Variation of Traits on Seeds and Germination Derived from the Hybridization between the Sections Tacamahaca and Aigeiros of the Genus Populus

Poplar is an important research organism, and species in sections Tacamahaca and Aigeiros, have advantages in terms of stress resistance, ease of propagation, and fast growth. Poplar species are widely distributed and well-adapted in the world, presenting a large potential for genetic improvement. Hybridization between different species allows us to generate offspring with a unique combinations of traits. This approach has a huge potential for breeding new poplar varieties that could aid in controlling desertification in the arid and semi-arid zones of the “Three-North” in China. In this study, we carried out a cross test scheme with nine female and thirteen male poplar trees. A total of 105,401 seeds were collected from 117 crosses. Flowering phenology and seed maturation differences of the hybrid progeny were monitored in greenhouses. For male trees, Populus deltoides had the longest flowering time. For female trees, Populus pseudo-simonii showed the longest seed maturity time. The number of carpals and ovules were not the same in different females. Meanwhile, three carpals were found in P. pseudo-simonii. A highly significant positive correlation was found between the seed size and the Thousand Kernel Weight, as well as the seedling cotyledon length. During seed germination, non-radicle and non-hypocotyl seedlings were observed. We also observed a number of cotyledon variants, including single and fused cotyledons, two cotyledons with one cotyledon cracking into two parts, three cotyledons, as well as four cotyledons. These results lay a favorable foundation for combining the research between the sections Tacamahaca and Aigeiros in future work.


Introduction
Desertification is one of the most serious ecological and environmental problems in the world [1][2][3][4][5].It often causes water and wind erosion, vegetation degradation, and soil salinization [2].Vast areas suffer from it in China, especially the "Three-North" (northeast, northwest, and north) regions (Figure 1).Inner Mongolia, for example, has seen a profound expansion of desertification and by 2008, it had reached nearly 185,000 km 2 [6].Desertification contributes to poor soil fertility and economic losses in these afflicted areas, particularly in arid, semi-arid, and dry sub-humid zones [7][8][9].Shelterbelt construction is one of the controlling measures for desertification.However, the vegetation of shelterbelt is not easily established [10] because of lack of tree varieties.Therefore, it is necessary to enhance variety breeding of afforestation trees in these regions and create new forest cultivars with a superior quality and high stress resistance.Poplar trees are widely distributed over the world and play a significant role in ecological and environmental construction [11].Taxonomically, poplar species are grouped into five sections, Populus, Aigeiros, Tacamahaca, Turanga, and Leucoides [12,13].More than 100 poplar species have been discovered in the world, and 53 species exist in China [14].The genus Populus L. has the characters of rapid growth, wide adaption, high fiber contents, and easy propagation [14][15][16].For their rapid growth property, these trees are often used as a source of fiber, fuel, lumber, and to prevent soil erosion, they are also planted into windbreaks and protective stands [17].Nevertheless, in the construction of shelterbelt in China, most of the poplar plantations have been showing low productivity and even more they are small, old trees, such aged ones lacking variety [18].In order to solve this problem, it is better to create new poplar cultivars.
Hybridization is an efficient way to create a new genotype and plays an important role in the genetic improvement of numerous crops and animals [19][20][21].Compared with other breeding methods, hybridization could recombine the parents' genes, thus, combining the superior genes with different characters controlled by the parents into an organic whole or accumulate different minor genes with the same characters controlled by the parents, and contribute to a new genotype with all the characters.That is so-called heterosis, a complex genetical phenomenon [22].It should be considered as the interaction of favorable combinations of genes at different loci for the parental lines [23].Hybrid vigor may play an essential role in maintaining the genetic variation in populations [24].Parental selection is very crucial during hybridization.Several principles should be considered, including the breeding objective, complementary characters, high combinability, heritability, far genetic relationship, and cross ability [19].
The history of poplar breeding spans more than a 100 year period [16].In the hybridization experiment, interspecies hybridization within a section is easier than that between sections without manipulative techniques [16,25], whereas the crossability between sections is not the same.Crosses between sections Aigeiros and Populus, and Tacamahaca and Populus are quite difficult, while those between Aigeiros and Tacamahaca are easy [12,26,27].The section Tacamahaca is the largest group of Populus, possessing an abundance of gene resources.Populus simonii and P. pseudo-simonii, are two widely distributed native tree species in China, and commonly appear in desertification regions, possessing features of easy rooting, drought tolerance, cold resistance, and wide adaptation [12,14,25].In poplar genetic improvement, the section Tacamahaca frequently plays a role of the complementary parent.However, in contrast to other poplar groups, the level of genetic improvement is lower.Therefore, it is essential to enhance the genetic improvement level of the section Tacamahaca tree species.The section Aigeiros is distributed in North America, Europe, and Western Asia [17].Among which, P. deltoides, Populus nigra, and their hybrids are important in the hybridization program of poplar in China because they have the characteristics of rapid growth, a straight stem, and a resistance to diseases and pests, contrasting to P. simonii, P. pseudo-simonii, and especially P. deltoides, which are generally difficult to be propagated by hardwood cuttings [17].Utilizing hybridization between section Tacamahaca and Aigeiros would allow for making the traits complementary and develop poplar varieties that are adaptable to use within shelterbelt programs in the "Three-North" regions.
In this work, we estimated the crosses compatibility among section Tacamahaca, section Aigeiros, and their hybrids.For this, we detected the seed and seedling traits of 117 crosses coming from a test cross scheme between 9 female parents and 13 male parents.The seed traits included seed length, seed width, and Thousand Kernel Weight (TKW).The seedling traits included the performance of the embryo root, the hypocotyl and cotyledon morphology, and length.In addition, seed formation rate and seed germination rate were also studied.Meanwhile, in order to describe the seed formation rate, the ovary observation and in vivo pollen germination test were performed.

