Characterization of Pollinators Associated with Cocoa Cultivation and Their Relationship with Natural Effective Pollination

Abstract

Cocoa (Theobroma cacao L.) relies on insect pollination for fruit production, making it essential to understand the pollinators involved. This study aimed to identify the insects associated with cocoa pollination and their relationship with effective natural pollination in a cocoa agroforestry system in Yopal, Colombia. Indirect (wood traps) and direct (manual aspiration in flowers) methods were used to capture pollinators. The number of captured insects was correlated with the percentage of pollinated, fertilized, aborted, and transitioning flowers using Pearson’s correlation coefficient. Additionally, the natural transfer of pollen was assessed by evaluating the coverage of near-opening flowers and staining pollen grains on the stigma under natural conditions. This study identified Diptera from the Ceratopogonidae family, specifically the genera Forcipomyia and Dasyhelea, as key pollinators. The findings highlight the presence of these pollinators in the Yopal cocoa agroforestry system and suggest the need for further pollinator capture and identification efforts on local farms. A positive correlation was found between the number of pollinating insects and the percentage of fertilized flowers, emphasizing the crucial role of these insects in cocoa pollination and the importance of promoting their presence to optimize fruit production.

1. Introduction

Cocoa (Theobroma cacao L.) (Sterculiaceae) is a native species of Central and South America, found especially in the Amazon region, where it has been cultivated for centuries as part of local agricultural traditions [1]. Ecuador, Brazil, Peru, and Colombia are the main cocoa-producing countries in the Americas, and together they account for a significant portion of its global production [2]. In Colombia, cocoa is grown in both monoculture and agroforestry systems, which differ in terms of their production strategies and sustainability approaches. According to [3], these agroforestry systems are often associated with enhanced biodiversity and improved soil quality, supporting more sustainable production methods. Colombia’s favorable climatic and humidity conditions make it suitable for cocoa cultivation across different regions. However, as noted by [4], 60% of the national production is concentrated in Santander, Norte de Santander, and Arauca, highlighting the regional importance of these areas in the country’s cocoa industry.

Although the cultivation conditions vary widely between regions at both the international and national levels, the challenges related to pollination, pollinator management, and crop sustainability are common across most production systems. Therefore, the study of pollinators in agroforestry systems in Colombia not only contributes to local knowledge but also provides valuable information for other regions with similar agroecological conditions. This facilitates the adoption of management strategies that support pollinator conservation and optimize crop yields.

Cocoa flowers are small and develop directly on the stems and branches, forming cauline inflorescences, and, like many other tropical plants, cocoa produces a surplus of flowers, a trait that influences its reproductive strategy [5]. It has been demonstrated that all cocoa genotypes depend on insect pollination for fruit development, as their anthers, covered with sticky pollen, are enclosed within a folded petal, preventing self-pollination and requiring pollinator intervention [1]. Studies have reported that, in certain cocoa-producing regions, approximately 90% of the pod production relies on pollination, highlighting the crucial role of pollinators in the sustainability of cocoa cultivation [6].

The characteristics of its flower (size, arrangement of floral structures, incompatibility in cocoa plants, and sticky pollen) make cocoa dependent on insect pollinators to maintain its production levels [7]. Likewise, according to [5], most cocoa varieties are self-incompatible and, therefore, depend on cross-pollination, mainly by mosquitoes (Diptera: Ceratopogonidae).

Approximately 10% or fewer of the flowers are naturally pollinated, with natural fruit set rates ranging between less than 5% and 40%, primarily due to fruit abortion [8,9,10]. In Ecuador, reference [11] reported a maximum of 12% of flowers being pollinated through natural pollination, while, in Ivory Coast, reference [12] reported an average fruit set rate of 28%.

