Annual, biennial, or perennial herbs or shrubs
[rarely scandent], succulent. Stems
mostly alike and all potentially flowering, (flowering stems annual, axillary, overtopping primary stem). Leaves
usually persistent, sometimes deciduous or caducous, cauline or sometimes in dense basal rosettes, alternate, opposite, or whorled, simple (pinnate in Bryophyllum
), (heteromorphic in Dudleya, Echeveria, Graptopetalum
); stipules absent; petiole usually absent; blade margins usually entire, sometimes toothed or lobed. Inflorescences
terminal or axillary, spikes, thyrses, or mostly cymes, branches commonly cincinni (scorpioid cymes) and at first commonly coiled, or rarely flowers solitary. Flowers
bisexual (sometimes unisexual in Rhodiola
), radially symmetric (except sometimes calyx); perianth and androecium hypogynous or weakly perigynous; hypanthium present; sepals (3-)4-5(-12)[-30+], distinct or connate basally (or into tube); petals (3-)4-5(-12)[-30+], distinct or connate basally or sometimes into tube, margins entire (fimbrillate in Jovibarba
, crenate in Phedimus
); nectaries as scales at base of carpels; stamens as many as sepals or 2 times as many and in 2 series, antipetalous if in 1 series, free or adnate to corolla base (or tube); anthers dehiscent longitudinally and laterally; pistils (3-)4-5(-12)[-30+] [rarely fewer than sepals], mostly distinct or nearly so, (3-)4-5(-12)[-30+]-carpellate; ovary superior or semi-inferior, 1-locular; placentation submarginal; ovules anatropous, bitegmic, crassinucellate or tenuinucellate; styles (3-)4-5(-12)[-30+], terminal, distinct; stigmas (3-) 4-5(-12)[-30], terminal, capitate. Fruits
usually whorls of follicles, each dehiscent along adaxial suture, rarely whorls of flimsy 1-seeded utricles (Sedella
) or capsules of basally connate pistils (Diamorpha, Jovibarba
1-20+ per carpel, brownish, fusiform or ellipsoid; embryo straight; endosperm oily, scant. Phylogenetic analyses (M. W. Chase et al. 1993; D. R. Morgan and D. E. Soltis 1993; D. E. Soltis and P. S. Soltis 1997; D. E. Soltis et al. 1997; M. E. Mort et al. 2001) confirm that Crassulaceae are monophyletic and place them in a Saxifragales clade. They are close to Saxifragaceae but differ in their predominantly succulent habit, their numerically regular flowers with the same number of sepals, petals, and pistils and as many or twice as many stamens, their separate nectaries, one at the base of each pistil, their mostly separate pistils, and their seeds with little or no endosperm. The only genus of questionable position has been Penthorum
Linnaeus, sometimes placed here but clearly discordant. On the basis of embryology, J. Mauritzon (1933) found it closer to Saxifragaceae. From the anatomy, M. L. Haskins and W. J. Hayden (1987) thought it distinct from both and probably best kept as a separate family, Penthoraceae, as here. Podostemaceae also are thought related to Crassulaceae, although these submerged aquatics are so different and distinct that there are no intermediates. In this family, as J. D. Hooker (1865) remarked, genera are poorly defined, and there has been much disagreement about them. Recent and continuing chemical and chloroplast-DNA studies by H. ´t Hart and associates (e.g., H. ´t Hart 1995) are shedding light on relationships and will continue to influence delimitation of genera, although thus far they are based on far too few American plants. U. Knapp´s (1994) seed studies also are helpful.
A. Berger´s (1930) division of the family into six subfamilies has been useful but is now seen as largely artificial. With new information, H. ´t Hart (1995) proposed a better system, with two subfamilies: the Crassuloideae as before, but the rest in the Sedoideae, with two tribes and one of these with two subtribes. Also, U. Knapp (1994) found the Crassuloideae unique in the jigsaw-puzzle testa of their seeds. ´t Hart found the very large, long-recognized genus Sedum to be paraphyletic but, in a compromise to avoid nomenclatural chaos, he kept Sedum mostly intact while still recognizing as genera several of the smaller, more distinct, related groups that seemed monophyletic, such as Phedimus in this treatment. There is still disagreement about which to recognize.
In a long and continuing series of cytological studies, C. H. Uhl (e.g., 1963) has dealt with most genera of Crassulaceae; particularly the American ones. Through many hundreds of crosses, he has directly or indirectly interconnected more than 200 species of Echeveria and relatives, to identify one of the largest comparia yet known, and has studied meiosis in hundreds of these hybrids. From the extent of chromosome pairing in diploid hybrids, he has learned much about the degree of relationship of the parents, sometimes assigned here to different genera. He has argued that, in contrast to most known polyploids of other families, natural polyploids of Mexican Crassulaceae generally are autoploid.
Many New World plants, as in present-day Dudleya, Echeveria, and Graptopetalum, were first named in Cotyledon, and are now recognized as being native only in the Old World, in Africa and Arabia.
Many herbarium specimens of Crassulaceae represent the plants poorly and lack the supplementary notes and photos or sketches that would make them most useful. Because succulent plants often die and dry very slowly in the plant press and so often grow out of shape or grow moldy, it is best to kill the tissues first so that they give up their moisture quickly. A good way, when possible, is to freeze the specimen solid, then, after thawing, to puncture the larger parts and milk the juice from them. Then it is best to change the papers or felts in the press several times in short order, to get rid of most moisture quickly and so avoid mold, then to arrange the specimen carefully in a natural position before final drying.