Parasitic and Robber Bees

In many groups of organisms that store food for themselves or their young, parasitic or robber individuals, species, or genera can be found. Such forms steal or feed upon the stored food, often starving or more directly killing the hosts. Bees well illustrate such tendencies; reviews were by Bischoff (1927: 397-401), Grutte (1935), Bohart (1970), and Iwata (1976).

This section consists of four parts, as follows: first, on nest usurpation and robbing; second, on social parasites that live in the nests of social host bees; third, on clep-toparasites that leave their eggs, cuckoo/cowbird fashion, in the cells of their host bees; and fourth, on some common attributes of social parasites and cleptoparasites.

Usurpation and robbing. Intraspecific usurpation is probably frequent. It has been little studied because, unless bees have been marked for individual recognition, it is likely to go unnoticed. When a solitary bee comes out of a nest hole and is the same species that was seen there yesterday, one naturally assumes that it is the same individual. Usually, this assumption is correct, but various studies of wasps and bees made with marked individuals show that intruders are often present, that they enter nests, often fight with nest owners, and sometimes win and take over nests made by other individuals. Barthell and Thorp (1995) showed that in Megachile apicalis Spin-ola the usurpers average larger than their victims, at least in highly competitive situations. Presumably, larger individuals are better able to win contests.

Wcislo (1987) listed 17 species of bees in which usurpation or intraspecific robbing has been reported. Perhaps to facilitate nest recognition by the owner as well as, among social species, to reduce admission of foreign individuals while allowing access by nestmates, many bees and wasps appear to have distinctive individual or nest odors. Nest guarding, too, is common, especially in social species. It is my clear impression that the commonest function ofsuch guards is to reject conspecific individuals that attempt to enter, although of course guards also react strongly to other insects, including parasites and predators. Summary accounts of individual and nest recognition in bees are by Michener and Smith (1987) and Breed and Bennett (1987).

As might be expected, the defenses implied in the preceding paragraph do not always succeed. In the Meliponini and Apini, robbing is frequent. Weak colonies may not be able to defend themselves against intra-specific or interspecific robbing, in spite of guards and constricted entrances. Two genera of Meliponini, the neotropical Lestrimelitta and the African Cleptotrigona, are specialist robbers. They have nests of their own but obtain food, not from flowers, but from nests of other species of Meliponini. Accounts of their robbing activity are by Portugal-Araujo (1958), Wittmann et al. (1990), and Sakagami, Roubik, and Zucchi (1993). They lack tibial corbiculae for carrying pollen and have other common features, as listed in Section 120 on the Meliponini, but these features appear to be convergent; the two genera evolved from different nonrobbing ancestors.

Social parasites. Parasitic bees, as distinguished from robbers, can be divided into two groups, social parasites and cleptoparasites. A female social parasite enters a nest of the social host and in some way replaces the queen, so that the workers ofthe host thereafter rear offspring of the parasite rather than of their own species. The host must be social but sometimes (as in Allodapini) is marginally so. The female parasites are found living in colonies of the hosts. Although derived from eusocial ancestors (see Sec. 5), the parasitic species lack a worker caste.

There are relatively few social parasites among bees (Table 8-1). In the genus Bombus, the subgenus Psithyrus is entirely parasitic and has lost the corbicula for carrying pollen. The subgenera Alpinobombusand Thoracobombus each have a parasitic species; they have normal or nearly normal but presumably functionless corbiculae. Thus in Bombus there have been three origins of social parasitism.

Allodapini is the only group other than Bombini that contains indisputably socially parasitic species. There are eight or nine such species, and four others considered as probable social parasites because of reduction in the pollen-carrying scopa (Table 8-1). (For an account of the parasites, see Michener, 1970.) Except for the two species of Eucondylops, two species of Exoneura subgenus In-quilina, and the two similar species of the Braunsapis bre-viceps group, each parasitic species appears to have arisen independently from nonparasitic relatives. There is one or more parasitic or probably parasitic species derived from or included in each of the following genera: Allo-dape, Allodapula, Braunsapis, Exoneura, and Macrogalea. Eleven independent origins of parasitic allodapine species are indicated, but in no case has a parasitic allo-dapine line undergone as much speciation to produce a group of parasitic species as has Bombus (Psithyrus); the largest such groups in the Allodapini contain only two known species.

In both Bombus (Psithyrus) and Braunsapis kaliago Reyes and Sakagami, females functionally replace queens of the host; the latter are sometimes killed but often remain alive. The female of the parasitic Braunsapis kaliago becomes unable to fly but participates in many nest activities such as the feeding of larvae (Batra, Sakagami, and Maeta, 1993).

