*2.9. Sensory Analysis*

The sensory evaluation of edible flowers was performed by 30 panellists (trained students thanks to the course Sensory analysis and trained academic staff). They were acquainted with the monitored materials and instructed on the principles of analysis in advance. The course of sensory evaluation and equipment of the room for sensory analysis met precisely defined conditions according to the international standard ISO 6658. The sensor room at Tomas Bata University in Zlín was equipped with 12 separate evaluation boxes, placed next to each other and modified to limit contact with other evaluators. The room temperature was approximately 21 ◦C and was lit by artificial lighting. The assessment took place at 10:00 am, approximately 1 h (6 samples). It was recommended to take a break of approximately two minutes between the evaluations of the individual samples. Individual samples (each 3 flowers from one species) were administered in order *Rosa, Hemerocalis*, *Calendula officinalis*, *Begonia* × *tuberhybrida*, *Tagetes patula* and *Tropaeolum majus*. Before tasting, the samples were stored in reusable plastic containers at 7 ◦C for 12 h. Between individual samples, participants could neutralise the taste in their mouths with common water and white bread. The following sensory attributes were evaluated: appearance, fragrance, consistency, acidity, bitterness, astringency, sweetness, spiciness, overall taste, juiciness, and overall flower evaluation. The panellists assessed each blossom attribute using a 9-point hedonic scale; 1 = dislike extremely, 2, 3, 4 = subjective sense of dislike (very much/moderately/slightly), 5 = neutral, 6, 7, 8 = like slightly/moderately/very much, 9 = like extremely for overall taste and overall evaluation. They also determined the perceived intensity of each taste (acidity, bitterness, astringency, sweetness, and spiciness); 1 = very strong, maximum, 5 = slightly, moderate, 9 = without the taste. The scales for the remaining descriptors were as follows: 1 = unacceptable, 5 = neutral, 9 = ideal for appearance (suitability for food decoration); 1 = very intense and unpleasant, 5 = odourless, 9 = very intense pleasant for fragrance; 1 = very stiff, 5 = ideally crispy, 9 = flowable for consistency; 1 = dry, 5 = moderately juicy, 9 = watery for juiciness. The results were expressed graphically as the mean values of all ratings for each component and the overall score.

## *2.10. Statistical Analysis*

Microsoft Office-Excel 2013 (Microsoft Corporation, Redmond, WA, USA) and STA-TISTICA CZ version 12 (StatSoft, Inc., Tulsa, OK, USA) were used for data analysis. The results were expressed by mean ± standard deviation (M ± SD). To establish statistically significant differences between individual species, Shapiro-Wilk test of normality and Levene's test of homogeneity of variances was performed. Since the conditions for the calculation by ANOVA analysis were not complied in any of the monitored data sets, a non-parametric Kruskal-Walllis test (α = 0.05) were performed. Correlation functions were calculated using statistic software Unistat 5.1 (Unistat Ltd., London, UK) and Microsoft Office-Excel 2010 (Microsoft Corporation, Redmond, WA, USA).

#### **3. Results and Discussion**

The results of the chemical analyses are shown in Tables 3–8. The results were expressed as an average of a two-year experiment (2018–2019) since there was no statistically significant difference between the years in any parameter researched.

**Table 3.** Total phenolic content (g of GAE/kg of FM), total antioxidant capacity (g of AAE/kg of FM–DPPH or as g GAE/kg of FM-FRAP) and total flavonoid content (g of RE/kg of FM) in 6 species of edible flowers.


Note: All values are expressed as the mean ± standard deviation (SD) (*n* = 10). Values in a column that do not share the same superscript letters (a,b,c,d) are significantly different at *p* < 0.05. TPC: total phenolic content; TFC: total flavonoid content; TAC: total antioxidant capacity; GAE: gallic acid equivalents; FM: fresh mass; RE: rutin equivalents.


