Abstract
Violets (Viola) are potential candidates for aroma-focused breeding research. Though most Viola species and modern hybrids lack fragrance, the genus contains a famously scented species, Viola odorata L. This species and its cultivars are genetic resources of aroma traits that could be used to investigate the selection for and transmission of fragrance during the breeding process. Despite its famous scent, however, the floral volatile emissions of V. odorata have not been characterized using modern headspace techniques. Using static and dynamic headspace volatile collection methods and gas chromatography–mass spectrometry, the floral volatile emissions of V. odorata were temporally and developmentally characterized. Floral volatiles were also sampled from 10 V. odorata cultivars, three Parma violet cultivars, five violet species, and one hybrid, and variation in scent among these violets was investigated. Total volatile emissions in V. odorata were highest from 0600 HR to 1900 HR, suggesting a diurnal pattern of emission. Volatile emissions also varied over the developmental lifespan of the flower, with the highest emission of individual and total volatiles occurring, in general, from stages 0 or +1 to stages +3 or +4. Floral scent qualitatively and quantitatively differed among assorted violets. The floral volatile emissions of V. odorata exhibit temporal and developmental variation. Compared with the other violet species in this study, sweet violets are intensely fragrant. The quantity and quality of floral scent differs among V. odorata cultivars, providing genetic variation from which selections could be made in a fragrance-focused breeding program.
Floral fragrance arises from mixtures of volatile organic compounds (volatiles or VOCs), primarily lipophilic, low molecular weight (<300 amu) compounds with high vapor pressures at ambient temperatures (Dudareva et al. 2013). Plants can also emit volatiles from their roots and vegetative organs; however, the most abundant diversity of volatiles is found in floral scent (Knudsen et al. 2006). These volatiles serve numerous functions, from attracting pollinators, acting as signaling molecules, mediating plant-plant interactions, and acting as defense compounds against predators or pathogens, among other functions (Dudareva and Pichersky 2000; Dudareva et al. 2006). Floral volatiles also serve as a sensory attractant to humans. Indeed, fragrant flowers have intrigued and enchanted humans throughout history and across cultures, inspiring verses and songs of praise by the likes of Shakespeare and Keats. More recently, a study on consumer preferences for the essential features of a flower product identified fragrance as one of the most important drivers of consumer liking (Levin et al. 2012). A lack of scent negatively affected overall consumer interest the most out of all the flower characteristics evaluated, indicating that consumers prefer fragrant flowers. But in the world of ornamental plant breeding, little to no selection for scent traits occurs; rather, priority is placed on the selection of traits like disease resistance, growth habit, and visual aesthetic characteristics such as flower form, size, and coloration (Pichersky and Dudareva 2007). The importance of fragrance to consumers and the desire to develop unique, value-added cultivars has spurred some interest in the selection of aroma traits in ornamental breeding programs; however, limited research has been conducted on breeding for fragrance in ornamentals.
Violets (Viola) are potential candidates for fragrance-focused breeding research. Containing more than 650 species, Viola is the largest genus in the Violaceae family (Marcussen et al. 2022; Yockteng et al. 2003). Hybridization is a common occurrence throughout the genus, and multiple instances of interspecific hybridization resulting in fertile hybrids has been documented (Ballard et al. 1999; Erben 1996; Ko et al. 1998; Marcussen and Borgen 2000; Neuffer et al. 1999; Stebbins et al. 1963). Although most Viola species and commercial hybrids—including the immensely popular garden pansies—are largely devoid of fragrance, the genus contains a famously scented species, Viola odorata L., also known as the sweet violet. The sweet violet is an acaulescent, stoloniferous, myrmecochorous, and perennial species that is native to southern Europe, northwestern Africa, and western Asia (Saint-Lary et al. 2014). Like other members of the Violaceae family, V. odorata is a cleistogamous species. Cleistogamy (CL) describes the formation of closed, self-pollinating flowers; it is the opposite of chasmogamy (CH), the formation of open flowers which may be self- or cross-pollinated (Culley and Klooster 2007). V. odorata exhibits true or dimorphic cleistogamy, meaning it produces both CL and CH flowers, and that the CL and CH flowers are morphologically different (Baskin and Baskin 2014). In sweet violets, the production of CL or CH flowers is controlled by photoperiod: CH flowers are produced under 11 h or less of daylight, CL flowers are produced under 14 h or more of daylight, and transitional flowers that are morphologically intermediate between CH and CL flowers are formed under 11 to 14 h of daylight (Mayers and Lord 1983a).
Sweet violets have been cultivated and used for a variety of purposes since ancient times (Marcussen 2006). They have an exceptionally extensive history of medicinal use, appearing in influential herbals and medical texts from ancient Greece and Rome, the Middle East, and Europe, and they are still used in traditional Iranian and Ayurvedic systems of medicine today (Lim 2014). Moreover, sweet violets have long been known and loved for their unique, characteristic fragrance. Indeed, Shakespeare says of the violet in the play King John: “To gild refined gold, to paint the lily,/To throw a perfume on the violet…/Is wasteful and ridiculous excess” (Shakespeare 1954). The sweet violet’s intense and distinctive fragrance made it a choice ingredient for the production of perfumes, ointments, and scented oils for more than a thousand years. During the 18th and 19th centuries, light floral fragrances—particularly rose, violet, and lavender—were “in with a vengeance” (Pybus and Sell 1999). However, only the most luxurious, high-end, and expensive violet perfumes contained true essence of violets, as violet flower essential oil was difficult to obtain and exorbitantly expensive to make (Maxwell 2017). Because of this high cost, efforts were made by the burgeoning synthetic perfume industry in the late 19th century to find a synthetic violet odorant, efforts that were ultimately successful in 1893 (Tiemann and Krüger 1893). After their discovery, the use of chemically synthesized violet odorants, especially α-ionone and β-ionone, exploded (Gautschi et al. 2001). These synthetic ionones decreased the price of the violet smell by a factor of a million, and it became one of the most popular smells of all by the turn of the century (Turin and Sanchez 2008). Because of their ability to blend well with virtually all other perfumery materials, the synthetic ionones “revolutionized perfumery” and are present in the vast majority of perfumes today (Gautschi et al. 2001). They are widely used in the fragrance of personal care products such as lotions, shampoos, and deodorants, and they are widely used as flavoring substances in the food industry (Lalko et al. 2007a, 2007b). Despite the sweet violet’s illustrious history and storied fragrance, however, the chemical composition of the fragrance actually emitted from the intact flowers has not been fully characterized using modern headspace techniques.
Sampling volatiles from the “headspace,” the air surrounding a plant or plant part, has become the preferred method for the analysis of plant volatiles, as it is a nondestructive form of sampling that gives a more realistic picture of the volatile profile emitted by a plant and experienced by a human or animal (Raguso and Pellmyr 1998; Tholl et al. 2006). Traditionally, volatiles were collected from plants via solvent extraction or steam distillation; however, these destructive methods are prone to produce artifacts, the latter mainly by degradation or rearrangement of heat-labile compounds, and the former mainly through the extraction of nonvolatile material from tissues (Knudsen et al. 1993). Consequently, the volatile profile of an essential oil or other concentrated plant extract may differ substantially from the volatile profile experienced by a person smelling the flower from which the oil or extract came (Tholl et al. 2006).
