A MONOGRAPHIC REVISION OF RETROPHYLLUM (PODOCARPACEAE)

Abstract

The living species of the genus Retrophyllum C.N. Page (Podocarpaceae) are revised. A key to the six species recognised is given. Retrophyllum filicifolium (N.E.Gray) R.R.Mill comb. nov. from New Guinea and the Moluccas is recognised as a species distinct from R. vitiense (Seemann) C.N. Page, which is restricted to material from Fiji and the Solomon Islands. Additional characters are given by which to separate Retrophyllum piresii (Silba) C.N. Page (Brazil) from R. rospigliosii (Pilg.) C.N. Page (Andes). The six species fall into three species-pairs that differ in reproductive characters: one pair in Papuasia and Melanesia, a second on New Caledonia and a third in South America. In the New Caledonian species-pair, the adult leaves are flattened in four ranks with only minimal heterofacial twisting, whereas in the species-pair inhabiting Fiji, New Guinea and neighbouring areas, as well as the pair inhabiting South America, the adult leaves are arranged in two ranks with significant heterofacial twisting. The names Podocarpus filicifolius N.E.Gray and Nageia minor Carrière, respectively the basionyms of Retrophyllum filicifolium and R. minus, are lectotypified, and the typification of Nageia minor comprehensively discussed in an appendix. Adult female epitypes are additionally designated for Podocarpus filicifolius, which was originally based on juvenile foliage of Retrophyllum mixed with a detached seed of Nageia wallichiana, and for Nageia minor, which Carrière described on the basis of sterile material. Two other appendices provide a list of accepted names and synonyms, and a list of exsiccatae. Illustrations and distribution maps are provided for each species.

Introduction

The genus Retrophyllum C.N. Page was erected by Page (1989 [‘1988’]) for a small group of species in the family Podocarpaceae. This group has been recognised as distinct ever since Bertrand (1874) used the name Podocarpus sect. Polypodiopsis C.E.Bertr. for two species that were then known as Podocarpus vitiensis Seem. and P. minor (Carrière) Parl. A similar concept was followed by Wasscher (1941), Buchholz & Gray (1948) and Gray (1962), among others. This sectional name appeared to be based on an earlier name, Polypodiopsis Carrière, but that name, and the name Polypodiopsis muelleri Carrière, were proposed for rejection by Mill & Weston (2004) on the grounds that the names are ambiguous and could refer to a species of a genus of Proteaceae, such as Beauprea Brongn. & Gris. These two nomenclatural proposals were subsequently approved (Barrie, 2006). By the time of Gray (1962), Podocarpus sect. Polypodiopsis comprised six named species, P. rospigliosii Pilg. having been described by Pilger (1923), P. comptonii J.Buchholz and P. palustris J.Buchholz by Buchholz (1949) and P. filicifolius N.E.Gray by Gray (1962). Shortly afterwards, de Laubenfels (1968) published a short note pointing out that the material Gray (1962) had used to describe her new species Podocarpus filicifolius was a mixture of juvenile foliage of what he regarded as Podocarpus vitiensis Seem. and detached fruits of what he called P. blumei Endl. [= Nageia wallichiana (C.Presl) Kuntze], and therefore he regarded Podocarpus filicifolius as a synonym of P. vitiensis [= Retrophyllum vitiense].

More recently, authors have realised that Podocarpus L'Hér. ex Pers. as conceived by Buchholz & Gray (1948) and earlier workers is heterogeneous, and have split it into numerous smaller genera. De Laubenfels (1969) erected the genus Decussocarpus de Laub., with three sections: Decussocarpus sect. Decussocarpus corresponded to Podocarpus sect. Polypodiopsis, D. sect. Dammaroides (Bennett) de Laub. to P. sect. Nageia (Gaertn.) Endl. and D. sect. Afrocarpus (J. Buchholz & N.E.Gray) de Laub. to P. sect. Afrocarpus J.Buchholz. Decussocarpus sect. Decussocarpus had four species: D. vitiensis (Seem.) de Laub., D. minor (Carrière) de Laub. and D. comptonii (J.Buchholz) de Laub. in the South Pacific, and D. rospigliosii (Pilg.) de Laub. in tropical South America. To these was subsequently added Decussocarpus piresii Silba, a close ally of D. rospigliosii that was rather unsatisfactorily described from Brazil (Silba, 1983). Later, on account of a change in the rules of the International Code of Botanical Nomenclature, the generic name Nageia Gaertn. had to be adopted for Decussocarpus, and de Laubenfels (1987) then referred to the three sections as Nageia sect. Nageia (= Decussocarpus sect. Dammaroides), N. sect. Polypodiopsis (C.E.Bertr.) de Laub. (= D. sect. Decussocarpus) and N. sect. Afrocarpus (J.Buchholz & N.E.Gray) de Laub. (= D. sect. Afrocarpus). De Laubenfels (1987) accepted five species in Nageia sect. Polypodiopsis, namely Nageia vitiensis (Seem.) O. Kuntze, N. comptonii (J.Buchholz) de Laub., N. minor Carr., N. rospigliosii (Pilg.) de Laub. and N. piresii (Silba) de Laub. In both that paper and his earlier one, he regarded Podocarpus filicifolius as a synonym of Nageia/Decussocarpus vitiensis and Podocarpus palustris as a synonym of N./D. minor.

Page (1989) decided that the three sections of Nageia as delimited by de Laubenfels (1987) were each worthy of independent generic status. For Nageia sect. Polypodiopsis, he had to propose the name Retrophyllum C.N. Page, because Decussocarpus de Laub. had been declared illegitimate and was (and still is) unavailable for use. Retrophyllum as delimited by Page (1989) had five species, corresponding to those recognised under Nageia by de Laubenfels (1987): Retrophyllum vitiensis (Seem.) C.N. Page on Fiji (the New Guinea etc. distribution was not accounted for), R. comptonii (J.Buchholz) C.N. Page and R. minus (Carrière) C.N. Page (by Page mis-spelled ‘minor’) on New Caledonia, and ‘at least two’ – R. rospigliosii (Pilg.) C.N. Page and R. piresii (Silba) C.N. Page – in tropical South America. Once again, Page (1989) included Podocarpus palustris under Retrophyllum minus, but unlike de Laubenfels (1969, 1987) he did not account for Podocarpus filicifolius nor cite Gray's 1962 paper in his bibliography. The possibility that there might be more than two species of Retrophyllum in South America was repeated in Page's later treatment of Podocarpaceae for The Families and Genera of Vascular Plants (Page, 1990), but the reason for this opinion was unexplained in both works.

As well as these five extant taxa, two Australasian fossils have been assigned to Retrophyllum as well as an undescribed fossil species from Argentina that is known from leafy branches bearing attached pollen cones (Wilf, 2012; Wilf et al., 2012, 2014; Merkhofer et al., 2015). A fourth fossil species may also belong to Retrophyllum, although the necessary combination has not yet been made. The Miocene fossil Retrophyllum vulcanense M.Pole is known from St. Bathans and Mata Creek in Otago, New Zealand (Pole, 1992, 1997; Hill & Brodribb, 1999; Pole, 2007), while R. australe R.S.Hill & Merrifield from West Dale in Western Australia is slightly older, of Middle Eocene or Oligocene age (Hill & Merrifield, 1993; Hill & Brodribb, 1999). Both these appear to be unequivocal members of Retrophyllum. However, on account of some morphological differences, Decussocarpus maslinensis D.T.Blackburn from South Australia (Blackburn, 1981) was later transferred to the extinct genus Willungia R.S.Hill & M.Pole as W. maslinensis (D.T.Blackburn) R.S.Hill & M.Pole ex R.R.Mill & R.S.Hill (Hill & Pole, 1992; Mill & Hill, 2004). Rather more contentious is Decussocarpus araucoensis (E.W.Berry) D.R.Greenw., an Eocene fossil from Coronel, Chile (Berry, 1922; Greenwood, 1987). This may well also belong to Retrophyllum but, unlike the other fossil species, it has not yet been compared with similar, Retrophyllum-like fossils that were described by Hill & Pole (1992) as the new fossil-genera Smithtonia R.S.Hill & M.Pole and Willungia. Until that is done, assignment of ‘Decussocarpus araucoensis’ to Retrophyllum should not be made, even though its current name is nomenclaturally incorrect because of the illegitimacy of Decussocarpus and should not be used. ‘Decussocarpus araucoensis’ was first described as a species of Araucaria Juss. by Berry (1922) but was transferred to Podocarpus by Florin (1940), who commented that it was near to P. rospigliosii (i.e. Retrophyllum rospigliosii). Hill & Pole (1992), based only on an examination of albeit excellent quality photographs in Florin's 1940 paper, considered that “probably some, but not all, of the specimens illustrated by Florin belong to this genus” [Retrophyllum]. More recently, Hill & Brodribb (1999: 667) repeated this assertion but were more definite in a footnote to a table in the same paper (Hill & Brodribb, 1999: 662), in which they wrote, “Hill & Pole (1992) considered this species to belong to Retrophyllum, but they did not formally transfer it to that genus”.

In recent molecular phylogenetic studies, Retrophyllum in its traditional sense appears close to Afrocarpus and Nageia, the three genera forming a monophyletic subclade that is sister to Podocarpus (Conran et al., 2000; Sinclair et al., 2002; Biffin et al., 2011, 2012; Knopf et al., 2012). Only the studies by Sinclair et al. (2002), Herbert et al. (2002) and Biffin et al. (2011, 2012) used species from all three informal groups A (South Pacific), B (New Caledonia) and C (South America) here recognised within Retrophyllum. Those by Sinclair et al. (2002) and Herbert et al. (2002) remain the most comprehensive in their sampling; Conran et al. (2000) sampled only the two species of group B, while Knopf et al. (2012) sampled only species from New Caledonia and South America (i.e. groups B and C) and Biffin et al. (2011, 2012) sampled only three of the six species treated here (one from each group). Herbert et al. (2002) sampled the same four species as Sinclair et al. (2002), i.e. excluding Retrophyllum piresii, for which DNA could not be obtained, and R. filicifolium (N.E.Gray) R.R.Mill, which at that time was not recognised as distinct from R. vitiense. Their results strongly supported the monophyly of Retrophyllum, despite the large disparities in morphology between the three morphological groups. For that reason, despite the large morphological differences, some of which appear to be quite fundamental, the genus is retained in its traditional circumscription in this paper. Herbert et al. (2002) additionally used random amplified polymorphic DNA (RAPD) and restriction fragment length polymorphism to assist in separation of the two New Caledonian species Retrophyllum comptonii and R. minus.

Retrophyllum has been placed controversially by Bobrov & Kostrikin (2000) and Melikian & Bobrov (2000) in Nageiaceae, in the order Cephalotaxales. Although a relationship with Nageia is suggested by other evidence (including molecular data: Conran et al., 2000; Sinclair et al., 2002; Biffin et al., 2011, 2012; Knopf et al., 2012), a removal from Podocarpaceae (even in a broad sense) is not, nor is its removal to a different order. Salter (2004) has pointed out that Bobrov & Melikian's removal of Retrophyllum to Cephalotaxales was based on a misinterpretation of the tissue layers of the mature seeds; they interpreted the seeds of Retrophyllum and Nageia as possessing two integuments, rather than a single integument and an epimatium.

Taxonomic Problems

Although a small group, the genus has been the subject of some debate. Four principal problems are addressed in this revision:

• • the status of the genus and whether its widely disjunct species form a monophyletic group.

• • species limits within Retrophyllum vitiense.

• • identification of the two New Caledonian species Retrophyllum minus and R. comptonii, which can be very difficult to separate in the herbarium.

• • whether Retrophyllum piresii is taxonomically distinguishable from R. rospigliosii, and if so at what rank.

Limits of the genus

Retrophyllum in the sense as originally defined by Page (1989, 1990) appears to be a morphologically unnatural genus that has been held together by the single character of heterofacially flattened shoots, which has been regarded as unique within Podocarpaceae. Although this is true for the extant genera of the family, two extinct genera with heterofacially flattened shoots (Smithtonia and Willungia) have since been described (Hill & Pole, 1992), demonstrating that the character has evolved more than once in the family in the past. In fact, the six living species recognised here fall into three natural groups, here denoted Groups A, B and C, that are geographically separated today, although in each of them the mechanism whereby the ‘retrophyllous’ leaf arrangement is achieved is similar. The Southwest Pacific species divide into those with distichous leaves in both adult and juvenile phases (Group A: Fiji, Papuasia) and those with 4-ranked leaves in the adult phase but distichous ones in the juvenile state (Group B: New Caledonia). The South American species-pair (Group C) also has distichous leaves in both adult non-reproductive and juvenile phases, but on reproductive shoots the leaves are not ‘retrophyllous’. In all three groups, some degree of internode torsion appears to be involved in bringing the leaves into one plane, just as it is in Amentotaxus Pilg. of Taxaceae (Tomlinson & Zacharias, 2001); this results in a characteristic zigzag pattern of decurrent leaf bases along the twigs that is particularly noticeable in twigs of the penultimate order and is especially pronounced in Group C. Each of the three Groups A, B and C possesses unique combinations of characters of male and female reproductive organs, the degree of importance of which is such that, on the basis of morphology alone, each group could probably be segregated at generic rank, as has been done for other small genera of Podocarpaceae such as Lepidothamnus Phil., Lagarostrobos Quinn and Manoao Molloy. Furthermore, in the two groups for which seedlings are known (Groups B and C), these are totally different in their morphology; seedlings of Group B have epigeal germination, about 10 ‘cotyledons’ arranged decussately on a naked axis with very short internodes between them, and decussate phyllotaxis of the first few foliage shoots, while Retrophyllum rospigliosii of Group C has apparently hypogeal germination with no visible cotyledons, and spiral phyllotaxis of the initial foliage shoots, which are separated by relatively long internodes along an axis clad with leaves. Nevertheless, the three groups always form a monophyletic clade in molecular phylogenetic studies (e.g. Conran et al., 2000; Sinclair et al., 2002). For that reason, the decision has been made to retain the present generic limits but to recognise three informal geographically defined Groups A, B and C within the genus, as outlined above and defined in detail below.

Group A species

Group A comprises the species occurring in the Southwest Pacific exclusive of New Caledonia. Most authorities have recognised only one species of Retrophyllum in this area, R. vitiense, with a range extending from Fiji through the Solomon Islands to New Guinea and the Moluccas (e.g. Eckenwalder, 2009; Farjon, 2010). Seemann (1862, 1863) provided the first descriptions of Retrophyllum vitiense (as Podocarpus vitiensis), based on material from Fiji collected by himself. Pilger (1903) also knew the species only from Fiji, citing Seemann's collection and one other. Wasscher (1941) was the first author to realise that a species of Podocarpus sect. Polypodiopsis [i.e. Retrophyllum] also grew in New Guinea and adjacent islands. Under what he regarded as Podocarpus vitiensis, he cited two specimens from the south-eastern part of New Guinea and one from New Ireland in the Bismarck Archipelago and gave a very full description (as is typical of his work) that was based on those three specimens and one female specimen from Fiji. Therefore, although he was the first to describe male material of what is here treated as Retrophyllum filicifolium, he did not see female material of the New Guinea plant. Since then, much more material of Retrophyllum has been collected from New Guinea, and it is now known from both the eastern half of the island (Papua New Guinea) and the western half (West Papua and Papua). Netta Gray, in her revision of the Pacific members of Podocarpus sect. Polypodiopsis [i.e. Retrophyllum], also regarded Podocarpus vitiensis (= Retrophyllum vitiense) as occurring in Fiji, the Solomon Islands and New Guinea, but she also described a new species, Podocarpus filicifolius N.E.Gray, from Morotai in the Moluccas, just west of New Guinea (Gray, 1962). This she separated from Podocarpus vitiensis on the basis of “the spreading scale leaves, thinner foliage leaves, the distinct receptacle 7 mm long supporting the seeds, and the spherical seeds” and used the last of those supposed differences (spherical seeds) to distinguish Podocarpus filicifolius from P. vitiensis. She also rather perceptively commented that the New Guinea specimens (all of them were by her included under Podocarpus vitiensis) “may still prove to represent a separate species”, but the immature ovules that she had seen were pear-shaped, as in Fijian Podocarpus vitiensis, and so she did not segregate them. De Laubenfels (1968) discovered that Gray's Podocarpus filicifolius was in fact based on a loose female cone of Nageia wallichiana (C.Presl) Kuntze and juvenile foliage of what he regarded in his paper as Podocarpus vitiensis. Consequently, most of the alleged differences that Gray had proposed between her new species and the Fijian Podocarpus vitiensis become irrelevant, and de Laubenfels (1968) therefore synonymised Gray's species with the latter. That remained his stance when in the following year he segregated the genus Decussocarpus from Podocarpus and regarded Podocarpus filicifolius as a synonym of D. vitiensis (de Laubenfels, 1969). Until now, no one has revisited Gray's comment that New Guinea material might be taxonomically separable from Fijian Retrophyllum vitiense, possibly because Gray's taxon from Morotai was founded on mixed material. Nevertheless, examination of ample material demonstrates that there are numerous other differences, not observed by either Gray or de Laubenfels, between the New Guinea plants that have been hitherto included in Retrophyllum vitiense and the Fijian ones, and that the material from mainland New Guinea agrees with that from Morotai (excluding the foreign cone material). Accordingly, Gray's name is here reinstated and applied in an emended sense to encompass all plants of Retrophyllum seen from New Guinea and the Moluccas. Because Gray's type material is deemed insufficient, consisting as it does of juvenile leaves belonging to Retrophyllum plus a cone that does not belong to the genus, an epitype is designated in this paper to fix the application of the name with regard to the characters of adult foliage and female cones. Gray's protologue of Podocarpus filicifolius is here emended to exclude all mention of female cone characteristics, specifically all the text in the Latin diagnosis after “strobilis masculis ignotis”, everything after “foliage leaves” in the paragraph beginning “This tree differs from Podocarpus vitiensis in. . .”, the last sentence of the paragraph on p. 71 beginning “The female strobili”, and the supposed difference in seed shape in couplet C of the key to species (p. 71). The description given here provides full details of the male and genuine female cones of Retrophyllum filicifolium – the female being here described for the first time although, as mentioned above, Wasscher (1941) had already described the male cones.

