Hymenoptera: Megachilidae - Semantic Scholar

PLANTÐINSECT INTERACTIONS

Nest Establishment, Pollination Efficiency, and Reproductive Success of Megachile rotundata (Hymenoptera: Megachilidae) in Relation to Resource Availability in Field Enclosures THERESA L. PITTS-SINGER1

AND

JORDI BOSCH2

Environ. Entomol. 39(1): 149Ð158 (2010); DOI: 10.1603/EN09077

ABSTRACT The alfalfa leafcutting bee, Megachile rotundata (Fabricius), is used to pollinate alfalfa, Medicago sativa L., for seed production in the United States and Canada. It is difÞcult to reliably sustain commercial M. rotundata populations in the United States because of problems with disease, parasites, predators, and unexplained mortality. One possible explanation for early immature mortality is that, relative to ßoral availability, superßuous numbers of bees are released in alfalfa Þelds where resources quickly become limited. Our objective was to determine how M. rotundata density affects bee nesting, pollination efÞciency, and reproductive success. Various numbers of bees were released into enclosures on an alfalfa Þeld, but only 10 Ð90% of released female bees established nests. Therefore, a “bee density index” was derived for each enclosure from the number of established females and number of open ßowers over time. As the density index increased, signiÞcant reductions occurred in the number of pollinated ßowers, number of nests, and number of cells produced per bee, as well as the percentage of cells that produced viable prepupae by summerÕs end and the percentage that produced adult bees. The percentage of cells resulting in early brood mortality (i.e., pollen balls) signiÞcantly increased as the density index increased. We conclude that bee nest establishment, pollination efÞciency, and reproductive success are compromised when bee densities are high relative to ßoral resource availability. Open Þeld studies are needed to determine commercial bee densities that result in sustainable bee populations and adequate pollination for proÞtable alfalfa seed production. KEY WORDS Medicago sativa, alfalfa leafcutting bee, bee density, reproductive success

The alfalfa leafcutting bee, Megachile rotundata (Fabricius) (Hymenoptera: Megachilidae), is a very effective pollinator of alfalfa, Medicago sativa L. (Fabaceae). This bee has been used as a commercial pollinator since the 1960s (Stephen 1961, Hobbs 1967, Bohart 1971, Richards 1984, Frank 2003), especially for alfalfa seed production in northwestern North America. As a gregarious cavity-nesting bee, females build nests in holes in wood or polystyrene bee boards that are provided for them in domiciles placed in Þelds planted for alfalfa seed production. Female bees line nest cavities with leaf pieces cut from alfalfa or other plants creating a series of brood cells. By foraging for pollen and nectar to make nest provisions, bees pollinate (i.e., “trip”) the alfalfa ßowers. Tripping occurs when the action of the bee releases pressure on interlocking keel petals, allowing the ßowerÕs fused reproductive column to abruptly snap upward from within the keel (Frank 2003). M. rotundata populations are partially bivoltine (Krunic and Hinks 1972, Tepedino and Parker 1986, Kemp and Bosch 2001). By 1 Corresponding author: USDAÐARS Bee Biology & Systematics Laboratory, Logan, UT 84326 (e-mail: Theresa.Pitts-Singer@ars. usda.gov). 2 CREAF, Universitat Auto ` noma de Barcelona, 08193 Bellaterra, Spain.

the end of the summer, most bee larvae already have entered diapause in the prepupal stage and wait to resume development in the following summer. Other individuals (so-called “second generation bees”) skip prepupal diapause and develop to adulthood to produce a new generation before the end of the summer. For commercial management, the diapausing prepupal cells are stored at cold winter temperatures and are artiÞcially incubated early in the following summer to synchronize adult emergence with alfalfa bloom (Richards 1984). The use of ⬃75,000 Ð99,000 bees per hectare was recommended in the early 1990s for seed production on established alfalfa Þelds in the United States (Peterson et al. 1992). Currently, larger populations of M. rotundata (100,000 Ð150,000 bees/ha) are released in Þelds by U.S. alfalfa seed growers (Peterson et al. 1992, unpublished data), and the seed yields obtained may exceed 1,100 kg/ha (unpublished data). Although the highly efÞcient pollination ability of M. rotundata leads to proÞtable seed yields, U.S. production of live bee progeny is seldom self-sustaining, with often ⬍50% replacement of released bees (Pitts-Singer and James 2008). Fewer bees (61,000 Ð74,000 bees/ha) are used in Canada (Frank 2003), where bee populations are increased in alfalfa Þelds and sold to U.S. growers.

