论文部分内容阅读
Abstract Studying the flowering development and seed germination of Eucalyptu dunnii and mastering the reproductive growth law can provide a theoretical basis for seed garden construction. In this study， a blooming plant growing in Zitong County was selected as the experiment subject. The results were as follows： florescence date of Eucalyptus dunnii was Mid-late February to mid to late October. Depending on morphologic change， flowering can be divided into 6 phases： Flower bud differentiation stage， inflorescence formation stage， opercle color change stage， capsule enlargement and shedding ring formation stage， flowering period， seed development stage. The development process of the buds on the same inflorescence was basically the same， the growth of buds near the base of the same branch was faster than that of the top buds， and the development of flower buds in different positions was random. There were 3-4 ovary cavities in a capsule. The number of shriveled seeds was far more than the number of effective seeds， and the average of each capsule contained five normal seeds which could develop into seedlings.
Key words Eucalyptus， reproductive growth， flower bud differentiation， pollen， seed germination
Eucalyptus is one of important timber forests and industrial raw material forest species in Sichuan. As early as the 1980s， Sichuan Forestry Academy began the regional introduction test of Eucalyptus. After many years， many types of eucalypts suitable for planting in Sichuan were selected， such as E. grandis and E. dunnii， E. maidenii， E. camaldulensis， etc. Among them， E. dunnii is a excellent pulp wood species and large diameter wood species[1]， but also has strong ability to withstand cold， which makes it possible to be planted in hilly areas in the central Sichuan Basin and in the basin-mountainous area north of Sichuan. Therefore， it is an ideal species for expanding the planting area of Eucalyptus[2].
E. dunnii is a tall arbor species of series viminales of Section Maidenaria in E. subgenus symphyomyrtus[3]. Compared with those eucalypts， which have been industrialized in production， such as E. urophylla × E. grandis， E. saligna， E. dunnii has problems in rooting difficulties regardless of tissue culture[4-5] or cutting propagation[6]. The technical bottleneck makes it unable to put into industrialized production[7]. At present， the seedling cultivation of E. dunnii is mainly based on seedlings. However， since the seeds are mainly from natural forests in New South Wales， Australia， and it takes E. dunnii trees at least 10 years to bloom， the seeds have low yield but high price. The reported E. dunnii seed orchards mainly are the E. dunnii seedling seed orchard in Huishui County of Guizhou[8]， the E. dunnii clone seed orchard in Huanjiang County of Guangxi[9]. However， it is still hard for E. dunnii trees to bloom and seed， which seriously restricts the development of seed orchards. Therefore， the key research focus of E. dunnii breeding lies in the study of the reproductive growth law of E. dunnii， and solving the difficulty of low seed yield of E. dunnii. In this study， 3 flowering and seeding E. dunnii were found at the end of 2015 was studied for its reproductive growth characteristics of the introduced species in northern Sichuan through observing the characteristics of flower bud differentiation and inflorescence development characteristics， pollen morphology， capsule and seed morphology， and seed germination ability， so as to lay the theoretical and practical foundation for the establishment of E. dunnii seed orchard.
Materials and Methods
Introduction to test forest
Sowing and seedling raising of E. dunnii was done in the spring of 2010， and planted in the mountains in 2011 with the planting density of 2 m× 3 m. In the fall of 2015， 3 plants of E. dunnii were observed to have flower bud differentiation. In 2015， the one with the shortest tree height and under crown height， which was easy to sample and observe， was chosen from the 3 strains of E. dunnii as the study material.
Overview of the test site
The introduced test forest of E. dunnii was located in Liangquan Village， Shuangban Town， Zitong County， Mianyang City. The soil there was purple-yellow soil， which was developed from the lower Cretaceous mother rock， and the soil was sticky of light yellow color. The soil structure was similar to that of the purple soil， of patch structure， tight texture， pH 7.6， slight alkalinity.
Test methods
Flowering period observation From mid-February to mid-to-late October， the changes of buds were observed every 30 d. When a significant change in the shape and color of the buds was observed， the observation interval was shortened to 15 days. The developmental rates of shoot buds within and between shoots were recorded separately.
