Bulletin British Museum | (Natural History ) OLUME 60 NUMBER 7 25 APRIL 1991 i : : 5 i ' | The Bulletin of the British Museum (Natural History ), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. The Entomology Series is produced under the editorship of the Keeper of Entomology: Dr L. A. Mound Publications Manager (Entomology): Dr P. C. Barnard Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff and by specialists from elsewhere who make use of the Museum’s resources. Many of the papers are works of reference that will remain indispensable for years to come. A volume contains about 288 pages, made up of two numbers: published Spring and Autumn. Subscriptions may be placed for one or more of the series on an Annual basis. Individual numbers and back numbers can be purchased and a Bulletin catalogue, by series, is available. Orders and enquiries should be sent to: Sales Department, Natural History Museum Publications, British Museum (Natural History), Cromwell Road, London SW7 5BD Telephone: 071-938-9386 Fax: 071-938-8709 World List abbreviation: Bull. Br. Mus. nat. Hist. (Ent.) © British Museum (Natural History), 1991 ISBN 0 565 06041 4 Entomology Series ISSN 0524-6431 Vol 60, No. 1, pp. 1-204 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 25 April 1991 Typeset by Computer Typesetting Services, Redhill, Surrey Printed in Great Britain by Henry Ling Ltd, Dorchester, Dorset Bull. Br. Mus. nat. Hist (Ent.) 60 (1):1—204 Issued 25 April 1991 The bumble bees of the Kashmir Himalaya (Hymenoptera: Apidae, Bombini) PAUL H. WILLIAMS clo Department of Entomology, British Museum (Natural History), Cromwell Road, South Kensington, London SW7 5BD CONTENTS WERTROXG ECCI Et Ar Gin RET caer On ree Ee ICR BICEP ct Roca Stoner ace cee at a ten eee 2 Materialexamined|and/depositonies! 5.2.2. 000. satase snes sens sae sss dgers ne 8 Diasnosis of MimalayanBOmbiM. 2. aaa ence | See ee eee fey eee ee 10 G@haracters Studied). 7. cs sever cays agate) ne esc yaeya coepevebaseis pies ener ase ayerd wie le 2 11 Relationships among species by ancestry ............ 0. cc cece c eee ee 13 GeneraOMmBOml Dually arcccn! dese ce tsrte rae «alle Mae acdelakebegara pele ele ache a cet iais ae 14 Discrimination of species and intraspecific variation ................2...0.000005 16 The ‘biological’ species and the Recognition Concept ................0---0e eee 17 Mate-searching behaviour of male bumble beesin Kashmir..................... 18 Relationship between male mate-searching behaviour and habitat structure ....... 25 Inference of allopatric, conspecific taxa... 2.2... ccc eens Pa Inference of allopatric, separate species ......... 0... cece cece eee eee eee ee 27 Characters of species not related to male mate-searching behaviour .............. 28 The description of intraspecific variation. ................. cee cece cece e eee eee 29 INomenclatuinalistimmmtiany) rac cits «is © Sd aay Sess) = a Oa aay. eS so > a NT —— CaaS 0) SS — lo) aot eR): es, SA ES [ Y¢s0 pxa | snxaauoo moyjom doyluysodpys snupiyIssijpupy s0j2afap — SMIIuDJSayin] ~—s snuissmuDUu snlimA nafioyou xppuaut snjjaisouran | SNAWOFOTXA SNIWOGlIVANAN, ‘TS SNEWOd INISSOTONA 16 more distant affinities for the generic concept, with the result that one genus, Bombus, is recognised for all bumble bees, including Psithyrus. Of course the use of Bombus for all of these species does bring certain names originally ascribed to the genus Psithyrus into secondary homonymy with names in the genus Bombus. It also requires that qualifications concerning the socially parasitic species will have to be made more often to any generalisations about the species of Bombus. But despite this break with the most widely-established bi-generic tradition, the use of one genus for all bumble bees has the advantage that it recognises a group for which the evidence of monophyly is particularly strong, so that the nomenclature is most likely to remain stable. This single genus is also very easy for non-specialists to recognise. Within the genus Bombus, the established subgeneric names (e.g. Richards, 1968) may be found convenient by some specialists who wish to label those assemblages of species that are more closely similar to one another. Unfortunately, Richards’s usage of subgeneric names does not always communicate ideas of relationship well, because some of these assemblages now appear to be paraphyletic (e.g. Mendacibombus) or even polyphyletic (e.g. Sibiricobombus, see the com- ments on B. asiaticus and B. oberti). Of course subgeneric names can simply be treated as synonyms of Bombus Latreille and ignored by those who prefer to do so. But when a comprehen- sive cladistic revision is available, then perhaps the nomenclature of the more reliably mono- phyletic species-groups could be revised, if the system is still found to be useful. In this review, the bumble bee species of Kashmir are listed in a sequence that is chosen to reflect their relationships (after the sequencing convention of Nelson, 1972) as these are currently envisaged. DISCRIMINATION OF SPECIES AND INTRASPECIFIC VARIATION Early work on the taxonomy of Asian bumble bees showed little consistency in the delimitation of the taxa named as ‘species’ in some species- groups. In part this has been due to the changing nature of the species category as applied to living organisms. The species concept adopted for this review is briefly summarised here so that it is explicit. A particularly visible part of the problem of discriminating species of bumble bees is that some PAULH. WILLIAMS of their populations are now known to include individuals with very different colour patterns. This was not always appreciated when only small samples were available from widely-scattered localities. Conspicuous colour patterns are likely to be advantageous because they advertise a warning to potential predators against the pain- ful sting that female bumble bees can inflict. Consequently, the theory of Miullerian mimicry may explain not only the regional convergences in colour pattern among unrelated species, but also how parts of the same species may have diverged in colour pattern in different regions of its distribution (see the comments on characters not related to male mate-searching and the discussion of the colour patterns of the Kashmir fauna). It is perhaps precisely because bumble bees are such large and colourful insects that, like butterflies, they have been widely collected and have received so much attention in the literature. The consider- able effort that has been invested in the descrip- tion of subtleties of colour variation among some bumble bees has left an unfortunate nomen- clatural burden of over 3000 formal names (i.e. about 12 names per species). A large part of the more general problem in the application of the category ‘species’ to living organisms arises from confusion of the different kinds and degrees of relationship among popula- tions (reviewed by e.g. Queiroz & Donoghue, 1988). Sexually reproducing organisms have been envisaged as living in populations that are united by interbreeding, so that in the long term relation- ships within the population tend to be reticulate. This provides the possibility of some evolutionary cohesion through gene flow, even though without concerted contraction and expansion of popula- tions, random ‘diffusion’ of genes may not be sufficient to curb divergence (Barton, 1988). In contrast, relationships among populations are envisaged as tending to be more consistently divergent. In the short term, parts of a population may show restricted interbreeding with neighbour- ing parts of the population and may even show genetic and apomorphic divergence from them. Only with the benefit of hindsight is it possible to know whether divergence will be maintained or whether it will subsequently be dissipated in the broader population, should interbreeding become reinstated when the parts of the population meet. At one extreme, each of these parts of a popula- tion that may be more or less reproductively distinct in the short term could be regarded as a terminal element and given a name. But it can be seen that there is a broad range of possible relationships among these parts of the popula- tion, which may involve various degrees of THE BUMBLE BEES OF THE KASHMIR HIMALAYA inter-breeding both past and present, in differing combinations with various degrees of genetic and apomorphic divergence. Consequently, no absolute criterion can be devised that can univer- sally define the species category. Faced with the apparent impossibility of finding a universal solution to the ‘species problem’, it is still clearly desirable to try to minimise the confusion in the application of the species category. For the practice of discriminating species in this review, I have attempted to explore the application of species-defining criteria that place a greater emphasis on the potential for interbreeding (see the following section) than on general apomorphic divergence. General apomorphic divergence may have the appeal that it can be considered to have resulted in an observable pattern of monophyletic taxa, the discovery of which may be independent of any ideas of the process that caused it, which may not be discover- able. However, this rejection of process robs the ‘morpho-species’ category of any particular theory-based significance and so allows it to be defined only by a choice of some degree of apomorphic divergence that, in consequence, must be essentially arbitrary. Hence the cladistic method is not appropriate for the discovery of biological species. In contrast, the interbreeding criterion should be most closely associated with the process of gene flow within populations over the long term, which is believed to be one of the processes that can maintain some evolutionary cohesion (see Templeton, 1989). Of course there are also considerable problems with applying the interbreeding criterion, which are discussed in the next section. In practice, the application of the interbreeding criterion to many cases requires a similar degree of subjective judgement to that required by the use of the general apomorphic divergence criterion. Often the only available evidence for interbreeding is very indirect and must still be inferred from patterns of variation in characters of morphology or of molecules. The ‘biological’ species and the Recognition Concept Early authors may have regarded species as convenient classes for the typological description of a particular level of variation, but since Darwin (1859), the evolutionary significance of biological species has been stressed. This biological concept defines species of sexually reproducing organisms by a lack of interbreeding between populations, where these overlap. One of the problems with the biological species concept is that it does not solve the problem of interpretation for populations 17 that do not overlap at present (reviewed by e.g. Splitter, 1982), because in its simple form it does not show how to obtain direct and representative evidence of the ‘potential’ for interbreeding be- tween individuals that do not meet under natural conditions. Dobzhansky (1937) suggested that isolating mechanisms have been selected for because they perform the function of preventing interbreeding between species. Isolating mechanisms were said to have the advantage of protecting their ‘more harmonious’ gene assemblages from disruptive recombination. But since isolating mechanisms might be selected for only when populations with such gene assemblages overlap, Dobzhansky’s Isolation Concept provides little help in the assessment of the potential for interbreeding among non-overlapping (allopatric) populations. Paterson (1980, 1985) criticised Dobzhansky’s Isolation Concept of the species in response to difficulties with the population genetics of the selection of isolating mechanisms. Paterson’s Recognition Concept defines a species by the common ‘specific-mate recognition systems’ that are shared by individuals in the population. From this viewpoint, individuals of different species are not expected to interbreed because they do not recognise one another as potential mates (thus ‘postmating isolating mechanisms’ [Mayr, 1963] are, strictly speaking, incidental to delimiting species). Paterson stressed that the breakdown in recognition at speciation might occur in allopatric populations, but only indirectly, as a consequence of other evolutionary divergences and not by reinforcement of hybrid disadvantage (i.e. ‘pre- mating isolating mechanisms’ [Mayr, 1963] evolved as effects, and were not selected for their isolating function). Evidence of this particular kind of divergence in allopatry might help in the assess- ment of the potential for interbreeding among non-overlapping populations. Paterson’s Recognition Concept is not entirely satisfactory in that any allopatric populations with effective postmating isolation but no premating isolation would apparently be included in the same species. If these populations subsequently made contact, there would then be a selective advantage for reinforcement by premating isola- tion, as envisaged in the Isolation Concept. Ultimately it is likely that both kinds of specia- tion process have occurred, so that a major contribution of the Recognition Concept may be its emphasis on the role of premating barriers or specific-mate recognition systems. Changes in these specific-mate recognition systems could evolve rapidly by sexual selection (see West- Eberhard, 1983) in allopatry, despite stabilising 18 selection. Furthermore, neither stabilising selec- tion nor sexual selection necessarily imply uniform selection pressures across a geographically exten- sive population and so they do not preclude the possibility of clinal variation in the recognition system (see Verrell, 1988). But if the characters that are crucial to specific-mate recognition were known to differ among individuals of at least some of the problematic disjunct populations, then the Recognition Concept might provide better grounds for discounting a potential for inter- breeding and gene flow in these cases (e.g. Vrba, 1985). In principle, the most widely-applicable morphological criterion for the discrimination of species might therefore be found where there are differences in the characters that function in specific-mate recognition. However, the Recogni- tion Concept of species does not require individuals of separate species necessarily to differ at all in morphology (or in ecology, see Hengeveld, 1988) and the search for key characters for the discrim- ination of species has indeed been unsuccessful (reviewed by Templeton, 1981). This search can be traced to the classical concept of species as typological classes, whose members should share some particular similarity or ‘essence’. More recently species have been viewed as individuals (Ghiselin, 1975), characterised instead by their internal organisation, within the population (Hull, 1980). This organisation could be provided by their shared specific-mate recognition systems and potential for interbreeding. Yet from a cladistic viewpoint, free interbreeding within a population is a shared ancestral characteristic and so does not define a taxon (Rosen, 1979). Cladists would therefore be obliged to recognise some para- phyletic terminal elements, because interbreeding populations are not always distinguished by the possession of convenient apomorphies by all individuals (e.g. Ackery & Vane-Wright, 1984; Queiroz & Donoghue, 1988). If apomorphic characters were always to become fixed through- out a population as a necessary part of the process of speciation, then these character states might fulfil the role of an essence. In contrast, character states correlated with specific-mate recognition systems are not essences because they may only be expected to evolve as secondary consequences of reproductive processes and they may occur in just one of the sexes. So these characters can provide no more than a guide for the inference of the potential for interbreeding among in- dividuals as parts of the same species in certain cases. PAULH. WILLIAMS Mate-searching behaviour of male bumble bees in Kashmir The mate-searching behaviour of male bumble bees represents some of the first stages in the specific-mate recognition systems of these insects (cf. Paterson, 1985: fig. 2). In Kashmir, differences in male mate-searching behaviour among some closely-related species are associated with dif- ferences in some of their morphological character states, such as eye size and shape, antennal length and thoracic shape. Two contrasting examples are summarised briefly here (see the comments on the inference of allopatric, conspecific taxa & on allopatric, separate species). By analogy, these differences might provide a particularly strong form of morphological evidence from which to discriminate certain other likely species, even among preserved specimens in collections. The use of characters related to specific-mate recognition systems for discriminating species still requires support from the study of the subsequent stages in the operation of these systems. Key components in this may involve certain volatile or contact pheromones. The head-gland secretions of males of European ‘patrolling’ species are believed to act as female arrestants (Svensson, 1979) and some of these secretions have been characterised (e.g. Bergstrom et al., 1981). This has not been undertaken for female head-gland secretions, which may be just as important in the subsequent stage of specific-mate recognition by males (see Free, 1971; van Honk et al., 1978). Some method to assay for behavioural discrimina- tion between secretions of different composition by the organisms themselves in the field is now needed. In 1985 I recorded the distribution of mate- searching activity in time and space from transect counts of numbers of male bumble bees around Gulmarg. Distribution with respect to altitude is described from altitudinal transects of the Pir Panjal range (the mountains on which Gulmarg is situated), between Tangmarg at 1800 m in the Vale of Kashmir, and the peak of Mt Apharwat above at 4143 m (13,592 ft) (total number of bumble bees identified > 2000, see Fig. 6). The habitat is described in more detail, together with the altitudinal distribution of foraging females, in the discussion of distribution with altitude. Spatial and temporal pattern in the distributions of mate- searching males can also be described at the smaller scale of a 30 m-high hill at Gulmarg, from vertical and horizontal transects (Figs 7 & 8). I also recorded the activities of males that had THE BUMBLE BEES OF THE KASHMIR HIMALAYA Altitude in feet (metres): LES 13500 (4115) Alpine 13000 (3962) 19 12500 (3810) 12000 (3658) 11500 (3505) 11000 (3353) 10500 (3200) Subalpine 10000 (3048) 9500 (2896) 9000 (2743) 8500 (2591) 8000 (2438) 7500 (2286) 7000 (2134) 6500 (1981) Upper forest Gulmarg meadow Lower forest Terraces trifasciatus pyrosoma hypnorum tunicatus FEMALES 13500 (4115) Alpine 13000 (3962) lucorum rufofasciatus himalayanus kashmirensis avinoviellus lepidus melanurus astaticus biroi keriensis 12500 (3810) 12000 (3658) 11500 (3505) 11000 (3353) 10500 (3200) 10000 (3048) 9500 (2896) 9000 (2743) 8500 (2591) 8000 (2438) 7500 (2286) 7000 (2134) 6500 (1981) Subalpine Upper forest Gulmarg meadow Lower forest Terraces @ Patrolling males, same colour pattern as females © Patrolling males, pubescence predominantly yellow @ Foraging females A Cruising males @ Territorial males * Racing males Fig. 6 Distribution of mate-searching males (above) and foraging females (below) of the social bumble bees (i.e. excluding species of the subgenus Psithyrus) with altitude on Mt Apharwat in the Pir Panjal Range around Gulmarg. Altitudes are estimated from the 500 ft contours of the 1 : 63,360 British Survey of India Map (Oxford University Library). Records are combined from transects made during July, August and September 1985. For a description of the vegetation in each habitat zone, see the discussion of distribution of the fauna with altitude. been individually labelled, using the plastic tags that were developed for use with queen honey bees (52 male bumble bees were labelled in the vicinity of the 30 m hill in 1985; 108 males labelled in the same area in 1986). Numbered tags are ideal for the many perching males, which can be approached for their tags to be read, but plain discs of different colours are all that can be seen on patrolling bees, which spend most of their time in rapid flight. Labelling individuals did not obviously affect them adversely. For instance, one B. tunicatus male had resumed patrolling within 10 minutes of being released. Activities of males were recorded in part to compare their time 20 PAULH. WILLIAMS ee oe Perches of racing B. asiaticus males (29.viii.1986). @ _ _ Contested perch area of two territorial B. rufofasciatus males (20.viii.1985). Cc] Circuit of places approached by a patrolling B. funicatus male (14.viii.1985). Fig. 7 Sketch map of the 30 m hill on Gulmarg meadow, showing the distribution of mate-searching activity by individually-labelled males of three species of bumble bees. Dotted lines show contours at 10 m intervals, black rectangles show the positions of two huts. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 30°C Air temperature 20°C sunrise 10°C 21 sunset B. ferganicus B. tunicatus B. rufofasciatus B. asiaticus Time: 5.00 hrs 10.00 a Marking Mate-searching Feeding 10 individuals 15.00 20.00 hrs Fig. 8 Distribution of activity by males of four species of bumble bees between dawn and dusk on the 30 m hill in Gulmarg meadow. Counts were made every 30 mins during 19.viii.1985, along a transect of 450 paces around the 30 m contour at the top of the hill (see Fig. 7). budgets, but also to record the nature and out- come of interactions between males. This included mapping the circuits followed on the 30 m hill by males of each of three species (B. ferganicus, B. melanurus and B. tunicatus), and mapping the perch positions of other species (B. avinoviellus, B. rufofasciatus and B. asiaticus). The reactions of males of all species to moving objects of different sizes were observed and tested. Young queens of B. asiaticus were tethered with lengths of thin black cotton at different places on the 30 m hill and the reactions of perching males were noted. 2D; Number of records oO fo secs 10 secs Number of records PAULH. WILLIAMS B. rufofasciatus 30 secs LJ Flight durations 20 secs Perching durations B Medians 30 secs ter Fig. 9 Duration of consecutive flight and perching periods for mate-searching males of B. rufofasciatus (above) and B. asiaticus (below). For each species, records of 20 consecutive flight and perching periods were made from each of five males on the 30 m hill in Gulmarg meadow, during 11—12.viii.1985. Four kinds of male mate-searching behaviour were observed during 1985. These are referred to here as ‘patrolling’, ‘racing’, true ‘territorial’ and ‘cruising’. The territorial behaviour can be considered as an example of what has been termed pure dominance or lek polygyny, whereas the other three categories of behaviour represent divergent forms of scramble competition polygyny (see Thornhill & Alcock, 1983). (A) PATROLLING behaviour has been described previously from British bumble bees and from most other European species (e.g. Darwin, see Freeman, 1968; Awram, 1970; Schremmer, 1972; Alford, 1975; Svensson, 1979). These males appear to mark objects early each day with an oral secretion, which presumably contains a pheromone. The objects are numerous, spaced by any distance from a few centimetres to some tens of metres, and chosen so that they can be patrolled as ‘approached places’ on a circuit flown regularly by the male. It has been suggested that receptive females are arrested at these pheromone ‘traps’ and that mating takes place when the male returns on his next patrol of the circuit (see Svensson, 1979). Examples of species whose males perform patrolling behaviour in Kashmir are B. pyrosoma and B. tunicatus (e.g. records from the 30 m hill, see Fig. 7). These males did not interact with one another directly nor did they respond to moving objects by pursuit, either near or between approached places. (B) RACING behaviour differs from patrolling in that each male perches in one small area and will then pursue potential mates from this look-out position, but without competing with other males for preferred perches. Racing behaviour is described here from B. asiaticus. These males perched on rocks, cowpats, THE BUMBLE BEES OF THE KASHMIR HIMALAYA leaves or fragments of wood (a maximum of 107 males were recorded at their perches on the 30 m hill on any one day, 29.viii.1986, Fig. 7). The perches were not associated with flowers (i.e. food resources) or nests (no nests of any species were found on the 30 m hill in either 1985 or 1986). Males would leave their perches in darting flights in pursuit of rapidly moving objects, which included conspecific queens and other males in flight, birds flying overhead and even stones thrown past the perch. Males rarely performed any slow inspec- tion flights around perches (as shown by the short durations of flights by B. asiaticus males in Fig. 9). Successful pursuit flights with conspecific queens ended with the male bringing the queen to the ground, where he grappled with her, and attempted to mount her and clasp her sting base with his genitalia. There was no obvious courtship. When a queen was tethered, males approached rapidly and hovered for a few seconds, antennae almost in contact, before attempting to clasp her. Once a male had succeeded in clasping an untethered queen with his genitalia, she ceased pushing at him with her legs and the pair immediately flew from the area in tandem. When a pursuit flight with a conspecific male resulted in an approach within a few centimetres, they diverged without clashing and returned to their perches without displacement of either male. Males were rarely found to have damaged limbs (2/108 labelled males, 1986, 30 m hill). Individual males returned to the same perches on many days (the maximum period spanned 27 days, 2-29. viii. 1986, male red-83, 30 m hill) and other males rarely usurped these perches after they had become vacant (occupancy is known to have changed for 5/170 perches 2. viii-11.ix.1986, 30 m hill, though only 108 males were individually labelled). Marking of perches with oral secre- tions was observed only occasionally, and only immediately after males arrived at their perches in the morning. Some of these males were found to spend the night torpid in tight clusters of up to 10 individuals, deep among the short plant stems and roots on the middle slopes of the 30 m hill. Some individuals (e.g. male red-18, 1986) switched between clusters on successive nights, although this could have been the result of disturbance. (C) TERRITORIAL behaviour differs from patrolling in that each male perches in one small area and will then pursue potential mates from this look-out position, but unlike racing behaviour, the males do compete for preferred perches. Territorial behaviour is described here from B. rufofasciatus (e.g. on the 30 m hill, see Fig. 7). 23 These males perched and pursued moving objects much as racing males did, but they also responded to conspecific males on their perches as well as in the air, and clashed audibly with them during extended chasing flights (Fig. 9). This may have contributed to the higher incidence of damage to appendages among these males (5/20 males labelled on the 30 m hill, 1985, had lost at least half of one antenna or leg) and did result in displace- ment from perches. For instance, on 18.viii.1985 male green-7 was perched above the entrance to a 3 m-long, 1 m-deep gully at the edge of the flat- topped 30 m hill. It left this perch to engage yellow-7 as it entered the gully, but after the ensuing chase, only yellow-7 returned to the gully, where it then perched within 10 cm of where green-7 had been. Later the same day green-7 returned, chased and replaced yellow-7 in this gully (occupancy changed at least 14 times for 9 perches 9—20.viii.1985, with or without contest, and 10 of these changes were for this gully). Yellow-7 then moved to a less frequently occupied perch area that was 10 m away on the side of the hill. These perches were not associated with nests or with concentrations of resources, but neverthe- less continued to be highly preferred despite a high turnover in individual males (only 9 perch areas, each 2-4 m in diameter, could be found on the entire 30 m hill, but 20 males were labelled in these areas between 9-20. vili.1985). Males of B. rufofasciatus marked their perches with oral secretions much more frequently than did males of B. asiaticus. The frequent inspection tours of B. rufofasciatus males around perch areas, such as the gully at the top of the 30 m hill, have a characteristic slow, stalling flight, with a low, interrupted buzz. Pursuit flights, like those of B. asiaticus, are much faster than the flight of females and have a higher- pitched buzz. (D) CRUISING behaviour differs from patrolling in that each male uses one small area as a look-out position from which to pursue potential mates. Unlike racing behaviour, the males hover almost stationary in the air rather than perch, and unlike territorial behaviour the males do not compete for preferred positions. Cruising behaviour was not seen on the 30 m hill and is described from observations of B. kashmirensis from just a few days (most on 23.vili.1985) near the peak of Mt Apharwat. These males maintained a very slow, almost hovering flight about 1 m above the ground, but each was confined within an area only 34 m across. From these aerial positions they pursued moving objects, much as racing and territorial .ntennal Shape 24 PAULH. WILLIAMS SEXUAL DIMORPHISM (males/queens) asiaticus melanurus elongate male antennae kashmirensis trifasciatus avinoviellus C) pyrosoma himalayanus 75 |e hypnorum rufofasciatus birot @ keriensis js Fens () 1.0 undifferentiated male antennae reduced male eyes narrowed LS male thorax enlarged male eyes Eye shape Fig. 10 Sexual dimorphism in some characters that may be related to male mate-searching behaviour among the social species of bumble bees (i.e. excluding species of the subgenus Psithyrus) of Mt Apharwat in the Pir Panjal Range. Shape dimorphism is measured as the ratio of mean shape of males/mean shape of queens, so values of 1.0 show no dimorphism in shape between males and queens. Eye shape is measured as the-ratio of the maximum right compound eye length/minimum dorsal inter-compound eye distance; antennal shape is measured as right flagellum length/right scape length; and thoracic shape is measured as distance between and including tegulae/right radial cell length. Sample sizes (males/queens): B. avinoviellus (10/10), B. himalayanus (10/3), B. trifasciatus (8/5), B. melanurus (10/10), B. kashmirensis (10/10), B. hypnorum (10/4), B. lepidus (2/6), B. biroi (6/5), B. tunicatus (10/10), B. lucorum (8/3), B. asiaticus (10/10), B. keriensis (4/8), B. pyrosoma (10/10), B. rufofasciatus (10/10). Symbols show the different kinds of male mate-searching behaviour and colour pattern dimorphism (see Fig. 6, strong divergence in male colour pattern from worker colour pattern is only apparent among patrolling species): filled circles — patrolling males with similar colour pattern to females; open circles — patrolling males with pubescence predominantly yellow; star — racing males; squares — territorial males; triangle — cruising males. America have also been seen to return to perches between intermittent pursuit flights (e.g. Schremmer, 1972; Haas, 1976; Alcock & Alcock, males did on the 30 m hill. Cruising males also interacted in flight, but no clashes with violent physical contact were detected, no males were Thoracic sha found to have damaged limbs (0/12 of the males collected) and no displacement of males was seen. Marking of vegetation with oral secretions within the cruised area was seen. Patrolling behaviour is well known for bumble bees, but the other three categories are distin- guished here for the first time. Males of a few species from central Europe and from North 1983). Pairs of males of some of these species were seen to engage in protracted interactions, at least occasionally, so their behaviour may correspond to the territorial behaviour described here (e.g. Schremmer, 1972; Alcock & Alcock, 1983). Sexual dimorphism in relative eye size, antennal length, wing shape and thorax shape is least strongly developed in patrolling (‘marathon’) species and more pronounced among the other ‘perch and THE BUMBLE BEES OF THE KASHMIR HIMALAYA sprint’ categories (e.g. review by Schremmer, 1972), especially for the racing species (Fig. 10). The species that patrol exclusively in the upper forest show pronounced sexual dimorphism in colour with similar, almost uniformly bright yellow males (see the discussion of colour patterns of the Kashmir fauna). Relationship between male mate-searching behaviour and habitat structure It seemed likely from a comparison of the male mate-searching behaviour among the bumble bees observed during 1985 that each of the four categories could have advantages for increasing a male’s number of matings. But which of these kinds of behaviour is most advantageous to an individual male for maximising his mating success is likely to depend on the particular conditions of density and dispersion in the distributions of receptive females. This scheme is summarised in Table 2. Each species’ characteristic density and dispersion of receptive females is expected to reflect the particular responses by its individuals to the different kinds of habitat in the mountains: Table 2 Summary of the likely characteristics of the distributions of receptive female bumble bees on Mt Apharwat (altitudes 2700-4143 m), for the most abundant species in the four principal kinds of habitats (1985 survey). These different distributions may each favour one of the four groups of male mate- searching behaviour (in bold). Dispersion high low (even) (clumped) meadow high alpine high RACING CRUISING e.g. B. asiaticus e.g. B. kashmirensis Density forest subalpine low PATROLLING TERRITORIAL e.g. B. pyrosoma e.g. B. rufofasciatus (A) PATROLLING. Males of all species found around Gulmarg to be mate- searching only within the forest (1900-3000 m) (e.g. B. pyrosoma), and males of the species most widespread among the different kinds of habitat (e.g. B. tunicatus), were patrolling in search of mates (Fig. 6). The densities of all bumble bees are expected to be especially low in the coniferous forest, where floral resources are scant (see the discussion of distribution with altitude). Dense forest has a complex physical structure, which reduces the visibility of potential mates and obscures any long 25 range visual cues that might be used by individuals to aggregate for pairing. A web of pheromone traps, formed by the approached places that these males patrol, should be especially well suited for collecting mates at low, relatively uniform densities. Patrolling behaviour is also likely to work especially well in open habitats where young queens are present at low density and may not be predictably aggregated (e.g. arctic environments). (B) RACING. Resources in the large meadow of Gulmarg are apparently much richer, because it appears to have the highest density of bumble bees among all of the areas surveyed on Mt Apharwat (see the discussion of distribution with altitude). For species specialising in this open habitat, patrolling behaviour might be expected to break down at high density if some males were to exploit the pheromone ‘traps’ at approached places by perching there and waiting for the more frequent females to arrive. At high density, males that persisted in patrolling would be wasting much of their time in flight between traps (when patrolling males may not be responsive to recep- tive females, see Free, 1971), without increasing their chances of finding a receptive female. Racing B. asiaticus were present at more uniformly high densities than patrolling species, in some open, grassy areas below the forest (1800— 1900 m), in parts of the meadow at Gulmarg (2650-2700 m) and just above the upper edge of the forest (3000-3100 m, Fig. 6). Males appeared to be less aggregated than those of the cruising or territorial species, which are associated with the higher relief, alpine habitats. This presumably reflects either a more even distribution of recep- tive females, or the unpredictable spatial pattern of this distribution in the meadow. Transect counts show a tendency for males to concentrate on steep ridges (vertical transects 16 & 23.viii. 1985: 53/86 males in steepest middle half of slope, 7.5—22.5 m up the 30 m hill; horizontal transect 18.vili.1985: 23/26 males on ridges rather than gullies) of south and east-facing slopes (i.e. facing the sun in the morning, see Fig. 7), but perch sites were still relatively widely and evenly spread. The mate-searching activities of B. asiaticus males were also spread over a longer period of weeks than those of the territorial or cruising species in subalpine and alpine habitats. This may reflect the longer season during which floral resources are available to fuel colony development and the rearing of young queens at the lower altitude of Gulmarg. (C) TERRITORIAL. Above the forest, in the subalpine zone, the season of profitable foraging 26 is shorter, the levels of resources may be lower than in the forest meadows, and so the bumble bee density is expected to be lower (see the discussion of distribution with altitude). The open subalpine zone covers a large area on Mt Apharwat, although it also provides abundant topological cues that could be exploited to facilitate the meeting of the sexes even if they are present at low density, for instance by local ‘hilltopping’ (hilltopping could arise through the evolution of female choice, see Kirkpatrick, 1982). The most abundant species above the forest is B. rufofasciatus. Males were indeed seen to be very closely aggregated in small groups on the tops of shoulders protruding from the subalpine slopes of Apharwat (3000-3800 m) and, more rarely, in the alpine zone (c. 3900 m) and on top of the 30 m hill in the meadow at Gulmarg (2700 m, two perch areas shown in Fig. 7). At any one time these males were few in number, but were all closely associated with even fewer perch areas. So although perch areas were widely spaced, they each had up to seven males flying around them. If this does reflect the pattern in the distribution of receptive females, then territorial defence is likely to be more advantageous for these males than for species at higher densities, because it could win exclusive access to females at an especially attrac- tive site, while the probability of missing a female during a male-male chase would be lower. (D) CRUISING. Habitat conditions in the alpine zone resemble those in the subalpine, but are even more extreme in features such as the short foraging season (see the discussion of distribution with altitude). However, this open habitat has the ultimate topological cues for potential mates at low density to meet by hilltopping — the mountain peaks. The most abundant high alpine specialist is B. kashmirensis. Workers forage throughout most of the alpine and subalpine zones (3000-4000 m, Fig. 6), but cruising males were found to be aggregated exclusively in the high alpine zone (4000-4100 m), in the lee of the peak of Mt Apharwat (4143 m). Moreover, males were seen on just a few of the days during which this area was visited, although in larger aggregations than the males of B. rufofasciatus, so that mate-searching activity may be highly concentrated in time as well as in space. Both factors could contribute greatly to enhance the encounter rate or effective local density of mates. If the female arrival rate were sufficiently high, then territorial chases between males might result in more females being missed, or lost to ‘sneaky’ males, than would be lost to tolerated competitors. The hovering flight of cruising males PAULH. WILLIAMS may be one way of gaining an advantage over perching competitors, both by saving time on take-off in pursuit of a potential mate and possibly by increasing the field of view. Thus it may be that, whereas racing males tolerate one another’s proximity because no one perch site is predictably much more attractive to females than any other, a male cruising over a particularly attractive site tolerates the proximity of another male because displacing the competitor might cost more in lost opportunities than could be gained through attempts at sole possession. The characteristics of density and dispersion in the distributions of receptive female bumble bees could not be measured directly. Young queens are not often seen before winter. More particularly, bumble bees have rarely been found in copula, despite their abundance. Preliminary searches had shown that male marking activity is greatest in the early morning, so I had expected that searching at this time might yield more observations of pairing. In fact this was seen only four times, all in the afternoon (e.g. Fig. 8). An explanation for the rarity with which bumble bees are seen paired, despite the length of time for which females remain clasped by the males (usually in excess of 20 minutes, pers. obs.), may lie with the observa- tion that paired females and males flew in tandem from the areas where males were mate-searching as soon as the males had clasped the females, which took just a few seconds. This scattering to inconspicuous places may help to avoid compet- itive interference from other males of the same species. Unfortunately it also precludes the use of any observations of paired bumble bees as a measure of the distribution of receptive females. Less direct methods may nonetheless reveal the distribution of receptive females. It can be assumed that it would be to the advantage of each indivi- dual male if the male were to invest its mate- searching efforts in time and space in a pattern that closely resembles the predictable component in the distribution of receptive females. Therefore the density and dispersion of mate-searching males should reflect the pattern for receptive females. Estimates of density and dispersion of mate- searching males were made from transect counts in each of the principal habitats at Gulmarg during the summer of 1986 (Tables 3 & 4). Only 54 mate- searching males of B. pyrosoma, B. asiaticus, B. rufofasciatus and B. kashmirensis were recorded during this transect census. The spring thaw had been unusually late and the summer was cold and wet. The peak of mate-searching activity by males of the alpine and subalpine bumble bee species THE BUMBLE BEES OF THE KASHMIR HIMALAYA Table 3 Estimates of mean/variance ratio in the number of mate-searching males among 160 m segments of transect on Mt Apharwat as a measure of male dispersion (12 & 14.ix.1986). The figures were calculated for each of the most abundant species only within its associated habitat (see Table 2). Dispersion high low (even) (clumped) B. asiaticus 0.51|B. kashmirensis 0.21 B. pyrosoma 0.61 | B. rufofasciatus 0.14 Table 4 Estimates of mean number of mate-searching males counted along each 1 km of transect on Mt Apharwat as a measure of male density (12 & 14.ix.1986). The figures were calculated for each of the most abundant species only within its associated habitat (see Table 2). high 8B. asiaticus 8.8 B.kashmirensis 4.4 Density —— low B. pyrosoma 3.1 B. rufofasciatus 1.7 was about 20 days later in 1986 than it had been in 1985 and their density also appeared to be lower in 1986 (e.g. males of B. rufofasciatus were found on the 30 m hill at Gulmarg during 1985 but not during 1986). Further counts were prevented by the first heavy fall of snow, which blanketed the subalpine and alpine zones the morning after the first complete transect census. Despite the small size of the sample in 1986, the figures in Tables 3 & 4 do appear to corroborate the expected characteristics of the common bumble bee distributions as postulated in Table 2, at least insofar as the figures in adjacent cells of the tables differ in the expected directions. Of course this result is only an apparent correlation based on observations of the bumble bees on one mountain and does not constitute a test of the relationship. The behaviour of one group of B. rufofasciatus males on a high, steep-sided shoulder of Mt Apharwat (c. 3700 m, 23.viii.85) provides stronger support for the importance of distribution, although it is still only circumstantial evidence. In this situation their behaviour was apparently not territorial but resembled the cruising of B. kashmirensis. Conforming with the pattern in Table 2, these B. rufofasciatus males had formed a large group of unusually high density (estimated to be 50-100 males). However, it must be noted that individuals of most species are apparently not this flexible in their behaviour. For instance, B. asiaticus males were never found patrolling, even where they were present at low density. No males of patrolling species were found racing, 27 either at localities with unusually high density or otherwise. If differences in habitat structure between separated (allopatric, peripatric or parapatric) parts of an ancestral population may affect the balance of advantage among the different kinds of mate-searching behaviour, then the mate- searching behaviour of the descendent populations could diverge very rapidly by sexual selection. This may be particularly likely to affect the specific-mate recognition system and could result in speciation (see Rice, 1987, for a similar model for sympatric speciation). Inference of allopatric, conspecific taxa Several nominal taxa that are closely similar in morphology show pronounced variation in the colour pattern of the pubescence among adjacent regions. One example is provided by individuals with the two principal colour patterns that are ascribed to B. asiaticus in Kashmir (see the comments on B. asiaticus). Individuals with both the unbanded yellow pattern (e.g. Figs 371-376) and the banded grey-white pattern (e.g. Figs 386, 387, 389-391) of B. asiaticus occupy similar open meadow or steppe habitats on either side of the Great Himalaya range (Map 48). The mate- searching behaviour of the males in the two regions could not be distinguished and consists of the racing behaviour that is unique to these nominal taxa among the bumble bees of Kashmir. This close similarity in mate-searching behaviour is reflected in the similarity of the associated morphological characters of the males, such as their much larger eyes and longer antennae in relation to those of the females (Fig. 10). Support- ing evidence for interbreeding between individuals with the different principal colour patterns is provided by specimens with intermediate colour patterns (Figs 377-382) from a few localities in high valleys of the Great Himalaya. The frequencies of the different states for two of the three variable colour characters examined from these localities is consistent with simple models of population genetics (see the discussion of variation of B. asiaticus in Kashmir). So as parts of what appears to be a single, interbreeding population, these nominal taxa are considered to be parts of a single species. Inference of allopatric, separate species Discontinuous geographical variation has also been found in the morphological characters that are associated with the different kinds of male mate-searching behaviour, between what are in 28 other respects very similar nominal taxa. B. wurflenii is widespread in Europe among the upper forests of mountains, where the males, which patrol their scent-traps within the forests in search of mates (pers. obs., Austrian Alps), have eyes similar in size to those of the females. The morphologically similar B. kashmirensis is wide- spread in the Himalaya and eastern Tibet (Map 31) among the open alpine zones. The principal difference from B. wurflenii is that the males of B. kashmirensis possess slightly-enlarged eyes relative to the females (Fig. 10). On Mt Apharwat in Kashmir at least, these males apparently aggregate in the neighbourhood of mountain peaks, where they ‘cruise’ in search of females by sight. If this behaviour is representative and constant, then individuals of the two taxa would be unlikely to interbreed even if they were to occur on the same mountain, because they appear to search for mates in different ways and in different kinds of habitat. They may therefore have diverged in at least the first stages of their specific-mate recognition systems and so can be considered to be separate species (see the comments on B. kashmirensis, and on another group of more or less parapatric sister-species, B. simillimus, B. pyrosoma and B. rufofasciatus). Characters of species not related to male mate-searching behaviour Unfortunately the mate-searching behaviour component of the specific-mate recognition system only differs sufficiently among species for diver- gence in associated morphological characters to be apparent in a few cases. So, of necessity, discrimination of most species still depends on the interpretation of even less direct evidence of interbreeding. This is usually inferred from a diagnosis by a combination of other morphological characters. Specimens are now available from many more localities in Kashmir so that a more comprehensive examination of the variation in these characters among individuals can be made. At the beginning of this study, cladistic methods were used to divide material in collections into what are likely to be monophyletic taxa by character state patterns of the male genitalia. Material was further sorted by sample sites. Patterns of variation in a broad range of characters were then examined for description, both within and among these groups of individuals. For the discrimination of species, particular attention was paid to any strictly coincident discontinuities in the patterns of character variation among. individuals, irrespective of whether they occur between the supposedly monophyletic taxa PAULH. WILLIAMS or within them. These coincident discontinuities may provide evidence for barriers to interbreed- ing between populations and so may show the presence of separate species, which need not correspond to monophyletic taxa. On the other hand, where individuals were found that show intermediate or mixtures of character states between the otherwise differen- tiated taxa, this is interpreted as evidence in favour of the occurrence of hybridisation between them, as parts of a single population and a single species. For instance, in Europe wide variation is known to occur between parapatric parts of the population of B. soroeensis (Fabricius). Individuals of this species from much of Russia, Scandinavia, Britain and Spain are banded yellow bumble bees with white ‘tails’. Individuals from central and south-eastern Europe are unbanded, red-‘tailed’ bees. But individuals from the intervening areas in both France and eastern Europe show a continuous range of variation in both characters between the two, more widespread, colour patterns (Reinig, 1939: fig. 10). In some areas, individuals with both principal colour patterns may even be reared from the same nest, apparently as the offspring of the same queen (Vogt, 1909). Therefore individuals with either of these two colour patterns have been considered to be interbreeding as parts of the more or less continuous population of a single species. Zones of hybridisation between nominal taxa can be very narrow, with steep clines in character frequencies (see e.g. Mallet, 1986; Hewitt, 1988). For instance, localised steep clines in hybrid zones are well known for groups of colour pattern characters within populations of certain European bumble bees (e.g. Reinig, 1970). Yet as long as some evolutionary cohesion could be maintained by the passage of some alleles across a hybrid zone through interbreeding, then the parts of the population could be considered to be parts of a single species. Among Sino-Himalayan bumble bees, some of the most difficult decisions concerning conspecific status involve bees from the montane forests (e.g. see the comments on B. lepidus, B. pyrosoma). These forests are confined to a long, narrow ribbon around the southern and eastern edges of the great Tibetan massif. In this situation, any hybrid zones across this ribbon are likely to be much shorter than those described for B. soroeensis in Europe and so more difficult to locate and study adequately. If some of the provisional decisions reached in this review in cases where the available information remains inconclusive seem unduly biassed in the direction of ‘lumping’, it is because the burden of proof of a divisive speciation event within an THE BUMBLE BEES OF THE KASHMIR HIMALAYA ancestral population is considered to rest with the ‘splitter’. It has been argued that evidence from subtle differences between two individuals is not sufficient. For example, it can be seen that the people from the different parts of Europe, of Asia, or even from different areas within Kashmir, also vary in shape and hair colour, and yet from their breeding behaviour they would all be placed in one biological species. But irrespective of the decisions reached here, as much of the available information as time has allowed is presented in the comments on the variation within Kashmir and in the figures and maps, so that readers with a different concept of the nature of species should be able to make their own decisions regarding this fauna. The description of intraspecific variation Many examples of character variation among samples of a single species can be related to where the samples were collected within the species’ distribution. Often this geographic variation in single characters can be regarded as belonging to a continuum between extreme states, so that defining two or more discrete states would require threshold criteria that must be essentially arbitrary. Furthermore, where comparisons are drawn between two or more characters, these may show geographically incongruent patterns of variation, so that most combinations of states can be found somewhere within the species’ distribution. Therefore, because variation of characters among individuals does not always delimit discrete sub- units within populations (Wilson & Brown, 1953; but see Mallet, 1986; Barton, 1988), previously described nominal taxa of subspecific rank are not attributed here with formal names. This removes any requirement for a general subspecies concept that would otherwise necessitate imposing more assumptions about the genetic structure of populations. The significance of intraspecific variation is better judged separately for the different cases that are presented by each species. In the place of the earlier systems of subspecific and infrasubspecific nomenclature, in this review the most obvious character states are referred to using shorthand descriptions. An example is an ‘unbanded (= no band of black pubescence between the wing bases) yellow (= colour of the pale pubescence, excluding any red hairs on the apical terga) queen (= large female) of Bombus asiaticus.. This terminology may appear less elegant, but it has the advantage that it is easily modified to give a more precise description of the recognised attributes of a particular individual. Intraspecific variation of selected characters is 29 described separately for each species as it is represented in the samples available, together with the frequencies of the various states in these samples. These frequencies may have been biassed at some sites by selective collecting, at least for the later samples (1985, 1986). But for the maps of the geographical distribution of variation in the bees’ colour patterns, it is necessary for clarity of presentation to resort to a more typological approach, referring instead to the predominant local phenotype among the workers (see the figures accompanying the maps). This emphasis is placed on the workers because they are usually the most abundant sex and caste to be seen by collectors. Species’ distributions within Kashmir are shown by spot maps rather than by grid maps, because sampling effort could not be even nearly uniform across the country (Map 4; see the comments on material examined and depositories). NOMENCLATURAL SUMMARY Where possible type specimens have been examined in order to assign specific nomenclature more precisely. References in the text to ‘Art. n’ allude to the numbered articles of the /nternational Code of Zoological Nomenclature (ICZN, 1985). Synonyms are listed here only in cases where I have examined the types or specifically discussed their identity. Names of taxa whose status as part of the same species is in particular doubt are cited as provisional synonyms. Unavailable names are cited in brackets. Nomenclatural details for taxa in the genus group are not repeated in this section. BOMBUS Latreille 1802a avinoviellus (Skorikov, 1914a) callophenax Cockerell, 1917 syn. n. [subtunicatus Richards, 1930, infrasubspecific] himalayanus (Skorikov, 1914a) varius (Skorikov, 1914a) (provisional synonym) marussinus Skorikov, 1910a afghanus Reinig, 1940 syn. n. bohemicus Seidl, 1837 nemorum (Fabricius, 1775) novus (Frison, 1933) comb. n. nepalensis (Tkalct, 1974b) (provisional synonym) branickii (Radoszkowski, 1893) comb. n. eriophoroides (Reinig, 1930) elisabethae (Reinig, 1940) ferganicus (Radoszkowski, 1893) comb. n. indicus (Richards, 1929a) morawitzianus (Popov, 1931) comb. n. redikorzevi (Popov, 1931) 30 skorikovi (Popov, 1927b) comb. n. [mesoxanthus (Richards, 1928c) infrasubspecific] gansuensis (Popov, 1931) trifasciatus Smith, 1852a montivagus Smith, 18785 syn. n. secundus Dalla Torre, 1890 syn. n. ningpoensis Friese, 1909 wilemani Cockerell, 1911 (provisional synonym) albopleuralis Friese, 1916 syn. n. maxwelli Pendlebury, 1923 (provisional synonym) mimeticus Richards, 1931 syn. n. albolateralis Richards, 1931 syn. n. gantokiensis Richards, 1931 syn. n. turneri Richards, 1931 syn. n. insidiosus Richards, 1931 syn. n. geminatus Richards, 1931 syn. n. magrettianus Richards, 1931 syn. n. minshanicus Bischoff, 1936 syn. n. quasibreviceps (Tkalct, 1968b) syn. n. atropygus (Tkalci, 1989) syn. n. haemorrhoidalis Smith, 1852a orientalis Smith, 1854 syn. n. buccinatoris Smith, 1879 syn. n. assamensis Bingham, 1897 syn. n. montivolans Richards, 19295 syn. n. khasianus Richards, 1929b syn. n. semibreviceps (Tkalcu, 19685) syn. n. semicoloricontrarius (Tkalct, 1968b) syn. n. cinnameus (Tkalct, 1989) syn. n. melanurus Lepeletier, 1836 tschitscherini Radoszkowski, 1862 syn. n. difficillimus Skorikov, 1912b (provisional synonym) subdistinctus Richards, 1928b syn. n. griseofasciatus Reinig, 1930 (provisional synonym) maidli Pittioni, 19396 syn. n. personatus Smith, 1879 roborowskyi Morawitz, 1886 syn. n. kashmirensis Friese, 1909 stramineus Friese, 1909 tetrachromus Cockerell, 1909 pulcherrimus (Skorikov, 1914a) meinertzhageni Richards, 1928b [albohirtus Richards, 1930, infrasubspecific] beresovskii (Skorikov, 19335) syn. n. hypnorum (Linnaeus, 1758) bryorum Richards, 1930 syn. n. fletcheri Richards, 1934 syn. n. subtypicus (Skorikov, 1914c) comb. n. leucopygus Morawitz in Fedtschenko, 1875 leucurus Bischoff & Hedicke, 1931 kohistanensis (Tkalcu, 1989) syn. n. Jemniscatus Skorikov, 1912b flavopilosus Friese, 1918 syn. n. peralpinus Richards, 1930 lepidus Skorikov, 1912b genitalis Friese, 1913 syn. n. tetrachromus Friese, 1918 syn. n. yuennanicola Bischoff, 1936 (provisional synonym) hilaris (Tkalct, 1989) syn. n. biroi Vogt, 1911 PAULH. WILLIAMS flavobistriatus Vogt, 1911 syn. n. [flavostriatus Vogt, 1911, infrasubspecific] [flavofasciatus Vogt, 1911, infrasubspecific] nursei Friese, 1918 syn. n. abbotti Cockerell, 1922 agnatus Skorikov, 1933b (provisional synonym) kotzschi Reinig, 1940 (provisional synonym) tunicatus Smith, 1852a gilgitensis Cockerell, 1905 syn. n. simlaensis Friese, 1909 fulvocinctus Friese, 1909 Jucorum (Linnaeus, 1761) cryptarum (Fabricius, 1775) (provisional synonym) modestus Cresson, 1863 moderatus Cresson, 1863 magnus Vogt, 1911 (provisional synonym) jacobsoni Skorikov, 19126 syn. n. lucocryptarum Ball, 1914 (provisional synonym) longipennis Friese, 1918 syn. n. alaiensis Reinig, 1930 syn. n. mongolicus Kriiger, 1954 syn. n. reinigi Tkalcu, 19746 syn. n. asiaticus Morawitz in Fedtschenko, 1875 longiceps Smith, 1878a regeli Morawitz, 1880 miniatocaudatus Vogt, 1911 [fuscocaudatus Vogt, 1911, infrasubspecific] [albocaudatus Vogt, 1911, infrasubspecific] [tenuifasciatus Vogt, 1911, infrasubspecific] falsificus Richards, 1930 syn. n. flavodorsalis (Skorikov, 1933b) oshanini (Skorikov, 1933b) oberti Morawitz, 1883 semenovi Morawitz, 1886 syn. n. duanjiaoris Wang, 1982 syn. n. ladakhensis Richards, 1928b phariensis Richards, 1930 syn. n. variopictus Skorikov, 19335 syn. n. bianchii Skorikov, 1933b reticulatus Bischoff, 1936 syn. n. semenovianus (Skorikov, 1914a) problematicus Bischoff, 1935 keriensis Morawitz, 1886 separandus Vogt, 1909 syn. n. kohli Vogt, 1909 syn. n. postzonatus Vogt, 1909 syn. n. kozlovi Skorikov, 19105 syn. n. incertoides Vogt, 1911 syn. n. tenellus Friese, 1913 syn. n. meridialis (Skorikov, 1914a) syn. n. alpivagus Richards, 1930 syn. n. karakorumensis (Tkalct, 1989) syn. n. simillimus Smith, 18526 grossiventris Friese, 1931 syn. n. oculatus (Frison, 1933) syn. n. tonsus (Skorikov, 1933b) syn. n. haemorrhous Richards, 1934 syn. n. pyrosoma Morawitz, 1890 flavothoracicus Bingham, 1897 (provisional synonym) miniatus Bingham, 1897 (provisional synonym) THE BUMBLE BEES OF THE KASHMIR HIMALAYA [canosocollaris Skorikov, 1912b, infrasubspecific] friseanus Skorikov, 1933a syn. n. formosellus (Frison, 1934) (provisional synonym) hoenei Bischoff, 1936 syn. n. [flavocorbicularis Tkalct, 1961, infrasubspecific] wutaishanensis (Tkalct, 1968a) syn. n. rufofasciatus Smith, 1852b prshewalskyi Morawitz, 1880 rufocinctus Morawitz, 1880 chinensis Dalla Torre, 1890 championi Richards, 1928a syn. n. [rufior Richards, 1928b, infrasubspecific] [intermedius Richards, 1930, infrasubspecific] waterstoni Richards, 1934 TAXONOMY OF THE KASHMIR BOMBINI BOMBINI Michener Bombini Michener, 1944: 290. Type genus: Bombus Latreille, 1802a: 437. In this review, formal descriptions of bumble bees are not repeated because the literature already contains many descriptions of particular individuals. Instead I describe just a few characters that may be useful to distinguish species. Tentative inter- pretations of some of these character-state patterns are included for the continuing discussion of relationships among species. In the keys, those characters that have been found more reliably diagnostic have generally been placed nearer the beginning of each couplet. Characters of the genitalia are used frequently in the key to males (see Figs 44 & 45), so male specimens should be prepared for examination by extraction of their genital capsules. Colour patterns of the pubescence are used in the final stages of the keys for the determination of species within a species-group, where the range of varia- tion does not overlap among these species within Kashmir. But after a reference collection has been prepared for a particular locality using the keys, it may often be possible to identify the majority of further specimens accurately by direct comparison, using only selected details of their colour patterns. Diagrams of the colour patterns are provided here to illustrate some of the variation (Figs 232-439). These diagrams should be coloured by hand to render them more immediately intelligible (see Fig. 258 for a colour key). Inevitably these diagrams are a compromise between portraying the general appearance of an insect and accurately showing the position of every coloured hair. For 31 instance, the presence of a minority of black hairs among pale pubescence could not be shown without reducing the general resemblance of the diagram to the insect. Consequently these diagrams provide only a guide that should not be used without the keys. Records of food-plants are included from notes made in the field during 1980, 1985 and 1986. These plant species were identified by reference to Polunin & Stainton (1984). The lists are by no means exhaustive, but they do include at least some of the preferred species. Key to species (females: queens and workers) 1 Outer surface of hind tibia flat or partially concave, without long hairs on posterior part of lower or distal half of outer surface (corbicula), but with a comb of stout spines (rastellum) along inner distal margin (Figs 225 & 226); gastral sternum VI without ventro- lateral keels; clypeus strongly protuberant, with lateral margins curved back towards occiput (Fig. ES) St. eee eres Mes Sone eet 2) — Outer surface of hind tibia convex, with moderate to long hairs throughout, but without a comb of stout spines along inner distal margin; gastral sternum VI with a pair of ventro-lateral keels (Figs 206-211); clypeus nearly uniformly flat, only apico-lateral corners curved back strongly towards occiput ..... 5 2 Outer (corbicular) surface of hind tibia usually smooth and shining or only weakly sculptured (retic- ulate coriaceous), without any long stout hairs aris- ing from posterior part of outer surface below upper or proximal quarter (Fig. 226); proximoposterior corner of hind basitarsus strongly and acutely pro- duced, projection longer than its own basal breadth (Figs 41 & 42); labrum with basal transverse depres- sion extending apically as a deep median furrow between pronounced lateral tubercles, displacing ridge between them to form a lamella that overhangs apical margin (Figs 25 & 26) ................... 10 — Outer (corbicular) surface of hind tibia coarsely sculptured (imbricate), appearing very rough, with widely-spaced long stout hairs or bristles arising from near middle of outer surface throughout upper or proximal half (Fig. 225); proximo-posterior corner of hind basitarsus rounded, projection no longer than its own basal breadth (Fig. 43); labrum with a narrow transverse basal depression, leaving a straight transverse ridge joining weak lateral tubercles, so that there is no median furrow and no apical lamella (Rigs: 2392205222) er omen faethe See choc 3 3 (Subgenus Mendacibombus.) Ridge between labral tubercles is, at its mid point, narrower than basal depression (Figs 23 & 220), with few punctures and shining; wings clouded with brown (moderately imkuscated) Wan teaser eee eee cea avinoviellus — Ridge between labral tubercles is, at its mid point, broader than basal depression, with few or many 32 punctures (Figs 221 & 222); wings nearly clear (subhyaline) 24555 Sjeeseee ee ere eee 4 Ridge between labral tubercles subsiding towards middle, where there are often many punctures (Fig. 221); pubescence of thoracic dorsum grey-white or lemon-yellow, but with a broad black band between wings (Figs 242-245, 247, 248, 250, 251) SE Re or ae Me et PR Ree) Sa EC, core himalayanus Ridge between labral tubercles convex for its entire length, with few punctures (Fig. 222); pubescence of thoracic dorsum lemon-yellow, with only a very few black hairs (Figs 254, 255, 257)......... marussinus (Subgenus Psithyrus.) Lateral keels of sternum VI projecting well beyond tergum VI from dorsal aspect, crests of keels sharply acute and blade-like in section, distinctly angled about midway along their length (Figs 207 & 208); pubescence of tergum V usually mostly orange-red, without yellow or white hairs, only occasionally entirely black............ 6 Lateral keels of sternum VI not clearly projecting beyond tergum VI from dorsal aspect, crests of keels broadly rounded, both in section and along their entire length (Figs 206, 209-211); pubescence of tergum V black, yellow or white, but without orange- red hairs Crests of lateral keels of sternum VI, beyond project- ing angle of mid-point, with a strongly concave margin (Fig. 208); sternum II with transverse ridge sharply- defined and nearly straight; labral furrow narrow, about a fifth of total basal breadth of labrum; clypeus with many large punctures spaced more closely than their own widths, except in a well-defined narrowly- unpunctured mid-apical area; pubescence of thoracic dorsum without black hairs, except for a few hairs adjacent to tegulae, tergum IT predominantly black (lips:26326552660) = ease eee branickii Crests of lateral keels of sternum VI, beyond project- ing angle of mid-point, almost straight (Fig. 207); sternum II with transverse ridge rounded and curved unevenly towards anterior margin in middle; labral furrow wide, about a third of total basal breadth of labrum (Fig. 24); most of clypeus with scattered large punctures spaced more widely than their own widths; pubescence of thoracic dorsum usually with a poorly- defined patch of black hairs between wings, tergum II yellow (Figs 259 & 261) Lateral keels of sternum VI strongly swollen almost to their apices, where they are separated by a groove that is narrower than their own breadths (Figs 209 & 210); basal depression of labrum deep and clearly delimited by an absence of punctures, apex of lamella pointed Lateral keels of sternum VI declining from near their mid-points so that strongly swollen parts are separated by more than their own breadths (Figs 206 & 211); basal depression of labrum shallow and with punctures, apex of lamella broadly rounded Tergum VI with many large punctures, but shining; 10 1 — PAULH. WILLIAMS basal keel of mandible weak or absent; thoracic dorsum without black pubescence (Fig. 267) “oC ivArmir aes Vag Aa e Se oe ee cee ferganicus Tergum VI with only fine punctures, strongly shining; basal keel of mandible well developed but interrupted before margin of mandible; pubescence of thoracic dorsum black with an admixture of yellow hairs as an anterior band (Fig. 269) ........... morawitzianus Lateral keels of sternum VI small and converging to a point well before apex, which is down-curved as a narrow spinose hook that projects strongly beyond apex of tergum VI (Fig. 211); basal keel of mandible present but interrupted before reaching margin; labral tubercles and lamella weakly developed and rounded in section; oculo-malar distance nearly equal to basal breadth of mandible; pubescence of thoracic dorsum and tergum I predominantly yellow, at most with a narrow band of black hairs between wings (Biss 271(& 275) see skorikovi [From material from Britain.] Lateral keels of sternum VI strongly swollen and converging to a point only just before apex of sternum, which is broadly triangular and not down-curved or particularly narrowed and does not project beyond apex of tergum VI (Fig. 206); basal keel of mandible weak or absent; labral tubercles pointed and lamella strongly swollen; oculo-malar distance less than two-thirds of basal breadth of mandible; pubescence of thoracic dorsum and tergum I black, with only a broad yellow band anteriorly onthorax............. bohemicus Apex of mandible broadly rounded, with a basal tooth, one pre-basal tooth and often an apical tooth (e.g. Fig. 34); longest of erect hairs near anterior margin of outer surface of hind basitarsus shorter than greatest breadth of basitarsus ............. 11 Apex of mandible with six teeth, though these are subject to wear (Fig. 33); longest of erect hairs near anterior margin of outer surface of hind basitarsus just longer than greatest breadth of basitarsus. (Subgenus Alpigenobombus, oculo-malar distance about half of basal breadth of mandible; pubescence long, with a broad black band between wings, terga III-V always with some reddish hair, remainder grey- white, yellow or black, Figs 295, 296, 298, 299, 301- 30373052506) 3085309) eee ee eee kashmirensis Hind basitarsus with dense pubescence of proximal margin (auricle) continuing onto outer surface of proximo-posterior projection as just a few sparse hairs (Fig. 229); oculo-malar distance shorter or longer than one and a fifth times basal breadth of mandible, but anterior part of malar area smooth, at most the larger individuals have only a narrow diagonal band with a few very small punctures... . 12 Hind basitarsus with dense pubescence of proximal margin (auricle) continuing onto outer surface of proximo-posterior projection as a dense long brush (Figs 227 & 228); oculo-malar distance more than one and a fifth times longer than basal breadth of mandible, and for larger individuals at least, anterior THE BUMBLE BEES OF THE KASHMIR HIMALAYA 12 13 14 15 part of malar area uneven with many small to moderately-large but widely-spaced punctures almost RERROOEL NOUR erates eee, Tee exc! aynins aie) ye, sate dan chewte' 3 23 Disto-posterior corner of mid basitarsus usually broadly rounded, never produced as a spine (Fig. 40); oculo-malar distance shorter or longer than one and a fifth times basal breadth of mandible . . 