PBIO 450 Lecture Notes

James L. Reveal

Norton-Brown Herbarium, University of Maryland
University of Maryland, College Park, MD 20742-5815

Selected Families of Angiosperms: Rosidae

Euphorbiales Lindl., 1833

As defined by Cronquist (1988) the Euphorbiales consist of four families and more than 8000 species, essentially all of which belong to the Euphorbiaceae. As he stated at the time, of the three satellite families, only Pandaceae (4 genera, 30 species) is unquestionably a member of the order, with the Buxaceae (5 genera, 60 species) and Simmondsiaceae (monospecific) dubious. Cronquist also noted that the placement of the order within the angiospermous dicotyledones was uncertain but that they were more likely members of the Rosidae than the Dilleniidae. This later uncertain has yet to be fully resolved, but at least the placement of the two dubious families has been resolved. Webster (Bot. J. Linn. Soc. 94: 3-46. 1987) assigned the order to what Thorne (1997) terms the Malvanae, a taxon that now contains a mixture of families from Cronquist's Hamamelididae (e.g., Urticales), Dilleniidae (e.g., Malvales) and Rosidae (e.g., Rhamnales). Based on the rbcL data, the family seems to be close to the Theanae (Ochnaceae) and Violanae (Violaceae, Passifloraceae) of the Dilleniidae, and the Rosanae (Chrysobalanaceae) and Geranianae (Linaceae, Humiriaceae, Erythroxylaceae, Malpighiaceae and Trigoniaceae) of the Rosidae. As these data do not include two important families, the Elaeocarpaceae and Flacourtiaceae (Violanae), it is difficult for me to judge the significance of the findings at this point in time.

Given the seemingly close molecular relationship between the Passifloraceae and the Euphorbiaceae, however, I have adopted Takhtajan's (1997) placement of the Euphorbianae in the Dilleniidae rather than retain the family in Cronquist's Rosidae. Nonetheless, I have positioned the superorder at the end of my Dilleniidae next to the Urticanae and near the Malvanae following somewhat the observations of Webster. In any linear arrangement of families, it is difficult to precisely express the anastomosing relationships among the taxa. Also, based on the information presented by Takhtajan (pp. 241-242), the majority of the anatomical and morphological information suggests a flacourtean point of origin for the Eurphoriaceae. He concluded his remarks by saying:

One may presume that the Euphorbiales have arisen either from some ancient group intermediate between the Flacourtiaceae and the Elaeocarpales and Malvales or directly from some Flacourtiaceae-like lower Violales. In both cases the Euphorbiales have only collateral affinities with the Elaeocarpales and Malvales.
In defense of Cronquist, he is correct that the "Euphorbiaceae are so diverse in vegetative and chemical features and in pollen-morphology that one could compare some members of the group to any of a wide range of families and orders of dicotyledons" (p. 393). He suggested that one should depend more upon floral features, and especially aspects of the gynoecium as these features "provide a better guide to the affinities of the group, and in these respects in Euphorbiaceae compare well with the Celastrales and Sapindales" although he goes on to declare that the "malvalean gynoecium is more or less comparable to that of the Euphorbiaceae, but the nectaries in the two groups are wholly different."

In an important but now dated paper, Hutchinson (Amer. J. Bot. 56: 738-758. 1969) characterized the ancestor of the Eurphorbiaceae as having:

hypogynous, bisexual flowers, provide with both free sepals and free petals (polypetalous), numerous to few free stamens, anthers 2-locular opening by longitudinal slits, some without, some with a [nectiferous] disk, and mostly with a 3-locular superior ovary with axile (perhaps never parietal) placentation of the ovules, and a capsular fruit and seeds with endosperm and a straight embryo.
He concluded that the only modern taxon that filled the above requirements were his Tiliales which he defined to included the Dirachmaceae, Scytopetalaceae, Tiliaceae, Sterculiaceae and Bombacaceae. Of these, Bombacaceae (anthers 1-locular) and Dirachmaceae (8-locular ovary) can be quickly dismissed, and the anthers of Scytopetalaceae open by terminal pores or transverse slits. Tiliaceae and Sterculiaceae are now more frequently placed (along with Dirachmaceae and Bombacaceae) in the Malvales, and thus it is to here that one might look for supporting data if one were to test this hypothesis. What may prove to be much more significant in Hutchinson's paper is his suggestion that while a portion of the Euphorbiaceae might have originated from a malvanean ancestory, other parts might have come from ancestral groups located elsewhere, namely Celastraceae and Rhamnaceae. That this is indeed the case was recently championed by Meeuse in a privately printed booklet entitled The Euphorbiaceae auct. plur.: An unnatural taxon published in Delft, Holland, in 1990. Not only did Meeuse divided Euphorbiaceae into numerous families, as done earlier by Airy Shaw (in Willis, Dict. Flow. Pl. Ferns, 1973), but he then dispersed them among several different orders. This aspect of the problem will be discussed in more detail under the family heading.

