Germplasm Innovation and Establishment of Comprehensive Evaluation System for Hedgerow Garden Chrysanthemum

Abstract

Garden chrysanthemums are a group of chrysanthemums (Chrysanthemum morifolium) that are mostly used in garden landscape collocations. Because most garden chrysanthemum varieties have low plants and poor space effects, they have certain limitations in garden landscape applications. In this study, we selected seven garden chrysanthemum varieties with excellent characteristics for artificial hybridization and intercross parentage to obtain new varieties with good traits we need. The phenotypic characteristics of the F1 offspring in terms of plant height, crown width, crown height ratio, number of main branches, branching intensity, plant form, inflorescence diameter, number of flowers per plant, number of ray florets, flower color, and days from planting period to coloring period were analyzed and systematically evaluated. Subsequently, a comprehensive evaluation system was established using the analytic hierarchy process (AHP) and k-means clustering method. We built a comprehensive analysis model, calculated the weighting of each evaluation factor, and multiplied the weight value by the score of the evaluation index standard of each factor to obtain the comprehensive score of each plant. All F1 plants were divided into four grades: excellent grade accounting for 13%, good grade accounting for 28%, medium grade accounting for 36%, and poor grade accounting for 23%. Then, we analyzed the differences of some quantitative traits between Group E (hybrids in excellent grade) and Group O (hybrids in good, medium, and poor grades). There were significant differences in plant height, crown height ratio, and the number of main branches but no significant difference in crown width. Combining with comprehensive score showed that Group E performed well overall. Finally, we selected five hybrid offspring with the highest overall scores in Group E as excellent variety materials of garden chrysanthemum for hedgerow. They were CH22, YQ73, HY08, CQ80, and HY07, respectively. We also found that plant height, lodging resistance, crown height ratio, plant form, crown width, and the number of main branches could be the main indicators in the AHP, which can be effectively applied to the comprehensive evaluation and breeding of garden chrysanthemums for hedgerows.

1. Introduction

Chrysanthemum (Chrysanthemum morifolium Ramat.), belonging to the Asteraceae family, is a kind of annual or perennial flowering herb that originated in China. Its cultivation history can be traced back to the 15th century BC and it has high ornamental and medicinal value [1,2,3]. After thousands of years of artificial domestication, chrysanthemum is a kind of highly heterozygous artificial species, and the seed setting rate of hybrid combinations is usually hard to determine, so it is generally propagated by cutting. Chrysanthemum has formed a highly diversified group and plays an important role in flowers in China and even the world. As one of the most beautiful commercial flowers in the world, chrysanthemum ranks second in the international cut flower trade [4]. Modern cultivated chrysanthemum is mainly hexaploid, and many new varieties have been produced through hybridization and artificial selection [5,6]. Garden chrysanthemum is an important cultivation and application type of chrysanthemum, and it is one of the most important decorative cultures in the world’s floriculture [7]. Its crown width is larger than that of traditional chrysanthemum, and its plant shape is round, with a consistent flowering period, bright colors, few pests and diseases, low maintenance cost, and strong adaptability to various environments [8]. It is widely used in courtyard decoration, garden decoration, and green belt, thus becoming a new category of urban beautification [9,10].

Garden chrysanthemum breeding is mainly aimed at creeping chrysanthemum breeding research to cover the ground to achieve the effect of paving. To date, chrysanthemum breeders at home and abroad have achieved remarkable results in the improvement of ornamental traits such as plant type and flower color through conventional cross-breeding methods using abundant germplasm resources [11]. For example, with ‘Meiaifen’ as the female parent and Chrysanthemum vestitum and Chrysanthemum zawadskii as the male parent, groundcover chrysanthemum varieties with low plants and lush flowers and leaves were selected [12]. A number of new groundcover chrysanthemum varieties with fully prostrate plant form and vigorous growth have been bred by mutagenizing chrysanthemum seeds [13]. Many dwarf transgenic materials have also been obtained through genetic engineering breeding, and some foreign scholars have cloned some Arabidopsis mutants. A few dwarf materials were obtained by transferring the rolC gene into chrysanthemum. The rolC gene was transferred into chrysanthemum to obtain a new material with dwarf plant type and compact branches and wider petals. The gai gene was transferred into chrysanthemum to obtain significantly dwarfed plants, and transgenic plants were found insensitive to exogenously applied GA in traits such as plant height and internode length; the plant growth was inhibited after the phytochrome phy-B1 gene was transferred into chrysanthemum [14,15,16].

