Webinar Date: July 23, 2026
Speakers: Natalie Bumgarner, Associate Professor and Extension Specialist in Residential and Consumer Horticulture, University of Tennessee; Celeste Scott, Commercial Horticulture Specialist for West Tennessee, UT Extension
Recorded introduction: Narrated by Bethany Harris, then Whitney Ottinger
Moderator: Gabrielle LaTora, Agriculture and Natural Resources Agent, Fulton County, Georgia
Duration: 30:30
Segment: 03 of 08
Series: Cut Flower Symposium
CEU Categories: [PENDING]
About This Session
Celeste Scott opens with how Tennessee got here. Interest in cut flower production kicked in during what she calls the all-famous COVID era. A grant through the UT Center for Profitable Agriculture let the team look into specialty crops gaining traction in the state and paid for travel, including a trip to Utah to meet Melanie, who speaks later in the same symposium. The meet and greets and workshops that followed packed the house and turned people away, the interest carried into the statewide Pick Tennessee Conference, the contact list of Tennessee growers passed 300, and a state association began forming around marketing, budgeting, and education. Three years ago Scott, Natalie Bumgarner, and their team won a specialty crop block grant and started the project this talk reports on. Bumgarner describes what it bought: the plant people wanted to put plants in the ground and count them and fertilize them differently, and that half runs alongside work with the farm management team and colleagues in agricultural economics on consumer perspective research, surveys, and enterprise budgets.
The trials were built around what Tennessee growers were already doing. Scott says warm season cut flowers are the gateway flower into cut flower production, easy from seed and a good group to get your footing with, and a large influx of new growers had arrived in exactly that space. Farm visits and conversations told the team where growers wanted support, and that shaped the trials. Replications ran in Knoxville and Jackson, about six hours apart, with different soils and temperature climates. Scott grew on fabric and Bumgarner on plastic, a split made to cut down on time. The first year covered focal flowers, filler flowers, and greenery at once, tracking stems per square foot and stem lengths, and Scott calls the data collection overwhelming to say the least.
Fertility was the main variable. Scott says the only guidance available for the region comes from the University of Maryland, which breaks fertility into three categories. They set low at one pound of nitrogen per thousand square feet, then two pounds for medium and three for high, delivered as a 20-10-20 soluble fertilizer through a Dosatron. Year one compared low against medium, which let them narrow to the plant types that were responding. Year two compared low against high. The current year runs a no-fertility control alongside low, medium, and high.
Zinnias came first and gave the clearest response. Bumgarner shows the Knoxville data: Benary’s Giants carry larger flowers on fewer stems per square foot at high quality, and there was potential for increase as they moved up the fertility scale, holding in year two at the high rate. Scott’s Jackson counts ran lower for two reasons she names. She was cutting much larger stems at the very beginning, before her harvest practice settled to something representative, and her plot was a brand new prepared site where Bumgarner had been planting into organic farm ground that had grown plants for years and carried existing nutrients and loosened soil. The trend was the same at both sites. The 2025 numbers show a dip on Queeny Lime Orange that Scott explains as a septic drainage field problem, which moved plots in year two and left lower quantities in the zone where the high fertility treatments sat.
The zinnia takeaways are practical. Increasing nitrogen produced differences, and the trend suggests higher levels could keep increasing stem counts, which is why the final year looks at the full picture. Zinderellas were their highest cull percentage. Queeny Limes are good but have to be harvested at peak for best color or they look dirty and dingy, which can turn a grower off in the middle of the warmest part of the season. Late season disease and insect pests remain a recurring problem on zinnias.
Celosia followed. The team started with wheat types, plume types, and coxcomb types, mixing industry standards with tried and trues as they got experience. Bumgarner reports the largest fertility response in the plume types, and notes that compared with zinnia, which had a more consistent fertility impact across two years, the more vigorous celosia showed more variability. Scott explains why there is no Jackson celosia data to compare. In 2024 she planted too early because the transplants were getting root bound, and both constricted root growth and cool night temperatures can induce early flowering, so the numbers were not reliable. In 2025 her plot, which sits at a production row crop research facility, took off-target herbicide damage from a neighboring plot, since rectified. Both years went out the window.