Insects from the order Diptera, particularly those in the families Ceratopogonidae (e.g., Forcipomyia sp. and Dasyhelea sp.) and Cecidomyiidae (e.g., Clinodiplosis sp.), have been identified as cocoa pollinators. Additionally, insects from the order Homoptera, such as Aphididae (Toxoptera sp. and Aphis sp.), and from the order Thysanoptera, specifically Thripidae, are recognized as potential pollinators across Asia, South America, and North America. Among these, tiny midges of the Ceratopogonidae family are considered the primary pollinators of cocoa due to their specialized morphology which makes them particularly suited for this role [13].

The primary cocoa pollinators, Diptera from the Ceratopogonidae family, have a flight capacity of up to 6 m. However, wind can extend their feeding range to approximately 35 m [14]. These insects fly around cocoa trees looking for flowers, landing on petals, sepals, bracts, and staminodes, where they tend to remain stationary [15].

On the other hand, the genus with the highest abundance and frequency reported in cocoa agroforestry systems is Forcipomyia. It has an excellent pollination capacity due to its small size (a thorax that is 0.16 mm wide and 1.0 mm long), which favors its entry into the interior of the flowers. In the process, the insect leaves a large amount of pollen grains in the flowers, which allows the pollination and effective fertilization of the cocoa flower [16].

Research on cocoa has highlighted the critical role of pollinator abundance in ensuring multiple visits per flower. However, further studies are needed to evaluate the impact of functional diversity on achieving optimal pollination outcomes [1].

It has been demonstrated that cocoa pollinator populations, particularly those of the Forcipomyia genus, are closely related to the leaf litter cover and the presence of decomposing fruit, which provide an essential habitat and essential food sources for their reproduction [17,18,19]. The removal of these substrates due to management practices such as excessive soil cleaning or organic matter removal can negatively impact pollinator density, reducing flower fertilization rates.

Additionally, it has been documented that using cocoa husks as mulch under trees can significantly increase flower production, suggesting a potential benefit of this practice for pollination and crop yields [20]. Similarly, in Ecuador, using banana/plantain pseudostems as a food substrate for pollinators has been shown to improve the fertilization efficiency and increase the pod production of cocoa [21].

Environmental conditions also play a fundamental role in pollinator activity. The temperature and light positively influence floral production, while rainfall is the most significant factor affecting cocoa phenology, directly impacting flower opening and pollinator availability [12,22]. However, in regions where temperature fluctuations are minimal throughout the year, such as in some Caribbean islands, the temperature does not significantly affect the pollinator presence. Nevertheless, prolonged droughts can reduce the pollinator abundance, making cocoa a crop that is vulnerable to climate change [21,23].

Although cocoa can be cultivated in monoculture systems, studies have shown that growing it within agroforestry systems (AFSs) or in association with fruit trees commonly found on cocoa farms—such as orange, banana/plantain, avocado, and tamarind—creates a more favorable environment for pollinator species. These systems offer ecological benefits by enhancing the structural diversity in the shade canopy, regulating wind exposure, and improving the soil quality, ultimately supporting pollinator abundance [21].

These findings highlight the importance of management strategies that promote pollinator habitat conservation and minimize the impact of agricultural practices that may disrupt their activity. Considering that natural pollination limits cocoa yields, it is crucial to recognize pollinator population levels in production systems and their relationship to crop productivity.

Despite the extensive research on cocoa pollination, significant knowledge gaps remain regarding the identity, abundance, and effectiveness of key pollinator species under different environmental conditions and management practices. While previous studies have identified Forcipomyia sp. and other Ceratopogonidae as the primary cocoa pollinators, little is known about their specific roles in cocoa agroforestry systems in Yopal, Colombia. Additionally, the influence of local agroforestry practices on pollinator activity and their impact on natural pollination rates have not been thoroughly evaluated.

Given these gaps, this study seeks to answer the following research questions: What insect species are associated with cocoa pollination in a cocoa agroforestry system in Yopal, Colombia? What is the relationship between the abundance of these pollinators and effective natural pollination? How do local environmental conditions influence pollinator activity and natural pollination rates?

Understanding the insects responsible for pollination in cocoa production systems is crucial for developing conservation strategies that enhance both pollinator populations and crop productivity. This study aimed to identify the insects associated with cocoa pollination and their relationship with effective natural pollination in a cocoa agroforestry system in Yopal, Colombia.

