifi cation by stroking the hair with a pinpoint artists paintbrush.Distribution of hair on diff erent parts of the thorax was thus carefully investigated.To avoid minor equipment vibrations photographs were taken on a table with a shock absorbing granite plate, using extension bellows and a macro-lens (Olympus Zuiko 1:1 Macro 1:4/80 mm).Th e thorax was mounted on stubs with conductive carbon cement, coated with gold and viewed with a Hitachi S-530 scanning electron microscope.

Gland preparation and GC/MS
Th e cephalic part of the labial glands were dissected from the heads of males (reared in artifi cial colonies) under freezing conditions and placed in vials (glands from 5 males per vial) containing 0.2 ml pentane.A Finnigan MAT TSQ700 gas chromatograph/tandem mass spectrometer was employed.Gas chromatography was carried out on a Hewlett Packard Ultra 1 (50 m, 0.2 mm i.d., 0.11 µm fi lm thickness) in the splitless mode with helium as carrier gas at an inlet pressure of 300 kPa.Initial temperature of 120 °C was held for 1 min, then increased at 8 °/min to 280 °C, at 3 °/min to 310 °C and at 1 °/min to 320 °C.Th is temperature was held for 10 min.Mass spectrometer conditions were: interface temperature 300 °C, source temperature 130 °C, electron energy 70 eV, emission current 0.2 mA, and electron multiplier 1400 V.In the positive ion chemical ionization mode ammonia CI gas pressure was 70 Pa.Dimethyl disulfi de adducts were prepared as described by BUSER et al. (1983).Compounds were identifi ed by comparing their mass spectra with those of the NIST '02 Library (National Institute of Standards and Technology, USA) and by coinjection with commercially available standards.
Character coding and data matrix for Cladogram (Fig. 14

R e s u l t s
The diagnostic character "coloration of the collare" and "border of collare at prono tallobus/episternum" of queens Queens of B. lucorum were carefully examined by KRÜGER (1939KRÜGER ( , 1951)), according to GER (1939GER ( , 1951)), according to GER his investigations the yellow hair of the collare always end at the border of the pronotallobus, only a few yellow hairs may be found at the dorsal border of the episternum.Queens of B. cryptarum show a characteristic S-shaped band of dark hair, which follows the border of the pronotallobus and separates a yellow collare from two patches of yellow hair at the upper border of the episternum (Fig. 1).In northern Germany many specimen of B. cryptarum are strongly melanised, and the collare is not bright yellow but often dark-brown due to the mixture of yellow and black hair.Only the two yellow patches at the upper border of the episternum remain unchanged by melanisation, this distinct characteristics makes a safe determination of queens especially easy (Fig. 3 & 4).).However discrimination is not easy and it is helpful to sharpen the eye for the characteristics by fi rst inspecting specimens of B. cryptarum with the characteristic S-shaped band at the border of the pronotallobus (Fig. 1).It is also very helpful to do fi eld work early in spring when only queens of B. cryptarum are in fl ight (see spring phenology).

Distribution of B. cryptarum and B. magnus in Berlin and Brandenburg
B. magnus in Berlin and Brandenburg B. magnus Distribution maps do not give information about the abundance of species, but they are good fi rst approximations for further studies.In Berlin and Brandenburg B. lucorum is frequent and can be found in great numbers everywhere.B. cryptarum is less frequent, but probably distributed over the whole Area (Fig. 11).Wherever we searched with the

Phenology of spring queens of B. cryptarum and B. magnus in Berlin and Branden burg
B. magnus in Berlin and Branden burg B. magnus Fig. 12 shows the distribution of catch-dates for all spring queens that could be safely determined.If we take the interval mean ± standard deviation (theoretically including about 95% of all observations) for both species, the biological diff erence in emergence of spring queens from winter sleep is about 2 1/2 weeks.Th e data also show that in early April the probability of seeing B. magnus is very low, which gives us the possibility of B. magnus is very low, which gives us the possibility of B. magnus studying the characteristics of B. cryptarum without interference with B. magnus.Th e amount of data available from Berlin and Brandenburg, however is too small for a rigorous statistical analysis.Further studies will thus be necessary to investigate this apparent biological diff erence.