Note: All values are expressed as the mean ± SD (*n* = 10). Values in the column that do not share the same superscript letters (a,b,c,d,e) indicate significant differences at *p* < 0.05. The content of macroelements is expressed as mg/kg of FM.



Note: All values are expressed as the mean ± SD (*n* = 10). Values in the column that do not share the same superscript letters (a,b,c,d,e) are significantly different at *p* < 0.05. The content of microelements is expressed as mg/kg of FM.

**Table 6.** Dry matter and the content of crude protein in 6 species of edible flowers.


Note: All values are expressed as the mean ± SD (*n* = 10). Values in a column that do not share the same superscript letters (a,b,c,d) are significantly different at *p* < 0.05. The content of dry matter is expressed as% *<sup>w</sup>*/*<sup>w</sup>*, and the content of crude protein is expressed as g/kg of FM.


**Table 7.** Correlation analysis between TPC, TAC, TFC and phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), natrium (Na), iron (Fe), manganese (Mn), copper (Cu), zinc (Zn) and molybdenum (Mo) of six edible flowers grown for two years.

Note: Mean values were used in the analyses of chemical parameters at levels of statistical significance (\* *p* < 0.05; \*\* *p* < 0.01).

**Table 8.** Selected correlation coefficients between TAC-DPPH and mineral content, TPC and TFC.


*3.1. Total Content of Phenolic Substances, Total Antioxidant Capacity and Total Flavonoid Content*

Table 3 provides TPC values of six edible flowers. These TPC values varied from 3.23 g GAE/kg in *T. majus* to 6.59 g GAE/kg in *Hemerocallis*, with almost double the difference. The results of *Hemerocallis* showed up to six times higher TPC than in some cultivars of daylilies presented by Muchová [48].

The TPC values for tested pink *Begonia* × *tuberhybrida* were 4.82 g GAE/kg of FM. They were similar for the white cultivar of Begonia (*B. semperflorens* Link et Otto) [49] and double higher when compared to another pink cultivar, 'Chanson' [48].

TPC in *T. patula* (French marigolds) was 4.78 g GAE/kg of FM, and this content is slightly higher than that reported by Rop et al. [14] but lower than in different orange cultivars of French marigold flowers [50].

According to Ashraf et al., for *C. officinalis*, the TPC values were slightly lower (by 0.83 g) than the values measured by us (3.65 g GAE/kg FM), which could be caused by the fact that other parts of the plants (leaves) were used or different growing conditions [51]. Ferreira et al. quantified TPC values of 2.30 g GAE/kg DW in calendula flowers [52]. This research found an aqueous methanol solution (70:30, Me-OH:H2O) more effective for extracting phenolic compounds than pure methanol; the values were probably higher because the solvent was highly polar and had the highest dielectric constant [52].

The content of total phenolic substances in *Rosa* was determined to be 4.45 g GAE/kg FM, which is approximately 1.12 times lower than that measured by Rop et al. [14]. Yang and Shin found the difference between TPC in ethanolic samples of roses, and their values ranged from 7.99 to 29.79 g/kg FM [1]. They also reported that yellow and pink roses had lower TPC than red roses. Despite different flower varieties and conditions of extraction, these values are lower than those reported by Gonçalves et al., where the orange rose cultivar had a slightly higher TPC content (17.60 g GAE/kg FM) than *Tagetes erecta L.* (17.47 g GAE/kg FM) [11]. The considerable variation could indicate that the composition of substances differs within the genus.

As mentioned above, the red cultivar of *T. majus* exhibited the lowest TPC value (3.23), which is significantly lower than the value for the orange cultivar exhibited by other authors [29,53]. Nevertheless, our TPC results are comparable to those reported by Rop et al. [14] and higher than those reported by Muchová [48]. The differences could be caused by using a different variety, growing conditions, the flower's measured parts, or its adjustment before analysis.