As most garden chrysanthemums have low plant types and strong creeping, which limit the application of certain garden landscapes such as hedge applications, we used artificial hybridization—the simplest and most effective method for selecting new chrysanthemum varieties to improve the ornamental characteristics of garden chrysanthemum and enrich its application range [17,18]. Some of the most important goals of this study were to increase plant height, branching intensity, and improve plant form, crown width, etc. In this study, varieties with high and hard branches were selected as parents to obtain new offspring, and a statistical analysis of the genetic diversity of crosses with dominant landscape features was performed on all F1 generations of individual plants. Finally, we established a comprehensive evaluation system with the growth characteristics of garden chrysanthemums as the main evaluation index and selected five hybrid offspring with prospective traits that provided a reference for the evaluation and variety selection of garden chrysanthemums for hedgerows. The evaluation system can be used or learned for garden chrysanthemum breeding in the future.

2. Materials and Methods

2.1. Plant Materials

The seven garden chrysanthemum varieties including ‘Zhongshan Yucheng’, ‘Zhongshan Yanghong’, ‘Zhongshan Yanhong’, ‘Orange-red Autumn Xiaoju’, ‘Qixia Scarlet’, ‘Qiu Fen’, and ‘Color Volcano’ were all provided by the Chrysanthemum Germplasm Resource Preserving Center of Nanjing Agricultural University, China. They all had relatively good performance of target traits in the field and had been selected by experts as materials for hybridization.

Based on the characteristics of the parents, the possibility of practice, and the hybridization conditions such as different flowering stages, we finally determined the 12 hybrid combinations listed in

Table 1. Match of combination.

Combination	Female Parent	Male Parent	Combination	Female Parent	Male Parent
1	‘Zhongshan Yucheng’	‘Zhongshan Yanhong’	7	‘Orange-red Autumn Xiaoju’	‘Qiu Fen’
2	‘Zhongshan Yucheng’	‘Zhongshan Yanghong’	8	‘Orange-red Autumn Xiaoju’	‘Zhongshan Yanghong’
3	‘Zhongshan Yucheng’	‘Orange-red Autumn Xiaoju’	9	‘Orange-red Autumn Xiaoju’	‘Color Volcano’
4	‘Zhongshan Yanghong’	‘Orange-red Autumn Xiaoju’	10	‘Qixia Scarlet’	‘Zhongshan Yanghong’
5	‘Zhongshan Yanghong’	‘Qixia Scarlet’	11	‘Qixia Scarlet’	‘Qiu Fen’
6	‘Zhongshan Yanghong’	‘Qiu Fen’	12	‘Zhongshan Yanhong’	‘Zhongshan Yanghong’

. The artificial hybridization method has been described by Wang et al. with modifications [19]. A total of 1166 inflorescences were hybridized, and after seed collection, round and full seeds amounting to 2454 were selected for sowing (

Table 2. Statistics of hybridization and F1 generation sowing.

Combination	Number of Pollination Inflorescences	Number of Seeds	Number of Seeds Sown	Number of Field Planting	Number of Survivors
CH	100	1585	200	81	80
CY	75	2314	200	82	80
CQ	125	1500	200	80	78
YQ	101	2100	200	80	76
YH	100	1300	200	80	78
YF	100	1164	200	81	76
QF	100	727	200	81	80
QY	110	144	144	39	33
QS	50	310	310	68	44
HY	75	1287	200	80	76
HF	100	1185	200	82	78
YY	130	450	200	87	81

). The 860 plants from the 12 combinations were planted with a row spacing of 45 cm × 45 cm under conventional field management. The first pinching was performed on the 20th day after planting [20], and the second was performed according to the growing conditions in the later period.

2.2. Phenotypic Trait Measurement and Statistics

The traits of hybrid offspring were observed and recorded during the flowering stage. Phenotypic traits recorded were divided into quantitative and qualitative traits. There were nine quantitative traits: plant height, crown width, crown height ratio, number of main branches, days from planting period to coloring period, number of flowers per plant, number of ray florets, inflorescence diameter, and days of florescence. Excel software was used to calculate the mean value, maximum value, minimum value, frequency, and coefficient of variation of quantitative characteristics in the F1 generation for each combination. Qualitative traits included plant form, branching intensity, and flower color. Plant form can be divided into four grades: lodging, loose, compact, and round. Branching intensity was divided into three grades: strong, medium, and weak. Color statistics for yellow, red, white, pink, purple, pink-purple, orange, orange-yellow, orange-red, etc. were recorded. Plant height and crown width were measured using a ruler, and inflorescence diameter was measured using a vernier caliper.

