Pollination Deficit: A Key Limitation of Fruit Set in Northward-Expanded Camellia Orchards

Abstract

Northward expansion of economically essential plants is a vital strategy for enhancing agricultural productivity; however, it often results in reduced yields. This study systematically assessed the impact of translocating the high-value oilseed species Camellia hainanica from its native tropical habitat Sanya to the temperate cultivation area of Changsha, focusing on its reproductive processes, including flowering, pollination, and fruit development. Our findings revealed a 45-day delay in anthesis at the transplanted location, which was associated with notably lower average daily temperatures (7.89 °C in Changsha compared to 24.63 °C in Sanya) during the anthesis period. While floral longevity, stigma receptivity, and pollen viability remained comparable between sites, anther dehiscence was markedly delayed by three days after transplanting. Crucially, pollinator visitation during peak flowering plummeted by 92% compared to the levels in Sanya, and a 57% reduction in pollen deposition on stigmas occurred. Consequently, natural fruit sets in Changsha collapsed to 0%, significantly lower than those in Sanya, despite artificial cross-pollination achieving an 11% fruit set rate. These results and the pollination deficit coefficient (D = 1.00) all demonstrate that severe pollination deficits are the key limitation causing reproductive failure in northward-expanded C. hainanica orchards. Addressing these yield constraints necessitates targeted breeding for earlier flowering genotypes and implementing pollination management strategies.

1. Introduction

Expanding the cultivation range of plants, particularly through the northward introduction of southern plants (termed “northward expansion”), represents a critical strategy for enhancing agricultural productivity and safeguarding food and oilseed security [1]. This approach enables the full utilization of northern land resources and significantly expands crops’ production potential [2]. Global warming has led to a tendency for some plants to expand their ranges northward, which is beneficial for expanding the cultivation of economic plants [3]. However, in contrast to the climate-driven northward expansion of natural plant distributions, human-mediated northward expansion of plants often subjects them to abrupt shifts in critical environmental conditions, such as temperature and photoperiod [4,5]. These changes pose substantial challenges to plant growth and reproduction.

Environmental shifts during transplant (temperature, photoperiod, and soil nutrients) can disrupt anthesis, floral morphology, and pollen development [6], potentially altering pollination networks for northward-expanded plants. This frequently causes phenological mismatches or reduced pollinator availability [7], leading to normal flowering but low or failed fruit set [8,9]. This challenge causes yield reduction or instability in many northward-expanding economic plants, including tea oil camellia [10]. Consequently, systematically deciphering how environmental changes affect the reproductive success of northward-expanded plants through impacts on sequential reproductive phases (from flowering and pollination to fruit set) holds critical theoretical and practical importance [11].

Tea oil camellia collectively refers to economically significant Camellia species (Theaceae) characterized by high seed oil content, including C. oleifera, C. hainanica, C. chekiangoleosa, and others [12,13]. The tea oil extracted from these seeds is nutritionally rich. Despite the continuous expansion of cultivation areas, current production remains inadequate to meet market demands. The northward expansion of cultivation areas is imperative to increase the total tea oil output and address the supply–demand imbalance [10,14]. Recently, some studies have found that the cultivation area of tea oil camellia is expanding northward [15,16]. A major concern is that current research on the northward expansion of tea oil camellia predominantly focuses on vegetative growth (e.g., branch and leaf development) while neglecting reproductive processes (flowering, pollination, and fruit set) [14]. This critical knowledge gap severely impedes crop production and industry development, particularly regarding the factors limiting seed set (which directly determines oil yield).

Against this backdrop, we selected C. hainanica (noted for its high oil yield, favorable fruit traits, and distinctive flavor) as our study species [17]. As a hermaphroditic but self-incompatible plant, it is highly dependent on pollinators, primarily Apis cerana, for successful fruit and seed production [18,19]. To investigate the impacts of northward expansion on its flowering, pollination, and fruit set, we transplanted C. hainanica from its native tropical habitat (Sanya) to a temperate zone (Changsha). Specifically, we assessed the effects of this latitudinal shift on the following: (1) anthesis, floral longevity, pollen viability, and stigma receptivity; (2) pollinator attraction (visitation rates, primarily by A. cerana), flower-visiting insect number, and stigma pollen deposition; (3) fruit set rate and development. Beyond providing critical guidelines for the northward expansion of tea oil camellias, this work establishes an empirical foundation for cross-latitude transplanting of other insect-pollinated economic plants.