GC/MS of the cephalic labial glands of males
Typical gas chromatograms (GC) for the cephalic labial gland secretions of males of B. lucorum, B. cryptarum and B. magnus from Teupitz (LKr.Dahme-Spreewald, Brandenburg) B. magnus from Teupitz (LKr.Dahme-Spreewald, Brandenburg) B. magnus are given in Fig. 13 and the compounds are summarised in Table 1.Th e major compound in B. lucorum is ethyl 9-tetradecenoate (Fig. 7 B. magnus minor amounts of ethyl-9-hexadecenoate (peak 12) and ethyl-hexadecanoate (peak 13) were also detected.Substantial amounts of dodecanoic acid (peak 3), 9-hexadecenoic acid (peak 11) and 9-octadecenoic acid (peak 19) were also identifi ed.Hexadecenol (peak 8) and hexadecanol (peak 9) are only found in B. lucorum, and are absent in B. cryptarum and B. magnus.9,12-Octadecadienol (peak 14) and 9,12,15-octadecatrienol (peak 15) are characteristic for the labial glands of B. lucorum and B. magnus, both substances are absent in B. cryptarum.Icosenol (peak 23 & 24), docosenol (peak 27 & 28), tetracosenol (peak 32) and hexacosenol (peak 37 & 38) were also identifi ed.Whereas for B. magnus a prominent peak of icosenol (peak 24) is characteristic, in B. cryptarum docosenol (peak 28) is the dominant alkenol besides two characteristic hexacosenol peaks (peak 37 & 38) at the end of the chromatogram (see Fig. 13 and Table 1).A complex mixture of about 20 wax type esters with carbon chain lengths of between 32 -38 was found in the chromatograms.Th e characteristic MS fragment ions of the alcohol and the acid part of these esters (PEPE et al., 1993) enable the detection of small amounts of the respective alcohols or acids, which are otherwise diffi cult to detect and identify in gas chromatograms.Characteristic components in the male labial gland secretions of many bumblebee species are primary alcohols, therefore it is likely that they play a major role in communication.Both the main qualitative and quantitative diff erences between the labial gland secretions of B. magnus and B. magnus and B. magnus B. cryptarum are diff erences in primary alcohols (9,12-octadecadien-1-ol, 9,12,15-octadecatrien-1-ol, icosen-1-ol, docosen-1-ol), the species recog-  Rasse latocinctus from Sylt/Germany), Rasse latocinctus again the characteristics "lateral border of the collare" were described.KRÜGER (1954) GER ( 1954) GER raised these taxa to species status.We carefully studied the corresponding type specimens in the collections at Amsterdam, Brussels and St. Petersburg.Th ese studies signifi cantly clarifi ed the descriptions in the literature and allowed us to use the character "lateral border of the collare" as a distinguishing characteristic.As already described for B. cryptarum (BERTSCH 1997a/b), contrary to many discussions in literature, the determination of spring queens in Germany is not diffi cult.Most individuals can easily be determined while feeding on plants.Th e strongly melanised dark collare, compared to the light yellow band of the abdomen characterises nearly all females of B. cryptarum, whereas the broad, bright collare and a similar broad bright band at the abdomen makes the females of B. magnus especially brilliant against the dark-B.magnus especially brilliant against the dark-B.magnus green background of Vaccinium myrtillus and the mosses of the forest fl oor.
Vaccinium myrtillus and the mosses of the forest fl oor.Vaccinium myrtillus B. magnus is B. magnus is B. magnus very shy and all females disappear immediately if the observer does not move slowly and avoids casting a shadow.If one queen has been disturbed all the others on the site will disappear within a short time.Inspection of the characteristic "border of the collare at the pronotallobus/episternum" while the bee is in a glass vial usually confi rms the fi rst judgement.Only a few specimens need closer inspection at the laboratory with stereomicroscope.Th e situation is more diffi cult in Scotland, where spring queens of all three species are large and bright coloured.Th erefore, inspection of the specimen in a glass vial is always necessary but, again, a closer inspection of the characteristic "border of the collare at the pronotallobus/ episternum" allows a safe determination of most females in the fi eld.Only a few specimens must be transferred to the laboratory for closer inspection under a stereomicroscope.On wet and rainy days, removing the specimen from the net in most cases results in completely soaked bees and an inspection of the diagnostic characteristics is impossible.Th e bees have then to be transferred to fl ight cages and fed honey solution.Th ere they warm-up, get dry and, after a short time, start cleaning and brushing their hair so that the diagnostic characteristics come out distinctively.Determination then becomes easy.Bombus magnus from Scotland was fi rst described as a form of Bombus magnus from Scotland was fi rst described as a form of Bombus magnus B. lucorum by VOGT (1911) and raised to rank of species by KRÜGER (1954) (2000) for instance investigated a series of 32 females from Scotland from which 6 had been morphologically determined as B. magnus by B. magnus by B. magnus P. RASMONT.After measuring "how far the collare extends (dorso-ventrally) below the tegula" and the "maximum breadth (antero-posteriorly)" and plotting the standardised values P. WILLIAMS concluded, that the material investigated showed a continuum and not, as he expected, a distinct gap separating the measurements for B. lucorum and B. cryptarum.Apart from the diffi culties involved in measurements on hair patches with diff use borders, attempts at quantifying subtle morphological diff erences often fails either because the database is insuffi cient or because B. cryptarum is not treated separately as for instance in the measurements of LØKEN (1973), PEKKARINEN (1979) andBAKER (1996).

