Higher northern latitude normalized difference vegetation index and growing season trends from 1982 to 1999

Normalized difference vegetation index data from the polar-orbiting National Oceanic and Atmospheric Administration meteorological satellites from 1982 to 1999 show significant variations in photosynthetic activity and growing season length at latitudes above 35°N. Two distinct periods of increasing plant growth are apparent: 1982–1991 and 1992–1999, separated by a reduction from 1991 to 1992 associated with global cooling resulting from the volcanic eruption of Mt. Pinatubo in June 1991. The average May to September normalized difference vegetation index from 45°N to 75°N increased by 9% from 1982 to 1991, decreased by 5% from 1991 to 1992, and increased by 8% from 1992 to 1999. Variations in the normalized difference vegetation index were associated with variations in the start of the growing season of –5.6, +3.9, and –1.7 days respectively, for the three time periods. Our results support surface temperature increases within the same period at higher northern latitudes where temperature limits plant growth. Received: 25 October 2000 / Revised: 20 August 2001 / Accepted: 22 August 2001     

Use of bioclimatic indexes to characterize phenological phases of apple varieties in Northern Italy

The research was designed to characterize the phenological behaviour of different apple varieties and to compare different bioclimatic indexes in order to evaluate their adaptability in describing the phenological phases of fruit species. A field study on the requirement for chilling units (winter chilling requirement) and the accumulation of growing degree hours of 15 native apple cultivars was carried out in a fruit-growing area in North West Italy (Cuneo Province, Piedmont). From 1991 to 1993, climatic data were collected at meteorological stations installed in an experimental orchard (Verzuolo, Cuneo). Four methods were compared to determine the winter chilling requirement: Hutchins, Weinberger-Eggert, Utah and North Carolina. The Utah method was applied to determine the time when the chilling units accumulated become effective in meeting the rest requirements. A comparison of the different methods indicated that the Weinberger-Eggert method is the best: as it showed the lowest statistical variability during the 3 years of observations. The growing degree hour requirement (GDH) was estimated by the North Carolina method with two different base temperatures: 4.4°C and 6.1°C. More difficulties were met when the date of rest completion and the beginning of GDH accumulation was determined. The best base temperature for the estimation of GDH is 4.4°C. Phenological and climatic characterizations are two basic tools for giving farmers and agricultural advisors important information about which varieties to choose and which are the best and the most correct cultivation practices to follow. Received: 25 October 2000 / Revised: 6 August 2001 / Accepted: 6 August 2001     

Spatial and temporal variability of the phenological seasons in Germany from 1951 to 1996

Various indications for shifts in plant and animal phenology resulting from climate change have been observed in Europe. This analysis of phenological seasons in Germany of more than four decades (1951–96) has several major advantages: (i) a wide and dense geographical coverage of data from the phenological network of the German Weather Service, (ii) the 16 phenophases analysed cover the whole annual cycle and, moreover, give a direct estimate of the length of the growing season for four deciduous tree species. After intensive data quality checks, two different methods – linear trend analyses and comparison of averages of subintervals – were applied in order to determine shifts in phenological seasons in the last 46 years. Results from both methods were similar and reveal a strong seasonal variation. There are clear advances in the key indicators of earliest and early spring (?0.18 to ?0.23 d y^?1) and notable advances in the succeeding spring phenophases such as leaf unfolding of deciduous trees (?0.16 to ?0.08 d y^?1). However, phenological changes are less strong during autumn (delayed by + 0.03 to + 0.10 d y^?1 on average). In general, the growing season has been lengthened by up to ?0.2 d y^?1 (mean linear trends) and the mean 1974–96 growing season was up to 5 days longer than in the 1951–73 period. The spatial variability of trends was analysed by statistical means and shown in maps, but these did not reveal any substantial regional differences. Although there is a high spatial variability, trends of phenological phases at single locations are mirrored by subsequent phases, but they are not necessarily identical. Results for changes in the biosphere with such a high resolution with respect to time and space can rarely be obtained by other methods such as analyses of satellite data.     

