Materials and methods

Fifteen cultivars and two unnamed accessions representing four species were chosen for the trial (Table 1). Plants were purchased from Nurseries Caroliniana (North Augusta, SC) in Apr. 2015 and immediately potted into 14.6-L containers (C2000; Nursery Supplies, Chambersburg, PA), with the exception of two large fortune’s osmanthus cultivars that were potted into 23.0-L containers (C2800; Nursery Supplies). Growing media consisted of pine bark amended with 6.6 kg·m⁻³ 19N–2.1P–7.4K controlled-release fertilizer (Osmocote Pro; Scotts-Sierra Horticultural Products Co., Maryville, OH), 0.6 kg·m⁻³ micronutrient fertilizer (Micromax; Scotts-Sierra Horticultural Products Co.), 0.6 kg·m⁻³ iron sulfate (Sprint 330; BASF Co., Florham Park, NJ), and 0.2 kg·m⁻³ magnesium sulfate (Epsom salt; PQ Corp., Joliet, IL).

Table 1.

Sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars represented in McMinnville, TN cultivar trial.^z

View Table

Plants were moved to the trial site 1 May 2015. Trial site consisted of a pot-in-pot production system with 23.0-L socket pots in 4-ft-wide rows separated by 4-ft grassy alleys. The 14.6- or 23.0-L pots containing the plants were placed inside the socket pots. Native soil was backfilled in socket pots containing 14.6-L pots to steady and insulate pots. The experimental design was a completely randomized design. Within-row spacing was 4 ft. Plants were top-dressed with 72 g of 19N–2.1P–7.4K controlled-release fertilizer (Osmocote Pro) in Aug. 2015 and Apr. and Aug. 2016 and 2017. Drip irrigation was used as needed throughout the growing season.

Six plants of each cultivar were included in the trial. Within each cultivar, one-half of the plants were assigned to a protected overwintering treatment. These plants were removed from the trial site to a 48-ft-long × 24-ft-wide × 10.8-ft-tall covered elliptical coldframe ≈15 Nov. each year of the study. The coldframe was covered with 6-mil clear plastic and provided with supplemental heat via a gas-fired unit heater to maintain at least 39 °F. Protected plants were watered once per week for 10 min; no supplemental irrigation was provided to exposed plants. Protected plants were brought back to the trial site 15 Apr. each year of the study. The study began 15 May 2015 and covered three growing seasons, ending 25 Oct. 2017. Weather data during the study period were collected from the National Environmental Satellite, Data, and Information Service McMinnville, TN Station [USC00405882 (lat. 35.6723°N, long. 85.7810°W, elevation 940 ft)].

Plants were measured for growth in 2015 (15 May and 20 Oct.), 2016 (16 May and 20 Oct.), and 2017 (22 May and 25 Oct.). Measuring took place in spring after winter damage was manifest in shoots and foliage. Shoot height was measured to the tallest node with foliage. A growth index (GI) was calculated for each plant according to the formula πhr², where h is shoot height in centimeters, r = 0.5d, and d is the mean of two perpendicular diameter measurements in centimeters (Lindstrom et al., 2001). Percent increase in GI was determined for each plant by the formula: [(GI_t − GI_initial) ÷ GI_initial] × 100%, where GI_intial in the GI value at the start of the study period and GI_t is the GI at the time of measurement. Necrotic stem was measured from the first node with foliage to the shoot tip of the tallest three shoots. Percent necrotic stem was determined for each shoot by the formula: percent necrotic stem ÷ total stem length × 100%. Total stem length was determined by measuring from the base of the plant to the shoot tip. Flowering was observed every 2 weeks from the onset of flowering of the earliest cultivar to the termination of flowering of the latest cultivar. Numbers of apical and lateral flowers were counted on each plant and flower morphology (male, female, or perfect) was noted.

Data analysis was performed using SAS software (version 9.4 for Windows; SAS Institute, Cary, NC). Fisher’s exact test of independence was used to determine whether winter mortality was independent of cultivar for winter-exposed accessions. For each treatment (winter-protected or winter-exposed), the general linear model procedure (PROC GLM) was used to partition variance in GI and percent necrotic stem into sources attributable to cultivar and error. Constancy of residual variance was checked using the Brown–Forsythe test. All variables satisfied analysis of variance assumptions except percent growth in 2017. Percent growth in 2017 data were transformed using square root, which alleviated nonconstant variance; nontransformed means are presented in tables and figures for clarity. Means for each cultivar were compared using Tukey’s Studentized range test with an α = 0.05 significance level.

