An assessment on durum wheat cultivars and lines suitable for frike making in southeast Turkey

doi:10.21203/rs.3.rs-2949048/v1

This study aimed to assess the most suitable durum wheat cultivars and lines for frike making through frike yield and quality, as well as marketing prices and profitability. Field trial was set up in Gölbaşı (Adıyaman) in 2021-22 growing season. A randomized complete block design with 20 entries and three replications were employed. Whole grain frike yield, some of frike quality traits such as frike color, humidity (%), hectoliter weight (g), protein content (%), wet gluten (%), 1000 kernel weight (g), ’L’ shininess value, ’a’ value for green (-) vs red (+) scale, ‘b‘value for blue (-) vs yellow (+), raw cellulose (%), marketing prices (US$ kg-1) and profitability (US$ 3 kg-1 fresh spikes) were scored. The results revealed that Tüten-2002, R5(Hom-) Edessa, R5 (Hom+), Artuklu, and Perre had the highest frike yields giving 679.91, 621.16, 620.03, 593.33 and 570.5 g per three kg of fresh roasted spikes. The first five highest marketing price generating entries were Saricanak-98, R112(Hom+), Burgos, Sumerli and Artuklu giving 1.449, 1.428, 1.275, 1.049 and 1.026 US$ kg-1, respectively. Sarıcanak-98, Artuklu, Burgos, R5 (Hom-) and Sumerli were the most profitable lines giving 0.653, 0.608, 0.581, 0.564 and 0.535 US$ from three kg of fresh spikes. Finally, considering all traits under this study, rank-sum analysis showed that Artuklu, Firat-93, Saricanak-98, R5 (Hom-), and R 112(Hom+) were the most convenient entries for frike making giving the least sum of ranks.

Triticum turgidum

frike yield

quality

marketing price

net return

According to food lone, frike consumption dates as far as 2300 BCE (Before Common Era) (US durum 2015). Frike became common in the east Mediterranean Basin and has long been a significant part of the food culture in some WANA countries such as Algeria, Egypt, Lebanon, Jordan, Morocco, Syria, Tunisia, and Turkey (Akgün et al. 2021; Carsanba et al. 2017; Bayram 2008). Great Ottoman traveler Evliya Çelebi indicated in his travelogue (1630-1676) that frike was consumed as ‘firik soup’ (Yerasimos 2011) in some alms houses, dervish lodges, caravanserais, and some monasteries. Frike (or Firik, Frik, Freekah, Freekeh, farik, fireek) is an ancient food from wheat grain that has been harvested early, while the grains are still tender and green. This stage ranges from 80 to 89 of the Zadoks scale (Zadoks et al. 1974). The kernels are then parched, roasted, dried, and rubbed. Frike is the name of the food process rather than a specific grain variety. However, it is made from wheat and generally durum wheat. Studies showed that frike has a low glycaemic index as well as a high content in iron, fiber, and lutein. It is also a good source of fructo-oligosaccharides and some functional compounds (Yılmaz and Yıldırım 2020). Frike has positive effects on the consumers health, and it is considered a functional food. Accumulation of functional compounds in the kernel starts from the spike formation to the end of milk stage of grain filling. Harvest at dough stage of kernel development results in capture of high amount of fructan groups and this groups help to increase calcium and iron use efficiencies (Huynh et al., 2008). Antioxidant, physico-chemical and nutritional values of frike obtained from immature grains varies upon variety and maturity period (Sramkova et al. 2009). Apart from antioxidant activity, the content of total phenolic compounds decreases due to maturity period (Özkaya et al. 2018). Grain composition changes during maturity period, which greatly affects the frike characteristics (Humeid et al. 1990). Due to the high sugar content of grains, spikes, harvested at the late milk stage of grain filling are preferred to full mature grains (Bayram, 2008). Harvest stage is crucial for optimum frike production. Early harvest might result in large number of carbonic spots on roasted grains derived from over effect of flames and contrary, at late harvest, yellow grains might come across instead of green ones. Growing stage of 75-77 of Zadoks scale seems to be best for spike harvest for frike making. Despite its huge marketing potential, farmers, and consumers of frike are not still taking advantages of its nutritional and food potential. Annual world frike production ranges between 250,000 and 300,000 tons (Akgün et al. 2021; Yılmaz and Yıldırım 2020; Bayram 2008;). This can be increased to a million tons via increasing international marketing opportunities to far East countries such as Japan. Frike is a lucrative product, its marketing price is about three-four times than that of bulgur (Yılmaz and Yıldırım 2020). Frike is produced in some specific locations in Turkey starting from Gölbaşı (Adıyaman) and spreads southwest direction to Yavuzeli, Oguzeli (Gaziantep) and jumps to Reyhanlı in Hatay province. There are no accurate figures for frike production in Turkey. Turkey annual frike production is estimated about 2,000 tons (Dagyudan, p.c.). Durum wheat (Triticum turgidum L. subsp. durum (Desf.) Husn.) is the raw material for frike making, and selection of genotypes with better frike quality is possible and desirable. In pre-historic era, people avoided wheat species of fragile spike at maturity and grain shattering such wild diploid einkorn (T. boeticum) or wild emmer (T. turgidum L. ssp. Dicoccoides) for frike making (Vallega 1996). There is no evidence that cultivated emmer (T. turgidum L. ssp. dicoccum) was used for frike making due to their thin seed coat affected from flame, severely resulting in large carbonic spots on grains during roasting (Nespit and Samuel 1996). Durum wheat is the most suitable wheat species for frike making, since the best frike is made from the largest and hardest grains (Bayram 2008). Cultivars Sarıçanak-98 in Gölbaşı, and Zenit and Diyarbakır-81 in Gaziantep, were the preferred by farmers for their required frike characteristics such as high frike yield, early grain filling ability, drought, cold tolerance, and suitability for roasting. But there was no systematic study for cultivars suitable for frike making in the region.

