Landscape distribution of food and nesting sites affect larval diet and nest size, but not abundance of Osmia bicornis

(1)

ORIGINAL ARTICLE

Landscape distribution of food and nesting sites affect larval diet and nest size, but not abundance of Osmia bicornis

Val ´erie Coudrain^1,2,3,†, Sarah Rittiner², Felix Herzog¹, Willy Tinner⁴ and Martin H. Entling³

1Research Station ART, Z ¨urich CH-8046, Switzerland;²Institute of Ecology and Evolution, University of Bern, Bern CH-3012, Switzerland;

3Institute for Environmental Sciences, University of Koblenz-Landau, Landau/Pfalz D-76829, Germany and ⁴Institute of Plant Sciences and Oeschger Centre for Climate Change Research, University of Bern, Bern CH-3013, Switzerland

AbstractHabitat fragmentation is a major threat for beneficial organisms and the ecosystem services they provide. Multiple-habitat users such as wild bees depend on both nesting and foraging habitat. Thus, they may be affected by the fragmentation of at least two habitat types. We investigated the effects of landscape-scale amount of and patch isolation from both nesting habitat (woody plants) and foraging habitat (specific pollen sources) on the abundance and diet ofOsmia bicornisL. Trap-nests ofO. bicorniswere studied in 30 agricultural landscapes of the Swiss Plateau. Nesting and foraging habitats were mapped in a radius of 500 m around the sites. Pollen composition of larval diet changed as isolation to the main pollen source,Ranunculus, increased, suggesting thatO. bicornisadapted its foraging strategy in function of the nest proximity to main pollen sources. Abundance of O. bicorniswas neither related to isolation or amount of nesting habitat nor to isolation or abundance of food plants. Surprisingly, nests ofO. bicorniscontained fewer larvae in sites at forest edge compared to isolated sites, possibly due to higher parasitism risk. This study indicates thatO. bicorniscan nest in a variety of situations by compensating scarcity of its main larval food by exploiting alternative food sources.

Key words diet; fragmentation; habitat amount; habitat isolation; Hymenoptera;

Megachilidae; pollen analysis

Introduction

Replacement of natural habitats by cropland during the last decades often resulted in simplified landscapes, where remaining natural habitat is highly fragmented (Fahrig, 1997; Tilmanet al., 2001; Benton et al., 2003; Poveda et al., 2005). This can reduce the abundances of bene- ficial organisms such as wild bees and the pollination services they provide (Rickettset al., 2008; Farwiget al., 2009; Garibaldiet al., 2011; Garibaldi et al., 2013). In contrast to many species that complete their life-cycle in

Correspondence: Val´erie Coudrain, Research Station ART, Reckenholzstrasse 191, Z¨urich CH-8046, Switzerland.

Tel:+33 664 55 21 84; email: vcoudrain@gmail.com

†INRA Ecosys, RD 10, 78026 Versailles Cedex, France.

one clearly defined habitat type, multiple-habitat users such as wild bees depend on both nesting and foraging habitat (Westrich, 1996). Thus, the fragmentation of at least two habitat types may affect occurrence and population abundances ofO. bicornis, something that has rarely been considered in empirical studies.

Habitat fragmentation changes landscape structure through two main mechanisms. On the one hand, the amount of habitat in a landscape is reduced and on the second hand, the distance among remaining habitat patches increases (an aspect of fragmentation per se) (Fahrig, 2003). These two mechanisms have rarely been distin- guished, but first evidence indicates that habitat amount and patch isolation can have different effects on species abundance, diversity, and ecosystem functions (Smith et al., 2009; Baileyet al., 2010; Mortellitiet al., 2011;

Sch¨ueppet al., 2014a,b). Here, we aim to determine the

(2)

effects of landscape-scale amount of and patch isolation from both foraging and nesting habitats on the abundance of the Red Mason beeOsmia bicornis(Linnaeus, 1758, Hymenoptera: Megachilidae, formerly Osmia rufa L.).

Assessing this question might be significant for agricultural purposes, sinceO. bicornishas been found to pol- linate crops such as rape, onion, and fruit trees (Kruni´c et al., 1995; Wilkaniec et al., 2004; Teper & Bilinski, 2009; Jaukeret al., 2012).

