STONE-CROP FAMILY HI 



Generic and specific criteria. — Both Bentham and Hooker (Genera Plantarum 1:659) and 

 Schonland in Engler and Prantl's Die Natiirlichen Pflanzenfamilien (3=":34) included the Amer- 

 ican species of DeCandolle's Echeveria in the genus Cotyledon. In a later treatment by Berger, 

 Cotyledon is restricted essentially to Africa (Pflanzenfamilien, ed. 2, 18a:471, — 1930). Tech- 

 nical grounds for the separation of the American Echeverias are largely insufficient. The sum 

 total of characters, however, helps to establish a sort of cleavage between the New World and 

 Old World forms. In Echeveria the leaves are alternate (rarely opposite) and usually produce 

 rosettes. In Cotyledon the leaves are opposite or alternate, rarely producing rosettes. In Eche- 

 veria the calyx is usually as long as the corolla-tube and commonly exceeds it. In Cotyledon the 

 calyx is usually much shorter than the corolla-tube. In Echeveria the petals are united usually 

 only at the very base or less than half way and are often ridged or angled. In Cotyledon the 

 petals are usually united half way or more, the tube being terete or 5-angled. The most constant 

 character seems to reside in the inflorescence: Echeveria has an inflorescence arising laterally, 

 while Cotyledon has an inflorescence arising terminally. Segregation of Echeveria from Cotyle- 

 don for geographic reasons is not necessarily a point of moment. Those botanists who accept 

 minor generic segregates hold, for example, that true Sphaeralceas and true Malvastrums grow 

 both in the Mexico-California region and in South Africa. 



The Californian representation of Echeveria has been set off by Eritton and Rose under 

 the genus name Dudleya. Echeveria with its "strongly 5-angled flowers" is contrasted with Dud- 

 leya's "tubular" flowers, but this distinction does not hold. Several Dudleyas, such as Echeveria 

 lanceolata and caespitosa, have strongly 5-angled flowers due to the markedly keeled petals. It is 

 implied by Britton and Rose that the petals in Dudleya are united more or less to the middle, 

 which is contrary to the facts. The petals in our Californian species are united only at base and 

 are often nearly distinct. For Echeveria Britton and Rose assign as a stamen character that 5 

 (of the 10) stamens are attached near the middle of the petals, but they describe the stamens in 

 E. australis as attached one-fourth the distance above the base. In garden plants we find various 

 Echeverias in which the antipetalous stamens are inserted near the base of th% nearly distinct 

 petals, just as in the native Californian species. The Californian forms as to geographical origin 

 nmst be thought of as derived from Mexico and we think the two groups quite congeneric. Species 

 defined as of the genus Dudleya by Britton & Rose are not limited to California and Lower Cali- 

 fornia, but are also found in Arizona, which belongs to the same floristie region as northern 

 Mexico. 



For the two Californias Britton and Rose have published a considerable number of new 

 species under their generic name Dudleya. Major divisions in their grouping (N. Am. Fl. 22:33) 

 are distinguished by such paired headings as "Calyx-segments lanceolate to triangular-lanceolate" 

 and "Calyx-segments ovate, ovate-lanceolate or triangular-ovate." But such characters depend, in 

 certain of our forms at least, on degree of development of the organ concerned j a single indi- 

 vidual may exhibit both ovate and lanceolate calyx-segments according to the age of the flowers. 

 These authors, too, make a good deal of inflorescence, but the inflorescence is very inconstant in 

 most of our species and lacks distinctive characters. In one small colony a species often shows 

 pronounced differences. Consider Echeveria lanceolata Nutt. At a station on Palomar Mt. we 

 studied two plants (Jepson 1530) growing side by side, wholly identic save that the first had 

 simple elongated divisions to the cyme, the second had the divisions branched. Other characters 

 used by these authors relate to inconsiderable or wavering features, so that we, in this work, 

 dispose of their proposed species with difficulty, a situation not lessened by the fact that so many 

 of them are represented by only a single station, often by only a single collection. In order to 

 provide satisfactory groundwork for classification, a study of garden cultures in reference to 

 variation in the field, carried on more extensively than hitherto, would seem necessary. Such 

 garden cultures to be useful must, however, bring together plants selected as the result of a 

 planned geographic survey, and not mere casual or incidental gatherings. In such a group as 

 this where differentiating or well-marked vegetative and floral characters are largely absent, 

 the use of such characters as are so commonly employed in many other groups does not answer. 

 It seems to us desirable that effort be directed towards the special biology of the species, with a 

 view of accumulating ample knowledge of the life history and distribution. 



Carpels erect, usually distinct; corolla cylindric or obconic. 



Rosettes very large (6 to 8 inches wide), raised on a stout caudex 2 to 6 inches high; plant 



densely white-mealy; leaves 2 to 2^ inches wide; seaward belt, S. Cal 



1. E. pulverulenta. 

 Rosettes medium-sized (^4 to 3 inches wide), white-glaucous to green, borne at the ground 

 on a short caudex; leaves % to 1% inches wide. 

 Pedicels stout, shorter than the flowers ; flowers yellow. 



Rosettes globose, very compact; leaves very thick, often slightly ridged towards 



base on upper side ; sea bluffs or headlands 2. E. cotyledon. 



Rosettes somewhat flat-topped, less dense; leaves less thick, often slightly cupped 

 towards base on upper side; seaward belt, San Francisco to San Luis 



Obispo 3. E. caespitosa. 



Pedicels slender, usually equaling or exceeding the flowers. 