The objective of this study was to assess the suitability of a set of durum wheat cultivars and lines for frike making ability through yield, quality, marketing price and profitability.

Material

A set of 18 durum wheat cultivars and lines were selected among a collection of 58 entries including some Italian landraces and durum wheat-wheat wild relatives (DW-WWR) near isogenic lines tested in previous growing season employing early maturing ability, good early drought and cold tolerance, high grain yielding ability, high yield components such as number of grains per spike, and grain weight per spike. Finally, high glume and palea thickness for roasting suitability (> 0.2 mm) were considered a breeding criterium for selection of candidate lines. Two landraces with relatively late maturing, thick and hairy glumes were also included in the set. Awned cultivars were also given priority for being facile flammable capability during roasting.

Methods

Field trials were set up in fallow land employing randomized complete block design with 20 entries and three replications in Gölbaşı location of Adiyaman province of southeast Anatolia (Turkey). They were planted on 10th November in 2021 by plot drill with a seed rate of 500 grain m-2. Plot size was 6 m ×6 rows (= 7,2 m²) at planting and 5m × 6 rows (= 6 m²) at harvest. 300 kg ha-1 20-20-0 as base fertilizer were applied, and 150 kg ha-1 of urea was applied at late joining and early stem elongation stage. No irrigation was practiced. All other agronomical measures were taken such as weed control and rodent management. Harvest maturity was observed daily by local farmers and local extension office involving in the study. Appropriate harvest time for each entry was detected by squashing the grains.

Confirming the presence of dough stage, 3 kg plot-1 spikes (Fig. 1) were trimmed and immediately roasted, keeping the green color of grain stable at the experimental area. Open air fire at soil surface was employed for frike roasting. In one-week time, all entries, except landraces achieved appropriate stage (75–77 of Zadoks scale) for harvest. Roasting was fulfilled by local farmers who are expert on this process. Roasted spikes kept green color stable, but the grain humidity was still high and all frike samples were dried under shadow shelter one week (Fig. 2).

Dried whole grain frike samples were threshed by single plant thresher at Olgunlar Seed Company in Adıyaman. The frike yield of the entries were scored. Frike samples of all replications were presented to a frike purchaser in local commodity market (Fig. 3). There was no other purchaser in the commodity market. Therefore, frike samples could not be presented to various purchasers.

Chemical analysis

Quality analyses were practiced at the seed quality laboratory of local commodity market and the field crop department of Faculty of Agriculture of the Harran University (Turkey). Frike color inspections of samples were first made visually, and they were identified employing a 1 to 5 scale (i.e., 1: yellow, 2: cream, 3: light green, 4: green and 5: dark green). Humidity (%), hectolitre weights (kg), whole frike grain protein content (%), wet gluten (%) were scored by Perten 9500 calibrated American Association of Cereal Chemistry (AACC) methods. L (100 = white;0 = Black),’a’ (-green; +red value) and ‘b’ (-blue; +yellow) values were scored by Minolta CR400 color meter. 1,000 kernel weights were scored by Uluöz (1965). Raw cellulose (%) analysis was also performed referring Haver-Boecker method (Crampton and Maynard 1938).

