Herbs, annual or perennial, succulent, glabrous, with caudices, not rhizomatous or stoloniferous.
Roots fleshy and/or fibrous, or taproots.
Stems decumbent to erect, simple or branched; nodes glabrous.
Leaves basal and/or cauline, glabrous, not articulate at base, somewhat to markedly clasping, attachment points linear; basal leaves in rosettes; cauline leaves alternate (usually absent in
C. umbellata and
C. tweedyi).
Inflorescences axillary, umbellate, paniculate, racemose, or cymose, sometimes scorpioid and/or secund, loose to dense, 1-many-flowered, leafy and/or bracteate; bracts (1-)2 subtending each flower, markedly unequal.
Flowers pedicellate to subsessile; sepals persistent, distinct, not keeled or angled, often unequal, herbaceous to scarious, glabrous; petals 2-9(-12), distinct; stamens 1-23; ovules 1-many; style present or absent; stigmas 2 or 3.
Capsules 2-3-valved, dehiscence loculicidal from apex, valves not deciduous, (except
C. tweedyi, which has circumscissile deshiscence and 3-4 valves splitting from base), margins not reflexed after dehiscence, margins not markedly involute.
Seeds 1-40, dark brownish red, dull gray, or black, usually elliptic to orbiculate, sometimes orbiculate-reniform, dull or shiny, smooth or sculptured, sometimes hairy, strophiolate in
C. maritima and
C. tweedyi.
x = 22, 23. Publications by R. C. Carolin (1987) and M. A. Hershkovitz (1991, 1991b) have required the addition to the North American flora of the genus
Cistanthe, which originally was erected to accommodate some Chilean species that were segregated from
Calandrinia. Recent investigations indicate that a considerably broader range of species belongs in the genus, including two North American species formerly placed in
Calandrinia, all the species formerly included in
Calyptridium, and one species previously classified in
Lewisia. While the current recognition of
Cistanthe originally rested on the cladistic work of Carolin and the studies of leaf morphology by Hershkovitz, who documented nine traits of leaf morphology distinguishing
Cistanthe, more recent molecular studies by M. A. Hershkovitz and E. A. Zimmer (2000) have provided general support for it. The inclusion of
C. tweedyi appears to be somewhat equivocal and it might best be treated as a distinct genus.