Because of a scarcity ofinformation on the often rather common parasitic halictids, the nature of their parasitism is often in doubt. Those that parasitize solitary hosts are of necessity cleptoparasitic (see below). Those that attack social halictines may also leave the nest promptly after oviposition, like other cleptoparasites. But some of those species that attack social halictines are more or less social parasites, staying in the host nest and possibly taking on qualities of the host queen. Field collecting and nest excavations suggest that in a species of Microsphecodes the females remain in the host (Lasioglossum subgenus Dial-ictus) nests and may be social parasites (Eickwort and Eickwort, 1972b). Knerer (1980) reported on two species of Sphecodes whose females were found in nests of hosts (Halictus maculatus Smith). Wcislo (1997) found similar

Table 8-1. Social Parasites.

The notations are as follows: (p) Probable social parasite, recognized by the reduced scopa but not known from host nests; (n) Found in host nests. Numbers in parentheses after subfamily names indicate the probable numbers of origins of social parasitism. An asterisk (*) in front of a generic or subgeneric name indicates that all species of that taxon are parasitic. Specific names linked by an and represent two species that probably diverged from a common parasitic ancestor, and therefore are considered to represent a single origin of parasitism. Note that bees known to be clep-toparasites are not included in this list; see Table 8.2.

Higher taxa and parasitic taxa Host taxa

Family Halictidae

Subfamily Halictinae (3) Tribe Halictinia *Sphecodes, n

* Microsphecodes, n Lasioglossum (Dialictus) b, n

Family Apidae

Subfamily Xylocopinae (11) Tribe Allodapini

Allodape greatheadi Michener, p Allodapula guillarmodi Michener, p Braunsapis bislensisMichener and Borges, Braunsapis breviceps (Cockerell), n and B. kaliago Reyes and Sakagami, n Braunsapis natalica Michener, n? Braunsapis pallidaMichenerc, p *Eucondylops konowi Brauns, n, and E. reducta Michener, n

* Effractapis furax Michener, p *Inquilina excavata (Cockerell), n, and I. schwarzi Michener, n Macrogalea mombasae (Cockerell), n *Nasutapis straussorum Michener, n Subfamily Apinae (3) Tribe Bombini

Bombus (*Psithyrus), n B. (Alpinobombus)arcticus(Quenzel), n B. (Thoracobombus) inexspectatus (Tkalcu), n a As indicated in the text, some species of Sphecodes and its relatives may be more like social parasites than cleptoparasites. They are included in this table as reminders that they may be social parasites; their behavior is too little known for assurance on either count.

b The parasitic species of Dialictus were formerly segregated as a parasitic genus Paralictus. c Some supposedly parasitic Australian Braunsapis species are now believed to be probably not parasitic.

Halictus

Lasioglossum (Dialictus) Lasioglossum (Dialictus)

Braunsapis pis pis

Exoneura Exoneura

Bombus Bombus n behavior in a parasitic species of Lasioglossum (Dialictus) (the former Paralictus). There may be intermediates between social parasites and cleptoparasites among the parasitic halictines.

Cleptoparasites. The remaining parasitic bees, the great majority, are cleptoparasites. A cleptoparasite enters the nest of a host and lays an egg in a cell. In most cases the adult parasite then leaves, although sometimes (e.g., in Hoplostelis s. str.) it ejects the host and stays in the nest. The parasite larva feeds on the food that had been provided for a host larva. Such bees are often appropriately called "cuckoo bees." Rarely recorded is intraspecific cleptoparasitism (Field, 1992), in which a bee opens a cell of another individual of its own species and replaces the egg. Most cleptoparasites belong to their own obligately parasitic species, genera, tribes, or subfamilies. The host is commonly solitary, although some social Halictinae are hosts of cleptoparasites. Table 8-2 lists the cleptoparasitic taxa. Grutte (1935) published a valuable account of such parasites. Some genera are recognized as cleptoparasites only by reduction or lack of pollen-manipulating and pollen-carrying structures, especially the scopa of females, and their probable association with solitary hosts. In Table 8-2, such forms are marked "p" for probable.

Females of some mostly cleptoparasitic genera such as Sphecodes (Halictini) destroy the egg of the host and replace it with their own. In such cases one never finds a cell with two or more eggs; the Sphecodes egg is not noticeably different from that of other halictine bees. It is failure to find cells with two or more eggs, combined with the ordinary (i.e., hostlike) structure of the young parasitic larvae, that leads to the conclusion that the host egg is de-

Hylaeus (Nesoprosopis)

Table 8-2. Cleptoparasites.

The notations are as follows: (p) Probable cleptoparasite; see Section 8. (n) Found in or reared from host nests, but the method of killing the host egg or larva is not known. (ad) Host egg or larva killed by adult female parasite; this is assumed if among many cells studied, none contained both host and parasite eggs. (lo) Host egg or young larva killed by active but otherwise rather ordinary parasite larva. (lm) Host egg or young larva killed by active young parasite larva with scle-rotized head and sickle-shaped mandibles. Numbers in parentheses after subfamily names represent the probable numbers of origins of cleptoparasitism. Generic names are omitted when all species of a subfamily or tribe are cleptoparasitic. An asterisk (*) in front of a name indicates that all species of that taxon are parasitic. Note that bees known to be social parasites are not included in this list; see Table 8.1.