The total phenolic range level is comparable with some berries *Vaccinium L. hybrids* and *Rubus L. species* [54,55], both of which are considered to be grea<sup>t</sup> sources of antioxidants [29]. This content is higher when compared to different vegetables like Ceylon spinach, white and red onions [56], lamb's lettuce [57], or other green leafy vegetables [58]. Flowers may be ideal for making salads more appealing to consumers, either adding colour or increasing the phenolic content of the food.

In this study, total flavonoids were another parameter studied in flowers, as seen in Table 3. Flavonoids are likely the most significant natural phenolics, and also they are one of the most varied and widespread natural chemicals [59]. The TFC in the flowers ranged from 1.29 to 3.76 g RE/kg FM. The highest TFC was determined in *Hemerocallis* (daylily), and the lowest was observed in the red variety of *T. majus*. For other flowers, the TFC was above 2 g of RE/kg of FM. In the Rosa' Gloria Dei', the total flavonoid contents were 2.43 g of RE/kg of FM, which is slightly higher than that observed in Rosa odorata [14]. According to a study by Yang and Shin, the TFC of edible roses ranged between 0.79 to 5.32 g/kg FM; our analysed cultivar is in this range [1]. The flowers reached significantly higher TFC values than some vegetables and fruit, such as tomatoes (0.133 to 0.474 g RE/kg FM) [60], watermelons (0.09 to 0.27 g RE/kg FM) [61], or fruit studied by Mirzaei et al., such as blackberry and black grape, whose TFC values ranged from 0.05 to 1.03 g RE/kg FM [62]. According to studies on 12 cultivars of EFs, the yellow blossoms had a higher content of flavonoids and suggested that they have a stronger antioxidant potential than other colours [59]. This correlates with our results because the flowers with the highest TFC content were *Hemerocallis* and *Rosa* with yellow colour petals. However, Garzón et al. analysed the *T. majus* flowers depending on their colour. The yellow cultivar had lower TPC values and antioxidant activity than the orange and red cultivars due to the low content of primary anthocyanins [29].

Further research could involve using high-performance liquid chromatography (HPLC) to identify and accurately quantify phenolic compounds in the sample. In the case of including HPLC analysis in this study, its length and complexity would exceed the proposed research framework.

The antioxidant potentials of flower extracts were estimated using two different colourimetric assays in vitro based on electron-transfer reactions. The first DPPH method was measured antioxidants' capacity to scavenge an organic radical; the results were expressed as ascorbic acid equivalents. The second FRAP method was estimated antioxidants' ability to reduce ferric to the ferrous ion, and findings were reported as reducing power per gallic acid equivalent. Combining these two approaches based on distinct mechanisms may provide more reliable and complex data for antioxidant capacity. A single spectrophotometric assay may not provide satisfactory results because of its deficiency of specificity and sensitivity [63]. Both methods are widely used because of their simplicity, speed, high reproducibility, and ability to measuring by simple instrumentation [41,64,65]. Each of them has some advantages and limitations. For example, the DPPH assay can detect weak antioxidants and thermally unstable compounds; however, DPPH might react with other radicals in the sample and is sensitive to light [43,64,65]. The FRAP method result may not positively correlate with the total antioxidant activity of the sample; because this assay is non-specific and has limitations in measuring slow-reacting polyphenolic antioxidants and thiols [43,66].

The total antioxidant capacity of samples ranged from 4.11 g AAE/kg FM in *C. officinalis* to 7.94 g AAE/kg FM in daylilies. TCA values above 6 g AAE/kg FM were measured in *T. patula*, *Rosa*, and *T. majus*. The *Hemerocallis* (daylily) achieved a higher antioxidant capacity than the edible flowers in the Rop et al. study; the TAC of edible flowers ranged from 4.21 to 6.96 g AAE/kg FM [14]. In addition, the strong antioxidant activities of daylilies extracts (aqueous and ethanolic) were observed by Que et al. [35]. These results exhibited lower AA than synthetic antioxidant (butylated hydroxyanisole) but higher than α-tocopherol. According to Fu et al., the highest antioxidant capacity and the highest proportion of phenolic substances is in the opening stage of daylilies [17]. Mao et al. found that the use of lyophilised daylily flowers is more suitable for obtaining an extract with higher AA and a higher TPC than the use of flowers dried with hot air [67]. The limitation of using daylily is that each flower only lasts one day. The flower extracts from *T. majus* are active, reducing agents, which indicates a good ability to scavenge radicals [68]. According to Pavithra et al., the methanol extracts of flowers have scavenging abilities dependent on their concentration (25 mg/mL and higher) but lesser than ascorbic acid [69].

