Horticulture, Environment, and Biotechnology 56(4): 487-497 (August 2015)
Crossability among Modern Roses and Heterosis of Quantitative and Qualitative Traits in Hybrids
Muhammad Nadeem, Adnan Younis, Atif Riaz, Ki-Byung Lim

Abstract Nine divergent popular hybrid rose cultivars and 22 of their F1 hybrids were studied to determine the extent of crossability and heterosis effects for several quantitative and qualitative traits in the climatic conditions of Faisalabad, Pakistan. The percentage of pollen germination differed across cultivars. Gruss-an-Teplitz exhibited the maximum value for pollen germination (46.5%), followed by Autumn Sunset (38.9%). Iceberg, being triploid, presented the lowest percentage of germination. All cultivars exhibited a strong correlation between pollen germination percentages in the lab and crossing success under field conditions. The cultivar Handel showed remarkable fruit-set and crossing success. In contrast, the cultivar Iceberg exhibited very poor crossing success rates. The setting of hips after crossing was maximal in the crosses Handel x Louise Odier and Autumn Sunset x Casino (83% each). The number of seeds in a single hip varied among cultivars. The potential for heterotic and heterobeltiotic was also variable and showed contrasting performance between F1 progenies for various qualitative and quantitative traits. The parent cultivars and progenies were sorted into five major groups by cluster analysis based on phenotypic variation. Phenotypic and genotypic coefficients of variability and heritability percentages varied among all parent cultivars for all morphological traits.


Roses are an important economic crop worldwide, cultivated for cut flowers, essential oils, and landscaping. The breeding of rose plants provides a reliable source for the development of new cultivars and unique germplasm for gardens, the cut-flower industry, and essential oil extraction (Younis, 2006). Many private companies are involved in this field to develop their own cultivars. Rose cultivars can lose their potential for successful hybridization programs because of harsh climates and biotic and abiotic stresses. Hence, there is always a need for genetically improved cultivars to combat these conditions. There is also a need for scientifically rigorous evaluation of the processes involved in inheritance, for proper documentation of rose breeding. In Pakistan, most of the cultivars used in gardens and city green belts have been introduced from other countries, but In Pakistan, due to high temperatures and low humidity indices-conditions that are unsuitable for the production of hybrid seeds and the exhibition of unique performance (Younis et al., 2006). Modern approaches are aimed at developing cultivars that fit the criteria of disease resistance, repeating blooms, color variety, and the ability to withstand the adverse impacts of different climatic conditions.

Crossing between closely related cultivars with common parentage results in a better output that is commercially desirable; however, compatibility barriers can limit crossing success (Younis et al., 2014a). Other challenges include sterility and decreased pollen viability in pollen-donor parents. Pollen and ovule abortion in the progenies can lead to sterile hybrids (Svejda, 1974). Some hybrids present less potential to germinate with limited reproductive performance, as many F1 progenies are unable to accomplish normal growth and development of physiological systems. In practice, many hybrid seeds show failure in seedling emergence, which ultimately leads to complete death of the seed. Even with healthy, vigorously growing seeds and seedlings, some flowers do not always produce desirable traits. In some cases, the F2 generation may produce non-viable seeds or may not produce seeds at all. Thus, any rose breeding program can be challenging, as there are many factors affecting its success. Some of these factors include the condition of the pollen before pollination, pollen growth and pollen tube development, compatibility barriers in the style and stigma, variations in ploidy levels, toxicity in gametes and zygotes, abortion of embryos, and splitting of ripened hips. Other factors contributing to the success of pollen germination, fruit setting, and maturity of hips include climatic conditions and storage environment (Khan, 1988; Farooq et al., 2013). Internal conditions and bio physiological activity of the plants also affect developmental stages of growth and maturation of the hips (Gudin et al., 1991; Gudin, 1992).

