Field Assessment of Floral Resources and Pollinator Abundance Across Six Vegetable Crops

Abstract

Pollinators play a crucial role in global biodiversity, providing essential ecosystem services such as crop pollination. However, their abundance and diversity have been gradually decreasing in recent years. Despite increasing interest in sustainable agriculture, information on vegetable crops that attract insect pollinators remains limited. We hypothesize that variation in floral traits among vegetable crop cultivars, especially nectar volume, nectar sugar concentration, and pollen characteristics, significantly influences visitation patterns and species composition. To test this, we evaluated multiple cultivars of six vegetable crops (cowpea, sweet potato, eggplant, green bean, mustard, and chickpea) over two years, focusing on five key pollinator groups (honey bees, bumble bees, carpenter bees, sweat bees, and wasps). Cowpea and sweet potato consistently attracted the most pollinators, whereas chickpea attracted the fewest. In 2022, nectar volume was highest in sweet potato (16.45 ± 0.37 µL) and lowest in chickpea (1.18 ± 0.75 µL). Similarly, in 2023, sweet potato recorded the highest nectar volume (8.33 ± 2.95 µL), and chickpea the lowest (0.02 ± 0.01 µL). However, chickpea (31.00 ± 1.58 °Bx) and mustard (30.10 ± 1.12 °Bx) recorded the highest nectar sugar concentration in both years, and chickpea and eggplant produced significantly more pollen grains. A significant positive correlation was observed between nectar volume and pollinator abundance. Comprehensively, this two-year study demonstrates the complex relationship between floral traits and pollinator preferences. These findings offer growers practical guidance on selecting vegetable intercrops that attract specific pollinators, thereby enhancing pollination services, supporting biodiversity, and improving the yield of pollinator-dependent crops.

1. Introduction

Pollinators play a vital role in supporting ecosystem services by enabling plant reproduction [1]. Globally, they pollinate about 30,000 plant species [2], with insects accounting for roughly 75% of cultivated flowering plants [3]. Among insect pollinators, the honey bee is the most significant, contributing an estimated USD 10 billion annually to the US economy [4]. However, the roles of other bee pollinators (bumble bees, carpenter bees, and sweat bees), flies, and wasps are also essential and should not be overlooked. Pollinator effectiveness often depends on the crop species, and the complementary effect of wild and managed bees enhances crop pollination, resulting in higher overall pollination effectiveness [5,6,7,8,9]. Despite their critical role, the abundance and diversity of insect pollinators have markedly declined in recent years [10]. Habitat loss, which causes a reduction in flower diversity and quantity, is considered one of the primary causes of pollinator decline, as it reduces foraging resources [11]. Other factors include improper pesticide use, which can harm non-target organisms [12]; the spread of pests and diseases, such as the Varroa mite that transmits disease to honey bees [13]; and the effects of climate change [14].

Pollinator decline has significantly impacted the production of pollinator-dependent crops (PDCs) such as squash (Cucurbita spp.), watermelon (Citrullus lanatus L.), apple (Malus spp.), cherry (Prunus avium L.), and pumpkin (Cucurbita pepo L.) [15,16,17], raising concerns as the global population is projected to reach 8.5 billion by 2030. Incorporating flowering plants into cropping systems that include pollinator-dependent crops has been shown to attract a greater diversity and abundance of pollinators, increase crop yield [17,18,19], and provide floral resources (nectar and pollen), which are essential food sources for pollinators [20]. Several popular vegetables grown and consumed in the US and globally, especially cowpeas (Vigna unguiculata [L.] Walp), sweet potato (Ipomoea batatas [L.] Lam), green beans (Phaseolus vulgaris L.), eggplant (Solanum melongena L.), chickpeas (Cicer arietinum L.), and mustard (Brassica juncea L.), attract diverse pollinators [15] even though they are not strictly pollinator-dependent for fruit or seed production. For example, in field studies, chickpea attracted honey bees, wasps, and moths [21], and cowpeas attracted diverse pollinators, with honey bees and sweat bees being the most abundant [22,23,24,25]. Similarly, sweet potatoes were recommended as an attractive resource for bumble bees [26,27], and green beans were reported to be highly attractive to honey bees [28], while mustard attracted large numbers of honey bees [29].

The attractiveness of plants to pollinators depends on several factors, including floral traits such as the volatiles released, nectar and pollen produced, as well as the morphological and phenological characteristics [30,31,32,33,34]. Even within the same species, different cultivars can vary significantly in their ability to attract pollinators because of differences in these traits. Among these traits, Fowler et al. [35] reported that greater nectar resources, such as increased nectar volume and concentration, positively influenced pollinator choice. Additionally, bumble bees were reported to collect pollen with higher protein content and more essential amino acids [36]. It has been reaffirmed that the combined effects of these traits can significantly influence pollinator abundance and, consequently, crop quality [6,37]. Nonetheless, the variability of these resources and their impact on insect foraging preferences are poorly understood, especially among crop cultivars. Studies on pollinator abundance and diversity reported results from experiments involving a single crop cultivar [17,23,25] rather than multiple cultivars. These limitations highlight the need for designing studies that capture the complex interactions that may be present in diverse farming systems.