Group B species

Three taxa belonging to Retrophyllum have been described from New Caledonia, but until 1949 only one of them was known: Podocarpus minor Carrière, now known as Retrophyllum minus. Carrière (1867: 641) in his protologue said that this grew “au sommet de très-hautes montagnes, où il fut découvert par un jardinier anglais nommé Richard”, but the only actual specimen cited, Vieillard 1275, was from a low-lying locality (Lac Arnaud in the Plaine des Lacs) in the south of the island. From the very start, therefore, Retrophyllum minus seems to have been used to apply to all Retrophyllum of the island. The discoverer of the high-altitude plant mentioned in Carrière's protologue was most likely to have been N. Richards, who along with W. H. Duncan was a collector in New Caledonia for Ferdinand von Mueller (Morat, 2010). Richards collected Araucaria biramulata J.Buchholz “on open stony land of a very high mountain” in or before 1862 (Richards 3, P, barcoded P00190084, originally named Araucaria cookii R.Br.). However, no specimen of Retrophyllum collected by him in New Caledonia has been traced at P by me, and none was cited by Guillaumin (1911), who did, however, include three specimens of Araucaria collected by him. Nageia minor is therefore based purely on the single specimen cited, Vieillard's collection from low land in the south, as discussed in Appendix I of this paper. However, much of what is now recognised as the distinct species Retrophyllum comptonii was originally named Podocarpus minor in herbaria. Indeed, when Brongniart & Gris (1869) published the first description of the female cones of what they called Podocarpus minor, that description was based on two then newly collected specimens: Balansa 186 from Prony, which is Retrophyllum minus, and Balansa 1381 from woods near Téné and Bourail, which belongs to what we now know as R. comptonii. Brongniart & Gris's description of the albumen and embryo of Podocarpus minor could only have been based on the seed of Balansa 1381, and so this part of their description of P. minor applies only to Retrophyllum comptonii in today's taxonomy of the group. Seed colour, which varies from red to whitish-grey and has been used to separate taxa within the group (notably Podocarpus palustris), was not mentioned by Brongniart & Gris. However, an early collection of Retrophyllum minus by Pancher (s.n., labelled Vieillard 1275; see discussion, Appendix I) from Prony described the fruit as red, although only very young seed cones, not long past receptivity, are present on the sheet.

Buchholz (1949) described three new species of Podocarpus sensu lato from New Caledonia, the first two of which, P. comptonii J.Buchholz and P. palustris J.Buchholz, belong to Retrophyllum (the third, Podocarpus sylvestris J.Buchholz, does belong to Podocarpus as currently circumscribed: de Laubenfels, 1972). Both Podocarpus comptonii and P. palustris were distinguished by him from P. minor [Retrophyllum minus], for which, however, he provided no comparable description. Podocarpus comptonii was separated by its habit (a tall tree), its thinner adult leaves with midrib less than one quarter the width of the lamina, and by differences in female seed shape and wood colour. Podocarpus palustris was described as being amphibious and was distinguished from P. minor by: (1) the large caudices (trunks enlarged below: an adaptation to the amphibious habit); (2) the whitish wood of lower density (0.3); (3) the thicker, shorter, strongly reflexed seeds; (4) the smaller bracts on the female peduncle; (5) the smaller receptacle only 2 mm long; (6) the beak of what he called the ‘nut’ [i.e. the seed proper, only visible by dissection of a cone] shorter, more broadly conical and recurved; (7) the leaves broader than usual for P. minor relative to the length, with the margins less thick in the dried state, and deciduous on the branches of the current year, the individual leaves (called by Buchholz ‘leaflets’) 2–6 × 1–3 mm. The seeds of Podocarpus comptonii were stated to be red, while those of P. palustris were described as having a grey epimatium when ripe, thereby distinguishing it from the red epimatium of P. minor (and P. comptonii as described in its protologue). Buchholz (1949) noted that both species sometimes grew at the same locality (e.g. 22 km Station) while maintaining their character differences.

The description of these two new species by Buchholz (1949) brought a deal of confusion into the taxonomy of New Caledonian Retrophyllum that is still not satisfactorily resolved. There are two separate problems: the distinction between Retrophyllum minus and R. comptonii, and whether or not Podocarpus palustris is a distinct taxon.

Over most of the island, there is no problem over species separation of Retrophyllum comptonii and R. minus, because R. comptonii is the only one present, but in the south both species can occur sympatrically, along with the taxon Podocarpus palustris if that is regarded as distinct from R. minus. Of the various differential characters cited by Buchholz (1949) to distinguish Retrophyllum comptonii and R. minus, the width of the midrib (or ‘central band’ of de Laubenfels, 1969) is the one that has more often been used (e.g. de Laubenfels, 1969, 1972). Herbert et al. (2002: 176), however, commented that this character was “exceedingly difficult to interpret” and that it was therefore “a poor arbiter where there is uncertainty over identifications” but provided no alternative morphological means of separating the two species. They did, however, find that the use of RAPD provided reliable molecular markers that could be employed to separate them. Gray (1962) had separated them on the basis of habit (large trees in Retrophyllum comptonii, small trees or shrubs in R. minus), foliage mostly flattened (R. comptonii) versus rarely flattened (R. minus) and twigs pinnately leaved (R. comptonii) versus foliage mostly decussate (R. minus). Farjon (2010) attempted to get round the problem by keying out Retrophyllum minus first on the basis of specimens of it being ‘shrubs or small, dwarfed trees’ as compared with the other four species, all of which he considered to be ‘potentially tall trees’, by the foliage branches of R. minus being ‘ascending or spreading’ versus ‘spreading or more or less pendulous’ in R. comptonii and the other species, and by the leaf apex ‘always obtuse’ versus ‘acute, apiculate or some obtuse’. The habit/stature characters used by both these authors are difficult or impossible to interpret correctly from herbarium specimens unless these are accompanied by detailed notes on habit, which is rarely the case; additionally, although Retrophyllum minus is normally short in stature (0.9–3.5 m), it can occasionally reach a height of 8 m, thereby causing confusion with R. comptonii (2.5–30 m).

No one since Gray (1962) has tackled the issue of Podocarpus palustris, which later workers such as de Laubenfels (1969, 1972), Gaussen (1976) and Farjon (2010) have all regarded simply as a synonym of Retrophyllum minus. Before Buchholz's work, Bernier (in sched., P) had differentiated two informal taxa within Podocarpus minor in southern New Caledonia: ‘type sylvestre’ and ‘type lacustre’, corresponding to P. minor sensu stricto (? and also southern, lowland plants of what Buchholz recognised as P. comptonii) and what Buchholz subsequently described as P. palustris. Most of Bernier's collections are labelled ‘type lacustre’ (Bernier 245, 246, 247, 248, 249, 250, 251; s.n. 7 iii 1948); of those seen, only Bernier 269 and 270 from Forêt Walker and Bernier 271 from Montagne des Sources are labelled ‘type sylvestre’. All of these have been determined for this revision as Retrophyllum comptonii. Bernier 271 was collected at an altitude of 900–1000 m at Montagne des Sources and is obviously that species. Bernier 269 and 270, however, were collected at only 200 m at Forêt Walker. Buchholz & Gray determined three sheets of Podocarpus comptonii at P as being possibly a new form or variety of Podocarpus comptonii; these were Bernier 203 (Forêt du Mois de mai), Buchholz 1350 (Walker's place) and Buchholz 1697 (Walker's place).

It is clear from determinavit slips (e.g. on Virot 658 at A, in 1953) that Gray originally intended to retain Podocarpus palustris as distinct from P. minor. However, that was not the case in her published revision (Gray, 1962). In that work, Gray justified her synonymisation of Podocarpus palustris with P. minor by saying that “Dr Buchholz (ms. data) believed there to be a related species Podocarpus palustris which he described, but the differences he listed do not fall outside the normal range of variation in P. minor, and the difference in wood density is no more than expected from the slightly different ecological habitats”. Other apparent differences between Podocarpus minor and P. palustris can be found when herbarium material is examined. However, insufficient is known about the variation or constancy of these differences, and in this revision Podocarpus palustris is therefore regarded as a synonym of Retrophyllum minus.

Group C species

Until 1983, only one species of Retrophyllum was considered to grow in South America. This was Retrophyllum rospigliosii, first described as Podocarpus rospigliosii by Pilger (1923) based on a collection from Peru bearing very young male cones. Female cones were apparently unknown until they were described, from Venezuelan material, by Schnee (1944). The species is now known to be widespread along the northern Andes chain, occurring in Bolivia, Peru, Colombia, Ecuador and Venezuela, where it reaches its northernmost limit in the state of Trujillo, whence it has been recently collected (Dorr et al., 2000; Hokche et al., 2008). Paradoxically, it is now the case that female cones have been much more often collected than mature male ones, and there is some variation in female cone shape, which might have led Page (1989, 1990) to suggest that there could be more than two species of Retrophyllum in South America. However, Retrophyllum rospigliosii is not divided here. Regional treatments of the species have been provided by Schnee (1944), Buchholz & Gray (1957) and de Laubenfels (1982) for Venezuela, Torres-Romero (1988) for Colombia and de Laubenfels (1994) for Peru.

The second South American species of Retrophyllum is R. piresii, first described as Decussocarpus piresii by Silba (1983). This is apparently a narrow endemic to the Serra Pacaás Novos in Rondónia, Brazil, and remains very poorly known from the type (a female specimen with ripe cones, represented in several herbaria) and two other wild-collected specimens as well as one of cultivated material. No male material appears to have been collected so far. As noted in the species account, there have been records of Retrophyllum piresii from similar habitats in Bolivia and Peru, but these cannot be confirmed at present. There is little published taxonomic information about the species apart from the protologue, a short treatment by Secco et al. (2009) and a more detailed one by Farjon (2010), who said that it should be provisionally kept distinct from Retrophyllum rospigliosii, a decision accepted here for reasons given later.

Graham (2008) documented the discovery of an unidentified Retrophyllum species from humid lowland swamp forest at Madre Selva Biological Station near Iquitos, Peru. Two male trees (nicknamed ‘Lonely George’ and ‘Lonely John’: Graham, 2012) were found but at that time no females, and so the field botanists could not determine whether it was a new species, although it appeared to be different from Retrophyllum rospigliosii. The photograph published in Graham (2008) shows a tree with leaves similar to those of Retrophyllum piresii in their elliptic outline. The pollen cones are borne very profusely in the leaf axils; they are brown and about twice the length of the leaves. One female tree has since been found, as well as at least 20 seedlings and other mature trees, all over 100 ft [c.30 m] tall (Graham, 2012). However, no herbarium material of this enigmatic taxon has been seen, and it is not known whether any exists. Cones from the single known female tree need to be collected to determine whether this taxon is conspecific with Brazilian Retrophyllum piresii or represents a third South American species of the genus.

In the Paris herbarium, there is also an undetermined specimen (10 xii 1875, Mosén 4393) from Sao Paulo, Brazil, labelled simply ‘Podocarpus’. Superficially it looks rather like a juvenile Retrophyllum, but closer examination reveals that it has net-veined, angiospermous leaves that have apiculate tips. Lindman (1898: 25) identified the Stockholm duplicate of this number as sterile material of Holocalyx balansae Micheli (Fabaceae); for a recent revision of that genus see Mansano & Vianno Filho (2010), although it does not cite the Mosén material.

Materials and Methods

This revision is based on examination of selected specimens from the herbaria (using standard codes according to Thiers, continuously updated) A, B (types), BM, E, GH, K and NY, as well as specimen images in the following online repositories and databases: COL (Instituto de Ciencias Naturales, Universidad Nacional de Colombia, 2004 and continuously updated, accessed 8 February 2013; Herbario Nacional Colombiano, no date, accessed 5 February 2015 etc.), JSTOR Global Plants (no date, accessed 3 September 2014 etc.: some types), L, MG (one isotype), MO (types), P, PNGplants (Conn et al., 2004 onwards, accessed 18 December 2014), R (one isotype), RB (JBRJ, no date, accessed 15 April 2014: one isotype), S (Krypto-S, no date, accessed July 2014 etc.: some types) and UDBC. Additional material at COL and UDBC was photographed for this research. The databases of COL, K (Kew, no date, accessed 3 September 2014 etc.), MEDEL, MO (TROPICOS), NOU (Herbier du centre IRD de Nouméa, no date, accessed 21 and 22 July 2014), S (Krypto-S, no date), WU (Vienna Herbaria) and Z (Zürich Herbaria, no date, accessed July 2014) were also consulted (that of MEDEL can no longer be found on the Internet so is not cited). Database records (indicated in specimen citations by the suffix ‘–database’) have been included only where there is no doubt as to correctness of identification, particularly if no duplicates of the specimen concerned have been examined either physically or as images. In specimen citations, the suffix ‘–photo’ denotes a photograph of a specimen, either mounted on a herbarium sheet (as in the case of much of Buchholz's material) or taken by a colleague, while the suffix ‘–image’ indicates a digital image in an online repository.

Herbarium specimens were examined using a stereo dissecting microscope and Leitz ×6 or ×8 loupes and photographed using a Panasonic Lumix TZ30 digital camera. Measurements were all made on dried material; when fresh, some organs can be larger, especially female cones, but in Retrophyllum there is not the same degree of lengthening of pollen cones with age as there is in genera such as Podocarpus. Some measurements were made on calibrated images using ImageJ version 1.49d (Rasband, 2014); calibration of Paris database images lacking scales was performed using the barcode label, measuring 50 × 20 mm, as a datum basis. Line drawings are provided for taxa that have not been previously illustrated or for which the known published illustrations are in some way unsatisfactory. Microsporophylls and microsporangia were, when possible, rehydrated for drawing by the artist. Where a species has been previously satisfactorily illustrated, an iconography is given listing known published illustrations of good quality and correctly identified.

References to Articles of the International Code of Nomenclature refer, unless stated otherwise, to the current Melbourne version (McNeill et al., 2012).

For each species, as part of the distribution summary, the Level 3 and 4 country names and codes of the Taxonomic Databases Working Group (TDWG) are listed according to Brummitt (2001). The names of World Wide Fund for Nature Bioregions and Ecoregions follow the schemes published for Latin America and the Caribbean by Dinerstein et al. (1995) and for the Indo-Pacific by Wikramanayake et al. (2002). Assignments to ecoregions were done using the Worldwide Fund for Nature's ecoregion map available at Wildfinder (no date).

In specimen citations, country names are in Small Capitals, first-order subdivisions of countries (divisions in Fiji; provinces in Solomon Islands, Indonesia, Papua New Guinea, New Caledonia and Ecuador; states in Brazil and Venezuela; departments in Bolivia and Colombia; and regions [formerly provinces, until 2002] in Peru) in Bold and second-order subdivisions of countries (provinces in Fiji, Papua New Guinea, Bolivia and Peru; regencies in Indonesia; districts in Papua New Guinea; communes in New Caledonia; municipalities in Brazil, Colombia and Venezuela; and cantons in Ecuador) in Bold Italic. Solomon Islands has no second-order administrative divisions. Some countries also have third-order divisions (e.g. districts in Peru and municipalities in Bolivia); where known, these are indicated in italic. Assignment to communes in New Caledonia follows the database of NOU (Herbier du Centre IRD de Nouméa, no date). Some peaks in New Caledonia are at the boundary between two or more communes (e.g. Mt. Dzumac: Dumbéa and Païta), so some records are assigned to one commune and some to another. For Fiji, it has seemed helpful also to indicate the main islands in Bold Small Capitals, because the first-order administrative divisions sometimes divide islands or are divided between islands.

The paper includes a list of accepted names and synonyms (Appendix II) and a list of exsiccatae (Appendix III). Collectors’ initials are given in Appendix III but not in the specimen citations unless omitting them would create ambiguity (e.g. A.C. Smith, D. Smith and D.N. Smith). The New Caledonian collector H.S. MacKee spelled his surname in earlier collections as McKee; all specimen listings have been standardised as the more frequent MacKee, following Morat (2010), although his wife Margaret styled herself McKee (cf. McKee, 1972, cited elsewhere in this paper). In Appendix II, specimens collected by one person but numbered in another person's series are listed twice for cross-referencing, under both the name of the actual collector and that of the person whose number sequence was used. This situation is most common with MacKee, in whose collection number series are specimens collected by numerous other individuals, but is also found in other numbering series, such as that of Vieillard, which includes material collected by Pancher and probably also Deplanche.

Maps are provided for all species. These were generated in ArcGis version 9.3 (ESRI, 2008) after geospatial coordinates had been added for specimens lacking such information. The main sources of geographical coordinates were GeoNames (no date) and Fuzzy Gazetteer (no date), as well as, for New Caledonia, the gazetteer coordinates in many specimen records in the database VIROT (Herbier du Centre IRD de Nouméa, no date; cited in square brackets after the locality). Phytogeography is indicated according to the scheme of Takhtajan (1986).

Version 3.1 (second edition) of the categories and criteria of the International Union for the Conservation of Nature (IUCN, 2012) and version 10.1 of the guidelines for applying these (IUCN Standards and Petitions Subcommittee, 2013) were used to make the provisional conservation assessments for certain taxa treated in this paper.

As part of the process for the lectotypification of the name Nageia minor Carrière, basionym of Retrophyllum minus (Appendix I), colour profiling was performed on the digital images of certain collections of R. minus at P using the image-processing and -analysis program ImageJ (Rasband, 2014). These included all those numbered Vieillard 1275 (the ‘type material’), as well as collections by Pancher and Deplanche (Appendix table 1). The original images were assumed to have been made using identical exposures, and no alterations were made to the original files. The colour profiles of portions of leaf measuring approximately 20 × 20 pixels (px, equivalent to c.1.7 × 1.7 mm on all the Paris specimens analysed, after calibration at 11.7 px/mm), as well as of a transverse line across the middle of that leaf, were generated using the ImageJ tool Analyze – Tools – Colour Profiler, which determines the proportions of red (R), green (G) and blue (B) in the image and provides a plot of intensity for each of the channels as well as tabulating the results (Ross, 2007). Profiles were generated for several leaves on each specimen on each sheet; usually one profile was generated per leaf, but on very small fragments with few leaves, or single leaves, separate profiles were generated for areas c.20 × 20 px at the base and apex of the leaf. Areas of obvious discoloration or damage of the leaf surface were avoided whenever possible. The profile values given in Appendix table 1 relate to 20 × 20 px areas; those from the transverse lines showed slightly more variation, related to the often greater changes in topography of the dried leaf across the line. The profile values and corresponding plots allowed the accurate assignment of specimen fragments to the different elements that make up the mixed gatherings numbered Vieillard 1275, as well as ones by Deplanche and Pancher.

Taxonomy

Retrophyllum C.N. Page, Notes Roy. Bot. Gard. Edinburgh 45(2): 379 (22 Feb. 1989 [‘1988’]). – Podocarpus L'Hér. ex Pers. sect. Polypodiopsis C.E.Bertr., Ann. Sci. Nat., Bot. Sér. 5, 20: 65 (1874; see comments). – Nageia Gaertn. sect. Polypodiopsis (C.E.Bertr.) de Laub., Blumea 32: 210 (3 Feb. 1987). – [Decussocarpus de Laub., J. Arnold Arbor. 50: 340 (1969) sect. Decussocarpus (op. cit. p. 341), now a nom. illegit. (Art. 52) although legitimate when published.] – Type: Retrophyllum vitiense (Seem.) C.N. Page, based on Podocarpus vitiensis Seem.; the latter name is also the type of Decussocarpus de Laub. and of Podocarpus sect. Polypodiopsis C.E.Bertr. (non Polypodiopsis Carrière).

?[Polypodiopsis Carrière, Traité Gén. Conif. ed. 2: 710 (1867), nom. utique rej. (see comments)]. – Type: Polypodiopsis muelleri Carrière, loc. cit. (type specimen not indicated).

