The oldest known fossil Spheciformes date from the early Cretaceous. (For relevant geological terminology, see Table 23-1.) Lomholdt (1982) speculated that they might be pemphredonine wasps, although Rasnitsyn (1980) suggested that two of the genera should be in the Ampulicidae. Lomholdt also speculated that sphecoid wasps must have existed in the late Jurassic.
Bees as we know them are dependent on products of angiosperm flowers (nectar, pollen, sometimes oil) for food. The group therefore is usually believed to have arisen at the same time as, or after, the angiosperms. Various authors (e.g., Baker and Hurd, 1968) have noted that the more primitive groups of angiosperms (e.g., Magno-liaceae) are largely pollinated by beetles. Hence it seemed likely that bees arose or at least became common with the subsequent evolution of angiosperms. In their review of angiosperm biogeography, Raven and Axelrod (1974) indicated that angiosperm fossils first appear in middle Early Cretaceous, that by early Late Cretaceous angiosperm pollen (Muller, 1970) is becoming more abun-
Table 23-1. Geological Time Scale and Approximate Age in Millions of Years to Base of Each Unit.
RECENT AND TERTIARY Recent and Pleistocene Pliocene Miocene Oligocene
Eocene Late Middle Early Paleocene CRETACEOUS Late Cretaceous Maastrichtian Campanian Santonian Coriacian Turonian Cenomanian Early Cretaceous Albian Aptian
Barremian-Hauterivian Valanginian-Berriasian JURASSIC
Late Jurassic (Maim) Middle Jurassic (Dogger) Early Jurassic (Lias) TRIASSIC
Notes relevant myBPa to text
30 Dominican amber, Florissant shale
38 Baltic amber
83 Trigona prisca
112 diverse angiosperms
124 sparse angiosperms
245 S. Hasiotis' cells a Million years before present; from Taylor and Taylor, 1993. 100
dant than spores of ferns and pollen of gymnosperms, and that by the end of the Cretaceous there was much diversity among angiosperms. Some plant families present at that time (e.g., Myrtaceae and Aquifoliaceae, genus //ex) are now visited extensively by bees, and it is probable that bees were present and perhaps abundant at that time (the Maastrichtian). More recent studies (Crane and Heren-deen, 1996; Crepet, 1996) established the abundance and diversity of angiosperms earlier in the Cretaceous with abundant floral remains of various angiosperm families in early to middle Albian floras (about 110 myBP) and with sparse angiosperm remains in early to middle Aptian.
Among the Albian floral structures recently recognized are some that suggest modern bee-pollinated flowers, according to Crepet (1996); it is likely that early bees utilized these flowers, and that the origin of bees was not delayed until the appearance of known bee-pollinated families of plants. Bee-pollinated plants tend to produce limited quantities of sticky pollen that does not blow extensively in the wind. It may be, therefore, that bee-pollinated plants became abundant, especially in dry areas, well before their pollens appeared abundantly in the fossil record. For these reasons one can postulate that bees arose before the middle Cretaceous. An important review of early bee and angiosperm evolution was by Grimaldi (1999); see also Engel (2000b), Grimaldi and Engel (2005), and Danforth et al. (2006). Engel (2004b) gave a necessarily speculative account of the geological history ofbees. For example, he said that the end-of-Mesozoic extinction had little effect on bees; perhaps he was right, but only one fossil specimen (Cretotrigona) is known from before that event.
Although the use of gymnosperm pollen by bees is not now usual, the possibility exists of pollen sources exploited by bees even before the Cretaceous. Some members of the Triassic to Cretaceous gymnosperm group Bennettitales had bracts, quite possibly colored and petallike, around their reproductive structures. Probably the bracts served, along with rewards such as pollen and perhaps nectar, to attract pollinators. Is there a possibility that ancestral bees could have been among such pollinators? If so, they radiated with the advent of abundant an-giosperms, but bees did not have to arise at so late a time.
The possible importance of gymnosperms for early bees is also suggested by the observations of Ornduff (1991) on the male cones of an Australian cycad that produces abundant pollen as well as a distinct fruity odor. Workers of Trigona carbonaria Smith collect the pollen. Ornduff notes that cycads may have arisen in the Paleozoic.
The oldest fossil bee and the only known Mesozoic bee fossil is a worker of the late Cretaceous Cretotrigona prisca (Michener and Grimaldi) from New Jersey amber, believed to be nearly 80 million years old. A cloud hangs over the dating, for the piece of amber was collected long ago and its label could be wrong; moreover, the bee closely resembles a modern genus of highly derived bees and is associated with certain insect fossils expected to be Tertiary. Analyses of the amber, however, show that it is similar to known Cretaceous amber from New Jersey. Relevant papers are those of Michener and Grimaldi (1988a, b), Grimaldi, Beck, and Boon (1989), and Rasnitsyn and Michener (1991). Rejection of the dating, as is done in the English summary of Radchenko and Pesenko (1994), is not appropriate. If the dating is correct, it strengthens the case for an origin of bees before the angiosperms became common, because the Meliponini to which it belongs contains extremely unwasplike insects. It was highly social; the fossil is a worker, to judge by its very small metasoma.
