Seeds and cuttings I collect around North Carolina, what they turned out to be, and how to grow them for a NC Piedmont native-plant nursery.
Chionanthus virginicus · Oleaceae
Ripe now, and the birds are already working it. Half the trees are male and will never carry a single drupe.
Community photos of Chionanthus virginicus to study before I head out: leaf, flower or bract, ripe fruit, bark, whole plant.
Native and secure across NC, no conservation status. The real constraint is not rarity but sex: it is functionally dioecious, and males carry the showier longer-petaled flowers. So the tree that stops you in April is statistically the one that gives you nothing in August. Flowering visit for the ID and the sex, fruiting visit for the seed
Best bet: Seed, and only seed. Clean the same day, since letting the pulp ferment delays germination. Then 4 months warm-moist at 21–26.5°C followed by 4 months cold-moist at 3–5°C: Flanagan and Witte showed 4 plus 4 more than doubles the germination of the standard 3 plus 3. Nicking a chip off the hilum end converts stalled radicles into actual shoots. Never freeze fresh seed. Cuttings are among the hardest in the trade, 0.5% to 20% from mature wood, so don't plan on them
Piedmont native and this is home range. Zones 3a–9b, tolerates clay, pH 6.0–8.0, and it flowers more freely in hot-summer regions, so our heat is an asset. Two real constraints. Water use is high and drought stress is the trigger for borers, so irrigate liners hard through July and August. And deer browse is the likely killer: NC Extension rates it only moderately resistant, and plants may die when more than a third of the annual growth is removed, which at 6–10 in a year is not much. Cage every liner. Also expect spring calls about a dead plant, since it is one of the last things to leaf out
Each look-alike with the one character that settles it. Tap a photo to zoom.
The highest-volume false positive here. Same family, opposite simple leaves, white panicles, clusters of black olive-like drupes, and it lines every Piedmont bottom. Flower size decides it: privet's corolla is a short 4-lobed tube a few mm across, against fringetree's thread-like straps 15–30 mm long, roughly ten times longer. Backup: privet leaves are 1–2 in, papery, semi-evergreen and still green in January; fringetree leaves are 3–8 in and fully deciduous
Strongest habitat overlap on the list: 8–20 ft understory shrub of the same upland hardwood forest and ravines, with drooping white racemes in Apr – May, and it fruits Aug – Sep alongside fringetree. Leaves are alternate, with star-shaped stellate hairs beneath; fringetree is opposite with simple hairs. Backup: Styrax flowers are 5-parted with reflexed petals and protruding yellow stamens, and its fruit is a dry brown drupe with 2–3 seeds rather than a fleshy blue one
Carolina Nature names this comparison on the fringetree page, so it is the real vegetative confusion: opposite or whorled, simple, entire leaves 4–6 in, the same leaf gestalt in the same size class. Pubescence decides it. Buttonbush leaves and young branches are smooth; fringetree carries dense white hairs on both. Backup: buttonbush makes 1.5 in spherical pincushion heads Jun – Aug and stands in swamp margins and standing water, guaranteed on the Mayo and Hyco shorelines
Same genus, same white fringe, same window, planted on campuses and in parking lots across the Piedmont, and it defeats the wild-provenance goal if collected by mistake. C. retusus holds its panicles terminally on the current season's shoots, above fully expanded leaves, blooming after leaf-out; C. virginicus hangs them laterally from last year's wood as the leaves unfold. Backup: retusus leaves are smaller, thick, glossy and notched at the tip, and its bark exfoliates rather than staying scaly-ridged
Multi-stem 15–25 ft with 3–7 in drooping clusters of white strap-petaled flowers in the same window, and from thirty yards the airy white haze reads the same. Leaves are alternate and serrate; fringetree opposite and entire. One node settles it, even on a bare twig. Backup: serviceberry has 5 petals against fringetree's 4 lobes, and its fruit is a red-to-purple pome ripening May – Aug, mostly gone before fringetree drupes color
