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.
Physocarpus opulifolius · Rosaceae
Papery peeling bark on scoured river rock. Nothing wild recorded anywhere near Roxboro.
Community photos of Physocarpus opulifolius to study before I head out: leaf, flower or bract, ripe fruit, bark, whole plant.
Person 0 iNat and 0 herbarium sheets. Caswell 0/0. Alamance has 3 iNat records but no wild-origin ones — two flagged cultivated at Elon, the third a duplicate log of the same planting. Of 635 NC iNat observations, 553 fall west of −81.0 and 82 east of it. Every central-Piedmont record is a planting.
Best bet: Semi-hardwood cuttings in late June, IBA talc, mist at 25–28 °C: 82–100% rooted at 8 weeks against 40–48% untreated.
Grows here happily as a planted shrub. NC Extension rates it zones 2a–8b and notes it appreciates some shade in its southernmost locations, so the Piedmont at 7b–8a is the warm edge. Give morning sun and afternoon shade. Powdery mildew and fireblight are the two problems NC Extension lists. The hard part is not growing it but sourcing wild Piedmont material.
Each look-alike with the one character that settles it. Tap a photo to zoom.
Leaves opposite, paired at every node, against ninebark's alternate. This is the default answer for any lobed-leaf shrub on a Person County slope — 1,082 wild records in the central Piedmont, 19 in Person. Backup: blue-black drupes and gray non-peeling bark.
Named for the same many-layered peeling bark, same size, same May–July white corymb. Leaves opposite against alternate, readable at arm's length and on bare winter twigs. Backup: unlobed pinnately-veined leaf and dry ribbed capsules. NC Extension flags it poisonous to people and pets.
Carries two of ninebark's headline characters, peeling strip bark and alternate lobed leaves. Tendrils opposite the leaves and a climbing habit settle it — ninebark is self-supporting with no tendrils. Backup: woolly leaf underside and fleshy berries.
Leaves never lobed, elliptic, 1–2 in. Blooms March–April on bare stems with double white buttons, while ninebark blooms in leaf with single five-petaled flowers. Spirea stems are shiny and never peel.
Its new growth flushes pink, red, purple or chartreuse, so a colony reads as a purple ninebark cultivar. Flowers deep rose in a flat corymb to 8 in against ninebark's white 1–2 in dome; leaves lanceolate and never lobed. NC Invasive Plant Council listed — do not collect or move it.
Strongest overall match here: arching multi-stemmed Rosaceae to 8 ft with alternate sometimes-lobed leaves, white clusters, follicle fruit, persisting at every old farmstead. Scrape a thumb-thick stem — vanhouttei stays slender, smooth and brown and never exfoliates.
A pure bark trap — coarsely shreddy exfoliating light-brown bark on a 3–10 ft arching shrub of rocky bluffs, native to the Piedmont, 3 records in Person County. Leaves opposite and unlobed. Backup: 4-petaled inch-wide flowers with 60–90 stamens and a woody capsule.
The only plant carrying both ninebark headline characters at once, deeply lobed leaves and exfoliating cinnamon bark. Leaves opposite against alternate. Backup: an erect 4–12 in panicle with large sterile florets. All 12 wild-flagged central Piedmont records are plantings.
Not around here, is the short answer, and I would rather write that down than keep walking Mayo Lake looking for it.
Person County has zero iNaturalist observations and zero herbarium sheets for this species, ever. Caswell the same. Alamance has three iNat records but no wild-origin ones: two are flagged cultivated at Elon University and the third is the same planting logged again two days later at the same coordinates. Statewide iNat holds 635 NC observations, and binning them by longitude puts 553 west of −81.0 against 82 east of it, led by Watauga 120, Clay 84, Ashe 70, Buncombe 62, Burke 46. SERNEC's 270 NC specimens, 1891 to 2023, tell the same story. Granville has one skeletal NCU sheet with a county name and nothing else.