The celosia cultivar takeaways survived anyway. Sunday Green gave reliably long, thin stems, which matters because fasciation can leave a grower with a beautiful bloom on a stem too thick to fit what it needs to fit. Sunday Orange, added the following year, bloomed well but the stem standards were not as comparable. Flamingo Feather was a favorite filler, with super long stems and a reliable re-bloom several weeks after cutting it way down. Pampas Plume was good on quality and versatility. Kurume, an older cultivar they added last year, gave a smaller coxcomb form with enough shape variability that it was always fun to see what it would produce.
Sunflowers changed direction. Scott ran them on fertility in year one and then realized every cut was a single stem sunflower, so fertility was unlikely to move the number of stems per plant. What she did see was a lot of variability in flower head size, and reliable single cuts from healthy transplants. From year two she switched to spacing, aiming to push head size in a predictable direction while using planting templates growers already have, since it is complicated to run different planting standards between cropping systems. Three treatments: one plant on six-inch spacing, close to the industry standard of four to six inches; two plants per planting hole on six-inch centers, which brings the effective population down to a three-inch pattern; and two plants per nine-inch centers, because many larger warm season annuals already go on nine-inch fabric templates.
The spacing results are the sharpest numbers in the talk. Last year ran two successions with high first-succession counts, which Scott explains as a moist spring that made single-stem plants branch after the first cut, producing stems still above the pencil-diameter threshold she counted as marketable. The diameter distribution is what she wanted. One plant on six inches gave a fairly even spread of heads below four inches and above four inches, plus a high percentage above six inches, which are less versatile for mixed bouquets. Two plants on six-inch centers produced the highest percentage of heads below four inches measured petal to petal, not just the disc. Two plants on nine-inch centers gave the highest four-inch-and-below percentage and the lowest percentage at six inches and above. She wants to replicate that this year so growers can plan how many plants each succession needs.
Her sunflower takeaways: no cultivar they trialed was one they were set against, they were reliable and resilient, and with a good planting schedule a grower gets harvest across a long stretch of the season, starting in some parts of Tennessee as early as Mother’s Day. Change the color palette to match warm or cool tones as the season moves. Insect damage was primarily foliage feeding, with flower head and petal damage on the lighter colored flowers. Scott then shows what she calls the And More, the other crops they trialed beyond celosia and zinnia, and says the trialing summary is being assembled now for publication as production guides.
Bumgarner takes the marketing half. A grower survey came first, then a consumer survey whose respondents had all bought cut flowers in the year prior, two-thirds of them buying local, so the audience is a real customer base rather than a general population. Most were Tennessee residents with a good share of the rest of the southeastern states. From the grower side, farmers markets and direct to consumer are the highest outlets, with pick your own and some CSAs behind them, and respondents could pick more than one. From the consumer side, asked about cut flowers generally, the grocery store number is very high while far fewer growers reported selling there. Add the word local and the grocery store share drops while mass merchandisers, wholesale clubs, local florists, and farmers markets pick up. Even so, a large share of local cut flower customers still buy at grocery stores, which Bumgarner reads as a point about accessible locations rather than preference.
The purchase decision is where she spends the most time. About 62 percent were buying for gifts, a little less than 40 percent in the growing-for-themselves category, and 40 percent or higher reported an impulse purchase. Asked at what point they decided to buy, a third of what she calls the prepared audience knew before entering the shopping area, 50 percent knew when they saw it, and 16 percent do not recall deciding at all. Those last two groups are the audience a grower can capture on placement. On attitudes, consumers want to support local producers and largely understand local cut flowers as grown within their state, but a number still say local flowers are not available where they live or are not something they think about. The barriers question puts high price first and low availability second, and Bumgarner treats availability as the one worth working on. Willingness-to-pay surveys show that as the share of local flowers in a bouquet rises, consumers will pay an increasingly large amount for it, with more of that work coming in the fall.