The development of modern hybrid rose varieties is the result of the long-term efforts of breeders. Hybrid Tea, Floribunda, Grandiflora, and Polyantha roses were prominent in the 20th century and maintained strain purity because they were propagated by vegetative means. Selected cultivars were known for their remarkably prominent flowers, long vase life, proliferative growth habits, and superior stem length. Breeding efforts result in both desirable traits and many undesirable changes in the progenies (Phillips and Rix, 1988). Cross breeding can also result in variations at the ploidy level, as present-day varieties of hybrid roses can present as triploids and tetraploids (Rout et al., 1999). As chromosome numbers increase, the chances for the successful development of seeds and viable embryos decrease. The aims of rose cultivation include introducing varieties with vibrant colors, large double flowers, longer shelf life of blooms, and enhanced environmental responses. Historically, many rose cultivars have been developed through classic breeding approaches; these cultivars certainly fulfill the breeder’s criteria for desirable traits, but have limitations. Gene combinations have been traditionally restricted and gene crossing has been limited. Consequently, there is less variation in cytology and decreased compatibility among candidate varieties. In addition, the uniformity of most traits in the processes of growth and flowering synchronization is controlled by many genes, the resulting gametes therefore do not demonstrate the coherence of the alleles for a specific trait (Rout et al., 1999).

Several studies on heterosis have been conducted in the cultivars of agronomic crops, including rapeseed (Radoev et al., 2008), rice (Rahimi et al., 2010), Jatropha (Islam et al., 2011), and wheat (Patil et al., 2011). In rose breeding, less attention has been paid to scientifically rigorous studies on the heterotic effect. In light of this, the present study was designed to select inbred hybrid progenies that perform relatively well, to achieve new lines, and to gain greater insight into the inheritance of traits and heterosis. Therefore, the aims of the present study were to compare crossability among popular hybrid rose cultivars and to assess heterosis of several quantitative and qualitative traits in their hybrids.

Materials and Methods

Plant Material and Hybridization

Nine hip-bearing hybrid rose cultivars, namely, Autumn Sunset, Iceberg, Paradise, Angel Face, Louise Odier, Casino, Grand Margina, Handel, and Gruss-an-Teplitz were selected on the basis of their superior performance under the climatic conditions of Faisalabad, Pakistan. Cultivars were planted in the floriculture field of the Horticulture Department at the University of Agriculture Faisalabad (Table 1). Plants of all cultivars were irrigated, and equal amounts of farmyard manure and NPK fertilizer were applied during their growth. Prior to hybridization, the pollen of all cultivars was tested for viability and germination potential in vitro. Crosses were performed in March 2008, using a diallel crossing scheme. Pollens were collected in the evening and stored at room temperature. Crossing was performed in the early morning and pollination was repeated twice a day. All crosses were properly labeled, tagged, and covered with paper bags.

Table 1. Percentage of successful crosses and fertility status of pollen donor parents

Pollen donor parent Pollen viability
No. of
crosses made
% Crossing
Autumn Sunset (V1) **** 35 80 36 45 Fertile
Iceberg (V2) *** 58 80 0 0 Sterile
Paradise (V3) **** 51 80 15 18.75 Low Fertility
Angel Face (V4) **** 60 80 5 6.25 Low Fertility
Casino V5) **** 63 80 12 15 Low Fertility
Louise Odier (V6) **** 42 80 13 16.25 Low Fertility
Grand Margina (V7) **** 51 80 17 21 Moderate Fertility
Handel (V8) **** 70 80 43 53.75 Fertile
Gruss-an-Teplitz (V9) **** 64 80 22 27.5 Moderate Fertility