Bremer et al. (1998, 1999) assign both the Euphorbiaceae and the Pandaceae to their broadly defined Malpighiales. As this order also contains the Flacourtiaceae, Ochnaceae, Passifloraceae and Violaceae, this is not surprising. Nonetheless, the Malpighiales are placed in their "eurosids I" whereas the Malvales (including Tilales) are assigned to their "eurosids II." Until the true nature of the Malpighiales, as defined by Bremer and his colleagues is resolved, the placement of Euphorbiaceae will remain in doubt.

While the closeness of Pandaceae to Euphorbiaceae is not in doubt, the relationships of Buxaceae and Simmondsiaceae to the euphorbs are a concern. The two families are related but not nearly so close as to include Simmondsia in the Buxaceae as was once commonly done. In accepting the Simmondsiaceae and then formally establishing the Simmondsiales, I felt the numerous differences in anatomy (Bailey, Amer. J. Bot. 67: 147-161. 1980; Carlquist, Flora 172: 463-491. 1982), pollen morphology (Köhler & Brückner, Wiss. Z. Friedrich-Schiller-Univ. Jena, Math.-Naturwiss. Reihe 32: 945-955. 1983; Nowicke & Skvarla, Amer. J. Bot. 71: 210-215. 1984), the types of sieve-element platids (Behnke, Pl. Syst. Evol. 139: 257-266. 1982), and more recently seed coat anatomy (Tobe & Oginuma, Bot. Mag. (Tokyo) 105: 529-538. 1992) were all sufficient to kept the two separated at the family and ordinal levels.

What I find difficult to accept is the referral of the Simmondsiaceae to the Caryophyllales as proposed by Bremer et al. (1999). Equally curious is the failure of these workers to place Buxaceae (assigned merely to the "eudicots" as a taxon basal to the Ranunculales!). As well be noted below, I believe the two can be associated into a single superorder.

For me at least it was Hutchinson (Fam. Flow. Pl., ed. 3: 228. 1973) who first provided some rationale for assigning Buxaceae (where he included Simmondsia) to the Hamamelididae. Ih his system the family was held within the Hamamelidales, a position he first suggested in the second edition of Families of flowering plants in 1959. Dahlgren (Nord. J. Bot. 3: 119-149. 1983) then aligned the Didymelaceae and Daphniphyllaceae with Buxaceae into the about to be validly published Buxales. Thorne (Bot. Rev. 58: 225-348. 1992) later added the Balanopaceae and took up the correct name Balanopales for the lot, keeping them near the Hamamelidales. Presently, Thorne (1997) assigns the Balanopales to his broadly defined Rosanae (including the lower hamamelids) among a series of orders Cronquist, Reveal and Takhtajan refer to the Hamamelididae. Takhtajan now assigns the Didymelales, Buxales and Simmondsiales to the about to be published superorder Buxanae. This he places in close proximity to the Barbeyales (Barbeyaceae) and Daphniphyllales (Daphniphyllaceae) in individual superorders of which only the name Daphniphyllanae is valid currently; the Barbeyanae is about to be published. In 1997, I retained all of the orders in the Daphniphyllane finding little reason to fragment these orders into a series of superorders when compared to other superorders in the Hamamelidade.

I firmly believe both Buxaceae and Simmondsiaceae will remain associated with the hamamelids and are now conclusively removed from the Euphorbiales. Having gained much of my initial understanding of families and their arrangment from John Hutchinson, it is always nice to see that he has been shown to be right on occasions.

Buxaceae Dumort., 1829

Evergreen or occasionally deciduous shrubs and subshrubs or suffrutescent herbs; leaves opposite (Buxus and Notobuxus) or occasionally alternate (Pachysandra and Sarcococca), simple, entire or tooth (Pachysandra), leathery, the stipules lacking; inflorescences arranged mainly in dense spikes, racemes or heads; flowers small, unisexual and the plants monoecious, rarely some individuals bisexual, actinomorphic, hypogynous, the sepals small, 4 and in two groups of 2, connate basally, imbricate, the petals lacking; stamens 4 and alternating with the sepals or in Notobuxus 6 and 2 opposite the outer sepals and 4 in two pairs opposite the inner sepals, free, the anthers tetrasporangiate and 2-locular, dehiscing by longitudinal slits; gynoecium superior, the carpels 2 (3) and united to form a compound ovary with as many locules as carpels, the styles free, each with an elongate, decurrent stigma, the ovules 2 per carpel and the placentation axile, pendulous, apotropous or anatropous to campylotropous, bitegmic and crassinucellar; fruits a loculicidal capsule, rarely a drupe (Sarcococca), the seed small, the embryo straight and axile to the abundant, oily endosperm; x= 10, 12-14. CA(4 CO0 A4 or 6 GS(2). As defined here, 4 genera and about 60 species. Nearly cosmopolitan in temperate and tropical regions but rather widely scattered.