She closes on the tools. Five warm season annual enterprise budgets are already on the UT website, built from grower visits and the team’s own field trial data, with Tennessee growers as reviewers and ground truthers, and a QR code on the slide points to them. She asks for feedback and additions and wants to add more plants to the palette. Gabrielle LaTora then brings two questions from the chat, both answered by Scott, with an addition from Bumgarner on the second.
Chapters (Video Timestamps)
0:00 Recorded Introduction: Bumgarner and Scott
0:56 A Multi-State Program, and How Tennessee Got Here
3:13 The Specialty Crop Block Grant and What It Funded
4:11 Building the Trials Around What Growers Grow
5:15 First-Year Scope and the Data That Overwhelmed Us
6:13 Fertility Treatments and the Maryland Baseline
7:19 Zinnias: Zinderellas, Benary’s Giants, Agaves, Queenys
8:59 Jackson Zinnia Results and Why They Read Lower
10:14 Zinnia Takeaways: Nitrogen, Culls, Harvest Timing
10:56 Celosia Types and the Knoxville Fertility Data
12:34 Why the Jackson Celosia Data Was Thrown Out
13:28 Celosia Cultivars: Sunday Green to Kurume
14:37 Sunflowers: From Fertility to Spacing
15:29 Three Spacing Treatments, Six-Inch to Nine-Inch
16:39 How Spacing Changed Flower Head Diameter
18:34 Sunflower Takeaways and the And More
19:40 Trialing Summary and Production Guides
20:01 Marketing: The Grower and Consumer Surveys
22:05 When and Where Consumers Buy
24:04 Barriers, Willingness to Pay, and the Budgets
26:03 Closing and Handoff to the Moderator
26:38 Q&A: Leaf-Footed Bugs on Sunflowers
28:04 Q&A: Where to Start, and What to Plant
Questions and Answers
Q: Why run the same trials in two locations instead of one?
A: Scott says Knoxville and Jackson sit about six hours apart, with different growing conditions, different soil types, and different temperature climates. Running replications in both was how they meant to support growers across the whole state rather than one corner of it. The two sites also grew under different mulch: Scott used fabric and Bumgarner used plastic, a choice they made to cut down on time.
Q: What fertility rates did they compare, and how were they delivered?
A: Scott says the only guidance they had for the area came from the University of Maryland, which breaks fertility into three categories. They set low at one pound of nitrogen per thousand square feet, medium at two pounds, and high at three. The nitrogen went out as a 20-10-20 soluble fertilizer through a Dosatron injector. Year one compared low against medium, year two compared low against high, and the current year adds a no-fertility control so all four levels run side by side.
Q: Did zinnias respond to added nitrogen, and what did the two sites show?
A: Bumgarner reports Knoxville trends toward more stems per square foot as they moved up the fertility scale, with those trends holding when they went to the high rate in year two. Benary’s Giants carry larger flowers and produce fewer stems per square foot, but the quality is high. Scott’s Jackson counts came in lower for two reasons she names: she was cutting much larger stems at the start of the trial before her harvest practice settled, and her plot was a brand new prepared site while Bumgarner was planting into organic farm ground that had been in production for years and carried existing nutrients and loosened soil. The trend was the same at both sites.
Q: Which zinnias gave them trouble, and why?
A: Scott says Zinderellas were their highest cull percentage. Queeny Limes are good flowers but have to be harvested at peak for best color, or they look dirty and dingy, which can turn a grower off, especially in the middle of the warmest part of the season. She also notes late season disease and insect pests as a recurring problem on zinnias generally.
Q: What happened to the Jackson celosia data?
A: Scott calls herself a poor celosia grower for those two years and gives the reasons. In 2024 she planted too early because the transplants were getting root bound, and both constricted root growth and cool night temperatures can induce early flowering, so the numbers were not reliable. In 2025 her plot, which sits at a production row crop research facility, took off-target herbicide damage from a neighboring plot, since rectified. Both years of Jackson celosia data were discarded. The celosia fertility numbers in the talk are Bumgarner’s Knoxville data.
Q: Which celosia cultivars performed, and what were they good for?