2. Materials and Methods

The research was conducted in a 2.5-hectare agroforestry system (AFS) of cocoa and acacia that is located on the Utopia campus of Universidad de La Salle (UNISALLE) in Yopal, Colombia (5°19′18.75″ N, 72°17′01.4″ W) at an altitude of 248 m above sea level. The AFS experiences a monomodal rainfall pattern, with annual precipitation ranging from 1500 to 2500 mm and temperatures between 20 °C and 36 °C. It is classified as a tropical rainforest biome [24].

The study area included cocoa clones CAU 39, CAU 43, CCN 51, FTA 2, FSA 12, ICS 1, ICS 95, and IMC 67, planted in a 3.5 × 3.5 m trellis system and intercropped with Acacia mangium Wild. Eleven plots, each consisting of six cocoa trees, were established for insect sampling (Figure 1).

Insect captures associated with cocoa pollination were conducted during the 2021 rainy season (October–November), through indirect and direct sampling in flowers. Indirect sampling was performed with a pyramid-type wooden trap [11] (Figure 2) placed on the ground at the center of the evaluated plots. This trap features an upper lateral opening fitted with a transparent plastic hose that connects to a plastic container filled with 70% alcohol. The setup collects potential adult pollinators emerging from the cocoa tree leaf litter. The trap’s base was sealed and covered with leaf litter or soil, and the trap remained in place for one month. Insects were collected every 15 days, with sampling sessions beginning on 15 October and 2 November 2021 (

Table 1. Climatological data from the weather station “El Recreo”, Yopal, Colombia.

Sampling Period	Maximum Temperature (°C)	Minimum Temperature (°C)	Accumulated Precipitation (mm)	Average Relative Humidity (%)
First: 15–29 October 2021	31.37	22.90	37.90	83.92
Second: 2–16 November 2021	30.00	22.54	43.80	85.42

Source. Data taken from “El Recreo” Climate Station, Fedearroz, Yopal, Colombia.

).

The second method of insect collection involved direct sampling from cocoa flowers. Between 7:00 and 11:00 a.m., three open flowers from three trees in each plot were vacuumed using manual aspirators (Figure 3). Each flower was sampled for 10 min to collect all insects that landed on the stamens, staminodes, pistils, or internal parts of the petals. The captured insects were then stored in 70% alcohol, following the methodology outlined by González [6]. The collection was conducted on cool mornings, as cocoa flowers open early in the day and release aromas that attract pollinators. This approach aligns with [25], which estimates that the peak activity of cocoa pollinators occurs between 7:00 a.m. and 12:00 p.m.

After collection, the insects captured through both indirect and direct methods were transported to the entomology laboratory at UNISALLE. There, the specimens were counted and taxonomically characterized. To facilitate detailed identification, the insects were mounted using Hoyer’s medium and photographed. These photographs were sent to Dr. Gustavo Spinelli at the Entomology Division of La Plata Museum for identification at the family and species levels.

Evaluation of Pollinating Insect Presence and Its Relationship with Natural Effective Pollination. To establish the relationship between pollinating insects and natural, effective pollination, five plots with the exact same cocoa clone (CAU 43) were selected from the 11 previously established plots. From each plant, 10 flower buds that were about to open were marked (Figure 4). The transition from bud to flower and from flower to fruit was monitored over 30 days. This monitoring period coincided with the times of both direct and indirect insect capture.

During the daily monitoring of the marked flower buds for each tree in each plot, the percentages of pollinated, fertilized, aborted, and transitioning flowers were determined, following the methodology proposed by [7]. According to [10], unpollinated flowers abscise after 24–36 h without showing signs of senescence. If pollination and fertilization are successful, the ovary increases in size, the pedicel enlarges, and the corolla withers and degenerates. Therefore, flowers that remained on the tree for 24 h were considered pollinated, those that lasted more than 36 h were considered fertilized, and those showing visible signs of fruit formation were classified as transitioning to fruit.