13 Disto-posterior corner of mid basitarsus spinose (Fig. 39), or for small workers, at least acutely produced; oculo-malar distance at least one and a fifth times longer than basal breadth of mandible Ocello-ocular area with unpunctured and shining areas large or small, but area anterior to ocelli unpunctured for less than breadth of an ocellus; labral lamella narrow, only about a third of basal breadth of labrum; clypeus with many large punctures scattered on flattened central area; pubescence of thoracic dorsum with or without much pale AME eft: Marts ctl css hacia pcte ete ayes wees wees 14 Ocello-ocular area with unpunctured and shining areas very large and including most of area anterior to ocelli for a distance of more than breadth of an ocellus (Fig. 215); labral lamella broad, more than half of basal breadth of labrum; clypeus with almost no large punctures in flattened central area; pube- scence of thoracic dorsum entirely black. (Subgenus Orientalibombus, large individuals with very dark wings; pubescence short and very even, terga I-II bright yellow, III-V orange-red, Figs 279 & DOU) Preeiictse cis crt seve dale Dees haemorrhoidalis Band of punctures along eye margin in oculo- ocellar area opposite lateral ocellus with few small punctures, large punctures separated by more than their own widths, whole band occupying about a third of distance between lateral ocellus and eye (Figs 216-219); tergum VI subapically nearly flat, or at most with a poorly-defined apical swelling or a narrow median ridge, but without a discrete subapical rounded boss; wings clear (hyaline), or at most only weakly clouded with brown (weakly AUUSCAted) Pe. aah wits was Beak S Saya dea dene oo 18 Band of punctures along eye margin in oculo- ocellar area opposite lateral ocellus occupying nearly half of distance between lateral ocellus and eye, with many small and large punctures all separated by less than widths of large punctures, or if punctures are mostly large and more widely spaced, then tergum VI has a discrete subapical rounded boss (e.g. Fig. 231, often weak or absent for small workers); wings clear (hyaline) to nearly black (strongly infuscated)................... 15 Oculo-malar distance about equal to or distinctly longer than basal breadth of mandible; tergum VI with a discrete subapical rounded boss (e.g. Fig. 231, often weak or absent for small workers); mandible with only a shallow notch (incisura) and tooth apically (Fig. 34) ...................4.. 24 Oculo-malar distance less than two-thirds of basal 16 17 18 19 20 33 breadth of mandible; tergum VI subapically nearly flat, without a rounded boss; mandible with a deep notch (incisura) and tooth apically Clypeus without large punctures in central area; antennal segment 4 longer than broad and more than three-quarters of length of segment 5; pube- scence of gaster white, cream, yellow, brown or black, but without orange-red ................ 17 Clypeus with many large punctures throughout; antennal segment 4 just shorter than broad and less than three-quarters of length of segment 5 (Fig. 17); pubescence of gaster with some orange-red. (Sub- genus Diversobombus, pubescence of thoracic dorsum black, terga I-II bright yellow, III-V predominantly orange-red, Figs 275, 276, 278) 6b JW ie AOS OS SE HURCIC Bey ee ORAS trifasciatus (Subgenus Subterraneobombus.) Central area of clypeus almost without punctures, at most with only a few very widely-spaced and fine punctures; pube- scence of head, lower sides of thorax, legs, all sterna and terga III-V black (Figs 282-284, 286-289) ra Seas eas aR aR Rea s wlare at melanurus Central area of clypeus with many fine punctures, often separated by a distance of only about twice their own diameter; pubescence of thorax almost entirely cream-yellow, except for a black band between wings, and with many pale hairs on head around bases of antennae, on coxae and hind femora, on gastral sterna and as posterior fringes of terga III-V (Figs 291 & 292) .......... personatus (Subgenus Pyrobombus.) Pubescence of thoracic dorsum grey-white or yellow, but not brown, often with many black hairs, terga [V-V orange-red . . .19 Pubescence of thoracic dorsum uniformly yellow- brown, only occasionally faded to grey, terga IV-V white (Figs 311, 312, 314,315) ......... hypnorum Vertex, postero-laterally to ocello-ocular area, shining, with few fine punctures between large punctures (Figs 216 & 217), punctures becoming more numerous and more closely spaced towards posterior margin of vertex (occiput); apex of tergum VI rounded, subacute (Fig. 214) .............. 20 Vertex, postero-laterally to ocello-ocular area, with many close, fine punctures between large punctures (Figs 218 & 219), punctures uniformly spaced and numerous towards posterior margin of vertex (occiput); apex of tergum VI broadly truncate (Figs 212 & 213) Labral lamella narrow, less than one-third of breadth of labrum; hind basitarsus with posterior margin nearly straight in distal half (Fig. 42), stout bristles on outer surface arising from sockets that are scarcely raised from surface; pubescence long, longest hairs of anterior margin of hind tibia much longer than its distal breadth, longest hairs of face at least three-quarters as long as segment 1 (scape) of antenna; pale pubescence of thoracic dorsum either yellow, grey-white or absent, but if it is yellow then 34 21 22 23 24 there are no black hairs intermixed (Figs 335, 336, 338, 339, 341, 342, 344-346) ............... biroi Labral lamella broad, more than one-third of breadth of labrum; hind basitarsus with posterior margin usually distinctly concave just proximal to disto-posterior corner (Fig. 41), stout bristles on outer surface arising from sockets that are strongly raised from surface on their proximal sides; pube- scence short, longest hairs of anterior margin of hind tibia only as long as its distal breadth, longest hairs of face only two-thirds as long as segment 1 (scape) of antenna; pale pubescence of thoracic dorsum yellow, with black hairs between wing bases (Figs 317, 319, 320, 322,323) ......... subtypicus Thoracic scutum mid-dorsally with an area as large as tegula around posterior end of longitudinal median groove smooth with few or no punctures (Fig. 224); oculo-malar distance shorter than basal breadth of mandible; hairs of ventral parts of thorax and gaster predominantly grey-white, top and front of head often with many pale hairs intermixed Ppa Re Nan PEERS Cate eat AONE eve SHOES ee nN lepidus Thoracic scutum mid-dorsally with punctures and sculpturing almost throughout area around posterior end of longitudinal median groove (Fig. 223); oculo- malar distance equal to or just longer than basal breadth of mandible (Fig. 15); hairs of ventral parts of thorax and gaster predominantly black, top and front of head without pale hairs ....... Jemniscatus (Subgenus Bombus.) Pubescence of thoracic dorsum grey-white, with a more or less well-developed black band between wings, tergum V red (Figs 347, 348, 350 435143537354) s eect cheese ae tunicatus Pubescence of thoracic dorsum black, with a broad, lemon-yellow or cream anterior band, sometimes with a narrow posterior band, and tergum V nearly white (Figs 356, 357, 359-361, 363, 364, 366, 368, BOO) oF tee cate ead swage cetera weet veneene ee Jucorum (Subgenus Sibiricobombus.) Oculo-malar area with many widely-spaced moderate to large punctures; antennal segment 4 less than half of length of segment 3 (Fig. 19); tergum VI coarsely sculptured but without amedian groove ........... asiaticus Oculo-malar area with many widely-spaced small punctures almost throughout, surface uneven; antennal segment 4 more than half of length of segment 3 (Fig. 20); tergum VI coarsely sculptured, with a well-defined long narrow median groove (FIge230) isos eee a ee eee oberti (Subgenus Melanobombus.) Ocello-ocular areas laterally shining between few large punctures along eye margin opposite ocelli, these punctures separated by distances greater than their own widths; oculo- malar distance longer than, or about same length as, basal breadth of mandible; long hairs of head interspersed with a dense cover of short branched hairs, which are especially evident between ocelli and occiput; pubescence of tergum V usually with some white hairs, or if these are absent, then wings are nearly black (strongly infuscated) .......... 25 25 26 27 28 PAULH. WILLIAMS Ocello-ocular areas laterally matt, with a dense band of large punctures along eye margin opposite ocelli, many of these punctures separated by distances less than their own widths; oculo-malar distance just shorter than basal breadth of mandible; only a few short branched hairs interspersed with long hairs on head; wings nearly clear (subhyaline) and pubescence of tergum V usually uniformly red, or if white hairs are present then there are very few black hairs intermixed with pale pubescence on anterior Pant ofthoracte COTSunN yey sn tebe eee 27 (rufofasciatus-group.) Outer (corbicular) surface of hind tibia with only a few short branched hairs, which are confined to margins near disto-posterior corner; pubescence of thoracic dorsum yellow or grey-white, but with a black band between wings; tergum V usually with some white hairs, at least apically; wings nearly clear or lightly clouded with brown (subhyaline to weakly infuscated) ....... 26 Outer (corbicular) surface of hind tibia with many short to moderately long, branched hairs, at least in proximal half (few or absent for small workers and subject to loss for older females, Fig. 226); pube- scence of thorax either entirely black or almost entirely cream-white dorsally and without a black band between wings, tergum V red without any white hairs (Figs 417, 418, 420); wings nearly black (strongly infuscated)................. simillimus Boss on tergum VI nearly circular and evenly convex, only narrowly pointed adjacent to apex of tergum; pubescence of head entirely black, thoracic dorsum and tergum I with white pubescence (Figs 431,432, 434, 435, 437, 438)........ rufofasciatus Boss on tergum VI nearly triangular, for some queens with a weak median groove, for workers it may be weakly ridged; short hairs of head yellow, thoracic dorsum and tergum I with yellow pube- scence (Figs 422, 423, 425, 426, 428, 429) wick Boatingnend. svahniisie, Geneon eee a eR ee pyrosoma (lapidarius-group.) Ocello-ocular area with a continuous broad band of fine punctures along eye margin; pubescence of thoracic dorsum uniformly lemon-yellow, without black hairs (Figs 397, 398, AN. AQ) ie a2 cea ue Sine ans ange era semenovianus Ocello-ocular area with band of fine punctures along eye margin narrowly interrupted opposite ocelli, so that there is a small shining area with only large punctures; pubescence of thoracic dorsum pale yellow, cream, or grey-white, but with a broad black band between wings ..................- 28 Apex of tergum VI usually broadly, but shallowly, notched; pubescence of tergum III predominantly black, often with a posterior fringe of pale hairs, and of tergum V predominantly red (Figs 403-409, ALI—AN5)). casei hey eRe tak eee eee keriensis Apex of tergum VI nearly straight; pubescence of tergum III predominantly orange-red and of tergum V white, although these hairs have very dark bases (Figs 394 & 396) .......... Jadakhensis THE BUMBLE BEES OF THE KASHMIR HIMALAYA Key to species (males) 1 Volsella always strongly sclerotised, inner corner usually much closer to apex than to base (to form an interio-apical corner), often bearing two inwardly- directed hooks or a single combined inwardly- directed process (Figs 136-165); gonostylus usually with a distinct interio-basal process or shelf, but not associated with long hairs (Figs 176-205); head of penis valve much modified, often curved strongly in towards mid-line of body, as a sickle-shaped hook (Figs 61-71, 76-85), otherwise curved outwards for some individuals (Figs 72-75), but if it is nearly straight from dorsal aspect (Figs 56-60), then volsella has pronounced inner hooks (Figs 136-140)...... 10 Volsella weakly or strongly sclerotised, inner corner located near mid-point of its length, without any inwardly-directed hooks, so that volsella is usually nearly triangular in the distal section (Figs 126-135); gonostylus without an interio-basal process (Figs 166-169), or if present (Figs 170-175), then it is associated with many long branched hairs; head of penis valve nearly straight from dorsal aspect (Figs AGS =SIN) Ge Soa OR ORC EC OE CREE Cee ee een 2 Volsella weakly sclerotised, yellowish in colour; gonostylus with a pronounced interio-basal process (Figs 170-175), associated with many long branched hairs; head of penis valve, as defined by an outer lateral ridge, less than a quarter of total length, not strongly curved but shaped like a slender arrowhead from lateral aspect (Figs 90-95) ................. 5 Volsella strongly sclerotised, dark brown in colour; gonostylus without an interio-basal process or hairs (Figs 166-169); head of penis valve, as defined by an outer lateral toothed ridge, accounting for nearly half of total length, ventrally curved and sabre- shaped from lateral aspect (Figs 86-89) .......... 3 (Subgenus Mendacibombus.) Penis valve strongly laterally flattened (Figs 46, 47, 86, 87); gonostylus dorso-ventrally flattened (Figs 166 & 167); mandible with basal tooth strongly developed (Fig. 35); wings clouded with brown (moderately infuscated) MRA Set st seicke Polar hive a) niet tN anwar avinoviellus Penis valve nearly tubular (Figs 48, 49, 88, 89); gonostylus stout and not strongly flattened (Figs 168 & 169); mandible without a distinct basal tooth; wings nearly clear (weakly infuscated or subhyaline) Volsella with interio-apical margin weakly concave (Fig. 128); gonostylus from outer dorso-lateral aspect with a distinct, though broadly-rounded, exterio-apical corner (Fig. 168); pubescence of thoracic dorsum anteriorly and of terga I-II grey- white and/or lemon-yellow, with a broad black band between wings (Figs 246, 249, 252, 253) FON eT RTA COO 6 OIC ome oie himalayanus Volsella with interio-apical margin strongly concave, so that apex resembles a curved finger (Fig. 129); gonostylus from outer dorso-lateral aspect with exterio-apical margin evenly rounded, without any 35 trace of an exterio-apical corner (Fig. 169); pube- scence of thoracic dorsum lemon-yellow, with few blackihairs (Rig256) 79" saueteisees 20: marussinus (Subgenus Psithyrus.) Volsella in its distal section, beyond inner corner, greatly narrowed, almost parallel-sided and finger-shaped (Fig. 135); gono- stylus with only afew shorthairs ......... skorikovi Volsella distally broader, in the form of a broad nearly triangular plate (Figs 130, 131, 133, 134), or if distal section is narrowed then inner corner is strongly produced (Fig. 132); gonostylus usually with many long hairs around interio-basal process. ..... 6 Ventro-basal angle of penis valve strongly and broadly produced ventrally and outwardly, so as to be clearly visible from dorsal aspect (Figs 51-54, 91- 94); inner corner of volsella well defined, strongly produced for some individuals (Figs 131-134); antennal segment 3 distinctly shorter than segment 5 (e.g. Fig. 22); pubescence of scutellum usually almost entirely yellow or grey, terga [V-V with red or yellow, or if terga IV-V are nearly white (Fig. 270) then tergum II has a narrow apical band of pale LA ETT RSE Tea Ro ae CR rar Ia ire an 7 Ventro-basal angle of penis valve much reduced and not visible from dorsal aspect (Figs 50 & 90); inner corner of volsella weak, almost unmarked for some individuals (Fig. 130); antennal segment 3 just longer than segment 5 (Fig. 21); pubescence of scutellum black, with only a narrow band of yellow hairs, apical margin of tergum II without a band of pale hair, terga IV-V nearly entirely white (Fig. 258) ... bohemicus Inner margin of distal section of volsella irregular but not predominantly concave (Figs 133 & 134); antennae very long, reaching back nearly to gaster; pubescence of terga V-VII black, with lateral patches OMGOMO WTO ECS sa a apace facaie ap cvaday ara alten cut secant 8 Inner margin of distal section of volsella predom- inantly concave, with exception of a weak subapical process (Figs 131 & 132); antennae short, reaching back only to tegulae; pubescence of terga V-VII OLAN PETER .Gracae eee a ee ae eR aie 9 Distal section of volsella as long as about one and a half times its maximum breadth from ventral aspect (Fig. 133); gonostylus broadly triangular (Fig. 173); pubescence of thoracic dorsum predominantly brownish-yellow, with only a few black hairs above tepulae(bicn268) een ae cctcean esate oF ferganicus Distal section of volsella only just longer than its maximum breadth from ventral aspect (Fig. 134); gonostylus reduced to a narrow transverse band at apex of gonocoxite, but retaining a well-developed interio-basal process (Fig. 174); pubescence of thoracic dorsum light yellow with a broad black band between Wines (HIG: 270). sects suri cee morawitzianus Distal section of volsella almost twice as long as its maximum breadth from ventral aspect (Fig. 132); pubescence of thoracic dorsum yellow, with black hairs intermixed between wings (Fig. 264)...... branickii 36 11 12 Distal section of volsella about as long as its maximum breadth from ventral aspect (Fig. 131); pubescence of thoracic dorsum grey-white, with a poorly-defined black band between wings (Figs 260 & 262) . .novus Penis valve from dorsal aspect turned inwards before apex and dorso- ventrally flattened in the form of a sickle (Figs 61-71, 76-85) ............ 11 Penis valve from dorsal aspect apically nearly straight, turned outwards, or at least not strongly inwardly directed and flattened as a sickle-shape (Figs:$6=6057/2=75))... J aoatSs sonmavewe tings he 14 Ventro-basal angle of penis valve much reduced, shown only as a weak curve, or absent (Figs 116- 125); spatha, beyond basal fusion with penis valves, more than three times longer than its breadth midway along this length, sides nearly parallel in proximal al figs iced cece cyencveusts w ae cseuevssts ae aie 12 Ventro-basal angle of penis valve produced as a basally-directed hook (Figs 101-111); spatha, beyond basal fusion with penis valves, less than three times longer than its breadth midway along this length, sides weakly diverging in proximal half (B. kashmirensis), or distinctly converging towards apex throughout theirlength ................. 13 Gonostylus just longer than its greatest breadth, excluding interio-basal process, which is pronounced as a broad shelf, and broadening apically (Figs 196— 199); shaft of penis valve more than three-quarters as broad from lateral aspect as greatest breadth of gonostylus, excluding its interio-basal process (Figs 116-119); head of penis valve with a broad outer ridge (Figs 76-79, 116-119) .................. 23 Gonostylus shorter than its greatest breadth, excluding interio- basal process, which is reduced to a small rounded projection or a right-angled corner, and narrowing apically (Figs 200-205); shaft of penis valve usually strongly narrowed from lateral aspect, its breadth less than two-thirds as broad as greatest breadth of gonostylus, excluding interio- basal process (Figs 120-125); head of penis valve with only a narrow outer ridge (Figs 80-85, 120— 125) Deze ony Serra eto evecesisl ane ota eae 24 Apex of mandible acutely pointed, with a basal tooth and a pre-basal tooth (Fig. 37); gonostylus with a pronounced interio-basal process (Figs 181— 183); volsella, from dorsal aspect, projecting beyond gonostylus apically by about its own apical breadth. (Subgenus Alpigenobombus, pubescence with a broad black band between wings, head predom- inantly black, terga III-VI with some reddish hair, remainder variously banded with grey-white, yellow or black, Figs 297, 300, 304, 307, 310) Weal phate GREET rad open b-t aoe nett a Bre kashmirensis Apex of mandible more or less pointed, otherwise with only a basal tooth; gonostylus without a strongly projecting interio-basal process (Figs 184— 191); volsella, from dorsal aspect, projecting beyond gonostylus apically by only about half of its own apical breadth {i455 4c ei 18 14 15 17 PAULH. WILLIAMS Penis valves strongly broadened in dorso-ventral plane so as to form two halves of a tube, ends splayed outwards as a broad funnel (Figs 72-75, 112-115); antennae short, reaching back only as far aSitepulae ey nae elt ee ere oe oct eee eee 22 Penis valves dorso-ventrally narrow, at least in apical third, which is further narrowed and more or less ventrally-directed (Figs 56-60, 96-100); anten- nae long, reaching back beyond tegulae ........ 15 Volsella apically not narrowed, with apical corner forming a right angle, not produced, a projection from inner corner terminating in a single small serrated process, which is directed apically, without any trace of a separate recurved basally-directed hook (Figs 138-140); ventro-basal angle of penis- valve strongly produced ventro-laterally, forming a flattened paddle-shape, transverse to main axis of penis valve (Figs 98-100); pubescence of thoracic dorsum cream, yellow or brown, some individuals with black bands between wings, terga III-VII without oranpe-Ted =<). 6s ec. ity eee 17 Volsella apically much narrowed, or with apical corner narrowly produced, inner corner terminating in a pair of serrated hooks, one directed apically, the other, which may be very much reduced for some individuals, directed basally (Figs 136 & 137); ventro-basal angle of penis valve produced ventrally, but hardly laterally, as a single tooth (Figs 96 & 97); pubescence of thoracic dorsum black, terga ITI-VII predominantly orange-red ................... 16 Gonostylus with interio-basal process strongly produced, with both an apically-directed hook and a more ventrally-directed hook (Fig. 176); volsella with inner hooks very close to apical corner, the large apically-directed hook nearly straight and spinose, the basally-directed hook strongly recurved and broad with many teeth, both hooks projecting beyond gonostylus from dorsal aspect, part of volsella immediately proximal to inner hooks much narrowed (Fig. 136); head of penis valve twice as broad as shaft, from dorsal aspect, with pronounced teeth along outer lateral ridge (Fig. 56) (Subgenus DiversODOMmbDUS) ase eee eee trifasciatus Gonostylus without an interio-basal process (Fig. 177); volsella with apical corner narrowly produced distal to inner hooks and curled inwardly, inner hooks close to inner corner and obscured from dorsal aspect by gonostylus, part of volsella immediately proximal to inner hooks broad (Fig. 137); head of penis valve scarcely broader than shaft, from dorsal aspect, without distinct teeth along lateral ridge (Fig. 57) (Subgenus Orientalibombus) sue aes lay ale. Sa Ges et ae vee haemorrhoidalis (Subgenus Subterraneobombus.) Curved head of penis valve, from dorsal aspect, longer than broad (Figs 58 & 59), paddle-like ventro-basal angle of penis valve clearly tridentate, with a spinosely- produced dorso-lateral tooth, which is longer than tooth that some individuals have on exterio-basal corner of penis valve head (Figs 98 & 99); tip of THE BUMBLE BEES OF THE KASHMIR HIMALAYA 19 20 aN interio-basal process of gonostylus not projecting apically as far as apex of gonostylus (Figs 178 & 179); pubescence of head around antennal bases, of ventral parts of thorax and of terga III-VII, predominantly black (Figs 285 & 290) .. melanurus Curved head of penis valve, from dorsal aspect, as broad as long (Fig. 60), paddle-like ventro-basal angle of penis valve almost rounded, with only a weak dorso-lateral tooth, which is shorter than spinose tooth of exterio-basal corner of penis valve head (Fig. 100); tip of interio-basal process of gonostylus projecting apically as far as apex of gonostylus (Fig. 180); pubescence of head around antennal bases, of ventral parts of thorax and of posterior fringes of terga III-VII predominantly PASI SS 295 GC D94) rate w asasine eine aie personatus (Subgenus Pyrobombus.) Inner margin of gono- stylus weakly convex basally, longitudinal sub- marginal groove arising close to apex of gonocoxite, separated from it by a distance usually only about equal to width of groove (Figs 185-191); pubescence of terga V-VII orange-red ................... 19 Inner margin of gonostylus with a deep concavity or notch basally, separating longitudinal submarginal groove from apex of gonocoxite by a distance nearly twice width of groove (Fig. 184); pubescence of terga V-VII white (Figs 313 & 316) ..... hypnorum Antennae short, reaching back only as far as tegulae; head of penis valve with recurved hook not distinctly narrower than adjacent penis valve shaft, from dorsal aspect, and apically broadly rounded (Figs 67-71); pubescence long, longest hairs arising from outer surface of proximal part of mid basitarsus nearly one and a half times breadth of mid basitarsus Antennae moderately long, reaching back to scutel- lum; head of penis valve with recurved hook distinctly narrower than adjacent penis valve shaft, from dorsal aspect, and apically acute (Figs 65 & 66); pubescence short, longest hairs arising from outer surface of proximal part of mid basitarsus shorter than breadth of mid basitarsus . subtypicus Penis valve head with recurved hook no greater in breadth than adjacent penis valve shaft, from dorsal aspect (Figs 67-69), and ventro-basal angle of penis valve close to mid-point between penis valve head and its base (Figs 107-109); pale pubescence with grey-white hair, or if this is lacking (i.e. pubescence predominantly yellow), then tergum III is yellow or orange-red with very few black hairs........... 21 Penis valve head with recurved hook more than one and a half times breadth of adjacent penis valve shaft, from dorsal aspect (Figs 70 & 71), and ventro- basal angle of penis valve at less than a third of length of penis valve shaft from its base to its head (Figs 110 & 111); pale pubescence without grey- white hair, tergum III predominantly black (Figs 331; SAN N343) ise Someries aes bale eae ea wee biroi Pubescence of thorax and tergum I predominantly 22 23 24 25 37 yellow, without grey-white hairs (Figs 329 & 332); volsella with interio-apical process narrow and separated from apex by a distance nearly equal to its own breadth (Figs 148 & 149)............ lepidus [From material from Uttar Pradesh, India.] Pubescence of thoracic dorsum and tergum I grey- white, without yellow hairs, but with a broad black band between wing bases; volsella with interio- apical process arising directly from apex (Fig. LATIONS. A Ss See tase oe adeno oon s lemniscatus (Subgenus Bombus.) Pubescence of thoracic dorsum grey-white, often with a more or less well-developed black band between wings, terga V-VI red (Figs FAQESS ZA SIME ae tee crite e aes see es ees tunicatus Pubescence of thoracic dorsum with lemon-yellow anteriorly, remainder with variable degrees of replacement of black by yellow, terga V-VI nearly white (Figs 358, 362, 365, 367,370) ...... Jucorum (Subgenus Sibiricobombus.) Eyes greatly enlarged, so that separation from lateral ocelli is by less than an ocellar diameter; antennae very long, reaching almost to gaster; volsella narrowed apically, interio- apical process narrow and elongate but exterio-apical corner scarcely acute (Figs 156-158); recurved hook of penis valve head broad, tapering to tip (Figs 16-78); pubescence/short .............- asiaticus [From material from the Pamir, U.S.S.R.] Eyes similar in relative size to those of females, not enlarged, and separated from lateral ocelli by more than two ocellar diameters; antennae long, but reaching only just past tegulae; volsella broadened subapically, interio-apical process very short and broad but exterio-apical corner strongly produced as a broad tooth (Fig. 159); recurved hook of penis valve head narrowly constricted (Fig. 79); pube- SCENCEION De re tenets Fates akc aa tas oberti (Subgenus Melanobombus.) Gonostylus with apical margin concave, leaving an exterio-apical corner and an interio-apical process, latter often with two apical corners or bidentate (Figs 203-205); volsella extending beyond gonostylus apically by at least nearly twice its own breadth at the point where it emerges from beneath gonostylus from dorsal EY 0c C ROE MOLI Oe On oe CIE Ce oy ere Te 25 Gonostylus with apical margin broadly convex (Figs 200-202); volsella extending beyond gonostylus by about its own breadth at the point where it emerges from beneath gonostylus from dorsal aspect ....27 (rufofasciatus-group.) Gonostylus about a third as long in middle as broad, excluding interio-basal process, and reduced to a transverse band, with interio-apical corner extended as a spinose or bilobed process (Figs 204 & 205); pubescence of tergum II extensively lemon-yellow ........... 26 Gonostylus more than half as long in middle as broad, excluding exterio-basal process, nearly rectangular, with interio-apical corner not strongly produced (Fig. 203); pubescence of tergum II either 38 nearly white or chocolate brown, but not yellow (Figsl9 G42) (se ees Apes ene ee simillimus 26 Eyes distinctly enlarged relative to those of females, separated from lateral ocelli by one ocellar diameter; interio-basal process of gonostylus with apex forming nearly a right angle, interio-apical process with sides diverging towards apex, which is broadly bilobed (Fig. 205); pale pubescence of thorax and tergum I grey-white (Figs 433, 436, 439) PSOE: o Stic D Maas ema UE wate rufofasciatus — Eyes not enlarged relative to those of females, separated from lateral ocelli by two ocellar diameters; interio-basal process of gonostylus with apex distinctly acute, interio-apical process with sides often converging towards apex (Fig. 204), which may be spinose even from outer lateral aspect; pale pubescence of thorax and tergum I yellow (Figs 424, 427, 430) ............ pyrosoma 27 (lapidarius-group.) Eyes slightly enlarged relative to those of females, separated from lateral ocelli by less than two ocellar diameters; pubescence of thoracic dorsum uniformly lemon-yellow, without black hair (Figs 399 & 402) ......... semenovianus — Eyes not enlarged relative to those of females, separated from lateral ocelli by about two and a half ocellar diameters; pubescence of thoracic dorsum pale yellow with a black band between wings . . . .28 28 Free, recurved hook of penis valve head about two- thirds as long as more apical part of penis valve head (Fig. 82), penis valve shaft narrower at its mid-point from lateral aspect (Fig. 122) than spatha is at its mid-point from dorsal aspect; pubescence of tergum III black, at most with only a narrow apical and lateral fringe of pale hairs, terga ITV-VII orange- red (Figs 410 & 416)................... keriensis — Free, recurved hook of penis valve head about one- third as long as more apical part of penis valve head (Fig. 80), penis valve shaft broader at its mid-point from lateral aspect (Fig. 120) than spatha is at its mid-point from dorsal aspect; pubescence of terga III-V predominantly orange-red, terga VI-VII predominantly pale, hairs nearly white with orange bases, black hairs mostly confined to basal margin of tergum III, otherwise intermixed on lateral quarters of remaining terga (Fig. 395) ......... ladakhensis Genus BOMBUS Latreille [Bremus [Jurine], 1801: 164. Type species: Apis terrestris Linnaeus, 1758: 578 = Bombus terrestris (Linnaeus), by subsequent designation of Morice & Durrant (1915: 429). Suppressed by the ICZN in Opinion 135 (ICZN, 1939).] [Bremus Panzer, [1801]: pls. 19-21. Type species: Apis agrorum Fabricius, 1787: 301 = Bombus pascuorum (Scopoli), by subsequent designa- tion of Sandhouse (1943: 532). Suppressed by the ICZN in Opinion 220 (ICZN, 1954).] PAULH. WILLIAMS Bombus Latreille, 1802a: 437 [redescribed 1802b: 385]. Type species: Apis terrestris Linnaeus, 1758: 578 [cited as Apis terrestris F.| = Bombus terrestris (Linnaeus), by monotypy. Specialists often find it useful to be able to refer to groups of closely related species by group-names. The use of the established subgeneric names as labels for assemblages of similar bumble bee species is continued here solely for the sake of nomenclatural continuity, because there is little advantage at present in modifying the usage of previous authors (see the introduction on genera of Bombini). When it is possible to revise all of the bumble bees thoroughly, this system could be replaced or at least simplified. For this review, only a few additions and revisions to a preliminary cladistic study of relationships (Williams, 1985) are included in the discussions of affinities for each species. As little as possible is added to the burden of subgeneric nomenclature, although the application of certain names does require some clarification. Richards (1968) described the characteristics of both sexes for species of the subgenera referred to here, with the exception of Psithyrus, which is described by Popov (1931) and by Lgken (1984). Subgenus MENDACIBOMBUS Skorikov Mendacibombus Skorikov, 1914a: 125. Type species: Bombus mendax Gerstaecker, 1869: 323, by subsequent designation of Sandhouse (1943: 572). Mendacibombus Skorikov; Kriiger, 1917: 62 (asa subgenus of Bombus Latrielle). Species of the subgenus Mendacibombus have been regarded as showing the most plesiomor- phic forms of the male genitalia among bumble bees (Williams, 1985). A brief review of most species of this group is now possible thanks to generous loans of material from other institutions (ITZ, MNHU, ZI). These nominal taxa have been described from female type specimens, which are all closely similar in morphology (e.g. mostly described as subspecies of B. mendax by Skorikov, 1910a). The application of the names to the more divergent males naturally depends on correct association of the sexes. For these associa- tions I have relied on reference material from A. S. Skorikov’s collection (ZI, with a few specimens in the BMNH), because he had access to relatively large samples of material and had described many of the taxa himself. The results of a cladistic analysis show that the subgenus Mendacibombus is likely to be THE BUMBLE BEES OF THE KASHMIR HIMALAYA paraphyletic with respect to all other bumble bees (Fig. 5, all other bumble bees represented by B. exil, see the comments on B. avinoviellus, B. himalayanus and B. bohemicus). This is a change in the interpretation of relationships within this species-group (see the introduction on generic nomenclature) that results from a change in the interpretation of the broadened state of the inner lateral shelf of the penis valve. This character is at best weakly developed in species of the out-group, Euglossini, which have the shaft of the penis valve much shorter. If the strong development of the shelf is regarded as plesiomorphic rather than apomorphic within the subgenus Mendacibombus, then its compatibility with the additional character state patterns in this analysis is maintained. Holophyletic groups within the paraphyletic sub- genus Mendacibombus are largely single species (monotypic) or unresolved. For a general descrip- tion of both sexes of the species of this subgenus see Richards (1968) and the keys (couplets 1 & 2). The North American B. (Bombias) nevadensis Cresson is similar to species of the subgenus Mendacibombus in the simple, straight form of the penis valves (see Milliron, 1971: pl. 11) and in the enlarged compound eyes of the male. How- ever, at present the weight of evidence does not support a particularly close relationship by descent. On the one hand, the shared form of the penis valves is interpreted as a symplesiomorphy among the bumble bees and the enlarged eyes of the males are likely to be a convergent character state. Alcock & Alcock (1983) related the enlarged eyes of male B. nevadensis to the particular kind of mate-searching behaviour that they employ. Among the Kashmir fauna, this behaviour is apparently related in turn more to habitat structure than to relationships by ancestry (see the introduction on the relationship of male eye-size to mate-searching behaviour and habitat structure, also see the com- ments on the species of the rufofasciatus-group: B. simillimus, B. pyrosoma and B. rufofasciatus). Thus neither the shape of the penis valve nor the shape of the eyes can be used as evidence of close relation- ship. On the other hand, B. nevadensis has a flattened form of the male gonostylus that has a distinct interio-basal process. More particularly this species has a strongly narrowed form of the volsella that has a narrowly subapical and broad, many-toothed, interio-apical process. B. nevaden- sis appears to share these character states (2synapomorphies) with species such as B. (Alpinobombus) arcticus (Quenzel) (oldest avail- able name for B. hyperboreus Schonherr), rather than with any species of the subgenus Mendaci- bombus. Consequently this species is not included here in the subgenus Mendacibombus. 39 The males of the European and central Asian B. (Confusibombus) confusus Schenck also have relatively simple, straight heads of the penis valves and enlarged compound eyes. Because of this similarity, B. confusus, like B. nevadensis, has sometimes been grouped with B. mendax (e.g. Kruseman, 1952). However, both B. confusus and B. nevadensis share the development of an inwardly-directed process from the base of the gonostylus with most bumble bees apart from species of the subgenus Mendacibombus (synap- omorphy of Psithyrus + all other subgenera of Bombus excluding Mendacibombus). Ito (1985) concluded from his detailed phenetic study that B. nevadensis is most closely similar to B. confusus, and then to B. (Kallobombus) soroeensis, followed by species of the subgenera Rhodobombus or Alpinobombus. Evidence from biochemical charac- ters would be particularly useful in these cases. Within the subgenus Mendacibombus, nests have been described in detail for only the European B. mendax (by e.g. Haas, 1976; see also a brief footnote on a nest of B. makarjini by Reinig in Bischoff, 1931). These nests appear to show a greater resemblance in their structure to nests of some stingless bees (cf. Wille & Michener, 1973) than do those of other bumble bees. In particular, the cocoons are almost completely torn down soon after the emergence of the adults, so that pollen and honey must be stored elsewhere. In contrast to other bumble bees, these food reserves are held exclusively outside the wax envelope of the brood nest and the honey may be stored in hexagonally arranged wax cells. Bombus (Mendacibombus) avinoviellus (Skorikov) (Figs 14, 16, 23, 27, 31, 35, 43-47, 86, 87, 126, 127, 166, 167, 220, 225, 232-241, Maps 5 & 6) Mendacibombus avinoviellus Skorikov, 1914a: 126. Lectotype by designation of Podbolotskaya (in press) [not seen]. Paralectotype queen, INDIA: Kashmir, Baltal, 11.vi.1912 (Jacobson) (ZI) [examined]. Bombus niveatus subsp. callophenax Cockerell, 1917: 122. Holotype queen by monotypy (see Note 2 below), INDIA: Kashmir, no further data (Woglum) (USNM) [examined]. Syn. n. Bombus (Mendacibombus) avinoviellus (Skorikov); Richards, 1930: 635. [Bombus (Mendacibombus) avinoviellus var. subtunicatus Richards, 1930: 635 [examined]. Infrasubspecific (see Note 1 below) (Art. 45g (ii)(1)), unavailable name (Art. 45e).] NOMENCLATURE. Note 1. Richards’s use (1930: 40 634 etc.) of the term ‘Subsp.’ elsewhere in his paper is taken to indicate that infrasub- specific rank is meant for B. avinoviellus var. subtunicatus. TYPE MATERIAL. Note 2. Cockerell’s description of B. niveatus callophenax gives no reason to believe that he had based his description on more than one specimen. Only a single queen in the USNM collection agrees with the original description and carries the data quoted (Dr B.B. Norden, in litt.) and this specimen bears a red label ‘Type’. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. B. avinoviellus is easily distinguished from the other species of the subgenus Mendac- ibombus (and from all other bumble bees) by the laterally compressed and scythe-like heads of the penis valves, which have the weakly-sclerotised lateral channel relatively narrow (Figs 46, 47, 86, 87) (plesiomorphic within the bumble bees, Bombini, and shared with the South American orchid bees, Euglossini). The closest relatives of B. avinoviellus are probably B. mendax Gerstaecker and B. makarjini Skorikov (Fig. 5). B. mendax has the head of the penis valve not laterally compressed as for B. avinoviellus, but rounded and finger-like, with a narrowly subapical shoulder on the inner margin from the dorsal aspect. The gonostylus is expanded ventrally along its interio-apical margin, so that it is particularly stout (synapomorphies of B. mendax with all species of Mendacibombus excluding B. avinoviellus; the gonostylus may be secondarily slightly narrowed for B. makarjini and the species of the defector-group, handlirschianus- group and waltoni-group, see the comments on B. himalayanus). B. mendax is known from the European Alps and the Pyrenees (BMNH, PW). B. makarjini has the head of the penis valve even more strongly dorso- ventrally compressed than for B. mendax and the penis valve head is relatively shorter. The head of the penis valve is defined by an outer toothed ridge (e.g. Fig. 86), which for B. makarjini extends for only half of the distance from the apex to the ventro-basal angle of the penis valve. The apex of the volsella is strongly concave on the inner margin so that the apex is narrowed and inwardly curved (synapomorphies of B. makarjini with all species of Mendacibombus excluding B. avinoviellus and B. mendax). B. makarjini is known from the region of the Pamir and the Tien Shan ranges, where it is uncommon (Reinig, 1930; Skorikov, 1931; Panfilov, 1957; ITZ, ZI). All the other species of the subgenus Mendacibombus have the penis valve narrowed PAULH. WILLIAMS basally, particularly by the reduction of the inner lateral shelf (Figs 48 & 49) (synapomorphy; see the comments on B. himalayanus). B. niveatus subsp. callophenax was described by Cockerell from a queen that is closely similar in morphology and in its banded white colour pattern to the paralectotype of B. avinoviellus. There is no reason to believe that Cockerell was aware of Skorikov’s description of B. avinoviellus. These two nominal taxa have not previously been recognised as conspecific because most authors have followed the opinion of Richards (1930: 652) that B. niveatus subsp. callophenax belonged to a taxon of the subgenus Sibiricobombus that has a similar, banded white colour pattern. This is a misidentification (see the comments on B. asiaticus). A queen from the Karakoram range (Fig. 232), kindly lent to me by B. Tkalci, resembles B. marussinus in the unbanded yellow colour pattern of the pubescence (Fig. 254). However, this queen is otherwise closely similar in its morphological characteristics (see the key), as well as in the stronger clouding of the wings with brown, to the paralectotype queen of B. avinoviellus (see the comments on B. marussinus). It is therefore considered likely to be part of the same species as B. avinoviellus. More material, especially of the males, is needed in order to establish its status definitively. DISTRIBUTION. B. avinoviellus is a west Himalayan species (Map 5). It is known only from Uttar Pradesh (BMNH), Himachal Pradesh (BMNH, PW) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. avinoviellus has been recorded from the southern side of the Great Himalaya range, and from the Pir Panjal and Karakoram ranges, primarily in mesic subalpine scrub and in meadows in the upper montane coniferous forest zone (Map 6, Fig. 6): 29 queens, 69 workers, 80 males, from 12 localities 1900-4600 m (BMNH, BT, IZ, LK, MNHN, NR, PW, SEMK, USNM, ZI, ZM). VARIATION WITHIN KASHMIR. The only specimen of B. avinoviellus examined from the Karakoram range, a queen from Banidas (Fig. 232), has the thoracic dorsum, tergum I and the basal margin of tergum II uniformly lemon-yellow, with only a few black hairs above the wing bases. Individuals from around the Kishanganga valley have the thoracic dorsum dull yellow, with the exception of a band of black hairs between the wing bases (Figs 233-235). Individuals from around the Vale of Kashmir have the pale pubescence of the thoracic pubescence grey-white rather than yellow THE BUMBLE BEES OF THE KASHMIR HIMALAYA (Map 6, Figs 236-241), although there is often some yellow on terga I-II (Figs 236 & 237). Both banded yellow individuals and banded white individuals were taken together at Chittakatha Sar (banded yellow: 2/4 males; banded white: 2/4 males) and at Lal Pani (banded yellow: 5/6 workers, 28/35 males; banded white: 1/6 workers, 7/35 males) in the upper Kishanganga valley, and at Nigagar (1 banded yellow worker, 1 banded white male, 2 banded cream males) in the upper Sind valley. The individuals with an unbanded yellow thoracic dorsum (Fig. 232) are particularly closely similar in colour pattern to those of B. marussinus (Figs 254 & 255), which may occur with them. They are also similar to some individuals of B. branickii (Fig. 265), B. subtypicus (Fig. 319-324), B. biroi (Figs 335-337), B. asiaticus (Figs 371- 376) and B. semenovianus (Figs 397-402). The banded yellow individuals from the Kishanganga valley (Figs 233-235) resemble the local, banded yellow B. subtypicus (Figs 317 & 318), B. asiaticus (Figs 377, 378, 382) and B. keriensis (Fig. 414— 416). The banded white individuals from around the Vale of Kashmir (Figs 239-241) are closely similar in colour pattern to the B. asiaticus that occur with them (Figs 386, 387, 389-391), but also show some similarity to some banded white individuals of B. kashmirensis (Figs 306-310), B. biroi (Fig. 342), B. tunicatus (Figs 350, 353, 355) and B. keriensis (Figs 408 & 409, see also the comments on B. subtypicus). B. avinoviellus can be recognised by the presence of pale pubescence on the face, by the clouded wings, by the sub- generic and specific characters (see the key, couplets 1—3) and by its small size. FOOD PLANTS. (Ranunculaceae) Aconitum heter- ophyllum Wallich ex Royle; (Balsaminaceae) Impatiens glandulifera Royle; (Hippocastanaceae) Aesculus indica (Colebr. ex Cambress.) Hook.; (Leguminosae) Trifolium repens L.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC., unidentified dandelion-like composites; (Oleaceae) Syringa emodi Wallich ex Royle; (Scropulariaceae) Scrophularia pauciflora Benth., Digitalis lanata Ehrh. [introduced]; (Labiateae) Phlomis bracteosa Royle ex Benth., Stachys sericea Wallich ex Benth., Prunella vulgaris L. Bombus (Mendacibombus) himalayanus (Skorikov) (Figs 48, 88, 128, 168, 221, 242-253, Maps 7 & 8) Mendacibombus varius Skorikov, 1914a: 125. Lectotype by designation of Podbolotskaya (in press) [not seen]. Paralectotype queen, INDIA: 41 Kashmir, Zoji La, above 3000 m, 12—15.vi.1912 (Jacobson) (Z1) [examined]. Junior secondary homonym in Bombus of Psithyrus varius Lepeletier, 1832 [= B. campestris (Panzer)]. Mendacibombus mendax subsp. himalayanus Skorikov, 1914a: 127. Holotype queen by monotypy (see Note below), INDIA: Kashmir, Ladakh, Khardung La, 4500 m, 3.vii.1912 (Jacobson) (ZI) [examined]. Provisional synonym. Bombus_ (Mendacibombus) mendax subsp. himalayanus (Skorikov); Richards, 1930: 635. TYPE MATERIAL. Note. Skorikov specifies that he made his original description of Mendacibombus mendax subsp. himalayanus from a single female specimen. Only a single queen in the ZI collection agrees with the original description and carries the data quoted. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. In addition to B. avinoviellus, B. mendax and B. makarjini (see the comments on B. avinoviellus), there are at least eight species of the subgenus Mendacibombus that are known from both sexes. These species share forms of the penis valve that are narrowed at the base, particularly by the reduction of the inner lateral shelf (Figs 48 & 49) (synapomorphy). This character state is also shared with all the other bumble bees. Three putative species, B. himalayanus, B. marussinus and B. turkestanicus Skorikov, are found in the mountain system of the western Himalaya, the Hindu Kush, the Pamir and the Tien Shan ranges (see below). Their relationships are not fully resolved (Fig. 5). No synapomorphies are known from which these bees can be recognised as a true, monophyletic group, yet they are morphologically closely similar and they lack the synapomorphies of the remaining species of the subgenus Mendacibombus. Among these other species may be recognised the defector-group, a possible handlirschianus-group and the waltoni- group. The latter three groups share with all other bumble bees (see the comments on B. bohemicus) a further narrowing and straightening of the penis valve base, with some development of a small semi-circular projection from the inner basal part of the penis valve by a more pronounced reduc- tion of the inner lateral shelf, together with a reduction of the outer lateral shelf (synapomor- phies, least well developed for males of the handlirschianus-group, also only weakly evident for some species of other subgenera of Bombus). The defector-group includes at least B. defector Skorikov, for which the penis valve is elongated beyond the volsella and the volsella is particularly 42 broad basally, so that its inner angle is only very broadly rounded (?autapomorphies). B. defector may be conspecific with similar females described by Skorikov under the name B. mendax subsp. margreiteri Vogt, which have the corbicular fringes orange rather than black, and possibly also with females described by Skorikov under the name B. mendax subsp. altaicus, which have the pale pubescence lighter in shade as well as the corbicular fringes orange. I have not seen any males that can definitely be associated with either of these two nominal taxa. Bees of the defector-group are known from the northern Pamir and Tien Shan ranges (Skorikov, 1931; Panfilov, 1957; BMNH, PW, ZI). They may also occur in some of the mountains that extend north-eastwards as far as the other localities known for these bees in Kamchatka (Bischoff, 1930; Skorikov, 1931). The handlirschianus-group may include two species, for which the basal angle of the penis valve is more pronounced and fully ventral rather than lateral, and the gonostylus is narrowed apically (synapomorphies of the handlirschianus- group + waltoni-group + all other subgenera of Bombus, for which they may be secondarily much modified). This is not a true group in that no exclusive synapomorphies are known for the included species, yet they lack the apomorphic character states of the waltoni-group and of the other subgenera of Bombus. Females of B. handlirschianus Vogt have the pale pubescence grey-white. Other females, described by Skorikov under the name B. mendax subsp. shaposhnikovi, are known from the same area of distribution but have the pale pubescence yellow. It is possible that the white females are part of the same species as those with yellow bands (cf. comments on colour variation of B. keriensis). However, the only male of this species- group that I have seen with the pale pubescence white rather than yellow (ITZ) also has the whole gonostylus and the apex of the volsella distinctly broader than it is for the yellow males (see also Skorikov, 1931: figs 14 & 15). Therefore the yellow and the white bees are likely to belong to two separate species, B. shaposhnikovi and B. handlirschianus respectively. Both of these species of the handlirschianus-group are known from Turkey, the Caucasus and the Elburz mountains of Iran (e.g. Reinig, 1971; BMNH, ITZ). Species of the waltoni-group share with most other groups of bumble bees the development of a dorsal toothed ridge on the volsella (see Skorikov, 1931: fig. 20), which, from the ventral aspect, projects beyond the inner margin (synapomorphy of the waltoni-group + all other subgenera of Bombus, possibly secondarily reduced for B. PAULH. WILLIAMS (Eversmannibombus) persicus Radoszkowski and for species of Psithyrus). The species of the waltoni-group share a very strongly curled apex of the volsella (synapomorphy). The male gono- stylus of B. waltoni Cockerell is uniquely much broader at the apex than at the base and is dorso- ventrally more flattened (autapomorphy). B. waltoni is conspecific with B. mendax subsp. chinensis Skorikov (junior secondary homonym in Bombus of Apathus rupestris var. chinensis Morawitz), and is probably also conspecific with B. rufitarsus Friese (see Skorikov, 1914a@), and with B. asellus Friese. B. waltoni is known from Tibet (Cockerell, 1910; Friese, 1918; Richards, 1930; Wang, 1982; BMNH), Qinghai (Morawitz, 1886; Skorikov, 1910a, 19126; Bischoff, 1936; Wang, 1982; ZI) and Sichuan (Wang, 1982). A record for Mongolia (Friese, 1924) needs confirma- tion. B. convexus Wang (oldest available name for B. lugubris Morawitz, 1880, syn. n., which is a junior secondary homonym in Bombus of Psithyrus lugubris Kriechbaumer, 1870 [= B. maxillosus Klug]) has the gonostylus narrow and finger-like (?plesiomorphic within this group), but the apex of the volsella is strongly elongated (autapomorphy). B. convexus is known from Tibet (Wang, 1979, 1982; PW), Qinghai (Wang, 1982), Sichuan (Morawitz, 1890), Gansu (Morawitz, 1880, 1890; Bischoff, 1936; MNHU, ZI) and Beijing (Panfilov, 1957). Another species that may belong to this group was described from a single female from the mountains of northern Mongolia (Mendacibombus superbus Tkalci, 1968a). This individual is unusually large for a species of the subgenus Mendacibombus. B. himalayanus, B. marussinus and B. turk- estanicus appear to differ from one another only in colour pattern (Skorikov, 1910a, 1914a), in subtle differences in the male genitalia (Figs 48, 49, 168, 169, 128, 129) and in details of female sculpturing (Figs 220-222). Nonetheless these differences are apparently maintained in combination at sites where they are recorded together. For instance, two queens and three workers from Chhantir Gah in the Hindu Raj range are assigned in this review to B. marussinus. If these workers were considered together with the material from the same locality that is ascribed to B. himalayanus, all of the workers could be interpreted as representing almost a continuum in colour variation between the extreme individuals with an unbanded yellow thorax (originally described under the name B. mendax subsp. marussinus) and those with a banded white thorax (described under the name Mendacibombus varius). However, the workers with a black band on the thoracic dorsum (both Mendacibombus_ varius and Mendacibombus THE BUMBLE BEES OF THE KASHMIR HIMALAYA mendax subsp. himalayanus) do have more punctures on the labral ridge, which, although broad, is also particularly weakly raised near its mid-point (Figs 220-222). This combination of character states is strictly maintained where the two taxa occur together at Holojut as well as at Chhantir Gah. Therefore there is no evidence of hybrids from interbreeding between B. himalayanus and B. marussinus, so they are considered to be separate species (see the comments on B. marussinus). There is a similar lack of obvious hybrids as evidence for interbreeding between B. himalaya- nus and B. turkestanicus, although I have not seen any samples from single localities that include both taxa. From the material available, B. himal- ayanus always retains the particularly weak labral ridge with many punctures in combination with a tendency towards black pubescence on the lower sides of the thorax and on the legs, both of which distinguish it from B. turkestanicus. A single male with a yellow thorax with black hairs between the wing bases from Kilian (Raskam range, Xinjiang, China) was ascribed by Skorikov (1914a :127) to Mendacibombus mendax subsp. himalayanus s.str. Unfortunately this specimen could not be found in the ZI collection (M. Podbolotskaya, in litt.). This could be the same individual from the ‘Raskemkette, Nordhang des Kilieng’ that Skorikov later listed under the name Mendacibombus makarjini (see Skorikov, 1931: 215). I have seen only two queens of B. himalayanus s.str. from Ladakh. These do differ from queens of B. makarjini, but resemble queens of Mendacibombus varius (Fig. 222), in the more coarsely punctured middle section of the labral ridge. A comparison of much more material of B. himalayanus s.str. with Mendacibombus varius, B. turkestanicus and B. makarjini, especially for the males, is still needed. From the meagre evidence available at present, I consider that B. himalayanus s.str. is most likely to be part of the same interbreeding population and species as Mendacibombus varius, which is likely to be separate from B. turkestanicus and B. makarjini. DISTRIBUTION. B. himalayanus is a west Himalayan species (Map 7). It is known from Himachal Pradesh (BMNH, PW) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. B. himalay- anus is widespread in Kashmir, where it is recorded from the Hindu Raj, Ladakh, Zanskar, Great Himalaya, and Pir Panjal ranges, in alpine scrub and steppe (Map 8, Fig. 6): 15 queens, 70 workers, 12 males, from 13 localities 2700-4800 m (BMNH, NR, PW, SEMK, ZI, ZM; 1 male photographed in the Sangisfaid valley, C. A. Chadwell). 43 VARIATION WITHIN KASHMIR. Two queens from the Ladakh and Zanskar ranges (Map 8) have the pale pubescence yellow, with a band of black hairs between the wing bases and black basally on tergum III. The queen from Nimaling has a pronounced black band between the wing bases (Fig. 242), but Skorikov’s queen from the Khardung La has only a poorly-defined black band on the thorax (cf. B. marussinus). Five workers from Chhantir Gah in the Hindu Raj range have the thoracic band well defined, but the black band on the gaster is much reduced (Fig. 243). In common with many workers in this species-group (but not among queens, except for B. turkestanicus), these workers have many pale hairs ventrally. Also from Chhantir Gah, another queen and eight workers have a similar pattern, although the yellow of the thorax, and to a variable extent on tergum II, is replaced by grey-white. These specimens, and other similar individuals from elsewhere in the Hindu Raj range (Figs 244, 245, 247, 248), have tergum III almost entirely orange- red. This banded white and yellow pattern is predominant among individuals in the samples from the mountains around the Kishanganga valley and the Vale of Kashmir (Figs 246 & 249). However, many of these individuals show some replacement of the pale band at the rear of the thorax by black hairs and black is also more extensive on tergum III, so that it may form a distinct black band on the gaster (Figs 250-252). The single male from Sangisfaid valley is even more extreme, in that all of the orange-red pubescence of the gaster appears to have been replaced by black (Fig. 253). The yellow individuals of this species (Figs 242 & 243) are similar in colour pattern to the B. kashmirensis (Figs 295-300), B. lepidus (Figs 327 & 328), B. oberti (Figs 392 & 393), B. ladakhensis (Figs 394 & 396) and B. keriensis (Figs 412 & 413) that occur with them. The white and yellow individuals (Figs 244-252) are similar to the B. novus males (Figs 260 & 262), B. kashmirensis (Figs 301-304), B. lemniscatus (Figs 325 & 326), B. lepidus (Figs 330, 331, 333, 334), B. biroi (Figs 338 & 339) and B. rufofasciatus (Figs 431-436, 438, 439) that occur with them, although the yellow pubescence is on tergum I for B. himalayanus and on tergum II for the other species. B. himalayanus is easily distinguished from all other species of the genus Bombus with similar colour patterns by using the subgeneric characteristics (see the key, couplets 1 & 2). FOOD PLANTS. Kashmir: unidentified pink clover- like legume; (Compositae) Cirsium falconeri (Hook.f.) Petrak, unidentified dandelion-like 44 composites; (Gentianaceae) Swertia petiolata D. Don; (Scrophulariaceae) Picrorhiza kurrooa Royle ex Benth.; (Labiateae) Thymus linearis Benth. ex Benth. Ladakh: (Leguminosae) Caragana versicolor (Wallich) Benth. Bombus (Mendacibombus) marussinus Skorikov (Figs 49, 89, 129, 169, 222, 254-257, Maps 9 & 10) Bombus mendax subsp. marussinus Skorikov, 1910a: 330. Lectotype queen by designation of Podbolotskaya (in press), U.S.S.R.: Tadzhikistan S.S.R., Shugnan, 3000 m, 24— 27.vi.1909 (Makarjin) (Z1) [examined]. Bombus (Mendacibombus) marussinus subsp. afghanus Reinig, 1940: 230. Syntype queen and male, presumed lost (Tkalct, 1969a, see also Note 4 on B. biroi), AFGHANISTAN: Badakhshan, Khwaja Muhammed, 3800-4000 m, 25.vii-10.viii.1936 (Kotzsch) [not seen]. Syn. n. AFFINITIES. B. marussinus is closely similar to B. himalayanus and B. turkestanicus in morphology (Fig. 5, see the comments on B. himalayanus). Female B. marussinus have fewer fine punc- tures on the labral ridge than do those of B. himalayanus (Figs 221 & 222) or B. turkestanicus. This difference appears to be retained in combina- tion with the unbanded yellow thoracic dorsum where B. marussinus and B. turkestanicus occur together in the Tarest mountains of the Hindu Kush (BMNH) and where B. marussinus and B. himalayanus occur together in the Hindu Raj range (see the comments on B. himalayanus). I infer from this precise coincidence between the discontinuities in the patterns of variation, for both colour and sculpturing, that individuals of these three taxa are unlikely to be interbreeding. However, it is possible that there are narrow hybrid zones, so their specific status is still not certain. More information would be desirable concerning their mate-searching behaviour, where they occur together, in order to ascertain more directly whether or not they interbreed. The distribution of B. turkestanicus includes the Tien Shan (Skorikov, 1910a, 1931; Panfilov, 1957; ZI), the Pamir and the Hindu Kush ranges (BMNH). B. turkestanicus may occur in Kashmir, but I have yet to see specimens from this area that can be attributed definitely to this species (but cf. yellow workers assigned to B. himalayanus). According to Tkalcu (1969a), the syntypes of B. marussinus subsp. afghanus appear to have been lost. The original description contains nothing to PAULH. WILLIAMS distinguish this material from ‘Chodja-Mahomed’ [= Khwaja Muhammed] in the Hindu Kush from a slight variation of the colour pattern of the lectotype of B. marussinus. Tkalci ascribed material with a similar colour pattern from the ‘Nachbargebiet Afghanistans (Karakorum)’ to B. afghanus Reinig. From his detailed description and from an examination of a queen from Banidas kindly lent to me by Dr Tkalci, it is likely that these specimens from the Karakoram are part of the same species as the banded white B. avino- viellus, despite their uniformly yellow thoracic dorsa (see the comments on B. avinoviellus). DISTRIBUTION. B. marussinus is a central Asian species (Map 9). It is known from the Hindu Kush (Reinig, 1940; Tkalct, 1969a; BMNH), the Pamir (Skorikov, 1910a; Reinig, 1930; Skorikov, 1931; ZI), Pakistan (PW) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. marussinus has been recorded only from the Hindu Raj and western Karakoram ranges, in the drier subalpine scrub (Map 10): 5 queens, 7 workers, 2 males, from 5 localities 2700-3200 m (BMNH). VARIATION WITHIN KASHMIR. All of the individuals of B. marussinus have some yellow hairs in the pubescence of terga I-II, although there is varia- tion in the extent of this pale pubescence on tergum II (Figs 254-257). Workers have more pale pubescence ventrally, as for workers and queens of B. turkestanicus. This species is closely similar in colour pattern to some of the B. avinoviellus (Fig. 232), B. branickii (Fig. 265), B. subtypicus (Figs 319-324), B. biroi (Fig. 335-337), B. asiaticus (Figs 371-376) and B. semenovianus (Figs 397-402) that may occur with it. It is, however, easily recognised using the subgeneric characteristics (see the key, couplets 1 & 2). B. marussinus can be separated from any B. avinoviellus with a similar colour pattern (Fig. 232) of the pubescence by its clear (hyaline) wings and by the specific characters (see the key, couplet 3). FOOD PLANTS. No records. Subgenus PSITHYRUS Lepeletier, stat. n. Psithyrus Lepeletier, 1832: 373. Type species: Apis rupestris Fabricius, 1793: 320 = Bombus rupestris (Fabricius), by subsequent designa- tion of Sandhouse (1943: 572). Apathus Newman, 1835: 404 [footnote k]. Replace- ment name for Psithyrus Lepeletier, incorrectly stated to be a junior homonym of Psithyros Hubner (Lepidoptera). THE BUMBLE BEES OF THE KASHMIR HIMALAYA Laboriopsithyrus Frison, 1927: 69 (as a subgenus of Psithyrus Lepeletier). Type species: Bombus laboriosus Fabricius, 1804: 352 [?= Emphoropsis laboriosa (Fabricius)] in the sense of Frison [= Bombus citrinus (Smith), a misidentification, see Milliron (1960: 99), requiring designation by the ICZN (see Note below)] by original fixation. Provisional synonym. Ashtonipsithyrus Frison, 1927: 69 (as a subgenus of Psithyrus Lepeletier). Type species: Apathus ashtoni Cresson, 1864: 42 = Bombus ashtoni (Cresson), by original designation [p. 70]. Synonymised with Psithyrus Lepeletier by Milliron (1961). Fernaldaepsithyrus Frison, 1927: 70 (as a sub- genus of Psithyrus Lepeletier). Type species: Psithyrus fernaldae Franklin, 1911: 164 = Bombus fernaldae (Franklin), by original designation. Synonymised with Psithyrus Lepeletier by Milliron (1961). Eopsithyrus Popov, 1931: 134 (as a subgenus of Psithyrus Lepeletier). Type species: Apathus tibetanus Morawitz, 1886: 202 = Bombus tibetanus (Morawitz), by original designation [p. 135]. Synonymised with Psithyrus Lepeletier by Milliron (1961). Metapsithyrus Popov, 1931: 135 (as a subgenus of Psithyrus Lepeletier). Type species: Apis campestris Panzer, 1801: 74 = Bombus camp- estris (Panzer), by original designation. Synon- ymised with Psithyrus Lepeletier by Milliron (1961). Allopsithyrus Popov, 1931: 136 (as a subgenus of Psithyrus Lepeletier). Type species: Apis bar- butella Kirby, 1802: 343 = Bombus barbutellus (Kirby), by original designation. Synonymised with Psithyrus Lepeletier by Milliron (1961). Ceratopsithyrus Pittioni, 1949: 271 (as a subgenus of Psithyrus Lepeletier). Type species: Psithyrus klapperichi Pittioni, 1949: 273 = Bombus cornutus (Frison) (see Tkalct, 1989), by original designation. Synonymised with Psithyrus Lepeletier by Milliron (1961). Citrinopsithyrus Thorp [in Thorp et al.], 1983: 50 (as a subgenus of Psithyrus Lepeletier). Type species: Apathus citrinus Smith, 1854: 385 = Bombus citrinus (Smith), by original designa- tion. Syn. n. Note. Application of Laboriopsithyrus is a matter that should mandatorially be referred to the ICZN (Art. 70b). Since this name is treated here as a synonym of Psithyrus, I intend to take no action. Descriptions of the diagnostic characters of both sexes for species of this subgenus are provided in the keys (couplet 1 for the females and couplet 2 45 for the males; for further details see Popov, 1931; Loken, 1984). The many separate subgeneric names for species-groups of parasitic bumble bees are not used here, because the male genitalia of their type species are more similar to one another than is general among the subgenera of the genus Bombus (e.g. Pittioni, 1939a; Ito, 1985; Williams, 1985). Knowledge of the systematics of species of the subgenus Psithyrus from Asia has been limited by the scarcity of material in collections. These bees are permanent social parasites in colonies of other species of the genus Bombus (e.g. reviews by Alford, 1975; Fisher, 1987; no records from colonies of species of the subgenus Mendaci- bombus). Consequently they are usually much less abundant than individuals of the host species. Many of the nominal taxa have been described from very small samples, collected at widely separated localities. There is therefore relatively little information available to describe the varia- tion of many of the nominal taxa from Asia. Bombus (Psithyrus) bohemicus Seidl (Figs 21, 50, 90, 130, 170, 206, 258, Maps 11 & 12) Apis nemorum Fabricius, 1775: 380. Lectotype female by designation of Lgken (1984: 8) (see Note 1 below), DENMARK: Copenhagen, no further data (BMNH) [examined]. Junior primary homonym of Apis nemorum Scopoli, 1763. Bombus bohemicus Seidl, 1837: 73. Type not found (Tkalct, 1969b) (see Note 2 below), CZECHOSLOVAKIA: Prague, no further data [not seen]. Synonymy with Apis nemorum Fabricius suggested by Blithgen (1918), con- firmed here. TYPE MATERIAL. Note 1. In her revision of the Scandinavian Psithyrus fauna, Lgken (1984) states that the ‘type’ of Apis nemorum F. is a female in the Banks collection (BMNH). This had been examined for her by I. H. H. Yarrow, who had ascribed it to the subgenus Ashtonipsithyrus. In this collection, above an original, handwritten drawer label that begins ‘Apis nemorum ?’, stand two female Psithyrus that carry Yarrow’s deter- mination labels. One is a female labelled Psithyrus sylvestris Lepeletier and the other a female labelled Psithyrus bohemicus by Yarrow, identifications that I confirm. Lg@ken’s use of the word ‘type’ in reference to the second specimen, which has had terga and sterna V-VI mounted on card on a separate pin, is taken to be a valid lectotype designation (Art. 74b). 46 Note 2. There is no reason to doubt the usual interpretation of B. bohemicus, so it is neither necessary nor valid to designate a neotype (Art. 75b). AFFINITIES. The rare Mongolian (NMP) and Siberian (BMNH) 8B. (Exilobombus) exil (Skorikov) (justified emendation of ‘Mucidobombus exiln. nov.’ Skorikov by Milliron, 1961: 56) shares a pronounced interio- basal process of the male gonostylus (Fig. 5) with all groups of bumble bees (synapomorphy), apart from the species of the subgenus Mendacibombus (secondarily partly reduced for species of the subgenera Orientalibombus, Kallobombus Dalla Torre, Alpinobombus Skorikov, Pyrobombus and Cul- lumanobombus Vogt). A general dorso-ventral flattening of the male gonostylus against the dorsal surface of the volsella is shared by most groups of bumble bees (?synapomorphy), apart from species of the subgenera Mendacibombus, Exilobombus_ Skorikov, Confusibombus and possibly Bombias (but see the comments on the subgenus Mendacibombus). The absence of an interio-apical process on the volsella is also shared by species of the subgenera Eversmannibombus and Psithyrus (possibly not a symplesiomorphy with Mendacibombus, see Williams 1985, but a secondary reduction and synapomorphy of Eversmannibombus + Psithyrus, Figs 130-132). The species of the subgenus Psithyrus share the development of a particularly dense brush of long, branched hairs associated with the interio-basal process of the male gonostylus (synapomorphy). Within the subgenus Psithyrus, B. bohemicus belongs to a group of species (the ashtoni-group, formerly subgenus Ashtonipsithyrus) that share a reduction of the ventro-basal angle of the penis valve (Fig. 90) (synapomorphy). Most females of this group have the keels of sternum VI broad, narrowing apically where they converge before the apex of the sternum (Fig. 206) (the keels are more pronounced and blade-like in the North American B. suckleyi Greene). Within the ashtoni-group, B. bohemicus shares with the North American (BMNH, PW) B. ashtoni (Cresson) and the Korean (Kim & Ito, 1987; BMNH) B. coreanus (Yasumatsu) the develop- ment of a pronounced shoulder on the outer side of the penis valve shaft, immediately behind the penis valve head (Fig. 50) (?synapomorphy). This shoulder is weakly developed in the European and North African (Popov, 1927a, 1931; Loken, 1984; BMNH, PW) B. vestalis (Geoffroy in Fourcroy), which can also be distinguished by its broader male gonostylus. The male of B. coreanus has the head of the penis valve shorter so that it is almost PAULH. WILLIAMS semi-circular in dorsal aspect, with the apex inwardly directed. B. bohemicus is closely similar to the North American B. ashtoni and it is possible that they are parts of the same species. However, subtle differences in shape and sculpturing could be taken to reflect a lack of interbreeding as separate species. These differences include the broader male gonostylus, relative to the apex of the gonocoxite, and a slightly smaller ventro-basal angle of the penis valve for B. ashtoni. DISTRIBUTION. B. bohemicus is one of the most widespread of all socially parasitic bumble bee species (Map 11). It is known from Europe (e.g. Popov, 1927a, 1931; Alford, 1975; Lg@ken, 1984; BMNH, PW), south-eastwards to Turkey and the Caucasus (BMNH), and north-eastwards across the U.S.S.R. (Popov, 1927a, 1931; BMNH) to Kamchatka (Popov, 1927a; Bischoff, 1930). In central Asia it reaches southwards to the Tien Shan (Popov, 1927a, 1931), the Pamir (Popov, 1931) and Kashmir; and in the far east it reaches southwards to Mongolia (Popov, 1927a, 1931; Tkalci, 1974a; BMNH, PW), Inner Mongolia (Reinig, 1936), Heilongjiang (Popov, 1931), North Korea (Kim & Ito, 1987), Shanxi (Maa, 1948), Gansu (Bischoff, 1936) and Sichuan (Popov, 1927a). MATERIAL EXAMINED FROM KASHMIR. B. bohemicus has been collected in Kashmir only at Lal Pani (Map 12): 10 males, 3000-3700 m, 2-3.ix.1953 (Schmid) (BMNH). In Britain this species is known to breed in the nests of B. lucorum (Alford, 1975; see the comments on B. lucorum). VARIATION WITHIN KASHMIR. The males from Kashmir are closely similar in colour pattern (Fig. 258) to individuals from Britain. Among the fauna of Kashmir, they are most similar to some males of B. morawitzianus (Fig. 270), but are easily distin- guished by the narrow volsella. The females are likely to have a similar colour pattern to the males, but with very much less of the yellow pubescence at the rear of the thorax and on tergum I. This pattern is quite distinctive among the Kashmir fauna. FOOD PLANTS. No records. Bombus (Psithyrus) novus (Frison), comb. n. (Figs 24, 28, 32, 36, 51, 91, 131, 171, 207, 259-262, Maps 13 & 14) Psithyrus (Psithyrus) novus Frison, 1933: 340. Holotype female by original designation [p. