This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
INTRODUCTION
The genus Delphinium, a member of the Ranunculaceae family, is widely distributed in the temperate regions of the Northern Hemisphere (Malyutin, 2001). Delphinium comprises about 350 species of perennial flowering plants, of which 173 species have been described in China (Wang and Warnock, 2001). In particular, most of the species are concentrated in Southwest China, which is one of the most important distribution centres and covers most of the morphological primitive representatives (Yuan and Yang, 2008). Yunnan Province, known as the ‘Kingdom of Plant’, is a centre of species diversity and endemism of Delphinium, consisting of 49 species and 13 varieties (Yuan and Yang, 2019). According to reports, there are 24 species with a high ornamental value that has potential for development and utilisation, such as D. yunnanense and D. grandiflorum (Guan and Li, 2002). Moreover, Delphinium is widely cultivated in the world for a long time, that is, Delphinium elatius flore plena was cultured by J. Parkinson in Western Europe as early as 1640 (Legro, 1961). In floriculture, the most important hybrid species is D. elatum, which developed as a series of 4,000 named cultivars from hundreds of different breeders and growers in the past few decades (Royal Horticultural Society, 1949). Besides, through hybridisation, mutagenesis, chemical and radiation treatment of different species within the genus Delphinium, a wide range of flower colours from white, yellow, red, violet to blue were introduced into the new cultivars as well as improved quality of cut flower, mildew resistance and toughness of the stems (Mehlquist et al., 1943; Wegulo and Vilchez, 2008; Chen et al., 2011; Lin et al., 2014). In addition to the ornamental purposes of showy flowers, the plants of Delphinium contain various alkaloids with spectacular medicinal values, such as diterpenoid alkaloids, trichodelphinines and flavonoids, which increase the utilisation and economic value of Delphinium (Chen et al., 2011; Agnihotri et al., 2014; Lin et al., 2014).
Delphinium yunnanense (D. yunnanense) is an endemic species in the Yunnan Province of China with blue flowers, which is considered a precious ornamental germplasm to be developed. However, Delphinium elatum (D. elatum) ‘Guardian Blue’ is one of the most popular flowers in the world as an attractive high-grade cut flower. Both species are the potential candidates for developing new Delphinium varieties with blue flowers through hybridisation. The chromosome number and the ploidy level of many species and varieties in the genus Delphinium have been investigated (Legro, 1961; Yuan and Yang, 2008). Most of the wild species, including D. yunnanense, are diploids with 16 chromosomes (2n = 2x = 16), whereas D. elatum was tetraploid (2n = 4x = 32). An intraspecific difference in chromosome number was detected neither in diploid nor in tetraploid species (Legro, 1961; Yuan and Yang, 2008). Nevertheless, the breeders have had success in interploidy crosses using different Delphinium species. For example, the frequently used tetraploid cultivars of D. elatum were widely crossed with diploid species to form sterile triploid hybrids, which, after the colchicine treatment, became fertile, produced seeds and viable seedlings (Lewis et al., 1951; Samuelson, 1957; Tjebbes, 2010). Moreover, the reproductive system with pollen– stigma interactions has been previously investigated in D. grandiflorum, D. cardinale, D. nudicaule, D. barbeyi and D. nuttallianum and their interspecific hybrids (Honda et al., 1999; Williams, 2007; Elliott and Irwin, 2009; Briggs et al., 2016). Some studies have shown that the breeding system of Delphinium is facultative xenogamy or obligate outcrossing, accompanying with partly self-compatibility requiring pollinators (Price and Waser, 1979; Williams et al., 2001; Zhang et al., 2014). Self-incompatibility (0–50%) and autogamy (1– 90%) in different degrees among different species in Delphinium have been reported (Macior, 1975; Varney, 1979; Powell and Jones, 1983; Bosch et al., 2001). To our knowledge, both of the studied species are capable of selfing; however, no previous study concerned the cross-compatibility in the species of D. yunnanense and D. elatum.
In this study, we explored and investigated the interspecific cross-compatibility of D. yunnanense and D. elatum ‘Guardian Blue’ to establish the mating ability of these two Delphinium species, which could be helpful in developing new varieties of these ornamental plants.
