EVALUATION AND POOLING OF SAFFLOWER (Carthamus tinctorius L.) ACCESSIONS FROM DIFFERENT WORLD ORIGINS USING AGRO-MORPHOLOGICAL TRAITS

Karim Houmanat 1,2 , Hamid Mazouz 2 , Mohamed El Fechtali 1 and * Abdelghani Nabloussi 1 . 1. INRA, Research Unit of Plant Breeding and Plant Genetic Resources Conservation, CRRA of Meknes, Po. Box 578, Meknes, Morocco. 2. University Moulay Ismail, Laboratory of Biotechnology and Molecular Biology, Faculty of Sciences, Meknès, Morocco. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History

About 734.000 tons of safflower are produced in an area of 937.000 ha worldwide. Mexico, USA, Kazakhstan, India and Argentina account for approximately 95% of the world's safflower production (FAOSTAT, 2014). Safflower is a multi-purpose plant cultivated for edible oil, birdseed, spices, dye of its flowers, medicinal and ornamental use (Mündel et al., 1992;Johnston et al., 2002).
In Morocco, safflower cultivation, for oil production, was initiated in 1965 but has been abandoned since 1992, mainly for commercial reason. In fact, the production sale prices were not fixed and guaranteed as was the case for sunflower and rapeseed. In addition, during safflower cultivation period, seed yield and oil content remained very low due to the lack of adapted cultivars and poor management techniques used by farmers (Nabloussi and Boujghagh, 2006). Recently, the Moroccan Ministry of Agriculture has called for the restart and development of safflower through a global action plan. In this context, a new safflower breeding program was launched. The objective is to release new spineless and productive safflower varieties with high technological value characterized by shorter vegetative cycle and higher seed yield and oil content than the existing varieties. Only three varieties have been registered, two of them are spineless, and could produce a quite high seed yield between 1.60 and 2.00 t/ha (Nabloussi and Boujghagh, 2006). However, all these varieties have a low oil content < 30%. The expected varieties should have oil content close to 40%. Furthermore, they must be early to avoid a late season drought and spineless to be suitable for traditional management practices, such as hand-harvesting.
A world safflower collection, comprising more than 200 accessions, was introduced to constitute the initial germplasm for our breeding program. A preliminary study was carried out to evaluate the adaptation and performance of those accessions under Moroccan conditions (Nabloussi et al., 2008). Among this material, 61 accessions were selected for further evaluation in other environments. The present study was conducted to assess the genetic diversity of 61 safflower genotypes from different geographic origins in the world, using agromorphological and technological traits during two seasons.

Materials and Methods:-Plant material:-
The plant material used in this research consisted of 61 accessions previously selected among a world safflower collection provided by the Department of Agriculture and Research, Agricultural Service in the United States (Nabloussi et al., 2008), and a cultivar from Spain, Namely 'Rancho', provided by CSIC, IAS-Cordoba, and used as a ckeck. All these accessions were planted during two seasons, 2013 and 2014, at the INRA Experimental Station of Douyet. Table 1 summarizes the INRA-codes, PI-codes and countries of origin of these accessions.

Planting design:-
The genotypes tested, long with the check, were planted late in February during 2013 and early in December during 2014. The field experiment was conducted during these two years using a completely randomized blocks design, with two replications. Each genotype was grown in a single 3 m-length row, with inter and intra row spacing of 80 and 30 cm, respectively.

Observations and Measurements:-
Data were gathered on morphological parameters (initial vigor, degree of spines presence, plant height, number of branches per plant, number of heads per plant, color of petals, homogeneity), phenological parameter (flowering time), agronomic parameters (number of heads per plant, NHP, number of seeds per head, HSN, thousand seeds weight (g), TSW, seed yield per plant (g), total seed yield per hectare (t/ha), TSY, technological parameter (seed oil content, SOC) and pathological parameters (degree of leaf rust presence, degree of broomrape presence). A random sample of five plants per genotype was taken for the different measurements.
The initial vigor (IV) was evaluated on a scale ranging from 1 (least vigorous accession) to 5 (most vigorous accession).The presence or absence of spines (SL) was assessed on a scale from 1 (spineless accession) to 4 (highly spiny accession). Similarly, homogeneity (HOM) was rated on a scale ranging from 1 (very homogeneous accession) to 4 (very heterogeneous accession). Number of days at flowering (DFS) was the number of days from planting to opening of the first flower on each sample plant. The reaction of accessions to rust (RR), was assessed according to a scale of attack severity ranging from 1 (no, or very light, attack) to 4 (very severe attack).Broomrape virulence (BV) was evaluated, only in the second year (2014),by accounting the number of broomrape stems infecting each sample plant. The oil content (SOC) was determined by Nuclear Magnetic Resonance (NMR, Oxford 4000) and was expressed in% of dry matter.
Statistical Analysis:-Descriptive analysis of gathered data and analysis of variance were performed using anova procedure of SAS program. Duncan's new multiple range test was applied to compare genotypes means. This enabled to distinguish which accessions were significantly different from others and, thus, to select those interesting ones according to an established selection threshold. This threshold was fixed on basis of the parameters studied and their relation with 929 agronomic and technological performances, the overall average of each parameter and the check performance (Houmanat et al., 2016).  Table  2. There were significant differences (p <0.001) between the 62 accessions for all the traits studied, indicating the existence of a very large genetic diversity expressed by wide ranges of variation of these traits. These results confirmed those obtained by Nabloussi et al. (2008) who studied a larger safflower collection from which the 62 accessions of the present study had been selected. Other authors found similar results, in their previous researches, namely Mahasi Figure 1 showed the inter-annual variation recorded for the parameters studied during two consecutive and contrasted cropping seasons. Analyzing the 'location' effect on the evaluated parameters, Nabloussi et al. (2008) had reported significant differences between locations studies, except for TSW. The variability observed between both years, for most of the parameters studied, could be explained by the contrasting environmental conditions, mainly rainfall and temperature, aggravated by planting date which was too late in 2013. On the other hand, the fact that both years were comparable for SL, TSW and SOC, might indicate that these characters were not significantly affected by those contrasting conditions. Also, the interaction Year × Accession had significant effect on all parameters, except TSY and NHP. Interestingly, this indicated that the better accessions, regarding seed yield and number of heads per plant, were confirmed from one environment to the other. In other words, they maintained their higher performance in terms of seed yield and number of heads per plant under contrasted environmental conditions. Previous study on the original safflower collection had shown significant effect of the interaction Location × Accession for all parameters, except TSW (Nabloussi et al., 2008).

