Reduction or Loss of Structures

Systematists devote a great deal of time and energy to finding shared characters that have not evolved conver-gently and that therefore characterize taxa and are useful in developing phylogenetic hypotheses. Some of the most interesting characters that organisms have, however, are those that are convergent and that therefore suggest some common behavior or common environmental challenge perhaps working through behavior. The following are some examples of characters that have arisen two or more times in the course of bee evolution, assuming our phy-logenetic and classificatory ideas to be reasonably correct. There are many more such characters; many were listed by Michener (1944), who recorded, in a section on comparative morphology, taxa that exhibit common features. Those selected for mention here are particularly prominent or functionally interesting.

Losses of structures are the commonest and often the most easily understood convergences. In phylogenetic studies, synapomorphies based on losses must always seem weak compared to those based on novel structures, because losses are frequent and often repetitive. Independent losses of the same structures usually have morphologically identical outcomes, but the losses are not homologous and therefore can be misleading for phylo-genetic analyses requiring recognition ofhomologies. It is not legitimate in an analysis based on parsimony to code losses as different characters simply because, according to a phylogenetic hypothesis, they must have arisen independently in different clades. Because they might be homologous and the phylogenetic hypothesis thus wrong, one must tolerate a reduced consistency index and say that if the phylogeny is correct, the losses were independent, i.e., not homologous.

Throughout zoology, losses of unused structures or organs, like eyes of cave animals, are usual. Losses of the nest-making and pollen-collecting and pollen-manipulating structures of parasitic and robber bees are of a similar nature; see Section 8. Such bees commonly lack anterior basitarsal brushes, which are common in other bees and used in removing pollen from anthers. Scopal reduction or loss is almost universal in parasitic and robber bees (Figs. 8-5 to 8-8), regardless of the family to which they belong, for such bees do not transport pollen loads. Other losses associated with the parasitic way of life include loss or reduction of the stipital comb of L-T bees, of the ba-sitibial plate (see next paragraph), and of the pygidial and prepygidial fimbriae (see next paragraph), and reduction in the size of or rarely (as in Melanempis) loss of the jugal lobe of the hind wing (Roig-Alsina and Michener, 1993).

Some of these same losses, i.e., of basitibial plates and pygidial and prepygidial fimbriae, also occur in some nonparasitic bees such as the corbiculate Apidae, and the jugal lobe is absent in Bombini and Euglossini. Information on the possible function of the jugal lobe and why it is sometimes reduced or absent is completely lacking. The pygidial plate and basitibial plates seem ancestral for bees, at least in females, because of the presence of these structures in basal clades, and the presence of pygidial plates in some sphecoid wasps. In parasitic bees these plates tend to disappear (Fig. 8-9). Associated with these plates, especially in females, are usually the pygidial and prepy-gidial fimbriae—dense fringes of hair different from any fringes that may be present on more anterior terga. All these structures are well developed in ground-nesting bees that excavate brood cells of uniform shape and with smooth walls. The basitibial plates are used to brace the bee while she tamps the cell wall with the apex of her metasoma—i.e., with the pygidial plate and associated fimbriae. Not surprisingly, since they do not construct nests, male bees frequently have the plates and fimbriae reduced or absent. In females of diverse taxa that do not construct cells in the soil (or sometimes in rotting wood), these plates and fimbriae are commonly reduced or lacking. Examples among nonparasitic bees include nearly all Hylaeinae, Megachilinae, and corbiculate Apidae, and the genus Colletes. The Xylocopinae mostly have reduced or much modified basitibial and pygidial plates. Females of cleptoparasitic Halictidae and Apidae mostly have py-gidial plates but lack or nearly lack the fimbriae and the basitibial plates.

The following paragraphs describe a few other interesting losses that have occurred during bee evolution. The basic number of segments of the maxillary palpus is six in Hymenoptera in general and in many bees. This is anomalous, since for insects generally the basic number is five; almost certainly the so-called basal segment in Hy-menoptera is in reality the palpifer. In diverse groups of bees the number of segments is reduced to five, four, three, two, perhaps one; and in a few cases the maxillary palpi are absent, as in Oxaea, Rhathymus, Melanempis, and Pasites maculatus Jurine. Reduction from six to five, four, three, or two segments is known in the Perditini (Pa-nurginae), Megachilinae, Apinae, etc.; and to three segments within the tribe Eucerini. The functional significance of this reduction or loss is unknown. In the genus Perdita, different species of which have from six to two segments, there is no evident difference in associated mouthpart structures that might suggest how the palpi function. However, except for the Perditini, reduction in the number of segments of the maxillary palpi is very rare in S-T bees but common in L-T bees. One can therefore speculate that the small divergent apical segments of the labial palpi of L-T bees may have the same (tactile?) function as the maxillary palpi. Increases above the basic number of segments are very rare and are not loss characters. Worth noting are Andrena grossella Grunwaldt, extraordinary in having nine-segmented maxillary palpi (Grunwaldt, 1976), and Xeromelissa wilmattae Cockerell, variable but with up to eight-segmented maxillary palpi.

