Designation of Lectotypes

Ridley (Reference Ridley1912, p. 7) listed the following specimens, all of which are therefore syntypes: Gunong Jerai (Ridley 5209, 5356 and 5357), Mount Ophir (Ridley 3310), Hullett 775 and Derry 608. We have been able to trace the following materials: Gunong Jerai (Ridley 5209 [SING0222424], 5356 [SING0222425] and 5357 [SING0222422]), Mount Ophir (Ridley 3310 [K000800732]), Hullett 775 (SING0222423) and Derry 608 (SING0222420 and SING0222421). Of these, Ridley 5209 (SING0222424) has only a single unopened flower bud; Derry 608 (SING0222420) and Ridley 5356 (SING0222425) have very few flowers and both have lost a number of their leaves; and Derry 608 (SING0222421) has lost parts of the stem and petioles and has few flowers. Of the remaining materials, Ridley 5357 (SING0222422) is in better condition, with more flowers than either Hullett 775 (SING0222423) or Ridley 3310 (K000800732) and is designated as the lectotype above. Two synonyms of Decaspermum montanum are discussed below.

Although this species has been quoted as being first published by Gagnepain in Lecomte, Fl. Indo-Chine 2: 846 (Reference Gagnepain1921) (May 1921, according to Stafleu & Cowan, Reference Stafleu and Cowan1979), this publication is pre-dated by Gagnepain (Reference Gagnepain1920). This is important, because the publications differ slightly in the information given. Gagnepain (Reference Gagnepain1921, p. 847) listed only a single collection: Pierre from “prov. de Thepong (Pierre)”, but with no collection number. Gagnepain (Reference Gagnepain1920) gives this information and adds the collection number “985”. The collection Pierre 985 bears the following information on the herbarium label, ‘Herb. L. Pierre. No. 985 Nelitris Arbuscula 2–4 m. Hab. Crescit in prov thepong in Cambodia 5/1870 Coll. LP’. Numerous duplicates of Pierre 985 exist (A00068988, K000261633, P00467145, P00467146 and P00467147). It is almost certain that these are multiple preparations from a single gathering and represent duplicates of the type gathering (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012: Article 8.3). To assume otherwise would be almost perverse, but to avoid all doubt, we believe that it is appropriate to lectotypify Decaspermum cambodianum from the available specimens. Of these, only P00467145 bears original drawings of this species and it is in the best physical condition; it is therefore selected as the lectotype above.

Ridley mentions only a collection of “Mahommed Nur 18912” from “Pulau Tioman, Gunong Kajang, 2,500 ft.” Two sheets of Nur 18812 from that locality exist at K (K000800733 and K000800734), but both from the slightly higher elevation of 2700 ft and collected on 2 May 1927. We treat these differences in altitude and collection number as simple typographic errors (McNeill et al., 2010: Article 60.1) by Ridley. K000800733 is the more complete specimen and is chosen as the lectotype above.

Dumont de Courset (Reference Dumont de Courset1802) provides a description but does not cite any specimens nor refer to any plates. Brophy et al. (Reference Brophy, Craven and Doran2013) follow Craven (Reference Craven2006) and treat this species as being first described by Curtis (Reference Curtis1794) as Meterosideros citrina, but there is no direct indication in Dumont de Courset's (Reference Dumont de Courset1802) description of the species that it is based on Curtis (Reference Curtis1794), and he makes no mention of Curtis in the list of works consulted. In the ‘Avertissment’ to the second edition of his work (Dumont de Courset, Reference Dumont de Courset1811), however, Dumont de Courset indicates that he has consulted “le magasin de Curtis”, and there is good evidence that it was his intent to base his name on that of Curtis; for example, the epithet chosen and the locality cited are the same. Furthermore, the full title of Dumont de Courset's work indicates that it follows the system of Jussieu (Reference Jussieu1789), who in his Appendix (p. 453) included Metrosideros under Melaleuca.

As no contrary evidence exists and as Dumont de Courset fulfilled the conditions for valid publication of Melaleuca citrina as the name of a new species, we believe that it should be treated under Article 41.4 of the International Code of Nomenclature (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) as based on Metrosideros citrina Curtis (Reference Curtis1794).