The measurement standards and methods are as follows: 1. Plant height: height of the aboveground part of the plant, measured per plant with a ruler; 2. Crown width: the average of two directions perpendicular to each other in the corolla, measured per plant using a ruler; 3. Crown height ratio: ratio of plant crown width to plant height; 4. Number of main branches: the number of branches on the main stem of the plant counted visually; 5. Branching intensity: the strength and toughness of the main branches of chrysanthemum at the blooming stage, divided into three grades; 6. Plant form: contour, roundness, and compactness of the crown plexus at the blooming stage, divided into four grades; 7. Days from planting period to coloring period: number of days from colonization to color development of chrysanthemum; 8. Number of flowers per plant: number of inflorescences per plant counted visually; 9. Flower color: ligulate flower front color (yellow, red, white, pink, purple, pink-purple, orange, orange-yellow, and orange-red), counted visually; 10. Number of ray florets: visually count the number of ray flowers, measure six inflorescences, and take the average; 11. Inflorescence diameter: three inflorescences were measured and averaged; 12. Coloring period: the date when a single flower bud opens and the color of the petals can be clearly distinguished and counted visually; 13. Blooming period: the date when the inflorescences of the whole plant were 70% fully open, counted visually; 14. Last flowering period: the date when 50% of the inflorescences of the whole plant completely withered, counted visually; 15. Days of florescence: number of days from coloring to the last flowering period.

2.3. Establishment of Comprehensive Evaluation Criteria for Hybrid Offspring of Hedgerow Garden Chrysanthemum

With reference to the evaluation indices of different chrysanthemum varieties and other flowers, combined with expert opinions, a 3-point scoring standard for each factor was formulated according to the actual situation (

Table 3. Appraisal criterion of all factors.

Factor	Score
	3	2	1
Plant height (cm)	>50	35–50	<35
Crown width (cm)	>75	39–75	<39
Crown height ratio	>1.56	1.23–1.56	<1.23
Number of main branches	>10	5–10	5
Days from planting period to coloring period	138–145	<138	>145
Number of flowers per plant	>1000	500–1000	<500
Flower color	Pure color and high chroma	High color heterochromaticity or low color pure chromaticity	Mixed color and low chroma
Number of ray florets	>150	100–150	<100
Inflorescence diameter	>5	3–5	<3
Days of florescence	>33	25–33	<25
Lodging resistance	Stiff branches and high resilience	Branches are relatively stiff and not easy to lodging	Branches are soft and prone to lodging
Plant form	Compact and rounded	Relatively compact rounding	Crown plexus is not full and has different shapes

). The AHP was used to evaluate scores and the grades of hybrid offspring plants were classified using IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp. Combining with cluster analysis, each factor was divided into three grades, with scores of 3, 2, and 1. Conduct comprehensive evaluation and analysis according to various evaluation indicators.

3. Results

In order to breed for target traits, we hybridized chrysanthemum varieties that express these traits and evaluated the hybrid offspring for improvements. We here present the most striking outcomes, whereas a complete report of the evaluated traits is shown in the Supplementary Material.

3.1. Genetic Statistical Analysis of F1 Generation Traits

3.1.1. Genetic Statistical Analysis of Quantitative Traits in F1 Generation

Genetic Statistical Analysis of Plant Height

A correlation analysis of nine quantitative characters in the F1 generation of the 12 cross combinations was carried out. Plant height was significantly positively correlated with crown width, number of main branches, number of ligulate flowers, number of whole plant flowers, etc. There was a significant negative correlation between plant height and crown height ratio and the number of days from planting period to coloring period (Table S1). The average plant height of hybrid offspring was 80.2% of the median parent value. Among the three groups of hybrids with ‘Zhongshan Yucheng’ as the female parent, the proportion of superhigh parent was 12.5%, and the proportion of ultralow parent was 80.9%; among the three groups of crosses with “Zhong Shan Yang Hong” as the female parent, 24.6 % was the proportion of the superhigh parent, and 67.9% was the proportion of the of ultralow parent. Among the three groups of hybrid offspring with ‘Orange-red Autumn Xiaoju’ as the female parent, 13.7% of the total number were of the superhigh parent, and 74.0% of the total number were of the ultralow parent. The maximum value of each combination was 3 to 8 times the minimum value, and the maximum value of QS was the largest; the minimum value of YQ was the smallest. The mean variation coefficient of plant height was 0.29, indicating a high genetic diversity. The hybrids of HY had the largest mean in plant height, it was concluded that hybrid offspring of HY are potential candidates to increase the height of garden chrysanthemum (Table S2).