Comparing the results obtained from TAC-FRAP with TAC-DPPH, it is evident that the extracts' ability to reduce Fe3+ has a different order than the ability to quench the DPPH• radical. Additionally, the values obtained by FRAP assay show that the highest antioxidant capacity corresponded to *T. patula* (5.62 g GAE/kg FM), followed by *Begonia* × *tuberhybrida* (5.15 g GAE/kg FM) and *T. majus* (4.98 g GAE/kg FM). The lowest AC was that of *C. officinalis* (3.44 g GAE/kg FM). The FRAP values displayed a 1.6-fold difference.

People generally do not consume as many edible flowers as carrots, radishes, cucumbers, tomatoes, and other vegetables. Because some edible flowers have a pungen<sup>t</sup> or strong aroma, it is advisable to use them sparingly to encourage food flavour [70]. The ornamental EFs evaluated in the study were non-toxic; nevertheless, it should also be considered that the daily limit for their ingestion is not determined for all samples, and no international authority has published the official list of EFs [12,71,72]. Consumption and culinary use of some EFs were documented in history before May 1997; consequently, these flowers are not defined as novel foods [73,74]. For example, none of the species analysed in our research was featured on official lists like the Novel Food Catalogue [72]. On the other hand, other blossoms that cannot prove their significantly large consumption by people before 15 May 1997 must be submitted to the European Food Safety Authority for their safety application as novel foods [73,75].

Lucarini et al. [75] examined available information in databases and bibliographies about the same plant genera as our study, and they discovered no scientific proof that these plants constitute potential allergens.

Even the most favourable herbs can have unpredictable effects, for example, the consumption of more than 39.5 g of fresh *T. majus* flowers surpassing the daily erucic acid allowance [70,76]. The number of blossoms consumed may be the limiting factor because of allergic and toxic reactions by sensitive persons to some of the flower unidentified compounds [14]. In addition, pollen from specific blossoms might induce an allergic response in humans with allergies or asthma [75]. Thus, it is important to study the toxicity of EFs with high antioxidant activity to establish their safety as food additives. Moreover, identification of the plant is critical because some toxic flowers could be readily confused with EFs, such as daylilies with true lilies, and confusing them might be dangerous [77].

#### *3.2. The Content of Mineral Elements*

Five macroelements (P, K, Na, Ca, Mg) and five microelements (Fe, Mn, Cu, Zn, Mo) were determined and quantified in the petals of diverse species of ornamental edible plants. These mineral elements are essential for the human body's vital functioning, but the available literature contains scant data about their content in EFs. The contents of minerals, expressed on a FM basis, are shown in Tables 4 and 5, and were in this order: K > Ca > P > Mg > Na > Zn > Mn > Fe > Cu > Mo. The macroelements amount ranged from 121 to

3623 mg/kg FM (Table 4), and the content of microelements then from 0.98 to 14.91 mg/kg FM (Table 5).

*Hemerocallis* had the greatest Ca, Mg, Fe, Mn, Cu, and Mo concentrations, whereas *T. patula* contained the highest amount of K, Na and Zn, and the highest P was detected in *T. majus*. In contrast, the lowest P, K, Mg, Na and Fe content was observed in *Begonia* × *tuberhybrida*; *Rosa* had the least amount of Ca, Mn and Zn, the lowest quantity of Cu and Mo was found in *T. majus*.