Finally, frike samples of each cultivar with replications were presented to the local purchaser dealing with frike for frike marketing price estimates. Net returns were calculated by multiplying purchasing price × frike yield.

Statistical analysis

Frike samples of all three replications were subjected to analysis of some quality traits under study. Visual frike color scores were transformed by square root before ANOVA analysis. Lower ‘a’ and ‘b’ values were better for green frike colour scored by Minolta CR 400. ANOVA for frike grain yield, frike colour (visual), humidity (%), hectolitre weights (kg), protein content (%), wet gluten (%), 1,000 kernel weights (g), L, a and b values, and marketing prices (US$ kg ^-1) were performed and statistically significant means were grouped by a post-hoc Tukey test.

Marketing price estimates of each entry were multiplied by frike yields, and profitability per 3 kg of green spikes was calculated for each entry and ranked. Finally Rank-sum method (Mulamba and Mock 1978) was performed for final selection of entries referring to all traits under study. JMP 11 software was used for statistical analysis.

The 2021/22 durum wheat growing season had an average rainfall season of 506 mm. Rainfall distribution throughout the season was not so regular, early drought resulted in late sowing about a month in autumn. Anova’s were not given here. Means and the groups of the characteristics under study were given in Table 1a,1b and 1c.

Table 1

a: Means separations of some quality traits.
Entry no (EN)	Entry name	EN	Visual frike Color/ Grp	EN	Humidity %	EN	Hectoliter Weight (g) / Grp	EN	Protein (%)	EN	Wet gluten (%) / Grp
1	Sarıbaşak	10	2.15a	9	9.36	12	78.16a	5	12.23a	5	25.4a
2	Simeto	9	2.11ab	6	9.33	8	77.53a	2	12.10a	2	24.96ab
3	Zühre	7	1.98ab	3	9.33	17	77.26a	12	11.90a	12	24.60ab
4	Güneyyıldızı	2	1.98ab	14	9.30	10	75.80a	10	11.63ab	10	24.03ac
5	Fırat-93	15	1.82ac	15	9.26	7	75.76a	15	11.43ab	15	23.53ac
6	R5(hom+)	19	1.82ac	10	9.23	6	75.49a	20	11.30ab	20	23.16ac
7	S.Çanak-98	18	1.79ac	17	9.23	4	75.06a	7	11.20ab	9	22.95ac
8	Artuklu	11	1.73ac	1	9.23	3	75.00a	18	11.20ab	18	22.93ac
9	R5(Hom-)	4	1.71ac	13	9.23	5	74.86a	9	11.19ab	7	22.90ac
10	R112(Hom+)	5	1.66ac	8	9.16	16	74.84a	8	11.16ab	8	22.83ac
11	Sümerli	13	1.60ac	16	9.16	13	74.40a	1	11.13ab	6	22.78ac
12	R112(Hom-)	14	1.60ac	11	9.13	11	74.26a	6	11.12ab	1	22.76ac
13	A.Kale-2000	20	1.57ac	2	9.10	1	74.06a	14	11.06ab	14	22.56ac
14	Edessa	8	1.47ac	4	9.10	2	73.60a	13	11.03ab	13	22.53ac
15	R23(Hom-)	6	1.39ac	5	9.10	9	73.34a	4	10.90ab	4	22.16ac
16	Tüten-2002	1	1.38ac	18	9.10	14	72.63a	11	10.86ab	11	22.16ac
17	Perre	3	1.38ac	12	9.06	15	71.13a	3	10.80ab	3	22.00ac
18	Burgos	12	1.13bc	7	8.96	20	71.06a	17	10.40ab	17	20.93bc
19	Devedişi	17	1.13bc	19	8.93	18	70.43a	16	10.14ab	16	20.45bc
20	Hacı Ali	16	1.11c	20	8.80	19	56.90b	19	9.8b	19	19.66c
Data were analyzed by columns with Tukey test (0.95). Figures that are followed by a letter in common are not significantly different.