To confirm that pollination occurred naturally by insects and to rule out self-pollination, a section of branch with at least 10 flower buds close to opening was selected from each tree. This branch section was enclosed in a cage made from tulle fabric (Figure 5) to prevent the entry of pollinating insects.

Finally, to ensure pollen transfer to the evaluated flowers, three flowers per tree were randomly removed after 48 h in each plot. The removed flowers were placed in bottles containing a 3:1 ratio of acetic acid and ethanol. In the laboratory, the staminodes and petals were removed from the collected flowers, leaving only the ovary with the stigma. To count the number of viable and non-viable pollen grains on the stigma under a microscope, a drop of Alexander’s stain was applied [7]. Alexander’s stain (1980) [26] allows for the differentiation between viable and non-viable (aborted) pollen, with non-viable pollen turning green and viable pollen turning purple.

Statistical analysis: The total number of specimens and the summary statistics for pollinator abundance were calculated for both indirect (wooden trap) and direct sampling methods across the 11 evaluated plots. A linear regression analysis was used to examine the relationship between the response variables—pollinated flowers, fertilized flowers, aborted flowers, and flowers transitioning to fruit—and the abundance of Ceratopogonidae pollinators (Forcipomyia sp. and Dasyhelea sp.), as well as the total number of specimens. The regression model used can be symbolically expressed as:

$$Y={\beta }_0+{\beta }_{1}X+ϵ$$

where:

• Y = represents the response variable (pollinated flowers, fertilized flowers, aborted flowers, and flowers transitioning to fruit);
• X = represents pollinator abundance;
• ${\beta }_0$ = is the intercept;
• ${\beta }_1$ = is the regression coefficient;
• $ϵ$ = is the error term.

Pollinator abundance was determined using absolute values from both collection methods (wooden traps and manual aspiration), which were combined for the analysis. Variables with a significant effect were identified based on a p-value < 0.05. The statistical analysis was conducted using Infostat software 2020I.

Finally, Pearson’s correlation coefficient (α = 0.05) was used to analyze the relationships between the abundance of potential Ceratopogonidae pollinators collected in the pyramid trap and the variables (pollinated flowers, fertilized flowers, aborted flowers, and flowers transitioning to fruit). This analysis aimed to determine the relationship between the abundance of potential pollinators and the potential productivity of the cocoa crop. The correlation also assessed the strength and direction (direct or inverse) of the relationships between these variables [27] (Figure 6).

3. Results

3.1. Capture of Potential Pollinators and Floral Visitors

This study is based on the hypothesis that the pollinator abundance (Diptera: Ceratopogonidae) directly influences the effectiveness of natural pollination in cocoa cultivation. To evaluate this relationship, key variables such as the number of pollinated, fertilized, aborted, and fruit-transitioning flowers were measured.

To capture the pollinators, two complementary methods were used: wooden traps (indirect sampling) and manual aspiration on flowers (direct sampling). This experimental design allowed us to quantify the presence of pollinators in the agroecosystem and analyze their relationship with cocoa reproduction.

Using the indirect sampling method (wooden box trap), 2774 Diptera specimens were collected, of which 612 belonged to the family Ceratopogonidae. Among these, 477 were identified as belonging to the genus Forcipomyia and 116 to the genus Dasyhelea, and 19 were classified as “other”. The “other” individuals lacked wings or antennae, key morphological traits that are necessary for genus identification. In eight of the analyzed plots, both Forcipomyia and Dasyhelea were found, while in the remaining three plots, only Forcipomyia or only Dasyhelea were present.

The box plot below visually represents the median, quartiles, and potential outliers, offering a detailed understanding of the central tendencies and dispersion of the Ceratopogonidae family pollinators collected across the two sampling periods in the 11 plots. The diagram highlights that the individuals from the genus Forcipomyia show better dispersion, particularly during the second sampling period, and exhibit a higher average count compared to the other genus across both collection dates (Figure 7).