The content of mineral elements is comparable to the flower mineral concentration mentioned by Rop et al. [14]. When compared to ordinary fruit and vegetables, EFs are a good source of minerals. This is evidenced by the higher K content than vegetables and fruit, which have an average K content of 1500–2100 mg/kg FM (Table 4) [78–80]. Several researchers observed a similar trend in which potassium content was highest in flowers [14,81–83]. Potassium content in flowers was higher than in leaves, roots, and stem of *Chrysanthemum indicum* L. [82]. According to Grzeszczuk et al., in other Hemerocallis species, the most abundant macroelement was K, which correlates with our results, but P content was higher than that of Mg [83]. However, Navarro-González et al. reported that *T. majus* and *Tagetes erecta* blossoms contain more zinc, iron, and manganese than potassium [53]. Flowers (100 g fresh weight) provided only 10.0–18.1% of the daily recommended K intake of 2000 mg for adults [47]. Potassium content is an important source for maintaining acid-base balance in blood and tissues and preventing cardiovascular or oncogenic diseases [84].

The content of other elements in flowers is comparable to vegetables [80], but some selected leafy vegetables had a higher content of sodium than potassium [58]. Compared to fruit, a two-fold increase in Ca and Mg contents and a fourfold rise in Na content can be observed [78,85,86]. In addition, the content of mineral elements in flowers can be compared with published minerals data about wild-growing and cultivated mushrooms. Calcium and sodium contents are two to four times higher than that of fungi, the content of other elements is approximately comparable, but the phosphorus one is twice lower [87,88].

Mineral elements perform several functions: as components of enzymes, regulation of cellular energy transduction, gas transport, antioxidant defense, membrane receptor functions, second-messenger systems, and integration of several physiological functions [89–91]. Furthermore, they can strengthen the immune system [92,93], form building blocks of the human skeleton [91,94] and are associated with anti-inflammatory [24,95], antibacterial [93,96], antifungal [97] and antiviral effects [98]. A few published research papers deal with the content of mineral elements in EFs regarding their relevance for human health [14,83,99,100].

Previous research has shown that iron concentrations in ornamental flowers are highly varied, compared to our results, for example, *Begonia boliviensis* with lower content of 2.65 mg/kg FM [14], *T. majus* with slightly lower content from 5.51 to 6.47 mg/kg FM [14,53], and *T. erecta* with slightly higher amount of 10.26 mg/kg FM [53]. Different species probably caused variations in mineral elements content between the flower samples because they were grown in the same location and with identical agricultural practices.

All analysed flower species have high molybdenum levels based on recommended daily intakes for adults (50 μg) since 100 g of fresh blossoms provides 64–196% of this value [47]. The concentration of Mo affects ascorbic acid level; for example, its deficiency can cause a decrease in AA content in some vegetables [101]. Tolerable upper intake level of Mo range from 0.1 to 0.6 mg/day [102]; therefore, consuming a slight amount of flowers is unlikely to be a risk for human health. *Hemerocallis* can be considered as a possible source of Cu (0.29 mg/100 g FM), Mn (0.88 mg/100 g FM), and Zn (1.15 mg/100 g FM), and these mineral elements can contribute to daily recommended dietary allowances for adults. For example, 100 g fresh *Hemerocallis* can provide 29.3% copper, 43.8% manganese and 11.5% zinc for dietary reference intakes [47]. EFs should not be overlooked as a source of mineral elements in the human diet; however, it is unlikely that somebody would eat

100 g of flowers in a single day. Edible flowers will most likely be used as a garnish to add colour and flavour to the food.

#### *3.3. Dry Matter and Content of Crude Protein*

The dry matter and the content of crude protein of the edible flowers are shown in Table 6. As can be seen from the results, the DM of these edible flowers ranged from 7.38 to 14.39%, and this amount is slightly higher than the average content in fruit and vegetables [103]. On the other hand, according to Montañés Millán et al. [104], the DM percentage in blossoms from the fruit tree was higher. When comparing our DM results to previous research for the same plant genus, *Begonia nelumbiifolia* ranged from 5.31 to 6.15%, which is lower than Begonias results in our experiment [105]. However, Rop et al. determined *Begonia boliviensis* (14.20%) with a higher DM [14]. In addition to the lastmentioned research, they determined a higher DM for *T. patula* (9.68%) and *T. majus* (11.27%) and lower DM for *Rosa odorata* lower DM (10.09%) [14]. De Lima Franzen et al. observed a higher DM percentage for rose (*Rosa* × *grandiflora*) and *C. officinalis* of 15.44% and 10.66%, respectively [106].