An examination of the relationship between flowering times and temperature at the national scale using long-term phenological records from the UK

This paper examines the mean flowering times of 11 plant species in the British Isles over a 58-year period, and the flowering times of a further 13 (and leafing time of an additional 1) for a reduced period of 20 years. Timings were compared to Central England temperatures and all 25 phenological events were significantly related (P<0.001 in all but 1 case) to temperature. These findings are discussed in relation to other published work. The conclusions drawn from this work are that timings of spring and summer species will get progressively earlier as the climate warms, but that the lower limit for a flowering date is probably best determined by examining species phenology at the southern limit of their distribution. Received: 21 October 1999 / Revised: 27 January 2000 / Accepted: 27 January 2000     

Phenology in central Europe – differences and trends of spring phenophases in urban and rural areas

In order to examine the impacts of both large-scale and small-scale climate changes (urban climate effect) on the development of plants, long-term observations of four spring phenophases from ten central European regions (Hamburg, Berlin, Cologne, Frankfurt, Munich, Prague, Vienna, Zurich, Basle and Chur) were analysed. The objective of this study was to identify and compare the differences in the starting dates of the pre-spring phenophases, the beginning of flowering of the snowdrop (Galanthus nivalis) and forsythia (Forsythia sp.), and of the full-spring phenophases, the beginning of flowering of the sweet cherry (Prunus avium) and apple (Malus domestica), in urban and rural areas. The results indicate that, despite regional differences, in nearly all cases the species studied flower earlier in urbanised areas than in the corresponding rural areas. The forcing in urban areas was about 4 days for the pre-spring phenophases and about 2 days for the full-spring phenophases. The analysis of trends for the period from 1951 to 1995 showed tendencies towards an earlier flowering in all regions, but only 22% were significant at the 5% level. The trends for the period from 1980 to 1995 were much stronger for all regions and phases: the pre-spring phenophases on average became earlier by 13.9 days/decade in the urban areas and 15.3 days/decade in the rural areas, while the full-spring phenophases were 6.7 days earlier/decade in the urban areas and 9.1 days/decade earlier in the rural areas. Thus rural areas showed a higher trend towards an earlier flowering than did urban areas for the period from 1980 to 1995. However, these trends, especially for the pre-spring phenophases, turned out to be extremely variable. Received: 21 October 1999 / Revised: 5 April 2000 / Accepted: 25 April 2000     

Starting dates and basic temperatures in phenological observations of plants

 Several methods have been used in plant phenology to find the best starting date in spring and the best threshold or basic temperature for growth and development of perennial plants. In the present paper the date giving the highest correlation coefficient for development to various phenophases, in relation to 24-hourly mean air temperatures was chosen as the best starting value in further analyses. For many woody plants this date was very often found to be 1 April based on phenological and climatological observations at about 60 sites in western Norway (at about 61°N). The early flowering species Corylus avellana and Salix caprea and the early leaf-bud breaking Prunus padus seemed to start development earlier in Spring, while the late sprouting Fraxinus excelsior showed the highest correlation coefficient using 15 April. If daytime temperatures were used in the calculations, the ”best” starting date was generally found to be later than for the 24-hour mean temperatures. This variation must be seen as resulting from the different basic temperatures for the development of various species. Estimates of basic temperatures in various species and periods may be given, for example by finding the value having the least variance in heat sums or by various regression analyses. A technique has been developed to minimise the influence of significance of correlation, using the intercept with the temperature axis by merging the two least squares rectilinear regression lines that can be found between plant development and mean air temperature (from the estimated best starting date) at r=+1 or –1. The basic temperature seemed to vary from –5.9°C for leaf-bud break of P. padus to 5.5°C for leaf-bud break of F. excelsior, with basic temperatures of several other woody plants having intermediate values. These values are compared with those found by other methods. Received: 26 May 1998 / Accepted: 7 September 1998     

Geographic pattern of genetic variation in photosynthetic capacity and growth in two hardwood species from British Columbia

