Syzygium cumini trees growing in Visakhapatnam (17° 43' 51.744" N, and 83° 20' 16.368" E), Andhra Pradesh, India, were used for the present study during April-August 2022. Field observations were made to record flowering period, floral biology, foraging activity and pollination, breeding systems, and fruiting and seed dispersal aspects. Thirty flowers from 10 trees were collected to describe morphological aspects of flowers briefly, because of reports of inconsistencies on these aspects in the literature. Fifty mature buds tagged on 10 trees were followed for recording the timing of anthesis and anther dehiscence. A 10X hand lens was used to confirm the dehiscence time and mode. Stigma receptivity was observed by using H₂O₂ test described in Dafni et al. (2005). In this test, the period of release of bubbles from the stigma surface following application of H₂O₂ was recorded as the total duration of stigma receptivity during flower life.

The presence of nectar was determined by gently pulling a flower from its calyx and firmly pressing its base against a hard surface. Twenty mature buds from five trees which were about to open were bagged before sunrise and removed on the evening of the same day to measure total nectar produced by each flower by inserting a micropipette into the flower base. Again, the flowers were bagged and removed on the next day for measuring the nectar in each flower. The same process was followed on the 3rd and 4th day. The flowers did not produce nectar in the 5th day. The average volume of nectar produced by all these flowers for four consecutive days was taken as the total volume of nectar/flower and expressed in µl. The nectar produced in these flowers was used for measuring nectar sugar concentration each day and then for calculating the mean sugar concentration. A hand sugar refractometer (Erma, Japan) was used for measuring nectar sugar concentration.

Nectar analysis for sugar types was carried out using the paper chromatography method described in Dafni et al. (2005). The sugar content/flower is expressed as the product of nectar volume and sugar concentration per unit volume according to Dafni et al. (2005). The protocols given in Sadasivam & Manickam (1997) were followed for the quantitative estimation of sucrose, glucose and fructose in mg/flower. The caloric reward of nectar/flower/day was measured as per the formula given in Heinrich (1975). Baker & Baker (1982) method was used for the calculation and classification of sugar ratios of nectar. Paper chromatography method described in Dafni et al. (2005) was followed for identifying the amino acid types present in the nectar. Nectar was spotted on Whatman No. 1 filter paper along with the standard samples of 21 amino acids. The paper was run ascendingly in chromatography chamber for 24 h with a solvent system of n-butanol-acetic acid-water in 4:1:5 ratios. The chromatogram was detected with 0.2% ninhydrin reagent and dried at 85°C in an electric oven for 15 min for the development of spots from the nectar and the standard amino acids. The developed nectar spots were compared with those of the standard amino acids to record the amino acid types present in the nectar. Lowry et al. (1951) method was used for measuring protein content in the nectar; 40 bagged fresh nectariferous flowers were used.

Twenty mature but undehisced anthers were collected from different trees and placed in a Petri dish. Later, each time a single anther from each flower was taken out and placed on a clean microscope slide and dabbed with a needle in a drop of lactophenol-aniline blue. The pollen mass was drawn into a band and the total number of pollen grains was counted under a compound microscope. Based on these counts, the mean number of pollen grains produced per anther was determined. The mean pollen output per anther was multiplied by the number of anthers in the flower for obtaining the mean number of pollen grains per flower. Another set of 10 dehisced anthers was collected in a Petri dish and the pollen grains removed from these anthers were examined under microscope for recording the pollen grain features.

The protocols described by Mondal et al. (2009) were followed for identifying amino acid types present in the pollen. Pollen was collected from mature anthers and filtered through sieving using meshes of different size (100, 200 and 300 µm) to remove the debris. Then, the pollen was rapidly dried over silica gel at 30°C and stored. Free amino acids were extracted from the pollen using the method described in Mondal et al. (2009). Later, the extract thus obtained was used for the qualitative analysis of the free amino acids of pollen using thin layer chromatography. The protocol described in Sadasivam & Manickam (1997) was followed for the extraction of protein from the pollen samples using phosphate buffer of pH 7.4 and then Lowry et al. (1951) protocol was followed for estimating the protein content in the sample.

The insects visiting the flowers were observed during 05:00-19:00 h to record foraging activity period of each insect species. The foraging schedule, forage collected and the flower probing behavior of each insect species were recorded. The number of foraging visits made by each insect species was recorded for 15 min at each hour during the entire length of the observation period to examine the pattern of foraging activity according to the availability levels of nectar and pollen. This field observation on foraging activity of insect species was repeated on four clear sunny days and the data thus collected were used to calculate the average number of visits made by each species at each hour of the day and the percentage of foraging visits by each category of insect species to record the foraging rate of individual insect species and each insect category. The foraging behavior of each insect species was observed with reference to its approach, landing, probing behavior employed for pollen collection and contact with essential organs in effecting pollination.