Higher taxa and parasitic taxa Host taxa

Family Colletidae

Subfamily Hylaeinae (1)

Hylaeus (Nesoprosopis). in part, n Family Halictidae

Subfamily Halictinae (9) Tribe Halictinia

*Echthralictus, p (Derived from Homalictus) *Eupetersia, Sphecodesclade, p Halictus (*Paraseladonia), p Lasioglossum

Dialictusb, in part, ad

* Paradialictus, p

*Microsphecodes, Sphecodes clade, ad *Nesosphecodes, Sphecodes clade, p *Parathrincostoma, p *Ptilocleptis, Sphecodes clade, p

* Sphecodesc, ad Tribe Augochlorini

Megalopta (Noctoraptor), p Megommation (*Cleptommation), p *Temnosoma, p Family Megachilidae

Subfamily Megachilinae (10) Tribe Osmiini

*Bekilia, position doubtful, p Hoplitis (*Bytinskia), n Tribe Megachilini

* Coelioxys, lm

*Radoszkowskiana, lm Tribe Anthidiini *Afrostelis, p *Euaspis, ad *Hoplostelis, ad *Larinostelis, p

(continues)

Lasioglossum (Dialictus) Halictini

Halictinae and othersd

Hoplitis

Megachile, less frequently various Apidae (see Sec. 84)

Euglossini stroyed by the adult parasite. Sick et al. (1994) saw a female Sphecodes in an observation nest of Lasioglossum (Evylaeus) malachurum (Kirby) enlarge the opening of a host cell, enter the cell head first, for two minutes probably destroying the host egg, then back out, turn, and back into the cell, remaining there for five minutes probably laying her egg. She then came out and closed the cell with soil, and the next day, on leaving the nest, she closed the nest entrance. Genera known to parasitize host cells in this way are marked "ad" (for adult) in Table 8-2.

Females of most cleptoparasites, however, lay eggs in host cells without destroying the host egg or larva. The egg of the parasite may be (1) inserted into and hidden in the cell wall of an as yet unclosed cell while the host is out of the nest, or (2) laid in a finished and closed host cell by the parasitic mother through a hole that she makes and later seals in the cell wall or closure.

Rozen and Ozbek (2003) and Rozen (2003a) described the mature oocytes or eggs of numerous clep-toparasitic megachilid and apid taxa. Relative to body size, eggs of cleptoparasites are smaller than those of related nonparasitic solitary bees. Moreover, larger num-

Table 8-2. Cleptoparasites (continued)

Higher taxa and parasitic taxa

Host taxa

*Dolichostelis, ad *Xenostelis, p *Tribe Dioxyini, lo Family Apidae

*Subfamily Nomadinae (1), lm

Subfamily Apinae (10) Tribe Ctenoplectrini

* Ctenoplectrina, p (Derived from Ctenoplectra) *Tribe Rhathymini, lm

*Tribe Ericrocidini, lm *Tribe Melectini, lm *Tribe Isepeolini, lm *Tribe Protepeolini, lm *Tribe Osirini, lo

Tribe Tetrapediini

* Coelioxoides, lm Tribe Euglossini

*Exaerete, ad, lo

Megachilinae Megachile

Megachilinae many groups, see Section 91

Epicharis Centridini Anthophorini Colletes, Canephorula Diadasia

Epeoloides on Macropis, others on Tapinotaspidini

Tetrapedia

Eulaema, Eufriesea Eulaema a As indicated in the text, some species of parasitic halictids may be more like social parasites than cleptoparasites.

b A few North American species of Dialictus formerly placed in Paralictus are parasitic on other Dialictus.

c Eupetersia, Microsphecodes, and Ptilocleptisare probably members of the Sphecodesphyletic line, although no analysis has been made.

n bers are ready to be laid than in nonparasitic solitary bees; sometimes there are more ovarioles than the plesiomor-phic number for the family. Compared with eggs of solitary bees, those of cleptoparasites often show thickening and elaboration of ornamentation of the chorion and elaboration of the micropyle. Rozen (2003a) provided a table showing, for many cleptoparasitic Apidae, that the eggs are significantly smaller relative to body size in those that hide their eggs in open host brood cells, often still being provisioned, than in those that open completed and closed host cells for oviposition and do not need to hide their eggs because the host is no longer about.