Plant Materials
Twenty-two cultivars were used in this study.Among which, nine were female and thirteen were pollen donors (Table 1).Their flower branches were collected from Tongliao city in Inner Mongolia, Wuwei city in the Gansu province, Yiling Plain Forest Farm in Sinkiang, and from other places in China during December 2012 (Figure 1).The flower branches were packaged tightly with a plastic film and put in low temperature specified storage.Male branches were cultured in water with all the flower buds and one leaf bud for each branch in the greenhouse of the Beijing Forestry University with an average temperature of 10-20 • C and 50%-70% humidity during January 2013.About two weeks later, the female branches were cultured, keeping four robust and full flower buds and two leaf buds for each branch.Three leaves remained when the buds unfolded.The water was changed two or three times a week.

Pollen Collection and Pollination
A total of 117 cross combinations were made using an entire 9 × 13 test cross mating design scheme; that is, every female parent crossed with every male parent.The pollen was collected during anthesis of male flowers and immediately stored at −20 • C until its utilization.A controlled pollination hybridization was carried out in February 2013.When the female's stigma became shiny, pollination was conducted by softly transferring pollen onto it with a brush and stopped when it was covered with pollen.In all cases, we performed three replicate pollinations for each stigma.We also strictly carried out work in isolation for all the materials before powdering the stamens and well bagging the pistils with paper bags after pollination until the sigma lost receptivity.
We noted the number of pollinated catkins per cross, and, meanwhile, five traits were recorded during the water cultured procedure, including the days before flowering (starting from water cultivation to first staminate flower anthesis), duration of pollen shedding, days before pollination (starting from water cultivation), days to seed maturation (starting from pollination), and number of capsules per catkin for each cross combination.

In Vitro Pollen Germination
The flowers were harvested 24 h after pollination and immediately fixed in Carnoy solution (3:1 ethanol/acetic acid) at 4 • C.Over a 24 h fixation, they were stored in 70% ethanol.The stigmas were removed from the fixed flowers, washed three times using distilled water, and softened in 8 M sodium hydroxide at room temperature for 4 h, rinsed briefly, then mounted in a drop of 0.1% aniline blue prepared with 0.2 M dipotassium phosphate staining for 5-10 min on a microslide.Then the coverslips were lowered onto the tissue and extruded with the appropriate strength.The slides were observed under a fluorescent microscope (Olympus BX51, Tokyo, Japan) and the photos were taken with an attached charge-coupled device video camera (Olympus DP70, Tokyo, Japan).

Ovary Observation
Ten fixed florets were randomly chosen for each female parent.The numbers of carpel and ovule per capsule were counted under a stereo anatomical lens (Olympus SZX12, Tokyo, Japan) and photos were taken with an attached image capture camera (Olympus U-CMAD3, Tokyo, Japan).

Seed Traits Measurements
Seed traits including the TKW, seed length and width, and seed germination rate were collected for the per cross combination in this research.For the TKW of every cross, three 100-seed samples were randomly counted and weighed, then converted into TKW.A sample of 30-seed was randomly chosen for measuring the seed length and width of each cross.Five-hundred seeds were randomly extracted for germination for each treatment.We placed the seeds in glass culture dishes with wet filter paper in a climate chamber (24 h at 25 • C under light and 50% relative humidity).After two days of germination, the number of germinated seeds was counted.The seed formation rate was defined as the percentage of seed formation from ovules in each capsule, which was calculated as follows: Seed formation rate = (Seed number per female catkin ÷ capsule number per catkin) ÷ ovule number per capsule × 100%.