Breeding systems were tested for different modes of self- and cross-pollination. Spontaneous autogamy was tested by bagging five complete inflorescences consisting of 210 flowers on the same tree. Fifty mature buds were bagged, opened the next day after the occurrence of anthesis, anther dehiscence and stigma receptivity; the stigma was pollinated with the pollen of the same flower using a brush and bagged again to test hand self-pollination. Eighty mature buds were bagged after emasculation, opened the next day after the commencement of stigma receptivity; the stigma was pollinated with the fresh pollen of a different flower of the same tree using a brush and bagged again to test geitonogamy. Sixty mature buds were bagged after emasculation, opened the next day after the occurrence of stigma receptivity; the stigma was pollinated with the fresh pollen from the flowers of a different tree using a brush and bagged again to test xenogamy. The bagged flowers were followed for 30 days for fruit set. Based on the flowers that produced fruits, the percentage of fruit set was calculated. Twenty inflorescences with 840 flowers from 10 trees were tagged prior to anthesis and followed for natural fruit set. The percentage of fruit set was calculated based on the number of fruited flowers. Fruits are characteristically 1-seeded and hence seed set rate was treated as equal to fruit set rate.

Fruit and seed characters were described in view of the reports of inconsistencies in these characters. Field observations were made on fruit dispersal agents and the role of them was briefly described.

Syzygium cumini is a tropical evergreen tree species which grows wild in many wild pockets in Simhachalam, Adavarivaram, Anandapuram, Madhurawada and Rushikonda in Visakhapatnam (Fig. 1A). It is also planted and cultivated for its edible fruits and ornamental value. The flowering occurs en masse during May-July. The inflorescences are intercalary polychasial cymes each with an average of 46 ± 8.9 flowers; they are borne at the end of each branch (Fig. 1B, C). The pattern of flowering in each cyme is that the central flower matures first while the lateral ones flower subsequently. The flowers are small, 4 mm diameter, creamy white, mildly odoriferous, bisexual and actinomorphic (Fig. 1D). The calyx consists of five creamy white sepals, united with thalamus and form a cup-like structure. The corolla consists of five petals, free, creamy white, delicate and inserted on the top of a deep cup-like receptacle. The petals form a cap in the bud condition and fall off as a calyptra due to the pressure of the growing stamens inside. The stamens are 46-50, arranged on the rim of the receptacle in several whorls. They are bent inwards in bud condition; straighten at the time of opening and extend outwards after anthesis. They are creamy white and tipped with versatile anthers. The ovary is bicarpellary syncarpous with 18-22 ovules on axile placentation. The style is terminal, 10-11 mm long and terminates into a simple stigma (Fig. 1E).

The flowers are open throughout the day with peak anthesis during 18:00-20:00 h. The buds while opening push the corolla cap upwards exposing the stamens and stigma. The petals soon fall off. At this stage, the stamens bend downward and gradually stretch out completely exposing the stamens. The anthers dehisce by longitudinal slits following anthesis. The pollen output per anther is 201 ± 17. The pollen grains are white, triangular, powdery and 15.75 × 18 × 13.5 µm in size. The style arises from the center of the cup, and shows growth after anthesis. The stigma attains receptivity 24 hours after anthesis and remains receptive until the evening of 4th day of flower life. The nectar secretion occurs continuously for a period of four days from the time of anthesis. A flower produces 3.71 ± 1.0 µl of nectar consisting of 43.1 ± 5.26% sugar concentration; the total sugar content per flower is 1.904 mg. The nectar energy per flower is 26.767 joules. The sugar types in the nectar include sucrose, glucose and fructose; their quantity per flower varies with sugar type. The sucrose is 0.22 mg, glucose 0.233 mg and fructose 0.224 mg. The nectar is hexose-rich and the sugar ratio is 0.481. The pollen analysis for amino acids showed that it has four essential and six non-essential amino acids for insects. The essential ones include arginine, histidine, isoleucine and lysine while the non-essentials include aspartic acid, cysteine, cystine, glycine, hydroxyproline and serine (Table 1). The total protein content per 1 mg of pollen is 0.15 mg. The nectar contains six each essential and non-essential amino acids for insects. The essential amino acids are arginine, histidine, lysine, phenylalanine, threonine and tryptophan. The non-essential amino acids include aspartic acid, cysteine, cystine, glycine, serine and tyrosine (Table 1). The protein content in the nectar is 0.213 mg/flower. The flowers fall off at the end of 5th day.