MATERIALS AND METHODS
Plant material and growth conditions
D. yunnanense and a commercial cultivar of D. elatum ‘Guardian Blue’ were used in this study. The seeds of D. yunnanense used in this study were randomly collected in natural habitats (N25°7′33″, E102°45′31″) in Kunming, Yunnan Province at an altitude of 2,139 m. The seeds of D. elatum ‘Guardian Blue’ were obtained from the local flower market with 500 seeds in the same bag. A total of 100 seeds of each species were first germinated in the petri dish covering with a moist filter paper. Then, 50 seedlings of each species were separately transferred and planted in 100-well plant trays. After the root system filled the holes, the plants were separately planted in pots (diameter = 8 cm, height = 8 cm) filled with substrate (red soil:peat = 1:1) and grown in the solar greenhouse (without heating) at the experimental farm of Yunnan Agricultural University. The flowering dynamics were continuously observed when the plants begin to bloom.
Pollen viability and stigma receptivity
The examination of pollen viability was conducted with three plants of each species during anthesis based on in vivo germination method (He et al., 2017). In a full flowering stage, fresh pollen was collected every 2 h from three flowers per plant and was incubated for 4 h at 25°C in the liquid medium (5% sucrose, 150 mg · L−1 boric acid and 20 mg · L−1 CaCl2) from 9:30 to 15:30. The total number of pollen grains and the number of germinating pollen grains were counted in three separate fields of view per slide under the light microscope. The percentage of germinating pollen was estimated by germinating pollen number/total pollen number × 100. The active pollen was identified by the length of the pollen tube longer than the diameter of pollen grain, otherwise recorded as inactive pollen (Rodriguez-Riano and Dafni, 2000).
Stigma receptivity was examined by the benzidine– hydrogen peroxide method (Li et al., 2018). During anthesis (from day 1 to day 8), three flowers at the full-open stage of each species were collected every day and were soaked in the benzidine–hydrogen peroxide solution (1% benzidine: 3% hydrogen peroxide: water = 4:11:22, v/v, Cato Research Chemicals, Inc.). The receptive stigma shows peroxidase activity and turns blue with a mass of bubbles surrounding it. The depth of the blue color and number of bubbles indicate the intensity of stigma receptivity.
Observation of pollen–stigma interaction
To observe the pollen–stigma interaction in interspecific hybridisation, the reciprocal crosses between D. yunnanense and D. elatum were conducted as well as intraspecific allogamic crosses. The D. elatum × D. yunnanense was treated as an orthogonal cross, and the D. yunnanense × D. elatum referred to reciprocal cross. To avoid self-pollination, the flowers were emasculated and isolated by plastic bags until the anther dehiscence. This treatment was applied in each cross combination with a random selection of 10 plants (three flowers per plant). Before pollination, the stigma was examined by a magnifier for a secretion check. The pollen sac was then gently dipped onto the stigma with secretions using tweezers. Each stigma was pollinated several times to ensure a sufficient amount of pollen and was covered by bags after pollination.
The pistils of each cross were collected at 2, 4, 6, 24, 72 and 120 h after pollination and fixed in Carnoy's solution (ethanol: glacial acetic acid = 3:1, v/v, Shanghai Fusheng Industrial Co., Ltd.) for 24 h at 4°C and then stored in 70% of ethanol. The samples were then treated with 6 mol · L−1 NaOH for 12 h. After repeated rinsing with distilled water, the pistils were immersed in a 0.1% water-soluble aniline blue solution (dissolved in 0.1 mol · L−1 K3PO4, ScyTek Laboratories) for 24 h. The stained ovaries were then observed with Leika fluorescent microscope under ultraviolet light to investigate pollen adherence, pollen germination, callose formation and pollen tube growth. For each treatment, six pistils in three replications were used to calculate the average number of pollen grains.
Calculation of fruit setting and germination rate
Thirty flowers were used to evaluate the fruit setting for each cross. The number of mature fruits and seeds per fruit were counted, and the thousand kernel weight (TKW) was measured. The fruit-setting rate was calculated by the formula: the number of mature fruits/the number of pollinated flowers × 100.
The F1 seeds of each cross were germinated in a presterilised culture dish (Guangzhou Jet Bio-Filtration Co., Ltd., TCD000060) on wet filter papers. The seeds were then cultivated in environmentally controlled growth chambers (LRH-400-GSI, 400 L; Zhujiang, China) under the condition of 12 h photoperiod at 25°C. The experiment was repeated in triplicate (100 seeds per repeat). The germination rate was calculated as germinated seeds/total seeds × 100, in which germinated seeds refer to the sprout length equal to or longer than the seed length.