O. bicornisis a solitary, polylectic (pollen generalist) bee species with a univoltine life cycle (Raw, 1974; Raw &

O’Toole, 1979; Westrich, 1990). It is widely distributed in Central Europe and its main flight period lasts from mid- April until the end of June (Steffan-Dewenter & Schiele, 2008). To complete its life cycleO. bicornisdepends on the presence of sufficient food resources for adults and larvae, as well as on the availability of nesting cavities, such as beetle burrows in dead wood, hollow stems in shrubs or crevices in buildings (Westrich, 1990; Bellmann, 1995;

M¨ulleret al., 1997).

To provision its brood,O. bicorniscollects pollen from different sources such as large trees, shrubs, and herba- ceous plants, but shows a preference for pollen from RanunculusandQuercus(Free & Williams, 1970; Tasei, 1973; Radmacher & Strohm, 2009). Some plants such as mass-flowering oilseed rape are not visited for pollen collection but provide abundant nectar sources for adult O. bicornis(Jaukeret al., 2012; Holzschuhet al., 2013).

Natural nesting habitats comprise woody vegetation types such as hedgerows, orchards, and forest (Bellmann, 1995).

In agricultural landscapes, nesting habitat can be a limited resource forO. bicornisas demonstrated by the strong in- crease in population sizes following the establishment of additional nesting cavities (Steffan-Dewenter & Schiele, 2008).

Foraging habitat ofO. bicornisis often spatially sep- arated from nesting habitat (Westrich, 1990). While the occurrence of floweringQuercustrees corresponds with potential nesting habitat in our study region, Ranuncu- lusandBrassica napusgrow mostly in open habitat that does not provide natural nesting cavities. Thus, a sepa- rate analysis of foraging and nesting habitat ofO. bicor- nis could reveal its spatial dependence on both habitat types. Since sufficient larval resources are a prerequisite for population build-up, we investigated pollen composition of brood cells and tested the hypothesis that (i) the proportion of specific pollen types increases with increasing amount and decreasing isolation of the corresponding food plants in the landscape. We then test our expectations that abundance ofO. bicornisincreases (ii) with increasing abundance and decreasing isolation of the most visited

pollen sources, and (iii) with increasing amount and decreasing isolation of woody nesting habitats.

Materials and methods Study sites and vegetation mapping

The study was carried out in summer 2010 on the Swiss Plateau between the cities of Bern, Solothurn and Fribourg. This region is characterized by agricultural areas interspersed with forest fragments. We studied 30 circu- lar landscape sectors, each with a radius of 500 m that corresponds to O. bicornis flying range (Gathmann &

Tscharntke, 2002; Guedot et al., 2009). The landscape sectors were distributed over an area of 23 km×32 km.

The study sites were situated on low-intensity grassland in the centers of each landscape and varied in altitude between 465 and 705 m above sea level. Centers of the landscapes sectors were located at least 750 m apart from each other, limiting overlap to <15% of the landscape area.

Moran’s I correlograms gave no indication for possible confounding effects of spatial autocorrelation of explanatory or response variables (Supporting Information S1).

Study sites were selected according to their percentage of woody habitat (3.6%–74.2%) within the 500 m radius and their level of isolation from woody habitat, as described in Farwiget al.(2009) and Sch¨uepp et al.(2011). The sites had three levels of isolation from forest. Ten sites were situated adjacent to forest, 10 sites laid in the open agricultural landscape but were connected to forest by hedgerows or trees, and 10 sites were isolated from all woody habitats by at least 100 m. There was no statistical dependency between the percentage of woody habitat in a landscape sector and the level of isolation of a study site (F2,27 =0.004,P=0.99). Maps showing details of the study location and design are available in Figs. 1 and 2 in Coudrainet al.(2014).

Local climate and altitude were measured at each site to test for possible confounding influences. Altitude was obtained from topographic maps. Air temperature and humidity were recorded at each site once per hour with a data logger (Hygrochron^TMiButtons, Maxim, Sunnyvale, USA) and the mean over the flying period (May to August 2010) was calculated and used in the analyses.

Potential food plants ofO. bicorniscould not be considered for study site selection but were mapped in a radius of 500 m around the study sites in early spring 2010.

