Pollinators, Flowering Plants, and Conservation Biology

Pollinators, Flowering Plants, and Conservation Biology Much remains to be learned about pollinators and plants M ore than a century ago, Darwin (lg59) observed that "the number of bumblebees in any district depends in a great measure upon the number of fieldmice, which destroy their combs and nests ... che number of mice is largely Jependent, as everyone knows, on the number of cats .. .it is quite credible that the presence of a feline animal in large numbers in a district might uetermine, through the intervention first of mice and then of bees, the frequency of certain flowers in that district!" (p. -125). Dar\'{in recoglllzed the importance of interactions among organisms, specifically the role that pollinators play as links in communities. Yet, almost 140 years later, our understanding of pollination interactions is still rudimentary. The extent of dependence and linkage in pollination sysCarol Ann Kt'arns is an instructor in Environmental, Population, and Organismic Biology and assistant director of the WilJiams Village Residential Academic Program in environmental science, Campus Box 334, University of Colorado, Boulder, CO 80309. Her interests are in pollination of montane and alpine plants, plant pOPlllation genetics, and conservation biis an associology. David William Inouy~ ate professor at the University of Maryland in both the Zoology andPJant Biology departments, College Park, MD 20742. He also directs the graduate program in Snst<"linable Development and Conservation Biology at the University of Maryland and conducts long-term studies of plants and insects at the Rocky f"vlountain Biological Laboratory, Crested Entre, CO 81224. © 1997 American Institute of Rioi(]gicai Sciences. May 1997 The ultimate fate of many plants may depend on preserving their mutualistic relationships with pollinators and with the web of organisms that affect both plant and pollinator terns is currently under scrutiny as an issue affecting conservation of biodlversity (.Figure 1). Plant-pollinator mutnalisrns date hack to the C":retaceous period, when insects began to acquire food from flowers, and flowers achieved higher reproductive stlccess through the movement of pollen by insects. T 0day, at least 67% of flowering plants depend on insects for pollination cpedino 1979), and others depend on birds and mammals. For these plants to persist, pollinators are as critical as light and water (Levin 1971). However, comparatively little is known about pollinators. In addition, the pollination requirement of most wild species remains unknown: For example, the pollinators of 14 of 16 endangered plant species nrar the United States-!vlexico border have not been determined C~abhn 1996), and little or nothing is known a bout the pollination requirements of approximately one-third of the Euro- er pean Union's crop plants (Williams 1995). In this article, wc review variation in dependence by plants on pollinatOfs, the potential for a cascade of effects in a communiry resulting from loss of a plant-pollinator mutualism, current threats to plants and pollinators, and possible management issues. l\,.fuch of our discussion will focus on bees, because bee life histories arc intimately linked \-vith those of flowering plants. Bees are active, constant foragers, and they are reliable pollinators because they m ust continually forage for their own nutritional needs and those of their progeny. Approximately 30% of human food is derived from beepollinated crops (O'T oole 1993), and thousands of wild plants depend on the services of bees for seed and fruit formation. Thus, reductions in hee abundance could have serious implications for both narural and agricultural ecosystems. Although bees are major pollinators, it is estimated that less than two-hirJ~ ofthe world's bee species have been named and described (Buchman and Nabhan J 996). Even in wcll-srudied areas, such ac; western Europe, estimates of rhe Ilumber of bee species vary from 2000 to 4500 (~'ila1s 1995).l'\cverrhelcss, large declines of certain ly pes of bees ha\'e been documented 111 Germany, Austri3, Rrirain, the former Soviet Union, PolalJd, Italy, and Canada (O'Toole 1993), and in Costa Rica (Janzen 1974, Vinson ct a1. 1993). In North America, \-vhere farmers tend to be heavily depen- 297 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 Carol Ann Kearns and David William Inouye fi gure 1. A mus - co id fl y p nJlinarof (Thrh- u ps spp .) co l- lects n ectar a nd po llen from aPoten tilla gracilis flower in montanc Colo (ad o , Flies are an example u f a li t rl ek nown I.:ategor)' of flower visitors rh a t Me pro bably impo(tanl ill all maj or ecosys tems, from tropi .;a l to alpine, 298 BioScie,zC(! Vol. 47 N o. 5 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 dent on introduced do mesticated hOIlcy bet: s for c r o p po ll inatio n, honeybee numbers ha ve de clin ed by a pp rox imately 2S ()/a since 1990 (Nabh an 1996 ). In addi- pollin aro rs Or disruption of po llin ation to declining bee numb ers l other tion systems may cause red uce d seed di sruptions of natural pollin at;on sys- and fru it production and , ultima rely , rem s have been docum ented in many plant extinct ion. Any of th ese event s will a ffec t the organi sms fh at con d ifferent parts of the world ( Aj / , e~ and Feiusinger 1994 a, Buc hmann :; umt: seed s, fruits, or pl a nts o r tha l and N abha n 1996, Cox 1983, Crop- use pl ants for nes t constructi o n. Alch ollgh the loss of ind iv idu a l per and Ca ldcr 1990, Gcss and Cess 1993 , G insburg 1983, Paw n 19 93, spel.:it::s is of great public co nce rn, Ro ubik et al. 1986, Sc haffer et al. the ex tinctio n of an ecologica l process may go unnoticed for som e t ime 1979, Sugden and Pyke 1991). PoJlinator los s can a ffec t plants in becau se it is a morc subtle Ins s whose severa l ways (Tep ed ino 1979) . Loss effec ts are difficult to predict. T he of, or reduced , seed sel is t he most ult imate fate of many pla ntS may ohviou s result. In addi tion, a scar- depend o n preser ving the ir mUlll a lci ty o f pollinalOrs may a ffeCl a plant's istic reJ ado ns hi ps with poll inators mati ng system, res ultin g in the pro- a nd w ith the weh of orga ni sms tha r J uct iun of less vigorous o ffsprin g. affect both p lant and po ll ina tor. Thi s is because, in the a bse nce of po llinators, a higher perce ntage of Threats to pollinators see ds may be set throu gh self-pollin ati o n, decrea sing he teroz ygosity A number of threats to po!lillarors and increasing rh e ex pression of del- have bee n identified. These include eteri o us trairs associated with in- ha bitat a lteration . inrroducrions o f pesricide poi breeding. Smaller po ll en lo ad s can ali en po llin ato rs, ~nd also red uce the opportuni ty fo r pol- sonin g (Bond 1994 ). len co mpetitio n; different rates of po ll en tube growth rhro ugh the st yle Habitat alteration. Agricultu re , gratcan re sult in comp eti tio n, with the ing, frag ment ation of native lan dmo re vigorous poll en gra in s fertiliz- scapes, a nd development o f a reas in g the ovules. In so me species, there that once supported v,rild veger<1tion is a demonstrabl e cor re lation be- all caust: t ht: loss of nari ve foo d tween the rare of po ll en tube g ro wth plants, r endezvous pl a nts , an d nestand vigor of rhe res ulting progeny ing sites used by polLinacors. Po lli (Snow and Spira 1991 ); th us, re- nators ma y d ~ p e nd on nati ve phllus du ced polJen load s- and a conse- beca use they are not always a ble to {lu t: nt redu ction in po llen co mpeti- access food rewards from introduced t io n- could potenti a ll y de crease flowers (O T oolc 1993 ). f...bn y bees fitness , On a broad er sca le, loss of not onl y req uire large numbers o f flo wers to prov ide nec tar and po llen, hm a lso need a vari ety of Howering plants to sus ta in them throughout th e growing seaso n. Oligolectic insects, su ch