We conducted a two-year field experiment with multiple cultivars of six vegetable crops (cowpea, sweet potato, eggplant, green bean, mustard, and chickpea) to examine how floral traits affect pollinator abundance and diversity. First, we recorded the most frequent pollinators (honey bees (Apis spp.), bumble bees (Bombus spp.), carpenter bees (Xylocopa spp.), sweat bees (Lasioglossum spp.), and wasps (Vespa spp.)) across all cultivars of the six crops. Then, we measured the nectar and pollen resources provided by each cultivar. Finally, we investigated how specific floral traits, such as nectar volume, nectar sugar concentration, and pollen characteristics, shape the preferences of different pollinators. We hypothesize that variation in floral traits among vegetable crop cultivars, especially nectar volume, nectar sugar concentration, and pollen traits, will significantly impact pollinator abundance and composition. Overall, this study aims to offer insights into how crop cultivar and floral resource traits can be strategically utilized to promote pollinator diversity and enhance pollination services in vegetable farming systems.

2. Materials and Methods

2.1. Field Experimental Setup

Field experiments were conducted during the summers of 2022 and 2023 at the North Carolina Agricultural and Technical State University (NCAT) Research Farm in Greensboro, North Carolina, USA (36.0586243° N, 79.7358932° W) in Greensboro, North Carolina, USA, within the Piedmont region, and were characterized by Ultisol soil, most commonly the Cecil series. The farm had a fine, mixed, active, thermic Ultic Hapludalfs Alfisol (58% sand, 13% silt, and 29% clay) soil type. Six vegetable crop species, each comprising at least two cultivars, were evaluated: sweet potato (Ipomoea batatas, var. Murasaki, Korean, Garnet, and Red Japanese), cowpea (Vigna unguiculata, var. Dixilee, Whippoorwill, Pinkeye Purple Hull, and Pennyrile), eggplant (Solanum melongena, var. Galine F1, Hansel F1, Nadia F1, and Black Beauty), green bean (Phaseolus vulgaris, var. Blue Lake Bush Bean, Provider Bush Bean, Bountiful Bush Bean, and Tendergreen Bush Bean), mustard (Brassica juncea, var. Florida Broadleaf and Southern Giant Curled), and chickpea (Cicer arietinum, var. Desi and Kabuli), totaling 20 cultivars. Due to germination failure of the Kabuli chickpea in 2022, only 19 cultivars were evaluated that year. The selected crops are widely grown across the United States and around the world, and are known to affect visitation patterns [15,38].

2.2. Greenhouse Propagation and Transplanting

Cowpea, green beans, and chickpeas were directly sown in the plots, whereas eggplant, mustard, and sweet potato were transplanted and prepared accordingly for transplanting. In both years, eggplant and mustard were sown in a greenhouse using commercial growing mix (Sun Gro Horticulture Distribution Inc., Agawam, MA, USA). Eggplant seeds were sown on 21 April 2022, and mustard on 29 May 2023. Seedlings were watered daily and fertilized biweekly with a fish-based natural organic fertilizer (5-1-1; Alaska Fish Fertilizer, Walnut Creek, CA, USA). After four weeks of growth, the seedlings were transplanted into the field. Sweet potato propagation began on 11th February each year. Sweet potato tubers were soaked in tap water under laboratory conditions for five weeks, while changing water every two days to promote sprouting. After sprouting, the tubers were transferred to pots filled with commercial growing mix (Sun Gro Horticulture Distribution Inc., Agawam, MA, USA) grown in the greenhouse for an additional six weeks to generate planting slips for field transplanting.

2.3. Experimental Design and Setup

In both years, each cultivar of the six vegetable crops was planted on two rows, each 5 m long, in a Randomized Complete Block Design (RCBD) with four replications. Planting was done on different dates to synchronize flowering based on phenological information from previous studies [17,18,19,39,40,41,42,43,44]. In 2022, sweet potato slips, mustard seedlings, cowpea seeds, and chickpea seeds were planted on 15 June, while eggplant and green bean were planted on 22 June. In 2023, adjustments were made based on observations from the previous year, particularly the delayed flowering of cowpea. Thus, cowpea was planted earlier on 29 May, followed by sweet potato and eggplant on 6 June, green bean and chickpea on 13 June, and finally mustard on 15 June. Fertilizer (NPK 10:10:10) was applied to eggplant, sweet potato, and mustard 21 days after planting at the rate of 10 g/plant. No insecticide application was made in either year. However, in 2023 the mustard plants were attacked by the Harlequin bugs (Murgantia histrionica (Hahn)) which were hand-picked to avoid plant damage.