The minimum plant height considered was 5 m for trees and 1 m for shrubs. We mapped all species of shrubs and broadleaf trees except forFagus sylvaticaL. because

(3)

Fig. 1 Overall proportion of the 6 most important pollen types in the brood cells ofOsmia bicornisin relation to the percentage of the corresponding plants across all landscapes. Plain and dotted lines represent correlation with (r=0.97) and without B. napus(r=0.58).

its pollen was not previously recorded in the diet ofO.

bicornis(Raw, 1974; Radmacher & Strohm, 2009). Black- berries (Rubusspecies) were included if they covered an area of at least 1 m². In addition to trees and shrubs, oilseed rape (Brassica napusL.) fields and grasslands containing Ranunculusspecies were mapped using Global Position- ing System (GPS). Landscape data were analyzed with Geographic Information System (ArcGIS, version 9.2, ESRI Redlands, CA, USA). We measured the distance of every mapped tree, shrub, oilseed rape field andRanun- culusstand to each trap-nest location (see below). Plant abundance was determined as its percent coverage in each landscape sector, assuming a crown radius of 5 m for trees and 1.5 m for shrubs.

Trap nests and pollen analysis

In March 2010, four trap nests (Tscharntkeet al., 1998) per study site were installed on wooden posts 1 m above the ground. The trap nests consisted of a plastic cylin- der (diameter 10 cm, length 20 cm) filled with about 170 common reed internodes (Phragmites australisTrin.).

The reed internodes were divided into three size classes (2–4, 5–7, and>7 mm) and the same weight of reed nodes of each size class was introduced into the trap nest. Two of the four trap nests had already been in the field the pre- vious year to enable local population development. They were stored at 4°C in a climate chamber from October 2009 to March 2010, after which they were brought to the same study sites along with two trap nests with new reed

Fig. 2 Proportion of pollen fromRanunculus() andQuer- cus() over all samples in relation to (A) distance from trap nests to the nearestRanunculusstand (Ranunculus:F1,22=4.4, P=0.05;Quercus:F1,22=3.9,P=0.06) and (B) distance from trap nests to the nearestSalixstand (Ranunculus:F1,22=1.7, P=0.20;Quercus:F_1,22=3.8,P=0.06).

internodes. During the main flight period ofO. bicornis, reed internodes containing newly completed nests were collected without replacement from two trap-nests per site every second week between mid-May until end of June 2010 (four sampling rounds). Collected reed tubes were opened in the laboratory and a small amount of pollen was sampled from one cell in eight randomly chosen reed tubes per study site. Pollen was prepared and determined following procedures described in Supporting Information S2.

After pollen collection, the reed tubes were closed again to allow bee emergence. The reed tubes were stored in individual glass tubes at ambient temperature (ca. 23°C) until October. In October we removed the two remaining trap-nests from each site and put them together with the reed tubes collected for pollen analysis in a cool chamber

(4)

(4°C) over winter. In March, all reed tubes were placed into individual glass tubes and brought in a greenhouse to allow bees to emerge and species to be determined. After species determination, reed tubes from the two trap-nests per site not used for pollen analysis were opened and the number of brood cells was counted.

Statistical analysis

Amount of foraging habitat was defined as the proportion of each mapped plant species within the landscape, while isolation of foraging habitat was defined as the minimum distance from the study site to the nearest locality of the plant species. Amount of nesting habitat was defined as the proportion of woody ele- ments (forest, hedgerows, and orchards) within the 500-m radius-landscape surrounding each study site. Although O. bicornis can nest in cavities in buildings (Westrich, 1990), we did not include building area in the calculation of nesting habitat, because area was small (5.9%±4.9%) and unrelated to the abundance ofO. bicornis (Pearson correlation test:n=30,r<0.001,P=0.999). Isolation from nesting habitat was characterized by the three levels of patch isolation from forest described above (adjacent, connected, isolated). Abundance ofO. bicorniswas defined as the number of brood cells per study site, which was strongly correlated to the number of occupied reed tubes (Pearson correlation test: n =30,r =0.99,P <

0.001). Additionally, individual nest size was defined as the mean number of brood cells per nest. Cell abundance and individual nest size were not correlated (Pearson correlation test:n=30,r=0.06,P=0.76).

To test our first hypothesis, pollen composition of brood cells in each site was characterized and only plants accounting for >1% of the pollen provisions were con- sidered for further analyses (highlighted in bold in Sup- porting Information S3), because pollen accounting for

<1% may represent accidental contamination, for ex- ample, from plants visited for nectar by the adult bee (Quiroz-Garcia et al., 2001). First, Pearson correlation was used to test the relation between the mean proportion of pollen types in the samples and the mean proportion of corresponding food plants in the landscape. Then, we constructed for each site a site×pollen matrix, using observed pollen proportion as matrix entries. Redundancy Analysis (RDA) was used to test if pollen composition of the larval diet could be explained by (1) the proportion of each mapped plant within the landscape, and (2) the nearest distance to the nest of each mapped plant. To further in- vestigate the relation among plant abundance or distance and pollen composition, the proportion of each pollen

type was regressed against the plants whose abundance or distance explained a significant amount of variance in pollen composition, using generalized linear models with quasi-binomial distribution (log it link).