**** = Tetraploid *** = Triploid

Data Collection

Hips were harvested in July 2008, and hip maturation days were recorded as the number of days elapsing between crossing and hip maturation. The successful crossing percentage was determined for each combination by dividing the number of successful combinations by the total number of combinations and multiplying by one hundred. After the hips matured and were harvested, the fresh weight of each hip was measured using an electric balance. After 20 days of drying at room temperature, seeds were extracted manually from the hips. For each combination, the number of seeds in selected hips was counted with three repeats and the average was calculated. Seeds were placed in paper bags and stored at ambient temperature (24 ± 2°C). For each combination, the weight in grams of 100 seeds was measured in three repeats. The length and width of the seeds (mm) and the width of the pericarp were also measured using a Vernier caliper (CD-6 CSX, Mitutiyo, Kawasaki, Japan). Seeds were subjected to various treatments prior to germination to break dormancy (data not shown), and sown in a peat moss and sand mixture. Seedlings were transplanted into pots and cultivated for one year. For morphological and growth comparisons, progenies and parents were budded on Rosa bourboniana rootstock under field conditions in January 2010. In November 2010, all stems were cut back to 9 inches (~23 cm) from ground level, leaving four main canes on the trunks of all bushes. From March to October 2011, data on various morphological traits were recorded for both parent cultivars and progenies to assess heterosis and to compare the performance of progenies over their parents.

Progeny data were calculated for the total number of blooms (all flowers were counted from the first bloom until the period just prior to pruning). Data for fully matured petals were recorded in March 2011, during the period of active growth of flowers and petals. The diameters of the flowers of progeny saplings were measured using a Vernier caliper. The length in centimeters of the pinnate was recorded from its point of emergence to the terminal point of its final leaf. The size of the plant was taken as the height from its base to its terminal bud.

Bush shape was classified as compact or regular, open or scattering, with intermediate rating levels. The scale for assessment of quality traits was designed to accommodate varying degrees: 1 = regular branching, compactness, and organized development; 2 = less regular branching, but compactness in growth; 3 = irregular branching and partial overall compactness; and 4 = highly irregular branching with an open framework.

To investigate the cumulative potential of the bushes of selected parent genotypes, another scale was designed with the following quality points: 1 = recurrent blooms, compact appearance, and excellent health; 2 = average blooming intensity, partial compactness and average health; 3 = satisfactory blooming intensity, average compactness with fair growth; 4 = flowering irregularity with minimal growth.

Floral scent was also estimated using the following rating scale: 1 = strong fragrance; 2 = average fragrance; 3 = minimal fragrance; and 4 = no discernible fragrance. The presence of flowers and prickles on branches and canes was judged using the following quality point scale: 1 = flowers without prickles; 2 = flowers with a few prickles; 3 = flowers with an average number of prickles; and 4 = flowers with a maximal number of prickles on canes and twigs.

Estimation of Heterosis and Data Analysis

Analysis of the data was conducted using the SPSS-15 software (SPSS, Inc., Chicago, IL). Pairwise comparisons between all means were subjected to the Duncan’s Multiple Range (DMR) test and differences were determined using analysis of variance (ANOVA). The least significant difference (LSD) test was applied to compare variation. Genotypic and phenotypic coefficients of variation, which are the square roots of genetic and phenotypic variance, respectively, expressed as a percentage of the mean, were calculated according to the method used by Burton and DeVane (1953). Broad-sense heritability was estimated as the value of the total variance due to genetic effects. Cluster analysis was conducted on the means of the variables in nine cultivars, according to Ward’s minimum variance method using the statistical software STATISTICA ver.8 (StatSoft, Inc., Tulsa, OK). To determine vigor-related effects on the parameters under study, mid-parent and better-parent heterosis estimations for the performance of F1 progenies were made by the following formula (Fonseca and Patterson, 1968):

Hbt% = [(F1-MP)/MP] x 100
Ht% = [(F1-BP)/BP] x 100

where Ht = Heterosis, Hbt = Heterobeltiosis, MP = Mid-parent value, and BP = Better-parent value.


Pollen Viability, Successful Crosses, Hip Sets, Maturity, and Number of Seeds

The Handel cultivar showed the highest pollen viability (70%) in vitro, and the Autumn Sunset showed the least (35%) in comparison to other cultivars. The selected parent genotypes were crossed in a diallel scheme, therefore, all possible combinations (72 x 10) were made by repeating each cross 10 times, resulting in variations of the hip set percentages among crosses. The success of all the crosses could be attributed to several factors. Among the parent genotypes, the Handel cultivar showed the greatest crossing success (53.7%), followed by the Autumn Sunset generating a 45% viable crossing rate (Table 1). The parent Iceberg cultivar did not exhibit potential crossing success, although it showed 58% pollen viability in vitro.