Results

The minimum temperatures experienced by unprotected plants in the first winter (2015–16) and second winter (2016–17) were 9 and 5 °F, respectively (Table 2). The first winter had 36 d in January and February during which the minimum temperature was below 32 °F. In the second winter, only 13 d in January and February were below 32 °F. There was no day during the study during which minimum temperatures went below 0 °F. Both winters were warmer than the 30-year normal (National Oceanic and Atmospheric Administration, 2018). In the first winter, mean maximum and mean minimum monthly temperatures exceeded the 30-year normal by an average of 4.2 and 6.0 °F, respectively. Mean maximum and minimum monthly temperatures in the second winter exceeded the 30-year normal by an average of 7.0 and 5.0 °F, respectively.

Table 2.

Surface air temperatures and precipitation in McMinnville, TN, during the winters of 2015–16 and 2016–17.^z

View Table

Cultivar was not independent of post-winter mortality (χ² = 51.0, P < 0.01). ‘Fodingzhu’ fragrant tea olive had significantly greater mortality than any other cultivar, with two of three winter-exposed accessions dying during the first winter (2015–16). One of three winter-exposed ‘Jim Porter’ sweet olive and one of three winter-exposed ‘Beni Kin Mokusei’ fragrant tea olive died during the 2017 growing season, likely from very wet conditions after heavy rainfall. Four of six ‘Variegatus’ holly tea olives died during the study period, split equally between winter-exposed and winter-protected plants. The two ‘Variegatus’ accessions that died were the only winter-protected plants that died during the study period.

There were significant differences in cold-hardiness among cultivars as measured by percent necrotic stem observed in May of each year (Table 3). Mean percent necrotic stem measured in May 2017 for sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive was 1.1%, 2.7%, 44.8%, and 20.2%, respectively. Each species except fragrant tea olive had at least one cultivar with 0% stem death each year.

Table 3.

Percent necrotic stem for sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars grown in a winter-exposed pot-in-pot production system in McMinnville, TN.^z

View Table

Initial GI ranged from 2170 cm³ for ‘Kaori Hime’ holly tea olive to 441,015 cm³ for the unnamed fortune’s osmanthus, reflecting the size of these plants at the time of purchase (Table 4). Final GI for winter-exposed plants ranged from 26,610 cm³ for ‘Variegatus’ holly tea olive to 1,723,232 cm³ for the fortune’s osmanthus seedling. Final GI for winter-protected plants ranged from 45,263 cm³ for ‘Variegatus’ holly tea olive to 2,811,047 cm³ for ‘Apricot Echo’ fragrant tea olive.

Table 4.

Growth index (GI) for sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars represented in McMinnville, TN, cultivar trial.

View Table

Percent increase in GI for winter-exposed sweet olive, fortune’s osmanthus, and fragrant tea olive after three growing seasons ranged from −18.9% for ‘Beni Kin Mokusei’ fragrant tea olive to 2871% for ‘Longwood’ fortune’s osmanthus (Table 5). Mean percent increase in GI after three growing seasons for winter-exposed accessions of sweet olive, fortune’s osmanthus, and fragrant tea olive were 867%, 1175%, and 155%, respectively. Percent increase in GI for winter-exposed holly tea olive after three growing seasons ranged from −79.3% for ‘Ogon’ holly tea olive to 27,121% for ‘Kaori Hime’ holly tea olive (Table 4). Mean percent increase in GI after three growing seasons for winter-exposed accessions of holly tea olive was 6361%. After 3 years in a winter-exposed pot-in-pot production system, there were significant differences in percent growth increase among cultivars (Table 4). Accessions with the largest growth increase were the holly tea olive cultivars Kaori Hime, Hariyama, Shien, and Head-Lee Fastigate, followed by San Jose fortune’s osmanthus, Rotundifolius holly tea olive, and Longwood sweet olive.

Table 5.

Percent increase in growth index (GI) for sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars represented in McMinnville, TN, cultivar trial.