Meiosis determination Sample fixation： After each observation of the flowering period， the flower buds were cut into halves along the shedding rings on the flower buds using a No. 3 scalpel equipped with a No. 11 blade， and the opercles with anthers were put into and fixed in the centrifuge tube containing Carnot fixed solution， which were then taken into the laboratory. Observation of microspore development： After taking the opercles from the centrifuge tube using a small forceps， anthers were picked from the opercles and then placed on a glass slide. The filaments were cut off， and 10-20 anthers were placed together. Then， a drop of acetic acid magenta staining solution was dropped to the glass side， and the anthers were repeatedly crushed with forceps， so that the microspores in the anthers were released as much as possible. After gently covering the cover glass， the glass side was fired for 1-2 s from below using a lighter to make the chromosomes fully enter into the cells. The side was then placed on a light microscope （BX51） for observation. Capsule structure observation In early January 2017， the capsules of E. dunnii were collected and brought back to the laboratory. The capsules were dried in an oven at 30 ℃ until cracked. The transverse and longitudinal photographs and the seed morphology map of the capsules were separately collected.
Seed germination All seeds and impurities in a capsule were sprinkled in a 90 mm × 90 mm square plastic dish containing absorbent cotton of 0.5 cm thick. After moistening the cotton， the culture dish was placed in an incubator at 25 ℃ for incubation. Distilled water was added to the dish every other day to maintain moisture. After the seeds germination， the seedlings were taken out to take pictures.
Results and Analysis
Flowering period observation and development node judgment
It was observed that after the differentiation， the flower buds of E. dunnii can be distinguished from the leaf buds in morphology. The development of flower buds can be divided into the following six stages in morphology.
（1） Flower bud differentiation stage （Fig. 1-A）： From middle February， flower buds began to germinate， and flower buds and leaf buds could be distinguished from the morphological view. The flower buds were full at the middle， significantly larger than the shuttle-type leaf buds， and throughout the differentiation period， the flower buds gradually grew up and did not differentiate into buds. This stage lasted until mid-to-late June， and took the longest time.
（2） Inflorescence formation stage （Fig. 1-B， C）： The scales on the surface of the buds began to crack and shed off. Each young flower grew out of the bud and stretched in all directions， forming a corymb inflorescence. This stage lasted for 20-30 days. Continued tablet observations revealed that the microspore mother cells appeared in the anther sac until the end of the stage.
（3） Opercle color change stage （Fig. 1-D）： In mid-July， the diameter of the flowers grew to 2 mm， and the color of the opercles began to turn yellow， when the opercles and capsules could be clearly identified. Cutting the flower bud in length using a scalpel found that the flower bud was full of anthers and filaments， and the filaments were curled together （Fig. 2-A）. Microscopic examination revealed that the microspore mother cells started meiosis at this time， mainly in the leptotene stage of early stage I （Fig. 2-B）， when the chromosomes were in a single， thin line， no double chromosomes were visible， and the arrangement of the chromosomes was disorderly filamentous. The nucleus was large in size with thick cytoplasm a very obvious nucleoli. Some chromosomes were connected to the nucleolus at one end and the overall shape resembled a bouquet. This showed that the change in the color of opercles was a morphological marker that indicated the onset of meiosis. （4） Capsule enlargement and shedding ring formation stage （Fig. 1-E， F）： In early August， the diameter of flower buds was as long as 3 mm. The color of the opercles continued to deepen， and some of the flower buds showed a preliminary formation of shedding rings. The opercles could be completely removed by gently ringing the buds at 1mm below the opercles， which could reveal the filaments and stigma. The filaments were arranged closely， the anthers were of diamond shape and transparent， and the style was light green with stigma of dark green color （Fig. 2-C） At this time， the meiosis of microspore mother cells was mainly concentrated in the mid-phase I. Microscopic examination showed that the chromosomes moved to the equatorial plate， and spindle fibers appeared in the cells and spindles were formed （Fig. 2-D）.
（5） Flowering period （Fig. 1-G）： In mid-September， the shedding ring cracked， and under the tension of the filaments， the opercle was lifted and dropped from the opening， and the filaments were stretched from the bud （Fig. 2-E）. Anthers were yellow or faint yellow， stigma was light green with mucus， which could trap wind pollen or pollen carried by insects. Microscopic examination revealed mature pollen grains with a tetrahedral structure and no obvious germination hole （Fig. 2-F）.