150

ENVIRONMENTAL ENTOMOLOGY

Therefore, one possible explanation for the inability to sustain commercially viable populations of M. rotundata in the United States is the inundation of Þelds with numbers of bees that are too high to be supported by the available ßoral resources (Strickler and Freitas 1999, Bosch and Kemp 2005, Maeta and Kitamura 2007). For two Þeld seasons (2003 and 2004), we performed an experiment to determine the effects of bee density (relative to ßower availability) on M. rotundata nesting, pollination efÞciency, and reproductive success. We predicted that nest establishment and individual bee reproductive success would decline as the number of foraging females increased relative to available ßoral resources but that pollination efÞciency would be greatest at higher densities before reaching a plateau. Reductions in reproductive success could be realized as an increase in nonviable brood, a more pronounced male-bias in sex ratio, a decrease in diapausing prepupae, or a failure of diapausing prepupae to emerge as adults in the next summer. An increase in pollination efÞciency could be evaluated from the percentage of ßowers that are tripped. Materials and Methods In 2003 and 2004, eight screened enclosures (length by width by height ⫽ 6.1 by 6.1 by 1.8 m) were established in an alfalfa hay Þeld (variety not speciÞed) that was allowed to ßower in North Logan, UT. In the center of each enclosure, one small portion of a commercial polystyrene bee board (range: length by height by depth ⫽ 10 by 4 by 9.2 cm to 34 by 15 by 9.2 cm) was housed in a small plywood domicile suspended ⬇1.5 m above the ground on a metal post. The number of nesting cavities in the bee boards equaled four cavities for each of the females released, ensuring that nest availability was not a limiting factor for nesting or reproduction in the enclosure. Paper straws of slightly smaller diameter than the nest cavities were inserted into each hole to facilitate removal of nests at the end of the study for x-radiographic analysis (Stephen and Undurraga 1976) and dissection of cells as necessary. Bees that had been purchased from a Canadian supplier (JWM Leafcutters, Nampa, ID) were incubated at 29⬚C in June, and they emerged from their cocoons in the laboratory after ⬃3 wk. They were released from small containers into the enclosures next to the bee boards. The number of bees released in each enclosure was varied to emulate bee densities between 9,100 Ð95,300 bees/ha to reßect a series of bee densities ranging from lower than would be used commercially up to a modest commercial level. Two male bees were released for every female to attain a 2:1 sex ratio, which is near the average sex ratio for Canadian M. rotundata commercial populations (Pitts-Singer and James 2005). Throughout the season at weekly intervals starting a few days after bee release, the following data were recorded concerning ßower production from within 12 quadrats (area of each ⫽ 0.3 m2) in each enclosure