Over the past 50 years, the chemical composition of essential oils and various concentrated extracts made from V. odorata were analyzed, including essential oils made from the flowers (Uhde and Ohloff 1972), “aerial parts” (Hammami et al. 2011), unspecified plant parts (Abdel-Rahim 2016), and leaves (Akhbari et al. 2012), as well as “absolutes” isolated from leaves (Saint-Lary et al. 2014), and an oil extracted from a “concrete” made from V. odorata leaves and flowers (Mohamed and Ghatas 2016). Aalthough essential oils are obtained through various distillation methods and are primarily composed of volatile compounds, concretes and absolutes are obtained through solvent extraction and contain nonvolatile compounds, too (Saint-Lary et al. 2016). Gautschi et al. (2001) mentioned a headspace analysis of violets in bloom, stating that the headspace contained “35.7% of α-ionones, 21.1% of β-ionone, and 18.2% of dihydro-β-ionone”; however, the citation for these percentages is a personal communication, and no details are provided about the violets studied or the methodology used to collect or analyze the volatiles. To date, the floral volatile profile of sweet violets has not been fully characterized using modern headspace techniques.
Thus, the specific objectives of this research were to 1) characterize the volatile emissions of V. odorata vegetative tissues (leaves, petioles, stolons) so that these vegetative volatiles could be subtracted from the floral volatile profile where necessary; 2) temporally characterize the floral volatile emissions of V. odorata to determine if floral volatiles are released in a time-dependent or rhythmic manner and identify at what time(s) peak emission occurs; 3) developmentally characterize the floral volatile emissions of V. odorata over the life span of the flower (i.e., from the day before anthesis to perianth wilting as the first sign of senescence) and determine on what day(s) emissions are at their peak; and 4) comparatively evaluate the floral volatile profiles of V. odorata, an assortment of sweet violet cultivars, three Parma violet cultivars, and a small selection of other Viola hybrids and species—some of which may eventually be candidates for fragrance-focused breeding—to gain a basic understanding of what compounds are present in those profiles. Taken together, these objectives provide modern chemical insight into a fragrance that has delighted untold numbers of people for more than 1000 years, and the comprehensive characterization of V. odorata volatile emissions contributes to the body of research on floral fragrance more broadly. In addition, this volatile survey establishes baseline fragrance profiles for violets that, in the future, could be chosen to make interspecific crosses. If these crosses are successful and one or more hybrids are produced, the fragrance profiles of the progeny could be characterized and compared with those of the parents, potentially providing insight into the transmission of volatile traits, as well as the possible consequences or tradeoffs of prioritizing such traits during the breeding process. Finally, this research builds a foundation for a fragrance-focused breeding program that has the potential to develop more fragrant flowers and ornamentals.
Materials and Methods
Plant material and growth conditions.
Violets, including the species V. odorata, assorted sweet violet cultivars, Parma violet [Viola alba subsp. dehnharditii (Ten.) W. Becker] cultivars, and other Viola were purchased as plants or seeds from one or more vendors (Supplemental Appendix Tables A-1 and A-2). The species Viola walteri House and Viola sororia Willd. were collected as plants with permission from an acquaintance’s yard in Gainesville, FL (lat. 29°38′60.0″N, long. −82°22′12.0″W). Violets obtained as plants were propagated by division and/or by taking cuttings from well-developed stolons. The sweet violet cultivar Empress Augusta was further propagated by seed harvested from cleistogamous seed pods. All violets were grown in a temperature-controlled air-conditioned house (minimum 18 °C, maximum 27 °C) covered with two layers of 30% density shadecloth. Violets were grown in standard 1-gallon nursery pots containing a high-porosity peat-based media (PRO-MIX HP; Premier Horticulture Ltd., Rivière-du-Loup, Quebec, Canada) and fertigated as needed via drip tubes with 150 ppm N of 15–5–15 fertilizer (Peters Excel 15–5–15 Cal Mag Special; ICL Specialty Fertilizers, Summerville, SC, USA).
Dynamic push-pull headspace volatile collection system.
Except for the temporal volatile collections, all volatiles were collected using a “push-pull” dynamic headspace collection system. Floral or vegetative tissue was excised and quickly loaded into glass tubes (2.5-cm internal diameter, 61-cm long, 300-mL volume). The glass tubes were connected to a dynamic push-pull headspace collection system. Purified air was pushed into one end of the tube and pulled via a vacuum pump over the plant material and through a glass column connected to the outlet end of the tube. Volatiles were captured within the glass column, which contained ∼50 mg of a porous polymer adsorbent (HaySep Q80-100; Hayes Separations Inc., Bandera, TX, USA). Collections occurred for 2 h at room temperature. Volatiles were eluted from the adsorbent polymer with 150 µL of methylene chloride [Chemical Abstracts Service (CAS) 75-09-2; Thermo Fisher Scientific, Waltham, MA, USA] spiked with ∼28.3 ng/µL of nonyl acetate (CAS 143-13-5; Tokyo Chemical Industry, Portland, OR, USA) as an elution standard. Samples were stored at –80 °C until analysis by gas chromatography–mass spectrometry (GC-MS).
Static headspace volatile collection system.
Analyzing temporal variation of scent emissions over the course of a day required a different headspace collection system than the one described previously, as violet flowers begin to wilt within 1 to 2 h after detachment, and the dynamic collection system was not equipped to sample volatiles from intact plant material. Instead, a static headspace collection system and method based on Balao et al. (2011), Dötterl et al. (2012), Keene et al. (2020), and Steenhuisen et al. (2010) were used. A single V. odorata flower was inserted into a gusseted, biaxially oriented polypropylene bag (5.08 × 3.81 × 12.7 cm; Restaurantware, Chicago, IL, USA). Following 1 h of volatile enrichment, a small cut was made in the end of the bag opposite from the opening, and a glass column containing adsorbent polymer (described in previous section) was inserted into the slit. A single-setting vacuum pump (Barnant Co., Barrington, IL, USA) was used to pull volatile-laden air through the adsorbent trap for 4 min. The bag was then removed to prevent carryover of volatiles from one hour to the next, and the column was stored in a closed container until elution. New bags were used for each sampling period, and volatiles were sampled from empty headspace bags to control for ambient contaminants. Volatiles were eluted from the adsorbent polymer within 12 h using 150 µL of methylene chloride spiked with ∼28.3 ng/µL of nonyl acetate as an elution standard as described previously. Samples were stored at –80 °C until analysis by GC-MS.
Characterization of vegetative volatiles.
Volatiles were collected from vegetative tissues—including leaves, petioles, and stolons—to discriminate between volatiles emitted by vegetative tissues and floral/reproductive tissues. Vegetative tissue was excised from plants of V. odorata, massed using an analytical balance, and loaded into the dynamic headspace collection system described previously. Volatiles were sampled from vegetative tissues for 2 h.
Temporal characterization of V. odorata floral volatile emissions.
To investigate potential temporal variation in the floral volatile emissions of V. odorata, volatiles were sampled individually from three fully opened violet flowers for 4 min every 3 h, starting at 0600 HR and ending at 0400 HR the next day, using the static headspace collection system described previously. Three violet flowers, all on separate plants, were loosely enclosed in headspace bags at 0600 HR, 0900 HR, 1200 HR, 1500 HR, 1800 HR, 2100 HR, 0000 HR, and 0300 HR. Following 1 h of volatile enrichment, volatile-laden air was sampled from each violet for 4 min at 0700 HR, 1000 HR, 1300 HR, 1600 HR, 1900 HR, 2200 HR, 0100 HR, and 0400 HR. The order in which violets were sampled was randomized for each time point, and new bags were used for each sampling period to avoid the accumulation of floral volatiles from one sampling period to the next. After the final collection time point, the violet flowers were harvested and massed using an analytical balance. Emission data over a 24-h period was thus collected for three individual flowers per experiment, and the entire experiment was repeated three times: 7–8 Feb 2022, 10–11 Feb 2022, and 16–17 Feb 2022.
