Taxus cuspidata

Rigid branch yew

The World Botanical Associates Web Page
Prepared by Richard W. Spjut
April 2003; photos added May 2006; updated August 2007

23. Taxus cuspidata Siebold & Zuccarini (Fig. 147–152, 279–282), Abh. Königl. Bayer. Akad. Wiss., Math.-Phys. München, III, 801. 1843 (in footnote: ref. to Siebold and Zuccarini, Fl. Jap. II. tab. 128 and to [a description] T. baccata in Thunberg(1784, Fl. Jap. 275, “foliis solitariis, linearibus, cuspidatis, approximatis”). No specimens cited. Taxus baccata (var.) cuspidata Carričre, Conifčrae 733. 1867. Taxus baccata ssp. cuspidata (Siebold & Zucc.) Pilger, Planzenreich 18 (iv, 5): 112. 1903. Type: Japan. Lectotype (designated in Spjut, 2007b) M! (Fig. 279) with the following data on labels “Legit in Japonia et Communicavit de Siebold, anno 1842, ”Herbarium Regium Monacense," “Herb. Zuccarini,” and “Herbar. Univ. Ludov. Maximil.,” the specimen on the right of two mounted on the same sheet (Fig. 279), with mature male cones; the left specimen also identified here as T. cuspidata is believed to have been taken from a different plant. Isolectotypes M!, K!, apparently from the same plant but with other data on labels.

Rigid branched yew. Distribution: Korea, Japan.

Shrub or tree; branchlets arising from a leader branch, subpinnately arranged branchlets occasionally with shorter recurved branchlets, appearing pendant in the type, or spreading revurved, reddish orange; bud scales persistent to the 2 nd or 3 rd yr, 3 or 4-seriate, paleaceous, brownish, lower scales closely adnate, deltoid, indurate, ca. 1 mm long, with a distinct narrow midnerve, the upper scales cucculate—folded inwards over an obscure horny nerve, especially towards the apex, thus, appearing cuspidate, 1.5–2 mm long. Leaves persistent on older twigs, dark green, mostly in two ranks, or secund near base of branchlets, or with an occasional upper or lower leaf spreading at right angles to lateral leaves, especially on young shoots, linear, abruptly and uniformly recurved near margins, convex above to a rounded or squarrose or acute midrib that forms a channel at the base, slightly concave below to a slightly rounded greenish or yellowish midrib; adaxial (upper) epidermal cells wider than tall, 10–18 µm tall, 25–40 µm wide, elliptic, thin-walled; abaxial (lower) epidermal cells epapillose in 8–20 rows from the margin, these 10–15 µm tall, 15–40 µm, 3–10× l/w, thin-walled, papillae absent or present on up to one-half of the marginal cells, absent or present on outer midrib cells, positioned submarginally or medially on subsidiary and accessory cells, in 1–2 rows, concrescent; stomata bands green or yellowish-green, broader than the marginal region, with 11–13 interrupted rows of stomata/band. Male cone scales 5 seriate; microsporophylls ca. 8 spreading in an umbrella-like manner, united below into a flattened ribbed column, each sporophyll with 5 cucculate microsporangia. Female cone scales decussate-imbricate in 5 ranks; seed ellipsoidal, 4-angled near apex, to 7 mm long, 5 mm diam, with a short nipple.

     Although T. cuspidata has been the name generally applied to any yew that occurs naturally in Japan, Korea, E Russia, and NE China, the type seems least representative.  It does have prominent carinate bud-scales considered diagnostic by Siebold and Zuccarini (1870), and by others (e.g., Carričre 1855; Endlicher 1847; Pilger 1903; Rehder 1940). However, this feature is difficult to evaluate due to rather subtle variation in bud-scale characteristics among Eurasian yews.  The characters involve persistence, size (0.5–2 mm), and curvature (Cope 1998).  Moreover, T. cuspidata has proved difficult to define because of lack of taxonomic studies in the genus, particularly in typification, interpreting relationships to cultivars, and to other temperate Eurasian species.

