Bees are principal contributors to pollination services in both the natural and anthropogenic habitats (Klein et al., 2007). Insect pollination is a key factor in seed and fruit yield, ensuring the reproduction of plant species and providing food for humans and wildlife. In the past decade, one of the fastest emerging traditions in cities has been urban agriculture, including beekeeping (Peters, 2012; Marìn et al., 2019). Urban agriculture provides such multiple benefits as food production, relaxation and the saving of cultural heritage (Marìn et al., 2019). However, such activities require proper “bee-friendly” management of urban areas, including the selection and use of plants for example valuable nectar and pollen resources. Urban areas contain patches of unused land including road verges or ruderal areas and such amenity areas as parks, lawns and roofs. This is a potential to alter the management of this land and to create an attractive flower-rich landscape thus increasing food pasture for insect pollinators, including bees (Matteson, Grace & Minor, 2013; Larcher et al. 2017). Ornamental plants can provide insects with considerable amounts of nectar and pollen and contribute to the diversity of food resources within cities (Garbuzov & Ratnieks, 2014; Masierowska et al., 2018; Jachuła et al., 2019).
The warming Urban Heat Island effect, increased water stress and pollution contribute to cities’ specific environmental conditions (Pickett et al., 2011) and may affect plant-pollinators relations. Spring-flowering plants tend to advance their blooming in urban habitats (Neil & Wu, 2006). Honeybees, queen bumblebees and solitary bees are active early in the year in cities (Hicks et al., 2016). Spring is a period of high nutritional demand for spring-emerging bees, which lose a lot of weight during over-wintering in response to warm winters (Fründ, Zieger, & Tscharntke, 2013). Thus, an adequate quality and quantity of early floral resources is vital for their development and activity. Spring forage limitation is currently considered as a possible driver of the negative impact of urbanization on bees (Harrison & Winfree, 2015). Early in the year (April-May), nectar and pollen are usually provided by native perennial weeds (Denisow, 2011), which are problematic from the aesthetic point of view. Thus, incorporating ornamental trees, shrubs or flowers into green space landscaping can enhance both the availability of early season resources to urban bees and the aesthetic value of the urban landscape.
The attractiveness of plants to insects depends on the traits of flower morphology and colour, pollen and nectar quality (Garbuzov & Ratnieks, 2014), flowering pattern and floral-display size (Thompson, 2001) and these should be considered when selecting plants for bee pasture. The aim of this study was to evaluate three shrubs
Since studies on the apiarian value of ornamental
The study was conducted in 2014–2015 in the UMCS Botanical Garden, Lublin, SE Poland (51°16′N, 22°30′E; 200 m a.s.l.). The local climate is characterized by an average annual air temperature of 7.9°C and average annual precipitation of 556 mm. The average length of the growing season is 209 days (Rysiak & Czarnecka, 2018).
The studied
The experimental plants were exposed to sun grown on loess-originated soil. Four shrubs per species were used in this study.
The onset and termination of blooming were recorded. The abundance of flowering (number of flowers · shrub−1) was estimated by multiplying the mean number of flowers · inflorescence−1 (n = 20 inflorescences recorded per plant), the number of inflorescences · shoot−1 and the number of shoots · shrub−1. The flower and inflorescence lifespan was established as a period between a flower bud opening and perianth closing, and between the first flower opening and the last flower closing in an inflorescence, respectively. The number of randomly chosen flowers and inflorescences is shown in Tab. 1. For the examination of the inflorescence flowering pattern in 2015, four inflorescences were randomly chosen on each studied plant and marked prior to the opening of the first flowers. On the consecutive days of blooming, the number of newly opened flowers in each inflorescence was counted until blooming terminated. The flowering pattern was expressed as the percentage of newly opened flowers on the successive flowering days in relation to the total number of flowers formed in the inflorescence.