Nomenclatural note

Polypodiopsis Carrière, typified by P. muelleri Carrière from New Caledonia, may be synonymous with Retrophyllum, or it could be a member of the Proteaceae, such as Beauprea balansae Brongn. & Gris, as was suggested by Hutchinson (1920). Polypodiopsis Carrière, an earlier homonym of the fern genus Polypodiopsis E.B.Copel., is now a nomen rejiciendum and cannot be used at genus rank (Mill & Weston, 2004; Brummitt, 2006 [‘2005’]; Barrie, 2006). However, when Bertrand (1874) apparently reduced Carrière's name to Podocarpus sect. Polypodiopsis, he designated a different name, Podocarpus vitiensis (now Retrophyllum vitiense), as type of the section, the name of which should therefore be attributed to Bertrand, not Carrière.

Etymology

Retrophyllum is derived from Latin, retro (‘backwards’), and Greek, phyllon, Latinised as phyllum (‘leaf’), on account of the orientation of the leaves, which are twisted so that along one side of the stem axis all the abaxial surfaces face uppermost, while along the other side of the axis all the adaxial surfaces are turned uppermost. The gender is neuter.

Trees or shrubs; dioecious. Trunk buttressed or not. Bark exfoliating in short vertical strips or small pieces (Pacific species: Groups A and B), or as large plates (South America: Group C). Vegetative shoots dimorphic with either both foliage and scale leaves or only scale leaves (Groups A and B) or monomorphic with foliage leaves only (Group C), heterofacially flattened at least in juvenile phase, in adult phase sometimes not fully so (New Caledonian species: Group B). Terminal buds developed on shoots, these either protected in some way (Groups A and B) or naked (Group C); bud scales present (Fiji and Malesia: Group A) or absent (Groups B and C). Leaves of juvenile phase always ± distichous, heterofacially turned at base so that all leaves along one side of the axis present their abaxial surface to the light and all those along the other side present the adaxial surface; those of adult phase either distichous and flattened in 2 ranks (Groups A and C) or wholly or partly in 4 ranks that show only minimal heterofacial turning (Group B). Adult leaves subsessile with bases decurrent and ± crossing over to a departure point on the opposite side and so creating a ± zigzag pattern along the branch (all groups, but particularly prominent in Group C); blades bifacially flattened, coriaceous, relatively thin (Groups A and C) or thick to very thick (Group B), with or without a visible midrib; resin canals 1 median below the vascular bundle (Groups A and B) or (1)3(5) below the vascular bundle plus 2–6 lateral bundles around the leaf margin (Group C); transfusion tissue present; accessory transfusion tissue absent. Pollen cones usually lateral but sometimes terminal in Retrophyllum vitiense, 1–8 together, potentially arranged in racemose inflorescences (Groups A and C) or groups (Group B) borne on a common peduncle; individual pollen cones ± pedicellate (Groups A and C) or sessile (Group B), cylindrical, ellipsoid or ovoid; microsporophylls imbricate, decussate (Group B: appearing spirally arranged) or spirally arranged (Groups A and C), triangular or ovate (Groups A and B) or lanceolate (Group C). Pollen 2-saccate. Female cones borne on current growth, as part of specialised reproductive shoots (Groups A and C) or terminating foliage shoots (Group B); cones sylleptic (Groups A and C) or proleptic (Group B), pedunculate; peduncle broadened distally (Group C) or not (Groups A and B), in the vegetative portion bearing adpressed scales (Group A) or spreading bracts (Groups B and C); carpophore absent; cone axis with either 1 or 2 (Group B) or 6–8 (Groups A and C) sterile bracts below the fertile bract; fertile bract free from the epimatium (Groups A and B) or with its basal half firmly adnate to it (Group C), distally navicular (Group C) or flat (Groups A and B). Prophylls absent. Receptacle vestigial, formed from the fertile bract and sometimes the last sterile bract (Groups A and C) or from 1 or 2 peduncle bracts plus the last sterile bract and fertile bract (Group B), cylindrical (Groups A and C) or obovoid (Group B). Ovule inverted. Epimatium present. Seed distally crested or not, with a beak at the micropylar end (Groups A and B) or lacking a beak (Group C). Germination epigeal (Group B) or hypogeal (Group C). Seedling phyllotaxis decussate (Group B), spiral (Group C) or currently unknown (Group A).

Present diversity and distribution

Six living species in three very distinct morphological lineages occurring respectively in: (A) Fiji, Solomons and eastern Malesia west to the island of Morotai in the Moluccas; (B) New Caledonia; and (C) tropical South America (eastern slopes of the Andes and extending into the Amazon Basin in Brazil).

Past diversity and distribution

Two (or more?) fossil species have also been included in the genus as traditionally circumscribed: †Australasia (New Zealand, Western Australia and Tasmania [undescribed material]) and southern South America (Argentina and ?Chile). There is apparently no record of the genus having occurred in Antarctica, which might be expected given the disjunct distribution of the living and known fossil species between the Australasia/Southwest Pacific area and South America.

Chromosome number and genomics

Diploid chromosome number: 2n = 20 (3 species cytologically known, from Groups A and B only). The karyotype of all known species comprises 10 pairs of metacentric chromosomes. The genome size is known only for Retrophyllum rospigliosii (of which the chromosome number is not yet known); at 12 pg, it is small for conifers but average within the Podocarpaceae, most species of which have small genome sizes of 8–18 pg, with only Falcatifolium (22 pg in F. taxoides) and Manoao colensoi (28 pg) being larger within the family (Zonneveld, 2012). Vieira et al. (2016) have very recently published the complete plastome sequence of Retrophyllum piresii, based on material cultivated at the Museu Paraense Emílio Goeldi in Belém; I have seen herbarium material of this plant as well as photographs kindly sent by Chad Husby, and have verified the identification.

Key to species

1. 1a. Vegetative shoots monomorphic, scale leaves absent; buds naked; female reproductive shoots subtended by a bract or by a modified (reduced) foliage leaf; female cones with the base of the fertile scale firmly adnate to the epimatium and with a navicular tip; bark exfoliating in large plates (Group C: South America) 2

2. 1b. Vegetative shoots dimorphic, scale leaves present; buds protected, either by true bud scales or by modified leaves; female reproductive shoots subtended by a normal foliage leaf; female cones with the fertile scale free from the epimatium and not navicular; bark exfoliating in short vertical strips or small pieces (Malesia, Papuasia and New Caledonia) 3

3. 2a. Peduncle of female cone bearing reduced leaves; foliage leaves lanceolate, acute to acuminate and frequently abruptly narrowed just below apex; adult shoots ± alternate (widespread from Venezuela to Peru) 5. R. rospigliosii

4. 2b. Peduncle of female cone bearing small bracts; foliage leaves elliptic, obtuse, never abruptly narrowed below the tip; adult shoots opposite (one definite locality in Brazil) 6. R. piresii

5. 3a. Female cones borne on specialised reproductive shoots that are either lateral or terminal; ultimate adult foliage shoots pendulous with leaves in 2 ranks; adult leaves with an evident midvein; buds protected by true bud scales (Group A: Malesia, Fiji, Solomons) 4

6. 3b. Female cones terminal at ends of ordinary foliage shoots; ultimate adult foliage shoots erect or suberect with at least some leaves in 4 ranks; adult leaves with no visible midvein; buds protected by very reduced leaves, lacking true bud scales (Group B: New Caledonia) 5

7. 4a. Leaves broadest near base; microsporophylls 44–80 in 11–20 whorls; female cone axis with 6–8 sterile scales that persist at least until the resting phase (Fiji and Solomons) 1. R. vitiense

8. 4b. Leaves broadest near middle; microsporophylls 20–32 in 5–8 whorls; female cone axis with 4 sterile scales that are very quickly caducous, only present during the youngest stages (New Guinea, Moluccas) 2. R. filicifolium

9. 5a. Adult leaves with the raised central area narrower than the marginal area; some adult leaves on ultimate shoots in 2 ranks, those towards shoot tips normally in 4 ranks; habit variable from shrub to tall tree; not aquatic 3. R. comptonii

10. 5b. Adult leaves with the raised central area broader than or at least equalling the marginal area; all adult leaves on ultimate shoots in 4 ranks; aquatic shrub or ‘treelet’ with trunk rapidly tapering from a wide base 4. R. minus

Species descriptions

Group A. Bark exfoliating in short strips or small pieces. Vegetative adult shoots dimorphic. Terminal buds protected by bud scales. Leaves heterofacially flattened in both juvenile and adult phases. Adult leaves distichous in 2 ranks, relatively thin. Resin canal 1, positioned below vascular bundle. Pollen cones in racemose inflorescences subtended by scale-like bracts (or sometimes terminal and apparently solitary, as in the type of Retrophyllum vitiense), the raceme branches if present alternate or spiral, individual cones pedicellate, subtended by smaller scale-like bracts. Female cones on specialised lateral reproductive shoots, sylleptic; peduncle not broadened distally, in its vegetative portion bearing subadpressed scales. Cone axis with 4–8 sterile bracts. Fertile bract free from epimatium, flat. Receptacle vestigial, formed from the fertile bract and distal part of peduncle which may be slightly swollen, cylindrical or obovoid. Seed beaked at micropylar end. Details of germination and seedlings not known.

1. Retrophyllum vitiense (Seem.) C.N. Page, Notes Roy. Bot. Gard. Edinburgh 45(2): 380 publ. (22 Feb. 1989 [‘1988’]). – Podocarpus vitiensis Seem., Bonplandia 10: 366 (25 Dec. 1862) & J. Bot. 1: 33, t. 2 (1863). – Nageia vitiensis (Seem.) Kuntze, Revis. Gen. Pl. 2: 800 (1891). – Decussocarpus vitiensis (Seem.) de Laub., J. Arnold Arbor. 50: 342 (1969). – Lectotype (Farjon, 2010: 942 & 943): Fiji, Seemann 576, sheet 2/2 (lecto K, isolectos BM [2 sheets], E, GH–image, S–images). Fig. 1D.

Fig. 1. Retrophyllum species, comparative drawings. A–C, Leaf arrangements × 2. A, Retrophyllum filicifolium (Group A) (Woods 345). B, R. comptonii (Group B) (Gardner et al. TNCA 5045). C, R. rospigliosii (Group C) (Wright 247). D–F, Terminal bud or equivalent, × 15. D, Retrophyllum vitiense (Group A) (Smith 7076). E, R. comptonii (Group B) (Gardner et al. TNCA 2036). F, R. rospigliosii (Group C) (García-Barriga 27584). G–J, Female cones, × 1.5. G, Retrophyllum filicifolium (Group A) (Woods 345). H, R. comptonii (Group B) (Gardner et al. 256). J, R. rospigliosii (Group C) (Bunting 4939). K–M, Male cones × 1.5. K, Retrophyllum filicifolium (Group A) (Carr 14160). L, R. minus (Group B) (McMillan 5139). M, R. rospigliosii (Group C) (R.T. Pennington 1433). Scale bars: A–C, 1 cm; D–F, 1 mm; G, H, J–M, 2 cm. Drawn by Claire Banks.

?[Torreya bogotensis Linden ex K.Koch, Wochenschr. Vereines Beförd. Gartenbaues Königl. Preuss. Staaten 14(25): 199 (1871), nom. utique rejic.: Mill, 2010; Brummitt, 2011].

Note on typification

Gray (1962: 73) indicated that the Kew example of Seemann 576 was the ‘holotype’ of Podocarpus vitiensis (basionym of Retrophyllum vitiense), with an isotype (not so indicated) at GH. This was incorrect because, as was noted by Smith (1979), Seemann (1862: 366; 1863: 33) cited two syntypes in both the original protologue and in another publication the following year; the first of these was collected by W.G. Milne on the voyage of HMS Herald, and the second was Seemann 576. (Smith, 1979: 101 in fact cites two Milne collections of this species, both unlocalised, one without number and the other Milne 33). However, de Laubenfels (1969: 344) repeated the designation of Seemann 576 at K as ‘holotype’ of Podocarpus vitiensis, with specimens at A and BM indicated as ‘isotypes’. These specimens should be regarded as lectotype (K) and isolectotypes (A, BM, GH) respectively, following Smith (1979: 100) and the more precise lectotypification by Farjon (2010). The choice of a Kew specimen as lectotype by both the above authors is difficult to understand, because it is one of the two BM sheets, not either of the K sheets, that bears Seemann's drawings that were used in his two publications (1863, 1868). However, there is no controversy over the identity of any of these sheets, and Farjon's recent designation, which supersedes the earlier ones by de Laubenfels (1969: 342) and Smith (1979: 100), must therefore stand. Both latter authors independently chose Seemann 576 at K but without realising that there are two sheets of that number at K, one of them (annotated ‘1/2’, i.e. sheet 1 of 2) being a seedling and the other (‘2/2’) a branch from a young tree, and they could not have come from the same plant. Sheet ‘1/2’ at Kew is therefore not an isolectotype.

Iconography

Seemann, J. Bot. 1: t. 2 (1863).

Etymology

Derived from Viti, alternative name for Fiji.

Vernacular names

ailumu (Solomon Is., Kwara'ae language: Mauriasi et al. BSIP 17025 in sched., K, L); dakua salusalu (Fiji: Parham, 1964, 1972; Damanu R10 in sched., K), ndakua salusalu or salu-salu (Fiji: Mead, 1928; Smith, 1979); ngapiru [Vanikoro (Solomon Is.: Nambalua – Piaito BSIP 7061 in sched., L); ?kau solo (Fiji: Seemann, 1862, 1863, 1868); tau solo (Fiji: Seemann, 1868, possibly a misprint for kau solo) Seemann's use (1862 etc.) of the name kau solo in fact refers to Dacrycarpus imbricatus var. patulus, according to Smith (1979), who also said that the most frequently used name for Retrophyllum vitiense is ndakuua salusalu.

Distinguishing features

Retrophyllum vitiense is most closely allied to R. filicifolium (N.E.Gray) R.R.Mill (see below) from New Guinea and the Moluccas. Both have ± flat, distichous adult leaves, female cones borne laterally near the base of an ultimate foliage shoot, and dimorphic shoots bearing either foliage leaves or small decussate scale leaves. The dimorphic shoots separate both these species from the two South American species of Retrophyllum, which have monomorphic shoots. The flat, distichous, non-imbricate adult leaves and lateral female cones separate the two Papuasian species from the New Caledonian species Retrophyllum comptonii and R. minus, both of which have terminal female cones and adult leaves more or less in 4 ranks and ± imbricate (although juvenile and transitional shoots are distichous). Vegetative material of Retrophyllum vitiense is best separated from its nearest ally R. filicifolium by its tapered leaves widest close to the base, not near the middle. For differences in reproductive characters between these two species, see under Retrophyllum filicifolium.

Tree, 15–47 m, dbh 25–450 cm. Trunk sometimes spur-buttressed at base on large trees, bole straight and when adult unbranched except near top. Bark smooth, finally with faint vertical fissures, flaking into small longitudinal crumbly pieces, fissured, greyish-brown or (Solomons) dark brown, whiter in mature trees; inner bark yellowish-brown or (Solomons) reddish-brown; wood cream-coloured. Crown compact but much branched, pyramidal and finally rounded, with pendulous branchlets and twigs. Buds ovoid or obovoid; bud scales present, in 2 series, decussate, all equal, ovate and keeled, acute, brownish with narrow paler entire margin, initially persistent but later slowly caducous, not leaving scars. Ultimate juvenile shoots alternate, subopposite or opposite, 130–280 mm, with 23–72 (usually 35–60) pairs of foliage leaves, occasionally biramulate. Juvenile foliage leaves 4–11 mm apart, opposite-distichous, heterofacially turned at base, diverging at 60–80°, shortly petiolate, the petiole 0.2–0.5(–1.5) mm, lamina lanceolate or narrowly lanceolate, those near middle of branchlet 18–35 × (3.8–)4–6(–7) mm. Adult trees with dimorphic shoots, most shoots bearing foliage leaves but leader shoots bearing scale leaves; shoot flattening heterofacial with the decurrent leaf bases running straight down the internodes. Primary adult branches opposite-decussate or alternate. Penultimate adult branches erecto-patent with respect to subtending branch (although topographically deflexed-patent), straight or slightly curved with flexuous internodes, alternately bearing axillary decussate lateral foliage shoots and decussate ovate to orbicular bracts c.2 mm. Ultimate adult foliage shoots (70–)100–230 mm with (7–)11–60 pairs of leaves, straight or curved with slightly flexuous internodes. Leaf scars absent. Adult foliage leaves 4–8 mm apart, opposite-distichous, not imbricate, heterofacially twisted, shortly petiolate; petiole decurrent straight along length of internode, free part 0.2–1.7 mm; lamina of most leaves spreading but not forming a shallow ‘V’ along branch, widely erecto-patent, diverging at 45–70(–80)°, deep green, narrowly lanceolate to lanceolate (widest just above the base), those near middle of branchlets (12–)16–27(–32) × 3.2–6(–7) mm but lowest ones shorter and more ovate, all straight, coriaceous but relatively thin and flexible; margin not thickened, narrowly hyaline; surfaces equally amphistomatic, lacking a raised central area but with longitudinal vein-like striae between the rows of stomata; midrib evident, raised beneath, obscure throughout or at least distally above; apex subacute or obtuse, never mucronate; base rounded, obtuse or cuneate.

Pollen cones terminal or lateral, subtended by a bract, foliage leaf or (when terminal) pair of foliage leaves, solitary or 1–3 together, occasionally in a compound ‘inflorescence’ of 2 pairs of cones and a terminal cone, the individual cones when more than one present each subtended by a bract and shortly pedicellate; cones linear-cylindrical to narrowly cylindrical when terminal, cylindrical to ovoid when lateral; microsporophylls 44–80 per cone, in 11–15(–20) whorls; lamina triangular to ovate or somewhat wedge-shaped with buff-coloured scarious margin; microsporangia reniform, equalling the microsporophyll; pollen with hyaline body and whitish to hyaline sacci.

Female cones lateral, arising from near the base of an ultimate foliage shoot, 1 or 2 per shoot, pedunculate. Peduncle unbranched (bearing 1 cone axis) or forked at apex (bearing 2 cone axes), 2–12 mm, equalling or shorter than cone, longer than vestigial receptacle, erect or erecto-patent when young, with 3–5 pairs of bracts below the cone axis; distal part of peduncle not shed with cone. Peduncle bracts adpressed or subadpressed, ovate, c.1 × 0.5 mm, scarious-margined. Cone axis formed of 6–8 sterile bracts and 1 or 2 fertile bracts. Sterile bracts persisting till resting phase and then slowly caducous, narrowly ovate or ovate-elliptic, subacute, larger than peduncle bracts, at first horizontally spreading, later deflexed and finally caducous. Receptacle cylindrical. Fertile bract initially erect and longer than the epimatium, free from the epimatium and last sterile bract, with a narrowly wedge-shaped, prominently keeled proximal portion and free, ovate-elliptic, acute, non-keeled distal portion. Cones obovoid or pyriform, 15–23 × 10–12 mm; epimatium initially blackish-violet, later turning bright magenta and finally becoming bluish-red when ripe, not crested, with a conical, slightly hooked beak at micropylar end; micropylar beak with two short prongs that point ± vertically downwards towards the last sterile bract.