For many years, this species was known as Callistemon citrinus (Curtis) Skeels. We follow Brophy et al. (Reference Brophy, Craven and Doran2013) and Craven (Reference Craven2006), who treat Callistemon as a synonym of Melaleuca. When this combination was first published, Skeels (Reference Skeels1913, p. 49) referred to the correct plate number but to Curtis's Botanical Magazine volume 7 rather than 8; this is a correctable error (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012: Article 41.3).

Curtis (Reference Curtis1794) states that his drawing was made from a “plant which blossomed toward the close of last summer at Lord Cremornes, the root of which had been sent from Botany-Bay”. We cannot trace this specimen, and therefore the sole remaining material available for typification is the plate in Curtis (Reference Curtis1794). As no one, including Skeels (Reference Skeels1913), who cited only Curtis (Reference Curtis1794), has named a type, the name requires lectotypification. We designate Curtis (Reference Curtis1794, t. 260) as the lectotype.

Other synonyms of Melaleuca citrina that appear to require typification, being often cited in synonymy and considered as legitimate names but which are in fact superfluous and illegitimate and automatically typified by the type of the name that ought to have been adopted (Article 7.5), include the following.

Psidium littorale will be treated by us in the Flora of Cambodia, Laos and Vietnam as a synonym of P. guajava and was first described in a preprint of the more commonly cited work by Raddi (Reference Raddi1823): Opuscoli Scientifici, volume 4 (p. 254, t. 7, Fig. 2). Unfortunately, Raddi (Reference Raddi1821) cited no specimens in the protologue and referred only to a single illustration of his own, indicating that he must have had a specimen or specimens available to him (J. McNeill, Royal Botanic Garden Edinburgh, personal communication). Later in the text, Raddi does refer to a sterile plant growing in the Botanical Gardens in Florence. M0146756 is labelled Psidium littorale and bears a label stating ‘Herbarium Regium Monacense & Herb Zuccarini & Herbar. Univ. Ludo. Maximil.’, with the written annotation by Otto Berg ‘hort. berol. Otto’; it has tentatively been verified by Dickoré as type material on JSTOR, but we have traced no associated publication. Given that Berg's dates are 1815–1866, it is reasonable to conclude that it was impossible for Raddi to have seen this material in 1821, and we can find no indication that Raddi saw this material or used it in constructing the description. As we can trace no specimens that might have been seen by Raddi, it therefore appears that Raddi (Reference Raddi1821) t. 1., Fig. 2, is the only remaining original material, and we designate it as the lectotype above.

Psidium cattleyanum Sabine will be treated by us in the Flora of Cambodia, Laos and Vietnam as a synonym of P. guajava. Sabine (op. cit.) coined the name based on a drawing by Hooker and an account written by William Cattley and sent to him. Cattley certainly sent specimens of fruits to Sabine, but unfortunately the original herbarium of the Royal Horticultural Society was auctioned off in the mid-nineteenth century to alleviate the Society's debts, and no relevant herbarium material of Cattley's has been traced by us. Hooker's drawing is, however, excellent and is nominated as the lectotype above.

Rhodamnia siamensis is a taxonomic synonym of Rhodamnia dumetorum (DC.) Merr. & L.M.Perry in J. Arnold Arbor. 19: 195 (Reference Merrill and Perry1938). Craib (Reference Craib1926) cites Kerr 2147 as the type but does not specify the herbarium where the holotype is located, so his nomination can be taken as a first-stage lectotypification. Duplicates of Kerr 2147 exist in ABD, BM (BM000944014) and K (K000800683). Of these, the material in ABD is by far the most complete, was obviously seen by Craib, who worked for many years in Aberdeen, and is nominated as the lectotype above (second stage). ABD does not, at present, use barcodes for its specimens.

Hance (Reference Hance1876, p. 259) cites a Pierre collection from “Phu kok, sinus Siamensis, a 1874”. BM000926077 ex Herb. Hance from Herb. Pierre no. 1440 and bearing the Pierre number ‘19212’ has all the above information, as does K000261634. The former is the larger and more complete specimen and so is designated as the lectotype above.