Vol. 39, no. 1

starting at ⬇1400 hours MDT: (1) number of alfalfa ßower racemes (from budding to withering); (2) number of open tripped ßowers per raceme; and (3) number of open untripped ßowers per raceme. By averaging data from the 12 quadrats in each enclosure, such counts revealed the estimated number of open ßowers and which of these ßowers had presumably been pollinated (tripped) or remained available for bees to visit for pollen (untripped). Starting 1 wk after the bees were released, an otoscope was used to look inside each nest cavity of each board at twilight on the same weekday on which ßowers were counted in the afternoon after bees had had the opportunity to forage. Both sexes of M. rotundata will rest in bee boards at night, and physical characteristics make the sexes easy to distinguish from each other. Females do not share nests, but males sometimes do. From these “night counts,” we scored the number of males and females to determine the number of live bees and presumably nesting females over time. For evaluation of nesting success and bee reproduction per enclosure, we collected all complete and partial nests produced by the end of the study period by pulling out the straw inserts containing nest materials. Nests were brought into the USDAÐARS Bee Biology & Systematics Laboratory, and the contents of all straw inserts were examined using x-radiography (30-s exposure at 25 kvp in Faxitron 43804N; Faxitron X-Ray, Lincolnshire, IL) to determine percentages of cells containing healthy prepupae or adults, “pollen balls” (provision with dead egg/young larva or no detectable brood; Pitts-Singer 2004), and other immature mortality (Stephen and Undurraga 1976). All cells containing live progeny were removed, placed into gelatin capsules (size 00; Eli Lilly and Co., Indianapolis, IN), and allowed to continue development in darkness in an incubator (26⬚C). Any adult that emerged from these cells by October was designated as a “second-generation adult bee,” and its sex was recorded. Cells with live, diapausing prepupae were stored for the winter at 4.5⬚C beginning 1 October and incubated for adult emergence at 29⬚C the following June for determination of adult number and sex. These bees were designated as “diapausing adult bees.” For the bee foraging period to be as long as possible, bee release occurred after the onset of bloom but not necessarily at peak bloom that often occurred later. The duration of bloom and timing for peak bloom within any enclosure could not be accurately predicted before assignment of bee density, although in 2004, ßowers were counted before bee release to assure that bloom was underway. As a result of using random assignment, high numbers of bees were sometimes paired with enclosures that ultimately contained relatively low numbers of ßowers. Therefore, because of uneven ßoral resources and failure of all released female bees to establish in the enclosures (see Results), a derived metric, “density index,” was used for data analysis and interpretation. Throughout the rest of this manuscript, the density index is deÞned as “the area under a plotted curve of female bees by date calculated from the number of female bees counted at

February 2010

PITTS-SINGER AND BOSCH: M. rotundata DENSITY AFFECTS SUSTAINABILITY

night once a week” per “the area under a plotted curve of estimated open ßowers by date in the enclosure harboring those females for the same time period.” This density index summarizes bee density relative to ßower availability over the season, given the dynamic interaction of female bees and ßower production. Spearman rank partial correlation analysis of pooled data from both years, but controlling for year, was used to determine if the highest number of female bees found in the cavities correlated with the highest number of ßowers counted during the study period and to evaluate the relationships between the density index and several dependent variables: (1) the percentage of open ßowers that were tripped, (2) the number of ßowers tripped per highest count of female bees, (3) the number of nests and cells produced per the highest count of females, (4) the percentage of all cells produced that became pollen ball cells, (5) diapausing prepupae, (6) adult progeny (male and female), and (7) the sex ratios of progeny produced. Additionally, the lengths and widths of all prepupal cocoons formed in cells from all enclosures that ultimately resulted in second generation or diapausing adults were measured from x-ray images to estimate prepupal masses (mg) according to the Bosch and Kemp (2005) model for this estimation: prepupal weight ⫽ 1.42 (cocoon length ⫻ width) Ð 10; R2 ⫽ 0.67. A three-way factorial mixed model analysis of variance (ANOVA) was performed to evaluate whether prepupal mass (mg) was affected by year, sex, or diapause condition (i.e., whether prepupae were destined to develop into second-generation or diapausing adults). Correlation analyses were computed using the CORR procedure and analysis of variance was computed using the MIXED procedure in the SAS System for Windows, Release 9.2. (SAS Institute, Cary, NC). Results The number of ßowers was inconsistent between enclosures because of variability in the plant stand, incidence of plant pests, and yearly differences in blooming time in relation to bee release (Fig. 1). In both 2003 and 2004, the total number of open ßowers declined by the second week after bee release. Compared with 2003, there were many fewer ßowers in 2004, and the number of untripped ßowers (that served as pollen and nectar sources for the bees) represented a much smaller proportion of open ßowers throughout the study (Fig. 1). The number of bees counted in cavities at night within enclosures was always lower than the number released, but a higher proportion of released bees established nests in 2003 when more ßowers were present (Fig. 1; Table 1). However, the highest number of female bees found in the cavities was not signiÞcantly correlated with the highest number of ßowers counted during study periods (R ⫽ ⫺0.20; P ⫽ 0.48; Table 1). Of the available open ßowers in the enclosures in 2003 and 2004, a much higher percentage of them were tripped in 2004, when overall ßower abundance was