Results and Discussion
As mentioned above, there was a large variability among the accessions for all parameters studied. Hereafter, description and discussion of averages and variations recorded for morphological, pathological, agronomic and technological traits are given (Table 3).

Morphological Parameters:-
The accessions showed a variation in the color of their petals (COP): red, white, yellow, orange and intermediate color (Fig.2). Some studies have highlighted the close correlation between color of safflower petals and their biochemical composition, mainly the total polyphenol content (carthamin) (Hiramatsu et al., 2008).This component, which is characterized by its high antioxidant properties, can be further explored and exploited for industrial and pharmaceutical purposes. Also, the accessions studied showed a large variation in the degree of spinessness (SL) at leaves and bracts, varying from strongly spiny to completely spineless. On average, the accessions were spiny (2.96); therefore, a large number of genotypes had fewer spines than the check variety (Rancho), which is too spiny (4). Comparable findings had been reported by Nabloussi et al. (2008). Narkhede and Deokar, (1990) reported that the spiny type is generally dominant over the spineless type in safflower genotypes.It was reported that cultivation of safflower was not widely practiced because of the presence of spines, which is a major constraint for some operations, such as harvesting and threshing (Nabloussi and Boujghagh, 2006;Sujatha, 2008). Thus, one of the main objectives of safflower breeding programs is to develop spineless varieties with high yield and high oil content (Nabloussi and Boujghagh, 2006; Golkaret al., 2010).   (Nabloussi, 2006).
Regarding rust resistance (RR), the accessions studied were in general moderately resistant, with an average score of 2.9.'Rancho'had 3 as resistance score, and the accession'216' was the only one with no rust symptoms, which might indicate that it was immune. For the rest of accessions, the resistance score ranged from 1 (most resistant) to 4 (most susceptible). Nabloussi et al. (2008) found similar results on related safflower material. In a previous study on ten safflower genotypes of different origins in field conditions, Karakaya et al. (2004) showed that all of them were attacked by Puccinia carthami, the causative agent of safflower rust, but with different degrees of severity. The average seed yield (TSY) was almost 2 t/ha. However, there was a very large variation among the accessions, from 0.231 to 4.490 t/ha. The accession'164' was the least productive due to its high susceptibility to broomrape infection, while the accession '123' was the most performant. However, several genotypes were more productive than the check variety which produced 1.3 t/ha. Narrower variation ranges were reported in previous studies, from 1.5 to 2.7 t/ha (Panahiet al., 2013) and from 1.1 to 2.2 t/ha (Amini et al., 2007).

932
Accessions pooling:-Overall, high genetic diversity was observed among the evaluated accessions, which belong to different geographical origins, for all the parameters studied. In particular, large variations were shown for some economically important traits, such as seed yield, seed oil content, flowering earliness, spinelessness, rust resistance and broomrape resistance. This would be potentially useful for our breeding program. As was suggested by Nabloussi et al. (2008), we could think of developing some genetic pools in the light of our findings. Six pools would be feasible, and each of them may comprise the most homogeneous, performant and stable accessions regarding a specific trait from the six ones mentioned before. Table 4 shows the proposed genetic pools with accessions selected according to a suggested threshold for each trait. The threshold established for the development of each pool was made taking into account the degree of variability observed among the safflower accessions and the performance of the check variety used in this study.
In the next future, crossings and interbreeding should be achieved between genitors belonging to different pools in order to combine various desirable traits in some particular safflower germplasms. These bred germplasms will be suitable and useful for developing highly productive and adapted safflower varieties to be cultivated in Morocco as well as other regions throughout the world.

Conclusions:-
In light of the results obtained, a large genetic diversity was noticed among the accessions evaluated for all the parameters studied. Also, there was a significant effect of the interaction environment (year) × accession on all the parameters, except TSY and NHP. This finding means that accessions were stable across both environments for these two traits. Thus, there will be a possibility to identify and select some stable and widely adapted accessions, with higher TSY and NHP than the check variety. This is very interesting since TSY is the most important breeding objective in safflower. Furthermore, it was possible to conceive and establish six genetic pools on the basis of single and economically important traits. These traits are spinelessness, early flowering, rust resistance, broomrape resistance, seed oil content and seed yield. The accessions comprised in each pool showed an interesting level of performance and stability in terms of each corresponding trait. The pools thus obtained will be multiplied, conserved and then used as valuable genetic resources in crossing and interbreeding for safflower breeding programs in Morocco as well as in other safflower cultivation areas in the world.