The default opposite-leaved understory tree here, blooming Mar – May in the same forest. The white is structurally different: dogwood shows 4 notched petal-like bracts around a tight green button of tiny true flowers, while fringetree shows a loose drooping panicle of many separate strap-lobed flowers. Backup: dogwood veins arc from base to tip paralleling the margin, and its fruit is a glossy bright red drupe in tight clusters of 3–4
In September the fruit picture is close: opposite leaves, 10–20 ft multi-stem tree of dry rocky Piedmont woods, dark blue drupes in drooping clusters. Margin and hair color decide it. Blackhaw leaves are finely serrate, glossy and leathery with rusty-brown hairs beneath; fringetree leaves are entire with soft white hairs. Backup: the old infructescence is flat-topped, since it flowered in cymes, never in a drooping panicle
Same family, same opposite branching, and a female ash hung with drooping fruit clusters in August reads like a fringetree in fruit at distance. Ash leaves are pinnately compound with 5–9 leaflets on one rachis; fringetree leaves are simple, one blade per node. Backup: ash fruit is a dry one-winged samara to 2 in, not a fleshy drupe, and ash twigs are stout with crescent-shaped leaf scars
Second target on the list that is ripe right now, and this one has a real deadline. Fruit starts coloring mid-July and is fully ripe by the first week of August at this latitude. Flanagan and Witte working at Knoxville flagged the Woody Plant Seed Manual's "September or October" as too late for the southern latitudes, so I should be treating mid-August as the target and scouting from late July. Deer, turkey, quail and songbirds all take the drupes, so there is no leisurely window.
The harder problem is not identification, it is sex. White fringetree is functionally dioecious, and the male plants carry the longer, showier petals. So the tree that stops you in April, the one that looks best in bloom, is statistically the one that gives you nothing in August. The right pattern is two visits: April to identify and flag the females, then July or August for the fruit. Or skip the flowering visit entirely and scout for fruit directly, which is what I'd do this month.
Person County has no obvious record concentration, but Carolina Nature documents it from Chapel Hill, Durham, Wake, Granville and Stanly, which brackets Person on three sides, and calls it fairly common across all three provinces. Walk the rich stream slopes and coves on the Hyco, Mayo and Flat, and the wooded edges around Mayo Lake.
One quirk worth knowing: it is one of the last trees to leaf out in spring. A fringetree in late April can look dead.
Leaves opposite, simple, entire, 3–8 in long, with dense white hairs on the underside and on young branches. That white pubescence is doing a lot of work, since it separates it from buttonbush and from rusty blackhaw's rust-colored hairs.
In flower, April into early May: corolla deeply 4-lobed with thread-like straps 15–30 mm long and about 3 mm wide, two stamens, in drooping panicles 4–10 in long. The panicles hang laterally from the previous year's wood, so the fringe drops below leaves that are still unfolding.
In fruit, July through September: ovoid fleshy drupe 1.5–2 cm, dark blue to purple-black, one seed, in loose drooping clusters.
Multi-stemmed, 12–30 ft. Bark scaly, dark brown ridges over reddish furrows. Growth is slow at 6–10 inches a year, and first flower comes at 5 to 8 years.
Before treating a tree as a seed source, check its health. White fringetree was confirmed in 2014 as the first non-ash host of emerald ash borer, and surveys in the Midwest found 26% of 178 trees showing signs, with infested trees at every site in four states. The character that ends the argument is a D-shaped exit hole about 3 mm across. Native clearwing borers leave round quarter-inch holes and roundheaded borers leave oval ones. Dieback and sprouting alone are not diagnostic, since drought-stress borers and Botryosphaeria do the same thing. Encouragingly, a 2020 study found fringetree health improved for trees that were not reinfested, so unlike ash it can recover.
Privet is the one that costs the most time. Same family, opposite, white panicles, black drupes, and it dominates every Piedmont bottom. Corolla lobe length settles it instantly in flower, and leaf size settles it any other time: privet is 1–2 in and semi-evergreen, fringetree is 3–8 in and deciduous.