as some bees and some butterfly larvae , depend on specific plants for Sl r v iv ~ l 8. nd persistence of their popul ati ons. In additi on to food requirements, po llinaring: o rga ni sms often ha ve specific nes tin g requirements. Some bee species nes t in cavities in the ground ) such as o ld ro dent burro\,,is, spaces und er roc ks, or holes ex c a ~ vated in sand o r so ft dirt. :Many other typ es of bees nest in hollow twigs. As land is developed for hus~ man a ctivity, the ava iJahili ty of twig rodent burrows, a nd suita ble nesting suhstrates typica lly decreases. Agriculture. In ag ri cul tural a rea s, large-sca le mo ooc ulture of crops re duces the amo unr o f land availabl e to support wil d vegetation. W'ith rhe increasin g mechani zation of agriculture, th e num ber and area of hedgerow s and uncultivated patches decrease, reduci ng the number o f native planes ava il a ble as poll en and nectar sources (O'T oole 199 3, Witlia m s -1_986 ). For t'X,a mple . in Grea t Britain, ar eas of low bumblebee de n~ sity correspo nd to areas of intensive agriculture. In such areas, land pr e~ viously l:onsidered of marginal valu e has been "improved" by draining grass lands and removing hedgerows; these cha nges have res ulted in decrea sed h a bitat d ive rsit y (Wil1iams 198 6). Compara ble cffe_cts have bee n repo rted fo r Germa ny (\'X' illi a ms 1986). Tn Po la nd, t he n umber of bumblebees o n cro ps is correlated with the cro p fie ld;s proximity to forests, meadows, and pastures (Williams 1986). Simil arly, in Canada, narive pollin arors h ~nre declined as alfalfa fields (.;()ver more and mo re prairies. In add it io n , in parts 01 Eu rope) the use o f insec ticides on agrj~ cultu ral la nd s has been blamed for decreasing bee ab undance (\Xljlliams 1986). Grazing. From th eir work in South Africa, Gess an d Gess (1993 ) determined that g r.lZin g li yestock alters habita r s ufficienrly to "Heer po llinato r ... Th ~y doc ume nted changes in ~\'ailbt )' of nesti ng sites, water resources, ~ lI d vegera tion that could ha ye dircC( negat ive d ft'l:ts on species diversity and po pulation sizes of May 1997 ested areas of Guanacaste Province pollinators, and the plants that rebecause it grows well and provides main may thus suffer reduced reprogood fodder for beef cattle. Each ductive success. Jennersten (19RR} year, ranchers set fire to the grass found that self-compatible Dianthus fields to clear them of trees and deltoides (Caryophyllaceae) pro~hrubs. The taller African pasture duced more seed in areas of continugrass supports a hotter fire than the om. rneadovv and forest than in patchy native grasses, and the fires often fragments separated by cultivated spread to forested reserves. These fields of barley and oats. Moreover, intense fires are strong enough to the crop fields, and the limited wild burn through the bark of most oil- spaces between them, contained few producing plant species, and those nectar plants to provision pollinatrees that du manage tu survive fires tors, whereas the continuous fields have reduced oil production. !vlore- and forest had a greater diversity of over, fire allows grasses to invade both flowering plants and anthofarther into the forest each year. The philous (flower-loving) insects. Visiloss of dead wood results in the loss of tation rates to flowers in the connesting sites for bees, and the loss of tinuous area were approximately oil plants results in decreased provi- three times higher than to flowers in siom. for bee development. Further- habitat fragments. Hand pollination more, as the forest becomes less dense, increased seed set in fragmented arground temperatures increase, becom- eas bur had no effect on seed set in ing too high for bees to develop nor- the continuous areas. In habitat fragments, seed set of flowers did not mally in any nests that remain. Habitat fragmentation. Develop- differ between natural flowers and ment can fragment natural habitats, £lowers that had been bagged isolating remnants of plant popula- throughout the reproductive period tions. Endangered plants often exist to prevent cross-pollination. This obin "ecological traps" (Stebbins 1979) servation suggests that seeds in the surrollnded by different habitats. remnant fragments were produced They may lack the genetic diversity largely through self-pollination. Therethat would allow them to colonize fore, fragmentation not only affected these different habitats. Small popu- the reproductive output but also aplations can also suffer from reduced peared to change the relative amount pollen delivery or reduced quality of of outcrossing as well. the pollen delivered. For example, Although habitat fragmentation Lamont et 31. (199.1) found that is a problem, preserving la rge tracts population fragmentation reduced of a particular vegetation type may fertility to zero in Banksia goodii not be enough to maintain pol1ina(Proteaceae). This species is a highly tor populations. Janzen and coloutcrossing species pollinated by leagues (1982) censused euglossine birds (honeyeaters) and mammals bee populations in parks and re(honey possums}. Flowers in small serves in Costa Rica and determined populations either receive fewer vis- that even within the same park, difits from pollinators or receive pollen ferent habitats vary dramatically in from sibling plants, which results in bee diversity. Many of the bee spelow seed production. Furthermore, cies travel lung distances to pollismall populations are sometimes nate plants that do not occur within bypassed by pollinators because some the habitats in which thev were coltypes of pollinators exhibit density- lected. This finding indcat~s that presdependent foraging behavior, pre- ervation of diverse patches within an ferring large floral displays to iso- area may be essential to maintain lated flowers (the Allee effect; adequate pollinator populations, Lamont et a1. 1993). Like plants, pollinators may per- Introductions of pollinators. Introsist in ecological traps as a result of ducti ons of polli nators can have borh habitat fragmentation caused by beneficial and detrimental consedevelopment and the associated loss quences, and have occurred both of nest sites and food plants. Re- intcntiona lly and accidentally. Honsearch conducted by Jennersten ey bees ha ve spread around the world, (1988) 111 Sweden illustrates that frag- ,,,,here they have become competimented regions may harbor fe\ver tors with native bees, birds, and 299 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 bees and wasps. Trampling of vegetation by livestock can directlv destroy th~ nests of ground-e~ti species and can compact the soil, constraining nest formation. Tn addition, the people who tend livestock in these areas of South Africa collect wood for fuel, thus reducing the avaiLlbility of hollow tv.rigs that provide nesting sites for some bee speCles. Grazing also affects bees by decreasing water availability. Both ground-nesting and cavity-nesting bees must collect water for use in nest construction. Most bees cannot obtain water from livestock water tanks ,vith steep sides, or even ponds without sloping edges, but need to stand at the edge of shallow water. Tampering \vith the natural water supply to provlsion cattle or produce crops often modifies water availabilitv for bees. areas, selective grazing In som~ by livestock has decreased the number of plant species (Gess and Gess 1983). One species, Galenia africana (Aizoaceae), which tends to become dominant, is not attr<lctive to pol1inators or livestock. As native plant species that were used originally as pollen and nectar sources disappear, the total number of bees that can be supported in the area decreases, in turn potentially decreasing seed set and funh~r reducing diversity. Dramatic reductions in bee numbers and species diversity have been documented in areas of the Guanacaste Province of Costa Rica that ,vere deforested to support cattle (Janzen 1974, Vinson et a1. 