Eggs hidden in cell walls are unusually small compared to those of other bees of the same size (see Sec. 4), and are quite diverse in structure, often differing widely from the usual bee egg with its soft chorion. Specialized eggs are laid by all Nomadinae and also by Protepeolini in the Apinae and Coelioxys in the Megachilini. Their eggs are inserted into the inside wall of the host cell or otherwise hidden before the cell is closed by the host, often before the provisioning is completed. Eggs inserted by cleptopara-sitic bees into finished, closed cells are of ordinary size and form, like those of Sphecodes. Such eggs are those of Melectini, Ericrocidini, and others.

Interesting diversity exists in the Nomadinae not only in egg form but also in the manner in which the eggs are inserted in the hosts' cell walls (Rozen, 1991a, 1992a).

Some (e.g., Doeringiella) are completely buried and at right angles to the cell wall; some are thrust into the wall only partway and left with one end projecting (Nomada) (see Radchenko, 1981); some are doubled over in the cell wall (Oreopasites); and some are placed in the wall almost parallel to its surface with one side exposed, that side hardened and roughened, in contrast to the usually soft chorion (Biastini, illustrated by Rozen, Roig-Alsina, and Alexander, 1997). Epeolus, which lays eggs in Colletes cells composed of two cellophane-like layers, places its egg between the two layers, with the anterior end exposed. Females of nomadine genera have distinctive structures, especially of S6, presumably for their particular methods of egg laying (Figs. 8-10f, 91-2, 95-3c, d).

Unlike nonparasitic bees that ordinarily lay one egg per cell, cleptoparasitic forms as different as Nomada and Coelioxys, i.e., forms in different families, frequently put two to several eggs into parasitized cells. The resultant larvae then kill not only the host egg or larva but also their conspecific competitors until there is only one left alive. Some cleptoparasitic larvae are active and able to kill the host egg or larva with ordinary-sized but sharp mandibles [e.g., Stelis, "lo" (for larva ordinary) in Table 8-2]. It is usually young larvae that do this, but Rozen (1987a) indicated that even last-stage larvae of Stelis may have modifications for killing hosts. In contrast, the young larva of Coelioxys and parasitic Apidae has a large, usually more or less prognathous, sclerotized head and sickle-shaped mandibles with which to kill the host egg or young larva [e.g., Coelioxys(Fig. 84-5a, d), "lm" (for larval mandibles) in Table 8-2]. "Prognathous" means that the head and mouthparts are directed forward, rather than more or less downward as in other larval bees. In the cleptoparasitic Apidae the first-stage larvae are those specialized for killing the host or their conspecific competitors, whereas in Coelioxys (Megachilidae) it is the still small second- or third-stage larvae that have the largest mandibles (Fig. 84-5c, d). Rozen (1991a) described and illustrated the known first-stage larvae of parasitic tribes of Apinae (except for the parasites in the Euglossini) and compared them with first-stage larvae of Nomadinae. Only in the Rhathymini, among first-stage parasitic Apinae, is the larval head incompletely sclerotized, more or less spherical, and hypognathous. Young larval morphology and presumably behavior are thus convergent in various groups independently derived from nonparasitic ancestors.

Rozen (2001) has given a key to the mature larvae of cleptoparasitic bees, along with a summary of host relationships.

The numbers after subfamily names in Table 8-2 indicate the numbers of independent origins of cleptopara-sitism within those subfamilies. These estimates may be high, since phylogenetic analyses remain to be done or are problematic. Thus Radoszkowskiana and Coelioxys may not represent separate origins of parasitism, and Afrostelis, Larinostelis, Stelis, and possibly Euaspis could have evolved from a single parasitic ancestor. The tribes Prote-peolini and Isepeolini, even perhaps Osirini, might be basal branches of Nomadinae and thus have the same parasitic ancestor as that subfamily. Melectini, Ericroci-dini, and Rhathymini might have a common parasitic ancestor. Thus the total number of origins of cleptopara-sitism could be less than the 31 indicated in Table 8-2. Alexander (1990), in a conservative list that omitted the then doubtful entries such as the Hylaeinae and Osmiini, nonetheless enumerated 17 origins of cleptoparasitism.

The parasitism by members of the Colletidae is based on five Hawaiian species of Hylaeus (Nesoprosopis) reported to be parasites of other species of the same subgenus by Perkins (1899). He evidently found parasites in host nests, and recognized parasites by the reduced pollen-gathering hairs on the front tarsi of females, but gave no data to clarify or verify his conclusions. His work, however, was usually dependable and is supported by Daly and Magnacca (2003).

In spite of the taxonomically diverse groups of bees that have evolved cleptoparasites, many large and sometimes old and widespread groups have not done so. Except as noted above, the Colletidae is such a group. Parasites of any sort are unknown in the Andrenidae, the Nomiinae and Rophitinae in the Halictidae, the Melittidae, and various large groups of Apidae such as the Ceratinini, Xylo-copini, and Eucerini.