（6） Seed development stage （Fig. 1-H）： After pollination， filaments and stigmas gradually faded， and the upper surface of the capsule gradually changed from green to dark gray， and the texture changed from soft to hard. Fertilized eggs began to develop in the ovary and matured in January of the following year.
Comparison of the development flower buds at different locations
The inflorescence of E. dunnii consisted of 5-7 flower buds. The diameter of the flower buds of the same inflorescence was not much different from each other. Microscopic examination showed that the development process of the microspores in the same flower buds was basically the same. However， the average diameters of flower buds in different locations were different as follows： the average diameters of flower buds gradually decreased with the location of inflorescences going from the base to the top. The maximum difference in diameter was greater than 2 mm. This showed that the development of the flower buds at the base was faster than the development of the top buds， which was proved by the early beginning of the flowering stage of basal buds.
The average diameters of the flower buds at different flowering branches of the same mother plant were random， that is， the overall development of the inflorescence had nothing to do with the location of the flowering branch. The average diameter of the flower buds of the flowering branch at the sunny position was significantly larger than the average diameter of the flower buds of the flowering branch at the shade. This indicated that the developmental speed of flower buds at the sunny position was relatively faster， and the sunlight had a benign stimulus to the development of flower buds. Observation of capsule shape and structure
The observation results showed that an inflorescence generally had only 2-4 capsules. Compared with the 5-6 buds in full bloom， it indicated that after pollination， about half of the buds or capsules would drop. This may be due to the fact that the first capsules to complete pollination would secrete hormones which led to the falling of flower buds， or due to the nutrient deficiencies in the mother trees that in order to ensure the normal development of some seeds， the mother trees form detached layers at the bases of other flower buds to promote shedding.
As shown in Fig. 3-A， B， each capsule contained 3-4 ovary cavities， corresponding to the tops cracking into "three diamond shape" and "cross shape". Most of E. dunnii capsules were of cross shape.
The longitudinal structures of the two types of capsules were identical， with the seeds arranged in an up-and-down arc in the cavity （Fig. 3-D）. The upper seeds were longer and could reach 1mm （Fig. 3-E）， while the seeds in the middle were malformed with the length of only about 0.5 mm. These were unfertilized empty seeds which could not develop into seedlings. There was a yellow germination trace at the bottom of the cavity （Fig. 3-F）. Since the seeds and empty seeds were small with similar color， it was difficult to count the number of seeds in the capsule.
Seed germination
Since the seed color and size were similar to those of empty seeds， the seed germination test used the method of sowing the complete contents of the capsule. A total of 6 petri dishes were sown， and a total of 30 seedlings were obtained， an average of 5 seedlings per capsule.
Conclusions and Discussions
In this study， the introduced test forest of E. dunnii was located in Zitong County， Mianyang City， which is the highest latitude of E. dunnii distribution， higher than that of Taizhou Bay of Taizhou City， Zhejiang[10-11]， and much higher than that of the north margin distribution area of E. dunnii in Hunan[12]. Although E. dunnii is widely introduced in southern China， there are few reports on its fertility， such as flowering and fruiting. Some plants with growth-vigor difference were reported to have bloom in Heilongtan Reservoir of Sichuan and Yunnan， and some plants were also reported to flower and fruit in Guizhou Academy of Forestry， but infertile[12]. In Guilin， the introduced E. dunnii flowered after 10 years， but there is no report on the fruiting[13]. Studies on the physiological and biochemical characteristics[11] and gibberellin treatment[14] of the parent plants were carried out in E. dunnii clone seed orchard in Guangxi， but it is still difficult for E. dunnii to blossom and fruit. Through the observation on the reproductive growth characteristics of E. dunnii in the introduced test forest， this study obtained the free pollinated offspring， as well as the reproductive growth characteristics of the introduced E. dunnii in the hilly areas of northern Sichuan， including the flowering habits， the time and morphological features of flower buds and capsules. Specifically include the following：
（1） E. dunnii began the reproductive growth in the middle and late February. The flowering period lasts for 8 m. According to the morphological classification， the development of flower buds can be divided into 6 stages. The meiosis initiation period of male flowers is highly related to the change of the color of the opercles. The change of colors can be used as the morphological marker for the beginning of meiosis. This phenomenon is consistent with the research result of the species in the same subgenus of E. dunnii， namely E. urophylla， E. urophylla × E. tereticornis， and E. urophylla × E. grandis[15-16].