Distribution and biology
Discussion about taxonomic status is often is restricted to morphological characteristics, and the application of modern biochemical or molecular methods is generally based on such morphological investigations.But species are biological entities and therefore the insight formulated by FRISON (1926) that good morphological species are also always different in biology should not be neglected.Much more expert fi eldwork has to be done.As the taxonomical status of B. cryptarum and B. magnus is still in dispute, it is prema-B.magnus is still in dispute, it is prema-B.magnus ture to discuss their biology by comparing distribution data.If we take only safely determined queens as data, the following picture is available: B. lucorum is abundant all over Western, Central and Northern Europe, data of PAMILO et al. (1997 Fig. 1) make it probable that it is less abundant in the far north of Scandinavia and Finland, where B. cryptarum might be more predominant.In Southern Europe B. lucorum is restricted to the mountainous regions.B. cryptarum is distributed all over Europe.It is abundant in Benelux, Northern and Middle Germany, Poland and White Russia, the eastern border of the distribution in Russia is unknown though specimens have been found at Moscow.Th e available data show that B. cryptarum is less abundant in the Southern parts of these countries, abundance and distribution in the Alps of France, Switzerland and Austria has still to be investigated.On the British Isles B. cryptarum is abundant in Dartmoor, Exmoor and all over Wales and Scotland, the exact distribution has still to be investigated.B. magnus is sometimes treated as an endangered species, which most probably is not B. magnus is sometimes treated as an endangered species, which most probably is not B. magnus the case.Th e distribution is more patchy, but locally B. magnus is abundant, some-B.magnus is abundant, some-B.magnus times even the predominant species.Specimen are available from Ireland, United Kingdom (Scotland, Wales and England), Benelux, France, Germany, Poland and Russia (St.Peters burg).In Scandinavia and Finland B. magnus is restricted to the southern coast  B. magnus LØKEN (1973) belong to the bright unmelanised form of B. cryptarum, so further studies will be necessary.So far no specimens from the Alps of Switzerland and Austria are available, however, the possibility that B. magnus might occur in some valleys of the Southern Alps has to be investigated.Th e distribution of the species in Southern France, the Pyrenées and the Cantabrian Mountains has also to be further investigated.