The hosts of cleptoparasitic bees are always other bees. Emery's rule (see Wilson, 1971) for parasitic aculeate Hy-menoptera indicates that parasites usually attack their close relatives. Cleptoparasitic bees that are similar to their nonparasitic relatives, so that both are in the same genus, tribe, or subfamily, are usually parasitic on members of that genus, tribe, or subfamily. Thus parasitic Hal-ictini mostly parasitize other Halictini, although as noted in the discussion of Sphecod.es (Sec. 66), a few Sphecod.es species parasitize other halictids and even bees in other families. Parasitic Euglossini parasitize other Euglossini. Parasitic Megachilinae are parasitic on other megachi-lines except that a few of the many Coelioxys species attack Anthophora,, Centris, Euglossa,, or, reportedly, Tetralonia (Bischoff, 1927: 398), and Hoplostelis s. str. attacks Eu-glossini. The parasitic tribes of Apinae are, so far as is known, all parasitic on other tribes of Apinae, except for Isepeolini, which parasitizes Colletinae and Epeoloides, which parasitizes Macropis (Melittidae). As noted elsewhere, the Isepeolini could be a basal nomadine tribe.

A common observation is that species of cleptopara-sitic bees vary greatly in size. Sometimes two size classes are evident; that they parasitize host species of different sizes is a common assumption, rarely verified. Such size classes of parasites can be explained (1) as probable cryptic species, each specializing on a host species of a certain size, (2) as "races" specializing on such hosts, or (3) as direct effects of food quantity on the growth and maturation of the larvae of the parasitic species. A study of Coe-

Figure 8-1. Proboscidial structures of parasitic and nonparasitic al-lodapine bees, maxilla at left, labium at right. a, b, The social parasites Eucondylops reducta Michener and Nasutapis straussorum Michener; c, The nonparasitic species Allodapula melanopus (Cameron). The bodies of these bees are roughly the same size; these drawings are to the same scale, thus showing the reduction of the proboscis in size as well as in palpal segmentation in the social parasites. Drawings modified from Michener, 1970.

Figure 8-1. Proboscidial structures of parasitic and nonparasitic al-lodapine bees, maxilla at left, labium at right. a, b, The social parasites Eucondylops reducta Michener and Nasutapis straussorum Michener; c, The nonparasitic species Allodapula melanopus (Cameron). The bodies of these bees are roughly the same size; these drawings are to the same scale, thus showing the reduction of the proboscis in size as well as in palpal segmentation in the social parasites. Drawings modified from Michener, 1970.

Ctenoplectrina Africa

Figure 8-2. Bodies of females of Halictini. a,The cleptoparasitic Sphecodes monilicornis (Kirby); b, The nonparasitic Lasioglossum malachurum (Kirby); hairs are omitted on the left half of each. Note the coarse head and thoracic punctation and propodeal sculpturing and the strong dorsolateral pronotal angles of the cleptoparasite. Drawing by D. J. Brothers, from Michener, 1978b.

lioxys funeraria Smith parasitizing two different-sized species of Megachile in Michigan supports the third explanation, for on the basis of 41 loci, the genetic difference between samples of large and small Coelioxys from the two Megachile hosts could be explained entirely by sampling error (Packer et al., 1995). Presumably, there was one panmictic population of Coelioxys parasitizing the two species of Megachile.

The largest group of cleptoparasites is the Nomadinae, which, although assuredly an apid subfamily, is so different from its nonparasitic relatives that its closest relatives are unrecognized. Genera of Nomadinae are often rather host-specific (e.g., Epeolus on Colletes, Triepeolus mostly on Eucerini), but as a whole the subfamily parasitizes a wide range of bees, including Colletidae (Colletinae, Diphaglossinae), Andrenidae (all major subfamilies), Halictidae (all subfamilies), Melittidae (Melittinae, Dasy-podainae), and Apidae (Anthophorini, Eucerini, Exoma-lopsini, Tapinotaspidini).

A curious finding is that in various species of Nomada the cephalic secretions of the males are chemically simi

Figure 8-2. Bodies of females of Halictini. a,The cleptoparasitic Sphecodes monilicornis (Kirby); b, The nonparasitic Lasioglossum malachurum (Kirby); hairs are omitted on the left half of each. Note the coarse head and thoracic punctation and propodeal sculpturing and the strong dorsolateral pronotal angles of the cleptoparasite. Drawing by D. J. Brothers, from Michener, 1978b.

lar to Dufour's gland volatiles of the females of the host species, Andrena or Melitta (Tengo and Bergstrom, 1976, 1977). The species of Nomada tend to be rather host-specific, and for each host-parasite pair studied, the chemical similarities mentioned are evident. The Dufour's gland product is used to line brood cells; its odor presumably characterizes the nests. Cephalic secretions of males are similar to those of females in most species ofAndrena, but the secretions of males and females of Nomada species are quite different; it seems likely that a mimetic relationship has evolved between host or host nest odor and male parasite odor. Attempted explanations for why it is the male parasite's odor that resembles the host female or nest odor are not yet convincing. Does the female No-