（2） The overall development of the inflorescence has nothing to do with the position of flowering branches， and the development of flower buds at the sunward position is faster. The development of inflorescence at the base of the flowering branch is faster than that at the top of the branch. These conclusions can lay the foundation for the construction of E. dunnii Orchard.
（3） There are 3-4 ovary cavities in the capsule of the introduced E. dunnii， which is slightly lower than the number of common Eucalyptus ovary cavities （4-5）. However， the development of seeds in capsule after pollination is consistent with other eucalyptus reproductive biology studies[3]. That is， there are many pseudo-ovules in the eucalypts. The pseudo-ovules grow into empty seeds， and the number of true seeds formed by the real ovules is small， about 4-10 seeds per capsule. The results of seed germination test obtained the results of an average of 5 seedlings per capsule， indicating that the seeds of the introduced E. dunnii can grow normally， and have the ability to develop into seedlings.
The study obtains preliminary data on the reproductive growth characteristics of E. dunnii， especially the change characteristics of phenotypic traits in flower buds， flowers， capsules and other stages. However， it has not thoroughly studied the internal mechanisms and regulatory measures of E. dunnii. The next step is not only to continue to study reproductive biology characteristics in depth， such as the development rules of long-to-medium-sized spores， pollination and pollen germination， and double fertilization process， but also to explore and study the environmental causes of reproductive growth， such as soil structure， nutrient composition， moisture temperature conditions， in order to find out the technical method which can effectively improve the flowering and seeding rates of E. dunnii， so as to realize the practical application of seed production of E. dunnii as soon as possible. References
[1] CHEN QW， YANG MH， LI ZH， et al. A study of the fiber form variation of cold-resistant Eucalyptus[J]. Journal of Central South Forestry University， 2002，22（04）：61-65.
[2] LUO JZ， CAO JG， LU WH. Cold-tolerance study on 6 eucalypt species[J]. Eucalypt Science & Technology， 2011， 28（02）： 34-38.
[3] WANG HR. A Chinese appreciation of Eucalypt[M]. Beijing： Science Press， 2010.
[4] SONG JY. Bud micropropagation and plantlet-transplanting techniques for breeding cold-resistant strain of Eucalyptus dunnii[J]. Scientia Silvae Sinicae， 2010， 46（6）： 138-145.
[5] SONG JY. Tissue culture of Eucalyptus dunnii seeds[J]. Journal of Central South University of Forestry and Technology， 2008，28（6）：75-80.
[6] LIN CL. Experiment of cutting propagation of Eucalyptus dunnii[J]. Journal of Fujian Forestry Science and Technology， 2005，32（2）：80-84.
[7] ZONG YC， ZHENG YQ， CHANG JC. Research progress of Eucalyptus dunnii introduction and raising technology in China[J]. Chinese Agricultural Science Bulletin， 2010，26（5）：82-86.
[8] LI TH， DENG BL， XIE YJ， et al. Study on management of Eucalyptus dunnii seed orchard[J]. Forestry Resources Management， 2007（02）：82-85.
[9] CHEN JB， LI CR， XIANG DY， et al. A method of selecting and establishing grafting seed orchard of Eucalyptus dunnii[J]. CN 104115677A， October 29， 2014.
[10] LIN Y， WENG Y， LI HP， et al. Effect of salt stress on growth and photosynthetic characteristics of Eucalyptus dunnii seedlings[J]. Journal of Zhejiang Forestry Science and Technology， 2014，34（03）：33-38.
[11] LI CR， GUO DQ， ZHOU W， et al. Comparison of the physiological and biochemical characteristics of flowering for grafted seed orchard of Eucalyptus dunnii[J]. Chinese Agricultural Science Bulletin， 2014， 30（25）：20-23.
[12] LIU YQ， LI ZH， MA Y， et al. Adaptability of Eucalyptus dunnii in the northern limit areas[J]. Journal of Central South University of Forestry and Technology， 2007，27（2）：66-69.
[13] ZHANG JM. The first flowering of introduced Eucalyptus dunnii in Guilin[J]. Guangxi Forestry， 2001（6）：21.
[14] ZHOU W， LI CR， GUO DQ， et al. The effect of GA3 on buds growth of grafting seed orchard of Eucalyptus dunnii[J]. Journal of Fujian Forestry Science and Technology， 2015，42（02）：16-19.