Geographic patterns of intraspecific variations in traits related to photosynthesis and biomass were examined in two separate common garden experiments using seed collected from 26 Sitka alder (Alnus sinuata Rydb.) and 18 paper birch (Betula papyrifera Marsh.) populations from climatically diverse locations in British Columbia, Canada. Exchange rates of carbon dioxide and water vapour were measured on 2-year-old seedlings to determine the maximum net instantaneous photosynthetic rate, mesophyll conductance, stomatal conductance, and photosynthetic water use efficiency. Height, stem diameter, root and shoot dry mass and fall frost hardiness data were also obtained. Mean population maximum photosynthetic rate ranged from 10.35 to 14.57 μmol CO₂ m^–2 s^–1 in Sitka alder and from 14.76 to 17.55 μmol CO₂ m^–2 s^–1 in paper birch. Based on canonical correlation analyses, populations from locations with colder winters and shorter (but not necessarily cooler) summers had higher maximum photosynthetic rates implying the existence of an inverse relationship between leaf longevity and photosynthetic capacity. Significant canonical variates based on climatic variables derived for the seed collection sites explained 58% and 41% of variation in the rate of photosynthesis in Sitka alder and paper birch, respectively. Since growing season length is reflected in date of frost hardiness development, an intrinsic relationship was found between photosynthetic capacity and the level of fall frost hardiness. The correlation was particularly strong for paper birch (r=–0.77) and less strong for Sitka alder (r=–0.60). Mean population biomass accumulation decreased with increased climate coldness. These patterns may be consequential for evaluation of the impact of climate change and extension of the growing season on plant communities. Received: 12 July 1999 / Accepted: 24 November 1999     

The phenological calendar of Estonia and its correlation with mean air temperature

A phenological calendar with 24 phenological phases was compiled for three meteorological stations in Estonia for the period 1948–1996. We analysed the length of the vegetation period, the order of the phenological phases, and the variability and possible changes for two incremental climate change scenarios (±2°C), and compared the results with examples of extreme years. The statistically significant linear trends show that the spring and summer-time phenological phases occurred earlier and the autumn phases moved later during the study period. The study of extreme (minimum and maximum) years shows that 70% of the earliest dates of the 24 phases studied have occurred during the last 15 years with an absolute maximum in 1990 with 8 extreme phases. The phenological spring has shortened (slope –0.23), the summer period has lengthened (slope 0.04), and the autumn has lengthened too. The length of the growing season, determined by the vegetation of rye, has shortened (slope –0.09), which could be the result of changing agricultural technology. The correlation between the starting dates of the phenological phases with the air temperature of the previous 2–3 months is relatively high (0.6–0.8). Studying the +2°C and –2°C scenarios and values for the extreme years shows that, in the case of short variations of air temperature, the phenological development remains within the limits of natural variation. Received: 29 November 1999 / Revised: 15 May 2000 / Accepted: 16 May 2000     

Temperature data for phenological models

In an arid environment, the effect of evaporation on energy balance can affect air temperature recordings and greatly impact on degree-day calculations. This is an important consideration when choosing a site or climate data for phenological models. To our knowledge, there is no literature showing the effect of the underlying surface and its fetch around a weather station on degree-day accumulations. In this paper, we present data to show that this is a serious consideration, and it can lead to dubious models. Microscale measurements of temperature and energy balance are presented to explain why the differences occur. For example, the effect of fetch of irrigated grass and wetting of bare soil around a weather station on diurnal temperature are reported. A 43-day experiment showed that temperature measured on the upwind edge of an irrigated grass area averaged 4% higher than temperatures recorded 200 m inside the grass field. When the single-triangle method was used with a 10°C threshold and starting on May 19, the station on the upwind edge recorded 900 degree-days on June 28, whereas the interior station recorded 900 degree-days on July 1. Clearly, a difference in fetch can lead to big errors for large degree-day accumulations. Immediately after wetting, the temperature over a wet soil surface was similar to that measured over grass. However, the temperature over the soil increased more than that over the grass as the soil surface dried. Therefore, the observed difference between temperatures measured over bare soil and those over grass increases with longer periods between wettings. In most arid locations, measuring temperature over irrigated grass gives a lower mean annual temperature, resulting in lower annual cumulative degree-day values. This was verified by comparing measurements over grass with those over bare soil at several weather stations in a range of climates. To eliminate the effect of rainfall frequency, using temperature data collected only over irrigated grass, is recommended for long-term assessment of climate change effects on degree-day accumulation. In high evaporative conditions, a fetch of at least 100 m of grass is recommended. Our results clearly indicate that weather stations sited over bare soil have consistently higher degree-day accumulations. Therefore, especially in arid environments, phenology models based on temperature collected over bare soil are not transferable to those based on temperature recorded over irrigated grass. At a minimum, all degree-day-based phenology models reported in the literature should clearly describe the weather station site. Received: 25 October 2000 / Revised: 10 July 2001 / Accepted: 10 July 2001     