     The authorship of the name itself (T. cuspidata) is even controversial because the official date of the volume for Siebold and Zuccarini’s Flora Japonica in which T. cuspidata appeared was not until 1870; yet, reference to this publication, including a figure (Tab. 129 instead of Tab. 128), was first made by Zuccarini (1843) 27 years earlier at which time Zuccarini cited the name in footnote with further reference to a description in Thunberg (1784, Flora Japonica).  Siebold and Zuccarini's (1846) synopsis of plant taxa in Japan, which many authorities have cited as the valid publication date, does not provide a plant description of T. cuspidata; technically, this is a “nomen nudum” (Nakai 1938); however, reference was made to Thunberg (1784).  Aside from Zuccarini (1843), the first direct description of T. cuspidata appears in Endlicher (1847). Others also described the species (e.g., Carričre 1855; Lawson et al. 1851; Parlatore 1868) before Siebold & Zuccarini (1870), who had noted these earlier references. However, Zuccarini's (1843) reference to a description in Thunberg (1784) satisfies the requirement for valid publication of “Taxus cuspidata Sieb. & Zuccar.” (Art. 32.1), indicating also the origin of the epithet (“cuspidatis”).

     Among specimens of T. cuspidata from the Zuccarini Herbarium (syntyptes), distributed to various herbaria, I have seen those at Harvard (GH), Kew (K), and München (M).  Not all appear from the same plant; indeed, three different plants may be represented, two of which may be mounted on one sheet. Data on labels and specimen sheets are also not the same; K has two sheets, one annotated by Zuccarini in 1843, and another indicated to be a duplicate that was apparently annotated by Maximowicz during 1860; GH has one sheet that has the same kind of label as the lectotype, but the year (of annotation) is somewhat illegible, appearing to be 1862, or 1842, and it also bears a branch from a female plant with handwritten data “Fructus,” “Yeso,” “W. P. Blake,” “1862”; M has three sheets, two of which lack reference to dates. The lectotype was selected based on leaf and bud-scales features that best matched the description. Additionally, the illustration in Siebold and Zuccarini (1870, Tab. 128) does not match any of the syntypes or fide ICBN terminology—uncited specimens (original material); rather, I consider it to be T canadensis var. adpressa.

     Taxus cuspidata is apparently rare except perhaps in arboreta.  Figure 280 shows one of many cultivars at the Secrest Arboretum in Ohio (Chadwick & Keen 1976) that I recognize as belonging to this species by the leader shoots that ascend upwards, and by the long and undivided thick branchlets that are either pendant or recurved with or without right angled recurved branchlets. The leaves of T. cuspidata are often distinct for their rigid erect orientation.  In addition to the type, two specimens I studied at the Arnold Arboretum Herbarium—that were collected by Jack from Hokkaido—are undoubtedly T. cuspidata.  In the typical specimens, the leaves are not as neatly two-ranked, the uppermost leaves often point in the direction of the branch axis.

     Taxus cuspidata can be difficult to distinguish from T. umbraculifera.  Leaves of T. cuspidata on young shoots appear shorter (oblong) and more radial in orientation.  However, as noted above, the recurved ± pinnately arranged branchlets distinguishes T. cuspidata.  Taxus umbraculifera has wide spreading, often ascending branchlets, except for the divaricately branched var. microcarpa, which can be recognized by the leaves with a ± strong two-ranked orientation near apex of branchlets as opposed to the upwardly secund orientation in other varieties and T. cuspidata.


     The habit of T. cuspidata, as shown above in the photos next to the type specimen, is similar to the European Dovaston yew (T. baccata var. dovastoniana Leighton)—known for its weeping branches—and to other European plants in horticulture known as T. baccata var. glauca Carričre (US: 1396503 “f. glauca = f. subglaucescens Jacq.;” ex Hillier's Arboretum BM; ex Herb. Petropolitani, Szovich 610, S; C. Baenitz US, “f. pendula”).  Thus, I wonder if T. cuspidata, as recognized by Siebold and Zuccarini (1870), is of horticultural origin.  Wilson (1916) commented that he was not sure whether the yew plants he saw in Japan were natural or cultivated, and since Siebold & Zuccarini (1870) noted that yew there occurred in horticulture, and in cultivation around temples, and spontaneously in the mountains, it seems quite possible that the original material came from horticulture.   However, at least one other specimen that was collected by Wilson from Korea, indicates this may be a natural species.