Gas Chromatogram
Th e components of the cephalic part of the male labial glands of bumblebees used for scent marking are species specifi c, following the discussion of PETERS (1998) about speciation it could be useful to see them more as cohesion mechanism of the species than as isolation mechanism.WILLIAMS 1991WILLIAMS , 1998) ) show extensive colour variance (PEKKARINEN 1979), a variation associated with the chemical composition of the marking pheromones produced by the labial glands (BERGSTRÖM et al. 1973).Pheromonal studies revealed two distinct types of males; one representing the "blond" and the other one the "dark" form.Later BERGSTRÖM et al. (1981) regarded these forms as two sibling species on the basis of their pheromonal diff erences.As the "blond" form was characterised by the component ethyl tetradecenoate, which was detected as the main component of B. lucorum by CALAM (1969), the "blond" form was B. lucorum.In an attempt to relate the fi gures of the publication (BERGSTRÖM et al.1973 Fig. 1-14) to taxa, RASMONT et al. (1986) (PAMILO et al. 1997) gives a peculiar result.Th e 28 males analysed individually come from Hanko (southern Finland), where all three species can be expected.Th e males have not been identifi ed morphologically.Again we are left with the "blond" and "dark" form and strange enough, in the summarising table (PAMILO et al. 1997, Fig. 3) 2001) the main component of B. lucorum is ethyl 9-tetradecenoate, so we are left with the assumption that somehow in this investigation names or specimen have been mixed.Most probably as already in the investigation of BERGSTRÖM et al. (1973) only males of B. lucorum and B. cryptarum have been analysed.Again the distinction of "blond" and "dark" form without further morphological information is not helpful in analysing critical taxa of bumblebees from fi eld collections and Fig. 3 (PAMILO et al. 1997) shows that "blond" and "dark" specimen are represented in the same category so the distinction does not really separate the taxa.Our GC/MS results from Brandenburg/Germany and Scotland/United Kingdom for males, reared in artifi cial colonies from unmistakable spring queens clearly prove that carefully determined specimens result in distinct diff erent GCs.Th is diff erence of compounds of male labial glands was confi rmed by material (BERTSCH et al. unpublished) from a range of places in Germany (Nuremberg/Bavaria), Russia (St.Petersburg), France (Col de la Croix Morand/Puy-de-Dôme) and the United Kingdom (Porlock/England, Abergavenny/Wales and Glenmore Forest/Scotland), where all three species live sympatric.B. cryptarum is not restricted to Central and Northern Europe, it is also a species of the bumblebee fauna of the British Isles.

Enzyme electrophoretic data and Mitochondrial Cytochrome oxidase 1 DNA sequences
Mitochrondrial cytochrome oxidase 1 (CO1) was used especially successfully in analysing diff erent Hymenoptera (Bombus;PEDERSEN 1996, 2002, Lasius;HASAGAWA 1998, HASAGAWA 1998, HASAGAWA Lasioglossum;DANFORTH 1999, Apis;TANAKA et al. 2001a/b) PEDERSEN (2002, page 382) concludes from his investigation "although the Bombus group of species seems to form a distinct monophyletic group, observed diff erences within the group indicate taxonomic problems so severe that likely only a closer study of morphology and molecular data from several localities in Europe will delimitate the species."We think, that will not be necessary.Th e genetic relationship between the taxa of the subgenus B. s. str.was studied using protein electrophoretic data (SCHOLL & OBRECHT 1983;SCHOLL et al. 1990;SCHOLL et al. 1992).Th e phenogram of the genetic relationships of the species (SCHOLL et al. 1992, Fig. 2) PEDERSEN (2002).Using enzyme electrophoretic data of SCHOLL et al. (1992), the main components, the alcohols of the labial glands and the morphological characters "border of the collare at the pronotallobus/episternum" and "cuticula sculpture of the 2. tergite" we calculated the shortest possible phylogenetic tree with B. terrestris as outgroup (Fig. B B. magnus has until now never been reported B. magnus from such subalpine localities.According to AMIET (1996) B. magnus is not a mem-B.magnus is not a mem-B.magnus ber of the fauna of Switzerland.However an exceptionally bright form of B. cryptarum (ssp. reinigianus RASMONT 1984, Fig. 1b) can be found in the Alps of Austria and Switzerland (RASMONT 1984, Map 3;AMIET 1996).Th e specimens named B. magnus by PEDERSEN (2002) should be conspecifi c with B. cryptarum.Without morphological inspection the specimens named B. cryptarum by PEDERSEN (GenBank AY181101 Danmark and AY181100 Austria) cannot be put in their proper place.Probably no specimens of B. magnus were included in the investigations of B. magnus were included in the investigations of B. magnus PEDERSEN.