Genetic Statistical Analysis of Crown Width

The average crown width of hybrid offspring was 97.0% of the median parent value. Among the three groups of hybrids of ‘Zhongshan Yucheng’ as the female parent, 36.5% of the total number was of the superhigh parent, and 31.7% of the total number were of the ultralow parent. Among the three groups of crosses with ‘‘Zhongshan Yanghong’’ as the female parent, 30.3% of the total number were of the superhigh parent, and 56.1% of the total number were of the ultralow parent. Among the three groups of hybrid offspring of ‘Orange-red Autumn Xiaoju’ as the female parent, 66.5% were of the total number of superhigh parent individuals, and 20.4% of the total number of ultralow parent individuals. The crown width of different hybrid combinations was widely separated, and the maximum of each combination was 3.0 to 7.6 times of the minimum, the difference of YF is the largest, and that the CY is the smallest. The average coefficient of variation of the crown width index of hybrid offspring was 0.30, and the hybrid offspring had high genetic diversity. The hybrid offspring of QY had the largest mean in crown width, which could be used to improve the crown width of garden chrysanthemum (Table S3).

Genetic Statistical Analysis of the Number of Main Branches

In Table S4, the average number of main branches of hybrid offspring was 52% of the median parent value. Among the three groups of hybrid offspring of ‘Zhongshan Yucheng’ as the female parent, 42.6% of the total number were superhigh parent individuals, and 18.3% of the total number of ultralow parent individuals. Among the three groups of hybrid offspring with ‘Zhongshan Yanghong’ as the female parent, 25.8% of the total number were of superhigh parent, and 36.6% of the total number of the ultralow parent. Among the three groups of hybrid offspring of ‘Orange-red Autumn Xiaoju’ as the female parent, 37.7% were superhigh parent, and 21.0% were ultralow parent. The maximum value of the number of main branches of each combination was 2.5–10 times the minimum value, among which the difference of combination of CH was the largest, and the combination of QY was the smallest. The average coefficient of variation of the number of main branches of hybrid offspring was 0.52, indicating high genetic diversity of the hybrid offspring. Generally, the hybrid offspring of CQ is a powerful candidate to increase the number of main branches of garden chrysanthemum with their large mean in the number of main branches (Table S4).

3.1.2. Genetic Statistical Analysis of Quality Traits in F1 Generation

Genetic Statistical Analysis of Plant Form

At the flowering stage, the hybrid plants were classified into four types—Grade I: rounder plant form, prosperous flowers, and lush branches; Grade II: round plant form and many flowers and branches; Grade III: loose plant form, sparse flower branches; and Grade IV: plant lodging, sparse flowers, and branches.

There was a big difference in the plant form among the hybrids in different combinations (Table S8). Among the hybrids of the three hybrid combinations with ‘Zhongshan Yucheng’, ‘Zhongshan Yanghong’, and ‘Orange-red Autumn Xiaoju’ as the female parent, 16.7%, 18.8%, and 8.1% of the individual plant forms were Grade I, respectively; 41.7%, 13.7%, and 30.1% were Grade II; 34.2%, 51.3%, and 44.6% were Grade III; and finally, the individuals with the plant form of class IV accounted for 7.5%, 16.2%, and 17.3%, respectively. In all hybrid combinations, the offspring of YH had the largest proportion at Grade I, which is a potential candidate to improve the plant form of the garden chrysanthemum (Table S8).

Genetic Statistical Analysis of Branching Intensity

The branching intensity can be determined by artificial pressing and observing the degree of rebound. Offspring were divided into three intensity grades. Level 1: After pressing, the plant has high resilience, stiffening, and strong lodging resistance. Level 2: After pressing, the plant is slightly deformed, with average rebound strength, stiffening, and common lodging resistance. Level 3: The plant fell after pressing, with poor lodging resistance. In all hybrid progeny populations, Level 1 accounted for 42.6%, Level 2 accounted for 39.1%, and Level 3 accounted for 18.3%. Judging from the branching intensity performance of the hybrid progeny of each combination, Levels 1 and 2 accounted for more. The branching intensity of progeny in YY was greatest in Level 1. The specific distribution is shown in Table S9.