With the second type of insect collection (direct aspiration on flowers), a total of 173 insects were captured, with 88.44% belonging to the order Diptera (

Table 2. Order of insects collected by direct sampling in cocoa flower (Yopal, Colombia 2021).

Order	N° of Individuals Captured	Percentage (%)
Diptera	153	88.44
Hemiptera	1	0.58
Hymenoptera	7	4.05
Psocoptera	10	5.78
Araneae	2	1.16
Total	173	100

). Among these, 61 individuals were identified as belonging to the family Ceratopogonidae. Within this family, 22 individuals were classified as Forcipomyia sp., and four were identified as Forcipomyia genualis.

3.2. Presence of Pollinating Insects and Relationship with Natural Effective Pollination

The results indicate that the pollinator abundance is strongly correlated with the reproductive success of cocoa. In the evaluated plots, positive correlations were found between the number of Forcipomyia sp. individuals and the flower fertilization rate (coefficients close to 1.0 in linear regression and Pearson’s correlation).

These findings support the hypothesis that pollinators are essential for cocoa reproduction and highlight the need for management practices that promote their conservation in agroforestry systems.

The data on pollinator abundance indicate that, in the four selected plots planted with the CCN51 clone (chosen because the 11 evaluated plots contained different clones, and only four included CCN51), Diptera: Ceratopogonidae were reported in all plots across both sampling methods (

Table 3. Abundance of pollinating insects (Ceratopogonidae) in the cocoa AFS (Yopal, Colombia 2021).

Type of Sampling	Abundance of Insects	n	Mean	S.D.	CV	Minimum	Maximum	Median
Indirect—Wood trap	Diptera	4	176.5	132.4	75.0	63.0	326.0	158.5
	Diptera: Ceratopogonidae	4	31.0	24.9	80.2	5.0	59.0	30.0
	Ceratopogonidae: Forcipomyia sp.	4	27.8	24.2	87.2	3.0	59.0	24.5
	Ceratopogonidae: Dasyhelea sp.	4	3.3	5.3	161.6	0.0	11.0	1.0
Direct—Flower Suction	Diptera	4	27.8	23.0	82.9	7.0	57.0	23.5
	Diptera: Ceratopogonidae	4	8.0	3.7	45.6	4.0	12.0	8.0
	Ceratopogonidae: Forcipomyia sp.	4	5.0	4.2	83.3	0.0	10.0	5.0

n: number of plots analyzed; S.D.: standard deviation; CV: coefficient of variation.

).

When analyzing the number of pollinated, fertilized, and aborted flowers (

Table 4. Percentage of pollinated, fertilized, and aborted flowers and flowers in transition from flower to fruit in the AFS of cocoa (Yopal, Colombia 2021).

N° Plot	Number of Flowers Evaluated	Percentage of Flowers (%)
		Pollinated	Fertilized	Aborted	Flower to Fruit
1	122	27.32	4.10	68.03	0.82
2	61	45.90	9.84	40.98	3.28
3	79	35.44	5.06	59.49	0.00
4	99	34.34	13.13	50.51	2.02

), it was observed that the percentage of pollinated flowers varied across the different plots, ranging from a maximum of 45.90% to a minimum of 27.32%. Furthermore, the flowers enclosed in tulle fabric cages were not pollinated, confirming the role of pollinating insects in the pollination process of the cocoa crop.

Laboratory staining of the flower stigmas confirmed the transfer of pollen to the evaluated flowers (Figure 8a–c). A highly variable number of viable and non-viable pollen grains was observed on the stigmas, with the viable pollen counts ranging from a minimum of 6 to 130 grains (mean = 31.9; CV = 124.02%).

Once the transfer of pollen by pollinating insects was confirmed, the abundance of pollinating insects (Diptera: Ceratopogonidae and Ceratopogonidae: Forcipomyia sp.) was correlated with the percentages of pollinated, fertilized, aborted, and flower-to-fruit transition flowers (

Table 5. Relationships between the abundance of pollinators (Ceratopopogonidae) and the percentages of pollinated, fertilized, aborted, and flower-to-fruit transition flowers in the cocoa AFS (Yopal, Colombia 2021).