Table 1

b : Means separations of some quality traits.
Entry no (EN)	Entry name	EN	1,000 kernels weight (g) /grp	EN	L value	EN	a value	EN	B value	EN	Raw cellulose % /grp
1	Sarıbaşak	6	59.24	19	35.897	20	7.82a	4	16.82a	13	2.76a
2	Simeto	12	55.06	17	34.58	16	6.90ab	17	16.30ab	8	2.50ab
3	Zühre	16	51.14	15	33.59	14	6.33ab	13	16.13ab	20	2.44ab
4	Güneyyıldızı	10	50.38	13	33.38	17	6.28ab	11	16.02ab	14	2.25ab
5	Fırat-93	15	49.76	7	33.34	6	6.11ac	16	15.96ac	12	2.14ab
6	R5(hom+)	9	49.09	4	33.17	12	6.09ac	3	15.73ac	16	2.12ab
7	S.Çanak-98	3	47.88	3	32.81	3	6.04ac	8	15.60ac	5	2.12ab
8	Artuklu	5	47.80	11	32.53	8	5.98ac	7	15.36ac	19	2.10ab
9	R5(Hom-)	2	47.43	20	32.19	5	5.73ac	18	15.29ac	15	2.08ab
10	R112(Hom+)	17	47.20	5	31.92	9	5.60ac	14	15.29ac	3	2.03ab
11	Sümerli	8	46.85	9	31.88	4	5.50ac	12	15.20ac	1	2.03ab
12	R112(Hom-)	20	45.30	18	31.77	13	5.46ac	15	14.98ac	18	2.01ab
13	A.Kale-2000	1	44.86	8	31.68	1	5.05ac	5	14.95ac	9	1.98ab
14	Edessa	13	44.83	12	31.33	15	5.02ac	9	14.91ac	2	1.95ab
15	R23(Hom-)	18	44.47	1	30.73	11	4.63ac	20	14.82ac	10	1.88ab
16	Tüten-2002	14	43.67	10	30.58	7	4.10ac	6	14.53ac	17	1.87ab
17	Perre	7	42.30	16	30.09	18	4.02ac	1	14.49ac	4	1.86ab
18	Burgos	4	40.76	14	29.45	2	3.88bc	19	14.11ac	11	1.85ab
19	Devedişi	11	40.20	6	29.03	10	3.54bc	10	13.72bc	7	1.59b
20	Hacı Ali	19	39.45	2	27.73	19	2.09c			6	1.58b
Data were analyzed by columns with Tukey test (0.95). Figures that are followed by a letter in common are not significantly different.

Table 1

b : Means separations of some quality traits.
Entry no (EN)	Entry name	EN	1,000 kernels weight (g) /grp	EN	L value	EN	a value	EN	B value	EN	Raw cellulose % /grp
1	Sarıbaşak	6	59.24	19	35.897	20	7.82a	4	16.82a	13	2.76a
2	Simeto	12	55.06	17	34.58	16	6.90ab	17	16.30ab	8	2.50ab
3	Zühre	16	51.14	15	33.59	14	6.33ab	13	16.13ab	20	2.44ab
4	Güneyyıldızı	10	50.38	13	33.38	17	6.28ab	11	16.02ab	14	2.25ab
5	Fırat-93	15	49.76	7	33.34	6	6.11ac	16	15.96ac	12	2.14ab
6	R5(hom+)	9	49.09	4	33.17	12	6.09ac	3	15.73ac	16	2.12ab
7	S.Çanak-98	3	47.88	3	32.81	3	6.04ac	8	15.60ac	5	2.12ab
8	Artuklu	5	47.80	11	32.53	8	5.98ac	7	15.36ac	19	2.10ab
9	R5(Hom-)	2	47.43	20	32.19	5	5.73ac	18	15.29ac	15	2.08ab
10	R112(Hom+)	17	47.20	5	31.92	9	5.60ac	14	15.29ac	3	2.03ab
11	Sümerli	8	46.85	9	31.88	4	5.50ac	12	15.20ac	1	2.03ab
12	R112(Hom-)	20	45.30	18	31.77	13	5.46ac	15	14.98ac	18	2.01ab
13	A.Kale-2000	1	44.86	8	31.68	1	5.05ac	5	14.95ac	9	1.98ab
14	Edessa	13	44.83	12	31.33	15	5.02ac	9	14.91ac	2	1.95ab
15	R23(Hom-)	18	44.47	1	30.73	11	4.63ac	20	14.82ac	10	1.88ab
16	Tüten-2002	14	43.67	10	30.58	7	4.10ac	6	14.53ac	17	1.87ab
17	Perre	7	42.30	16	30.09	18	4.02ac	1	14.49ac	4	1.86ab
18	Burgos	4	40.76	14	29.45	2	3.88bc	19	14.11ac	11	1.85ab
19	Devedişi	11	40.20	6	29.03	10	3.54bc	10	13.72bc	7	1.59b
20	Hacı Ali	19	39.45	2	27.73	19	2.09c			6	1.58b
Data were analyzed by columns with Tukey test (0.95). Figures that are followed by a letter in common are not significantly different.