2.4. Sampling for Pollinator Abundance and Diversity

At the onset of flowering, three sampling methods, including direct visual counts, pan traps, and vacuum, were used in 2022, while in 2023, direct visual counts, pan traps, and sticky cards were used to evaluate pollinator abundance and diversity. In 2023, yellow sticky cards were used instead of vacuum sampling because, from the identified samples collected in 2022, we realized that vacuum sampling was not effective in trapping pollinators. Five key and common pollinators from the insect order Hymenoptera, including honey bees (Apis spp.), bumble bees (Bombus spp.), carpenter bees (Xylocopa spp.), sweat bees (Lasioglossum spp.), and wasps (Vespa spp.), were recorded on all cultivars of the six vegetable crops (sweet potato, cowpea, eggplant, green bean, mustard, and chickpea) using three sampling methods, as these groups were the most frequently observed and documented in our investigations over the past five years and they significantly enhanced yield and promoted sustainable agriculture [17,18,19].

2.4.1. Pollinator Sampling Using Direct Visual Counts

During each census, two adjacent 5 m rows were observed sequentially, with each row surveyed for 60 s, resulting in a total of 120 s of observation per census, and the number of each pollinator type was counted and recorded. In both years, observations were carried out every Monday from 9:00 am to 12:00 noon for five weeks.

2.4.2. Pollinator Sampling Using Pan Traps and Sticky Cards

Pan traps were set up every Tuesday at 8:00 am and removed on Wednesday at 8:00 am, and sticky cards were set up on Thursday at 8:00 am and removed on Friday at 8:00 am as described by [17]. The trapped insects were transferred into vials containing 70% alcohol and stored at room temperature in the laboratory pending identification. The process was repeated every week for five weeks. In the laboratory, the trapped pollinators were counted, identified, and recorded.

2.4.3. Pollinator Sampling Using a Vacuum

In 2022, a RYOBI Rechargeable handheld vacuum (Bioquip products, Rancho Dominguez, CA, USA) was moved in a zigzag pattern for 60 s on the two rows of each cultivar. The trapped pollinators were immediately transferred into a Ziplock bag (size 6 X 82 MIL) and stored in the refrigerator at 4 °C before identification. Vacuum sampling was done weekly for five weeks. Pollinators were identified and recorded using a microscope (AmScope Stereozoom trinocular microscope, SZMT2 Series, WF10X/20; United Scope LLC, Irvine, CA, USA).

Visually sampled insects were identified by their common names, while those sampled with pan traps, vacuum, and sticky cards were identified using dichotomous keys [45] and other freely accessible online resources, namely Bug Guide (https://bugguide.net/), Discover Life (https://www.discoverlife.org), and also from GBIF (GBIF.org (2022, 2023, 2024), GBIF Home Page. Available from: https://www.gbif.org). All websites were accessed continuously as needed from 10 December 2022 through to 30 May 2024.

2.5. Sampling for Nectar Quantity and Nectar Sugar Concentration

For each vegetable cultivar, five plants were randomly selected, and two flower buds were bagged per plant. At bloom, a total of 40 flowers per cultivar were bagged. Bags were removed from the flowers, and the flowers were harvested within an hour from 9.00 am, and nectar was collected by removing the calyx and inserting a microcapillary tube into the nectaries at the base of the corolla. A caliper was placed next to the capillary tube (from bottom to top), and the height of the nectar column in the microcapillary tube was measured in millimeters. The volume of the nectar in the capillary tube drawn by capillary action was determined using the formula πr²h (π = 22/7, r = radius of capillary tube, and h = height of nectar in the tube) and recorded. Nectar sugar concentration (quality) was measured by dropping 20 µL of nectar per cultivar on a handheld analog refractometer (Cole Parmer Instruments, Vernon Hills, IL, USA), and the Brix reading (°Bx) was recorded. However, because of the small flower size of the mustard and chickpea cultivars, 5 µL of nectar was used instead. Nectar extraction from eggplant flowers was attempted, but the very low volume per flower made measurement challenging. The handheld analog refractometer requires a minimum liquid volume for accurate readings, and squeezing the flowers to obtain this volume could introduce contaminants; therefore, nectar data for eggplant were not included.

2.6. Sampling for Pollen Quantity and Quality

For each crop cultivar, five random plants were selected, and two flower buds were bagged per plant. At bloom, a total of 40 flowers per cultivar were removed and placed in bags. Flowers were removed from the bags, and the anthers of each flower were harvested using a pair of scissors. Anthers were placed in gelatin capsules and stored in Ziploc bags containing silica desiccant to prevent mold growth, then refrigerated at 4 °C [46] for analysis. In total, anthers were harvested from 40 flowers per cultivar. For each crop cultivar, all the anthers collected from each flower were suspended in 3 mL of Ampha buffer. The suspension was gently shaken by hand to release the pollen grains into the solution. The suspension was filtered using a 50 µm Ampha filter for cowpea, eggplant, sweet potato, mustard, and chickpea, and a 100 µm Ampha filter for green beans. The pollen grains in the filtrate were counted using the Ampha Z32 Neutec pollen counter (Neutec Group Inc., Farmingdale, NY, USA) according to the manufacturer’s instructions. The Ampha Z32 uses impedance flow cytometry (IFC) to measure cells’ electrical properties. The IFC system uses a microfluidic chip that enables measurements in the radio frequency range from 0.1 to 30 MHz using alternating current (AC). At a chosen frequency, data of cell size, membrane capacitance, cell concentration, and cytoplasmic conductivity of single cells are simultaneously obtained and related to biological key parameters such as cell viability and membrane permeability [47]. Thus, Ampha Z32 data acquisition and processing algorithms display results on a scatterplot as counts of non-viable and viable pollen, with the percentage of viable pollen to the top right and non-viable pollen to the top left. The number of pollen grains per flower and the percentage of viable pollen for each flower in each cultivar were recorded.

