The high alpine bee fauna (Hymenoptera: Apoidea) of the Zillertal Alps, Austria

Abstract Bees from the Zemmgrund area in the Zillertal Alps (Austria, Tyrol) were collected and determined to investigate the species composition of the area. A total of 61 specimens were collected over a two year period; they represent 24 species from 8 genera. Building on these records, the first commented checklist for the area is presented, with notes on habitats and visited flowers.


Introduction
Faunistic research on bees (Hymenoptera: Apoidea) in Tyrol enjoys a long tradition. Its foundations were laid by the great works of Dalla Torre (e.g. Dalla Torre 1873, Dalla Torre 1877a, Dalla Torre 1877b, Dalla Torre 1879, Dalla Torre 1882) and Schletterer (1887) that are related to the former County of Tyrol. Numerous complementary studies followed by a large number of authors concerning different parts of the region, for example the Ötztal Alps (Schedl 1982), Lower Inn Valley (Schuler 1982), Tiroler Mittelland (Stöckl 1996), Upper Inn Valley (Stöckl 1998), tyrolean Lech area (Schmid-Egger 2011) and Silvretta Alps together with Kleinwalsertal in Vorarlberg (Kuhlmann and Tumbrinck 1996). Bellmann and Hellrigl (1996) assembled a species list for South Tyrol which has been updated several times (Hellrigl 2006, Hellrigl and Franke 2004, Hellrigl 2003. Furthermore, Kopf (2008) provided an excellent treatise on the bees of the Schlern region. Overviews for different taxonomic groups of bees have been conducted by Gusenleitner (1985) regarding Andrena in Northern Tyrol and records of Halictidae from Northern Tyrol were prepared by Ebmer (1988). Stöckl (2000) worked on the Megachilinae in North and South Tyrol and Neumayer and Kofler (2005) evaluated the bumblebee fauna of eastern Tyrol. Additional faunistic data was assembled within the framework of the "GEO-Tage der Artenvielfalt" in Tyrol in 2005. Aside from the data in Kopf et al. (2010), records for the Zillertal Alps are very rare, and for the Zemmgrund area only a list of bumblebee species for a small transect study (Penninger 2008) could be found in the literature. The aim of this study is to help close this knowledge gap and contribute the first extensive faunistic data set concerning the high alpine bee fauna of the Zillertal Alps.

Materials and methods
Collections were conducted near the Berliner Hütte in three periods: July 4-10, 2012, July 3-9, 2013 and August 6-10, 2013. The focus within the area was on bees around and above the Berliner Hütte which is located on 2042 m above sea level. All sighted wild bees were collected manually and were transfered into an ethyl acetate killing jar. The majority of specimens were collected between an altitudinal range of 1850 m to 2400 m a.s.l. Four specimens were collected at lower altitudes as accidental findings during ascent and descent. GPS coordinates and altitudes were logged. The habitat of each collection site was categorized. Determinations were conducted using the identification keys of Amiet (Amiet 1996, Amiet et al. 1999, Amiet et al. 2001, Amiet et al. 2004, Amiet et al. 2007, Amiet et al. 2010, Dathe (Dathe 1977, Dathe 1980, Ebmer (Ebmer 1969, Ebmer 1971, Ebmer 1984, Gokcezade et al. (2010), Mauss (1994), Schmid-Egger and Scheuchl (1997) and Scheuchl (Scheuchl 1995, Scheuchl 2006. Critical specimens were sent to experts for examination: the author is indebted to Fritz Gusenleitner for helping with the specimens belonging to the genus Andrena and Maximilian Schwarz for specimens of Nomada. The suprageneric classification follows Michener (2007). All specimens are kept in the collection of the author. If the specimens were collected on flowers, the respective plant species was recorded. The plant species were determined using Rothmaler (2009) andFischer et al. (2008). A list of the recorded plant species in the supplementary material (Suppl. material 1) provides further informations about the bee species visiting the respective flowers.
One collected specimen belongs to a cryptic bumblebee species group, the so-called Bombus lucorum complex. The status of three distinct species within the complex is widely accepted today (Bertsch et al. 2004, Bertsch et al. 2005, Murray et al. 2008, Bertsch 2009. In contrast, there is heavy doubt about the species identification based on morphology and there are implications that the species might be morphologically indistinguishable (Williams 2000, Waters et al. 2011. Therefore, a partial sequence of the mitochondrial COI gene from the specimen was ascertained to ensure the morphological determination. The specimen was stored in pure ethanol and a single crushed midleg was used for the analysis. DNA was extracted using a Proteinase K digestion prior to a phenol-chloroform protocol (Sambrook et al. 1989). The so-called "Folmer region" was amplified with Polymerase chain reactions (PCR) using the primers LCO1490 and HC02198 (Folmer et al. 1994). The remaining PCR components were provided using the DreamTaq™ PCR Mastermix (2x) (Thermo Fisher Scientific Inc., Waltham, MA, USA) and the amplification profile was conducted following the manufacturer´s protocol. The product was purified using the GeneJET™ PCR Purification Kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) and sequencing was carried out by the VBC-Biotech Service GmbH (Vienna, Austria). The obtained sequence was checked manually using BioEdit 7.2.5 (Hall 1999). A BLAST search (Altschul et al. 1990), as implemented in GenBank, was conducted to estimate the query cover and identity to other sequences deposited in the databank.
The climate map of the study area ( Fig. 1) is based on the recently updated Austrian digital climate atlas from 1971-2000 (Hiebl et al. 2011). Given the importance of dependable data concerning the altitudinal climate changes, the high resolution Austrian climate maps for 1971-2000 consider altitudinal changes by a digital elevation model. The GIS grids were kindly provided by Alexander Orlik from the Zentralanstalt für Meteorologie und Geodynamik (ZAMG) and handled with QGIS 2.2 (QGIS Development Team 2014).
The microscope images were created using a SMZ25 stereomicroscope and a DS-Ri1 U3 microscope camera (Nikon Corp., Tokyo, Japan). A temperature-based climate map of the study area and its localization in Austria. The map shows the mean annual air temperature for the period 1971 -2000 with linearized color interpolation and is based on the data of Hiebl et al. (2011). The white circles indicate the collection localities. The mean annual air temperature of these localities, based on the years 1971 -2000, range from -0.7 to 3.7 °C.