Frike yield

ANOVA revealed that entries were significant different statistically (p ≤ 0.006**). Tuten-2002, Edessa, R5 hom+, Artuklu and Perre took up the first five positions with 679.91, 621.16, 620.03, 593.33 and 570.5 g/ 3 kg green spikes, respectively. The CV was quite reliable with a 15.66%.

Quality traits

Frike color (visual inspection)

Square root transformed data was subjected to analysis of variance and both replications (p ≤ 0.0001***) and entries (p ≤ 0.0006***) were found to be significant. CV was 16.6%. Names of the entries with original means ranked at first five were R112 hom+ (4.66), R5 hom- (4.41), Sarıçanak-98 (3.92), Simeto (3.92) and R23 hom-(3.31) respectively. Entries of Italian origin DW-WWR gave dark green frikes.

Humidity (%)

ANOVA analysis indicated that none of the source of variation turned out to be significant statistically. Means ranged from 8.8–9.36%.

Hectolitre weight (kg)

Analysis of variance indicated the presence of statistically significant differences among entries (p ≤ 0.0001***). A low value of 4.58% of CV indicated the reliability of data. The five entries with higher hectolitre weight were R112 hom+, Artuklu, Perre, R112 hom- and Saricanak-98 with 78.16,77.53, 77.26, 75.8 and 75.76 kg, respectively.

Protein (%)

Entries in the ANOVA turned out to be significant (p ≤ 0.006**). A 5.48% of CV indicated the high reliability of the results of this trait. Entries Firat-93, Simeto, R112hom-, R112 hom + and R23 hom- were the top 5 giving protein percentages values of 12.2, 12.1, 11.9, 11.63 and 11.43, respectively.

Wet gluten (%)

Performed ANOVA showed that the entries were significant statistically (p ≤ 0.0048**). A 6.26% of CV indicated the reliability of the results. Firat-93, Simeto, R 112 hom-, R 112 hom + and R23 hom-, were top five ranking entries giving 25.4, 24.96, 24.6, 24.03 and 23.53% of wet gluten, respectively.

1000 kernel weights (g)

Entries turned out to be non-significant unexpectedly in the ANOVA. CV was relatively high for this trait with 15,39%. R5 hom + had the highest thousand kernel weight generating entry with a 59.24 g, whereas Devedişi (local landrace) gave the lowest with a 39.45 g.

L (shininess)

While a reliable CV of 8.35%, ANOVA indicated the absence of any significant source of variation. Devedişi gave the highest L value (35.89) and Simeto gave the lowest (27.73).

‘a’ value

Although the presence of a high CV with 23.7%, entries turned out to be significant statistically (p ≤ 0,0014**) in the ANOVA. The lower values were the higher ranks. Devedişi, R112 hom+, Simeto, Burgos and Saricanak-98 were the top 5 ranking entries with 2.09, 3.54, 3.88, 4.02 and 4.10 respectively.

‘b’ value

With a reliable CV of 6.22%, ANOVA indicated that entries were statistically significant with an p ≤ 0.039*. The lower values were the higher ranks. The entries Simeto, R112 hom+, Devedisi, Saribasak and R5 hom + placed the first five positions giving 12.93, 13.72, 14.11, 14.49 and 14.53, respectively.

Row cellulose (%)

CV of this trait was quite high with 17.67%. This resulted in the non- significant entries in the ANOVA with a p ≤ 0.08s. Scores ranged from 1.58 to 2.76. Akcakale-2000 gave the highest score with 2.76 and R5 (Hom-), Haci Ali, Edessa and R112 (Hom+) placed at second, third, fourth and fifth ranks giving 2.50, 2.44, 2.25 and 2.14, respectively.

Marketing prices (US$ kg^-1 )

Although the CV was quite high (28.9%), there were statistically significant differences (p ≤ 0.0001***) between entries under study. TL (Turkish lira) kg-1 was converted to US$ kg-1 and the first five highest marketing price generating entries were Saricanak-98, R112(hom+), Burgos, Sumerli and Artuklu giving 1.449, 1.428, 1.275, 1.049 and 1.026 $kg-1 respectively.

