Shrubs, glabrous. Stems apparently jointed and fleshy when young, becoming woody and not jointed, not armed; some stems terminated by an inflorescences, others entirely vegetative. Leaves opposite, united at base, petiolate, decurrent forming fleshy joints (sterile segments) fleshy on stem, fleshy; eventually deciduous; blade fleshy triangular projections at tips of joints, edges with narrow scarious margins. Inflorescences terminal and lateral, spikelike, apparently jointed, each joint (fertile segment) consisting of 2 opposite, axillary, 3(-5)-flowered cymes embedded in and adnate to fleshy tissue of distal internode; flowers in each cyme arranged in transverse rows, central flower separating lateral flowers, slightly larger; flowers in each cyme partially separated by flaps of tissue which persist on stem when flowers have fallen. Flowers bisexual or unisexual, ± radially symmetric; perianth segments persistent in fruit, 3-4, united except at tip, fleshy; stamens 1-2; styles 2-3. Fruits utriclelike, ellipsoid; pericarp membranous. Seeds vertical, ellipsoid; seed coat light brown, membranous, pubescent; hairs strongly curved or hooked and slender, or conic, straight or slightly curved; perisperm absent. x = 9. Sarcocornia is taxonomically difficult and has never been the subject of a taxonomic revision for the Northern Hemisphere. Although it is possible to identify dry specimens to some extent, by comparison, it is impossible to obtain from dried specimens data that can be used in a taxonomic revision. Characters that may be taxonomically useful are lost on drying, especially flower and inflorescence characters and those derived from the fleshy vegetative segments. Habit appears to be useful, but few specimen labels note the habit of the living plant, and the parts collected rarely allow for a reliable determination of habit. Some species, such as S. perennis are prostrate, with the woody stems readily rooting in the substrate. Others such as S. pacifica are procumbent to erect shrubs in which the woody stems usually do not root. This apparently obvious habit difference is sometimes confounded by external factors, erect species becoming procumbent due to water movement, trampling, or burial by silt or sand deposits. Conversely, prostrate rooting species can be disturbed by erosion and appear to be procumbent plants of a nonrooting species. One of the most useful characters, the indumentum on the testa of the seeds, is rarely present in dried specimens because of the lateness of the plants´ flowering season. Plants collected in August and September rarely have even immature seeds present, so most herbarium specimens do not display this character.
The consequence of these problems is that most accounts of Sarcocornia in North America recognize only one species, frequently using the name Salicornia virginica Linnaeus for the collective entity. The type specimens of S. virginica were collected by John Clayton, presumably from Virginia, which are immature annuals and not flowering. The name S. virginica cannot be applied to a species in this genus.