2.7. Data Analysis

The two study years (2022 and 2023) were analyzed separately due to significant differences in sampling methods, floral composition, and environmental conditions, all of which can significantly affect pollinator community structure and behavior. Year-to-year variability in pollinator abundance and diversity is well-documented and may be driven by fluctuations in climate, floral resource availability, land-use changes, and other ecological factors [48]. Furthermore, in 2022, pollinator abundance was assessed using a vacuum sampling method. Although this approach was initially chosen to enhance the specificity of pollinator capture, we later found it to be less effective in representing the full spectrum of pollinators. Therefore, in 2023, we adopted yellow sticky cards, which provided a more consistent and representative measure of pollinator activity, which can bias species detection rates and distort community composition measures [49]. Additionally, due to an unforeseen germination failure in 2022, only one chickpea cultivar was available for evaluation that year, whereas two cultivars were successfully established and assessed in 2023. Differences in crop cultivar representation (i.e., one cultivar in 2022 vs. two in 2023) may further alter floral traits, thereby influencing pollinator visitation patterns. Despite taking utmost care in experimental planning and execution, these discrepancies necessitated analyzing each year’s data independently. This approach preserves the methodological rigor within each year and minimizes the risk of confounding treatment effects with sampling artifacts or inconsistencies in crop representation.

Means were separated using the ZINB regression and the Kruskal–Wallis + Bonferroni correction factor. The nectar volume, nectar sugar concentration, pollen count, and pollen viability were summarized as means per flower for each variety. For factors with only two levels (such as the comparisons within crops for mustards and chickpeas), the Wilcoxon test (W) was adopted. Linear regression was used to assess relationships between floral traits and pollinator abundance. Pollinator diversity and evenness indices were computed using Shannon’s diversity index (H′ = −∑ [Pi × ln(Pi)]) and evenness (E = H′/ln k), where Pi is the proportion of each pollinator species and k is the total number of pollinator types recorded per crop. All data were analyzed in R (R Core Team 2023). All plots were generated using Prism (version 10.5.0 for Windows, GraphPad Software Inc., San Diego CA, USA).

3. Results

3.1. Pollinator Counts

Throughout the 2022 sampling period, a total of 1372 pollinators were counted on the six vegetable crops using the three sampling methods. Among the pollinators, bumble bees constituted the highest percentage of pollinators (41.5%), followed by wasps (28.1%), sweat bees (25.6%), honey bees (3.8%), and carpenter bees (0.9%). The distribution of pollinator types across crops was as follows: honey bees were most abundant on cowpea (82.7%), followed by sweet potato (7.7%), eggplant (5.8%), and green bean (3.8%). Bumble bees were highest on sweet potato (70.9%), followed by cowpea (17.4%), eggplant (7.2%), green bean (4.0%), and mustard (0.5%). Carpenter bees were most abundant on cowpea (46.2%), followed by green bean (30.1%), mustard (15.4%), and chickpea (7.7%). Sweat bees were highest on cowpea (51.6%), followed by eggplant (15.4%), sweet potato (13.1%), green bean (9.9%), mustard (6.3%), and chickpea (3.7%). Wasps were predominantly found on cowpea (91.9%), followed by green bean (3.3%), eggplant (3.1%), sweet potato (1.0%), and mustard (0.5%).

Overall, no significant difference in pollinator counts was observed among the cultivars of each crop when the three sampling methods were combined (Figure 1b–g). However, total pollinator counts across the crops vary significantly (χ² = 249.1, df = 5; p < 0.001), with cowpea recording the highest count (35.5 ± 1.93) and chickpea the lowest (13.1 ± 1.56) (Figure 1a).

Among various individual sampling methods, the distribution of pollinators across the six vegetable crops varied significantly when using visual counts (χ2 = 365.8, df = 5; p < 0.001) (Figure 2a), pan traps (χ2 = 93.4, df = 5; p < 0.001) (Figure 2b), and vacuum sampling (χ2 = 117.7, df = 5; p < 0.001) (Figure 2c) in 2022. From direct visual counts, pollinator abundance among the four eggplant cultivars was significantly different (χ2 = 12.0, df = 3; p < 0.01), with the highest counts recorded on the cultivar Hansel F1 (1.20 ± 0.42). Data collected from pan traps and vacuum sampling showed no significant difference among the cultivars of all the other crops.