Net returns (US$ 3 kg-1 fresh spikes)

Means of marketing prices for each entry was multiplied by frike yields obtained from 3 kg of fresh spikes of each entry. Net returns were calculated as US$ 3 kg-1 fresh spikes. Sarıcanak-98, Artuklu, Burgos, R5 (Hom-) and Sumerli placed at top five positions giving 0.653, 0.608, 0.581, 0.564 and 0.535 US$ from 3 kg fresh spikes.

Rank-Sum analysis

Ranks of the eight quality traits under study were added and the rank of the sums was given in Table 2a. Rank of added of quality traits, ranks of frike yield, marketing price and net return were further tested giving equal weights and employing rank-sum analysis (Table 2b).

Table 2

a. Rank –Sum analysis of some quality traits of entries.
Entry no (EN)	Entry name	Rank frike visual color	Rank Hlt. weight	Rank 1,000 kernel weight	Rank protein ratio	Rank wet gluten	Rank L	Rank ‘a’	Rank ‘b’	Rank raw cellulose	Sum of ranks	Rank sum of rank (A)
1	Sarıbaşak	16	13	13	11	12	15	8	4	11	103	13
2	Simeto	4	14	9	2	2	20	3	1	14	69	4
3	Zühre	17	8	7	17	17	7	14	15	10	112	15
4	Güneyyıldızı	9	7	18	15	15	6	10	20	17	117	16
5	Fırat-93	10	9	8	1	1	10	12	8	7	66	3
6	R5(hom+)	15	6	1	12	11	19	16	5	20	105	14
7	S.Çanak-98	3	5	17	7	9	5	5	13	19	83	8
8	Artuklu	14	2	11	10	10	13	13	14	2	89	9
9	R5(Hom-)	2	15	6	9	7	11	11	7	13	81	7
10	R112(Hom+)	1	4	4	4	4	16	2	2	15	52	1
11	Sümerli	8	12	19	16	16	8	6	17	18	120	17
12	R112(Hom-)	18	1	2	3	3	14	15	10	5	71	5
13	A.Kale-2000	11	11	14	14	14	4	9	18	1	96	11
14	Edessa	12	16	16	13	13	18	18	11	4	121	18
15	R23(Hom-)	5	17	5	5	5	3	7	9	9	65	2
16	Tüten-2002	20	10	3	19	19	17	19	16	6	129	20
17	Perre	19	3	10	18	18	2	17	19	16	122	19
18	Burgos	7	19	15	8	8	12	4	12	12	97	12
19	Devedişi	6	20	20	20	20	1	1	3	8	79	6
20	Hacı Ali	13	18	12	6	6	9	20	6	3	93	10

Table 2

b. Rank –Sum analysis of major agricultural traits of entries giving equal weights.
Entry no (EN)	Entry name	Rank of quality traits (A)	Rank of frike yield (B)	Rank of marketing Price (C)	Rank of net return (D)	Sum of ranks	Rank of rum of ranks	Rank std deviation of A, B, C and D
1	Sarıbaşak	13	10	9	11	43	12	1.70
2	Simeto	4	15	7	8	34	7	4.65
3	Zühre	15	7	20	19	61	19	1.73
4	Güneyyıldızı	16	13	18	18	65	20	2.36
5	Fırat-93	3	6	8	6	22	2	2.06
6	R5(hom+)	14	3	11	7	35	8	6.23
7	S. Çanak-98	8	18	1	1	28	4	8.04
8	Artuklu	9	4	5	2	20	1	2.56
9	R5(Hom-)	7	8	6	4	25	3	1.70
10	R112(Hom+)	1	20	2	9	32	5	8.75
11	Sümerli	17	12	4	5	38	9	6.13
12	R112(Hom-)	5	11	14	14	44	13	4.24
13	A. Kale-2000	11	9	16	15	51	16	3.30
14	Edessa	18	2	12	10	42	11	6.60
15	R23(Hom-)	2	17	19	20	58	18	8.42
16	Tüten-2002	20	1	17	13	51	15	8.34
17	Perre	19	5	13	12	41	10	5.73
18	Burgos	12	16	3	3	34	6	6.55
19	Devedişi	6	19	10	16	51	14	5.85
20	Hacı Ali	10	14	15	17	56	17	2.98

Results indicated that Artuklu, Firat-93, R5 (Hom-), Saricanak-98, and R112(Hom+) were top five genotypes and the most convenient entries for frike making giving the least sum of ranks. Furthermore, Artuklu, Firat-93 and R5 (Hom-) were also stable giving lower rank standard deviations.