Taxonomic notes

Seemann was convinced that he was dealing with a new genus allied to Podocarpus (see Seemann, 1863: 33 as well as his annotation on the type, and Wasscher, 1941: 424), although he nevertheless provisionally placed the species in Podocarpus. His perceptive taxonomic insight has since proved to be correct, because there is no doubt that Retrophyllum, in whatever sense it is defined, is distinct from Podocarpus.

Podocarpus filicifolius N.E.Gray, based on material from Morotai in the Moluccas, has usually been regarded as a synonym of this species, but examination of ample material both from Fiji and from New Guinea and Morotai has demonstrated that the plants from the latter areas are taxonomically distinct from Fijian material. They are treated below as Retrophyllum filicifolium.

Seeds of a fossil species allegedly closely allied to Retrophyllum vitiense were found in Pliocene beds in New Zealand by Evans (1937: 192, note; quoted by Florin, 1940: 47). The fossil was not named, and the only other New Zealand fossil that has been definitely assigned to Retrophyllum is the Miocene species R. vulcanense M.Pole (Pole, 1992). The possibility that Evans's seeds may have belonged to an extinct allied genus, such as Smithtonia or Willungia, should not be discounted.

Nomenclatural notes

Page (1989) used Seemann (1863) as the basionym reference, being presumably unaware of the earlier Bonplandia one (Seemann, 1862).

Linden [1869, nomen nudum (see Mill, 2010); in Thurber (1870, nomen nudum); Linden ex K. Koch (1871)], described Torreya bogotensis from Bogotá, Colombia; this was considered to be synonymous with Podocarpus vitiensis (Retrophyllum vitiense) by Bertrand (1874), but this is surely incorrect. If it really is a Retrophyllum at all (and there is some doubt), it is far more likely to be R. rospigliosii, which grows in Colombia. Linden's name was proposed for rejection under Art. 56 of the ICBN (McNeill et al., 2006) by Mill (2010), and this was unanimously recommended by the Nomenclature Committee for Vascular Plants (Brummitt, 2011). Reveal (2012) considered that the 1869 French description by Linden validated the name but Mill (2010) did not, and Brummitt (2011) reported that even the more detailed descriptive matter given by Koch (1871) was “just sufficient to validate the name”.

Phenology

Fl. Mar–Dec; fr. Jun–Feb (Smith, 1979: 99).

Distribution

Southwest Pacific: Solomon Is. (Santa Cruz group), Vanuatu (only Banks Islands, Prov. Torba: absent from the main archipelago), Fiji. The Santa Cruz group (Solomons) and Banks Islands (Vanuatu) are adjacent. TDWG: 60 FIJ SCZ VAN. Map: Fig. 2.

Fig. 2. Global distribution of Retrophyllum vitiense.

Specimens seen and other records. Fiji. Unloc., vii 1860, Seemann 576 (K, 2 sheets annotated ‘1/2’ and ‘2/2’, sheet ‘2/2’ being lecto); isolecto BM–2 sheets, E, S–database; 2nd Seemann collection at S (S–database) without collector's number may be another isolectotype, as may a third “fragment ex Berlin” that lacks any information regarding collector or locality). Unloc., no date, Graeffe s.n. (BM, K). ‘Albizzi Levu’: otherwise unloc., Graeffe (K: Graeffe's locality has not been traced). Unloc., anon. (Department of Agriculture, Fiji) 5536 (L–image). Unloc., 300 m, 10 v 1962, Damanu L12 (K–database). Viti Levu: Western: Nausori Highland intermediate montane zone, 1900 ft [579 m], 8 iii 1962, Damanu NH15 (K). Ba: Nandarivatu, seedlings, 2800 ft [853 m], 21 ii 1927, Mead 1964 (K). Nandarivatu, 2800 ft [853 m], 26 ii 1927, Mead 1974 (K); ibid., 840 m, 9 iii 1927, Mead 1982 (K–database). Nandarivatu, valley of the Sigatoka river, Colo-north (= Tholo North), 800 m, 17 xi 1927, Gillespie 3865 (K); ibid., 900 m, 17 xi 1927, Gillespie 3865 (NY; note altitude discrepancy cf. K sheet). Nadarivatu, 2700 ft [823 m], v 1947, d'Espeissis 1460 (E, 2 sheets). Mt. Lomalagi (Nadarivatu), 3000 ft [914 m], 6 viii 1946, Vaughan 3254 (BM). Nadarivatu, 2700 ft [823 m], ix 1907, Gibbs 674 (BM). Nandarivatu, Easter [2–5 iv] 1926, Teulon s.n., comm. (J.D. & B.H.) Tothill 844 (K). Tholo North, vicinity of Nandarivatu, 750–900 m, Nauwanga, 4 ii–26 iii 1941, Degener 14483 (A, L–image, NY, S–database), 11484 (S–database). Tholo North, vicinity of Nandarivatu, 4 ii–26 iii 1941, Degener 14496 (NY). Mba, Tavua, 2.5 miles (by road) W of Nandarivatu, 2800 ft [853 m], 15 vii 1968, Webster & Hildreth 14270 with I. Kuruvoli (GH). Mba, Mandrongo, Nausori Highlands, W slopes of Mt. Mandrongo [Mangondro], 2000 ft [610 m; but the summit is only c.547 m fide GeoNames], 16 vii 1968, Webster & Hildreth 14277 with I. Kuruvoli (GH). Central: Serua: Naboutini, 4 vii 1962, Damanu R10 (K); ibid., 240 m, 5 xii 1962, Damanu R15 (K–database); ibid., 150 m, 28 iii 1963, Damanu R32 (K–database). Galoa, c.500 ft [152 m], 27 vii 1962, Damanu G7 (K). Near Korovisilou, Serua, 300 ft [91 m], 2 vii 1973, de Laubenfels P 508 (A, L–image); Korovisilou, 900 m, 12 xii 1968, Redrodro K110 (K–database). Road to Nambukilevu from Namboutini, c.1000 ft [305 m], 1 xi 1964, de Laubenfels P 309 (A). Tailevu: hills east of Wainimbuka River, in vicinity of Ndakuivuna, 100–200 m, 15–27 iv 1953, A.C. Smith 7076 (GH, L–image, NY, S–database, K–database). Tailevu/Naitasiri: Upper Rewa River valley, 3 vi 1992, Cox 1360 (NY, BRY–database). Naitasiri: Navolau, 17°52′S 178°24′E, 50 m, vi 1936, Peni 636 (K–database). Kadavu: Eastern:Kadavu: Kadavu Naikorokoro, 18°59′S 178°12′E, 240 m, 12 ix 1962, Damanu KU22 (K–database). Navutulevu, 18°14′S 177°48′E, 90 m, 4 vi 1962, Damanu NL10 (K–database). Vanua Levu: Northern: Cakaudrove: Thakaundrove, Yanawai River region, Mount Kasi, 300–430 m, 10–11 v 1934, A.C. Smith 1796 (GH, NY, K–database, S–database). Macuata: Labasa Sarava, 18°02′S 178°31′E, 300 m, 10 v 1962, Damanu L10 (K–database). Vanuatu. Torba: Banks Islands: Ureparapara, near Lehali village, 13°32′S 167°20′E, 300 m, 6 ix 1982, Hammer s.n. (K–database). Solomon Islands: Temotu: Santa Cruz I.: N.E. Santa Cruz, Carlisle Bay area, 900 ft [274 m], 22 x 1969, Mauriasi et al. BSIP 17025 (K, L–image). Ndeni I.: Graciosa Bay, 130 m, 11 xi 1969, Mauriasi et al. BSIP 17794 (K–database, L–image). Vanikoro I.: Lawrence River, 22 iii 1963, Whitmore BSIP 1580 (K–database, L–image); Vanikolo, 11°37′S 166°54′E, iii 1965, Piaito BSIP 7061 (K–database, L–image); nr Lemon River [11°39′S 166°54′E], 27 xi 1945, Walker BSIP 212 (A, K–database). Fiji (cultivated). Agr. exp. Garden, Naduruloulou, 215 i 1949, Lam 6932 (L–image). Australia (cultivated). New South Wales: Botanic Gardens, Sydney, viii 1908, Boorman s.n. (A). Java (cultivated). Jardin Botanique de Tjiboudas, viii 1959, Schnell 10647 (2 sheets, P–images).

Bioregions: Oceania; New Guinea & Melanesia (only the Santa Cruz and Vanuatu populations)

Ecoregions: OC0105 Fiji tropical moist forests, AA0126 Vanuatu rain forests.

Ecology

Dense lowland wet zone rain forest, montane rain forest or ridge forest, where it typically grows as scattered trees on slopes and in gullies in areas with rainfall of approximately 5000 mm per annum; 50–915 m. According to Mead (1928), this rain forest develops on the southeastern or ‘weather’ sides of the mountains. Associates include species of Agathis Salisb., Calophyllum vitiense Turrill, Gironniera celtidifolia Gaudich. (in the understorey), Parinari insularum A.Gray, Semecarpus vitiensis (A.Gray) Engl. and Pagiantha thurstonii (Horne ex Baker) A.C.Sm. (Keppel, 2005; Mead, 1928, gives additional rain forest components). It can form a unique community in which it and Calophyllum vitiense are codominant (Keppel et al., 2011).

Chromosome number: 2n = 20 (Hair & Beuzenberg, 1958, as Podocarpus vitiensis)

Phytochemistry

The wood of Retrophyllum vitiense contains podocarpic acid, totarol, 19-hydroxytotarol and 4β-carboxy-19-nortotarol (Cambie et al., 1983), a unique combination of phenolics among the 10 species of Podocarpaceae studied.

Plant–animal interactions

The beetle Xyleborus rameus Schedl is known to bore into the seeds of Retrophyllum vitiense in Fiji (Greenwood, 1977). The drywood termite Cryptotermes domesticus (Haviland) is known to attack timber of Retrophyllum vitiense in Fiji (Gray, 1974).

IUCN conservation assessment (global: IUCN 3.1)

The most recent assessments are as LC (Farjon, 2010: 943; Thomas, 2013), but these both referred to the species in a wide sense, including what is here recognised as the separate species Retrophyllum filicifolium. Doyle (1999), using an earlier IUCN grading system, evaluated the species as circumscribed here as LRlc on both Fiji and the Solomon Islands, also equivalent to IUCN 3.1: LC. Therefore it is considered that an assessment of LC is appropriate for the species as more narrowly circumscribed here, although it should be noted that Farjon commented that, in the wider geographical area of his circumscription, the species was subject to logging throughout its range.

Uses

Retrophyllum vitiense is one of the most highly prized timber trees of Fiji, being used for high-quality furniture, etc. and for flooring, exterior work and construction (Damanu 15 in sched., K). Mead (1928) commented that in his time it was one of the commonest and most important timber trees of Fiji. He also mentioned its use in boat-building. However, use of the species today should be avoided, because most imports are from non-sustainable forests (Rainforest Information Centre, no date, accessed 16 September 2015). Moreover, the wood is not very decay-resistant and is therefore unsuitable for use in some situations, such as ground contact (Osborne, 1967). The resin is inflammable and is used by Fijians to start fires (Degener 14483 in sched., A, NY and Degener & Degener, 1953).

2. Retrophyllum filicifolium (N.E.Gray) R.R.Mill, comb. nov. – Podocarpus filicifolius N.E.Gray, J. Arnold Arbor. 43: 74 (15 Jan. 1962) quoad descr. foliorum tantum. – Lectotype designated here: Moluccas, Morotai, “tree, 15 m”, 1949, A. Kostermans s.n. (lecto L, foliage shoots only – image, originally determined by van Steenis as Podocarpus vitiensis Seem.; isolecto A, K). The lectotype specimen at L is a mixture of juvenile foliage of Retrophyllum filicifolium and detached fruits of Nageia wallichiana; the latter are here excluded as a result of lectotypification by the foliage element. The isolectotypes cited comprise foliage only. Epitype designated here: New Guinea, Northern District, ridge above Doma, 1500 m, 14 xi 1962, P.J.B. Woods 345 (epi E, barcode E00094518: comprising adult foliage, several attached young female cones, one attached older female cone and another in packet; isoepi K–database). Figs 1A, G, K, 3A–J.

Fig. 3. Retrophyllum filicifolium. A, Branching arrangement (Woods 345). B, Juvenile foliage shoot (Woods 241). C, Adult male shoot (Carr 15666). D, Young male cones (Carr 15666). E, Mature male cone (Carr 15666). F, Microsporophyll and microsporangia, adaxial view (Carr 15666). G, Microsporophyll and microsporangia, abaxial view (Carr 15666). H, Adult female branch (Woods 345). I, Young female cone (Woods 345). J. Mature female cone (Woods 345). Magnifications: A, B, × 0.67; C, H, × 1.5; D, E, I, J, × 3; F, G × 15. Scale bars: A, B, 6 cm; C, H, 2 cm; D, E, I, J, 1 cm; F, G, 1 mm. Drawn by Claire Banks.

Podocarpus vitiensis sensu Wasscher (1941: 425–427), Gray (1962: 72–74), p.p. quoad pl. novoguineensis, non Seemann (1862).

Decussocarpus vitiensis sensu de Laubenfels (1969: 342) p.p. excl. typ., quoad syn. Podocarpus filicifolius N.E.Gray (p.p.).

Nageia vitiensis (Seem.) Kuntze sensu de Laubenfels (1988: 394–395) p.p. excl. typ., quoad syn. Podocarpus filicifolius N.E.Gray (p.p.).

Nageia vitiensis sensu Takeuchi (2010: 20) non (Seem.) Kuntze (1891).

Retrophyllum vitiense sensu Li (2007: 346), Abdillahi et al. (2010: 7) p.p. quoad pl. novoguineensis.

Etymology: filicifolium, Latin from filix, ‘a fern’, and folium, ‘a leaf’

“The name refers to the fern-like appearance of the pinnately leaved twigs, reminiscent of the royal fern, Osmunda regalis” (Gray, 1962: 74).

Vernacular names: lehil (New Britain: Frodin NGF 26292), moegò (New Guinea – Wissel Lakes, Kapaukoe language, Vink & Schram BW8730 as Retrophyllum vitiense)

Distinguishing features. Retrophyllum filicifolium has for long been confused with and included within R. vitiense, as explained in the Taxonomic Problems section of this paper

Male material of Retrophyllum filicifolium can be distinguished from R. vitiense by its pollen cones with fewer microsporophylls (typically 20–32 in 5–8 whorls as opposed to 44–80 in 11–20 whorls). Terminal pollen cones, which occur in some examples of Retrophyllum vitiense, including its type specimen, have not so far been observed in R. filicifolium. Material with young female cones can be separated from Retrophyllum vitiense by the fewer sterile bracts on the female cone axis (about 4, not 6–8) that are ephemeral and quickly caducous, not persisting till the cones reach resting phase, by the fertile bract with the proximal half lacking a keel, and by the obovoid, not cylindrical, receptacle. The ripe cones also appear to be slightly larger on the basis of the few that have been seen, being c.26 × 17 mm rather than 15–23 × 10–12 mm, although this may not hold when additional material is examined. Vegetative differences include the more quickly caducous bud scales that leave pale scars on the branchlet, the typically shorter juvenile and adult ultimate foliage shoots with fewer pairs of leaves (although there is overlap) and especially the broadest point of both adult and juvenile leaves being near the middle of the lamina rather than towards the base as in Retrophyllum vitiense. There is no geographical overlap between the two species, so if the provenance is known, identification is automatic.

Tree, 27–60 m, dbh 50–130 cm. Trunk sometimes spur-buttressed at base, when adult unbranched except near top. Bark smooth, flaking, fissured, dark brown or blackish on mature trees, light or medium brown on juveniles; inner bark reddish-brown, salmon or pink; wood pale orange, straw-coloured, yellow or white. Crown not widely spreading. Terminal buds on primary branches globose or ovoid; bud scales present, in 4 series, decussate, the lowest pair longer and slightly wider than the other three pairs, all ovate or ovate-rhombic and slightly keeled, acute, brownish with very narrow paler entire margin, caducous, leaving cream, transversely narrowly rhombic scars much paler than the branchlet; those on foliage shoots usually remaining dormant, ovoid. Ultimate juvenile shoots alternate, subopposite or opposite, 65–210 mm, with 7–35 (usually 11–20) pairs of foliage leaves, occasionally biramulate or sending off a lateral scale-bearing shoot. Juvenile foliage leaves 4–7 mm apart, opposite-distichous, heterofacially turned at base, diverging at 60–85°, shortly petiolate, the petiole 0.2–1 mm, lamina lanceolate or narrowly lanceolate with the lowest ones elliptic, oblong-elliptic, ovate or narrowly ovate and considerably smaller, those near middle of branchlet 18–40 × 4.2–6 mm, the apex often indistinctly mucronate. Adult (and juvenile) trees with dimorphic shoots, most shoots bearing foliage shoots but leaders bearing scale leaves; shoot flattening heterofacial with the decurrent bases running straight down the internodes. Primary adult branches not observed. Penultimate adult branches erecto-patent to widely patent with respect to subtending branch, straight or slightly curved, bearing opposite-decussate axillary lateral foliage shoots interspersed with remote pairs of decussate subadpressed to spreading ovate scales. Ultimate adult foliage shoots 20–90 (–120) mm with 4–18(–25) pairs of leaves, straight or curved, internodes slightly flexuous. Leaf scars absent. Adult foliage leaves 2–7 mm apart, opposite-distichous, pectinate, not imbricate, heterofacially twisted, subsessile to shortly petiolate; petiole decurrent straight along whole length of internode, free part 0.2–1.5(–2) mm); laminas spreading and at least sometimes forming a shallow ‘V’ along the branchlet (cf. label of epitype), widely erecto-patent, diverging at (45–)50–80(–90)°, mid to deep green, narrowly elliptic to elliptic or oblong (widest just below the middle), those near middle of branchlets 7–25(–30) × 3–4.5(–5) mm but lowest ones shorter, all straight, coriaceous and moderately thick; margin not thickened, very narrowly hyaline; abaxial surface flat to ± convex, adaxial surface ± flat, both equally amphistomatic, lacking a raised central area but with vein-like striae alternating with the stomatal rows; midrib evident, raised beneath but obscure throughout or at least distally above; apex subacute, obtuse or sometimes broadly rounded, usually not mucronate, the base rounded, obtuse or cuneate.

Pollen cones terminal on lateral shoots arising just above bud scales of previous season, subtended by a bract, 1–several together in groups that sometimes form a raceme-like inflorescence, ellipsoid or ovoid; microsporophylls 20–32(–40) per cone, in 5–8(–10) whorls; lamina triangular to ovate with buff-coloured scarious margin; microsporangia reniform; pollen hyaline.