[15] YANG J， LAN J， YAO P， et al. Comparative microsporogenesis and flower development in Eucalyptus urophylla × E. grandis[J]. Journal of Forestry Research， 2016， 27（2）： 257-263.
[16] YANG J， YAO P， LI Y， et al. Induction of 2n pollen with colchicine during microsporogenesis in Eucalyptus[J]. Euphytica， 2016， 210（1）：69-78.
Key words Eucalyptus， reproductive growth， flower bud differentiation， pollen， seed germination
Eucalyptus is one of important timber forests and industrial raw material forest species in Sichuan. As early as the 1980s， Sichuan Forestry Academy began the regional introduction test of Eucalyptus. After many years， many types of eucalypts suitable for planting in Sichuan were selected， such as E. grandis and E. dunnii， E. maidenii， E. camaldulensis， etc. Among them， E. dunnii is a excellent pulp wood species and large diameter wood species[1]， but also has strong ability to withstand cold， which makes it possible to be planted in hilly areas in the central Sichuan Basin and in the basin-mountainous area north of Sichuan. Therefore， it is an ideal species for expanding the planting area of Eucalyptus[2].
E. dunnii is a tall arbor species of series viminales of Section Maidenaria in E. subgenus symphyomyrtus[3]. Compared with those eucalypts， which have been industrialized in production， such as E. urophylla × E. grandis， E. saligna， E. dunnii has problems in rooting difficulties regardless of tissue culture[4-5] or cutting propagation[6]. The technical bottleneck makes it unable to put into industrialized production[7]. At present， the seedling cultivation of E. dunnii is mainly based on seedlings. However， since the seeds are mainly from natural forests in New South Wales， Australia， and it takes E. dunnii trees at least 10 years to bloom， the seeds have low yield but high price. The reported E. dunnii seed orchards mainly are the E. dunnii seedling seed orchard in Huishui County of Guizhou[8]， the E. dunnii clone seed orchard in Huanjiang County of Guangxi[9]. However， it is still hard for E. dunnii trees to bloom and seed， which seriously restricts the development of seed orchards. Therefore， the key research focus of E. dunnii breeding lies in the study of the reproductive growth law of E. dunnii， and solving the difficulty of low seed yield of E. dunnii. In this study， 3 flowering and seeding E. dunnii were found at the end of 2015 was studied for its reproductive growth characteristics of the introduced species in northern Sichuan through observing the characteristics of flower bud differentiation and inflorescence development characteristics， pollen morphology， capsule and seed morphology， and seed germination ability， so as to lay the theoretical and practical foundation for the establishment of E. dunnii seed orchard.
Materials and Methods
Introduction to test forest
Sowing and seedling raising of E. dunnii was done in the spring of 2010， and planted in the mountains in 2011 with the planting density of 2 m× 3 m. In the fall of 2015， 3 plants of E. dunnii were observed to have flower bud differentiation. In 2015， the one with the shortest tree height and under crown height， which was easy to sample and observe， was chosen from the 3 strains of E. dunnii as the study material.
Overview of the test site
The introduced test forest of E. dunnii was located in Liangquan Village， Shuangban Town， Zitong County， Mianyang City. The soil there was purple-yellow soil， which was developed from the lower Cretaceous mother rock， and the soil was sticky of light yellow color. The soil structure was similar to that of the purple soil， of patch structure， tight texture， pH 7.6， slight alkalinity.
Test methods
Flowering period observation From mid-February to mid-to-late October， the changes of buds were observed every 30 d. When a significant change in the shape and color of the buds was observed， the observation interval was shortened to 15 days. The developmental rates of shoot buds within and between shoots were recorded separately.
Meiosis determination Sample fixation： After each observation of the flowering period， the flower buds were cut into halves along the shedding rings on the flower buds using a No. 3 scalpel equipped with a No. 11 blade， and the opercles with anthers were put into and fixed in the centrifuge tube containing Carnot fixed solution， which were then taken into the laboratory. Observation of microspore development： After taking the opercles from the centrifuge tube using a small forceps， anthers were picked from the opercles and then placed on a glass slide. The filaments were cut off， and 10-20 anthers were placed together. Then， a drop of acetic acid magenta staining solution was dropped to the glass side， and the anthers were repeatedly crushed with forceps， so that the microspores in the anthers were released as much as possible. After gently covering the cover glass， the glass side was fired for 1-2 s from below using a lighter to make the chromosomes fully enter into the cells. The side was then placed on a light microscope （BX51） for observation. Capsule structure observation In early January 2017， the capsules of E. dunnii were collected and brought back to the laboratory. The capsules were dried in an oven at 30 ℃ until cracked. The transverse and longitudinal photographs and the seed morphology map of the capsules were separately collected.