The CP content of EFs samples was estimated by the Kjeldahl method, and the results ranged between 2.89 to 4.56 g/kg of FM (Table 6).

The highest values were reached for *T. majus* (4.56) and *Begonia* × *tuberhybrida* (4.51). Comparing these CP values with results obtained by Rop et al. [14], *T. majus* had slightly lower values (4.74 g/kg FM), and Begonia had one and a half times higher than another cultivar. However, the CP contents of *T. majus* and other varieties of Begonia and roses cultivated in Japan were significantly higher than in our research [9]. The difference may be caused by different cultivars, place and growth conditions. A similar CP was observed in EF *Fernaldia pandurate* with 3.0 g/kg FM [107]. This protein content can be comparable to some fruit and vegetables but not to cereals because they have an order of magnitude higher content [79]. Similar proportion content was observed in some fruit, for example, plum with 3.9 g/kg FM [108] or red banana (*Musa acuminata*) [109]. The CP content in fresh vegetables was higher than in our experiment, for example, radishes with 5 to 15.5 g/kg, beetroot with 13.22–14.43 g/kg [110], celery with 6.91 g/kg, carrot with 5.64 g/kg, and turnip with 4.88 g/kg [108]. Thus, flower petals could not be regarded as good protein sources because of their low CP levels [106]; also, people consume fewer EFs than radishes, carrots and other popular types of vegetables.

#### *3.4. Correlation Analysis between Mineral Elements and Bioactive Compounds*

The correlation coefficients of mineral elements and bioactive compounds in edible flowers are shown in Tables 7 and 8. Significantly strong positive correlations were observed between some mineral elements contents; for example, the correlation of Na-K (*r* = 0.92 \*\*), Zn-K (*r* = 0.96 \*\*), Zn-Na (*r* = 0.92 \*\*) and Zn-Fe (*r* = 0.83 \*\*). Furthermore, considerable high positive correlations between TFC-Mo (*r* = 0.93 \*\*) and TFC-Cu (*r* = 0.81 \*) were found. From a different point of view, negative relationships were noticed between the contents of M and P (*r* = −0.88 \*\*), between TFC and P (*r* = −0.69 \*), and also between Cu and P (*r* = −0.59 \*).

Table 8 shows the selected correlation coefficients between TAC-DPPH and mineral elements, TPC or TFC. These relationships are studied to assess if these components contribute to the TAC of the flowers and if they have any potential benefits for human metabolism.

In accordance with some research studies [111–114], significant correlations between TAC, TPC and TFC were commonly achieved in our results as well, from *r* = 0.57 \* to 0.94 \*\*. The results imply that blossoms with a higher amount of polyphenols have a stronger antioxidant activity, and flavonoids comprise an important group of phenolic compounds. Some authors also found a strong positive correlation between TPC and FRAP assay values [43,71,115]. The antioxidant activity could be attributed to some mineral elements like copper, iron and manganese [116]. In our case, AC correlates with the Ca (*r* = 0.68 \*) and Fe (*r* = 0.61 \*), which means a moderate positive correlation; some authors also described these relationships [117,118]. Their articles state the importance of nutrition by given elements on the content of bioactive substances. On the other hand, the correlations between TAC and the remaining mineral elements were weak or negligible. Aside from polyphenols, the antioxidant activity of floral extracts may be affected by other biological compounds, including vitamins, pigments such as carotenoids, mineral elements, nitrogenous compounds, and other metabolites. [17,119–121].