151

low (Fig. 1; Table 1). Correlation analysis showed a signiÞcant positive relationship between the percentage of ßowers that were tripped and the density index (R ⫽ 0.53; P ⫽ 0.04; Fig. 2A), but a negative relationship between the number of ßowers tripped per female and the density index (R ⫽ ⫺0.83; P ⬍ 0.01; Fig. 2B). Production of bee nests and also bee cells per female were negatively correlated with the density index (nests: R ⫽ ⫺0.855; P ⬍ 0.01; bee cells: R ⫽ ⫺0.861; P ⬍ 0.01; Fig. 3A and B). However, bee cell production did not translate directly to reproductive success because many of the cells produced (20 Ð100%) were pollen ball cells. The correlation between the percentage of cells that were pollen balls and the density index was signiÞcant and positive (R ⫽ 0.62; P ⫽ 0.01; Fig. 3C). No death caused by other known M. rotundata morality factors, such as hymenopteran parasitoids and chalkbrood (a fungal disease of larvae), was observed; however, some large larvae were found dead from unknown causes. The density index was not signiÞcantly correlated with the percentage of diapausing prepupae (R ⫽ ⫺0.32; P ⫽ 0.24; Fig. 4A). However, a signiÞcant negative correlation existed between the density index and percentage of cells that survived to adulthood (total of second generation and diapausing adults of both sexes; R ⫽ ⫺0.64; P ⬍ 0.01; Fig. 4B). Both the percentage of all cells that resulted in live female progeny (R ⫽ ⫺0.55; P ⫽ 0.03; Fig. 4C) and male progeny (R ⫽ ⫺0.66; P ⬍ 0.01) were signiÞcantly and negatively correlated with the density index. The density index did not correlate with sex ratio (R ⫽ ⫺0.37; P ⫽ 0.18; Table 2). Fewer second generation bees were produced in 2004 than in 2003. Analysis of pooled data from all enclosures in each year revealed differences in prepapal mass between years and according to sex and diapause status (Table 3). Prepupal mass was signiÞcantly greater for females than males (F ⫽ 40.05; df ⫽ 1, 331; P ⬍ 0.01), was greater in 2003 than 2004 (F ⫽ 14.24; df ⫽ 1, 331; P ⬍ 0.01), and was greater for diapausing than second generation individuals (F ⫽ 5.82; df ⫽ 1, 331; P ⬍ 0.02; Table 3). No signiÞcant interaction was found between any of the independent variables. Discussion Clearly, bee reproductive success is limited by the availability of ßoral resources, yet resource availability is hard to both predict and measure because of the complex interplay between foraging bees and alfalfa ßower production. By monitoring the numbers of nesting bees (rather than released bees) as well as open ßowers throughout our study, we were able to determine the contribution of only the surviving bees to pollination and reproduction. In our alfalfa Þeld enclosures, relatively high numbers of foraging M. rotundata could not be sustained, and the number of established females was never as high as the number of females released even in low numbers. Because

152


Vol. 39, no. 1

Fig. 1. By date of data collection for 2003 and 2004, the left y-axis gives the number of M. rotundata female bees released (Þrst data point) and then found nesting in Þeld enclosures and the right y-axis gives estimated numbers of total open (tripped ⫹ untripped) alfalfa ßowers and only untripped ßowers. Graphs represent data for enclosures with relatively low, moderate, and high numbers of female bees released in each year.

bees were trapped with limited ßoral resources in the enclosures, we surmised that bees found at the highest night counts represented the number of bees that could survive (at least at one point in time) with the amount of resources in the enclosures and that the missing bees had died. These Þndings further suggest that, in commercial open-Þeld environments, high densities of bees may not be sustained; bees in openÞeld situations may die or simply leave the release site (Bosch and Kemp 2005). Overall, the small Þeld (⬇0.25 ha) used for this trial was spatially heterogeneous and did not offer consistently ideal conditions for survival and nesting of many female bees in some enclosures. Extrapolation of ßower counts in our enclosures showed a range of 23.2Ð50.8 mil ßowers/ha while bees were nesting. This

can be compared with data from a 2.2-ha alfalfa Þeld planted for seed production that had a maximum ßower count of 68.7 mil ßowers/ha during the nesting season (unpublished data). In each year of this study, the alfalfa stand endured various and different problems that reduced bloom in some of the enclosures, especially in 2004. Such problems included pests (alfalfa weevil, lygus bugs, grasshoppers, mites, meadow voles, and weeds) that damaged plants, ßowers, or seeds. Using the number of bees that were known to be alive in the enclosures, a derived measurement of bee density, the density index, was used to encapsulate the dynamic interaction of bees and ßowers over time. This metric allowed for analysis and interpretation of a complex system, despite the variability between en-

February 2010 Table 1.