Study area
The Zemmgrund is a valley located in the Zillertal Alps in Tyrol (Fig. 1) and is part of the Nature Park Zillertal Alps. It is located close to the main ridge of the Alps in northern direction and the southern boundary of the Upper Zemmgrund is the present-day border to Italy. Characteristic feature of the area are three glaciers, which are rapidly retreating at present (Gereben-Krenn et al. 2011): the Waxeggkees, Hornkees and the Schwarzensteinkees. The glaciers greatly influenced the geomorphology of the area and their moraines and remaining waters caused mosaics of diverse small-scaled habitats (Fig.  2). The predominant habitats of the area are alpine meadows and alpine pastures, especially above the treeline (Fig. 3). Other habitats are Swiss pine forests (Pinus cembra L.), aggregations of mountain pines (Pinus mugo Turra), tall forb meadows, dwarf shrub communities and wet meadows (Fig. 4). Great comprehensive information of the area, concerning the history, anthropogenic usage, climate and geology are available in Luzian and Pindur (2007). In addition, this compendium contains an extensive study about the flora of the area by Niklfeld and Schratt-Ehrendorfer (2007), a very useful source for melittologists! Figure 2.
A photograph of the small scaled habitats that are characteristic for study area: A part of a Swiss pine forest (left side), aggregations of mountain pines (in the middle of the picture), an alpine pasture (in the foreground) and alpine meadows (upper right side).  Wet meadows are present in the study area but are rarely frequented by wild bees.

Notes:
The species is oligolectic on Ericaceae (Gusenleitner et al. 2012). Distribution: According to Zettel et al. (2008), A. rogenhoferi is a high-alpine species distributed all over the European Alps.

Andrena rogenhoferi
Notes: Only few records of A. rogenhoferi have been reported so far (Zettel et al. 2008). Distribution: According to Warncke (1981), the species is distributed between 43°and 70° north latitude in Europe and probably reaches Asia.

Notes:
A. ruficrus is a rare species and Gusenleitner (1985) solely reports one single record of the species in Northern Tyrol.

Distribution:
The specis has an altimontane distribution in the western and central Palaearctic (Ebmer 1988 Distribution: Northern, western and central Europe (Scheuchl 1995).

Bombus bohemicus
Notes: According to Amiet (1996), Bombus lucorum is the host species of Bombus bohemicus. It is presently not known if the other closely related species of the so-called Bombus lucorum-complex serve as host species as well.  The two specimens of Nomada panzeri collected during the study. Note the great variation in size and color.

Distribution:
The species seems to have a boreal distribution in great parts of the Palaearctic and even reaches western North America .

Notes:
The specimen belongs to a cryptic species complex consisting of B. cryptarum, B. lucorum and B. magnus but could positively be determined as B. cryptarum with the analyses of the nucleotide sequence of the COI gene. For details, see the discussion.

Distribution:
The species occurs in the Pyrenees, European Alps and on the Balkan Peninsula (Amiet 1996). Further it has been mentioned for the Carpathian and Caucasus Mountains (Ponchau et al. 2006).

Analysis
In total, 61 specimens were collected, representing 24 species from 8 genera. The list of bumblebee species provided in Penninger (2008) can be complemented with 4 species. The note about B. lucorum from this source is not evaluated due to the current unreliability of morphological identification of the species. Combining these records, 30 bee species have been recorded for the area. Of these, 15 are representatives of the genus Bombus.