View Table

Flowers were observed at least once during the study period for all cultivars except Beni Kin Mokusei, Goshiki, and Head-Lee Fastigate. Flowering of trial cultivars lasted from September to November for winter-exposed plants and from September to March or April for winter-protected plants (Fig. 1). Cultivars of sweet olive and fortune’s osmanthus bloomed most reliably after winter exposure. Three of four fragrant tea olive cultivars bloomed after the first winter, whereas only one of four bloomed after the second winter. Only one of eight holly tea olive cultivars flowered after winter exposure. Winter exposure greatly reduced the length of the flowering period and the number of flowers. Length of flowering for winter-exposed sweet olive and fortune’s osmanthus was reduced by an average of 45% while flowering time of fragrant tea olive was reduced by an average of 75% compared with winter-protected plants (Fig. 1). Number of flowers was also greatly reduced in winter-exposed plants compared with winter-protected plants for all cultivars except San Jose sweet olive (data not shown).

Flowering of sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars represented in McMinnville, TN, cultivar trial. Length of flowering period is represented by horizontal bars for winter-protected (dark blue) and winter-exposed (light blue) accessions. Winter-protected plants were placed in a heated coldframe in November and returned to the pot-in-pot production system in April of each year. Winter-exposed plants remained in the pot-in-pot production system year-round; n = 3 plants/cultivar/winter location except ‘Fodingzhu’ (n = 1 winter-exposed accession), ‘Beni Kin Mokusei’ and ‘Jim Porter’ (n = 2 winter-exposed accessions), and ‘Variegatus’ (n = 1 winter-exposed accession and n = 1 winter-protected accession).

Citation: HortTechnology hortte 29, 1; 10.21273/HORTTECH04166-18

• Download Figure

View Full Size

Flowering of sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars represented in McMinnville, TN, cultivar trial. Length of flowering period is represented by horizontal bars for winter-protected (dark blue) and winter-exposed (light blue) accessions. Winter-protected plants were placed in a heated coldframe in November and returned to the pot-in-pot production system in April of each year. Winter-exposed plants remained in the pot-in-pot production system year-round; n = 3 plants/cultivar/winter location except ‘Fodingzhu’ (n = 1 winter-exposed accession), ‘Beni Kin Mokusei’ and ‘Jim Porter’ (n = 2 winter-exposed accessions), and ‘Variegatus’ (n = 1 winter-exposed accession and n = 1 winter-protected accession).

Citation: HortTechnology hortte 29, 1; 10.21273/HORTTECH04166-18

• Download Figure

Flowering of sweet olive, fortune’s osmanthus, fragrant tea olive, and holly tea olive cultivars represented in McMinnville, TN, cultivar trial. Length of flowering period is represented by horizontal bars for winter-protected (dark blue) and winter-exposed (light blue) accessions. Winter-protected plants were placed in a heated coldframe in November and returned to the pot-in-pot production system in April of each year. Winter-exposed plants remained in the pot-in-pot production system year-round; n = 3 plants/cultivar/winter location except ‘Fodingzhu’ (n = 1 winter-exposed accession), ‘Beni Kin Mokusei’ and ‘Jim Porter’ (n = 2 winter-exposed accessions), and ‘Variegatus’ (n = 1 winter-exposed accession and n = 1 winter-protected accession).

Citation: HortTechnology hortte 29, 1; 10.21273/HORTTECH04166-18

• Download Figure

Flower morphology was variable both between and within species. Both sweet olive cultivars produced perfect flowers, whereas all three fortune’s osmanthus cultivars produced male-only flowers. Flowers of ‘Hariyama’, ‘Rotundifolius’, and ‘Shien’ holly tea olive were perfect, whereas ‘Kaori Hime’, ‘Ogon’, and ‘Variegatus’ holly tea olive produced male-only flowers. Morphology and function was difficult to determine for fragrant tea olive, where pistils can range from slightly to completely reduced. Male-only flowers were observed on ‘Fodingzhu’ throughout the study period, whereas ‘Apricot Echo’ produced male-only flowers the first winter and perfect flowers the second winter.

Discussion

McMinnville, TN, is in a transition zone between USDA Hardiness Zones 6 and 7, and results from evaluations made in transition zones often can be used over a wide geographic and climatic area (Fare, 2017). The plants in this pot-in-pot production system likely experienced colder conditions than in-ground plants, so these interpretations may be considered conservative for this region.