3.2. Construction of an Evaluation System of Garden Chrysanthemum for Hedgerows

3.2.1. Constructing the Hierarchical Structure Model

Referring to the construction standards of the analytic hierarchy process and previous chrysanthemum evaluation system [20], the weighting of each trait index was determined, and a three-level comprehensive analysis model was established based on the actual traits of the hybrid offspring of garden chrysanthemums. The first layer was the target layer (A), which was the F1 generation of garden chrysanthemum with good comprehensive characteristics. The second layer is the constraint layer (C), which represents the growth characteristics, ornamental characteristics, and adaptability of plants. The third layer was the standard layer (P), which contained 12 specific trait evaluation indices (

Table 4. The hierarchy model of evaluation system.

Target Layer (A)	Constraint Layer (C)	Standard Layer (P)
F1 generation single plant with excellent comprehensive characters of garden chrysanthemum for hedges (A)	Growth characteristics (C1)	Plant height (P1)
		Crown width (P2)
		Crown height ratio (P3)
		Number of main branches (P4)
	Ornamental characteristics (C2)	Days from planting period to coloring period (P5)
		Flower number per plant (P6)
		Flower color (P7)
		Number of ray florets (P8)
		Inflorescence diameter (P9)
		Days of florescence (P10)
	Adaptability (C3)	Lodging resistance abilities (P11)
		Plant form (P12)

).

3.2.2. Construction of a Comparative Judgment Matrix

The importance of each evaluation factor is relative, and the importance of the factor is measured by comparing the other two factors. According to each influence factor’s contribution to the integrated properties evaluation of hybrid offspring and the relative importance degree of each factor, integrating the suggestions of several chrysanthemum breeding experts, and referring to the ratio of the 1–9 scaling method (

Table 5. The implication of ratio scale of 1–9.

Scale	Implication
1	Compared to the two elements, they have the same importance
3	Compared to the two elements, the former is slightly important than the latter
5	Compared to the two elements, the former is more important than the latter obviously
7	Compared to the two elements, the former is more important than the latter strongly
9	Compared to the two elements, the former is extremely important than the latter
2, 4, 6, 8	The median of above judgment
1/bij	bij represents the judgment value of comparison between index i and j, then 1/bij represents the judgment value of comparison between index j and i

) to determine the importance degree of each influencing factor to build A-C, C-P four matrices (

Table 6. The results of matrix and its consistency test between Target layer(A) and Constraint layer (C₁–C₃).

A	C1	C2	C3	W
C1	1	4	3	0.608
C2	1/4	1	1/3	0.12
C3	1/3	3	1	0.272

λ max = 3.074, C.I. = 0.037, R.L. = 0.58, C.R. = 0.064 < 0.10.

,

Table 7. The results of matrix and its consistency test between Constraint layer (C₁). and Standard layer (P₁–P₄).

C1	P1	P2	P3	P4	W
P1	1	3	5	6	0.570
P2	1/3	1	1/2	2	0.147
P3	1/5	2	1	3	0.205
P4	1/6	1/2	1/3	1	0.077

λ max = 4.153, C.I. = 0.051, R.L. = 0.9, C.R. = 0.057 < 0.10

,

Table 8. The results of matrix and its consistency test between Constraint layer (C₂). and Standard layer (P₅–P₁₀).

C2	P5	P6	P7	P8	P9	P10	W
P5	1	1/6	1/3	1/3	1/2	1/4	0.051
P6	6	1	4	2	3	1/2	0.278
P7	3	1/4	1	1/2	2	1/3	0.113
P8	3	1/2	2	1	2	1/2	0.161
P9	2	1/3	1/2	1/2	1	1/3	0.085
P10	4	2	3	2	3	1	0.313

λ max = 6.212, C.I. = 0.042, R.L. = 1.24, C.R. = 0.034 < 0.10

and

Table 9. The results of matrix and its consistency test between Constraint layer (C₃). and Standard layer (P₁₁–P₁₂).

C3	P11	P12	W
P11	1	2	0.667
P12	1/2	1	0.333

λ max = 2, C.I. = 0, R.L. = 0, C.R. = 0 < 0.10

), and carry on the consistency check; the inspection results are listed below in the matrix.

3.2.3. Consistency Test of the Judgment Matrix

Based on the AHP and YAAHP software statistical calculation, the matrix of each influencing factor was processed, and the consistency test of the three constructed matrices was carried out. Taking the random consistency ratio of the judgment matrix C.R. (<0.10) as the standard, that is, the ratio of the general consistency index C.I. of the judgment matrix to the average random consistency index R.I., the R.I. values of orders 1–9 are shown in

Table 10. The values of R.I. in list of 1–9.

n	1	2	3	4	5	6	7	8	9
R.I.	0	0	0.58	0.90	1.12	1.24	1.32	1.41	1.45

. The calculation results show that when the above matrix C.R. is within the allowable error range (<0.10), it means that the judgment matrix has satisfactory consistency, indicating that the weighting distribution of each factor meets the requirements.