Discussion
With increasing altitude, the climatic conditions in alpine environments become more extreme (Franz 1979). Especially the decreasing temperature ( Fig. 1) and increasing insolation are of importance for terrestrial arthropods above the timberline (Sømme 1989). This also applies for bees: due to the short and cool summers in alpine regions in the European Alps, Neumayer and Paulus (1999) conclude a phenological window of solely three months for bumblebees to complete their life cycle. With respect to the nearly completed snowmelt in the study area by mid of June, it can be safely assumed that the three investigation periods were extensive enough to collect the majority of bee species that may occur in the area. Nonetheless, the species list above cannot be assumed to be complete with certainty. Therefore the investigation periods were not evenly distributed throughout the season since no collections have been conducted in September. Also the study area is almost completely restricted to the Upper Zemmgrund and species which potentially occur below 1900 m altitude are absent from the species list. This becomes particularly apparent when comparing the records with the species list of Kopf et al. (2010). This species list is based on collections from July 17-19, 2009 by four persons on eight collection sites approx. 13 to 17 km from the Zemmgrund Area as the crow flies. Therefore it is comparable by place, time and collection effort but differs in the altitudinal range: the collections were conducted between 1760 m and approx. 950 m a.s.l. Only seven of 47 species collected by Kopf et al. (2010) can be found in both lists, namely the widely distributed bumblebees Bombus hortorum, B. monticola, B. pratorum, B. pyrenaeus and the widespread Lasioglossum albipes, L. fratellum and L. morio. Therefore a great number of additional species can be expected at lower altitudes of the Zemmgrund area. Further the comparison of the lists indicates a decreasing species diversity along the rising altitudinal gradient. This is in line with the described species decrease of terrestrial arthropods at the timberline (Sømme 1989). However, the species composition in the Upper Zemmgrund clearly reflects the high altitudes of the study area, since the majority of species have at least a montane distribution. Several records belong to explicit highmountain species, such as Hylaeus nivalis (Dathe 1980, Dathe 2000, Andrena rogenhoferi (Gusenleitner 1985, Zettel et al. 2008, Panurginus montanus (Patiny 2003), Lasioglossum alpigenum (Ebmer 1988), Dufourea alpina (Ebmer 1988) and the bumblebee species Bombus mendax, B. monticola and B. wurflenii (Neumayer 1998).
Some species determinations must be discussed: Since females from P. montanus cannot be separated from females of the closely related Panurginus sericatus Warncke, 1972 with the key of Amiet et al. (2010), the collected females are marked with "cf". The species status of P. sericatus has been doubted (Ebmer 2001), but is valid after Patiny (2003). However, since the males can easily be assigned by the shape of the gonostylus and both sexes were observed in the same area, it seems likely that the females belong to P. montanus.
Another species with difficult determination is Bombus cryptarum. An identification based on the characteristic color patterns of queens was shown to be unreliable , and an examination on the reliability of traits described in the common keys is urgently needed since several characters overlap. In contrast, sequence analyses of the COI gene represents a confident method for identification (e.g. Bertsch et al. 2005, Murray et al. 2008, Bertsch 2009. A BLAST search of the obtained 609 bp long sequence from the collected specimen (Suppl. material 2, GenBank acc. no. KJ787691) revealed an identity of 99% with a query cover of 100% to a B. cryptarum voucher (GenBank acc. no. JQ843372.1) and the next 40 hits by total score were assigned to B. cryptarum. Therefore it can safely be assumed that the specimen belongs to B. cryptarum. The specimen shows the 'S'-shape of black hairs in the first collar ( Fig. 6), which has been considered a characteristic trait for queens in the literature (Rasmont 1981, Rasmont 1984, Bertsch 1997, Bertsch et al. 2004). After Rasmont (pers. comm.), specimens showing the "S" belong to the subspecies Bombus cryptarum cryptarum. However, since  could show that this trait seems to be unreliable, further discussion about species identification of the cryptic species of the Bombus lucorum complex based on color patterns cannot be conducted until more safely determined specimens are accessible.
As with many species of the genus, Nomada panzeri shows a great variation in color and size (Schwarz 1986). This also applies to the two collected specimens from this study which vary considerably (Fig. 5). The specimens were determined and labeled by the European expert for this group, Maximilian Schwarz, as "Nomada glabella auct." (Nomada glabella Thomson, 1870) which is a junior synonym to N. panzeri (Schwarz 1986). Burger (2005) disagrees with the synonymy and argues with the clear distinguishability described in Stoeckhert (1930). Further he proposes differences in distributional patterns between N. panzer and N. glabella and solely mentions Andrena lapponica and A. fucata as host species of N. glabella. In contrast, the majority of authors agree with the synonymy (e.g. Schwarz et al. 1996, Smit 2004, Straka et al. 2007, Westrich et al. 2008, Nilsson 2010, Gusenleitner et al. 2012). However, it can safely be assumed that the host species of N. panzer in the study area is A. lapponica. Figure 6.
The voucher of the collected queen of Bombus cryptarum. The black "S"-shape in the first collar is clearly visible.