153

Number of M. rotundata females and open flowers

Females released versus highest count (% released) 2003

2004

10 versus 5 (50%) 20 versus 17 (85%) 30 versus 25 (83%) 40 versus 24 (60%) 50 versus 45 (90%) 60 versus 47 (78%) 90 versus 75 (83%) 105 versus 80 (76%)

16 versus 14 (88%) 32 versus 23 (72%) 48 versus 39 (81%) 64 versus 49 (77%) 80 versus 42 (53%) 96 versus 22 (23%) 112 versus 14 (13%) 128 versus 13 (10%)

Range and mean count of open ßowers (⫻1,000) 2003

2004

Range

Mean

Range

Mean

18.3Ð168.1 6.8Ð111.3 0.0Ð46.2 0.4Ð27.7 1.3Ð61.8 0.4Ð44.2 3.0Ð33.2 6.4Ð20.9

82.5 59.2 18.3 9.6 26.0 17.1 15.0 13.2

2.5Ð22.6 2.8Ð27.3 1.8Ð39.7 3.4Ð39.7 6.8Ð32.9 3.4Ð60.5 2.3Ð13.3 0.7Ð32.3

9.4 14.7 15.1 18.3 15.9 20.1 6.9 11.3

In a 2-yr study with bees in enclosures on an alfalfa Þeld, the no. M. rotundata females released into Þeld enclosures versus the highest night count of female bees (i.e., surviving females) in nest cavities in enclosures, and the range and mean of open ßowers while bees were present. The ßower data shown here were collected while bees were present in the enclosures.

closures in ßoral abundance. Results supported our hypothesis that pollination efÞciency and individual bee reproductive success are impaired by a relative

increase in M. rotundata density. There was a significant, but not steep, increase in the percentage of ßowers tripped as bee density index increased. How-

Fig. 2. In alfalfa Þeld enclosures in 2003 and 2004 studies, according to density index, (A) the percentage of the ßowers (estimated for each enclosure) that were tripped and (B) the number of ßowers tripped per the highest count of M. rotundata female bees.

154


Vol. 39, no. 1

Fig. 3. In alfalfa Þeld enclosures in 2003 and 2004 studies, according to density index, (A) the number of nests made per M. rotundata female, (B) the number of cells made per female, and (C) the percentage of all cells that contained pollen balls.

February 2010


155

Fig. 4. In alfalfa Þeld enclosures in 2003 and 2004 studies, according to density index, (A) the percentage of all M. rotundata cells that contained diapausing prepupae, (B) the percentage of all cells that produced adult M. rotundata progeny (second generation ⫹ diapausing), and (C) the percentage of all cells that produced female adults (second generation ⫹ diapausing).

156


Table 2. Year 2003

2004

Number and sex ratio of M. rotundata progeny Density indexa

Total no. live progeny

Adult progeny sex ratio

0.03 0.12 0.47 1.05 1.48 1.60 3.27 3.95 0.63 0.73 0.92 1.03 1.27 1.32 2.03 2.45

15 40 52 97 14 29 4 1 37 26 21 31 None 2 15 6

0.67:1 1.35:1 1.00:1 1.94:1 0.27:1 0.71:1 0.33:1 No males 1.10:1 0.44:1 0.75:1 0.24:1 No males No males 0.25:1 2.00:1

In a 2-yr study with bees in enclosures on an alfalfa Þeld, according to density index, the total M. rotundata progeny (both sexes of second generation ⫹ diapausing) produced that survived to adulthood and the sex ratios (M:F) of those progeny. a Area of female bees/area of open ßowers over study period; see detailed description in Materials and Methods.

ever, per female contribution to the number of tripped ßowers was negatively and dramatically affected by an increase in the bee density index, thus showing a pronounced reduction in individual bee pollination efÞciency. As bee density indices increased, reproductive success decreased. At the higher indices, each bee produced fewer nests and cells and generated a larger proportion of cells that were pollen balls. Additionally, the percentage of live adult progeny of both sexes (including adults that emerged at the end of summer or after winter) was negatively correlated with increases in bee density index, although the index did not correlate with the percentage of diapausing prepupae (that would be available for pollinating in the subsequent growing season). In a study of the effects of limited ßoral resources on M. rotundata reproduction, Peterson and Roitberg (2006) also found increased numbers of bee cells, nests, and progeny made by bees released in tents with relatively high ßoral resources compared with those with low or medium Table 3.