To avoid preparing frike with the whole plot of the field trial, only an amount of 3 kg of fresh spikes of each entry in each replication was sampled for frike making. Spikes were harvested including peduncle. A total of 9 kg of spikes from each entry in three replications seemed to be adequate for frike roasting. Average frike/ fresh spike ratio was 1:6 (weight: weight) due to the fully dry weighting. A 1:2 and 1:3 ratio is thought to be acceptable.

Waste was quite high due to contaminations such as chaff and small stones during roasting, and the presence of carbonized kernels due to over roasting. Harvest time for each entry was detected by local frike maker farmers. Unfortunately, harvest for some entries were a little late. Higher frike yield generating entries were little late harvesting entries for frike making. Farmers preferred early maturing varieties with longer stay green period. Escaping from early drought and allowing silage making of green herbage retained in the field after fresh spike harvest for frike making and cotton plantation at the same growing season (Kozak, p.c.).

Five classes for frike color were identified visually. Timely harvested entries generated shine or pale green frikes. Harvest time is very important for frike making. As soon as spikes are harvested, they must be roasted immediately. Otherwise, they lose their original color rapidly. Moreover, after roasting they must be further dried under shadow and dark place until threshing. Italian and Turkish landraces were a little later maturing because they have an alternative growing habit. They performed better for frike color. R112 hom+ (4.66), R5 hom- (4.41), Sarıçanak-98 (3.92) seem to be good green color generating entries for frike making.

Although similar results were pointed out earlier (Yıldırım and Atasoy 2020), hectoliter and thousands kernel weights were quite high as if they were fully mature grains. Data for those two traits revealed that harvest was a little late for frike making. Farmers would prefer a little late harvest and tolerate a little loss for green frike color and marketing price, but they can compensate their income by higher frike yield. Entries for protein (%) and wet gluten (%) ranked similarly. There was a high rank correlation between these two traits. Wet gluten ratios were a little lower than those found by Yıldırım and Atasoy (2020) and for protein ratios. Protein (%) was not so high, which is somehow contrary to findings by Hamit and Omari (1986), Yıldırım and Atasoy (2020), and Akgün et al. (2021). They were a little lower than those of fully mature grains. This revealed that specific taste of frike came from other components such as sugar content raher than high protein percentage (Ozkaya et al. 2018; Bayram 2008; Gambus et al. 2004; Nardi et al. 2003; Maskan 2001).

L,’a’ and ‘b’ values indicated that landraces and DW-WWR crosses were promising entries. Positive L, negative ‘a’ and b values indicate the brightness and green frike color. Late maturing varieties resulted in shiny, relatively green with increasing L and decreasing red (a) and yellow scores (b). L value of Saricanak-98 was quite good for this trait (33.34) confirming the results of Yıldırım and Atasoy (2020). They found Saricanak as the best for L value with 53.50. In our study, Saricanak-98 and Burgos were good for ‘a’ value generating entries and this was confirmed by Yıldırım and Atasoy (2020). But there was no agreement between the present study and their research finding for ‘b’ values.

Although non-significant, raw cellulose (%), lower values were obtained from released cultivars. Raw cellulose tends to be relatively high in landraces and the crosses with the wild relatives. High amount of soluble raw cellulose is considered beneficial for health (Yılmaz and Yıldırım 2020). The highest raw cellulose ratio was 2.7%, contrary to that of Carsanba et al. (2017) indicating a little late harvesting of material for frike making. Yang et al (2012) proved that immature grains include higher amount of raw cellulose than those of mature grains.

Shine green frike color with the smoked flavor is given the highest marketing price. Large and harder grains receive better marketing price offer (Bayram 2008). Presence of carbonized grains, small soil pieces, small stones with same density that of frike, presence of debris, increasing yellow grain color reduce marketing prices sharply. The best purchasing price for shiny green frike was 1.87 US$ kg-1 and sale price was 2.41 US$ kg-1 on 4th of November in the local market. This was three times of bulgur (cracked wheat grains) marketing price confirming previous results (Yılmaz and Yıldırım 2020). Cesare, Levante and Zenit (Bayram 2008) were the preferred varieties for frike making in Sanliurfa local commodity market.

Net returns of entries indicated that neither frike yield alone nor marketing price is good selection criteria for variety preference. Although Tuten-2022 ranked first in frike yield, due to give low marketing price, it ranked as 13th in net return.

Rank-sum method for variety selection for frike making gave equal opportunity to some traits such as frike yield, frike quality, marketing price and net return for each entry under study. Finally, a clear picture was presented to the stakeholders that Artuklu, Firat-93, R5 (Hom-), Saricanak-98 and R 112(Hom+) were the promising varieties with lower sum of ranks for frike making. Furthermore, R5 (Hom-), Firat-93 and Artuklu were quite stable with the least rank standard deviations.