Changes in phenology of the locust tree (Robinia pseudoacacia L.) in Hungary

 Intense research is being carried out on climate variability and change and the estimation and detection of anthropogenic effects. In addition to statistical methods, the use of plants, as biological indicators is becoming more popular as they are sensitive to environmental conditions. In this article we compare maps of the flowering dates of the locust tree (Robinia pseudoacacia L.) for three different time intervals between 1851 and 1994. The maps revealed noticeable shifts of dates, of approximately 3–8 days, towards earlier flowering. This change is related to the average temperature of spring (15 March–15 May), via a simple statistical model that is accurate enough to be able to quantify phenological changes and to calculate the corresponding warming. The model developed can estimate spring mean temperature using phenological data from R. pseudoacacia L. with an accuracy of 0.2° C. Estimates of mean temperature based on phenological changes are compared to climatic series. This comparison emphasizes the possibility of using R. pseudoacacia. L. as a bio-indicator. Estimates of temperature changes are also given. Received: 5 August 1996 / Revised: 14 April 1997 / Accepted: 11 November 1997     

Phenology and the changing pattern of bird migration in Britain

 Britain has a huge mass of data on the timing of bird migration, although much of this remains in a form that is inaccessible for immediate scientific study. In this paper, I undertake a preliminary examination of data from a number of historical and current sources. Among these are the Marsham family records from Norfolk, dating back to 1736, and post-World War II records from coastal bird observatories. The majority of the examined time series displayed a negative relationship to temperature indicating a tendency for the earlier arrival of the studied birds in warmer springs. In addition to temperature effects, trends through time and some sampling effects (through population size) have become apparent. Identification and curation of data sources and further analysis is still required to produce a clearer picture of climate effects on bird migration timing and on subsequent bird population dynamics. Received: 26 May 1998 / Accepted: 28 October 1998     

Phenological stages of natural species and their use as climate indicators

 The objectives of this paper are to: (1) present 10 years of phenological data for nine natural species growing in a Mediterranean-type climate, (2) present threshold temperatures that were derived for the computation of cumulative degree-days (CDD), and (3) evaluate the sensitivity of the nine natural species to weather variability. The study was conducted at the Phenological Research Garden of Oristano, Sardinia, Italy, during the period 1986–96. The observations were made on five typical Mediterranean species and four species that are typical of higher latitudes. The mean annual pattern of phenological events and the CDD from 1 January are given for each development stage. Temperature thresholds were evaluated by comparing the standard deviation about the mean number of days in the development period for each species. A good relationship between timing of phenophase occurrence and temperature was observed for the Mediterranean species, which were little affected by variations in rainfall. Phenological development of the non-native species was affected by springtime rainfall. Accepted: 28 October 1998     

Growing seasons of Nordic mountain birch in northernmost Europe as indicated by long-term field studies and analyses of satellite images

The phenophases first greening (bud burst) and yellowing of Nordic mountain birch (Betula pubescens ssp.tortuosa, also called B. p. ssp. czerepanovii) were observed at three sites on the Kola Peninsula in northernmost Europe during the period 1964–2003, and at two sites in the trans-boundary Pasvik-Enare region during 1994–2003. The field observations were compared with satellite images based on the GIMMS-NDVI dataset covering 1982–2002 at the start and end of the growing season. A trend for a delay of first greening was observed at only one of the sites (Kandalaksha) over the 40 year period. This fits well with the delayed onset of the growing season for that site based on satellite images. No significant changes in time of greening at the other sites were found with either field observations or satellite analyses throughout the study period. These results differ from the earlier spring generally observed in other parts of Europe in recent decades. In the coldest regions of Europe, e.g. in northern high mountains and the northernmost continental areas, increased precipitation associated with the generally positive North Atlantic Oscillation in the last few decades has often fallen as snow. Increased snow may delay the time of onset of the growing season, although increased temperature generally causes earlier spring phenophases. Autumn yellowing of birch leaves tends towards an earlier date at all sites. Due to both later birch greening and earlier yellowing at the Kandalaksha site, the growing season there has also become significantly shorter during the years observed. The sites showing the most advanced yellowing in the field throughout the study period fit well with areas showing an earlier end of the growing season from satellite images covering 1982–2002. The earlier yellowing is highly correlated with a trend at the sites in autumn for earlier decreasing air temperature over the study period, indicating that this environmental factor is important also for autumn phenophases.     