To test our second and third hypotheses, linear regres- sion models were used with (1) brood cell abundance and (2) individual nest size of O. bicornisas response variables. Explanatory variables were as follows: the total proportion of food plants per site (1 univariate model), the proportion of and isolation from each individual food plant in the landscape (7 individual plants leading to 14 univariate models), the amount of and isolation from woody habitat (2 univariate models). We further tested the interaction among the total proportion of food plants per site and the amount of and isolation from woody habitat.Post hocpairwise comparisons among the three levels of isolation from nesting habitat were performed using Tukey tests. Abundance ofO. bicorniswas square-root- transformed to improve normality of the residuals. For significant models, possible confounding effects of abiotic variables on landscape variables were tested by enter- ing each single abiotic variable—temperature, humidity, and altitude—as first variable in the models and using type I sum of squares. Model fit was evaluated by visual inspection of residual plots. All analyses were performed with R version 2.14.1 (R Core Team, 2011).

Results

Composition of pollen provisions

We found 41 different pollen types in the nesting provisions ofO. bicornisin accordance with the wide variety of available flowering plant species during the nesting period (Supporting Information S3). The most abundant pollen type wasRanunculuswith 58.6% of all pollen grains collected, followed by Quercus with 23.4%. The remaining 39 pollen types were only found in small amounts (<4.0% over all samples). The total proportion of the most common pollen types in the samples (>1%) correlated with the total proportion of the corresponding food plants in the landscape except forB. napus(Pearson correlation test with resp. withoutB. napus:n=6,r=0.58, P=0.22;n=5,r=0.97,P<0.01, respectively; Fig. 1).

Papaver rhoeascould not be considered for this analysis because it had not been mapped in the field due to late appearance of the plants.

The composition of the pollen provisions could be explained by the distance to the plants in the landscape (RDA:k=999,λ=0.30,F2,20=4.4,P<0.01), specif- ically by the distance to the nearest Ranunculus stand

(5)

Fig. 3 Mean nest size ofO. bicornisin relation to the 3 levels of trap nest isolation from forest (F2,26=8.5,P<0.01). Different letters show significant differences. Error bars show standard error of the mean.

(F1,20=5.1,P=0.01) and by the distance to the nearest Salixstand (F1,20=4.4,P=0.01). The proportion ofRa- nunculuspollen correlated negatively with the distance to the nearestRanunculusstand (F1,22=4.4,P=0.05), but was not related to the distance to the nearestSalixstand (F1,22 =1.7, P =0.20). In contrast, the proportion of Quercuspollen per sample marginally increased with increasing distance to the nearestRanunculusstand (F1,22

=3.9,P=0.06) and with decreasing distance to the next Salixstand (F1,22=3.8,P=0.06) (Fig. 2).

O. bicornis abundance

We recorded a total of 11 220 cells ofO. bicornisdis- tributed over the 30 study sites (range: 0–1229). Abun- dance ofO. bicorniswas neither related to the amount nor isolation of foraging habitat, as no significant relation- ships were detected with the total amount of food plants or any of the individually modeled food plants. Similarly, O. bicornis abundance was neither related to isolation from forest (F2,26=0.9,P=0.40), nor to the amount of woody habitat in the surrounding landscape (F1,26=0.0, P = 0.99). Mean nest size was neither related to the amount nor isolation of foraging habitat, but was significantly lower in sites adjacent to forest compared to connected or isolated sites (F2,26 = 8.5, P < 0.01, Fig. 3). In addition, nest size marginally increased with the amount of nesting habitat in the landscapes (F1,26 =3.4, P=0.08). Interactive effects among nesting and foraging habitats were not significant. None of the three abiotic variables (temperature, humidity, or altitude) significantly explained variation in the nest size ofO. bicornis.