The flowers were foraged by bees, ants, flies, butterflies and diurnal hawkmoths (Table 2). The data collected on the foraging visits of all these groups of insects showed that butterflies made 53%, bees 28%, flies 9% and ants and hawkmoths, each 5% of total visits paid to flowers (Fig. 2). The bees were Apis dorsata Fabricius, 1793 (Fig. 3A, B), A. cerana Fabricius, 1793 (Fig. 3C), A. florea Fabricius, 1793 (Fig. 3D) and Trigona iridipennis Smith, 1854 (Fig. 3E) (Apidae) and Halictus sp. (Halictidae) (Fig. 3F). The ants represented only one species, Camponotus sp. (Formicidae) (Fig. 3G). The flies were Helophilus sp. (Syrphidae) (Fig. 3H) and Chrysomya megacephala (Fabricius, 1794) (Calliphoridae) (Fig. 3I). The butterflies included 18 species representing Papilionidae, Pieridae, Nymphalidae, Lycaenidae and Hesperiidae families. The Papilionidae and Pieridae each was represented by two species, Nymphalidae 10 species, Lycaenidae three species and Hesperiidae one species. The Papilionids were Pachliopta aristolochiae (Fabricius, 1775) and P. hector (Linnaeus, 1758). The Pierids were Catopsilia pomona (Fabricius, 1775) and C. pyranthe (Linnaeus, 1758). The Nymphalids were Hypolimnas bolina (Linnaeus, 1758), Tirumala limniace (Cramer, 1775), T. septentrionis (Butler, 1874) (Fig. 3J), Parantica aglea (Stoll, 1782) (Fig. 3K), Danaus chrysippus (Linnaeus, 1758) (Fig. 3L), D. genutia (Cramer, 1779) (Fig. 3M), Precis iphita Cramer, 1782 (Fig. 3N), Euploea core (Cramer, 1780) (Fig. 3O), Melanitis leda Linnaeus, 1758 and Acraea violae Fabricius, 1775. The Lycaenids were Castalius rosimon (Fabricius, 1775), Everes lacturnus Fruhstorfer, 1924 (Fig. 3P) and Jamides celeno (Cramer, 1775). The Hesperiid was Hasora chromus (Cramer, 1780) (Fig. 3Q). The diurnal hawkmoths belonged to Sphingidae and they were Macroglossum gyrans Walker, 1856 and Cephonodes hylas Linnaeus, 1771 (Sphingidae) (Fig. 3R). Further, one unidentified nocturnal moth species (Fig. 3S) had also occasionally collected nectar after sunset for a brief period.

The bees, flies and ants collected forage regularly from 07:00/08:00 to 17:00 h with peak activity at 11:00-13:00 h (Fig. 4). All butterflies collected nectar from 08:00 to 17:00 h with peak activity at 10:00-12:00/13:00 h (Figs. 5-7). Hawkmoths collected nectar from 06:00 to 08:00 h and again from 16:00 to 19:00 h (Fig. 8). The cup-shaped calyx with exposed nectar facilitated all insect foragers to collect nectar with great ease. The flowers with copious amount of nectar presented in numerous polychasial cymes at each branch level enabled butterflies to remain on the same tree for a long time and such a foraging activity was considered to be promoting geitonogamy. While collecting nectar, the insects came in contact with both the stigma and stamens invariably ensuring the transfer of pollen and pollination of stigmas. Among all insects, only bees collected pollen along with nectar in the same or in another foraging visit. Since the bees were involved in pollen collection, they tended to stay mostly on the same flower/polychasial cyme or on the same tree and such a foraging activity appeared to be effecting mostly geitonogamy rather than xenogamy. The sole ant species being a resident forager was also found to be effecting mostly or exclusively geitonogamy. Flies, butterflies, the diurnal hawkmoths and the moth being exclusive nectar foragers tended to fly swiftly between flowers of the same or conspecific trees growing nearby effecting both geitonogamy and xenogamy. The possibility for the occurrence of vector-mediated autogamy is almost ruled out due to non-receptive nature of stigma on the day of anthesis and the falling of stamens on the morning of the 2nd day of anthesis.

Hand-pollination tests showed that the flowers set fruit through autogamy, geitonogamy and xenogamy. Individual flowers that were bagged did not produce any fruits but fruit set was 8% when entire inflorescences were bagged. Fruit set was 8% in manipulated autogamy, 46% in geitonogamy, 78% in xenogamy and 51% in open-pollinated flowers (Table 3). The fruits mature within 3-5 weeks and drop off during July-August. The ripe fruit is a globose dark purplish one stony-seeded berry with fleshy, sweet, juicy and slightly acidic taste pulp; it shows different colors from green to light purple to dark purple during its growth and development (Fig. 9A-D). Local people collect fruits due to their edible and commercial value and probably contribute to seed dispersal. Frugivorous birds such as Pycnonotus cafer (Linnaeus, 1766), P. jocosus (Linnaeus, 1758), Iole indica (Jerdon, 1839) (Passeriformes: Pycnonotidae), Acridotheres tristis (Linnaeus, 1766) (Passeriformes: Sturnidae), Zosterops palpebrosus (Temminck, 1824) (Passeriformes: Zosteropidae), Megalaima viridis (Boddaert, 1783) and Megalaima haemacephala (Müller, 1776) (Piciformes: Megalaimidae) were found feeding on the fruits occasionally. These birds fed on the fruit pulp either by spitting out and leaving the seed attached to the parent tree. In seed-spitting mode, the birds carried the fruit with their beak to other trees of the area, landed on the branch and fed on the pulp and dropped the seed under the tree. In seed-leaving mode, the birds simply ate the pulpy part without removing the fruit from the parent tree. In both modes, birds dispersed seeds either under the parent tree or in the vicinity of the parent tree or in distant areas. The ripe fruits that were not collected by locals remained on the ground under the tree canopy. Fruit dispersal occurs through feeding of fleshy part of the fruits by birds opportunistically and by humans until fruit stock is exhausted.