Dependent Variable	Independent Variable	Relationship	p Linear Regression	R2 of Linear Regression	Pearson Correlation Coefficient
Pollinated	Diptera: Ceratopogonidae	+	0.45	0.31	0.55
	Ceratopogonidae: Forcipomyia sp.	+	0.60	0.16	0.40
Fertilized	Diptera: Ceratopogonidae	+	0.01	0.99	0.99
	Ceratopogonidae: Forcipomyia sp.	+	0.02	0.96	0.98
Aborted	Diptera: Ceratopogonidae	-	0.18	0.68	−0.82
	Ceratopogonidae: Forcipomyia sp.	-	0.30	0.49	−0.70
Flower to Fruit	Diptera: Ceratopogonidae	+	0.27	0.53	0.73
	Ceratopogonidae: Forcipomyia sp.	+	0.42	0.33	0.58

). The data reveal that the strongest correlation is between the number of pollinating insects and the percentage of fertilized flowers, for which coefficients close to 1.0 for both the linear regression and Pearson’s correlation were obtained. This indicates a positive relationship between these two variables. Additionally, a negative correlation (−0.82) was found between the percentage of aborted flowers and the number of Diptera: Ceratopogonidae. The third most significant correlation is a positive relationship (0.73) between the percentage of flowers transitioning from flower to fruit and the number of Diptera: Ceratopogonidae.

4. Discussion

4.1. Assessing Abundance of Potential Pollinators and Floral Visitors

Both types of sampling (indirect and direct in-flower) confirmed the presence of Diptera: Ceratopogonidae as potential pollinators of the cocoa crop. In particular, direct flower captures revealed that the most abundant order was Diptera, comprising 88% of the captures. This value is significantly higher than the 13.4% reported by [17] using emergence boxes, where Diptera was the second most abundant order, including pollinators and other Diptera. Their study recorded a range from 0.5% to 65.4% in cocoa agroforestry systems (AFSs). Similar findings were obtained by [15] in Ecuador, where Diptera was the third most abundant group, comprising 7.09% of the collected insects.

In collections conducted by [28] in Peru, 7% of floral visitors were from the Ceratopogonidae and Cecidomyiidae families, a finding consistent with [29], conducted in Bolivia. Their research found that, although Diptera was the second most abundant order captured, the most prevalent families were Chloropidae (11%), Phoridae (12%), and Ceratopogonidae (6%).

In Ecuador, various studies have identified Diptera genera associated with cocoa crop pollination. Reference [17], through specimen captures from the Ceratopogonidae family, reported the presence of Forcipomyia sp. (5.4%), Dasyhelea sp. (2.2%), and Atrichopogon sp. (0.23%). Similarly, reference [21] in a faunistic analysis of pollinating insects, identified Ceratopogonidae grouped into 27 morphospecies, including the genera Forcipomyia sp., Dasyhelea sp., and Culicoides sp. Finally, reference [11] captured pollinating insects from the Ceratopogonidae family representing the genera Forcipomyia sp., Culicoides sp., and Dasyhelea sp.

In Panama, [18] identified Forcipomyia sp. as the most frequent genus, followed by Dasyhelea sp. and Atrichopogon sp. In Nicaragua, reference [30] captured Diptera as visitors to cocoa flowers, which included a population of 175 Cecidomyiidae, 19 Ceratopogonidae, 5 Chironomidae, and 4 Sciaridae. In Indonesia, reference [31] captured, through sticky traps, Forcipomyia sp. in more significant numbers in the rainy season.

In contrast, a study conducted in Indonesia [19] found no Ceratopogonidae visiting cocoa flowers, instead capturing only ants and other dipterans. Similarly, research in Costa Rica [32] reported a low presence of this family, with only two specimens being observed in organic cocoa plantations. Their findings highlighted that the most abundant insects across three production systems—conventional, organic, and traditional—were Thysanoptera and Diptera: Cecidomyiidae.