Seed germination All seeds and impurities in a capsule were sprinkled in a 90 mm × 90 mm square plastic dish containing absorbent cotton of 0.5 cm thick. After moistening the cotton， the culture dish was placed in an incubator at 25 ℃ for incubation. Distilled water was added to the dish every other day to maintain moisture. After the seeds germination， the seedlings were taken out to take pictures.
Results and Analysis
Flowering period observation and development node judgment
It was observed that after the differentiation， the flower buds of E. dunnii can be distinguished from the leaf buds in morphology. The development of flower buds can be divided into the following six stages in morphology.
（1） Flower bud differentiation stage （Fig. 1-A）： From middle February， flower buds began to germinate， and flower buds and leaf buds could be distinguished from the morphological view. The flower buds were full at the middle， significantly larger than the shuttle-type leaf buds， and throughout the differentiation period， the flower buds gradually grew up and did not differentiate into buds. This stage lasted until mid-to-late June， and took the longest time.
（2） Inflorescence formation stage （Fig. 1-B， C）： The scales on the surface of the buds began to crack and shed off. Each young flower grew out of the bud and stretched in all directions， forming a corymb inflorescence. This stage lasted for 20-30 days. Continued tablet observations revealed that the microspore mother cells appeared in the anther sac until the end of the stage.
（3） Opercle color change stage （Fig. 1-D）： In mid-July， the diameter of the flowers grew to 2 mm， and the color of the opercles began to turn yellow， when the opercles and capsules could be clearly identified. Cutting the flower bud in length using a scalpel found that the flower bud was full of anthers and filaments， and the filaments were curled together （Fig. 2-A）. Microscopic examination revealed that the microspore mother cells started meiosis at this time， mainly in the leptotene stage of early stage I （Fig. 2-B）， when the chromosomes were in a single， thin line， no double chromosomes were visible， and the arrangement of the chromosomes was disorderly filamentous. The nucleus was large in size with thick cytoplasm a very obvious nucleoli. Some chromosomes were connected to the nucleolus at one end and the overall shape resembled a bouquet. This showed that the change in the color of opercles was a morphological marker that indicated the onset of meiosis. （4） Capsule enlargement and shedding ring formation stage （Fig. 1-E， F）： In early August， the diameter of flower buds was as long as 3 mm. The color of the opercles continued to deepen， and some of the flower buds showed a preliminary formation of shedding rings. The opercles could be completely removed by gently ringing the buds at 1mm below the opercles， which could reveal the filaments and stigma. The filaments were arranged closely， the anthers were of diamond shape and transparent， and the style was light green with stigma of dark green color （Fig. 2-C） At this time， the meiosis of microspore mother cells was mainly concentrated in the mid-phase I. Microscopic examination showed that the chromosomes moved to the equatorial plate， and spindle fibers appeared in the cells and spindles were formed （Fig. 2-D）.
（5） Flowering period （Fig. 1-G）： In mid-September， the shedding ring cracked， and under the tension of the filaments， the opercle was lifted and dropped from the opening， and the filaments were stretched from the bud （Fig. 2-E）. Anthers were yellow or faint yellow， stigma was light green with mucus， which could trap wind pollen or pollen carried by insects. Microscopic examination revealed mature pollen grains with a tetrahedral structure and no obvious germination hole （Fig. 2-F）.
（6） Seed development stage （Fig. 1-H）： After pollination， filaments and stigmas gradually faded， and the upper surface of the capsule gradually changed from green to dark gray， and the texture changed from soft to hard. Fertilized eggs began to develop in the ovary and matured in January of the following year.