Bigleaf snowbell is the one that costs a collection, because it fruits in August in the same upland woods. Alternate leaves with stellate hairs beneath, 5-parted flowers, and a dry brown drupe rather than a fleshy blue one.
Buttonbush is the leaf-only twin, and Carolina Nature names it on the fringetree page. Opposite or whorled entire leaves in the same size class. The tell is that buttonbush is smooth where fringetree is densely white-hairy.
Chinese fringetree matters for provenance, not for confusion in the woods. It blooms after leaf-out, holds its panicles terminally on current-season shoots, and has smaller notched leathery leaves on exfoliating bark. Never take seed from a planted fringetree without confirming the species.
Then the quick ones. Alternate leaves rules out serviceberry, snowbell, silverbell and sourwood. Compound leaves rules out ash and both buckeyes. Toothed margins rules out blackhaw. Bracts around a button rules out dogwood.
Name traps: "old man's beard" is a common name for this tree and also for Clematis virginiana, whose feathery seed heads are conspicuous August through October, exactly when I'd be out collecting. It also means Usnea lichen and Spanish moss. "Fringe flower" usually means Loropetalum chinense. Grancy gray beard, or the binomial, is unambiguous.
Hand-strip ripe purple-black drupes from the branches. Not off the ground, since fallen fruit has usually begun to ferment. About 630 fruits to the pound, giving roughly 590 clean seeds, and it fruits every year with no crop failure.
Clean the same day. Carpenter found that letting the fleshy pericarp ferment before removal delayed germination, which contradicts the usual "soak drupes a few days to soften" advice. Macerate, rub over hardware cloth, wash the pulp away.
Do not float-test. The bony endocarp traps air and sound seed floats initially. Cut-test instead.
Storage: fresh seed sits at 40–41% moisture and is killed outright at −15°C. Never freeze it. Store moist at 5°C, where it holds one to two years.
Germination is a two-phase dormancy: the radicle emerges under warm conditions while the shoot stays dormant, and only subsequent cold releases the epicotyl. In the wild it germinates in the second spring after seedfall. The standard trade protocol is 3 months warm plus 3 months cold, but Flanagan and Witte tested six combinations and 4 warm plus 4 cold more than doubled the full germination of 3 plus 3, and their follow-up went to 5 and 5. So: 4 months at 21–26.5°C in moist peat in sealed bags, then 4 months at 3–5°C, then sow a quarter to half inch deep and hold at 25°C. That 25°C figure is sharp — Carpenter got 82% at 25°C, 38% at 20°C and 3% at 30°C.
Across every treatment, far more seeds rooted than shot. Nicking a chip off the hilum end with a knife, opposite the radicle end, converts stalled radicles into shoots. Germination is erratic and can come a week after stratification or a year later, so do not dump a flat.
Cuttings: don't. Surveys in 1988 and 1995 ranked white fringetree among the hardest plants in the trade to propagate. One paper spans 0.56% to 90% rooting in a single study, and the variable is stock-plant juvenility, not technique. The 90% came only from forced 4-year-old container seedlings. Wild mature wood roots at essentially zero.
Honest timeline: seed takes 2 years to germinate, 2 more in the seedbed, then 3–4 years in the field for a saleable 3–4 ft plant. Six to eight years total. That is exactly why it is expensive in the trade, which is the opportunity, but it is not a crop that funds early cash flow.
Person County: rich, well-drained stream slopes and coves on the Hyco, Mayo and Flat drainages, plus lower slopes and forest edges around Mayo Lake Park and the game lands. It wants moisture nearby, so skip the dry ridges.
Denser records sit in the Eno River corridor and Duke Forest in Orange and Durham, and along the Haw bottoms near Saxapahaw. Morrow Mountain State Park in Stanly is a documented site if a longer trip makes sense.
Practical plan for this week: scan for the drupes rather than the tree. A 12–30 ft multi-stem understory tree with opposite entire leaves and loose drooping clusters of dark blue fruit is worth walking to. Check the leaf underside for white hairs, then check the trunk for D-shaped holes before deciding it is a mother plant.