Bee Dufour Gland

Figure 8-3. Bodies of females of Megachilini. a, The cleptoparasitic Coelioxys octodentata Say; b, Its host, Megachile brevis Say; hairs are omitted on the left half of each. Note the coarse punctation and pointed axillae (on each side of the scutellum) of the cleptoparasite. Drawing by D. J. Brothers.

mada learn from the male with whom she mated the odor needed to find an oviposition site? Or, if the female needs the host nest odor, perhaps to facilitate her entrance past defending owners into host nests, might she acquire the needed odor from the male at the time of copulation? Neither of these explanations seems likely!

Social parasites and cleptoparasites. The following paragraphs take up topics that relate to both types of parasitic bees.

Parasitic bees have many morphological features not or rarely found in nonparasitic bees; see the remainder of this section and see Sections 27 and 28. Social parasites have reduced scopae (Fig. 8-8). In the case of parasitic species of Bombus, this means that the corbiculae are reduced. Michener (1970) listed convergent features in socially parasitic Allodapini. Among such features is the reduction in the proboscis (Fig. 8-1); these parasites feed in the host nests rather than on flowers.

A surprising number of the cleptoparasitic bees are wasplike in appearance, partly because of reduced hairiness and loss of the scopa, features frequently enhanced by slender form, red coloration (especially of the meta-soma) or yellow-and-black wasplike coloration, as in many species of Nomada (Pl. 2). Figures 8-2, 8-3, and 8-

Figure 8-3. Bodies of females of Megachilini. a, The cleptoparasitic Coelioxys octodentata Say; b, Its host, Megachile brevis Say; hairs are omitted on the left half of each. Note the coarse punctation and pointed axillae (on each side of the scutellum) of the cleptoparasite. Drawing by D. J. Brothers.

4 show in a general way the reduction of hairiness in parasites, as contrasted with their hosts, which in the case of Figures 8-2 and 8-3 are members of the same tribes as the parasites. The loss of the pollen-carrying scopa, the most decisive morphological characteristic of parasitic bees, is illustrated in Figure 8-5, which compares the hind leg of a cleptoparasitic Sphecodes with that of an ordinary, nonparasitic bee of the same tribe, the Halictini. As is often seen among cleptoparasites, the outer surface of the tibia has coarse setae or spines that probably help the parasite to push through a burrow, against an opposing host bee, and the basitarsus has lost not only scopal hairs but also the apical process and brush (penicillus) used by nest-making halictids in spreading secreted material on cell walls. Among cleptoparasites, various degrees of scopal reduction can be found. For comparison with Figure 85, Figure 8-6 shows degrees of reduction of the femoral and tibial scopa of other cleptoparasitic Halictini. In the

Diagram Tetrapedia

Figure 8-4. Bodies of females of Apidae. a, A cleptoparasite in the Nomadinae, Triepeolus concavus (Cresson); b, Its host in the Ap-inae, Svastra obliqua (Say). Hairs are omitted on the left half of each. Note the large, angular axillae at the sides of the scutellum of the cleptoparasite. Drawings by D. J. Brothers.

cleptoparasitic Megachilidae and Apidae, scopal reduction similar to that of Halictini is found; in the Nomad-inae and some Megachilidae (Fig. 8-7), the reduction is complete. Partial reduction in the parasitic Allodapini (Apidae) is shown in Figure 8-8.

Other structures of female bees that are commonly reduced in cleptoparasites include the pygidial plate. Figure 8-9 shows that of the female of an ordinary species of Halictini, and the reduced plates of two cleptoparasitic species. The pygidial plate's usual function is probably the tamping of cell surfaces; cleptoparasites do not make cells and have reduced plates. Similarly, the basitibial plate, which probably helps nest-making bees to brace themselves while digging or tamping, is reduced in cleptopar-asites (Figs. 8-9, 66-17d). In the Halictinae, the labrum of females ordinarily has an apical process with a strong median keel of unknown function. The process becomes broad and flat in cleptoparasites, thus more like the labrum of males (Fig. 8-9), or at least the keel is lost (Fig. 66-17c).