Female cones borne laterally from near the base of an ultimate foliage shoot of the previous year's growth, the reproductive shoot subtended by a bract. Peduncle 5–10 mm, equalling or shorter than unripe cone and always shorter than ripe cone, longer than vestigial receptacle, erect or erecto-patent when young, becoming pendulous, with 3–5 pairs of bracts below the cone axis. Peduncle bracts adpressed or subadpressed, ovate to ovate-rhombic, c.1.5 × 0.3–0.5 mm, very narrowly scarious-margined. Cone axis formed of c.4 sterile bracts and 1 or 2 fertile bracts. Sterile bracts ephemeral, present at receptivity but then quickly caducous, elliptic, obtuse to rounded, larger than peduncle bracts, at first spreading but soon becoming deflexed and falling. Receptacle obovoid, purplish-brown, glaucous. Fertile bract initially erect, its proximal half fused to the epimatium and lacking a keel. Cones obovoid or pyriform, c.26 × 17 mm; epimatium green when young becoming red when ripe, glaucous when unripe, not crested, with conical, slightly hooked beak at micropylar end.

Distribution

Malesia: Moluccas (Morotai: type locality of Podocarpus filicifolius), New Guinea, Bismarck Archipelago (New Ireland, New Britain). Morotai marks not only the westernmost limit of this species, but also of the genus Retrophyllum in Asia. TDWG: 42 MOL 43 BIS NWG-IJ NWG-PN. Map: Fig. 4.

Fig. 4. Global distribution of Retrophyllum filicifolium.

Specimens seen and other records. Indonesia. North Maluku: Morotai Island Regency: Morotai, without detailed locality, 1949, tree, 15 m, Kostermans s.n. (lecto L–image, excl. fruits of Nageia wallichiana in packet; isolecto A, K). Papua: Jayapura: Dist. Hollandia, Cycloop Mountains, 1220 m, 28 vi 1961, van Royen & Sleumer 6073 (A, K, L–image, P–image). Res. Hollandia, Cycloop Mts (Ifar–Ormu), 1100 m, 2 xi 1954, Versteegh B.W. 913 (A, CANB–n.v., K, L–2 sheets, images). Paniai: Wissel Lake Region, slope and summit of Mt. Barara, 4 ix 1939, Eyma 5155 (L–image). Motito, Wissel Lakes, 1800 m, 18 v 1960, Vink & Schram BW 8730 (CANB–n.v., L–image, collection date wrongly entered in database as ‘18 vi’). Wamena: 6 km SW of Bernhard Camp, Idenburg River, 1200 m, ii 1939, Brass 12787 (A, BM, K–database, L–image); ibid., juvenile foliage of no. 12787, from a plant 4 m high, 1200 m, ii 1939, Brass 12787a (A, L–image). 6 km SW of Bernhard Camp, Idenburg [Taritatu] River, 1200 m, 17 ii 1939, Brass & Versteegh 12534 (A, L–image). ibid., “further material of no. 12787,” 1200 m, ii 1939, Brass 12912 (A, L–image). Papua New Guinea. Northern [Oro] (border with Central): ridge above Doma, 1500 m, 14 xi 1962, Woods 345 (epitype E, isoepitype K–database); near Doma village, 900–1500 m, 2 xi 1962, Woods 241 (E). Northern [Oro]: Sohe: Lala River, c.5000 ft [1524 m], 20 ii 1936, Carr 15666 (BM–2 sheets, CANB–n.v., L–image, NY, S–database). Alola, c.6000 ft [1829 m], 3 i 1936, Carr 14160 (A, BM–2 sheets, L–image, NY). Eastern Highlands: Lufa: Herowana, Crater Mt. WMA, N of airstrip, 6°39′3′′S 145°11′30′′E, 21 xi 2005, Kuria LAE 87216 & Oliver (LAE–n.v., NSW–n.v.). Hela: Koroba-Kopiago: Southern Highlands, c. 6 miles [9.5 km] E of Koroba station, 5°42′S 142°48′E, c.80 ft. [24.5 m] tall, 26 viii 1961, Pullen 2840 (LAE 96292–n.v.: de Laubenfels, 1969: 343). West New Britain: Talasea: NE slope of Mt. Tangis, SW of Airagilpua, Talasea, 5°40′S 148°25′E, 3500 ft [1067 m], 17 xi 1965, Frodin NGF 26292 (A, CANB–n.v., L–image, NY). Mt. Tangis, NW slope, Talasea, 5°39′S 148°18′E, 2400 ft [732 m], 30 v 1966, Frodin NGF 26917 (A, CANB–n.v., L–image, NY). Kandrian: Umai River, Fullerborn Harbour, 6°06′S 150°40′E, 100 ft. [30.5 m], 5 iii 1965, Hammermaster & Sayers NGF 21842 (L–2 sheets, images); 7 miles SE of Benim village, Wariai subdist., Kandrian dist., 1000 ft [305 m], 5°47′S 148°57′E, 23 iii 1966, Henty & Frodin NGF 27359 (A, CANB–n.v., K, L–image, NY). New Ireland: Namatanai: Namatanai subprovince, Hans Meyer Range, 4°24′S 152°58′E, c.8 km (map distance) WNW of Taron on E coast, 1350 m, 29 x 1975, Sands 2381 et al. (A, CANB–n.v., K–database, L–image). Northern Hans Meyer Range, 70 km SE of Namatanai, 850 m, 3°03′S 152°40′E, 1 xi 1984, Gideon & Obedi LAE 77181 (BISH–n.v., CANB–n.v., E, K–database, L–image).

Bioregions: Wallacea; New Guinea & Melanesia

Ecoregions: AA106 Halmahera Rain Forests, AA0112 New Britain–New Ireland Montane Rain Forests, AA0116 Northern New Guinea Montane Rain Forests.

Ecology

According to field notes on specimen labels, Retrophyllum filicifolium is a typical species of tall primary montane rain forest or ridge forest, associated with Dacrycarpus (Endl.) de Laub., emergent palms such as Gulubia Becc. and other trees including Agathis, Podocarpus L'Hér. ex Pers., Serianthes Benth., Syzygium R.Br. ex Gaertn., Casuarina rumphiana Miq., and sometimes in shade among bamboos. It occurs at altitudes of 300–1800 m, mostly between 900 and 1500 m. It would seem to occur chiefly at higher altitudes than Retrophyllum vitiense which, in Fiji and the Solomon Islands, grows mainly between 100 and 900 m.

Provisional IUCN conservation assessment: LC

This species is here reinstated for the first time after a lapse of more than 40 years (when it was synonymised with Podocarpus vitiensis by de Laubenfels, 1968), and there has therefore been no previous conservation assessment for it as a separate species. The species appears to be widespread on New Guinea, and it is therefore regarded as Least Concern, although it is, or has been, subjected to logging.

Uses

None have been recorded. However, they are likely to be similar to those of Retrophyllum vitiense.

Group B. Bark exfoliating in short strips or pieces. Vegetative shoots dimorphic. Terminal buds protected by reduced foliage leaves; true bud scales absent. Leaves heterofacially flattened in both juvenile and adult phases. Adult leaves in 4 ranks, thick to very thick; juvenile ones distichous. Resin canal 1, situated below vascular bundle. Pollen cones in groups at branch tips, on a common peduncle (sometimes ± sessile), individual cones sessile, subtended by scale-like bracts and the lower ones in the group axillary to leaves; microsporophylls decussate (appearing spirally arranged), triangular or ovate. Female cones terminating ordinary foliage shoots, proleptic; peduncle not broadened distally, in its vegetative portion bearing spreading bracts; cone axis with 1 or 2 sterile bracts below fertile bract. Fertile bract free from epimatium, flat. Receptacle formed from 1 or 2 peduncle bracts plus the last sterile bract and the fertile bract, obovoid. Seed beaked at micropylar end. Germination epigeal. Seedling phyllotaxis decussate. – New Caledonia.

3. Retrophyllum comptonii (J.Buchholz) C.N. Page, Notes Roy. Bot. Gard. Edinburgh 45(2): 380 (22 Feb. 1989 [‘1988’]). – Podocarpus comptonii J.Buchholz, Bull. Mus. Hist. Nat. Paris sér. 2, 21: 284 (30 Jul. 1949). – Decussocarpus comptonii (J.Buchholz) de Laub., J. Arnold Arbor. 50: 344 (15 Jul. 1969). – Nageia comptonii (J. Buchholz) de Laub., Blumea 32(1): 211 (3 Feb. 1987). – Type: New Caledonia, Mt. Mou, J. Buchholz 1684 (holo ILL–image, photos A, P; iso ILL–image, K–image, MO, NY, S–image [S-C-1294]). Paratypes Mt. Mou, Buchholz 1085 (ILL–image; isoparatypes elsewhere as cited below); Mt. Mou, Buchholz 1449 (ILL–image; isoparatypes elsewhere as cited below); Mt. Humboldt, Buchholz 1697 (ILL–image), 1697a and 1697s. Fig. 1B, E, H.

Iconography

De Laubenfels, Fl. Nouv. Caléd. 4: 51, pl. 11 f. 1–8 (1972), as Decussocarpus comptonii.

Etymology

This species is named after Robert Harold Compton (1886–1979), Cambridge botanist who, after the First World War, became director of Kirstenbosch Botanical Garden (South Africa) in 1919. Compton collected Retrophyllum comptonii several times during a field expedition to New Caledonia in 1914, although he was not the first to do so; a much earlier collection by Vieillard has been seen.

Vernacular names: palissandre (Chevalier, 1957: 114; Suprin, 2011b)

Distinguishing features

Retrophyllum comptonii in most of its wide range has the capacity to become a tall tree up to at least 30 m, although in the far south of the island it tends to be much shorter. It is rarely if ever an aquatic. The adult foliage is mostly opposite decussate, with only the leaves in the upper part of the tree in 4 ranks. The adult leaves have a raised central area that is always narrower than the leaf margins and when dried is marked either by two (not three) parallel ridges or longitudinal wrinkles. The surface of the seed is scalloped and ridged, these low and not prominent.

Tree or shrub, 2.5–30 m tall, 12–80 cm dbh. Trunk apparently not buttressed, the bole straight and lacking branches for 10 m or more in large trees. Bark smooth when young, ± rough on older trees, peeling in short vertical strips or ragged pieces, fissured, dark grey, greyish-brown, tan or reddish-brown; inner bark brownish; wood cream or whitish. Crown pyramidal, finally rounded. Buds ovoid, protected by very reduced foliage leaves; true bud scales absent. Juvenile shoots with 1 or 2 very small leaves at base and 22–35 pairs of foliage leaves in up to 3 growth increments. Juvenile leaves 5–7 mm apart, opposite, distichous and pectinate, heterofacially turned at base, diverging at 55–65°, sessile or subsessile, lamina narrowly lanceolate or linear-lanceolate, 20–35 × c.2.5 mm on seedlings, somewhat smaller on older saplings, gradually tapered to a subacute, narrow and frequently slightly mucronate tip. Adult trees with dimorphic shoots, most shoots bearing foliage leaves but a few leader shoots bearing scale leaves and these either emitting lateral foliage shoots or terminating in foliage shoots; shoot flattening fully heterofacial on juvenile and transitional shoots but only partly so on adult branches. Primary adult branches irregularly arranged, divaricate or widely spreading, opposite, alternate or in whorls of 3. Penultimate adult branchlets erecto-patent to divaricate, ± curved, initially clad with distant, alternate, small, elliptic-ovate subadpressed scales with decurrent bases, their free laminas finally deciduous. Ultimate adult foliage shoots ± clustered at tips of penultimate branches, erecto-patent to suberect, flexuous, 40–70 mm with 8–15 pairs of leaves, subtended by 1 or 2 pairs of very reduced scale-like foliage leaves. Leaf scars absent. Lowest true adult foliage leaves smaller and relatively broader than the others; adult leaves in 4 ranks or some (on transitional shoots) opposite-decussate and distichous, crowded, 3–4 mm apart, sessile with partly twisted bases, most showing heterofacial twisting, scarcely or not imbricate, erecto-patent to suberect, mid to dark green; lamina of leaves in middle of shoot elliptic, oblong-elliptic, narrowly lanceolate or lanceolate, 6–15 × 2–5 mm, straight, thick and coriaceous; margin not thickened, narrowly hyaline; surfaces unequally amphistomatic, with more stomata on adaxial surface; midrib not visible on either surface but both surfaces with a raised, narrowly elliptic central area narrower than the margins and either longitudinally wrinkled or marked by two parallel ridges when dried; apex obtuse or broadly rounded; base obtuse or cuneate.

Pollen cones lateral (axillary) or terminal on lateral shoots of current growth, subtended by a leaf or scale-leaf, 1–5 together, pedunculate (individual cones sessile when > 1 together); peduncle shorter than cone(s), erecto-patent, bearing 1 pair of ± adpressed scales; peduncle scales greenish, keeled, ovate with decurrent base, acute, with very narrow, buff scarious margin; cones ellipsoid or ovoid, 4–6(–12) × 2.5–3 mm, straight, reddish, or greenish with a pink tinge; microsporophylls decussate but appearing spirally arranged, 16–20(–28) per cone in 4 or 5 short/3 or 4 long ‘spirals’; lamina greenish sometimes tinged pink or red, broadly triangular or deltate and abruptly narrowed to an acute, apiculate apex, with very narrow whitish hyaline margin, longer than the microsporangia; microsporangia pinkish, elliptic; pollen white or cream.

Female cones terminal on lateral foliage shoots of current growth, shortly pedunculate. Peduncle in line with shoot or erecto-patent, shorter than remainder of cone but longer than receptacle, bearing 1 (occasionally 2) cone(s), bearing 8–10 bracts in 4 or 5 decussate pairs; distal part of peduncle shed with cone. Peduncle bracts light green turning brownish green, not keeled but convex abaxially, elliptic or ovate-elliptic, 1.5–4 × c.1–1.5 mm, subacute to obtuse. Proximal sterile bracts (cone axis) absent other than the one subtending the fertile bract. Cone subtended by a non-fleshy receptacle formed from 1 or 2 peduncle bracts plus either one sterile and one fertile bracts or two fertile (or part-fertile) bracts. Receptacle narrowly obovoid or narrowly pyriform, light green at first becoming orange. Sterile bract normally 1, light green becoming pale orange, stiffly spreading to erecto-patent, elliptic to ovate-elliptic, obtuse, with scarious margin. Fertile bract light green becoming pale orange, initially erect, with median longitudinal groove, wholly free from epimatium, with reflexed, ovate free tip with acute apex. Cones pyriform, 20–25 × 13–20 mm; epimatium green becoming scarlet or deep red when ripe (occasionally blue-black, fide Munzinger et al. 1680 in sched., P), with short conical apical crest; seed beaked at micropylar end, the beak usually straight, occasionally curved. Seed surface with two ridges leading to the micropylar beak, not porous and not adapted to water dispersal.

Germination epigeal. Seedlings with two brown, erect, soon deciduous cataphylls that envelop the base of the stem and have greenish linear lanceolate tips that protect the first leaves; initial leaves of main axis cotyledon-like, c.6 pairs, separated by short, naked internodes, widely spreading and the lowest pair becoming recurved, their blades linear, 21–28 × 1.5–2 mm, longer than the leaves on the initial branchlets, basally not twisted, opposite-decussate, amphistomatic.

Distribution

New Caledonia (throughout the island except the northern half of the western coast). TDWG: 60 NWC. Map: Fig. 5.

Fig. 5. Global distribution of Retrophyllum comptonii.