Developmental characterization of V. odorata floral volatile emissions.
To investigate potential variation in floral volatile emissions over the life span of V. odorata flowers, volatiles were sampled from the mature flower bud 1 d before opening (stage −1), on the first day of anthesis (stage 0), and each day thereafter until the first sign of senescence (i.e., perianth wilting). Volatiles were collected using the dynamic push-pull system described previously. A minimum of two flowers of the same developmental stage were loaded into each tube, and the number and mass of the flowers in each tube was recorded. Volatiles were sampled for 2 h, and volatile collections occurred between 1300 HR and 1600 HR. This time frame was chosen based on the results of the temporal characterization of volatile emissions. Volatiles were collected from each developmental stage a minimum of six times over the course of nine volatile collections between 15 Feb 2022 and 10 Mar 2022.
Comparative volatile analysis of assorted Viola.
To gain a basic understanding of what compounds were present in the floral volatile profiles of assorted Viola and assess potential variation among them, the floral volatile emissions of V. odorata, 10 V. odorata cultivars, three Parma violet cultivars, the hybrid ‘Governor Herrick’ (reportedly a cross between V. odorata and V. sororia), and five violet species were qualitatively investigated using the dynamic headspace collection system described previously (Supplemental Appendix Tables A-1 and A-2). In addition, the floral volatile emissions of five V. odorata cultivars and three V. odorata genotypes from different sources were quantitatively investigated using the dynamic headspace collection system described previously. At least two fully opened flowers of the same genotype were loaded into each tube, and the mass and number of flowers in each tube was recorded. Volatiles were collected from each genotype at least in triplicate, with the following exceptions: Viola blanda, Viola tricolor, and the sweet violet cultivar Fair Oaks were sampled in duplicate, while the species V. alba was only sampled once. Depending on flower abundance, volatiles were either collected from multiple replicates of a genotype during a single collection, or they were sampled from a genotype over subsequent collections.
Analysis of volatiles via GC-MS.
where pACOI = peak area of compound of interest in sample; RF = response factor for compound obtained from standard curve; EV = sample elution volume; pANAStnd = peak area of nonyl acetate in collection control; pANASamp = peak area of nonyl acetate in sample; and M = biological sample mass.
Statistical analysis of volatile data.
For all emission replicates, means and standard errors were calculated using JMP Pro (version 14.1.0; SAS Institute, Cary, NC, USA). For temporal volatile data, total volatile emissions were analyzed using analysis of variance (ANOVA) (α = 0.05) and Tukey’s range test with a P value of 0.05 in JMP. For developmental volatile data, emission of five volatiles for the cultivar Rosea and six volatiles for V. odorata plus total volatile emissions for both were analyzed using ANOVA (α = 0.05) and Tukey’s range test with a P value of 0.05 in JMP.
Results
Vegetative volatiles
Five volatiles were emitted from the vegetative tissue of V. odorata: (Z)-3-hexenal (CAS 6789-80-6), hexanal (66-25-1), (Z)-3-hexen-1-ol (928-96-1), (Z)-3-hexenyl acetate (3681-71-8), and methyl benzoate (93-58-3).
Temporal characterization of the floral volatile emissions of V. odorata ‘Rosea’
Because of plant availability and flowering abundance, the V. odorata cultivar Rosea was used to investigate temporal changes in floral volatile emissions. The static headspace collection system was not sensitive enough to capture low-abundance volatiles, so five high-abundance volatiles were used for quantitative analysis: benzaldehyde, 1,4-dimethoxybenzene, dihydro-β-ionone, α-ionone, and β-ionone. The emission of the first three volatiles was not significantly different across the sampling periods; however, the emission of α-ionone (F7,64 = 2.9466, P = 0.0097), β-ionone (F7,64 = 10.4444, P < 0.0001), and total volatile emissions (F7,64 = 7.5613, P < 0.0001) were significantly different (P < 0.05) among sampling periods (Figs. 1−3). The emission of α-ionone was highest between 0600 HR and 1900 HR, and significantly higher (P < 0.05) than emissions between 0300 HR and 0400 HR (Fig. 1). The emission of β-ionone was highest between 0600 HR and 1900 HR, and significantly higher (P < 0.05) than emissions between 2100 HR and 0400 HR (Fig. 2). Total volatile emissions were highest between 0600 HR and 1900 HR, and significantly higher (P < 0.05) than emissions between 0000 HR and 0400 HR (Fig. 3).
Temporal pattern of emission of α-ionone (ng·g FW−1) in V. odorata ‘Rosea’ starting at 0600 HR and ending at 0400 HR the next day. Each bar represents the overall mean ±SE amount of α-ionone emitted during the sampling period (n = 9; three individual flowers sampled per experiment; experiment replicated three times). Different letters indicate significantly different means according to Tukey’s honestly significant difference (P < 0.05).
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Temporal pattern of emission of β-ionone (ng·g FW−1) in V. odorata ‘Rosea’ starting at 0600 HR and ending at 0400 HR the next day. Each bar represents the overall mean ±SE amount of β-ionone emitted during the sampling period (n = 9; three individual flowers sampled per experiment; experiment replicated three times). Different letters indicate significantly different means according to Tukey’s honestly significant difference (P < 0.05).
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Temporal pattern of total volatile emissions (ng·g FW−1) in V. odorata ‘Rosea’ starting at 0600 HR and ending at 0400 HR the next day. Each bar represents the overall mean ±SE amount of total volatiles emitted during the sampling period (n = 9; three individual flowers sampled per experiment; experiment replicated three times). Different letters indicate significantly different means according to Tukey’s honestly significant difference (P < 0.05).
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Developmental characterization of the floral volatile emissions of V. odorata
Developmental variation in floral volatile emissions was investigated in both the V. odorata cultivar Rosea and the species V. odorata. For ‘Rosea’, the emission of five volatile compounds—benzaldehyde, 1,4-dimethoxybenzene, α-ionone, dihydro-β-ionone, and β-ionone—and the total emission of these compounds were examined. For V. odorata, the emissions of these same five compounds, plus the compound p-cresol, and the total emissions of these six compounds were examined.
In ‘Rosea’, the emission of individual compounds and the total emission was high from the first day of anthesis (stage 0) or stage +1 until stage +4 or wilt (Table 1, Fig. 4). Most compounds had a general emission pattern of little to no emission at stage −1, followed by an increase in emission at stage 0, maximal emission from stages +1 to +4, and a decrease in emission at wilt; however, significant differences between all stages were not found for all compounds. In V. odorata, most compounds had a similar general emission pattern as those in ‘Rosea’, but emissions tended to decrease after stage +3 (Table 1, Fig. 5). More significant differences between stages were found for more compounds as well.
Results of the analysis of variance for the effects of developmental stage on the emission of individual floral volatiles and of total volatile emissions in Viola odorata ‘Rosea’ and Viola odorata. Statistically significant results are bolded.