During the 2023 sampling period, a total of 2680 pollinators, including honeybees, bumble bees, carpenter bees, sweat bees, and wasps, were counted from the six vegetable crops from all sampling methods. Among these, sweat bees accounted for the largest proportion (46.8%), followed by wasps (25.6%), bumble bees (20.0%), and honey bees (7.4%) and carpenter bees (0.3%). The distribution of pollinator types across crops was as follows: honey bees were most abundant on cowpea (88.3%), followed by sweet potato (3.6%), eggplant (3.6%), mustard (2.5%), and green bean (2.0%). Bumble bees were highest on sweet potato (54.5%), followed by cowpea (29.7%), green bean (6.3%), eggplant (5.9%), chickpea (2.2%), and mustard (1.3%). Carpenter bees were most abundant on eggplant (42.9%), followed by cowpea (28.6%), and mustard (28.6%). Sweat bees were highest on cowpea (34.2%), followed by sweet potato (22.5%), eggplant (14.0%), mustard (12.7%), green bean (12.1%), and chickpea (4.3%). Wasps were predominantly found on cowpea (52.7%), followed by green bean (16.3%), eggplant (11.8%), sweet potato (11.2%), mustard (6.3%), and chickpea (1.7%). Overall, and similar to 2022, pollinator counts differed significantly among the six vegetable crops (χ² = 323.3, df = 5; p < 0.001), with cowpea having the highest number of pollinators (14.07 ± 0.95) and chickpea the lowest (1.95 ± 0.28) (Figure 3a). Similar to 2022, no significant differences were recorded in the number of pollinators among cultivars within each crop when all three sampling methods were combined (Figure 3b–g).

Across individual sampling methods in 2023, pollinator counts again differed significantly among the six vegetable crops when using direct visual counts (χ² = 282.1, df = 5; p < 0.001) (Figure 2d), pan traps (χ² = 78.2, df = 5; p < 0.001) (Figure 2e), and sticky cards (χ² = 80.6, df = 5; p < 0.001) (Figure 2f). From direct visual counts, pollinator abundance differed significantly between the two mustard cultivars (W = 129.5, p < 0.05), with the Southern Giant Curl cultivar recording higher numbers (1.75 ± 0.51) than Florida Broad Leaf (0.55 ± 0.21). Pollinator counts from pan traps and sticky cards showed no significant differences among the cultivars of the other vegetable crops.

In terms of diversity, the highest diversity and evenness indices in 2022 were recorded for cowpea (H′ = 0.8; E = 0.5), followed by eggplant (H′ = 0.6; E = 0.4), green bean (H′ = 0.4; E = 0.3), sweet potato (H′ = 0.2; E = 0.3), and mustard (H′ = 0.07; E = 0.04). Similarly, in 2023, cowpea had the highest diversity and evenness indices (H′ = 1.5; E = 0.9), followed by eggplant (H′ = 1.2; E = 0.7), sweet potato (H′ = 1.0; E = 0.6), and mustard (H′ = 0.9; E = 0.6).

3.2. Nectar Volume and Nectar Sugar Concentration

Eggplant was excluded from this calculation because it produced very small quantities of nectar, insufficient for accurate measurement with the handheld refractometer. Nectar volume differed significantly among the vegetable crops in 2022 (χ² = 824.7, df = 4; p < 0.01) (Figure 4a). Sweet potato produced the highest nectar volume (16.45 ± 0.37 µL), while chickpea had the lowest (1.18 ± 0.75 µL). Significant differences in nectar volume were also observed among the four sweet potato cultivars (χ² = 17.8, df = 3; p < 0.001) (Figure 5g), the four cowpea cultivars (χ² = 28.6, df = 3; p < 0.001) (Figure 5h), and the four green bean cultivars (χ² = 22.9, df = 3; p < 0.001) (Figure 5i). In 2023, nectar volume again varied significantly (χ2 = 737.7, df = 4; p < 0.001) among all vegetable crops, with the highest in sweet potato (8.33 ± 2.95 µL) and the lowest in chickpea (0.02 ± 0.01 µL) (Figure 4c). There were significant differences in nectar volume among the four green bean cultivars (χ2 = 38.3, df = 3; p < 0.001) (Figure 5c) and between the two chickpea cultivars (W = 1144.5, p = 0.0002) (Figure 5e). However, the remaining crop cultivars showed no significant variation among each other in both 2022 (Figure 5f) and 2023 (Figure 5a,b,d).

The nectar sugar concentration in 2022 varied significantly among the six vegetable crops (χ2 = 603.2, df = 5; p < 0.001), with chickpea (31.00 ± 1.58 °Bx) and mustard (30.10 ± 1.12 °Bx) recording the highest nectar sugar concentration (Figure 4b). Among the crop cultivars, there was a significant variation in nectar sugar concentration among the four cowpea cultivars (χ2 = 13.8, df = 3; p = 0.003) (Figure S1b), the four cultivars of sweet potato (χ2 = 3.7, df =3; p = 0.3) (Figure S1a) and the four cultivars of green bean (χ2 = 8.4, df =3; p = 0.04) (Figure S1c). Similarly, in 2023, there was a significant difference (χ2 = 118.9, df = 5; p < 0.001) in nectar sugar concentration among the crops with mustard (30.11 ± 1.68 °Bx) and chickpea (30.96 ± 0.94 °Bx) recording the highest concentration (Figure 4d). Among the crop cultivars, nectar sugar concentration varied significantly among the cultivars of sweet potato (χ2 = 12.8, df = 3; p < 0.05) (Figure S2a), cowpea (χ2 = 20.5, df = 3; p < 0.001) (Figure S2b), and green bean (χ2 = 13.3, df = 3; p = 0.004) (Figure S2c). The remaining crop cultivars showed no significant variation among each other in both 2022 (Figure S1a,d) and 2023 (Figure S2d,e).

