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.
Lobelia cardinalis · Campanulaceae
A scarlet spike standing in wet mud, readable from a hundred yards. Peak bloom is this week.
Community photos of Lobelia cardinalis to study before I head out: leaf, flower or bract, ripe fruit, bark, whole plant.
Native, G5 / N5 / NC S5, not tracked by NCNHP, no collection restriction. The NC Biodiversity Project calls it "throughout the state; probably in all 100 counties" and "generally fairly common to common statewide." iNaturalist counts pulled 2026-08-23: Person 3, Caswell 4, Granville 9, Alamance 67, Chatham 56, Durham 244, Orange 290, NC statewide 3,573. Person is observer effort, not rarity. Herbarium sheets are thinner still: Person 1, Granville 3, Alamance 4, Chatham 7, Durham 12, Orange 27
Best bet: Surface-sow and never cover; the light requirement is the one point every source agrees on. Dry stalks 1–2 weeks in open paper, crush, screen. Preferred route: surface-sow flats Nov–Dec and overwinter outdoors in a shaded rodent-proof frame, germination mid-Mar to mid-Apr. Fallback: 60 days cold-moist at 33–40°F from mid-Dec, then surface-sow at 65–72°F under mist, germination 7–21 days. Saleable quarts by May–Jun, a short spike possible in year one, full 3–4 ft spikes in year two
Native throughout NC, hardy 3a–9b, so 7b/8a sits mid-range. Heavy clay is an asset here because it holds water; amend for organic matter and structure, not drainage. The limit is water, not heat. Drought tolerance is rated low and the soil should never dry out, so site it in a low spot, rain garden or drip line and plan on irrigation. Morning sun with afternoon shade unless the ground is genuinely wet. Failure modes in rank order: winter smothering of the basal rosette, drying out, crown rot in stagnant compacted soil, and losing the colony when mulch blocks reseeding. At this latitude I rake litter off the crowns in November rather than mulching over them. All parts are toxic; gloves for cutting and seed cleaning
Each look-alike with the one character that settles it. Tap a photo to zoom.
Conspicuously hairy stem and leaves, worst toward the base, with blue-violet flowers, against cardinal flower's smooth stem and deep crimson. Backup: puberula leaves are sessile, blunt-tipped and only about 3 in, and it tolerates dry upland edges and powerline cuts rather than mud. 308 iNat records across the six-county block, 6 of them in Person.
Blue to violet-blue flowers with white bumps and stripes at the throat of the lower lip; cardinal flower is crimson with no white throat markings. Backup: siphilitica packs flowers densely all around the stem where cardinalis holds a narrower, more one-ranked column. Wild NC range runs Mountains east only to Rockingham and Caswell, so most of the 78 records in the block are plantings.
Violet-blue flowers, and vegetatively the leaves are thick, parchment-textured, glaucous blue-green and whitened beneath, against cardinal flower's thin bright green leaves on both faces. Backup: gland-tipped teeth on some calyx segments and the uppermost leaves. Only 2 records in the block, but it's the right answer for a big smooth blue lobelia standing in the same seep.
The base of each spent flower swells into an inflated sphere, so there are always football-shaped capsules low on the raceme while the top still flowers. Cardinal flower capsules never inflate like that. Backup: flowers tiny, about 0.3 in and pale blue, on a branched hairy plant, and it's an upland ruderal rather than a wetland species. 257 records in the block.
Flowers run pink, magenta or purple-red instead of pure deep crimson, on a shorter, stouter raceme over a fleshy rosette. Backup: bronze or maroon foliage means cultivar, and anything in mulch, in a bed, or in a row is horticultural. Sixth most common iNaturalist misidentification of cardinal flower at 25 cases. Don't collect seed from it.
Opposite leaves and five radially symmetric petals, each strongly notched at the tip, against cardinal flower's alternate leaves and a bilateral corolla split along the top with the anther column thrust out. Backup: fire pink blooms Apr–Jul on dry rocky roadbanks and is long finished before cardinal flower peaks.
Paired opposite ovate leaves clasping the stem, and the corolla opens into five short equal lobes showing bright yellow inside the tube. Cardinal flower has alternate leaves and no yellow anywhere. Backup: blooms May–Jun, and it isn't wild in the Piedmont (S1, extreme southwestern mountains), so all 88 regional records are plantings or escapes.
Square stem, opposite aromatic leaves, and one flat-topped terminal head of flowers radiating from a common axis under reddish bracts, against cardinal flower's round scentless stem, alternate leaves and elongated raceme. Backup: crush a leaf, it smells of oregano and thyme. Wild NC range is essentially Mountains only; the 109 records in the block are plantings.
Flowers dangle horizontally from thread-thin curved stalks out of the leaf axils; cardinal flower holds its flowers erect in a terminal raceme. Backup: the color is orange with darker spots, on a branched succulent translucent stem 3–5 ft.
Habitat does most of the sorting before I ever see the plant. Wet ground with direct sun on it, which in practice means the seam between bottomland canopy and open water. Gravel and cobble bars inside river bends. The scour strip between normal water level and the flood line. Seeps and spring runs where they cross a road or trail. Reservoir drawdown mud in late summer as the pool falls. Beaver ponds, oxbows, backwater sloughs, marsh edges, wet roadside ditches, powerline corridors crossing bottomland. Not upland slopes, not closed canopy, and not a dry sunny field either.