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 341], INDIA: Kashmir, Nagaberan, 10000- 10500 ft [3000-3200 m] (Bion) (Calcutta) [not seen]. Psithyrus (Psithyrus) novus subsp. nepalensis Tkalct, 1974b: 318. Holotype female by original designation, NEPAL: Khumbu, Pheriche, 4300 m, 5.vii.1964 (Dierl) (ZS) [examined]. Provisional synonym. AFFINITIES. B. novus belongs to a group of species (the rupestris-group, formerly subgenus Psithyrus s.str.) for which the volsella has an acutely produced inner corner (Figs 131 & 132) (synapo- morphy). Females of this group have the keels of sternum VI blade-like and converging close to the apex of the sternum (Figs 207 & 208). The females have the mandibles conspicuously narrowed in the distal half (Fig. 32). Within the rupestris-group, males of B. rupestris and B. branickii (see the comments on B. branickii) are distinguished by a slight elongation of the penis valve shaft between the penis valve head and the ventro-basal angle (Figs 51 & 52) and by the narrower distal part of the volsella (Figs 131 & 132) (?synapomorphies). The females have the clypeus closely and coarsely punctured almost throughout. In contrast, other material from Kashmir runs to the name Psithyrus morawitzi Friese in Popov’s (1931) key, although it agrees closely with the more recent description of B. novus. Tkalct (1974b) considered B. novus to belong to the ‘chinensis-Gruppe von Psithyrus s. str.’. From Popov’s key this includes B. chinensis (Morawitz) and B. morawitzi, which were both described from Chinese material (from Gansu & Sichuan respectively). I have seen no specimens of either of these nominal taxa for comparison with B. novus. Bischoff (1936) also described material belonging to this group from Sichuan and Yunnan. His material showed the colour patterns both with and without yellow bands, as described for B. chinensis and B. morawitzi respectively, although he found that these individuals scarcely differed in morphology (‘Morphologisch ist diese Form [Psithyrus chinensis subsp. hoenei Bischoff, with yellow pubescence] kaum von dem typischen chinensis zu trennen, weicht aber in der Farbung sehr stark ab.’) and Bischoff identified them all as Psithyrus chinensis. More recently, Tkalct (1987) has regarded Psithyrus morawitzi as a subspecies of Psithyrus chinensis. The holotype of Psithyrus novus subsp. nepalensis differs from B. novus s.str. not only in colour (which could be due in part to the fading of any yellow or orange pubescence to whitish for this abraded, old specimen), but also in several 47 morphological details. In particular, none of the specimens from Kashmir has the basal keel of the mandible reaching the margin (Fig. 32, including the holotype, see Frison, 1933: fig. Sb). The weak incision in this part of the margin is present for the specimen from Nepal, as noted by Tkalci, but is not at all distinct for the specimens from Kashmir. The dorsal surface of the labral tubercles of the Nepalese specimen has many large punctures, which are few or lacking for the specimens from Kashmir. The unpunctured part of the ocello- ocular area extends nearly three-quarters of the breadth of the ocello-ocular distance for the specimen from Nepal but no more than half of the distance for the specimens from Kashmir. The weakly-arched transverse ridge of sternum II shows a median anterior deviation for the specimens from Kashmir that is not marked for the specimen from Nepal. However, both nominal taxa share a reduction in the density of the punctures on the clypeus and on antennal segment 1 (scape). With so little material avail- able from the eastern Himalaya and from the intervening areas, especially of the males, I consider it premature to conclude that these nominal taxa represent separate species. More material is needed to establish the relationships of all of the nominal taxa in this group. DISTRIBUTION. The present, provisional concept of B. novus is of a Himalayan species (Map 13). It is known from Nepal (Tkalct, 1974b; ZS), Himachal Pradesh (BMNH, PW) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. novus is recorded from the southern side of the Great Himalaya range and from the Pir Panjal range, in mesic subalpine scrub (Map 14): 13 females, 31 males, from 2 localities 3000-3700 m (BMNH, BT, NR, PW). During July and August 1986, I saw many females investigate gaps among large boulders that were lying in gullies in the subalpine zone of Mt Apharwat. Queens of several species of bumble bees had been seen at this site, many apparently searching for nest sites. Most abundant among these bees were queens of B. rufofasciatus. VARIATION WITHIN KASHMIR. The colour pattern of only one of the females agrees closely with the original description, in that it has a broad black band between the wing bases (Fig. 261). Other females have only a few black hairs along the posterior margin of the scutum (Fig. 259). The putative males are associated with these females because, among material from Kashmir, they also belong to the rupestris-group but, unlike B. branickii, resemble female B. novus in having 48 some form of black band between the wing bases (Figs 260 & 262). The males differ from the females in that the pale pubescence of the thorax and tergum I is grey-white, as for the female from Nepal (Tkalci, 19745). The colour pattern of the pubescence of the males is black, with the sides of the thorax as far as the leg bases, the rather poorly- defined anterior and posterior bands on the thoracic dorsum and tergum I grey-white; terga II-III pale yellow; terga IV-VII orange- red, with variable degrees of replacement by black basally on tergum IV (Fig. 260), which may extend over the entire tergum (Fig. 262); sterna and legs predominantly black. The wings are weakly to moderately strongly clouded with brown (infuscated). The colour pattern of the males (Figs 260 & 262) is similar to that of some males of B. himalayanus (Figs 246, 249, 252), B. kashmirensis (Fig. 304), B. rufofasciatus (Figs 433, 436, 439), and probably B. lemniscatus (cf. Figs 325 & 326), that occur in the same localities. FOOD PLANTS. (Ranunculaceae) Aconitum laeve Royle; (Gentianaceae) Swertia petiolata D. Don; (Scrophulariaceae) Scrophularia paucifiora Benth. Bombus (Psithyrus) branickii (Radoszkowski), comb. n. (Figs 52, 92, 132, 172, 208, 263-266, Maps 15 & 16) Psithyrus Branickii Radoszkowski, 1893: 241. Lectotype female by designation of Tkalct (1969a: 204), U.S.S.R.: ?Kirgiziya S.S.R., ‘Kara-Karyk’ mountains, 10000 ft [3000 m] (Barei) (MNHU) [examined]. Psithyrus rupestris var. eriophoroides Reinig, 1930: 110 (see Note 1 below). 4 syntype males, U.S.S.R.: Tadzhikistan S.S.R.; KirgiziyaS.S.R.; CHINA: Xinjiang (ITZ) [not seen]. Synon- ymised with Psithyrus branickii Radoszkowski by Tkalct (1969a). Psithyrus (Psithyrus) rupestris subsp. elisabethae Reinig, 1940: 231. Holotype female by monotypy (see Note 2 below), AFGHANISTAN: Hindu Kush, north side of ‘Nuksan’ pass, 3500-4000 m, mid vii.1936 (Kotzsch) (ZS) [examined]. Synonymised with Psithyrus _ branickii Radoszkowski by Tkalct (19692). NOMENCLATURE. Note 1. Reinig’s use (1930: 83 etc.) of the term ‘ssp.’ elsewhere in his paper is taken to indicate that infrasubspecific rank is meant for Psithyrus rupestris var. eriophoroides. However, this name has subsequently been PAULH. WILLIAMS treated as a name in the species group by Reinig (1935), so it is deemed to be of subspecific rank (Art. 45g(11)(1)). TYPE MATERIAL. Note 2. Reinig’s description of Psithyrus rupestris subsp. elisabethae specifies that only a single female was examined. A single female in the ZS collection agrees with the original description and bears labels with the data quoted. It also bears a red typed label “Holotypus’ and a handwritten label ‘Psithyrus (Ps.) / rupestris Fabr. / ssp. elisabethae / [female] n. Reinig / det.W.F.Reinig 1939’. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. B. branickii belongs to the rupestris- group and is closely similar to B. rupestris (see the comments on B. novus), a Eurosiberian species (e.g. Popov, 1931; Reinig, 1935; Lgken, 1984). B. rupestris is known from an area that reaches south- eastwards from Europe to Turkey and the Caucasus (BMNH); in central Asia it reaches south to the Altai (Popov, 1931); and in the far east it reaches southwards to Mongolia (Popov, 1931; Tkalcu, 1974a; BMNH), Inner Mongolia (Reinig, 1936), and Gansu (Popov, 1931; Bischoff, 1936). B. branickii has shallower and less well-defined post- ocular punctures, has a weaker post-ocular groove on the vertex (the ‘dorsal furrow of the gena’ in Richards, 1968) and has longer pubescence. This is especially evident in the posterior fringe of the hind basitarsus, where the length of the hairs exceeds the breadth of the basitarsus for B. branickii, but not for B. rupestris. The wings of B. branickii are also only moderately, rather than very strongly, darkened (infuscated). DISTRIBUTION. B. branickii is a central Asian and Tibetan species from the mountains around the deserts of China (Map 15). It is known from Transbaikal, Mongolia and the Altai (Popov, 1931), the Tien Shan (Radoszkowski, 1893; Morawitz, 1894; Popov, 1931; Maa, 1948; MNHU), the Pamir (Reinig, 1930), the Hindu Kush (Reinig, 1940; Tkalci, 19692; BMNH, ZS), Kashmir, Himachal Pradesh (BMNH, PW), Tibet (BMNH), Sichuan (Reinig, 1935), Shanxi (Yasumatsu, 1951) and North Korea (Kim & Ito, 1987). Reinig (1935: chart 1) mapped part of this distribution under the name Psithyrus rupestris subsp. eriophoroides. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. branickii is recorded from the Hindu Raj, Great Himalaya and Zanskar ranges, in the drier alpine scrub and steppe (Map 16): 21 females, 71 THE BUMBLE BEES OF THE KASHMIR HIMALAYA males, from 9 localities 1500-4800 m (BMNH, IZ, MI, NM, PW). During July 1980, I saw females of this species investigate gaps under boulders and holes in dry- stone walls on Nimaling plain. Although queens of 8 species of the genus Bombus were flying in the area, only B. asiaticus, B. kashmirensis and B. keriensis had workers flying at this time. VARIATION WITHIN KASHMIR. One of the females from Atro Sar (Fig. 266) and another from Sonamarg agree closely in colour pattern with the original description, in that the pubescence of terga V-VI is entirely black. The remaining females have red pubescence on these terga (Figs 263, 265). All females have tergum I yellow and often have a few pale hairs apico-laterally on tergum II. Two females from Nimaling have more of these pale hairs on tergum II, so as to form an apical band that is only interrupted in the middle by black. Another two females from the same locality also have a pale, apical band on tergum III (Fig. 263). The males associated with these females belong to the rupestris-group, but unlike B. novus, they lack a black band in the pubescence between the wing bases, although they do have a few black hairs intermixed on the thoracic dorsum (Fig. 264). Some individuals (Fig. 265) are closely similar in colour pattern to some B. avinoviellus (Fig. 232), B. marussinus (Figs 254-257), B. subtypicus (Figs 319-324), B. biroi (Figs 335-337), B. asiaticus (Figs 371-376) and B. semenovianus (Figs. 397-402) that may occur with them. B. branickii is easily distinguished by the subgeneric characters (see the key, couplets 1 & 2). FOOD PLANTS. (Leguminosae) Caragana versicolor (Wallich) ex Benth.; unidentified yellow thistle-like composites. Bombus (Psithyrus) ferganicus (Radoszkowski), comb. n. (Figs 22, 53, 93, 133, 173, 209, 267, 268, Maps 17 & 18) Psithyrus ferganicus Radoszkowski, 1893: 241. Lectotype female by designation of Tkalct (1969a: 206), U.S.S.R.: ?Kirgiziya S.S.R., “Kara-Karyk’ mountains (MNHU) [examined]. Psithyrus indicus Richards, 1929a: 139. Holotype male by original designation [p. 141], INDIA: Kashmir, Gulmarg, 1913 (Thomson) (BMNH) [examined]. Synonymised with Psithyrus fergan- icus Radoszkowski by Griitte (1937). AFFINITIES. B. ferganicus belongs to a group of species (the campestris-group, formerly subgenus 49 Metapsithyrus) for which the apical part of the volsella is greatly broadened (Figs 133 & 134) and the ventro-basal angle of the penis valve is apically rounded, although this angle retains its curved, hook-like form (Figs 53 & 54) (synapomorphies). Females of this group have the keels of sternum VI broad throughout their length, although they converge distinctly before and below the apex of the sternum (Figs 209 & 210). Within the campestris-group, B. ferganicus shows the least broadening of the volsella (Figs 133 & 134) (see the comments on B. morawitzianus). The most similar European species is the Euro- siberian (e.g. Popov, 1931; Alford, 1975; Lgken, 1984) B. campestris (Panzer). From Europe B. campestris reaches south-eastwards to Turkey (BMNH), the Caucasus (Popov, 1931; BMNH) and the Elburz (BMNH); from central Asia it reaches southwards to the Tien Shan (Popov, 1931); and in the far east it reaches southwards to Mongolia (Popov, 1931; Tkalct, 1974a; BMNH), Inner Mongolia (Reinig, 1936), Ningxia (Popov, 1931), and Gansu (Popov, 1931; Bischoff, 1936). The unique female of B. susteraianus (Tkalct) from North Korea is closely similar to B. campestris, but has slightly less pronounced keels of sternum VI (Tkalci, 1959) and slightly more pronounced labral tubercles. The male of this nominal taxon remains unknown. The Chinese B. pieli (Maa) (synonymised with Psithyrus tajushanensis Pittioni by Tkalci, 1987) is known from Sichuan (Tkalcu, 1987; BMNH), Hubei (Sakagami, 1972), Anhui and Zheijiang (Maa, 1948) and Fujian (Pittioni, 1949; Tkalct, 1987; BMNH). The male genitalia of B. ferganicus, in ventral aspect, have the distal half of the volsella narrower than for B. campestris or B. morawitzianus. I have not seen males of B. pieli, but Maa (1948: fig. 16) described its volsella as slightly shorter than for B. campestris. Females of B. ferganicus can be distinguished from those of the other three species by the shallower punctures of the post-ocular area. These punctures are widely spaced for B. ferganicus, as for B. morawitzianus, but in contrast they are very shallow and poorly-defined, so that the area is nearly smooth and strongly shining. The females of B. campestris and B. pieli have smaller punctures in the post-ocular area, separated by less than their own widths, though these punctures are slightly closer and better defined for B. pieli. B. pieli can also be distinguished from B. campestris by their labral tubercles, which are more strongly angled. DISTRIBUTION. B. ferganicus is a central Asian species (Map 17). It is known from the Tien Shan (Radoszkowski, 1893; Morawitz, 1894; Griitte, 1937; MNHU), the Hindu Kush (Reinig, 1937; 50 Tkalci, 19692; BMNH, PW), Pakistan (PW), Kashmir and Himachal Pradesh (BMNH, PW). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. ferganicus is recorded from the western Karakoram, Great Himalaya and Pir Panjal ranges, in both montane coniferous forest and subalpine scrub (Map 18): 14 females, 115 males, from 5 localities 2400-3700 m (BMNH, BT, NR, PW, ZM, ZS). VARIATION WITHIN KASHMIR. Only minor variation is evident in the replacement of pale hair on the gaster by black, especially on terga I-II, and in the colour of the pale hair, which varies from lemon- yellow to brown (Figs 267 & 268). This species is quite distinctive in colour pattern among the fauna of Kashmir. FOOD PLANTS. (Leguminosae) Trifolium repens L.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC. Bombus (Psithyrus) morawitzianus (Popov), comb. n. (Figs 54, 94, 134, 174, 210, 269, 270, Maps 19 & 20) Psithyrus (Metapsithyrus) morawitzianus Popov, 1931: 148, 183 (see Note 1 below). Lectotype female by designation of Podbolotskaya (in press), U.S.S.R.: Kirgiziya S.S.R., Naryn, 19. vii.1904[?] (ZI) [examined]. Psithyrus (Metapsithyrus) redikorzevi Popov, 1931: 160, 181. 5 syntype males, U.S.S.R.: Tadzhikistan S.S.R.; Kirgiziya S.S.R. (ZI) [not seen]. Synonymised with Psithyrus mora- witzianus Popov by Griitte (1937) (see Note 2 below). NOMENCLATURE. Note 1. Popov described Psithyrus morawitzianus with particular reference to a female that bore Vogt’s manuscript label “Ps. campestris var. morawitzianus typicus O.V. Cotype’ (Popov, 1931: 183, although the label actually reads ‘Ps. campestris / var. morawitzianus O.V./typicus O.V. / Cotype’). Therefore Popov is deemed to be the author (Art. 50a). Note 2. Following the Principle of the First Reviser (Art. 24), B. morawitzianus is accepted in precedence to B. redikorzevi, by the action of Griitte (1937). AFFINITIES. B. morawitzianus, like B. ferganicus, belongs to the campestris-group (see the comments on B. ferganicus). The male genitalia of B. morawitzianus clearly differ from those of B. ferganicus, B. campestris and B. pieli by the great breadth of the distal half of the volsella, in ventral PAULH. WILLIAMS aspect (Figs 133 & 134), and by the extremely short, or transverse, gonostylus (Figs 173 & 174). The punctures of the post-ocular area of the female B. morawitzianus are larger, more distinct and more widely spaced than for individuals of B. ferganicus, B. campestris or B. pieli. The punctures of tergum VI are finer for B. morawitzianus than for individuals of B. ferganicus, B. campestris or B. pieli. DISTRIBUTION. B. morawitzianus is acentral Asian species (Map 19). It is known from the Tien Shan (Morawitz, 1875; Popov, 1931; Griitte, 1937; ZI), the Pamir (Popov, 1931; Griitte, 1937), the Hindu Kush (Reinig, 1940; Tkalci, 1969a) and Kashmir. A record from the Kopet Dag (Rasmont, 1983) needs to be confirmed. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. morawitzianus is recorded from the Hindu Raj and Great Himalaya ranges (Map 20): 1 female, 7 males, from 4 localities 2700-4000 m (BMNH). VARIATION WITHIN KASHMIR. The colour patterns of both sexes are similar to those described by Popov (1931) (Figs 269 & 270). There is some variation in the colour of the pale pubescence of the male gaster between white and dull yellow. This species is quite distinctive in colour pattern among the fauna of Kashmir. FOOD PLANTS. No records. Bombus (Psithyrus) skorikovi (Popov), comb. n. (Figs 55, 95, 135, 175, 211, 271-274, Maps 21 & 22) Psithyrus skorikovi Popov, 1927b: 267. Holotype female by original designation [p. 268], CHINA: Gansu [?or Qinghai], northern slopes of the Xining mountains, before 1.vii.1910 (Grum- Grzhimailo) (Z1) [examined]. [Psithyrus skorikovi var. mesoxanthus Richards, 1928c: 360 [examined]. Infrasubspecific (see Note 1 below) (Art 45g(ii)(1)), unavailable name (Art. 45e).] Psithyrus (Fernaldaepsithyrus) gansuensis Popov, 1931: 168, 202. Holotype male by monotypy (see Note 2 below), CHINA: Qinghai, ‘Ui-ju’ to ‘Choto’, 15-16.viii.1908 (Kozlov) (ZI) [not seen]. Synonymy with Psithyrus skorikovi suggested by Popov (1931), confirmed here. NOMENCLATURE. Note 1. In the introduction to his paper, Richards (1928c: 345) contrasts his use of the term ‘varieties’, for ‘sporadically’ occurring colour variations, with the term ‘races’, which he used for ‘. . . populations inhabiting definite, usually THE BUMBLE BEES OF THE KASHMIR HIMALAYA continuous, areas and characterised by peculiarities of sculpture, hair-length or average colour- pattern.’. This is taken to show that infrasub- specific rank is meant for Psithyrus skorikovi var. mesoxanthus. TYPE MATERIAL. Note 2. Popov’s description of Psithyrus gansuensis specifies that only a single male was examined. If only a single male with the data quoted can be found, then this specimen should be regarded as the holotype (Art. 73a(ii)). AFFINITIES. B. skorikovi belongs to a group of morphologically similar species (the fernaldae- group, formerly subgenus Fernaldaepsithyrus) of mainly Eurosiberian distribution (e.g. Popov, 1931; Lg@ken, 1984), with one North American species (B. fernaldae (Franklin)). These species share a great reduction in the breadth of the male volsella (Fig. 135) and a narrowing of the apex of the gonocoxite (synapomorphies). Females of this group have the keels of sternum VI narrowed apically and converging a long way before the apex of the sternum (Fig. 211). This apex is itself narrowed into a ventrally-curved spine, which projects distinctly beyond the apex of tergum VI. Within the fernaldae-group, B. skorikovi is most similar to the Eurosiberian (e.g. Popov, 1931; Loken, 1984) B. flavidus Eversmann in that the ventro-basal angle of the penis valve remains acute and strongly hook-shaped (Figs 55 & 95). B. flavidus is otherwise unique within the fernaldae- group for its restricted distribution in the sub- alpine and subarctic zones (see Pittioni, 1942: map 4; Loken, 1984). In southern Europe, B. flavidus is known from the Pyrenees, the Alps and western Turkey (Pittioni, 1942; BMNH). In a separate northern area of distribution it reaches eastwards from Scandinavia to Kamchatka (Popov, 1931; Pittioni, 1942), without records from further south than Lake Baikal. The females of B. skorikovi differ from those of B. flavidus in the weaker basal keel of the mandible, in the less extensively and less densely punctured ocello-ocular area and in the more strongly marked labral tubercles and lamella. The male of B. skorikovi has the ventro- basal angle of the penis valve much narrower. The single male from Qinghai described by Popov under the name Psithyrus gansuensis appears from the description of the penis valve (see Popov, 1931: fig. 26D) to be a male of B. skorikovi. A single female from Qinghai described by Tkalci (1961) under the name Psithyrus kuani also appears to be closely similar to B. skorikovi. I have not seen either of these specimens. The females from Kashmir that are ascribed here to B. skorikovi differ from specimens from Qinghai and Gansu in that tergum VI has, at most, 51 only a very weak median ridge and there are more large punctures with fewer small punctures. For some of the females from Kashmir the sculpturing of the surface of tergum VI is weaker so that it appears more strongly shining. However, some of the females from southern Tibet appear to be intermediate, at least in the reduction of the median ridge (Richards, 1928c; BMNH). It is considered premature to regard the Kashmir material as part of a separate species. DISTRIBUTION. B. skorikovi is a peri-Tibetan species (Map 21). It is known from Kashmir, southern Tibet (Richards, 1928c; BMNH), Qinghai (Popov, 1931; Tkalci, 1961; ZS) and Gansu (Popov, 1927b, 1931; Bischoff, 1936; PW, ZI). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. skorikovi is recorded from the southern side of the Great Himalaya range and from the Pir Panjal range (Map 22): 15 females, 301 males, from 2 localities 2700-3700 m (BMNH, NR, PW). VARIATION WITHIN KASHMIR. Females from Mt Apharwat are similar in colour pattern to the holo- type, although the apical margin of tergum II is yellow, whereas the median part of tergum IV and the entirety of tergum V are black-haired (Fig. 273). The majority of the females from Lal Pani have the black pubescence of the thoracic dorsum reduced to a small, indistinct patch at the rear of the scutum and have more yellow hairs on the lateral parts of terga II and IV (Fig. 271). The putative males are associated with these females because they are the only males of the fernaldae- group of species among the material from Kashmir and occur together with the known females at Lal Pani, in the Kishanganga valley. The colour pattern of the pubescence of the male (Figs 272 & 274) is dull yellow, except for black on the front and sides of the head, in a poorly-defined band between the wing bases, on terga V-VII and as a very few hairs in the middle of the posterior margin of tergum IV; the dark hairs of terga VI-VII are pale-tipped and often distinctly orange; the hairs of the legs are predominantly black, with some pale hairs inter- mixed. The wings are weakly clouded with brown (infuscated). Some males have the black band between the wing bases much reduced (120/301 males) (Fig. 272), as described for the female. A few males have a strong admixture of black hairs basally on tergum II (25/301 males, Fig. 274). This species is quite distinctive in colour pattern among the fauna of Kashmir. FOOD PLANTS. (Compositae) Cirsium falconeri (Hook.f.) Petrak; (Scrophulariaceae) Scrophularia pauciflora Benth. 52 Subgenus DIVERSOBOMBUS Skorikov Diversobombus Skorikov, 1914d: 406 (as a sub- genus of Bombus Latreille). Type species: Bombus diversus Smith, 1869: 207, by sub- sequent designation of Sandhouse (1943: 546). Diversibombus Skorikov; Skorikov, 19385: 1. Unjustified emendation (Art. 33b(i, iii). For a general description of both sexes see Richards (1968). Bombus (Diversobombus) trifasciatus Smith (Figs 11, 17, 39, 56, 96, 136, 176, 275-278, Maps 23 & 24) Bombus trifasciatus Smith, 1852a: 43. LECT- OTYPE queen by present designation (see Note 1 below), CHINA: Zhejiang, Zhoushan (Fortune) (BMNH) [examined]. Bombus montivagus Smith, 1878b: 168 [pos- thumous re-proposal in error, 1879: 131]. Lectotype queen by designation of Tkalct (1968b: 23), BURMA: Karen, Moolaiyet, 3000-6000 ft [900-1800 m] (BMNH) [examined]. Syn. n. Bombus_ secundus Dalla Torre, 1890: 139. Replacement name for Bombus montivagus Smith, 1879, junior primary homonym of Bombus montivagus Smith, 1878b. Synonymised with Megabombus montivagus montivagus (Smith, 1878b) by Tkalct (19685). Syn. n. Bombus ningpoénsis Friese, 1909: 676. 2 syntype workers, CHINA: Zhejiang, Ningbo (MNHU) [not seen]. Synonymised with Bombus trifasci- atus Smith by Tkalct (1961). Bombus wilemani Cockerell, 1911: 100. Holotype worker by original designation [p. 101], TAIWAN: Arizan, 7500 ft [2300 m], 11.ix.1906 (Wileman) (BMNH) [examined]. Provisional synonym. Bombus haemorrhoidalis var. albopleuralis Friese, 1916: 108. Lectotype queen by designation of Tkalct’ (1974b: 344), INDIA: Uttar Pradesh, Kumaun, 21.vi.1910 (MNHU) [examined]. Syn. n. Bombus maxwelli Pendlebury, 1923: 67. Holotype queen by monotypy (see Note 2 below), WEST MALAYSIA: Pahang, Gunong Berumbun, 4800 ft [1500 m], 4.ix.1922 (BMNH) [examined]. Provisional synonym. Bombus (Hortobombus) mimeticus Richards, 1931: 529. Holotype queen by original designation [p. 530], INDIA: West Bengal, Darjiling, Tukvar, 4000 ft [1200 m], iv.1894 (Bingham) (BMNH) [examined]. Syn. n. Bombus (Hortobombus) mimeticus var. albol- PAULH. WILLIAMS ateralis Richards, 1931: 530. Holotype worker by monotypy (see Note 3 below), NEPAL: no further data (BMNH) [examined]. Syn. n. Bombus (Hortobombus) mimeticus var. ganto- kiensis Richards, 1931: 530. Holotype queen by monotypy (see Note 4 below), INDIA: Sikkim, Gantok, 24.vi.1903 (BMNH) [examined]. Syn. n. Bombus (Hortobombus) mimeticus var. turneri Richards, 1931: 530. LECTOTYPE queen by present designation (see Note 5 below), INDIA: Meghalaya, Shillong, v.1903 (Turner) (BMNH) [examined]. Junior secondary homonym in Bombus of Psithyrus turneri Richards, 1929a [= B. turneri (Richards)]. Syn. n. Bombus (Hortobombus) mimeticus var. insidiosus Richards, 1931: 531. Holotype queen by original designation, PAKISTAN: Baltistan, no further data (BMNH) [examined]. Syn. n. Bombus (Hortobombus) mimeticus var. geminatus Richards, 1931: 531. Holotype worker by monotypy (see Note 6 below), INDIA: Kashmir, Gulmarg, 1913 (Thomson) (BMNH) [exam- ined]. Synonymised with Megabombus albo- pleuralis albopleuralis (Friese) by Tkalct, 1974b. Syn. n. Bombus (Hortobombus) mimeticus var. magret- tianus Richards, 1931: 531. Holotype male by original designation [p. 533], 7BURMA [not Sikkim (see Note 7 below)]: no further data (MNHU) [examined]. Syn. n. Bombus (Diversobombus) ningpoensis subsp. minshanicus Bischoff, 1936: 19. LECTOTYPE worker by present designation (see Note 8 below), CHINA: Gansu/Sichuan border, ‘Ma- tou-shan’, 600-1000 m, 2.vi.1930 (Hummel) (MNHU) [examined]. Syn. n. Megabombus (Diversobombus) montivagus subsp. quasibreviceps Tkalci, 1968b: 27. Holotype queen by original designation, CHINA: Yunnan, no further data (Genf) [not seen]. Syn. n. Megabombus_ (Diversobombus) albopleuralis subsp. atropygus Tkalct, 1989: 58. Holotype worker by original designation, BURMA: Nam Tamai valley, 27°48’N 97°48’E, 3500 ft [1067 m], 12.ix.1938 (Kaulback) (BMNH) [examined]. Syn. n. TYPE MATERIAL. Note 1. There are two queens of B. trifasciatus in the BMNH collection labelled ‘Shang / hai’ that may have been among Smith’s syntypes in 1852. The original description quotes neighbouring ‘Chusan’ [= Zhoushan] as the type- locality. Much of Fortune’s material is labelled ‘Shang / hai’, although Smith described it as from ‘Chusan’ on Fortune’s advice (Smith, 1852a: 33, 43), even when some of it must have originated THE BUMBLE BEES OF THE KASHMIR HIMALAYA from India (see the comments on B. haemor- rhoidalis). In this case there is no reason to doubt Zhoushan as the area of origin of the material. The queen that agrees most closely with the description, by the absence of a distinct black band between the wing bases, carries a purple- edged label (1) ‘Lecto- / type’; (2) a blue label ‘Shang / hai’, reverse side ‘52 / 28’; (3) ‘trifasciatus / Type Sm.’, in handwriting identical to that of Smith; (4) ‘Bombus / trifasciatus / TYPE. Smith’; (5) a red-edged, printed label ‘Type’; (6) ‘B.M. TYPE / HYM. / 17B.1028.’. According to the accessions catalogue, the number 52/28 refers to 17 Hymenoptera collected at ‘Shanghai’ by Fortune and purchased from him between February 28th and March 6th, 1852, which predates the reading of Smith’s paper at the meeting (3.v. 1852) of the Entomological Society of London by two months. This specimen, which lacks both hind legs, the distal joints of all the remaining tarsi and part of the right antennal flagellum, is designated as lectotype (Art. 74a). Note 2. Pendlebury stated that he had seen two specimens of B. maxwelli: a queen, which he described, and a ‘similar’ worker. The “Type and paratype’ were then said to have been deposited in the Federated Malay States Museum, Kuala Lumpur. A single pre-1923 queen in the BMNH agrees with the original description and bears labels with the data quoted by Pendlebury, together with a label ‘Bombus / maxwelli Pendl. / Type. Pahang.’ and an accession number 1926-167. This number refers to a collection that is specified to have included eight Hymenopteran types, which was presented to the BMNH from the museum at Kuala Lumpur in 1926. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 3. Richards’s description of B. mimeticus var. albolateralis specifies that only one worker was examined. A single pre-1931 worker in the BMNH collection agrees with the original descrip- tion and bears labels with the data quoted. It also carries a label with the name B. mimeticus var. albolateralis and ‘TYPE’, in handwriting identical to that of Richards. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 4. Richards’s description of B. mimeticus var. gantokiensis specifies that only a single queen was examined. A single pre-1931 queen in the BMNH collection agrees with the original descrip- tion and bears labels with the data quoted. It also carries a label with the name B. mimeticus var. gantokiensis and ‘TYPE’, in handwriting identical to that of Richards. I believe that this is the single 53 specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 5. The original description of B. mimeticus var. turneri lists one queen and one worker as typical. A single pre-1931 queen in the BMNH collection bears a purple-edged label (1) ‘Lecto- / type’; (2) ‘Assam / R. Turner. / 1905-175’; (3) ‘Shillong / 5.03’; (4) ‘B. orientalis var.’; (5) ‘B. MIMETICUS / RICHARDS / VAR. TURNERI / RICHARDS. / Type [female]’ in handwriting identical to that of Richards; (6) a red-edged label ‘Type’; (7) “B.M: TYPE / HYM. / 17B.1027.’. This specimen, which lacks the distal tarsal joints of all legs, the tibia of the right mid leg, all of the right front leg and most of both antennal flagella, is designated as lectotype (Art. 74a). Note 6. Richards’s description of B. mimeticus var. geminatus lists two workers, the second of which is specified to be the paratype. A single pre- 1931 worker in the BMNH collection agrees with the original description and bears labels with the data quoted for the first worker. It also carries a label with the name B. mimeticus var. geminatus and ‘TYPE’, in handwriting identical to that of Richards. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 7. Tkalci (1968b: 27) questioned the validity of the locality data ‘Sikkim’ associated with the holotype of B. mimeticus var. magret- tianus Richards. The other material with this colour pattern that Tkalci had examined bears labels for “Tenasserim’ in Burma. Note 8. Bischoff’s description of B. ningpoensis subsp. minshanicus lists two workers from Gansu, dated ‘14.5’ and ‘2.6’. A worker in the MNHU collection bears Bischoff’s labels with this name, the date 14.5.1930 and a red label ‘Para-/Typus’. A second worker in the MNHU collection agrees with the description and bears a printed label (1) “Sven Hedins / Exp. Ctr. Asien / Dr Hummel’; (2) ‘Kina/S. Kansu’; (3) ‘ningpoensis / minshanicus / n. subsp. [worker] / det. Bischoff’; (4) ‘(Bombus) loubouen- / sis Friese m. minshani- / cus (Bisch.) / Skorikov det.’; (5) a red label ‘68 / 61’; (6) a pencilled label ‘loubouensis Friese / var. nov.’; (7) a pencilled label ‘2/6’; (8) ‘M. (D.) tri- / fasciatus (Sm. / Tkalci det.’; (9) ‘Zool. Mus. / Berlin’. This second specimen, which lacks only the tarsi of the left mid leg, is designated as lectotype (Art. 74a). AFFINITIES. The species of the subgenera Mega- bombus Dalla Torre, Senexibombus Frison and Diversobombus share a displacement of the two hooks on the inner margin of the volsella to a narrowly subapical position (Fig. 136), and the interio-basal process of the male gonostylus is 54 extended as a strongly recurved cusp (Fig. 176) (synapomorphies). The species of the subgenus Diversobombus share an enlargement of the inner hooks of the volsella, of which the inner basal hook has become broadened and carries many small teeth (?synapomorphy). Within the subgenus Diversobombus, two species-groups have been recognised, the trifasci- atus-group and the diversus-group (Sakagami, 1972). Species of the diversus-group have the recurved part of the cuspate interio-basal process of the gonostylus reduced to a single, long, slender spine (?synapomorphy) (see Tkalct, 1965: figs 46 & 47). These species include the Japanese B. diversus Smith (distribution mapped by Sakagami, 1975: fig. 14; see also Ito, 1987), the Chinese B. longipes Friese (synonymised with B. hummeli Bischoff by Tkalci, 1987, and possibly also con- specific with Diversobombus malaisei Skorikov [not seen]) and the more widespread east Asian B. ussurensis Radoszkowski (distribution mapped by Sakagami, 1975: fig. 14). For males of the trifasciatus-group, the recurved part of the cuspate interio-basal process of the gonostylus bifurcates near its base, so that it forms two shorter, strongly curved spines (Fig. 176) (?plesiomorphy within the subgenus Diversobombus). Females can be distinguished from those of the diversus-group by their well- defined punctures of the post ocello-ocular area, whereas these punctures are embedded in coarse sculpturing for the species of the diversus-group. Three widespread nominal taxa have been recognised in the trifasciatus-group, which have been described under the names B. trifasciatus s.str., B. montivagus and B. haemorrhoidalis var. albopleuralis. They differ principally in colour pattern, possibly as members of regional groups of Miillerian mimics (Fig. 11, see the discussion of the colour patterns of the Kashmir fauna). These nominal taxa of the subgenus Diversobombus are very similar in colour pattern to the individuals of B. (Orientalibombus) haemorrhoidalis, B. (Alpigenobombus) breviceps Smith and B. (Pyro- bombus) rotundiceps Friese that occur with them (see the comments on B. haemorrhoidalis, B. kashmirensis and B. lemniscatus respectively; Sakagami & Yoshikawa, 1961; Tkalci, 1968b, 1989). B. haemorrhoidalis var. albopleuralis is the oldest available name for the Himalayan nominal taxa of the trifasciatus-group ( the ‘montivagus- Gruppe’ of Tkalci, 1974b), which were first described as a species separate from B. haemor- rhoidalis by Richards (1931), under the name B. mimeticus. The colour pattern of the pubescence on the thoracic dorsum is usually entirely black. PAULH. WILLIAMS There is variation in the amount of grey-white pubescence on the sides of the thorax (pleura). This pale pubescence is most often extensive in males. Among the material examined in the BMNH, the highest proportions of individuals with extensively pale-haired pleura appear to be from localities at lower altitudes in the eastern Himalaya. The pubescence of gastral terga I-II is usually bright lemon-yellow. The remaining gastral terga III-VI are predominantly red for individuals from the western Himalaya to as far east as Nepal, although the same extensively red pattern is also shown by two workers labelled ‘Manipur’ (females with terga III-VI red and without the white pubescence on the thorax were described under the name B. mimeticus var. insidiosus by Richards, whereas he described similar workers with the white pubescence under the name B. mimeticus var. geminatus). The red pubescence is replaced by black on tergum III or even on tergum IV for individuals from eastern Nepal eastwards to Arunachal Pradesh (these females with more extensive black on terga III-VI and without the white pubescence on the thorax were described under the name B. mimeticus var. gantokiensis by Richards for those with terga I-II pale yellow, or B. mimeticus s.str. for those with terga I-II creamy-white; Richards described similar females with the white pubescence under the name B. mimeticus var. turneri for those with terga I-II pale yellow, or B. mimeticus var. albolateralis for those with terga I-II creamy- white). Further east, the pubescence of terga III- VI is entirely black for individuals from northern Burma (described under the name Megabombus albopleuralis subsp. atropygus by Tkalcu, 1989). Material from further south in the highlands of South East Asia was described under the name B. montivagus by Smith (1878b). These individuals usually have the pubescence of the thorax bright orange, often paler yellow or grey-white at the sides, with a black band between the wing bases. The pubescence of tergum I and of the base of tergum II is bright lemon-yellow. Some specimens from above Hanoi have gastral terga III-VI black (described under the name Megabombus mont- ivagus subsp. quasibreviceps by Tkalci, 1968b), as for the north Burmese Megabombus albopleuralis subsp. atropygus, but individuals from most of South East Asia have red pubescence on the apical terga, as for the east Himalayan B. haemor- rhoidalis var. albopleuralis. Material from further to the north and east, in China, was described by Smith (1852a) under the name B. trifasciatus s.str. These individuals also have pale pubescence on the thorax and on terga I-II, although for this nominal taxon it is uniformly THE BUMBLE BEES OF THE KASHMIR HIMALAYA light yellow. There is a general trend for the black band between the wings to cover most of the thoracic dorsum for individuals from western China and to be reduced or absent for individuals from near the east coast. The apical terga of all of these individuals have at least some red pubescence. B. albopleuralis, B. montivagus and B. trifasci- atus s.str. have been regarded as three separate species (e.g. Tkalct, 1968b, 1974b, 1989). No two of these three principal nominal taxa of the trifasciatus-group are known to occur together at any locality without intermediates, so that direct evidence for barriers to interbreeding is not avail- able. Tkalct (19685) stated that in comparison with the most similar taxon, B. albopleuralis, B. montivagus differs slightly by its longer oculo- malar distance and by the stronger protrusion of its clypeus. However, individual differences between selected specimens may not be easy to interpret. For instance, the shape of the malar area at least is subject to large allometric variation with body size (Sakagami, 1972). Therefore large random samples of material from a range of localities across Asia are needed in order to examine trends in these differences in detail. Other characters of the morphology (especially of the sculpturing of the ocello-ocular areas of the females and of the sizes of the terminal process and inner hooks of the male volsellae) also vary considerably within each of these nominal taxa as well as among them, so I have been unable to find reliable morphological criteria to distinguish them. Thus in practice all three nominal taxa are still recognised by differences in the colour patterns of their pubescence. Some individuals from intermediate areas, particularly from western China and eastern Burma, appear to show intermediate colour patterns of the pubescence between those of B. albopleuralis, B. montivagus and B. trifasciatus s.str. (Fig. 11). Sakagami (1972) described varia- tion in the colour pattern of B. trifasciatus s.str. from western China towards that of B. montivagus, by replacement of the yellow hairs by black hairs at the apex of tergum II (Sakagami, 1972: fig. 6bD), and towards that of B. albopleuralis by replacement of the pale hairs by black hairs on the thoracic dorsum (Sakagami, 1972: fig. 6aA; a similar colour pattern was also described under the name B. ningpoensis subsp. minshanicus by Bischoff, 1936). Furthermore Richards (1931) described individuals with a colour pattern (B. mimeticus var. magrettianus, probably from central Burma, see Note 7 on the type material) that is intermediate between B. montivagus and B. albopleuralis, in that both the thoracic dorsum and most of tergum II are black-haired. I consider 55 it likely that at least some of these intermediate colour patterns are evidence of clines between the more widespread regional colour patterns (Fig. 11). The general pattern is for individuals from the west of the range in the Himalaya to have most black hair on the thorax, to have least black hair on the gaster and to have the darkest wings, whereas individuals from the east of the range generally have the least black hair on the thorax, more on the gaster and have lighter wings. This interpretation would place all the individuals within a single interbreeding population as parts of a single species. The names B. wilemani and B. maxwelli have been applied to populations that are isolated from other B. trifasciatus s.1. by sea and by tropical lowland forest respectively (Fig. 11). Both nominal taxa were originally described primarily using colour pattern characters and without comparison to any species of the subgenus Diversobombus. B. wilemani is characterised by an absence of yellow or orange-brown pubescence and by its orange-brown to nearly clear (subhyaline) wings. The tendency of individuals of B. maxwelli towards nearly uniformly orange pubescence converges closely with another very distantly- related species that also occurs in the Cameron Highlands of Malaysia, B. (Pyrobombus) rufoflavus Pendlebury (possibly conspecific with the east Himalayan B. flavescens Smith). The pubescence of the three workers of B. maxwelli in the BMNH varies from orange, with dark brown hairs inter- mixed on tergum II, to orange with most of terga II-III black. This dark specimen is therefore more similar in appearance to the lectotype of B. montivagus. Otherwise it differs by the lighter brown sclerites and slightly shallower punctures of the sculpturing of the head. I know of no evidence from which to suggest that individuals from these disjunct populations on either the island of Taiwan or in the Cameron highlands would not interbreed with those from the more widespread, northern population if they were to occur together. So B. maxwelli and even B. wilemani could be parts of a single species, B. trifasciatus s.1. However, I have not yet seen any males of B. wilemani from which to examine their characters. DISTRIBUTION. B. trifasciatus is a widespread Oriental species (Map 23, Fig. 11). It is known from Taiwan (Cockerell, 1911; Skorikov, 1933a; Frison, 1934; Chiu, 1948; BMNH, PW, SEMK), Jiangsu (BMNH), Zhejiang (Smith, 1852a; Friese, 1909; Tkalci, 1960; BMNH, SEMK), Fujian (Pittioni, 1949; BMNH), Guangdong (Sakagami, 1972), Jiangxi (MNHN, PW), Hubei (Tkalci, 1960; Sakagami, 1972; PW), Gansu (Bischoff, 56 PAULH. WILLIAMS Fig. 11 Distribution of the colour patterns of the ¢rifasciatus-group (for a colour key see Fig. 258). These individuals may all be considered to be parts of a single species, B. trifasciatus, depending on which species-defining criterion is accepted. The dashed line shows the 1000m contour above sea level and the solid line shows the 4000m contour. Most records are from material examined, supplemented from descriptions by Sakagami (1972: Leizhou) and by Tkalcii(1968b: 5 localities in Southeast Asia). Si THE BUMBLE BEES OF THE KASHMIR HIMALAYA LP 58 1936; MNHU), Sichuan (Panfilov, 1957; Tkalci, 1960; Sakagami, 1972; BMNH), Yunnan (Panfilov, 1957; Tkalci, 1968b), Vietnam (Tkalct, 1968), Laos (Tkalci, 1968b; BMNH), Thailand (Tkalct, 1968b; BMNH, PW, SEMK, ZM), Peninsular Malaysia (Pendlebury, 1923; BMNH, PW), Burma (Smith, 18785; Friese, 1918; Richards, 1931; Frison, 1935; Tkalci, 1968b, 1989; BMNH), Manipur (BMNH), Meghalaya (Richards, 1931; Frison, 1933, 1935; BMNH), Arunachal Pradesh (BMNH), south-eastern Tibet (Wang, 1982), Sikkim (Friese, 1918; Richards, 1931; Frison, 1935; BMNH, MCSN, MNHU, UM), Darjiling Bengal (Richards, 1931; BMNH), Nepal (Richards, 1931; Tkalci, 1974b; BMNH, NMS, PW), Uttar Pradesh (Friese, 1916; Frison, 1935; BMNH, MNHU, PW, SEMK), Himachal Pradesh (Richards, 1931; Frison, 1933, 1935; BMNH, SEMK), Kash- mir and Pakistan (Richards, 1931; Frison, 1933). A disjunct distribution between Burma and the mountains of Peninsular Malaysia is also known for B. (Pyrobombus) flavescens (Tkalct, 1974b; see the comments on B. lemniscatus). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. trifasciatus is recorded from the lower slopes in the Vale of Kashmir and from the foothills of the Pir Panjal range, in the more mesic habitats in and around the lower montane coniferous forests (Map 24, Fig. 6): 6 queens, 72 workers, 9 males, from 6 localities 1000-2700 m (AB, BMNH, IZ, NR, PW, ZM, ZS). VARIATION WITHIN KASHMIR. Almost all of the specimens from Kashmir that have been examined have the sides of the thorax black-haired and terga III-VI red. There is a small patch of yellow hairs in the middle of the basal margin of tergum III, often with a few black hairs intermixed (Figs 275-277). The only exception to this colour pattern is the holotype worker of B. mimeticus var. geminatus, labelled ‘Gulmarg’, which has grey-white hairs on the sides of the thorax (Fig. 278). Among the sample I collected at and around this locality during 1985 and 1986 (5 queens, 29 workers, 5 males), all specimens have the pleura entirely black-haired. This species is closely similar in colour pattern to the local B. haemorrhoidalis, which occur with it at some localities outside the Vale of Kashmir (Figs 279-281, see the comments on B. haemor- rhoidalis). It can usually be recognised by the presence of a few yellow hairs basally on tergum III, and for the females in particular, by the spinose mid basitarsus (Fig. 39) and by the broad band of punctures in the ocello-ocular area of the head. There may also be some similarity in colour pattern to some individuals of B. biroi (Figs 344 & PAULH. WILLIAMS 345), although the two species have not been found together. B. biroi has the oculo-malar distance much shorter and lacks a disto-posterior spine on the mid basitarsus. FOOD PLANTS. (Balsaminaceae) Impatiens glandu- lifera Royle; (Leguminosae) Lupinus sp. [intro- duced], unidentified pink clover-like legume; (Compositae) Cirsium wallichii DC.; (Acanth- aceae) Pteracanthus urticifolius (Kuntze) Bremek.; (Labiateae) Stachys sericea Wallich ex Benth. Subgenus ORIENTALIBOMBUS Richards Orientalibombus Richards, 1929b: 378 (as a sub- genus of Bombus Latreille). Type species: Bombus orientalis Smith, 1854: 402 = Bombus haemorrhoidalis Smith, by original designation. Orientalobombus Richards; Kruseman, 1952: 102 (as a subgenus of Bombus Latreille). Unjustified emendation (Art. 33b(i, iii)). For a general description of both sexes see Richards (1968). Bombus (Orientalibombus) haemorrhoidalis Smith (Figs 18, 40, 57, 97, 137, 177, 215, 279-281, Maps 25 & 26) Bombus haemorrhoidalis Smith, 1852a: 43. Types presumed lost (see Note 1 below) [not seen]. Bombus orientalis Smith, 1854: 402. Lectotype queen by designation of Richards (1929b: 383) (see Note 2 below), INDIA: West Bengal, Darjiling (Pearson) (BMNH) [examined]. Change of status to Orientalibombus haemor- rhoidalis orientalis (Smith) by Tkalcé (1974b). Syn. n. Bombus buccinatoris Smith, 1879: 132. Lectotype worker by designation of Richards (1929b: 384), INDIA: Himachal Pradesh, Kinnaur (BMNH) [examined]. Synonymised with Bombus orient- alis Smith by Bingham (1897). Syn. n. Bombus assamensis Bingham, 1897: 550. Lectotype male by designation of Richards (1929b: 384), INDIA: Assam, Margherita (Bingham) (BMNH) [examined]. Change of status to Orientalibombus haemorrhoidalis _ subsp. assamensis (Bingham) by Tkalct (1989). Syn. n. Bombus (Orientalibombus) montivolans Richards, 1929b: 382. Holotype male by original designa- tion [p. 383], LAOS: ‘Tintoe’, 1.xii.1918 (Salvaza) (BMNH) [examined]. Syn. n. Bombus (Orientalibombus) orientalis var. khasi- anus Richards, 1929b: 384. Holotype worker THE BUMBLE BEES OF THE KASHMIR HIMALAYA by original designation, INDIA: Meghalaya, Khasi Hills (BMNH) [examined]. Synonymised with Orientalibombus haemorrhoidalis pectoralis (Friese) by Tkalct (19745). Syn. n. Orientalibombus montivolans subsp. semibreviceps Tkalci, 19686: 10. Holotype worker by original designation, VIETNAM: Tonkin, Chiem Hoa, Viii-ix (Fruhstorfer) (Ziirich) [not seen]. Syn. n. Orientalibombus montivolans subsp. semicolor- icontrarius Tkalci, 1968b: 10. Holotype worker by original designation [p. 11], BURMA: ?Karen, ‘Cheba’, 900-1100 m, vi.1888 (Fea) (Zurich) [not seen]. Syn. n. Orientalibombus haemorrhoidalis subsp. cinnameus Tkalct, 1989: 47. Holotype queen by original designation, BURMA: Arakan Yoma, Mount Victoria, 1400 m, iv.1938 (Heinrich) (MNHU) [examined]. Syn. n. TYPE MATERIAL. Note 1. The abundant material from the western Himalaya agrees with the original description of B. haemorrhoidalis, although the type-locality is stated to be ‘Chusan’ [= Zhoushan, Zhejiang, eastern China]. I have not seen any specimens of the subgenus Orientalibombus from the region of Zhoushan. However, by analogy with the pattern of geographical variation described here for B. trifasciatus between Kashmir and eastern China (which closely resembles the Kashmir species of Orientalibombus throughout its known distribution between Kashmir and Vietnam, see the comments on B. trifasciatus), | would not expect any species from Zhoushan to agree in colour pattern with the original descrip- tion of B. haemorrhoidalis. Therefore the type- locality cited is considered to be inconsistent with the description given by Smith. Smith described B. haemorrhoidalis from syn- types in the Fortune collection, which was sold in 1872 without any material going to the BMNH or UM, and all syntypes are believed to be lost (D. B. Baker, pers. comm.). The Indian specimens, referred to separately by Smith in the original publication as being ‘In the British Museum’, cannot be considered as syntypes (Art. 72b(vi)). There is a single queen in the BMNH collection that had been acquired early enough to be one of these specimens. It bears a label ‘Shang / hai’, reverse side ‘51 / 26’, and a pale blue (drawer) label ‘haemorrhoidalis Smith / MSS’, in hand- writing identical to that of Smith. This accession number refers to 19 Hymenoptera that had been obtained in exchange with Baly for duplicates in March 1851. The area of origin is given in the accession catalogue as ‘China (Shanghai)’ (close to Zhoushan), but this has been crossed out and ‘Northern Ind.’ added in identical handwriting. 59 Therefore this specimen is almost certainly of Indian origin and agrees with the original descrip- tion and the previous interpretation of the taxon. It lacks the distal joints of all tarsi except that of the left mid leg, the apical gastral segments are curled ventrally and the pubescence has been abraded from the central area of tergum III. At some earlier stage, the gaster has also been glued back into place. Despite the inconsistency within the original description of B. haemorrhoidalis, the more recent authors such as Richards (19296, 1931), Frison (1933, 1935) and Tkalci (1968b, 1974b, 1989) have agreed on the identity of Smith’s B. haemor- rhoidalis. Richards (1929b) and Frison (1933, 1935) continued to apply B. haemorrhoidalis var. albopleuralis Friese to certain individuals of B. (Orientalibombus) haemorrhoidalis, although Tkalct’s (1974b) lectotype of B. haemorrhoidalis var. albopleuralis belongs to a species of the subgenus Diversobombus (conspecific with B. trifasciatus). Nonetheless, I regard the identity of B. haemorrhoidalis s.str. as not in doubt, so that it is neither necessary nor valid to designate a neotype (Art. 75b). Note 2. Two queens in the BMNH collection are candidates for having been among Smith’s syntypes of B. orientalis in 1854. One is simply labelled ‘India’. The other is labelled ‘Darjeeling’, reverse side ‘Dr. / Pearson’, with a label ‘orientalis / Type Sm.’, in handwriting identical to that of Smith, and with two identical accession labels ‘60— 15 / E.I.C.’. The accessions catalogue lists this material as presented in 1860 by the Secretary of the India Board and as having included 94 Hymenoptera from Java. This might have explained the type-locality statement ‘East Indies’, although it has a colour pattern that is characteristic of the eastern Himalaya and the species is not known from the East Indies. However, this dona- tion also contained 83 Hymenoptera from ‘India + asiatic Islands’ that is specified to have included ‘Several of the Hymenoptera type specimens described by Mr Smith’. According to D. B. Baker (pers. comm.), not all of Smith’s ‘B.M.’ syntypes were deposited in the BMNH at the time of publication. I agree with Richards (1929b: 383 footnote) that it is reasonable to regard the Darjiling queen as one of Smith’s syntypes. Richards’s use of the word ‘type’ in reference to this specimen is then taken to be a valid lectotype designation (Art. 74b). AFFINITIES. The species of the subgenus Oriental- ibombus share a reduction of the interio-basal process of the male gonostylus (Fig. 177) and an extended and twisted apex of the volsella (Fig. 60 137) (synapomorphies). This is a morphologically divergent group with no obviously close relatives. The present interpretation is that its affinities lie with the more northern species of the subgenus Thoracobombus Dalla Torre rather than with those of the subgenus Diversobombus (in contrast to the interpretation in Williams, 1985). Species of the subgenus Orientalibombus share with the species of the subgenus Thoracobombus an apical extension of the part of the volsella beyond the pair of hooks on its inner margin (?synapomorphy). The volsella of both is also relatively broad and short, with the inner hooks placed basally on the inner apical margin, close to the inner corner. Apart from B. haemorrhoidalis, the only other species of the subgenus Orientalibombus from the Himalaya is B. funerarius Smith, which is known from Sichuan (Frison, 1935; Panfilov, 1957), Yunnan (Panfilov, 1957), Burma (Skorikov, 19385; Tkalct, 1989; BMNH), Darjiling Bengal and Sikkim (Friese, 1918; Richards, 1929b; Frison, 1935; BMNH, UM), Nepal (PW) and Uttar Pradesh (BMNH). The males of B. funerarius can be distinguished by their gonostyli, which are much shorter than broad in dorsal aspect, whereas those of B. haemorrhoidalis are nearly square, with a very much reduced interio-basal process (Fig. 177) (autapomorphy). The females of B. funerarius can be distinguished by the tubercles of the labrum (Richards, 19296), which are unusually pronounced so that they form a transverse ridge that is only very narrowly interrupted by the median furrow. The apical impression of the clypeus is also deeper and less strongly punctured than for B. haemorrhoidalis. There may be just one other species in the subgenus Orientalibombus, B. braccatus Friese (synonymised with Bremus metcalfi Frison by Tkalct, 1987). This is known only from Sichuan and is similar in colour pattern to the B. trifasciatus that also occur in Sichuan, in that the thorax and terga I-II are extensively yellow-haired (Friese, 1905). The gonostylus of the male genitalia (holo- type male of Bremus metcalfi examined) is inter- mediate in outline between that of B. funerarius and that of B. haemorrhoidalis s.1. (see Frison, 1935: fig. 7c). The volsella is unique for species of this subgenus because the interio-basal hook is longer than broad and broadens slightly towards its serrated apex. I have not examined the females of B. braccatus. Three principal nominal taxa have been recog- nised among the remaining haemorrhoidalis- group of the subgenus Orientalibombus and these were described under the names B. montivolans, B. assamensis and B. haemorrhoidalis. They differ from each other in colour pattern, possibly PAULH. WILLIAMS as members of regional groups of Millerian mimics (see the discussion of the colour patterns of the Kashmir fauna). These nominal taxa are very similar in colour pattern to individuals of B. (Diversobombus) trifasciatus, B. (Alpigeno- bombus) breviceps Smith and B. (Pyrobombus) rotundiceps Friese that occur with them (see the comments on B. trifasciatus, B. kashmirensis and B. lemniscatus respectively; Sakagami and Yoshikawa, 1961; Tkalct, 1968b, 1989). B. haemorrhoidalis is the name that has been used most recently for all of the Himalayan nominal taxa in the haemorrhoidalis-group (the ‘haemorrhoidalis-Gruppe’ of Tkalct, 1974b, 1989). The pubescence of the thoracic dorsum is usually entirely black. As in the case of B. trifasciatus, the replacement of black by grey-white pubescence on the sides of the thorax appears to be most common among males from lower altitudes in the eastern Himalaya, at least among BMNH material. The pubescence of terga I-II is usually pale yellow. Individuals from the western Himalaya to as far east as Nepal have terga III-VI pre- dominantly red. Individuals from eastern Nepal to as far east as Arunachal Pradesh have black pubescence on tergum III and even on tergum IV (females with more extensive black on terga III- IV and without the white pubescence on the thorax were described under the name B. orientalis by Smith for those with terga I-II cream-yellow, or B. buccinatoris for those with terga I-II more distinctly lemon-yellow; whereas similar females with yellow on terga I-II and white pubescence on the thorax were described under the name B. orientalis var. khasianus by Richards). Tkalci (1989) described a queen from Burma under the name Orientalibombus haemorrhoidalis subsp. cinnameus that has the pubescence of the thoracic dorsum black, of the sides of the thorax white, of terga I-II nearly white, of tergum III predominantly black, and of the legs orange. The name B. assamensis has been applied to some specimens from Assam, which like those from northern Burma have very little or no red pubescence on the gaster. These individuals have the thoracic dorsum black, the sides of the thorax grey-white and terga I-II pale yellow. Material from further south in the highlands of South East Asia was described under the name B. montivolans by Richards (1929b). These individuals usually have the pubescence of the thorax bright orange, though often paler at the sides, with a black band between the wing bases. The pube- scence of tergum I and of the base of tergum II is pale yellow. Some specimens from above Hanoi have gastral terga III-VI black (described under the name Orientalibombus montivolans subsp. THE BUMBLE BEES OF THE KASHMIR HIMALAYA semibreviceps by Tkalci, 1968b) as for B. assam- ensis, but individuals from most of South East Asia have red pubescence on the apical terga, as for the east Himalayan B. orientalis. B. haemorrhoidalis and B. montivolans have been regarded as separate species (e.g. Tkalct, 19686, 1974b, 1989). However, males of B. haemorrhoidalis s.str. differ only slightly from those of B. assamensis and B. montivolans, for instance in the shape of the inner hooks of the volsella (Richards, 1929b: figs 4-7, although another B. assamensis male (BMNH) has the volsella as shown in Richards’s fig. 5) and of the distal margin of the gonostylus. These characters also vary within each nominal taxon. The three principal nominal taxa are not known to occur together at any locality, so that direct evidence for interbreeding is not available. Yet Tkalct (19685) described an individual from central Burma under the name Orientalibombus montivolans subsp. semicoloricontrarius that could represent an inter- mediate between B. orientalis var. khasianus of India and B. montivolans of Laos. This specimen has the pubescence of both the thoracic dorsum and of tergum II almost entirely black. Further- more, some individuals of B. orientalis var. khasianus from Assam have the red hair of terga III-VI largely replaced by black, so that they appear similar to B. assamensis. Therefore I consider it likely that these intermediate colour patterns are evidence of clines between the more widespread regional colour patterns. This inter- pretation places all of the individuals within a single interbreeding population as parts of a single species. DISTRIBUTION. B. haemorrhoidalis is a Himalayan and South East Asian species (Map 25). It is known from Vietnam (Tkalci, 19685), Laos (Richards, 1929b; Tkalci, 19685; BMNH), Thailand (Sakagami & Yoshikawa, 1961; Tkalct, 1968b; BMNH, PW, SEMK, ZM), Burma (Frison, 1933, 1935; Skorikov, 1938; Tkalci, 1968), 1989; BMNH, MNHU), Yunnan (Wang, 1987), Meghalaya (Richards, 1929b; Frison, 1935; BMNH), Assam (Bingham, 1897; Richards, 1929b; BMNH), south-eastern Tibet (Wang, 1988), Arunachal Pradesh and Bhutan (BMNH), Sikkim (Friese, 1918; Richards, 1929b, 1930; Frison, 1935; BMNH, MCSN, PW, UM), Darjiling Bengal (Friese, 1918; Richards, 1929b; Frison, 1933, 1935; BMNH), Nepal (Richards, 1929b; Frison, 1935; Tkalct, 1974b; BMNH, NMS, PW), Uttar Pradesh (Richards, 19295; Frison, 1935; BMNH; PW, UM, ZM), Himachal Pradesh (Smith, 1879; Richards, 1929b; Frison, 1933, 1935; BMNH), Kashmir and Pakistan (Richards, 1929b; Frison, 61 1933). There is also a dubious record from the Indian plains at Calcutta (Dover, 1922). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. haemorrhoidalis is recorded from the foothills of the Pir Panjal range (at Patnitop it was found in humid gullies in the coniferous forest) (Map 26): 8 queens, 102 workers, 1 male, from 3 localities 1000-2000 m (BMNH, NR, PW). VARIATION WITHIN KASHMIR. All of the females from Kashmir have the sides of the thorax black- haired and terga III-VI red, with only a few black hairs intermixed in the middle of tergum III, near the base (Figs 279-281). Only the male has grey- white hairs intermixed on the lower sides of the thorax (Fig. 281). This species is closely similar in colour pattern to the B. trifasciatus that occur with it (Figs 275— 277). It can usually be recognised by the absence of any yellow hairs at the base of tergum III, and for the females in particular, by the absence of the spine on the mid basitarsus (Fig. 40), and by the broad area without punctures on the ocello-ocular area of the head (Fig. 215). At the lower altitude of Muzaffarabad (1000 m, 15.x.1953), B. haemor- rhoidalis was apparently far more abundant (93/94 workers) than B. trifasciatus (1/94 workers). Higher up, at Patnitop (2000 m, 6-8.ix.1986), B. haemorrhoidalis was less abundant (8/31 workers) than B. trifasciatus (23/31 workers). FOOD PLANTS. (Balsaminaceae) /mpatiens gland- ulifera Royle; (Acanthaceae) Pteracanthus urtici- folius (Kuntze) Bremek. Subgenus SUBTERRANEOBOMBUS Vogt Subterraneobombus Vogt, 1911: 62 (as asubgenus of Bombus Latreille). Type species: Apis subterranea Linnaeus, 1758: 579 = Bombus subterraneus (Linnaeus), by subsequent desig- nation of Frison (1927: 68). Subterraneibombus Vogt; Skorikov, 1938a: 145. Unjustified emendation (Art. 33b(i, iii)). For a general description of both sexes see Richards (1968). Bombus (Subterraneobombus) melanurus Lepeletier (Figs 58, 59, 98, 99, 138, 139, 178, 179, 282-290, Maps 27 & 28) Bombus melanurus Lepeletier, 1836: 469. Lecto- type queen by designation of Tkalci (1969a: 202), SYRIA: no further data (UM) [examined]. 62 Bombus Tschitscherini Radoszkowski, 1862: 591 [by indication of Radoszkowski, 1859: 485]. Holotype queen by monotypy (see Note 1 below), U.S.S.R.: ‘Transoural’, no further data (ZI) [not seen]. Recombined as Bombus melanurus tschitscherini Radoszkowski by Vogt (1909). Syn. n. Bombus difficillimus Skorikov, 1912b: 609. Lecto- type queen by designation of Podbolotskaya (in press), U.S.S.R.: Tadzhikistan S.S.R., Pamir, 3700 m, 15.vi.1909 (Makarjin) (Z1) [examined]. Change of status to Bombus melanurus difficil- limus Skorikov by Reinig (1934). Provisional synonym. Bombus (Subterraneobombus) melanurus subsp. subdistinctus Richards, 1928b: 333. Holotype queen by original designation, INDIA: Kashmir, 8000-9000 ft [2400-2700 m], vi.1901 (Nurse) (BMNH) [examined]. Syn. n. Bombus (Subterraneobombus) melanurus subsp. griseofasciatus Reinig, 1930: 83. 36 syntype queens, 38 syntype workers, 16 syntype males, U.S.S.R.: Tadzhikistan S.S.R., Pamir, 3700— 4500 m, 5.vii- 29.viii.1928 (ITZ) [not seen]. Synonymised with Bombus melanurus difficil- limus Skorikov by Reinig (1934). Provisional synonym. Bombus (Subterraneobombus) maidli Pittioni, 1939b: 246. LECTOTYPE worker by present designation (see Note 2 below), MONGOLIA: north, no further data, 1892 (Leder) (NM) [examined]. Synonymised with Megabombus subdistinctus (Richards) by Tkalct (1969a). Syn. n. TYPE MATERIAL. Note 1. Radoszkowski described B. tschitscherini from a single queen sent to him by E. Eversmann that had a band of black pube- scence between the wings. If just a single specimen can be found in the ZI collection that bears labels with the appropriate data, this would be regarded as the holotype (Art. 73a(ii)). Note 2. Pittioni described B. maidli from one queen, one male and three workers. There is a single worker in the NM collection with (1) a red printed label ‘Type’; (2) a printed label ‘N. Mongolei/ Leder 92’; (3) ‘melanurus / det. Kohl.’; (4) ‘Subterraneob. / maidli Pitt. [worker] / det. Pittion’’ in handwniting identical to that of Pittion1; (5) ‘B. MAIDLI / MAIDLI PITT. / DET. PITTIONI, 1936’; (6) ‘LECTOTYPE / B. (S.) / maidli Pitt. / Tkalci det.’ [designation not published]; (7) a red printed label ‘Lecto- / typus’. This specimen, which lacks the right mid basitarsus, is designated as lectotype (Art. 74a). AFFINITIES. A strongly inwardly-curved apex of the penis valve head is shared by a large group of PAULH. WILLIAMS bumble bees, which in Kashmir includes the species of the subgenera Subterraneobombus, Alpigenobombus, Pyrobombus, Sibiricobombus and Melanobombus (Figs 58-71, 76-85) (synap- omorphy, secondarily much modified for Bombus s.str., Figs 72-75). The species of the subgenus Sub- terraneobombus share a particularly pronounced broadening of the penis valves in lateral aspect, in which the ventro-basal angle is produced and broadened across the penis valve axis, into a toothed paddle-like form (Figs 98-100) (synap- omorphy). The male gonostylus has the interio- basal process broadened antero-posteriorly (synapomorphy) and, for most species (?second- arily reduced for B. fedtschenkoi Morawitz), curved ventrally (Figs 178-180). The male genitalia of B. melanurus may be unique for males of the subgenus Subter- raneobombus, in that the head of the penis valve lacks a strong posteriorly-directed process from the dorsal exterior part of its base (Figs 58 & 59) and the interio-basal process of the gonostylus does not become broader distally (Figs 178 & 179) (plesiomorphies within the subgenus Subterrane- obombus) (see Reinig, 1930: fig. 5). This distin- guishes B. melanurus from B. fragrans Pallas, which also occurs in the central Asian steppes, but with a more western and northern distribution (distribution maps for both are provided by Skorikov, 1931: figs 5, 9; see also Panfilov, 1984: map 192). The head of the penis valve is not shortened and broadened for either of these species as it is for B. fedtschenkoi Morawitz, B. personatus (Fig. 60) and the species of the subterraneus-group (synapomorphy of B. fedtschenkoi + B. personatus + subterraneus- group, see the comments on B. personatus). Neither is the head of the penis valve strongly directed ventrally as it is for species of the subterraneus-group alone (synapomorphy of the subterraneus-group). The females of B. melanurus can be distinguished from B. fragrans, B. amurensis Radoszkowski from Siberia and from B. fedtschenkoi (which is closely similar in colour pattern) from the Tien Shan ranges, by its virtually unpunctured central area of the clypeus (Skorikov, 1914b). The oculo- malar distance is much longer than the breadth of the mandible at its base for B. melanurus, but scarcely longer than this breadth for B. fragrans or B. amurensis. I have seen no males of B. amurensis. B. flaviventris is another species that has been placed in the subgenus Subterraneobombus together with these species with a relatively short oculo-malar distance (Richards, 1930), although the material in the BMNH collection shows many character states in common with species of the THE BUMBLE BEES OF THE KASHMIR HIMALAYA subgenus Sibiricobombus (see the comments on B. oberti). B. melanurus s.str. was originally described from a queen with a yellow thoracic dorsum that lacks a band of black hair between the wing bases. Individuals with this colour pattern, which also have the wings clouded with brown (infuscated), occur in Turkey, the Tien Shan ranges and Kashmir. The pale pubescence varies from brown to lemon- yellow in colour (e.g. Figs 283 & 288). Richards (1928b) described the individuals from Kashmir with this unbanded colour pattern, but with wings that ‘are, on the average, much less dark than usual’ (see Tkalci, 1969a: plate facing page 200), under the name B. melanurus var. subdistinctus. Tkalct (1969a) concluded that this is a separate species, noting the presence of a small posteriorly- directed tooth on the posterior part of the penis valve head as another distinguishing character. However this tooth is not present for all individuals (Figs 58 & 59), or even for both penis valves of some individuals. Pittioni (1939b) also described some unbanded individuals from Mongolia that have the yellow pubescence particularly extensive on the sides of the thorax under the name B. maidli. All of these individuals from Kashmir and Mongolia are otherwise closely similar to other B. melanurus, so there is no reason to believe that they are not all parts of the same species. Individuals of the ‘melanurus-Gruppe’ (Tkalcu, 1974a, 1974b) from further east in Mongolia and Gansu (e.g. Skorikov, 1931: fig. 9) usually have a band of black hair between the wing bases. The name B. tschitscherini was applied to these banded individuals by Radoszkowski (1859, 1862). The pale pubescence and the wings are similar in shade to the darker individuals of B. melanurus s.str. The male genitalia (Reinig, 1930: fig. 5; Tkalci, 1974a: figs 28-32) are very similar to those of B. melanurus s.str., but the females have a broader band of punctures in the oculo-ocellar area, there are more micropunctures in the middle of the clypeus, and the pubescence is particularly short and even. Some of the individuals from Mongolia and from further south in the Tibetan massif and the higher Pamir ranges resemble B. tschitscherini in that they have a black band between the wings, although the pale pubescence is lemon-yellow or cream and the wings are nearly clear (subhyaline). These individuals were first described under the name B. difficillimus by Skorikov (resembling Fig. 282). Reinig apparently redescribed this taxon under the name B. melanurus subsp. griseofasciatus, because he subsequently synon- ymised the two (Reinig, 1934). I have not seen any males, but the male genitalia (Reinig, 1930: fig. 5, 63 under the name ‘B. mel. griseofasciatus’) appear to be very similar to those of B. melanurus s.str. The females have the band of punctures in the ocello-ocular area intermediate in breadth between the other two nominal taxa. However, the inner corners of the labral tubercles are usually more narrowly acute so that the sides of the longitudinal median furrow converge throughout its length, the micropunctures are completely absent from the middle of the clypeus, the disto-posterior angle of the mid basitarsus is less strongly produced, and the pubescence is particularly long and uneven. There is considerable and possibly continuous variation among individuals within each of these nominal taxa in the colour pattern, in the shade of the pale pubescence and of the wings, and in the morphological characters. Therefore there is no good evidence as yet from which to suggest that these nominal taxa are not just divergent parts of a single interbreeding population and so parts of a single species. B. melanurus and B. difficillimus have not been found together in the Pamir ranges (Reinig, 1930: fig. 2) or in Kashmir. However, they do occur at low density on opposite sides of at least one of the ridges of the Zanskar mountains (at Chogdo & Nimaling), where B. difficillimus occupies the higher valley. Of course more information is needed concerning interbreeding between them in order to establish their status more definitely. Nests from the Pamir were described by Reinig (1930) and by Bischoff (1931). DISTRIBUTION. B. melanurus is a widespread, but primarily central Asian species (Map 27). It is known from Mongolia (Pittioni, 19396; Skorikov, 1933a; Bischoff, 1936; Tkalct, 1974a; BMNH, NM, PW), the Altai (BMNH), the Tien Shan (Skorikov, 1931; Bischoff, 1936; Panfilov, 1957; BMNH), the Pamir (Skorikov, 1912b, 1931; Reinig, 1930, 1934; Bischoff, 1931; ZI), the Hindu Kush (Reinig, 1940; Richards, 1951; Tkalci, 1969a; BMNH), Pakistan (Frison, 1935; BMNH, PW), Kashmir, Xinjiang [Kunlun Shan] (Morawitz, 1886), Himachal Pradesh (PW), Uttar Pradesh (BMNH), Nepal (Tkalci, 1974b; BMNH, NMS), Sikkim (Richards, 19286, 1930; BMNH), Tibet (Richards, 1928b, 1930; Wang, 1982; BMNH), Qinghai (Morawitz, 1886; Skorikov, 19125; Panfilov, 1957; Tkalct, 1961; Wang, 1982; ZS), Gansu (Bischoff, 1936; MNHU), Inner Mongolia (Skorikov, 1933a) and Shanxi (Yasumatsu, 1951). In the west it is also known from western Iran (BMNH), the Caucasus (Skorikov, 1931), Turkey (Skorikov, 1931; Reinig, 1971; BMNH), Syria (UM) and Lebanon (BMNH). Part of this distribu- tion is mapped by Skorikov (1931: fig. 9). A 64 similar disjunct distribution between the Hindu Kush ranges on the one hand and the Elburz and Armenian highlands on the other is shown by B. keriensis. MATERIAL EXAMINED FROM KASHMIR. B. melanurus is widespread in Kashmir, recorded from the Hindu Raj, Karakoram, Ladakh, Zanskar, Great Himalaya and Pir Panjal ranges, in montane coniferous forest, alpine scrub and steppe and in high, subtropical semi-desert (Map 28, Fig. 6): 85 queens, 186 workers, 135 males, from 29 localities 1800-4800 m (AB, BMNH, FA, MI, NR, PW, RH, ZM). VARIATION WITHIN KASHMIR. Queens and workers are often poorly differentiated in size. Although the colour pattern of this species is apparently stable over large areas of its distribution, there is considerable variation within Kashmir. Queens from the upper forest on Mt Apharwat (2700— 3400 m, 9/9 queens) have the pale pubescence brown rather than yellow (Fig. 288). These queens are all likely to be old, over-wintered individuals because they have very torn wing margins. Young queens seen at Gulmarg, all of the queens from Leh (3500 m, 5/5 queens, Fig. 283), and workers generally (Figs 284, 287, 289), have the pale pubescence slightly paler and more sand-yellow to lemon-yellow. Queens from Nimaling (4800 m, 3/3 queens) have the pale pubescence lemon-yellow to almost cream. These queens, and another from Burzil Chauki, also have well-defined and broad bands of black hairs between the wing bases (Fig. 282, the pattern described under the name B. difficil- limus, Map 28). The presence of a few black hairs on the thoracic dorsum, especially just above the wing bases, is common among workers from Gulmarg. The black hairs sometimes form a diffuse band between the wings, which is particularly obvious for a few, fresh workers from Gulmarg (3/26 workers, Fig. 287), and for a queen from Chittakatha Sar (4600 m, Fig. 286), above the Kishanganga valley. There is also some varia- tion in the extent of the yellow pubescence on the sides of the thorax. This yellow pubescence usually occupies the dorsal half of the area between the wing bases and the leg bases, but may occupy nearly all of this area for some workers and even extends onto the legs in males. This species is distinctive in its appearance at most localities. At high altitudes, workers with a black band between the wing bases (cf. Fig. 282) may resemble B. personatus (Fig. 292, see the comments on B. personatus). At a few localities in the Great Himalaya range, it may be closely similar in colour pattern to some individuals of B. PAULH. WILLIAMS asiaticus that lack red pubescence on the apical terga (e.g. Fig. 278), but it is recognisable by the shiny outer surface of the hind tibia and by the lack of dense, short hair at the base of the hind basitarsus (see the key, couplet 11). FOOD PLANTS. Kashmir: (Ranunculaceae) Aconi- tum hookeri Stapf; (Balsaminaceae) Impatiens glandulifera Royle; (Leguminosae) Trifolium repens L.