2. Materials and Methods

2.1. Plant Materials

The plant materials used in this study were grown in orchards located in Sanya (18.25° N, 109.50° E) and Changsha (28.20° N, 112.97° E), separated by a distance of approximately 1458.47 km. Over 500 C. hainanica plants in Changsha were transplanted from the Sanya population before March 2022. Plants were carefully excavated with intact root balls and soil; subsequently, these trees were transplanted at a spacing of approximately 2.5 m apart. Plants in both orchards were healthy and vigorous and had no pests or diseases.

2.2. Cultivation Site Environment

To compare environmental conditions across different cultivation sites, geographical coordinates (latitude and longitude) were first acquired from the National Platform for Common GeoSpatial Information Services (https://www.tianditu.gov.cn/, accessed on 19 June 2025). Soil samples were collected at each site to determine soil type. Finally, the 2024 annual mean temperature, 2024 annual precipitation, climate type, and climatic characteristics for each site were obtained from the National Meteorological Information Center (https://data.cma.cn/, accessed on 19 June 2025) and their respective local meteorological bureaus.

2.3. Anthesis and Temperature During Anthesis

The anthesis of plants in both orchards was recorded to assess the impact of northward expansion on C. hainanica anthesis. Beginning in November, daily (9:00–11:00) observations of the flowering stage were conducted for individuals of C. hainanica in the Sanya and Changsha orchards. The dates marking the start time of flowering (defined as when 5% of flowers in the orchard were open) and end time of flowering (defined as the wilting of >90% flowers in the orchard) were recorded for the C. hainanica populations at both sites. Additionally, the daily mean temperature during anthesis was recorded at each site to compare the temperatures between the two sites during flowering.

2.4. Flower Longevity Recording

Five healthy, pest-free plants with similar growth rates were selected at each site to compare the effects of northward expansion on flower longevity. Three unopened flower buds were randomly selected and tagged on each plant. All tagged buds (n = 15 per site) were monitored daily. The flowering start and end times were recorded for each flower, from which the individual flower longevity (in days) was determined.

2.5. Observation of Anther Dehiscence

To compare the effects of northward expansion on anther dehiscence, three healthy C. hainanica plants were selected at each site. Accounting for differences in flower number on per tree canopy between sites, 15 unopened flower buds per plant were randomly tagged in Sanya (n = 45), while 10 buds per plant were tagged in Changsha (n = 30). To assess anther dehiscence across the 5-day flower longevity, tagged flowers were evenly allocated for observation on flowering days 1 to 5. Tagged flowers were monitored daily. Flowers were collected when they reached the target day.

Collected flowers were transported to the laboratory. The androecium was dissected for each flower to separate the inner and outer round anthers. Each round anther was examined separately under a stereomicroscope (OLYMPUS DP73, Tokyo, Japan). The number of dehisced and non-dehisced anthers per round anther was recorded, and the dehiscence percentage was calculated as follows:

Dehiscence percentage (%) = (Dehisced anther number/Total number of anthers) × 100

2.6. Pollen Viability and Stigma Receptivity Assays

(1) Pollen viability: First, five healthy C. hainanica plants were selected from each cultivation site, and two flowers were randomly selected and marked on each plant to determine their pollen viability on different flowering days. Considering that the flower longevity of C. hainanica is five days, seven flowers needed to be marked on each tree to determine pollen viability from day 1 to 5 of flowering. Subsequently, tagged flowers were monitored daily to record flowering time. Ultimately, 56 flowers were assayed due to variations in flowering time. Upon reaching the target flowering day, flowers were collected. Anthers were excised, placed into centrifuge tubes, and covered with TTC (2,3,5-triphenyltetrazolium chloride) staining solution [0.5% MTT: 0.5 g TTC (Sigma, Darmstadt, Germany) dissolved in 100 mL 95% alcohol]. Samples were incubated in the dark at 37–40 °C for 15–20 min [20,21]. The solution was then vortexed, a droplet was placed on a glass slide, covered with a coverslip, and observed under an optical microscope (OLYMPUS CX31, Tokyo, Japan) for counting. Viable pollen grains were stained red, while non-viable grains remained unstained.