Mass of M. rotundata prepupae Females

Year

Second generation

2003

47.48 ⫾ 0.92 n ⫽ 47 42.57 ⫾ 1.07 n ⫽ 15

2004

Males

Diapausing

Second generation

Diapausing

49.76 ⫾ 0.93 n ⫽ 55 46.81 ⫾ 0.68 n ⫽ 61

41.29 ⫾ 0.61 n ⫽ 65 39.04 ⫾ 2.24 n⫽5

43.33 ⫾ 0.76 n ⫽ 52 39.44 ⫾ 1.27 n ⫽ 39

In a 2-yr study with bees in enclosures on an alfalfa Þeld, the estimated mass (mg; mean ⫾ SE) of M. rotundata prepupae that were produced and developed to adults from second generation or diapausing prepupae. Prepupal mass differed signiÞcantly (P ⬍ 0.05) between years, sexes, and destinies (whether development to adult occurred in summer ⫽ second generation, or after winter ⫽ diapausing).

Vol. 39, no. 1

resources; however, they did not see a signiÞcant difference in the number of cells that failed to produce adult bees. Population density and its effects on resource availability play a role in determining progeny investment and sex ratio. A study of Osmia cornifrons (Radoszkowski) (Megachilidae) showed that an increase in bee density resulted in a decrease in number of eggs laid and number of female offspring produced per female (Maeta et al. 2005). The proportion of female bee progeny was reduced as density increased, but density had no clear effect on cocoon weights. For O. cornuta, progeny sex ratio was more male-biased, and production of undersized offspring increased when resources were scarce (Bosch 2008). Kim (1999) found that Megachile apicalis increased maternal investment in daughters, as well as in total investment per offspring, when more ßoral resources were available. For beewolves, Philanthus triangulum F. (Hymenoptera: Sphecidae), low food availability resulted in a severely male-biased sex ratio (Strohm and Linsenmair 1997). In their resource allocation study in small tents, Peterson and Roitberg (2006) found that the sex ratio of M. rotundata diapausing progeny (M: F ⫽ 2.4:1) was unaffected by resource allocation. Their results suggest that all resource allocation treatments in their tents were somewhat limiting, and thus they concluded that the production of the “cheaper” (smaller) male progeny occurred across treatments because of those resource limitations. However, our sex ratios that included both summer-emerging and diapausing progeny were less male-biased (range, 0.2: 1Ð2:1) than in the aforementioned study under our low resource conditions. The sex ratios of only the diapausing progeny (range, 0.3:1Ð1.65:1) were even less male-biased than the sex ratio of the summeremerging bees. Our low sex ratios may be explained by the production of female cells Þrst in the nesting cycle when there were enough resources in our enclosures for making these cells, but a lull in the subsequent production of male cells because of the later paucity of resources. Progeny investment was also shown in prepupal mass as well as sex ratio. Although we lacked enough progeny from each enclosure for a comparison by bee density index, we could compare the estimated weights of nondiapausing and diapausing prepupae by pooling samples across enclosures in each year. A reduction in M. rotundata offspring investment from 1 yr to the next is shown by the 2003 prepupae being heavier than the 2004 prepupae, possibly because of the fact that the Þeld in 2003 had about twice the number of ßowers as the Þeld in 2004. We also found in both years that prepupae destined to remain in diapause through the winter were signiÞcantly heavier than second-generation prepupae, as we may have predicted because M. rotundata adults that develop from diapausing prepupae are heavier than summer-emerging adults (Tepedino and Parker 1988). A positive relationship between body size and propensity to diapause has been reported in other insects (Nesin 1985, Danforth 1999, Menu and Desouhant