3.3. Pollen Grain Count

Pollen grain counts per flower varied significantly among the six vegetable crops (χ2 = 187.5, df = 5; p < 0.001) in 2022, with eggplant producing significantly more pollen grains (Figure 6a). Significant differences in pollen count were recorded among the eggplant cultivars (χ2 = 38.49, df = 3; p < 0.001) (Figure 6d) and mustard cultivars (W = 380, p < 0.001) (Figure 6f). The pollen grain count per flower in 2023 varied significantly among the six vegetable crops (χ2 = 279.9, df = 5; p < 0.001), with eggplant recording the highest (10,889.28 ± 1063.93) (Figure 7a). Among the vegetable crop cultivars, pollen grain count was significantly different among the sweet potato cultivars (χ2 = 59.2, df = 3; p < 0.001) (Figure 7b), cowpea cultivars (χ2 = 21.3, df = 3; p < 0.001) (Figure 7c), and between the two mustard cultivars (W = 295.5, df = 3; p = 0.01) (Figure 7f). The remaining crop cultivars showed no significant variation among themselves in both 2022 (Figure 6b,c,e) and 2023 (Figure 7d,e,g).

Pollen viability was significantly different (χ2 = 14.6, df = 5; p = 0.01) among the six vegetable crops in 2022, with eggplant recording a higher percentage of viable pollen grains (82.21 ± 1.69%) (Figure 8a). There was a significant difference in the percentage of viable pollen among the four green bean cultivars (χ2 = 8.5, df = 3; p = 0.04) (Figure 8e) and the two mustard cultivars (W = 90, p = 0.002) (Figure 8f). In 2023, the percentage of viable pollen varied significantly among crops (χ2 = 61.1, df = 5; p < 0.001) (Figure 9a). The percentage of viable pollen varied significantly among cowpea cultivars (χ2 = 21.1, df = 3; p < 0.001) (Figure 9c) and eggplant cultivars (χ2 = 15.4, df = 3; p < 0.001) (Figure 9d). However, the other crop cultivars in both 2022 (Figure 8b–d) and 2023 (Figure 9b,e–g) showed no significant differences among themselves.

3.4. Relationship Between Nectar Traits and Pollinator Abundance

Using pollinator counts from all sampling methods in 2022, there was a significant positive relationship between nectar volume and pollinator counts (r = 0.395, p < 0.001) (Figure 10a). In contrast, no significant relationship was observed between nectar sugar concentration and pollinator counts (r = 0.122, p = 0.295) (Figure 10b). Similarly, in 2023, a positive relationship was found between nectar volume and pollinator counts (r = 0.480, p < 0.001) (Figure 10c). Consistent with the previous year, nectar sugar concentration again showed no significant relationship with pollinator counts (r = 0.040, p = 0.729) (Figure 10d).

3.5. Relationship Between Pollen Traits and Pollinator Abundance

There was no significant relationship between the number of pollen grains and pollinator abundance in 2022 (r = −0.189, p = 0.102) (Figure 10g). The relationship between the percentage of viable pollen and the number of pollinators was positive but non-significant (r = 0.046, p = 0.694) in 2022 (Figure 10e). In 2023, pollen grain count had a negative relationship (r = −0.298, p < 0.01) with pollinator abundance (Figure 10h), and pollen viability showed a positive relationship (r = 0.238, p < 0.05) with pollinator counts (Figure 10f).

4. Discussion

In this study, we examined multiple cultivars of six vegetable crops (cowpea, sweet potato, eggplant, green bean, mustard, and chickpea) over a two-year period to understand the field-level interactions between pollinators and the crops. Although the literature suggests that all our experimental crops provide floral resources for pollinators, a few studies reported that some of these crops actually attract pollinators or are involved in crop pollination [15,21,22,23,24,25,26,27,50,51,52]. However, in most of these studies, only single cultivars were evaluated [15,23,27,50,51,52]. Here, for the first time, a field study was conducted to assess pollinator preferences when multiple crop cultivars were planted alongside each other. We also investigated how specific floral traits, including nectar volume, nectar sugar concentration, and pollen quality, affect the foraging preferences of pollinators.

