two publications (1974 and 1981) Leach sought to establish the validity of Euphorbia candelabrum Welwitsch. The epithet was first used by Welwitsch in a letter dated 2 March 1854 and written in German, to Richard Kippist Esq., who read out a translation at a meeting of the Linnaean Society in June 1854. This translation was subsequently published in the Linn. Soc. Proceedings (1855). Halfway through the translation (p. 328) is a reference to a Euphorbia: Euphorbia I have already found near Loanda a gigantic species, with a stem 21/2ft. in diam. and upwards of 30ft. high, forming woods as Pinus sylvestris does with us!. Towards the end (p. 329) is another reference: In the woods of Euphorbia (Candelabra, n. sp.) is found a wonderfully beautiful terrestrial Orchidea. Leach (1981) maintains that these two references constitute a valid publication and adequate description by Welwitsch. However, Welwitsch obviously had not intended his letter to be published, and at the time he wrote it, in accordance with Article 34.1(b) of ICBN (1983), the name was 'merely proposed in anticipation of the future acceptance of the group concerned'. As such it did not constitute a valid publication, but at best, may possibly have been effectively published as Euphoria candelabrum Welw. (Candelabra), or perhaps E. candelabrum Welw. ex Kippist. Furthermore, the dimensions of the plant briefly given by Welwitsch in the friendly context of his letter, can hardly constitute a description, certainly not a diagnosis when many other such trees were already known, and thus, also, the requirements of Article 32.1(c) are not satisfied. addition, there is no positive indication that this particular Euphorbia species should be associated with the Euphorbia (Candelabra, n. sp.) referred to at the end of the letter. The name was next mentioned in a list published the following year in Coimbra (1856). This consisted of information attached to a consignment of 200 living plants collected by Welwitsch and sent by him from Loanda (Angola) on 21 Aug. 1854. His accompanying letter clearly shows that this information was not intended by him for publication, but simply as an aid in the cultivation of the various cuttings, bulbs and plants at Coimbra and other horticultural establishments in Lisbon. Of the names in this list, 'Euphorbia spec. (Euphorbia candelabrum Welw. mspt.)' appeared under item No. 5, of which there were 20 individuals. The addition, in brackets, of an alternative manuscript name to 'Euphorbia spec.' clearly indicated Welwitsch's proposal only of the name, 'in anticipation of the future acceptance of the group concerned', and thus was once more invalidly published according to Article 34.1(b). It is also confirmation that Welwitsch had not prepared his new species for publication by the time he wrote his letter to Kippist. That the description in the list sent to Coimbra (translated from the Portuguese) of 'a candelabriform tree 30-45 feet high ... with purplish flowers in innumerable quantity' could perhaps be considered as diagnostic, does not affect the validity of the publication of the name; nor does the mention of 'Euphorbia candelabrum' in the notes under item No. 18 Orchidea terrestre alter the position. Although it is shown here to be a later homonym of Kotschy's name of 1857, the name for Welwitsch's species was eventually validly published by Hiern (1900) when a full description was given with the Welwitsch collection number 641 cited. Several localities from the Loanda district were included under this number (Cacuaco Jan. 1854, Mutollo 23 July 1854 and Loanda 17 June 1858), but Hiern's description, translated from Welwitsch's notes written in Latin and attached to the existing duplicated specimens, state that the plant was flowering and fruiting at Loanda in July and August 1858. It thus appears, in the absence of other specimens, that no herbarium material was preserved (or has survived) until this date. The correct authorship of the valid (but illegitimate) name was therefore E. candelabrum Welw. ex Hiern, the holotype of the species being the specimen in the British Museum of Welwitsch 641, with isotypes elsewhere (COI, G, K, LISU). Despite this and the discrepancy in dates, the Lisbon specimen has been designated as a neotype of the invalid E. candelabrum Welw. by Leach (1981). Later, N. E. Brown (1912) reduced the name, with both authorships, to synonymy under his validly published legitimate name of E. conspicua, which has been universally accepted until Leach's publication of 1974.