The central Piedmont records are plantings. Of 67 in the box, 26 are flagged cultivated and the other 41 are unflagged plantings at the NC Botanical Garden, Duke Gardens, the UNC-Chapel Hill campus, JC Raulston bed E33, the NC Zoo and Elon. Zero natural-area records anywhere in the box. Almost all are named cultivars, so they are worthless as seed for a wild-provenance nursery. Still worth one trip to learn the shape and the bark.
The closest real Piedmont record in genuine habitat is Durham County, 2 July 1931, H. R. Totten, "meadow near Eno River, grounds of Camp Sacarusa." Ninety-five years old, 36 km from Roxboro, never relocated. Two Virginia sheets sit closer at around 60 km, in Pittsylvania and Mecklenburg counties, but both are skeletal with no locality or date.
I first wrote the habitat as scour bar only and that was too narrow. Of the seven genuine non-planted Piedmont sheets, two read rocky river bank, the Deep at High Falls and the Yadkin above the NC 49 bridge. The other five read swamp-forest in Anson and Davidson, low woods in Mecklenburg, and that Durham meadow. Searching only for cobble would miss most of the records. Full sun either way, heliophily 7. Impounded lake shorelines are still the weakest bet, but landform alone does not rule Mayo and Hyco out.
Bark is the name and the character. On any cane thicker than a pencil it peels in long papery vertical strips over layered inner bark that runs yellow to orange to red-brown. Usually it is a fountain of thin arching stems with no real trunk, 5–8 ft by NC Extension's numbers and up to about 10 ft. Older plants can go tree-like and then the bark peels in broader sheets, so the no-trunk rule is a tendency rather than a test.
Leaves alternate and simple, 3 and sometimes 5 shallow palmate lobes, blade 6–8.5 cm, blunt crenate teeth, base cuneate to truncate, both surfaces essentially hairless. The narrow 6–10 mm stipules drop early, so look on the youngest shoots. Flowers are what read from thirty yards, an upright dome 2.5–5 cm across of 30–50 small white to pale-pink flowers. Fruit is 3–5 inflated papery beaked follicles per flower, 100 to 250 to a cluster.
Purple, burgundy, gold or chartreuse new growth means a cultivar. Green foliage does not prove the opposite, which is a trap I had wrong at first: NC Extension notes purple selections fade toward green in sun, so a washed-out 'Diabolo' in an old planting reads perfectly wild. Provenance comes from where the plant is standing.
In the Uwharrie or Pee Dee area a lens on the follicles sorts the varieties: var. opulifolius glabrous, var. intermedius persistently stellate-hairy. The hairy one is ranked S1 in NC, so document it and take almost nothing.
One caveat on abundance. Four attempts failed the Cloudflare gate on the Weakley site, so the rating is not from FSUS. Carolina Nature says common in the Mountains, occasional in the Piedmont. Coastal Plain absence is independent: no Coastal Plain county among the 272 NC GBIF specimens, and BONAP agrees.
Half of this list falls to alternate or opposite. Mapleleaf viburnum, wild hydrangea, mock-orange and oakleaf hydrangea are opposite, ninebark alternate, and that works at arm's length and on bare winter twigs. The other four are alternate too. For those, bridalwreath and Japanese meadowsweet come down to lobing, since neither has a lobed leaf. Vanhoutte is the exception even there, so for that one it is the stem. For summer grape it is the tendrils.
Wild hydrangea is the interesting one, called sevenbark for the same many-layered peeling bark, same size, same moist rocky places, white corymb at the same time of year. Leaf arrangement decides it, then the unlobed pinnately-veined leaf and the dry ribbed capsules. Summer grape carries peeling strip bark and alternate lobed leaves, two of the three headline characters, and there are 122 Vitis records in the Person County box alone.