Conclusions
GC/MS of the male cephalic labial gland is an excellent tool for clarifying the taxonomic rank of critical species.For species diffi cult to separate by morphology the species recognition signals of the male labial gland secretions gives clear and unequivocal evidence.Such results can, however, only be expected if the material involved has been carefully identifi ed.Males taken for GC/MS from artifi cial colonies are useful to clarify problems in critical taxa and should be preferred to males collected in the fi eld.Fig. 14: Cladogram (branching information) calculated using morphological characteristics, data from GC/ MS analysis and enzyme electrophoretic data from Scholl et al (1992).
For coding see methods.
Th e broad, bright collare of the B. magnus queen is never melanised.It reaches far down B. magnus queen is never melanised.It reaches far down B. magnus below the tegula to both sides of the thorax (Fig. 2), with both the lower and anterior border often characteristically diff used by long yellow hair stroking in parallel to the lateral parts of the thorax (Fig. 7 & 8).Th e combination of broad bright collare and the broad bright band on the abdomen make the queen of B. magnus both especially brilliant B. magnus both especially brilliant B. magnus and unmistakable.In Scotland B. lucorum, B. magnus and B. magnus and B. magnus B. cryptarum queens are large and bright, melanisation nearly never occurs.Th e characteristic dark S-shaped band of B. cryptarum, which follows the border of the pronotallobus is very faint (Fig. 5 & 6), sometimes it is completely absent.Nevertheless the characteristic curved patch of yellow hair at the upper border of the episternum is always present, and helps to discriminate B. cryptarum from B. magnus, where the yellow hair of the collare extend much lower on both sides of the thorax (Fig. 9 & 10

Fig. 11 :
Fig. 11: Map showing the distribution of B. cryptarum (FABRICIUS) and B. magnus ( B. magnus ( B. magnus VOGT) in Berlin and Brandenburg with UTM 10x10 km grid.(Queens only, for details see appendix I).