3.3. Comprehensive Evaluation of Quality Traits of Garden Chrysanthemum Used for Hedgerow

3.3.1. Comprehensive Ranking of Evaluation Indicators

The weight value of the constraint layer (C) to the target layer (A) and the weight value of the standard layer (P) to the constraint layer (C) are obtained by the analytic hierarchy process, and the comprehensive weight value of the standard layer (P) to the target layer (A) can be calculated using the weight value of the constraint layer (C). Excel 2010 software was used for statistical analysis and sorting. The relative importance of each evaluation factor can be obtained by calculating the weight value to screen the F1 generation of garden chrysanthemum for hedgerows with comprehensive and accurate characteristics. According to the weighting size of the individual plant of the F1 generation of garden chrysanthemum with excellent comprehensive characteristics, the influence effect was sorted into plant growth characteristics, adaptability, and ornamental characteristics from large to small, and the weight values were 0.608, 0.272, and 0.120, respectively, indicating that the application of AHP in the analysis of new garden chrysanthemums for hedgerow use, and the growth characteristics of the garden chrysanthemum as the most influential characteristics of comprehensive effect can be used as a reference. Among the 12 indicators selected, the top five were plant height (P1), lodging resistance (P11), crown height ratio (P3), plant form (P12), and crown width (P2), with weight values of 0.374, 0.181, 0.125, 0.091, and 0.089, respectively. The results showed that plant height should meet the standards of the hedgerow, and lodging resistance should be considered. The compactness and fullness of the plant form and the crown width are also key factors in improving the standard of the hedgerow. However, the weight values of the number of days from the planting stage to the color stage (P5), inflorescence diameter (P9), and flower color (P7) were 0.006, 0.01, and 0.014, respectively, revealing that these characteristics were not considered important selection conditions in the selection of garden chrysanthemums for the hedgerow.

3.3.2. Grade Determination of Hybrid Offspring and Selection of Excellent Individual Plants

The weight values of the 12 evaluation factors (

Table 11. The comprehensive weight value of each layer.

Standard	W(A-Ci)	W(C-Pi)	W(A-Pi)	Order
P1	0.608	0.570	0.347	1
P2		0.147	0.089	5
P3		0.205	0.125	3
P4		0.077	0.047	6
P5	0.120	0.051	0.006	12
P6		0.278	0.033	8
P7		0.113	0.014	10
P8		0.161	0.019	9
P9		0.085	0.010	11
P10		0.313	0.038	7
P11	0.272	0.667	0.181	2
P12		0.333	0.091	4

) were multiplied by the score values of the evaluation index standards of each factor (Table 3), and the comprehensive score of each individual plant was calculated by cumulative calculation. K-means clustering analysis was used to classify the 860 individual plants of the F1 generation into four grades: excellent, good, medium, and poor. In the F1, the proportion of excellent grade was the lowest (13%), good and medium grades synthesis accounted for 64%, and poor grade accounted for 23% of the total (

Table 12. Grading result of individual plants.

Grade	Center of Clustering	The Number of Individual Plants in F1	Proportion (%)
Excellent	2.247	115	13
Good	1.902	240	28
Medium	1.610	307	36
Poor	1.285	196	23

).

The comprehensive score of all single plants in the F1 generation was made into a scatter plot (Figure 1). Then, we integrated the offspring of the other three groups except excellent into Group O and analyzed the significant differences of quantitative traits that with high weighting between Group E (hybrids in excellent grade) and Group O (hybrids in good, medium, and poor grades). The results showed that there were significant differences in plant height, crown height ratio, and the number of main branches between Group E and Group O but no significant difference in crown width (

Table 13. Significant difference analysis data of four traits in the two groups.

Traits	Group	Sample Size	Mean ± SEM
Plant height	E	115	42.617 ± 0.624 *
	O	745	30.634 ± 0.314
Crown width	E	115	67.104 ± 1.213
	O	745	45.223 ± 0.529
Crown height ratio	E	115	1.590 ± 0.026 *
	O	745	1.541 ± 0.021
The number of main branches	E	115	6.800 ± 0.330 *
	O	745	4.789 ± 0.091

* It means that there are significant differences in this trait among the two group.