In Honduras, reference [33] identified the genus Forcipomyia as the most frequent during the summer, followed by Atrichopogon and Dasyhelea. The study also highlighted the significant variability in pollinator abundance, reporting a minimum of 3.45 pollinators per square meter and a maximum of 231.45.

Although variations in the abundance of pollinating insects have been observed compared to other studies, it is important to highlight that this research was conducted during a period of high humidity and rainfall, which likely had a positive influence on the number of pollinators present. Reference [15] reported a moderate positive correlation between insect presence and relative humidity, with 52% of the observed variability in insect numbers being explained by this factor. They further noted that the abundance of dipteran pollinators tends to be higher during rainy periods and decreases during droughts. Nonetheless, further research is needed to examine the complete life cycle of these pollinators and their relationship with climatic variables.

Effective pollination is highly dependent on the synchronization between the dynamic populations of dipteran pollinators and the flowering cycles of cocoa trees. Furthermore, the abundance of midges is closely associated with the number of flowers present [34].

While techniques such as wood traps and manual aspiration allow for the capture of pollinators and the generation of detailed scientific data, we acknowledge that these methods may not be directly relatable to farmers, who typically rely on the direct observation of insect visits to flowers during blooming periods. For this reason, we believe that the best way to bridge this gap is by conducting direct sampling on farms with active farmer participation, enabling them to understand and validate the collected information. Additionally, it is crucial to develop training programs focused on pollinator conservation and the implementation of proper agronomic practices for their management. This approach would enhance informed decision-making in the field and encourage the adoption of sustainable management strategies in cocoa production systems.

4.2. Impact of Pollinating Insect Presence on Natural Pollination Effectiveness

The findings highlight the essential role of insects in the pollination of cocoa flowers, as the covered flowers showed no evidence of pollination. These results align with studies conducted by [18] in Panama, as well as [35,36] in Ecuador. These studies similarly report significant differences between naturally pollinated flowers and those that are covered, noting that pollination and fertilization occurred exclusively in the flowers exposed to pollinating insects.

In the present study, the average percentage of fertilized flowers was 8.03%. In addition, no fertilized flowers were found in the absence of pollinators. This result agrees with those obtained by [18], which reported a 4% pollination/fertilization of flowers in the presence of pollinators and 2% in the absence of pollinators in an AFS of cocoa from Panama. However, it should be noted that the indirect capture methodology used in this study (with an emergence box) may influence the number of insects collected, since in AFSs of cocoa where pollinators are not found, it does not imply their absence, but rather that they are less abundant than in AFSs of cocoa where they were found.

As for the percentage of flowers in transition to fruit, an average of 1.53% was obtained for the clone studied (CAU-43). This result agrees with the literature, where reference. [37] states that the percentage of flowers that bear fruit in cocoa is usually very low, with percentages between 0.5 and 5%.

These low percentages may be related to the fact that the arrival of pollen grains at the stigma does not imply fertilization. For this reason, the count of pollen grains on the flower stigma should not be extrapolated to reflect fruit production. However, it is considered that there is a possibility that a flower could become a fruit when it has received at least 25 pollen grains [38]. In our case study, the results showed that the number of total pollen grains was variable, with a mean of 31.9. This result is superior to the study developed by [39], where the number of pollen grains deposited on the stigma ranged from 10 to 29 in the seven clones evaluated.

In ae study conducted in northern Peru [28], an average of 31 ± 1.2 (mean ± S.D.) pollen grains deposited per flower was found, where only 0.8% formed fruit; in those flowers that formed fruit, an average of 111 ± 19.2 pollen grains were deposited, while, in the flower styles that did not bear fruit, an average of 30.7 ± 1.2 pollen grains were deposited. Because of the very low fruit set rates, it was not possible for them to relate the fruit set to pollen quantities measured directly on cocoa trees.

According to [8], 200 or more compatible pollen grains deposited on flowers can indicate the formation of pods containing from 50 to 60 seeds. It is important to keep in mind the existence of genetic variability among clones in the number of pollen grains produced per flower, which allows breeding for this trait and could reduce yield gaps due to poor pollination [40].