Discussion

As expected, the two plant genera Ranunculus and Quercus strongly dominated the larval provisions ofO.

bicornis(Free & Williams, 1970; Tasei, 1973; Radmacher

& Strohm, 2009). The total proportion of pollen in the larval diet ofO. bicornisscaled with the overall availability of the corresponding plant in the landscape, in accordance with the polylectic habits ofO. bicornis. As a main devia- tion from this pattern, pollen of the mass-flowering oilseed B. napuswas rare in the diet ofO. bicorniscompared to its abundance in the landscape. AlthoughB. napus can provide valuable nectar for adult bees and enhanceO. bi- cornis population sizes (Jaukeret al., 2012; Holzschuh et al., 2013), our results confirm that the pollen of this mass-flowering crop is of minor importance for larval diet (i.e., Jaukeret al., 2012). Possibly,B. napuspollen is of poor quality for the larval development ofO. bicornis, as individuals raised on pureB. napuspollen showed behavioral failures (Dobsonet al., 2012). In addition, the short flowering time ofB. napusmay constrain the amounts of pollen stored within the brood cells.

Our first hypothesis predicted a relation between pollen proportions in the larval diet and the amount of or distance to the corresponding food plants in the landscape. The proportion ofRanunculuspollen significantly decreased in the larval diet with increasing distance of the nests from this plant, while it was not related to the amount of Ranunculuspollen available in the landscape. This result suggests that O. bicornislimits its foraging distance by adapting its foraging strategy to closely available pollen sources. Long foraging distances can lead to substantial reproductive costs for solitary bees (Zurbuchen et al., 2010), and the developing larvae and pollen provisions are a desirable resource for parasitoids and cleptoparasites (Kruni´cet al., 2005; Seidelmann, 2006). The probability of brood cell attack by parasitoids depends on the time the cell is open and unguarded by the female bee (Ulbrich

& Seidelmann, 2000; Bosch, 2008). Thus, by feeding in proximity to its nest,O. bicornismight not only save en- ergy by flying over shorter distances but also reduce parasitism risk. Interestingly, the amount ofQuercuspollen was not related to the distance from nest toQuercustrees, but increased with the isolation of the nests fromRanun- culusstands. Therefore,O. bicornismay have preferably exploitedRanunculusplants when they were close to its nest and switched toQuercusonce the distance increased.

Although Quercus is a wind-pollinated plant that pro- duces no nectar (Chambers, 1945), it belongs to the most protein-rich anemophilous pollen (about 40%) (Roulston et al., 2000) and may therefore provide a valuable alter- native resource. The proportion ofQuercuspollen in the

(6)

samples was positively related to the proximity of Salix stands. BecauseSalixproduces a high amount of nectar early in the season (Gaudin, 1828), Salixmay act as a complementary nectar source forO. bicornisfemales that collect pollen fromQuercus.

Our second and third hypotheses stated thatO. bicornis abundance was related to the amount and isolation of its foraging and nesting habitats. Although abundance of food resources is a main limiting factor for solitary bees (Roulston & Goodell, 2011), our results suggest that the abundance of O. bicorniswas neither related to the amount nor to the isolation of any food plant within its foraging range. Similarly, O. bicornis population sizes were not affected by the distance to abundant foraging resources in urban areas, indicating that food limitation is uncommon inO. bicornis(Everaarset al., 2011). Possibly, the observed adaptability to available pollen sources as a function of distance to food plants enablesO. bicornisto successfully forage in a wide variety of situations.

Abundance ofO. bicornisdid not significantly respond to the amount of nesting habitat in the landscape. This is surprising, given that O. bicornis populations can be strongly constrained by nesting site availability (Steffan- Dewenter & Schiele, 2008). Characterization of natural breeding locations and population sizes would help un- derstanding our results but is difficult to achieve. Like- wise, the isolation distances tested in our study appeared to be no significant barrier forO. bicornis. This result con- trasts with the strong decline in population size recorded in isolated habitats for the spider-hunting waspTrypoxy- lon figulus from the same trap nests (Coudrain et al., 2013), as well as the reported negative effects of isolation on smaller species such as Hoplitis adunca or Chelostoma sp. (Gathmann & Tscharntke, 2002;

Zurbuchenet al., 2010).

Surprisingly, despite similar abundance irrespective of isolation, nest size (i.e., the number of brood cells per reed tube) was smaller in sites adjacent to forest than in connected or isolated sites. In solitary bees, the number of brood cells per nests can be affected by parasitism (Good- ell, 2003). As parasitism was highest in sites adjacent to forest (results not shown), this could explain why fewer cells per nest had been built near forest edges. Alterna- tively, nest size may decline with declining availability of food resources (Peterson & Roitberg, 2006). However, we found no significant relation between nest size and main pollen sources or isolation from nesting habitat in our study, but additional unmeasured resources, such as adult food and nest building material, may have contributed to the recorded difference in nest size.