). In general, the offspring in Group E had significantly good traits and were representative. Therefore, it was reasonable that we selected the five offspring with the highest comprehensive score as the candidate materials of garden chrysanthemum for hedgerow in Group E. They were CQ80, HY08, CH22, YQ73, and HY07, which performed excellent overall, with higher plant height, round plant form, high lodging resistance, and huge crown width (Figure 1). Different individuals had notable differences. CQ80 had the largest crown height ratio; HY08 had the largest number of main branches; CH22 and YQ73 performed relatively well balanced; and HY07 had the biggest crown width but softer branches. The number of days from the planting period to the coloring period and flower color also showed a great difference, as the flowering period of YQ73 was shorter than that of HY07, and all five single plants had different flower colors (Figure 2A). In addition, other materials, excluding the chosen ones, also had some good characteristics (Figure 2B) and could be used as garden chrysanthemums for hedgerows.

4. Discussion

4.1. Genetic Analysis of Quantitative Traits in Hybrid Offspring

Garden chrysanthemum has high landscape value; however, most garden chrysanthemum plants are not tall and are mostly used as ground cover in landscape applications [21], which limits their landscape value. The hybrid progeny of the ‘Orange-red’ series and the ‘Qixia’ series as the female parents had more superhigh parent individuals in plant height than that in the other hybrid combinations, indicating that F1 generation plant height has the characteristics of partial maternal inheritance (Table S2).

For garden chrysanthemum varieties with strong ornamental characteristics, the crown width of the plant is closely related to the number of flowers and is an important growth indicator that affects the ornamental quality of the variety. The crown width of the hybrid offspring in this experiment tends to be greater than that of the parents, which also shows that the hybrid offspring inherits the characteristics of excellent traits (Table S3); The branching systems of plants are closely related to plant morphogenesis [22]. In this experiment, the number of main branches of the progeny of the 12 hybrid combinations of garden chrysanthemum was higher than that of the parents in different degrees compared with the values of the male and middle parents. The variation of F1 generation plants had a better performance in terms of the genetic characteristics of the number of main branches (Table S4). The inflorescence diameter of the hybrid progeny generally tended to increase compared to that of the male parent and the middle parent (Table S5).

The number of flowers is an important trait affecting the aesthetic value of ornamental plants [23]. Compared with the male parent and middle parent, the number of flowers in the whole plant of the progeny of the 12 hybrid combinations in this experiment showed a better multivariate variation, and an increasing trend (Table S7). In terms of the heredity of the number of ray florets, double-flowered varieties are favored by consumers and the market because of their long flowering period [24]. In this experiment, the number of ray florets in hybrid progenies did not increase apparently compared with the male parent and the median parent, but most of them still showed double petals (Table S6).

4.2. Genetic Analysis of Quality Traits of Hybrid Offspring

The individual ornamental effect and group landscape construction of garden chrysanthemums are closely related to plant form and flower color. Plant form is a significant characteristic of ornamental plants [22]. Chrysanthemum has different plant form requirements under different application scenarios [25]. Therefore, plant form is an important indicator for the selection of new varieties of garden chrysanthemums. The performance of the plant form of the hybrid offspring for each combination was quite different. ‘Orange-red Autumn Xiaoju’ is a taller plant with soft branches and loose plant form, and more than half of the hybrid offspring with it as the female parent are loose and not round. The offspring of the 6 hybrid combinations of the ‘Zhongshan’ series of cultivars as the female parent are mostly compact and round, and most of them inherit the excellent traits of the female parent’s compact and round plant form (Table S8).

Branching intensity is one of the most important agronomic traits affecting plant lodging resistance, plays a crucial role in plant quality and yield [26], and is the key to determining plant lodging resistance [27]. In this experiment, the branching intensity of the progeny of the 12 hybrid combinations per plant was higher than that of the parent and middle parent, and they all showed good genetic characteristics (Table S9).

Flower color is one of the most important characteristics of ornamental plants [28], and it is also one of the most important ornamental traits of chrysanthemums. In this experiment, the flower colors of the female parents were mostly yellow and red. Although the hybrid progeny showed great diversity, the proportion of orange-red and yellow was the highest (Table S10). Indeed, some studies have shown that the color inheritance of chrysanthemums is generally a partial maternal inheritance, but the performance of different cross combinations is not completely consistent [29,30,31]. Owing to the complex genetic background of chrysanthemum, further genetic analysis of flower color variation is required under more specific cross combinations [32].