Flowering period and abundance, lifespan of flowers and inflorescences in the studied species throughout the years of study
Species | Year | Flowering period (days) | No. of flowers:* | Lifespan (days) of: * | ||
---|---|---|---|---|---|---|
inflorescence−1 | shrub−1 (thous.) | flower | inflorescence | |||
2014 | 17 April–9 May (23) | 7.6 ± 1.8 (20) | 8.7 ± 0.9 (4) | 7.2 ± 1.2 (20) | 15.7 ± 3.8 (15) | |
2015 | 25 April–12 May (28) | 8.0 ± 1.7 (20) | 19.5 ± 2.3 (4) | 5.4 ± 0.7 (20) | 10.1 ± 1.3 (18) | |
mean | (26) | 7.8C ± 1.8 (40) | 14.1AB ± 6.0 (8) | 6.3B ± 1.3 (40) | 12.6B ± 3.9 (33) | |
2014 | 25 March–14 April (20) | 19.1 ± 4.0 (20) | 5.2 ± 2.5 (4) | 8.6 ± 1.2 (20) | 19.3 ± 4.4 (18) | |
2015 | 11 April–10 May (30) | 18.5 ± 4.0 (20) | 5.6 ± 1.7 (4) | 6.0 ± 0.7 (20) | 14.4 ± 1.8 (18) | |
mean | (25) | 18.8B ± 4.4 (40) | 5.4B ± 1.9 (8) | 7.3A ± 1.6 (40) | 16.9A ± 4.1 (36) | |
2014 | 23 April–20 May (28) | 30.5 ± 8.0 (20) | 18.5 ± 9.7 (4) | 7.5 ± 1.5 (20) | - | |
2015 | 28 April–26 May (29) | 40.0 ± 19.3 (20) | 92.9 ± 2.0 (4) | 6.8 ± 2.4 (20) | 19.0 ± 3.4 (20) | |
mean | (29) | 35.2A ± 15.4 (40) | 55.7A ± 4.2 (8) | 7.1AB ± 2.1 (40) | 19.0A ± 3.4 (20) | |
Mean for the year of study | 2014 | (24) | 19.1a ± 10.7 (60) | 10.8a ± 7.9 (24) | 7.7a ± 1.4 (60) | |
2015 | (29) | 22.2a ± 17.5 (60) | 39.3a ± 41.4 (24) | 6.0b ± 1.6 (60) |
Data are means ± SD (numbers in brackets indicate number of flowers, inflorescences and shrubs, respectively). Means in columns with the same letter do not differ significantly among species (capital letters) (Kruskall-Wallis ANOVA, H - test) and between years of study (small letters) (Kolmogorov-Smirnov test); p > 0.05.
In 2014 and 2015, during the peak blooming of the plants, a spectrum and abundance of bee visitors were noted. Five-minute observations were made three times every hour from 9.00 to 17.00 (GMT + 2 hrs) for two to three consecutive days in sunny and non-windy weather. The observations were performed in four marked areas on a shrub canopy, each sized 0.12 m2.
Nectar production was studied using the Jabłoński pipette method (2002). Prior to sample collection, flower buds in the middle part of randomly selected inflorescences were marked and then whole inflorescences were isolated with tulle isolators to exclude insect visits. Accumulated nectar was sampled from the marked
Nectar and pollen production in flowers of the studied species throughout the years of study
Species | Year | Nectar amount · flower−1 (mg) | Nectar sugar concentration (% w/w) | Nectar sugar amount · flower−1 (mg) | Pollen amount · flower−1 (mg) |
---|---|---|---|---|---|
2014 | 5.3 ± 1.8 (22) | 28.3 ± 5.2 (22) | 1.5 ± 0.6 (22) | 0.4 ± 0.0 (6) | |
2015 | 4.6 ± 2.1 (11) | 31.7 ± 9.1 (11) | 1.4 ± 0.6 (11) | 0.4 ± 0.1 (6) | |
mean | 5.0A ± 1.9 (33) | 29.5A ± 6.8 (33) | 1.5A ± 0.6 (33) | 0.4B ± 0.1 (12) | |
2014 | 4.4 ± 1.5 (10) | 38.5 ± 4.5 (10) | 0.9 ± 0.3 (10) | 0.4 ± 0.1 (6) | |
2015 | 2.3 ± 0.9 (11) | 30.1 ± 3.2 (11) | 0.7 ± 0.3 (11) | 0.4 ± 0.0 (6) | |
mean | 3.2B ± 1.5 (21) | 34.1A ± 4.9 (21) | 0.8B ± 0.3 (21) | 0.4B ± 0.0 (12) | |
2014 | 4.6 ± 1.7 (21) | 35.4 ± 11.4 (21) | 1.5 ± 0.4 (21) | 0.8 ± 0.1 (6) | |
2015 | 5.6 ± 2.5 (19) | 30.7 ± 7.4 (19) | 1.7 ± 0.8 (19) | 0.8 ± 0.1 (6) | |
mean | 5.1A ± 2.2 (40) | 33.2A ± 28.6 (40) | 1.6A ± 0.6 (40) | 0.8A ± 0.1 (12) |
Data are means ± SD (numbers in brackets indicate number of collected samples). Untransformed data are presented. Means in columns with the same letter do not differ significantly among species (HSD Tukey's test, p > 0.05)
Pollen production was determined using the protocol of Warakomska (1972). Each year, six samples for each species were collected and eighteen samples in total. A single sample contained fifty mature stamens, which were placed on previously tarred watch glasses and dried in a SUP-65G dryer (Wamed, Poland) at 30°C for several days. Pollen was rinsed from open anthers using 70% ethanol. The pollen samples were then dried and weighed on a balance (WA 34 PRL T A14 MERA-KFM, Poland), which allowed the calculation of the mass of air-dried pollen. The results were expressed in mg · flower−1.
Moreover, sugar and pollen yield · shrub−1 (g) offered to insects was estimated with the use of the number of flowers · shrub−1 already formed during a given growing season on four plants of each species and mean nectar sugar amount · flower−1 and pollen amount · flower−1. Each year, twelve samples of sugar and pollen yields calculated for four plants of the three studied species were used to analyse the between-year variation.