Seedling Traits Measurement
After two days of germination in the chamber, we found abnormal development of the radicle and cotyledon, and counted the number of normal rooting seedlings, non-radicle, and non-hypocotyl seedlings.The cotyledon length of the 30 seedlings was measured for each cross after one month of growth in the soil under the greenhouse condition.

Statistical Analysis
The seed formation rate and seed germination rate were subjected to arcsine root square transformation for statistical analysis.The analyses of variance (ANOVA), Pearson's correlation coefficients, and Duncan's multiple range test were performed for the days before flowering, the duration of pollen shedding, the days before pollination, the days to seed maturation, the days for seed development, the ovule number of per capsule, the seed formation rate, the seed length, the seed width, the seed germination rate, the seedling no radicle rate, and the seedling cotyledon length, using SPSS 16.0 (SPSS Inc., Chicago, IL, USA).

Seedling Traits Measurement
After two days of germination in the chamber, we found abnormal development of the radicle and cotyledon, and counted the number of normal rooting seedlings, non-radicle, and non-hypocotyl seedlings.The cotyledon length of the 30 seedlings was measured for each cross after one month of growth in the soil under the greenhouse condition.

Statistical Analysis
The seed formation rate and seed germination rate were subjected to arcsine root square transformation for statistical analysis.The analyses of variance (ANOVA), Pearson's correlation coefficients, and Duncan's multiple range test were performed for the days before flowering, the duration of pollen shedding, the days before pollination, the days to seed maturation, the days for seed development, the ovule number of per capsule, the seed formation rate, the seed length, the seed width, the seed germination rate, the seedling no radicle rate, and the seedling cotyledon length, using SPSS 16.0 (SPSS Inc., Chicago, IL, USA).

Figure 2.
The flowering phenology on days before flowering and for duration of pollen shedding for the 13 male parents under the condition of water culturing in a greenhouse (mean + s.e.).The bars followed by the same letter are not statistically different (p < 0.05, n = 3).Duncan's multiple range test was taken.
The days before pollination was significant (p = 0.0057), it was about 26 days for female parents (Figure 3), while the time for the crosses from pollination to seed maturation was not exactly the same.ANOVA revealed that there was a highly significant effect (p = 0.0001) on the days to seed The days before pollination was significant (p = 0.0057), it was about 26 days for female parents (Figure 3), while the time for the crosses from pollination to seed maturation was not exactly the same.ANOVA revealed that there was a highly significant effect (p = 0.0001) on the days to seed maturation according to the choice of females.The crosses that took P. pseudo-simonii as the female parent required the longest time for seed maturation, which was above 40 days (Figure 3).However, non-significant differences were observed in seed maturation for different male parents (p = 0.7324).maturation according to the choice of females.The crosses that took P. pseudo-simonii as the female parent required the longest time for seed maturation, which was above 40 days (Figure 3).However, non-significant differences were observed in seed maturation for different male parents (p = 0.7324).

Figure 3.
The phenology of the days before pollination and seed maturation for the 9 female parents under the condition of water culturing in a greenhouse (mean + s.e.).The bars followed by the same letter are not statistically different (p < 0.05, n = 13).Duncan's multiple range test was taken.

Variation in Ovary Traits of Female Parents
The anatomical observation revealed that there was a highly significant difference in the number of carpels and ovules per capsule for the female parents (p = 0.0001 both for the carpel and ovule).Most of the capsules had two carpels per capsule (Figure 4a), while three carpels were found in P. pseudo-simonii (Figure 4b).Cultivar "XQ-2" was at an 80% proportion of tree-carpel.The number of ovules per capsule for P. simonii was minimum, especially for the cultivar "XY-2", which merely had 5.2 ± 0.3 ovules per capsule.Higher numbers were observed in other females.