1993). The numher of oil-collecting hees is decreasing as the forest is destroyed (Vinson et al. 1993). Many of the large, solitary Centris bees (' Centridini; Anthophoridae) in the dry forest of Costa Rica need dead wood for nests. Female bees provision thei r nests \vith the pollen of several tree species and with oil collected largely from Byrsonima crassifolia trees (Ma Iphigiaceae). Oil collectors are major po11ioators of B. crassifoii<J (Vinsol1 et a1. 1993). Male and femaleCentris bees arc generally pollinators of different plant species, and thus many species are linked by pollination associations. African pasture grass (Hyparrhenia rufa) has been introduced to defor- other pollinators, and bumblebees have also been moved to islands and continents where they did not occur naturally. These introductions have sometimes benefited agriculture, but their consequences for native plants and animals can be deleterious. 300 BioScience Vol. 47 No. S Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 The introduction of the honeybee. The honeybee, Apis mellifera, has a positive popular image, and it has generally been considered a valuable pollinator. It is native to Europe and has been introduced throughout much of the world for crop pollination. Although honeybees are important for pollination of some crops, the contributions of wild bees on the same crops and the economic value of these other bee species are often ignored. Honeybees have been studied extensively, often at the expense of these other bees. Of the 20,000-30,000 bee species (Neff and Simpson 1993), information on all but a few species is scanty. This information imbalance is a problem because issues of pollination and biodiversity are probably tied more closely to the thousands of wild species th'an to the domesticated honeybee. In addition, evidence is mounting that honeybees can compere \vith native pollinators, disrupting natural pullination systems (see below). The introducdon of honeybees throughout the world must nece<;sarily alter resource availability for native pollinators, and the impact of honeybee introduction warrants further study, Honeybees pollinate a wide variety of plants, and they are easy to manage and transport, However, <;ince the 1980s, the number of managed honeybee colonies in the United States has declined as a result of several serious problems: the establishment of bee mites (Acarapis /'uoodi and Varroa /acobsoni), the establishment of the aggressive African honeybee (Apis melIifera scutellata,l, the reduction of available floral resources outside of agricultural areas, the use of pesticides, and the importation of honey (Torchio 1990), The resulting decline in honeybee availability has produced an increased awareness of the potential role of native bee species in crop pollination and stimulated interest in learning to manage them. lVloreover, some researchers have expressed coo- cerns about the wisdom of depending mercial operations import honeyon a single species of bee to pollinate bees in an attempt to ensure fruit set. crops (Torchio 1990) and have indi- HO\vever, honeybees are not efficated a grm.ving need to domesticate cient pollinators of blueberry, as ,"vas demonstrated one vear in New other pollinating species. Another problem with honeybees Brunswick, Canada, ~fter a poorly is that, despite thelI popular image timed pesticide application wiped as productive pollinators, they are out many native pollinators of blueinefficient pollinators of some crops berries (O'Toole 1993). To compenand many \vild plants (O'Toole sate for the loss of native pollinators 1993). Many native North Ameri- and ensure fruit set, growers imcan flowers as well as some crop ported colonies of honeybees. HowflO\·vers are not attractive to intro- ever, because honeybees cannot viduced honeybees and depend on na- brate the anthers, they are less able tive bees for pollination. Some flow- than vibratile species to access the ers are the wrong size for honeybees pollen, and they soon learn to pierce to enter. Bee species range in size holes in the back of the flowers to rob from 2 mm to 32 mm-a 3000-fold nectar. These nectar robbers never difference in body mass-and hon- contact the reproductive parts of the eybees are approximately 12 mm flowers, and fruit set inNc...v Brunswick long. A bee that is too small or too remained low that year. large for a flower will transfer few, Another crop plant that is unif any, pollen grains. Other flowers suited to pollination by honeybees is have small amounts of nectar, mak- the West Indian cherry, Malphigia ing them unattractive to honeybees, glabra (Malphigiaceae). This tree and some flowers have specialized was introduced to Hawaii as a crop mechanisms for pollen release that plant (O'Toole 1993) because its can deter honeybees.lVloreover, flow- fruit lS an excellent source of vitaers may have . unusual features that min C. I Io,"vever, fruit set in IIa,"vaii deter h-oneybees but not their "nor- was poor, even a fter colonies ofhonmal" pollinators. for example, eybees were brought in to pollinate Zygadenus elegans (Liliaceae), the the flowers. In this case, the problem death ca mas, has a nectar that is toxic was that the honeybee is not atto honeybees but is eagerly t:ollected tracted to the oil that fhnvers of by native bees (Tepedino 1979). the West Indian cherry, like many Honeybees are also unable to pol- plants in the Malphigiaceae, prolinate those plants that rely on "huzl'. duce rather than nectar. In its napollination." Such plants, including tive habitat, the oil is collected by several of economic importance, Centris bees that simultaneously have poricidal anthers (anthers that serve as pollinators. release pollen from a small hole at The introduction of honeybees the tip rather than splitting open to thus appears to have had a variety release pollen) from which pollen is of consequences, both unintended released only when tbe anther is and intended. The large numher of vibrated at a characteristic fre- honeybees in some habitats may quency. These flo,\vers are pollinated make them important competitors most efficiently by vibratile pollina- for a great diversity of native tors. Several groups of bees and a flower visitors. 1£ native pollinafew fly species can buzz pollinate tors decline in number, honeybees flowers, but this behavior is not in may not visit or pollinate the same the honeybees' repertoire. Buzz-pol- range of flowers, hecause of their linated crops include cranberries, nectar chemistry, small nectar or blueberries, tomatoes, eggplant, and pollen rewards, flowering phenolpeppers. Some of these plants do not ogy, or floral morphology. Alproduce ne(tar and are attractive though honeybees can be useful in only to bees foraging for pollen. pollination of some crops, native Others, like hlueherry, produce nec- insect~ can often do as well or tar as well as pollen and attract a better if they are present in suffidiverse group of pollinators, includ- cient numbers. ing honeybees. A small blueberry Domestication of bees other than stand can be pollinated effectively the honeybee. Farmers are beginby native pollinators, but large com- ning to appreciate that certain types May 1997 tablishment of feral colonies of alien species {e.g., in Japan, \vhere the European species Bombus terreslris has recently been found in the wild).1 Competition among honeybees and native pollinators. Several studies have indicated that introuut:eu honeyhees decrease the foraging success of native pollinators by Outcompeting them for resources (Ginsburg 1983, Paron 1985, 1993, Pyke and Balzar 19R5, Roubik et al. 19B6, Schaffer et al. 1979, Sugden and Pyke 1991). A case in point is provided by honeybees in Australia. Honeybees \vere introduced to Australia approximately 150 years ago, and until recently they were considered beneficial to the native flora. However, new evidence indicates that honevbees may actuallv be harmful to the· native fiora: they may displace