Although plants in this experiment were all 1-year-old rooted cuttings, there were large differences in initial sizes, as reflected in the initial GI of each plant measured at the onset of the experiment. By the end of the experiment, GI rank among cultivars had changed such that many of the smallest accessions were in the midrange of size. There was no correlation between initial plant size and percent growth increase, supporting the hypothesis that growth increase or lack thereof was due to adaptability in this environment rather than inherent differences in growth rate.

As expected, fragrant tea olive was the least cold-hardy species in the trial. Several of the winter-exposed fragrant tea olive accessions died, and all had some stem death from winter injury. However, there was significant variation among cultivars. On the basis of mortality and stem death, fragrant tea olives in order of least to most cold-hardy were ‘Fodingzhu’, ‘Apricot Echo’, ‘Beni Kin Mokusei’, and the unnamed seedling accession. While two of the three ‘Fodingzhu’ died the first winter, the single remaining ‘Fodingzhu’ was the only fragrant tea olive accession to flower each year under winter-exposed conditions even though stem mortality was observed each spring. These differences underscore the need to import and test more fragrant tea olive from China and Japan, where more than 160 cultivars have been named (Xiang and Liu, 2008). Considering that there are only a dozen or so cultivars available in the U.S. market, substantial progress in recommending cold-hardy selections could be made simply by importing and testing more cultivars. In addition, the hardiness of the seedling over the cultivars indicates that improvement for cold-hardiness can be made by breeding and selecting within the species.

Accessions with the largest growth increase in winter-exposed conditions were the holly tea olive cultivars Kaori Hime, Hariyama, Shien, and Head-Lee Fastigate, followed by San Jose fortune’s osmanthus, Rotundifolius holly tea olive, and Longwood sweet olive. Of these cultivars, Kaori Hime, San Jose, and Longwood flowered under winter-exposed conditions. ‘Beni Kin Mokusei’ fragrant tea olive and ‘Ogon’ holly tea olive both showed negative increases, meaning stem death and lack of growth made their GI values smaller by the end of the study period than at the beginning.

Three kinds of breeding systems are found within the Oleaceae: hermaphroditism, androdioecy, and dioecy (Hao et al., 2005; Xu et al., 2014). Flowers of sweet olive, fortune’s osmanthus, and holly tea olive were perfect (i.e., hermaphrodites possessing functional male and female organs) at all observation points during the study period, whereas fragrant tea olive flowers were perfect or male-only at different times depending on the cultivar. The male-only flowers contained pistillodes with different degrees of degradation compared with true pistils, often making it difficult to classify flower morphology without a microscope (Xu et al., 2014). Evidence for cryptical dioecy, where the hermaphrodite is functionally female, was observed as viable pollen did not naturally dehisce from the anthers of hermaphrodite flowers of ‘Apricot Echo’ fragrant tea olive. Cultivars producing male-only flowers appear to be more numerous than cultivars that produce hermaphroditic flowers. Nine of 62 fragrant tea olive cultivars studied by Duan et al. (2013) produced hermaphroditic flowers and the remainder were male-only. The presence of both male and hermaphrodite flowers in fragrant tea olive has been interpreted as a step in the evolutionary pathway from hermaphroditism to dioecy (Xu et al., 2014).

Further breeding improvement is necessary in fragrant tea olive to increase cold-hardiness. Many of the 120 cultivars described by Xiang and Liu (2008) (including cultivars that produce hermaphroditic flowers) should be tested at multiple locations in the United States. Producing new fortune’s osmanthus hybrids as well as hybrids of fragrant tea olive and sweet olive will provide variation for multisite testing. The close relationship between tea olive and fringetree (Chionanthus) indicates the possibility of wide hybridization as a breeding strategy (Yuan et al., 2010). Tools to select parents and verify hybrids, including molecular markers and a transcriptome data set for identification of important genes, are now available for fragrant tea olive and its relatives (Alexander, 2016; Alexander et al., 2017; Duan et al., 2013; Mu et al., 2014).

In summary, the most-cold tolerant cultivars (based on mortality and stem death) were Kaori Hime, Hariyama, Shien, and‘Head-Lee Fastigate holly tea olive, San Jose fortune’s osmanthus, Rotundifolius holly tea olive, and Longwood sweet olive. Of these cultivars, Kaori Hime, San Jose, and Longwood flowered under winter-exposed conditions. Fragrant tea olive cultivars were the least cold-hardy species based on winter mortality, winter stem damage, and low to negative growth increases. Evaluation and breeding efforts are underway to extend the range for production and growth of this genus.

Units

View Table