Pollen viability (%) = (Number of viable pollen grains/Total pollen grains counted) × 100

(2) Stigma receptivity: Following the flower selection and tagging procedure described for pollen viability assessment, 50 flowers (n = 25 per site) were ultimately obtained. Pistils were excised entirely using a scalpel. The dissected stigmas were immersed in benzidine–hydrogen peroxide reaction solution (1% benzidine: 3% H₂O₂: distilled H₂O = 4:11:22, v/v/v) within concave glass slides. After 1 min of reaction, the samples were examined and photographed under an optical microscope (OLYMPUS CX31, Tokyo, Japan). Receptive stigmas contain active peroxidase, which oxidizes benzidine to produce a blue product and catalyzes H₂O₂ decomposition, releasing O₂ bubbles [22]. Stigma receptivity was therefore assessed semi-quantitatively based on the extent of blue coloration and bubble production, and assigned a score on an ordinal scale (Figure S3) ranging from 0 (no receptivity) to 5 (strong receptivity).

2.7. Pollen Number and Nectar Volume Quantification

(1) Pollen number per anther: Three healthy C. hainanica plants were selected at each site. Ten fully open flowers were selected from these plants. All anthers were excised from each flower. From the anthers of each flower, 30 anthers were randomly selected and divided into three samples of 10 anthers each. Each sample (n = 30 per site) was placed in a 1.5 mL centrifuge tube. Then, 1 mL of distilled water was added to each tube, followed by thorough vortexing. A 2 μL aliquot of the pollen suspension was pipetted onto a hemocytometer, covered with a coverslip, and pollen grains were counted under an optical microscope (OLYMPUS CX31, Tokyo, Japan). The pollen number per anther (grain/per anther) was calculated using the following formula:

Pollen number per anther (grain/per anther) = (Count number × 1000/10/2)/Number of anthers

(2) Nectar volume per flower: Three healthy C. hainanica plants were selected at each site. Five unopened flower buds per plant were tagged and individually bagged with 40-mesh nylon nets to exclude pollinators (n = 15 per site). Following flower opening, nectar was collected from each flower at 8:00 using a microsyringe, and the volume per flower was recorded (Figure S1).

2.8. Observation of Flower-Visiting Preferences of A. cerana

To investigate the impact of northward expansion on the pollinator attractiveness of C. hainanica, the primary flower visitor of this species (A. cerana) was selected for a behavioral two-choice assay and Y-tube olfactometer tests (Figure S2). Considering the geographical differences of the two orchards, we first collected 50 unopened C. hainanica flowers from each site and transported them to Jinhua City, Zhejiang Province (29.08° N, 119.63° E). The flowers were placed in sterile containers with mineral water and maintained hydroponically under greenhouse conditions (26 °C and 70% relative humidity) until flowering for use in two experiments.

(1) Wild behavioral two-choice assay: Blooming flowers were positioned 0.5 m from the entrance of A. cerana hives, ensuring an unobstructed path between the hive and flowers. Flowers from two sites were spaced 15–20 cm apart. Flowers exhibiting visible senescence were promptly replaced with fresh specimens to avoid affecting the experimental results. Observations were conducted from 10:00 to 12:00 on sunny days, with flower positions systematically alternated every 30 min to mitigate positional bias. The floral choices of A. cerana individuals were recorded, excluding no-selection bees (n = 30). A selection was recorded when a bee visited a flower. The selected flower was scored 1, and the non-selected flower was scored 0.