4.3. Comprehensive Evaluation of Quality Traits of Garden Chrysanthemum for Hedgerows

As one of the evaluation methods for ornamental plants, AHP can simplify complex problems and obtain the best scheme. This method has been used in the evaluation and screening of tulip, chrysanthemum, Prunus mume, Cymbidium ensifolium and other ornamental plants [33,34,35,36]. In the variety evaluation of different ornamental plants, the AHP method can be used to assign the index weight value, whereas the selection index and matrix assignment should be flexible according to the specific variety evaluation target to establish a suitable evaluation system.

In contrast to the traditional evaluation criteria of garden chrysanthemums for ornamental purposes, our research set the plant growth characteristics as the main evaluation index and established a different evaluation criterion, which provided a reference for the evaluation and screening of garden chrysanthemums for hedgerows. Six indices of growth characteristics and adaptability, including plant height, lodging resistance, crown height, plant form, crown breadth, and the number of main branches, were highly ranked in the test evaluation index system. A higher plant height and crown height ratio can broaden the application of garden chrysanthemums. The compact plant form, round crown width, and large number of main branches can improve ornamental quality and the ability to resist pruning, so that the new varieties have a better advantage for hedgerows in landscape configuration. The other traits were low-ranking qualitative indicators despite and in part because the weighting of ligulate flower number, flower color, and inflorescence diameter between the two traits in the population of the tested varieties and the similarity in traits was caused by long-term directional selection inbreeding. The weight value of the number of days from the planting stage to the color stage was the lowest, indicating that early and late flowering did not affect the ornamental value of the hedgerow varieties, and the flowering stage was only one of the ornamental stages for hedgerows. In conclusion, the selection of cultivars using AHP is also suitable for breeding new varieties of garden chrysanthemums for hedgerows, which can provide a reference for production practices and garden applications.

5. Conclusions

5.1. Establishment and Application of Comprehensive Evaluation System for Hedgerow Garden Chrysanthemum

In this study, a comprehensive evaluation system for F1 generation of garden chrysanthemums for hedgerow was established by combining AHP and cluster analysis. Weights derived from the AHP showed that the six most effective drivers for garden chrysanthemum used in hedgerow were plant height (0.347) followed by lodging resistance (0.181), crown height ratio (0.125), plant form (0.091), crown width (0.089), and the number of main branches (0.047). The C.R. for the studied parameters and indices were within the acceptable limit (<0.10), indicating the high accuracy of the pairwise comparisons and the weight assignment of each factor met the requirements. The three-point scoring standard was used to comprehensively score and evaluate the 12 traits to obtain ideal evaluation results and to screen out the hybrid offspring of garden chrysanthemum with good comprehensive characteristics. Employing K-means clustering analysis, 860 individual plants were divided into four grades: excellent, good, medium, and poor. Combining with the results of significant difference analysis, excellent hybrid offspring were selected for hedgerow garden chrysanthemums, which were characterized by high plants, hard branches, and compact plant form that could be selected in future breeding research.

5.2. Expectation for Breeding Research in Future

Compared with the traditional evaluation system for ornamental garden chrysanthemums, this study took growth characteristics as the targets, provided a variety of criteria for classification and screening of garden chrysanthemums for hedgerow, and established multiple evaluation criteria. Garden chrysanthemum plants are generally shorter and have softer branches. The results of this study indicate that the traits of growth characteristics and adaptability of garden chrysanthemum for hedgerow ranked high in the evaluation index system of this experiment, so they were used as the most critical evaluation indices. High weightings of plant height, crown height ratio, and lodging resistance show that when breeding new varieties, higher plant height and crown height ratio can broaden the use of garden chrysanthemum for hedgerow in landscaping. Compact plant shape, round crown width, and many main branches are important, which can expand advantages in landscape decoration by improving the ornamental quality and enhancing the ability of tolerance for pruning. Meanwhile, the weight value of the number of days from planting period to coloring period is the lowest, suggesting that the flowering time does not seriously affect their ornamental value as hedgerows, and flowering is only one of the ornamental stages. Finally, we selected five garden chrysanthemum varieties with good characteristics that are suitable for hedgerow use and achieved the desired effects.

In future breeding work, researchers should make corresponding adjustments according to the breeding target to obtain traits that are practical and have decisive factors to make a more accurate comprehensive evaluation. For example, when selecting multiflower garden chrysanthemum, it needs to pay attention to the display effect of the group and the indicators related to the ornamental quality like flower density should be considered emphatically. This research will provide a theoretical basis for parental selection in garden chrysanthemum breeding and provide ideas for further breeding new varieties of garden chrysanthemum by in-depth analysis and understanding of the relationship and genetic characteristics of hybrids and their parents.