Application of Article 9.9 (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) to Various Taxa

Article 9.9 deals with the use of “a term as defined in the Code as denoting a type, in a sense other than that in which it is so defined”. The article indicates that this is to be “treated as an error to be corrected (for example, the use of the term lectotype to denote what is in fact a neotype)”. We deal with five examples where we apply this Article.

Bean (Reference Bean1997, p.248) states that P. Osbeck s.n. LINN 505.1 from China (Guangzhou [Canton], Guangdong Province) is the holotype. Based on Bean (op. cit.), Jarvis (Reference Jarvis2007), in accordance with Article 9.9 of McNeill et al. (Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012), indicates that the above is in fact a lectotype. LINN 505.1 comprises at least five separate fragments, mounted on a single sheet, that were probably derived from a single specimen, and therefore second-stage lectotypification (McNeill et al. Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012: Article 9.17) does not apply.

Scott's designation of a holotype is correctable under Article 9.9 (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) to lectotype.

Blume (Reference Blume1825–1826, p. 1100) gives no collection details beyond stating “in cacumine montis Gede”. Numerous collections by Blume of this taxon exist. These include L0009496 (comprising two specimens, this is labelled as the holotype but with no details appended on collection locality; based on the handwriting, the annotation is almost certainly by J. Thompson). Thompson's designation of a holotype is correctable under Article 9.9 (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) to lectotype.

The specimens on L0009496 are somewhat different in pubescence; the left-hand specimen lacks flowers and is pubescent on its uppermost branchlets and along the abaxial midrib of the leaves, whereas the right-hand one lacks these features and has pubescent hypanthia. Because of these differences, it is not absolutely certain that the specimens form part of the same collection, but we prefer to treat them as such in the absence of conclusive evidence to the contrary, thereby obviating the need for second-stage lectotypification (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012: Article 9.17).

Scott's designation of a holotype is correctable under Article 9.9 (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) to lectotype.

Scott's designation of a holotype is correctable under Article 9.9 (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) to lectotype. ‘Singkara’ must be a misspelling of Sinkara in Sumatra and is not Singkara in the Philippines.

Blume (Reference Blume1850, p. 78) describes Rhodamnia cinerea and cites a number of previous publications by way of reference but does not cite any collections or specimens. Among the literature references, Blume (Reference Blume1850) states “Ejusd. Icon. Tab. 524. – Cum Var b(beta) in montanis Sumatrae.”, by which he refers to Wight's Icones (Reference Wight1843) and may be indicating that this illustration refers to his var b(beta), which is var. concolor. Wight's tab. 524 and its accompanying description are of Rhodamnia dumetorum (as Monoxora spectabilis), but there is no indication in his description of the species that the leaves are concolorous. Indeed, the preceding generic description suggests that the leaves are always discolorous.

Later, Miquel (Reference Miquel1860, p. 315) adds that this taxon (as Rhodamnia concolor Miq.) has been found near “Singkara”, which, as indicated above, must be a misspelling. The collector is indicated by Miquel (Reference Miquel1860) as T. (= Teijsmann). A.J. Scott (in herb. U in 1977) indicates that U0005195 is the holotype of Rhodamnia concolor Miq. (the label on this specimen reads: ‘841 HB’ [= 841 Herbarium Bogorense] ‘Rhodamnia concolor Mara’ ?’paejan’?’ [the latter is probably a vernacular name] ‘Sinkara’ [Sinkara is at decimal degree –0.683333S, 100.6E] ‘WK’ [= West Coast (Kust in Dutch)]. ‘Sum.’ [= Sumatra]). But in 1979, Scott cites Korthals s.n. (L0009541) as the holotype of Rhodamnia cinerea var. concolor Blume, and this attribution must be followed.

Scott's designation of a holotype is correctable under Article 9.9 (McNeill et al., Reference McNeill, Barrie, Buck, Demoulin, Greuter, Hawksworth, Herendeen, Knapp, Marhold, Prado, Prud'homme van Reine, Smith, Wiersema and Turland2012) to lectotype.