February 2010


2002). Studies of the adult-wintering bee O. cornuta (Megachilidae) provided further support of this idea: larger O. cornuta were more likely to survive the winter and live longer than smaller bees (Bosch and Kemp 2004, Bosch and Vicens 2005, Bosch 2008). It is interesting, however, that the diapausing prepupae in 2004 weighed less on average than the 2003 second-generation prepupae. Therefore, other factors besides prepupal weight must be important for the determination of whether prepupae will enter diapause. Previous studies have shown that such factors as photoperiod, weather conditions throughout immature development, and maternal inheritance signiÞcantly affect the ratio of diapausing and nondiapausing individuals in M. rotundata (Parker and Tepedino 1982, Tepedino and Parker 1986, Kemp and Bosch 2000). Overall, our study conforms to, but is not completely explained by, the idea that nondiapausing prepupae lack the fat stores to survive the winter because their larval provisions were inadequate because of the low ßoral resource availability and the resulting competition between foraging bees for those sparse resources (Strickler and Freitas 1999, Bosch and Kemp 2005). By conÞning bees in enclosures, we eliminated dispersal as a survival and reproduction strategy for M. rotundata females. Although we unexpectedly found that the caged situation presented a dilemma for assigning bee density, our results using a density index support the idea that the high numbers of bees released for pollination on U.S. alfalfa seed Þelds could negatively affect each beeÕs pollination ability and reproductive success, which in turn results in poor population sustainability. For further understanding of M. rotundata population sustainability in the United States, an open-Þeld study currently is underway to examine the effects of releasing low, moderate, and standard bee densities (range, 37,000 Ð124,000 bees/ ha) in open alfalfa seed Þelds. Again, it is imperative to examine the number of nesting bees (including second-generation adults) and ßoral resources available throughout the nesting season. Along with the basic density-dependent relationships for bee reproductive success found in this study, an open-Þeld study performed over several years will allow for a commercial-scale assessment of both M. rotundata reproduction and alfalfa seed production.

Acknowledgments We thank ARS technicians S. Kalaskar, E. Klomps, and G. Trostle and Utah State University undergraduates M. Barker, J. Beattie, N. Boehme, C. Gorrell, E. Sharp, and M. J. Watterson for their combined efforts over the years in setting up and executing experiments. We also thank statisticians M. West and especially S. Durham for advice in project planning and analysis. We appreciate critical manuscript reviews by S. Durham, W. Goerzen, S. Peterson, and three anonymous reviewers whose comments and suggestions were very helpful.

157

References Cited Bohart, G. E. 1971. Management of habitats for wild bees. Proc. Tall Timbers Conf. Ecol. Anim. Control Habitat Manage. 3: 253Ð256. Bosch, J. 2008. Production of undersized offspring in a solitary bee. Anim. Behav. 75: 809 Ð 816. Bosch, J., and W. P. Kemp. 2004. Effect of pre-wintering and wintering temperature regimes on weight loss, survival, and emergence time in the mason bee Osmia cornuta (Hymenoptera: Megachilidae). Apidologie 35: 469 Ð 479. Bosch, J., and W. P. Kemp. 2005. Alfalfa leafcutting bee population dynamics, ßower availability, and pollination rates in two Oregon alfalfa Þelds. J. Econ. Entomol. 98: 1077Ð1086. Bosch, J., and N. Vicens. 2005. Sex allocation in the solitary bee Osmia cornuta: do females behave in agreement with FisherÕs theory? Behav. Ecol. Sociobiol. 59: 124 Ð132. Danforth, B. N. 1999. Emergence dynamics and bet hedging in a desert bee, Perdita portalis. Proc. Roy. Soc. Lond. B 266: 1985Ð1994. Frank, G. 2003. Alfalfa seed & leafcutter bee: productions and marketing manual. Irrigated Alfalfa Seed Producers Association, Brooks, Alberta, Canada. Hobbs, G. A. 1967. Domestication of alfalfa leaf-cutter bees. Canada Department of Agriculture, Ottawa, Canada. Kemp, W. P., and J. Bosch. 2000. Development and emergence of the alfalfa pollinator Megachile rotundata (Hymenoptera: Megachilidae). Ann. Entomol. Soc. Am. 93: 904 Ð911. Kemp, W. P., and J. Bosch. 2001. Postcocooning temperatures and diapause in the alfalfa pollinator Megachile rotundata (Hymenoptera: Megachilidae). Ann. Entomol. Soc. Am. 94: 244 Ð250. Kim, J. 1999. Inßuence of resource level on maternal investment in a leaf-cutter bee (Hymenoptera: Megachilidae). Behav. Ecol. 10: 522Ð556. Krunic, M. D., and C. F. Hinks. 1972. The effect of temperature and of temperature pretreatment on diapause and on the synchronization of adult emergence in Megachile rotundata (Hymenoptera: Megachilidae). Can. Entomol. 104: 889 Ð 893. Maeta, Y., and K. Kitamura. 2007. EfÞciency of seed production in breeding of ladino clover by three species of wild bees, Ceratina flavipes Smith, Megachile rotundata (Fabricius) and Megachile spissula Cockerell (Hymenoptera, Apidae and Megachilidae). Chugoku Kontyu 21: 35Ð53. Maeta, Y., S. Yoshida, S. Gotoˆ , and K. Kitamura. 2005. Relationship between seed yield of Chinese milk vetch and density of female bees, Osmia cornifrons (Radoszkowski) (Hymenoptera, Megachilidae). Chugoku Kontyu 19: 45Ð 61. Menu, F., and E. Desouhant. 2002. Bet-hedging for variability in life cycle duration: bigger and later-emerging chestnut weevils have increased probability of prolonged diapause. Oecologia 132: 167Ð174. Nesin, A. P. 1985. Contribution to the knowledge of the diapause of some pests of cones and seeds of conifers. Entomol. Rev. 68: 38 Ð 43. Parker, F. D., and V. J. Tepedino. 1982. Maternal inßuence on diapause in the alfalfa leafcutting bee (Hymenoptera: Megachilidae). Ann. Entomol. Soc. Am. 75: 407Ð 410. Peterson, J. H., and B. D. Roitberg. 2006. Impact of resource levels on sex ratio and resource allocation in the solitary bee, Megachile rotundata. Environ. Entomol. 35: 1404 Ð 1410.