Developmental emission (ng·g FW−1) of floral volatiles over the floral lifespan of V. odorata ‘Rosea’. Volatiles were collected on the day before anthesis (stage −1), the first day of anthesis (stage 0), and on each day thereafter until the first sign of senescence (i.e., perianth wilting). Each bar represents the mean ±SE amount of volatile emitted at a given developmental stage. Each developmental stage was sampled a minimum of six times. Different letters indicate significantly different means according to Tukey’s honestly significant difference (P < 0.05).
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Developmental emission (ng·g FW−1) of floral volatiles over the floral lifespan of V. odorata (species purchased as plants from Crimson Sage Nursery). Volatiles were collected on the day before anthesis (stage −1), the first day of anthesis (stage 0), and on each day thereafter until the first sign of senescence (i.e., perianth wilting). Each bar represents the mean ±SE amount of volatile emitted at a given developmental stage. Each developmental stage was sampled a minimum of six times. Different letters indicate significantly different means according to Tukey’s honestly significant difference (P < 0.05).
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Comparative volatile analysis of assorted violets
Qualitative comparison of floral volatile emissions among sweet violets, a violet hybrid, and Parma violets.
Volatiles were sampled from the Viola hybrid ‘Governor Herrick’, V. odorata, 10 sweet violet cultivars, and three Parma violet cultivars, and 18 volatile compounds were identified (Tables 2 and 3). Qualitative variation was found among all violets, as well as among the sweet violet cultivars, and to a lesser extent, among Parma violet cultivars. Notably, V. odorata and the sweet violet cultivars Clive Groves, Lianne, Mauve, Princess of Wales, Royal Wedding, and Queen Charlotte were found to emit p-cresol. In addition, the compounds 4-methylanisole and estragole were only identified in the floral profiles of the three Parma violet cultivars.
Volatile compounds qualitatively identified in the floral headspace of assorted violets. Compounds are listed in ascending order of GC-MS retention time.
Compound class of volatile compounds identified in the headspace of violets. Compound classification is according to Knudsen et al. (2006).
Qualitative comparison of floral volatile emissions among select violet species.
Volatiles were sampled at least twice from the four violet species V. blanda, V. sororia, V. tricolor, and V. walteri, and once from V. alba, and 17 volatile compounds were qualitatively identified (Tables 3 and 4). V. alba was only sampled once due to plants not being grown from seed until late 2021, and flowering not occurring until Spring 2022. Overall, emissions from these violets were low to extremely low. V. alba emitted the three ionones associated with the classic violet fragrance (α-ionone, dihydro-β-ionone, and β-ionone), but none of the other violet species emitted all three of these compounds. Only V. tricolor emitted small quantities of β-ionone.
Volatiles qualitatively identified in the floral headspace of Viola species. Compounds are listed in ascending order of GC-MS retention time.
Quantitative comparison of floral volatile emissions among selected sweet violets.
Volatiles were sampled at least five times from five sweet violet cultivars and three geno types of the species V. odorata (each obtained from a different source), and five volatile compounds were quantitatively compared (Fig. 6). Qualitative comparison of floral volatiles for the species and its cultivars is shown in Table 2. In addition, the relative ratio of the three major ionones, α-ionone, dihydro-β-ionone, and β-ionone, emitted by the sweet violet cultivars, species genotypes, and Parma violet cultivars were examined (Fig. 7).
Comparison of select volatiles among V. odorata and its cultivars. Each bar represents the overall mean ±SE amount (ng·g FW−1) of a given volatile that was emitted by the violet during the sampling period. Each genotype was sampled at least five times.
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Percent contribution of individual ionones to total ionone emission in select violets. For each violet, the percent of the three ionones sums to 100. Standard errors are indicated by vertical bars.
Citation: HortScience 59, 7; 10.21273/HORTSCI17847-24
Discussion
Vegetative volatiles
Four of the compounds emitted from the vegetative tissues of V. odorata—(Z)-3-hexenal, hexanal, (Z)-3-hexen-1-ol, and (Z)-3-hexenyl acetate—are well-known “green leaf volatiles” (GLVs) formed via lipoxygenase cleavage of fatty acids (Ameye et al. 2018; Unsicker et al. 2009). Methyl benzoate is a C6-C1 ester most commonly found as a component in floral scent (Knudsen et al. 2006). However, it has also been detected in the volatile profile of undamaged leaves of Chamaerops humilis (Caissard et al. 2004) and Camellia sinensis (Wang et al. 2019), as well as from insect-damaged rice leaves (Zhao et al. 2010) and mechanically wounded Populus tremula leaves (Portillo-Estrada et al. 2017). Most relevantly, Saint-Lary et al. (2014) identified all these compounds except (Z)-3-hexenal in absolutes made from V. odorata leaves. (Z)-3-Hexenal is a relatively unstable compound that is prone to either enzymatic or nonenzymatic isomerization, and thus often rearranges to form the more stable (E)-2-hexenal (Ameye et al. 2018). This more stable compound was detected in the absolutes of V. odorata leaves (Saint-Lary et al. 2014). Violet leaf extracts are still widely used in the perfume industry to impart fresh, green, and cucumber-like notes (Saint-Lary et al. 2016). The four GLVs identified here contribute to these notes.
Temporal characterization of the floral volatile emissions of V. odorata ‘Rosea’
The emission of β-ionone was highest between 0600 HR and 1900 HR, and significantly higher than emissions between 2100 HR and 0400 HR, and total volatile emissions were highest between 0600 HR and 1900 HR, and significantly higher than emissions between 0000 HR and 0400 HR, suggesting a diurnal pattern of emission. This emission pattern may be related to the primary pollinators of sweet violets: a variety of bees and hoverflies (Beattie 1969b, 1971). β-Ionone, an irregular terpene in the apocarotenoid group that is formed via enzymatic cleavage of several carotenoid substrates, is known to attract different types of bees (Knudsen et al. 2006; Rabeschini et al. 2021; Zhang et al. 2022). In addition, the emission of β-ionone was found to be highest during the day in Petunia hybrida (Simkin et al. 2004) and Osmanthus fragrans (Baldermann et al. 2010). Plants that are pollinated by diurnally active insects—such as bees—tend to exhibit diurnal rhythmicity in their volatile emissions (Burdon et al. 2015; Jürgens et al. 2014). Aside from reducing the energy costs related to scent production, varying the quality and/or quantity of floral volatile emissions to correspond with the activity of potential pollinators can improve the reproductive success of a plant (Kolosova et al. 2001). However, although insect visitations to the CH flowers of V. odorata have long been observed and documented, the CH flowers frequently fail to set seed (Beattie 1969a). Indeed, this phenomenon was discussed in Charles Darwin’s correspondence with several individuals including George Bentham, who said that botanists had no explanation for why the showy spring flowers of V. odorata were sterile while the “very inconspicuous and petalless” flowers alone produced seed (Bentham 1868). Eventually, an explanation for this phenomenon was determined: Madge (1929) and Mayers and Lord (1984) observed that before anthesis in V. odorata, a portion of the pollen germinated precociously in the anthers of CH flowers, creating a tangled mass of germinating and ungerminated pollen within the locules that limits or prevents pollen dispersal despite normal anther dehiscence, thus rendering CH flowers functionally sterile much of the time. However, spontaneous hybrids resulting from natural interspecific hybridization between V. odorata and other violet species have been reported in Europe, suggesting that insect visitation to the CH flowers of V. odorata can result in successful pollen transfer, fertilization, and seed set (Marcussen and Borgen 2000). Moreover, Mayers and Lord (1983b) achieved successful fertilization in CH flowers of V. odorata by transferring mature, ungerminated pollen to the stigma by hand. Thus, although the primary form of sexual reproduction in V. odorata—self-pollination of CL flowers—occurs without the intervention of pollinators, the species retains the capacity for outcrossing via CH flowers. Consequently, the temporal variation in floral volatile emissions observed in this study, which corresponds to the activity of potential pollinators, could be a byproduct of the maintenance of an outcrossing system (Oakley et al. 2007).