Specimens seen and other records. New Caledonia. Unloc., Baudouin 542 (P–image); unloc., no date, Bernier s.n. (P–image, barcode P00188269); unloc., Pancher s.n. (P–image, barcode P00188209); unloc., Sarlin 228 (P–image) and 238 (P–image); collection details obscured, Bernier 268 (K–database, P–image, possibly a composite sheet of up to three separate gatherings?; Bernier 269 is from Plaine des Lacs); Province Nord: Hienghène: Contrefort est du Mt. Panié, 1000–1400 m, 10 ix 1966, MacKee 15594 (P–image); ibid., 800 m, 10 ix 1966, MacKee 15639 (P–image, seedlings); Mont Panié [20°35′S 164°46′E], 700–900 m, 19 ix 1966, Schmid 1422 (NOU–database). Mont Panié, 800 m, 3 viii 1978, MacKee 35577 (P–image); Mt. Panié, c.20 air-km NW of Hienghène, 950 m, 6 xii 1979, McPherson 2191 (MO–database, P–image, NOU–database); Mont Panié [20°35′S 164°46′E], 950 m, 3 iii 1983, Lauri 45 (NOU–database). Mt. Panié, above Haut Coulna, on SW slopes, 20°36′82[sic]′′S 164°44′40′′E, 1060–1250 m, 30 x 1999, McPherson & van der Werff 17834 (P–image, NOU–database); Mt. Panié, NE slopes on summit route via eastern trail, 20°35′00′′S 164°46′14′′E, 1158 m, 7 xii 2002, Gardner et al., Third New Caledonia Araucaria Exped. (TNCA) 5045 (E). Kaala-Gomen: Mont Taom (Crête Ouest), 900 m, 24 iv 1982, MacKee 40376 (MO–database, P–image, NOU–database), 40377 (P–image). Kouaoua: décharge Montmartre, 750 m, 14 ii 1978, Veillon 3514 (P–image, NOU–database, Z–database). Poindimié: Wagap, Vieillard 3264 [p.p.] (P–image, original label bearing manuscript name “Podocarpus Guillainii”). Sommet du Mt. Grandie (Haute Amoa), 900–960 m, 14 v 1968, MacKee 18807 (NOU–database, P–images, 2 sheets), 18808 (NOU–database, P–image); massif de Grandie, pente SE, c.700 m, 31 iii 1988, Veillon 6783 (P–image, NOU–database). Dans une cuvette au massif du Tchingou, c.1250 m, 18 iv 1951, Hürlimann 1220 (P–image, Z–database); Massif de Tchingou, 750 m, 13 xii 1983, Morat 7621 (P–image); Massif de Tchingou, 20°53′36′′S 165°00′42′′E, 1160–1175 m, 1 iv 2001, McPherson & Munzinger 18100 (P–image, MO–database, NOU–database); ibid., 20°54′15′′S 165°01′19′′E, 760–850 m, 5 iv 2001, McPherson & Munzinger 18149 (MO–database). Pouébo: Bois du montagne à Balade, 1855–1860, Vieillard 1265 (P–image). Mont Colnett [20°29′S 164°41′E], c.1000 m, 1 x 1997, Suprin 2628 (NOU–database); sur la crête SW du Mt. Colnett, c.1200 m, 12 ix 1951, Hürlimann 1964 (P–image, S–database), 1966 (P–image, Z–database); Mont Colnett (pente est), 1100 m, xi 1981, Nasi s.n. (MacKee 39997) (P–image); ibid., 1100 m, xi 1981, Nasi (MacKee 40011) (P–image); Mt. Colnett, E slopes, 20°30′00′′S 164°42′52′′E, 1000 m, 29 x 2003, McPherson 19070 et al. (P–image, MO–database); ibid., 1000 m, 29 x 2003, McPherson 19073 et al. (P–image, MO–database, S–database). ‘Ignambi, 18 iii 1914’, Compton 608 (S–database under Retrophyllum minus, 3 accessions in spirit coll., nos. S-A-3669, S-A-3681, S-A-3684; the locality, date [only given for S-A-3681] or collection number of all must be wrong because Compton collected at Mt. Mou in the S of the island on that date, see Compton 607 below); Ignambi, > 3500 ft [1067 m], 30 vii 1914, Compton 1524 (A–photo, BM, K); Ignambi, Compton 1527 (S–database under R. minus); Ignambi, 3000 ft [914 m], 1 viii 1914, Compton 1587 (BM, A–photo); am Gipfel des Ignambi, 4 vi 1925, Däniker 2901a (Z–database); en montant vers le sommet de l'Ignambi, c.1250 m, 17 viii 1951, Hürlimann 1842 (P–image, Z–database); sur la crête du massif Ignambi le long de la route de Gomen, c.1170 m, 25 viii 1951, Hürlimann 1832 (P–image, Z–database); summit of Ignambi, on trail from Oubatche to Gomen, c.1000 m, 12 vi 1956, Foster 160 (P–image); near top of Mt. Ignambi, c.10 km S of Panébo, c.1300 m, 7 xii 1963, Green 1783 (A); Mont Ignambi, 1100–1200 m, 19 viii 1965, Bernardi 10347 (L–image, P–image, Z–database); E slopes of Mont Ignambi, 20°26′39′′S 164°34′58′′E, 1200 m, 6 iii 1999, Cretinon & Gardner, ICCP New Caledonia Exped. (1999): 73 (E); Ignambi [20°27′S 164°35′E], 1250–1300 m, 2 v 2002, Lowry 5747 (MO–database, NOU–database). Pouembout: Kopéto [21°10′S 165°00′E], 940 m, 8 viii 2006, Rigault & Barrière s.n. (Barrabé 365) (NOU–database). Ad montem Paéoua metalliferum, 900 m, 13 viii 1965, Bernardi 10131 (K, L–image, P–image, Z–database); ibid., 900–950 m, 13 viii 1965, Bernardi 10149 (K–database, L–image, P–image, S–database, Z–database); Mt. Paéoua, 900–1100 m, 4 vii 1967, MacKee 17032 (P–image), 17056 (K, NOU–image, P–image, Z–database); Mt. Paéoua, c.1100 m, 16 x 1986, Veillon 6062 (P–image, NOU–database); Paéoua, 21°10′36.0′′S 165°05′14.2′′E, 1077 m, 5 xii 2005, Gardner et al. 256 (E). Poya (N): Mt. Boulinda, c.1102 m, 26 iv 1965, Veillon 120 (P–image, NOU–database; P sheet has ‘Blanchon’ blacked out and ‘Veillon’ substituted below, with Blanchon's initials J.P. remaining; Veillon's are J.M.). Boulinda [21°14′S 165°09′E], 1102 m, 26 iv 1965, Veillon 122 (NOU–database). Mont Boulinda, c.1200 m, 28 iv 1965, Schmid 137 (P–image); Mt. Boulinda, pente nord, 1150–1300 m, 28 viii 1967, MacKee 17354 (P–image), 17357 (NOU–database, P–image, Z–database), 17358 (P–image). Monte Boulinda, 450–900 m, 12–15 iv 1968, Bernardi 12754 (K, P–image, Z–database); Massif de Boulinda, N of Poya, c.850 m, 28 viii 1981, McPherson 4121 (GH, P–image, MO–database, NOU–database, Z–database). Aoupinié [21°10′S 165°18′E], 1000 m, 16 ix 1981, Suprin 1407 (NOU–database). Province Sud: Bouloupari: Dent du St Vincent, vii 1909, Le Rat 71 (P–image, seedlings, also duplicate without locality details, P–image). Mont Do, 1000 m, 28 xi 1966, Veillon 949 (P–image, NOU–database); c.15 km towards NW, Mont Do, upper region towards top, 700–1014 m, 28 xi 1966, Ehrendorfer 6600-138-38 (WU–image). Mont Do [21°45′S 165°59′E], 1000 m, 24 v 1977, Musselman 5415 (NOU–database). Bourail: au-dessus de Téné, près de Bourail, 18 iii 1869, Balansa 1381 (P–image; duplicate without locality information, P–image). Pente boisée d'un contrefort du massif Mé Maoya (Haute Houaïlou), 800–900 m, 28 xii 1962, MacKee 9886 (CANB–database, L–image, P–image, K–database). Poya, Mé Maoya Massif, southern edge of Këiyoümê in disused mine, 21°26′37.8′′S 165°21′43.5′′E, 1012 m, 2 xi 2008, Hollingsworth et al. 194 (E, P–image). Mé Aoui, 850 m, 18 vi 1981, Cherrier (MacKee 39235) (NOU–database, P–image). Canala: Crête Mt. Nakada, 21°37′S 166°03′E, 1050 m, 6 x 1977, MacKee 33961 (MO–database, NOU–database). Near Prokoméo radio tower, NW of Canala, 21°30′S 165°51′E, 700 m, 3 ii 1983, McPherson 5477 (MO–database). Mt. Nakada, 21°37′35′′S 166°03′22′′E, 800–900 m, 17 iv 2001, McPherson & Munzinger 18254 (P–image, MO–database); ibid., 21°37′82 [sic]′′S 166°03′35′′E, 780 m, 18 iv 2001, Munzinger 764 & McPherson (P–image). Dumbéa: Mt. Dzumac, 1000 m, 28 iv 1951, Guillaumin & Baumann-Bodenheim 12717 (Z–database), 12725 (P–image, Z–database), 12727 (P–image, Z–database); N face of Mt. Dzumac, c.1000 m, 27 ix 1963, Green 1216 (A, K–database, NOU–image). Pic de Kouwehé [Couvélé], 700–1000 m, 4 ix 1882, Brousmiche 697 (P–image). Au fond de la vallée derrière la mine Sunshine, c.730 m, 15 iii 1951, Hürlimann 1062 (P–image, K–database, Z–database). Near Pic du Rocher, on mountain crest along trail to Pic du Rocher, Plateau Montagne des Sources, nr Pic du Rocher, 950 m, 15 x 1947, Buchholz 1222 (NY, P–image, K–database, S–database; photo of trees in situ, A). Un vallon au dessous de la route vers la Montagne des Sources au Pic Buse, c.530 m, 11 vii 1951, Hürlimann 1573 (P–image, Z–database). Mont Dzumac, 900 m, 25 vii 1981, McPherson 3958 (NOU–database). Mont Dzumac [22°02′S 166°28′E], 880 m, 5 i 1982, Suprin 1612 (NOU–database, 2 sheets); ibid., 1050 m, 4 v 1982, Suprin 1830 (K–database, NOU–database). Mont Dzumac, ‘0 m’ [sic: from map, must be c.600 m], 11 ix 1984, Jaffré 2585 (NOU–database, 2 sheets). Le Mont-Dore: Thy [22°10′S 166°31′E], 980 m, 19 x 1982, Brinon 1359 (NOU–database). Mont Koghis, [22°10′S 166°31′E], 1000 m, 26 x 1982, Zlarnik 30 (NOU–database). Inland from Baie des Pirogues, 16 x 1923, White 2120 (A, 2 sheets; S–database). Moindou: Mé Ori, 900–1000 m, 3 xii 1969, MacKee 21219 (P–image, K–database, MO–database, NOU–database, Z–database). Along old lumber road to top of Mt. Mé Ori, above Katrikoin, 21°31′S 165°39′E, 8 ix 1980, McPherson 3066 (MO–database, 2 sheets). Bourail, E slopes and summit area of Mé Ori (ascent starting from Katrikoin), 21°31′31.7′′S 165°39′56.4′′E, 31 x 2008, Hollingsworth et al. 153 (E, P–image). Me Adeo pente O, c.750 m, 24 iii 1988, Veillon 6747 (P–image, NOU–database). Païta: Mont Humboldt, 1400 m, 16 ix 1902, Schlechter 15331 (A–photo of BM sheet, BM, E, K–2 sheets, L–image [currently as R. minus, May 2015], P–image, S–database, Z–database); ibid., 1400 m, Schlechter 15332 (P–image); Mt. Humboldt, 1200–1300 m, Buchholz 1578 (A–photo, seed; NY, S–database); Mt. Humboldt, 1400 m, 20 ix 1951, Baumann-Bodenheim 15393 (P–image, Z–database); ibid., 1400 m, 20 ix 1951, Baumann-Bodenheim 15411 (P–image, Z–database); Mt. Humboldt, c.1100 m, xi 1973, Schmid 4838 (P–image, NOU–database); Mont Humboldt, Réserve Spéciale Botanique, 29 xi 2005, Gardner et al. 108 (E); Mont Humboldt [21°53′S 166°24′E], 1380 m, 2 xii 2009, Grignon et al. 559 (NOU–database). Mt. Dzumac, 800 m, 1 xii 1957, de Laubenfels P153 (S–database); Mont Dzumac, [22°02′S 166°27′E], 950 m, 6 xii 1964, Blanchon 1246 (NOU–database); Mt. Dzumac, 760 m, 29 xi 1964, de Laubenfels P 415 (K); Mt. Dzumac, 8 vii 1965, Aubréville & Heine 229 (P–image); Mont Dzumac, 22°06′–22°08′S 166°27′–166°32′E, 800–900 m, 8 vii 1965, Bernardi 9520 (P–image). Crête entre Mt. Dzumac et Mt. Ouin, c.1000 m, 22 xii 1962, Barets 8 (P–image); ibid., 900 m, 11 x 1967, MacKee 17670 (P–image); ibid., 900 m, 17 iv 1968, MacKee 18694 (P–image, seedlings); ibid., 900 m, 7 vii 1969, MacKee 20244 (P–image); ibid., 6 i 1982, MacKee 40185 (P–image, K–database, NOU–database, Z–database). Haute Ouinné, côte orientale, 750 m, 18 i 1948, Bernier 267 (P–image); Ouinné sup., 700 m, 29 iv 1951, Guillaumin & Baumann-Bodenheim 12815 (P–image, Z–database), 12843 (P–image, Z–database), 12861 (P–image, Z–database); ibid., 900 m, 30 iv 1951, Guillaumin & Baumann-Bodenheim 12910 (Z–database), 12960 (P–image). Monts Dzumac and down to the Rivière Ouinné, 800 m, 22°02′54′′S 166°28′47′′E, 9 iii 1999, Cretinon & Gardner, ICCP New Caledonia Exped. 1999: 90 (E); Monts Dzumac road down to the Rivière Ouinné, 800 m, 22°02′54′′S 166°28′47′′E, 3 vi 2001, Gardner et al., New Caledonia Araucaria Exped. 2001: 1026 (E, 2 sheets). Mont Mou, 1861–1867, Vieillard 3064 (GH, P–image, original labels of both sheets bearing manuscript name “Podocarpus Guillainii”); sommet du Mont Mou, 1866, Vieillard 3264 [p.p.] (P–image, with manuscript name “Podocarpus Guillainii”). Mt. Mou, 1910–11, Godefroy s.n. (P–image); Mont Mou, 3500 ft [1067 m], 18 iii 1914, Compton 607 (BM, A–photo; Compton 608, at S and said to be collected from Ignambi, could not have been collected on 18 iii 1914 as the S database record says); summit of Mt. Mou, 3500 ft [1067 m], x 1923, White 2033 (E, P–image, K–database); Gipfelcrête des Mt. Mou, 21 ii 1926, Däniker 2902 (P–image, Z–database); crête sommitale du Mont Mou, c.1150 m, 19 vi 1938, Virot s.n. (A, P–image); ibid., c.1100 m, 4 ix 1938, Virot 38 (A, P–image); crête sommitale du Mont Mou, 1150 m, 19 ii 1939, Résineux no. 8, coll. Virot (A, P–image); Mont Mou, 1100 m, 28 ix 1947, Buchholz 1085 (paratypes ILL–2 sheets, images; isoparatypes A–photo, GOET–image, NY, P–image, K–database, S–database); Mt. Mou, 1100 m, 4 xii 1947, Buchholz 1447 (A–photo), 1449 (paratype ILL–image, isoparatypes A–photo [seed kernel]), K, NY, P–image, S–database: paratype and isoparatypes of Podocarpus comptonii); Mont Mou, 1100 m, 4 xii 1947, Buchholz 1452 (K–database, NY, S–database); Mt. Mou, 1100 m, 9 ii 1948, Buchholz 1684 (holo ILL–image, A and P–photos of holotype, iso ILL–image, K–image, MO, NY, S–image); Mt. Mou, 1100 m, 9 ii 1948, Buchholz 1684s (small seedlings, ILL–image, P–image); Mt. Mou, 1100 m, 5 vi 1948, Buchholz 1791 (NY, S–database); Mont Mou, Pic des Mousses, 1000–1100 m, 23 i 1949, Skottsberg 202 (P–image); Mont Mou, Pic des Mousses, 23 i 1949, Selling 202 (S–database); Mont Mou, 1000 m, 22 xi 1949, MacDaniels 2323 (P-1041) (NY–database). Mt. Mou, 1200 m, 23 viii 1950, Baumann-Bodenheim 5654B (Z–database); Mt. Mou, 1200 m, 13 iii 1951, Guillaumin & Baumann-Bodenheim 11257 (P–image, Z–database); ibid., 1200 m, 13 iii 1951, Guillaumin & Baumann-Bodenheim 11261 (P–image, Z–database), 11282 (P–image, Z–database), 11299 (P–image, Z–database), 11301 (P–image, Z–database); Mount Mou, c.20 miles NW of Nouméa and c.5 miles N. of Païta, 26 vi 1952, McMillan 5015 (P–image); Mt. Mou, below N. summit, 1100 m, 25 xi 1955, MacKee 3516 (A, E, P–image); sommet du Mont Mou, Blanchon 341 (P–image); Mont Mou, 1200 m, 9 x 1957, de Laubenfels P129 (S–database, spirit coll.; SBT–n.v., cited by de Laubenfels, 1969: 345); ibid., 9 ix 1960, Erdtman s.n. (3 collections, S–database); Mt. Mou (c.25 km N of Nouméa), c.1200 m, 7 x 1963, Green 1268 (A, K–database, NOU–database); Mt. Mou, 1140 m, 18 xi 1964, de Laubenfels P 360 (A); ibid., juvenile, 1140 m, 18 xi 1964, de Laubenfels P 361 (A); Mont Mou, 28 v 1965, Boisseau (hb. MacKee 12725) (P–image); Mont Mou, 22°03′–22°04′S 166°19′–166°22′E, 30 vii 1965, Bernardi 9879 (L–image, P–image, Z–database); Mt. Mou, along trail starting near Sanatorium c.7 air-km NNW of Païta, 1100 m, 8 ix 1979, McPherson 1867 (P–image, MO–database, NOU–database); Mt. Mou, along trail from near Sanatorium to summit, 1150 m, 8 viii 1981, McPherson 4020 (P–image, K–database, MO–database, NOU–database); Mont Mou, close to summit on N ridge slope, 1000 m, 22°04′S 166°20′E, 27 v 2001, Gardner et al., New Caledonia Araucaria Exped. 2001: 1014 (E). Poya (S): Contrefort Ouest du Mé Maoya au dessus de la Mine Emma, 1350 m, 11 vii 1965, Corbasson (hb. MacKee 13037) (P–image, K–database); ibid., 1400–1450 m, 2 x 1965, MacKee 13492 (K, P–image); Thio: Mont Poueari (SE Thio) [21°40′S 166°14′E], 1140 m, 13 x 1994, Suprin (hb. MacKee 46368) (P–image, NOU–database). Forêt de Sailles, 21°39′58′′S 166°14′49′′E, 1100 m, 12 iv 2001, Munzinger 1200, Suprin & Carriconde (P–image, det. Suprin incorrectly as Prumnopitys ferruginoides); ibid., forêt au bas du col juste avant le mont Pwénari, 21°39′58′′S 166°14′49′′E, 1200 m, 8 xii 2001, Munzinger 1307, Suprin & Carriconde (MO–n.v., P–image, NOU–database). Forêt de Sailles vers Thio, parcelle 3 no. 25909, 21°39′56′′S 166°14′38′′E, 849 m, 18 xi 2009, Grignon et al. 584 (P–image, NOU–database). Mont Ninga [21°45′S 166°08′E], 19 vi 1975, Sévenet 958 (NOU–database). Pic de Na Kado, Brousmiche s.n. (P–image, named as Decussocarpus minor (Carr.) de Laub. by de Laubenfels). Mt. Nekandi, 1200–1300 m, 7 xi 1967, MacKee 17908 (P–image). Nékando [21°51′S 166°26′E], 1100 m, 28 xi 1983, Veillon 5659 (NOU–database). Yaté: Montagne des Sources, Pic des Conifères et Campement “Bernier”, 900–1000 m, x 1947, Bernier 271 (P–image and duplicate without locality information, P–image). Montagne des Sources, 850 m, 9 vii 1949, Selling 208 (S–database); au dessous du Camp Bernier à la Montagne des Sources, c.800 m, 21 ii 1951, Hürlimann 931 (NY, P–image, Z–database); Montagne des Sources, 22°05′–22°09′ S 166°32′–166°38′ E, 6 vii 1965, Bernardi 9445 (L–image, P–image, Z–database); ibid., 8 ix 1960, Erdtman & Chevalier s.n. (S–database); ibid., 8 ix 1960, Erdtman s.n. (3 sheets, S–database); Montagne des Sources [22°07′S 166°35′E], 600 m, 6 vii 1965, Veillon 281 (NOU–database). Crête sommet Montagne des Sources, 22°07′S 166°36′E, 900 m, 7 x 1969, MacKee 20928 (NOU–database). Montagne des Sources [22°07′S 166°35′E], 850 m, 24 i 1982, Suprin 1636 (NOU–database). Montagne des Sources, Parc Territorial de la Rivière Bleue, highest summit area to which trail leads, 22°06′52.06′′S 166°36′19.68′′E, 1011 m, 11 xii 2002, Gardner et al., Third New Caledonia Araucaria Exped. (TNCA) 2036 (E). Forêt du Mois de mai, 150 m, viii 1947, Bernier 203 (P–3 sheets, images, one juvenile and lacking locality or date); Forest du Mois de mai, Plaine des Lacs near Walker's place, 4 xi 1947, Buchholz 1350 (K–database, P–image, NY, S–database); Plaine des Lacs, Mois de mai, 4 xi 1947, Buchholz 1350a (NY, S–database); Forêt Walker, Haute Yaté, Plaine des Lacs, c.200 m, 6 xi 1947, Bernier 269 (P–image, seedlings); ibid., same details, Bernier 270 (P–image); Mois de mai forest, Plaine des Lacs, 160–200 m, 6 xi 1947, Buchholz 1359 (A, P–image, S–database), 1359a (P–image, MO–database, S–database); Forest du Mois de mai, 180–250 m, 13 ii 1948, Buchholz 1697 (A–photo, seed kernel; ILL–image, K, NY, S–database; paratype and isoparatypes), 1697a (S–database), 1697s (ILL–image, NY); forest du Mois de mai near Walker's place, Plaine des Lacs, 180–200 m, no date, Buchholz 1697 (P–image, isoparatype, label details differ slightly from others of this number); forêt “Mois de mai” (Riv. blanche), 13 vii 1951, Baumann-Bodenheim 14057 (P–image, Z–database); ibid., 14 viii 1951, Baumann-Bodenheim 15178 (NY, P–image, Z–2 sheets, database); ibid., 14 viii 1951, Baumann-Bodenheim 15197 (P–image, Z–database); Rivière Blanche, 200 m, 24 vii 1954, H.D. Ingle I. 66 (L–image, MEL–n.v.); Mois de mai, 300 m, 22°07′S 166°39′E, 15–29 iii 1987, Watt 519 (NY–image, juvenile, fragment). Parc Territorial de la Rivière Bleue, Mois de mai, 22°07′11.2′′S 166°39′36.8′′E, 230 m, 27 xi 2005, Gardner et al. 68 (E). [Forêt] des Electriques (Rivière blanche sup.), [c.350 m], 17 vii 1958, Hürlimann 3173 (Z–database, 2 sheets); Rivière Bleue, 5 viii 1951, Baumann-Bodenheim 15028 (NY, P–image, Z–2 sheets, database); Rivière Bleue [22°06′S 166°38′E, c.200 m], 4 viii 1963, Blanchon 399 (NOU–database, det. Blanchon, needs confirmation); forêt de la Rivière Bleue, 150 m, 1 vi 1994, Pintaud 51 (juv., P–image); Plaines des Lacs, sources de la Kuébini, 24 vi 1980, Cherrier (hb. MacKee 39255) (P–image); margin of Rivière Bleue at Pont Germain, 22°06′05.2′′S 166°39′28.9′′E, 111 m, 27 xi 2005, Gardner et al. 85 (E). Kouakoué, piste Haute Ouinni, 22 ix 1977, Veillon 3305 (P–image, NOU–database); Kouakoué [21°58′S 166°31′E], 1300 m, 22 ix 1977, Morat 5676 (NOU–database). Haute Yaté, Rivière Bleue, 200 m, 1 x 1981, MacKee 39724 (A, E, NOU–database, P–image); Mt. Kouakoué, 21°57′29′′S 166°32′20′′E, 26 xi 2002, Munzinger 1661, Tronchet et al. (P–image, NOU–database); Mt. Kouakoué, 21°57′31′′S 166°32′11′′E, 26 xi 2002, Munzinger 1680, Tronchet et al. (P–image, NOU–database); Kouakoué, along trail SW of camp, 21°57′49′′S 166°32′07′′E, 12 v 2006, Lowry II et al. 6821 (P–image, MO–database, NOU–database); Kouakoué [21°57′S 166°31′E], 950 m, 3 xi 2009, Grignon et al. 470 (NOU–database). Forested slopes above a tributary of the Rivière Ni, 22°00′17′′S 166°28′42′′E, 950–1000 m, 10 xi 2003, McPherson 19209 & Mouly (P–image, MO–database). Mamié, between Ointapoin and Tiova, 22°06′23.8′′S 166°53′03′′E, 552 m, 27 x 2008, Hollingsworth et al. 90 (E, P–image).

