Abstract. The following new species, new variety, new name, and 12four new combinations are published for the forthcoming accounts of Micranthes Haworth and Saxifraga L. in the Flora of China, Volume 8: Micranthes dungbooi (Engler & Irmscher) Gornall & H. Ohba, comb. nov., M. gageana (W. W. Smith) Gornall & H. Ohba, comb. nov., M. laciniata (Nakai & Takeda) Gornall & H. Ohba, comb. nov., M. manchuriensis (Engler) Gornall & H. Ohba, comb. nov., M. pallida (Wallich ex Séringe) Losinskaja var. yunnanensis (Franchet) Gornall & H. Ohba, comb. nov., M. paludosa (J. Anthony) Gornall & H. Ohba, comb. nov., M. zekoensis (J. T. Pan) Gornall & H. Ohba, comb. nov., Saxifraga aristulata J. D. Hooker & Thomson var. macrostigma (Franchet) J. T. Pan & Gornall, comb. et stat. nov., S. epiphylla Gornall & H. Ohba, sp. nov., S. gemmigera Engler var. gemmuligera (Engler) J. T. Pan & Gornall, comb. nov., S. heterotricha Marquand & Airy-Shaw var. anadena (H. Smith) J. T. Pan & Gornall, comb. et stat. nov., S. hypericoides Franchet var. aurantiascens (Engler & Irmscher) J. T. Pan & Gornall, comb. nov., S. hypericoides var. rockii (Mattfeld) J. T. Pan & Gornall, comb. et stat. nov., S. sinomontana J. T. Pan & Gornall, nom. nov., and S. sinomontana var. amabilis H. Smith ex J. T. Pan, var. nov. In addition, the following names are lectotypified here: S. macrostigma Franchet, the names S. mengtzeana Engler & Irmscher and, S. mengtzeana var. cordatifolia Engler & Irmscher are lectotypified here, S. micrantha Edgeworth var. yunnanensis Franchet, and S. montana H. Smith, non (Small) Fedde.

The following new species, new variety, new name, and four new combinations are required for the forthcoming account of Micranthes Haworth and Saxifraga L. (Pan et al., in press) in the Flora of China, Volume 8. The segregation of Micranthes from Saxifraga is justified on the basis of certain morphological differences (described by Webb & Gornall (1989) and summarized in Table 1) as well as gene sequence data (Soltis et al., 1996 and pers. comm.).

Table 1. Characters distinguishing Micranthes from Saxifraga.

Saxifraga laciniata was published in 1914 with a very brief diagnosis in a commentary under S. takedana Nakai, with which it was compared. The key passage from this commentary runs "Habitus nostrae S. laciniatae tamen tantisper discrepans. Nulli soboles emittit et acaulis est." A more complete description of the species was subsequently provided by Takeda (1915: 235–236). We here validate an earlier publication of the combination Micranthes laciniata by Hara (1939: 74, as "Micranthes laciniata (Nakai et Takeda) Hara, nom. altern."), which was published as one of two synonyms, both explicitly cited as such, of S. laciniata, and was therefore invalid under Art. 34.1(c) of the St. Louis Code (Greuter et al., 2000).

In the protologue, Franchet cited one gathering, J. M. Delavay 68, but he did not specify the herbarium where it is preserved, nor how many duplicate specimens it comprises. A lectotype is here designated from among the several specimens of this gathering in Delavay’s primary set of duplicates at P.

The protologue and label both give the type locality as being in Yunnan Province, but maps show the co-ordinates as being in Sichuan.

Since the robust Saxifraga macrostigma and the slender S. aristulata intergrade in the Himalaya, particularly in vegetative features such as stature and leaf size, separation of these two taxa at specific level is unjustified.

Two Delavay gatherings were cited in the protologue of Saxifraga macrostigma: nos. 2100 and 3726, both under the same locality and date. There are two specimens of the former gathering and one of the latter preserved at P. Therefore, one of them is here designated as the lectotype. The plate in Plantae Delavayanae cited by Franchet in the protologue ("52, A") was never published.

Species haec a Saxifraga mengzteana Engler & Irmscher differt sinu basalium foliorum embryonibus foliaceis instructis.

Saxifraga epiphylla and S. mengtzeana Engler & Irmscher (1913: 36) are similar species sharing the following features: stolons absent; stem usually hardly branched; basal leaves with blade peltate to broadly ovate, rarely reniform, hairy and spotted at least abaxially, base cordate; cauline leaves few, lanceolate to ovate; flowers zygomorphic, disc obscure; petals not callose, usually arcuate veined; filaments clavate. However, S. epiphylla differs from S. mengtzeana chiefly in that it produces a foliar embryo in the sinus of the basal leaf blades. The new species also has basal leaf blades glandular hispid on both surfaces, whereas S. mengtzeana has blades glabrous adaxially.