Fig. 12 :
Fig. 12: Spring phenology of B. cryptarum (FABRICIUS) and B. magnus ( B. magnus ( B. magnus VOGT) queens in Berlin and Brandenburg.Each dot represents a date of catch.(For details see appendix I).
B. magnus is restricted to the southern coast B. magnus of Finland, Sweden and the southern and south-western coast of Norway.Contrary to B. lucorum and B. cryptarum the distribution shows a distinct border to the north.Many of the specimens determined as B. magnus by B. magnus by . terrestris as outgroup (Fig. B. terrestris 14).Th e cladogram clearly separates B. lucorum from B. cryptarum and B. magnus, which should be treated as sister species clearly distinct from B. lucorum.Th e specimens of B. magnus investigated by B. magnus investigated by B. magnus PEDERSEN (2002) were collected from Switzer land and Austria, at localities of ± subalpine/alpine habitats (Sölkpass 1790 m, Austria and Julier-Pass 2284 m, Switzerland).B. magnus has until now never been reported BERTSCH, A., et al.: Discrimination of B. cryptarum and B. magnus by morph , peak 6) compared to ethyl dodecanoate (peak 4) in B. cryptarum and B. magnus.All three gland secretions contain the usual mixture of straight-chain fatty acid derivatives (alcohols, esters, aldehydes and hydrocarbons) normally detected in the GC of the subgenus Bombus s. str.( BERTSCH, A., et al.: Discrimination of B. cryptarum and B. magnus by morph B. magnus by morph B. magnus ology nition-signals diff er signifi cantly, proving the specifi c status of both taxa.Th e chemistry of the labial glands show that B. lucorum is distinct from the species pair B. magnus and B. cryptarum, which share the same main component but diff er in alcohols.B. lucorum var.lucocryptarum from Belgium).Both descriptions refer to the lateral border of the collare as a diagnostic characteristic.RASMONT (1981) made a detailed description of var.lucocryptarum BALL and raised the taxon to species status.Th e special characteristics of B. magnus were fi rst described by TRAUTMAN & TRAUTMAN (1915; B. terrestris var. who carefully described the GER (1954) who carefully described the GER diagnostic characteristics of this species.Nevertheless the dispute surrounding the taxonomic status of B. magnus and its delimitation from B. magnus and its delimitation from B. magnus B. lucorum is on-going.WILLIAMS B. cryptarum diff ers biologically from B. lucorum by the phenology of the spring queens because B. cryptarum emerges earlier than B. lucorum (BERTSCH 1997a).KRÜGER (1939) GER (1939) GER reported for B. magnus from Sylt, that queens appear distinctly later in spring than those B. magnus from Sylt, that queens appear distinctly later in spring than those B. magnus of B. lucorum, a fact which could be confi rmed for B. magnus from Belgium ( B. magnus from Belgium ( B. magnus RASMONT1984).It therefore seems plausible that the time diff erence between the early B. cryptarum and the late B. magnus should be substantial.Fig.12shows this diff erence in the B. magnus should be substantial.Fig.12shows this diff erence in the B. magnus emergence of spring queens, though, it is still a rather sketchy picture and much more fi eld observations are necessary.It is already obvious, however, that in the fl owering period of early Salix species and Th ese compounds have been successfully used to discriminate or to recognise diffi cult bumblebee taxa, as for instance B. lapponicus and B. lapponicus and B. lapponicus B. monticola (SVENSSON 1979) and B. lucorum and B. cryptarum (BERTSCH 1997a/b).In this investigation the species recognition signals of the labial glands are used to separate B. cryptarum and B. magnus.Besides the main component ethyl 9-tetradecenoate, the gas chromatogram of B. lucorum is unique in the occurrence of considerable amounts of hexadecan-1-ol, which is absent in both B. magnus and B. cryptarum only dodecanoic acid.Correspondingly B. lucorum has a larger pattern of esters compared to B. cryptarum and B. magnus.Males of B. lucorum in the broadest sense (B.lucorum s. lat.-see came to the conclusion that most of the specimens illustrated belong to B. lucorum, some might belong to B. cryptarum and, probably, no B. magnus was included. and text B. lucorum is characterised by the presence of ethyl dodecanoate.As CALAM (1969) has shown and as was confi rmed by URBANOVÁ URBANOVÁ URBANOV et al.
. Recently PEDERSEN (2002) sequenced CO1 from B. lucorum, B. cryptarum and B. magnus.His results (PEDERSEN 2002, Fig. 4 and 5) show B. cryptarum as very close to B. lucorum (from the continent) and B. magnus as distinctly diff erent but surprisingly close to B. magnus as distinctly diff erent but surprisingly close to B. magnus B. lucorum from the United Kingdom.
based on a similarity matrix (NEI coefficient of genetic identity I), was constructed by average linkage cluster analysis (UPGMA), it clearly separates both B. cryptarum and B. magnus from B. magnus from B. magnus B. lucorum.Th is view is corroborated by our results of the GC/MS of the labial gland secretions.It is therefore very improbable that the genetic distance between B. cryptarum (Austria) and B. lucorum (Austria) is so much smaller (6 base substitutions) than the genetic distance (43 base substitutions) between B. magnus (Austria) and B. magnus (Austria) and B. magnus B. lucorum (Austria), as supposed by 20 years ago P. RASMONT by careful morphological work established the species B. cryptarum and B. magnus as part of the Central European bumblebee fauna, this judgement could now B. magnus as part of the Central European bumblebee fauna, this judgement could now B. magnus be confi rmed by biochemical methods.