native pollinators, they may be ineffective at pollinating native flowers, and they may interact in complex ways with native pollinators to reduce che amount and effkiency of pollen transfer (Paton 1993). Evidence of the deuimencal effects of honeybees on a native bee comes from studies in Kosciusko ~ationl Park, Australia (Pyke and Ba!zar 19S5, Sugdcn and Pykc 1991). Both the honeybee and the native bee, Exoneura asimiliima (Anthophoridae), are highly polylectic (use a wide variety of flowers as food source~), and there is a good deal of overlap in the kinds of flowers that they use. To test for competition among honeybees and the native generalist bee, colonies of honeybees were introduced to alpine areas of the park. After six 1110nths, experimental areas contained significantly fewer adult Exoneura, although the number of preauult Exoneura bees within colonies was the same in experimental areas and in the undisturbed areas that lacked honeybees. Thus, adult Exoneura either died or left the area, presumably as a result of competition with honeybees. These findings imply that competition disrupts the natural composition of the community and is likely to have effects on the reproduction of the native vegetation. IM. Ono, 1997, personal communication. T;lmagawa University, Tokyo, Japan. IIoneybees can also compete with bird pollinators. In communities in Australia where honeybees thrive, Paton (1985) found that honeybees remove one-third to one-half of the available nect;.) r and compete directly with honeyeaters, a group of native pollinating birds. Such areas have reduced honeyeater species diversity and lower floral visitation rate~ by honeyeaters. Presumably to compensate for the diminished food supply, the honeveaters increase their territory siz~. The competition with honeyeaters lowers the pollination success of at least one plant, the largely self-incompati.ble Callistemon rugulosus (1vlyrtaceae), which both honeyeaters and honeybees visit (Paton 1993). Honeyeaters contact stigmas on more than 50% of visits to the flowers, whereas nectar-gathering honeybees do so on only 4.4% of their visits, and pollen-collecting honeybees do so on 16.7% of their visits. Individual honeybees were watched for 9.9 hours; and not a single between-plant movement was recorded, compared with 7.3 inrerplant movements per hour for honeyeaters (PatorI 1993). These differences have consequences for fruit production. Plants enclosed in mesh cage~ (and which therefore receive only honeybee visits) have significantly ]o\ver fruit set than those serviced by birds as well as bees (Paton 1985). Introduced honeybees in Australia also affect beetle pollinators. Many species of flower-visiting jewel beetles (Buprestidae) in southwestern Australia ha ve suffered from the presence of honeybees. Whereas ten species of beetles could once be observed on a single flo-'wering plant, now only a few may be found (Knowles 1981/4). Since these findings, some reserves have been set aside to help protect the beetles. Possible effects of introduced honeybees on native bees can be seen in other continents as well. For example, Aizen and Fcinsinger (1.9.94a) found an inverse relationship between the number of native bees and the llllmQCr of honeybees visiting two polyphilic trees (i.e., trees attracting a wide range of flower visitors, such as flies, bees, butterflies, wasps, and beetles) in ,>ubtropical dry forests of Argentina. Honeybees were more prevalent in fragmented 301 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 of bees may be more effective than honeybees at pollinating specific crops, and some farmers are learning to encourage and manage other alien hee specieI'>. For example, honeybees are poor pollinators of alfalfa because alfalfa flowers must be tripped mechanically to discharge their pollen on an insect visitor. After a few experiences of having pollen slapped onto their undersides, honeyhees avoid tripping the flowers and instead Jearn to pierce holes in the backs of the flowers to rob them of nectar. However, cultivation of alternative bee species, such d'> Megachile rotundata (A.fegachilidae) and Nomia melanderi (Halictidae), has resulted in succe'>sful pollination of alfalfa (Bohart 1972, O'Toole 1993), N, me/ander; is a gregarious ground-nesting bee that is native to the western United States. Its management involves providing "bee beds"; areas of soil within and adjacent to alfalfa fields that have proper levels of moisture and <;alinitv for nesting. M. rotundata is a Eurasi~n species that nests in premade cavities in wood. Beekeepers provideM. rotundata with "bee boards": they predrill holes of the proper diameter into boards that are mounted on stands with wheels. The beekeepers can move these to sites in the field where pollinators are needed. Several other Eurasian pollinators have been suggested for use on alfalfa, a plant native co that region. Bumblebees have also become pollinators of some crops in regions where they have recently been introduced. In Ne"v Zealand, introduced red clover forage is pollinated by three of the four imported Emopean bumblebee species that have become naturalized to the region. The native hees arc all short-tongued species that are ineffective pollinators of this crop (O'Toole 1993). Bumblebee culture on a commercial scale has only recently become feasible. Although bumblebees are more expensive-to culture than honeybees, it is still economically feasible to domesticate bumblebees for use in greenhouse culture of tomatoes, which are buzz pollinated (Thomson 1993). Domesticated bumblebees are also now being used for greenhouse pollination of stra\vberries, peppers, eggplant, and squash (GriHiths and Roberts 1996); an unintended but predictable consequence has been the es- 302 search for food a nd successfully recruitcolony members (Roubik 1980). Ho\vever, evidence of superior competitive abilities of Africanized bees under natural conditions remains equivocal. Total nectar and pollen collections by native bees do not diminish in the presence of Africanized bees, indicating that native hees may be capable of switching host plants in response to competition (Roubik et aL 1986). Even in cases where there are clear indications that honeybees interfere with the foraging of native pollinators, their effect on the population sizes of native hees is still ambiguous (Aizen and Feinsinger 1994a, Roubik et al. 1986). Pesticides. Pesticides are another major problem for pollinators. Chemicals applied to crop plants and to rangelands can cause high bee mortality. In the United States, pesticide use has created local problems since the late 1800s, but the problems increased dramatically after World War 11, when there was a su bstanrial increase in the use of pesticides on crops, range lands, and forests (Johansen 1977). Herbicides have also been applied extensively to control weeds in crops and along roads ides, thus reducing the availability of the native \-vild plams that provide food for pollinators. Foraging on pesticide-treated plants is a major source of bee mortality, yet honeybees are often expected to pollinate crops that have been treated with pesticides. The susceptibility of bees to chemical poisoning is llSually related to their surface area-to-volume ratio. Bumblebees are often more tolerant of pesticides than honeybees because of rheir smaller surface area-to-volume ratio. and honeybees are 111 turn more rolerant than most small native bees. Chemical poisoning results in abnormal communication dances and mistakes in indicating distance and direction to food sources, in addition to direct mortality. A mortality rate of 100 dead bees per day in a honeybee colony is considered normal, and only the loss of more than 1000 hees per day is considered serious poisoning (honeybee colonies may contain 40,000-70,000 bees; Johansen 1977). Bees produce relatively few offspring and invest a large amount of energy in each. Lab studies of several solitary bee