Developmental characterization of the floral volatile emissions of V. odorata
The overall high emission of total and individual floral volatiles from stages 0 or +1 to stages +3 or +4 may suggest that the flowers are receptive to pollination during these times (Burdon et al. 2015). Typically, the highest emission of volatiles occurs when a flower is ready for pollination, and the levels decrease during senescence (Muhlemann et al. 2014). Indeed, two of the compounds with high emission from stages +1 to +3, 1,4-dimethoxybenzene in both ‘Rosea’ and V. odorata and p-cresol in the latter, are known to elicit positive behavioral responses in bees, one of the primary pollinators of sweet violets (Beattie 1969b; Dötterl and Vereecken 2010). However, given that sweet violets primarily reproduce via seeds from self-pollinated CL flowers, the variation in floral volatile emissions over the life span of CH flowers could be due to other factors besides floral receptivity and pollinator activity, or, as discussed previously, such variation could be a byproduct of the maintenance of an outcrossing system (Oakley et al. 2007).
Comparative volatile analysis of assorted violets
Qualitative comparison of floral volatile emissions among sweet violets, a violet hybrid, and Parma violets.
Floral volatile emissions qualitatively varied among all the violets. Notably, the compound p-cresol was only detected in the volatile profiles of V. odorata and six of its cultivars. p-Cresol has been identified in the floral scent of numerous plants, and it is one of the major volatile constituents of herbivore dung, particularly cattle and swine (Knudsen et al. 2006; Schiestl and Dötterl 2012). Though it is often a minor component of floral scent, it can pack an odorant punch due to its low odor threshold and impart unpleasant fecal and urinous characters to a flower’s aroma, which for some plants serves as an attractant to pollinators (Kite 1995; Shuttleworth 2016). It has also been shown to elicit positive behavioral responses in bees (Dötterl and Vereecken 2010). Given its somewhat unpleasant and musky notes, a violet with a fragrance that lacks p-cresol may be preferred, and thus p-cresol should potentially be selected against in a fragrance-focused Viola breeding program. However, sensory panel evaluations of violet fragrances with and without p-cresol should first be conducted.
The violet hybrid ‘Governor Herrick’ is reportedly the result of a cross between V. odorata and V. sororia (2n = 54) (Coombs 2003). ‘Governor Herrick’ inherited some of its growth characteristics from V. sororia, namely its leaf shape and the fact that it does not produce runners/stolons (Coombs 2003). However, it seems to have inherited at least a tiny bit of fragrance from its other parent: two of the major ionones, α-ionone and β-ionone, were indeed found in trace quantities in its floral volatile profile. The presence of these compounds suggests that volatiles can be introduced into aroma profiles through breeding.
The compound 4-methylanisole was only identified in the headspace of the three Parma violet cultivars and was not detected in the volatile profile of any other violet. This compound was identified by Chervin et al. (2019) in the floral headspace of violets belonging to the V. alba cluster, and its presence allowed for distinction between violets belonging to V. alba and V. suavis M. Bieb. The mostly sterile Parma violets were long thought to originate from V. odorata, given the similarities in their fragrance; however, Parma violets were determined to originate from the V. alba complex and are best included in the Mediterranean subsp. dehnhardtii (Malécot et al. 2007). Parma violets are considered to have a sweeter fragrance than V. odorata (Coon 1977). This sweeter fragrance may be due in part to the presence of 4-methylanisole, as this compound has a sweet, floral, and fruity odor character (Johnson et al. 2019; Zhou et al. 2023); however, other factors, such as the absence of p-cresol, and/or the ratio of the three major ionones (α-ionone, dihydro-β-ionone, and β-ionone) may play a role as well. If it is determined that the regular species V. alba emits 4-methylanisole, and if sensory panel evaluations determine that this compound imparts favorable odor characteristics, then this may be a compound of interest to incorporate into a fragrance-focused Viola breeding program.
Qualitative comparison of floral volatile emissions among select violet species.
Compared with V. odorata, most of the other violet species had unremarkable fragrance profiles and low to extremely low emissions. While the three ionones associated with the classic violet fragrance (α-ionone, dihydro-β-ionone, and β-ionone) were identified in the headspace of V. alba, the compound 4-methylanisole was not detected. Chervin et al. (2019) reported that 4-methylanisole was a distinctive component of the volatile profile of violets belonging to V. alba; however, it is unclear whether they identified the compound in the profile of the pure species, V. alba, or if it was only detected in the profiles of V. alba subsp. dehnhardtii, more commonly known as Parma violets. Additional volatile collections must be conducted with V. alba to determine if 4-methylanisole is indeed a constituent of its floral profile. Aside from V. alba, V. tricolor was the only violet found to emit any ionones (β-ionone). Anca et al. (2009) identified β-ionone in essential oils made from fresh and dried aerial parts of V. tricolor. Compared with the other violet species, V. tricolor had a noticeable and pleasant floral fragrance, with a profile dominated by terpenes and benzenoids. V. tricolor (2n = 26), also known as heartsease, has almost as long and storied of a history as V. odorata, and it is one of the progenitorial species of the modern cultivated garden pansy (Coon 1977; Dalbato et al. 2013).
Although V. sororia and V. walteri had unremarkable fragrance profiles with virtually no detectable floral scent, they were included in the analysis because they are both native to the southeastern United States and grow well in Florida, despite the heat and humidity. As mentioned previously, ‘Governor Herrick’ supposedly arose from a cross between V. odorata and V. sororia in the early 20th century. Although the resulting hybrid appears to have inherited some aromatic characteristics from V. odorata, its fragrance is still very low—indeed, almost negligible. To increase the fragrance further, backcrosses with V. odorata would need to be performed; however, hybrids between V. sororia (2n = 54) and V. odorata (2n = 20) may be sterile, or mostly infertile at best, as CL flowers of ‘Governor Herrick’ wither without forming seed capsules (Marcussen and Karlsson 2010). Thus, V. walteri may be a particular species of interest for future interspecific hybridization efforts with V. odorata. Aside from V. walteri’s greater tolerance for heat and humidity, both species have chromosome number 2n = 20, and V. walteri has been known to successfully hybridize with other 2n = 20 violets, including Viola labradorica, Viola conspera, and Viola striata (Ballard 1992). Similarly, spontaneous hybrids with varying degrees of fertility have resulted from natural interspecific hybridization between V. odorata and at least four violet species with 2n = 20, including V. alba, Viola collina, Viola hirta, and Viola reichenbachiana (Marcussen and Borgen 2000).
Quantitative comparison of floral volatile emissions among selected sweet violets.