(2) Y-tube olfactometer tests in laboratory: Blooming flowers were carefully positioned using forceps at the air inlets of both arms of a Y-tube olfactometer (central stem: 20 cm length, 3 cm diameter; arms: 15 cm length, 2.5 cm diameter, 60° angle), ensuring symmetrical placement without obstructing airflow [23]. The olfactometer base was connected to a vacuum pump generating a unidirectional downwind airflow at 0.3 L/min. Fifty A. cerana individuals were starved and introduced singly into the entry port, and their choices were recorded. A choice was registered when a bee entered an arm and either landed on the flower, remained stationary near it (≤5 cm), or flew within a 5 cm radius; the selected arm was scored as one, and the non-selected arm as zero. Bees remaining in the central stem section for more than 5 min without choosing were excluded. A total of 41 valid observations were obtained.

2.9. Observation of Flower-Visiting Insects

Observations were conducted during the full flowering stage of C. hainanica at both Sanya in 2024 and Changsha in 2025. On five sunny days per site, insect visitation was recorded daily from 10:00 to 12:00 on three randomly selected C. hainanica plants per orchard [24]. Following the methodology of Yuan et al. [25], floral visitors were classified into five categories: bees, wasps, flies, hoverflies, and other insects.

2.10. Quantification of Stigma Pollen Deposition

To compare pollen transfer efficiency in C. hainanica before and after northward expansion, we quantified pollen deposition on stigmas during the full flowering stage at both sites. Five healthy C. hainanica trees were selected per site, with six freshly opened flowers sampled per tree (n = 30 per site). Pistils (excluding ovaries) were excised and individually fixed in FAA solution (70% ethanol/distilled water/formaldehyde = 18:1:1) within glass vials for transport to the laboratory. Fixed pistils were transferred to 1.5 mL centrifuge tubes containing 1 mL NaOH (8 mol/L) for 4 h softening. Pistils were then triturated, diluted to 1.5 mL with distilled water, and centrifuged (800 rpm, 10 min). The supernatant was discarded. The pellet was diluted to 0.5 mL of distilled water. A 2 μL aliquot of the suspension was loaded onto a hemocytometer, covered with a coverslip, and pollen grains were counted under optical microscopy (OLYMPUS CX31, Tokyo, Japan). Pollen deposition per stigma was calculated as follows:

Pollen grains per stigma (grain) = Counted pollen grains × 500/2/10

2.11. Pollination Treatments and Fruit Development Monitoring

(1) Pollen collection: Healthy plants (n = 5 per site) were selected at Sanya, Wenchang (19°37′01″ N, 110°45′18″ E), and Changsha. Ten unopened flower buds per plant were collected. Anthers were excised and dried in a greenhouse (26 °C, 75% relative humidity) for 6–8 h to release pollen. Debris was removed by sieving through a 40-mesh screen. Purified pollen was aliquoted into 1.5 mL centrifuge tubes and stored at −20 °C until use.

(2) Pollination treatments: In Sanya, we carried out two pollination treatments—natural pollination and artificial crossbreeding (with pollen from Wenchang C. hainanica). In Changsha, we carried out three pollination treatments: natural pollination, self-pollination (with pollen from the same tree), and artificial cross-pollination (with pollen from Sanya C. hainanica). During the full flowering stage of C. hainanica, artificial pollination was conducted daily between 8:00 and 11:00. Firstly, all anthers of flower buds (in each treatment, 45 flowers from different trees were treated) were removed carefully to avoid damaging the pistil. Pollen from the designated treatment was applied to stigmas using a pollination rod until a visible yellow coating was achieved. Pollinated flowers were immediately isolated with parchment paper bags, and the bags were removed 4–5 days post-pollination. In Sanya, the fruit set rate was recoded after pollination for 3 months. To explore the fruiting process of transplanted C. hainanica, the fruit set in Changsha was recorded at 7-day intervals for all pollination treatments until April 25. The fruit set rate was calculated as follows:

Fruit Set Rate (%) = (Retained fruit number/Pollinated flower number) × 100

(3) Fruit observation: After pollination, plastic mesh bags (90-mesh, 35 cm × 20 cm) were placed over fruiting branches to collect abscised fruits. Fallen fruits were collected according to their pollination treatment and immediately fixed in Carnoy’s fluid (absolute ethanol/glacial acetic acid = 3:1). Samples were vacuum-infiltrated for 30 min and stored at 4 °C. Following complete fruit abscission in the natural pollination treatment, any remaining fruits still attached to branches were collected. The retained fruits from branches and all abscised fruits across pollination treatments were dissected. To assess embryo development status, ovaries were examined under a stereomicroscope (OLYMPUS DP73, Tokyo, Japan).