Regarding the correlation between the abundance of Ceratopogonidae insects and that of pollinated and fertilized flowers, a positive correlation was found between the abundance of pollinating insects and the percentage of fertilized flowers. This finding aligns with [18], who reported that the total abundance of pollinating insects significantly correlates with the percentage of pollinated flowers. Specifically, the genus Forcipomyia shows a positive relationship with the number of pollinated flowers, reinforcing its importance in crop pollination. Reference [36] observed that the percentages of pollinated flowers, fertilized flowers, and fruit set under natural pollination are low, with 9% of flowers being pollinated, 4% being fertilized, and 4% forming fruits.

It is important to consider that, while effective pollination influences fertilization and future seed formation, this process is also affected by various factors such as fruit abortion due to intrinsic tree conditions, the presence of diseases and pest insects, crop nutrition, and compatibility phenomena among cocoa clones that can lead to fruit abortion. As a result, up to 72% of pollinated flowers may not develop into mature harvestable fruits, leading to fruit mortality and a reduction in potential cocoa yields [41].

For reference [28], despite the large amounts of pollen deposited on flowers, the low pollination rate may be due to other factors such as pollen viability, pollen compatibility, and resource availability, all of which may limit fruit set. For this reason, cocoa is often designated as under-pollinated because of the small fraction of pollinated flowers that is often observed [42,43]. Although the proportion of pollinated flowers of the wild tree has not yet been established, it is considered desirable to study the rate of pod formation of wild cocoa in the Amazon rainforest, as this will clarify whether there are natural pollination deficits of cocoa crops [25].

The results indicate that Diptera: Ceratopogonidae plays a crucial role in the pollination of cocoa crops, accounting for 88.44% of the collected flower visitors. However, reference [28] observed variations in flower visitation patterns across different study locations. Consequently, the taxonomic identity of the primary pollinators remains a topic of debate, and it is likely that other arthropod taxa, beyond Diptera, also contribute to cocoa pollination. Therefore, examining floral visitor patterns in different cocoa-growing regions is essential to better understand the pollination potential of various insects and enhance pollination services.

Similarly, reference [44] considers that there is still a need for research on pollination services in cocoa agroforestry systems to identify other pollinators, as cocoa trees are visited daily by many insects that are pollinators or potential pollinators. It is important to combine pollen estimation from macro photographs with controlled insect visits, an ideal strategy to confirm pollen loads, visitation frequencies, and, ultimately, the identities of pollinators and flower visitors [28]. Therefore, it is considered that, in the case of the cocoa AFSs, the abundance of pollinators favors the formation of future fruits, so there must be clear strategies that allow the conservation of potential pollinators of this crop.

5. Conclusions

In this study, insects associated with pollination in the cocoa agroforestry system (AFS) at the Utopia Campus in Yopal, Colombia, were identified using two methods. The indirect method, employing wood traps, revealed that Forcipomyia sp. and Dasyhelea sp. of the Ceratopogonidae family were the most abundant genera. In contrast, the direct method, which involved capturing insects on flowers, did not capture Dasyhelea sp. This absence does not imply that Dasyhelea sp. was not present at the sampling sites, but instead suggests that it occurs in lower abundance.

The results highlight the importance of insect activity for cocoa flower pollination. Flowers exposed to natural (open) pollination exhibited pollen transfer to the stigma, whereas covered flowers did not show signs of pollination. Furthermore, a positive correlation was found between the percentage of fertilized flowers and the number of insects associated with pollination in the studied agroforestry system

These findings underscore the need for sustainable farming practices that support pollinator conservation. To enhance pollination efficiency, cocoa farmers should implement strategies such as using soil cover with mulch, preserving water sources, and adopting integrated pest management approaches that reduce pollinator exposure to pesticides. Promoting these conservation practices can help mitigate pollination variability and improve the cocoa productivity in agroforestry systems.