A: Scott says Sunday Green gave reliably long, thin stems, which matters with celosia because fasciation can produce a beautiful bloom on a stem too thick to fit what you need. Sunday Orange, added the following year, bloomed well but the stem standards were not as comparable. Flamingo Feather was a favorite filler with very long stems and re-bloomed several weeks after a hard cut. Pampas Plume was good for quality and versatility. Kurume, an older cultivar they added last year, gave a smaller coxcomb form with variability in shape. Bumgarner adds that the largest fertility response across the celosia types showed up in the plume types, and that the more vigorous celosia showed more variability than zinnia did.
Q: Why did they stop testing fertility on sunflowers?
A: Scott says they ran sunflowers on fertility in year one and then realized every cut was a single stem sunflower, so fertility was unlikely to change the number of stems they got per plant. What they did see was a lot of variability in flower head size. From there they switched to spacing treatments, aiming to push head size in a predictable direction while using plant spacing templates growers already have for other crops.
Q: Which sunflower spacing produced the most usable flower heads?
A: Scott compared one plant on six-inch spacing, two plants per hole on six-inch centers, and two plants per hole on nine-inch centers. The single plant on six inches gave a fairly even distribution of heads below four inches and above four inches, plus a high percentage above six inches in diameter, which are less versatile for mixed bouquets. Two plants on six-inch centers, which brings the effective population down to a three-inch pattern, gave the highest percentage of heads under four inches measured petal to petal. Two plants on nine-inch centers gave the highest percentage under four inches and the lowest percentage at six inches and above. She wants to replicate that result this year.
Q: Where do Tennessee growers sell, and where are consumers actually buying?
A: Bumgarner puts the two survey sides next to each other. From the grower survey, farmers markets and direct to consumer are the highest outlets, with pick your own and some CSAs behind them, and respondents could select more than one option. From the consumer survey, asked about cut flowers generally, the grocery store number is very high, while far fewer growers reported selling at grocery stores. When the word local is added to the question the grocery store share drops and mass merchandisers, wholesale clubs, local florists, and farmers markets pick up, but a large share of local cut flower customers are still buying from grocery stores, which she reads as a point about highly accessible locations.
Q: What did the consumer survey say about how the purchase decision gets made?
A: About 62 percent were buying for gifts. Bumgarner puts a little less than 40 percent in the growing-for-themselves category, and 40 percent or higher reported the purchase as an impulse. Asked at what point they decided to buy, a third knew what they wanted before entering the shopping area, 50 percent knew they wanted it when they saw it, and 16 percent do not recall deciding at all. She treats the middle and last groups as the audience a grower can capture on accessibility. Willingness-to-pay surveys separately show that as the share of local flowers in a bouquet rises, consumers are willing to pay an increasingly large amount for it.
Q: My sunflowers are getting hit by leaf-footed bugs. Any organic options?
A: Scott answers this one from the chat. Most of the insect feeding damage she saw on sunflowers was on foliage, and flower head or petal damage showed up mainly on the lighter colored flowers. If blooms are harvested at the right time the petals should not be fully expanded yet, so a grower may be able to forgo treatment even with pest insects in the field. If pressure warrants control, she points to horticultural oils and insecticidal soaps, with three conditions: coverage has to be excellent for those products to work, they leave little to no residual so they need reapplying regularly following label directions, and they can cause phytotoxic issues in warm season temperatures.
Q: I’m starting cut flowers in middle Georgia. What size garden should I start with, and what should I plant?
A: Scott repeats that warm season cut flowers are the gateway for new growers, mostly easy from seed, and that they fit different setups. Her team started transplants indoors because the trials needed certain plant population counts, but many Tennessee growers direct seed warm season annuals. She recommends a combination of filler, foliage, and focal: zinnias as a start, celosia as either focal or filler depending on cultivar, gomphrena as a light filler with the caution that it takes time to harvest, Frosted Explosion as an annual panicum grass for foliage and filler, ornamental millets that come into their own later in autumn, and basil, which has to be cut after the stems begin to feel woody or it wilts past reviving. Bumgarner adds that the UT enterprise budgets are set up to grow about a quarter acre across five different cultivars, on the principle that nobody starting out needs an acre of zinnias.