Representative Specimens—Korea: Monastery and over the top of Hallai-san down to Mushroom House, bush and small trees abundant as undergrowth and on upper slopes, Wilson 9484 (A); Quelpaert Island, In-Cho 4387 (F). JapanHokkaido: Sapporo, 21 Aug 1905, Jack s.n. (A). [Honshu] locality? 1862, Siebold s.n., three specimens on one sheet, upper left is T. cuspidata, one showing underside of leaves is probably part of type collection (original material), Herb. Zuccarinii label, p.p., with fragment on lower right from Yeso [Hokkaido], is T. umbraculifera var. microcarpa (GH); Insula Nippon, ubiquitous in cult. and in forest of Yezo regions, Dec 1904 Faurie s.n. (A). CultivarsSecrest Arboretum, Ohio: ‘Hunnewell’ forms: ‘Anglo-Japanese Yew’ A31-711; ‘Siebold’ A30-245; ‘Media Burr’ A30-16; ‘Thayer Yew’ A30-209; ‘Hunnewell Yew’ A30-149; ‘Cuspidata Adams’ A31-9.


JapanHokkaido: Sapporo, 21 Aug 1905, Jack s.n. (A).

JapanHokkaido: Sapporo, yr 1905, Jack s.n. (A).

Comparison of Taxus baccata with T. cuspidata (M)

Cultivated: US National Arboretum, collected by Doren, det. by Keen as T. x hunnewelliana

Korea: Monastery and over the top of Hallai-san down to Mushroom House, bush and small trees abundant as undergrowth and on upper slopes, Wilson 9484 (A)

Cultivated: Arnold Arboretum, type for T. cuspidata f. thayerae.

Cultivar: Secrest Arboretum, OH: 30-182.  Observe the many recurved branchlets; compare with specimen collected by Jack from Hokkaido shown above.

Taxus cuspidata lectotype (M). Specimens collected by von Siebold, possibly during the years 1823–1829 when von Siebold was in Japan (near Nagasaki), thus it would seem to have been received by Zuccarini in 1842 (either sent by von Siebold or obtained from the Rijksherbarium during his visit there in 1842).  However, the lectotype and other specimen compare with specimens in the Arnold Arboretum Herbarium (A) collected by Jack from Hokkaido. Thijsse (2005) has indicated von Siebold did obtain specimens from Hokkaido even though it was impossible for him to have visited Hokkaido.

Cultivars: Secrest Arboretum, OH: Top left photo shows male cones clustered in development near ends of branchlets as also seen in left specimen on type sheet at M.  The lower left photo shows more scattered cones as in the lectotype.  The top right photo appears to be a hybrid, T. cuspidata x T. caespitosa var. latifolia.  The lower right photo shows arillocarpia of T. cuspidata

Japan—Original Material (GH).  The upper left specimen compares favorably with the lectotype (M) in leaf characteristics and conspicuous scales at base of branchlets, but lacks males cones as seen in the lectotype.  This specimen, which may have been distributed by Miquel, Musei Botanici Lugduno-Batavi [“Herbarium Japonicum” Thijsse (2005)], was possibly collected by von Siebold from Japan.  Also, the Herb. Zuccarini label indicates the specimen above it was collected by Von Siebold in Japan; thus, it would suggest original material but the year of collection (or preparation of the label) for one of the male specimens appears to be 1862, whereas the lectotype was clearly indicated to be received by Zuccarini in 1842.  However, Zuccarini died in 1848 so he could not have prepared the label in 1862, but he did visit the Rijksherbarium in 1842 in Leiden (Thijsse 2005).   The small female branchlet is evidently an annotation by W. P. Blake indicating the specimen was collected from Yeso [Hokkaido] in 1862.  It may be further noted that Maximowicz in 1862 collected near Yokohama as seen on a US specimen. This female branchlet and the larger adjacent male specimen belong to T. umbraculifera var. microcarpa.  The two species are distinguished by leaf characters as indicated in the key. Additionally, note that in T. cuspidata the longest leaves on the branchlet are towards the base of the branchlet; the leaves, thus, appear to increase in length during the growing season.  In T. umbraculifera var. microcarpa, the reverse occurs, the shortest leaves are seen near the base of the branchlet (or current season growth).

Cultivar: Private residence, Andover Hills in Laurel MD.  Shrub estimated to be around 35 years of age at the time of photograph, 12 October 1998.


Thijsse, G. (2005). The history of the Herbarium Japonicum Generale in Leiden. In The Botanical Collections: Proceedings of the symposium 'Siebold in the 21st Century' held at the University Museum, the University of Tokyo, in 2003. edited by H. Ohba and D. E. Boufford, Bull. 41, Tokyo