Overall, this study shows the ability ofO. bicornisto maintain large populations despite fragmentation of both

nesting and foraging habitats at the scale of our study.

The observed shifts in pollen provisioning of their larvae may contribute to the ability of this species to successfully breed in a wide variety of landscape situations.

Acknowledgments

We are grateful to Jacqueline van Leeuwen and Pim van der Knaap (from the Institute of Paleoecology of the Uni- versity of Bern) for the help in the pollen determination and to Florencia Oberli for the help in pollen analysis.

Sincere thanks to Jonas Winizki and Erich Szerencsits for their support in GIS analysis, to Roman Bucher and Sandra Krause for the collaboration, to Christof Sch¨uepp, Marc Benkemoun and two anonymous reviewers for use- ful comments. We thank the 30 farmers for providing land for our studies. This study was supported by the Swiss National Science foundation under grant number 3100A0-114058 to Felix Herzog and Martin Entling.

Disclosure

The authors do not have conflict of interest.

References

Bailey, D., Schmidt-Entling, M.H., Eberhart, P. Herrmann, J.D., Hofer, G., Kormann, U. and Herzog, F. (2010) Effects of habitat amount and isolation on biodiversity in fragmented traditional orchards.Journal of Applied Ecology, 47, 1003–

1013.

Bellmann, H. (1995)Bienen, Wespen, Ameisen. Hautfl¨ugler Mit- teleuropas. Frankh-KosmosVerlags-GmbH and Co, Stuttgart.

Benton, T.G., Vickery, J.A. and Wilson, J.D. (2003) Farmland biodiversity: is habitat heterogeneity the key?Trends in Ecol- ogy and Evolution, 18, 182–188.

Bosch, J. (2008) Production of undersized offspring in a solitary bee.Animal Behaviour, 75, 809–816.

Chambers, V.H. (1945) British bees and wind-borne pollen.

Nature, 155, 145–145.

Coudrain, V., Herzog, F. and Entling, M.H. (2013) Effects of habitat fragmentation on abundance, larval food and parasitism of a spider-hunting wasp.PLoS ONE, 8, e59286.

Coudrain, V., Sch¨uepp, C., Herzog, F., Albrecht, M. and Entling, M.H. (2014) Habitat amount modulates the effect of patch isolation on host-parasitoid interactions. Frontiers in Envi- ronmental Science, 2, 27.

Dobson, H.E.M., Ayasse, M., O’Neal, K.A. and Jacka, J.A.

(2012) Is flower selection influenced by chemical imprinting to larval food provisions in the generalist beeOsmia bicornis (Megachilidae)?Apidologie, 43, 698–714.

(7)

Everaars, J., Strohbacha, M.V., Grubera, B. and Dormann, C.F.

(2011) Microsite conditions dominate habitat selection of the Red Mason bee (Osmia bicornis, Hymenoptera: Megachili- dae) in an urban environment: A case study from Leipzig, Germany.Landscape and Urban Planning, 103, 15–23.

Fahrig, L. (1997) Relative effects of habitat loss and fragmentation on population extinction.Journal of Wildlife Manage- ment, 61, 603–610.

Fahrig, L. (2003) Effects of habitat fragmentation on biodiversity.Annual Review of Ecology Evolution and Systematics, 34, 487–515.

Farwig, N., Bailey, D., Bochud, E., Herrmann, J.D., Kindler, E., Reusser, N., Sch¨uepp, C. and Schmidt-Entling, M.H. (2009) Isolation from forest reduces pollination, seed predation and insect scavenging in Swiss farmland.Landscape Ecology, 24, 919–927.

Free, J.B. and Williams, I.H. (1970) Preliminary investiga- tions on occupation of artificial nests by Osmia rufa L.

(Hymenoptera, Megachilidae).Journal of Applied Ecology, 7, 559–566.

Garibaldi, L.A., Steffan-Dewenter, I., Kremen, C., Morales, J.M., Bommarco, R., Cunningham, S.A., Carvalheiro, L.G., Chacoff, N.P., Dudenh¨offer, J.H. and Greenleaf, S.S. (2011) Stability of pollination services decreases with isolation from natural areas despite honeybee visits.Ecology Letters, 14, 1062–1072.