(4) Coefficient of pollination deficit (D): To determine whether C. hainanica experienced pollen transfer constraints under open-field conditions at the two sites, we estimated the coefficient of pollination deficit (D) using the method of Layek et al. [26] as follows:

D = 1 − Ro/Rs

where Ro denotes reproductive success (fruit set rate) in natural pollination, and Rs means reproductive success in artificial cross-pollination. D has a value between 0 and 1. Based on the calculated value of D, four categories were defined: (a) high pollination deficit (D > 0.5), (b) medium pollination deficit (D = 0.3–0.5), (c) low pollination deficit (0.3 > D ≥ 0.1), and (d) negligible pollination deficit (D < 0.1).

2.12. Statistical Analysis

Data are expressed as the mean ± standard error of the mean (SEM). One-way analysis of variance (ANOVA) and an independent samples t-test were used to determine significant differences among the experimental groups, with statistical significance defined at p < 0.05. All analyses were conducted using SPSS Statistics version 27.0 (SPSS Inc., Chicago, IL, USA), while figures were generated using Origin 2024 (OriginLab, Northampton, MA, USA).

3. Results

3.1. Environmental Conditions of Different Sites

All experimental plants were cultivated in managed orchards. Still, the two sites differed climatically (

Table 1. Environmental characteristics of native and northward-expanded cultivated habitats of Camellia hainanica.

Site	Sanya	Changsha
Coordinates	18.25° N, 109.50° E	28.20° N, 112.97° E
Soil type	Red soil	Red soil
Annual temperature in 2024	26.7 °C	18.8 °C
Annual precipitation in 2024	2080.1 mm	1736.1 mm
Climate type	Tropical monsoon	Humid subtropical monsoon
Climate	High temperature and humidity, no frost	Hot summer and old winter, synchronized rain–heat

and Figure 1A). The orchard in the native habitat was located in a tropical monsoon climate zone (Sanya: 18.25° N, 109.50° E) with red soil. In 2024, Sanya experienced a mean annual temperature of 26.7 °C and a yearly precipitation of 2080.1 mm, characterized by high temperatures, high humidity, and no frost. The orchard at the transplanted site was situated in a humid subtropical monsoon climate zone (Changsha: 28.20° N, 112.97° E) with red soil. In 2024, Changsha recorded a lower mean annual temperature of 18.8 °C and an annual precipitation of 1736.1 mm, featuring hot summers and cold winters. The 10° latitude difference between the sites resulted in distinct photoperiods and differences in the thermal regime. Despite these differences, the soil conditions at both locations were suitable for cultivating Camellia species.

3.2. Impact of Northward Expansion on the C. hainanica Flower

3.2.1. Anthesis and Flower Longevity

The anthesis of C. hainanica was delayed at the transplanted site compared to that in its native habitat. In Sanya, it occurred from 18 November to 11 December 2024 (24 days). In contrast, in Changsha, it occurred from 2 January to 15 February 2025 (45 days, Figure 1B). The start date of anthesis differed by 45 days between the sites. Additionally, there was a significant difference in the mean daily temperature between the two sites during anthesis (p < 0.001). The mean daily temperature during anthesis for the Changsha plants (7.89 ± 3.16 °C) was significantly lower than that for the Sanya plants (24.63 ± 1.35 °C; Figure 1C, p < 0.001). However, flower longevity did not differ significantly between the two sites (Figure 1D, p > 0.05), lasting approximately five days at both.