The sweet violets included in the quantitative analysis were opportunistically sampled during dynamic volatile collections from 31 Jan 2022 to 10 Mar 2022 using two or more fully opened flowers from the same plant; however, the exact floral age for most of the flowers was not known at the time of sampling. Given the variation in floral volatile emissions at different developmental stages, volatile emission means were not statistically compared, only presented. Regardless, some of the most striking visual differences were seen among the three major ionones: α-ionone, dihydro-β-ionone, and β-ionone. Consequently, the relative ratio of ionones emitted among the sweet violet cultivars, species, and Parma violet cultivars were further examined. For the Parma violets, ‘Empress Augusta’, and the three species genotypes, the percent emission of α-ionone was highest. In contrast, the percent emission of dihydro-β-ionone was highest for ‘Rosea’ and ‘Rubra’. The percent emission of β-ionone varied widely as well, making up the second highest percentage for the three Parma violets and V. odorata genotypes, but it had the lowest emission for the five sweet violet cultivars. The differences in relative emission of these ionones among violets is intriguing and may have real-world impacts on fragrance perception and preference. These compounds have extremely low odor thresholds (3.2, 1.7, and 0.12 ng/L of air for α-ionone, dihydro-β-ionone, and β-ionone, respectively) and are thus considered important flavor and aroma contributors (Brenna et al. 2002). In addition, although these compounds collectively are described as having violet-like odors, they impart different odorant properties. For example, although both α- and β-ionone have characteristic violet-like odors, β-ionone possesses more fruity and woody characteristics, whereas α-ionone has stronger floral notes (Burdock 2010; Cataldo et al. 2016; Werkhoff et al. 1991). Moreover, although dihydro-β-ionone has a “β-ionone-type odor,” it also possesses green and earthy undertones and an “orris-type tonality” (Brenna et al. 2002; Cataldo et al. 2016). Another factor that may affect the perceptions of and preferences for violet fragrances is the high rates of specific anosmia for β-ionone (Paparella et al. 2021). Specific anosmia refers to the inability to perceive a particular odor by individuals who otherwise have a normal sense of smell (Amoore 1967). Brenna et al. (2002) stated that ∼34% of the European population cannot smell β-ionone, and Plotto et al. (2006) found that 50% of panelists in their study could not perceive β-ionone; however, these same panelists perceived α-ionone normally. Specific anosmia for dihydro-β-ionone has not been investigated. Regardless, the perception of an aroma that contains β-ionone may differ substantially between those who can detect β-ionone and those who cannot (Plotto et al. 2006). Ultimately, however, the differences in relative emission of the three ionones among violets and their impact on fragrance perception and preference must be further investigated.
Conclusions
The floral volatile emissions of V. odorata were temporally and developmentally characterized, and the floral volatile emissions of assorted Viola were investigated. Vegetative volatiles were also identified. Emission of α-ionone, β-ionone, and total volatile emissions in V. odorata ‘Rosea’ were highest from 0600 HR to 1900 HR, suggesting a diurnal pattern of emission that corresponds with its primary pollinators. Volatile emissions of sweet violets also showed variation over the developmental life span of the flower, with the highest emission of individual volatiles and total volatiles occurring, in general, from stages 0 or +1 to stages +3 or +4. This developmental emission pattern suggests that flowers may be receptive during these stages, although further investigation is warranted. Comparison of floral volatile emissions among Parma violets, sweet violets, violet species, and hybrids revealed both qualitative and quantitative variation.
Directions for future study include additional characterization of the floral volatile emissions of V. alba to determine if its volatile profile includes the compound 4-methylanisole. The presence and absence of the compound p-cresol in violet fragrance should be evaluated by sensory panels to determine if this compound should be selected against in a fragrance-focused Viola breeding program. Further sensory panel evaluation of sweet violets with varying ratios of major ionones also should be conducted to determine which ratios of ionones are preferred. Finally, floral volatile emissions should be rigorously investigated in V. odorata as the floral form transitions from fully chasmogamous, through its various intermediary forms, and finally to fully cleistogamous to gain a better understanding of the floral volatile emissions of a species whose primary form of reproduction does not rely on insect pollinators.
References Cited
Abdel-Rahim IR. 2016. Control of Alternaria rot disease of pear fruits using essential oil of Viola odorata. J Phytopathol Pest Manage. 3(2):71–84.
Akhbari M, Batooli H, Kashi FJ. 2012. Composition of essential oil and biological activity of extracts of Viola odorata L. from central Iran. Nat Prod Res. 26(9):802–809.
Ameye M, Allmann S, Verwaeren J, Smagghe G, Haesaert G, Schuurink RC, Audenaert K. 2018. Green leaf volatile production by plants: A meta-analysis. New Phytol. 220(3):666–683.
Amoore JE. 1967. Specific anosmia: A clue to the olfactory code. Nature. 214:1095–1098.
Anca T, Philippe V, Ilioara O, Mircea T. 2009. Composition of essential oils of Viola tricolor and V. arvensis from Romania. Chem Nat Compd. 45(1):91–92.
Balao F, Herrera J, Talavera S, Dötterl S. 2011. Spatial and temporal patterns of floral scent emission in Dianthus inoxianus and electroantennographic responses of its hawkmoth pollinator. Phytochemistry. 72:601–609.
Baldermann S, Kato M, Kurosawa M, Kurobayashi Y, Fujita A, Fleischmann P, Watanabe N. 2010. Functional characterization of a carotenoid cleavage dioxygenase 1 and its relation to the carotenoid accumulation and volatile emission during the floral development of Osmanthus fragrans Lour. J Exp Bot. 61:2967–2977.
Ballard HE Jr. 1992. Systematics of Viola, section Viola, in North America north of Mexico (PhD Diss). Central Michigan University, Mount Pleasant, MI, USA.
Ballard HE Jr, Sytsma KJ, Kowal RR. 1999. Shrinking the violets: Phylogenetic relationships of infrageneric groups in Viola (Violaceae) based on internal transcribed spacer DNA sequences. Syst Bot. 23(4):439–458.
Baskin CC, Baskin JM. 2014. Seeds: Ecology, biogeography, and evolution of dormancy and germination (2nd ed). Academic Press, San Diego, CA, USA.
Beattie AJ. 1969a. Studies in the pollination ecology of Viola. I. The pollen-content of stigmatic cavities. Watsonia. 7(3):142–156.
Beattie AJ. 1969b. The floral biology of three species of Viola. New Phytol. 68(4):1187–1201.
Beattie AJ. 1971. Pollination mechanisms in Viola. New Phytol. 70(2):343–360.
Bentham G. 1868. Darwin correspondence project, letter no. 6147. https://www.darwinproject.ac.uk/letter/?docId=letters/DCP-LETT-6147.xml. [accessed 8 May 2022].
Brenna E, Fuganti C, Serra S, Kraft P. 2002. Optically active ionones and derivatives: Preparation and olfactory properties. Eur J Org Chem. 2002:967–978.
Burdock GA. 2010. Fenaroli’s handbook of flavor ingredients (6th ed). CRC Press, Boca Raton, FL, USA.
Burdon RCF, Raguso RA, Kessler A, Parachnowitsch AL. 2015. Spatiotemporal floral scent variation of Penstemon digitalis. J Chem Ecol. 41:641–650. https://doi.org/10.1007/s10886-015-0599-1.
Caissard J-C, Meekijjironenroj A, Baudino S, Anstett M-C. 2004. Localization of production and emission of pollinator attractant on whole leaves of Chamaerops humilis (Arecaceae). Am J Bot. 91(8):1190–1199.