Garibaldi, L.A., Steffan-Dewenter, I, Winfree, R, Aizen, M.A., Bommarco, R., Cunningham S.A., Kremen, C., Carvalheiro, L.G., Harder, L.D., Afik, O., Bartomeus, I., Benjamin, F., Boreux, V., Cariveau, D., Chacoff, N.P., Dudenhöffer, J.H., Freitas, B.M., Ghazoul, J., Greenleaf, S., Hipólito, J., Holzschuh, A., Howlett, B., Isaacs, R., Javorek, S.K., Kennedy, C.M., Krewenka, K., Krishnan, S., Mande- lik, Y., Mayfield, M.M., Motzke, I., Munyuli, T., Nault, B.A., Otieno, M., Petersen, J., Pisanty, G., Potts, S.G., Rader, R., Ricketts, T.H., Rundlöf, M., Seymour, C.L., Schüepp, C., Szentgyörgyi, H., Taki, H., Tscharntke, T., Vergara, C.H., Viana, B.F., Wanger, T.C., Westphal, C., Williams, N. and Klein, A.M. (2013) Wild pollinators enhance fruit set of crops regardless of honey bee abundance.Science, 339, 1608–

1611.

Gathmann, A. and Tscharntke, T. (2002) Foraging ranges of solitary bees.Journal of Animal Ecology, 71, 757–764.

Gaudin, J.F. (1828)Flora Helvetica. Orell et F¨ussli, Zurich.

Goodell, K. (2003) Food availability affects Osmia pumila (Hymenoptera: Megachilidae) foraging, reproduction, and brood parasitism.Oecologia, 134, 518–527.

Guedot, C., Bosch, J. and Kemp, W.P. (2009) Relationship between body size and homing ability in the genus Osmia (Hymenoptera; Megachilidae).Ecological Entomology, 34, 158–161.

Holzschuh, A., Dormann, C.F., Tscharntke, T. and Steffan- Dewenter, I. (2013) Mass-flowering crops enhance wild bee abundance.Oecologia, 172, 477–484.

Jauker, F., Peter, F., Wolters, V. and Diek¨otter, T. (2012) Early reproductive benefits of mass-flowering crops to the solitary beeOsmia rufaoutbalance post-flowering disadvantages.

Basic and Applied Ecology, 13, 268–276.

Kruni´c, M., Pinzauti, A. and Staisavljevi´c, L. (1995) Further ob- servations onOsmia cornutaLatr. andO. rufaL. as alternative fruit pollinators, domestication and utilization.Archives Biological Sciences Belgrade, 47, 59–66.

Kruni´c, M., Staisavljevi´c, L., Mauro, P. and Antonio, F. (2005) The accompanying fauna ofOsmia cornutaandOsmia rufa and effective measures of protection.Bulletin of Insectology, 58, 141–152.

Mortelliti, A., Amori, G., Capizzi, D., Cervone, C., Fagiani, S., Pollini, B. and Boitan, L. (2011) Independent effects of habitat loss, habitat fragmentation and structural connectivity on the distribution of two arboreal rodents.Journal of Applied Ecology, 48, 153–162.

M¨uller, A., Krebs, A. and Amiet, F. (1997)Bienen – Mitteleu- rop¨aische Gattungen, Lebensweise, Beobachtung. Naturbuch Verlag, Augsburg.

Peterson, J.H. and Roitberg, B. (2006) Impact of resource levels on sex ratio and resource allocation in the solitary bee, Megachile rotundata.Environmental Entomology, 35, 1404–

1410.

Poveda, K., Steffan-Dewenter, I., Scheu, S. and Tscharntke, T.

(2005) Effects of decomposers and herbivores on plant per- formance and aboveground plant-insect interactions.Oikos, 108, 503–510.

Quiroz-Garcia, D.L., Martinez-Hernandez, E., Palacios-Chavez, R. and Galindo-Miranda, N.E. (2001) Nest provisions and pollen foraging in three species of solitary bees (Hymenoptera: Apidae) from Jalisco, Mexico.Journal of the Kansas Entomological Society, 74, 61–69.

R Core Team (2011)R: A Language and Eenvironment for Sta- tistical Computing. R Foundation for Statistical Computing, Vienna, Austria.

Radmacher, S. and Strohm, E. (2009) Factors affecting offspring body size in the solitary beeOsmia bicornis(Hymenoptera, Megachilidae).Apidologie, 41, 169–177.

Raw, A. (1974) Pollen preferences of three Osmia species (Hymenoptera).Oikos, 25, 54–60.