The next significant date for fossil bees is the late Eocene, because of fossils in the Baltic amber, some 3540 myBP. For a review of this material, amounting to 36 species, none ofwhich belongs to an extant genus, see Engel (2001b). The following remarks are largely based on that work. The few S-T bees in the Baltic amber are one halictid and two melitted-like species, a possible third being Salt's "Andrena" wrisleyi (see also Michener, 2000). (Other unplaced names, based on badly preserved fossils or inadequate descriptions, may represent other melittid-like forms.) The 34 remaining species were L-T bees, Megachilidae and Apidae. Thus by the late Eocene there had been substantial radiation of L-T bees, a great part of which is indicated by 21 species belonging to four tribes (three extinct) of corbiculate Apidae recognized in the Baltic amber. The scarcity of S-T bees except melitted-like forms, including complete absence of the Andrenidae and Colletidae, is in sharp contrast to later fossil material and to the Recent fauna. Thus the known fossil record is consistent with an early development of the melittid/L-T bee clade relative to other S-T bees. In further support of the antiquity of this clade is the widely disjunct distribution of some of its basal members among the dasypo-daine melittids, as indicated in Sections 26 and 70 and Michener (1981a). Moreover, the African-Chilean desert distribution of the Fideliini and the pantropical distribution of the Meliponini, both L-T tribes, suggest great antiquity. However, Danforth et al. (2004), using nu-cleotide date sets from a nuclear gene, estimated the divergence of halictid subfamilies as in the Cretaceous. If accurate, this is not necessarily in conflict with an early Cretaceous origin of bees; the Cretaceous was a very long period (Table 23-1).
A possible explanation for the prevalance of L-T bees in the Baltic Amber is that such bees are the ones that often use resin in nest construction, and so sometimes become entrapped and thus preserved in amber. However, later amber fossils do not lack the major families of S-T bees. By Oligomiocene times, 20-30 myBP, the bee fauna had a more modern aspect. It appears that after the late Eocene there must have been extinction of many taxa including most tribes of the rich Eocene fauna of corbicu-late Apidae, perhaps surviving since the Cretaceous. It is likely that Engel (2004b) was right in associating this extinction with the global cooling that characterized the Eocene-Oligocene border. There appears to have been a second radiation, including the major S-T families, to produce the Oligomiocene to Recent fauna. A sample of bees, from the Oligocene of Florissant, Colorado, shows S-T and L-T bees nearly equally abundant, with about 13 species of S-T bees plus one melittid, and 16 L-T bees. Another sample, possibly of middle Miocene age although commonly attributed to the Oligomiocene, unfortunately consists offew species but is more nearly comparable to the Baltic amber because the fossils are also in amber; it is from the Dominican Republic. It includes three meliponine species (L-T) and seven species of S-T bees: five Halictinae, one Colletidae (Xeromelissinae), and one Andrenidae (Panurginae) (Michener and Poinar, 1997; Engel, 1997b). Here, too, unlike the Eocene, S-T families are well represented, as they are in the Florissant and Recent faunas. It is reasonably clear that in the late Eocene, S-T families other than melittids were scarce or absent, whereas probable melittids and L-T bees were rather diverse. By the Oligomiocene, however, the principal S-T families were all present.
Evidence from fossil bee nests may conflict with that from fossil bees themselves. The South American ichno-genus Uruguay is based on clusters ofbee cells in paleosoil, separated from the surrounding matrix by a space (see photographs by Genise and Bown, 1996). The spiral cell closures and form of the cells and clusters are extremely similar to those of certain Augochlorini such as Pseudau-gochlora (Halictidae). They are dated as latest Cretaceous or early Tertiary, and thus should antedate the late Eocene Baltic amber, which contains only one recognizable hal-ictid, as noted above. This may mean that halictids were present at that time, at least in South America, or that some other bees, probably now extinct, made cell clusters similar to those of some Recent halictids. Such an explanation is not improbable, since subterranean cell clusters of a more or less similar nature appear to have arisen in species of Halictini, Augochlorini, Nomiinae, and even one subgenus (Proxylocopa) in the Xylocopini.
S. Hasiotis and associates (for references, see Engel, 2001b) reported brood cells, some supposedly of social bees, in Triassic paleosoils and petrified wood. These ich-nofossils were not made by bees. The order Hymonoptera does not appear in the fossil record until the Triassic, when Xyelidae in the suborder Symphyta are found. The earliest fossils of Aculeata are in the upper Jurassic; the earliest members of the Apoidea, in this case Spheciform wasps, in the Early Cretaceous. Bees should have arisen at that time or after that. But even if all these dates are meaningless, Hasiotis' cells lack features that identify them as made by bees. None have the spiral pattern in the cell closure that characterizes bee cells made of soil or wood particles.
Was this article helpful?