The spireas are the ones I would actually mis-call. Vanhoutte is the closest match of all, an arching multi-stemmed Rosaceae to 8 ft that the extension description even calls sometimes 3–5 lobed, with white clusters and follicle fruit, persisting at every old farmstead. Scrape one thumb-thick stem. It stays smooth and brown and never exfoliates. Japanese meadowsweet flushes pink and purple in new growth, so a colony reads exactly like a purple ninebark cultivar. It is NC Invasive Plant Council listed, so nothing of it should be moved.
One thing to be careful of if I do end up on a mountain river scour bar: Spiraea virginiana lives in that exact habitat and is federally threatened — do not collect it.
Cuttings are the route. Semi-hardwood, late June, one or two nodes, 2 cm deep in peat:perlite 2:1 at pH 4.9–5.1, misted at 25–28 °C, shaded two weeks. IBA talc at 10,000–20,000 ppm gave 82–100% rooted at 8 weeks against 40–48% untreated. The closer climate match is the Beltsville Shenandoah-ecotype protocol: early-to-mid July softwood, 6 in, trimmed to one leaf pair, 3,000 ppm IBA talc, straight perlite, mist, roots in a few weeks, about 18 months to a field-ready quart. Hardwood works with no mist bench, cut December and direct-stuck February through April, 48.5% untreated and 57–65.5% with hormone or mycorrhizal inoculum.
Seed is the second choice and it needs volume. Collect whole follicle corymbs into paper bags once the cluster has gone tan-brown and papery but before more than 10 or 20% have split. A ripe follicle crackles rather than squashing, and the seed inside should be 1.8–2.5 mm, hard, smooth and shiny yellow, one to four per follicle. No pulp, no maceration, no float test — dry 7–14 days on screened trays, rub over quarter-inch mesh, sieve at about 1 mm and winnow gently. Then 60 days cold-moist at 1–5 °C, sow about 3 mm deep with a grit or sawdust cap. The Missouri nursery has seedlings up by late March.
Viability is usually under 50%, so oversow five to ten times and cut-test 30–50 seeds before deciding the rate. The causes are indeterminate ripening, self-incompatibility, clonal root-sprouting stands that set poor seed, and two seed-specialist wasps. One Missouri nursery result: sowing about 1,770 seeds per linear metre gave 110–130 seedlings per square metre, roughly 1.5 to 2% field emergence.
I had the seed window backwards in the first version of this guide. The iNaturalist fruiting annotations peak in June and July and I read that as maturity. That peak is the red inflated display stage, photogenic and tagged the moment it shows. The Woody Plant Seed Manual says mature follicles turn brown, reddish or coppery and burst at both sutures. NC Extension gives fruit display as July through September, and the NC curve has a second bump of 11 records in October, persistent dry follicles. So the target is mid August into October. Ripening is indeterminate and one bush carries both stages at once, so go by the physical test.
Nowhere near Roxboro, wild. The realistic options are a mountain trip to Watauga, Ashe, Clay, Buncombe or Burke, or a scouting run to a river bank on the Yadkin, Deep, Rocky, Broad or Catawba. Work the swamp forest and low woods along those rivers as well as the scour bar. The nearest wild record in natural habitat to the north is Grassy Hill State Natural Area Preserve near Rocky Mount in Franklin County, Virginia, about 107 km away.
Go in May. The NC herbarium months put the Piedmont and foothill peak at May with 11 sheets against 5 in April, and the mountains at June with 32 and July with 22. Full bloom is the only stage that reads from thirty metres across a river, and Piedmont plants run 3 to 6 weeks ahead of Boone-area plants, so one trip will not catch both provinces at peak.
Grow-out is easy and the caution is the summer. NC Extension lists powdery mildew and fireblight as the problems, and mildew is the one that bites in humid Piedmont heat. I had also written that a ten-cultivar trial found only the green-leaved entry resistant, and I could not re-source that, so it stays unverified. Space wide, water at the base and early, morning sun with afternoon shade, which is what NC Extension means by appreciating some shade in its southernmost locations. Prune right after bloom.