All analyses were performed using STATISTICA v.13 (StatSoft Poland, Cracow). Two-way ANOVA tested for differences in nectar amount, nectar sugar concentration, total sugar amount in the nectar · flower−1, nectar sugar yield · shrub−1, pollen amount · flower−1, and pollen yield · shrub−1 between the taxa and years of the study. The percentage values of the nectar sugar concentration were transformed by arcsines. When significant differences were found, the ANOVA was followed by the HSD Tukey test at p = 0.05. Differences in the number of flowers · inflorescence−1, number of flowers · shrub−1, and flower lifespan among the species were tested using the Kruskal-Wallis non-parametric ANOVA and H-test. The Kolmogorov-Smirnov test was applied for seasonal differences in these characteristics.
Detailed data of the seasonal flowering period and abundance are shown in Tab. 1. In the climatic conditions of Lublin, SE Poland, blooming of the studied species lasted from late March to late May. The time of blooming differed among the taxa and between the years of the study. The
The species significantly differed in both the mean number of flowers · inflorescence−1 (H2,109 = 86.18, p < .001) and the mean number of flowers · shrub−1 (H2, 24 = 16.96, p < .001). The highest number of flowers developed in the inflorescences of
The lifespan of the flowers and inflorescences differed significantly among the species (H2, 120 = 7.33, p = .003 and H2, 89 = 29.08, p < .001). The flowers and inflorescences of
Blooming in the inflorescences started with a small number of flowers but peaked quickly between the second and sixth days after first flower opening with a few local maxima before a decline, leading to a skewed distribution (Fig. 1). During the peak in all the species, the daily portion of newly opened flowers exceeded 22% of all flowers developed in an inflorescence.
The observations revealed the presence of honeybees, bumblebees, solitary bees, and dipterans. The proportion of insect groups visiting the flowers of the plant species in the years of the study is shown in Fig. 2. Hymenopterans accounted for more than 79% of all visits. Bumblebees were the principal visitors to
Bees collected both nectar and pollen from the flowers but were more interested in nectar. The microscopic analysis of orange corbicular pollen loads of honeybees foraging on
The characteristics of nectar produced by the flowers of the studied species in the consecutive growing seasons (mean values and number of samples) are shown in Tab. 2. All the parameters of mean nectar amount, nectar sugar concentration and nectar sugar amount · flower−1 found for the three species did not vary significantly between the years of the study (F1, 88 = 0.12, p = .91; F1, 88 = 3.53, p = .06, and F1, 88 = 1.37, p = .26, respectively). However, the investigated species differed significantly in the mean nectar amount · flower−1 (F2, 88 = 14.13, p < .01) and the mean nectar sugar amount · flower−1 (F2, 88 = 14.35, p < .01).
The mean pollen amount produced by a single flower differed significantly among the species (F2, 30 = 202.61, p < 0.01) but not between the years of the study (F1, 30 = 2.13, p = .16) (Tab. 2). Again, the highest amount of pollen was produced by
Early spring is when both managed and wild bees have a high nutritional demand (Dick et al., 2015; Moquet et al., 2015), but also when there is often a shortage of nectar and pollen food (Lipiński, 2010; Denisow, 2011). However, thanks to the proper selection of plants, such urban flora as ornamental shrubs and trees (Somme et al., 2016) can provide bee species with a variety and continuity of floral resources and create good conditions for urban agriculture, including beekeeping. The results of this study expand the list of early-spring flowering shrubs that are useful for urban pollinators, including honeybees and bumblebees.
In all the plants observed in this study, the onset and length of blooming depended on weather conditions prevailing during the growing seasons and were probably affected by urban warming – a phenomenon recently confirmed for Lublin by Rysiak & Czarnecka (2018). At high temperatures, the investigated shrubs, in particular
The studied species produced thousands of flowers gathered in inflorescences. In mass-flowering plants, the inflorescence rather than a single flower is a fundamental unit of pollinators’ attraction, and the display of multiple flowers in the inflorescence increases plants’ attractiveness to flower visitors (Harder et al., 2004). Despite the considerable differences in the number of flowers, the inflorescences of the studied species exhibit quite a similar pattern of flower presentation. They are characterized by continuous flower opening with a peak during the first days of blooming. By displaying numerous flowers from the beginning of bloom, the plant attracts inexperienced bees and encourages experienced ones to revisit (Makino & Sakai, 2007). Thus, it obtains the services of faithful visitors that will continue to visit despite subsequent decreases in the rate of flower production (Thomson, 1985). Moreover, displaying many flowers might increase the likelihood of the plant being discovered by pollinators quickly (Makino & Sakai, 2007).
The flowers in the studied species persisted for five to nine days on average. Their inflorescences lived up to 2.7 weeks. A prolonged flower and inflorescence lifespan appears to be of a great value to floral visitors, as it increases the size and length of floral display and thus extends the food supply (Willmer, 2011).
The studied species bloomed abundantly, in particular
Besides their high aesthetic value, both
The amount of pollen available for insects from a single flower in all the species was small (0.4–0.8 mg) but similar to the values reported for the flowers of
The studied species produced a considerable number of flowers which resulted in a relatively high estimated sugar yield · plant−1 in the range of 4.3–92.9 g and pollen yield · plant−1 of 1.8–44.0 g (on average in both cases,
In conclusion,