In publishing the new combination in Rhodomyrtus, Hasskarl (Reference Hasskarl1842, p. 35) wrote “146. Rhodomyrtus tomentosa mihi (Myrtus Bl. Bijdr. 1051. DC. Prdr. III. 240.)”. His citation of Blume's Bijdrage 1051 is an error for 1081, where Myrtus tomentosa is ascribed to “Ait.”, as it is also in de Candolle (Reference Candolle1828, p. 240), who included the reference “hort. kew 2. p. 159.”

Comments on Other Names Yet to be Typified

Scott (Reference Scott1979b) indicates that the “Type” is from the Moluccas and was collected by Roxburgh with the holotype remaining unlocated in Herb. Lambert. The disposal of Lambert's herbarium is complex. Miller (Reference Miller1970) indicates that the final resting place of some of the material remains obscure, with the rest scattered throughout at least eighteen European institutions. The first author had hoped that relevant material would have been found owing to the recent availability of imaged material from many of the main institutions. Unfortunately, this is not yet the case; however, the search has been narrowed, as the material does not appear to be in AAH, B, BM, BR, FI, G, K, L, OXF or P.

Powell (Reference Powell1809, pp. 21–22) stated, “This oil was supposed to be the produce of Melaleuca Leucodendron (Med. Bot. t. 229), but it appears from specimens of the tree yielding the true Cajuputi, sent home by Mr. Christopher Smith, that the species is different, and referable to tab. 17, of Rumphius's Herbarium Amboinense (vol. ii), and not to that author's ‘Arbor alba,’ tab. 16. After a careful examination of specimens in Sir Joseph Banks's, and other collections by Dr. Maton, and of those in the Linnean Herbarium by Dr. Smith, we are authorised to consider the tree which yields the above oil as a new species, and, from the name of its medicinal product, these gentlemen have agreed to give it the appellation of (Melaleuca) Cajuputi.” Powell (op. cit.) gave no description or diagnosis, referring only to Rumphius's tab. 17, which consists of two very different figures. They were both considered by Rumphius to illustrate a single species, and tab. 17 was directly associated (in the “Explicatio”, p. 77) with adequate descriptive material to satisfy the requirement of Article 38.1(a) for a “previously and effectively published description or diagnosis”. Blake (Reference Blake1968), in adopting Powell's name, indicated that Roxburgh (Reference Roxburgh1832) only considered plate 1 to be relevant in the case of Melaleuca cajaputi (Blake's reference, on p. 26 of his paper, to “p. 347” in Roxburgh is an error for p. 397). As Roxburgh did not use the word “type” (Article 7.10), this cannot be taken as typification, but it appears to have been the general application ever since. Therefore Craven, the world expert on the group, was justified in his use of the name Melaleuca cajaputi (Craven, Reference Craven2006), recently repeated in the latest monograph of Melaleuca (Brophy et al., Reference Brophy, Craven and Doran2013). This is fortunate, as the name Melaleuca cajaputi is used widely in Asia because the essential oil extracted from plants with this name forms the basis of a large industry.

Melaleuca minor Sm. in Rees (Reference Rees1812), sometimes treated as a synonym of M. cajaputi, notably by Blake (Reference Blake1968), has an extensive description in which Smith mentions receiving material from Mr Christopher Smith and in which is cited “Rumph. Amboin. V.2.76. t.17.”, with the appended comment that “Rumphius's plates are by no means calculated to give a just idea of the foliage of either, especially of the present, but his descriptions are excellent.” So, as Smith (1812) pointed out, there is a link between Rumphius's plate and his text. Material in the Smith herbarium in LINN (LINN-HS 1243.3) may be that received from Christopher Smith.

Craven (personal communication) worked for many years on Melaleuca and, at the time of his early and untimely death, was reluctant to lectotypify M. cajuputi. We are similarly hesitant to undertake a task that Craven indicated was difficult, and so believe that lectotypification is best carried out by a specialist working exclusively on Melaleuca. As McNeill (personal communication) points out, only material available to Rumphius is eligible to be selected as type of Melaleuca cajuputi (Article 7.7), and this must mean that Rumphius's plate 17 is the obligate lectotype, although presumably a selection of Fig. 1 or Fig. 2 should be made.