Names of the Indigenous Neotropical Species of Luffa Mill. (Cucubitaceae)

Abstract

For many years, indigenous Luffa Mill. in the New World was generally considered to be represented by only one widespread species, e.g., by Cogniaux & Harms (1924) and Jeffrey (1984). However, Heiser et al. (1988) proposed that three species be recognized as indigenous in the New World. In particular, they segregated the Central American (western Mexico to Nicaragua) populations, for which they accepted the name L. quinquefida (Hook. & Arn.) Seem., as specifically distinct from the widespread South American species, for which they accepted the name L. operculata (L.) Cogn., previously used in a broad sense to refer to all neotropical indigenous Luffa. Jeffrey (1984) proposed a specimen in the Linnaean herbarium (LINN 1150/4) as the lectotype of the name Momordica operculata L., Syst. Nat. ed. 10: 1278 (1759) (and therefore of Luffa operculata (L.) Cogn., of which it is the basionym). In this choice he was followed by Heiser et al. (1988), but both overlooked the fact that a previous choice of lectotype had been made. Wijnands (1983) lectotypified the names on the plate of Commelin, P1l. Rar. & Exot.: 22, t. 22 (1706), reproduced as plate 53 in his work, and cited by Linnaeus in his protologue. In justification, Wijnands stated that while the Linnaean diagnosis referred to the fruits, these were not exhibited by the specimens in the Linnaean herbarium, which he cited as 1150/3 (sic) and 1150/4, but were depicted in the Commelin plate. In fact, the Linnaean specimens of the species in question are 1150/4 and 1150/5; the latter bears a fruit and the former, annotated "HU" (Hortus Upsaliensis) by Linnaeus, bears on its reverse an extended description of the fruit in Linnaeus' hand. It is clear that 1150/4 was taken from a cultivated specimen that had fruited in the Hortus Upsaliensis. However, a detailed description of the fruit was first published by Linnaeus only in Sp. Pl. ed. 2: 1433 (1763) and there is no evidence that this material was available to Linnaeus before the valid publication of his name in 1759. In the protologue, the Commelin plate is the sole cited element. There are therefore no grounds for rejecting Wijnand's earlier (1983) choice of lectotype in favour of Jeffrey's (1984) later one. There is no serious conflict with the protologue (Art. 8.1(b)), indeed no conflict at all, and there are no