The color carries fifty to a hundred yards. A one- to two-foot scarlet vertical against green sedge and mud isn't anything else here in late August. Hummingbird traffic hovering repeatedly at one spot is its own locating cue; ruby-throated is very nearly the only effective pollinator.
Then walk up and check three things. Alternate leaves. A red staminal tube arching out through a slit in the top of the corolla and nodding forward. Milky white latex from a snapped leaf. The latex confirms Lobelia, the arching stamen column confirms the species, and the alternate leaves throw out fire pink, beebalm and Indian pink in one glance.
Timing note, and it changes the plan. This week is peak flower, not seed. NC fruiting peaks in September, and capsules have to go brown, dry and withered before they release anything. So this is a finding trip: locate colonies while they're unmistakable, drop a waypoint on every stand, and come back mid-September through October. By late October the plant is a nondescript brown stick and I won't find it again without the pin.
On the return, clip whole stalks below the capsules once about half the pods on a stem are brown and seed pours out when one is tipped down. The raceme opens bottom to top, so ripe pods sit under flowers still in bloom. Either make two or three passes about two weeks apart or cut the stalk when the lower half is brown and let the top finish in the bag. Stalks go upside down into paper, never plastic. The seed is 0.4–2 mm powder at 350,000–400,000 per ounce, so bag whole capsules and don't try to shake seed loose in the field. One robust stalk with 40–60 capsules is plausibly a whole season's sowing, and since this is a short-lived perennial that persists by self-seeding, most of the capsules stay on the colony.
Dry the stalks one to two weeks in open paper, crush, screen through quarter-inch then a fine mesh. Surface-sow and never cover it. Light is the one germination requirement every source agrees on. My preferred route is to surface-sow flats in November or December and overwinter them outdoors in a shaded, rodent-proof frame; germination comes mid-March to mid-April. Fallback is 60 days cold-moist at 33–40°F from mid-December, then surface-sow at 65–72°F under mist, with germination in one to three weeks.
Stratification is where sources actually disagree. USDA NRCS says the seed germinates without cold and only needs light. Prairie Nursery says 30 days, Prairie Moon and Illinois Wildflowers say 60, LBJ says three months. That pattern reads like shallow light-mediated conditional dormancy, so I'll give it 60 days because it costs nothing. Worth knowing that the only real lab data anywhere is Tallgrass Prairie Center's: 95% pure live seed, 4% germination, 70% dormant in a standard dark unstratified test. Viability is high and almost all of it sits behind the light requirement. Nobody publishes a post-treatment germination percentage, so my expected rate is genuinely unknown.
Vegetative route worth a try on the same trip. Layering yields the most per parent plant and it's a mid-July to mid-August job: bend a stem down, pin it, cover with a quarter inch of soil, keep it moist. Roots and rosettes form at the buried leaf nodes and get severed and potted in October. A tall stem with eight to twelve buried nodes can give most of those. Division of rooted rosettes works fall or spring, but pot the offsets and grow them on — planting small divisions straight into the ground kills them.
Hazards. All parts of cardinal flower are toxic, nicotine-like alkaloids, and the sap has caused skin irritation. Gloves for cutting and cleaning, and label the grown-out stock. The bigger risk is the company it keeps: spotted water-hemlock, Cicuta maculata, stands in these same lake margins and ditches in all 100 NC counties, blooms May–August, and is lethally poisonous. Four to five feet, branched, white compound umbels, twice-divided leaves, hollow purple-mottled stem. Never handle its roots bare-handed.
Person County first, with the record admittedly thin. The one herbarium sheet is NCU, September 1957, "Hyco River, 12 mi N of Roxboro." That ground is now under the Hyco Lake basin, so I'd work the creek arms, creek mouths and drawdown zone, particularly South Hyco Creek and the Mayo Creek inflow. There's a 2023 iNaturalist record near Semora on the north and west arms, one from 2021 a few miles east-northeast of Roxboro, and a research-grade Person County record from 2026-08-17 with coordinates obscured. Mayo Lake Park returns zero records for the park polygon and the game lands return nothing at all, which is absence of observers rather than absence of plant — the shoreline drawdown and the Mayo Creek inflow are exactly right.
Twenty-five minutes south is the Rougemont cluster on the Person/Durham line, in the Flat and Little River headwaters: New Sharon Church Rd, State Forest Rd near Hill Demonstration Forest, Cothran Rd, Hobgood Rd, Bahama.
The strongest documented public site in reach is Lower Haw River State Natural Area in Chatham, 15 iNaturalist records including one from 2026-08-16. Then the Haw corridor at Saxapahaw and Graham, Cane Creek Mountains Natural Area at Snow Camp (a 2022 NCU label puts it at the seepy stream crossing the road running west from the parking area), Eno River State Park at Fanny's Ford and Cole Mill Rd, and Duke Forest. Skip Occoneechee Mountain; it's a dry monadnock and the habitat is wrong except along the Eno at its base.