Seed Formation Rate and Seed Morphological Variation
A total of 117 cross combinations were made by us and a total of 105,401 seeds were harvested.Most of the combinations hybridized between section Tacamahaca and P. deltoides only obtained shriveled seeds (Figure 5a,b).However, the germination of pollen on the stigma was quite well done (Figure 6a,b).The combinations which took P. pseudo-simonii as the female parent and "ZY3", "EBY", or hybrid "BJLY3" as the male parent could obtain full seeds (Figure 5c,d).The results of the hybridization between section Tacamahaca as well as their hybrids and P. nigra were generally fine with plump seeds (Figure 5e-h

Variation in Ovary Traits of Female Parents
The anatomical observation revealed that there was a highly significant difference in the number of carpels and ovules per capsule for the female parents (p = 0.0001 both for the carpel and ovule).Most of the capsules had two carpels per capsule (Figure 4a), while three carpels were found in P. pseudo-simonii (Figure 4b).Cultivar "XQ-2" was at an 80% proportion of tree-carpel.The number of ovules per capsule for P. simonii was minimum, especially for the cultivar "XY-2", which merely had 5.2 ± 0.3 ovules per capsule.Higher numbers were observed in other females.
maturation according to the choice of females.The crosses that took P. pseudo-simonii as the female parent required the longest time for seed maturation, which was above 40 days (Figure 3).However, non-significant differences were observed in seed maturation for different male parents (p = 0.7324).

Figure 3.
The phenology of the days before pollination and seed maturation for the 9 female parents under the condition of water culturing in a greenhouse (mean + s.e.).The bars followed by the same letter are not statistically different (p < 0.05, n = 13).Duncan's multiple range test was taken.

Variation in Ovary Traits of Female Parents
The anatomical observation revealed that there was a highly significant difference in the number of carpels and ovules per capsule for the female parents (p = 0.0001 both for the carpel and ovule).Most of the capsules had two carpels per capsule (Figure 4a), while three carpels were found in P. pseudo-simonii (Figure 4b).Cultivar "XQ-2" was at an 80% proportion of tree-carpel.The number of ovules per capsule for P. simonii was minimum, especially for the cultivar "XY-2", which merely had 5.2 ± 0.3 ovules per capsule.Higher numbers were observed in other females.

Seed Formation Rate and Seed Morphological Variation
A total of 117 cross combinations were made by us and a total of 105,401 seeds were harvested.Most of the combinations hybridized between section Tacamahaca and P. deltoides only obtained shriveled seeds (Figure 5a,b).However, the germination of pollen on the stigma was quite well done (Figure 6a,b).The combinations which took P. pseudo-simonii as the female parent and "ZY3", "EBY", or hybrid "BJLY3" as the male parent could obtain full seeds (Figure 5c,d).The results of the hybridization between section Tacamahaca as well as their hybrids and P. nigra were generally fine with plump seeds (Figure 5e-h

Seed Formation Rate and Seed Morphological Variation
A total of 117 cross combinations were made by us and a total of 105,401 seeds were harvested.Most of the combinations hybridized between section Tacamahaca and P. deltoides only obtained shriveled seeds (Figure 5a,b).However, the germination of pollen on the stigma was quite well done (Figure 6a,b).The combinations which took P. pseudo-simonii as the female parent and "ZY3", "EBY", or hybrid "BJLY3" as the male parent could obtain full seeds (Figure 5c,d).The results of the hybridization between section Tacamahaca as well as their hybrids and P. nigra were generally fine with plump seeds (Figure 5e-h).A few combinations did not obtain seed, especially the combinations which took hybrids "ZY3" or "EBY" as the female parent and P. deltoides as the male parent with a Forests 2018, 9, 516 7 of 15 serious problem of falling catkins and fruits.Additionally, this situation also occurred in the crosses taking P. simonii as the female parent and cultivars "137", "144", or hybrid "BJY" as the male parent.ANOVA indicated that the parents had a highly significant impact on the seed formation rate (Table 2).The crosses which took "EBY" or "BJY" as parents expressed a low seed formation rate.The highest seed formation rate was the inter-specific cross "XY-3 × XQ-4", reaching 79.26% while the intersectional cross, "XY-3 × OH-1", only had a seed formation rate of 55.93%.According to the analysis of different cross combinations, the ones taking P. simonii cultivars "XY-2", "XY-3" or "XY-4" as the female parent showed higher seed formation rate.
Forests 2018, 9, x FOR PEER REVIEW 7 of 15 combinations which took hybrids "ZY3" or "EBY" as the female parent and P. deltoides as the male parent with a serious problem of falling catkins and fruits.Additionally, this situation also occurred in the crosses taking P. simonii as the female parent and cultivars "137", "144", or hybrid "BJY" as the male parent.ANOVA indicated that the parents had a highly significant impact on the seed formation rate (Table 2).The crosses which took "EBY" or "BJY" as parents expressed a low seed formation rate.The highest seed formation rate was the inter-specific cross "XY-3 × XQ-4", reaching 79.26% while the intersectional cross, "XY-3 × OH-1", only had a seed formation rate of 55.93%.