The apex of the female metasoma of cleptoparasitic forms is frequently modified for the placement of the eggs, but this is not the case for cleptoparasites whose

Figure 8-4. Bodies of females of Apidae. a, A cleptoparasite in the Nomadinae, Triepeolus concavus (Cresson); b, Its host in the Ap-inae, Svastra obliqua (Say). Hairs are omitted on the left half of each. Note the large, angular axillae at the sides of the scutellum of the cleptoparasite. Drawings by D. J. Brothers.

adults destroy and replace the host egg (ad in Table 8-2), as shown by Figure 8-10a, b, which illustrates nearly identical apical sterna of nonparasitic and cleptoparasitic Hal-ictini. Cleptoparasites that insert their eggs into cells or cell walls instead of replacing the host egg (lo and lm in Table 8-2) commonly have apical modifications, as shown in Figures 8-3, 8-4, 8-10d, f, 84-6b-e, 91-1, 101-2a-c, and 104-4. Even in social parasites, whose egg-laying should differ little if at all from that of the host, the apex of the metasoma is more pointed in the subgenus Psithyrus than in other Bombus, and the last tergum is perhaps more often scoop-shaped in parasitic Allodapini than in nonparasitic species.

Parasitic taxa, especially those in the Apidae, are structurally very different from one another and from their probable nonparasitic antecedents. The result, in the No-

Figure 8-5. Hind legs of female halictine bees. a, The clep-toparasitic Sphecodes monil-icornis (Kirby), showing the lack of a scopa and penicillus and the presence of large tibial spicules; b, Lasioglossum malachurum (Kirby), showing the strong scopa from the trochanter to the basitarsus and the distal process and penicillus on the basitarsus. Drawing by D. J. Brothers, from Michener, 1978b.

Figure 8-5. Hind legs of female halictine bees. a, The clep-toparasitic Sphecodes monil-icornis (Kirby), showing the lack of a scopa and penicillus and the presence of large tibial spicules; b, Lasioglossum malachurum (Kirby), showing the strong scopa from the trochanter to the basitarsus and the distal process and penicillus on the basitarsus. Drawing by D. J. Brothers, from Michener, 1978b.

Figure 8-6. Posterior femora and tibiae of females of clep-toparasitic halictids, showing scopal reduction as compared to Lasioglossum malachurum (Fig. 8-5b). a, Lasioglossum (Dialictus) asteris (Mitchell), a parasitic species; b, Echthralic-tus extraordinarius (Kohl), a parasitic relative of Homalictus. Drawings by M. McCoy, from Michener, 1978b.

Figure 8-7. Side views of metasoma of female Megachili-nae. a, The cleptoparasite Coelioxys octodentata Say; b, Its host, Megachile brevis Say. Note the hairiness and especially the ventral scopa of the latter. Drawings by D. J. Brothers.

Megachile Ventral

Figure 8-8. Scopal reduction in parasitic Apidae. a, Hind leg of a cleptoparasitic species of No-madinae, Triepeolus concavus (Cresson), showing complete lack of scopal hairs; b, c, Hind tibiae of the nonparasitic species of Allodapini, Braunsapis simillima (Smith), with sparse but functional scopal hairs, and of the social parasite B. breviceps (Cockerell), with the scopa reduced. a, Drawing by D. J. Brothers; b, c, from Reyes, 1991b.

Figure 8-8. Scopal reduction in parasitic Apidae. a, Hind leg of a cleptoparasitic species of No-madinae, Triepeolus concavus (Cresson), showing complete lack of scopal hairs; b, c, Hind tibiae of the nonparasitic species of Allodapini, Braunsapis simillima (Smith), with sparse but functional scopal hairs, and of the social parasite B. breviceps (Cockerell), with the scopa reduced. a, Drawing by D. J. Brothers; b, c, from Reyes, 1991b.

madinae and Apinae, is recognition of numerous parasitic tribes. Their morphological diversity relative to probable ancestral taxa (when recognized) leads to the theory that the morphological features of parasitic taxa evolve relatively rapidly during and after the acquisition of obligatory parasitic behavior. Support for this idea comes from Echthralictus, a parasitic derivative of Homa-lictus (Halictinae). Echthralictus is found only in Samoa, where it must have evolved from its presumed hosts, the local Homalictus, during the relatively short history of Samoa, probably less than 2.6 million years. Yet Echthral-ictus differs in numerous morphological features from related halictids (see Sec. 66). If molecular characteristics evolved at a more uniform rate than morphological characteristics, molecular studies should indicate relationships of the parasitic taxa more clearly.

In females of nonparasitic bees there must be strong selection for the maintenance of such nest-making and food-collecting structures as characteristic mandibles, other mouthparts, basitibial plate, pygidial fimbria and plate, pollen-manipulating brushes, and the scopa. But

Figure 8-9. Structures of female halictine bees, nonparasitic species in lefthand column, otherwise cleptoparasitic forms. a-c, Pygidial plates. a, Lasioglossum (Evylaeus) malachurum(Kirby); b, Lasioglossum (Dialictus) asteris (Mitchell), a member of the cleptoparasitic group formerly called Paralictus; c, Ptilocleptis polybioides Michener. d-f, Labra. d, L. malachurum (Kirby); e, Sphecodes monil-icornis (Kirby); f, Ptilocleptis tomentosa Michener. g-j, Basitibial plates. g, L. malachurum (Kirby); h, Echthralictus extraordinarius (Kohl); i, S. monilicornis (Kirby); j, S. chilensis Spinola. Drawings by M. McCoy, from Michener, 1978b.