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC.; (Gentiana- ceae) Swertia petiolata D. Don; (Scrophulariaceae) Verbascum thapsus L., Digitalis lanata Ehrh. [introduced], D. purpurea L. [introduced]; (Labiateae) Stachys sericea Wallich ex Benth., Thymus linearis Benth. ex Benth.; (Polygonaceae) Bistorta viviparia (L.) Gray. Ladakh: (Leguminosae) Medicago falcata L., Caragana versicolor (Wallich) Benth.; (Composi- tae) Echinops cornigerus DC.; (Scrophulariaceae) Verbascum thapsus L.; (Labiateae) Stachys tibetica Vatke, Nepeta podostachys Benth. Bombus (Subterraneobombus) personatus Smith (Figs 60, 100, 140, 180, 291-294, Maps 29 & 30) Bombus personatus Smith, 1879: 132. Lectotype queen by designation [Art. 74b] of Richards (1930: 656), INDIA: Himachal Pradesh, Kinnaur (BMNH) [examined]. Bombus Roborowskyi Morawitz, 1886: 197. Lecto- type queen by designation of Podbolotskaya (in press), CHINA: Qinghai, Burhan Budai Shan, 14000 ft [4300 m] (Prshewalski) (ZI) [exam- ined]. Syn. n. AFFINITIES. Within the subgenus Subterraneo- bombus (see the comments on B. melanurus), B. personatus belongs to a group of species for which the penis valve head is shortened and broadened (Fig. 60) and the interio-basal process of the gonostylus becomes broader distally (Fig. 180) (synapomorphies of B. fedtschenkoi + B. per- sonatus + subterraneus-group; this entire group is the broader, more inclusive ‘subterraneus- Gruppe’ described from other characters by Tkalci, 1974a). However, for B. personatus and B. fedtschenkoi the penis valve head is not as strongly directed ventrally as for members of the subterraneus-group (Fig. 100) (synapomorphy of the subterraneus-group). B. fedtschenkoi is known from the Tien Shan region (Skorikov, 19146; Skorikov, 1931; BMNH). B. fedtschenkoi lacks an apically directed interio- basal process of the male gonostylus (autapo- morphy; see Reinig, 1930: fig. 5; Skorikov, 1931: THE BUMBLE BEES OF THE KASHMIR HIMALAYA fig. 35). B. personatus, like B. melanurus, has no large punctures in the middle of the clypeus, but there are more micropunctures. These punctures are not nearly as large as those of B. fragrans, B. amurensis, or B. fedtschenkoi. The oculo-malar distance of females of B. personatus is relatively longer than that of even the larger B. melanurus (see the comments on B. melanurus) and has fewer punctures. Morawitz described queens from Qinghai under the name B. roborowskyi. The lectotype is closely similar to the lectotype of B. personatus and is believed to be part of the same species. It is possible that B. personatus may even be conspeci- fic with B. difficillimus var. pamirus Skorikov from the Pamir ranges (Skorikov, 1912b, 1931), although the oculo-malar distance is apparently shorter for this nominal taxon (for a discussion of the identity of this nominal taxon, see also Reinig, 1930, 1934; and the comments on B. oberti). I have seen no material of B. difficillimus var. pamirus. DISTRIBUTION. B. personatus is a Tibetan species (Map 29). It is known from Kashmir, Himachal Pradesh (Smith, 1879; BMNH), Tibet (Richards, 1930; Wang, 1982; BMNH), Qinghai (Morawitz, 1886; Panfilov, 1957; Tkalci, 1961; Wang, 1982; ZI, ZS) and Gansu (Morawitz, 1890). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. personatus is recorded from the Zanskar ranges, in the high, dry alpine steppe (Map 30): 12 queens, 2 workers, 2 males, from 2 localities 3500-4800 m (BMNH, FA, PW). I found what could have been the site of a nest on Nimaling plain (4800 m, 23.ix.1986). A queen was seen to fly from, and later to return to, a hole at the base of a dry-stone wall in a shepherd’s summer hut. VARIATION WITHIN KASHMIR. The pale pubescence of the thorax and of terga I-II varies from lemon- yellow (only on terga I-II) to pale sand-yellow to almost cream. For queens and males, the pubes- cence of terga ITI-V has many cream-white hairs along the posterior margin (Figs 291, 293, 294), but pale hairs may be very few in this position in workers (Fig. 292). This species is usually distinctive in its appear- ance, although workers (Fig. 292) may resemble the individuals of B. melanurus with a black band between the wing bases, which are known to occur with them (cf. Fig. 282). These dark workers of B. personatus can be recognised by their pale hairs on the leg bases and on the sterna. FOOD PLANTS. (Leguminosae) Caragana versicolor (Wallich) Benth.; (Labiateae) Stachys tibetica Vatke. 65 Subgenus ALPIGENOBOMBUS Skorikov Alpigenobombus Skorikov, 1914a: 128. Type species: Alpigenobombus __ pulcherrimus Skorikov, 1914a: 128 = Bombus kashmirensis Friese, by present designation. Mastrucatobombus Kriiger, 1917: 66 (as a sub- genus of Bombus Latreille). Type species: Bombus mastrucatus Gerstaecker, 1869: 326 = Bombus wurflenii Radoszkowski, by monotypy. Synonymised with Alpigenobombus Skorikov by Richards (19285). Alpigenobombus Skorikov; Frison, 1927: 64 (as a subgenus of Bombus Latreille). [Nobilibombus Skorikov, 1933a: 62. Published without fixation of type species, unavailable name (Art. 13b).] [Nobilibombus Bischoff, 1936: 12 (as a sub- genus of Bombus Latreille). Type species: Nobilibombus morawitziides Skorikov, 1933a: 62 = Bombus nobilis Friese, by monotypy. Published as a junior synonym of Alpigenobom- bus Skorikov, unavailable name (Art. 11e).] Alpigenibombus Skorikov; Skorikov, 1938b: 1. Unjustified emendation (Art. 33b(i, iii)). [Nobilibombus Milliron, 1961: 54 (as a subgenus of Pyrobombus Dalla Torre). Type species: Bombus nobilis Friese, 1905: 513 [cited as Bombus nobilis Skorikov], by original designa- tion. Published as a junior synonym of Pyro- bombus Dalla Torre, unavailable name (Art. lle).] Nobilibombus Richards, 1968: 222 (as a subgenus of Bombus Latreille) (see Note 2 below). Type species: Bombus nobilis Friese, 1905: 513, by original designation. Syn. n. Note 1. Skorikov (1914a) originally included ’B. lefebvrei Lep. (= B. mastrucatus auct.)’ and A. pulcherrimus Skorikov in his genus Alpigeno- bombus. Unfortunately he did not clearly and unambiguously designate a type species (Art. 67c) and B. alpigenus cannot be the type species of Alpigenobombus by absolute tautonomy. Frison’s (1927: 64) subsequent designation of Bombus alpigenus Morawitz as type species is invalid because this species was not specifically included by Skorikov in the original publication (Art. 67g). Skorikov’s original reference to B. lefeb- vrei Lepeletier is a misidentification, because Lepeletier’s holotype belongs to a species (oldest available name B. pomorum (Panzer)) of the distantly related subgenus Rhodobombus Dalla Torre (Lgken, 1973), whereas Skorikov cites B. mastrucatus auct. (conspecific with B. mastrucatus Gerstaecker, for which the oldest available name is B. wurflenii Radoszkowski) as a synonym. 66 Designation of Skorikov’s Alpigenobombus lefeb- vrei as the type species would require the ICZN to decide the identity of this nominal taxon (Art. 70b). Therefore I designate Alpigenobombus pulcherrimus Skorikov as the type species of Alpigenobombus Skorikov because its identity is clear and this fixation maintains the stability of usage of Alpigenobombus Skorikov. Note 2. Skorikov (1933a) originally included Nobilibombus nobilis (Friese), Nobilibombus validus (Friese) and Nobilibombus morawitziides Skorikov in his genus Nobilibombus, but did not designate a type species. Richards (1968) consi- dered B. nobilis Friese to be the type species of Nobilibombus Skorikov by virtual tautonymy, although this is not sufficient under the present code, which requires absolute tautonymy (Art. 68e). However, Richards’s reference to B. nobilis Friese as the type species is sufficient as an original designation for the subgenus Nobilibombus Richards (Art. 68b). Unfortunately no putative syntype of B. nobilis has yet been found that agrees with the original description in having four or five teeth on the mandibles (see the comments on B. pyrosoma; Richards, 1968: 222). Nonethe- less the identity of B. nobilis is not in doubt, so the designation of a neotype is neither necessary nor valid (Art. 75b). For a general description of both sexes see the descriptions of the subgenera Alpigenobombus and Nobilibombus by Richards (1968). These two groups were originally distinguished primarily on the basis of the greater oculo-malar distance for the individuals ascribed to Nobilibombus. I agree with Bischoff (1936) that placing all the species together in a single group, by considering Nobilibombus Richards as a junior subjective synonym of Alpigenobombus Skorikov, empha- sises their many similarities (e.g. Ito, 1985; Williams, 1985). Bombus (Alpigenobombus) kashmirensis Friese (Figs 25, 29, 33, 37, 61-63, 101-103, 141-143, 181-183, 295-310, Maps 31 & 32) Bombus mastrucatus var. kashmirensis Friese, 1909 [September, see Tkalct, 1974b]: 673 [redescribed by Friese & Wagner, 1910: 47]. Lectotype queen by designation of Tkalcu (1974b: 327), INDIA: Kashmir, 8000-9000 ft [2400-2700 m], vi.1901 (Nurse) (MNHU) [examined]. Bombus mastrucatus var. stramineus Friese, 1909: 673 [redescribed by Friese & Wagner, 1910: 47]. Type worker presumed lost (Tkalcu, 1974b: 327), INDIA: Kashmir, no further data PAULH. WILLIAMS [not seen]. Synonymised with Alpigenobombus kashmirensis (Friese) by Tkalct (1974b) (see Note 1 below). Bombus tetrachromus Cockerell, 1909 [November, see Tkalci, 19746]: 397. Holotype queen by original designation, PAKISTAN: Baltistan (BMNH) [examined]. Synonymised with AI- pigenobombus kashmirensis (Friese) by Tkalca (1974b). Alpigenobombus pulcherrimus Skorikov, 1914a: 128. Holotype queen by monotypy (see Note 3 below), INDIA: Kashmir, Zoji La, above 3000 m, 12-15.vi.1912 (Jacobson) (ZI) [not seen]. Synonymised with Bombus tetrachromus Cockerell by Richards (1930). [Alpigenobombus (Alpigenobombus) beresovskii Skorikov, 1922: 156. Published without descrip- tion or indication, unavailable name (Art. 12a).] Alpigenobombus (Alpigenobombus) kashmirensis (Friese); Skorikov, 1922: 156 Bombus (Mastrucatobombus) mastrucatus subsp. meinertzhageni Richards, 1928b: 335. Holotype queen by monotypy (see Note 4 below), INDIA: Kashmir, Ladakh, Chushul, 1925 (Meinertz- hagen) (BMNH) [examined]. Synonymised with Alpigenobombus kashmirensis (Friese) by Tkalci (19745). [Bombus (Alpigenobombus) tetrachromus vat. albohirtus Richards, 1930: 637 [examined]. Infrasubspecific (see Note 2 below) (Art. 45g(i1) (1)), unavailable name (Art. 45e).] [[Alpigenobombus] beresovskii Skorikov, 1931: 204. Published without description, unavailable name (Art. 13a).] Alpigenobombus_ beresovskii Skorikov, 1933b: 248. Lectotype by designation of Podbolotskaya (in press) [not seen]. Paralectotype queen, CHINA: Sichuan, ‘Yach Zhou’, 28.i1i—S.iv. 1893 (Potanin) (ZI) [examined]. Syn. n. NOMENCLATURE. Note 1. Following the Principle of the First Reviser (Art. 24), B. kashmirensis is accepted in precedence to B. stramineus by the action of Skorikov (1922). Note 2. Richards’s use (1930: 634 etc.) of the term ‘Subsp.’ elsewhere in his paper is taken to indicate that infrasubspecific rank is meant for B. tetrachromus var. albohirtus. No subsequent author has either adopted the name albohirtus for a taxon in the species group or regarded it as a senior homonym, so Richards’s taxon is deemed to be of infrasubspecific status (Art. 45g(ii)(1)). TYPE MATERIAL. Note 3. Skorikov’s description of Alpigenobombus pulcherrimus s.str. specifies that only a single queen was available. If a single queen with the appropriate data can be found then THE BUMBLE BEES OF THE KASHMIR HIMALAYA this should be regarded as the holotype (Art. 73a(il)). Note 4. Richards described B. mastrucatus meinertzhageni from ‘One female from Shusal’ [= Chushul]. A single queen in the BMNH collection agrees with the original description and carries the data quoted. It also bears a label ‘B. tetrachromus / var. meinertz- / -hageni, / Type. Richards’, in handwriting identical to that of Richards. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. A dorso-ventrally flattened and sickle- shaped head of the penis valve is shared by a large group of bumble bees, which in Kashmir includes the species of the subgenera Alpigenobombus, Pyrobombus, Sibiricobombus, and Melanobom- bus (Figs 61—71, 76-85) (synapomorphy, second- arily much modified for Bombus s.str., Figs 72-75). Females of species of the subgenus Al- pigenobombus share the development of six teeth on the distal margin of the mandible (Fig. 33, ?synapomorphy). Within the subgenus Alpigenobombus, females of the breviceps-group share a particularly pro- nounced development of the two pre-apical teeth of the mandible, which are acute, and the adjacent incisura is deeper than its own width (?synapo- morphy of the breviceps-group). This group includes B. grahami (Frison), B. genalis Friese and B. breviceps. The males of the Chinese and east Himalayan (Richards, 1930; Frison, 1935; Skorikov, 1938b; Wang, 1987, 1988; BMNH, IZ, PW) B. genalis are not known. The males of B. breviceps have the gonostylus reduced apically to a transverse band, although this retains a strongly produced, interio-basal process and a strongly produced interio-apical corner (see Frison, 1935: fig. 5). The recurved hook of the penis valve head is short, as for B. nobilis, but more broadened. The males of the Chinese and east Himalayan (Frison, 1933, 1935; Wang, 1982; BMNH) B. grahami, which I have not seen, are apparently intermediate in these character states between B. nobilis and B. breviceps (Frison, 1935: fig. 6). The Chinese B. breviceps Smith is very similar in morphology to the west Chinese and Himalayan B. dentatus Handlirsch, which has the orange pubescence of the thoracic dorsum replaced by black. Individuals with intermediate colour patterns have been described (see Tkalct, 1968b). There- fore these nominal taxa are likely to be parts of a single species that is widespread from China throughout much of South East Asia and the Himalaya (records in e.g. Smith, 1852a; Frison, 1935; Bischoff, 1936; Skorikov, 19385; Tkalct, 67 1960, 1968b; Sakagami, 1972; Wang, 1987; BMNH, MCSN, MNHN, PW, SEMK, UM). This species, which resembles B. trifasciatus and B. haemorrhoidalis in colour pattern (see the com- ments on these species and Fig. 11), may also be present in the foothills of south-eastern Kashmir (oldest available name for B. orichalceus Friese, unconfirmed record from Kashmir by Skorikov, 1933b; there is also a queen in the BMNH from Dalhousie, just across the border from Kashmir in Himachal Pradesh). The females of B. wurflenii, B. kashmirensis and B. nobilis have the incisura wider than deep and the two pre-apical teeth of the mandible are weak and obtuse (Fig. 33, ?plesiomorphic among species of the subgenus Alpigenobombus). Males of these species have the interio-apical process of the volsella reduced (Figs 141-143) (?plesiomorphic among species of the subgenus Alpigenobombus). The males of B. nobilis can be distinguished from those of B. kashmirensis by their shorter gono- stylus, which has the apical margin concave rather than straight or convex (Figs 181-183), and by the shorter recurved hook of the penis valve head (these character states appear to be intermediate between B. kashmirensis and B. breviceps). The females of B. nobilis have the oculo-malar dis- tance just greater than the breadth of the man- dible at its base, rather than distinctly shorter than this breadth as for B. kashmirensis and B. wurflenii. I have seen material of the banded yellow B. nobilis Friese from Sichuan (BMNH, PW), Yunnan, northern Burma, south-eastern Tibet and Sikkim (BMNH, UM), Nepal (PW) and Uttar Pradesh (BMNH). B. validus Friese (synony- mised with Nobilibombus morawitziides Skorikov by Tkalci, 1987), which has the pale pubescence of the thorax grey-white rather than yellow, is otherwise closely similar to B. nobilis and may be part of the same species. I have seen material of this nominal taxon from Gansu (ZI), northern Burma, south-eastern Tibet and Sikkim (BMNH) and Nepal (PW). Females of two further nominal taxa from Qinghai, Tibet and Yunnan (B. xizangensis and B. chayaensis), both very similar to B. nobilis, have been described by Wang (1979, 1982, 1987, 1988) from slight differences in the colour pattern (not seen). Apart from species of the breviceps-group and B. nobilis, the subgenus Alpigenobombus may otherwise include just two species that show great variation in the colour patterns of the pubescence. B. wurflenii Radoszkowski [the correct original spelling of B. wurfleini of authors] is a European species that reaches eastwards as far as the Caucasus and the Urals (e.g. Reinig & Rasmont, 1988: fig. 1; BMNH). The females of B. wurflenii are closely 68 similar to those of B. kashmirensis in morphology. The clypeus of B. wurflenii has widely-spaced, moderate or small punctures throughout, whereas for B. kashmirensis the punctures are slightly more widely-spaced in the middle of the clypeus, so that there is an almost unpunctured, shining area. However, the males can be separated more easily because, from the anterior aspect, the eyes of B. wurflenii (not noticeably enlarged relative to the female) are at their broadest in their ventral half, whereas the (enlarged) eyes of B. kashmiren- sis are broadest in their dorsal half. The genitalia of males of B. kashmirensis and of B. wurflenii are closely similar, but the recurved hook of the penis valve head is more narrowed to a point for B. wurflenii and the apex of the gonocoxite is narrower. Differences in eye size between males of B. kashmirensis and B. wurflenii are associated with differences in mate-searching behaviour (see the introduction on male mate-searching behaviour). B. wurflenii is widespread in Europe among the upper montane forests, where males patrol cir- cuits of scent-marked sites within the forests to find mates (pers. obs.). B. kashmirensis is wide- spread in the Himalaya and Tibet among the high alpine zones. On Mt Apharwat in Kashmir, the males hover near the peak and watch for potential mates. Individuals from these populations of B. wurflenii and B. kashmirensis would be unlikely to interbreed even if they were to occur on the same mountain, because they appear to search for mates in different ways and in different kinds of habitat (cf. comments on B. rufofasciatus, B. asiaticus and B. hypnorum). The queen from Ladakh described by Richards (19285) under the name B. mastrucatus subsp. meinertzhageni differs from B. kashmirensis s.str. principally in that the pale pubescence is entirely yellow (resembling Fig. 295), rather than both white and yellow (Fig. 302). The male genitalia of the bees of this group from Kashmir show limited variation, although this is more pronounced in the shape of the gonostylus (Figs 181-183). The apex of the gonostylus is formed from a thin plate, which has an irregular margin. The variation in the extent of this plate is not associated with the variation in colour pattern and appears to be of only individual significance. The yellow males are otherwise closely similar to those of B. kashmiren- sis S.str., even in the size and shape of their eyes. All of the workers collected at Gumri near the Zoji La in the Great Himalaya range are inter- mediate between the B. mastrucatus subsp. meinertzhageni of Ladakh and the B. kashmiren- sis s.str. from the mountains around the Vale of Kashmir in that they have the pale pubescence of PAULH. WILLIAMS the thorax grey-white and that of both abdominal terga I-II bright yellow (Fig. 301). Skorikov described females from Sichuan that have the pale pubescence reduced to a few grey- cream hairs mixed with the black pubescence on the anterior thoracic dorsum under the name Alpigenobombus_ beresovskii. The pattern of punctures on the female clypeus does not associ- ate them definitely with either B. wurflenii or B. kashmirensis, but Skorikov’s description of the male genitalia (Skorikov, 1933b: 248) agrees more closely with B. kashmirensis. From the meagre evidence available at present, I consider that Alpigenobombus beresovskii is most likely to be part of the same species as B. kashmirensis. DISTRIBUTION. B. kashmirensis is primarily a Tibetan species (Map 31). It is known from Gansu (Morawitz, 1880; Bischoff, 1936; PW, ZS), Qinghai (Morawitz, 1886; Bischoff, 1936; Wang, 1982; PW, ZS), Sichuan (Morawitz, 1890; Skorikov, 1933b; Wang, 1982; ZI), Guangxi [no precise locality cited] (Wang, 1982), Tibet (Richards, 1930; Wang, 1982; BMNH, PW), Nepal (Tkalci, 1974b; NMS), Uttar Pradesh (BMNH), Himachal Pradesh (BMNH, PW) and Kashmir. A record from Mongolia (MNHU) needs to be confirmed. MATERIAL EXAMINED FROM KASHMIR. B. kash- mirensis is widespread in Kashmir, recorded from the Hindu Raj, Ladakh, Zanskar, Great Himalaya and Pir Panjal ranges, in alpine scrub and steppe (Map 32, Fig. 6): 34 queens, 172 workers, 67 males, from 15 localities 2400-4800 m (BMNH, FA, ITZ, MI, MNHN, MNHU, NR, PW, TL; 1 worker photographed in the Sangisfaid valley, C. A. Chadwell). A colony was found at the western end of Nimaling plain (4800 m, 15.vii.1980). The entrance to the nest was under a 2 m boulder in the derelict terminal moraine. VARIATION WITHIN KASHMIR. Individuals from the Ladakh and Zanskar ranges have the pale pubes- cence yellow, with only a narrow band of black hairs basally on tergum III (Figs 295-300). The pubescence of terga III-VI is orange-red, with at most only a few hairs of terga V-VI pale-tipped. Workers from Gumri in the Great Himalaya range are similar, but the pale hairs of the thorax are grey-white rather than yellow (terga I-II yellow, or 1/4 workers with some grey-white hair antero-laterally on tergum I) and the orange-red hairs of terga IV-VI are white-tipped (Fig. 301). Males and females from the Hindu Raj range are similar, but the pale hairs of tergum I are also grey-white (tergum II yellow). Material from around the Kishanganga valley includes individuals with THE BUMBLE BEES OF THE KASHMIR HIMALAYA both yellow and white pubescence (4/10 queens, 16/123 workers, 10/43 males), and some individuals for which the yellow is completely replaced by white, even on tergum II (6/10 queens, 107/123 workers, 33/43 males). On Mt Apharwat in the Pir Panjal range, the yellow and white individuals predominate (9/10 queens, 24/27 workers, 16/17 males). Another worker from Apharwat has very few pale hairs on the thorax and many black hairs intermixed on terga I-III (cf. B. beresovskii). The yellow and white individuals usually have a dis- tinct black band at the base of tergum III that may extend over most of tergum III for queens and males (Figs 302-304). These queens have the yellow very pale and the white-tipped hairs of terga III-VI are only orange-red at their bases. In contrast, the white individuals often have few black hairs on tergum III, the hairs of terga III-VI are often completely orange-red and the white pubescence at the rear of the thorax may be almost completely replaced by black hairs (Figs 305-310). The yellow individuals of this species (Figs 295-300) from Ladakh are closely similar in colour pattern to B. oberti (Figs 392 & 393), which is known to occur with them. These individuals may also resemble some of the B. himalayanus (Figs 242 & 243) and B. ladakhensis (Figs 394 & 396) that occur with them. B. kashmirensis can be recognised by its 6-toothed mandibles (Fig. 32). The white and yellow individuals (Figs 302-304) from the mountains around the Vale of Kashmir are particularly closely similar in colour pattern to the B. rufofasciatus (Figs 434436, 438, 439) that occur with them. B. kashmirensis can be recog- nised by the longer and white-tipped, red hairs on terga III-VI, where the hairs of B. rufofasciatus are short and either red or white. These indivi- duals may also resemble some B. himalayanus (Figs 244-252), B. novus males (Figs 260 & 262), B. lemniscatus (Figs 325 & 326), B. lepidus (Figs 330, 331, 333, 334) and B. biroi (Figs 338 & 339). The white individuals of this species (Figs 305— 310) from the mountains around the Kishanganga valley and the Vale of Kashmir are particularly closely similar in colour pattern to some of the B. biroi (Figs 341 & 342) that occur with them at some localities. These individuals may also resemble some B. avinoviellus (Figs 238-241), B. himalayanus (Figs 250-252), B. tunicatus (Figs 350, 353, 355), B. asiaticus (Figs 386 & 387) and B. keriensis (Figs 408 & 409). B. kashmirensis is easily recognised by its 6-toothed mandibles (Fig. 32): FOOD PLANTS. Kashmir: (Ranunculaceae) Aconi- tum laeve Royle, A. heterophyllum Wallich ex 69 Royle, A. hookeri Stapf, A. violaceum Jacquem. ex Stapf; (Compositae) Lactuca lessertiana (DC.) C. B. Clarke, unidentified dandelion-like compo- sites; (Gentianaceae) Swertia petiolata D. Don; (Scropulariaceae) Scrophularia pauciflora Benth.., Digitalis lanata Ehrh. [introduced], Pedicularis oederi Vahl, P. pectinata Wallich ex Benth., P. punctata Decne. Ladakh: (Leguminosae) Caragana versicolor (Wallich) ex Benth.; (Labiateae) Stachys tibetica Vatke. Subgenus PYROBOMBUS Dalla Torre Pyrobombus Dalla Torre, 1880: 40 (as a subgenus of Bombus Latreille). Type species: Apis hypnorum Linnaeus, 1758: 579 = Bombus hypnorum (Linnaeus), by monotypy. Poecilobombus Dalla Torre, 1882: 23 (as a sub- genus of Bombus Latreille). Type species: Bombus sitkensis Nylander, 1848: 235, by sub- sequent designation of Sandhouse (1943: 589). Synonymised with Pyrobombus Dalla Torre by Milliron (1961). [Pyrrhobombus Dalla Torre; Dalla Torre, 1882: 28 (as a subgenus of Bombus Latreille). In- correct subsequent spelling, unavailable name (Art. 33c).] Pyrrhobombus Dalla Torre; Dalla Torre, 1896: 503 (as a subgenus of Bombus Latreille). Un- justified emendation (Art. 33b(i, iii)). Pratobombus Vogt, 1911: 49 (as a subgenus of Bombus Latreille). Type species: Apis pra- torum Linnaeus, 1761: 424 = Bombus pratorum (Linnaeus), by subsequent designation of Frison (1927: 67). Synonymised with Pyrobombus Dalla Torre by Milliron (1961). Anodontobombus Kriiger, 1917: 61, 65 (as a subgenus of Bombus Latreille). Type species: Apis hypnorum Linnaeus, 1758: 579 = Bombus hypnorum (Linnaeus), by present designation (see Note below). Synonymised with Pyro- bombus Dalla Torre by Milliron (1961). Uncobombus Vogt in Kriger, 1917: 65 (as a subgenus of Bombus Latreille). Type species: Apis hypnorum Linnaeus, 1758: 579 = Bombus hypnorum (Linnaeus), by present designation (see Note below). Synonymised with Pyro- bombus Dalla Torre by Milliron (1961). Lapponicobombus Quilis-Pérez, 1927: 19 (as a subgenus of Bombus Latreille). Type species: Apis lapponica Fabricius, 1793: 318 = Bombus lapponicus (Fabricius), by subsequent desig- nation of Milliron (1961). Synonymised with Pyrobombus Dalla Torre by Milliron (1961). [Hypnorubombus Quilis-Pérez, 1927: 19 (as a 70 subgenus of Bombus Latreille). Incorrect original spelling (Art. 32c(iv)), unavailable name (Art. 19a).] [Laponicobombus Quilis-Pérez, 1927: 63 (as a subgenus of Bombus Latreille). Incorrect original spelling (Art. 32c(iv)), unavailable name (Art. 19a).] Hypnorobombus Quilis-Pérez, 1927: 97 (as a subgenus of Bombus Latreille). Type species: Apis hypnorum Linnaeus, 1758: 579 = Bombus hypnorum (Linnaeus), by monotypy. Synony- mised with Pyrobombus Dalla Torre by Milliron (1961). Pratibombus Vogt; Skorikov, 1938a: 146. Unjus- tified emendation (Art. 33b(i, iii)). Note. The names Anodontobombus Kriger and Uncobombus Vogt were originally pro- posed for a section of the genus Bombus and a group of subgenera respectively, but they must both be treated as subgeneric names (Art. 10e). Milliron (1961: 53) considered Anodontobom- bus Kriiger to be synonymous with his concept of the subgenus Pyrobombus Dalla Torre (with- in his genus Pyrobombus Dalla Torre). Milliron also considered Uncobombus Vogt to ‘corres- pond’ to Pyrobombus Dalla Torre. The type species of both subgenera are designated as Apis hypnorum Linnaeus, in common with Pyrobombus Dalla Torre, in order to stabilise an unnecessarily complex nomenclature. For a general description of both sexes see Richards (1968). His descriptions include many exceptions for the characters of B. festivus Smith, which was subsequently placed in a separate subgenus Festivobombus by Tkalci (1972). Bombus (Pyrobombus) hypnorum (Linnaeus) (Figs 64, 104, 144, 184, 311-316, Maps 33 & 34) Apis hypnorum Linnaeus, 1758: 579. Lectotype queen by designation of Day (1979: 64), SWEDEN: no further data (LSL) [examined]. [Bombus hypcorum (Linnaeus); Walckenaer, 1802: 148. Incorrect subsequent spelling, un- available name (Art. 33c).] Bombus (Pratobombus) hypnorum var. bryorum Richards, 1930: 650 (see Note below). Holo- type queen [not the male] by original designa- tion [p. 651], INDIA: Kashmir, 8000-9000 ft [2400-2700 mJ], vi.1901 (Nurse) (BMNH) [examined]. Syn. n. Bombus fletcheri Richards, 1934: 90. Holotype queen by original designation, INDIA: Kashmir, Gulmarg, 7.vi.1931 (Fletcher) (BMNH) [exam- ined]. Syn. n. PAULH. WILLIAMS Pyrobombus (Pyrobombus) bryorum (Richards); Tkalct, 1974b: 328. NOMENCLATURE. Note. Richards’s use (1930: 634 etc.) of the term ‘Subsp.’ elsewhere in his paper could be taken to indicate that infrasubspecific rank is meant for B. hypnorum var. bryorum, although his comment ‘this Indian form’ does associate the nominal taxon with a particular geographical area (Art. 45f(ii)). However, this has subsequently been treated as a name in the species group by Tkalci (1974b), so Richards’s taxon is therefore deemed to be of subspecific status (Art. 45g(ii)(1)). AFFINITIES. Species of the subgenus Pyrobombus share a reduction and ventral displacement of the interio-basal process of the male gonostylus, with an apically-directed extension that forms a longi- tudinal ridge, just ventral and parallel to the inner margin of the gonostylus (Figs 184-191) (synapo- morphy). An intermediate form of this structure may be seen in the gonostylus of B. (Festivobom- bus) festivus Smith, which has the interio-basal process only partially reduced and rotated ven- trally, but with a weakly sclerotised fold running apically, below the inner margin of the gonosty- lus. The subgenus Pyrobombus has more species (at least 42) than any other subgenus of bumble bees and many of these species are morphologi- cally closely similar. Consequently ideas of the relationships among the species are at present particularly tentative. Himalayan B. hypnorum have often been con- fused with the workers of B. (Festivobombus) festivus, which have a similar colour pattern. B. festivus is common in southern China and the eastern Himalaya to as far west as Himachal Pradesh (BMNH, PW, UM). It may also be present in the foothills of south-eastern Kashmir (unconfirmed record by Skorikov, 1933b). The workers of B. festivus have the disto-posterior angle of the hind tibia spinosely produced, where- as for B. hypnorum this process is no longer than broad. The pubescence of B. festivus is generally shorter and more even, with a particularly dense covering of short, branched hairs on the head, posterior to the eyes. The male genitalia of B. festivus show similarities to those of B. hypnorum, but have a distinct interio-basal process of the gonostylus (see Tkalci, 1974b: fig. 114), lack a well-defined ventro-basal angle of the penis valve (see the comments on B. asiaticus), and have the volsella strongly constricted just before its apex. Within the subgenus Pyrobombus, the rare B. abnormis (Tkalct) from Sikkim (Tkalcu, 19684; BMNH) also resembles B. hypnorum in colour pattern, but with fewer pale hairs on the apical THE BUMBLE BEES OF THE KASHMIR HIMALAYA terga. The ocello-ocular area is, however, almost completely and densely covered with fine punc- tures for both sexes of B. abnormis. The male genitalia differ from those of B. hypnorum in that the volsella is weakly, but distinctly, constricted before its apex (more like B. festivus), of which the outer corner is acute, and the recurved hook of the penis valve head is shorter than for B. hypnorum (more like B. festivus). The interio- apical angle of the gonocoxite of B. abnormis is broadly rounded, rather than acutely pointed, and the shape of the gonostylus differs (see Tkalci, 1968a: fig. 52-56). B. hypnorum may be one of the most plesio- morphic species of the subgenus Pyrobombus, at least insofar as the the gonostylus retains a well- defined inner basal notch (Fig. 184) and the recurved hook of the penis valve head remains broad but narrowed to a rounded point (Fig. 64). At present I am not aware of any synapomorphic characters from which to recognise a mono- phyletic ‘hypnorum-group’ (cf. the ‘hypnorum- Gruppe’ of Tkalci, 19682, 1989). Many other Old World species of the subgenus Pyrobombus have the recurved hook broadened apically and more spatula-shaped (e.g. the pratorum-group, see the comments on B. biroi, Figs 70 & 71), or more or less shortened (e.g. the parthenius-group, see the comments on B. lepidus, Figs 67 & 68). B. hypnorum is closely similar to the North American B. perplexus Cresson. In the Old World, its closest relative may be B. haematurus Kriechbaumer, which is known from south- eastern Europe, Turkey and the Caspian coast of Iran (D. B. Baker, pers. comm.; BMNH, PW). Females of B. haematurus have the posterior half of the thoracic dorsum and tergum I black-haired and the post ocello-ocular area is more smooth and more finely punctured than for B. hypnorum. The distal margin of the male gonocoxite is more nearly at a right angle to its longitudinal axis for B. haematurus (se Kriiger, 1943: p. 540 figs 1-12) and the gonostylus is shorter than for B. hyp- norum (see Krier, 1943: p. 538 figs 1-7). B. hypnorum appears to be widespread between northern Europe and Kashmir. Both females and males differ slightly in colour pattern between the two regions (by extension of the area of pale pubescence onto at least tergum II among indivi- duals from Asia; see Reinig, 1939: fig. 8). Indivi- duals from the Himalaya have been regarded as a separate species, B. bryorum, by Tkalct (1974b), although I can find no reliable morphological differences. The queen described under the name B. fletcheri by Richards is closely similar to other faded queens of B. hypnorum from Gulmarg. Individuals of all of these nominal taxa occupy 71 similar habitats in the upper montane coniferous forest. The males search for mates by patrolling the forest in a similar way and I know of no evidence to suggest that individuals from Kashmir would not interbreed with those from Europe if they were to occur on the same mountain (cf. comments on B. asiaticus, B. rufofasciatus and B. kashmirensis). Therefore they are all considered to be parts of the same species. DISTRIBUTION. B. hypnorum is widespread in the Old World (Map 33). It is known from Europe (but not Britain e.g. Reinig, 1939; Lgken, 1973) and eastwards across the U.S.S.R. (Reinig, 1939; Panfilov, 1984; BMNH) to as far as Kamchatka (Bischoff, 1930; Reinig, 1939; BMNH) and Hokkaido in Japan (Tkalct, 1965; Sakagami & Ishikawa, 1969; BMNH, PW). In central Asia it reaches southwards to the Caucasus and the Altai (Panfilov, 1984) and Mongolia (Skorikov, 1931, 1933a; Reinig, 1939; PW); and in the south-east it is recorded from North Korea (Kim & Ito, 1987), Gansu (Bischoff, 1936; Reinig, 1939), Sichuan (Morawitz, 1890), Yunnan (Wang, 1987), Burma and Arunachal Pradesh (BMNH), Tibet (Richards, 1930; Wang, 1987), Sikkim (Friese, 1918), Nepal (Tkalct, 19746; BMNH, NMS) and Kashmir. Part of this distribution is mapped by Reinig (1939: fig. 8) and also by Panfilov (1984: map 186). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. hypnorum is recorded from the Hindu Raj range, from the southern side of the Great Himalaya range and from the Pir Panjal range, in the upper montane coniferous forest (Map 34, Fig. 6): 6 queens, 43 workers, 22 males, from 5 localities 2400-3700 m (BMNH, NR, PW). VARIATION WITHIN KASHMIR. The colour pattern of the females resembles that of individuals from northern Asia, in that pale brown hairs often cover most of tergum II (Figs 311, 314, 315). Most workers have some pale hairs intermixed on tergum III as well, and extreme workers from Mt Apharwat have just a few black hairs intermixed on terga III-IV (Fig. 312, 5/26 workers). On the thorax the pale brown pubescence extends down to the leg bases and femora in most individuals. The males are much lighter in colour pattern than those from Europe (but see the discussion of the colour patterns of the Kashmir fauna). Their pubescence is almost uniformly lemon yellow, with the exception of a few black hairs intermixed on the head and on terga IV-V, and of white hairs on terga V-VII (Figs 313 & 316). Females of this species are distinctive in their appearance. The yellow males with white hairs on the apical terga (Figs 313 & 316) may be similar in 2 colour pattern to the males of B. lucorum (Figs 365 & 370), which occur with them at some localities. B. hypnorum males can be recognised by their smaller size and by the inwardly hooked apices of the penis valves (Fig. 64). Host PLANTS. (Ranunculaceae) Cimicifuga foetida L., Aconitum laeve Royle; (Balsaminaceae) Impatiens glandulifera Royle, unidentified small yellow balsam; (Sambucaceae) Lonicera quinque- locularis Hardw.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC.; (Scrophulari- aceae) Scrophularia pauciflora Benth., Digitalis lanata Ehrh. [introduced]; (Labiateae) Prunella vulgaris L., Salvia hians Royle ex Benth. Bombus (Pyrobombus) subtypicus (Skorikov), comb. n. (Figs 41, 65, 66, 105, 106, 145, 146, 185, 186, 216, 317-324, Maps 35 & 36) Bombus leucopygus Morawitz in Fedtschenko, 1875: 3. Lectotype female by designation of Skorikov (1914c: 293) (see Note 2 below), U.S.S.R.: Tadzhikistan $.S.R. or Uzbekistan S.S.R. (ZI) [not seen]. Junior primary homo- nym of Bombus leucopygus Illiger, 1806 [= B. hypnorum (Linnaeus)]. [Pratobombus leucopygos Skorikov, 1914c: 293. Incorrect subsequent spelling, unavailable name (Art. 33c).] Pratobombus leucopygos var. subtypicus Skorikov, 1914c: 294. Lectotype by designation of Pod- bolotskaya (in press) [not seen]. Paralectotype queen, U.S.S.R.: ?Uzbekistan S.S.R., ‘Bukan- Khapa’, 27.v.1911 (Topbek) (ZI) [examined]. Bombus leucurus Bischoff & Hedicke, 1931: 391. Unnecessary replacement name for B. leucopy- gus Morawitz. Synonymised with Pyrobombus subtypicus (Skorikov) by Tkalcti (1968a). Pyrobombus (Pyrobombus) subtypicus (Skorikov); Tkalct,, 1968a: 27 (see Note 1 below). Pyrobombus (Pyrobombus) kohistanensis Tkalci, 1989: 49. Holotype queen by original designa- tion, PAKISTAN: Indus Kohistan, Kagan valley, Naran, 2400-3000 m, 3—13.vi.1977 (de Freina) (ZS) [examined]. Syn. n. NOMENCLATURE. Note 1. Bombus subtypicus (Skorikov) replaces Bombus leucopygus Morawitz because it is the oldest available name (Art. 60b) (Tkalcti, 1968a, 1969a). TYPE MATERIAL. Note 2. Skorikov (1914c) des- cribed the variation among Morawitz’s syntypes of B. leucopygus in the ZMMU. He stated that he considered that the first specimen bearing a large label with the name of the species was the type. PAULH. WILLIAMS This use of the word type in reference to a single specimen is taken to be a valid lectotype designa- tion (Art. 74b). AFFINITIES. Within the subgenus Pyrobombus (see the comments on B. hypnorum), close rela- tives of B. subtypicus, as judged from the similar- ity in the form of the male genitalia, may include both B. haematurus and B. lemniscatus (see the comments on B. hypnorum and B. lemniscatus). However, for B. subtypicus the recurved hook of the penis valve head is more acutely pointed (Figs 65 & 66) than for either of the other two species (e.g. Fig. 67) (see Kriiger, 1943: p. 541 figs 1-11). Unlike B. haematurus, B. subtypicus has virtually no interio-basal notch on the gonostylus above the interio-apical corner of the gonocoxite near their point of contact (Figs 185 & 186). The females of B. haematurus can be distinguished by their long pubescence and by their many close and regular fine punctures in the post ocello-ocular area of the head. Skorikov (1914c) described the local variation in the colour pattern of the pubescence among Morawitz’s series of B. leucopygus Morawitz from ‘Turkestan’. Some of these individuals had the black band between the wing bases partially or completely replaced by yellow pubescence. Recently Tkalci (1989) described a single queen under the name Pyrobombus kohistanensis from across the Pakistan border in the Indus- Kohistan range that has a colour pattern similar to the banded grey- white individuals of B. avinoviel- lus. This specimen (Tkalcu, 1989: figs 14, 18, 22) is closely similar in morphology to B. subtypicus and is likely to be part of the same species. It differs from individuals of B. subtypicus from Kashmir that have a black band between the wings prin- cipally in that the pale pubescence is nearly white, rather than yellow. However, a single queen from nearby on the Kashmir side of the border at Astor is intermediate in these characters insofar as it has both black hairs intermixed between the wing bases and the pale pubescence is a particularly pale yellow. More specimens, especially of the males, are needed in order to establish definitively the relationships of this nominal taxon. DISTRIBUTION. B. subtypicus is a central Asian species (Map 35). It is known from the Tien Shan (Morawitz, 1880; Skorikov, 1914c, 1931; Kriiger, 1943; BMNH, Z1), the Pamir (Reinig 1930, 1934; Skorikov, 1931; BMNH), the Hindu Kush (Reinig, 1934, 1940; Richards, 1951; Tkalct, 1969a; BMNH, PW), Pakistan (Tkalci, 1989; BMNH, PW, ZS) and Kashmir. The record from Gansu (Bischoff, 1936) is based on a misidentification (Tkalci, 19682). THE BUMBLE BEES OF THE KASHMIR HIMALAYA MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. subtypicus is recorded from the Hindu Raj range and from the Great Himalaya range, in dry subalpine scrub (Map 36): 2 queens, 36 workers, 108 males, from 9 localities 2700-3700 m (BMNH, NR, PW). A colony was found near Dras (3200 m, 10. viii. 1986). The nest entrance was at the base of a dry- stone wall. VARIATION WITHIN KASHMIR. Material from Kashmir is relatively homogeneous in colour pattern, at least insofar as all specimens have the pubescence of the apical terga orange-red rather than white (Figs 317-324). Females from the Hindu Raj range and from Matayan and Dras in the Great Himalaya range have yellow hairs on the head and to a more variable extent on terga I-II (Figs 319, 320, 323). These individuals have only a few black hairs on the thoracic dorsum. The single queen from Astor has the head, the ventral parts of the thorax and terga I-II black (Fig. 322). Black hairs are also more numerous between the wing bases, where they form an indistinct band. The yellow pubescence of this individual is particularly pale. The single worker from Lal Pani has a broad, black band between the wing bases (Fig. 317). The hairs of its head are black but those of terga I-II are yellow. Some of the males from the same locality also have a black band between the wing bases (Fig. 318, 10/25 males). The individuals of this species that lack a black band between the wing bases (Figs 319-324) are closely similar to some B. biroi (Figs 335-337), although they are not known to occur together. They are also similar to some of the B. avinoviel- lus (Fig. 232), B. marussinus (Figs 254-257), B. branickii (Fig. 265), B. asiaticus (Figs 371-376) and B. semenovianus (Figs 397-402) that may occur with them at some localities. Unbanded B. subtypicus can be recognised by the characters given in the key (couplets 19 & 20). The indivi- duals with the black thoracic band (Figs 317 & 318) resemble some of the B. avinoviellus (Figs 233-235), B. asiaticus (Figs 377 & 378) and B. keriensis (Fig. 414416) that occur with them. B. subtypicus can be recognised by its small size, by its shiny outer surface to the hind tibia, and by the small number of fine punctures in the ocello- ocular area of the head (Fig. 216). FOOD PLANTS. (Leguminosae) Melilotus officinalis (L.) Pallas, Trifolium repens L.; (Scrophulariaceae) Verbascum thapsus L.; (Labiateae) Nepeta podo- stachys Benth., Prunella vulgaris L. 73 Bombus (Pyrobombus) lemniscatus Skorikov (Figs 15, 67, 107, 147, 187, 212, 218, 223, 325, 326, Maps 37 & 38) Bombus lemniscatus Skorikov, 19126: 606. Holo- type queen by monotypy, CHINA: Qinghai, northern slopes of the Xining mountains, 30.v. 1890 (Grum-Grzhimailo) (ZI) [examined]. Bombus nursei var. flavopilosus Friese, 1918: 84. LECTOTYPE queen by present designation (see the Note below), no data (Bingham) (MNHU) [examined]. Syn. n. Bombus (Lapidariobombus) peralpinus Richards, 1930: 646. Holotype queen [not a worker] by original designation [p. 647], CHINA: Xizang [= Tibet], Rongshar valley, Tasam, 12000 ft [3700 ml], 20.vi.1924 (Hingston) (BMNH) [examined]. Synonymised with Pyrobombus lemniscatus (Skorikov) by Tkalct (19746). TYPE MATERIAL. Note. Tkalct' (1968) described Pyrobombus mirus from a queen that is the only one of Friese’s three syntype queens of B. nursei var. flavopilosus to bear one of the red ‘Type’ labels. Tkalct (1968a: 39) had noted that there are some discrepancies between this indivi- dual and Friese’s original description, but con- trary to Tkalct’s conclusion (‘kann infolge der Nachtiibereinstimmung mit der Urbeschreibung nicht die Lectotype dieses Taxons [B. nursei var. flavopilosus| darstellen’), this does not prevent this specimen from being a syntype (Art. 73b), so it remains a candidate for designation as the lectotype of B. nursei var. flavopilosus. However, another one of Friese’s syntypes is designated here as lectotype of B. nursei var. flavopilosus in order to maintain the current usage of the name B. mirus (Tkalca). A queen in the MNHU collection agrees with the original description of B. nursei var. flavopilosus and carries a label (1) ‘Coll. Bingham’; (2) ‘B. nursei / flavopilosus / [female] 1914 Friese det. / Fr.’; (3) a red printed label ‘Paratypus’; (4) ‘LECTOTYPUS/ B. nursei var. / flavopilosus / Friese Tkalct det [female]’ [designa- tion not published]. This specimen, which is badly damaged with Anthrenus holes and lacks the left front tarsi and the right mid and hind tarsi, is designated as lectotype (Art. 74a). AFFINITIES. Within the subgenus Pyrobombus (see the comments on B. hypnorum), B. lemnis- catus is likely to be related to the parthenius-group of species. All other species of this group, with the possible exception of B. mirus (Tkalct), have the ventro-basal angle of the penis valve displaced closer to the penis valve apex than to its base (Figs 107-109), with the shoulder between this angle and the base of the penis valve head strongly 74 curved, and the head of the penis valve broadened basally (Figs 67-69, 107-109) (synapomorphies). This group includes Tkalcu’s (1974b, 1989) ‘parthenius-Gruppe’ together with the additional species B. lepidus, B. infirmus (Tkalca) and probably B. lemniscatus and B. mirus (see the comments on B. lepidus). There are two more-distantly related species of the subgenus Pyrobombus that occur at lower altitudes in the eastern Himalaya, to at least as far west as Uttar Pradesh (BMNH): B. rotundiceps Friese and B. flavescens. It is possible that B. rotundiceps may also be present in Kashmir (unconfirmed record by Skorikov, 1933b). This species closely resembles the local B. trifasciatus and B. haemorrhoidalis in the colour pattern of the pubescence (see the comments on those species). Females of both B. rotundiceps and B. flavescens from Uttar Pradesh have the pubes- cence of the thoracic dorsum entirely black. The legs are extensively bright orange-brown for B. flavescens but black for B. rotundiceps. Another Himalayan species, which resembles B. lemniscatus more closely and which may belong to the parthenius-group in its broadest sense, is B. mirus. Females of B. mirus have a colour pattern similar to some of the darker, white B. biroi from Kashmir (Fig. 341), but with the pubescence of tergum III predominantly black. They can be distinguished from most species by the profuse punctures that cover more than half of the ocello- ocular area of the head (though not as extensive as for B. abnormis, see the comments on B. hypnorum). Females of B. mirus are known from the alpine zone of Nepal (BMNH, PW), Sikkim (Tkalci, 19684; BMNH, MNHU) and Tibet (Wang, 1982). Wang also described a male from Zayii in Tibet (not seen). This male is apparently similar in its genitalia (Wang, 1982: fig. 32) and in its white and yellow colour pattern to another male from Tibet described by Friese (1913) under the name B. pratorum var. tibetanus (junior secondary homonym in Bombus of Apathus tibetanus Morawitz). The genitalia of both appear to differ from those of B. lemniscatus primarily in their strongly concave interior margin of the gonostylus. B. lemniscatus (redescribed independently under the name B. nursei var. flavopilosus by Friese, see the Note above, and under the name B. peralpinus by Richards) is closely similar to B. lepidus not only in morphology, but also in the colour pattern of the pubescence and in its dis- tribution. Skorikov (19125) distinguished females of B. lemniscatus by their longer pubescence, slightly longer oculo-malar distance, more sharply defined band of black hairs between the wing PAULH. WILLIAMS bases and by their generally more extensive black pubescence. Tkalct has listed several other subtle differences between the females (Tkalci, 1989: table 3) and noted that the scutum, on the thoracic dorsum, is punctured almost throughout for B. lemniscatus (Fig. 223) but broadly unpunctured in the middle for B. lepidus (Fig. 224). A single male in the BMNH collection from ‘Rata, 11,000 ft.’ [= 3350 m], Uttar Pradesh, shows similar hair length, colour pattern and sculpture of the scutum to females of B. lemniscatus. Although the genitalia of this male are closely similar to those of B. lepidus, they do differ in that, unlike B. lepidus from Kashmir, the recurved head of the penis valve is slightly narrower (Fig. 67), the ventro- basal angle of the penis valve is just closer to the base of the penis valve shaft than to the penis valve head (Fig. 107), and the interio-apical process of the volsella is broader and extends without a concave angle to the exterio-apical corner of the volsella (Fig. 147). All of the available males of B. lepidus can be distinguished by their extensively yellow colour pattern of the pubescence, which has few white or black hairs (see the discussion of convergent colour patterns). More material, especially of the males, of B. mirus, B. lemnis- catus and B. lepidus is needed from Nepal, where all three species may occur in proximity, in order to clarify their relationships. DISTRIBUTION. B. lemniscatus is primarily a Tibetan species (Map 37). It is known from Gansu (Bischoff, 1936), Qinghai (Skorikov, 19126; PW, ZI, ZS), Sichuan (BMNH, PW), Tibet (Richards, 1930; Wang, 1982, 1988; BMNH), Sikkim (BMNH, MNHU), Nepal and Uttar Pradesh (BMNH) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. lemniscatus is recorded only from Mt Apharwat in the Pir Panjal range, from above the montane coniferous forest (Map 38): 2 workers, Apharwat, above 3000 m, viii-ix.1986 (Williams) (PW). VARIATION WITHIN KASHMIR. The workers from Mt Apharwat differ from one another in the extent of the pale pubescence on the scutellum, at the rear of the thoracic dorsum, and on tergum III (Figs 325 & 326). This species is particularly closely similar in colour pattern to some of the white and yellow B. lepidus (Figs 330 & 334) and B. biroi (Figs 338 & 339) that occur with it. They also resemble some of the B. himalayanus (Figs 247-252), B. novus males (Figs 260 & 262), B. kashmirensis (Figs 302-304) and B. rufofasciatus (Figs 434-436, 438, 439) from the same area. B. lemniscatus can be recognised by the combination of the dense patch THE BUMBLE BEES OF THE KASHMIR HIMALAYA of small punctures in the ocello-ocular area of the head (Fig. 218) together with the very narrowly unpunctured median area of the thoracic scutum (Fig. 223; see the key, couplets 19 & 21). FOOD PLANTS. No records. Bombus (Pyrobombus) lepidus Skorikov (Figs 68, 69, 108, 109, 148, 149, 188, 189, 213, 219, 224, 327-334, Maps 39 & 40) Bombus lepidus Skorikov, 1912b: 606. Lectotype by designation of Podbolotskaya (in press) [not seen]. Paralectotype queen, CHINA: Qinghai, northern slopes of the Xining mountains, 30.v. 1890 (Grum-Grzhimailo) (ZI) [examined]. Bombus genitalis Friese, 1913: 85. Holotype male by monotypy (see Note 1 below), CHINA: Tibet, no further data (Staudinger) (MNHU) [examined]. Syn. n. Bombus nursei var. tetrachromus Friese, 1918: 85. LECTOTYPE queen by present designation (see Note 2 below), INDIA: Kashmir, 8000— 9000 ft [2400-2700 m], vi.1901 (Nurse) (MNHU) [examined]. Junior primary homonym of Bombus tetrachromus Cockerell, 1909 [= B. kashmirensis Friese]. Syn. n. Bombus (Pratobombus) yuennanicola Bischoff, 1936: 7. Holotype queen by orginal designation, CHINA: Yunnan, Lijiang, 15.v.1934 (Hone) (MNHU) [examined]. Change of status to Pyro- bombus lepidus subsp. yuennanicola (Bischoff) by Tkalct (1989). Provisional synonym. Bombus (Pyrobombus) yunnanicola Bischoff; Wang, 1982: 435. Unjustified emendation (Art. 33b(i, iii)). Pyrobombus (Pyrobombus) lepidus subsp. hilaris Tkalci, 1989: 48. Holotype queen by original designation, PAKISTAN: Gilgit, Banidas, 36°11'N 74°33’E, 2600 m, 30.vi-2.vii.1959 (Lobbichler) (ZS) [examined]. Syn. n. TYPE MATERIAL. Note 1. Friese’s description of B. genitalis specifies that only a single male was examined. A single male in the MNHU collection agrees with the original description and carries the data quoted, together with a red printed label ‘Type’, and one of Friese’s determination labels ‘Bombus / genitalis / [male] 1912 Friese det / n. Fr.’. This is believed to be the single individual on which the original description is based and is regarded as the holotype. Note 2. Friese’s description of B. nursei var. tetrachromus lists two queens. A single queen in the MNHU collection agrees with the original description and carries a label (1) ‘Kashmir / 8-9000ft / 6.01’; (2) ‘Bombus / tetrachromus / 75 [female] 1915 Friese det. / Fr.’; (3) a red printed label ‘Type’. This specimen, which lacks both front tarsi and the mid left tarsus, is designated as lectotype (Art. 74a). AFFINITIES. Within the subgenus Pyrobombus (see the comments on B. hypnorum), B. lepidus belongs to the parthenius-group of species (see the comments on B. lemniscatus). Within the parthenius-group, the recurved hook of the penis valve head remains broad and scarcely shortened for males of B. lepidus (Figs 68 & 69) (plesiomorphy within the parthenius- group), although males of most of the other species have the recurved hook much shortened (synapomorphy of B. infirmus + B. parthenius + B. luteipes). Individuals of B. infirmus (Tkalct) are banded with yellow pubescence like indivi- duals of B. lucorum, but have fewer white hairs on the apical terga. B. infirmus is known from northern Burma (BMNH) and Sichuan (BMNH, PW). Females of the two Himalayan species, B. parthenius Richards (14.iv.1934) (which is probably conspecific with Bremus sonani Frison, 30.iv.1934) and B. luteipes Richards (conspecific with Pyrobombus signifer Tkalci, 1989, syn. n.), can be recognised by their generally dull grey- yellow pubescence (lighter for the males), with no distinct black band between the wing bases. The apical parts of the sterna and the distal sclerites of the legs are usually a lighter orange-brown colour for females of B. luteipes (see Tkalci, 1989: fig. 26) than for females of B. parthenius. Males of B. luteipes have very distinctive large genitalia, which have the interio-apical corner of the gono- coxite spinosely produced rather than forming a right angle as for B. parthenius. B. luteipes and B. parthenius are known from China and the eastern Himalaya to as far west as Nepal (BMNH, NMS, PW) and Uttar Pradesh (BMNH). Individuals of B. parthenius tend to occur at higher altitudes in the montane forests than do those of B. luteipes. Another nominal taxon, Pyrobombus infre- quens, has been described recently (Tkalcu, 1989) from four workers from Burma and China. These are similar to workers of B. parthenius, but the oculo-malar distance is described as slightly longer. Other workers from near Gongga Shan in Sichuan (PW) are similar to the holotype of Pyrobombus infrequens in that they have few large punctures near the middle of the clypeus, although these workers have most of the orange- red pubescence of terga IV-V replaced by black. A male from the same locality also differs from Nepalese B. parthenius in that the orange-red pubescence of terga IV-V is replaced by black. 76 The genitalia of this male are closely similar to those of males from Nepal, although the gonosty- lus projects inwards distinctly beyond the interio- apical corner of the gonocoxite, whereas this corner of the gonocoxite is longer for both B. parthenius and B. luteipes, and the volsella is slightly broader from the ventral aspect. Much more material of the nominal taxa of the parthenius- group is required to establish their relationships definitively. Most females of B. lepidus from Kashmir have the pale bands of the thoracic pubescence grey- white and have relatively few black hairs on the gastral terga (e.g. Fig. 333; redescribed indepen- dently under the name B. nursei var. tetrachromus by Friese). A queen from north-western Kashmir with the pale pubescence of the thorax and tergum I yellow as well as on tergum II (Fig. 327) was described by Tkalci (1989) under the name Pyrobombus lepidus subsp. hilaris. Some indivi- duals of both sexes from further to the east in Tibet (Wang, 1982, 1987; BMNH, PW), Yunnan (Bischoff, 1936; Wang, 1987; BMNH), Sichuan (BMNH) and Shanxi (Yasumatsu, 1951) also have all of the pale pubescence yellow. These indivi- duals were described by Bischoff (1936) under the name B. yuennanicola, which Tkalcu (1989) has regarded as a subspecies of B. lepidus. For females of B. yuennanicola, the many close, fine punctures of the post-ocular area of the head do not extend as far anteriorly among the large punctures of the lateral part of the ocello-ocular area, although the difference is subtle. The male genitalia have the shoulder of the penis valve particularly convex and the inner margin of the gonostylus is more deeply and broadly concave (Wang, 1982: fig. 39; PW). Both the yellow individuals and the white individuals apparently occur together at the same localities in the far east of Tibet (Wang, 1982: Jomda & Chagyab) and in Yunnan (Wang, 1987: Dongchuan). However, a male from Tibet described by Friese (1913) under the name B. genitalis is more extensively yellow like the males of B. lepidus, although the genitalia are more similar to those of B. yuennanicola. It is possible that these bees are all part of the same interbreeding population, in which the shape of the gonostylus and the colour of the pale pubes- cence show more or less correlated patterns of variation. Much more information on interbreed- ing and variation is needed to establish the status of the nominal taxa in this group. DISTRIBUTION. B. lepidus is a peri-Tibetan species (Map 39). It is known from Shanxi (Yasumatsu, 1951), Gansu (Skorikov, 19126; Bischoff, 1936; Wang, 1982; ZS), Qinghai (Skorikov, 19125; PAULH. WILLIAMS Wang, 1982; BMNH, ZI), Yunnan (Bischoff, 1936; Wang, 1987; BMNH, MNHU), Tibet (Friese, 1913; Wang, 1982; BMNH, MNHU, PW), Nepal (Tkalct, 19745; BMNH), Himachal Pradesh (BMNH, PW) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. lepidus is recorded from the Hindu Raj range, from the southern side of the Great Himalaya range and from the Pir Panjal range, primarily from around the upper limit of the montane coniferous forest (Map 40, Fig. 6): 23 queens, 79 workers, 22 males, from 11 localities 2400-4600 m (BMNH, MNHU, NR, PW, ZS). VARIATION WITHIN KASHMIR. Females from the Hindu Raj and Karakoram ranges have the pale pubescence of the thorax and terga I-II yellow and have very few black hairs on tergum III (Figs 327 & 328). Queens and workers from the Kishanganga valley have a variable amount of grey-white hair intermixed on the head. The pale pubescence of the thorax and of tergum I is grey-white (tergum IT remains yellow), and there are very few or no black hairs on tergum III (Figs 330 & 331). Females from Mt Apharwat are similar, although the lateral margins of tergum II are fringed with more of the white hairs (Figs 333 & 334), and one queen has all of the pubescence of tergum II cream-white. Yellow females (Figs 327 & 328) are similar to some of the B. himalayanus (Figs 242 & 243) and B. keriensis (Fig. 404 & 407) that may occur with them. White and yellow workers of this species are particularly closely similar in colour pattern (Figs 331 & 334) to the B. lemniscatus (Figs 325 & 326) and B. biroi (Fig. 339) workers that occur with them at some localities in the Kishanganga valley and around the Vale of Kashmir. White and yellow females also resemble some of the B. himalayanus (Figs 247-249), B. novus males (Fig. 262), B. kashmirensis (Fig. 203) and B. rufofasci- atus (Figs 435 & 436). B. lepidus can be recognised by the characters given in the key (couplets 19 & 21). The yellow males with red hairs on the apical terga may be similar in colour pattern (Figs 329 & 332) to the males of B. pyrosoma (Fig. 430), which occur with them at some localities. B. lepidus males can be recognised by their small size and by their rounded gonostyli (Figs 188 & 189). FOOD PLANTS. (Ranunculaceae) Aconitum laeve Royle, A. heterophyllum Wallich ex Royle; (Balsaminaceae) Impatiens glandulifera Royle; (Scrophulariaceae) Scrophularia pauciflora Benth.., Digitalis lanata Ehrh. [introduced]. THE BUMBLE BEES OF THE KASHMIR HIMALAYA Bombus (Pyrobombus) biroi Vogt (Figs 42, 70, 71, 110, 111, 150, 151, 190, 191, 214, 217, 335-346, Maps 41 & 42) Bombus (Pratobombus) biroi Vogt, 1911: 51. Lectotype queen by designation of Tkalci (1969a: 195), U.S.S.R.: Kirgiziya S.S.R., Alai mountains, 1905 (TM) [examined]. Bombus (Pratobombus) biroi [var.] flavobistriatus Vogt, 1911: 52 (see Note 1 below). LECTO- TYPE queen by present designation (see Note 2 below), U.S.S.R.: Kirgiziya S.S.R., Naryn, 5.vi.1904 (ITZ) [examined]. Syn. n. [Bombus (Pratobombus) biroi flavostriatus Vogt, 1911: 52 [examined]. Infrasubspecific (see Note 1 below) (Art. 45g(ii)(1)), unavailable name. | [Bombus (Pratobombus) biroi flavofasciatus Vogt, 1911: 52 [examined]. Infrasubspecific (see Note 1 below) (Art. 45g(ii)(1)), unavailable name. ] Bombus nursei Friese, 1918: 84. Holotype queen by monotypy (see Note 3 below), INDIA: Kashmir, 8000-9000 ft [2400-2700 ml], vi.1901 (Nurse) (MNHU) [examined]. Syn. n. Bombus abbotti Cockerell, 1922: 2. Holotype queen by original designation [p. 3], CHINA: Xinjiang, Taghdumbash, 13000 ft [4000 ml], 11.vii.1894 (Abbott) (USNM) [examined]. Syn- onymised with Pyrobombus biroi (Vogt) by Tkalct (1969a). Bombus agnatus Skorikov, 1933b: 248. Lectotype queen by designation of Podbolotskaya (in press), INDIA: Kashmir, Ladakh, Khardung La, 2-4.vii.1912 (Jacobson) (ZI) [examined]. Junior primary homonym of Bombus lapponi- cus var. agnatus Skorikov, 1912a [not infrasub- specific after Tkalci, 1969a (Art. 45g(ii)(1))]. Provisional synonym. Bombus (Pratobombus) kotzschi Reinig, 1940: 227. Holotype male by monotypy, presumed lost (see Note 4 below), AFGHANISTAN: Baghlan, Andarab, 4000-4500 m, viii.1936 (Kotzsch) [not seen]. Synonymised with Bombus agnatus Skorikov by Tkalct, 1969a. Provisional synonym. NOMENCLATURE. Note 1. Vogt states (1911:50, footnote 1) that he intends ‘Varietas geographica’ to be equivalent to subspecies in rank and “Aberratio extrema’ to be of infrasubspecific rank. He uses ‘Forma’ for taxa that he could not assign to either rank with certainty. In the second paragraph on page 52 he describes ‘Die Varietat von Narynj und dem Boro-Chorogebirge . . .’ and the first name used in this paragraph is ‘(flavobis- triatus)’. This is deemed to be the subspecific name that was intended for individuals from this area with this colour pattern, which Vogt charac- terised by a greater tendency towards a yellowish TW colour for the pale pubescence. The four names that follow in the same paragraph are all applied to what can be regarded as variants of this local colour pattern, even though only the first two (dyspostzonatus and nigrozonatus) have their infrasubspecific status specified by the use of the term ‘ab.’. Therefore the taxon named flavobistri- atus is deemed to be of subspecific rank, whereas flavostriatus and flavofasciatus are deemed to be of infrasubspecific rank (Art. 45g(1i)(1)). TYPE MATERIAL. Note 2. Vogt’s description of B. biroi var. flavobistriatus does not specify the number of specimens examined. A single queen in the ITZ collection agrees with the original des- cription and carries a label (1) ‘Narynj / 5.6.04’; (2) a red, handwritten label ‘flavobistriatus / Type’; (3) ‘Collectie /C. et O. Vogt / Acq. 1960”. This specimen, which lacks only the left antenna, is designated as lectotype (Art. 74a). Note 3. Friese’s description of B. nursei speci- fies a single queen. A single queen in the MNHU collection agrees with the original description and carries labels with the data quoted, together with a red, printed label “Type’ and one of Friese’s determination labels ‘Bombus / nursei’, dated 1909. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 4. Reinig’s description of B. kotzschi specifies a single male from Andarab. This speci- men could not be found either in the ITZ collec- tion (W. Hogenes in litt.) or in the ZS collection (E. Diller in litt.), which are the collections with many of Reinig’s bumble bee type- specimens. According to P. Rasmont (in /itt.), any of Reinig’s pre-war material not in the ITZ collection must be considered to have been destroyed by soldiers during the winter of 1945. The types of P. rupestris subsp. elisabethae Reinig, 1940, from Afghanistan in the ZS collection may have survived because they were on loan to a colleague. AFFINITIES. Among the Old World species of the subgenus Pyrobombus (see the comments on B. hypnorum), the male genitalia of B. biroi are very distinctive, because the sickle-shaped, recurved hook of the penis valve head is strongly broadened and apically rounded (Figs 70 & 71; Kriiger, 1943: p. 530 figs 30-45). The other species that share this character state belong to the North American flavifrons-group (?synapomorphy of the broad hook of the penis valve head) (the ‘Flavifrons group’ of Thorp, 1969). This includes at least B. flavifrons Cresson, B. centralis Cresson, B. vandykei (Frison), B. caliginosus (Frison) and B. vagans Smith. B. sandersoni Franklin and B. 78 frigidus Smith are similar but have the hook of the penis valve head less strongly broadened. The closest relatives of B. biroi in the Old World may include the species of the pratorum- group, which are primarily Eurosiberian (see e.g. Lgken, 1973; BMNH). Among the species of the subgenus Pyrobombus, they have relatively undif- ferentiated male genitalia. But for B. jonellus (Kirby), B. pyrenaeus Pérez and B. brodmannicus Vogt in particular, not only is the ventro-lateral angle of the penis valve just nearer to the base than to the head of the penis valve (as for species of the flavifrons-group and B. biroi, Figs 110 & 111), but the recurved hook of the penis valve head is also weakly broadened (see Kriiger, 1943: 530). B. jonellus is widespread in the northern parts of Europe and of the U.S.S.R. (e.g. Panfilov, 1982: map 150) and reaches eastwards to western Canada (where it occurs together with B. flavi- frons; PW). B. pyrenaeus is known from the mountains of southern Europe, including the Pyrenees, the Alps, the Carpathians and the Balkans (e.g. Tkalci, 19695; BMNH, PW). B. brodmannicus is known from both the western Alps and from Turkey and the Caucasus (e.g. Tkalct, 1973; BMNH, PW). The females of B. biroi can be distinguished by their few, scattered, small punctures among the large punctures of the post ocello-ocular areas of the head, where the other species have these areas closely and evenly covered with small punctures. The lectotype queen of B. biroi from the Alai mountains has the pale bands on the thorax and on terga I-II grey-white, whereas the pubescence of terga I-II is yellow for the queen described by Vogt (1911) under the name B. biroi var. flavobis- triatus. Queens from the Himalaya are closely similar in morphology, but the grey-white pubes- cence is often extensively replaced by black, especially on the thorax and on tergum I (des- cribed under the name B. nursei by Friese, 1918). Although the single male described under the name B. kotzschiis apparently no longer available (see Note 4), Reinig’s description of the broader recurved hook of the penis valve head (Reinig, 1940: 228 ‘ist der Uncus etwas kraftige entwickelt als bei B. pratorum.’) and of the unbanded yellow colour pattern, similar to B. semenovianus, agrees most closely with two males from the Zanskar mountains that are otherwise similar to B. biroi (Figs 70 & 71). Variation in the male genitalia of B. biroi is greatest in the breadth of the recurved head of the penis valve, but this is not associated with the variation in the colour pattern of the pubescence. It appears to have no more than in- dividual significance and the males from Zanskar are within the range of variation for banded B. PAULH. WILLIAMS biroi (e.g. Kriiger, 1943: p. 530 figs 30-45). Females with a similar, unbanded yellow colour pattern were described from Ladakh by Skorikov (1933b) under the name B. agnatus. More recently, Tkalct (1969a) synonymised B. kotzschi with B. agnatus, but listed them as a separate species from Pyrobombus biroi without further comment. From an examination of the available material, the unbanded yellow females of B. agnatus seem to be closely similar in morphology to the banded yellow and white females of B. biroi from Kashmir, as well as to the banded yellow females of B. biroi from further north in central Asia. There is some variation in the extent of the fine punctures in the ocello-ocular area and in the breadth of the labral lamella of the females, especially among individuals of B. nursei. These characters also vary considerably within each nominal taxon so that they do not appear to be diagnostic. Thus B. biroi and B. kotzschi (as the oldest available name for B. agnatus) may be two separate species, but until there is better evidence against interbreeding than just differences in colour pattern, I consider that they are most likely to be parts of the same species, B. biroi. More material is needed from the possible hybrid zone to the north of the Great Himalaya range in order to resolve the problem. DISTRIBUTION. B. biroi is a central Asian species (Map 41). It is known from Mongolia (BMNH, PW), the Tien Shan (Vogt, 1911; Skorikov, 1931; Kriiger, 1943; BMNH, ITZ, TM), the Pamir (Cockerell, 1922; Reinig, 1930, 1934; Skorikov, 1931; USNM), the Hindu Kush (Reinig, 1934, 1940; Tkalci, 19692; BMNH), Kashmir and Himachal Pradesh (BMNH, PW). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. biroiis recorded from the Hindu Raj, Ladakh, Zanskar, Great Himalaya and Pir Panjal ranges, in subalpine scrub (Map 42, Fig. 6): 14 queens, 38 workers, 12 males, from 13 localities 2300-4600 m (BMNH, MNHU, NR, PW, ZI). During July 1986, I saw queens investigate gaps among the large boulders that were lying in gullies in the subalpine zone of Mt Apharwat. However, I could not find any colonies there in August or September. VARIATION WITHIN KASHMIR. Females from the Hindu Raj range have the thoracic dorsum uni- formly yellow and terga I-III black (5/S queens, 1/1 workers, Fig. 335), like the lectotype queen of B. agnatus from Ladakh. A single worker from Rumbak in the Zanskar ranges has yellow hairs intermixed on tergum I (Fig. 336). Two males from the Zanskar ranges at Panichar near Suru THE BUMBLE BEES OF THE KASHMIR HIMALAYA and at Tungri (Fig. 337) have the pubescence of the thorax, of tergum I and of the basal three- quarters of tergum II uniformly yellow, with the apical and lateral margins of tergum II, all of tergum III and for the male from Suru, the basal margin of tergum IV black. This colour pattern is otherwise known for this species only from the Hindu Kush ranges (BMNH). The species may not vary much in colour pat- tern on the southern side of the Great Himalaya range, where light individuals occur (Figs 338- 340), but it is very variable in the Pir Panjal range, where dark individuals (Figs 341-346), similar to those from Himachal Pradesh, also occur. On Mt Apharwat the variation is chiefly in the form of (a) a reduction of the pale thoracic bands to an almost completely black thorax, with only a few pale hairs intermixed anterior to the wing bases; (b) replacement of the red with black progres- sively across terga III-VI, the darkest queen retaining red only on the posterior median part of tergum IV and on the anterior median part of tergum V; and (c) replacement of the yellow with white on tergum II. Reduction of the pale thoracic pubescence is pronounced for all 4 queens (Figs 341, 344, 346), but less common among workers (2/18 workers, Fig. 345). Complete replacement of yellow with white on tergum II may be common among queens (2/4 queens, Figs 341, 346) and among workers with broad bands of pale pubes- cence on the thorax (6/16 workers, Fig. 342). The unbanded yellow individuals of this species (Figs 335-337) are closely similar to some B. subtypicus (Figs 319-324), although they have not been found together. They are also similar to some B. avinoviellus (Fig. 232), B. marussinus (Figs 254-257), B. branickii (Fig. 265), B. asiati- cus (Figs 371-376) and B. semenovianus (Figs 397-402) that may occur with them. Unbanded yellow B. biroi can be recognised by the charac- ters given in the key (couplets 19 & 20). The light, white and yellow individuals (Figs 338 & 339) are closely similar in colour pattern to the B. lemnis- catus (Figs 325 & 326) and B. lepidus (Figs 330, 331, 333, 334) that occur with them at some localities. They also resemble some of the B. himalayanus (Figs 247-252), B. novus males (Figs 260 & 262), B. kashmirensis (Figs 302-304) and B. rufofasciatus (Figs 435 & 437) that occur with them. B. biroi can be recognised by the characters given in the key (couplets 19 & 20). The queens with a dark thorax and a white tergum II (Figs 341 & 346) resemble some of the B. himalayanus (Fig. 250) and B. kashmirensis (Figs 308 & 309) that occur in the same area. B. biroi can be distin- guished from these by its combination of a shiny outer surface to the hind tibia and by its simple 79 mandibles, which lack the many, pronounced teeth of B. kashmirensis. Some of the darkest individuals of B. biroi (Figs 344 & 345) show some similarity in colour pattern to some individuals of B. haemorrhoidalis or B. trifasciatus but have the oculo-malar distance much shorter. These species are completely segregated by altitude. The banded yellow males (Figs 340 & 343) are similar in colour pattern to the males of B. keriensis (Figs 410 & 416) that occur with them, but can be recognised by the broader recurved hooks of the head of the penis valve (Figs 70 & 71). FOOD PLANTS. Unidentified dandelion-like com- posites; (Gentianaceae) Swertia petiolataD. Don; (Scrophulariaceae) Scrophularia pauciflora Benth.., Picrorhiza kurrooa Royle ex Benth. Subgenus BOMBUS Latreille Leucobombus Dalla Torre, 1880: 40 (as a sub- genus of Bombus Latreille). Type species: Apis terrestris Linnaeus, 1758: 578 = Bombus ter- restris (Linnaeus), by subsequent designation of Sandhouse (1943: 564). Synonymised with Bombus Latreille s.str. by Milliron (1961). Terrestribombus Vogt, 1911: 55 (as a subgenus of Bombus Latreille). Type species: Apis ter- restris Linnaeus, 1758: 578 = Bombus terrestris (Linnaeus), by subsequent designation of Frison (1927: 67). Synonymised with Bombus Latreille s.str. by Skorikov (1931). Authors have disagreed concerning the correct application of the name Apis terrestris Linnaeus (see the comments on B. lucorum). This is currently under consideration by the ICZN but does not affect the application of the name Bombus s.str. For a general description of both sexes see Richards (1968). Bombus (Bombus) tunicatus Smith (Figs 72, 73, 112, 113, 152, 153, 192, 193, 347-355, Maps 43 & 44) Bombus tunicatus Smith, 1852a: 43. Lectotype queen by designation of Tkalci (1974b: 325), INDIA: north, no further data (BMNH) [examined]. Bombus gilgitensis Cockerell, 1905: 223. Holo- type queen by monotypy (see Note below), PAKISTAN: Gilgit (BMNH) [examined]. Change of status to Bombus tunicatus [Rasse] gilgitensis Cockerell by Tkalct (1969a). Syn. n. Bombus terrestris var. simlaénsis Friese, 1909: 80 674. Lectotype queen by designation of Tkalct (1974b: 325), INDIA: Kashmir, ‘2—3000’ [?m], vi.1904 (MNHU) [not seen]. Synonymised with Bombus tunicatus subsp. tunicatus Smith by Tkalct (19745). Bombus terrestris var. fulvocinctus Friese, 1909: 674. Type worker not found (Tkalct, 1974b), INDIA: Himachal Pradesh, Simla [not seen]. Synonymised with Bombus tunicatus Smith by Tkalct (1974b). TYPE MATERIAL. Note. Cockerell’s description of B. gilgitensis specifies a single queen. A single queen in the BMNH collection agrees with the original description and carries a label ‘Gilgit’ and another label ‘Bombus / gilgitensis / Ckll TYPE’, in handwriting identical to that of Cockerell. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. The species of the subgenera Festivo- bombus Tkalct,, Rufipedibombus Skorikov, Pressi- bombus Frison and most Bombus s.str. share a subapical constriction of the volsella (?synapo- morphy). The species of the subgenera Rufipedi- bombus, Pressibombus and Bombus share an enlargement of the outer ridge of the penis valve head and a reduction of the exterio-apical corner of the gonostylus (synapomorphies). The species of the subgenera Pressibombus (which includes only the east Himalayan B. pressus (Frison)) and Bombus share a ventrally and apically directed extension of the outer ridge of the penis valve head (synapomorphy). The species of the sub- genus Bombus share a loss of the recurved hook of the penis valve head and a great broadening of the ventrally directed outer ridge, so that the penis valve heads are flared outwards to form a funnel (Figs 72-75, 112-115) (synapomorphy). The vol- sella shows a secondary distal broadening by the extension of its broad interio-apical process in- wards towards the midline of the body (Figs 152— 155) (?synapomorphy of the subgenus Bombus). Within the subgenus Bombus there are a very large number of nominal taxa, although there is relatively little variation in morphology in comparison with some apparently monophyletic subgenera with fewer species, such as Subterraneo- bombus. The male genitalia of the Eurosiberian B. sporadicus Nylander (distribution mapped by Tkalci, 1967: 45) have broader and straighter penis valves in lateral aspect than do those of B. (Pressibombus) pressus (synapomorphy of the subgenus Bombus). However, for B. sporadicus the subapical inner margin of the penis valve is nearly straight or convex in dorsal aspect, the interio-apical corner of the gonostylus is produced PAULH. WILLIAMS beyond the exterio-apical corner by a distance along the axis of the gonocoxite that is nearly equal to the breadth of the volsella below its interio-apical process, the interio-basal process of the gonostylus is particularly weakly sclerotised and the volsella projects beyond the gonostylus in dorsal aspect by more than its own breadth below its interio-apical process (?plesiomorphies within the subgenus Bombus, showing some similarity to B. pressus). The females of B. pressus and B. sporadicus share a lack of large punctures in the posterior part of the ocello-ocular area (for B. sporadicus see Tkalcu, 1967: fig. 1). The male genitalia of all species of the subgenus Bombus except B. sporadicus have the subapical inner margin of the penis valve concave in dorsal aspect (usually with a small rounded projection interio-basally, particularly Figs 72 & 73), the gonostylus has the interio-apical corner produced beyond the exterio-apical corner by a distance that is much less than the breadth of the volsella (Figs 192-195), the interio-basal projection of the gonostylus is strongly sclerotised, and the exterio- apical corner of the volsella projects beyond the gonostylus in dorsal aspect usually by much less than its own breadth (synapomorphies of B. tunicatus + B. franklini + B. affinis + terrestris- group). Species of the ferrestris-group (see the comments on B. lucorum) have the ventro-basal angle of the penis valve marked with a much more pronounced right angle in lateral aspect (Figs 114 & 115) (synapomorphy). B. tunicatus is unique among the Old-World species of the subgenus Bombus for the combina- tion of a reduced projection of the interio-apical corner of the gonostylus (Figs 192 & 193) and a broadly obtuse form of the ventro-basal angle of the penis valve (Figs 112 & 113). The other species with these character states, B. franklini (Frison) and B. affinis Cresson, are restricted to a small area near the west coast of North America (see Thorp etal. , 1983) and to the eastern side of North America (see Milliron, 1971) respectively. Males of B. tunicatus are unique among all species of the subgenus in that they have the gonostylus extended by the development of a distal shelf, so that the interio-apical corner is often particularly broadly curved (Figs 192 & 193). The queens can be distinguished by a distinct band of fine punc- tures in a slight groove at the midline of the clypeus. Cockerell described a queen under the name B. gilgitensis from north-western Kashmir that differs from those from the Vale of Kashmir and from those of the eastern Himalaya by an exten- sive replacement of the pale pubescence by black on the scutellum and on terga I-II. These dark THE BUMBLE BEES OF THE KASHMIR HIMALAYA females from the lower Indus valley are otherwise closely similar to B. tunicatus s.str. and are considered to be part of the same species (see below). DISTRIBUTION. B. tunicatus is a Himalayan species (Map 43). It is known from Sikkim (Bingham, 1897; Friese, 1918; Skorikov, 1933b), Nepal (Tkalci, 1974b; BMNH, NMS), Uttar Pradesh (Bingham, 1897; Frison, 1935; BMNH, MNHN, UM), Himachal Pradesh (Friese, 1909; Cockerell, 1917; Frison, 1935; BMNH, MNHN, PW), Kashmir, Pakistan (Frison, 1933, 1935; PW) and the Hindu Kush (Tkalct, 1969a). It may also have been genuinely recorded from the Indian plains, at Calcutta (Dover, 1922). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. tunicatus is recorded from the Hindu Raj, Karakoram, Ladakh, Great Himalaya and Pir Panjal ranges, where it is widespread among habitats, if most abundant around the lower montane coniferous forest (Map 44, Fig. 6): 149 queens, 569 workers, 195 males, from 37 localities 1600-4100 m (AB, BMNH, BPBM, IZ, NR, PW, RH, TL, ZM, ZS). A colony was found in woodland near Harwan (1700 m, 10.ix.1985). The nest entrance was a hole 3 cm wide in the soil amongst grass, 1m from the foot of a tree. VARIATION WITHIN KASHMIR. Queens from the southern side of the Great Himalaya range and from the Pir Panjal range have a lighter colour pattern, with the pubescence of the scutellum, tergum I and most of tergum II grey-white (Figs 350 & 353). The workers differ in that the black band between the wing bases is more often reduced or replaced by grey-white hairs, and the pale pubescence of tergum II is usually chocolate- brown (Figs 351 & 354). Males may have tergum II either grey-white (Fig. 355) or chocolate-brown (Fig. 352). Queens from the northern side of the Great Himalaya and from the Karakoram and Hindu Raj ranges are darker, with the pale pubescence of the scutellum and of terga I-II almost com- pletely replaced by black (Fig. 347). The workers and males associated with these queens retain a pronounced pale band on the scutellum, but the black band between the wing bases always extends over at least one-third of the length of the thorax (Figs 348 & 349). The chocolate-brown pubes- cence is replaced by black on all but a narrow crescent in the middle of the base of tergum II. Most individuals from all localities can be identified with one or other of these two extremes of the variation within Kashmir, so that there 81 appear to be regional differences in colour pattern without much overlap. For this species of the low to mid montane zones, the segregation of the light individuals in the Jhelum valley watershed (Vale of Kashmir and the Kishanganga valley), from the dark individuals of the Indus valley watershed (Map 44), is consistent with regional colour variation as parts of a single interbreeding popu- lation and a single species. Opportunities for dispersal of individuals over the intervening barrier of high mountain ridges may perhaps have been restricted, which could have resulted in the apparently narrow colour cline (cf. the comments on B. lucorum). Light queens of this species (Figs 350 & 353) are most similar in colour pattern to the B. avinoviel- lus (Figs 238-241) and B. asiaticus (Figs 383-389) that occur with them at some localities in the Vale of Kashmir. They also resemble some B. kash- mirensis (Figs 305-307), B. biroi (Figs 341 & 342) and B. keriensis (Figs 408 & 409). Queens of B. tunicatus can be recognised by their large size, by the short oculo-malar distance and by the strongly arched posterior margin of the hind basitarsus. Workers that lack a black band on the thorax (Fig. 354) can be closely similar in colour pattern to workers of B. simillimus (Figs 418 & 420), but can usually be recognised easily because they have tergum I white rather than chocolate-brown, which is very rare for B. simillimus. Males can be recognised by their outwardly flared heads of the penis valves (Figs 72, 73, 112, 113). FOOD PLANTS. (Ranunculaceae) Aconitum hetero- phyllum Wallich ex Royle, A. hookeri Stapf; (Balsaminaceae) Impatiens glandulifera Royle; (Leguminosae) Lupinus sp. [introduced]; (Sam- bucaceae) Lonicera quinquelocularis Hardw.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC., Centaurea iberica Trevir ex Spengel, unidentified yellow thistle-like compo- sites, unidentified dandelion-like composites; (Gentianaceae) Swertia petiolata D. Don; (Scrophulariaceae) Digitalis lanata Ehrh. [intro- duced], D. purpurea L. [introduced], Pedicularis punctata Decne.; (Acanthaceae) Pteracanthus urticifolius (Kuntze) Bremek.; (Labiateae) Stachys sericea Wallich ex Benth., Prunella vulgaris L. Bombus (Bombus) lucorum (Linnaeus) (Figs 74, 75, 114, 115, 154, 155, 194, 195, 356-370, Maps 45 & 46) [Apis terrestris Linnaeus, 1758: 578. Lectotype queen by designation of Day (1979: 74), ‘in Europa’, no further data (LSL) [examined]. Application temporarily suspended (see Note 3 below) (Art. 80c).] 82 Apis lucorum Linnaeus, 1761: 425. Lectotype male by designation of Day (1979: 66), SWEDEN: no further data (LSL) [examined]. Synony- mised with Bombus terrestris (Linnaeus) by Day (1979). Apis cryptarum Fabricius, 1775: 379. Lectotype queen by designation of Lgken (1966: 199), DENMARK: Copenhagen (Kiel) [not seen]. Change of status to Bombus terrestris var. cryptarum (Fabricius) by Thomson (1872); syn- onymised with Bombus lucorum (Linnaeus) by Léken (1966). Provisional synonym. Bombus lucorum (Linnaeus); Fabricius, 1804: 350. Bombus modestus Cresson, 1863: 99. Holotype queen by monotypy (see Note 4 below), U.S.A./CANADA: Yukon river, no further data (Kennicott) [not seen]. Junior primary homonym of Bombus modestus Eversmann, 1852. Synonymised with Bombus lucorum [subsp.] /ucorum (Linnaeus) by Milliron (1971). Bombus moderatus Cresson, 1863: 109. Replace- ment name for B. modestus Cresson. Synony- mised with Bombus lucorum [subsp.] lucorum (Linnaeus) by Milliron (1971). [Bombus lucorum ab. mongolicus Vogt, 1909: 42. Infrasubspecific (see Note 1 below) (Art. 45f(iv)), unavailable name (Art. 45e).] Bombus (Terrestribombus) lucorum forma magnus Vogt, 1911: 56. Syntype queens, BRITAIN: northern Scotland; Orkney Isles (ITZ) [not seen]. Synonymised with Bombus lucorum (Linnaeus) by Pekkarinen (1979). Provisional synonym. [Bombus (Terrestribombus) lucorum mongolicus Vogt, 1911: 56. Published without description or indication, unavailable name (Art. 12a).] Bombus lucorum subsp. jacobsoni Skorikov, 1912b: 610. Lectotype queen by designation of Podbolotskaya (in press), INDIA: Kashmir, Maru, Fargabad, 19.v.1910 (Trubetskoy) (ZI) [examined]. Syn. n. Bombus terrestris var. lucocryptarum Ball, 1914: 82. Lectotype queen by designation of Rasmont (1981: 151), BELGIUM: Brabant, Molenstede, 6.v.1913 (Brussels) [not seen]. Synonymised with Bombus cryptarum (Fabricius) by Rasmont (1983a). Provisional synonym. Bombus pratorum var. longipennis Friese, 1918: 83. LECTOTYPE worker by present designa- tion (see Note 5 below), INDIA: Sikkim, no further data (MNHU) [examined]. Syn. n. Bombus (Terrestribombus) lucorum var. alaiensis Reinig, 1930: 107 (see Note 2 below). 3 syntype queens, 7 syntype workers, U.S.S.R.: Kirgiziya S.S.R.; Tadzhikistan S.S.R., 1950-3700 m, 21.vi-28.vii.1928 (ITZ) [not seen]. Synony- PAULH. WILLIAMS mised with Bombus lucorum subsp. jacobsoni Skorikov by Tkalct (1969a). Syn. n. [Bombus lucorum mongolicus Skorikov, 1933a: 57. Published without description or adequate indication (only Vogt, 1911, cited), unavailable name (Art. 12a).] Bombus (Terrestribombus) magnus Rasse mon- golicus Kriiger, 1954: 276 [by indication of Vogt, 1909: 42]. LECTOTYPE queen by present designation (see Note 6 below), MONGOLIA: north, no further data, 1892 (Leder) (ITZ) [examined]. Change of status to Bombus lucorum [subsp.] mongolicus Kriger [cited as Vogt] by Tkalct (1974a). Syn. n. Bombus reinigi Tkalct, 19746: 322. Holotype queen by original designation, NEPAL: Ringmo to Phoksumdo lake, 3700 m, 10.vi.1973 (Martens) (NM) [examined]. Syn. n. NOMENCLATURE. Note 1. Vogt’s (1909) term ‘ab.’ must be taken to indicate infrasubspecific status (Art. 45f(i1, iv)). For his subsequent paper, Vogt states explicitly (1911: 50 footnote 1) that he intends “Varietas geographica’ to be equivalent to subspecies in rank and ‘Aberratio extrema’ to be of infrasubspecific rank. He uses ‘Forma’ for taxa that he could not assign to either rank with certainty and these are treated as being of sub- specific rank (Art. 45g(ii)). Note 2. Reinig’s use (1930: 83 etc.) of the term ‘ssp.’ elsewhere in his paper is taken to indicate that infrasubspecific rank is meant for B. lucorum var. alaiensis. However, this name has subse- quently been treated as a name in the species group by Kriiger (1951), so it is deemed to be of subspecific rank (Art. 45g(1i)(1)). TYPE MATERIAL. Note 3. Both Lg@ken (1973) and Day (1979) noted that the only admissible syntype material of A. terrestris Linnaeus did not agree with the traditional interpretation of the taxon (which corresponds to Harris’s Apis audax, 1776: 130, not of page 137), but instead may be a queen of the same species as B. lucorum. This syntype queen has since been identified as conspecific with B. cryptarum (Fabricius) by Rasmont (1988: 52). If this is the only admissible syntype, then the many citations of B. terrestris in its traditional sense must refer to misidentifications of Linnaeus’s taxon. This is unfortunate because the name B. terrestris appears particularly widely in the litera- ture on physiology, behaviour and ecology. To minimise any possible confusion an application has been submitted to the ICZN to use its plenary powers for the conservation of the traditional application of B. terrestris (L¢ken, Pekkarinen & Rasmont, unpublished). Note 4. Cresson’s description of B. modestus THE BUMBLE BEES OF THE KASHMIR HIMALAYA (1863) specifies that a single female was examined. If found, this specimen which was originally in the ‘Coll. Smith. Inst.’ should be regarded as the holotype (Art. 73a(ii)). Note 5. Friese’s description of B. pratorum var. longipennis (1918) lists three queens and one worker from Sikkim. I have seen three workers from the MNHU collection that agree with the original description and carry labels with the data ‘Sikkim / Coll. Bingham’ and Friese’s determina- tion labels. Large workers of species of the subgenus Bombus are often similar in size to small queens of species of the subgenus Pyrobombus. One of these workers in the MNHU collection carries the labels (1) ‘Sikkim / Coll. Bingham’; (2) ‘B. prat./v. longipennis / [female] 1914 Friese det. /n. n.’; (3) a handwritten label ‘selbstandige Art’; (4) ‘LECTOTYPE / Bombus prato- / rum var. longi / pennis Tkalcu det. / Friese’ [designation not published, see Tkalct, 19746: 324, footnote 4]; (5) ‘Zool. Mus. / Berlin’. This specimen, which lacks the left hind leg, is designated as lectotype (Art. 74a). Note 6. Kriiger (1954) elevates Vogt’s B. lucorum ab. mongolicus to subspecific rank as part of the nominal taxon B. magnus. Vogt’s original description of B. lucorum ab. mongolicus refers to more than one specimen. A queen in the ITZ collection agrees with the original description and carries a label (1) ‘mongolicus Vogt’ in hand- writing identical to that of Vogt; (2) ‘terrestris / det. Kohl.’; (3) ‘.gelb’; (4) a label with two illegible lines written in pencil, possibly ‘B terres- tris v. / or lucorum’; (5) ‘N. Mongolei / Leder 92’; (6) ‘Coll. Zoologisch / Museum / Amsterdam’; (7) ared printed label ‘TYPE’; (8) ‘LECTOTYPUS / B.lucorum ab. / mongolicus Vogt / [female] Tkalct’ det.’ [designation not published]. This specimen, which lacks the distal joints of the right hind tarsus, is designated as lectotype (Art. 74a). AFFINITIES. B. lucorum is one of the least well understood bumble bee taxa at present and appli- cation of the name has been a matter of opinion. There is both a biological problem in determining how many species of the tferrestris-group are present in Europe and Asia, and a nomenclatural problem (see Note 3 above) that has arisen from what appears to be the perpetuated misapplica- tion of Linnaeus’s name Apis terrestris (currently under consideration by the ICZN). The biological limits of the species for which, at present, A. lucorum is the oldest available name (Art. 80c) remain very much a question of the interpretation of indirect evidence. Within the subgenus Bombus, species of the terrestris-group (see the comments on B. tunica- 83 tus) have the ventro-basal angle of the penis valve marked with a strongly pronounced right angle in lateral aspect (Figs 114 & 115) (synapomorphy). Only the most southerly species of the terrestris- group in the lowlands of eastern Asia, B. ignitus Smith (distribution mapped by Tkalci, 1962: 89; Sakagami, 1975: fig. 13), can be distinguished easily by the shape of the male genitalia. In particular, the apex of the penis valve head of B. ignitus is more elongated dorsally and finger- like, whereas the ventro-apical corner is rounded rather than acutely pointed. The remaining taxa within the ferrestris-group are less strongly diver- gent in their morphology. Those differences in the form of the male genitalia that are known may not always be reliably discrete, particularly among samples from widely separated areas. Any cladis- tic interpretations of relationship based on them would be premature. In the northern part of its range B. ignitus overlaps with the east Asian B. hypocrita Pérez in northern China and Japan (distribution mapped by Tkalcti, 1962: 89), and with the Siberian and Chinese B. patagiatus Nylander in northern China (distribution mapped by Tkalci, 1967: 52). In North America the terrestris-group is represented by the widespread B. terricola Kirby (probably conspecific with B. occidentalis Greene, see e.g. Milliron, 1971, for distribution records) and by the north-western B. moderatus Cresson (probably conspecific with B. lucorum s.1.). In Britain, at least two species of the terrestris- group coexist over a large area: B. terrestris (of most authors after Linnaeus, but not of Linnaeus, see Note 3 above) and B. /ucorum. They can be distinguished only by subtle differences in their morphology, by the length of the pubescence and by their colour patterns. For instance, British queens of B. terrestris of authors can be recog- nised by their larger size, by their shorter pubes- cence, which has broader, more golden-yellow bands (rather than lemon-yellow), and especially by the brownish or ‘buff’- coloured pubescence of terga IV-V. Throughout the rest of the species’ distribution at low altitudes in central and southern Europe, the pubescence of the apical terga is usually more nearly white, like that of B. lucorum. There, B. terrestris of authors must be identified using characters such as the number of small punctures in an antero-lateral cluster in the ocello-ocular area of the head (see Kriiger, 1951; Loken, 1973; Pekkarinen, 1979; Rasmont, 1984). B. terrestris of authors can probably be considered conspecific with individuals of this group with other colour patterns that are known from the Canary Islands (described under the name B. ter- restris var. canariensis Pérez) and Madeira 84 (described under the name B. maderensis Erlandsson), because there is no reason to believe that these very similar nominal taxa would not interbreed if they were to occur together. B. terrestris of authors is otherwise known from Morocco (BMNH, PW), Algeria and Tunisia (Kruger, 1956; L¢ken, 1973; BMNH) and from across Europe, Turkey and the Caucasus (Kriiger, 1954, 1956; Lgken, 1973; Panfilov, 1981; BMNH) to Syria (BMNH, PW), Lebanon (BMNH) and the Elburz (Panfilov, 1981; BMNH); and from across the central U.S.S.R. to as far as the Tien Shan (Kruger, 1956; Panfilov, 1981) and Hindu Kush ranges (BMNH). It is perhaps because of the difficulty in distin- guishing continental European specimens of B. lucorum and B. terrestris of authors that a particu- lar effort has been made to describe subtle differences among individuals of the terrestris- group. Even excluding B. terrestris of authors in all of its colour patterns (i.e. considering only B. lucorum in its broadest sense), at least 187 names have been published, although the majority are deemed to be of infrasubspecific rank (for a discussion of the status of the major groups of nominal taxa, see e.g. Kriiger, 1951, 1954, 1956, 1958; Léken, 1973; Pekkarinen, 1979; Erlandsson, 1979; Delmas, 1981; Obrecht & Scholl, 1984; Rasmont, 1984; Rasmont et al., 1986 and refer- ences therein). These samples of the terrestris- group from Europe show colour variation that is only relatively slight for females in comparison with many other species. Colour variation is more pronounced for males, with the frequencies of light and dark individuals differing among local- ities (e.g. Pekkarinen, 1979: fig. 7). This variation in the colour patterns of the males shows some correlation with variation in the compositions of their volatile marking secretions (Bergstrom et al., 1973) and with variation in the mobilities of the enzymes phosphoglucomutase and esterase-1 under electrophoresis (e.g. Scholl & Obrecht, 1983; Pamilo et al., 1984). In northern Europe, samples of the terrestris- group that exclude B. terrestris of authors do show correlated patterns of variation in the colour of the pubescence and in the different phosphoglu- comutase mobilities, although these combinations of character states are not strictly segregated between the two groups of individuals (Pekkarinen, 1979; Pamilo et al., 1984). This apparent recom- bination of character states in some individuals could arise by interbreeding between the different parts of a population of a single species (B. lucorum s.\.). Furthermore, males with both light and dark colour patterns have been reared from the same nest (Lgken, 1961). PAULH. WILLIAMS In central Europe, Scholl & Obrecht (1983) found that individuals differed in the mobilities of not only phosphoglucomutase but also of esterase-1. As in northern Europe, the males in the two groups also tended to differ in the extent of the pale pubescence. Scholl & Obrecht interpreted the lack of individuals with double or hybrid enzyme bands expected from hetero- zygotes, and the lack of the alternative banding pattern between the two enzymes expected from recombinants, as evidence against interbreeding and concluded that two species were present. However, the expected heterozygote individuals with both phosphoglucomutase bands were also absent from the samples from northern Europe (Pamilo et al., 1984), where interbreeding was believed to be occurring (Pekkarinen, 1979). In fact phenotypes with two phosphoglucomutase bands were found in samples from both the Alps and from the Pyrenees, although at least one of the bands had a lower mobility than expected fora heterozygote. Furthermore, the close association between the states of the two enzymes could possibly result from close genetic linkage. But in any case, a definitive interpretation of the differ- ences between these electromorphs can only emerge from studies of genetic inheritance. Scholl & Obrecht (1983: fig. 2) went on to describe differences in body shape between the queens with the two enzyme phenotypes, al- though there was considerable overlap in the variation. Pekkarinen (1979, see fig. 12) con- cluded that similar differences in shape among individuals from northern Europe showed a con- sistent allometric pattern as though they had all been taken from a single population. Scholl & Obrecht (1983; Obrecht & Scholl, 1984) also associated one of the two enzyme phenotypes (form A) particularly with highland samples, both from the Alps (where it was identified as the nominal taxon B. cryptarum) and from the Pyrenees (where it was identified as the nominal taxon B. magnus). However, covariation of size, shape, colour and enzyme mobilities is also known from across Europe within the single species of western honey bee, Apis mellifera Linnaeus (see Ruttner, 1988). I agree with Pamilo et al. (1984), who concluded from a comparison of their results with those of Scholl & Obrecht (1983), that a closer study of the pattern of associations between the colour characters, the enzyme mobilities and the male pheromone compositions of B. lucorum s.]. is still needed to solve the problem. There is no direct evidence to say whether or not males or young queens with any one of the colour patterns of B. lucorums.1|. avoid those with any other colour pattern as potential mates in THE BUMBLE BEES OF THE KASHMIR HIMALAYA their natural habitats. Information on assortative mating in the field is central to the concept of the nature of species that is used for this review. Unfortunately the results of experiments on mating under laboratory conditions may not reflect the discrimination by specific-mate recog- nition systems under natural conditions (see the introduction on discrimination of species). De Jonghe (1982) showed this with bumble bees in cages, where males of B. terrestris of authors would even mate with young queens of B. luco- cryptarum Ball, a nominal taxon that is more closely similar to B. lucorum s.str. Therefore under these artificial conditions neither positive nor negative results necessarily have implications for conspecific status. De Jonghe & Rasmont (1983) attempted to test the fertility (i.e. post- mating isolation) of five artificial crosses between three of the nominal taxa by laboratory experi- ment, using 11 queens that had survived the winter, of which 6 laid eggs, although none hatched. Control crosses were carried out within each of two nominal taxa. A total of 5 of these queens survived the winter, but only one queen laid eggs, which hatched normally. As de Jonghe & Rasmont stated in the discussion of their results, it is perhaps premature to attribute the failure to hatch of the eggs from the experimental crosses to genetic incompatibility on the basis of so few replicates. Nevertheless, more extensive experiments may confirm that such incompati- bility exists. At present it can be concluded that there is a broad set of individuals that appears to be discrete (B. lucorum in its broadest sense). Within this set there is variation in pubescence colour, morph- ology and enzyme mobilities. What is not resolved is whether the patterns of variation reveal entirely discrete subsets or, if they do, whether such patterns necessarily imply a lack of interbreeding between separate populations. Therefore, until more direct evidence against shared specific-mate recognition systems and interbreeding is avail- able, I shall follow many previous authors and continue to treat these bees as parts of a single species, B. lucorum s.1. (see the comments on B. hypnorum). Of course this interpretation does not preclude the idea of some restriction of gene flow among some parts of the population (see Barton, 1988) of B. lucorum s.|., perhaps between the lowland and highland bees in particular. Thus B. lucorum is treated for the present as conspecific with the European taxa described under the names Apis cryptarum (infra-subspecific accord- ing to Lgken, 1966, 1973; Tkalct, 1969b, 1975), B. terrestris var. lucocryptarum and B. lucorum f. magnus (conspecific according to Pekkarinen, 85 1979; Pekkarinen et al., 1981); and conspecific with the Asian taxa described under the names B. magnus Rasse mongolicus, B. lucorum var. alaiensis, B. lucorum subsp. jacobsoni, B. reinigi and B. pratorum var. longipennis. DISTRIBUTION. The present broad concept of B. lucorum leads to the conclusion that it is the most widespread of all bumble bee species, even though, unlike some (see Richards, 1973), it is not quite circumpolar (Map 45). It may include individuals recorded from Europe (e.g. Kriiger, 1951; Lgken, 1973; Alford, 1975; Rasmont, 1984; BMNH, PW), Turkey and the Caucasus (Kriger, 1951; Panfilov, 1981; Rasmont, 1984; BMNH) to the Elburz (Rasmont, 1984; BMNH); and from across the northern U.S.S.R. (Kriger, 1951; Panfilov, 1981; BMNH) to Mongolia (Morawitz, 1880; Vogt, 1909; Kriiger, 1954, 1958; Tkalci, 1974a; Panfilov, 1981; ITZ, PW), Inner Mongolia (Reinig, 1936; Panfilov, 1981; Wang, 1982), Heilongjiang (Kriger, 1954, 1958), North Korea (Kim & Ito, 1987), Hokkaido in Japan (Tkalcu, 1962; Sakagami & Ishikawa, 1969) and Kamchatka (Bischoff, 1930; Kriiger, 1951; Panfilov, 1981; BMNH); and from Alaska almost to Hudson Bay (Milliron, 1971; BMNH, PW). In central Asia its distribution reaches southwards to the Tien Shan (Morawitz, 1880; Skorikov, 1931; Kriiger, 1954; Panfilov, 1957, 1981), the Pamir (Reinig, 1930, 1934; Skorikov, 1931; Kriiger, 1951, 1958; Panfilov, 1981), the Hindu Kush (Reinig, 1940), Pakistan (BMNH, PW), Kashmir, and eastwards to Himachal Pradesh (BMNH, PW), Uttar Pradesh (BMNH), Nepal (Tkalct, 1974b; BMNH, NMS), Sikkim (Friese, 1918; BMNH, MNHU), Tibet (Kriiger, 1951, 1958; Panfilov, 1957; Wang, 1982, 1988; BMNH), northern Burma (BMNH), Yunnan (Wang, 1987), Sichuan (Wang, 1982; PW), Gansu (Bischoff, 1936; Kriiger, 1951, 1958) and Shaanxi (BMNH). Part of this distribution is mapped by Panfilov (1981: map 96). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. lucorum is recorded from the Hindu Raj and Karakoram ranges, and from the southern side of the Great Himalaya range and from the Pir Panjal range, in the upper montane coniferous forest and in subalpine scrub (Map 46, Fig. 6): 14 queens, 78 workers, 93 males, from 12 localities 2100-4600 m (BMNH, NR, PW, ZI). VARIATION WITHIN KASHMIR. As in Europe there is some variation in the extent of the pale pubes- cence and in its hue. Some of the females from Kashmir (Figs 359, 363, 364, 368, 369) broadly resemble British B. Jucorum in colour pattern, but have the yellow band across the anterior of the 86 thorax extending down the sides of the thorax from a third to nearly half of the distance to the leg bases. The pubescence of the face, of the scutellum, and of tergum I is almost completely black and tergum II is often black in the apical quarter. The pale pubescence of the anterior thoracic dorsum of the queen may be lemon- yellow (similar to the holotype of B. reinigi) or grey-cream (lectotype of B. lucorum subsp. jacob- soni and one paralectotype queen, BMNH: both specimens show little wing wear, which is ex- pected to be related to age and fading of yellow pubescence). The possibility that these cream- banded, dark individuals are semi-melanic B. patagiatus has been considered, but the punctures anterior to the lateral ocelli are more widely spaced (cf. Tkalct, 1967: fig. 2). The dark indivi- duals are rare in the Hindu Raj and Karakoram ranges (1/4 queens, 0/26 workers, Fig. 359), but are more common further south in the Great Himalaya range (1/2 queens, 12/27 workers, Figs 363 & 364) and predominate in the Pir Panjal range (7/7 queens, 23/24 workers, Figs 368 & 369). Other females from Kashmir (Figs 356, 357, 360, 361, 366) are lighter in colour, sometimes with pale pubescence on the face, but more generally on the scutellum, tergum I and almost all of tergum II (similar to B. magnus Rasse mongolicus, but with more pale pubescence). The possibility that these extensively pale individuals are B. sporadicus has been considered, but they have more of the large punctures in the ocello- ocular area of the head, especially posteriorly (cf. Tkalci, 1967: fig. 1). The pale hair of the thorax and of tergum II is always lemon-yellow, although it is susceptible to fading among older individuals. Most individuals that have been examined from the Hindu Raj and Karakoram ranges (2/4 queens, 26/26 workers, Figs 356, 357, 360) have this light colour pattern, although they are rare in the Great Himalaya range (0/2 queens, 0/27 workers, Fig. 361) and in the Pir Panjal range (0/7 queens, 1/24 workers collected, Fig. 366, — the frequency is actually much lower: during more than two months of field work on Mt Apharwat in each of two years, many more of the dark workers were seen, but no more of the light workers). There is also variation in details of the morph- ology among these females, though this appears not to correlate closely with variation in colour. For instance, there is subtle variation in the shape and sculpturing of the labrum. These differences allow individual specimens to be recognised, but I found no discrete or reliable morphological differ- ences in this character between individuals with the two extreme colour patterns. Variation in the sculpturing of the ocello-ocular areas also appears PAULH. WILLIAMS to be continuous rather than being ordered in discrete states, and there is no strict association between the states of these morphological and colour characters among individuals. Generally the light individuals have many fine punctures antero-laterally in the ocello-ocular area of the head, where they often form a cluster similar to that of B. terrestris of authors (L@ken, 1973: fig. 12B; Rasmont, 1984: fig. 11). There are often few or none of these fine punctures for the yellow- banded dark individuals, although rather more for the cream-banded dark individuals. But then the four queens from the Hindu Raj range all have these fine punctures, even though only two of them have a light colour pattern (like most of the workers), one of them can be described as inter- mediate (only a very few pale hairs intermixed in the black pubescence of the scutellum and tergum I), and the other has a dark colour pattern (there are slightly fewer punctures on the two darker queens). In contrast, the single light worker from Mt Apharwat in the Pir Panjal range has none of these fine punctures, like most of the dark workers. Much of the material from the Great Himalaya range is more or less intermediate in these characters. For instance, half of the material from Lal Pani (0/1 queen, 11/21 workers) can be described as intermediate in colour pattern and has fine punctures on the head, whereas the remainder (1/1 queen, 10/21 workers) has a dark colour pattern and lacks these punctures. Variation in the colour pattern of the males may be related to the two extreme colour patterns described from the females from Kashmir. The majority of males from the Pir Panjal range (6/9 males, Fig. 370) and some from the Great Himalaya range (38/82 males, Fig. 365) are exten- sively yellow. Other males from the Pir Panjal (3/9 males, Fig. 367), from the Great Himalaya (44/82 males, Fig. 362) and from the Hindu Raj and Karakoram ranges (2/2 males, Fig. 358) are darker, with a broad, black band on the thorax and another on tergum III. Thus the dark males are most frequent where the light or intermediate females predominate in the Karakoram and Great Himalaya ranges, whereas the light males are restricted to localities with the dark females in the Pir Panjal and Great Himalaya ranges. The gonostylus of these males varies considerably in its exterio-lateral length (Figs 194 & 195), but this variation appears to be continuous and uncor- related with the colour pattern of the pubescence so that it is of only individual significance. One interpretation is that the individuals from Kashmir with the extreme light or dark colour patterns of the pubescence could represent at least two separate species, with some variation within THE BUMBLE BEES OF THE KASHMIR HIMALAYA each (Rasmont et al., 1986:678). The bees of this group from north-western Kashmir are more similar to some of those from central Asia, whereas those from south-eastern Kashmir re- semble those from the eastern Himalaya. But my present interpretation of the specimens available from Kashmir is that a large proportion can be classed as intermediates between the extremes of variation (the intermediate female pattern is similar to that described under the name B. lucorum var. alaiensis) without any apparent discontinuity in this pattern of variation. For these bees of the subalpine and upper montane forest zones, the tendency for the frequencies of light and dark colour patterns to change between north- western and south-eastern Kashmir (Map 46) is consistent with regional variation among individuals from a single interbreeding population and a single species. There is perhaps relatively free movement of individuals along the mountain chains, which has resulted in extensive hybridisa- tion and a broad colour cline (cf. comments on B. tunicatus). Some of the factors that may influence the variation in the male colour pattern are discussed in the section on the colour patterns of the Kashmir fauna. Females of B. lucorum are distinctive in their appearance among the fauna of Kashmir. The yellow males with white hairs on the apical terga (Figs 365 & 370) may be similar in colour pattern to the males of B. hypnorum (Figs 313 & 316), which occur with them at some localities. B. lucorum males can be recognised by their larger size, by the numerous black hairs intermixed on the thorax and on terga III-IV and by their outwardly flared heads of the penis valves (Figs 74, 75, 114, 115). FOOD PLANTS. (Leguminosae) Trifolium repens L.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC., unidentified dandelion- like composites; (Gentianaceae) Swertia petiolata D. Don; (Scrophulariaceae) Digitalis lanata Ehrh. [introduced], Pedicularis punctata Decne.; (Labiateae) Prunella vulgaris L. Subgenus SIBIRICOBOMBUS Vogt Sibiricobombus Vogt, 1911: 60 (as a subgenus of Bombus Latreille). Type species: Apis sibirica Fabricius, 1781: 478 = Bombus sibiricus (Fabricius), by subsequent designation of Sandhouse (1943: 599). Obertobombus Reinig, 1930: 107 (as a subgenus of Bombus Latreille). Type species: Bombus oberti Morawitz, 1883: 238, by monotypy. 87 Synonymised with Sibiricobombus Vogt by Richards (1968). Obertibombus Reinig; Reinig, 1934: 167 (as a subgenus of Bombus Latreille). Unjustified emendation (Art. 33b(i, iii)). Sibiricibombus Vogt; Skorikov, 1938a: 145. Un- justified emendation (Art. 33b(i, iii)). For a general description of both sexes see Richards (1968). Bombus (Sibiricobombus) asiaticus Morawitz (Figs 19, 76-78, 116-118, 156-158, 196-198, 227, 371-391, Maps 47 & 48) Bombus hortorum var. asiatica Morawitz in Fedtschenko, 1875: 4. Incorrect original spell- ing (Art. 32c(i)), without separate availability in this form (Art. 32c). LECTOTYPE worker by present designation (see Note 4 below), U.S.S.R.: ?Tadzhikistan $.S.R., “‘Tschiburgan’, 4530-8650 ft [1400-2600 m], 26.vi.1871 (ZMMU) [examined]. Bombus longiceps Smith, 1878a: 8. Syntype workers and male, INDIA: Kashmir, Ladakh, Dras; Kargil; Leh (Stoliczka) (Calcutta) [not seen]. Synonymy with Bombus asiaticus Morawitz suggested by Reinig (1940), con- firmed here. Bombus asiaticus Morawitz; Dalla Torre, 1896: 512. Justified emendation (Art. 33b(ii)). Bombus Regeli Morawitz, 1880: 337. Lectotype queen by designation of Podbolotskaya (in press), CHINA: Xinjiang, Borohoro Shan, near Yining [= Gulja] (Regel/) (ZI) [examined]. Synonymised with Sibiricobombus asiaticus (Morawitz) by Skorikov (1922). [Bombus regeli ab. miniatocaudatus Vogt, 1909: 50 [footnote 1], not of Vogt, 1909: 56 [examined]. Infrasubspecific (Art. 45f(iv)), unavailable name (Art. 45e).] Bombus (Sibiricobombus) regeli miniatocaudatus Vogt, 1911: 61 (see Note 1 below). Holo- type male by monotypy (see Note 5 below), MONGOLIA: ‘Septentrionalis’, no further data (ITZ) [examined]. Junior secondary homonym in Bombus of B. soroeensis var. miniatocaudatus Vogt, 1909. Synonymy with Bombus asiaticus Morawitz suggested by Reinig (1940), confirmed here. [Bombus (Sibiricobombus) regeli form. fusco- caudatus Vogt, 1911: 61 [examined]. Infrasub- specific (see Note 2 below) (Art. 45g(ii)(1)), unavailable name. | [Bombus (Sibiricobombus) regeli form. albocauda- tus Vogt, 1911: 61 [examined]. Infrasub-specific 88 (see Note 2 below) (Art. 45g(ii)(1)), unavail- able name. |] [Bombus (Sibiricobombus) regeli form. tenuifas- ciatus Vogt, 1911: 61 [examined]. Infrasub- specific (see Note 2 below) (Art. 45g(ii)(1)), unavailable name. ] Bombus (Sibiricobombus) miniatocaudatus race falsificus Richards, 1930: 652 (see Note 3 below). Holotype queen by original designation, CHINA: Xizang [= Tibet], Tingri, 14000 ft [4300 ml], 4.vii.1924 (Hingston) (BMNH) [examined]. Syn. n. Sibiricobombus flavodorsalis Skorikov, 1933b: 248. Syntype queens, INDIA: Kashmir, Ladakh, Suru valley (ZI) [not seen]. Junior secondary homonym in Bombus of B. sonorus subsp. flavodorsalis Franklin, 1913 [= B. pen- sylvanicus (Degeer), the correct original spell- ing of B. pennsylvanicus of authors]. Synonymy with Bombus asiaticus Morawitz suggested by Reinig (1940); synonymised with Pyrobombus longiceps (Smith) by Tkalct (1969a). Sibiricobombus oshanini Skorikov, 1933b: 248. Syntype queens and male, INDIA: Kashmir, Ladakh, Nubra valley (ZI) [not seen]. Syn- onymy with Bombus asiaticus Morawitz sug- gested by Reinig (1940); synonymised with Pyrobombus longiceps (Smith) by Tkalcu (1969a). NOMENCLATURE. Note 1. Vogt states (1911: 50, footnote 1) that he intends ‘Varietas geogra- phica’ to be equivalent to subspecies in rank and ‘Aberratio extrema’ to be of infrasubspecific rank. He uses ‘Forma’ for taxa that he could not assign to either rank with certainty. No particular status is specified for the trinomen B. regeli miniato- caudatus. Therefore it is deemed to be of sub- specific rank (Art. 45f(i)). Note 2. The other forms of B. regeli described by Vogt (1911) are all from the same series from ‘Siebenstrom’ and can be regarded as variants of One principal local colour pattern. These indivi- duals were collectively referred to as ‘typischen regeli’. Therefore they are deemed to be of infrasubspecific rank (Art. 45g(ii)(1)). Note 3. Although Richards (1930: 634 etc.) used the term ‘Subsp.’ as well as ‘var.’ and ‘race’, he states that the ‘race’ falsificus is characteristic of a particular geographical area (p. 652: ‘In the Himalayas a much paler form is found, for which I propose the name falsificus, race nov.’). This nominal taxon is therefore deemed to be of subspecific rank (Art. 45f(1i)). TYPE MATERIAL. Note 4. Morawitz’s description of B. hortorum var. asiatica lists several localities and a range of body lengths so that he must have PAULH. WILLIAMS had a series of specimens. I have seen a series of three queens, one large worker and three smaller workers from the ZMMU collection that agree with the original description (all have the pubes- cence of tergum II entirely yellow). The largest of these workers carries (1) a blue label ‘26’; (2) [Chiburgan]; (3) ‘Bombus / hortorum L. / varietas asiatica.” in handwriting identical to that of Morawitz; (4) ‘F. Morawitz / det. 1875.’; (5) ‘26. VI.1871’; (6) [Moscow / Zoological / Museum MGU]. This specimen, which bears data labels consistent with the original description (and is the only specimen to bear the full name on the third label) and is the least abraded and faded, but which lacks the tarsi of both mid legs and of the right hind leg, is designated as lectotype (Art. 74a). Note 5. Vogt’s description of B. regeli ab. miniatocaudatus is rather oblique, but is adequate to imply that it differs from B. regeli s.str. in that the pubescence of the apical terga is red. A single male in the ITZ collection agrees with this descrip- tion (Vogt, 1909, specified that he had a single male with this colour pattern from central Asia) and carries a handwritten red label ‘regeli= Gruppe / miniatocaudatus / Type’. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. A reduction of the ventro-basal angle of the penis valve is shared by a large group of the bumble bees with sickle-shaped, recurved hooks of the penis valve head (synapomorphy). In Kashmir this includes the species of the subgenera Sibiricobombus and Melanobombus (Figs 116— 125). The ventro-basal angle of the penis valve is also reduced for the east Himalayan B. (Festivo- bombus) festivus, although in this case it is likely to be a convergent character state (as opposed to the interpretation in Williams, 1985). In contrast to the species of the subgenera Sibiricobombus and Melanobombus, B. festivus shares a reduction and ventral torsion of the interio-basal process of the gonostylus with the species of the subgenus Pyrobombus, and B. festivus also shares a sub- apical constriction of the volsella with species of the subgenus Rufipedibombus (?synapomorphies, see the comments on B. hypnorum and B. tunicatus). Within the large group of bumble bees that have a reduced ventro-basal angle of the penis valve, the temperate steppe species of the subgenus Cullumanobombus (see Panfilov, 1951) share a reduced outer ridge of the penis valve head, a reduced interio-basal process of the gonostylus, a reduced interio-apical process of the volsella and THE BUMBLE BEES OF THE KASHMIR HIMALAYA an extended outer apical corner of the volsella (synapomorphies). Many of the remaining species that have a reduced ventro-basal angle of the penis valve appear to share a more pronounced narrowing of the volsella towards the apex (?syna- pomorphy, possibly secondarily expanded for B. oberti and for species of the lapidarius-group). These species include two species-groups. Species of the first group share a concave distal margin of the recurved hook of the penis valve head (Figs 79-85) (?synapomorphy). This group includes the oberti-group of the subgenus Sibiricobombus and the species of the subgenus Melanobombus. The centre of diversity for this group is near Kashmir, with divergent species- groups in the Oriental region and in Europe and northern Asia (see the comments on B. oberti, B. ladakhensis and B. simillimus). Species of the second group share a thickening of the volsella in its apical part, which is delimited by a strongly raised ridge. This ridge rises below the inner dorsal margin and curves back across the ventral surface, often forming a pronounced and coarsely-sculptured swelling or tubercle midway along the volsella (Figs 156-158) (synapomorphy). This second group includes the Mongolian B. sibiricus and the central Asian montane asiaticus- group of the subgenus Sibiricobombus, which reach westwards in the Caenozoic mountains to south-eastern Europe, as well as the large fraternus-group (in its broadest sense) from North America and the mountains of western South America (see below). If this interpretation of these character states is accepted as the most likely estimate of relation- ships available at present, then one part of the subgenus Sibiricobombus would appear to be paraphyletic with respect to one monophyletic clade, whereas the other part would appear to be paraphyletic with respect to another monophy- letic clade. Consequently the subgenus Sibirico- bombus may be a polyphyletic assemblage, which would account for its apparent lack of exclusive synapomorphies. Yet females of the species that are currently placed in this subgenus do show some phenetic similarity because they usually have particularly many of the short, branched hairs that continue onto the outer surface of the hind basitarsus, below the postero-proximal corner (Figs 227 & 228, see the key, couplet 11). The use of the subgeneric name Sibiricobombus for what appears to be an artificial group with both B. asiaticus and B. oberti is continued here only for the sake of stability. I consider it undesirable to take any nomenclatural action until the males of B. flaviventris and B. tanguticus are available to clarify the relationships among the species- 89 groups (see the comments on B. oberti and B. simillimus). Within the group of species that have both a reduced ventro-basal angle of the penis valve and a thickened volsella, B. sibiricus has an expanded gonostylus (plesiomorphic among the species of Sibiricobombus) (see Skorikov, 1931: fig. 27; Tkalct, 1974a: fig. 4) and the volsella is only weakly thickened. B. sibiricus (of the ‘sibiricus- Gruppe’ of Tkalci, 1974a) is known from the area around Mongolia (Skorikov, 1931: fig. 7). The eyes of the male are similar in relative size to those of the female. Apart from B. sibiricus and B. oberti, those males that are known from the other species of the subgenus Sibiricobombus all have very much enlarged eyes relative to those of the females. Also within the group of species that have both a reduced ventro-basal angle of the penis valve and a thickened volsella, the species of the New World fraternus-group in its broadest sense share a reduction of the inner apical corner of the gonostylus and the thickened part of the volsella is usually more pronounced (?synapomorphies). These New World species are currently separ- ated among what seems to be an unnecessarily large number of subgenera, most of which have very few species: Fraternobombus Skorikov, Separatobombus Frison, Crotchiibombus Franklin, Brachycephalibombus Williams, Robustobombus Skorikov and Rubicundobombus Skorikov; the volsella is further modified by a pronounced reduction of the exterio-apical corner for the other species of this group, all from Central America and the mountains of western South America, that are currently placed in the sub- genera Dasybombus Labougle & Ayala (1985: fig. 5), Funebribombus Skorikov and Coccineobombus Skorikov (which includes B. handlirschi Friese) (?synapomorphy). Also within the group of species that have both a reduced ventro-basal angle of the penis valve and a thickened volsella, the species of the asiaticus-group share a concavity of the outer margin of the middle part of the volsella, and the thickened part of the volsella is more pronounced particularly at the basal end, where it forms a rounded swelling or tubercle (Figs 156-158) (2synapomorphies). This tubercle may increase the contact area of the distal, coarsely-sculptured area of the male volsella for gripping the female sting base during copulation. This asiaticus-group includes the species of the ‘niveatus-Gruppe’ of Tkalcii (1974a), the species of the ‘P. asiaticus- Gruppe’ of Tkalcti (1969a, or ‘longiceps-Gruppe’ of Tkalci, 1974a, 1974b) and B. obtusus of the ‘morawitzi-Gruppe’ of Tkalct (1969a, 1974a, but 90 not B. morawitzi Radoszkowski, see the com- ments on B. oberti). Among the nominal taxa of the asiaticus-group, B. obtusus Richards from the Hindu Kush ranges has a broadly rounded gonostylus with a broad interio-basal process, and the inner margin of the head of the penis valve has only a single, straight dorsal ridge (plesiomorphic within Sibiricobom- bus). Three worker paratypes (BMNH) that are from the same locality as the holotype male (Richards, 1951) have few of the short, branched hairs arising from the outer surface of the hind basitarsus, so that this area appears to be more shining than for B. asiaticus, and the pubescence is generally shorter. These characteristics of B. obtusus remain segregated where it is known to occur with B. asiaticus, without any evidence of recombination in the small samples that are available. The males of the remaining species of the asiaticus-group share a slight reduction of the outer apical corner of the gonostylus, and the interio- dorsal ridge at the base of the penis valve head curves from its dorsal origin to the ventral margin before it joins the recurved hook (synapomorphies of B. asiaticus + B. niveatus [+ B. vorticosus] + B. sulfureus). Males of the many nominal taxa that are similar to B. asiaticus share a narrowing of the volsella towards the apex from the inner margin, so that it appears even more strongly inwardly- curved (Figs 156-158) (apomorphy). The males of B. niveatus Kriechbaumer, B. vorticosus Gerstaecker and B. sulfureus Friese share the ventrally-curved form of the interio- dorsal ridge at the base of the penis valve head, but also show an unusually pronounced develop- ment of an interio-ventral ridge at the base of the penis valve head (synapomorphy) (see Skorikov, 1931: fig. 29). B. niveatus, B. vorticosus and B. sulfureus also retain the tubercle of the volsella despite a subapical expansion of the volsella (synapomorphy) (see Skorikov, 1931: fig. 29), which is most pronounced for B. sulfureus (?auta- pomorphy). The only known differences between B. niveatus and B. vorticosus are in colour (Vogt, 1909; Pittioni, 1938). B. niveatus has the bands of light pubescence grey-white, whereas they are yellow for B. vorticosus. B. vorticosus is broadly distributed between Krasnovodsk and the Elburz mountains in the east to Yugoslavia in the west (e.g. Vogt, 1909, 1911; Pittioni, 1938; Reinig, 1967, 1971, 1974; Tkalci, 1969b; BMNH, PW). B. niveatus apparently occurs only within the more central part of this area, where it is less abundant (e.g. Vogt, 1909; Pittioni, 1938; Reinig, 1967, 1971, 1974; D. B. Baker, pers. comm.; BMNH, PW). Reinig (1967) described some PAULH. WILLIAMS individuals as possible hybrids and it may be that the banded white individuals and the banded yellow individuals are both parts of a single species, B. niveatus (cf. the comments on yellow or white B. keriensis). B. sulfureus is a rare species that occurs from the Elburz mountains to Turkey (e.g. Reinig, 1971, 1974; BMNH). Reinig (1940) suggested that the many similar nominal taxa of the central Asian asiaticus-group (apart from B. obtusus) might actually be indivi- duals with different colour patterns within a single species, B. asiaticus, but he lacked sufficient material, especially of the males, to draw firm conclusions. Individuals of the asiaticus-group from Ladakh have an unbanded, yellow thorax (Figs 371-376, described under the names B. longiceps Smith; Sibiricobombus flavodorsalis Skorikov; Sibiricobombus oshanini Skorikov). Individuals from the the other side of the Great Himalaya range in the Vale of Kashmir usually have a grey-white thorax with a black band between the wing bases (Figs 383-391, referred to as B. callophenax Cockerell by Richards, 1930: 652, a misidentification, see the comments on B. avinoviellus). There is also variation in the extent of pale pubescence on terga I-II and in the colour of the wings, which are usually more clouded with brown (infuscated) among specimens from the Vale of Kashmir. There is variation in the shape of the male genitalia, especially as to whether the gonostylus has the apical margin convex or concave and whether or not the interio-basal process is constricted near its base (Figs 196-198). However, individuals with both principal colour patterns show all forms of the gonostylus, so the variation appears to be of only individual significance. The male mate-searching behaviour of the unbanded yellow individuals and of the banded white individuals could not be distinguished (see the introduction on male mate-searching be- haviour and on the inference of allopatric, con- specific taxa). Individuals with both regional colour patterns occupy similar open habitats. The males all perch to watch for potential mates in a similar way and race in pursuit of moving objects (but without holding exclusive territories, cf. the comments on B. rufofasciatus). At a few localities between the two regions where individuals with each of the principal colour patterns predominate, other individuals have been collected that have colour patterns with combinations of the charac- ter states of both regional colour patterns (Figs 377-382, see the comments below on variation within Kashmir). This pattern of variation is consistent with hybridisation between individuals with divergent colour patterns within a single THE BUMBLE BEES OF THE KASHMIR HIMALAYA population (see the comments on variation below and the discussion section on the variation of B. asiaticus in Kashmir). Therefore these, and probably the many other nominal taxa of the central Asian asiaticus-group for which inter- mediates are known (i.e. apart from B. obtusus), including the banded and pale-yellow B. miniato- caudatus race falsificus from southern Tibet, are likely to be interbreeding as parts of a single species. Three further nominal taxa (B. huangcens, B. heicens and B. xionglaris), similar to B. miniatocaudatus race falsificus, have been des- cribed by Wang (1982) from Tibet (not seen). DISTRIBUTION. B. asiaticus is a central Asian and Tibetan species (Map 47). It is known from Mongolia (Vogt, 1909; ITZ, PW), the Altai (Skorikov, 1931), the Tien Shan (Morawitz, 1875, 1880; Vogt, 1911; Skorikov, 1931; Panfilov, 1957; BMNH, ZI, ZMMU), the Pamir (Friese, 1913; Reinig, 1930, 1934; Skorikov, 1931; BMNH), the Hindu Kush (Reinig, 1940; Richards, 1951; Tkalci, 1969a; BMNH), Pakistan (BMNH, PW), Kashmir, Himachal Pradesh (BMNH, PW), Uttar Pradesh (BMNH), Nepal (Tkalct, 19746; BMNH, NMS), Tibet (Richards, 1930; Wang, 1982; BMNH), Qinghai (Bischoff, 1936; Tkalct, 1961; Wang, 1982; BMNH, ZS) and Gansu (Bischoff, 1936). MATERIAL EXAMINED FROM KASHMIR. B. asiaticus is widespread in Kashmir, recorded from the Hindu Raj, Karakoram, Ladakh, Zanskar, Great Himalaya and Pir Panjal ranges, down to the margins of the valley floor in the Vale of Kashmir. Although it is widespread among open habitats, it is most abundant in large meadows in the montane forest zone (Map 48, Fig. 6): 156 queens, 428 workers, 434 males, from 42 localities 1800— 4800 m (AB, BMNH, BPBM, FA, IZ, MI, MNRHN, NR, PW, RH, SEMK, TL, ZM). Two colonies were found at Leh (3500 m, 8.vii. 1980, 19.ix.1986). The nest entrances were between rocks in dry-stone walls. VARIATION WITHIN KASHMIR. Even within Kashmir there is great variation in the colour pattern of this species (Figs 371-391, Map 48), often even within samples from a single locality. Only the principal aspects of this variation are described here. All of the individuals from the Hindu Raj, Karakoram, Ladakh and Zanskar ranges, and from across the Great Himalaya range from the northern side to at least as far as Gumri, have the pubescence of the thorax yellow without a black band between the wing bases (Figs 371-376). These individuals from the arid region show relatively little variation, except in the replace- 91 ment of black pubescence with yellow on terga I-II (Figs 374-376), which is especially pronounced for all of the specimens from central Zanskar (2 queens, 1 worker, 3 males). Individuals with this unbanded yellow colour pattern are also known from the Hindu Kush ranges, but not from Tibet, where individuals have the thoracic pubescence yellow but with a black band between the wing bases. Most of the individuals from the Kishanganga valley (16/16 queens, 80/83 workers, 105/114 males) and from around the Vale of Kashmir (24/26 queens, 132/198 workers, 42/82 males) have an obvious band of black hairs between the wing bases (Figs 377-382, 386-391). Most of these specimens have some pale pubescence on terga I-II (Figs 377-389). Some of the individuals from the Kishanganga valley (9/16 queens, 35/83 workers, 4/114 males, Fig. 381) and almost all of those from around the Vale of Kashmir (24/26 queens, 197/198 workers, 37/82 males, Figs 383— 387, 389-391) have the pale pubescence grey- white rather than yellow. Individuals from the more mesic region of the western Himalaya to at least as far east as Nepal also have the pale pubescence grey-white with a black band between the wing bases. Individuals from a few localities in the higher valleys of the Great Himalaya range show com- plex variation with intermediates between the unbanded yellow colour pattern that is common in Ladakh, and the banded white colour pattern that is common in the Vale of Kashmir. This is especially evident among the material from Lal Pani in the upper Kishanganga valley (Table 7, e.g. Figs 380-382), and from Nigagar in the upper Sind valley (e.g. Figs 377-379). These two samples are also unusual because they include specimens that have the red pubescence of the apical terga largely or completely replaced by black (Lal Pani: 2/5 queens, 4/17 workers, 48/92 males, Figs 380 & 381; Nigagar: 0/2 queens, 2/12 workers, 2/6 males, Fig. 379). Individuals with this colour pattern are otherwise known only from the Hindu Kush ranges (BMNH). For an analysis of these data, see the discussion of the variation and genetics of B. asiaticus in Kashmir. The individuals with an unbanded yellow thor- acic dorsum (Figs 371-376) are closely similar in colour pattern to some of the B. avinoviellus (Fig. 232), B. marussinus (Figs 254-257), B. branickii (Fig. 265), B. subtypicus (Figs 319-324), B. biroi (Figs 335-337) and B. semenovianus (Figs 397— 402) that occur with them at some localities. The banded white individuals of this species (Figs 381, 386, 387, 389-391) are closely similar in colour pattern to some of the B. avinoviellus (Figs 238- 92 241), B. kashmirensis (Figs 305-307), B. biroi (Fig. 342), B. keriensis (Figs 408 & 409) and to the queens and some males of B. tunicatus (Figs 350, 353, 355) that occur with them at some localities in the Vale of Kashmir. Females of B. asiaticus can be recognised by their very long oculo-malar distance, by the many fine punctures between the compound eye and the base of the mandible and by the many short hairs at the base of the hind basitarsus (Fig. 227, see the key, couplet 11). The males are easily recognised by their combination of greatly enlarged eyes and very long antennae. FOOD PLANTS. Kashmir: (Balsaminaceae) Impatiens glandulifera Royle; (Leguminosae) Trifolium repens L., Lupinus sp. [introduced], unidentified pink clover-like legume; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC.., unidentified green-flowered thistle-like compo- site; (Scrophulariaceae) Digitalis lanata Ehrh. [introduced], D. purpurea L. [introduced], Pedi- cularis punctata Decne.; (Labiateae) Prunella vulgaris L. Ladakh: (Leguminosae) Medicago falcata L., Trifolium repens L., Caragana versicolor (Wallich) Benth.; (Compositae) Echinops cornigerus DC.; (Scrophulariaceae) Verbascum thapsus L.; (Labia teae) Stachys tibetica Vatke, Nepeta podostachys Benth., Prunella vulgaris L. Bombus (Sibiricobombus) oberti Morawitz (Figs 20; 795 119° 1595 1995 228, 230, 392, 393; Maps 49 & 50) Bombus Oberti Morawitz, 1883: 238. LECTO- TYPE male by present designation (see the Note below), U.S.S.R.: Kazakhstan S.S.R., mountains near Alma Ata (Kuschakewitsch) (ZI) [examined]. Bombus Semenovi Morawitz, 1886: 198. Lecto- type queen by designation of Podbolotskaya (in press), CHINA: Qinghai, valley of the Huang He [= Yellow River], 13500 ft [4100 m] (Prshewalski) (Z1) [examined]. Syn. n. Bombus (Subterraneobombus) duanjiaoris Wang, 1982: 444. Holotype worker by original desig- nation [not seen]. Paratype worker, CHINA: Xizang [= Tibet], Rutog, 5100-5400 m, 23.viii. 1976 (Huang) (IZ) [examined]. Syn. n. TYPE MATERIAL. Note. Morawitz described B. oberti from males from [p. 240] ‘Im Hochgebirge bei Wernoye’. At the end of this description he noted that a queen, which he also described, was most probably of the same species (p. 240: “Als Weib- chen gehort h6échst wahrscheinlich zu vorstehend beschriebenem Mannchen folgendes.’). How- PAULH. WILLIAMS ever, because this specimen was only doubtfully included in the taxon, it cannot be considered to be a syntype. This is unfortunate, because Skorikov (1931) subsequently used the term ‘Typus’ in reference to a ‘defekte’ queen of Morawitz’s B. oberti from ‘Vernyj’ [= Alma Ata, in the Tien Shan region], which might otherwise have been taken to be a valid lectotype designa- tion (Art. 74b). Skorikov also mentions four males with the same locality labels “Vernyj’, but states that this species was never collected there again. The existence of several syntype males is indicated in the original description of B. oberti by the range of size measurements. I have seen a male from the ZI collection that agrees with the original description, carries a label ‘Werkoye / Oberti’ and lacks the right antennal flagellum, which is designated as lectotype (Art. 74a). AFFINITIES. Within the subgenus Sibiricobombus (see the comments on B. asiaticus), B. oberti shares with B. morawitzi (see Skorikov, 1931: fig. 28) a constriction of the recurved hook of the penis valve head that is shown to a lesser extent by many species of the subgenus Melanobombus (Figs 79-85) (?synapomorphy of B. morawitzi + B. oberti + Melanobombus). But only B. mora- witzi and B. oberti share the strongly curved form of this narrowed hook (Fig. 79) (?synapomorphy). B. morawitzi (of the ‘morawitzi-Gruppe’ of Tkalci, 1969a, 1974a, in part) is known from the Tien Shan (Morawitz, 1883; Skorikov, 1931; BMNH), the Pamir (Reinig, 1930, 1934; Skorikov, 1931; BMNH) and the Hindu Kush ranges (Reinig, 1940; Tkalct,, 1969a). Unlike B. oberti, the males have greatly enlarged eyes relative to the females. Another species that may be closely related has been recorded from just across the Tibetan border from Ladakh by Wang (1982: Shazia). Friese (1905) described some females from Qinghai and Gansu that have the pubescence of terga I-III yellow and of terga IV-VI black, under the name B. flaviventris. I have seen one of Friese’s queens labelled ‘Kukunor’ (MNHU), but not the putative male he described later (Friese, 1909) under the same name. Richards (1930) described a series of females from southern Tibet under the name B. (Subterraneobombus) flaviventris subsp. ochrobasis. Friese’s queen (MNHU) and all of Richards’s material in the BMNH appears to belong to the subgenus Sibiricobombus, although no males are present from which to confirm this. These females have the malar area extensively punctured, the lateral ocellus is separated from the dorsal margin of the compound eye by less than two ocellar diameters, the outer surface of the hind tibia is coarsely sculptured, the many THE BUMBLE BEES OF THE KASHMIR HIMALAYA short hairs at the base of the hind basitarsus extend onto the outer surface, which has a particu- larly dense covering of short branched hairs, and tergum VI has a shallow apical notch and a subapical rounded boss. In all of these characters they resemble females of B. morawitzi most closely, but differ in that the clypeus is nearly smooth, with only a few, very widely-spaced punctures and the subapical boss of tergum VI is rounded rather than V-shaped. B. flaviventris has also been recorded from Tibet by Wang (1982) and from Qinghai by Skorikov (1931) and Panfilov (1957). B. oberti is morphologically very distinct within the subgenus Sibiricobombus, both in the charac- teristics of the female (see the key) and in the form of the male volsella, which has a produced exterio- apical corner but is broadly rounded interio- apically with a very reduced interio-apical process (Fig. 159; Reinig, 1930: fig. 19; Skorikov, 1931: fig. 30). Reinig (1930) separated this species from the other species of the subgenus Sibiricobombus in the monobasic subgenus Obertobombus (the ‘oberti-Gruppe’ of Tkalct, 1974a) (the identity of Reinig’s material was discussed by Skorikov, 1931; Reinig, 1934), although this name has since been regarded as a synonym of Sibiricobombus (e.g. Richards, 1968; Ito, 1985). Any further nomenclatural action is considered undesirable at least until the males of B. flaviventris and B. tanguticus (see the comments on B. simillimus) are available to elucidate the relationships of the oberti-group to the species of the subgenus Melanobombus. Skorikov (1931) found scarcely any difference between morphometric characters of B. oberti and B. semenovi. I can find no evidence that these nominal taxa are not parts of the same species. A worker from just across the Chinese border from Nimaling in western Tibet (IZ), which was des- cribed under the name B. duanjiaoris by Wang (1982), has a broad black band on tergum III like queens of B. oberti from Qinghai (ZI). The Tibetan worker has a shorter antennal segment 3 than for these queens, and the punctures of the malar area are weaker. In these characters, as well as in the colour pattern of the pubescence, it resembles the queens of B. tanguticus from southern Tibet (see the comments on B. simil- limus), although they have the clypeus much more coarsely sculptured, the labral tubercles are strongly pointed and only narrowly separated, and the oculo-malar distance is distinctly shorter than the breadth of the mandible at its base. Otherwise B. duanjiaoris is closely similar to the queens of B. oberti from Ladakh and Qinghai in characters such as the pubescence of the hind 93 basitarsus. Thus from the evidence available at present, B. duanjiaoris is most likely to be part of the same species as B. oberti, despite the small morphological differences, which may be related allometrically to body size. Wang (1982: 447) described another queen from western Tibet under the name B. zhadaensis as ‘closely similar’ to the type specimens of B. duanjiaoris, but with tergum III red (not seen). Individuals of this nominal taxon may also be conspecific with B. oberti. DISTRIBUTION. B. oberti is a central Asian and Tibetan species (Map 49). It is known from the Tien Shan (Morawitz, 1883; Skorikov, 1931; ZI), the Pamir (Reinig, 1930, 1934; Panfilov, 1957; PW), Kashmir, Tibet (Wang, 1982; IZ) and Qinghai (Morawitz, 1886; Panfilov, 1957; ZI). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. oberti is recorded only from the Zanskar mountains, in the high, arid alpine steppe (Map 50): 18 queens, Nimaling plain terminal moraine, 4800 m, 15—24.vii.1980 (Williams) (BMNH, IZ, PW). VARIATION WITHIN KASHMIR. The queens from Nimaling generally have the black pubescence of tergum III largely replaced by orange-red, so that only a very few black hairs remain at the sides (Fig. 392). At most the black pubescence on tergum III is intermixed with the orange-red, except in a narrow, basal band (2/18 queens, Fig. 393). This species is similar in colour pattern to the B. himalayanus (Figs 242 & 243), B. kashmirensis (Figs 295-300) and B. ladakhensis (Figs 394-396) that are known to occur with it. It is easily recognised by its large size and by the characters given in the key. FOOD PLANTS. (Leguminosae) Caragana_ versi- color (Wallich) Benth. Subgenus MELANOBOMBUS Dalla Torre Melanobombus Dalla Torre, 1880: 40 (as a sub- genus of Bombus Latreille). Type species: Apis lapidaria Linnaeus, 1758: 579 = Bombus lapi- darius (Linnaeus), by subsequent designation of Sandhouse (1943: 569). Lapidariobombus Vogt, 1911: 58 (as a subgenus of Bombus Latreille). Type species: Apis lapi- daria Linnaeus, 1758: 579 = Bombus lapidarius (Linnaeus), by subsequent designation of Sandhouse (1943: 562). Synonymised with Melanobombus Dalla Torre by Milliron (1961). Kozlovibombus Skorikov, 1922: 152. Type 94 species: Bombus kozlovi Skorikov, 1910b: 413 [= Bombus keriensis Morawitz] in the sense of Skorikov, 1922 [based on misidentified males = Bombus pyrosoma Morawitz, see Reinig (1934: 169), requiring designation by the ICZN (see Note below)], by subsequent fixation of Sandhouse (1943: 561). Synonymised with Lapidariobombus Dalla Torre by Bischoff (1936). Kozlowibombus Skorikov; Bischoff, 1936: 10 (as a subgenus of Bombus Latreille). Unjustified emendation (Art. 33b(i, iii). Lapidariibombus Vogt; Skorikov, 1938a: 145. Unjustified emendation (Art. 33b(i, i1i)). Tanguticobombus Pittioni, 1939c: 201 (as a sub- genus of Bombus Latreille). Type species: Bombus tanguticus Morawitz, 1886: 200, by original designation. Synonymised with Melano- bombus Dalla Torre by Richards (1968). Note. Application of Kozlovibombus is a matter that should mandatorially be referred to the ICZN (Art. 70b). Since this name is treated here as a synonym of Melanobombus (a position that is not affected by either interpretation of the type species), I intend to take no action. For a general description of both sexes see Richards (1968). Bombus (Melanobombus) ladakhensis Richards (Figs 80, 120, 160, 200, 394-396, Maps 51 & 52) Bombus (Lapidariobombus) rufofasciatus var. ladakhensis Richards, 1928b: 336 (see Note 1 below). Holotype queen [not a worker] by monotypy (see Note 2 below), INDIA: Kashmir, Ladakh, Chushul, vi.1925 (Meinertzhagen) (BMNH) [examined]. Bombus (Lapidariobombus) rufofasciatus var. phariensis Richards, 1930: 642 (see Note 1 below). Holotype queen by original designa- tion, CHINA: Xizang [= Tibet], Phari to Gyangze, 13000-15000 ft [4000-4600 ml, vi. 1904 (Walton) (BMNH) [examined]. Change of status to Pyrobombus ladakhensis subsp. phariensis (Richards) by Tkalct (19745). Syn. n. Bombus variopictus Skorikov, 1933b: 248. Lecto- type by designation of Podbolotskaya (in press) [not seen]. Paralectotype queen, CHINA: Qinghai, valley of the Huang He [= Yellow River], 13500 ft [4100 m] (Prshewalski) (ZI) [examined]. Synonymised with Pyrobombus ladakhensis subsp. phariensis (Richards) by Tkalct (19746). Syn. n. Bombus variopictus subsp. bianchii Skorikov, 1933b: 248. Syntype queen and worker, INDIA: PAULH. WILLIAMS Kashmir, Ladakh, Rupshu (ZI) [not seen]. Synonymised with Pyrobombus ladakhensis subsp. ladakhensis (Richards) by Tkalct (19745). Bombus (Pratobombus) reticulatus Bischoff, 1936: 7. Lectotype queen by designation of Tkalct (1974: 336) [not seen]. Two paralecto- type workers, CHINA: Gansu, Min Shan, ‘Drakana’, 3100-4000 m, 27 & 29.vii.1930 (MNHU) [examined]. Synonymised with Pyro- bombus ladakhensis subsp. phariensis (Richards) by Tkalct (1974b). Syn. n. Bombus (Lapidariobombus) rufo-fasciatus [subsp.] ladakhensis Richards; Tkalcu, 1961: 353. Pyrobombus (Melanobombus) ladakhensis (Richards); Tkalct,, 1974b: 335. NOMENCLATURE. Note 1. Richards’s use (19285: 333 etc.; 1930: 634 etc.) of the term ‘Subsp.’ in both papers is taken to indicate that infrasub- specific rank is meant for B. rufofasciatus var. ladakhensis and for B. rufofasciatus var. pharien- sis. However, these names have subsequently been treated as names in the species group by Tkalct’ (1961, 1974b), so Richards’s taxa are therefore deemed to be of subspecific status (Art. 45g(ii)(1)). TYPE MATERIAL. Note 2. Richards’s description of B. rufofasciatus var. ladakhensis specifies that a single female was examined. A single queen in the BMNH collection agrees with the original descrip- tion and carries the data quoted, together with a label ‘B. rufofasciatus. Sm. / var. ladakhensis Richards / female. type’ in handwriting identical to that of Richards. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). AFFINITIES. The species of the subgenus Melano- bombus share a reduction in the outer ridge of the penis valve head (Figs 80-85) and a shortened form of the gonostylus with a reduced interio- basal process (Figs 200-205) (synapomorphies, see the comments on B. asiaticus and B. oberti). Females often have short, branched hairs below the postero-proximal corner of the hind basitarsus (e.g. Fig. 229), although these hairs are not as long, erect or numerous as for species of the subgenus Sibiricobombus (Figs 227 & 228, see the key, couplet 11). Within the subgenus Melanobombus, the species of the /apidarius-group (the ‘lapidarius- Gruppe’ of Tkalct, 1974b, 1989) share a further reduction of the interio-basal process of the gono- stylus to a small, rounded knob (Figs 200-202) (?synapomorphy). In contrast, the species of the rufofasciatus-group share a further reduction of the apex of the gonostylus (Figs 203-205) (syna- pomorphy) (see the key, couplet 24). THE BUMBLE BEES OF THE KASHMIR HIMALAYA Within the /apidarius-group, all of the species with the exception of B. ladakhensis share a much narrowed form of the penis valve shaft, in lateral aspect (Figs 120-122), and a narrower interio- apical process of the volsella (Figs 160-162) (synapomorphies). Males of B. ladakhensis differ from those of all other species of the subgenus Melanobombus by the recurved hook of the penis valve head, which is broadly fused to the shaft (Figs 80-82) (autapomorphy). A queen of B. ladakhensis from Uttar Pradesh (BMNH) shares the yellow colour pattern with individuals from Kashmir and from neighbouring parts of the far west of Tibet (Wang, 1982). I have seen much more material from southern Tibet, for which the yellow of the pubescence is usually replaced by cream or grey-white (described under the name B. rufofasciatus var. phariensis Richards). This grey-white colour pattern re- sembles B. rufofasciatus. Skorikov (1933b) in- dependently redescribed individuals with the yellow colour pattern under the name B. vario- pictus subsp. bianchii and redescribed those with the white colour pattern under the name B. variopictus s.str. Individuals with the white colour pattern were again redescribed shortly afterwards under the name B. reticulatus by Bischoff (1936). All of these females are closely similar in morph- ology and are considered to be parts of the same species. DISTRIBUTION. B. ladakhensis is a Tibetan species (Map 51). It is known from Kashmir, Uttar Pradesh (BMNH), Nepal (Tkalct, 1974b; NMS), Sikkim (BMNH), Tibet (Richards, 1930; Wang, 1982; BMNH, MNHU), Qinghai (Morawitz, 1886; Skorikov, 1933b; Wang, 1982; ZI) and Gansu (Skorikov, 1933b; Bischoff, 1936; MNHU). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. ladakhensis is recorded from the Zanskar and Ladakh ranges, in the high, arid alpine steppe (Map 52): 3 queens, 1 male, from 3 localities 3600-4800 m (BMNH, PW). VARIATION WITHIN KASHMIR. The queen from Chushul has the pale bands of the thorax and of terga I-II yellow (Fig. 394). The hairs of tergum III are orange, whereas those of tergum IV are orange with white tips and those of tergum V are white. There are few black hairs on these terga and those that are present are mainly on the lateral parts of tergum III. The two queens from Nimaling share a similar yellow pattern, but the longer hairs of terga IV-V are pinkish brown to black basally with white tips, intermixed with many black hairs (Fig. 396). Black hairs are also more numerous for these two individuals on the 95 apical part of tergum II and laterally on tergum III. The male has orange-red pubescence on terga III-V (Fig. 395). This species is similar in colour pattern to some of the B. himalayanus (Figs 242 & 243), B. kashmirensis (Figs 295-300) and B. oberti (Figs 392 & 393) that are known to occur with it. It is easily recognised by the mixture of grey-white and black hairs, rather than orange-red hairs, on terga IV-VI. FOOD PLANTS. (Leguminosae) Caragana_ versi- color (Wallich) Benth.; (Labiateae) Mentha longifolia (L.) Hudson. Bombus (Melanobombus) semenovianus (Skorikov) (Figs 81, 121, 161, 201, 229, 331, 397-402, Maps 53 & 54) Lapidariobombus semenovianus Skorikov, 1914a: 127. Lectotype by designation of Podbolotskaya (in press) [not seen]. Paralectotype queen, INDIA: Kashmir, Ladakh, Stakmo pass, 13.vii. 1912 (Jacobson) (Z1) [examined]. Bombus (Lapidariobombus) lapidarius subsp. problematicus Bischoff, 1935: 255. Holotype worker by monotypy, INDIA: Kashmir, Ladakh, Lamayuru Gompa, 3300 m, 1.viii. 1930 (7MNHUV) [not seen]. Synonymised with Bombus semenovianus (Skorikov) by Reinig (1940). Bombus semenovianus (Skorikov); Reinig, 1935: S22. AFFINITIES. Within the subgenus Melanobombus, B. semenovianus belongs to the lapidarius-group of species (see the comments on B. ladakhensis). Within the /apidarius-group, B. semenovianus may be most closely related to B. ladakhensis (see the comments on B. ladakhensis). Both of these species have the hook of the penis valve head elongated and recurved to the shaft at an angle of less than 45° (Figs 80 & 81). All the other species of the /apidarius-group have the penis valve head slightly reduced in size (smallest for B. lapidarius) with the recurved hook forming an angle with the shaft of 45° or more (Fig. 82) (?synapomorphy). Males of B. semenovianus can be distinguished from males of all the other species of the lapidarius-group, including B. ladakhensis, by their enlarged eyes (relative to those of the females) and by the shape of the gonocoxite, which is nearly parallel-sided, rather than con- stricted, just proximal to the apex. The females of B. semenovianus have the ocello-ocular area more densely punctured than for the other species of the 96 lapidarius-group, with a continuous, broad band of fine punctures along the eye margin. Bischoff (1935) apparently redescribed B. seme- novianus under the name B. lapidarius subsp. problematicus from a single worker from Ladakh. DISTRIBUTION. B. semenovianus is a central Asian species (Map 53). It is known from the Hindu Kush (Reinig, 1940; Richards, 1951; Tkalci, 1969a; BMNH), Pakistan (Frison, 1935; BMNH, PW) and Kashmir. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. semenovianus is recorded from the Hindu Raj, Karakoram, Ladakh, Zanskar and Great Himalaya ranges, in dry subalpine steppe and in high, subtropical semidesert (Map 54): 72 queens, 202 workers, 368 males, from 21 localities 2400-3700 m (BMNH, NR, PW, ZI, ZS). A colony was found near Dras (3200 m, 10.viii. 1986). The nest entrance was under the edge of the corrugated iron roof of a hut. VARIATION WITHIN KASHMIR. This species shows obvious variation only in the extent of the replace- ment of black by yellow in the pubescence of terga I-II (Figs 397-402). Extensive yellow on these terga is very common among males from all localities (Figs 399 & 402), but among the females it is only shown by a few queens from Gilgit (3/51 queens have an obvious yellow band on tergum I, Fig. 397). This species is similar in colour pattern to some of the B. avinoviellus (Fig. 232), B. marussinus (Figs 254-257), B. branickii (Fig. 265), B. sub- typicus (Figs 319-324), B. biroi (Figs 335-337) and B. asiaticus (Figs 371-376) that may occur with it at some localities. It can be recognised by the combination of the presence of many punctures in the ocello-ocular area of the head, the shining outer surface of the hind tibia and the lack of many short hairs at the base of the hind basitarsus (Fig. 229, see the key, couplet 11). The males can be recognised by their short antennae, by their slightly enlarged eyes and by the recurved hooks on the heads of the penis valves (Fig. 81). FOOD PLANTS. (Leguminosae) Melilotus officinalis (L.) Pallas; (Compositae) Echinops cornigerus DC.; (Labiateae) Stachys tibetica Vatke, Mentha longifolia (L.) Hudson, Perovskia abrotanoides Karelin. Bombus (Melanobombus) keriensis Morawitz (Figs 82, 122, 162, 202, 403-416, Maps 55 & 56) Bombus keriensis Morawitz, 1886: 199. Lectotype queen by designation of Podbolotskaya (in PAULH. WILLIAMS press), CHINA: Xinjiang, Kunlun Shan, mountains near Yutian [= Keriya], 9000 ft [2700 m] (Prshewalski) (Z1) [examined]. Bombus separandus Vogt, 1909: 58, 61 [footnote 1]. LECTOTYPE queen by present designa- tion (see Note 3 below), CHINA: Xinjiang, Borohoro Shan (ITZ) [examined]. Change of status to Bombus keriensis f.g. separandus Vogt by Reinig (1935). Syn. n. Bombus kohli Vogt, 1909: 41, 61 [footnote 2]. Syntype worker (see Note 4 below), MONGOLIA: north, no further data, 1892 (Leder) (ITZ) [examined]. Junior primary homonym of Bombus kohli Cockerell, 1906 [= B. morio (Swederus)]. Syn. n. Bombus kohli var. postzonatus Vogt, 1909: 61 [footnote 2]. LECTOTYPE queen by present designation (see Note 5 below), MONGOLIA: north, no further data, 1892 (Leder) (ITZ) [examined]. Syn. n. Bombus kozlovi Skorikov, 1910b: 413. Replace- ment name for B. kohli Vogt. Change of status to Bombus keriensis f.g. kozlovi Skorikov by Reinig (1935). Syn. n. Bombus (Lapidariobombus) separandus incer- toides Vogt, 1911: 58 [by indication of Vogt, 1909: 61 (footnote 1)] (see Note 1 below). Holotype queen by monotypy (see Note 6 below), MONGOLIA: north, no further data, 1892 (Leder) (ITZ) [examined]. Syn. n. Bombus lapidarius var. tenellus Friese, 1913: 86. 2 syntype females and 1 syntype male, U.S.S.R.: ?Zapadnyy Sayan, ‘Arasagun-gol’ (Staudinger) (MNHVU) [not seen]. Synonymised with Lapid- ariobombus incertoides (Vogt) by Skorikov (1931). Syn. n. Lapidariobombus separandus subsp. meridialis Skorikov, 1914a: 127. Holotype queen by monotypy (see Note 7 below), INDIA: Kash- mir, Sind valley above Sonamarg, 2400-3000 m, 9-10.vi.1912 (Jacobson) (ZI) [not seen]. Syn. n. Bombus (Lapidariobombus) tenellus var. alpiva- gus Richards, 1930: 639 (see Note 2 below). Holotype queen by original designation, CHINA: Xinjiang, Taghdumbash, 14000 ft [4300 ml], 18.vi.1913 (Hingston) (BMNH) [examined]. Syn. n. Bombus (Lapidariobombus) kozlowi Skorikov; Bischoff, 1936: 9. Unjustified emendation (Art. 33b(i, iii)). Pyrobombus (Melanobombus) keriensis subsp. karakorumensis Tkalct, 1989: 57. Holotype queen by original designation, PAKISTAN: Gilgit, Banidas 36°11’N 74°33’E, 2600 m, 30.vi-2.vii.1959 (Lobbichler) (ZS) [examined]. Syn. n. THE BUMBLE BEES OF THE KASHMIR HIMALAYA NOMENCLATURE. Note 1. Vogt’s (1911: 58) use of the name incertoides under ‘Meine Separandus- tiere ...’, for a female that he had described previously (1909: 61 [footnote 1]), is deemed to be of subspecific rank even though he did not state its rank explicitly (Art. 45f(i)). Skorikov subse- quently (1922) treated incertoides as an available name and adopted it as the name of a species (Art. 45g(i1)(1)). Note 2. Richards’s use (1930: 634) of the term ‘Subsp.’ elsewhere in his paper could be taken to indicate that infrasubspecific rank is meant for B. tenellus var. alpivagus (despite the comment ‘probably a geographical race’). However, this has subsequently been treated as a name in the species group by Reinig (1935: 333 [footnote 1: ‘f.g. means forma geographica (=subsp.)’ ]), so Richards’s taxon is deemed to be of subspecific status (Art. 45f(ii)). TYPE MATERIAL. Note 3. Vogt’s original descrip- tion of B. separandus mentions at least one female and one male from the Alai Mountains and from ‘Siebenstromgebiet’. Vogt subsequently (1911: 58) refers to the same material ‘aus der Umgebung des Festungswerks Narynj und dem Boro- Chorogebirge [= Borohoro Shan] im Sieben- stromgebiet’. A single queen in the ITZ collection agrees with the original description and carries a label (1) ‘Boro-Chorogeb / Tisilikau’; (2) a white, handwritten label ‘separandus O.V.’; (3) a red printed label ‘Type’. This specimen is designated as lectotype (Art. 74a). Note 4. Vogt’s description of B. kohli mentions two queens (one of B. kohli var. postzonatus) and a number of workers. However, the description is specified to have been made primarily from a queen that had the pubescence of tergum III black. I have seen a single worker from the ITZ collection that agrees with this description and carries a label ‘N. Mongolei/ Leder 92’, and a red printed label “Type’. I believe that this is one of the workers referred to at the end of footnote 2 and regard it as a syntype. It would be preferable to designate as lectotype the queen on which Vogt concentrated for the original description, when this queen is found. Note 5. Vogt described B. kohli var. postzona- tus as like the type, the queen of B. kohlis.str., so he probably based his description on the second of the two queens that he mentions at the end of footnote 2, rather than on any of the workers. A single queen in the ITZ collection agrees with the original description and carries a label (1) ‘N. Mongolei/ Leder 92’; (2) a white, handwritten label ‘kohli, ab postzonatus’; (3) a white hand- written label ‘13’; (4) a red printed label ‘Type’. 97 This specimen is designated as lectotype (Art. 74a). Note 6. Vogt’s description of B. separandus incertoides specifies that only one specimen (refer- ring to a second queen described separately under the name B. separandus in 1909: 61 [footnote 1]) was available. A single queen in the ITZ collec- tion agrees with the original description and carries a white, handwritten label ‘separandus- / incertoides-’ and a red, printed label ‘Type’. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 7. Skorikov’s description of Lapidario- bombus separandus subsp. meridialis specifies that only a single queen was available. If a single queen with the appropriate data can be found then this should be regarded as the holotype (Art. 73a(ii)). AFFINITIES. Within the subgenus Melanobombus, B. keriensis belongs to the lapidarius-group of species (see the comments on B. ladakhensis). Within the /apidarius-group, B. keriensis shares with B. incertus Morawitz, B. lapidarius (Lin- naeus) and B. sichelii Radoszkowski a slight reduction of the recurved head of the penis valve (Fig. 82) (?synapomorphy, see comments on B. semenovianus). All of these species, except B. incertus, have the volsella narrowed near its mid point in ventral aspect (Fig. 162) (synapomorphy of B. lapidarius + B. keriensis + B. sichelii). B. incertus and B. lapidarius apparently do not occur east of the Caspian Sea (distributions summarised by Reinig, 1935: chart 3, 1939: fig. 15), except for a couple of records of B. lapidarius from the northern U.S.S.R. (Panfilov, 1981: map 92). The closest relative of B. keriensis is probably B. sichelii Radoszkowski [the correct original spelling of B. sicheli of authors]. Unlike B. lapidarius, females of these two species share an apical notch on tergum VI and the males appear to share a slight displacement of the weak ventro- basal angle of the penis valve nearer to the mid- point of the shaft (Fig. 122). B. sichelii has a broad Eurosiberian distribution, although this is frag- mented in Europe among the southern mountains (mapped in Reinig, 1935: charts 2 & 5, 1939: fig. 4). The distributions of the two species overlap in the Altai (Reinig, 1935; BMNH), Mongolia (Reinig, 1935; Tkalct, 1974a; BMNH, PW) and Gansu (Bischoff, 1936). Unlike B. sichelii, the males of B. keriensis have the apex of the penis valve head slightly less acutely pointed (Fig. 82) and the interio-apical process of the volsella is narrower, with a less prominent interio-basal corner (Fig. 162). There is considerable variation 98 in the sculpturing of the ocello-ocular area of the females, but females of B. keriensis often differ from those of B. sichelii in that the band of fine punctures is more distinctly interrupted, so that there is a narrow shining area with only large punctures adjacent to the eye margin. Skorikov (1931) and Reinig (1935) both sug- gested that the cream individuals that were des- cribed under the name B. separandus Vogt, which often also have pale hairs on the face and a pale fringe in the pubescence of tergum III (=‘ciliated’, e.g. B. kohli var. postzonatus Vogt and Lapidariobombus separandus subsp. meri- dialis Skorikov), are actually conspecific with the more distinctly yellow and often unciliated B. keriensis (also described under the names B. kohli Vogt, B. lapidarius var. tenellus Friese, B. tenellus var. alpivagus Richards and Pyrobombus kerien- sis subsp. karakorumensis Tkalct). Reinig (1935) also included the unciliated and white queen described under the name B. separandus incer- toides Vogt in this group. Individuals with the very pale and ciliated colour patterns occur throughout much of the total distribution of this group of nominal taxa, but are most frequent in the region of the Pamir (Reinig, 1935: chart 4, 1939: fig. 23). The inference that they are all likely to be interbreeding as parts of a single population and a single species is supported by the apparently continuous variation between the two extreme colour patterns among the material from the Hindu Raj range (BMNH). A nest from the Pamir was described by Bischoff (1931). DISTRIBUTION. B. keriensis is a widespread, but primarily central Asian and Tibetan species (Map 55). It is known from Mongolia (Vogt, 1909, 1911; Tkalct, 1974a; BMNH, ITZ, PW), the Altai (Morawitz, 1880; Friese, 1913; Skorikov, 1931), the Tien Shan (Morawitz, 1880; Vogt, 1909, 1911; Skorikov, 1931; Panfilov, 1957; BMNH, ITZ, TM), the Pamir (Richards, 1930; Reinig, 1930, 1934, 1935; Bischoff, 1931; Skorikov, 1931; BMNH), the Hindu Kush (Reinig, 1940; Richards, 1951; Tkalct, 1969a; BMNH), Pakistan (BMNH, PW), Kashmir, Xinjiang [Kunlun Shan] (Morawitz, 1886; TM, ZI), Tibet (Richards, 1930; Wang, 1982; BMNH), Qinghai (Morawitz, 1886; Panfilov, 1957; Tkalct, 1961; Wang, 1982; ZS) and Gansu (Bischoff, 1936). In the west it is also known from the Elburz (Skorikov, 1931; BMNRH), Turkey and the Caucasus (Reinig, 1935; BMNRH). Reinig (1935: chart 2, 1939: figs 4 & 23) mapped a summary of this distribution. A similar disjunct distribution between the Hindu Kush ranges on the one hand and the Elburz and PAULH. WILLIAMS Armenian highlands on the other is shown by B. melanurus. MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. keriensis is recorded from the Hindu Raj, Zanskar, Great Himalaya and Pir Panjal ranges, in alpine scrub and steppe (Map 56, Fig. 6): 73 queens, 155 workers, 38 males, from 18 localities 2600-5100 m (BMNH, FA, NR, PW, ZS). VARIATION WITHIN KASHMIR. When these bees are seen in flight, the two most obvious characters of the colour pattern to vary are (a) the yellow or white colour of the pale pubescence on the thorax and on terga I- II; and (b) the presence or absence of a fringe of pale hairs at the posterior and lateral margins of tergum III (present in the ‘ciliated’ state). Females from the Hindu Raj range generally have the pale pubescence nearly white if queens, but yellow if workers, although both are strongly ciliated (cream-white: 49/50 queens, 0/68 workers, no males available; strongly ciliated: 47/50 queens, 43/68 workers; Figs 403-407). In the Great Himalaya range, workers are usually paler but unciliated, both around the Kishanganga valley (cream-white: 2/7 queens, 39/49 workers, 0/33 males; strongly ciliated: 0/7 queens, 0/49 workers, 0/33 males; Figs 408-410), and at the head of the Sind valley (cream-white: no queens available, 32/33 workers, 0/3 males; strongly ciliated: 0/32 workers, 1/3 males). Individuals from the Zanskar ranges usually have the pale pubescence yellow and are uncili- ated (cream-white: 1/5 queens, 1/2 workers, no males available; strongly ciliated: 1/5 queens, 0/2 workers [more yellow and unciliated queens were seen but not collected]; Figs 411-413) like those from Tibet. Individuals from the Pir Panjal range are similar, but with slightly more extensive black pubescence on the thorax (cream-white: 0/8 queens, 0/5 workers, 0/2 males; strongly ciliated: 0/8 queens, 1/5 workers, 0/2 males; Figs 414-416). This species is similar in colour pattern to some of the B. himalayanus (Fig. 242), B. kashmirensis (Figs 298-300, 305-307), B. subtypicus (Figs 317 & 318), B. biroi (Figs 340, 342, 343), B. tunicatus queens (Fig. 350), B. asiaticus (Figs 386-388) and B. oberti (Fig. 393) that occur with it at some localities. Females can usually be recognised by their combination of an apical notch on tergum VI with a lack of many short hairs at the base of the hind basitarsus (see the key, couplet 11). The males are closely similar in colour pattern (Figs 410 & 416) to some of those of B. subtypicus (Fig. 318) and of B. biroi (Figs 340 & 343), which may occur with them at some localities, but can be recognised by the narrow recurved hooks of the head of the penis valve (Fig. 82). THE BUMBLE BEES OF THE KASHMIR HIMALAYA FOOD PLANTS. Kashmir: (Ranunculaceae) Aconi- tum hookeri Stapf; (Balsaminaceae) Impatiens glandulifera Royle; (Leguminosae) Trifolium repens L.; (Compositae) Cirsium falconeri (Hook.f.) Petrak; (Gentianaceae) Swertia petio- lata D. Don; (Scrophulariaceae) Picrorhiza kurrooa Royle ex Benth. Ladakh: (Leguminosae) Caragana versicolor (Wallich) Benth.; (Labiateae) Stachys tibetica Vatke. Bombus (Melanobombus) simillimus Smith (Plate 1, Figs 83, 123, 163, 203, 226, 417-421, Maps 57 & 58) Bombus_simillimis Smith, 18526: 48. LECTO- TYPE queen by present designation (see Note 3 below), INDIA: north, no further data (BMNH) [examined]. Incorrect original spell- ing (see Note 1 below) (Art. 32c(i)), without separate availability in this form (Art. 32d). [Bombus similis Smith; Smith, 1854: 403. Incor- rect subsequent spelling, unavailable name (Art. 33c). Junior primary homonym of Bombus similis Fabricius, 1804 [= Centris similis (F.)].] Bombus simillimus Smith; Dalla Torre, 1896: 548. Justified emendation (see Note 2 below) (Art. 33b(ii)). [Bombus tonsus Skorikov, 1922: 160. Published without description or indication, unavailable name (Art. 12a).] [Bombus tonsus Skorikov, 1931: 202. Published without description or indication, unavailable name (Art. 12a).] Bombus terrestris var. grossiventris Friese, 1931: 303. LECTOTYPE worker by present designa- tion (see Note 4 below), INDIA: Kashmir, Srinagar, Shalimar, 1800 m, 7.x.1923 (Fletcher) (MNHU) [examined]. Syn. n. Bremus (Sibiricobombus) oculatus Frison, 1933: 335. Holotype male by original designation [p. 338], INDIA: Himachal Pradesh, Baghi, 8800 ft [2700 m], 7-8.x.1921 (Kemp) (Calcutta) [not seen]. Syn. n. Sibiricobombus tonsus Skorikov, 1933b: 248. LECTOTYPE queen by present designation (see Note 5 below), INDIA: Kashmir, Kishtwar, Datgash, 12.vi.1910 (Trubetskoy) (ZI) [exam- ined]. Syn. n. Bombus (Lapidariobombus) oculatus var. haemor- rhous Richards, 1934: 87. Holotype worker by original designation [p. 88], INDIA: Himachal Pradesh, Dalhousie, 7.vii.1906 (Barrow) (BMNH) [examined]. Syn. n. NOMENCLATURE. Note 1. Smith’s original spelling (1852b) is ‘simillimis’, which may have been 99 intended as the superlative of similis (comment after the original description: “This species is very like the Lapidarius of Linn.;’). There is no evi- dence that this spelling is incorrect from the original publication, except (Art. 32c(i)) in its termination (Art. 31b). Therefore the subsequent spelling by Smith (1854), which is not an emenda- tion (Art. 33b), is deemed to be incorrect (Art. 33c), although the termination of simillimis must still be corrected (Art. 32d(ii)) to simillimus. Note 2. Dalla Torre’s change in the subsequent spelling of ‘simillimis’ can be accepted as inten- tional (Art. 33b(i)), even though he only quotes ‘simillimus’, because of his introduction (1896: v): ‘So ist es z.B. wissenschaftlich gewiss nicht gerechtfertigt, wenn jeder Elementarlateiner befahigt und — berechtigt ist, WOrter, wie laevis, coelestis, sylvaticus zu corrigiren, ohne dass man in wissenschaftlichen Kreisen je davon Notiz genommen hat, und ahnlich verhalt es sich auch mit den Geschlechtsendungen, wo auch alle erdenklichen grammatikalischen Fehler gemacht und fortwahrend colportirt werden.’ TYPE MATERIAL. Note 3. Smith’s description of B. simillimus is of a queen. A queen in the BMNH collection agrees with the original description and bears a purple-edged label (1) ‘Lecto- / type’; (2) ‘N. / India’, reverse side ‘48 / 132’ (this accession number refers to 73 Hymenoptera from the Boyes collection received in 1848); (3) ‘N. China / Baily 11/7/55’; (4) ‘60-15 / E.I.C.’ (see Note 2 on B. haemorrhoidalis); (5) ‘similis / Type Sm.’; (6) ‘Bombus / similis / Smith Trans. / Ent. Soc. 1852’; (7) a red-edged label ‘Type’; (8) ‘B.M. TYPE / HYM. / 17B.981’. Other Smith syntypes might have existed and may since have become dispersed to other collections. Therefore this queen, which lacks both fore tarsi, the left front basitarsus and both hind tarsi, is designated as lectotype (Art. 74a). Note 4. Friese described B. terrestris var. grossi- ventris from three workers. I have seen one of these from the MNHU collection that agrees with the original description and carries a label (1) ‘Kashmir 200ft / Shalimar / Srinagar / 7 Oct. 1923 / Fletcher coll’; (2) a handwritten label ‘Bombus / grossiventris / Fr. / [worker] Friese det. 25’; (3) a printed label “Zool.Mus. / Berlin’. This specimen, which lacks the distal joints of the left hind tarsus, most of the flagellum of the right antenna and the distal joints of the left antenna, is designated as lectotype (Art. 74a). Note 5. Skorikov’s description of Sibiricobom- bus tonsus is of a queen. A queen in the ZI collection from near the village of Datgash, with a label ‘B. tonsus type’, agrees with the original 100 description. Because other Skorikov syntypes may exist, this specimen, which has two lateral patches of orange-brown pubescence on the anterior part of the scutum, is designated as lectotype (Art. 74a). AFFINITIES. Within the subgenus Melanobombus (see the comments on B. ladakhensis), B. simillimus belongs to a group of species, the rufofasciatus-group (the ‘flavothoracicus-Gruppe’ + ‘rufofasciatus-Gruppe’ of Tkalct, 1974b), for which the male gonostylus is shortened with dis- tinct inner and outer apical corners (Figs 203-205) (synapomorphy). Skorikov (1922) recognised a subgenus Kozlovibombus on the basis of the male genitalia of what was probably a specimen of B. pyrosoma s.str. (see Reinig, 1934, 1935). Females of all of the species in this group apparently vary in colour with size (see the comments on B. pyrosoma) in a manner resembling that described for the Central American B. (Pyrobombus) ephippiatus Say by Owen & Plowright (1980) and for the Himalayan B. (Festivobombus) festivus by Ito et al. (1984). This variation is not only in the extent of the pattern of replacement of black by pale hairs in the pubescence, but smaller indivi- duals often have at least one additional colour. This is usually yellow or brown in the pubescence on tergum II. Hence queens, workers and males may appear strikingly different (see Plate 1). Within the rufofasciatus-group, queens of B. simillimus share with those of B. tanguticus Morawitz a distinct and complete sulcus obliquus of the mandible. They also share the fine, branched hairs on the outer surface of the hind tibia (Fig. 226) and the dense (often black) short pubescence on the hind basitarsus. B. tanguticus is known from Sikkim and Tibet (Richards, 1930; BMNH) and Qinghai (Morawitz, 1886). It could also be present in the high mountain ranges of north-eastern Kashmir (unconfirmed record from Kashmir by Skorikov, 19335). Queens of B. tanguticus are very easily distinguished from those of B. simillimus by their acutely and deeply notched tergum VI, which is straight or only shallowly indented for B. simillimus; by their labral furrow, which is only one-quarter of the total breadth of the labrum, as opposed to nearly half of the breadth for B. simillimus; by their oculo-malar distance, which is about equal to the breadth of the mandible at the base, whereas it is much longer for B. simillimus; by their ocello- ocular area, which is much less densely punctured; by their clypeus, whichis less strongly swollen, but more coarsely punctured; and by their nearly clear (subhyaline) wings. B. tanguticus was considered sufficiently distinctive by Pittioni (1939c) to PAULH. WILLIAMS warrant the description of a monobasic subgenus, Tanguticobombus. The male remains undescribed and apparently unknown, so that the precise relationships of this species are difficult to resolve at present. Within the rufofasciatus-group, some females have a straight or only shallowly indented apex to tergum VI as well as a distinct sulcus obliquus of the mandible. Queens and workers with these characters from the western Himalaya were des- cribed under the names B. simillimus and B. terrestris var. grossiventris respectively. They have not been associated as castes of the same species because of their very different colour patterns. Queens of B. terrestris var. grossiventris and workers and males of B. simillimus were un- known. But within the rufofasciatus-group from the Himalaya, only these nominal taxa share the uniformly red pubescence of terga IV-V and the intensively darkened (infuscated) wings. One of the older queens (BMNH) has the pubescence of terga I-II distinctly lighter than on tergum III, so that it is dark brown rather than black (although the brown is still not as light as for B. terrestris var. grossiventris). The queens have particularly many of the fine, short, branched hairs on the outer (corbicular) surface of the hind tibia (Fig. 226), although these hairs are also present for some of the workers. These hairs are seldom numerous in this position among bumble bees, except for species of the subgenus Psithyrus. Among the social species, the presence of these hairs may be associated with particularly large body size because they are otherwise very numerous for B. (Melanobombus) tanguticus Morawitz, B. (Rufipedibombus) rufipes Lepeletier and B. (Rufipedibombus) eximius Smith, although they are also present for some queens of the European B. lapidarius. Almost certain confirmation that B. simillimus (queens, Fig. 417), B. terrestris var. grossiventris (workers, Fig. 418) and males (Fig. 419) described under the name Bremus oculatus by Frison (1933) are the castes and sexes of a single species comes from a colony discovered near Harwan, in the Vale of Kashmir (see Plate 1 and the description of material examined from Kashmir). The alternative interpretation is that a colony of one species, for which the queen must be unknown, had been usurped by a queen of another species, which is closely similar in morph- ology to the workers of the first queen. Males of B. simillimus can be distinguished from those of the other taxa of the rufofasciatus- group by their greatly enlarged eyes; by the large, sub-rectangular gonostylus, which is scarcely reduced and retains a narrow interio-basal process (Fig. 203) (plesiomorphic within the rufofasciatus- THE BUMBLE BEES OF THE KASHMIR HIMALAYA group); and by the twisted interio-apical process of the volsella (Fig. 163). The form of the gonostylus is most similar to that of B. richardsiel- lus (Tkalct, 1968a: figs 84-86; see the comments on B. pyrosoma). A queen with two light patches on the anterior dorsum of the thorax was described under the name Sibiricobombus tonsus by Skorikov (19335). Two workers were redescribed under the name B. oculatus var. haemorrhous by Richards (1934). These individuals are closely similar in morph- ology to those described under the names B. simillimus and B. terrestris var. grossiventris respectively and are considered to be part of the same species. Females (probably workers) of another nominal taxon from Tibet (B. trilineatus), which is closely similar to B. simillimus, have been described by Wang (1982) from slight differences in colour pattern (not seen). DISTRIBUTION. B. simillimus is a west Himalayan species (Map 57). It is known only from Himachal Pradesh (Frison, 1933; Richards, 1934; BMNH), Kashmir and Pakistan (BMNH). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. simillimus is recorded from the lower slopes in the Kishanganga valley, in the Vale of Kashmir and along tributaries to the Chenab valley, in open scrub around the lower montane coniferous forest (Map 58): 9 queens, 102 workers, 8 males, from 9 localities 1600-3000 m (AB, BMNH, MNHU, IZ, NR, PW, ZI). I found a colony near the edge of open wood- land near Harwan (1700 m, 10.ix.1985). A fresh hole, 50 cm deep by 30 cm wide, in the bank of an irrigation channel contained 6 queens (including an old and very abraded individual), 39 workers and 5 males. Returning workers found their way into the cavity via a separate tunnel, nearly 2 m long. The remains of only 4 cocoons were present in the cavity. I had seen bears (probably Solenarctos thibetanus (Cuvier)) in the vicinity earlier on the same day. VARIATION WITHIN KASHMIR. There is little varia- tion among the queens (Plate 1 and Fig. 417). The number of black hairs on the thoracic dorsum of workers does vary, but none of them has even a weak black band between the wing bases (Plate 1 and Fig. 418). One large worker (severely abraded) from the colony near Harwan has the hairs of tergum I white with dark bases, while tergum II is very dark brown, with a posterior fringe of white hairs (Fig. 420). The males vary from having terga I-II entirely nearly white (5/5 males, Harwan, Plate 1 and Fig. 419), to having tergum I white and tergum II chocolate-brown, with the exception of 101 a posterior fringe of white hairs (3/3 males, Banihal, Fig. 421). Queens of this species are distinctive in their appearance. Workers are similar in colour pattern to some of those of B. tunicatus (Figs 351 & 354), which occur with them. Workers of B. simillimus can usually be recognised by their chocolate- brown pubescence of tergum I (Fig. 418), rather than white as for B. tunicatus (Fig. 354). Males are also similar in colour pattern to some of those of B. tunicatus (Figs 352 & 355), or even some B. asiaticus (Fig. 385) that occur with them. They can be distinguished from B. tunicatus by their enlarged eyes and from B. asiaticus by their short antennae. FOOD PLANTS. (Balsaminaceae) Jmpatiens glandu- lifera Royle; (Leguminosae) unidentified pink clover-like legume; (Compositae) Cirsium fal- coneri (Hook.f.) Petrak, C. wallichii DC., Centaurea iberica Trevir ex Spengel, unidentified yellow thistle-like composite; (Gentianaceae) Swertia petiolata D. Don. Bombus (Melanobombus) pyrosoma Morawitz (Plate 1, Figs 84, 124, 164, 204, 422-430, Maps 59 & 60) Bombus pyrosoma Morawitz, 1890: 349. Lecto- type queen by designation of Podbolotskaya (in press), CHINA: ‘Gansu’, ‘Utai’, 8900 ft [2700 m], vi.1884 (Potanin) (ZI) [examined]. Bombus pyrrhosoma Morawitz; Dalla Torre, 1896: 544. Unjustified emendation (Art. 33b(i, iii)). Bombus flavothoracicus Bingham, 1897: 552. Lectotype queen by designation of Tkalcu (19746: 338), INDIA: Sikkim, Lintu, 12500 ft [3800 m], v.1894 (Bingham) (BMNH) [exam- ined]. Junior secondary homonym in Bombus of Psithyrus campestris var. flavothoracicus Hoffer, 1889 [= B. campestris (Panzer)]. Pro- visional synonym. Bombus miniatus Bingham, 1897: 553. Holotype male by original designation, INDIA: Sikkim, Lintu, 12500 ft [3800 m], v[!].1894 (Bingham) (BMNH) [examined]. Synonymised with Pyro- bombus flavothoracicus (Bingham) by Tkalci (19745). Provisional synonym. [Bombus pyrrhosoma var. canosocollaris Skorikov, 1912b: 608 [not seen]. Infrasubspecific (Art. 45g(ii)(1)), unavailable name (Art. 45e).] Bombus friseanus Skorikov, 1933a: 62. Holotype queen by monotypy (see Note 1 below), CHINA: Sichuan, Songpan, above 9500 ft 102 [2900 m], 1894 (Beresovski) (ZI) [examined]. Syn. n. Bremus (Lapidariobombus) formosellus Frison, 1934: 163. Holotype male by original designa- tion [p. 166], TAIWAN: ‘Roeichi’, 15.ix.1924 (Shiraki & Sonan) (INHS) [examined]. Pro- visional synonym. Bombus (Lapidariobombus) pyrrhosoma subsp. hénei Bischoff, 1936: 10. LECTOTYPE queen by present designation (see Note 2 below), CHINA: Yunnan, Lijiang, 23.v.1934 (Hone) (MNHU) [examined]. Syn. n. [Bombus (Lapidariobombus) pyrrhosoma f. flavo- corbicularis Tkalct, 1961: 353 [not seen]. Infra- subspecific (Art. 45g(1i)(1)), unavailable name (Art. 45e).] Pyrobombus (Lapidariobombus) wutaishanensis Tkalct,, 1968a: 39. Holotype queen by original designation [p. 41], CHINA: Shanxi, Wutai Shan, 3200 m, 15.vii.1936 (Héne) (MNHU) [examined]. Syn. n. TYPE MATERIAL. Note 1. Skorikov’s description of B. friseanus specifies that only a single queen was examined. A single queen in the ZI collection agrees with the original description and bears a label with the data quoted. I believe that this is the single specimen on which the original description is based and regard it as the holotype (Art. 73a(ii)). Note 2. Bischoff’s description of B. pyrrhosoma subsp. hoenei shows that several females, col- lected ‘20.4’ and ‘1.5.—23.6.34’, were examined. A queen in the MNHU collection bears the label (1) ‘Li-kiang. (China). / Provins Nord-Yuennan. / 23.5 1934.H.Ho6ne.’; (2) ‘pyrrhosoma / hoenei n. sp. / [female] / det. Bischoff’; (3) a red printed label ‘Typus’; (4) ‘LECTOTYPE / Bombus / pyrrhosoma / Tkalci det. / hénei Bischoff / [female]’ [designation not published]; (5) “Zool. Mus. / Berlin’. This specimen, which is complete, is designated as lectotype (Art. 74a). AFFINITIES. Within the subgenus Melanobombus (see the comments on B. ladakhensis), B. pyro- soma belongs to the rufofasciatus-group of species (see the comments on B. simillimus). Within the rufofasciatus-group, the known males, other than those of B. simillimus, are characterised by pronounced reductions (shorten- ing) of the gonostylus and a broadening of its interio-basal process (Figs 204 & 205) (synapo- morphies). Apart from B. simillimus and B. rufofasciatus (see the comments on B. rufofasci- atus), four principal nominal taxa have been recognised in this group. These were described under the names Lapidariobombus richardsiellus, B. pyrosoma, B. friseanus and B. flavothoracicus. PAULH. WILLIAMS Of these, B. pyrosoma, B. friseanus and B. flavothoracicus are morphologically closely simi- lar and differ from each other primarily in colour pattern. B. richardsiellus was described (Tkalci, 1968; BMNH) from 2 queens, 3 workers and 1 male from northern Burma and from the neighbouring parts of Tibet. The queens are similar to the B. friseanus from the same region, but have the clypeus slightly more coarsely punctured. Their white-banded colour pattern has a much more restricted distribution of pale pubescence. The ‘allotype’ male has the exterio-apical corner of the gonostylus more strongly reduced than for B. simillimus (synapomorphy of B. richardsiellus + B. pyrosoma [+ B. friseanus]| + B. rufofasciatus), so that it is intermediate in form between the gonostyli of B. simillimus and of B. friseanus. Another 4 males from near the Burmese border of Tibet (BMNH) may belong to B. richardsiellus, but show a broad range of forms of the gonostylus, from similar to that of the ‘allotype’, to more like that of B. friseanus. At a few localities in Kashmir (e.g. Mt Apharwat [1986], Daksum), B. simillimus and B. rufofasciatus not only occasionally occur together, but also with another taxon of the rufofasciatus- group. Males of B. simillimus and B. rufofasciatus have colour patterns similar to their respective workers, although the males have much larger eyes. The males of the third taxon are often predominantly yellow (Figs 424, 427, 430), unlike the females (even when reared from the same maternal colony, see Plate 1 and the comments on the material examined from Kashmir), and have eyes that are not enlarged relative to those of the females. These yellow males differ in their mate- searching behaviour from males of B. rufofasci- atus and presumably from those of B. simillimus (see the introduction on male mate-searching behaviour, and the comments on B. rufofasci- atus). There are no specimens with intermediate colour patterns or morphology as evidence that they interbreed with either B. rufofasciatus or B. simillimus. Therefore the three taxa are believed to represent three separate species. B. flavothoracicus is the name that has been used most recently for the third Himalayan taxon of the rufofasciatus-group that is found in Kashmir (e.g. Tkalci, 1974b). This name was used by Bingham for a queen, which has light olive-yellow pubescence at the front and rear of the thorax and on tergum I. Gastral terga IV-VI have the pubes- cence extensively white. The predominantly yellow males were also described by Bingham, but under the name B. miniatus. The workers differ from the queens in that they have most of THE BUMBLE BEES OF THE KASHMIR HIMALAYA tergum II dull olive-yellow or brown and most of the pubescence of terga III-V red, which replaces most of the white and even some of the black hairs. This caste dimorphism may be size- dependent, because the red pubescence is less extensive in large workers (decreasing with in- creasing body size in Figs 426, 428, 429) and red may occasionally be intermixed on terga III-IV of small queens. This nominal taxon is known from as far east in the Himalaya as the area of the Sikkim/Tibet/Bhutan border (BMNH), to as far west as Kashmir. Queens of two further nominal taxa from Tibet (B. eurythorax and B. steno- thorax), both very similar to B. flavothoracicus, have been described by Wang (1982) from slight differences in colour pattern (not seen). Material from further to the east in Yunnan and Sichuan was described under the names B. pyr- rhosoma subsp. hoenei by Bischoff (1936) and B. friseanus Skorikov (1933a) respectively. Both the queens and the workers from this region are similar in colour pattern to the workers of B. flavothoracicus from the Himalaya, although the pale pubescence (including tergum II) is usually a deeper orange-yellow and there are fewer black hairs intermixed with it on the thorax. Thus both castes have the pubescence of terga III-VI red, although without any distinctly white hairs. Workers of B. friseanus are known to occur as far west as the lower Zangbo [= Tsangpo] valley of Tibet and the Char Chu valley at the extreme eastern end of the Himalaya, adjacent to Arunachal Pradesh (BMNH, no queens available). Material from further to the north in Gansu was originally described under the name B. pyrosoma s.str. by Morawitz (1890). The queens have the thorax and tergum I black with just a very few grey-white hairs intermixed. The pubescence of gastral terga II-VI is predominantly red. Morawitz also described the workers from Gansu (listed as Var. b and Var. c), which have the pubescence of the thoracic dorsum grey-white with a black band between the wing bases, with tergum I grey-white, tergum II brown and terga III-VI red. Similar material is known from the hills that reach eastwards to as far as Beijing and Hebei. Tkalci (1968a) described two queens under the name Pyrobombus wutaishanensis that have distinctly fewer fine punctures in the ocello- ocular area than B. pyrosomas.str. At least one of these queens is from Shanxi. I have seen another queen of B. pyrosoma s.str. from neighbouring Beijing (PW) that also shows some reduction in the number and extent of these fine punctures, so that it is intermediate in this character between Pyrobombus wutaishanensis and, for example, B. friseanus. These specimens are otherwise similar 103 in morphology and in colour pattern to the lectotype of B. pyrosoma. Until more evidence to the contrary is available, I consider that they are most likely to be part of the same species. From the available material, each of the three principal colour patterns is relatively constant among the individuals of each caste from across extensive regions, with transitions in colour pattern apparently occurring over short distances. Yet from Qinghai, between the regions occupied by B. friseanus and B. pyrosoma s.str., Tkalci (1961) described a worker (B. pyrrhosoma f. flavocorbicularis, infrasubspecific under Art. 45¢(i)) that has both the yellow hairs on the head and legs, as for B. friseanus, and the pale pubes- cence of the dorsum grey-white (not seen), as for workers of B. pyrosoma s.str. Skorikov (1912b) even described banded grey queens from the area of Gansu and Inner Mongolia, similar in colour pattern to most of the workers rather than to the queens with a predominantly black thoracic dorsum from north-eastern China, under the name B. pyrrhosoma var. canosocollaris (not seen). I consider it likely that these intermediate colour patterns are evidence of clines from hybrid zones between the more widespread regional colour patterns (see the comments on B. trifasci- atus). This has also been recognised previously for the Chinese nominal taxa by Bischoff (1936) and Tkalct (1961). Evidence of intermediates between B. flavo- thoracicus and B. friseanus is not as strong, at least partly because so little material is available from where these nominal taxa occur in close proximity in the eastern Himalaya. However, the few workers and males that are available from the far east of the Himalaya (BMNH) are difficult to assign to either nominal taxon with any confi- dence. The only available queens are two speci- mens from between Gyangtse [=Gyangzé] and Phari [=Pagri] in Tibet (BMNH). These large individuals share the white-‘tailed’ colour pattern of the west Himalayan B. flavothoracicus queens, although they also have red hair extensively intermixed in the pubescence of terga III-IV. Overall, from west to east there may be a trend towards an extension of the red pubescence on the gaster, especially for queens, in which it com- pletely replaces the white hair. There is also a parallel decrease in the extent of the pale pubes- cence on the thorax and on terga I-II, which is accompanied by a change in colour from light yellow to orange-yellow to grey-white. The males that are associated with the three principal nominal taxa, B. flavothoracicus, B. friseanus and B. pyrosoma s.str., all have closely similar genitalia, but vary in the form of the 104 interio-apical corner of the gonostylus. This is narrowly bilobed or spinose (similar to some putative B. richardsiellus) for B. miniatus (the male associated with B. flavothoracicus, Fig. 204) and also for some B. pyrosoma s.str., but may be broadly bilobed (similar to B. rufofasciatus, Fig. 205) for other B. pyrosoma s.str. All of these males have the pubescence predominantly yellow and their eyes are not enlarged relative to those of the females. Both of these states are characteristic of males that patrol circuits of scent-marked sites in the upper montane forest habitats to find mates (see the introduction on male mate-searching behaviour and the discussion of the colour pat- terns of the Kashmir fauna). Therefore from the evidence available at present I consider it likely that the three principal nominal taxa are based on the regional colour patterns of the females from a single interbreeding population. If this is the case, they would be considered to be parts of a single species. More definite conclusions concerning the status of these nominal taxa must await more direct information on their mate-searching be- haviour and on interbreeding between them, particularly from the eastern Himalaya. The name Bremus formosellus was applied by Frison (1934) to a population that is isolated from other B. pyrosoma s.\. by sea on the island of Taiwan. He distinguished this nominal taxon from B. miniatus by colour pattern alone, although he had apparently not seen any specimens from the Himalaya (p. 166: ‘It [“This new species’] differs from the male (type) and workers of miniatus as described by RICHARDS (1930) in the more broader and distinct black band on the thorax between the wings and lack of bright yellow pubescence on first and second dorsal tergites.’). The grey females and the extensively yellow males of Bremus formosellus actually resemble closely some of the B. pyrosoma s.str. from Qinghai and Gansu in both colour pattern and morphology. I know of no evidence to suggest that individuals from the disjunct population on Taiwan would not interbreed with those from the continental popu- lation if they were to occur together, although I have not yet had the opportunity to examine any queens. So Bremus formosellus may also be a part of the species B. pyrosoma s.1. DISTRIBUTION. B. pyrosoma is primarily a peri- Tibetan species, although it also occurs in other highland areas in northern and southern China (Map 59). It is known from Inner Mongolia (Skorikov, 1912b, 1933a; Tkalct, 1960), Hebei (Yasumatsu, 1951; PW), Beijing (PW), Shanxi (Yasumatsu, 1951; Tkalci, 1968a; MNHU), Shaanxi (BMNH), Hubei (Sakagami, 1972), PAULH. WILLIAMS Gansu (Morawitz, 1890; Skorikov, 1912b; Bischoff, 1936; ZI), Qinghai (Skorikov, 1912b; Panfilov, 1957; Tkalci, 1961), Sichuan (Skorikov, 1933a; Sakagami, 1972; Wang, 1982; BMNH, PW, ZI), Yunnan (Bischoff, 1936; Wang, 1987; BMNH, MNHU), Taiwan (Frison, 1934; Chiu, 1948; INHS, PW), Tibet (Richards, 1930; Wang, 1982, 1988; BMNH, PW), Sikkim (Bingham, 1897; Friese, 1918; Richards, 1930; BMNH, UM), Nepal (Tkalci, 19746; BMNH, NMS, PW), Uttar Pradesh (Frison, 1935), Himachal Pradesh (BMNH, PW) and Kashmir. A record from Guangdong [Leizhou peninsula] (Sakagami, 1972) needs to be confirmed with more material. A similar disjunct distribution between the moun- tains of south-western China and Taiwan is known for B. parthenius (synonym of B. sonani (Frison), see the comments on B. lepidus) (Wang, 1982). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. pyrosomais recorded from the southern side of the Great Himalaya range and from the Pir Panjal range, in montane coniferous forest (Map 60, Fig. 6): 56 queens, 172 workers, 62 males, from 10 localities 1900-4000 m (AB, BMNH, IZ, MNHN, NR, PW, RH). A colony was found on Mt Apharwat, in the forest below Gulmarg (2500 m, 2.ix.1985). The entrance tunnel was in the bank of a stream and lead 40 cm to the nest cavity, 80 cm vertically below the surface of the bank. The nest contained 478 small cocoons (33 occupied, 7 that were opened all contained males), 126 large cocoons (106 occupied), 20 queens, including one old and very abraded individual, 36 workers and 31 males. No younger brood was present. VARIATION WITHIN KASHMIR. This species is very variable in colour pattern, but among the females from Kashmir the trends in variation appear to be associated more with body size than with locality (e.g. increasing body size in Figs 426, 428, 429, 425). In particular, small queens show the partial replacement of black or white pubescence by red on terga III-IV, so that they appear similar to the largest workers (Fig. 429). Most workers have, in addition, yellow (Fig. 426) to chocolate-brown (Figs 428 & 429) hair basally on tergum II, so that black hair remains only apically on tergum II and basally on tergum III (Plate 1). Some of the smallest workers have most of terga I-II dull olive- yellow and most of terga III-V orange-red (Fig. 426). The male colour pattern is predominantly yellow (Plate 1 and Figs 424, 427, 430). The pubescence of terga III-VII varies from entirely yellow (Figs 424 & 427) to largely black on tergum III and orange-red on terga IV-VII (Fig. 430). The THE BUMBLE BEES OF THE KASHMIR HIMALAYA most frequent pattern is for the pubescence of tergum III to be yellow with a band of orange-red hairs and black hairs intermixed basally, whereas terga IV-VII are orange-red with a few black hairs laterally and a few yellow hairs apically and laterally. Only one male (from Daksum) has a well-defined black band between the wing bases. The only pronounced geographical variation is that queens from Kel, in the Kishanganga valley, have unusually extensive patches of yellow hair on the basal half of tergum II (10/10 queens, Fig. 422). Workers from Kel (26 workers) have few black hairs on terga II-III, irrespective of size (Fig. 423). Another queen with the same data is severely abraded, so this series may represent the members of a single colony that had been col- lected by the Schmid expedition. There are also very few black hairs on the gasters of workers from 4 other sites in the Kishanganga valley (5/5 medium-sized workers, no queens available). Queens and workers of this species are distinc- tive in their appearance. The yellow males with red hairs on the apical terga may be similar in colour pattern to the males of B. lepidus (Figs 329 & 332), which occur with them at some localities. Males of B. pyrosoma can be recognised by their larger size and by the short gonostylus with an interio-apical spine (Fig. 204). FOOD PLANTS. (Balsaminaceae) /mpatiens glandu- lifera Royle; (Leguminosae) Trifolium repens L., Lupinus sp. [introduced]; (Sambucaceae) Lonicera quinquelocularis Hardw.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC., unidentified dandelion-like composite; (Ericaceae) Rhododendron anthopogon D. Don; (Gentianaceae) Swertia petiolata D. Don; (Scrophulariaceae) Scrophularia pauciflora Benth.., Digitalis lanata Ehrh. [introduced], D. purpurea L. [introduced], Pedicularis rhinanthoides Schrenk, P. punctata Decne.; (Labiateae) Prunella vulga- ris L. Bombus (Melanobombus) rufofasciatus Smith (Plate 1, Figs 26, 30, 34, 38, 85, 125, 165, 205, 431-439, Maps 61 & 62) Bombus rufo-fasciatus Smith, 1852b: 48. Lecto- type queen by designation of Tkalct (1974b: 340), INDIA: north, no further data (BMNH) [examined]. Bombus Prshewalskyi Morawitz, 1880: 342. Syn- type worker [not a queen] and male, CHINA: ‘Gansu’, no further data (Prshewalski) (ZI) [not seen]. Synonymised with Bombus rufofasciatus var. championi Richards by Richards (1930); 105 synonymised with Pyrobombus_ rufofasciatus (Smith) by Tkalct (19745). Bombus rufocinctus Morawitz, 1880: 343. Lecto- type queen by designation of Podbolotskaya (in press), CHINA: ‘Gansu’ hills, viii.1871 (Prshewalski) (ZI) [examined]. Junior primary homonym of Bombus rufocinctus Cresson, 1863. Synonymised with Bombus rufofasciatus Smith by Handlirsch (1888). Bombus chinensis Dalla Torre, 1890[June 25]: 139. Replacement name for B. rufocinctus Morawitz. Junior secondary homonym in Bombus of Apathus rupestris var. chinensis Morawitz, 1890 [April 30] [= B. chinensis (Morawitz)]. Synony- mised with Bombus rufofasciatus Smith by Richards (1930). Bombus rufofasciatus var. championi Richards, 1928a: 107. Holotype queen by original desig- nation, INDIA: Kashmir, 8000-9000 ft [2400- 2700 mJ], vi.1901 (Nurse) (BMNH) [examined]. Syn. n. [Bombus (Lapidariobombus) rufofasciatus var. rufior Richards, 1928b: 335 [examined]. Infra- subspecific (Art. 45g(ii)(1)), unavailable name (Art. 45e).] [Bombus (Lapidariobombus) rufofasciatus var. intermedius Richards, 1930: 643 [examined]. Infrasubspecific (Art. 45g(ii)(1)), unavailable name (Art. 45e).] Bombus (Lapidariobombus) waterstoni Richards, 1934: 88. Holotype queen by original designa- tion [p. 89], ‘Himalayas’, no further data (BMNH) [examined]. Synonymised with Pyro- bombus rufofasciatus (Smith) by Tkalct (19745). AFFINITIES. Within the subgenus Melanobombus (see the comments on B. ladakhensis), B. rufo- fasciatus belongs to the rufofasciatus-group (see the comments on B. simillimus). Within the rufofasciatus-group, B. rufofasciatus can be distinguished from B. simillimus by its strongly reduced male gonostylus (Fig. 205; see the key, couplet 25). B. rufofasciatus is similar to some B. pyrosoma s.\. in that it has a broadly bidentate, interio-apical corner of the gonostylus, although the apex of the interio-basal process usually forms a right angle rather than being distinctly acute (see the comments on B. pyro- soma). The females of B. rufofasciatus can be distinguished by their nearly clear (hyaline) wings, which are more strongly darkened (infus- cated) for the other taxa of the rufofasciatus- group, and by their weaker punctures in the ocello-ocular area of the head. The males can be distinguished from males of B. richardsiellus or B. pyrosoma by the size and shape of their eyes, which are moderately enlarged relative to those of 106 the females (i.e. similar in shape to those of B. simillimus). The enlargement of the eyes of male B. rufo- fasciatus in comparison with the females reflects differences in their mate-searching behaviour that may contribute to a failure to interbreed with individuals of the other species where individuals of B. rufofasciatus occur with them (see the introduction on male mate-searching behaviour). For instance, B. rufofasciatus and B. pyrosoma overlap extensively in their foraging areas around the upper limit of the montane coniferous forest on Mt Apharwat (Fig. 6). However, workers and queens of B. pyrosoma are very much in the minority on the subalpine slopes above the forest, whereas those of B. rufofasciatus are rare within the forest. The males of B. pyrosoma, with eyes similar in size to those of the females, patrol circuits of scent-marked sites only within the upper montane forest. The males of B. rufofasci- atus, with enlarged eyes, watch for potential mates from perches on open, subalpine hilltops, where they usually establish exclusive territories. So individuals from these two populations are searching for mates in different ways and in different kinds of habitat (cf. the comments on B. asiaticus). No intermediates could be found in samples containing both B. rufofasciatus and B. pyrosoma s.\. [= B. flavothoracicus| from Kashmir or Nepal (BMNH, PW); B. rufofasciatus and B. richardsiellus from northern Burma (BMNH); B. rufofasciatus and B. pyrosoma s.1. [= B. friseanus] from Yunnan (BMNH) or Sichuan (PW); and B. rufofasciatus and B. pyro- soma s.str. from Gansu (Bischoff, 1936). A queen from Kashmir that has a broad band of yellow pubescence on tergum II was described under the name B. rufofasciatus var. championi by Richards (1928a). This individual is closely similar in morphology to those queens of B. rufofasciatus that have tergum II entirely black and it is considered to be part of the same species. DISTRIBUTION. B. rufofasciatus is a peri-Tibetan species (Map 61). It is known from Gansu (Morawitz, 1880; Bischoff, 1936; ZI, ZS), Qinghai (Morawitz, 1886; Tkalct, 1961; Wang, 1982; PW, ZS), Tibet (Friese, 1918; Richards, 1928a, 1928), 1930; Wang, 1982, 1988; BMNH), Sichuan (Morawitz, 1890; Wang, 1982; PW), Yunnan and northern Burma (BMNH), Arunachal Pradesh (Friese, 1918), Sikkim (Bingham, 1897; Friese, 1918; BMNH, UM), Nepal (Tkalci, 1974); BMNH, NMS, PW), Uttar Pradesh (Richards, 1928a, 1930; BMNH), Himachal Pradesh (BMNH, PW) and Kashmir. A record from Inner Mongolia [Ordos] (Morawitz, 1880) is dubious and needs PAULH. WILLIAMS to. be confirmed (cf. Skorikov, 1912b, on B. pyrosoma). MATERIAL EXAMINED FROM KASHMIR. In Kashmir, B. rufofasciatus is recorded from the Hindu Raj, Great Himalaya and Pir Panjal ranges, in mesic alpine scrub (Map 62, Fig. 6): 58 queens, 351 workers, 260 males, from 28 localities 1900— 4600 m (AB, BMNH, BPBM, MI, MNHN, NM, NR, PW, RH, SEMK, TL; 3 males photographed in the upper Wangat valley, 1 worker photo- graphed in the Sangisfaid valley, C. A. Chadwell). A colony was found on Mt Apharwat, at the lower edge of the subalpine pasture at Khilanmarg (3000 m, 4.ix.1985). The nest entrance was among boulders, mostly larger than 1 m across, that were lying several deep in a gully. This nest could not be excavated. I saw many queens investigate similar sites in the subalpine zone during July 1986, presumably in search of suitable nest sites. VARIATION WITHIN KASHMIR. Most queens from Mt Apharwat in the Pir Panjal range have a few yellow hairs intermixed with the black near the base of tergum II (Plate 1), but some individuals (10/32 queens from Apharwat, also 2/2 queens from around the Kishanganga valley) have an obvious band of yellow pubescence on the basal half of this tergum (Fig. 434). Queens from the Hindu Raj range (12/12 queens), at the western limit of the species’ distribution, are unusual in that the pubescence of tergum II is predominantly yellow, with at most only a narrow band of black hairs apically (Fig. 431). Workers and males from all parts of Kashmir generally have little black pubescence on tergum II (Plate 1 and Figs 432, 433, 435, 436, 438, 439), especially among indivi- duals from the Hindu Raj range (Figs 432 & 433). At the opposite extreme, some of the largest workers from Apharwat have only a narrow yellow band on tergum II (Fig. 438), or even just a few yellow hairs at the base of the tergum, so that it appears generally black (2/92 workers). Queens from further to the east in the Great Himalaya range generally share the darker colour pattern (Fig. 437), whereas workers generally retain yellow pubescence on tergum II (Figs 435 & 438). A male from Batakush has the pale bands on the thorax and tergum I yellow (1/5 males). The queens without obvious yellow pubescence on tergum II are quite distinctive in their appear- ance (Fig. 437). The yellow and white individuals of both castes and both sexes (Figs 431-436, 438, 439) can be particularly closely similar in colour pattern to the B. kashmirensis (Figs 302-304) that occur with them at some localities. B. rufofasci- atus can be recognised by the hairs of terga II-VI, which are shorter and either red (most of terga THE BUMBLE BEES OF THE KASHMIR HIMALAYA III-IV in workers and males; only tergum III in queens) or white (most of terga IV-VI in queens, terga V-VI in workers, terga V-VII in males), rather than red with white tips. They also resemble some of the B. himalayanus (Figs 247-249), B. novus males (Figs 260 & 262), B. lemniscatus (Figs 325 & 326), B. lepidus (Figs 330, 331, 333, 334) and B. biroi (Figs 338 & 339) that occur with them. Females of B. rufofasciatus can be recognised easily because they have terga V-VI with white hairs rather than with orange-red hairs. FOOD PLANTS. (Ranunculaceae) Aconitum laeve Royle, A. heterophyllum Wallich ex Royle, A. violaceum Jacquem. ex Stapf; (Balsaminaceae) Impatiens glandulifera Royle; (Leguminosae) Trifolium repens L., Lupinus sp. [introduced]; (Sambucaceae) Lonicera quinquelocularis Hardw.; (Compositae) Cirsium falconeri (Hook.f.) Petrak, C. wallichii DC., Lactuca lessertiana (DC.) C. B. Clarke, unidentified dandelion-like composites; (Ericaceae) Rhododendron anthopogon D. Don; (Gentianaceae) Swertia petiolata D. Don; (Scrophulariaceae) Scrophularia pauciflora Benth. ., Digitalis lanata Ehrh. [introduced], Pedicularis oederi Vahl, P. roylei Maxim., P. pectinata Wallich ex Benth., P. rhinanthoides Schrenk, P. punctata Decne.; (Labiateae) Phlomis bracteosa Royle ex Benth., Nepeta podostachys Benth., Prunella vulgaris L., Thymus linearis Benth. ex Benth. DISTRIBUTION OF THE FAUNA Distribution with altitude The slopes of Mt Apharwat in the Pir Panjal range (Map 4 locality 1) show many of the more mesic kinds of habitat in which bumble bees occur in Kashmir. Differences in the distributions of the bumble bee species with altitude can be described from transect walks of the north-eastern flank of Apharwat (see the introduction on mate- searching behaviour of male bumble bees in Kashmir). These transects ran from Tangmarg at 1800 m in the Vale of Kashmir, through Gulmarg at 2700 m, to the peak of Apharwat above at 4143 m (13,592 ft). More than 2000 bumble bees were examined in the field along these transects during July, August and September 1985. Fig. 6 shows the records for each of the social species between the 500 ft contours of the 1 : 63,360 British Survey of India map (Oxford University Library). These patterns of distribution are simi- lar to those found along the same transects in 1980 107 and 1986, and on other mountains around the Vale of Kashmir in all three years. Away from the foot of the mountain, the middle of the Vale of Kashmir around Dal Lake (1500 m) is marshy and much of this area is used for the cultivation of rice. No bumble bees were found there. As the land slopes gently towards the mountains, terraces of rice give way on drier ground to apple orchards and higher still, near Tangmarg, to maize. Above Tangmarg, the lower coniferous forest (1900-2700 m) is dominated by Pinus wallichiana A.B. Jackson, with Cedrus deodara (Roxb. ex D. Don) G. Don near the lower margin. The forest is protected by law, although its lower edge now largely coincides with a pronounced increase in slope just above Tangmarg. The understorey is very disturbed and although few trees are felled, their branches are still collected for firewood. Flowers used by bumble bees are scarce in the lower forest (species listed in Table 5) but are more abundant around buildings and along roadsides. Bumble bee densi- ties are very low, with B. tunicatus and B. trifasciatus the most abundant species both within the forest and around its lower edge. Gulmarg (= ‘flower-meadow’, 2700 m) oc- cupies a basin on a shoulder of the mountain. This natural, marshy meadow is grazed intensively by cattle, sheep and horses, yet the flowers from which bumble bees feed (Table 5) remain abun- dant. In the extensive wet areas of the meadow, Prunella vulgaris is a major nectar source and Pedicularis punctata a major pollen source for much of the summer. Large stands of the thistle Cirsium falconeri, which is often over 2 m tall, occur on disturbed slopes around the edges of the meadow. Impatiens glandulifera is abundant around water outflows from houses. The density of bumble bees is high (e.g. 89 bees/km, from 2.7 km transect, 14.ix.1986), with B. asiaticus and B. tunicatus the most abundant species, although there are fewer species than in the other zones (Fig. 6). The upper coniferous forest is dominated in turn by Pinus wallichiana (2700 m), Picea smithi- ana (Wallich) Boiss. (2700-2900 m) and Abies spectabilis (D. Don) Mirbel (2900-3000 m). Prominent members of the shrub layer are bushes of Sambucus wightiana Wallich ex Wight & Arn. and Euonymus hamiltonianus Wallich. Flowers used by bumble bees (Table 5) are few, as in the lower forest. Digitalis purpurea, and more com- monly D. Janata, are naturalised just above Gulmarg. Bumble bee density is low (23 bees/km, from 4.2 km, 14.ix.1986), with B. pyrosoma and B. tunicatus the most abundant species. The lower slopes (3000-3200 m) of the sub-alpine 108 PAULH. WILLIAMS Table 5 Lists of the plants that were visited most frequently by foraging bumble bees on Mt Apharwat during 1985 and 1986 in the different altitudinal zones (LF, lower forest; GM, Gulmarg meadow; UF, upper forest; SA, subalpine; A, alpine; +, visited frequently; +, major food source). Family Binomen Altitudinal zone LF GM UF SA A Ranunculaceae Balsaminaceae Leguminosae Caprifoliaceae Compositae Ericaceae Oleaceae Gentianaceae Scrophulariaceae Labiateae Aconitum heterophyllum Wallich ex Royle Aconitum hookeri Stapf Impatiens glandulifera Royle =F Lupinus sp. Trifolium repens L. Lonicera quinquelocularis Hardw. Cirsium falconeri (Hook. f.) Petrak Cirsium wallichii DC. * Lactuca lessertiana (DC.) C.B. Clarke yellow dandelion-like composites Rhododendron anthopogon D. Don Syringa emodi Wallich ex Royle Swertia petiolata D. Don Scrophularia pauciflora Benth. Digitalis purpurea L. Digitalis lanata Ehrh. + Pedicularis oederi Vahl Pedicularis roylei Maxim. Pedicularis pectinata Wallich ex Benth. Pedicularis punctata Decne. + Phlomis bracteosa Royle ex Benth. Stachys sericea Wallich ex Benth. Prunella vulgaris L. + Thymus linearis Benth. ex Benth. ++ * * ++ * ++ ++ +++ Salvia hians Royle ex Benth. Polygonaceae Bistorta vivipara (L.) Gray zone (Fig. 1) are dominated by species of Salix, Betula utilis D. Don and lilac, Syringa emodi. Large areas are strewn with boulders. Spaces between the rocks provide nest sites for bumble bees and refuges from grazing for Scrophularia pauciflora, which, despite its green flowers, is very attractive to bumble bees. Much of this zone is used as pasture for goats, sheep and cattle, especially in the more gently sloping area, known as Khilanmarg, that is immediately above the forest. Pasture areas have many bushes of Euphorbia wallichii Hook. f. and flowers of Anemone obtusiloba D. Don. The steeper middle slopes (3200-3400 m) are dominated by B. utilis and Rhododendron campanulatum D. Don. The upper slopes (3400-3800 m) are dominated by R. anthopogon, Juniperus recurva Buch.