Determining the growing season of land vegetation on the basis of plant phenology and satellite data in Northern China

The objectives of this study are to explore the relationships between plant phenology and satellite-sensor-derived measures of greenness, and to advance a new procedure for determining the growing season of land vegetation at the regional scale. Three phenological stations were selected as sample sites to represent different climatic zones and vegetation types in northern China. The mixed data set consists of occurrence dates of all observed phenophases for 50–70 kinds of trees and shrubs from 1983 to 1988. Using these data, we calculated the cumulative frequency of phenophases in every 5-day period (pentad) throughout each year, and also drew the cumulative frequency distribution curve for all station-years, in order to reveal the typical seasonal characteristics of these plant communities. The growing season was set as the time interval between 5% and 95% of the phenological cumulative frequency. Average lengths of the growing season varied between 188 days in the northern, to 259 days in the southern part of the research region. The beginning and end dates of the surface growing season were then applied each year as time thresholds, to determine the corresponding 10-day peak greenness values from normalized difference vegetation index curves for 8-km² pixels overlying the phenological stations. Our results show that, at the beginning of the growing season, the largest average greenness value occurs in the southern part, then in the northern, and finally the middle part of the research region. In contrast, at the end of the growing season, the largest average greenness value is measured in the northern part, next in the middle and lastly the southern part of the research region. In future studies, these derived NDVI thresholds can be applied to determine the growing season of similar plant communities at other sites, which lack surface phenological data. Received: 29 November 1999 / Revised: 14 March 2000 / Accepted: 15 March 2000     

Determining degree-day thresholds from field observations

 This paper compares several methods for determining degree-day (°D) threshold temperatures from field observations. Three of the methods use the mean developmental period temperature and simple equations to estimate: (1) the smallest standard deviation in °D, (2) the least standard deviation in days, and (3) a linear regression intercept. Two additional methods use iterations of cumulative °D and threshold temperatures to determine the smallest root mean square error (RMSE). One of the iteration methods uses a linear model and the other uses a single triangle °D calculation method. The method giving the best results was verified by comparing observed and predicted phenological periods using 7 years of kiwifruit data and 10 years of cherry tree data. In general, the iteration method using the single triangle method to calculate °D provided threshold temperatures with the smallest RMSE values. However, the iteration method using a linear °D model also worked well. Simply using a threshold of zero gave predictions that were nearly as good as those obtained using the other two methods. The smallest standard deviation in °D performed the worst. The least standard deviation in days and the regression methods did well sometimes; however, the threshold temperatures were sometimes negative, which does not support the idea that development rates are related to heat units. Received: 26 May 1998 / Accepted: 28 October 1998     

Effects of seasonal rainfall on radial growth in two tropical tree species

 Seasonal drought may limit primary productivity in most of the tropics, but the determinants of tree growth are not well known. A 10-year study of the deciduous trees Cochlospermum vitifolium (Willd.) Spreng. (Cochlospermaceae) and Cnidoscolus spinosus Lundell (Euphorbiaceae) in southwestern México showed radial growth to be highly correlated (both r>0.85) only with precipitation during an interval of <2 months in the mid-wet season. Growth was not affected by total annual precipitation or by an early starting or late ending of the wet season, or by heavy rainfall in the dry season. Annual mean girth increments ranged from 0.03 to 3.31 cm and −0.1 to 2.01 cm, respectively. The best model for growth (r ²>0.85) was a linear combination of mid-summer precipitation (positive coefficient) and total precipitation over the previous 2 years (negative coefficient). Comparison with other species showed heterogeneous responses of wood production to climate variation, and suggests that the range of functional types of dry forest trees is still unknown. Received: 9 September 1996 / Accepted: 4 March 1997     

The impact of growing-season length variability on carbon assimilation and evapotranspiration over 88 years in the eastern US deciduous forest

 Recent research suggests that increases in growing-season length (GSL) in mid-northern latitudes may be partially responsible for increased forest growth and carbon sequestration. We used the BIOME-BGC ecosystem model to investigate the impacts of including a dynamically regulated GSL on simulated carbon and water balance over a historical 88-year record (1900–1987) for 12 sites in the eastern USA deciduous broadleaf forest. For individual sites, the predicted GSL regularly varied by more than 15 days. When grouped into three climatic zones, GSL variability was still large and rapid. There is a recent trend in colder, northern sites toward a longer GSL, but not in moderate and warm climates. The results show that, for all sites, prediction of a long GSL versus using the mean GSL increased net ecosystem production (NEP), gross primary production (GPP), and evapotranspiration (ET); conversely a short GSL is predicted to decrease these parameters. On an absolute basis, differences in GPP between the dynamic and mean GSL simulations were larger than the differences in NEP. As a percentage difference, though, NEP was much more sensitive to changes in GSL than were either GPP or ET. On average, a 1-day change in GSL changed NEP by 1.6%, GPP by 0.5%, and ET by 0.2%. Predictions of NEP and GPP in cold climates were more sensitive to changes in GSL than were predictions in warm climates. ET was not similarly sensitive. First, our results strongly agree with field measurements showing a high correlation between NEP and dates of spring growth, and second they suggest that persistent increases in GSL may lead to long-term increases in carbon storage. Received: 26 May 1998 / Accepted: 6 July 1998     