USA (Cultivated). Georgia: Atlanta Botanical Garden, Fuqua Conservatory, 16 xii 2009, Silba B-640 (accession 9040015/1996-0015, coll. A. Watt in New Caledonia) (NY–image).

A specimen collected by Bernard Suprin (in herb. MacKee, no. 46368, P, barcoded P001655140) very closely resembles Retrophyllum minus (treated below) in branching pattern but is from a high-altitude locality in the north of New Caledonia and belongs to R. comptonii, with which it agrees in leaf characters. Vieillard 1265 from Balade (extreme NE New Caledonia) also has to belong to Retrophyllum comptonii on geographical and habitat grounds, although it was determined as Podocarpus minor Parl. by Guillaumin in 1943 – this was 6 years before Podocarpus comptonii (now Retrophyllum comptonii) was segregated from P. minor (now R. minus). The later annotation slip “= Nageia minor Carrière cf. de Laubenfels, Blumea 32: 211, 1987” is misleading and represents an uncritical up-dating of the nomenclature on the 1943 determinavit slip rather than a new determination; this is also a feature of several other specimens in the P herbarium. The specimen not unexpectedly has the characteristics of Retrophyllum comptonii. On the other hand, a specimen from Port Boisé (Le Mont-Dore) that de Laubenfels in 1972 determined as Decussocarpus comptonii has in this revision been reassigned to Retrophyllum minus, along with others from the same locality that had previously been identified as Decussocarpus minor. Gray (1962: 75) cited an example of Vieillard 1275 from Mt. Mou at P under Podocarpus comptonii, but I have not seen such a specimen and it is not among those imaged in the P database: all examples of Vieillard 1275 that I have seen are from low-lying localities and are Retrophyllum minus. It is possible that the specimen Gray (1962) cited is the one now barcoded P00188118, which is an unnumbered Pancher collection but was given the number Vieillard 1275 and mentions Mt. Mou among several collecting localities, which may not all relate to that number.

Veillon 2962 from Ouégoa is currently determined at P as Decussocarpus comptonii, but the corresponding sheet at NOU is indicated as having been identified as Falcatifolium taxoides (Brongniart & Gris) de Laub. by Veillon on the date of collection. After examining images of both sheets, on account of its leaf and pollen cone morphology I concur with Veillon's determination.

Bioregion: New Guinea & Melanesia

Ecoregion: AA0113 New Caledonia Rain Forest.

Ecology

Dense humid rain forest, moss forest or dry montane forests on mountain slopes and by rocky riverbanks; chiefly on ultrabasic rocks but also occurring on gneiss (Mt. Ignambi, Mt. Colnett) and limestone (Mt. Panié); 150–1600 m. The lower altitudinal range was given as 600 m by Farjon (2010), even though de Laubenfels (1972) had cited numerous collections from locations such as Mois de mai that are at much lower altitudes; more recent collections from as low as 150 m in the Plaine des Lacs area have also been made, and the species descends to near sea level at Baie des Pirogues in the extreme south of Grande Terre. In these low-lying locations, the species grows together with Retrophyllum minus and some individuals are very difficult to separate from it. Indeed, it is possible that some plants of Retrophyllum minus, including the epitype collection, with female cones maroon-red or greenish red, are in fact fertile hybrids between R. comptonii and the lacustrine form of R. minus, equivalent to Buchholz's Podocarpus palustris, although there is as yet no evidence to support this hypothesis other than their intermediate morphology. Associates of Retrophyllum comptonii in its more typical montane forest habitats include other gymnosperm trees such as Acmopyle pancheri Pilg., Araucaria humboldtensis J.Buchholz, A. laubenfelsii Corbasson, A. montana Brongn. & Gris, A. muelleri (Carrière) Brongn. & Gris, Austrotaxus spicata Compton, Falcatifolium taxoides (Brongn. & Gris) de Laub., Podocarpus sylvestris J.Buchholz and various angiosperm trees such as species of Nothofagus Blume.

Little or nothing is known about the dispersal of this species. The attractively coloured red ripe female cones suggest an adaptation to dispersal by vertebrates such as birds, as postulated by Farjon (2010), although the flesh has been described as tasteless (McPherson 2191 in sched.). Pigeons such as Ducula goliath (Gray), the New Caledonian Imperial Pigeon, might act as dispersers, although no fruits of Podocarpaceae were recorded in the diet of that species in one limited study in the months of March and April (Barre et al., 2003).

Chromosome number: 2n = 20 (Hair & Beuzenberg, 1958, as Podocarpus comptonii)

Phytochemistry

Cambie et al. (1984) recorded the presence of the phenolics totarol and 19-hydroxytotarol in the wood of Retrophyllum comptonii. These were the only two phenolic compounds isolated; podocarpic acid and 4β-carboxy-9-nortotarol were both absent, in contrast to their earlier report on Retrophyllum vitiense (Cambie et al., 1983).

Mycological associations

The ascomycete fungus Caliciopsis podocarpi Huguenin was first described from two specimens; the host of the holotype was Retrophyllum minus and that of the other collection was R. comptonii (Huguenin, 1969: 300). This was the first instance of Caliciopsis Peck occurring on Retrophyllum or indeed any member of ‘Podocarpus sensu lato’ (in the sense of Buchholz & Gray, 1948), although the genus was later recorded from Podocarpus nubigenus (Butin, 1970).

IUCN conservation assessment: LC (Farjon, 2010: 938; Jaffré et al., 2010: 1489; Thomas, 2010a)

The species is one of the most widespread of all New Caledonia's conifers, occurring throughout the island.

Uses

According to Farjon (2010), the species has been extensively logged, especially at lower altitudes, for its high-quality wood used locally for flooring, furniture and similar purposes.

4. Retrophyllum minus (Carrière) C.N. Page, Notes Roy. Bot. Gard. Edinburgh 45(2): 380 (22 Feb. 1989 [‘1988]’, ‘minor’). – Nageia minor Carrière, Traité Gén. Conif. ed. 2, 2: 641 (15 Jan. 1867). – Podocarpus minor (Carrière) Parl. in A. DC., Prodr. 16(2): 509 (1868). – Decussocarpus minor (Carrière) de Laub., J. Arnold Arbor. 50: 352 (15 Jul. 1969). – Type citation: [New Caledonia] “Habite dans la Nouvelle-Calédonie, au sommet de très-hautes montagnes, où il fut découvert par un jardinier anglais nommé Richard. – D'après M. Vieillard (herbier de la Nouvelle Calédonie no. 1275 – in Herb. Mus. Par.) on le trouve aussi au bord du lac Arnaud.” – Lectotype designated here: New Caledonia, Yaté, “bord du Lac Arnaud”, Vieillard 1275 (P: sheet barcoded P00118110, top left sterile specimen, here designated 110-A: see Appendix I). Isolectotypes (all Vieillard 1275, P): sheet barcoded P00188109, sterile specimens at top centre right here designated 109-C and at bottom right here designated 109-G; sheet barcoded P00188110, unbranched sterile specimen c.26 cm long at right here designated 110-C and shoot in packet with c.9 larger leaves here designated 110-G. Epitype designated here: New Caledonia, “from banks of River des Lacs [sic] at site of former village below mine excavations of Madeliene [sic] mine Plaine des Lacs, 140 m or probably less. Tree 2–3 m high; wood sample with wood yellowish density near 0.5. Seeds maroon red, elongated pear-shaped, larger than coll. made of var. at 22 km Sta.”, 22 ii 1948, J.T. Buchholz 1729 (epi NY, isoepi ILL–n.v. [photos A, P], K, MO–n.v. [fide TROPICOS database], S–database). – For full details of the very difficult and complex typification of this name, see Appendix I. Fig. 1L.

[Podocarpus minor (Carrière) Parl. var. palustris J.Buchholz, in sched. Buchholz 1421, nom. nud.]. – Podocarpus palustris J.Buchholz, Bull. Mus. Natl. Hist. Nat. (Paris) sér. 2, 21: 284 (1949). – Type: New Caledonia, “Small tree growing in water, with stems flaring at base, dark gray bark, brownish beneath. 2–3 m high, some stems 40–50 cm in diameter at water line. Wood white soft of density about 0.3. Trees subject to flooding at times the leaves becoming covered with ferrugenus [sic] silt. Seeds smaller than in species coll. on Rio des Lacs”, 27 xi 1947, J. Buchholz 1421 (holo ILL–image barcoded ILL00010021, iso ILL–2 sheets–images barcoded ILL00010019 and ILL00010020, A–photo, K, MO–image (stamped 1712087), NY, P–image barcoded P00188158, RSA–image, S–image, TEX–image barcoded 00370048, WIS–image barcoded 0057120).

Iconography

De Laubenfels, Fl. Nouv. Caléd. 4: 51, pl. 11 f. 9–15 (1972), as Decussocarpus minor.

Etymology

Retrophyllum minus was originally described by Carrière (1867) in the genus Nageia as N. minor Carrière. The other six species that he enumerated in that genus all had leaves measuring 40–160 mm in length, whereas he gave the length of the leaves of Nageia minor as 12–17 mm. The epithet minor (Latin for ‘smaller’; minus when neuter) therefore most probably alludes to the much smaller leaves of Retrophyllum minus compared with the leaves of the other six Nageia species, although there are no clues in his protologue as to why the epithet was chosen. Although less likely, it might also refer to the smaller stature of Retrophyllum minus compared with those six Nageia species, for which Carrière only gave actual height details for two. The epithet has nothing whatever to do with Retrophyllum minus being the smallest species in the genus Retrophyllum, as stated by Farjon (2010: 938) and implied also by Gray (1962: 76), because, with the exception of R. vitiense (1863), of which Carrière was apparently unaware as he gave no treatment of it under any genus, all other species of Retrophyllum were first described between 1923 and 1983, long after Carrière described Nageia minor. Buchholz's epithet palustris alludes to the fact that the species is very frequently aquatic.

Vernacular name: bois bouchon (Jaffré, 1988; Suprin, 2011a) – French for ‘cork wood’, referring to the very light, corky texture of the wood (an adaptation to its aquatic habitat)

Distinguishing features

Retrophyllum minus is a low shrub or small tree that is typically aquatic. The adult foliage is all in 4 ranks. The very thick adult leaves have a raised central area that when dried is marked by longitudinal wrinkles or by three (not two) ridges, and is typically wider than the leaf margins, although this is variable and plants with a narrower central area can be confused with Retrophyllum comptonii unless other characters are also examined. The seed surface is rough, porous and very buoyant, an adaptation to the aquatic habitat.

Small tree or shrub, frequently aquatic, normally 0.9–3.5 m tall but rarely attaining 8 m; 20–50 cm diam. at water-line. Trunk rapidly tapering from a wide, buttressed base, sometimes split from near base. Bark rough when old, peeling in short vertical strips or ragged pieces, fissured, dark grey, greyish brown or tan (colour often masked by ferruginous silt); inner bark brownish or reddish-brown; wood cream or whitish, very light. Crown open, irregular. Buds shortly ovoid to shortly ellipsoid, protected by very reduced foliage leaves; true bud scales absent. Primary branches irregularly arranged, divaricate, widely spreading or ascending; foliage branches clustered at tips of branches, erecto-patent to erect, straight, curved or flexuous. Shoots dimorphic but appearing mostly monomorphic, with only occasional scale-bearing shoots that soon pass into foliage shoots. Shoot flattening heterofacial on juvenile shoots but adult shoots ± lacking flattening. Juvenile foliage shoots with (1–)2–4 small leaves at base and 17–35 pairs of foliage leaves in 2–3 growth increments. Juvenile leaves (3.5–)4.5–6(–7.5) mm apart, opposite, distichous and pectinate, heterofacially turned at base, diverging at (40–)50–60(–70)°, sessile or subsessile, lamina narrowly lanceolate, linear-lanceolate or lanceolate-elliptic, 25–33 × 2.8–4 mm, gradually tapered to a narrow, obtuse tip. Adult foliage shoots subtended by 1–3 pairs of very reduced scale-like foliage leaves, the lowest true foliage leaves smaller and relatively broader than the others; leaves opposite-decussate in 4 ranks, crowded, sessile with bases not twisted and all with abaxial surfaces outermost, the upper leaves of each shoot partially imbricate, erecto-patent to erect, either yellow- to mid-green or grey-green; lamina of leaves in middle of shoot narrowly elliptic, elliptic, narrowly lanceolate-elliptic, lanceolate-elliptic or narrowly lanceolate, straight, very thick and coriaceous; margin thickened, extremely narrowly hyaline; midrib not visible on either surface but both surfaces with a raised elliptic central area as wide as or wider than the margins and either longitudinally wrinkled or marked by three longitudinal lines when dried; apex obtuse, broadly rounded or occasionally subacute; base cuneate or shortly attenuate.

Pollen cones lateral (axillary) or terminal on lateral shoots of current growth, subtended by a leaf or scale-leaf, 1–7 together, pedunculate (individual cones sessile when > 1 together); peduncle shorter than cone(s), erecto-patent, bearing 1–3 decussate pairs of spreading scales; peduncle scales green or grey-green, not keeled, ovate or ovate-lanceolate with ± long decurrent base and ± patent free part, subacute, with or without a narrow scarious margin; cones ellipsoid, ovoid or subglobose, 4–8 × 2–3 mm, straight; microsporophylls decussate but appearing spirally arranged, 24–40 per cone, the lamina green, ± heavily tinged violet, broadly triangular, deltate or triangular-ovate, abruptly narrowed to an acute, apiculate apex, with extremely narrow whitish hyaline margin, shorter than the microsporangia; microsporangia free, pinkish, elliptic or semicircular, stomium longitudinal, abaxial; pollen white to hyaline.

Female cones terminal on lateral foliage shoots of current growth, shortly pedunculate. Peduncle in line with shoot or erecto-patent, shorter than remainder of cone but longer or shorter than receptacle, bearing either 8–10 or 4 bracts in 4–5 or 2 decussate pairs respectively; distal part of peduncle shed with cone. Peduncle bracts light green or grey-green, keeled, elliptic, ovate-elliptic or ovate-rhombic, subacute to obtuse. Cone subtended by a non-fleshy to slightly fleshy receptacle formed from 1 or 2 peduncle bracts plus the sterile and fertile bracts. Receptacle narrowly obovoid, narrowly pyriform, narrowly obconical or cylindrical, light green or greyish-violet. Sterile bract normally 1, stiffly spreading to erecto-patent, oblong-elliptic, ovate-elliptic or elliptic, obtuse or subacute, with very narrow scarious margin. Fertile bract light green or greyish-violet, initially erect, with median longitudinal groove, wholly free from epimatium, with spreading, ovate free tip with acute apex. Cone obovoid (sometimes shortly so), ellipsoid or pyriform; epimatium either green at first, then blushing red and finally maroon-red when ripe, or grey-green and not turning red, with a shortly conical (sometimes indistinct) apical crest; seed beaked at micropylar end, the beak straight or curved. Seed with rugulose, porous texture, adapted to floating in water. Germination epigeal. Seedlings: initial leaves of main axis c.5 pairs, linear, 15–18 × c.1.5 mm excluding decurrent base, basally not twisted, opposite-decussate in a spiral arrangement, amphistomatic, separated by short, naked internodes, the youngest ones narrowly linear, apiculate and 23–31 × 1.5–2.2 mm, older ones becoming opposite in one plane, linear-elliptic or narrowly elliptic, shorter (11–20 × 1.9–3.0 mm) and bluntly mucronate at tip.