The basic number of segments of the labial palpus is four. This number is not often reduced as it is for the maxillary palpus, but the labial palpi of Eulaema (Euglossini) and Hoplitis (Coloplitis) (Osmiini) are two-segmented, and of Efractapis (Allodapini), Neffapis (Panurginae), and Xeromelissa (Xeromelissini), three-segmented. Increase above the basic number is extremely rare, but is

Figure 27-1. Reduction of the sting; the sting apparatus is artificially extruded. a, Hesperapissp.? (Melittidae); b, Perdita albipen-nis Cresson (Panurginae); c, Melipona sp., worker (Meliponini). (s indicates the sting stylus, i.e., the fused second valvulae.) The large sclerites with setae are T6 and S6; the palplike structures with hairs are gonostyli or third valvulae. See also Figure 10-14.

known, for Leioproctus (Hexantheda) missionicus (Og-loblin) (Colletidae), six, seven, or nine segments; and for Andrena grossella Grunwaldt (Andrenidae), nine segments, the same number as for its maxillary palpi (Grunwaldt, 1976).

Most bees have an arolium between the claws of each leg. Some bees have lost arolia (Fig. 10-10); examples include Leioproctus (Urocolletes) in the Colletidae; Oxaeinae in the Andrenidae; some species of Trachusa and Hypan-thidiod.es, as well as all species of Anthidium and allied genera in the Anthidiini; and nearly all Megachilini. In addition, Amegilla, Centris, Pachymelus (Pachymelopsis), Pachysvastra, Ptilothrix, Svastrina, Zacosmia,, and the Eu-glossini and Xylocopini in the Apidae lack arolia. Relatives in all cases have arolia. This is by no means a comprehensive list of bees that lack arolia, but it shows that their loss occurred many times and so now characterizes some species, some genera, and some tribes in higher taxa that otherwise possess arolia. No known characteristics of behavior or nest structure are associated with arolial loss.

Although aculeate Hymenoptera are especially known for the stings of females, which for bees are commonly used for defensive purposes (Fig. 10-14), the reduction or loss of stings is surprisingly common and occurs at least in Andrenidae, Megachilidae, and Apidae. In Andren-idae, stings are somewhat reduced, often lacking the valve of the first valvula (Fig. 27-1b) that pumps the venom into the wound (Ruz, 1986; Michener, 1986c). The sting is much more reduced in the Meliponini (Fig. 27-1c), quite incapable of puncturing anything, and is the most reduced of all in the Dioxyini. The last is a cleptoparasitic megachilid group. In most such parasites, however, the sting is well developed, presumably for defense against irate hosts.

Many additional repetitive, i.e., convergent, loss characters, for example of distal veins of the wings in small bees (Danforth, 1989a), are indicated elsewhere in this

Figure 27-1. Reduction of the sting; the sting apparatus is artificially extruded. a, Hesperapissp.? (Melittidae); b, Perdita albipen-nis Cresson (Panurginae); c, Melipona sp., worker (Meliponini). (s indicates the sting stylus, i.e., the fused second valvulae.) The large sclerites with setae are T6 and S6; the palplike structures with hairs are gonostyli or third valvulae. See also Figure 10-14.

book as well as by Michener (1944). There is no need to repeat here an account that could fill many pages.

It has long been believed that, once lost, the same (homologous) structure cannot reappear at a later date. Although for a phylogenetic analysis this is probably a reasonable assumption, there is increasing evidence that the genetic mechanism may persist, inactivated, for long periods and may then be reactivated, causing the reappearance of the structure. When this happens, the clade may have as a synapomorphy the potential for a character state, but not all members demonstrate this potentiality. A possible example is the loss and possible resurrection of submarginal cross veins. Certainly in two subgenera of Perdita, a genus that normally has two submarginal cells, an intercalary cell exists; it is small and triangular, thus unlike the three-celled plesiomorphic condition, but it shows that veins can split or arise de novo. In some other bees with three ordinary submarginal cells, this condition may be derived from two-celled ancestors. Some of Roig-Alsina and Michener's (1993) analyses of L-T bees indicated such development; perhaps these authors were wrong to reject such hypotheses. The resurrection of a structure is easy to understand when it disappears in only one sex (or stage), for example the male, but is retained in the female, and later in phylogeny reappears in the male. Genes for the character would be continuously present and expressed in females, and could be reactivated in males to cause reappearance of the character in that sex.

Was this article helpful?

0 0

Post a comment