Cataldo VF, López J, Cárcamo M, Agosin E. 2016. Chemical vs. biotechnological synthesis of C13-apocarotenoids: Current methods, applications, and perspectives. Appl Microbiol Biotechnol. 100:5703–5718. https://doi.org/10.1007/s00253-016-7583-8.
Chervin J, Talou T, Audonnet M, Dumas B, Camborde L, Esquerré-Tugayé M-T, Roux C, Guillaume C, Guillaume M. 2019. Deciphering the phylogeny of violets based on multiplexed genetic and metabolomic approaches. Phytochemistry. 163:99–110. https://doi.org/10.1016/j.phytochem.2019.04.001.
Coombs RE. 2003. Violets: The history & cultivation of scented violets (2nd ed). B T Batsford, London, United Kingdom.
Coon N. 1977. The complete book of violets. A.S. Barnes and Company, Cranbury, NJ, USA.
Culley TM, Klooster MR. 2007. The cleistogamous breeding system: A review of its frequency, evolution, and ecology in angiosperms. Bot Rev. 73:1–30. https://doi.org/10.1663/0006-8101(2007)73[1:TCBSAR]2.0.CO;2.
Dalbato AL, Kobza F, Karlsson LM. 2013. Effect of polyploidy and pollination methods on capsule and seed set of pansies (Viola x wittrockiana Gams). Hortic Sci (Prague). 40(1):22–30.
Dötterl S, Jahreiß K, Jhumur US, Jürgens A. 2012. Temporal variation of flower scent in Silene otites (Caryophyllaceae): A species with a mixed pollination system. Bot J Linn Soc. 169:447–460.
Dötterl S, Vereecken NJ. 2010. The chemical ecology and evolution of bee-flower interactions: A review and perspectives. Can J Zool. 88:668–697. https://doi.org/10.1139/Z10-031.
Dudareva N, Klempien A, Muhlemann JK, Kaplan I. 2013. Biosynthesis, function and metabolic engineering of plant volatile organic compounds. New Phytol. 198(1):16–32. https://doi.org/10.1111/nph.12145.
Dudareva N, Negre F, Nagegowda DA, Orlova I. 2006. Plant volatiles: Recent advances and future perspectives. Crit Rev Plant Sci. 25:417–440. https://doi.org/10.1080/07352680600899973.
Dudareva N, Pichersky E. 2000. Biochemical and molecular genetic aspects of floral scents. Plant Physiol. 122:627–633.
Erben M. 1996. The significance of hybridization on the forming of species in the genus Viola. Bocconea. 5:113–118.
Gautschi M, Bajgrowicz JA, Kraft P. 2001. Fragrance chemistry—milestones and perspectives. Chimia. 55(5):379–387.
Hammami I, Kamoun N, Rebai A. 2011. Biocontrol of Botrytis cinerea with essential oil and methanol extract of Viola odorata L. flowers. Arch Appl Sci Res. 3(5):44–51.
Johnson BO, Golonka AM, Blackwell A, Vazquez I, Wolfram N. 2019. Floral scent variation in the heterostylous species Gelsemium sempervirens. Molecules. 24(15):2818.
Jürgens A, Glück U, Aas G, Dötterl S. 2014. Diel fragrance pattern correlates with olfactory preferences of diurnal and nocturnal flower visitors in Salix caprea (Salicaceae). Bot J Linn Soc. 175(4):624–640.
Keene SA, Johnson TS, Sigler CL, Kalk TN, Genho P, Colquhoun TA. 2020. A survey of the floral volatile profiles of daylily species and hybrids. J Am Soc Hortic Sci. 145:120–130. https://doi.org/10.21273/JASHS04833-19.
Kite GC. 1995. The floral odour of Arum maculatum. Biochem Syst Ecol. 23:343–354. https://doi.org/10.1016/0305-1978(95)00026-Q.
Ko MK, Yang J, Jin YH, Lee CH, Oh BJ. 1998. Genetic relationships of Viola species evaluated by random amplified polymorphic DNA analysis. J Hortic Sci Biotechnol. 73(5):601–605.
Kolosova N, Gorenstein N, Kish CM, Dudareva N. 2001. Regulation of circadian methyl benzoate emission in diurnally and nocturnally emitting plants. Plant Cell. 13(10):2333–2347. https://doi.org/10.1105/tpc.010162.
Knudsen JT, Eriksson R, Gershenzon J. 2006. Diversity and distribution of floral scent. Bot Rev. 72(1):1–120. https://doi.org/10.1663/0006-8101(2006)72[1:DADOFS]2.0.CO;2.
Knudsen JT, Tollsten L, Bergström LG. 1993. Floral scents—a checklist of volatile compounds isolated by head-space techniques. Phytochemistry. 33:253–280. https://doi.org/10.1016/0031-9422(93)85502-I.
Lalko J, Lapczynski A, McGinty D, Bhatia S, Letizia CS, Api AM. 2007a. Fragrance material review on β-ionone. Food Chem Toxicol. 45:S241–S247. https://doi.org/10.1016/j.fct.2007.09.052.
Lalko J, Lapczynski A, Politano VT, McGinty D, Bhatia S, Letizia CS, Api AM. 2007b. Fragrance material review on α-ionone. Food Chem Toxicol. 45:S235–S240. https://doi.org/10.1016/j.fct.2007.09.046.
Levin LA, Langer KM, Clark DG, Colquhoun TA, Callaway JL, Moskowitz HR. 2012. Using Mind Genomics® to identify essential elements of a flower product. HortScience. 47(11):1658–1665. https://doi.org/10.21273/HORTSCI.47.11.1658.
Lim TK. 2014. Viola odorata, 795–807. In: Lim TK (ed). Edible medicinal and non medicinal plants, vol. 8, flowers. Springer, Dordrecht, Netherlands. https://doi.org/10.1007/978-94-007-7395-0.
Madge MAP. 1929. Spermatogenesis and fertilization in the cleistogamous flower of Viola odorata, var. praecox, Gregory. Ann Bot. 43(171):545–577.
Malécot V, Marcussen T, Munzinger J, Yockteng R, Henry M. 2007. On the origin of the sweet-smelling Parma violet cultivars (Violaceae): Wide intraspecific hybridization, sterility, and sexual reproduction. Am J Bot. 94(1):29–41.
Marcussen T. 2006. Allozymic variation in the widespread and cultivated Viola odorata (Violaceae) in western Eurasia. Bot J Linn Soc. 151(4):563–571.
Marcussen T, Ballard HE, Danihelka J, Flores AR, Nicola MV, Watson JM. 2022. A revised phylogenetic classification for Viola (Violaceae). Plants. 11:2224. https://doi.org/10.3390/plants11172224.
Marcussen T, Borgen L. 2000. Allozymic variation and relationships within Viola subsection Viola (Violaceae). Plant Syst Evol. 223:29–57. https://doi.org/10.1007/BF00985325.
Marcussen T, Karlsson T. 2010. Violaceae, p 12–52. In: Jonsell B, Karlsson T (eds). Flora nordica, vol. 6. Swedish Royal Academy of Sciences, Stockholm, Sweden.
Maxwell C. 2017. Scents & sensibility: Perfume in Victorian literary culture. Oxford University Press, Oxford, United Kingdom.
Mayers AM, Lord EM. 1983a. Comparative flower development in the cleistogamous species Viola odorata. I. A growth rate study. Am J Bot. 70(10):1548–1555.