species from North America mdicate that under optimal conditions, that is, with" plenty of food, a bee will produce 15-20 offspring per year (Tepedino 1979). However, under natural conditions, the number of offspring is likely to be considerably lower because of predation, parasites, and competition. Due to this low fecundity, it may take several years for bee populations to recover after an area has been sprayed with pesticides (Karron 1991); recovery time is even longer if weather conditions are unfavorable or if more pesticide is applied. One source of pesticides that affects pollinators is the broad-spectrum insecticides used to control grasshoppers on rangelands in the south\,.,restern United States. When grasshoppers are dense, the rangelands are sprayed with these insecticides to save the grasses for cattle forage. The sprays kill many other insects in addition to grasshoppers, including native pollinators. The grasshopper-spraying campaigns (generally from mid-April to late lYlay) overlap the flowering period of a 11l1mber of endemic rangeland plants that grow among the grasM~ several of which are listed as endangered or threatened (Bowlin et al. 1993); they also overlap the period of emergence and active foraging of most native bee species (Peach et a1. 1993). Several rangeland plant species that are listed as endangered depend on insect pollinators for seed set (e.g., Eriogonum pelinophilum, Polygonaceae; Bowlin et al. 1993; Pediocactus silcri, Cactaceae; Peach et a1. -1993), and some of these endangered species show indications of pollinaror limitation (e.g.,Purshia subintegra, Rosaceae; Pitts et aL 1993). At sites where endangered plants are known to occur, an area with a radius of 3 miles around the patch of plants is not sprayed (Fitt,<, et al. 1993). However, the size of this area is arbitrary, and the effectiveness of the procedure is questionable. Relatively little is known about the foraging flight range of most native bees (Peach et al. 1993), and it is not clear ho\\' closely this protected area overlaps with the probable nesting site and foraging range of the BioScience Vol. 47 No. 5 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 areas, whereas wild bees were more common and more diverse in larger tracts of forest. The number of honeybees lTlcreascd with decreasing fragment size. Aizen and Feinsinger (19943) suggested that fragmentation could ultimately cause local extinction of some native bee species and replacement by a single species, the introduced honeybee. However, these researchers did not find evidence of direct interactions between native bees and honeybees, and it was not dear that floral resources were in fact limiting native bees in fragmented areas. The shift in pollinators may be more directly rdated to reduced availability of nesting sites for native bees (a result of habitat fragmentation) and to the foraging patterns of honeyhees (whose relative frequency increased with decreasing fragment size:). In the small fragments, the polyphilic tree species showed some decline in seed production, despite high levels of visitation by Africanized honeybee,'" which rnay have compensated for a decline in visits by native pollinators (Aizen and Fein,<,inger 1994h). The increase in number of visits by the honeybees was not sufficient to replace the decrease in pollen transfer hy native bees, suggesting that honeybees are poorer pollinators. Aizen and Feinsinger (1994b) concluded that monitoring pollination and seed production could provide a useful, integrated measure of the health of disturbed plant communities. In cases where honeybees do outcompete native bees,' \vh}' are honeybees more ,-.ucccssful? Honeybees appear to be adept at finding and monopolizing nectar- and pollen-rich resources, often at the expense of native bees. Although there is no indication of aggressive interactions among the honeybees and other foragers, temporal and spatial analyses indicate that honeybees dominate large, resource-rich patches of flowers (Ginsburg 1983, Schaffer et al. 1979). In the l.:ase of feral Afncanized honeybees in South America, the ability to outcompete native bees appears to rely on the capability of Africanized hees to locate and dominate food sources quickly, presumably because of the large number of workers that can me May 1997 Figure 2. A bum blebee quee n (Bo m hus apposilus ) fo ragi ng on Corydalis ((ls ea na b r(1lld eg ci in Co lo rado nea r th e Rock}' ylo untaill Bi ologica l Labora to r y. Thi s unus ual plan[ spec ies, a member of the fa mily Fumari aceae that is class ified as a sta te r<l re plant, is locally abu ndant bUl wi th a very patchy and disjunct distribution. The ro le of insects in its reprod uc ti o n i ~ still under study, but Lhe plan ts req uire insc:(;t. ...·isits t o set seed and may dcpc:n J u n twu species o f bu m t.,kbees as t hl: only pol lina to rs. Ph o to: N. Wa ser , U ni ve rsit y of Califo rni a , Ri verside, supre ~s grasshopper popu la tions is being investi gate d (Peach et al. 1994 ), but th ere may be no such alternati ve a pplica tion methods for some pesticides. Degrees of dependence on po\linalors Although we ca n doc umt m threats to poll ina to rs, pr edicting the acrual cffnt of po llin a to r loss for a partic ular pl a nt species is di ff ic ult. The signi fica nce to a p la nt o f th e loss o f its pollina to rs depend s o n whethe r the po llination rel ati o nship is fac ultati ve or o bli gate (Bond 1994 ;. Tn assessi ng the potential th reat to a given plant, une mmt a.1 50 con si der the imp ortan ce of see d produc tion in the demo grap hy of th e pl a nt. Some plants thriv e as a reslllt of vegetative reproduction a nd will therefore not be affected by po llina tor loss . Oth· ers either have large seed ba nks or are long lived and. thus , may no t be in imminent danger of extinction wh en a po Hina tord isa ppc::a rs. Rat her th a n strlt:t o ne-poJlinato r-o ne- planr rel a tionshi ps, most pla nes ha ve multiple poll ina to rs, and most pollina- tors pollinate mOre th an o ne plant species (Figure 2). Co mmunity co mposition varies among ha bita ts, a nd what appear to be spec iali zed re lationships between a pla nt species and a pollinator spec ies may c hange among communities. The p lants most at risk from loss of a pollinawr are th ()~e th a t a re dioe cio us and se lf-inco mpa tibl e. [hose that ha ve a single po llin ato r, and [hose tll a r propaga te o nl }' by seeds . Plants form a cominuum: at one end are pl a nts tha t depe-nd stri ctl y on a singlt: pollinato r fo r rep rodu ction, and at the o th er en d are plan ts that have multiple co mp ens3 ti on mechanisms to ensurt! su rvi va l an d reproduction. Bond (1994 ) developed a vulnerability index to ra nk the threat of extin ction o f a p la nr species due to loss of po llin aro r o r disperser mutu alisl11s . The in dex as sesses the vuln era bilit y of Cl sp/:'c it:s by considel·in g the fo ll owing vari a bles: the numb er o f poJl in a tor and disperser sp eci es needed, the lev tl u f vegetati ve propaga t ion, w heche r the s pecie s is se lf- c.o mpa ri b le, a nd 303 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 enda ngered pla nts' po llinators ..Many addi tion al pla nts th at are I..:andidates for listing do no t receive protection fr om spraying a nd may also be affected bv the loss of native bees. Anothe r exa mpl e of ho w pesticide ap pl icati o n can affect plant reproducti ve success through its actio ns o n po lli nato rs co mes fr o m Th omson et a /" s (1985) studies of fores ts in N ew Brunswick, Canada, th at w ere spraye d with Mataci! (amin oca rb in secti cide) to I..:ontrol spruc e bud worm, Choristoneura fumiferana Clem (Tortricidae; Lepidoptera). Na ti ve b ee,", in the families Andrenid ae, H a1i.c tidae, and Anthopho rida e a nd sy rphid fl y populations \,,'crc adve rsely a ffected by the pes tic ide; severa l insects di splayed convulsions fo ll owed by death. One o f th e na ti ve u nd e rs cor y lili es , Maianthemum canadensis. showed s ignifica nt reduc tio ns in f~ c undit y, a s did a seco nd und e rstory species, Cunm s canadens is (Co rn acea e). Comm e r cia l blu e be rry field s in N ew Br unsw ic k a lso suffered redu ced fru it set whe n adjacent fore st l an d s w ere s pra y ed \vith Fe nitr oth io n to con trol spruce budwor m (Keva n ·197 5) . During pestic id e application, honeybee colonies can be removed from field s, but na ti ve pollinators may still ht ex.posed to toxins. To minimiz e harm to these pollinators, insectici des shou ld nor be applied durin g a plant 's fl owe d n g p eriod. H o w eve r, man y na tj ve bees produu: multiple br oods per summer. and a lth o ugh proper timin g of pesti cide spr ay ing may ensure pollination of a pa rticula r flower spec ies, the net effec t o f spray in g atrer flo wer ing may be simpl y to d ec rea se t he pollinaror populati on for subsequent year (Fitts et 31. 