Variation in Seed Traits
There was a very significant difference in the seed length and width for different cross combinations according to the variance analysis (p = 0.0001 both for seed length and width).Female parents had a highly significant effect on seed length, while there was no significant difference for male parents (Table 2).However, for seed width, both parents had a highly significant effect.The seed width of the combinations which took "EBY" or "ZY3" as the female was the largest (p = 0.05).No significant difference was detected for it, contrasting to the combinations taking "XY-4" as the female parents.Nonetheless, it was significantly bigger than the combinations taking the other tree species as the female parent.For the effect of the male parents, crosses that took "DZY" or "BJLY3"  ** Significant at 0.01% levels; ns, non-significant result; "-" no data.

Variation in Seed Traits
There was a very significant difference in the seed length and width for different cross combinations according to the variance analysis (p = 0.0001 both for seed length and width).Female parents had a highly significant effect on seed length, while there was no significant difference for male parents (Table 2).However, for seed width, both parents had a highly significant effect.The seed width of the combinations which took "EBY" or "ZY3" as the female was the largest (p = 0.05).No significant difference was detected for it, contrasting to the combinations taking "XY-4" as the female parents.Nonetheless, it was significantly bigger than the combinations taking the other tree species as the female parent.For the effect of the male parents, crosses that took "DZY" or "BJLY3" had the largest seed width (p = 0.05).The combination "ZY3 × BJLY3" represented the longest seed length and width, while the shortest seed length occurred in the cross "XY-1 × OH-2", and the shortest seed width in cross "XY-1 × 137".There was a highly positive significant correlation between seed length and seed width with different combinations according to the correlation analysis (r = 0.4892, n = 99, p < 0.0001).
TKW ranged from 0.1550 to 0.8900 g among the crosses.There was no significant difference for the crosses taking "XY" or "XQ" as the female parent (p = 0.1045 for "XY" and p = 0.3438 for "XQ", respectively).The TKW ranged from 0.1550 to 0.7155 g when the female was P. simonii, especially the combination "XY-4 × XY-5" having the largest TKW.Similarly, when the female was P. pseudo-simonii, it ranged from 0.1515 to 0.6433 g, particularly the combination "XQ-3 × DZY" which had the largest TKW.Furthermore, there was a highly positive significant correlation between TKW and the seed length (r = 0.4325, n = 88, p < 0.0001) as well as the seed width (r = 0.7893, n = 88, p < 0.0001).

Seed Germination Characteristics
The germination rate of hybrid seeds which were obtained from the section Tacamahaca crossing with P. deltoides was very low, even as low as zero.A non-significant difference was present in the seed germination rate for female parents while the male parents had a very significant effect on it (Table 2).The seed germination was generally good for the combinations between section Tacamahaca along with its hybrids and P. nigra but presented differences.The seed germination rate was above 45% when the hybrids of section Tacamahac and P. pseudo-simonii crossed with "BJLY3" and "DZY"; while, the seed germination rate was below 25% when the female was instead P. simonii.The cross combinations among section Tacamahaca had a high seed germination rate, and the crosses which took P. pseudo-simonii as the female parent had a better seed germination situation compared with P. simonii as the female parent.

Variations in Seedling Cotyledon Length and Morphology
The parents' combination had a highly significant effect on the seed cotyledon length (Table 2).It was not significant on the seed cotyledon length for the combinations which took P. simonii or Forests 2018, 9, 516 9 of 15 P. pseudo-simonii as the female parent (p = 0.0770 for the former and p = 0.3836 for the latter).The seed cotyledon length of the crosses which took "ZY3" as the female parent was longer than the others and there was a tendency that the crosses taking hybrids as parents had a longer cotyledon length.
Seedlings without a visible root organ and hypocotyl were also observed.When the radicle of the normal seedlings was elongated (Figure 7a), in the others it had not yet been seen (14.24%) (Figure 7b).Moreover, a few seedlings only presented with two cotyledons without a hypocotyl (Figure 7c).There was no regular segregation ratio of non-rooting to normal rooting seedlings.The seedlings without a hypocotyl accounted for a percentage of 1.13%.When it came to the individual cross, the results were rule less.ANOVA indicated that there was no significance on the non-radicle rate of seedlings for female parents (Table 2), whereas it was highly significant for the male parents (Table 2).The non-hypocotyl rate of crossbred seedlings was generally low.It was relatively higher for the combinations which took the hybrids as one of the parents; such as the crosses "EBY × OH-2" and "XQ-3 × BJLY3" (11.43% for the former and 9.34% for the latter).Furthermore, during the germination stages, the color of the hypocotyl was not the same for different crosses.Parts of them were peak green (Figure 7b) and the others were red (Figure 7a,d-g).However, the peak green would turn red due to illumination during later growth.