In Pan (1992: 74--75, pl. 14), the name Saxifraga aculeata I. B. Balfour (1916: 70) was misapplied to the species here described as S. epiphylla. However, the holotype of S. aculeata, A. Henry 10316B (E), which lacks foliar embryos, is also an isosyntype of S. mengtzeana var. cordatifolia Engler & Irmscher (1913: 37), a taxon that we regard as synonymous with S. mengtzeana var. mengtzeana (see below). Indeed, there is no reference to foliar embryos in any of the protologues. The taxon with foliar embryos therefore constitutes a new species, which we here formally describe as S. epiphylla, the epithet referring to the position of foliar embryos.

In their protologue for Saxifraga mengtzeana, Engler and Irmscher divided the species into two varieties: var. cordatifolia and var. peltifolia Engler & Irmscher (1913: 37). They did not cite an autonym, such as a variety "typica," which was their usual procedure, e.g., in their monograph of Saxifraga (Engler & Irmscher, 1916, 1919). Three specimens were cited under S. mengtzeana: two under variety cordatifolia (A. Henry 10316 and A. Henry 10316B, syntypes both at B) and the third under variety peltifolia (A. Henry 9118, holotype at B). No specimens were cited under the species before the two varieties. We regard variety cordatifolia as constituting the type variety (autonym) because it is cited first in sequence (as "1. Var. cordatifolia"). In the accounts of polymorphic species in their monograph of the genus, Engler and Irmscher (1916, 1919) routinely treated the type variety first, usually as "var. a . typica." We can infer that in this particular case the authors wanted to make a nomenclatural contrast between the common plant, with ovate-cordate leaves, and the less common variant, with peltate leaves. To clarify this taxonomic ambiguity, we are lectotypifying both S. mengtzeana and its variety cordatifolia on the same specimen. The three specimens at B have not been seen and are probably destroyed. Given this, we here designate as the lectotype of S. mengtzeana and S. mengtzeana var. cordatifolia an extant isosyntype, Henry 10316B (E). This specimen is also the holotype of the later name S. aculeata, which now becomes a nomenclatural synonym of S. mengtzeana and a superfluous name. However, S. aculeata remains legitimate because it was not nomenclaturally superfluous when first published (Saint Louis Code, Art. 52 Note 2, Greuter et al., 2000).

Finally, it is worth pointing out that the holotype of Saxifraga mengtzeana var. peltifolia (A. Henry 9118, at B) is (or was) part of the same gathering as the holotype (at E) of S. henryi I. B. Balfour (1916: 72). To our knowledge there are no problems surrounding the application of these names, other than the question of whether the peltate-leaved variant is worth recognizing as distinct from S. mengtzeana.

Engler and Irmscher (1916: 141, 150--151) distinguished Saxifraga gemmigera from S. gemmuligera (Engler) Engler & Irmscher on the grounds that the former, which occurs in Shaanxi Province, has linear cauline leaves, whereas the latter, which occurs in Gansu, Qinghai, and Sichuan provinces, has oblong to subovate-oblong, median and distal cauline leaves. Although this is true, the variation is such that we can find no discontinuity worthy of specific separation. We therefore propose to treat the taxa as varieties, representing the two ends of a morphological continuum.

Engler and Irmscher (1912: 601) commented on the identity of the two syntypes of Saxifraga unguiculata var. gemmuligera at LE. They explicitly identified as S. gemmuligera the Przewalski specimen of 1872 (which they dated as July and numbered as 221) and as S. unguiculata sensu stricto the Przewalski specimen of 1880 (which they dated as June and numbered as 360). Formal lectotypification will have to await study of both syntypes at LE.

It should be noted that the name Saxifraga gemmifera Persoon (Syn. Pl. 1: 490. 1805), whose epithet some might think is sufficiently similar to S. gemmigera as to constitute an earlier homonym under Article 53.3 of the Saint Louis Code (Greuter et al., 2000), is invalid under Article 34.1(c) (i.e., it was published as a synonym) and consequently has no nomenclatural status.

This taxon is identical to Saxifraga heterotricha in every way except that the stem and cauline leaf margin have eglandular rather than glandular hairs.

The variety aurantiascens agrees in its vegetative morphology with Saxifraga hypericoides rather than with S. aristulata var. macrostigma, particularly in having adaxially pubescent leaves (rather than glabrous ones as in S. aristulata var. macrostigma). Saxifraga hypericoides var. aurantiascens is distinguished from variety hypericoides chiefly in having a solitary flower.

Rock’s plant differs from variety hypericoides in having petals that are rounded or cuneate at the base, rather than cordate at the base.

This new name is needed because Saxifraga montana H. Smith is a later homonym of S. montana (Small) Fedde, a validly published combination based on Micranthes montana Small, described from California, U.S.A. We are grateful to James L. Zarucchi (MO) for drawing this problem to our attention and for suggesting the replacement name, which we think is most apposite.

Varietas haec a var. sinomontana differt paginis foliorum basalium pilis rufescentibus crispulis instructis.

This variety differs from the polymorphic Saxifraga sinomontana var. sinomontana in having crisped, rufous, eglandular hairs on the adaxial surface of the basal leaves. This taxon was recognized at the rank of species by Smith (in sched., H. Smith 11262 and H. Smith 12278, both at BM), but the name was never published.

Acknowledgment. We thank Nicholas J. Turland (MO) for much help in preparing the manuscript.