| Vicia unijuga is a valuable forage legume in Gannan Tibetan autonomous prefecture in Gansu province. Researchers at Lanzhou University started the domestication and breeding of V. unijuga several years ago and found that the application and promotion of this species was restricted by its low seed setting rate. In this study, we focused on the reproduction strategy of V. unijuga and investigated the breeding system, pollination ecology, fruit set patterns and reproductive allocation that may affect the seed production. The main results are as follows:1. Breeding system. The mating system of V. unijuga showed a self-compatible mode; however, outcrossing was the dominated form and pollinators were needed. Moreover, anther-stigma distance (ASD) was absent in around 30% flowers. Opening inflorescences lasted 7-12 days and opening flowers lasted 6-10 days in V. unijuga of different ages. The pollen viability was generally higher than 70%. The pollen-ovule ratio (P/O) was 79365.1 and 72222.2 for the two-year-old and four-year-old population, respectively. The outcrossing index (OCI) was estimated to be three or four. On the 10th day after the pollination, the fruit set of each treatment was about 30% and there was no significant difference (P>0.05) between the results obtained from the selfing and outcrossing treatments. During the seed maturation stage, the fruit set of each treatment decreased by 10%. The fruit set of the autonomous flower selfing treatment was markedly lower (P<0.05) than that of the manual selfing, manual inflorescences selfing, manual geitonogamous selfing, manual outcrossing, autonomous outcrossing treatments and the control. Under a fluorescence microscope, we observed that the number of germinated pollen of the outcrossing treatment was more than that of the selfing treatment.2. Pollinators and pollination behaviors. Among the 13 insect species visiting the flowers, eight Hymenoptera insects were pollinators and they are Bombus Lepidus, B. sichelii, B. patagiatus, B. pyrosoma, B. personatus, B. rufofasciatus, Apis cerana and a kind of wild honey bees (Apis spp.). According to the pollinators’ visitation rate and appearance during the flowering stage, Bombus lepidus was considered to be the most efficient pollinator. It is found that the diurnal pollinator activity displayed a similar trend to that of the diurnal flowering dynamics, which was a double-peak curve. The presence of other native flowering plants near the study site may affect the pollination and reproduction of V. unijuga. The fruit set of insect pollinated flowers (2.8) was found to be significantly lower (P<0.05) than that of hand cross-pollinated flowers (4.2) on the 20th day after the pollination. However, there was no pronounced difference (P>0.05) between the fruit sets of the two groups (2.2 and 3.0) during the seed maturation stage. In addition, the difference was not significant (P>0.05) in pod length and mass, seed number, mass per pod, aborted seeds and ovule number per pod between the two treatments.3. Fruit set patterns. In terms of time, the highest fruit set was observed for the two-year-old population (28.8%) at the full-bloom stage and for the five-year-old population (41.6%) at the early flowering stage. For both populations, the lowest fruit set (3.0% and 1.1%) was found at the end of flowering stage. There was no appreciable difference (P>0.05) in the seed number per pod between the early flowering and full-bloom stage for the two populations, and the seed number per pod at the end of flowering stage were zero. In terms of plant tissues, the fruit set of the inflorescence in the middle was the highest in both populations, with 43.9% for the two-year-old and 41.6% for the five-year-old population; however, the seed number per pod exhibited no significant difference (P>0.05) in each site of inflorescence in the two populations. The ovules closing to the stigma and pedicel were more likely to abort. The ovules in the 2nd,3rd,4th and 5th positions starting from the top of the fruit would have higher possibilities to form mature seeds and show alternating abortion patterns. The number of fruit per inflorescence was lower than eight, which was not affected by the number of flowers.4. Reproduction biomass allocation. The vegetative biomass allocation was dominated and the reproductive biomass allocation (at the bud stage and full-bloom stage) was below 10% for V unijuga of different ages. Throughout the growing season, the reproductive biomass allocation was in the order of pod-setting stage> full-bloom stage> bud stage. The reproduction biomass allocation was five-year-old> four-year-old> two-year-old population at the full-bloom and bud stage, and two-year-old> five-year-old> four-year-old population at the pod-setting stage. The conversion ratios from flower buds to flowers and from flowers to pods for the two-year-old plants were 1:1.17 and 1:1.36, respectively, while those of the five-year-old plants were 1:0.94 and 1:1.31, respectively. It indicates that the reproduction utilization rate was higher for the two-year-old plants than the five-year-old plants.5. Seed yield. The potential seed yield per plant, the expressive seed yield per plant, the actual seed yield per plant, the proportion of the actual seed yield in the potential seed yield, the proportion of seed yield loss during the seed maturation stage and the rate of the relative reproductive success for the two-year-old group were 42.2 g,18.6 g,1.91 g,4.53%,89.71% and 23.57%, respectively, while those for the five-year-old group were 183.6 g,42.2 g,4.60 g,2.51%,89.11% and 11.41%, respectively. The top three factors that are closely correlated to the seed yield of the two-year-old group are podding rate/inflorescence, seed number/pod and fertile stems/plant, whereas those factors for the five-year-old population are thousand seed weight, podding rate/inflorescence and fertile stems/plant. Path analysis showed that the parameters affecting the seed yield per plant were in the order of mature pods/plant> thousand seed weight> fertile stems/plant for V. unijuga of different ages. Among the density treatments of 87500,50000,37500 and 30000 plants/hm2, the 87500 and 37500 plants/hm2 treatments showed the highest actual seed yields, which was 699.6±11.3 kg/hm2 and 619.9±44.0 kg/hm2, respectively. The thousand seed weight was 7.73±0.22 g and 8.80±0.12 g for the above two treatments, respectively. There was no significant difference (P>0.05) between the two seed yields whereas the two thousand seed weight exhibited significant difference (P<0.05). Since actual seed yield reflects seed numbers and thousand seed weight reflects seed quality, the treatment of 37500 plants/hm2 might be more beneficial for the seed production, due to the high actual seed yield and thousand seed weight.In summary, the low seed yield of V. unijuga at Gannan Tibetan region was related to the selfing mechanism in the breeding system, the pollinator behaviors, the selective abortion of fruits and seeds as well as the low reproductive allocation. The seed production could be improved with reasonable plant density. |
PDF Full Text Request