Statistical analysis
Data analysis and statistics were performed using Microsoft Excel 2016 and Data Processing System (Tang and Zhang, 2013). One-way and two-way ANOVA with Tukey's HSD post-hoc test was used to compare multiple samples at the 5% significance level.
RESULTS
Stages of flower development
The inflorescence of D. yunnanense and D. elatum is racemose, in which the development of individual flowers is from bottom to top. Eight flowering stages could be identified in the two species throughout the entire flower developmental period, including flower bud stage, stem elongation stage, bud expansion stage, sepal open stage, petal open stage, flower semi-open stage, flower full-open stage, and anther cracked stage (Figure 1). The flower of D. yunnanense contains five large and bright bluish violet sepals, and the real petals are very small and hidden between the staminode and the upper sepal. Besides, there are two blue staminodes with yellow barbs on the ventral surface. For D. elatum, there are two layers of hyacinthine ovate sepals, and petals and staminodes are white. The flowering period of D. yunnanense was about 10 days, whereas the flowers of D. elatum can maintain a longer period of 15 days. This may be affected by the number of petals since D. elatum has double petal flowers.
Figure 1
Stages of flower development of D. yunnanense (A) and D. elatum ‘Guardian Blue’ (B). Bud stage (a), stem elongation stage (b), bud expansion stage (c), sepal open stage (d), petal open stage (e), flower semi-open stage (f), flower full-open stage (g) and anther cracked stage (h). Bar = 1 cm.
Pollen viability and stigma receptivity
The pollen viability, estimated in pollen germination in vitro test, of D. elatum was significantly weaker than that of D. yunnanense during the day, and the diurnal variation displayed a trend of increase in the morning but decrease in the afternoon (Figure 2). Moreover, the pollen viability of D. yunnanense and D. elatum showed a peak at noon (11:30 AM) with germination ratio at 40.89% ± 15.65% and 18.78% ± 5.18%, respectively. This suggested that Delphinium is better to pollinated around noon, at least in these two tested species.
Figure 2
The diurnal variation of pollen viability in D. yunnanense and D. elatum. One-way ANOVA was used to assess statistical significance, and p values were calculated with Tukey's HSD test (α = 0.05). Multiple comparisons indicated by small letters were performed in different collection time within the same species. The data represent the mean value ± standard deviation from three replicates.
The observation of stigma receptivity revealed that the stigma of D. yunnanense was not receivable at the beginning of the flowering period, but displayed a positive reaction with pollen from 6th day to 8th day (Table 1). This positive reaction was strengthened at the 7th day of flowering and then declined until the end of the measurement. In contrast, the stigma receptivity of D. elatum presented one day earlier than that of D. yunnanense and had a strong receptivity the next day. However, the stigma receptivity of D. elatum had a shorter duration than that of D. yunnanense, since it became only partial receivable at the 7th day of flowering.
Stigma receptivity of the full-open stage during the flower opening process in D. yunnanense and D. elatum. ‘−’, stigma not receivable; ‘±’, partial positive reaction; ‘+’, positive reaction; ‘++’ strong positive reaction.
Species
1st day
2nd day
3rd day
4th day
5th day
6th day
7th day
8th day
D. yunnanense
−
−
−
−
−
+
++
+
D. elatum
−
−
−
−
+
++
±
−
The data were obtained from 10 flowers.
Pollen–stigma interaction
The number of pollen grains adhering to the surface of stigma and the germination rate was measured to reflect and evaluate the cross-compatibility of each cross combination.
The results showed that pollination time (F = 73.3260, p = 0.0000) and cross combination (F = 110.5710, p = 0.0000) had significant effects on the number of attached pollen grains in Delphinium, and their interaction was significant as well (F = 18.9170, p = 0.0000) (Table 2). The number of attached pollen grains significantly increased at 24 h after pollination in four crosses. The number of attached pollen grains of D. yunnanense was significantly higher than that of D. elatum after pollination on any type of stigma (Figure 3). This indicated that the pollen survival rate of D. yunnanense was higher than that of D. elatum, which was in line with the result of pollen viability. However, there was no statistically significant difference in the number of pollen grains attached to the stigma between interspecies crosses of D. elatum × D. yunannense and intraspecific crosses of D. yunanense, similarly as in the cross of D. yunannense × D. elatum and intraspecific crosses of D. elatum.