Variations in Seedling Cotyledon Length and Morphology
The parents' combination had a highly significant effect on the seed cotyledon length (Table 2).It was not significant on the seed cotyledon length for the combinations which took P. simonii or P. pseudo-simonii as the female parent (p = 0.0770 for the former and p = 0.3836 for the latter).The seed cotyledon length of the crosses which took "ZY3" as the female parent was longer than the others and there was a tendency that the crosses taking hybrids as parents had a longer cotyledon length.
Seedlings without a visible root organ and hypocotyl were also observed.When the radicle of the normal seedlings was elongated (Figure 7a), in the others it had not yet been seen (14.24%) (Figure 7b).Moreover, a few seedlings only presented with two cotyledons without a hypocotyl (Figure 7c).There was no regular segregation ratio of non-rooting to normal rooting seedlings.The seedlings without a hypocotyl accounted for a percentage of 1.13%.When it came to the individual cross, the results were rule less.ANOVA indicated that there was no significance on the non-radicle rate of seedlings for female parents (Table 2), whereas it was highly significant for the male parents (Table 2).The non-hypocotyl rate of crossbred seedlings was generally low.It was relatively higher for the combinations which took the hybrids as one of the parents; such as the crosses "EBY × OH-2" and "XQ-3 × BJLY3" (11.43% for the former and 9.34% for the latter).Furthermore, during the germination stages, the color of the hypocotyl was not the same for different crosses.Parts of them were peak green (Figure 7b) and the others were red (Figure 7a,d-g).However, the peak green would turn red due to illumination during later growth.(a) Seedling of hybrid between P. simonii "XY-4" and P. nigra "OH-1"; note: two opposite and approximate cotyledons; (b) seedling of hybrid between P. simonii "XY-1" and P. nigra var.italica "ZTY"; note: no radical and peak green hypocotyls; (c) seedling of hybrid between P. nigra var.thevestina × P. simonii "EBY" and P. nigra "OH-1"; note: no hypocotyls; (d) seedling of hybrid between P. pseudo-simonii "XQ-3" and "XQ-4"; note: fused cotyledons; (e) seedling of hybrid between P. pseudo-simonii × P. nigra var.lica "ZY3" and P. nigra "OH-1"; note: two cotyledons with one cracking into two parts; (f) seedling of hybrid between P. simonii "XY-4" and P. nigra "OH-1"; note: three cotyledons; and (g) seedling of hybrid between P. pseudo-simonii "XQ-2" and "XQ-4"; note: four cotyledons.(a) Seedling of hybrid between P. simonii "XY-4" and P. nigra "OH-1"; note: two opposite and approximate cotyledons; (b) seedling of hybrid between P. simonii "XY-1" and P. nigra var.italica "ZTY"; note: no radical and peak green hypocotyls; (c) seedling of hybrid between P. nigra var.thevestina × P. simonii "EBY" and P. nigra "OH-1"; note: no hypocotyls; (d) seedling of hybrid between P. pseudo-simonii "XQ-3" and "XQ-4"; note: fused cotyledons; (e) seedling of hybrid between P. pseudo-simonii × P. nigra var.lica "ZY3" and P. nigra "OH-1"; note: two cotyledons with one cracking into two parts; (f) seedling of hybrid between P. simonii "XY-4" and P. nigra "OH-1"; note: three cotyledons; and (g) seedling of hybrid between P. pseudo-simonii "XQ-2" and "XQ-4"; note: four cotyledons.
Seedlings with normal cotyledons generally had two opposite and approximate ones in dicot species (Figure 7a).However, there were various other types in our study, such as the single and fused cotyledons (Figure 7d), two cotyledons with one cotyledon cracking into two parts (Figure 7e), three cotyledons (Figure 7f), and four cotyledons (Figure 7g).Among them, the three cotyledons type appeared frequently.The four cotyledons type only appeared in the combinations taking P. pseudo-simonii as the female parent and occupied quite a low ratio (0.01%).The parents were not significant on the seedling cotyledons variation (p = 0.1212 for female and p = 0.4244 for male).The seedling cotyledons variation ratio was higher for the combinations which took P. pseudo-simonii as the female parent rather than the P. simonii ones.Between them were the combinations which took the hybrids as the female parents.The combinations taking "BJLY3" as the male parent had a high cotyledon variation rate.The crosses "XQ-2 × XMH", "XY-4 × OH-2", "XY-4 × XY-5", and so forth had a higher cotyledon variation than others (2.55%, 2.54%, and 2.31%, respectively).