3.2.2. Stigma Receptivity and Pollen Viability

Transplantation did not obviously affect stigma receptivity or pollen viability in C. hainanica, and their temporal dynamics during flowering remained consistent between the sites (p < 0.05).

Stigma receptivity in both Sanya and Changsha followed a unimodal pattern over the flower longevity (Figure 1E). On day 1 of flowering, stigma receptivity did not differ significantly between the sites (p > 0.05). By day 2, stigma receptivity peaked at both sites with no significant difference (p > 0.05). Subsequently, stigma receptivity declined steadily at a similar rate in both locations, reaching non-receptive levels by day 5.

Pollen viability declined gradually with flower longevity at both sites (Figure 1F). Although significantly higher in Sanya than in Changsha on flowering day 1 (p < 0.05), viability exceeded 50% at both sites. It then declined in parallel, reaching approximately 10% by day 5, with no significant difference between sites at this stage (p > 0.05).

3.2.3. Anther Dehiscence

Complete anther dehiscence in transplanted C. hainanica was delayed by three days compared to plants in the native habitat (Figure 1G). In Sanya, the anthers of C. hainanica dehisced completely on day 1 of flowering. However, in Changsha, only 3% of outer-round anthers dehisced on day 1, which was significantly lower than in Sanya (Figure 1G(i), p < 0.05). The proportion of dehisced outer-round anthers in Changsha plants increased progressively with flower longevity, reaching 100% by day 4. In contrast, inner-round anthers in Changsha C. hainanica showed no dehiscence during days 1–3 but reached 100% dehiscence by day 4 (Figure 1G(ii)).

3.2.4. Nectar Volume and Pollen Number

Transplantation significantly increased the pollen number of C. hainanica (Figure 1H, p < 0.05). Sanya plants produced significantly fewer pollen grains per anther (1335 ± 466.40 grains) compared to Changsha plants (1675 ± 730.61 grains; p < 0.05). Additionally, nectar volume was significantly higher in Changsha plants (17.59 ± 1.96 μL) than in Sanya plants (0.49 ± 0.61 μL, Figure 1I, p < 0.05).

3.3. Impact of Northward Expansion on Pollination of C. hainanica

3.3.1. Bee Attractiveness of C. hainanica Flower

Sanya C. hainanica exhibited greater attraction to A. cerana than plants from Changsha. A field behavioral two-choice assay demonstrated that 70% of bees visited flowers of Sanya C. hainanica compared to only 30% visiting Changsha C. hainanica flowers (Figure 2A(i), p < 0.05). Subsequent Y-tube tests indicated that transplantation altered the floral scent, resulting in significantly lower bee preference for Changsha C. hainanica flowers. Specifically, 63% of bees selected Sanya C. hainanica flowers, considerably exceeding the proportion choosing Changsha C. hainanica flowers (Figure 2A(ii), p < 0.05).

3.3.2. Flower-Visiting Insects

During the full flowering stage, Sanya C. hainanica attracted significantly more flower-visiting insects than those in Changsha (Figure 2B, p < 0.05). Observations conducted during 10:00–12:00 at the full flowering stage revealed an average daily count of 155.6 flower-visiting insects on C. hainanica in Sanya, compared to only 12.2 insect individuals in Changsha, which is reduced by 92% compared to Sanya C. hainanica.

The diverse insect taxa were observed in Sanya orchards, including bees, wasps, flies, hoverflies, and other flower visitors (Figure 2C(i)). Bees dominated the visitors, with significantly higher numbers than the other four insect categories (p < 0.05), indicating their role as primary pollinators for C. hainanica in Sanya. Conversely, Changsha orchards exhibited limited pollinator diversity, with only sparse occurrences of flies, hoverflies, and other flower visitors (Figure 2C(ii)). Among these, flies dominated the visitors, showing significantly greater numbers than other insect categories (p < 0.05).

3.3.3. Pollen Deposition on Stigmas

Pollen deposition on stigmas corroborated the above results (Figure 2D). Under natural conditions, Sanya stigmas received significantly more pollen grains (446 ± 86.51) than Changsha stigmas (194 ± 74.59 grains, p < 0.05). Based on these, the stigmatic pollen deposition after the northward expansion was ultimately reduced by 57%.