Figure 3
The number of attached pollen grains of each cross combination after different pollination time. One-way ANOVA was used to assess statistical significance, and p values were calculated with Tukey's HSD test (α = 0.05). Multiple comparisons indicated by small letters were compared in different cross combinations within the same time after pollination. The data represent the mean value ± standard deviation from three replicates.
Results of statistical analysis about pollination time and cross combination on the number of attached pollen grains in Delphinium.
Source of variation
Sum of square
Degree of freedom
Mean square
F value
p value
Pollination time
27,746.7778
5
5549.3556
73.3260
0.0000
Cross combination
25,104.2778
3
8368.0926
110.5710
0.0000
Time × Combination
21,474.2222
15
1431.6148
18.9170
0.0000
error
3632.6667
48
75.6806
Total variations
77,957.9444
71
From the overview of the dynamics of attached pollen grains in four crosses, some pollen grains had been attached to the stigma after 2 h of pollination (Figure 3). The number of attached pollens in the orthogonal cross (D. elatum × D. yunnanense) was higher than that of other three crosses, which were at the same low level in the early period after pollination. This trend continued until 6 h after pollination, in which the attached pollen grains of D. yunnanense allogamic cross (D. yunnanense × D. yunnanense) dramatically increased over ten-fold (from 10.67 ± 1.53 to 118.67 ± 11.02, n = 3). The same result as multiple comparisons, the attached pollens reached the peak after 24 h of pollination with a significant difference in each cross combination. Then, the attached pollens gradually decreased, except for that of D. yunnanense allogamic cross showing a sharp decrease.
The frequency of germinated pollen grains in different crosses after pollination displayed the same trend as the attached pollen grains in the corresponding cross combination (Figure 4). There was also a significant interaction of pollination time and cross combination on the germinated pollen number revealed by the multiple comparisons (F = 6.0460, p = 0.0000, Table 3). Moreover, the number of germinated pollen grains reached its peak 24 h after pollination. The frequency of germinated pollen of D. yunnanense was significantly higher than that of D. elatum in four crosses. This was in line with the variation of attached pollens, indicating that the amount of germinated pollen grains was positively correlated with the amount of attached pollens in the species of Delphinium.
Figure 4
The number of germinated pollen grains on each cross combination after different pollination time. One-way ANOVA was used to assess statistical significance, and p values were calculated with Tukey's HSD test (α = 0.05). Multiple comparisons indicated by small letters were compared in different cross combinations within the same time after pollination. The data represent the mean value ± standard deviation from three replicates.
Dynamics of pollen tube growth in different cross combinations
In the orthogonal cross (D. elatum × D. yunnanense), most of the pollen grains germinated on stigma 2 h after pollination, showing pollen tubes entered the style and few callose deposited in the style (Figure 5A). After 4 h, callose depositions appeared in large number in the style, some pollen tubes had entered the ovary at this stage. Six hours after pollination, pollen tubes arrived near the ovule and started to bend and twine. Finally, pollen tubes entered the ovule 24 h after pollination. In the reciprocal cross, pollen tube growth differs that in the orthogonal cross (Figure 5B). Only a small amount of pollen grains germinated on the stigma, whereas a large amount of callose deposition was visible in the style 2 h after pollination. Four hours after pollination, pollen tubes twined, knotted and spiraled, accompanied by callose formation. As a result, only a small number of pollen tubes reached the ovule 24 h after pollination, but none of them entered the ovule as observed.
Figure 5
Dynamics of pollen tube growth of the interspecific and allogamic crosses between D. yunnanense and D. elatum. A:D. elatum × D. yunnanense. A-2h: A bundle of pollen tubes (Pt) in the style of pistil, pollen grains (Pg) visible on stigma. A-4h: Large number of callose (Ca) depositions appeared in the style. A-6h: Pollen tubes arrived near the ovule (Ou) and started to bend and twine. A-24h: Pollen tubes entered the ovule. B:D. yunnanense × D. elatum. B-2h: Few pollen grains germinated on the stigma with a large amount of callose depositions visible in the style. B-4h: The twined, knotted and spiraled pollen tubes with callose. B-6h: Six hours after pollination, the state of pollen tube remains the same as 4 h after pollination. B-24h: Few pollen tubes were close to the ovule. C:D. yunnanense × D. yunnanens. C-2h: A number of pollen tubes in style of pistil, pollen grains also visible on stigma. C-4h: A bundle of pollen tubes grow along the style. C-6h: Pollen tubes arrived near the ovule. C-24h: Pollen tubes entered the ovule. D:D. elatum × D. elatum. D-2h: A small amount of pollen germinated and callose deposition appeared. D-6h: Pollen tubes were visible in the style. The pollen tubes were twined and knotted in the style. D-24h: Pollen tubes reached the ovary with visible callose deposition. D-72: Pollen tube entered the ovule. ‘Pg’ refers to pollen grain, ‘Pt’ means pollen tube, ‘Ca’ is the abbreviation of callose and ‘Ou’ represents ovule. Bar = 200 μm.