Our findings indicate that, overall, bumble bees were the primary and most abundant pollinators on sweet potatoes, supporting previous reports that bumble bees are the primary pollinators of sweet potatoes [26,27]. Although sweet potato rank as the seventh most important food crop worldwide [53], this study is among the first to document the pollinator abundance and diversity of sweet potatoes. Cowpeas, on the other hand, consistently attracted the highest number of pollinators, primarily honeybees, aligning with our earlier results [17,25]. On the other hand, the contributions of carpenter bees, sweat bees, and wasps should not be overlooked, as these pollinators have been shown to be efficient in pollination and can enhance crop yield [6,8,9,18]. For instance, Borchardt et al. [9] reported that wasps are comparable to bees in terms of plant interactions, body pollen load, and single-visit pollen deposition, suggesting that wasps can carry and potentially deliver substantial pollen to specific plants. In the present study, wasps were also among the most abundant pollinator groups, and we speculate that they may contribute significantly to overall pollination effectiveness alongside other pollinator groups. In our previous investigation, all five recorded pollinator groups significantly increased squash yield by 155% in intercropping systems [18].

In this study, chickpea attracted no pollinators according to direct visual counts and had the lowest number of pollinators from all sampling methods combined. Yet, it recorded the highest abundance of Apis spp. when grown as a single cultivar [21]. These findings suggest that pollinator preferences vary depending on whether foraging is restricted to a single crop or involves multiple species or cultivars. It has been suggested that when grown as a single crop, pollinator preference is mainly driven by within-crop variation in floral traits, with pollinators maximizing energy gain and efficiency through repeated foraging on similar flowers (flower constancy) [54,55]. Conversely, when multiple attractive crops are grown side by side, active selection among species occurs, and preferences are influenced by factors such as reward quality, accessibility, and nutritional needs [56,57]. This discrepancy may be due to the broader context: while earlier studies examined chickpea grown as a single cultivar, our experiment placed it next to more attractive crops, likely decreasing its visitation. However, the detection of pollinators through pan traps, vacuum sampling, and sticky cards shows that, although less appealing than cowpea and sweet potato, these less attractive crops (eggplant, green bean, and mustard) still add to the overall pollinator diversity [58,59,60]. For example, the higher Shannon diversity (H′) and evenness (E) indices recorded for eggplant indicate that this crop supported a more balanced and diverse pollinator community.

It is worth noting that environmental factors such as temperature and relative humidity can affect pollinator abundance. Data obtained from the State Climate Office of North Carolina [61] indicate that the average temperature during the sampling period in both years was similar, with a record of 23.5 °C in 2022 and 23.7 °C in 2023; however, the average relative humidity was higher (78.2%) in 2022 compared to 76.1% in 2023. The elevated relative humidity in 2022 was accompanied by reduced pollinator abundance compared to 2023, when relative humidity was 2.1% lower. It is most likely that the relative humidity during the sampling period in 2023 could have been optimal for pollinator foraging activity.

Nectar, a key attractant for pollinators and an essential energy source supporting foraging and pollination, varied among the six crops in our study. We found that floral visits were highest in sweet potato and cowpea, which had the highest nectar volumes. In contrast, chickpeas and mustard exhibited a nectar trade-off, characterized by higher sugar concentrations but lower volumes. In addition, nectar volume, rather than concentration, was positively correlated with the abundance of pollinators. We can deduce that this difference suggests that pollinators may prioritize nectar availability over concentration, with foraging decisions being more influenced by the quantity of nectar than its quality. These findings support the conclusion by [38] that pollinator abundance is more strongly influenced by total nectar availability than by nectar sugar concentration alone. Similarly, ref. [62] reported that the amount of nectar produced by onion flowers was the most important factor influencing honey bee attraction, not the sugar concentration in the nectar. Pollinators’ preference is generally influenced by a complex trade-off between the energetic reward of the nectar and the physical effort required to acquire it, which is primarily determined by nectar viscosity. The reduced number of pollinators recorded on chickpea and mustard, despite their nectar having the highest concentration, may be due to the viscosity of their nectar, which makes it harder for uptake, in addition to the relatively small size and color of their flowers (yellow for mustard and white/pink for chickpea) compared to those of cowpeas and sweet potatoes, with larger flowers characterized by purple or violet flowers, a trait previously shown to attract bee pollinators [17,63]. Furthermore, the specific mechanics of a pollinator’s feeding style dictate which characteristic yields the best return on its energy investment. Thus, flowers with larger nectar volumes but low to moderate concentration can be acquired faster and with less energy expenditure, thereby increasing foraging efficiency [64]. Overall, these results highlight that nectar quantity, rather than concentration alone, is the primary driver of pollinator preference and floral attractiveness.

Although the abundance of pollinators did not differ significantly among crop cultivars, except for mustard in 2022, which experienced pest attack and damage, there was a significant difference in nectar sugar concentration and volume among crop cultivars, but not among the two mustard cultivars. Our findings contrast with earlier studies that reported intra-specific variation in pollinator abundance among strawberry and onion cultivars [62,65]. Despite differences in nectar rewards observed among cultivars, the abundance of pollinators may not have varied significantly due to several interacting factors. For instance, pollinators often rely on visual and olfactory cues such as color, shape, and scent rather than nectar traits alone, and these floral signals may be relatively conserved across cultivars [66]. As a result, pollinators perceive them as functionally equivalent resources and distribute their foraging more evenly [67,68]. Additionally, many bees exhibit floral constancy and foraging thresholds, whereby once a resource exceeds a minimal reward level, pollinators would not discriminate among available options [54,69].