1993). Bees ' hairs easily p ic k up insl;:(.;tici J e duston flowers and bring it back to thei r colonies (Johansen 1977), . . vhere it can kill larval and adult bees. An other way to minimize harm to pollinators is to avoid spraying pes tici des under climatic condition s that enhance roxicity. Hot days and coo l nights ca use condensation of insecticides on pla nt leaves, thus prolonging rhe effects of insecticides and killing more bees. Me thods of pesticide applica tion tha t ""void spraying flowers and fo liage a re even safer for hees. The use of ca r baTyl bran ba its to Examples of disrupted pollinatorplant interactions. The repercussions from the disruption of particular pollination systems indicate wide variation in vulnerability among different plants and communities. The following examples illustrate the variation in vulnerability among different types of plants. • Frevcinetia arborea (Pandanaceae), the ieie vine, is native to Hawaii. It produces la'rge, colorful inflorescences that arc visited by an introduced bird, the Japanese white eye (Zosterops japonica), which serves as its major pollinator. Cox (198:)) wondered what the original pollinators of this plant might have been and studied the writings of early Hawaiian ornithologists to see whether any rare or extinct endemic birds visited this plant. Finding oc..::asl0nal referencel; indicating that such visits occurred, he examined pollen loads on the feathers of museum specimens of two extinct Hawaiian birds, the Ou (Psittirostra psittacea) and Laxiodes kano (DrepaniJidae), and of the endangered Hawaiian crow (Corl/us tropicus). The two extinct species carried large amounts of Freycinetia pollen, and the endangered bird car- 304 riell moderate amounts. ::-';-one of these birds carried other species of pollen, and I;reycinetia pollen was not found on other endemic birds specimens that \vere tested. Cox (1983.) thus suggests that Freycinetia received less pollen once these three species began to decline and that the plant then reproduced largely by donal grmvth until the Japanese white eye was introduced to Hawaii in 1929. This introduced bird seems to have filled effectively the vacant niche, and once again these plants reproduce sexually. • Johnson and Bond (1993) conducted research in South African fynbos that indicates that a single species of Satyrid butterfly (Meneris tulba?,hia) is virtually the only pollinator of a guild of 14 red-flowered plants belonging to four different families (lridaceae, Orchidaceae, Am ary llidaceae, Cras s ulace ae). 1tlost of the species require poll in ators, and more than half lack compensatory mechanisms to IDsure their persistence in the absence of pollinators, although some have bulbs or produce hundreds of tiny seeds ...vhen they afe pollinated. Most of the species are rare, possibly because of this need for a specific pollinator. At least one of these species, Disa unit/ora (Orchidaceae), sets significantly less seed in habitats that the butte'rfly does not fa vor, as a direct result of low-visitation rates (Johnson and Bond 1992). • Thelymitra epipactoides is an endangered orchid found on the coastal heaths of southeastern Australia (Cropper and Calder 1990). Although it is visited by many bee species, it appears to be pollinated largely by bees of the genus Nomia (Halictidae). The flowers produce a sweet scent but do not offer nutritional rewards; thus, the relationship is one of "deceit pollination." Extant populations of the orchid produced no seeds, and few pollinia had he en removed from the flowers, although a small number of Nomia bees were observed ca rrying pollinia (Cropper and Calder 1990). The rar· ity of the plant has been blamed on orchid lovers who collect the plants and on general habitat destruction. However, Cropper and Calder (1990) suggest that habitat management and a consequent reduction in fires has contributed to the plant's rarity. The lack of fires results in a climax community with low species diversity and with a decreased total floral abundance. Because Nomia is (] polylectic species, the reduced floral abundance probahly makes the habitat less attractive to pollinatoro;, resulting in poor pollination. • A rare fynbos shrub of the African Ca pe, Ixianthes retzioides (Scrophulariaceae), grows in rapids and waterfalls, with its roots anchored in rock crevices (Stciner 199.1). Tts £lowers produce oil as a pollinator attractant. Its nearest relatives (Bowkeria and Anastrabe) also have oil flowers and are pollinated by bees of the genus Redil/iva. female bees collect the £10-ral oil to provision nests or as a precursor for a material for lining nest cells. These same bees visit other flowers for nectar. Over the course of four years, Stein er (1993) observed only a few imects visiting the flowers (a honeybee, a halictid bee, and one small unidentified beel, and fruit set was extremely low. However, when the fJO\'-/ers were hand pollinated, they set seven times as many fruits. Based on the oil composition, flower morphology, and size of the plant, Steiner (1993) suggests that it is adapted to pollination by a large, oil-collecting, specialist bee, although no oil-collecting bees have been observed visiting the flowers. An undescribed bee of the appropriate size, morphology, and phenology-probably of the genus Redh'iva-has been collected at three sites outside the known range of thE' plant. Despite this apparent "uncoupling" of plant and pollinator through loss of the local pollinator population, Ixianthes persists through vegetative propagation. Cascade effects. Tn some systems, factors that affect a pollinator have the potential to create a cascade of events affecting multiple species. The following examples illustrate this POlDt: • Male euglossine bees pollinate Illany neotropical orchids that are obligate outcrossers. The male bees have brushes on their front legs that are used to collect fragrances and a modified hind leg: for scent storage. They visit the flowers to collect scents, which appear to be important for mating success. ApproxiBioScience Vo!. 47 No. 5 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 whether the species depends on set:dling recruitment. Based on this index, as many as 50% of the 300 species of Proteaceae of the South African Cape can he clao;sified as vulnerable (Bond 1994). Most of these potentially vulnerable species produce few, large seeds, arc seed limited, and will not res pro ut after fire. Manv are dioecious. However. Bond (1994) found that many of th~ species that appear vulnerahle have compensatory mechanisms that buffer them in part from pollinator failure: The bird-pollinated species can be pollinated by heetles, and the dioecious ones can be wind pollinated. A greater risk to these plilntli is the threat imposed by the introduction of an Argentine ant (Tridomyrmex humilis) to parts of the region. Many of the species in this area depend exclusively on native ant species for seed dispersal. The introduced ants do not disperse the seeds, and