The maximum hip set (83%) was recorded for the two cross combinations Autumn Sunset x Casino (V1 x V5) and Angel Face x Gruss-an-Teplitz (V4 x V9). The minimum hip set (30%) was recorded for Louise Odier x Paradise (V6 x V3). The maximum time taken for hip maturation (110 days) was recorded for the genotype combination Paradise x Autumn Sunset (V3 x V1), followed by Grand Margina x Gruss-an-Teplitz (V7 x V9) (109 days). Early hip maturation was recorded for the combination Gruss-an-Teplitz x Grand Margina (V9 x V7). Variations in the average number of seeds in a single hip were observed in the cross Louise Odier x Paradise (V6 x V3), resulting in an average number of 33 seeds per hip. The lowest seed count per hip (15) was recorded for the combination Handel x Gruss-an-Teplitz (V8 x V9) as presented in Table 2.

Plant Characteristics

The genotypes and resulting F1 progenies revealed significant differences in bush height (P < 0.001) in both parent cultivars and hybrid progenies (Tables 3 and 4). Among all parent genotypes, the maximum heights (0.89 and 0.83 m) were recorded for Handel (V8) and Gruss-an-Teplitz (V9), respectively. The maximum heights among hybrid progenies were recorded for H19, the hybrid of the cultivars Grand Margina x Casino (0.73 m), followed by H21 (Handel x Gruss-an-Teplitz) yielding heights of 0.72 m (Table 4).

The values for mid-parent heterosis (Ht) and better-parent heterosis (Hbt) also showed variation. Heterosis was evident in 27.7% of the hybrid progeny of Casino x Autumn Sunset (H8), and in 18.1% of the hybrid progeny of Angel Face x Casino (H15) (Table 5). The greatest progeny hybrid vigor over the better parent was exhibited by the F1 progeny of Gruss-an-Teplitz x Autumn Sunset (H10), yielding 15% heterobeltiosis.

The presence or absence of prickles on the bushes was estimated by quality points, i.e., 4 for the greatest number of prickles and 1 for the least. The prickle density varied in both the resulting hybrids and in the parent genotypes. The parent genotype Grand Margina exhibited the maximum density of prickles per plant (3.5 quality points), followed by Paradise and Louise Odier (2.75 quality points each). The average numbers of prickles exhibited by all parent genotypes are presented in Table 3. Among the F1 progenies, H19, the hybrid of Grand Margina x Handel exhibited the maximum density of prickles (3.66 quality points), followed by H1, the progeny of Autumn Sunset x Paradise (3 quality points). H10, the hybrid genotype of Gruss-an-Teplitz x Autumn Sunset and H12, the hybrid of Paradise x Grand Margina, each exhibited 1.55 quality points for prickle density. The performance of all genotypes for prickle density is presented in Table 4.

The F1 progenies H1 (Autumn Sunset x Paradise) and H19 (Grand Margina x Gruss-an-Teplitz) exhibited the greatest heterosis (20%) for prickle density (Table 5). Parent genotypes and their progenies exhibited significant differences (p < 0.001) in bush shape. The parent genotype Iceberg had excellent bush shape, with compact architecture. Excellent bush shape was also exhibited within hybrid progenies of Iceberg (V2) (4.58 quality points out of 5), followed by Autumn Sunset (2.39 quality points). The hybrid genotype H11 (Paradise x Angel Face) scored 2.06 quality points, and H4 (Autumn Sunset x Casino) and H13 (Paradise x Handel) each scored 2.08 quality points. The hybrid genotype H18 (Grand Margina x Handel) showed poor performance for bush shape with 3.75 quality points. Similarly, mid-parent and better-parent heterosis also varied among F1 progenies for bush shape. The progenies H15, (Angel Face x Casino), H14 (Angel Face x Handel), and H18 (Grand Margina x Handel) all had high scores for bush shape. Maximum Hb (20%) was observed in the F1 genotype H15 (Angel Face x Casino).