Proteins are essential to pollinators for their productivity and vitality, especially among bee species. In this study, significant variation in pollen grain count was observed among the six vegetable crops. Additionally, a non-significant and negative relationship was observed between pollen grain count and pollinator abundance, while a positive relationship was found between the percentage of viable pollen and pollinator abundance. The observed negative relationship between pollen grain count and pollinator abundance could reflect differences in floral reward strategies and pollinator foraging dynamics. Consequently, crops that produce abundant pollen per flower may require fewer visits to meet pollinator nutritional needs, resulting in lower visitation rates despite high pollen output [57]. This is evident in our findings, as eggplant and chickpea, which produced the most pollen grains, recorded the least pollinators. Conversely, crops with lower pollen production, such as cowpea in this study, attracted more frequent visitation as pollinators need to visit more flowers to meet foraging requirements [70]. We speculate that pollinators may avoid flowers with excessive pollen, as it could interfere with their ability to collect nectar. This could be a disadvantage for plants, as it may limit the quantity of pollen transferred and potentially decrease their reproductive success.

Viable pollen is often associated with healthier flowers and plants, which can produce stronger floral signals that enhance the attractiveness of pollinators [65]. In this study, variations in pollen viability were observed among crop cultivars, a finding similar to reports on different strawberry cultivars, which were suggested to be relevant to pollinator interactions and, in turn, influence visitation rates [71,72]. However, in this study, the variation in pollen viability among cultivars did not influence the abundance of pollinators on the cultivars. However, a positive relationship was observed between pollen viability and pollinator abundance across crops planted side by side, likely reflecting the pollinators’ ability to detect and respond to higher-quality reproductive rewards. Our findings support the notion that pollinators prioritize pollen quality over quantity during foraging [64,71,73,74,75]. Sweet potato and cowpea, the crops that attracted the highest number of pollinators in this study, were among those with the highest percentage of viable pollen. Pollen with greater viability often correlates with higher nutritional content (e.g., proteins, lipids, amino acids), which is critical for bee development and reproduction [57,74]. Pollinators may preferentially forage on crops offering viable, nutrient-rich pollen, as this increases their foraging efficiency and colony fitness [76]. Thus, it is plausible that, based on our findings, crops producing a large volume of non-viable pollen are less attractive than those offering fewer but more viable pollen grains. The spatial arrangement of the six vegetable crops grown adjacent to each other may also have contributed to this pattern. The presence of adjacent floral resources can affect pollinator distribution, behavior, and foraging decisions. Studies have shown that the type and arrangement of neighboring crops influence the composition and foraging dynamics of pollinator communities. For example, Reynolds et al. [77] reported that the abundance and diversity of pollinators in natural vegetation remnants varied depending on whether adjacent crops were insect-attracting (such as canola) or not (such as wheat). Such landscape-level effects could partly explain the observed pollinator distribution patterns in this study. Moreover, pollinators, particularly bees, are equipped with highly sensitive olfactory systems that enable them to detect floral traits, including pollen viability, through scent cues [78,79,80]. This supports the hypothesis that the five pollinator groups observed in this study may have been capable of detecting and preferring viable pollen, reinforcing the observed association between pollen viability and pollinator abundance.

Our results indicate moderate correlations between floral traits (nectar volume, nectar concentration, and pollen availability) and pollinator abundance. While these trends are consistent with the idea that floral rewards can influence pollinator behavior, the moderate strength of the correlations suggests that additional factors may also contribute. One potential limitation of our study is the relatively small size of the experimental plots. Larger plots or multiple experimental sites could provide more extensive floral displays, potentially enhancing detection of floral traits by pollinators and yielding stronger or more consistent patterns in pollinator abundance. Future research should focus on multi-location studies with larger plots while controlling for other variables that may influence pollinator abundance and diversity. This is an area we are currently exploring.

5. Conclusions

This study aimed to examine how floral traits influence pollinator abundance and diversity in vegetable crops and to provide new insights into how crop cultivars and floral resource traits can be strategically utilized to promote pollinator diversity and enhance pollination services in vegetable farming systems. Our two-year field findings underscore the intricate relationship between crops and pollinators, demonstrating that specific floral traits, particularly nectar and pollen characteristics, strongly influence pollinator abundance and diversity. Cowpea and sweet potato consistently attracted the highest numbers of pollinators, whereas chickpea was the least visited, likely due to its less favorable floral traits. These results suggest that prioritizing highly pollinator attractive crops such as cowpea and sweet potato, both as standalone crops and within intercropping systems, can enhance pollination services, support pollinator populations, and improve crop yields. Incorporating such pollinator-friendly strategies provides a practical pathway toward resilient, biodiverse, and productive agroecosystems, ultimately contributing to long-term food security in the face of global pollinator declines.