Discussion and Conclusions
Flowering asynchronization can lead to a barrier to hybridization, while flowering synchronization is a stronger selector in dioecious plants [28][29][30].Hence, mastering the flowering phenology plays a guiding role in the breeding program.Flowering phenology also strongly affects reproductive success [31].Changes in the flowering time and flowering duration may modify the overall ontogeny and the architecture of plant individuals [32].In our study, there was a significant difference in the flowering time for different kinds of poplars in the greenhouse.The results revealed that it took 44 days from water planting to powdering and 22 days for the powdering period for male P. deltoides, while Zhao et al. [33] showed that P. deltoides took five days from water planting to powdering, and three days for the powdering period in a greenhouse.Broeck et al. [29] also indicated that the powdering period was three and 14 days, respectively, in 1999 and 2000.The flowering phase was connected with the environment, genes, and other endogenous factors [29,31,[34][35][36].Due to diversified impact factors, it should be a reasonable arrangement of the hydroponic time to make sure the female and male flowering phase encounter each other when carrying out cross breeding work between section Tacamahaca and Aigeiros.
Seed size is an essential functional character in flowering plants [37][38][39], influenced by environmental factors and genetic traits [40].In our work, we found that the influence of female parents for seed length was significant, while for seed width, both females and males had a significant influence on it.This suggests that the seed length is mostly controlled by females in poplars.The TKW ranged from 0.3333 to 0.7155 g for section Tacamahaca in our research.However, species in section Populus ranged from 0.1250 to 0.1990 g [41,42] and in North America, Aigeiros ranged from 2.2222 to 3.3333 g [41].These data showed that there is a wide variation in the TKW for Populus L. Zhang et al. [43] pointed out that several kernel traits, including kernel length and kernel width, controlled the TKW in wheat.Similarly, there was a highly positive significant correlation between TKW with the seed length and seed width in our study.According to the correlation analyses results, the seed width was more relevant to the TKW than the seed length.
Root apical meristems are at the bottom of the hypocotyl and can lead to the formation of a radical or embryonic root [44].The normal development of the hypocotyl and radicle is essential to the seedling's growth.During embryogenesis, the embryonic root and its shoot apical meristems are established in angiosperms [28,[45][46][47] and they initiate post-embryonic development by producing a visible root organ.Zhang et al. [48] discovered that the separation ratio of the non-rooting to normal rooting was 1:3 in F1 offspring for P. tomentosa.However, we did not find the same segregation ratio in our study.Zhang et al. [48] indicated that it was a qualitative trait for the radical development character and a non-rooting trait due to the allelic mutation in P. tomentosa.However, some Arabidopsis embryo defective mutants lack a primary root but are capable of forming roots (hobbit and bodenlos mutants) and hypocotyl (monopteros and fackel mutants) in the post-embryonic stage, which changes the function of some housekeeping genes, the metabolic pathway, secretory pathway, or homeldomain transcription factor [49][50][51][52].These mutant genes have an effect on the formation of the apical-basal polarity pattern for zygotic embryo [53].Apical-basal pattern controls important elements of the seedling body, such as the shoot meristem, root meristem, embryonic root, hypocotyl, and cotyledons [44].It is clear to see that the regulation of genes has a huge effect on the growth of the plant.The traits of the seedling radicle and hypocotyl constitute a reference for the reselection of the parents.
Cotyledons provide nutrition and are able to carry out photosynthesis for some young plants in angiosperm [54].Generally, there are two bilaterally symmetric and almost equally sized cotyledons for most dicot species.However, research has revealed that there are numerous cotyledon mutants in many plant species.The partial or complete fusion of the cotyledons was presented in species of Calophyllum, Swietenia, and Guarea [55].Karschon [56] reported that the seedlings of the Chenopodiaceae occurred schizocotyly in order to give hemitrocotyly, where one of two cotyledons might be a cleft.Moreover, some species of Acer, Coffea, Raphanus, and Sesamum have three cotyledons [57][58][59][60].Additionally, there are four cotyledons in some Ranales and Persoonia [55,61].Similar morphological variations of cotyledons were shown in our study and the seedlings with three cotyledons occurred at a high frequency among the variable seedlings.It was the same for the results of triploid P. tomentosa [62], whereas, when it comes to Arabidopsis, there were more morphological variations of the cotyledons, such as cup-shaped cotyledons, malformed cotyledons, and cotyledon absence [63][64][65].Kang et al. [62] pointed out that the equilibrium of cell chromosomes might be related to these variations in triploid P. tomentosa, as genes were the main factor of controlling cotyledons variations in Arabidopsis [54,65].Atthesame time, auxin, which is essential for organ formation, may also be the ultimate reason.The variation of cotyledons may have an influence on the nutrient uptake and preliminary photosynthesis in epigeal seedlings in which the cotyledons appear above ground.As a consequence, it finally affects the growth of seedlings.
Intersectional hybridization plays an important role in the genetic improvement of the species of the genus Populus and wide hybridization easily contributes to hybrid incompatibility and hybrid sterility, but it is a way to produce transgressive hybrids.Cross-incompatibility usually occurs in wide hybridization and it generally takes place as the dropping of flowers and fruits, low seed set, shriveled seeds, and low seed formation percentage [66][67][68].In our study, the combinations which had P. deltoides as the male parent had a low seed formation percentage and seed set, shriveled seeds, and low seed germination rate, some having no collected seed.Similarly, the crosses involving male P. deltoides and female P. balsamifera produced few seeds while these seeds revealed a low viability [67].Meanwhile, Mahama et al. [68] pointed out that the crosses involving male P. deltoides and female P. maximowiczii were less successful and rarely acquired seedlings.However, when P. deltoides was taken as the female and some species of section Tacamahaca were taken as male, the crosses showed a high success rate, obtaining many seedlings [33,66,67].The results in our research suggested that cross-incompatibility was occurring.Many stages in the hybridization process, including the germination of the pollen grains and the failure of the pollen tube penetrating the stigma and ovary as well as the failure of seed development and hybrid sterility, may contribute to cross-incompatibility [68,69].These stages could be classified into prezygotic and postzygotic stages [70,71].The prezygotic stage is closely influenced by pollen tube behavior [68,69,72] and, if the cross-barrier occurred in the postzygotic stage, the embryo rescue and tissue culture technique could overcome it [73][74][75].In our study, numerous germinated pollen grains were clearly observed and penetrated into the stigma according to aniline blue stain observation in crosses taking P. deltoides as the pollen donor, and a few shriveled seeds were collected.Therefore, the reason for the incompatibility most probably occurred in the postzygotic stage.Further work is still needed to elucidate what factors led to the incompatibility.
Hybridization breeding plays an important role in poplar breeding engineering and lays the foundation for polyploidy breeding.The parental selection is essential for hybridization.At the same time, it is beneficial to improve the effectiveness of polyploidy breeding in the genus Populus, according to the experience of parent selection in hybridization breeding [76].Polyploidization is essential in Forests 2018, 9, 516 12 of 15 speciation, evolution, and creating elite varieties for poplar plantations [77][78][79].Moreover, triploid Populus is more widely used for the production of lumber, fiber, and fuel, in contrast to diploids [80,81].At the same time, it is also an essential part of the control of desertification.Since the application of polyploid Populus is important, it is necessary to conduct a study of polyploid induction.Therefore, we will carry out research of general and special combing ability in the next study.