Flower Characteristics

There were significant differences in petal number for the selected parents and progenies. The parent cultivar Angel Face produced the greatest number of flowers (57), and the cultivar Handel yielded the greatest number of petals per flower (42) (Table 3). Among the hybrids the maximum number of petals per flower (40) was produced by H19 (Grand Margina x Gruss-an-Teplitz), followed by the hybrid H21 (Handel x Gruss-an-Teplitz) (39 petals per flower) (Table 4). Heterosis and heterobeltiosis also varied, as comparatively high values for both heterosis (51%) and heterobeltiosis (49.2%) were evident in the hybrid progeny H2. In addition, the hybrid progeny H11 (Paradise x Angel Face) exhibited heterosis of 53% and heterobeltiosis of 18.6%. Heterosis and heterobeltiosis values for other hybrid progenies are presented in Table 5.

Parent genotypes and progenies exhibited significant differences (p < 0.001) in petal length. Among the parent genotypes, the cultivar Grand Margina yielded comparatively long petals (15.1 mm) (Table 3). Among the F1 genotypes, the hybrids H22 (Gruss-an-Teplitz x Grand Margina) and H19 (Grand Margina x Gruss-an-Teplitz) performed better, yielding petals 16.0 mm in length, followed by H12 (Casino x Grand Margina) generating average petal lengths of 15.5 mm (Table 4). The hybrid progeny H20 (Handel x Grand Margina) yielded the highest heterosis value (24.9%), followed by the hybrid H2 (Autumn Sunset x Angel Face), which scored 3.8% for petal length traits. Similarly, the hybrid progeny H20 (Handel x Grand Margina) exhibited the highest value for heterobeltiosis (4.6%). Other F1 progenies exhibited a mixed blend of heterosis and heterobeltiosis with both negative and positive values. Flower diameter also exhibited significant variations among parent cultivars, but little differences amongst all progenies (Table 4). There were also positive and negative heterosis and heterobeltiosis values for flower diameter (Table 5).

Significant differences were exhibited by the parent genotypes and F1 hybrid progenies for fragrance intensity (Tables 3 and 4). The parent cultivars Grand Margina and Gruss-an-Teplitz scored the greatest fragrance intensities, whereas the cultivar Casino exhibited the least intense fragrance. Among all hybrid progenies, the F1 genotype H22 (Gruss-an-Teplitz x Grand Margina) exhibited the most intense fragrance. The hybrid progenies H9 (Grand Margina x Autumn Sunset) and H19 (Grand Margina x Gruss-an-Teplitz) exhibited heterotic performance over their parents for the fragrance trait. Similarly, hybrid progeny also scored favorable heterobeltiotic values for the fragrance trait (Table 5).

Means analysis revealed that the cultivar Angel Face exhibited a flower persistence life (FPL) of 16.5 days followed by the cultivar Paradise. In comparison to other progenies, the hybrid progeny of Angel Face and Casino exhibited comparatively higher FPL values (15.5 days) in the field from the first day of blooming. The values of flower persistence life in the field for other progenies are presented in Table 4. Among all hybrid progenies, H13 (Paradise x Handel) exhibited maximal heterosis values, followed by the hybrid progeny H8 (Louise Odier x Autumn Sunset). The values of heterosis and heterobeltiosis are presented in Table 5.

The overall performance of the bushes varied significantly among the F1 progeny. The F1 hybrids H17 (Louise Odier x Grand Margina), H20 (Handel x Louise Odier), and H21 (Handel x Gruss-an-Teplitz) had the greatest scores (3.6 quality points each), followed by H8 (Louise Odier x Autumn Sunset) (3.5 quality points). Overall, the performance of the hybrid progeny H21 (Handel x Gruss-an-Teplitz) exhibited superior heterotic and heterobeltiotic effects.

Cluster Analysis


Literature Cited