Raw, A. and O’Toole, C. (1979) Errors in the sex of eggs laid by the solitary beeOsmia rufa(Megachilidae).Behaviour, 70, 168–171.

Ricketts, T.H., Regetz, J., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C., Bogdanski, A., Gemmill-Herren, B., Greenleaf, S.S., Klein, A.M., Mayfield, M.M., Morandin, L.A., Ochieng, A., Potts, S.G. and Viana, B.F. (2008)

(8)

Landscape effects on crop pollination services: Are there gen- eral patterns?Ecology Letters, 11, 1121–1121.

Roulston, T.H., Cane, J.H. and Buchmann, S.L. (2000) What governs protein content of pollen: Pollinator preferences, pollen-pistil interactions, or phylogeny? Ecological Mono- graphs, 70, 617–643.

Roulston, T.H. and Goodell, K. (2011) The role of resources and risks in regulating wild bee populations.Annual Review of Entomology, 56, 293–312.

Sch¨uepp, C., Herrmann, J.D., Herzog, F. and Schmidt-Entling, M.H. (2011) Differential effects of habitat isolation and landscape composition on wasps, bees, and their enemies.

Oecologia, 165, 713–721.

Sch¨uepp, C., Herzog, F. and Entling, M.H. (2014) Disentangling multiple drivers of pollination in a landscape-scale experi- ment.Proceedings of the Royal Society B-Biological Sciences, 281, 20132667.

Sch¨uepp, C., Uzman, D., Herzog, F. and Entling, M.H. (2014) Habitat isolation affects plant–herbivore–enemy interactions on cherry trees.Biological Control, 71, 56–64.

Seidelmann, K. (2006) Open-cell parasitism shapes mater- nal investment patterns in the Red Mason beeOsmia rufa.

Behavioral Ecology, 17, 839–848.

Smith, A.C., Koper, N., Francis, C.M. and Fahrig, L. (2009) Confronting collinearity: comparing methods for disentangling the effects of habitat loss and fragmentation.Landscape Ecology, 24, 1271–1285.

Steffan-Dewenter, I. and Schiele, S. (2008) Do resources or natural enemies drive bee population dynamics in fragmented habitats?Ecology, 89, 1375–1387.

Tasei, J.N. (1973) Le comportement de nidification chezOsmia cornutaLatr. etOsmia rufaL. (Hymenoptera: Megachilidae).

Apidologie, 4, 195–225.

Teper, D. and Bilinski, M. (2009) Red Mason bee (Osmia rufa L.) as a pollinator of rape plantations.Journal of Apicultural Science, 53, 115–120.

Tilman, D., Fargione, J., Wolff, B., D’Antonio, C., Dobson, A., Howarth, R., Schindler, D., Schlesinger, W.H., Sim-

berloff, D. and Swackhamer, D. (2001) Forecasting agricul- turally driven global environmental change. Science, 292, 281–284.

Tscharntke, T., Gathmann, A. and Steffan-Dewenter, I. (1998) Bioindication using trap-nesting bees and wasps and their natural enemies: community structure and interactions.Journal of Applied Ecology, 35, 708–719.

Ulbrich, K. and Seidelmann, K. (2000) Modeling population dynamics of solitary bees in relation to habitat quality.Web Ecology, 2, 57–64.

Westrich, P. (1990)Die Wildbienen Baden-W¨urttembergs. 2nd edn., Eugen Ulmer Verlag, Stuttgart.

Westrich, P. (1996) Habitat requirements of central European bees and problems of particular habitats.The Conservation of Bees (eds. A. Matheson, S.L. Buchmann, C. O’Toole, P. Westrich & I.H. Williams), pp. 1–16. Academic Press, London.

Wilkaniec, Z., Giejdasz, K. and Pr´oszy˜nski, G. (2004) Effect of pollination of onion seeds under isolation by the Red Mason bee (Osmia rufaL.) (Apoidea, Megachilidae) on the setting and quality of obtained seeds.Journal of Apicultural Science, 48, 35–41.

Zurbuchen, A., Cheesman, S., Klaiber, J., Muller, A., Hein, S.

and Dorn, S. (2010) Long foraging distances impose high costs on offspring production in solitary bees. Journal of Animal Ecology, 79, 674–681.

Accepted April 23, 2015

Supporting Information

Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:

Fig. S1.Spatial autocorrelation.

Fig. S2.Determination of pollen samples.

Fig. S3.The 41 different pollen types recorded in the brood cells ofO. bicornis.