ForestsFigure 1 .
Figure 1.Distribution map of "Three-North" region and location of flower branches collection.Figure 1. Distribution map of "Three-North" region and location of flower branches collection.

Figure 1 .
Figure 1.Distribution map of "Three-North" region and location of flower branches collection.Figure 1. Distribution map of "Three-North" region and location of flower branches collection.

DaysFigure 2 .
Figure 2.The flowering phenology on days before flowering and for duration of pollen shedding for the 13 male parents under the condition of water culturing in a greenhouse (mean + s.e.).The bars followed by the same letter are not statistically different (p < 0.05, n = 3).Duncan's multiple range test was taken.

Figure 3 .
Figure 3.The phenology of the days before pollination and seed maturation for the 9 female parents under the condition of water culturing in a greenhouse (mean + s.e.).The bars followed by the same letter are not statistically different (p < 0.05, n = 13).Duncan's multiple range test was taken.

Table 1 .
The original source of the materials used in the crosses, including the 9 female and 13 male cultivars belonging to section Tacamahaca and Aigeiros.

Table 2 .
The analysis of variance showing the mean squares of the agronomic related parameters for hybrids crossing between section Tacamahaca and Aigeiros.

Table 2 .
The analysis of variance showing the mean squares of the agronomic related parameters for hybrids crossing between section Tacamahaca and Aigeiros.