Resources Named on Camera
- UT enterprise budgets for five warm season annuals, on the UT website, built from grower visits and the team’s own trial data with Tennessee growers as reviewers. Bumgarner shows a QR code at 25:36 and asks for feedback and additions. No web address is spoken on the recording, so the link has to be supplied by the speakers or the symposium organizers before this page is published.
- Production guides from the trialing work, described at 19:48 as being assembled and headed for publication. Not yet available at the time of the talk.
- University of Maryland cut flower fertility guidance, named at 6:16 as the only guidance available for the region and the basis for the low, medium, and high structure used in the trials.
- UT Center for Profitable Agriculture, source of the first grant, which funded the specialty crop scouting and the travel that included the visit to Utah.
- Pick Tennessee Conference, the statewide conference the local cut flower interest carried into.
- A Tennessee cut flower growers association, described at 2:59 as in development, to support growers on marketing, budgeting, and education.
Publication Notes
Two presenters, and the site each number came from is part of the claim. Scott ran the Jackson plots, the trial design, the cultivar takeaways, and the sunflower spacing work. Bumgarner ran the Knoxville data, the grower and consumer surveys, and the enterprise budgets. Every result on this page is attributed to the speaker who reported it, and the celosia section says outright that the fertility numbers are Knoxville’s, because the Jackson celosia data was discarded for both years. Anyone editing this page should keep that split intact.
One rate was left as spoken in the corrected transcript. At 6:29 Scott sets low fertility at one pound of nitrogen per thousand square feet. At 6:34 she gives medium as two pounds and high as three pounds “per square foot.” Stage 1 preserved that rather than normalizing it. This page states all three rates on the per-thousand-square-feet basis established in the preceding sentence, which is the reading the passage supports.
One phrase in the spacing section is still pending audio confirmation. At 15:18 Scott refers to spacing templates people may already be using in “burnt fabrics,” preserved at Stage 1 as probable trade usage for pre-burned landscape fabric. Nothing on this page rests on it. The spacing answers use her clean phrasing at 15:55, “fabric templates that they already have.”
All flagged terms are resolved. Seven items in the corrected transcript were marked for audio verification at merge and all seven were closed against the recording on 2026-08-09, before any derived file existed. Four of the seven were cases where both engines agreed and the agreement did not read; two of those four turned out wrong. Two resolutions are load bearing on this page: the cull percentage at 10:36 is a trial result about Zinderellas, and Kurume is the celosia cultivar named at 12:03 by Bumgarner and again at 14:19 by Scott.
Processed for UGA Center for Urban Agriculture / Cut Flower Symposium Archives
Source: Stage 2 Archive Package, GTBOP_Archive_Summary_2026-07-23_BumgarnerScottUTResearch.md
Chain of custody: corrected SRT (Stage 1), GTBOP_Transcript_2026-07-23_BumgarnerScottUTResearch.srt (402 blocks, 6,226 words), verified against Natalie and Celeste.mp4 at 1830.719 s
Scaling Note
Reference point. Segment 02 ran 16:37 and produced a 17.3 KB website file from a 15.7 KB archive summary. This segment runs 30:30.7, which is 1.84 times as long, with an 18.2 KB archive summary.
Delivered: 25.6 KB, a factor of 1.48 on segment 02 against a runtime factor of 1.84. Sub-linear in runtime, which is the direction Stage 3 is supposed to move. Where it went:
| Section | Bytes | Note |
|---|---|---|
| About This Session | 10.3 KB | About twice segment 02’s. Three crops, two research sites, and a two-part survey |
| Questions and Answers | 8.2 KB | Twelve pairs carried from Stage 2 unchanged |
| Chapters | 1.2 KB | 23 markers, unchanged |
| Resources, Publication Notes, frontmatter | 5.6 KB | Comparable to segment 02 |
About This Session is the section that grew, and it grew about twofold, not fourfold. It carries the crop-by-crop account for a reader who will not watch 30 minutes, and it is the only section where a longer talk legitimately buys more words. Everything else scales with the Stage 2 package or not at all.
Nothing was added. Every claim on this page traces to the Stage 2 package, which traces to the corrected SRT.