3.4. Impact of Northward Expansion on C. hainanica Fruiting

3.4.1. Fruit Set Rate

In both sites, the fruit set rate of natural pollination was significantly lower than that of the artificial cross-pollination treatment (Figure 3A, p < 0.05). Moreover, after the northward expansion, the natural fruit set rate of Changsha C. hainanica was 0%, which was significantly lower than the fruit set rate of Hainan C. hainanica by natural pollination (p < 0.05). Moreover, using the final fruit set rate to calculate the pollination deficit coefficient, we found that C. hainanica at both sites experienced pollination deficits, with Sanya C. hainanica showing a medium pollination deficit (0.30 < D = 0.46 < 0.50) and Changsha C. hainanica showing a high pollination deficit (D = 1.00 > 0.50).

3.4.2. Fruit Development of Changsha C. hainanica

After pollination, all three pollination treatment groups exhibited similar fruit drop patterns (Figure S4 and Figure 3A). After 7 March, each group underwent a rapid fruit drop stage, losing exactly 13% of its fruits. The fruit drop rate declined after 14 March and stabilized after 18 April. The intraspecific cross-pollination treatment showed the least reduction in fruit set (36%), achieving a final rate of 11%. The self-pollination treatment retained only one fruit (4% fruit set), while the natural pollination treatment experienced complete fruit abortion (0% fruit set).

We further dissected both retained and abscised fruits. Ovaries of abscised fruits from natural and self-pollination treatment showed no seed development (Figure 3B(i)). In contrast, some abscised fruits from intraspecific cross-pollination contained aborted seeds (Figure 3B(ii)). All six normally developed fruits from intraspecific crosses and self-pollination contained fully formed translucent seeds (Figure 3B(iii)).

4. Discussion

The northward expansion of plants is a common strategy employed to enhance plant potential productivity, playing a crucial role in safeguarding food security and promoting dietary diversification. However, significant differences in biotic (e.g., pollinator communities) and abiotic (e.g., climate) factors between the transplanted habitat and the native habitat often constrain reproductive success, even leading to complete failure [27]. This study focused on C. hainanica. Despite the northward transplanted C. hainanica exhibiting normal pollen tube growth (Figure S5), fertilization, and fruit development after artificial cross-pollination, it experienced complete reproductive failure (0% fruit set) under natural pollination levels, driven by a severe pollination deficit (D = 1.00) [28]. Specifically, post-transplantation anthesis was delayed by 45 days, resulting in the full flowering stage occurring under low-temperature conditions (7.89 ± 3.16 °C). Concurrently, anther dehiscence was delayed by three days. Then, the number of flower-visiting insects plummeted by 92% compared to the native site, and effective pollen deposition on stigmas accounted for only 43% of the native level.

The synchrony between anthesis and peak pollinator activity is a key prerequisite for ensuring effective pollination services [29]. Post-transplantation, plant anthesis may become desynchronized with insect activity periods of the transplanted site, leading to a low flower visitation by insects. Indeed, northward expansion often induces changes in anthesis through alterations in photoperiod: long-day plants tend to flower earlier, while short-day plants exhibit delayed flowering; day-neutral plants primarily maintain or alter anthesis in response to other environmental factors, such as temperature [30,31]. Previous research indicates that the flowering of C. oleifera, a close relative of C. hainanica, is regulated by photoperiod, exhibiting a response pattern characteristic of short-day plants [32]. Given the phylogenetic relationship between C. hainanica and C. oleifera, the observed 45-day anthesis delay following northward expansion in this study was likely induced by photoperiod changes resulting from increased latitude. For autumn-flowering species, delayed flowering means reproductive activities are forced into the winter low-temperature period. However, low temperatures significantly suppress the activity capacity of various pollinating insects (e.g., bees, flies, moths), directly explaining the drastic decline in flower-visiting insect numbers post-transplantation [33]. Although this study did not directly test low-temperature thresholds, an extensive literature confirms that temperatures below 10 °C severely inhibit the flower-visiting behavior of most insects [34,35]. Furthermore, while A. cerana (the primary pollinator of C. hainanica) was observed to be active in Changsha during November (corresponding to the flowering period of C. hainanica), its visitation frequency decreased significantly with declining temperatures [25]. When cold temperatures severely inhibit insect activity, even insect-pollinated plants that increase floral rewards (e.g., nectar volume, pollen number) or adjust floral scent signals also find it difficult to carry out their ecological function in attracting pollinators. This consequently limits the contribution of flower opening to reproductive success.