Results of statistical analysis about pollination time and cross combination on the number of germinated pollen grains in Delphinium.
Source of variation
Sum of square
Degree of freedom
Mean square
F value
p-value
Pollination time
7034.9028
5
1406.9806
29.7860
0.0000
Cross combination
6178.2639
3
2059.4213
43.5980
0.0000
Time × Combination
4283.8194
15
285.5880
6.0460
0.0000
error
2267.3333
48
47.2361
Total variations
19,764.3194
71
In allogamic crosses of D. yunnanense, pollen tubes entered the ovule 24 h after pollination (Figure 5C). However, in the allogamic cross of D. elatum, a small amount of pollen germinated and callose deposition appeared 2 h after pollination. Six hours after pollination, the pollen tubes were visible in the style and reached the ovary 24 h after pollination (Figure 5D). Three days after pollination, the pollen tube entered the ovule. The pollen of D. yunnanense displayed a significantly strong interaction with stigmas of D. yunnanense and D. elatum, indicating that the pollen activity of D. yunnanense was stronger than that of D. elatum, which was in line with the results of pollen viability. It also suggested that D. elatum is more suitable as a maternal parent for hybridisation breeding.
Fruit and seed setting of F1 hybrids
After crossing, the fruit set rate was evaluated in each cross according to the between the fruit number and the number of previously pollinated flowers. The results showed differences in fruit set rate between the orthogonal and reciprocal cross of D. yunnanense and D. elatum (Table 4). In the orthogonal cross (D. elatum × D. yunnanense), the fruit set rate was the highest (90.0%) and generated 17.1 seeds per fruit. In contrast, there was no seed obtained in the reciprocal cross. This was consistent with the observation of microscopic dynamics that no pollen tube was found entering the ovule of D. yunnanense × D. elatum. On the other hand, there was no significant difference in fruit set rate between the allogamic crosses of D. yunnanense and that of D. elatum, showing the rate of 86.7% and 83.3%, respectively. Moreover, the seed number per fruit of D. elatum in allogamic cross was the highest, suggesting that there were no disturbances in the sexual reproductive system of D. elatum since both male and female lines were fertile. Therefore, the cross-incompatibility probably caused the sterility of D. yunnanense × D. elatum F1 hybrids. Furthermore, the seed germination rate of D. elatum allogamic cross was the highest, while seeds derived from the cross of D. elatum × D. yunnanense displayed the lowest germination rate of 16.7% as well as the lowest thousand kernel weight. This may relate to the formation of abnormal seeds detected in the orthogonal cross of D. elatum × D. yunnanense.
Characteristics of F1 seeds in the interspecific and allogamic cross of D. yunnanense and D. elatum.
Cross combination
Pollinated flowers number
Fruit number
Fruit set rate
Seed number
Seed number per fruit
Thousand kernel weight (g)
F1 seed germination rate (%)
D. elatum × D. yunnanense
30
27
90.0%
463
17.1
1.3
16.7 ± 2.08 a
D. yunnanense × D. elatum
30
0
0.0%
/
/
/
/
D. yunnanense × D. yunnanense
30
26
86.7%
421
16.2
1.7
24.7 ± 2.52 b
D. elatum × D. elatum
30
25
83.3%
527
21.1
2.0
80.0 ± 4.00 c
One-way ANOVA was used to assess statistical significance, and p values were calculated with Tukey's HSD test (α = 0.05).
DISCUSSION
The genus of Delphinium contains several species with a high ornamental value, which is crucial for the development of new floricultural varieties. In this study, we investigated the cross-compatibility of two Delphinium species, D. yunnanense, a wild endemic species of Yunnan province and a commercial variety D. elatum ‘Guardian Blue’. Strong cross-incompatibility was observed in the cross of D. yunnanense × D. elatum, whereas less cross-incompatibility was detected in the cross of D. elatum × D. yunnanense, which provided a better understanding of the hybridisation affinity in Delphinium.