once parasitism is established, such selection pressure vanishes and these structures are reduced or eliminated, just as eyes are reduced or lost in cave animals (see Sec. 27). Advantages probably accrue when nutrients and genetic machinery are not devoted to unneeded structures and associated behavior. At the same time, novel structures characteristic of parasites frequently develop, e.g., a strong cuticle and spines, lamellae, or carinae that probably protect the neck and petiolar regions (see Sec. 28). Thus there is a tendency for parasitic bees to be well defended against the jaws and perhaps stings of irate hosts. A possibly alternative strategy is seen in the cleptoparasite Osiris, which has relatively thin but smooth and shiny cuticle and no protective spines or carinae, but an enormous sting. Parasites commonly have stronger stings than their nonparasitic relatives, although, curiously, the cleptopar-asitic megachilid tribe Dioxyini has the most reduced sting of any bee. The tendency for structures of female parasitic bees to resemble structures of males was emphasized by Wcislo (1999b).

The number of mature oocytes (i.e., eggs nearly ready

Figure 8-10. Sixth sterna of certain female bees, showing modifications related to egg placement by cleptoparasites that insert their eggs into cell walls. a, Lasioglossum malachurum (Kirby), nonparasitic; b, Sphecodes monilicor-nis (Kirby), a cleptoparasite that does not insert eggs but merely replaces the host egg; c, Megachile brevis Say, non-parasitic; d, Coelioxysocto-dentata Say, cleptoparasitic; e, Svastra obliqua (Say), nonparasitic; f, Triepeolus con-cavus (Cresson), cleptoparasitic (see also Fig. 8-4a ). Species shown in e and f are in different subfamilies; nonetheless, e probably represents the type of sternum from which f evolved through various lessextreme steps found in the No-madinae. Drawings by D. J. Brothers.

Megachile Brevis

to be laid) in the ovaries in parasitic bees is frequently greater than the number in nonparasitic relatives, a convergent feature among parasitic bees. Presumably, a female solitary bee, which has to construct a cell and provision it before laying an egg, needs to have only one egg ready to be laid at a time. A parasite, by contrast, might find several cells in succession ready to receive its eggs, and therefore needs to have several eggs ready at any time. Some parasites simply have more than one mature oocyte per ovariole, whereas others have more than the usual number of ovarioles per ovary, the usual number being three for most bees, four for Apidae. In the whole subfamily Nomadinae, the number of ovarioles per ovary is variable but greater than four, most often five but up to ten (Alexander, 1996) and 17 in Rhopalolemma (Rozen, Roig-Alsina, and Alexander, 1997). In Ericrocis but not in other Ericrocidini dissected, the number is five instead of four. In the social parasite Bombus (Psithyrus) the number ranges from 6 to 18. In other parasitic bees that have been dissected, the number of ovarioles is the same as in the related nonparasites, i.e., three or four (Alexander and Rozen, 1987).

It is reasonably clear that in the Northern Hemisphere the percentage of parasitic species in a bee fauna tends to increase with distance from the Equator (Wcislo, 1987; Petanidou, Ellis, and Ellis-Adam, 1995). In northern Ellesmere Island there are two species of bees (Bombus), one of which is a social parasite in nests of the other, but the trend is more reliably indicated in the larger faunas of middle and tropical latitudes. Possible explanations for this trend in percentages of parasitic species include (1) the synchronization caused by cool seasonal climates, such that many host nests are in the right condition for attack at the same time, (2) the competition for nest sites when synchronization is intense, some individuals thus being unable to find their own good sites, and (3) the short summer at high latitudes, leaving an individual that is delayed in nesting unable to produce offspring ready to overwinter, either because of the seasonal lack of the right flowers to provide food or because of autumnal cold weather. A delayed individual thus might profit from laying eggs in an already established nest. Obviously, these ideas are not mutually exclusive. Petanidou, Ellis, and Ellis-Adam (1995) believed that the unpredictability of the xeric warm-temperate Mediterranean climates contributed to their reduced percentage of parasitic species relative to percentages in cooler and less xeric areas. Another possibility is that the whole pattern is simply a result of the abundance of Andrena and its cleptoparasites in the genus Nomada,, and of Bombus and its congeneric social parasites, in the cooler latitudes of the Northern Hemisphere, and that general ecological explanations such as are enumerated above are only marginally relevant, or not relevant at all. It remains to be determined whether percentages of parasitic species increase with latitude in bee faunas of southern Africa and South America. In Australia there are but few parasitic forms at any latitude.

Was this article helpful?

0 0

Post a comment