158


Peterson, S. S., C. R. Baird, and R. M. Bitner. 1992. Current status of the alfalfa leafcutting bee, Megachile rotundata, as a pollinator of alfalfa seed. Bee Sci. 2: 135Ð142. Pitts-Singer, T. L. 2004. Examination of “pollen balls” in nests of the alfalfa leafcutting bee, Megachile rotundata. J. Apic. Res. 43: 40 Ð 46. Pitts-Singer, T. L., and R. R. James. 2005. Emergence success and sex ratio of commercial alfalfa leafcutting bees, Megachile rotundata Say, from the United States and Canada. J. Econ. Entomol. 98: 1785Ð1790. Pitts-Singer, T. L., and R. R. James. 2008. Do weather conditions correlate with Þndings in failed, provision-Þlled nest cells of Megachile rotundata (Hymenoptera: Megachilidae) in western North America? J. Econ. Entomol. 101: 674 Ð 685. Richards, K. W. 1984. Alfalfa leafcutter bee management in Western Canada. Agriculture Canada, Ottawa, Canada. Stephen, W. P. 1961. ArtiÞcial nesting sites for the propagation of the leaf-cutter bee, Megachile (Eutricharaea) rotundata, for alfalfa production. J. Econ. Entomol. 54: 989 Ð993.

Vol. 39, no. 1

Stephen, W. P., and J. M. Undurraga. 1976. X-radiography, an analytical tool in population studies of the leafcutter bee, Megachile pacifica. J. Apic. Res. 15: 81Ð 87. Strickler, K., and S. Freitas. 1999. Interactions between ßoral resources and bees (Hymenoptera: Megachilidae) in commercial alfalfa seed Þelds. Environ. Entomol. 28: 178 Ð 187. Strohm, E., and K. E. Linsenmair. 1997. Low resource availability causes extremely male-biased investment ratios in the European beewolf, Philanthus triangulm F. (Hymenoptera, Sphecidae). Proc. R. Soc. Lond. B. 264: 423Ð 429. Tepedino, V. J., and F. D. Parker. 1986. Effect of rearing temperature on mortality, second-generation emergence, and size of adult in Megachile rotundata (Hymenoptera: Megachilidae). J. Econ. Entomol. 79: 974 Ð977. Tepedino, V. J., and F. D. Parker. 1988. Alternation of sex ratio in a partially bivoltine bee, Megachile rotundata (Hymenoptera: Megachilidae). Ann. Entomol. Soc. Am. 81: 467Ð 476. Received 9 March 2009; accepted 28 September 2009.