The measurement of pollen viability has an important theoretical and practical significance in hybrid breeding. The level of pollen viability affects the success and efficiency of hybrid breeding. Therefore, it is essential to measure the pollen viability before hybridisation, which could provide a useful reference for the formulation of hybrid combinations. Our results showed that the diurnal variation of pollen viability displayed a trend of increased in the morning but decreased in the afternoon in both D. yunnanense and D. elatum. This was consistent with previous independent observations of D. grandiflorum, in which the pollen viability increased at the early stage and decrease later in daily variation (from 9:30 to 15:30) with a peak at 11:30 AM (Zhang et al., 2014). Moreover, in Lilium davidii var. unicolor, the highest pollen vigor appeared at 08:00–14:00 with the highest germination rate appeared at 10:00 AM (Sun et al., 2019). The pollen vigor of D. yunnanense was significantly higher than that of D. elatum. This is probably because the pollen vigor of D. elatum has decreased after numerous artificial selections during the process of breeding.
Furthermore, we investigated the pollen–stigma interaction between D. yunnanense and D. elatum. The results showed that the number of pollen grains attached or germinated of D. yunnanense were relatively higher in the reciprocal or allogamic crosses. This was in line with the large pollen quantity and high pollen vitality of D. yunnanense. Moreover, it was detected that no pollen tube entered the ovule in the reciprocal cross, suggesting that there was a strong cross-incompatibility in the reciprocal cross of D. yunnanense × D. elatum. This was related to the cell recognition of the pollen–stigma interaction, which could prevent pollen with genetic variation from germinating on stigma and entering into ovule (Bedinger et al., 2017). It was reported that pre-fertilisation obstacles of interspecific hybridisation in plants were mainly caused by the abnormal growth of pollen tubes, the callose formed on stigma surface and excessive growth of pollen tubes (Heslop-Harrison, 1982; Dumas and Knox, 1983). In our study, the abnormal growth of pollen tubes was observed in the cross of D. yunnanense × D. elatum, such as winded, knotted and spiraled pollen tubes. The formation of callose in stigma mastoid cells was also observed after pollination, which prevented the pollen tube from entering the ovule and consequently resulted in the lack of fertilisation and seed production. Moreover, barriers to reproduction are the most important factor for cross-incompatibility, which can be effective before (prezygotic barriers) and after (postzygotic barriers) the formation of a zygote (Pease et al., 2016). Prezygotic barriers are arguably the most important and effective barriers that contribute to reproductive isolation acting before hybrids formation (Henrich and Kalbe, 2016). In the cross of D. yunnanense × D. elatum, although some pollen grains germinated, the pollen tubes were arrested in the pistils and were unable to reach the ovules. Thus, the pre-fertilisation reproductive barriers are the main factor affecting the cross of D. yunnanense × D. elatum. This phenomenon was also observed in the crosses of different petal phenotypes in Jasminum sambac Aiton (Deng et al., 2017).
Although the fruit set ratio of D. elatum × D. yunnanense was the highest and plenty of F1 seeds were generated, yet the germination ratio of F1 seeds was really low. In addition, we did not harvest any seed or fruit from the cross of D. yunnanense × D. elatum. These results demonstrated that the incompatibility indeed existed between D. yunnanense and D. elatum. One possible theory is that the development of embryo was disturbed and generated immature seed with normal appearance but containing aborted embryos, which reduced the seed vitality (Zeng et al., 2010; Chai et al., 2011; Dai et al., 2014; Chong et al., 2018). This also indicated by the thousand kernel weight of F1 seeds in D. elatum × D. yunnanense, which displayed the lowest thousand kernel weight. Moreover, the germination traits study of D. chefoense showed that the seed germination rate was about 38.0% and 7.3% under 20°C and 25°C, respectively (Qian, 2012). In our study, the seed germination rate of D. yunnanense was about 24.7% at 25°C. The similar low seed germination rate of the wild type in Delphinium species indicated that reproductive strategy of this genus was relatively conserved, which probably provides high recruitment rates for the existing populations with numerous, but the genetic variation and regeneration of natural population may not be able to maintain in the long term (Orellana et al., 2009). Thus, it is necessary to protect the species in Delphinium, since it was also reported that the species number of Delphinium decreased in China (Yuan, 2006). In conclusion, our study provided a fundamental information for the cross-incompatibility in D. yunnanense and D. elatum related to hybridisation. These findings may help ornamental breeders overcome reproductive barriers to develop new cultivars in the species.