Sex-differentiated migration patterns, protandry and phenology in North European songbird populations

This study aims to investigate causes and mechanisms controlling protandrous migration patterns (the earlier breeding area arrival of males relative to females) and inter-sexual differences in timing of migration in relation to the recent climate-driven changes in phenology. Using standardised ringing data from a single site for eight North European migratory passerines collected throughout 22 years, we analysed sex-differentiated migration patterns, protandry and phenology of the entire populations. Our results show protandrous patterns for the first as well as later arriving individuals for all studied species. Males show more synchronous migration patterns compared to females and, hence, first arriving females followed males more closely than later arriving individuals. However, we found no inter-sexual differences in arrival trends as both sexes advance spring arrival over time with the largest change for the first arriving individuals. These findings seem in support of the “mate opportunity” hypothesis, as the arrival of males and females is strongly coupled and both sexes seem to compete for early arrival. Changes in timing of arrival in males and females as a response to climatic changes may influence subsequent mating decisions, with subsequent feedbacks on population dynamics such as reproductive success and individual fitness. However, during decades of consistent earlier spring arrival in all phases of migration we found no evidence of inter-sexual phenological differences.     

Phytophenological trends in Switzerland

Nation-wide phenological observations have been made in Switzerland since 1951. In addition to these observation programmes, there are two very long phenological series in Switzerland: leaf bud burst of horse-chestnut trees has been observed in Geneva since 1808 and full flowering of cherry trees in Liestal since 1894. In addition to the presentation of these two long phenological series, trends for 896 phenological time series have been calculated with national data from 1951 to 1998. The earlier bud burst of horse-chestnut trees in Geneva can be attributed mainly to the city effect (warmth island). This phenomenon was not observed with the cherry tree flowering in Liestal. A clear trend towards earlier appearance dates in spring and a weak tendency towards later appearance dates in autumn could be shown with data from the national observation network. It must be noted that different phenophases and plant species react differently to various environmental influences. Received: 25 October 2000 / Revised: 9 May 2001 / Accepted: 4 June 2001     

Resource overlap and possible competition between honey bees and wild bees in central Europe

Evidence for interspecific competition between honey bees and wild bees was studied on 15 calcareous grasslands with respect to: (1) foraging radius of honey bees, (2) overlap in resource use, and (3) possible honey bee effects on species richness and abundance of flower-visiting, ground-nesting and trap-nesting wild bees. The grasslands greatly differed in the number of honey bee colonies within a radius of 2 km and were surrounded by agricultural habitats. The number of flower-visiting honey bees on both potted mustard plants and small grassland patches declined with increasing distance from the nearest apiary and was almost zero at a distance of 1.5–2.0 km. Wild bees were observed visiting 57 plant species, whereas honey bees visited only 24 plant species. Percentage resource overlap between honey bees and wild bees was 45.5%, and Hurlbert’s index of niche overlap was 3.1. In total, 1849 wild bees from 98 species were recorded on the calcareous grasslands. Neither species richness nor abundance of wild bees were negatively correlated with the density of honey bee colonies (within a radius of 2 km) or the density of flower-visiting honey bees per site. Abundance of flower- visiting wild bees was correlated only with the percentage cover of flowering plants. In 240 trap nests, 1292 bee nests with 6066 brood cells were found. Neither the number of bee species nor the number of brood cells per grassland was significantly correlated with the density of honey bees. Significant correlations were found only between the number of brood cells and the percentage cover of shrubs. The number of nest entrances of ground-nesting bees per square metre was not correlated with the density of honey bees but was negatively correlated with the cover of vegetation. Interspecific competition by honey bees for food resources was not shown to be a significant factor determining abundance and species richness of wild bees. Received: 22 March 1999 / Accepted: 24 September 1999