as they displace native ant species, seedling recruitment decreases. Preservation of plantpollinator systems Manage me nt of pollination sy stems is a relati vel y new and umried strategy fo r t heir conservat ion. For both pl an ts and pollinacors, preservarion Ma)' 1997 Figure 3. A bumble·· bee queen (Dnmbus oed.lenta /is) visits ~ ma le catk in [0 co llect polJen. Willow fl owe rs (Sa/ ix spe ci es) are important so urces of necr:H and po ll en for fl ow e r vis itors earl v in rhe ~ w s pro sp ring. Wil vide th e poll en re qu ired for some spe cies of bumblebee ro initiate ne stS hut arc soon rep la ced ill Ih e bees ' di ets by a series of later-blo om ing species. ma y require lit tle more than protection o r ma na ge ·· ment of appropr.ia te habi t.at. For exa mpl e, lea ving unpl owed strips of lan d hetween agri cultural fields ma y encourage nes ring by native bees, such a~ bumh lebees. Similarly, fragm entation o f forests may res ult in decreased pollination and seed production (e.g., Aizen and Fein -singer 1994b), and protecrion of for est habi t;:. t may help to prevent such a dcdine . Preservation of habitat for pl a nts that provide necr.ar and po lJen throughout the ~ eason d uring whic h a poll in:1tor species is active may a lso bt:: importa nt in conserv in g pollination systems. T o ensure prov ision of a season-long suppl y o f neeta r and pollc n. kn owled ge o f t he phenology of fl ow ering is c'\Scnria l (Figure 3). In other cases, more active measu res, such as controlled burns, ma y be required to exclude wo ody pl ants and maintain communi ties of h erb aceou~ plants to provide appropria Te floral resources for pollinators. Reintrodu crions of loc ally extinct popuiations of pl ants or pollinators are also 3 possible strate gy for rtestabli shing pla nr- po llinaror relationship s. For example, an effort is und c(war in Japan to res[Q rc n:ui ve bumb le bee p o lin~[Qrs to an are a w he re seed product ion by a rar e primrose spec ies, Pri11l 1tla sieboldii, has hee ll grcHly rt!ul1ced. 2 In the absence of a native po ili nacor, it may be possible, as an interim measure, to us e hand pollination for seed production. In those c ases in which introdu ced hon ey bees ha ve outcompeted n at.ive pollinaro rs, rc sulting in reduced seed prodUl.:t ion by na tive pla nt species, then it ma y he possi hle to redu ce t he density o f the introduced bees by removi ng apiaries or tra ppin g fe ral co lon ies. The bio logy of s ma ll pop ula tio ns presents spe c ia l p robl t:: ms fur management of pla nt- p oJlinat o r systems. Small planr popular io n s may have reduced seed set bccau,. e of a lower den si ry of flow e ring planrs, genetic effecrs such as inbreeding depr es':I io n, or Ld lur e to set seed due to inco mpa ri b ili ty problems. For examp le, De .\1 a ur o (1993) found that Cl s;~ ld d um: population o f lhe ra rc species Hymenoxys acaHlis (Ast era(." elle ) was ~J . T ho mp son , 1995 , lK' r:ouJltl l co mmunica(i on. State Univcrsir r o f :-.I ~·w Yo rk , Sro.··y Br ook. NY . 305 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 marely half of the euglo ssine-on:hiJ mutllalism s invol ve a si ngle bee sp ecies, and ha lf invol ve from 2 to J 2 euglossine species per orchid spccics (Roubi k 199 2 ). In co ntrast to males, fema le euglossines norm.dly travel lo ng di sra nces along predi cra hle r outes, vis itin g wide ly di spersed woody plants that occur at low densiti es in rhl! rropica l for est. As the number of th ese trees h as decreased due ro lumb ering, grazing, and development, bee nesring sites have be~n d~sro yc d an d the total number of bees has dec reased. Female bees have sh ifred to feeding on weed)' shrubs in disturbed sites, and ir is unlikel), th at they will continue to serve as reliable outcrosse rs of distantly spaced Aowe rin g t.rees . Reduced Ilis itation by femal e bees will further d ecrease the c hances of surv ival of [he tree spec ies, and the relationship will spiral into dcclin e. Tht! fart!s of [he orchid s, rh e bees, a nd (h e nativc woody p lanrs are a ll inrimatel}' co nnected (Janze" 1974 ). e species in • Fig trees an. ~ k ~ys ton tro pical co mmunities . ror exa mple ~ in Cocha Cashu, Peru, several primate spcciL.:s, pr()f.;yonids, marsupials, touc a ns, and other birds depend on fi gs as a sfaple food (Terborgh 1986). Worldwid e, there are approxim a tely 800 species of fig, each of \-vhich depends on a unique species of w asp for po llination. The fig wasps arc tiny, approximately:1 mm in length. Asy nc hronous flowering is essenti a l for che year-round survival of t he fi g ,.vasps, whic h die if no flow ers are av a ilab le within a few da ys a fter ha rchin g (Thomson et al. 1 991). Thu s, a reduction in tree num bers cou ld negati vely a ffcct wasp population viabili t Y'; simu la tion model s s uggeST thar a minimum \, jable popu lation size of 300 trees is need ed t o m<linr<lin a wasp population (Thomson et al. 1991). If \-va sp pollillator populations are lo st, then th e food base of many sp eci es could co llapse (LaSalle and Ga uld 1991 ). References cited AlZen MA, Feinsinger 1'. 1994a. Habitat fragmentation, native ;nsect pollinators, and f~r:-t! honey bees in Argentine Chaco Serrano. Ecological Applications 4: 378392. ___ . 1994b. r orest fr.'lgl11entation, pollination, and plant reproductiun ill a chaco dry forest, Argentma. Ecology 75: 33U-351. Roh;ln GE. 1972. M,wagel11ent of wild he~ for the pollination of crops. AllIlU<l1 Review of Entomology 17: 2S7-312.. Bond W]. 1 ')94. Do mutualisms nutter? Assessing the impflct of pollinatnr and disperser disruprJoTl on pIa nt extl nction. PhIlosophical Transacrions of the Royal Society, London R Rinlogical Sciences 344: S3-.90. Bowlin WR, T ept'dino VJ, Griswold TL. 1993. The reproductive biology of ErioKol1um pelinophilum iPolygonaceae.\.l'ages 296302 in Sivinski R, Liglltfoot K, ecls. South western rare and cndangen:d plants. Nliscellaneous publteationnr 1. -"anta Fe (NM): New 11exico Forestry and Resources ConsCfvatioJI Division. Buchmann SI, Nabhan GP. 1996. The forg{)[ten pollinators. Washington (DC): Island Press. Cox PA. 1983. Extil!(;tion of the HawaIian avifauna resulted in a change of poltinators for tlu' ieie, rreyrilletia arborea. Oikos 41: 195-199. Cropper se, Caldel' DM. 1990. The floral hiology of Thelymitra epipactozdes (Ordlidaceae), and the implications of pollination by deccit on the snrvival of th is rarc orchid. Plant Systematics and Evolution 170: .11-27. 306 Darwin CR. [fl59. On the origin of species hy means of natural ~dection. London (UK): John Murray. De.Mauro MM. 1993. H.c1arionship of breedillg sptem to ranty in the lakeside daisy (Hymenuxy; acauii;' vu .?,fabra). Cor:ser vatlOn Biology 7: 221-223. Fitts RD, Tepedino VJ, Griswold Tl.. 199.1. The pollination biology of Ariwlla diffrose (Purshia submtegra'l, including a report on experimental hybridization with its sympatric congener P. strlnShUliana (Rosaceae). Pages 268-359 in Slv\TJskl R, Llghtfoot K, eds. Southwe,tern rare and endangered plants. Miscellaneous publication nr 2. Sant;') Fe (KM): New Mexico Forestry ancl Re sources ConservatJon Division. Cess FW, G<,~ SK. lY9.1. EHects of increasing la (lJ utihzatlOfI on species representa rion and diversity of aculeate wasps and bees in the semi-,uid areas of Southern Africa. Page~ 83-1 t4 in LaS,llle J, Gallld [1), eds. Hymenoptera and biodiversity. Oxon {UK): C.A.B.lnternational. GimhurgHS. 