Mayers AM, Lord EM. 1983b. Comparative flower development in the cleistogamous species Viola odorata. II. An organographic study. Am J Bot. 70(10):1556–1563.
Mayers AM, Lord EM. 1984. Comparative flower development in the cleistogamous species Viola odorata. III. A histological study. Bot Gaz. 145(1):83–91.
Mohamed YFY, Ghatas YAA. 2016. Effect of mineral, biofertilizer (EM) and zeolite on growth, flowering, yield and composition of volatile oil of Viola odorata L. plants. J Hortic Sci Ornam Plants. 8(3):140–148.
Muhlemann JK, Klempien A, Dudareva N. 2014. Floral volatiles: From biosynthesis to function. Plant Cell Environ. 37(8):1936–1949.
Neuffer B, Auge H, Mesch H, Amarell U, Brandl R. 1999. Spread of violets in polluted pine forests: Morphological and molecular evidence for the ecological importance of interspecific hybridization. Mol Ecol. 8(3):365–377.
Oakley CG, Moriuchi KS, Winn AA. 2007. The maintenance of outcrossing in predominantly selfing species: Ideas and evidence from cleistogamous species. Annu Rev Ecol Evol Syst. 38(1):437–457.
Paparella A, Shaltiel-Harpaza L, Ibdah M. 2021. β-Ionone: Its occurrence and biological function and metabolic engineering. Plants. 10(4):754.
Pichersky E, Dudareva N. 2007. Scent engineering: Toward the goal of controlling how flowers smell. Trends Biotechnol. 25(3):105–110.
Plotto A, Barnes KW, Goodner KL. 2006. Specific anosmia observed for β-ionone, but not for α-ionone: Significance for flavor research. J Food Sci. 71(5):S401–S406.
Portillo-Estrada M, Kazantsev T, Niinemets U. 2017. Fading of wound-induced volatile release during Populus tremula leaf expansion. J Plant Res. 130(1):157–165.
Pybus DH, Sell CS. 1999. The chemistry of fragrances. Royal Society of Chemistry, Cambridge, United Kingdom.
Rabeschini G, Bergamo PJ, Nunes CEP. 2021. Meaningful words in crowd noise: Searching for volatiles relevant to carpenter bees among the diverse scent blends of bee flowers. J Chem Ecol. 47:444–454. https://doi.org/10.1007/s10886-021-01257-y.
Raguso RA, Pellmyr O. 1998. Dynamic headspace analysis of floral volatiles: A comparison of methods. Oikos. 81:238–254. https://doi.org/10.2307/3547045.
Shakespeare W. 1954. The life and death of King John. In: Honigmann EAJ (ed). Arden Shakespeare Second Series. http://shakespeare.mit.edu/john/full.html. [accessed 5 May 2022].
Saint-Lary L, Roy C, Paris JP, Martin JF, Thomas OP, Fernandez X. 2016. Metabolomics for the authentication of natural extracts used in flavors and fragrances: The case study of violet leaf absolutes from Viola odorata. Chem Biodivers. 13(6):737–747.
Saint-Lary L, Roy C, Paris J-P, Tournayre P, Berdagué J-L, Thomas O-P, Fernandez X. 2014. Volatile compounds of Viola odorata absolutes: Identification of odorant active markers to distinguish plants originating from France and Egypt. Chem Biodivers. 11(6):843–860.
Schiestl FP, Dötterl S. 2012. The evolution of floral scent and olfactory preferences in pollinators: Coevolution or pre-existing bias? Evolution. 66(7):2042–2055.
Shuttleworth A. 2016. Smells like debauchery: The chemical composition of semen-like, sweat-like and faintly foetid floral odours in Xysmalobium (Apocynaceae: Asclepiadoidae). Biochem Syst Ecol. 66:63–75. https://doi.org/10.1016/j.bse.2016.03.009.
Simkin AJ, Underwood BA, Auldridge M, Loucas HM, Shibuya K, Schmelz E, Clark DG, Klee HJ. 2004. Circadian regulation of the PhCCD1 carotenoid cleavage dioxygenase controls emission of β-ionone, a fragrance volatile of petunia flowers. Plant Physiol. 136(3):3504–3514.
Stebbins GL, Harvey BL, Cox EL, Rutger JN, Jelencovic G, Yagil E. 1963. Identification of the ancestry of an amphiploid Viola with the aid of paper chromatography. Am J Bot. 50(8):830–839.
Steenhuisen S-L, Raguso RA, Jürgens A, Johnson SD. 2010. Variation in scent emission among floral parts and inflorescence developmental stages in beetle-pollinated Protea species (Proteaceae). S Afr J Bot. 76(4):779–787.
Tiemann F, Krüger P. 1893. Über Veilchenaroma [Translation from German: About violet aroma]. Ber Dtsch Chem Ges. 26(3):2675–2708.
Tholl D, Boland W, Hansel A, Loreto F, Röse USR, Schnitzler USR. 2006. Practical approaches to plant volatile analysis. Plant J. 45(4):540–560.
Turin L, Sanchez T. 2008. Perfumes: The guide. Profile Books Ltd., London, United Kingdom.
Uhde G, Ohloff G. 1972. Parmon, eine Phantomverbindung im Veilchenblütenöl. [Translation from German: Parmone, a phantom compound in violet flower oil]. Helv Chim Acta. 55(7):2621–2625.
Unsicker SB, Kunert G, Gershenzon J. 2009. Protective perfumes: The role of vegetative volatiles in plant defense against herbivores. Curr Opin Plant Biol. 12(4):479–485.
Wang X, Zeng L, Liao Y, Zhou Y, Xu X, Dong F, Yang Z. 2019. An alternative pathway for the formation of aromatic aroma compounds derived from L-phenylalanine via phenylpyruvic acid in tea (Camellia sinensis (L.) O. Kuntze) leaves. Food Chem. 270:17–24. https://doi.org/10.1016/j.foodchem.2018.07.056.
Werkhoff P, Bretschneider W, Güntert M, Hopp R, Surburg H. 1991. Chirospecific analysis in flavor and essential oil chemistry. Part B. Direct enantiomer resolution of trans-α-ionone and trans-α-damascone by inclusion gas chromatography. Z Lebensm Unters Forsch. 192(2):111–115.
Yockteng R, Jr Ballard HE, Mansion G, Dajoz I, Nadot S. 2003. Relationships among pansies (Viola section Melanium) investigated using ITS and ISSR markers. Plant Syst Evol. 241(3/4):153–170.
Zhang H, Wang J-Y, Chen Y-J, Siemann E, Ji X-Y, Jiang JX, Wan N-F. 2022. Beta-ionone is a functional plant volatile that attracts the parasitic wasp, Microplitis pallidipes. BioControl. 67(1):1–13.
Zhao N, Guan J, Ferrer J-L, Engle N, Chern M, Ronald P, Tschaplinski TJ, Chen F. 2010. Biosynthesis and emission of insect-induced methyl salicylate and methyl benzoate from rice. Plant Physiol Biochem. 48(4):279–287.
Zhou T, Ning K, Han W, Zhou Y, Li Y, Wang C, Zhang F, Chong X, Zhang D, Chen H. 2023. Floral scent components of the hybrids between Lagerstroemia fauriei and Lagerstroemia ‘Tuscarora’. Scientia Hortic. 309:111670. https://doi.org/10.1016/j.scienta.2002.111670.