Buxaceae is traditionally defined as a family of five genera (e.g., Cronquist, Dahlgren, Takhtajan and Thorne, as well as Hutchinson and Bremer et al.), but I find it troubling to include the genus Styloceras, a taxon of three species found in the Andes from Colombia to Bolivia. In my opinion, the Stylocerataceae are to be excluded not only from Buxaceae but the Buxales, and placed in the Daphniphyllales next to the Daphniphyllaceae. An examination of a comparison of the three families might be instructive. The six to many stamens and the number of sepals, along with an array of similar anatomical features suggest the recognition of the family and its reassignment.

The genus Buxus (30 species) is the home of boxwood, that all important Maryland foundation and hedge plant so necessary for the proper colonial landscape image. The wood of several species has been in use by different cultures, but probably none so significant as the arborescent forms of B. sempervirens found in Great Britain that have been used by artisans and craftsmen for centuries. Even more common in Maryland is Pachysandra (6 species, 5 eastern Asia, 1 eastern North America), a ground-covering suffrutescent herb seemingly required to be anywhere there is a shady glen. Sarcococca (14 species) is also cultivated as a ground-cover. It is a plant of southeastern Asia but not commonly seen in our area.

Simmondsiaceae Tiegh., 1899

Xerophytic shrubs or small trees; leaves opposite, simple, entire, leathery, the stipules lacking; inflorescences of axillary, solitary female flowers and capitate cluster of male flowers; flowers small, unisexual and the plants dioecious, actinomorphic, hypogynous, the sepals small, (4) 5 (6), imbricate, fimbriate, those of the male flower short and greenish, longer and accressent in the female flowers, the petals lacking; stamens (8) 10 (12), free, the anthers tetrasporangiate and 2-locular, dehiscing by longitudinal slits; gynoecium superior, the carpels 3 and united to form a compound ovary with as many locules as carpels, the styles free and soon deciduous, the ovules 1 per carpel and the placentation axile, pendulous, apotropous or anatropous, bitegmic and crassinucellar; fruits a loculicidal capsule, the solitary seed large, the embryo straight and axile, without endosperm; x= 13. CA5 CO0 A10 GS(3). A monospecific genus found in the deserts of the Southwestern United States of southern California, southern Arizona and northwestern Mexico.

Although the family name was first suggested by van Tiegham in the late 1890s, and used occasionally even in the scientific literature (e.g., Airy Shaw, Cronquist, Dahlgren, Takhtajan, Thorne), the late Ruurd D. Hoogland and I could not discover a valid place of publication for the name when we were reviewing family nomenclature for the International Association for Plant Taxonomy. Thus, in 1990, we validated the name. In 1998, however, I found this name, and a series of others proposed by van Tieghem, validate in published reviews of his papers. As a result, the family name now stands attributable to van Tieghem alone.

The disposition of this family in my system of classification is reviewed above under the ordinal name Euphorbiales. The Simmondsiaceae are clearly unrelated to Euphorbiaceae and now assigned to their own order, Simmondsiales, and placed in the Hamamelididae rather than the Rosidae.

This shrub or small tree, known best by the common name "jojoba" or less often as "goatnut," is the source of an important oil used in fine lubricants. The cotyledons and the peripheral layers of the hypocotyl are rich in a waxy liquid that can be readily obtained without much difficulty. The oil of jojoba is now found in a wide variety of products, with small farmers growing the plant as a cash crop, particularly in Arizona. It is not particularly attractive, especially when compared to boxwood, but it is having some success as a cultivated plant in the warmer, more arid areas of the United States.

An interesting nomenclatural problem can be illustrated here. The plant was originally found by Thomas Nuttall in 1836 on the foothills east of San Diego in what is now California. Nuttall's labels are noted among western botanists for their brevity, and his crabbed handwriting does not add to their usefulness. Thus it was that when Johann Link described the plant as a new species of Buxus he named it B. chinensis. This was not a "slip-of-the-pen" but clearly intended as Link thought the label said "China." when in fact - if you know Nuttall's hand - it says "N Calif." When Nuttall proposed Simmondsia he changed the name of the species to S. californica, and so it remained for many years. However, the rules of nomenclature dictate that a name may not be changed, even if inappropriate, and therefore C. Karl Schneider proposed the combination S. chinensis by which this western American plant is known today.

Next page

Lecture Notes

Posted: 10 Jan 1998; last revised 28 Feb 1999