Acknowledgements

We thank to Nevzat Kozak, local farmer and head of the farmer union of Golbasi (Adiyaman), for his valuable advice. Thanks to the Golbası directorate of Ministry of Agriculture and to Murat Olgun of Olgunlar seed company for their help in harvest, drying and threshing. Many thanks go to the Sanlıurfa local community market quality laboratory staff. Finally, thanks to Ali Dagyudan (frike purchaser) for his unvaluable assistance.

Conflict of interest

The authors declare that they have no conflict of interest.

Akgün I, Özbek E, Sener A (2021). The effect of harvest time and nitrogen fertilizer application on grain yield and some quality properties of durum wheat and triticale varieties. -Turkish Journal of Nature and Science, Volume 10(2):53-63
Bayram M (2008). An analysis of scorched immature wheat. Cereal Foods World, May-June2008, Vol.53(3):134-138
Carsanba E, Akca I, Timur M ( 2017). Examination of firik produced in Hatay region in terms of nutritional aspect. Gıda, 42(6), 726-730
Gambus H, Gümül D, Juszczak L (2004). Rheological properties of pastes obtained from starches derived from ımmature cereal kernels. Starch, 56: 225-231
Hamid MA, Omari MAA (1986). Study of automatization of frekeh industry in Jordan. Dirasat, 13:31-50
Humeid MA, Takruri HR, Umari MAH (1990). Amino acid content of durum wheat protein at different maturity levels of the grains suitable for frekeh making. Dirasat. 3: 54-61
Huynh BL, Palmer L, Mather DE, Wallwork H, Graham RD, Welch RM, Stangoulis JCR (2008). Genotypic variation in wheat grain fructan content revealed by a simplified HPLC method. Journal of Cereal Science, 48(2): 369-378
Maskan M (2001). Effect of maturation and processing on water uptake characteristics of wheat. Journal of Food Engineering, 47: 51-57
Mulamba N, Mock J (1978). Improvement of yield potential of the Eto Blanco maize (Zea may L.) population by breeding for plant traits. Egyptian Journal of Genetics and Cytology 7: 40-51
Nardi ., Calcagno C, Zunin P, D'egidio MG, Cecchini C, Boggia R (2003). Nutritional benefits of developing cereals for functional foods. Cereal Research Communications, 31: 445-452
Nesbitt M, Samuel D(1996). Promoting the conservation and use of underutilized and neglected crops. Pages 41–100 in: Hulled Wheats. Padulosi, S., Hammer, K., Heller, J.eds. IPGRI Press, Rome, Italy
Özkaya B, Türksoy S, Özkaya H, Baumgartner B, Özkeser İ, Köksel H (2018). Changes in the functional constituents and phytic acid contents of firiks produced from wheats at different maturation stages. Food Chemistry, 246: 150-155
Sramkova Z, Gregova E, Sturdik E (2009). Chemical composition and nutritional quality of wheat grain. Acta Chimica Slovaca, 2(1): 115-138
Uluöz M (1965). Wheat flour and bread Analysis. EUZF Pub. No:57, Izmir, Turkey,95 pp
Usdurum (2015). http://www.usdurum.com/www.usdurum.com/ancient-grains-flours/freekeh/index.html
Vallega V (1996). The quality of Triticum monococcum L. in perspective. Pages 129–146 in: Hulled Wheats. Padulosi, S., Hammer, K., Heller, J. eds. IPGRI Press, Rome, Italy
Yang D, Shin JA, Gan LJ, Zhu XM, Hong ST, Sung CK (2012). Comparison of nutritional compounds in premature green and mature yellow whole wheat in Korea. Cereal Chemistry, 89(6): 284-289
Yerasimos M (2011). Evliya Çelebi seyahatnamesinde yemek kültürü. Pp:127-129 Kitap Yayınevi, Istanbul
Yıldırım A, Atasoy A F (2020). Quality characteristics of some durum wheat varieties grown in southeastern Anatolia region of Turkey. Harran Gıda ve Tarım Bilimleri Dergisi 20(4):420-431
Yılmaz MS, Yıldırım A (2020). Production techniques and functional properties of firik. HRU.Muh.Der.5(2):109-121

An assessment on durum wheat cultivars and lines suitable for frike making in southeast Turkey

Status:

Version 1

Abstract

Figures

Introduction

Material and methods