-Ham. ex D. Don and Bergenia stracheyi (Hook. f. & Thoms.) Engl. Overall, the subalpine zone has the largest number of bumble bee species (Fig. 6), although their total density (32 bees/km, from 4.2 km, ++ 12.ix.1986) is less than at Gulmarg. The most abundant species throughout the subalpine zone is B. rufofasciatus. The alpine zone (3800-4143 m) is marked by an absence of the low bushes of R. anthopogon and by the presence, at least in exposed areas, of a silver-grey mat vegetation with Anaphalis triplinervis (Sims) C.B. Clarke and species of Leontopodium (Edelweiss). Bistorta vivipara is conspicuous, but the major food-plants for bumble bees are species of Pedicularis, Swertia petiolata etc. (Table 5). The most abundant bumble bee species are B. rufofasciatus and B. kashmirensis (density of all bumble bees 31 bees/km, from 2.3 km, 12.1x.1986). These two species are closely similar in colour pattern and general appearance, but all of the individuals with this colour pattern that were caught visiting flowers of Compositae were of B. rufofasciatus, whereas those visiting Aconitum hookeri were all of B. kashmirensis. THE BUMBLE BEES OF THE KASHMIR HIMALAYA The overall pattern in the distribution of the species’ foraging activities with altitude on Mt Apharwat (Fig. 6) is very similar to that found on other mountains around the Vale of Kashmir (e.g. near Pahalgam and near Daksum). The overlaps in the species’ altitudinal ranges on these moun- tains were more extensive than those measured along the floor of the Sind valley (Fig. 12). For instance, species such as B. trifasciatus and B. tunicatus are restricted to the lower end of the Sind valley (e.g. Kangan, 1900 m) and other species such as B. rufofasciatus and B. keriensis are restricted to the upper end (e.g. Sonamarg, 2700 m), although they have all been collected from the same meadow on Apharwat (2700 m). The Sind river runs from deep in the Great Himalaya to the Vale of Kashmir, so the slope of the valley floor is far less steep than the north- eastern flank of Mt Apharwat. It is more likely that individuals could wander or commute to forage outside their optimal altitudinal zones on Apharwat, simply because the zones are so much closer together horizontally. B. simillimus was common in the lower Sind valley (e.g. at Kangan, 1900 m) but was not found on Apharwat during 1980 or 1985. Three workers of this species were found at 3000 m on Apharwat in 1986. In general, the distributions of bumble bee species are most likely to be governed by the combined effects of climate and of local levels of food resources (discussed in Williams, 1988, 1989). The environment in Kashmir is particularly complex, because the regional climate may be greatly modified locally by the effects of altitude and exposure (e.g. Troll, 1972). This is likely to have three principal direct effects on bumble bees. First, lower average temperatures at higher alti- tudes may increase thermoregulatory expenditure of energy by bumble bees, especially for queens, which must incubate their first brood alone in the spring. Second, a shorter annual season of temperatures above freezing point in the high mountains constrains the time available for colony development. However, daytime temperatures even in the alpine zone of Kashmir may be higher and more favourable than further north (e.g. in the Alps), although the day lengths are shorter in Kashmir. But third, at lower altitudes, bumble bees may actually suffer from heat stress. For instance, foraging activity by bumble bees (chiefly B. trifasciatus and B. haemorrhoidalis) at Patnitop (2000 m) in the Jammu foothills ceased during the warmest part of the day, between 10 a.m. and 5 p.m. (6-8.ix.1986), when the patches of Impatiens glandulifera and Pteracanthus urtici- folius from which they were foraging were no longer in the shadows of the forest trees. The 109 bumble bees could not be found visiting any other flowers elsewhere during these periods. A less pronounced trend towards bimodality in diurnal foraging patterns is well known for some temper- ate bumble bees (e.g. Heinrich, 1979; Willmer, 1983) and is also reported for the large carpenter bees of the genus Xylocopa at a subtropical lowland site (Abrol, 1987). In both of those situations, foraging activity by the bees was in- versely related to air temperature and intensity of direct solar radiation, but was positively associ- ated with nectar availability. At Patnitop, small and medium-sized anthophorine bees continued to forage from the same flowers throughout the day, so presumably at least some nectar was still available. A restriction of diurnal foraging periods for bumble bees is likely to reduce the daily rate of food supply to colonies and so slow their develop- ment within the limited foraging season. This has at least the potential to limit a species’ distribution at the southern edge of its range. In common with many other organisms, bumble bees tend to occur at higher altitudes in Kashmir than they do further north, but in zones with similar climate. B. lucorum is most abundant on Mt Apharwat around the upper edge of the montane forest zone (c. 3000 m) and is recorded from 2100-4600 m at other sites in Kashmir. Further north in the Alps, it is common on the forested middle slopes (1300-2600 m recorded by Pittioni, 1937). In cool temperate Britain, this species is common in virtually all of the available altitudinal zones down to sea level (0—-1200 m). Climate may also affect bumble bees indirectly through its effects on food resources. The lower parts of the valleys in Kashmir often appear to receive less rain or snow in summer than the mountain ridges. In some of the drier areas the availability of bumble bee food-plants in late summer may be very restricted and bumble bees are sometimes either absent or present at only very low density (e.g. at Baramula and Anantnag in the Vale of Kashmir). Food sources are also scant in the shade of the evergreen, coniferous forests on the hillsides. But higher in the mesic subalpine zone, the density of nectar-rich food- plants is often high and their diversity may be at its highest. In the alpine zone, the frequent freezing and overcast conditions throughout the summer are likely to reduce plant productivity and hence resources for bumble bees. The favour- able balance of these climatic and food resource factors in the subalpine zone may help to account for its large number of species of bumble bees (Fig. 6). Although the meadow at Gulmarg has much the highest density of bumble bees on Mt Apharwat, 110 @® Type ia] yellow banded/ @ Type 1B @ Type al grey-white © Type 2B @ Type 3 HIGH oberti ALPINE personatus ladakhensis ALPINE kashmirensis himalayanus keriensis =) MESIC rufofasciatus (=) @ SUBALPINE _ biroi @ g avinoviellus S ARID subtypicus SUBALPINE — semenovianus STEPPE asiaticus on en ee) melanurus @ @ UPPER lepidus g @ FOREST lucorum hypnorum @ pyrosoma @ & LOWER tunicatus iS) < (=) < S < FOREST simillimus O O trifasciatus g ® ® haemorrhoidalis < | s 3 a Predominant local < colour patterns of workers: S oD Ks) a ‘ = iS oun, unbanded thorax 7S i) & / roy isi =I eS & N ZANSKAR LADS S RANGE SI RANGE banded thorax s nS GREAT HIMALAYA Z ee PANJAL RANGE 3 $ black thorax RANGE 3 = Sind R. Dras R. — y PAULH. WILLIAMS SES OES Z @ oo @ @<« Qs @ @« ®9 8 S&S @ ; o ®@ Oo ® 6 ti) : 3 AZ. /a\ Martselang Tiggur Fig. 12 Distribution of the social species of bumble bees (i.e. excluding species of the subgenus Psithyrus) and their local colour patterns across the Himalaya. Only the locally predominant patterns among the workers are represented. Triangles show the most abundant species at each locality. B. lemniscatus is not included for lack of information. it has rather few species (Fig. 6). This may be partly because the abundance of food-plants is particularly uneven, although the relationship between diversity of food-plants and diversity of bees may not be straightforward (Williams, 1989). Pedicularis punctata and Prunella vulgaris are by far the most abundant of the food-plants used by bumble bees in the meadow. The meadow is often extensively flooded and some areas around streams are permanently marshy, although these plants are apparently tolerant of water-logged soil. Most of the more abundant species of bumble bees in the meadow, such as B. tunicatus and B. pyrosoma, are also among the most common species in the forest. Nests of both species have been found in the forest, but not in the meadow, so it may be that many of their foragers commute from nests in the forest to the richer resources in the meadow. B. asiaticus is unusual because it is both abundant on Gulmarg meadow and gener- ally associated with open habitats. The locality at the highest altitude in Kashmir from which I have collected a large sample of bumble bees is on Nimaling plain (4800 m). This is a high glacial valley, surrounded by large areas of arid alpine steppe that have very few if any bumble bees. Visits for 13 days in July 1980 and for three days in September 1986 found nine species of bumble bees (including the social para- site B. (Psithyrus) branickii) visiting the flowers of only a single plant species, Caragana versicolor. This plant forms low, thorny bushes on the old Altitude 6000 5000 4000 3000 2000 1000 0 THE BUMBLE BEES OF THE KASHMIR HIMALAYA lateral and terminal moraines (Fig. 3), from which the glacier snout has retreated up the valley by about 10 km to the south-east. If the foragers do all genuinely use what appears to be a single food- resource, and there were no other obvious poten- tial food-plants to be seen, then this situation deserves further study. It would follow from the application of competition-based ideas of ‘com- munity structure’ to these bumble bees (e.g. Inouye, 1977) either that the flowers of Caragana are very variable in the accessibility of their rewards to the different bumble bee species, or that other factors are preventing the local assem- blage of bumble bee species from reaching equili- brium by the local extinction of most of the species (cf. Williams, 1989). Fig. 6 shows the narrower range of altitudes and habitats in which the males search for mates on Mt Apharwat, in comparison with the range of habi- tats in which the females forage. The possible relationship between the different kinds of male mate-searching behaviour and the different kinds of habitat is summarised in the introduction. Distribution across the Himalaya The distribution of bumble bees across the Kashmir Himalaya can be described from the samples collected at selected sites in the Pir Panjal, Great Himalaya, Zanskar and Ladakh ranges during August and September 1986. These data are supplemented with records from speci- mens collected at Nimaling in 1980 and with others examined in collections from Muzaffarabad (BMNH), from the Khardung-La above Leh (FA, ZI) and from Tiggur near the Nubra/Shyok confluence (BMNH). Together this covers a hori- zontal distance of about 430 km from west to east, at about 34° latitude. Fig. 12 shows the distribu- tions of the bumble bee species in these samples and their locally predominant colour patterns. The most widespread species in Kashmir are the steppe species B. melanurus and B. asiaticus. B. asiaticus shows a major change in colour pattern between Nigagar and Gumri in the Great Himalaya range, with some intermediate individuals present at Nigagar (see the discussion of variation of B. asiaticus in Kashmir). The alpine element in the fauna is also widespread between the Vale of Kashmir and Zanskar. Like B. asiaticus, B. kashmirensis and B. himalayanus show a pro- nounced change in colour pattern, but in this case the major transition in the thoracic pattern is to the east of Gumri (and to the west of Nimaling). The principal boundary in the composition of the fauna is associated with the Great Himalaya range. This region of transition was studied in 111 more detail between 9-11.viii.1986 in samples from Sonamarg, Nigagar, Baltal, Gumri, Matayan and Dras (a gradient-directed transect, see Map 4 & Table 1). The forest element is represented at Sonamarg, Nigagar and Baltal but, like the forest, it is absent to the north-east, beyond the Zoji-La (La = pass). There is also a large change in the subalpine element across the Great Himalaya. Of the subalpine species found on Mt Apharwat, B. rufofasciatus reaches across the Great Himalaya as far as Matayan, although B. biroi may be more widely distributed in Zanskar and Ladakh (Map 42). At Matayan, B. rufofasciatus occurs together with B. subtypicus, although B. rufofasciatus is no longer the most abundant species, as it is in the subalpine zones around the Vale of Kashmir. The three high alpine species found at Nimaling are distributed widely across the Tibetan massif but are apparently absent from the mountains around the Vale of Kashmir. This abrupt discontinuity in the faunal composition reflects the transition in climate and vegetation from the more mesic, wooded environment of the Vale of Kashmir to the arid (= xeric), steppe environment of the high Tibetan massif (Map 3). The environmental dif- ferences between these two regions in Kashmir are more extreme than those across the eastern Himalaya (Chang, 1981), where the Tibetan plateau is generally lower, more humid and dissected by wooded valleys. This is reflected in the distribution of bumble bees such as B. rufo- fasciatus, which penetrates much more deeply into Tibet in the east (Wang, 1982; BMNH). A second boundary in the composition of the fauna is located on the outer slopes of the Pir Panjal range. The low altitude fauna has not been studied thoroughly here, largely for reasons of inaccessibility (see the comments on material examined). It is dominated by a lower montane forest element (B. haemorrhoidalis and B. trifas- ciatus). This includes many more species in the more humid foothills of the central and eastern Himalaya, where the effects of the monsoon are more pronounced. From the limited amount of material available (BMNH, NMS, PW), the general pattern in the distribution of bumble bees across the Himalaya of Kashmir is similar to that found across the Himalaya of western Nepal (Table 6). The major difference is that the same arid subalpine element is not known to be represented in Nepal or in the adjacent parts of Tibet. The Kashmir-western Nepal region of the Himalaya spans about 1000 km, but shares a broadly similar patterfiof climate and vegetation (see the maps in Schweinfurth, 1957). The pattern of vegetation formations is highly convoluted locally because it 112 PAULH. WILLIAMS Table 6 Distribution of the social bumble bees of Kashmir (i.e. excluding species of the subgenus Psithyrus) outside Kashmir. X — present in the regional lists from world summary grid-squares (see Map 1): AP, western Alps; AR, Armenia and the Caucasus; MG, northern Mongolia; TS, central Tien Shan; HK, Hindu Kush; NP, Nepal; GS, southern Gansu; BM, central Burma; TW, Taiwan. Region Habitat in Kashmir: AP AR MG _ TS High alpine x Alpine x x x Mesic subalpine x xX Arid subalpine x Steppe x x x ».4 x Upper forest x x x x x x x Lower forest tends to follow contours, but throughout this region it consists essentially of bands parallel to the Great Himalaya (e.g. Map 3). So for bumble bees, as for the vegetation and for many other groups of organisms (e.g. Mani, 1986), the fauna apparently changes to a greater extent in a small distance across the Himalaya than it does along them. This supports the familiar idea that distri- butions are restricted less by the capacity of Organisms to spread through a habitat than by their capacity to colonise a different kind of habitat (e.g. Fretwell & Lucas, 1970). Distribution outside Kashmir Only the broadest patterns in the distribution of bumble bees outside Kashmir can be discussed with- out an analysis of the distribution of all bumble bee species (in prep.). Table 6 summarises the distribu- tions of the social bumble bees of Kashmir (i.e. excluding species of the subgenus Psithyrus) across some of the mountainous areas of Europe and Asia. The species are listed in a sequence to show their altitudinal distribution and habitat associa- tion within Kashmir (see Figs 6 & 12). HK mmx KM g Kashmir BM oberti personatus ladakhensis kashmirensis himalayanus keriensis rufofasciatus biroi lemniscatus avinoviellus ~~ Km Km KM RM ~ Km KM KR MR marussinus subtypicus semenovianus asiaticus melanurus lepidus lucorum hypnorum pyrosoma KKK MK mM tunicatus simillimus trifasciatus haemorrhoidalis aM MRM KM * ~*~ The alpine and high alpine species of Kashmir are primarily associated with the Tibetan massif (Table 6: reaching Nepal, Gansu). The distribu- tions of B. keriensis and B. oberti also extend westwards and northwards through the Pamir to the Tien Shan ranges and, for B. keriensis, further to Armenia in the west and to the mountains of northern Mongolia in the north-east. The subalpine and steppe species of Kashmir are principally associated with the arid mountains to the north in central Asia (Table 6: Hindu Kush, Alai-Tien Shan s.1.). There are also a couple of very widespread, cool temperate and upper montane forest species (B. lucorum and B. hypnorum). Species of the arid subalpine element are at the south-eastern limits of their distribu- tions in Kashmir. Species of the steppe element are also more widespread in central Asia, but reach beyond Kashmir eastwards into the Tibetan massif. Together with the alpine species these correspond to the ‘western hypsobiont element’ described from the Himalayan butterfly fauna by Mani (1986). The remaining species are found in and around the forests at lower altitudes and are more THE BUMBLE BEES OF THE KASHMIR HIMALAYA southern and eastern in their distribution outside Kashmir. They are either restricted to the western Himalaya (B. tunicatus and B. simillimus) or are at the western limit of their distributions (Table 6: reaching eastwards to Nepal and Taiwan). B. haemorrhoidalis and B. trifasciatus reach east- wards and southwards into the Oriental tropics in the mountains of northern Thailand. Together this group of eastern species corresponds to the ‘eastern forest element’ described from butterflies by Mani (1986). Four species are narrowly endemic to the west- ern Himalaya, B. avinoviellus, B. himalayanus, B. tunicatus and B. simillimus. It is possible that B. himalayanus may prove to be conspecific with the central Asian taxa B. turkestanicus and B. marussinus (see the comments on B. marussinus and B. himalayanus). The three remaining species are among the more plesiomorphic species within their respective species-groups. If they are relics of a relatively ancient bumble bee fauna that was once more widespread, then it now persists in a region that has been subject to some of the most active mountain building during the last few million years (see Bally, 1983; Mani, 1986). But although the western Himalaya may not appear to have been a particularly stable refuge in this period, the high relief could have provided some buffering against any climatic fluctuations by allowing altitudinal displacement of bumble bee populations. Overall, the bumble bee fauna of Kashmir includes particularly few of the relatives of the European B. hortorum (only B. trifasciatus) or B. pascuorum (only B. haemorrhoidalis), but par- ticularly many of the relatives of the European B. lapidarius (8 species of the subgenera Sibirico- bombus and Melanobombus). This pattern (9% Megabombus + Diversobombus + Thoraco- bombus + Orientalibombus : 35% Sibiricobombus + Melanobombus, Kashmir fauna excluding Psithyrus) contrasts markedly with the repre- sentation of these groups in the fauna of the European Kalsbach Alps (31% : 6%, Pittioni, 1937), in the fauna of the whole of Europe (34% : 8%, Reinig, 1981) or in the fauna of the western Palaearctic (32% : 10%, Rasmont, 1983). Relationships between patterns at different spatial scales Among the social bumble bees, those species that are more widespread world-wide are not generally found at more localities within Kashmir (Fig. 13, which excludes the species of the subgenus Psithyrus, Kendall coefficient of rank correlation 0.08, P > 0.5 n.s.). It is possible that these data 113 might not be representative because the sampling effort was far from uniform at both spatial scales and the sites in Kashmir were not randomly located but selected by kind of habitat and for ease of access. Nonetheless, there is pattern in the data insofar as species with similar habitat associations (Figs 6 & 12) do tend to occur closer together in Fig. 13. One interpretation is that this pattern in Fig. 13 may show a common limiting effect within the groups of species with similar habitat require- ments, which could be just a consequence of the limited availability of the most suitable kinds of habitat for these species at both spatial scales. From a study of British bumble bees, patterns in local distributions appear to be related to how close the species are to the centres of their world- wide distributions (Williams, 1988). Among the Kashmir fauna, those species that are nearer to the centres of their world-wide distributions (from Maps 5-62: B. avinoviellus, B. himalayanus, B. melanurus, B. tunicatus, B. asiaticus, B. semeno- vianus, B. keriensis and B. simillimus) do also tend to occur at more localities within the country (Maps 5-62; Fisher-Pitman randomisation test, W = 181, single-tailed P < 0.01, see Krauth, 1988). More precisely, those species that are nearer to the centres of their distributions in Britain have been found in a broader range of kinds of habitat. For the fauna of Kashmir, this greater breadth of habitat use for those species that are nearer to the centres of their world-wide distributions is apparent in the greater altitudinal ranges of these species within the country (exclud- ing single altitude records of B. oberti and B. lemniscatus, Fisher-Pitman randomisation test, W = 18400, single-tailed P < 0.05). Fig. 13 also appears to contain information as to how close the species are to the centres of their world-wide distributions. For these data in which the variance is similar in the log(world) and log(kashmir) distribution values, the central/ marginal position of a species appears to be related to log(kashmir)-log(world). Since all species tend to occupy more localities nearer to the centres of their distributions, then if the more widespread species world-wide were generally found at more localities within comparable parts of their distributions, perhaps the deviation of the Kashmir species from any underlying positive relationship in Fig. 13 might show some influence of how far Kashmir is from their centres of world- wide distribution. The relationship between distribution patterns and patterns of local abundance found among British bumble bees (Williams, 1988) is not consis- tently evident in the samples from Kashmir. Those species that were chosen from the maps to 114 PAULH. WILLIAMS 100 distribution centre WIDESPREAD near Kashmir SPECIES oO aSialticus tunicalus e rufofasciatus A C] melanurus EGIL OY Bs kashmirensis (a keriensis ow 28 i: A ee biroi VIN ie himalayanus 7 aris O lucorum £5 10 avinoviellus GI pyrosoma aM ; 2s simillimus subtypicus 25 3 = wy rifasciatus s 3 & marussinus L\ © hypnorum ladakhensis we haemorrhoidalis personatus Vv RESTRICTED distribution margin 1-] SPECIES oberti OY e lemniscatus near Kashmir 1 10 100 distribution world-wide (number of world grid-squares) O Palaearctic & Oriental, cool temperate/upper montane forest species ial Central Asian & Tibetan, temperate/montane steppe species A Central Asian, montane steppe species vV Tibetan, alpine steppe species A Oriental (peri-Tibetan), alpine steppe species @~ Oriental, upper montane forest species W Oriental, lower montane forest species p endemic West Himalayan species Fig. 13 Plot of how widespread the species of social bumble bees (i.e. excluding species of the subgenus Psithyrus) are within Kashmir against how widespread they are world-wide (logarithmic axes). Data are taken from occupancy of the number of localities within Kashmir (see Map 4) and from the number of world summary grid-squares (see Map 1) respectively. The distributions of B. tunicatus and B. lemniscatus are anomalous. THE BUMBLE BEES OF THE KASHMIR HIMALAYA be nearer to the centres of their world-wide distributions in Kashmir do not appear to be generally more abundant locally in these samples (local abundance for a species estimated as the total numbers of individuals examined from Kashmir / numbers of localities in Kashmir, Fisher-Pitman randomisation test, W = 145, single-tailed P = 0.19 n.s.). However, there is still a tendency for the more widespread species within Kashmir to be more abundant locally (Kendall coefficient of rank correlation 0.29, single-tailed P < (00.025). It is possible that even if there were a relationship with the world-wide pattern, it may have been obscured in these data because the later collections at least were strongly biassed against including many specimens of the common species from each site, whereas the rare species were deliberately sought. Relationships between the patterns of distribu- tion at different spatial scales form the basis of a simple model for the interaction of those ecologi- cal factors that are most likely to govern bumble bee distributions (discussed in Williams, 1988, 1989). It follows from this that where some disturbance of sites causes a major reduction in food availability, it is the species that are nearest to the centres of their world-wide distributions that are most likely to persist. This may provide one explanation for the observation that B. tuni- catus and B. asiaticus are the species that are often found alone in severely over-grazed areas around the edge of the Vale of Kashmir and in the Sind valley. COLOUR PATTERNS OF THE KASHMIR FAUNA The colour patterns of bumble bees often vary greatly among individuals of one species (e.g. Fig. 11), but may be closely similar between indivi- duals of unrelated species where these species occur together (Fig. 12; see e.g. Vogt, 1909, 1911; Sakagami & Yoshikawa, 1961; Tkalci, 1968), 1989; Pekkarinen, 1979; Plowright & Owen, 1980). Genetic determination of characteristics of the colour pattern with simple Mendelian inheritance has been demonstrated for the North American B. melanopygus and B. rufocinctus (Owen & Plowright, 1980; Plowright & Owen, 1980; Owen, 1986). But in the face of such marked regional convergence, several authors have ques- tioned whether variation among colour patterns of individuals within a population is always deter- mined genetically, or whether it is perhaps more 115 often determined directly by environmental factors, such as diet, temperature or humidity during development (e.g. Vogt, 1909, 1911; Reinig, 1935; Pittioni, 1942, 1943). The pubescence of bumble bees develops during the pupal stage and the colours appear soon after the adults emerge from their cocoons. Adult queen and worker bees invest much time and energy ensuring that the conditions in the nest remain nearly constant, for instance by incubating brood or by fanning air over the nest to regulate temperature as necessary (e.g. Heinrich, 1979). Hence the brood is shielded from many of the possible environmental effects on colour pattern during its development, so colonies usually pro- duce individuals with relatively homogeneous colour patterns. However, if a nest is exposed and moved to a nest box, unusual colour patterns are more often found among the adults that sub- sequently emerge from pupation (pers. obs.). The few examples of pronounced caste dimor- phisms show that colour patterns are not solely determined by genes, but must also be influenced by the environment. Two examples have been described previously from subtropical environ- ments (B. ephippiatus from the mountains of Costa Rica, see Owen & Plowright, 1980; B. festivus from the eastern Himalaya, see Ito et al., 1984), in which workers and males from the same colonies share similar colour patterns that differ strongly from those of the queens. Similar differ- ences exist within species of the rufofasciatus- group from Kashmir (see Plate 1). For B. pyrosoma, males usually also differ in that they have the pubescence predominantly yellow (Figs 424, 427, 430), a characteristic that can be regarded as sex-linked (see Stiles, 1979). But colour variation among the females from Kashmir appears to be related to body size (which increases through Figs 426, 428, 429, 425, see the comments on B. pyrosoma) and so to larval food intake. From the material available it seems most likely that the white-tailed queens from Kashmir belong to the same population as the nominal taxon B. friseanus from Yunnan (synonymised with B. pyrosoma s.1.). These queens from Yunnan have the pubescence of terga IV-V red, so they closely resemble their own workers from Yunnan and also many of the workers from Kashmir. Thus caste dimorphism in colour pattern may exist in only part of the population of B. pyrosoma s.1. Variation of B. asiaticus in Kashmir The asiaticus-group is particularly well repre- sented in collections for an investigation of the variation in the colour pattern of the pubescence. 116 In Kashmir, two principal nominal taxa occur on opposite sides of the Great Himalaya range (Map 48), one with an unbanded yellow thorax (Figs 371-376), the other usually with a black band between the wing bases and an otherwise grey- white thorax (Figs 383-391). These nominal taxa are believed to be parts of a single species (see the comments on B. asiaticus). The basis of this colour variation can be exam- ined where the regional colour patterns occur together. At a few sites at the upper end of the higher valleys in the Great Himalaya, the B. asiaticus are not uniformly of one or other of the two principal colour patterns, but instead many individuals show a variety of combinations of their differing character states (Figs 377-382). The largest sample with this kind of variation was collected by F. Schmid from Lal Pani in the upper Kishanganga valley between 2-3.ix.1953 (BMNH: Table 7). This is likely to be a relatively unbiassed sample insofar as he specialised in collecting Trichoptera and not particular bumble bee taxa. I have also collected a small sample that shows similar variation, from Nigagar in the upper Sind valley (9.viii & 17.ix.1986, PW). The occur- rence of the black tail-colour state (Figs 379-381) at both of these localities is curious. It is not known for individuals from the adjacent regions of Ladakh or the Vale of Kashmir, although it is common in parts of the Hindu Kush ranges (BMNH). The ability of simple population genetic models to account for the observed patterns of variation Table 7 Numbers of individuals of B. asiaticus collected by Schmid from Lal Pani 2-3.ix.1953 with each combination of states of three characters of the colour pattern of the pubescence. The scoring was simplified prior to analysis in order to present discrete states despite some limited intergradation. The black thoracic band is deemed to be present if it appears well defined from the dorsal aspect. The pale colour ‘white’ includes very pale cream. The ‘tail’ colour is deemed to be red if the pubescence of at least two of the apical terga is predominantly red. Thoracic band Pale bands Tail colour Females Males present white red 7 1* absent white red - 1 present white black 2 1 absent white black 1 - present yellow black 3 34 absent yellow black - 13 present yellow red 9 37 absent yellow red - Se * B. callophenax of Richards, not of Cockerell ** B. longiceps Smith PAULH. WILLIAMS Table 8 Association between pairs of character states of the colour pattern among the males of B. asiaticus from Lal Pani (see Table 7). None of these comparisons shows significant association (chi-square two-sample test, all comparisons P>0.05). Pale bands Tail colour white yellow red black Thoracic present 2 71 38 35 band absent 18 6 13 if Tail red 2 42 colour black 1 among the specimens from these boundary sites can be tested if certain assumptions are made. Three aspects of the variation are chosen because they appear to be among the most important elements of the bees’ colour pattern as they are viewed flying among flowers. It is assumed initially that each character shows either of two states and that these are each determined by one of a pair of alleles at a single locus. In fact the three characters selected do not show entirely discrete states. However, the minority of intermediate specimens from the possible hybrid zone can be assigned to these states for a preliminary analysis (Table 7), at least until further information on modifier effects is available. There is no association between the states of the three colour characters among individuals from the top of the Kishanganga valley (Table 8). This apparent independent assortment of the charac- ters shows an absence of the disequilibrium (e.g. Mallet, 1986) that would be expected if gene flow across the coincident character clines in the hybrid zone were large. Hybrid zones are expected to be captured by linear habitat features where popula- tion density is particularly low (e.g. Mallet, 1986; Hewitt, 1988). For B. asiaticus in Kashmir, this role may be fulfilled by the high ridges and passes of the Great Himalaya range. For instance, where the upper Sind valley reaches the main ridge of the Great Himalaya just above Nigagar, the valley floor rises abruptly by about 700 m at the Zoji-La pass. My small samples from either side of Nigagar, at Sonamarg (2700 m) and above Gummi (3700 m), each consist of material with a single colour pattern (i.e. the alleles may be nearly fixed), so the hybrid zone at the top of the Sind valley may be less than 16 km wide. Such a narrow cline is consistent with the interpretation that there is little gene flow across the hybrid zone at the top of the Kishanganga valley (Table 8). The hybrid zone at the top of the Kishanganga may, nonethe- less, be much broader (Map 48, see the comments on variation within Kashmir for B. asiaticus). THE BUMBLE BEES OF THE KASHMIR HIMALAYA GLY? Table 9 Observed numbers of individuals of B. asiaticus from Lal Pani with each state of three colour pattern characters (see Table 7), together with the expected numbers from simple genetic models. Frequencies from a small sample (n = 20) from Nigagar are included for comparison. Deviation from model significant at P<0.05 *, P<0.001 ***, or not significant ns. Observed: females (diploid) Thoracic band present 21 absent 1 Frequency (absent) Lal Pani 0.05 [Frequency(absent) Nigagar 0.00 (Chi-square one-sample test, df = 1) Expected: males females females (haploid) —nodominance -simple dominance 73 17 [est(p) = 73/92] 21 19 5 [est(q) = 19/92] 1 0.21 0.66] (4.14) (0.00) P>>0.05 ns Probability that deviation of observation from expectation is due to chanceP<0.05 * Pale bands white 10 yellow 12 Frequency (yellow) Lal Pani 0.55 [Frequency (yellow) Nigagar 0.50 (Chi-square one-sample test, df = 1) 3 1 [est(p) = 3/92] 1 89 21 [est(q) = 89/92] 21 0.97 0.66] (84.86) (84.86) P<<0.05 *** Probability that deviation of observation from expectation is due to chanceP<<0.05 *** Tail colour red 16 black Frequency (black) Lal Pani 0.27 [Frequency (black) Nigagar 0.14 (Chi-square one-sample test, df = 1) Probability that deviation of observation from expectation is due to chanceP<0.05 * The frequency of alleles can be estimated and the dominant allele can be determined from the frequencies of the character states, because although female bumble bees are diploid, the males are usually haploid (see Plowright & Pallett, 1979). The frequency of the character states among males in the large sample from Lal Pani is assumed to be representative of the allele frequencies in that local part of the population (which assumes that even if the material was collected in the proximity of only a very few colonies, it still represents a random sample of the local gene pool). These frequencies may differ at other localities in the Kishanganga valley or at Nigagar, so these samples cannot be combined for analysis. If there is simple, complete dominance by one of the two alternative alleles at each locus, then the frequencies of character states should differ between males and females, with higher frequencies of expression for the recessive allele among males. The higher frequencies among males of the lack of a black thoracic band, of yellow pale bands and of black ‘tails’ at both Lal Pani and Nigagar (Table 9) are consistent with their determination by recessive alleles. Further- more, the frequencies of characters among males 44 11 [est(p) = 44/92] 16 48 11 [est(q) = 48/92] 6 0.52 0.33] (4.55) (0.00) P>>0.05 ns can be used to calculate the frequencies among females that would be expected if there were no dominance (which might result from Hardy- Weinberg equilibrium with co-dominance of both alleles, or from some mechanism other than straightforward Mendelian genetics). These cal- culated frequencies are found to differ signifi- cantly from those observed for all three characters (Table 9). The observed frequencies of the states of the black thoracic band and of the tail colour among females are consistent with those expected from the Hardy-Weinberg equilibrium model with simple, complete dominance between two alleles (Table 9). However, this is rejected for the pale band colour character, the distribution of which was not consistent with the dominance-free model either. It has to be noted that the chi-square one- sample test is not reliable when an expected frequency is less than 5 (e.g. Siegel, 1956), but the values of chi-square that result here are so large (84.8, df = 1) that the conclusion that the model should be rejected is unlikely to be altered. The reason for the failure of the models must be that their assumptions are violated. In addition to those mentioned above, assumptions of the 118 equilibrium model include insignificant dispersal, random mating and unselective predation. Firm conclusions as to the genetic basis of the patterns of colour variation among these bumble bees cannot be drawn without breeding experi- ments. This would allow ideas of the pattern of inheritance to be tested, although this is not particularly easy with bumble bees (but see Owen & Plowright, 1980). However, the pattern of variation for two out of three of the characters that have been analysed is consistent with simple genetic models. It is therefore also consistent with the assumption that individuals with these colour patterns are interbreeding as parts of a single population and a single biological species. Thus local climate or diet does not appear to be the proximate factor governing the close convergence in colour pattern between this species and the regionally covarying groups of species on either side of the Great Himalaya. Convergent colour patterns It is noticeable that only a few of the possible colour patterns are actually found among bumble bees world-wide, given the range of colours shown by the pale pubescence of bumble bees and the observed patterns of disposition of bands of pale and black pubescence on their bodies (Plowright & Owen, 1980; BMNH, PW). For instance, at least half of all species of bumble bees have some individuals with a ‘red-tailed’ colour pattern. Other common elements in the colour pattern are a band of black hairs on the thoracic dorsum between the wing bases, and a black band on tergum III. The most widespread colour pattern of the dorsal pubescence among all bumble bees of the world appears to be for the thorax to be yellow with a band of black hairs between the wing bases and for the gaster to be yellow — black — red, from base to apex (cf. Figs 233-235, 242, 243, 298-300, 317, 318, 377, 378, 382, 388, 393, 407, 412-416, 426). The three principal regions of Kashmir have groups of bumble bee species with characteristic colour patterns of the dorsal pubescence (Fig. 12): (1A) In the high alpine zone of Ladakh, the most common colour pattern of the pubescence is for the thorax to have yellow transverse bands anteriorly and posteriorly, with a black band between the wing bases. The gaster is yellow (terga I-II) and red (terga III-V), although often with a narrow black band between them (basal quarter or less of tergum III). (1B) At lower altitudes in the arid steppe of PAULH. WILLIAMS Ladakh, the common colour pattern is for the pubescence of the thoracic dorsum to be uniformly yellow and for the gaster to have a yellow band (terga I-II yellow, or just tergum I yellow, or terga I-II completely black) — black (at least tergum III) — red (usually terga IV-V). (2A) Around the Vale of Kashmir, above the forests, the most common pattern is for the pubescence of the thoracic dorsum to be grey-white with a black band between the wing bases. The gaster is grey-white (tergum I) — yellow (usually tergum II) —red (terga III-I[V, sometimes with white on terga IV-V). (2B) Inthe Vale of Kashmir and the surrounding upper montane forest, the common colour pattern is banded and white as for type 2A, but the gaster lacks any yellow and is grey- white (tergum I, sometimes tergum IT) — black (at least tergum III) — red (usually terga IV-V). (3) Inthe lower montane forests of the Jammu foothills, the common colour pattern is for the pubescence of the thorax to be black and for that of the gaster to be yellow (terga I-II) and red (terga III-V), but without an obvious black band. The type 1A colour pattern deviates only slightly from the widespread bumble bee pattern of three yellow bands and a red ‘tail’ by the replacement of much of the black on tergum III by red. This type 1A pattern is otherwise most predominant among individuals from the Pyrenees in Europe. The type 1B pattern differs from the wide- spread pattern principally in the absence of a black band on the thorax, although it retains an extensive black area on the gaster. It is predomi- nant in the Hindu Kush and occurs elsewhere in the Caucasus (B. (Melanobombus) lapidarius) and quite independently in South America (B. (Fervidobombus) bellicosus Smith). The type 2A pattern differs from the wide- spread pattern by a change in the pale pubescence from yellow to grey-white on all but tergum II. The type 2A pattern is perhaps most predominant in Kashmir, but is also found around the southern and eastern margins of the Tibetan massif. The type 2B pattern completes the transition of all of the yellow pubescence of the widespread pattern to grey-white. Individuals with this colour pattern occur widely in the Himalaya, the central Asian mountains and westwards in the Caenozoic mountains of southern Europe to as far west as the Alps (B. (Pyrobombus) brodmannicus) and the THE BUMBLE BEES OF THE KASHMIR HIMALAYA Atlas mountains of North Africa (B. lapidarius). Males from populations with the type 2B pattern often retain the common banded yellow pattern (e.g. B. wurflenii and B. lapidarius from the Caucasus, B. brodmannicus from the Alps, B. lapidarius from the Atlas, BMNH). A variant of the type 2B pattern has tergum II or terga I-II dark chocolate-brown. This is seen in the Himalaya only among the males and workers of B. tunicatus (Figs 348, 349, 351, 352, 354) and B. simillimus (Figs 418, 420, 421), but it is also known from the Caucasus mountains for B. (Thoracobombus) mlokosievitzii Radoszkowski (BMNH, PW). The type 3 pattern differs from the widespread pattern by the replacement of the pale pubescence on the thoracic dorsum by black and by the reduction of the black band on the gaster. In the western Himalaya individuals with this pattern are restricted to the outer foothills. Further east in Uttar Pradesh the same pattern is shared by B. (Alpigenobombus) breviceps and B. (Pyrobom- bus) rotundiceps. So what is responsible for the apparent con- straints on the range of bumble bee colour patterns that are actually observed? Why do these colour patterns appear to have converged es- pecially closely among many species within each geographical region? These species are often not closely related, as shown by differences in their male genitalia. Colour pattern had no effect on acceptance of females by patrolling males of British B. pratorum (L.) (Free, 1971). Stiles (1979) suggested that colour affects radiative properties that may be crucial to thermoregula- tion in extreme climates. Sakagami & Ito (1981) challenged this interpretation of sexual dimorph- ism in colour pattern, because the two faunas compared by Stiles were dominated by different groups of species. Within each of these subgenera, Fervidobombus and Pyrobombus, no regional trends were found. However, there may be differ- ences in the kind of habitat and in mate-searching behaviour between the species of these two groups that could also affect the relative advan- tages of the two kinds of colour patterns. The peculiar behaviour of males of some species of the subgenus Fervidobombus has been described by Lloyd (1981) and by Villalobos & Shelly (1987). It could be that many of these differences between the two groups, including differences in colour pattern, are each related to an underlying differ- ence in physiology that is linked to their regional patterns of distribution more directly. Many species of bumble bees commonly pro- duce individuals that are entirely black, but these are most often predominant in tropical environ- ments (Franklin, 1913). These species include B. 119 (Rufipedibombus) rufipes, B. (Megabombus) melanopoda Cockerell and workers and males of B. (Senexibombus) senex Vollenhoven in the East Indies (BMNH), and B. (Fervidobombus) brevivillus Franklin, B. (Fervidobombus) niger Franklin [= B. atratus Franklin, preoccupied name], B. (Fervidobombus) pullatus Franklin and B. (Fervidobombus) morio (Swederus) in South America (BMNH, PW). An extensively black body is likely to have a lower reflectance even in the infra-red part of the spectrum and so should have a higher rate of approach to thermal equi- librium with the environment by an exchange of long-wave radiation. This may be an advantage to actively foraging bumble bees in a warm environ- ment, because they produce a lot of excess heat from the flight muscles and so might otherwise suffer from heat stress. The subtropical black bees listed above may often be forest-dwelling. In the shade they can escape the additional problem of rapid overheating that could arise from an equally efficient absorption by black pigments of the short-wave radiation in direct sunlight. So this likely advantage of the black pubescence for bees foraging in a shaded warm environment may account for why the type 3 pattern, with a black thorax, the source of the endogenous heat, pre- dominates among bumble bees at the lowest altitudes at which they occur in Kashmir (Fig. 12). It is very noticeable that near their lower altitudi- nal and latitudinal limits in Kashmir these bees (B. trifasciatus and B. haemorrhoidalis) forage par- ticularly both in the shade and during the cooler periods of the day (see the discussion of the distribution of the fauna with altitude). However, further south in their distributions in South East Asia, these same species have extensively orange- brown colour patterns of the thoracic pubescence (Fig. 11). More information is needed on the foraging behaviour of these bees in South East Asia, where there is a tendency among the few large nocturnal bees and wasps to have an orange- brown colour pattern (M. C. Day, pers. comm.; Roubik, 1989). There is also clearly a need for measurements of the radiative properties of bumble bee colour patterns in general. Male bumble bees that patrol in the higher forest on Mt Apharwat tend to be almost uni- formly bright yellow (Fig. 6 and see the comments on B. pyrosoma, B. lucorum and B. hypnorum). This may also be true of those males that patrol in forests, but not in open habitats, at higher latitudes (see Pekkarinen, 1979: fig. 7, cf. Loken, 1973: fig.2). In contrast to black, a pale, highly reflective body colour should reduce the rate of approach to thermal equilibrium with the environ- ment by an exchange of long-wave radiation, so 120 that following an active warm-up, a sufficiently high excess thoracic temperature for flight can be reached quickly and maintained in a shaded cool environment. In contrast, males of those species that patrol in the forests at low altitudes, in open areas outside the forest or in a broad range of habitats in Kashmir tend to be darker and share the colour pattern of the females (Fig. 6). Uni- form bright yellow may be disadvantageous in open areas beyond the forest, for instance in alpine and arctic zones, because it is very con- spicuous to predators. This could be a particular problem for the more sedentary ‘perch and sprint’ males (e.g. B. asiaticus and B. rufofasciatus, see the introduction on mate-searching behaviour of male bumble bees in Kashmir) and those males that aggregate at nest entrances (which includes several species of Fervidobombus, e.g. review by Schremmer, 1972; Lloyd, 1981; Villalobos & Shelly, 1987). Yet the most uniformly light- coloured female bumble bees in Kashmir, which belong to the type 1A pattern-group, are found at the highest altitudes sampled for bumble bees, in Ladakh. This trend towards lighter colours in cooler environments is the reverse of that des- cribed by Pekkarinen (1979) from European species. Regional convergences in the finer details of the colour pattern between bumble bee species are likely to be the result of selective predation. Predators with perception of colour and great visual acuity, perhaps certain birds, may avoid the familiar common colour patterns of these insects because the females can inflict a painful sting, and preferentially attack any individuals with unusual colour patterns. This kind of selection could bring about Millerian mimicry among the colour pat- terns on the dorsa of the females (Plowright & Owen, 1980) and Batesian automimicry by the colour patterns on the dorsa of the stingless males (Stiles, 1979). The putative models for this process would be the colour patterns of the females of the most abundant species. In Kashmir, these species tend to be regionally restricted and relatively invariant in their colour patterns (type 1A, B. oberti, Figs 392, 393; type 1B, B. semeno- vianus, Figs 397, 398, 400, 401; type 2A, B. rufofasciatus, Figs 431, 432, 434, 435, 438; type 2B, B. tunicatus, Figs 350, 353; type 3, B. haemorrhoidalis, Figs 279, 280; see Fig. 12). Many of the mimics occupy the middle classes of abun- dance, but are often more widespread among the regions of Kashmir. These species covary in their colour patterns among the regions so that they show the protective livery of the local model (e.g. B. avinoviellus, Figs 232-241; B. himalayanus, Figs 242-253; B. kashmirensis, Figs 295-310; B. PAULH. WILLIAMS lepidus, Figs 327, 328, 330, 331, 333, 334; B. biroi, Figs 335-346; B. asiaticus, Figs 371-391). A subtle difference between the colour patterns of B. kashmirensis and B. rufofasciatus could be interpreted as further circumstantial evidence in support of the mimicry hypothesis for colour convergence. Workers of these two species ap- pear to be identical (type 2A colour pattern) when they are seen manoeuvring slowly from flower to flower in the alpine and subalpine zones of the mountains around the Vale of Kashmir. Most workers of B. rufofasciatus have the hairs of gastral terga III-[V almost uniformly red and those of tergum V uniformly white (Fig. 435). The B. kashmirensis from the same region often have all three terga (III-V) covered with hairs that are red at the base, but which become more exten- sively white-tipped towards the apex of the gaster (Fig. 303). This achieves the same white-tailed effect as for B. rufofasciatus, because the apex of the gaster curves slightly towards the ventral for bumble bees. Colour variation within many of the bumble bee species of Kashmir does appear to be much greater than within the bumble bee species of Britain, or even within the more variable species of the Alps. This may be blamed on taxonomic ‘lumping’ in this review. But if bumble bee colour patterns have converged into separate and distinc- tive mimetic groups within each of the neighbour- ing regions of Asia, then the position of Kashmir at the frontier between the large and divergent Oriental and Palaearctic faunas (which are still to some extent separated even inside Kashmir by the divide of the Great Himalaya) may account in part for the great diversity of colour pattern within some of the Kashmir species. Acknowledgements. I am very ‘prateful to all of those who gave advice, with special thanks to B. Bolton and M. C. Day. D. Agosti, N. H. Barton, R. L. Blackman, S. A. Corbet, A. O. Divers, W. G. Eberhard, P. J. Eggleton, K. Gaston, C. J. Humphries, L. S. Kimsey, R. P. Macfarlane, A. Pekkarinen, R. C. Plowright, P. Rasmont, M. J. Scoble and R. I. Wane-Wright commented on parts of the manuscript. C. R. Betts, G. R. Else, N. D. M. Fergusson, L. Ficken, M. G. Fitton, P. Gilbert, P. M. Hammond, J. M. Harvey, B. R. Pitkin, A. Polaszek, J. Spratt, N. D. Springate, D. A. Sutton, B. C. Townsend, C. R. Vardy, S. L. Ward and D. B. Williams helped with equipment, library searches and translations. M. A. Mir (No. 1 Guide, Gulmarg) provided support in the field and C. A. Chadwell helped with certain plant determina- tions. I am very grateful to the Keeper of Entomology for providing facilities for this work. THE BUMBLE BEES OF THE KASHMIR HIMALAYA I am also much indebted to the following who helped by lending material: A. V. Antropov (ZMMU), K. Arakaki (BPBM), E. Diller (ZS), M. Fischer (NM), W. Hogenes (ITZ), F. Koch and I. Wegener (MNHU), J.-P. Kopelke (NMS), J. Macek (NMP), K. C. McGiffen (INHS), R. J. McGinley and B. B. Norden (USNM), C. D. Michener (SEMK), J. Papp (T,), B. Petersen (ZM), V. Raineri (MCSN), P. Rasmont (FA), B. Tkalct, S.-f. Wang (IZ), J. C. Weulersse (MNHN) and especially Y. A. Pesenko and M. V. Podbolotskaya (ZI). N. R. Bogatyrov, W. Budenburg, C. A. Chadwell, C. S. Cockell, R. Desmier de Chenon, A. O. Divers, R. W. Husband, A. Hutchings, M. Ito, L. W. Macior, S. Martin, H. Osmaston, B. Petersen, N. Saville, G. S. Vick and S.-f. Wang kindly donated or exchanged other Asian specimens used here. The work was carried out during a Postdoctoral Research Fellowship funded by the Natural Environment Research Council, which also awarded a grant towards publication costs. Addi- tional assistance for the purchase of a microscope was provided from the Entomological Fund, and assistance for the purchase of computer hard- ware for the automated mapping system was pro- vided by the Systematics Association. REFERENCES Abrol, D. P. 1987. Influence of thermal and energetic con- straints on the pollination activity of carpenter bee Xylocopa pubescens. Environment and Ecology 5: 90-93. Ackery, P. R. & Vane-Wright, R. I. 1984. Milkweed butterflies, their cladistics and biology, being an account of the natural history of the Danainae, a subfamily of the Lepidoptera, Nymphalidae. ix+425 pp. London. Alcock, J. & Alcock, J. P. 1983. Male behaviour in two bumblebees, Bombus nevadensis auricomus and B. grisei- collis [sic] (Hymenoptera: Apidae). Journal of Zoology 200: 561-570. Alford, D. V. 1975. Bumblebees. xii+352 pp. London. Awram, W. J. 1970. Flight route behaviour of bumble bees. 173 pp. Ph.D. thesis, University of London. Ball, F. J. 1914. Les bourdons de la Belgique. Annales de la Société entomologique de Belgique 58: 77-108. Bally, A. W. 1983. Geological and ecological studies of Qinghai-Xizang (Tibet) Plateau. [Review.] American Scientist 71: 189-190. Barton, N. H. 1988. Speciation. Jn Myers, A. A. & Giller, P. S.: Analytical biogeography, an integrated approach to the study of animal and plant distributions, pp. 185-218. London. Bergstrom, G., Kullenberg, B. & Stallberg-Stenhagen, S. 1973. Studies on natural odoriferous compounds. VII. Recogni- tion of two forms of Bombus lucorum L. (Hymenoptera, Apidae) by analysis of the volatile marking secretion from individual males. Chemica Scripta 4: 174-182. Bergstrém, G., Svensson, B. G., Appelgren, M. & Groth, I. 1981. Complexity of bumble bee marking pheromones: bio- chemical, ecological and systematical interpretations. In Howse, P. E. & Clement, J.-L.: Biosystematics of social 121 insects, pp. 175-183. The Systematics Association, Special Volume No. 19. London. Bingham, C. T. 1897. The fauna of British India, including Ceylon and Burma. Hymenoptera. — Vol. I. Wasps and bees. xxix+579 pp. London. Bischoff, H. 1930. Entomologische Ergebnisse der schwedis- chen Kamtchatka-Expedition 1920-1922. 29. Bombinae. (Hymen.) Arkiv for zoologi 21: 1-6. —— 1931. Entomologische Ergebnisse der deutsch-russischen Alai-Pamir-Expedition 1928 (II). 4. Hymenoptera VII. Zur Kenntnis einiger Hummelnester aus dem Pamir. Mitteilungen aus dem Zoologischen Museum in Berlin 16: 860-864. —— 1935. Genus Bombus (Apidae, Hymen.). Jn: Wissen- schlaftliche Ergebnisse der niederlandischer Expedition in den Karakorum und die angrenzenden Gebiete in den Jahren 1922, 1925 und 1929/30 1: 255-256. Leipzig. — 1936. Schwedisch-chinesische wissenschaftliche Expedition nach den nordwestlichen Provinzen Chinas, unter Leitung von Dr. Sven Hedin und Prof. Si Ping-chang. Insekten gesammelt vom schwedischen Arzt der Expedition Dr. David Hummel 1927-1930. 56. Hymenoptera. 10. Bombinae. Arkiv for zoologi 27: 1-27. Bischoff, H. & Hedicke, H. 1931. Uber einige von TIlliger beschriebene Apiden (Hym.). Sitzungsberichte der Gesell- schaft naturforschender Freunde zu Berlin 1930: 385-392. Bliithgen, P. 1918. Psithyrus vestalis Geoffr. und bohemicus Seidl (distinctus Pérez) (Hym.). Entomologische Mitteilungen 7: 188-197. Boulangé, H. 1924. Recherches sur l'appareil copulateur des hyménopteéres et spécialement des chalastogastres. Mémoires et Travaux des facultés Catholiques de Lille 28: 444 pp. Bowers, M. A. 1985. Bumble bee colonization, extinction, and reproduction in subalpine meadows in northeastern Utah. Ecology 66: 914-927. Chang, D. H. S. 1981. The vegetation zonation of the Tibetan Plateau. Mountain Research and Development 1: 29-48. Chiu, S. C. 1948. Revisional notes on the Formosan bombid- fauna (Hymenoptera). Notes d’entomologie chinoise 12: 57-81. Cockerell, T. D. A. 1905. Descriptions and records of bees. — I. Annals and Magazine of Natural History (7) 16: 216-225. —— 1906. Descriptions and records of bees. — XII. Annals and Magazine of Natural History (7) 18: 69-75. 1909. Descriptions and records of bees. — XXIII. Annals and Magazine of Natural History (8) 4: 393-404. 1910. Some bees from high altitudes in the Himalaya mountains. Entomologist 43: 238-239. —— 1911. The humble-bees of Formosa. Entomologist 44: 100-102. — 1917. New social bees. Psyche 24: 120-128. 1922. Bees in the collection of the U.S. National Museum. Proceedings of the United States National Museum 60: 20 pp. Cresson, E. T. 1863. List of the North American species of Bombus and Apathus. Proceedings of the Entomological Society of Philadelphia 2: 83-116. — 1864. Descriptions of several new species of North American Apidae. Proceedings of the Entomological Society of Philadelphia 3: 38-43. Dalla Torre, K. W. von 1880. Unsere hummel-(Bombus) Arten. Der Naturhistoriker 2: 40-41. — 1882. Bermerkungen zur Gattung Bombus Latr., II. 3. Zur Synonymie und geographischen Verbreitung der Gattung Bombus Latr. Bericht des Naturwissenschaftlich-medizinischen Vereins in Innsbruck 12: 14-31. — 1890. Hymenopterologische Notizen. Wiener entomolo- gische Zeitung 9: 139. — 1896. Catalogus hymenopterorum hucusque descriptorum systematicus et synonymicus. Volumen X: Apidae (antho- phila). vitit643 pp. Lipsiae. Darwin, C. 1859. On the origin of species by means of natural 122 selection, or the preservation of favoured races in the struggle for life. ix+490 pp. London. Day, M. C. 1979. The species of Hymenoptera described by Linnaeus in the genera Sphex, Chrysis, Vespa, Apis and Mutilla. Biological Journal of the Linnean Society 12: 45-84. Delmas, R. 1981. Systematics and geographical variation in the Bombinae. Jn Howse, P. E. & Clement, J.-L.: Biosystematics of social insects, pp. 223-229. The Systematics Association, Special Volume No. 19. London. Dobzhansky, T. 1937. Genetics and the origin of species. xvit364 pp. New York. Dover, C. 1922. A note on bees of the genera Xylocopa and Bombus in the Indian Museum. Record of the Indian Museum 24: 85-89. Eberhard, W. G. 1985. Sexual selection and animal genitalia. xviit+244 pp. London. Erlandsson, A. 1979. Bombus canariensis Pérez, 1895 n. stat and Bombus maderensis n. sp. from the Macaronesian Islands. Entomologica scandinavica 10: 187-192. Eversmann, E. 1852. Fauna hymenopterologica Volgo-Uralensis. (Continuatio). Familia anthophilarium seu apidarum. /zvéstiya Moskovskago éntomologicheskago obshchestva 3: 3-137. Fabricius, J. C. 1775. Systema entomologiae, sistens insectorvm classes, ordines, genera, species, adiectis synonymis, locis, descriptionibvs, observationibvs. 32+832 pp. Flensbyrgi & Lipsiae. 1781. Species insectorvm exhibentes eorvm differentias specificas, synonyma avctorvm, loca natalia, metamorphosin adietis observationibvs, descriptionibvs. 1: vilit552 pp. Hambvrgi & Kilonii. 1787. Mantissa insectorvm sistens eorvm species nvper detectas adiectis characteribvs genericis, differentiis specificis, emendationibvs, observationibvs. 1: xx+348 pp. Hafniae. — 1793. Entomologia systematica emendata et aucta. Secun- dum classes, ordines, genera, species adjectis synonimus, locis observationibus, descriptionibus. 2: viiit519 pp. Hafniae. — 1804. Systema piezatorum secundum ordines, genera, species adiectis synonymis, locis, observationibus, descrip- tionibus. xiv+439 pp. Brunsvigae. Fisher, R. M. 1987. Queen-worker conflict and social parasitism in bumble bees (Hymenoptera: Apidae). Animal Behaviour 35: 1026-1036. Franklin, H. J. 1911. New North American Bombidae. Trans- actions of the American Entomological Society 37: 157-168. 1913. The Bombidae of the New World. Transactions of the American Entomological Society 38(1912): 177-486. Free, J. B. 1971. Stimuli eliciting mating behaviour of bumble- bee (Bombus pratorum L.) males. Behaviour 40: 55-61. Freeman, R. B. 1968. Charles Darwin on the routes of male humble bees. Bulletin of the British Museum (Natural History) (Historical) 3: 179-189. Fretwell, S. D. & Lucas, H. L. 1970. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta biotheoretica 19: 16-36. Friese, H. 1905. Neue oder wenig bekannte Hummeln des tussischen Reiches (Hymenoptera). Ezhegodnik Zoologi- cheskago muzeya 9(1904): 507-523. 1909. Neue Varietaten von Bombus (Hym.). Deutsche entomologische Zeitschrift 1909: 673-676. 1913. Uber einige neue Apiden (Hym.). Archiv fiir Naturgeschichte 78(1912)(A,12): 85-89. 1916. Uber einige neue Hummelformen (Bombus), besonders aus Asien (Hym.). Deutsche entomologische Zeitschrift 1916: 107-110. —— 1918. Uber Hummelformen aus dem Himalaja. Deutsche entomologische Zeitschrift 1918: 81-86. 1924. Uber auffallende Hummelformen. (Hym. Apid.). Deutsche entomologische Zeitschrift 1924: 437-439. 1931. Uber Bombus und Psithyrus. Konowia 10: 300-304. Friese, H. & Wagner, H. von 1910. Zoologische Studien PAULH. WILLIAMS an Hummeln. I. Die Hummeln der deutschen Fauna. Zoologische Jahrbucher, Systematik, Okologie und Geo- graphie der Tiere 29(1909): 1-104. Frison, T. H. 1927. A contribution to our knowledge of the relationships of the Bremidae of America north of Mexico (Hymenoptera). Transactions of the American Entomological Society 53: 51-78. 1933. Records and descriptions of Bremus and Psithyrus from India (Bremidae: Hymenoptera). Record of the Indian Museum 35: 331-342. — 1934. Records and descriptions of Bremus and Psithyrus from Formosa and the asiatic mainland. Transactions of the Natural History Society of Formosa 24: 150-185. — 1935. Records, notes and descriptions of Bremus from Asia (Bremidae: Hymenoptera). Record of the Indian Museum 37: 339-363. Gerstaecker, A. 1869. Beitrage zur naheren Kenntniss einiger Bienen-Gattungen. Stettiner entomologische Zeitung 30: 315— 367. Ghiselin, M. 1975. A radical solution to the species problem. Systematic Zoology 23(1974): 536-544. Griitte, E. 1937. Zur Kenntnis zentralasiatischer Arten von Psithyrus Lep. (Hym. Apid.). Mitteilungen der Deutschen entomologischen Gesellschaft 7: 103-109. Gurcharan Singh & Kachroo, P. 1976. Forest flora of Srinagar and plants of neighbourhood. x+278 pp. Dehra Dun. Haas, A. 1976. Paarungsverhalten und Nestbau der alpinen Hummelart Bombus mendax (Hymenoptera: Apidae). Ento- mologica germanica 3: 248-259. Handlirsch, A. 1888. Die Hummelsammlung des k. k. natur- historischen Hofmuseums. Annalen des Naturhistorischen Museums in Wien 3: 209-250. Harris, M. 1776. An exposition of English insects, with curious observations and remarks, wherein each insect ts particularly described; its parts and properties considered; the different sexes distinguished, and the natural history faithfully related. vilit 166 pp. London. Heinrich, B. 1979. Bumblebee economics. viiit245 pp. Massachusetts. Hengeveld, R. 1988. Mayr’s ecological species criterion. Syste- matic Zoology 37: 47-SS5. Hennig, W. 1981. Insect phylogeny. [Translated by Pont, A. C.; revisionary notes by Schlee, D.] xxii+514 pp. Chichester. Hewitt, G. M. 1988. Hybrid zones — natural laboratories for evolutionary studies. Trends in Ecology and Evolution 3: 158-167. Honk, C. G. J. van, Velthuis, H. H. W. & Roseler, P.-F. 1978. A sex pheromone from the mandibular glands in bumblebee queens. Experientia 34: 838-839. Hull, D. L. 1980. Individuality and selection. Annual Review of Ecology and Systematics 11: 311-332. Illiger, J. C. W. 1806. William Kirby’s Familien der bienenartigen Insekten, mit Zusatzen, Nachweisungen und Bemerkungen. Magazin fiir Insektenkunde 5: 28-175. Inouye, D. W. 1977. Species structure of bumblebee communi- ties in North America and Europe. Jn Mattson, W. J.: The role of arthropods in forest ecosystems, pp. 35-40. New York. International Commission on Zoological Nomenclature 1939. Opinion 135. The suppression of the so-called ‘Erlangen List’ of 1801. Opinions and Declarations rendered by the Inter- national Commission on Zoological Nomenclature 2: 7-12. 1954. Opinion 220. Validation, under the plenary powers, of the generic name Bombus Latreille, 1802 (class Insecta, order Hymenoptera), in so far as the use of those powers is required to provide that name with the status of availability. Opinions and Declarations rendered by the International Commission on Zoological Nomenclature 4: 103-114. — 1985. International code of zoological nomenclature. 3rd edn. xx+338 pp. Berkeley. Ito, M. 1983. Supraspecific classification of bumblebees based on oo ee THE BUMBLE BEES OF THE KASHMIR HIMALAYA the characters of male genitalia. 247 pp. Ph.D. thesis, Hokkaido University. —— 1985. Supraspecific classification of bumblebees based on the characters of male genitalia. Contributions from the Institute of Low Temperature Science, Hokkaido University (B) 20: 143 pp. — 1987. Geographic variation of an east Asian bumblebee Bombus diversus in some morphometric characters (Hymen- optera, Apidae). Kontyti 55: 188-201. Ito, M., Matsumura, T. & Sakagami, S. F. 1984. A nest of the himalayan bumblebee Bombus (Festivobombus) festivus. Kontyti 52: 537-539. Ito, M. & Sakagami, S. F. 1985. Possible synapomorphies of the parasitic bumblebees (Psithyrus) with some nonparasitic bumblebees (Bombus) (Hymenoptera: Apidae). Sociobiology 10: 105-119. Jonghe, R. de 1982. Copulations interspécifiques en captivité d’espéces du genre Bombus Latreille (sensu stricto) (Hymen- optera, Apidae, Bombinae). Bulletin et Annales de la Société royale entomologique de Belgique 118: 171-175. Jonghe, R. de & Rasmont, P. 1983. Kreuzungsexperiment mit Hummeln des Genus Bombus Latreille sensu stricto (Hymen- optera, Apidae). Phegea 11: 7-10. [Jurine] 1801. Nachricht von einen neuen entomologischen Werke des Hrn. Prof. Jurine in Geneve. /ntelligenzblatt der Literatur-Zeitung No. 21: 160-165. Kim, C.-W. & Ito, M. 1987. On the bumblebees from the Korean Peninsula (Hymenoptera, Bombidae). Entomological Research Bulletin 13: 1-42. Kimsey, L. S. 1984. A re-evaluation of the phylogenetic relationships in the Apidae (Hymenoptera). Systematic Entomology 9: 435-441. Kirby, W. 1802. Monographia apum angliae; or, an attempt to divide into their natural genera and families, such species of the Linnean genus Apis as have been discovered in England: with descriptions and observations. 2: 388 pp. Ipswich. Kirkpatrick, M. 1982. Sexual selection and the evolution of female choice. Evolution 36: 1-12. Kopelke, J.-P. 1982. Funktion der Genitalstrukturen bei Bombus- Arten am Beispiel von B. lapidarius (Linnaeus 1758) und deren Bedeutung fiir die Systematik (Insecta: Hymenoptera: Apidae). Senckenbergiana biologica 62(1981): 267-286. Krauth, J. 1988. Distribution-free statistics: an application- oriented approach. Techniques in the behavioral and neural sciences 2: xiv+381 pp. Oxford. Kriiger, E. 1917. Zur Systematik der mitteleuropaischen Hummeln (Hym.). Entomologische Mitteilungen 6: 55-66. —— 1943. Uber die Genitalanhange einiger Mannchen der Untergattung Pratobombus O. Vogt. Zeitschrift fiir Morph- ologie und Okologie der Tiere 39: 527-545. — 1951. Phanoanalytische Studien an einigen Arten der Untergattung Terrestribombus O. Vogt (Hymen. Bomb.). I. Teil. Tijdschrift voor Entomologie 93(1950): 141-197. — 1954. Phaenoanalytische Studien an einigen Arten der Untergattung Terrestribombus O. Vogt (Hymenoptera, Bombidae). II. Teil. Tijdschrift voor Entomologie 97: 263- 298. —— 1956. Phaenoanalytische Studien an einigen Arten der Untergattung Terrestribombus O. Vogt (Hymenoptera, Bombidae). II. Teil. Tijdschrift voor Entomologie 99: 75-105. — 1958. Phaenoanalytische Studien an einigen Arten der Untergattung Terrestribombus O. Vogt (Hymenoptera, Bombiidae). III. Teil. Tijdschrift voor Entomologie 101: 283— 344. Kruseman, G. 1952. Subgeneric division of the genus Bombus Latr. Transactions of the 9th International Congress of Entomology, Amsterdam 1:101-103. Labougle, J. M. & Ayala, R. 1985. A new subgenus and species of Bombus (Hymenoptera: Apidae) from Guerrero, Mexico. Folia Entomolégica Mexicana 66: 47-55. 