Undoubtedly, pollinator scarcity is a major factor in the complete reproductive failure of northward-expanded C. hainanica. However, the reduced pollen transfer efficiency due to delayed anther dehiscence should not be overlooked, because it directly shortens the effective duration for a single flower to release pollen during the flower’s longevity. In addition, insect visitation itself does not equate to effective pollination; successful reproduction depends on the efficient transfer of high-viability pollen to highly receptive stigmas. Delayed anther dehiscence directly impedes this process because pollen viability typically declines over floral longevity or follows a unimodal pattern but is always the highest in the early stage of flower opening [36,37]. This study found that pollen viability was highest on the first day of flowering at both sites and subsequently decreased. Crucially, northward transplant delayed the timing of anther dehiscence entirely. In Camellia species, anther dehiscence relies on the dehydration of anther stomium cells (pseudopollen) under suitable temperatures [38]. Low temperatures during the flowering period in the transplanted habitat likely inhibit this mechanism. While the dehiscence mechanism in Camellia has specific characteristics, the inhibitory effect of low temperature on anther dehiscence is general [39]. Here, C. hainanica anthers did not dehisce completely after flowering for three days, by which time pollen viability had fallen below 40% and stigma receptivity was only maintained at a moderate level (levels 2–3). This prevents flower-visiting insects from collecting high-viability pollen, further diminishing their potential pollination service. Additionally, low temperatures in the transplanted habitat may indirectly impair reproductive success by inhibiting processes such as pollen tube growth and ovule development [40]. The observed decline in the reproductive potential of transplanted individuals may be related to this, although differences in soil microorganisms and cultivation management could also contribute to this effect.

In conclusion, to overcome the yield bottleneck of C. hainanica after northward expansion, future horticultural practices require multi-dimensional strategies, as follows: (1) introducing cold-tolerant pollinator insects (e.g., Bombus spp.), although their pollination efficacy may be constrained by delayed anther dehiscence, and the risk of competition with native pollinators needs assessment [41]; (2) applying assisted pollination technologies (e.g., drone pollination), which can circumvent adverse climatic conditions and the impact of anther behavior, ensuring stable fruit set rates in commercial production [42]; (3) breeding genotypes with early flowering or strong anther dehiscence capability under low temperatures, which is essential for achieving sustainable agriculture [43]. This case highlights that during plant northward expansion, the plant–pollinator interaction network and its adaptation to the microclimate must be evaluated simultaneously. Future research should integrate pollinator management, mechanistic analysis of plant–pollinator interactions, and targeted breeding to construct a comprehensive framework for the ecologically sustainable expansion of woody oilseed crops.

5. Conclusions

This study compared the flowering, pollination, and fruiting processes of C. hainanica in its native tropical habitat (Sanya) and a translocated temperate habitat (Changsha). The key findings indicate that while the translocated population possesses reproductive potential, it suffers from complete reproductive failure (0% fruit set) driven by a severe pollination deficit (D = 1.00), critically limiting fruit yield. Translocation delayed flowering by 45 days, shifting bloom to colder winter temperatures (7.89 ± 3.16 °C), which suppressed pollinator activity. Concurrently, delayed anther dehiscence (3 days) shortened the pollen release period of a single flower to just 2 days and diminished the transfer of high-viability pollen (viability was <40% with full anther dehiscence). To enhance productivity in temperate regions, targeted pollinator management strategies (e.g., managed pollinators or drone-assisted pollination) or breeding for low-pollination-dependent/cold-tolerant cultivars are essential for sustainable cultivation.