Diodia tricocca T. & G., Fl. N, Am. 2: 30. 1841.
Crusea allococca A. Gray, Proc. Am. Acad. 19: 78. 1883, nom. superfl., illeg.
Crusea tricocca (T. & G.) Heller, Contr. Herb. Franklin & Marshall College 1: 97. 1895.
Richardia tricocca ranges from southern Texas and Louisiana south to the states of Mexico, Morelos, and Puebla. When Gray transferred the species from Diodia to Crusea, using the new (superfluous) name Crusea allococca, he did so without discussing his reasons for placing it in Crusea rather than in Richardia. However, his synopsis in the Synoptical Flora of North America (1884, p. 22) makes it clear that he was well aware of the similarity between the two genera. The only real difference between them, as he treated them, is that Crusea has the carpels separating from a persistent axis, while Richardia has no persistent axis. Apart from this difference, I would guess that Gray was influenced in his decision by the fact that his only North American species of Richardia was R. scabra L., which has cocci that are very different from those of R. tricocca. In R. tricocca the cocci fit together with smooth, flat ventral faces, attached slightly below the middle to a thin central axis. The seed does not completely fill the coccus. In R. scabra the cocci are strongly ingrown at the edges where they meet; their ventral faces do not fit together at maturity but are withdrawn into narrow grooves. They are attached near their bases to a short, stout axis, which is inconspicuous but does persist, Gray's statement to the contrary notwithstanding. The seeds completely fill the cocci at maturity and are rather tightly adherent to the thick coccus walls. These differences are impressive when only the two species are compared, and they apparently influenced Gray so much that he was willing to expand the circumscription of Crusea in order to accommodate R. tricocca there. On the other hand, there are strong similarities between R. tricocca and R. scabra. In growth habit they are identical; both are sprawling perennials with a woody rootstock. There is a tendency in both species for the calyx lobes to vary in number from four to six, even on the same plant, although R. scabra probably averages more lobes than R. tricocca. The calyx in both is gamosepalous with a broad, almost radiate limb, and in the base of this calyx is an unlobed annular disc, which persists there in fruit; the calyx is circumscissile and deciduous as in most species of Crusea. The corollas in both R. scabra and R. tricocca are white and broadly funnelform. The ovary seems to always have at least three carpels, very often four, and not rarely five or even six, especially in R. scabra. The "persistent axis" is longer in R. tricocca because the cocci are attached nearer the middle of the axis, but in neither species is it anything more than an undifferentiated central column of tissue; nothing like the flat, bifid carpophore of Crusea is present. On the basis of these similarities alone I would be quite ready to exclude R. tricocca from Crusea and consign it to Richardia, as was done by Loesener in 1922 and Standley in 1931. This inclination is strongly supported by comparison of the two IMexican species with the South American species of Richardia. IMost of the South American specimens that I have seen are vegetatively more simHar to R. scabra than to R. tricocca, but their cocci and central axes are strikingly like those of R. tricocca, so much so that if R. tricocca were retained in Crusea, these species (especially R. brasiliensis and R. humistrata) would also have to be transferred to Crusea. Some South American specimens have cocci intermediate between those of R. scabra and R. tricocca, but for the most part it seems that the species which least fits with the rest of the genus Richardia is the widespread type R. scabra. There can be no question of retaining R. tricocca in Crusea without expanding Crusea to accomodate most of Richardia, and there is no reason to do that, either as a matter of convenience or in the interests of natural taxonomy.