123 Latreille, P. A. 1802a. Histoire naturelle des fourmis, et recueil de mémoires et d’observations sur les abeilles, les araignées, les faucheurs, et autres insectes. xvit445 pp. Paris. 18026. Histoire naturelle, générale et particuliére des crustaces et des insectes. 3: xiit+467 pp. Paris. Lepeletier de Saint-Fargeau, A. L. M. 1832. Observations sur louvrage intitulé: “‘bombi scandinaviae monographice trac- tato, etc., a Gustav. Dahlbom.’. Annales de la Société Entomologique de France 1: 366-382. — 1836. Histoire naturelle des insectes. Hyménopteres. 1: 547pp. Linnaeus, C. 1758. Systema naturae per regna tria naturae, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis. 823 pp. Holmiae. 1761. Fauna svecica sistens animalia svecica regni: Mammalia, Aves, Amphibia, Pisces, Insecta, Vermes. Dis- tributa per classes & ordines, genera & species, cum differ- entiis specierum, synonymis auctorum, nominibus incolarum, locis natalium, descriptionibus insectorum. [45]+578 pp. Stockholmiae. Lloyd, J. E. 1981. Sexual selection: individuality, identification, and recognition in a bumblebee and other insects. Florida Entomologist 64: 89-118. Lgken, A. 1961. Observations on Norwegian bumble bee nests (Hymenoptera, Apidae, Bombus). Norsk entomologisk Tidsskrift 11: 255-268. 1966. Notes on Fabrician species of Bombus Latr. and Psithyrus Lep., with designations of lectotypes (Hym., Apidae). Entomologiske Meddelelser 34: 199-206. —— 1973. Studies on Scandinavian bumble bees (Hymen- optera, Apidae). Norsk entomologisk Tidsskrift 20: 1-218. —— 1984. Scandinavian species of the genus Psithyrus Lepeletier (Hymenoptera: Apidae). Entomologica scandinavica suppl. 23: 45 pp. Maa, T. 1948. On some eastern asiatic species of the genus Psithyrus Lepel. (Hymenoptera: Bombidae). Notes d’ento- mologie chinoise 12: 17-37. Mallet, J. 1986. Hybrid zones of Heliconius butterflies in Panama and the stability and movement of warning colour clines. Heredity 56: 191-202. Mani, M. S. 1986. Butterflies of the Himalaya. x+181 pp. Lancaster. Mayr, E. 1963. Animal species and evolution. xiv+797 pp. Massachusetts. Medler, J. T. 1962. Morphometric analyses of bumblebee mouthparts. Transactions of the 11th International Congress of Entomology, Vienna 2: 517-521. Michener, C. D. 1944. Comparative external morphology, phylogeny, and a classification of the bees (Hymenoptera). Bulletin of the American Museum of Natural History 82: 157— 326. Michener, C. D. & Amir, M. 1977. The seasonal cycle and habitat of a tropical bumble bee. Pacific Insects 17: 237-240. Milliron, H. E. 1960. Recognition of bumblebee type specimens, with notes on some dubious names (Hymenoptera: Apidae). Bulletin of the Brooklyn Entomological Society 55: 87-99. 1961. Revised classification of the bumblebees — a synopsis (Hymenoptera: Apidae). Journal of the Kansas Entomological Society 34: 49-61. — 1971. A monograph of the western hemisphere bumble- bees (Hymenoptera: Apidae; Bombinae). I. The genera Bombus and Megabombus subgenus Bombias. Memoirs of the Entomological Society of Canada 82: 80 pp. Morawitz, F. 1875. [Bees.] (Mellifera). In Fedtschenko, A.: Reise in Turkestan. II Zoologischer Teil. 303 pp. Berlin. — 1880. Ein Beitrag zur Bienen-Fauna Mittel-Asiens. /zvéstiya Imperatorskoi akademii nauk 26: 337-379. — 1883. Neue russisch-asiatische Bombus-Arten. Trudy Russkago éntomologicheskago obshchestva 17: 235-245. — 1886. Insecta in itinere cl. N. Przewalskii in Asia centrali 124 novissime lecta. I: Apidae. Trudy Russkago éntomologi- cheskago obshchestva 20: 195-229. — 1890. Insecta a cl. G. N. Potanin in China et in Mongolia novissime lecta. XIV. Hymenoptera Aculeata. III. Apidae. Trudy Russkago éntomologicheskago obshchestva 24: 349- 385. — 1894. Supplement zur Bienenfauna Turkestans. Trudy Russkago éntomologicheskago obshchestva 28: 1-87. Morice, F. D. & Durrant, J. H. 1915. The authorship and first publication of the ‘Jurinean’ genera of Hymenoptera: being a reprint of a long-lost work by Panzer, with a translation into English, an introduction, and bibliographical and critical notes. Transactions of the Entomological Society of London 47(1914): 339-436. Morse, D. H. 1982. Behavior and ecology of bumble bees. In Hermann, H. R.: Social insects, 3: 245-322. London. Nelson, G. 1972. Phylogenetic relationship and classification. Systematic Zoology 21: 227-231. Newman, E. 1835. Attempted division of British insects into natural orders. Entomological Magazine 2(1834): 379-431. Nylander, W. 1848. Adnotationes in expositionem monographi- cam apum borealium. Meddelanden af Societatis pro fauna et flora fennica 1: 165-282. Obrecht, E. & Scholl, A. 1984. Bombus lucorum auct. ein Arten- komplex — Enzymelektrophoretische Befunde (Hymenoptera, Bombidae). Verhandlungen der Deutschen zoologischen Gesellschaft 77: 266. Owen, R. E. 1986. Gene frequency clines at X-linked or haplodiploid loci. Heredity 57: 209-219. Owen, R. E. & Plowright, R. C. 1980. Abdominal pile color dimorphism in the bumble bee, Bombus melanopygus. Journal of Heredity 71: 241-247. Pamilo, P., Pekkarinen, A. & Varvio, S.-L. 1987. Clustering of bumblebee subgenera based on interspecific genetic relation- ships (Hymenoptera, Apidae: Bombus and Psithyrus). Annales zoologici fennici 24: 19-27. Pamilo, P., Varvio-Aho, S.-L. & Pekkarinen, A. 1984. Genetic variation in bumblebees (Bombus, Psithyrus) and putative sibling species of Bombus lucorum. Hereditas 101: 245-251. Panfilov, D. V. 1951. [Bumble bees of the subgenus Cullumano- bombus Vogt (Hymenoptera, Apoidea).] Trudy Vsesoyuznogo entomologicheskogo obshchestva, Akademiya nauk SSSR 43: 115-128. — 1957. [On the geographical distribution of bumble bees (Bombus) in China.] Acta geographica sinica 23: 221-239. — 1981. [Maps 91-97.] In Gorodkov, K. B.: [Provisional atlas of the insects of the European part of the U.S.S.R. Maps 73-125.] Leningrad. — 1982. [Maps 147-150.] In Gorodkov, K. B.: [Provisional atlas of the insects of the European part of the U.S.S.R. Maps 126-178.] Leningrad. —— 1984. [Maps 186-192.] In Gorodkov, K. B.: [Provisional atlas of the insects of the European part of the U.S.S.R. Maps 179-221.] Leningrad. Panzer, G. W. F. [1801]. Favnae insectorvm germanicae initia oder Deutschlands Insecten gesammelt und herausgegeben. 85: 24 pls. Nurnberg. Paterson, H. E. H. 1980. A comment on ‘mate recognition systems’. Evolution 34: 330-331. — 1985. The recognition concept of species. In Vrba, E. S.: Species and speciation, pp. 21-29. Transvaal Museum Monograph No. 4. Pretoria. Pekkarinen, A. 1979. Morphometric, colour and enzyme varia- tion in bumblebees (Hymenoptera, Apidae, Bombus) in Fennoscandia and Denmark. Acta zoologica fennica No. 158: 60 pp. Pekkarinen, A., Teras, I., Viramo, J. & Paatela, J. 1981. Distribution of bumblebees (Hymenoptera, Apidae: Bombus and Psithyrus) in eastern Fennoscandia. Notulae Entomologi- cae 61: 71-89. PAULH. WILLIAMS Pendlebury, H. M. 1923. Four new species of Bombus from the Malay Peninsular. Journal of the Federated Malay States Museums 11: 64-67. Pittioni, B. 1937. Die Hummelfauna des Kalsbachtales in Ost- Tirol. Ein Beitrag zur Okologie und Systematik der Hummeln Mitteleuropas. Festschrift zum 60. Geburtstag von Professor Dr. Embrik Strand 3: 64-115. 1938. Die Hummeln und Schmarotzerhummeln der Balkan-Halbinsel mit besonderer Beriicksichtigung der Fauna Bulgariens. I. Allgemeiner Teil. [zvestiya na Tsarskite priro- donauchni instituti v Sofiya 11: 12-69. — 1939a. Die Hummeln und Schmarotzerhummeln der Balkan-Halbinsel. II. Spezieller Teil. Izvestiya na Tsarskite prirodonauchni instituti v Sofiya 12: 49-115. — 1939b. Neue und wenig bekannte Hummeln der Palaarktis (Hymenopt., Apidae). Konowia 17(1938): 244-263. — 1939c. Tanguticobombus subg. nov. (Hymenopt., Apidae). Zoologischer Anzeiger 126: 201-205. — 1942. Die boreoalpinen Hummeln und Schmarotzerhum- meln (Hymen., Apidae, Bombinae). I. Teil. Izvestiya na Tsarskite prirodonauchni instituti v Sofiya 15: 155-218. — 1943. Die boreoalpinen Hummeln und Schmarotzerhum- meln (Hymen., Apidae, Bombinae). II. Teil. Izvestiya na Tsarskite prirodonauchni instituti v Sofiya 16: 1-77. 1949. Beitrage zur Kenntnis der Bienenfauna SO-Chinas. Die Hummeln und Schmarotzerhummeln der Ausbeute J. Klapperich (1937/38). (Hym., Apoidea, Bombini). Eos 25: 241-284. Plowright, R. C. & Laverty, T. M. 1984. The ecology and sociobiology of bumble bees. Annual Review of Entomology 29: 175-199. Plowright, R. C. & Owen, R. E. 1980. The evolutionary significance of bumble bee color patterns: a mimetic inter- pretation. Evolution 34: 622-637. Plowright, R. C. & Pallett, M. J. 1979. Worker-male conflict and inbreeding in bumble bees (Hymenoptera: Apidae). Canadian Entomologist 111: 289-294. Plowright, R. C. & Stephen, W. P. 1973. A numerical taxonomic analysis of the evolutionary relationships of Bombus and Psithyrus (Apidae: Hymenoptera). Canadian Entomologist 105: 733-743. Podbolotskaya, M. V. in press. [The types of the species of bumble bees described by A. S. Skorikov and F. Morawitz in the collection of the Zoological Institute, Leningrad.] Polunin, O. & Stainton, A. 1984. Flowers of the Himalaya. xxx+580 pp. Oxford. Popov, V. B. 1927a. Zur geographischen Verbreitung von Psithyrus vestalis Fourcr. und P. distinctus Pér. (Hymen- optera, Psithyridae). Russkoe entomologicheskoe Obozrenie 21: 128-132. — 1927b. New forms of the genus Psithyrus Lep. Konowia 6: 267-274. —— 1931. Zur Kenntnis der palaarktischen Schmarotzerhum- meln (Psithyrus Lep.). Eos 7: 131-209. Queiroz, K. de & Donoghue, M. J. 1988. Phylogenetic systema- tics and the species problem. Cladistics 4: 317-338. Quilis-Pérez, M. 1927. Los apidos de Espafia genero Bombus Latr. Anales del Instituto Nacionale, Valencia 15: 1-119. Radoszkowski, O. 1859. Sur quelques hyménoptéres nouveaux ou peu connus de la collection du Musée de l’Académie des Sciences de St. Pétersbourg. Byulleten’ Moskovskogo obshchestva ispytatelei prirody 32: 479-486. — 1862. Sur quelques hyménoptéres nouveaux ou peu connus. Byulleten’ Moskovskogo obshchestva ispytatelei prirody 32: 589-598. —— 1884. Révision des armures copulatrices des males du genre Bombus. Byulletin’ Moskovskogo obshchestva ispytatelei prirody 59: 51-92. 1893. Descriptions d’hyménoptéres nouveaux. Revue d’entomologie 12: 241-245. THE BUMBLE BEES OF THE KASHMIR HIMALAYA Rasmont, P. 1981. Redescription d’une espéce méconnue de bourdon d’Europe: Bombus lucocryptarum Ball, 1914 n. status (Hymenoptera, Apidae, Bombinae). Bulletin et Annales de la Société royale entomologique de Belgique 117: 149-154. — 1983a. Notes taxonomiques sur les bourdons (Hymen- optera, Apidae). Bulletin et Annales de la Société royale entomologique de Belgique 119: 167-170. — 1983b. Catalogue commenté des bourdons de la région ouest-paléarctique (Hymenoptera, Apoidea, Apidae). Notes Fauniques de Gembloux 7: 71 pp. — 1984. Les bourdons du genre Bombus Latreille sensu stricto en Europe occidentale et centrale (Hymenoptera, Apidae). Spixiana 7: 135-160. — 1988. Monographie écologique et zoogéographique des bourdons de France et de Belgique (Hymenoptera, Apidae, Bombinae). 309+1xi pp. Ph.D. thesis, Faculté des Sciences agronomiques de |’Etat, Gembloux. Rasmont, P., Scholl, A., Jonghe, R. de, Obrecht, E. & Adamski, A. 1986. Identité et variabilité des males de bourdons du genre Bombus Latreille sensu stricto en Europe occidentale et centrale (Hymenoptera, Apidae, Bombinae). Revue suisse Zoologie 93: 661-682. Reinig, W. F. 1930. Untersuchungen zur Kenntnis der Hummel- fauna des Pamir-Hochlandes. Zoologische Ergebnisse der deutsch-russischen Alai-Pamir-Expedition der Notgemein- schaft der Deutschen Wissenschaft und der Akademie der Wissenschaften der U.d.S.S.R. Zeitschrift fiir Morphologie und Okologie der Tiere 17: 68-123. — 1934. Entomologische Ergebnisse der deutsch-russischen Alai-Pamir-Expedition, 1928 (III). 7. Hymenoptera VIII (Gen. Bombus Fabr.). Nachtrag. Deutsche entomologische Zeitschrift 1933: 163-174. — 1935. On the variation of Bombus lapidarius L. and its cuckoo, Psithyrus rupestris Fabr., with notes on mimetic similarity. Journal of Genetics 30: 321-356. — 1936. Beitrage zur Kenntnis der Hummelfauna von Mandschukuo (Hym. Apid.). Mitteilungen der Deutschen entomologischen Gesellschaft 7: 2-10. — 1937. Entomologische Sammelergebnisse der deutschen Hindukusch-Expedition 1935 der deutschen Forschungs- gemeinschaft. Hymenoptera. Apidae. Arbeiten tiber morph- ologische und taxonomische Entomologie aus Berlin-Dahlem 4: 190. — 1939. Die Evolutionsmechanismen, erlautert an den Hummeln. Verhandlungen der Deutschen zoologischen Gesellschaft Suppl. 12: 170-206. — 1940. Beitrage zur Kenntnis der Hummelfauna von Afghanistan. (Hym., Apid.). (Ergebnisse der Reise von H. und E. Kotzsch in den Hindukusch im Jahre 1936.) Deutsche entomologische Zeitschrift 1940: 224-235. — 1967. Zur Kenntnis der Hummelfaunen einiger Gebirge West-Kleinasiens (Hym., Apidae). Nachrichtenblatt der Bayerischen Entomologen 16: 81-90. — 1970. Bastardierungszonen und Mischpopulationen bei Hummeln (Bombus) und Schmarotzerhummeln (Psithyrus). (Hymenopt., Apidae). Mitteilungen der Miinchener ento- mologischen Gesellschaft 59: 1-89. —— 1971. Zur Faunistik und Zoogeographie des Vorderen Orients. 3. Beitrag zur Kenntnis der Hummeln und Schmar- otzerhummeln Anatoliens (Hym., Apidae). Veroffentlichun- gen der Zoologischen Staatssammlung 15: 139-165. — 1974. Faunistische und zoogeographische Studien in Kleinasien. 5. Auf Hummelfang im Taurus (Bombus Latr., 1802 et Psithyrus Lep., 1832; Hym., Apidae). Nachrichtenblatt der Bayerischen Entomologen 23: 67-80. 1981. Synopsis der in Europa nachgewiesenen Hummel- und Schmarotzerhummelarten (Hymenoptera, Bombidae). Spixiana 4: 159-164. Reinig, W. F. & Rasmont, P. 1988. Beitrag zur Kenntnis der Bergwaldhummel Alpigenobombus wurfleini [sic] 125 (Radoszkowski, 1859) (Hymenoptera, Apidae, Bombinae). Spixiana 11: 37-67. Rice, W. R. 1987. Speciation via habitat specialization: the evolution of reproductive isolation as a correlated character. In Rosenzweig, M. L.: Habitat selection and evolutionary processes. A symposium. Evolutionary Ecology 1: 301-314. Richards, K. W. 1973. Biology of Bombus polaris Curtis and B. hyperboreus Schonherr at Lake Hazen, Northwest Territories (Hymenoptera: Bombini). Quaestiones entomologicae 9: 115— 157. Richards, O. W. 1927a. Sexual selection and allied problems in the insects. Biological Reviews 2: 298-364. —— 1927b. The specific characters of the British humblebees (Hymenoptera). Transactions of the Entomological Society of London 75: 233-268. — 1928a. Bombus and Volucella in the Himalayas. Ento- mologist’s Monthly Magazine 64: 107-108. — 1928b. On a collection of humble-bees (Hymenoptera, Bombidae) made in Ladakh by Col. R. Meinertzhagen. Annals and Magazine of Natural History (10) 2: 333-336. — 1928c. A revision of the European bees allied to Psithyrus quadricolor, Lepeletier (Hymenoptera, Bombidae). Trans- actions of the Entomological Society of London 76: 345-365. 1929a. On two new species of humble-bees in the collection of the British Museum, constituting a new group of the genus Psithyrus, Lep. (Hymenoptera, Bombidae). Annals and Magazine of Natural History (10) 3: 139-143. —— 1929b. A revision of the humble-bees allied to Bombus orientalis, Smith, with the description of a new subgenus. Annals and Magazine of Natural History (10) 3: 378-386. 1930. The humble-bees captured on the expeditions to Mt. Everest (Hymenoptera, Bombidae). Annals and Magazine of Natural History (10) 5: 633-658. 1931. A new species of Indian humble-bee in the collection of the British Museum (Hymenoptera, Bombidae). Annals and Magazine of Natural History (10) 8: 529-533. 1934, Some new species and varieties of oriental humble- bees (Hym. Bombidae). Stylops 3: 87-90. — 1951. The 3rd Danish expedition to Central Asia. Zoological results 5. Bombidae (Insecta) from Afghanistan. Videnskabelige Meddelelser fra Dansk naturhistorisk Forening i Kjgbenhavn 113: 191-199. —— 1968. The subgeneric divisions of the genus Bombus Latreille (Hymenoptera: Apidae). Bulletin of the British Museum (Natural History) (Entomology) 22: 209-276. Rosen, D. E. 1979. Fishes from the uplands and intermontane basins of Guatemala: revisionary studies and comparative geography. Bulletin of the American Museum of Natural History 162: 267-375. Roubik, D. W. 1989. Ecology and natural history of tropical bees. x+514 pp. Cambridge. Ruttner, F. 1988. Biogeography and taxonomy of honeybees. xii+284 pp. London. Sakagami, S. F. 1972. Bumble bees collected by the California Academy-Lingnan dawn-redwood expedition to central west China, 1948. (Hymenoptera: Apidae). Pan-Pacific Ento- mologist 48: 153-174. —— 1975. Some bumblebees from Korea with remarks on the Japanese fauna (Hymenoptera, Apidae). Annales historico- naturales Musei nationalis hungarici 67: 293-316. — 1976. Specific differences in the bionomic characters of bumblebees. A comparative review. Journal of the Faculty of Science, Hokkaido University (6) 20: 390-447. Sakagami, S. F. & Ishikawa R. 1969. Note préliminaire sur la répartition géographique des bourdons japonais, avec des- criptions et remarques sur quelques formes nouvelles ou peu connues. Journal of the Faculty of Science, Hokkaido University (6) 17: 152-196. Sakagami, S. F. & Ito, M. 1981. Specific and subgeneric varia- tions in tibial corbiculation of male bumblebees (Hymenoptera: 126 Apidae), an apparently functionless character. Entomologica scandinavica Suppl. 15: 365-376. Sakagami, S. F. & Yoshikawa, K. 1961. Bees of Xylocopinae and Apinae collected by the Osaka City University Biological Expedition to Southeast Asia 1957-58, with some biological notes. Nature and Life in Southeast Asia 1: 409-444. Sandhouse, G. A. 1943. The type species of the genera and subgenera of bees. Proceedings of the United States National Museum 92: 519-619. Schmid, F. 1958. Trichoptéres du Pakistan. Tijdschrift voor Entomologie 101: 181-221. Schmiedeknecht, H. L. O. 1882. Apidae europaeae (die Bienen Europa’s) per genera, species et varietates, dispositae atque descriptae. 1(4): pp. 235-314. Berlin. Scholl, A. & Obrecht, E. 1983. Enzymelektrophoretische Untersuchungen zur Artabgrenzung im Bombus lucorum — Komplex (Apidae, Bombini). Apidologie 14: 65-78. Schremmer, F. 1972. Beobachtungen zum Paarungsverhalten der Mannchen von Bombus confusus Schenck. Zeitschrift fiir Tierpsychologie 31: 503-512. Schweinfurth, U. 1957. Die horizontale und vertikale Verbrei- tung der Vegetation im Himalaya. xii+372 pp. Bonn. Searle, M. P., Cooper, D. J. W. & Rex, A. J. 1988. Collision tectonics of the Ladakh-Zanskar Himalaya. Philosophical Transactions of the Royal Society of London (A) 326: 117- 150. Seidl, W. B. 1837. Die in Bohmen vorkommenden Hummelarten. Beitrage zur gesammten Natur-und Heilwissenschaft 2: 65-73. Shapiro, A. M. & Porter, A. H. 1989. The lock-and-key hypothesis: evolutionary and biosystematic interpretation of insect genitalia. Annual Review of Entomology 34: 231-245. Siegel, S. 1956. Nonparametric statistics for the behavioral sciences. xviit+312 pp. London. Skorikoy, A. S. 1910a. [Bombus mendax Gerst. and its varia- tions (Hymenoptera, Bombidae).] Russkoe éntomologiches- koe Obozrénie 9(1909): 328-330. — 1910b. [New forms of bumble bees (Hymenoptera, Bombidae).] Russkoe éntomologicheskoe Obozrénie 9(1909): 409-413. — 1912a. Bombus lapponicus (F.) et ses formes ([queens] et [workers]) (Hymenoptera, Bombidae). Russkoe éntomologi- cheskoe Obozrénie 12: 95-102. — 1912b. Neue Hummelformen (Hymenoptera, Bombidae). IV. Russkoe éntomologicheskoe Obozrénie 12: 606-610. — 1914a. Les formes nouvelles des bourdons (Hymenoptera, Bombidae). VI. Russkoe éntomologicheskoe Obozrénie 14: 119-129. — 1914b. Subterraneobombus fedtschenkoi (F. Mor.), un bourdon de Turkestan peu connu (Hymenoptera, Bombidae). Russkoe éntomologicheskoe Obozrénie 14: 287-292. —— 1914c. Pratobombus leucopygos [sic] (F. Mor.) et ses variations (Hymenoptera, Bombidae). Russkoe éntomologi- cheskoe Obozrénie 14: 293-294. —— 1914d. Contribution a la faune des bourdons de la partie méridionale de la province Maritime. Russkoe éntomologi- cheskoe Obozrénie 14: 398-407. — 1922. [Palaearctic bumble bees. Part I. General biology (including zoogeography).] Izvestiya Severnoi oblastnoi stantsii zashchity rastenii ot vreditelei 4: 1-160. 1931. Die Hummelfauna Turkestans und ihre Beziehungen zur zentralasiatischen Fauna (Hymenoptera, Bombidae). Abhandlungen der Pamir-Expedition 1928 8: 175-247. —— 1933a. Zur Hummelfauna Japans und seiner Nachbar- lander. Mushi 6: 53-65 —— 1933b. Zur Fauna und Zoogeographie der Hummeln des Himalaya. Doklady Akademii nauk SSSR 5: 243- 248. —— 1938a. Zoogeographische Gesetzmassigkeiten der Hummel- fauna im Kaukasus, Iran und Anatolien (Hymenoptera, Bombinae). Entomologicheskoe Obozrenie 27: 145-151. PAULH. WILLIAMS —— 1938b. Vorlaufige Mitteilung uber die Hummelfauna Burmas. Arkiv for zoologi 30B: 1-3. Smith, E. L. 1970. Evolutionary morphology of the external insect genitalia. 2, Hymenoptera. Annals of the Entomological Society of America 63: 1-27. Smith, F. 1852a. Descriptions of some new and apparently undescribed species of hymenopterous insects from north China, collected by Robert Fortune, Esq. Transactions of the Entomological Society of London 2: 33-4S. — 1852b. Descriptions of some hymenopterous insects from northern India. Transactions of the Entomological Society of London 2: 45-48. — 1854. Catalogue of hymenopterous insects in the collection of the British Museum. Part II. Apidae. 266(199-465) pp. London. —— 1878a. Scientific results of the Second Yarkand Mission. Hymenoptera. 22 pp. Calcutta. — 1878b. List of the Hymenoptera obtained by Mr Ossian Limborg east of Maulmain, Tenasserim Provinces, during the months of December 1876, January, March and April 1877, with descriptions of new species. Journal of the Asiatic Society of Bengal 47: 167-169. — 1879. Descriptions of new species of Hymenoptera in the collection of the British Museum. xxi+240 pp. London. Snodgrass, R. E. 1941. The male genitalia of Hymenoptera. Smithsonian Miscellaneous Collections 99(14): 86 pp. Splitter, L. J. 1982. Natural kinds and biological species. 289 pp. D.Phil. thesis, University of Oxford. Stiles, E. W. 1979. Evolution of color pattern and pubescence characteristics in male bumblebees: automimicry vs. thermo- regulation. Evolution 33: 941-957. Svensson, B. G. 1979. Patrolling behaviour of bumble bee males (Hymenoptera, Apidae) in a subalpine/alpine area, Swedish Lapland. Zoon 7:67-94. Templeton, A. R. 1981. Mechanisms of speciation—a population genetic approach. Annual Review of Ecology and Systematics 12: 23-48. —— 1989. The meaning of species and speciation: a genetic perspective. Jn Otte, D. & Endler, J. A.: Speciation and its consequences, pp. 3-27. Sunderland. The Times Atlas of the World 1987. Comprehensive edn, 7th edn. xlvii+123 pls+228 pp. London. Thomson, C. G. 1872. Hymenoptera scandinaviae. (Apis Lin.). 2: 286 pp. Lund. Thornhill, R. & Alcock, J. 1983. The evolution of insect mating systems. 1x+547 pp. London. Thorp, R. W. 1969. The identity of Bombus vandykei (Hymen- optera: Apidae). Pan-Pacific Entomologist 45: 87-96. Thorp, R. W., Horning, D. S. & Dunning, L. L. 1983. Bumble bees and cuckoo bumble bees of California (Hymenoptera: Apidae). Bulletin of the California Insect Survey No. 23: Vili+79 pp. Tkalcu, B. 1959. Eine neue Art der Gattung Psithyrus Lep. aus Korea (Hymenoptera, Bombinae). Casopis Ceskoslovenské spolecnosti entomologické 56: 251-254. —— 1960. Remerques sur quelques espéces de bourdons de Chine (Hymenoptera, Bombinae). Bulletin de la Société entomologique de Mulhouse 1960: 66-71. —— 1961. Zur Hummelfauna der Umgebung Kuku-Nors (Hymenoptera, Bombinae). Casopis Ceskoslovenské spolec- nosti entomologické 58: 344-379. — 1962. Contribution a l’étude des bourdons du Japon (1) (Hymenoptera, Apoidea). Bulletin de la Société entomolo- gique de Mulhouse 1962: 81-100. 1965. Contribution a |’étude des bourdons du Japon (II) (Hymenoptera, Apoidea). Bulletin de la Société entomolo- gique de Mulhouse 1965: 1-14. — 1967. Sur deux espéces de bourdons décrites par William Nylander (Hymenoptera, Apoidea: Bombus). Bulletin de la Société entomologique de Mulhouse 1967: 41-58. THE BUMBLE BEES OF THE KASHMIR HIMALAYA — 1968a. Neue Arten der Unterfamilie Bombinae der palaarktischen Region (Hymenoptera, Apoidea). Sbornik Entomologického oddeleni Narodniho musea v Praze 65: 21-S1. — 1968b. Revision der vier sympatrischen, homochrome geographische Rassen bildenden Hummelarten SO-Asiens (Hymenoptera, Apoidea, Bombinae). Annotationes Zoologi- cae et Botanicae 52: 1-31. — 1969a. Beitrage zur Kenntnis der Fauna Afghanistans (Sammelergebnisse von O. Jakes 1963-64, D. Povolny 1965, D. Povolny & Fr. Tenora 1966, J. Simek 1965-66, D. Povolny, J. Geisler, Z. Sebek & Fr. Tenora 1967). Bombinae, Apoidea, Hym. Casopis Moravského musea v Brne 53(1968): 189-210. — 1969b. Ergebnisse der Albanien-Expedition 1961 des Deutschen Entomologischen Institutes. 78. Beitrag. Hymen- optera: Apidae IV (Bombinae). Beitrage zur Entomologie 19: 887-916. — 1972. Arguments contre l’interprétation traditionelle de la phylogénie des abeilles (Hymenoptera, Apoidea). Premiére partie, introduction et exposés fondamentaux. Bulletin de la Société entomologique de Mulhouse 1972: 17-28. — 1973. Taxonomie von Pyrobombus brodmannicus (Vogt) (Hymenoptera, Apoidea, Bombinae). Acta entomologica bohemoslovaca 70: 259-268. — 1974a. Ergebnisse der 1. und 2. mongolisch-tschechos- lowakischen entomologisch-botanischen Expedition in der Mongolei. Nr. 29: Hymenoptera, Apoidea, Bombinae. Sbornik faunistickych praci Entomologického oddeleni Narodntho musea v Praze 15: 25-S7. — 1974b. Eine Hummel-Ausbeute aus dem Nepal-Himalaya (Insecta, Hymenoptera, Apoidea, Bombinae). Senckenber- giana biologica 55: 311-349. 1975. Beitrag zur Kenntnis der Hummelfauna der franz6- sischen Basses-Alpes (Hymenoptera, Apoidea, Bombinae). Sbornik Slovenského Narodného Muzea 20(1974): 167-186. — 1987. Nouveaux synonymes chez les Bombinae (Hymen- optera, Apoidea). Bulletin de la Société entomologique de Mulhouse 1987: 59-64. —— 1989. Neue Taxa asiatischer Hummeln (Hymenoptera, Apoidea). Acta entomologica bohemoslovaca 86: 39-60. Troll, C. 1972. The three-dimensional zonation of the himalayan system. Jn Troll, C.: Geoecology of the high-mountain regions of Eurasia, pp. 264-275. Proceedings of the Symposium of the International Geographical Union Commission on High- Altitude Geoecology November, 20-22, 1969 at Mainz, Band 4. Wiesbaden. Verrell, P. A. 1988. Stabilizing selection, sexual selection and speciation: a view of specific-mate recognition systems. Systematic Zoology 37: 209-215. Villalobos, E. M. & Shelly, T. E. 1987. Observations on the behavior of male Bombus sonorus (Hymenoptera: Apidae). Journal of the Kansas Entomological Society 60: 541-548. Vogt, O. 1909. Studien tiber das Artproblem. 1. Mitteilung. Uber das Variieren der Hummeln. 1. Teil. Sitzungsberichte 127 der Gesellschaft naturforschender Freunde zu Berlin 1909: 28-84. 1911. Studien iiber das Artproblem. 2. Mitteilung. Uber das Variieren der Hummeln. 2. Teil. (Schluss). Sitzungs- berichte der Gesellschaft naturforschender Freunde zu Berlin 1911: 31-74. Vrba, E. S. 1985. Introductory comments on species and speciation. /n Vrba, E. S.: Species and speciation, pp. 1x—xvili. Transvaal Museum Monograph No. 4. Pretoria. Walckenaer, C. A. 1802. Faune Parisienne, insectes. Ou histoire abrégée des insectes des environs de Paris, classés d’apres le systéme de Fabricius; précédée d’un discours sur les insectes en général, pour servir d’introduction a l'étude de l’entomologie. 2: xxiit+438 pp. Paris. Wang, S.-f. 1979. Three new species of bumble bees from Tibet. Acta entomologica sinica 22: 188-191. —— 1982. Hymenoptera: Apidae — Bombus. Insects of Xizang 2: 427-447. — 1987. Bombus. In Huang, F.-s.: Forest insects of Yunnan, pp. 1378-1381. Yunnan. —— 1988. Hymenoptera: Apidae — genus Bombus. In Huang, F.-s.: Insects of Mt. Namjagbarwa region of Xizang, pp. 553-557. Beijing. West-Eberhard, M. J. 1983. Sexual selection, social com- petition, and speciation. Quarterly Review of Biology 58: 155-183. Wille, A. & Michener, C. D. 1973. The nest architecture of stingless bees with special reference to those of Costa Rica (Hymenoptera, Apidae). Revista de biologia tropical 21 suppl. 1: 278 pp. Williams, P. H. 1981. On the evolution and ecology of bumble- bees (Hymenoptera: Apidae). 60 pp. Undergraduate disserta- tion, University of Cambridge. —— 1985. A preliminary cladistic investigation of relationships among the bumble bees (Hymenoptera, Apidae). Systematic Entomology 10: 239-255. —— 1988. Habitat use by bumble bees (Bombus spp.). Ecologi- cal Entomology 13: 223-237. 1989. Why are there so many species of bumble bees at Dungeness? Botanical Journal of the Linnean Society 101: 31-4. Willmer, P. G. 1983. Thermal constraints on activity patterns in nectar-feeding insects. Ecological Entomology 8: 455- 469. Wilson, E. O. & Brown, W. L. 1953. The subspecies concept and its taxonomic application. Systematic Zoology 2: 97-111. Winston, M. L. & Michener, C. D. 1977. Dual origin of highly social behavior among bees (Hymenoptera/Apidae/phylogeny). Proceedings of the National Academy of Sciences of the United States of America 74: 1135-1137. Yasumatsu, K. 1951. Bombus and Psithyrus of Shansi, N. China (Hymenoptera, Apidae). Mushi 22: 59-62. Zander, E. 1900. Beitrage zur Morphologie der mannlichen Geschlechtsanhange der Hymenopteren. Zeitschrift fiir wis- senschaftliche Zoologie 67: 461-489. Plate 1 Representatives of the rufofasciatus-group from the Vale of Kashmir. In this part of their distribution, all three species show caste-dependent colour dimorphism among females. There is also gender-dependent colour variation among Bombus pyrosoma and B. simillimus. Top row: B. rufofasciatus (left to right: queen, worker, male), from a nest on Mt Apharwat, in subalpine pasture. Middle row: B. Pyrosoma (queen, worker, male), from a nest on Apharwat, in the coniferous forest. Bottom row: B. simillimus (queen, worker, male), from a nest near Harwan, in valley scrub. Photographed by Harry Taylor, BMNH photographic unit. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 129 ocellus ocello-ocular area compound eye antenna clypeus oculo-malar distance malar area mandibular breadth labrum 15 mandible se 2 ocellar breadth ocello-ocular distance Figs 14-22 Morphology of the head: (14) anterior aspect of the head of a banded white queen B. avinoviellus, Gulmarg; (15) left lateral aspect of a worker B. Jemniscatus, Apharwat, showing measurements of the malar area; (16) dorsal aspect of a banded white queen B. avinoviellus, Gulmarg, showing measurements of the ocello-ocular area; dorsal aspect of segments 3—5 of the left antenna of queens of (17) B. trifasciatus, Gulmarg; (18) B. haemorrhoidalis, Muzaffarabad; (19) banded white B. asiaticus, Gulmarg; (20) B. oberti, Nimaling; the same, males of (21) B. bohemicus, Lal Pani; (22) B. ferganicus, Gulmarg. 130 basal depression transverse ridge apical depression PAULH. WILLIAMS median furrow lamella lateral tubercle 23 24 25 a) 26 29 30 | | / { | 31 Ki 33 34 basal tooth | apical tooth 7 Yi basal keel sulcus obliquus incisura novus intercalary tooth avinoviellus kashmirensis rufofasciatus Figs 23-38 Morphology of the head: anterior aspect of the labrum of females of (23) banded white B. avinoviellus, Gulmarg; (24) B. novus, Apharwat; (25) yellow and white B. kashmirensis, Apharwat; (26) B. rufofasciatus, Apharwat; anterior aspect of the labrum of males of (27) banded white B. avinoviellus, Gulmarg; (28) B. novus, Apharwat; (29) yellow and white B. kashmirensis, Apharwat; (30) B. rufofasciatus, Apharwat; outer lateral aspect of the left mandible of females of (31) banded white B. avinoviellus, Gulmarg; (32) B. novus, Apharwat; (33) yellow and white B. kashmirensis, Apharwat; (34) B. rufofasciatus, Apharwat; outer lateral aspect of the left mandible of males of (35) banded white B. avinoviellus, Gulmarg; (36) B. novus, Apharwat; (37) yellow and white B. kashmirensis, Apharwat; (38) B. rufofasciatus, Apharwat. THE BUMBLE BEES OFTHE KASHMIR HIMALA YA 131 proximal margin femur i tibia oD S E i=] & corbicular area 50 = I = E Z z 3 S a. i= 2 i= s proximo-posterior distal margin corner basitarsus 43 gonobase gonocoxite spatha penis valve gonostylus volsella 44 Figs 39-45 Morphology of the thorax and gaster: outer lateral aspect of the basitarsus of the left mid leg of workers of (39) B. trifasciatus, Gulmarg; (40) B. haemorrhoidalis, Patnitop; outer lateral aspect of the basitarsus of the left hind leg of queens of (41) unbanded B. subtypicus, Dras; (42) banded yellow and white B. biroi, Apharwat; (43) outer lateral aspect of the left hind leg of a banded white female B. avinoviellus, Gulmarg; (44) dorsal aspect of the genital capsule of a banded white male B. avinoviellus, Gulmarg; (45) ventral aspect of the same (weakly sclerotised areas stippled). 132 PAULH. WILLIAMS 60 Se ee, 51 61 a ee) 62 63 Baa 55 Cre ia 65 Figs 46-65 Morphology of the male genitalia: postero-dorsal aspect of the right penis valve of (46) banded yellow B. avinoviellus, Lal Pani; (47) banded white B. avinoviellus, Gulmarg; (48) yellow and white B. himalayanus, Apharwat; (49) B. marussinus, Batakush; (50) B. bohemicus, Lal Pani; (51) B. novus, Apharwat; (52) B. branickii, Chogdo; (53) B. ferganicus, Gulmarg; (54) yellow B. morawitzianus, Batakush; (55) B. skorikovi, Lal Pani; (56) B. trifasciatus, Harwan; (57) B. haemorrhoidalis, Muzaffarabad; (58) unbanded B. melanurus, Chogdo; (59) unbanded B. melanurus, Gulmarg; (60) B. personatus, Chogdo; (61) yellow B. kashmirensis, Chogdo; (62) yellow and white B. kashmirensis, Apharwat; (63) white B. kashmirensis, Lal Pani; (64) B. hypnorum, Apharwat; (65) banded B. subtypicus, Lal Pani. THE BUMBLE BEES OF THE KASHMIR HIMALA YA 133 (eee l- —_ 7 ; 73 fo y , Figs 66-85 Morphology of the male genitalia: postero-dorsal aspect of the right penis valve of (66) unbanded B. subtypicus, Lal Pani; (67) B. lemniscatus, Uttar Pradesh; (68) B. lepidus [with yellow females], Dalti; (69) B. lepidus [with yellow and white females], Nigagar; (70) unbanded B. biroi, Suru; (71) banded B. biroi, Apharwat; (72) dark B. tunicatus, Shigar; (73) light B. tunicatus, Gulmarg; (74) dark B. lucorum, Gilgit; (75) light B. lucorum, Apharwat; (76) unbanded yellow [red-tailed] B. asiaticus, Leh; (77) banded yellow black-tailed B. asiaticus, Nigagar; (78) banded white [red-tailed] B. asiaticus, Gulmarg; (79) B. oberti, Pamir; (80) B. ladakhensis, Hemis; (81) B. semenovianus, Hemis; (82) B. keriensis [with yellow females], Apharwat; (83) B. simillimus, Banihal; (84) B. pyrosoma, Apharwat; (85) B. rufofasciatus, Apharwat. 75 134 86 87 88 89 95 = a PAULH. WILLIAMS 96 97 99 100 101 102 104 105 Figs 86-105 Morphology of the male genitalia: outer lateral aspect of the right penis valve of (86) banded yellow B. avinoviellus, Lal Pani; (87) banded white B. avinoviellus, Gulmarg; (88) yellow and white B. himalayanus, Apharwat; (89) B. marussinus, Batakush; (90) B. bohemicus, Lal Pani; (91) B. novus, Apharwat; (92) B. branickii, Chogdo; (93) B. ferganicus, Gulmarg; (94) yellow B. morawitzianus, Batakush; (95) B. skorikovi, Lal Pani; (96) B. trifasciatus, Harwan; (97) B. haemorrhoidalis, Muzaffarabad; (98) unbanded B. melanurus, Chogdo; (99) unbanded B. melanurus, Gulmarg; (100) B. personatus, Chogdo; (101) yellow B. kashmirensis, Chogdo; (102) yellow and white B. kashmirensis, Apharwat; (103) white B. kashmirensis, Lal Pani; (104) B. hypnorum, Apharwat; (105) banded B. subtypicus, Lal Pani. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 135 2 eee a 106 —- 116 —— 107 117 a 108 — 118 en a= 112 122 yi 113 123 Figs 106-125 Morphology of the male genitalia: outer lateral aspect of the right penis valve of (106) unbanded B. subtypicus, Lal Pani; (107) B. lemniscatus, Uttar Pradesh; (108) B. lepidu [with yellow females], Dalti; (109) B. lepidus [with yellow and white females], Nigagar; (110) unbanded B. biroi, Suru; (111) banded B. biroi, Apharwat; (112) dark B. tunicatus, Shigar; (113) light B. tunicatus, Gulmarg; (114) dark B. lucorum, Gilgit; (115) light B. lucorum, Apharwat; (116) unbanded yellow [red-tailed] B. asiaticus, Leh; (117) banded yellow black-tailed B. asiaticus, Nigagar; (118) banded white [red-tailed] B. asiaticus, Gulmarg; (119) B. oberti, Pamir; (120) B. ladakhensis, Hemis; (121) B. semenovianus, Hemis; (122) B. keriensis [with yellow females], Apharwat; (123) B. simillimus, Banihal; (124) B. pyrosoma, Apharwat; (125) B. rufofasciatus, Apharwat. 136 PAULH. WILLIAMS 126 Gee ae 128 cay 138 131 (5 132 142 133 143 134 144 135 145 Figs 126-145 Morphology of the male genitalia: interio-ventral aspect of the right volsella of (126) banded yellow B. avinoviellus, Lal Pani; (127) banded white B. avinoviellus, Gulmarg; (128) yellow and white B. himalayanus, Apharwat; (129) B. marussinus, Batakush; (130) B. bohemicus, Lal Pani; (131) B. novus, Apharwat; (132) B. branickii, Chogdo; (133) B. ferganicus, Gulmarg; (134) yellow B. morawitzianus, Batakush; (135) B. skorikovi, Lal Pani; (136) B. trifasciatus, Harwan; (137) B. haemorrhoidalis, Muzaffarabad; (138) unbanded B. melanurus, Chogdo; (139) unbanded B. melanurus, Gulmarg; (140) B. personatus, Chogdo; (141) yellow B. kashmirensis, Chogdo; (142) yellow and white B. kashmirensis, Apharwat; (143) white B. kashmirensis, Lal Pani; (144) B. hypnorum, Apharwat; (145) banded B. subtypicus, Lal Pani. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 137 156 146 157 147 158 148 159 149 160 3 NS 150 161 v NS 151 162 152 163 153 164 ¢ \ 154 165 J Figs 146-165 Morphology of the male genitalia: interio-ventral aspect of the right volsella of (146) unbanded B. subtypicus, Lal Pani; (147) B. lemniscatus, Uttar Pradesh; (148) B. lepidus [with yellow females], Dalti; (149) B. lepidus [with yellow and white females], Nigagar; (150) unbanded B. biroi, Suru; (151) banded B. biroi, Apharwat; (152) dark B. tunicatus, Shigar; (153) light B. tunicatus, Gulmarg; (154) dark B. lucorum, Gilgit; (155) light B. lucorum, Apharwat; (156) unbanded yellow [red-tailed] B. asiaticus, Leh; (157) banded yellow black-tailed B. asiaticus, Nigagar; (158) banded white [red-tailed] B. asiaticus, Gulmarg; (159) B. oberti, Pamir; (160) B. ladakhensis, Hemis; (161) B. semenovianus, Hemis; (162) B. keriensis [with yellow females], Apharwat; (163) B. simillimus, Banihal; (164) B. pyrosoma, Apharwat; (165) B. rufofasciatus, Apharwat. j= iS) co PAULH. WILLIAMS /\ 166 C 4 176 = 186 (g 196 /\ 167 fA 177 a 187 & 197 Nw ON hn ee 178 (| 169 ae ee 189 (yh 199 179 (ee 170 CY 180 a 190 CS 200 Ou 2 181 EY 191 ee. 201 Nex 182 i) 192 a. 202 172 (A (SS 183 193 a. 203 173 Figs 166-205 Morphology of the male genitalia: dorsal aspect of the right gonostylus of (166) banded yellow B. avinoviellus, Lal Pani; (167) banded white B. avinoviellus, Gulmarg; (168) yellow and white B. himalayanus, Apharwat; (169) B. marussinus, Batakush; (170) B. bohemicus, Lal Pani; (171) B. novus, Apharwat; (172) B. branickii, Chogdo; (173) B. ferganicus, Gulmarg; (174) yellow B. morawitzianus, Batakush; (175) B. skorikovi, Lal Pani; (176) B. trifasciatus, Harwan; (177) B. haemorrhoidalis, Muzaffarabad; (178) unbanded B. melanurus, Chogdo; (179) unbanded B. melanurus, Gulmarg; (180) B. personatus, Chogdo; (181) yellow B. kashmirensis, Chogdo; (182) yellow and white B. kashmirensis, Apharwat; (183) white B. kashmirensis, Lal Pani; (184) B. hypnorum, Apharwat; (185) banded B. subtypicus, Lal Pani; (186) unbanded B. subtypicus, Lal Pani; (187) B. lemniscatus, Uttar Pradesh; (188) B. lepidus [with yellow females], Dalti; (189) B. lepidus [with yellow and white females], Nigagar; (190) unbanded B. biroi, Suru; (191) banded B. biroi, Apharwat; (192) dark B. tunicatus, Shigar; (193) light B. tunicatus, Gulmarg; (194) dark B. lucorum, Gilgit; (195) light B. lucorum, Apharwat; (196) unbanded yellow [red-tailed] B. asiaticus, Leh; (197) banded yellow black-tailed B. asiaticus, Nigagar; (198) banded white [red-tailed] B. asiaticus, Gulmarg; (199) B. oberti, Pamir; (200) B. ladakhensis, Hemis; (201) B. semenovianus, Hemis; (202) B. keriensis [with yellow females], Apharwat; (203) B. simillimus, Banihal; (204) B. pyrosoma, Apharwat; (205) B. rufofasciatus, Apharwat. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 139 = 184 194 fh 204 195 205 | L 174 a 175 X VA 206 207 208 ~ 7 209 210 211 212 213 214 Figs 206-214 Morphology of the gaster: ventral aspect of gastral sternum VI of females of (206) B. bohemicus, Britain; (207) B. novus, Apharwat; (208) B. branickii, Nimaling; (209) B. ferganicus, Gulmarg; (210) B. morawitzianus, Lal Pani; (211) B. skorikovi, Lal Pani; dorsal aspect of the apex of gastral tergum VI of queens of (212) B. lemniscatus, Nepal; (213) yellow and white B. lepidus, Apharwat; (214) banded yellow and white B. biroi, Apharwat. 140 PAULH. WILLIAMS eR nee Figs 215-219 Morphology of the head (scanning electron micrographs, pubescence partially removed): antero-dorsal aspect of the left ocello-ocular area of the head of a queen of (215) B. haemorrhoidalis, Muzaffarabad; dorsal aspect of the left ocello-ocular area of the head of queens of (216) unbanded B. subtypicus, Afghanistan; (217) banded yellow and white B. biroi, ‘Krishnye’ Kashmir; (218) B. lemniscatus, Nepal; (219) yellow and white B. lepidus, Apharwat. THE BUMBLE BEES OF THE KASHMIR HIMALAYA 141 Figs 220-224 Morphology of the head and thorax (scanning electron micrographs): antero-ventral aspect of the labrum of queens of (220) banded yellow and white B. avinoviellus, Gulmarg; (221) yellow and white B. himalayanus, Atro Sar; (222) B. marussinus, Chhantir Gah; postero-dorsal aspect of the middle of the scutum, pubescence removed, of queens of (223) B. lemniscatus, Nepal; (224) yellow and white B. lepidus, Apharwat. PAULH. WILLIAMS Ws Ors \ q y Figs 225-231 Morphology of the legs and gaster (scanning electron micrographs): outer lateral aspect of the corbicular area of the hind tibia of queens of (225) banded white B. avinoviellus, Gulmarg; (226) B. simillimus, Srinagar [the circular structures are attachment cement from the peri-anal glands of uropodid mites, det. D. Macfarlane]; outer lateral aspect of the proximo-posterior corner of the hind basitarsus of queens of (227) unbanded yellow B. asiaticus, Nimaling; (228) B. oberti, Nimaling; (229) B. semenovianus, Baltit; dorsal aspect of gastral tergum VI of queens of (230) B. oberti, Nimaling; (231) B. semenovianus, Baltit. a "T) . \ Bombus (Md. ) avinoviellus ae @ Specimen @ Literature © Unconfirmed Equal area grid Map 5 Summary of the world-wide distribution of B. avinoviellus (for an explanation of the grid see Map 1). ‘Specimen’ records are from material examined by the author and ‘Literature’ records are from published data. ‘Unconfirmed’ records include both dubious identifications in the literature and records without precise locality data. The grid-square that includes Kashmir is shown with a solid outline. Map 6 Distribution of B. avinoviellus in Kashmir (for a comparison with the full range of localities sampled see Map 4). All records are from specimens that have been examined. The different symbols represent the principal colour patterns that predominate among workers in each local sample (key to colour pattern symbols see Figs 232-241). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 145 QUEENS @ WORKERS MALES Figs 232-241 Colour patterns of B. avinoviellus from localities in Kashmir: (232) Banidas 30.vi—2.vii.1959, (233) Shardi 1-13.viii.1953, (234) Lal Pani 2-3.ix.1953, (235) Lal Pani 2-3.ix.1953, (236) Gulmarg vii.1986, (237) Gulmarg viii.1985, (238) Gulmarg viii.1985, (239) Gulmarg vii.1986, (240) Gulmarg viii.1985, (241) Gulmarg viii.1985 (for the colour key see Fig. 258). 146 PAULH. WILLIAMS Bombus (Md. ) himalayanus @ Specimen G@ Literature © Unconfirmed Map 7 Summary of the world-wide distribution of B. himalayanus (for an explanation of the symbols see Map 5). 100km Land above 3000m. Map 8 Distribution of B. himalayanus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 242-253). THE BUMBLE BEES OF THE KASHMIR HIMALA YA 147 QUEENS ) WORKERS LEER me) ae id a wy 253 MALES Figs 242-253 Colour patterns of B. himalayanus from localities in Kashmir: (242) Nimaling 17.vii.1980, (243) Chhantir Gah 5-7. viti.1954, (244) Atro Sar 11.viii.1954, (245) Atro Sar 11.viii.1954, (246) Gumri 10—11.viii.1986, (247) Atro Sar 11 .viii.1954, (248) Holojut 13-14. viii.1954, (249) Apharwat 23.viii.1985, (250) Apharwat 30.viii.1985, (251) Apharwat vili.1985, (252) Apharwat 15.viii.1980, (253) Sangisfaid 1985 (for the colour key see Fig. 258). 148 PAULH. WILLIAMS Bombus (Mad. ) marussinus Che @ Specimen G@ Literature © Unconfirmed Equal area grid Map 9 Summary of the world-wide distribution of B. marussinus (for an explanation of the symbols see Map 5). ——— 1! 0 100km Land above 3000m. Map 10 Distribution of B. marussinus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 254-257). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 149 QUEEN WORKERS MALE Figs 254-257 Colour patterns of B. marussinus from localities in Kashmir: (254) Chhantir Gah 5-7.viii.1954, (255) Holojut 13-14. viii.1954, (256) Batakush 8.ix.1954, (257) Chhantir Gah 5—7.viii.1954 (for the colour key see Fig. 258). 150 PAULH. WILLIAMS Bombus (Ps.) bohemicus @ Specimen @ Literature © Unconfirmed Equal area grid Map 11 Summary of the world-wide distribution of B. bohemicus (for an explanation of the symbols see Map 5). | ee | 0 100km | Land above 3000m. Map 12 Distribution of B. bohemicus in Kashmir (see Map 6). THE BUMBLE BEES OF THE KASHMIR HIMALA YA 151 COLOUR KEY grey-white/cream-white LJ LC lemon yellow/golden yellow light brown chocolate brown orange-red dense mixture of black hairs black Biv MALE Fig. 258 Colour pattern of B. bohemicus from Kashmir: Lal Pani 2-3.ix.1953 152 PAULH. WILLIAMS Bombus (Ps.) novus @ Specimen @ Literature © Unconfirmed Map 13 Summary of the world-wide distribution of B. novus (for an explanation of the symbols see Map 5). | ee | 0 100km Land above 3000m. Map 14 Distribution of B. novus in Kashmir (see Map 6). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 153 CAD |) | | 7 a ae 259 FEMALES & MALES Figs 259-262 Colour patterns of B. novus from localities in Kashmir: (259) Apharwat 25.vii.1986, (260) Apharwat 23.vili.1985, (261) Apharwat 25.vii.1986, (262) Apharwat 23.viii.1985 (for the colour key see Fig. 258). 154 PAULH. WILLIAMS Bombus (Ps.) branickii iS @ Specimen @ Literature © Unconfirmed Equal area grid Map 15 Summary of the world-wide distribution of B. branickii (for an explanation of the symbols see Map 5). 100km Land above 3000m. Map 16 Distribution of B. branickii in Kashmir (see Map 6). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 155 FEMALES MALE Figs 263-266 Colour patterns of B. branickii from localities in Kashmir: (263) Nimaling 19.vii.1980, (264) Chogdo 24.ix.1986, (265) Nimaling 17.vii.1980, (266) Atro Sar 11.viii.1954 (for the colour key see Fig. 258). 156 PAULH. WILLIAMS Bombus (Ps.) ferganicus @ Specimen G@ Literature © Unconfirmed Equal area grid Map 17 Summary of the world-wide distribution of B. ferganicus (for an explanation of the symbols see Map 5). m=" 0 100km Land above 3000m. Map 18 Distribution of B. ferganicus in Kashmir (see Map 6). THE BUMBLE BEES OF THE KASHMIR HIMALAYA US i7/ FEMALE 268 MALE Figs 267, 268 Colour patterns of B. ferganicus from localities in Kashmir: (267) Gulmarg viii.1985, (268) Gulmarg viii.1985 (for the colour key see Fig. 258). 158 PAULH. WILLIAMS Bombus (Ps.) morawitzianus ie @ Specimen @ Literature © Unconfirmed Equal area as Map 19 Summary of the world-wide distribution of B. morawitzianus (for an explanation of the symbols see Map 5). 100km Land above 3000m. Map 20 Distribution of B. morawitzianus in Kashmir (see Map 6). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 159 c CORREA 4 p—“ | a _/ W 269 FEMALE MALE Figs 269, 270 Colour patterns of B. morawitzianus from localities in Kashmir: (269) Lal Pani 2—-3.ix.1953, (270) Lal Pani 2-3.ix.1953 (for the colour key see Fig. 258). 160 PAULH. WILLIAMS Bombus (Ps.) skorikovi ss @ Specimen @ Literature © Unconfirmed Equal area grid Map 21 Summary of the world-wide distribution of B. skorikovi (for an explanation of the symbols see Map 5). | a | 0 100km Land above 3000m. Map 22 Distribution of B. skorikovi in Kashmir (see Map 6). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 161 273 FEMALES ELEN ey , | ay 274 MALES Figs 271-274 Colour patterns of B. skorikovi from localities in Kashmir: (271) Lal Pani 2-3.ix.1953, (272) Lal Pani 2-3.ix.1953, (273) Apharwat viii.1985, (274) Lal Pani 2-3.ix.1953 (for the colour key see Fig. 258). 162 PAULH. WILLIAMS Bombus (Dv.) trifasciatus o @ Specimen @ Literature © Unconfirmed Equal area grid Map 23 Summary of the world-wide distribution of B. trifasciatus (for an explanation of the symbols see Map 5). | ee | 0 100km Land above 3000m. Map 24 Distribution of B. trifasciatus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 275-278). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 163 QUEEN WORKERS MALE Figs 275-278 Colour patterns of B. trifasciatus from localities in Kashmir: (275) Gulmarg 28.vii.1985, (276) Patnitop 6.ix.1986, (277) Harwan 24.viii.1980, (278) Gulmarg 1913 (for the colour key see Fig. 258). 164 PAULH. WILLIAMS Bombus (Or.) haemorrhoidalis @ Literature @ Specimen © Unconfirmed Equal area grid Map 25 Summary of the world-wide distribution of B. haemorrhoidalis (for an explanation of the symbols see Map 5). | ee | 0 100km Land above 3000m. Map 26 Distribution of B. haemorrhoidalis in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 279— 281). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 165 QUEEN WORKER MALE Figs 279-281 Colour patterns of B. haemorrhoidalis from localities in Kashmir: (279) Muzaffarabad 10-12.v.1954, (280) Patnitop 6—8.ix.1986, (281) Muzaffarabad 15.x.1953 (for the colour key see Fig. 258). 166 PAULH. WILLIAMS Bombus (St.) melanurus eo @ Specimen @ Literature © Unconfirmed Equal area grid 100km Land above 3000m. Map 28 Distribution of B. melanurus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 282-290). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 167 QUEENS @ WORKERS MALES Figs 282-290 Colour patterns of B. melanurus from localities in Kashmir: (282) Nimaling 18.vii.1980, (283) Leh 9.vii.1980, (284) Leh 3.vii.1980, (285) Chogdo 24.ix.1986, (286) Chittakatha Sar 19-24.viii.1953, (287) Gulmarg vili. 1986, (288) Gulmarg 13.viii.1980, (289) Gulmarg 16.viii.1980, (290) Gulmarg viii.1985 (for the colour key see Fig. 258). 168 PAULH. WILLIAMS Bombus (St.) personatus es @ Specimen G@ Literature © Unconfirmed Equal area grid Map 29 Summary of the world-wide distribution of B. personatus (for an explanation of the symbols see Map 5). | 0 100km Land above 3000m. Map 30 Distribution of B. personatus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 291-294). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 169 QUEEN WORKER MALES Figs 291-294 Colour patterns of B. personatus from localities in Kashmir: (291) Nimaling 16.vii.1980, (292) Nimaling 21.vii.1980, (293) Chogdo 24.ix.1986, (294) Chogdo 24.ix.1986 (for the colour key see Fig. 258). 170 PAULH. WILLIAMS Bombus (Ag.) kashmirensis aS @ Specimen @ Literature © Unconfirmed Equal area grid Map 31 Summary of the world-wide distribution of B. kashmirensis (for an explanation of the symbols see Map 5). 100km Land above 3000m. Map 32 Distribution of B. kashmirensis in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 295-310). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 171 & Nas fore ® °. Ty wh»n My wy vy 295 we 298 QUEENS WORKERS Pua is ee Bo 4 4 Tn (CuI SSE 307 MALES Figs 295-310 Colour patterns of B. kashmirensis from localities in Kashmir: (295) Nimaling 20.vii.1980, (296) Nimaling 15.vii.1980, (297) Chogdo 24.ix.1986, (298) Nimaling 20.vii.1980, (299) Nimaling 19.vii.1980, (300) Chogdo 24.ix.1986, (301) Gumri 10-11. viii.1986, (302) Apharwat 24.vii.1986, (303) Apharwat viii.1985, (304) Apharwat 30. viii.1986, (305) Apharwat 24.viii.1986, (306) Apharwat viii.1986, (307) Apharwat 23.viii.1985, (308) Nigagar 17.ix.1986, (309) Nigagar 17.ix.1986, (310) Lal Pani 2-3.ix.1953 (for the colour key see Fig. 258). 72} PAULH. WILLIAMS Bombus (Pr.) hypnorum _—S @ Specimen @ Literature © Unconfirmed Equal area grid Map 33 Summary of the world-wide distribution of B. hypnorum (for an explanation of the symbols see Map 5). 100km Land above 3000m. Map 34 Distribution of B. hypnorum in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 311-316). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 173 QUEENS WORKERS fo Eee] Caan, EE /, 313 MALES Figs 311-316 Colour patterns of B. hypnorum from localities in Kashmir: (311) Apharwat 25.vii.1986, (312) Apharwat viii.1986, (313) Apharwat 16.viii.1980, (314) Apharwat 25.vii.1986, (315) Apharwat viii.1985, (316) Apharwat vili. 1985 (for the colour key see Fig. 258). 174 PAULH. WILLIAMS Bombus (Pr.) subtypicus ae @ Specimen G@ Literature © Unconfirmed Equal area grid Map 35 Summary of the world-wide distribution of B. subtypicus (for an explanation of the symbols see Map 5). | 0 100km Land above 3000m. Map 36 Distribution of B. subtypicus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 317-324). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 175 ie Ww —_ v=) QUEENS | Oe | ie. ® Lo) — _ =) 12°) Ne —) WORKERS .- S “ ’ ie hee io) — foe) vs) i) _ MALES Figs 317-324 Colour patterns of B. subtypicus from localities in Kashmir: (317) Lal Pani 2-3.ix.1953, (318) Lal Pani 2-3.ix.1953, (319) Dras 10-11 .viii.1986, (320) Dras 10-11. viii. 1986, (321) Lal Pani 2—-3.ix.1953, (322) Astor 29.v.1954, (323) Dras 10-11.viii.1986, (324) Rawat 20-21.viii.1954 (for the colour key see Fig. 258). 176 PAULH. WILLIAMS Bombus (Pr.) lemniscatus eS @ Specimen @ Literature © Unconfirmed Equal area grid Map 37 Summary of the world-wide distribution of B. lemniscatus (for an explanation of the symbols see Map S). 0 100km Land above 3000m. Map 38 Distribution of B. lemniscatus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 325 & 326). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 177 WORKERS Figs 325, 326 Colour patterns of B. lemniscatus from Kashmir: (325) Apharwat viii-ix.1986, (326) Apharwat viii-ix.1986 (for the colour key see Fig. 258). 178 PAULH. WILLIAMS Bombus (Pr.) lepidus oe @ Specimen @ Literature © Unconfirmed Equal area grid Map 39 Summary of the world-wide distribution of B. lepidus (for an explanation of the symbols see Map 5). |__| 0 100km Land above 3000m. Map 40 Distribution of B. lepidus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 327-334). THE BUMBLE BEES OF THE KASHMIR HIMALAYA a <°,%, SV Qtr, ey cao rate ats QUEENS WORKERS MALES 179 TW TOW wy WH WV WV Figs 327-334 Colour patterns of B. lepidus from localities in Kashmir: (327) Banidas 30. vi-2. vii.1959, (328) Bulandtar 3.vili.1954, (329) Dalti 7.ix.1954, (330) Lilam 17.viii.1953, (331) Surgun 29-30.vii.1953, (332) Nigagar 17.ix.1986, (333) Apharwat 25.vii.1986, (334) Apharwat viii.1986 (for the colour key see Fig. 258). 180 PAULH. WILLIAMS Bombus (Pr.) biroi @ Specimen G@ Literature © Unconfirmed 100km Land above 3000m. Map 42 Distribution of B. biroi in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 335-346). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 181 cy nw ee SB i CWE? | WOM UU Wy Vv Vv 335 338 341 344 QUEENS Th 4 @ 336 vw] WORKERS Sw A___A | A MALES Figs 335-346 Colour patterns of B. biroi from localities in Kashmir: (335) Chhantir Gah 5—7.viii.1954, (336) Rumbak 12.viii.1980, (337) Tungri, (338) Gumri 10-11. viii.1986, (339) Gumri 10-11. viii.1986, (340) Gumri 30.ix.1986, (341) Apharwat 25. vii. 1986, (342) Apharwat viii-ix.1986, (343) Apharwat 23.viii.1985, (344) Apharwat 25.vii.1986, (345) Apharwat 15.viii.1980, (346) Apharwat 29. vii.1986 (for the colour key see Fig. 258). 182 PAULH. WILLIAMS Bombus (Bo.) tunicatus @ Specimen @ Literature © Unconfirmed 100km Land above 3000m. Map 44 Distribution of B. tunicatus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 347-355). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 183 QUEENS WORKERS 349 MALES Figs 347-355 Colour patterns of B. tunicatus from localities in Kashmir: (347) Luskum 30.v.1954, (348) Gilgit vi.1954, (349) Shigar 1-3.x.1953, (350) Gulmarg vii—viii.1986, (351) Gulmarg viii.1986, (352) Gulmarg viii.1985, (353) Gulmarg vii—viii.1986, (354) Gulmarg 10.viii.1980, (355) Gulmarg viii-ix.1986 (for the colour key see Fig. 258). 184 PAULH. WILLIAMS Bombus (Bo.) lucorum @ Specimen @ Literature © Unconfirmed Equal area grid Map 45 Summary of the world-wide distribution of B. Jucorum (for an explanation of the symbols see Map 5). | es | 0 100km Land above 3000m. Map 46 Distribution of B. lucorum in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 356-370). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 185 QUEENS ‘Wle> <1 3e» Wie» «le» | - 361 @ 364 ) Y 369 | eee Lee fe fo) es. ee ee ez, ay: eer / ag 362 365 7370 MALES Figs 356-370 Colour patterns of B. lucorum from localities in Kashmir: (356) Baltit 14-17.vi.1954, (357) Baltit 14-17.vi.1954, (358) Baltit 15.viii.1913, (359) Baltit 14-17.vi.1954, (360) Atro Sar 11.viii.1954, (361) Chittakatha Sar 19-24. vili.1954, (362) Lal Pani 2-3.ix.1953, (363) Maru 19.v.1910, (364) Lal Pani 2-3.ix.1953, (365) Lal Pani 2-3.1x.1953, (366) Apharwat 30.viii.1986, (367) Apharwat viii.1985, (368) Apharwat 17.vii.1986, (369) Apharwat 16. viii. 1980, (370) Apharwat viii.1985 (for the colour key see Fig. 258). 186 PAULH. WILLIAMS Bombus (Sb.) asiaticus a @ Specimen @ Literature © Unconfirmed Equal area grid Map 47 Summary of the world-wide distribution of B. asiaticus (for an explanation of the symbols see Map 5). 100km Land above 3000m. Map 48 Distribution of B. asiaticus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 371-391). THE BUMBLE BEES OF THE KASHMIR HIMALAYA — 87 | oe qq a a es | —" w — P= we sa si os) oe So QUEENS Oe Io» | |e | OO «f < < © we a | nN © 17%) ~ nn @ w ~~ oo ® w So —) WORKERS p-) es oo hc qi ice hor v*) —I io) ies) XQ nN Lo) | =) ae & =) — MALES Figs 371-391 Colour patterns of B. asiaticus from localities in Kashmir: (371) Leh 19-27.ix.1986, (372) Leh 8.vii.1980, (373) Leh 19-27.ix.1986, (374) Thonde 1980, (375) Leh 3. vii. 1980, (376) Thonde 1980, (377) Nigagar 17.ix.1986, (378) Nigagar 9.viii.1986, (379) Nigagar 17.ix.1986, (380) Lal Pani 2-3.ix.1953, (381) Lal Pani 2-3.ix.1953, (382) Lal Pani 2-3.1x.1953, (383) Gulmarg viii.1985, (384) Gulmarg viii.1985, (385) Gulmarg 29.viii.1986, (386) Gulmarg viii.1985, (387) Gulmarg viii.1985, (388) Gulmarg 12.viii.1980, (389) Yusmarg 8.ix.1985, (390) Gulmarg viii.1985, (391) Gulmarg 12.viii.1980 (for the colour key see Fig. 258). 188 PAULH. WILLIAMS eke Lg = okt Bombus (Sb.) oberti | @ Specimen @ Literature © Unconfirmed Equal area grid Map 49 Summary of the world-wide distribution of B. oberti (for an explanation of the symbols see Map 5). | ee | 0 100km | Land above 3000m. Map 50 Distribution of B. oberti in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 392 & 393). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 189 & = im TOW WM wv 392 QUEENS @ @ Figs 392, 393 Colour patterns of B. oberti from Kashmir: (392) Nimaling 15.vii.1980, (393) Nimaling 20.vii.1980 (for the colour key see Fig. 258). 190 PAULH. WILLIAMS Bombus (MI.) ladakhensis @ Specimen @ Literature © Unconfirmed Equal area grid Map 51 Summary of the world-wide distribution of B. ladakhensis (for an explanation of the symbols see Map 5). EE —, 0 100km Land above 3000m. Map 52 Distribution of B. ladakhensis in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 394-396). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 191 QUEENS MALE Figs 394-396 Colour patterns of B. ladakhensis from localities in Kashmir: (394) Chushul vi.1925, (395) Hemis 25.ix.1986, (396) Nimaling 23.vii.1980 (for the colour key see Fig. 258). 192 PAULH. WILLIAMS Bombus (MI.) semenovianus es @ Specimen G@ Literature © Unconfirmed Equal area grid Map 53 Summary of the world-wide distribution of B. semenovianus (for an explanation of the symbols see Map 5). | ee | 0 100km Land above 3000m. Map 54 Distribution of B. semenovianus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 397-402). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 193 (p> QUEENS qo» dio © wu oO oO © PS —) — | |e WORKERS he. ac io) N=) v=) PS i=) N MALES Figs 397-402 Colour patterns of B. semenovianus from localities in Kashmir: (397) Baltit 14-17.vi.1954, (398) Baltit 14-17.vi.1954, (399) Martselang 25.ix.1986, (400) Baltit 14-17.vi.1954, (401) Dras 10~-11.viii.1986, (402) Hemis 25.ix.1986 (for the colour key see Fig. 258). 194 PAULH. WILLIAMS Bombus (ML. ) keriensis @ Specimen @ Literature © Unconfirmed Land above 3000m. Map 56 Distribution of B. keriensis in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 403-416). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 195 SX “es oC = = 403 405 QUEENS \wmm Nem @® ~ 404 WORKERS fe | 416 MALES Figs 403-416 Colour patterns of B. keriensis from localities in Kashmir: (403) Atro Sar 11.viii.1954, (404) Atro Sar 11.vili.1954, (405) Atro Sar 11.viii.1954, (406) Atro Sar 11.viii.1954, (407) Rawat 20—21.viii.1954, (408) Lal Pani 2-3.ix.1953, (409) Lal Pani 2-3.ix.1953, (410) Lal Pani 2-3.ix.1953, (411) Nimaling 15.vii.1980, (412) Nimaling 15.vii.1980, (413) Nimaling 24.vii.1980, (414) Apharwat viii.1985, (415) Apharwat viii.1985, (416) Apharwat 23.vili.1985 (for the colour key see Fig. 258). 196 PAULH. WILLIAMS Bombus (MI.) simillimus NA @ Specimen @ Literature © Unconfirmed Equal area grid Map 57 Summary of the world-wide distribution of B. simillimus (for an explanation of the symbols see Map 5). | ———— | 0 100km Land above 3000m. Map 58 Distribution of B. simillimus in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 417-421). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 197 QUEEN WORKERS MALES Figs 417-421 Colour patterns of B. simillimus from localities in Kashmir: (417) Harwan 10.ix.1985, (418) Harwan 10.ix.1985, (419) Harwan 10.ix.1985, (420) Harwan 10.ix.1985, (421) Banihal 3.ix.1986 (for the colour key see Fig. 258). 198 PAULH. WILLIAMS Bombus (MI.) pyrosoma Ae @ Specimen @ Literature © Unconfirmed Equal area grid Map 59 Summary of the world-wide distribution of B. pyrosoma (for an explanation of the symbols see Map 5). ——————————, 0 100km Land above 3000m. Map 60 Distribution of B. pyrosoma in Kashmir (see Map 6; for a key to the colour pattern symbols see Figs 422-430). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 199 QUEENS WORKERS LEE LEAN fF) CO | | ee eee) CLA EL EL CN Gizy Gay YU 94 "7 430 MALES Figs 422-430 Colour patterns of B. pyrosoma from localities in Kashmir: (422) Kel 16.viii.1953, (423) Kel 16. viii.1953, (424) Kel 16. viii.1953, (425) Gulmarg 2.ix.1985, (426) Gulmarg viii.1985, (427) Apharwat 14.ix.1986, (428) Gulmarg vili.1985, (429) Gulmarg viii.1985, (430) Apharwat 14.ix.1986 (for the colour key see Fig. 258). 200 PAULH. WILLIAMS Bombus (ML) rufofasciatus @ Specimen @ Literature © Unconfirmed Map 61 Summary of the world-wide distribution of B. rufofasciatus (for an explanation of the symbols see Map 5). aT 0 100km Land above 3000m. = NI: Map 62 Distribution of B. rufofasciatus in Kashmir (see Map 6; for a a key to the colour pattern symbols see Figs 431-439). THE BUMBLE BEES OF THE KASHMIR HIMALAYA 201 QUEENS WORKERS MALES Figs 431-439 Colour patterns of B. rufofasciatus from localities in Kashmir: (431) Atro Sar 11.viii.1954, (432) Atro Sar 11.viti.1954, (433) Rawat 20-21.viii.1954, (434) Apharwat 23.viii.1985, (435) Apharwat viii-ix. 1986, (436) Apharwat viii. 1985, (437) Apharwat 24.vii.1986, (438) Apharwat viii.1985, (439) Apharwat viii.1985 (for the colour key see Fig. 258). 202 PAUL H. WILLIAMS INDEX This index includes references to the names of bumble bees used in the text and figures, but not to those in the nomenclatural summary (pages 29-31), in the keys to species (pages 31-38), or in the lists of similar species at the end of the comments on each species. Invalid or unavailable names are in italics; principal references affecting the application of names are in bold; references to the figures that follow the text are in (parentheses). abbotti 77 abnormis 70-71, 74 affinis 80 afghanus 44 agnatus (Skorikov, 1912a) 77 agnatus (Skorikov, 1933b) 77-78 agrorum 38 alaiensis 82, 85, 87 albocaudatus 87 albohirtus 66 albolateralis 52-54 albopleuralis 52, 55, 59 Allopsithyrus 45 Alpigenobombus 36, 62, 65-69 Alpigenibombus 65 alpigenus 65 Alpinobombus 39, 46 alpivagus 96-98 altaicus 42 amurensis 62, 65 Anodontobombus 69-70 Apathus 44 arcticus 39 ashtoni 45-46 Ashtonipsithyrus 45-46 asiaticus 13, 19-27, 29, 49, 87-92, 107, 110-118, 120 (129, 133, 135, 137-138, 142, 186-187) assamensis 58, 60-61 assellus 42 atratus (Friese) 14 atratus (Franklin) 14, 119 audax 82 avinoviellus 15, 19, 21, 24, 39-41, 72, 110, 112-114, 120 (129-132, 134, 136, 138, 141-144) barbutellus 45 bellicosus 118 beresovskii 66, 68-69 bianchii 94-95 biroi 6, 19, 24, 74, 77-79, 110, 112, 114, 120 (131, 133, 135, 137, 138, 139, 140, 180-181) bohemicus 45—46 (129, 132, 134, 136, 138-139, 150-151) Bombias 39, 46 Bombini 10, 31 Bombus 14-16, 31, 38, 62, 67, 79-87 braccatus 60 Brachycephalibombus 89 branickii 47, 48-49, 110 (132, 134, 136, 138-139, 154-155) Bremus 38 breviceps 54, 60, 67, 119 brevivillus 119 brodmannicus 78, 118-119 bryorum 70-71 buccinatoris 58 caliginosus 77 callophenax 39-40, 90 campestris 41, 45, 49-50, 101 canariensis 83 canosocollaris 101, 103 centralis 77 Ceratopsithyrus 45 championi 105-106 chayaensis 67 chinensis (Morawitz) 42, 47, 105 chinensis (Dalla Torre) 105 chinensis (Skorikov) 42 cinnameus 59-60 Citrinopsithyrus 45 citrinus 45 Coccineobombus 89 Confusibombus 39, 46 confusus 39 convexus 15, 42 coreanus 46 cornutus 45 Crotchiibombus 89 cryptarum 82, 84-85 Cullumanobombus 46, 88 Dasybombus 89 defector 15, 40-42 dentatus 67 difficillimus 62-65 Diversibombus 52 Diversobombus 52-60, 113 diversus 52, 54 duanjiaoris 92-93 dyspostzonatus 77 elisabethae 48, 77 Eopsithyrus 45 ephippiatus 100, 115 eriophoroides 48 eurythorax 103 Eversmannibombus 42, 46 exil 14-15, 46 exiln 46 Exilobombus 15, 46 eximius 100 falsificus 88, 91 fedtschenkoi 62, 64-65 ferganicus 21, 49-50 (129, 132, 134, 136, 138-139, 156-157) fernaldae 45, 51 Fernaldaepsithyrus 45, 50-51 Fervidobombus 119-120 Festivobombus 70, 80, 88, 100 festivus 70-71, 88, 100 flavescens 55, 74 flavidus 51 flavifrons 77-78 flaviventris 62, 89, 92-93 flavobistriatus 77-78 flavocorbicularis 102-103 flavodorsalis (Franklin) 88 flavodorsalis (Skorikov) 88, 90 flavofasciatus 77 flavopilosus 73-74 flavostriatus 77 flavothoracicus (Hoffer) 101 flavothoracicus (Bingham) 100, 101-103, 106 fletcheri 70-71 formosellus 102, 104 fragrans 62, 65 franklini 80 Fraternobombus 89 fraternus 89 frigidus 78 friseanus 101-103, 106, 115 fulvocinctus 80 Funebribombus 89 funerarius 60 fuscocaudatus 87 gansuensis 50-51 gantokiensis 52-54 geminatus 52-54, 58 genalis 67 genitalis 75-76 gilgitensis 79-80 grahami 67 griseofasciatus 62-63 grossiventris 99-101 haematurus 71-72 haemorrhoidalis 6, 52, 54, 58-61, 67, 74, 109-114, 119-120 (129, 131-132, 134, 136, 138, 140, 164-165) haemorrhous 99, 101 handlirschi 89 handlirschianus 15, 40-42 heicens 91 hilaris 75-76 himalayanus 15, 19, 24, 41-44, 110-114, 120 (132, 134, 136, 138, 141, 146-147) hoenei (Bischoff, 1936:10) 102-103 hoenei (Bischoff, 1936:26) 47 Hortobombus 52 hortorum 13, 87, 113 huangcens 91 hummeli 54 hyperboreus 39 hypcorum 70 Hypnorobombus 70 Hypnorubombus 69 hypnorum 19, 24, 69, 70-72, 110, 112, 114, 119 (132, 134, 136, 138, 172-3) hypocrita 83 ignitus 83 THE BUMBLE BEES OF THE KASHMIR HIMALAYA incertoides 96-98 incertus 97 indicus 49 infirmus 74-75 infrequens 75 insidiosus 52, 54 intermedius 105 Jacobsoni 82, 85-86 jonellus 78 Kallobombus 39, 46 karakorumensis 96, 98 kashmirensis 19, 23-28, 49, 65, 66-69, 75, 108, 110-112, 114, 120 (130, 132, 134, 136, 138, 170-171) keriensis 19, 24, 49, 94, 96-99, 109-110, 112-114 (133, 135, 137-138, 194-195) khasianus 58, 60-61 klapperichi 45 kohistanensis 72 kohli (Cockerell) 96 kohli (Vogt) 96-98 kotzschi 77-78 kozlovi 94, 96 Kozlovibombus 93-94, 100 kozlowi 96 Kozlowibombus 94 kuani 51 Laboriopsithyrus 45 laboriosus 45 ladakhensis 7, 94-95, 110, 112, 114 (133, 135, 137-138, 190-191) Lapidariibombus 94 Lapidariobombus 93-94 lapidarius 13, 89, 93-97, 99-100, 113, 118-119 Laponicobombus 70 Lapponicobombus 69 lapponicus 69, 77 lefebvrei 65-66 lemniscatus 72, 73-75, 112, 114 (129, 133, 135, 137-141, 176-177) lepidus 19, 24, 74, 75-76, 110, 112, 114, 120 (133, 135, 137-141, 178-179) Leucobombus 79 leucopygos 72 leucopygus (Illiger) 72 leucopygus (Morawitz) 72 leucurus 72 longiceps 87-90 longipennis 82-83, 85 longipes 54 lucocryptarum 82, 85 lucorum 13, 19, 24, 46, 75, 81-87, 109-110, 112, 114, 119 (133, 135, 137-138, 184-185) lugubris (Kriechbaumer) 42 lugubris (Morawitz) 42 luteipes 75—76 maderensis 84 magnus 82, 84-86 magrettianus 52-53, 55 maidli 62-63 makarjini 15, 39-41, 43 malaisei 54 margreiteri 42 marussinus 15, 40-43, 44, 112-114 (132, 134, 136, 138, 141, 148-149) Mastrucatobombus 65 mastrucatus 65-68 maxillosus 42 maxwelli 52-53, 55 Megabombus 53, 113 meinertzhageni 66-67 Melanobombus 62, 67, 88-89, 92, 93-107, 113 melanopoda 119 melanopygus 115 melanurus 19, 21, 24, 61-65, 110-114 (132, 134, 136, 138, 166-167) Mendacibombus 14-16, 38-46 mendax 15, 38-44 meridialis 96-98 mesoxanthus 50 Metapsithyrus 45, 49-50 metcalfi 60 mimeticus 52, 54—55 miniatocaudatus (Vogt, 1909) 87 miniatocaudatus (Vogt, 1911) 87-88, 91 miniatus 101-102, 104 minshanicus 52-53, 55 mirus 73-74 mlokosievitzii 119 moderatus 82-83 modestus (Eversmann) 82 modestus (Cresson) 82 mongolicus 82-83, 85-86 montivagus 52, 54-55 montivolans 58-60 morawitzi (Radoszkowski) 89-90, 92-93 morawitzi (Friese) 47 morawitzianus 50 (132, 134, 136, 138-139, 158-159) morawitziides 65-66, 67 morio 96, 119 Mucidobombus 46 mucidus 14 nemorum 45 nepalensis 47 ningpoensis 52, 55 nevadensis 14, 39 niger 14, 119 nigrozonatus 77 niveatus 39, 89, 90 Nobilibombus 65-66 nobilis 65—66, 67 novus 46-48 (130, 132, 134, 136, 138-139, 152-153) nursei 73, 75, 77-78 oberti 7, 87, 89-90, 92-93, 110, 112, 114, 120 (129, 133, 135, 137-138, 142, 188-189) Obertibombus 87 Obertobombus 87, 93 obtusus 89-91 occidentalis 83 ochrobasis 92 oculatus 99-100 orichalceus 67 Orientalibombus 46, 58-61, 113 Orientalobombus 58 orientalis 58-61 oshanini 88, 90 203 pamirus 65 parthenius 71, 74, 75-76 pascuorum 38, 113 patagiatus 83, 86 pectoralis 59 pennsylvanicus 88 pensylvanicus 88 peralpinus 73-74 perplexus 71 persicus 42 personatus 7, 62, 64-65, 110, 112, 114 (132, 134, 136, 138, 168-169) Phariensis 94-95 pieli 49-50 Poecilobombus 69 pomorum 65 postzonatus 96-98 Pratibombus 70 Pratobombus 69 pratorum 69, 71, 78, 82, 119 Pressibombus 80 pressus 80 problematicus 95-96 prshewalskyi 105 Psithyrus 14-16, 39, 44-51 pulcherrimus 65, 66 pullatus 119 pyrenaeus 78 Pyrobombus 46, 62, 67, 69-79, 119 pyrosoma 19, 22, 24-27, 94, 100, 101-107, 110, 112, 114-115, 119 (128, 133, 135, 137-138, 198-199) Pyrrhobombus 69 pyrrhosoma 101-102 quasibreviceps 52, 54 redikorzevi 50 regeli 87-88 reinigi 82, 85-86 reticulatus 94-95 Rhodobombus 39, 65 richardsiellus 101-102, 104-106 roborowskyi 64-65 Robustobombus 89 rotundiceps 54, 60, 74, 119 Rubicundobombus 89 Rufipedibombus 80, 88, 100 rufior 105 tufipes 100, 119 rufitarsus 42 rufocinctus (Cresson) 105, 115 rufocinctus (Morawitz) 105 rufofasciatus 6, 13, 19-27, 47, 94-95, 100, 102, 104, 105-112, 114-115, 120 (128, 130, 133, 135, 137-138, 200-201) rufoflavus 55 rupestris 42, 44, 47, 48, 105 sandersoni 77 secundus 52 semenovi 92-93 semenovianus 7, 78, 95-96, 110, 112-114, 120 (133, 135, 137-138, 142, 192-193) semibreviceps 59, 61 semicoloricontrarius 59, 61 senex 119 Senexibombus 53 separandus 96-98 204 Separatobombus 89 shaposhnikovi 15, 42 Sibiricibombus 87 Sibiricobombus 113 sibiricus 87, 89 sicheli 97 sichelii 97-98 signifer 75 similis 99 simillimis 99 simillimus 99-102, 105-106, 109-110, 112-114, 119 (128, 133, 135, 137-138, 142, 196-197) simlaensis 79 sitkensis 69 skorikovi 50-51 (132, 134, 136, 138-139, 160-161) sonani 75 sonorus 88 soroeensis 28, 39, 87 sporadicus 80, 86 stenothorax 103 stramineus 66 subdistinctus 62-63 Subterraneibombus 61 Subterraneobombus 61-65, 80 subterraneus 61-62, 64 subtunicatus 39 16, 62, 67, 87-94, subtypicus 72-73, 110-112, 114 (131-138, 140, 174-175) suckleyi 46 sulfureus 90 superbus 42 susteraianus 49 tajushanensis 49 Tanguticobombus 94, 100 tanguticus 89, 93-94, 100 tenellus 96, 98 tenuifasciatus 88 Terrestribombus 79 terrestris 38, 79-80, 81-86, 99 terricola 83 tetrachromus (Cockerell) 66, 75 tetrachromus (Friese) 75-76 Thoracobombus 60, 113 tibetanus (Morawitz) 45, 74 tibetanus (Friese) 74 tonsus 99, 101 trifasciatus 6, 19, 24, 52-60, 67, 74, 107, 109-114, 119 (129, 131-132, 134, 136, 138, 162-163) trilineatus 101 tschitscherini 62-63 tunicatus 19-22, 24, 79-81, 107, 109-110, 112-115, 119-120 (133, 135, 137-138, 182-183) PAUL H. WILLIAMS turkestanicus 15, 41-44, 113 turneri (Richards, 1929a) 52 turneri (Richards, 1931) 52-54 Uncobombus 69-70 ussurensis 54 vagans 77 validus 66, 67 vandykei 77 variopictus 94-95 varius (Lepeletier) 41 varius (Skorikov) 15, 41-43 vestalis 46 vorticosus 90 waltoni 15, 40-41, 42 waterstoni 105 wilemani 52, 55 wurfleini 67 wurflenii 28, 65, 67-68, 119 wutaishanensis 102-103 xionglaris 91 xizangensis 67 yuennanicola 75-76 yunnanicola 75 zhadaensis 93 \ 1 The bumble siete of the Kashmir Himalaya (Hymenoptera: Apidae, Bombini). Paul H. Williams ENTOMOLOGY SERIES Vol. 60, a. is April 1991