1983. Foraging ecology of bees in an old field . .l:.cology 64: 165-175. [;riHiths D, Robherts EJ. 1996. Bumble bees as pollinaton of g,lasshouse crops. Pages 33-39 in Matheson A, ed. Bumble bees for ple3sure 3nd for proflt. Cardiff (UK); Interllatiml.'11 Bee Rp~earch As;ociation. Janzm DH. 1974. The deflowering of Central America. Natural History 83: 4~-53. Janzen DH, OeVJ"les PT, Iip;gn~ M1, Kil11sey LS. 1982. Seasonal and sitt' v,lfialion in Costa Rican euglossine bees at chemical hait, in lowland deciduous and evergreen forests. Ecology 63: 6(,-74. Jcnnersten O. 1988. Pollination In Dianthus deltoides (Caryophyllaceae,l: effects of hahitat fragmenlalion on visitation and seed set. Conservation Biology 2: 359366. Johanst'n CA. 1977. l'esticides and pollinaton. Annual Review of Entomology 22: 177-192 . .TohnsOll SD, Bond WJ. 1992. IT.'Ihitat dependent pullination success in a Cape orchid. Occologia 91: 455-456. _ _ .1993. Red flowers and butterfly pollinatjon in rhe fynbos. P.'Iges 137-[48 in Arianourwu M, Crovt's RH, cd,. Plant animal interaction;, in \1editerranean-rype ecosystems. Dord[(;cht (the Nerherbnds): Kluwer Academic Publishers. Karron JD. 1991. Patterns of genetIC variation and hreeding systems in rare plant species. Pages 87-98 in Falk DE, Holsinger KE, eds. Genetics and conservation of rare plants. New York: Oxford University Press. Kevan PG. 1975. Forest application of the in" sectlcide FcnitrothiQn and Its effect on \"'ild hee pollinators (Hymenoptcra:Apoidea) of luwbush bhleberries (Vaccimum spp.) in southern .\Jew Brunswick, Canada. Biological Conservati()n 7: .101-309. KnowlesD. 1983/4. Flying jewels. Geo5:46-57. Lamont BB, Klinkhamer PGl, W"itkowski t.Tt. ] 993. Population fragmentation may ft:'duce fertility to ;-.ero in B,mksia goodiia oemomlration of the Allee eHect. Oecologla 94; 446-4.50. LaSalle.l, (;3uld ID. 199.1. I IYl11enoprera: their diversit), anJ their impact OIl the diversity of other organisms. P.~ges 1-26 in LaSalle .1, G;,wld ID, eds. Hymenoptera and hiodiversitv. Oxon (UKj: C.A.B.lnternational. . Levin DA. 1971. The origin 01 repodutlv~ isolaling mechanisms ill flowering plants. Taxon20;~1-3. Nahhan GP. 1996. The para hIe of the pOppy and the hee. Nature COilSerV<lm'y 46: 11-15. NeH JL, Simpson BB. 1993. Becs, pollination systems and phnt diversity. Pa~es 14.116fl in LaSalle J, Gauld ID, eds. Hymenoptera and hiodiversity. Oxon (UK): C.A.B. International. O'Toolc C. 1993. Diversity of native hees and agroecosystems.Pages 169-196 ill LaSalle J, Gauld lD, eds. Hyrnenoplera and biodiversity. Oxon (UK): C.A.H. International. PaLOn DC. 1985. Food ~uply, populalion structure, and behaVIOur of New Holland honeyeate r~ Phylidonyris novllelmllandiae in woodlands Ilt'ar Hur~ham, ViLtoria. Pages 222-230 in Keast t\, Recher HF, Ford H, Saunders D, eds. Birds of eucalypt forest5 and woodlands: ecology, conservation, and managemcnt. Sydney (Australia): Royal Australian Ornithologlsts Union and Surrey Beatty & Sons. 1993. HCJIlevbees in the Australian environment. BioScience 4:): 95-10.1. Peach Mt, Tepedino YJ, Alston DG, Griswold TL. 1993. Insecticide treatments for rangeland grasshoppers: potential dfects on the reproduction of Pedi()rartlfs si/eri (Englel11.) Benson (Cactaeeae). Pages 30'1-333 inSivinski R, Llghtfoot K, eds. Southwl.'stern rare and endangered plants. Miscellaneous publication nr 2. S<lflta Fe (NM): New l\fexico Forestrv ancl Re~ourcs Conservation Division. . Peach ML, Alston DG, TepeJino VJ. 1994. Bees dnd bran bait: is carb:Hyl bran bait lethal to alfalfa leafcutting b~e (Hymenoptera: Megachilidae) adnlts or larvae? Journal of Economic "Entomology 87: 311-317. I'yke GH, Balzer 1.. 19tiS. The dfem of the introduucd honey hee ':Apism{lera)onu~­ tralian nati\-~ bee,. OLca~il1 paper nr 7. Sydney (Australia): Ncw South Wales Nationall'arh Wildlife Service. Roubik DW. 1980. Foraging blCh,-,vior ofcnn1peting t\fri~2nzed honey bees and s[ingle~ bees. ELology 61: B36-84S. ___ .1992. Loost' niches in tropical cOmmunities: why ar~ there so few bees and so many trees? Pages 327-354l11 Hunter MD, Ohgushi T. Price P, eds. Effects of resource di 5trj bu tion on animal-pIa ut inreractions. Ncw York: Academic Pre~s. RoubikDW, MorenoJE, VergaraC, \Y.,'lttm:mn D. 1986. Sporadic food competition with the African honey bee: projen~d impact on tleotropical social hees . .Journ .. l of Tropical "Ecology 2: 97- [ 1 [ . Schaffer W,\l, .lensen DB, Hobbs DE, GurevitchJ, Todd J H., Schaffer .:\ItV. 1979. Competition, foraging energctics and the cost of ~oli.dty in lhre~ species of hees. Ecology 60: 976-987. Snow AA, Spira TP. 1991. Pollen vigor and the potential for sexual selection in pl.lms. Nalure 352: 796-797. Stebbins GL. 1979. Kare species as examples of plant evolnrion. Pages J 1 3-11 Rin Great Basin natuT<llist rnemoi~ nr 3: the endan- BioScience Vol. 47 No. 5 Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 effectively extinct because it contained only a single self-compatibility type a nd could no longer produce seeds. Tn this case, the population could potentially be rescued-a lheit at the cost of introducing foreign genes from another population-by importing plants of a compatible mating type. 1vlanagement of habitat to prevent fragmentation of plant populations is another means to forestall reproductive problems resulting from small population Size. Extinctions of ecological interactions may be more difficult to notice than extinctions of individual species. One of the partners may persist for several years after the other is gone because of long-liycd individuals or compensation mechanisms (J anzen 1974). Predicting the effects of loss of a particular pollinator is extremely difficult, bLlt it is important to remember that no species eXists in isolation. Each is part of an ecological web, and as we lose more and more pieces of that web, the remaining structure must eventually collapse. Support America's colleges. Because college is more than a place wnere young people are preparing for their future. It's where America is preparing for its future. If our country's gOing to get smarter, stronger-and more competitive-our colleges and universities simply must become a national priority. It's an investment we all share in. Government. Private citizens. And the business community. After all, the future of American business depends on it. So help America prepare for the future with a corporate gift to the college of your choice-and you'll know your company has done its part. Give to the college of your choice. May 1997 COUNCll 'OR AlOTOIDU(ATION ~ Downloaded from https://academic.oup.com/bioscience/article/47/5/297/222685 by guest on 12 December 2021 gert'ci spcit'~: a syUlpo~iurn; 7-8 Dl'c J 978. Provo (UT): Brigham Young University. Steiner KE. 1993. Has Ixianthes IScrophulariaceae) lost its special bee? Plant Systemat icsand Evolution 185: 7-16. Sugden EA, Pyke GII. J991. Effects of honey bees on colonies of Exoneura t1simillima, an Australian native bee. Australian Journal of Ecology 16: 171-181. Tepedino V]. 1979. The importance of bees and other insect pollinators in maintaining floral species composition. Pages 39-150 ill Gre;H flasin natIJralistmemoirs Ilr 3: the r:nciangcreJ ~pe(;is: a ~ymposiL1lJ\; 7-8 Dc.:.: 1978. Provo (LT): Brigham Young University. T erborgh]. 1986. Keystone plant resources ill the tropical forest. Pages 330-344 in SouJ6 ME, ed. Conservation biology. Sunder rand (MA): Sinauer. Thomson]D. 1993. The queen of forage and the bumblebee revolution: a conference with an attitude. Trends in Ecology & Evolution 8: 41-42. ThomsonJD, Plowright RC, Thaler CR. 1985. Maracil insecticide spraying, pollillator mortality, and plant feo::undity in New Brunswick forests. Canadian Journal of Botany 63: 2056-2061. Thomson .ID, Herre EA, Hamrick .lr., Stone JL. 1991. Genetic mosaics in strangler fig trees: implications for tropic.11 cOlserv~­ lioll. ScitTKt 254: 1214-1216. Torchio PF. 1990. Diversification of pollination strategies for U.S. crops. Environmental Entomology 19: 1649-1656. Vinson SB, Frankie GW, Barthell J. 1993. Threats to the diversity of solitary bees in a neotropical dry forest in Central America. Pages 53-82 in LaSalle J, Gauld ID, eds. Hymenoptera and biodiversity. Oxon (UK): CA.B. International. Williams CS. 1995. Conserving Europt's bees: wby all the buzz? Trends in Ecology & Fvolution 10:309-110. \~'ilaI PH. 1986. En,'iwn!Tlental (;hange and the distributions of British bumble bees (Bombus LHr.). Bee World 67: 50-61.