Comparison of the development flower buds at different locations
The inflorescence of E. dunnii consisted of 5-7 flower buds. The diameter of the flower buds of the same inflorescence was not much different from each other. Microscopic examination showed that the development process of the microspores in the same flower buds was basically the same. However， the average diameters of flower buds in different locations were different as follows： the average diameters of flower buds gradually decreased with the location of inflorescences going from the base to the top. The maximum difference in diameter was greater than 2 mm. This showed that the development of the flower buds at the base was faster than the development of the top buds， which was proved by the early beginning of the flowering stage of basal buds.
The average diameters of the flower buds at different flowering branches of the same mother plant were random， that is， the overall development of the inflorescence had nothing to do with the location of the flowering branch. The average diameter of the flower buds of the flowering branch at the sunny position was significantly larger than the average diameter of the flower buds of the flowering branch at the shade. This indicated that the developmental speed of flower buds at the sunny position was relatively faster， and the sunlight had a benign stimulus to the development of flower buds. Observation of capsule shape and structure
The observation results showed that an inflorescence generally had only 2-4 capsules. Compared with the 5-6 buds in full bloom， it indicated that after pollination， about half of the buds or capsules would drop. This may be due to the fact that the first capsules to complete pollination would secrete hormones which led to the falling of flower buds， or due to the nutrient deficiencies in the mother trees that in order to ensure the normal development of some seeds， the mother trees form detached layers at the bases of other flower buds to promote shedding.
As shown in Fig. 3-A， B， each capsule contained 3-4 ovary cavities， corresponding to the tops cracking into "three diamond shape" and "cross shape". Most of E. dunnii capsules were of cross shape.
The longitudinal structures of the two types of capsules were identical， with the seeds arranged in an up-and-down arc in the cavity （Fig. 3-D）. The upper seeds were longer and could reach 1mm （Fig. 3-E）， while the seeds in the middle were malformed with the length of only about 0.5 mm. These were unfertilized empty seeds which could not develop into seedlings. There was a yellow germination trace at the bottom of the cavity （Fig. 3-F）. Since the seeds and empty seeds were small with similar color， it was difficult to count the number of seeds in the capsule.
Seed germination
Since the seed color and size were similar to those of empty seeds， the seed germination test used the method of sowing the complete contents of the capsule. A total of 6 petri dishes were sown， and a total of 30 seedlings were obtained， an average of 5 seedlings per capsule.
Conclusions and Discussions
In this study， the introduced test forest of E. dunnii was located in Zitong County， Mianyang City， which is the highest latitude of E. dunnii distribution， higher than that of Taizhou Bay of Taizhou City， Zhejiang[10-11]， and much higher than that of the north margin distribution area of E. dunnii in Hunan[12]. Although E. dunnii is widely introduced in southern China， there are few reports on its fertility， such as flowering and fruiting. Some plants with growth-vigor difference were reported to have bloom in Heilongtan Reservoir of Sichuan and Yunnan， and some plants were also reported to flower and fruit in Guizhou Academy of Forestry， but infertile[12]. In Guilin， the introduced E. dunnii flowered after 10 years， but there is no report on the fruiting[13]. Studies on the physiological and biochemical characteristics[11] and gibberellin treatment[14] of the parent plants were carried out in E. dunnii clone seed orchard in Guangxi， but it is still difficult for E. dunnii to blossom and fruit. Through the observation on the reproductive growth characteristics of E. dunnii in the introduced test forest， this study obtained the free pollinated offspring， as well as the reproductive growth characteristics of the introduced E. dunnii in the hilly areas of northern Sichuan， including the flowering habits， the time and morphological features of flower buds and capsules. Specifically include the following：
（1） E. dunnii began the reproductive growth in the middle and late February. The flowering period lasts for 8 m. According to the morphological classification， the development of flower buds can be divided into 6 stages. The meiosis initiation period of male flowers is highly related to the change of the color of the opercles. The change of colors can be used as the morphological marker for the beginning of meiosis. This phenomenon is consistent with the research result of the species in the same subgenus of E. dunnii， namely E. urophylla， E. urophylla × E. tereticornis， and E. urophylla × E. grandis[15-16].
（2） The overall development of the inflorescence has nothing to do with the position of flowering branches， and the development of flower buds at the sunward position is faster. The development of inflorescence at the base of the flowering branch is faster than that at the top of the branch. These conclusions can lay the foundation for the construction of E. dunnii Orchard.
（3） There are 3-4 ovary cavities in the capsule of the introduced E. dunnii， which is slightly lower than the number of common Eucalyptus ovary cavities （4-5）. However， the development of seeds in capsule after pollination is consistent with other eucalyptus reproductive biology studies[3]. That is， there are many pseudo-ovules in the eucalypts. The pseudo-ovules grow into empty seeds， and the number of true seeds formed by the real ovules is small， about 4-10 seeds per capsule. The results of seed germination test obtained the results of an average of 5 seedlings per capsule， indicating that the seeds of the introduced E. dunnii can grow normally， and have the ability to develop into seedlings.
The study obtains preliminary data on the reproductive growth characteristics of E. dunnii， especially the change characteristics of phenotypic traits in flower buds， flowers， capsules and other stages. However， it has not thoroughly studied the internal mechanisms and regulatory measures of E. dunnii. The next step is not only to continue to study reproductive biology characteristics in depth， such as the development rules of long-to-medium-sized spores， pollination and pollen germination， and double fertilization process， but also to explore and study the environmental causes of reproductive growth， such as soil structure， nutrient composition， moisture temperature conditions， in order to find out the technical method which can effectively improve the flowering and seeding rates of E. dunnii， so as to realize the practical application of seed production of E. dunnii as soon as possible. References
[1] CHEN QW， YANG MH， LI ZH， et al. A study of the fiber form variation of cold-resistant Eucalyptus[J]. Journal of Central South Forestry University， 2002，22（04）：61-65.
[2] LUO JZ， CAO JG， LU WH. Cold-tolerance study on 6 eucalypt species[J]. Eucalypt Science & Technology， 2011， 28（02）： 34-38.
[3] WANG HR. A Chinese appreciation of Eucalypt[M]. Beijing： Science Press， 2010.
[4] SONG JY. Bud micropropagation and plantlet-transplanting techniques for breeding cold-resistant strain of Eucalyptus dunnii[J]. Scientia Silvae Sinicae， 2010， 46（6）： 138-145.
[5] SONG JY. Tissue culture of Eucalyptus dunnii seeds[J]. Journal of Central South University of Forestry and Technology， 2008，28（6）：75-80.
[6] LIN CL. Experiment of cutting propagation of Eucalyptus dunnii[J]. Journal of Fujian Forestry Science and Technology， 2005，32（2）：80-84.
[7] ZONG YC， ZHENG YQ， CHANG JC. Research progress of Eucalyptus dunnii introduction and raising technology in China[J]. Chinese Agricultural Science Bulletin， 2010，26（5）：82-86.
[8] LI TH， DENG BL， XIE YJ， et al. Study on management of Eucalyptus dunnii seed orchard[J]. Forestry Resources Management， 2007（02）：82-85.
[9] CHEN JB， LI CR， XIANG DY， et al. A method of selecting and establishing grafting seed orchard of Eucalyptus dunnii[J]. CN 104115677A， October 29， 2014.
[10] LIN Y， WENG Y， LI HP， et al. Effect of salt stress on growth and photosynthetic characteristics of Eucalyptus dunnii seedlings[J]. Journal of Zhejiang Forestry Science and Technology， 2014，34（03）：33-38.
[11] LI CR， GUO DQ， ZHOU W， et al. Comparison of the physiological and biochemical characteristics of flowering for grafted seed orchard of Eucalyptus dunnii[J]. Chinese Agricultural Science Bulletin， 2014， 30（25）：20-23.
[12] LIU YQ， LI ZH， MA Y， et al. Adaptability of Eucalyptus dunnii in the northern limit areas[J]. Journal of Central South University of Forestry and Technology， 2007，27（2）：66-69.
[13] ZHANG JM. The first flowering of introduced Eucalyptus dunnii in Guilin[J]. Guangxi Forestry， 2001（6）：21.
[14] ZHOU W， LI CR， GUO DQ， et al. The effect of GA3 on buds growth of grafting seed orchard of Eucalyptus dunnii[J]. Journal of Fujian Forestry Science and Technology， 2015，42（02）：16-19.
[15] YANG J， LAN J， YAO P， et al. Comparative microsporogenesis and flower development in Eucalyptus urophylla × E. grandis[J]. Journal of Forestry Research， 2016， 27（2）： 257-263.
[16] YANG J， YAO P， LI Y， et al. Induction of 2n pollen with colchicine during microsporogenesis in Eucalyptus[J]. Euphytica， 2016， 210（1）：69-78.