Taxonomy

Retrophyllum minus is a variable species. Some states, including the type, have ascending branches, leaves yellow- to mid-green when fresh, and female cones initially green turning red or finally maroon, as in Retrophyllum comptonii. Other states, including virtually all the aquatic material assigned by Buchholz to Podocarpus palustris and collected since, have more widely divaricate or patent branches, leaves grey-green, and female cones that are grey-green or whitish and apparently not turning red when ripe. Such plants also appear to have more microsporophylls per pollen cone (30–40) compared with plants at the other end of the spectrum of variation, which have only about 25 microsporophylls per cone. The plants with almost white, extremely glaucous ripe female cones are very distinct and easily recognisable in the field as belonging to Retrophyllum minus. This is not the case with the state having pinkish-green to red ripe cones, which often shows Retrophyllum comptonii-like characteristics. The variation within the species clearly influenced Buchholz to describe one extreme, that with cones not turning red when ripe, as Podocarpus palustris. Plants assignable to this and to typical Retrophyllum minus often occur together at the same site, however. The two states appear to be very distinct, especially when reproductive organs are present, and the fact that many of the distinguishing features are reproductive rather than vegetative also makes the case for recognising varieties within the species stronger than it might otherwise be. Matters are complicated, however, by the occurrence of low-altitude populations of Retrophyllum comptonii sympatrically with R. minus in the Plaine des Lacs area, and it is therefore possible that hybridisation may occur. Indeed, as mentioned under Retrophyllum comptonii, the morphological characters of ‘typical’ Retrophyllum minus (i.e. including its type material) are intermediate between those of Podocarpus palustris and Retrophyllum comptonii, and a hybrid origin of those populations of R. minus agreeing with its type therefore needs testing. For this reason, varieties are not here recognised within the species.

Material from the Port Boisé area (hb. MacKee 19303, 19304, 19305) was determined at Paris as both Decussocarpus minor and D. comptonii by de Laubenfels in 1972. All three sheets seem identical in foliage characters and are more similar to Retrophyllum comptonii than they are to R. minus. However, plants from the same population were analysed using RAPD by Herbert et al. (2002), who found that the Rivière Trou Bleu population, coded by them as RTB, grouped with Retrophyllum minus, not R. comptonii, using this technique. Hence, all the specimens from this locality are here assigned to Retrophyllum minus.

Distribution

New Caledonia (Grande Terre, Province Sud: Prony, Baie du Sud, Lac en Huit, Rivière des Lacs, Plaine des Lacs). Map: Fig. 6.

Fig. 6. Global distribution of Retrophyllum minus.

In the Rivière des Lacs area Retrophyllum minus coexists with plants that are assignable to Retrophyllum comptonii and as noted above some plants possess characters intermediate between the two species, suggestive of hybridisation. Most of the area of the two lakes forming Lac en Huit, one of the classic localities for Retrophyllum minus, is in Yaté commune, but the southern third of the eastern lake is in Le Mont-Dore, as well as the RM9 road (‘Route de l'Aérodrome’) along the south side of both lakes and all of the CR7 road (‘Route du Déversoir’ – the Nouméa–Yaté main road). Records from this area have therefore been assigned to one or other commune, or an unassignable boundary zone, depending on field notes and their accuracy. The whole of Grand Lac (Lac Arnaud) is in Yaté.

Specimens seen and other records. NEW CALEDONIA. Unloc., Le Rat 253A (S–database). Unloc., no date, Petit 138 (P–image); Bernier 245/1 (seedlings; P–image, barcode P00188175); Bernier 245/4 (P–image, barcode P00188172); Aubréville & Heine 241 (P–image); unloc., 1950–1951, Baumann-Bodenheim 6378 (P–image, juvenile and adult); unloc., 1861, Deplanche 170 (P–image); unloc., 1863, Deplanche 170 (P–image, central area of leaf rather narrow and transitional to R. comptonii but determined in 1950 as Podocarpus minor by Buchholz & Gray; A–photo). Without definite locality, “Hauteur un mètre sur les plateaux couverts du marais fort arides. 8 m sur les flancs escarpees des Montagnes du même sol, fruit rouge,” Pancher (Vieillard 1275) (BM–also seemingly (re?)numbered Hance 17247, K). Province Sud: Le Mont-Dore: Baie du Sud, 1861–1867, Vieillard 1275 (A, GH, K–2 sheets, P–2 sheets, images and also A–photos). Prony, without date, Franc 207 (BM [photo A], GH, K: note discrepancy in numbering and dating of other sheets, see below). Bords d'un torrent situé au fond de la Baie de Prony, ix 1868, Balansa 186 (2 sheets, P–images). Prony, Baie de Kué, 3 m, x 1903, Cribs 1493 (P–image, K–database); Prony, Le Rat 149 (P–image). Prony, x 1913, Franc 207 (P–image); ibid., x 1913, Franc 207 (Série A) (2 sheets, P–images); ibid., i 1914, Franc 207 (A); Prony, undated, Franc 207 (Série A) (A); Prony, 100 m, Pancher s.n. (P–image: label bears manuscript names ‘Podocarpus nana Parl.’, ‘Dacrydium austrocaledonicum Vieillard’ and ‘Podocarpus Guillainii Vieillard’, as well as the inscription “C'est la même plant que Vieillard 1275”, the first five words of which are in different handwriting to the rest of the label, which also gives locality information for specimens collected at Cougui 800 m and Mt. Mou 1200 m that are, from the altitudes, assignable to Retrophyllum comptonii). Prony, where road crosses the Rivière Bleue, very close to the river margin, 50 m, 22°18′35′′S 166°49′55′′E, 28 iii 1999, Cretinon & Gardner, ICCP New Caledonia Exped. 1999: 17 (E). Rivière des Pirogues, 31 x 1923, White 2261 (E, K, P–image). Le Carénage, just above sea, 8 iv 1955, MacKee 2373 (E, L–image, P–image); c.5 miles from Baie du Carénage, on road between Baie du Carénage and highway from Nouméa to Yaté, 600 ft [183 m], 22 vii 1952, McMillan 5139 (A, E, K, L–image, P–image). Port Boisé, embouchure du Port Boisé, vallée de Koué, 100 m, 28 ii 1976, Viratelle (hb. MacKee 32355) (P–image). Port Boisé, embouchure du ruisseau Trou Bleu, 20 viii 1968, Lavoix (hb. MacKee 19303 & 19304, both det. de Laubenfels as Decussocarpus minor) (P–images, 19303 also NOU–database, 19304 also K–database); ibid., Lavoix (hb. MacKee 19305, det. as D. comptonii or R. comptonii) (P–image, K–database, NOU–database). Port Boisé, Rivière Trou Bleue, Cretinon & Gardner, ICCP New Caledonia Exped. 1999: 50 (E; also specimen collected 25 iv 2001 from E cultivated accession 19990384C, E); ibid., 1 iii 1999, Cretinon & Gardner, ICCP New Caledonia Exped. 1999: 50A (E, seedling). Lac en Huit [22°16′S 166°52′E], 7 x 1957, de Laubenfels P115 (SBT–n.v., fide de Laubenfels, 1969: 345; S–database); ibid., 19 v 1977, Musselman 5280 (NOU–database). Rivière Bleue, route de Prony [22°17′S 166°46′E], xi 2007, Munzinger 4830 (NOU–database). Le Mont-Dore / Yaté boundary: Plaine des Lacs, 22°14′–22°16′S 166°50′–166°57′E, ad ripam lacus “en 8” (Lac en Huit) dicti, 250 m, 3 vii 1965, Bernardi 9369 (L–image, P–image, Z–database); along banks of La Madeleine (Rivière des Lacs), S of Nouméa–Yaté road, edge of Lac en Huit, c.250 m, 30 viii 1980, McPherson 2996 (P–image); Lac en Huit, 150 m, 15 xi 1955, MacKee 3382 (A, L–image, P–image). Yaté: Bords du lac Arnaud, 1855–1860, Vieillard 1275 p.p. (lectotype and isolectotypes as specified above and in Appendix I, P–images); ibid., Vieillard 1275 p.p. (non-lectotype material as specified in Appendix I, P–image). Grand Lac (Plaine des Lacs), 250 m, 30 iii 1942, Virot 658 (A, P–2 sheets, images); western margin of Grand Lac, 22°15′52′′S 166°54′30′′E, 250 m, 1 iii 1999, Cretinon & Gardner, ICCP New Caledonia Exped. 1999: 44 (E). Plaine des Lacs, Lac en Huit, au niveau du déversoir (berge du lac), 250 m, 8 i 1987, Jérémie & Tirel 1572 (P–image). Lac en Huit, 22°16′16.0′′S 166°53′13.5′′E, 256 m, 24 xi 2005, Gardner et al. 02 (E). Lac en Huit, northern shore-line of second lake, 22°15′20.3′′S 166°53′51′′E, 241 m, 24 xi 2005, Gardner et al. 04 (E). Plaine des Lacs, prior to 19 xii 1889 (date of receipt at P), Raoul s.n. (P–image); Plaine des Lacs, Le Rat 607 (BM and photo of this sheet A; P–image), 751 (P–image), 1040 (P–image); Bord des rivières de la Plaine des Lacs, ix 1905, Le Rat 2587 (P–image). Plaine des Lacs, bords de la rivière Pouéné, ix 1905, Le Rat 2621 (2 sheets, P–images, one without locality information). Plaine des Lacs, 22 km [station], c.750 m, viii 1947, Bernier 204 (P–image); 22 km station, 31 x 1947, Bernier 245/3 (female, P–image, barcode P00188173); Plaine des Lacs, 22 km Station, 4 xi 1947, Buchholz 1347 (paratypes of Podocarpus palustris ILL, 2 sheets (ILL00010017 and ILL00010018)–images, isoparatypes of Podocarpus palustris K, NY, P–image, RSA–image, S–database; photo of trees in situ, A); ibid., same date, Buchholz 1348 (paratype of Podocarpus palustris ILL–image (ILL00010016), isoparatypes of Podocarpus palustris GOET–image, K, NY, P–image, RSA–image, S–database, WIS–image); ibid., 22 xi 1947, Buchholz 1421 (holotype of Podocarpus palustris ILL–image (ILL00010021), isotypes of Podocarpus palustris ILL–images (ILL00010019, ILL00010020), K, MO–image, NY, P–image, RSA–image, S–image, TEX–image; photo of seeds, A); Buchholz 1474 (paratypes of Podocarpus palustris ILL–2 sheets (ILL00010015 and ILL00010022)–images [photo of latter specimen at A], isoparatype of Podocarpus palustris S–database (photo)]; ibid., 14 ii 1948, Buchholz 1705 (K, seedling, P–image, 2 seedlings; S–database). Guépyville, radier du 22 km, 150 m, 27 v 1948, Bernier 251 (P–image, K–database). Route de Yaté, km 22, viii 1949, Sarlin 73 (P–image). Route Bon Secour, 22 km ck., R[ivière] Blanche, 200 m, 19 vii 1954, Ingle I.44 (L–image, MEL–n.v.). 22 km station, 10 vii 1957, de Laubenfels P112 (SBT–n.v., fide de Laubenfels, 1969: 347; S–database). Plaine des Lacs, Mine Madeleine, 140 m or less, 22 ii 1948, Buchholz 1729 (epitype NY, designated here; isoepitype ILL–n.v. [photos A, P; A example with somewhat different text on label from others], K, MO–n.v. [fide TROPICOS], S–database). Plaine des Lacs, on road from Nouméa to Yaté, 10 vii 1956, “this specimen is from same tree as that from which Buchholz obtained his type”, Foster 200 (topotype of Podocarpus palustris, P–image). Plaine des Lacs, village Anna Madeleine, delta du Dacrydium n.sp., 230 m, 7 iii 1948, Bernier 246 (P–image); ibid., juvenile, 7 iii 1948, Bernier s.n. (P–image); ibid., 6 vi 1948, Bernier 249 (P–image, K–database, Z–database). Madeleine, Plaine des Lacs, 30 iii 1951, Guillaumin & Baumann-Bodenheim 11749 (P–image, Z–database); ibid., 30 iii 1951, Guillaumin & Baumann-Bodenheim 11811 (P–image, Z–database). Plaine des Lacs, Rivière des Lacs, la Chute, 239 m, 13 iii 1964, Staufer & Blanchon 5807 (L–image, P–image, K–database, NOU–database, Z–database); Plaine des Lacs – Bords du lac de Yaté (sous ligne haute tension après riv. Madeleine) – la Chute, 13 iii 1964, Blanchon 736 (P–image, NOU–database). Along banks of La Madeleine (Rivière des Lacs), S of Nouméa–Yaté road, near bridge, 23 xi 1979, McPherson 2132 (P–image, K–database, NOU–database); ibid., beside Chute de la Madeleine [22°13′S 166°50′E], 29 v 1980, McPherson 2748 (P–image, NOU–database); riverbanks below Chutes de la Madeleine, 10 viii 1981, Stone 14836 (P–image). Rivière des Lacs (rive droite) 5 km en Aval de la Chute, 200 m, 29 vi 1986, MacKee 43169 (CANB–database, P–image, K–database, NOU–database, Z–database). Forêt “Mois de mai”, Riv. blanche, 22 vii 1951, Baumann-Bodenheim 13923 (P–image, Z–database). Serpentine area, Rivière Blanche c.30 miles E of Nouméa on road to Yaté, Plaine des Lacs, 600 ft [183 m], 21 vii 1952, McMillan 5120 (A, E, L–image, P–image, K–database). Ex-radier Anna-Mad.[eleine] (locality given on composite label on sheet 245/4), Bernier 245/2 (male; P–image, P00188174). Plaine des Lacs, Anna Madeleine, 150 m, 22 i 1967, Aymard (hb. MacKee 16325) (P–image). Anna-Madeleine, rivière [22°09′S 166°54′E], 100 m, 23 xi 1979, Hartley 15064 (CANB–database, L–image, NOU–database). La Madeleine, 8 xii 1981, Dickison 223 (NOU–database). Rivière des Lacs (La Madeleine) en amont des chutes, 20 ix 2007, Dagostini et al. 1340 (P–image, NOU–database); Trois Bras [22°09′S 166°54′E, c.320 m], 18 ix 1980, Suprin 693 (NOU–database). Near bridge over ‘River des Lacs’ [sic] on road to Yaté, 21–22 ii 1948, Buchholz 1719 (P–image, packet only, its contents not visible; seeds similar to Buchholz 1729, fide label); Plaine des Lacs, Campement 22, 5 vii 1949, Selling 78 (S–database); Plaine des Lacs, i 1914, Franc 207 (Z–database: specimens seen of this number have all been from a different locality, Prony); Plaine des Lacs, Rivière des Lacs, 6 x 1924, Däniker 228 (Z–database); ibid., 11 x 1924, Däniker 228 (P–image, Z–database); Plaine des Lacs, 8 ii 1926, Däniker 228a (Z–database); Plaine des Lacs, Rivière des Lacs, Plaine des Lacs, rades de la Rivière des Lacs, c.100 m, 6 vi 1948, Bernier 250 (P–image); Plaine des Lacs, c.250 m, 14 xii 1949, MacDaniels 2544 (GH); Vallée des Lacs, 5 x 1950, Guillaumin & Baumann-Bodenheim 6511 (Z–database, 2 sheets); ibid., pont, 3 x 1950, Guillaumin & Baumann-Bodenheim 6580 (P–image, Z–database, 2 sheets); ibid., 5 x 1950, Guillaumin & Baumann-Bodenheim 6582 (P–image, juvenile; Z–database); ibid., 5 x 1950, Guillaumin & Baumann-Bodenheim 6594 (P–image; Z–database, 2 sheets); ibid., 7 x 1950, Guillaumin & Baumann-Bodenheim 6766 (P–image, Z–database). Rivière des lacs, route de Yaté, c.100 m, Bernier 245/2 (P–image). Plaine des Lacs, route de Yaté, near new bridge over Madeleine River, 5 xi 1959, Thorne 28565 (L–image, P–image). Bassin de la Rivière Yaté, Laverie Lafleur du 22ème km, 19 viii 1957, Service des Eaux et Forêts, Nouméa 225 (S–database); au bord de la Rivière des Lacs, au S du pont, 150 m, 9 vii 1958, Hürlimann 3113 (Z–database); Rivière des Lacs, 22°10′50′′S 166°50′56′′E, 200 m, 1 iii 1999, Cretinon & Gardner, ICCP New Caledonia Expedition 1999: 26 (E). Plaine des Lacs, bords de la Riv. à Chute, 25 vi 1963, Blanchon 208 (P–image, NOU–database); Plaine des Lacs, 29 vi 1965, Aubréville & Heine 130 (P–image); Tournée des Lacs, 3 vii 1965, Aubréville & Heine 170 (P–image). Süd-Bai, zwischen der Bai N'Go und Touaourou, viii–xii 1903, Rohrdorf 178 (Z–database); Touaourou, 350 m, x 1903, Cribs 1752 (P–image; highest altitude known for this species). Vallée de Toémo (entre Goro et Port-Boisé), 30 m, 4 v 1990, Cherrier (hb. MacKee 44895) (P–image, NOU–database as Retrophyllum comptonii). Bord de la Kuébini, 14 ii 1984, Lauri & Gay 165 (P–image, NOU–database). Rivière des Kaoris, Baie des Requins, 30 m, 7 iii 1948, Bernier 247 (P–image). Cascade de N'Goro, 50 m, 7 v 1948, Bernier 248 (2 sheets, P–images). Marais Kiki, 28 ix 1950, Baumann-Bodenheim 6370 (Z–database, 2 sheets); Marais Kiki, Plaine des Lacs, 150 m, 22 x 1954, MacKee 1118 (A, E); ibid., MacKee 1119 (E, P–image); bassin de la Rivière Yaté, Marais “Kiki”, 19 viii 1957, Service des Eaux & Forêts, Nouméa 226 (S–database, spirit coll.); au bord du Marais Kiki, 148 m, 17 vii 1958, Hürlimann 3157 (Z–database). Near forks of Yaté river, 3 xii 1957, de Laubenfels P160 (SBT–s.n., fide de Laubenfels, 1969: 347; S–database). Along Riv. des Lacs near beginning of Route du Carénage, 22°10′S 166°50′E, 150 m, 11 xii 1973, Webster 19205 (GH, P–image, NOU–database). Carènage – Creek Pernod [22°12′S 166°49′E, c.190 m], 1 xii 1964, Blanchon 1160 (P–image, NOU–database). Plaine des Lacs, Crique Pernod, 16 xi 1950, Guillaumin 8339 (P–image, S–database, Z–database); ibid., 16 xi 1950, Guillaumin 8345 (P–image, Z–database). Plaine des Lacs–Carénage road at crossing of Creek Pernod, 200 m, 15 xi 1955, MacKee 3377 (A, L–image, P–image; E, dated 15 xi 1953). Rivière des Lacs, near the junction with the Pernod Rau., c.160–170 m, 24 ix 1963, Green 1187 (A, P–image, K–database); Creek Pernod [22°12′S 166°49′E], 15 i 1970, Jaffré 352 (P–image, NOU–database); along road to Plaine des Lacs, at crossing of Creek Pernod, 22°11′01′′S 166°50′39′′E, 5 xii 1996, Lowry II et al. 4683 (MO–database, P–image, mis-determined as Decussocarpus comptonii); Pernod Creek, 4 xi 1984, Page 22031 (E = no. 22032 at ORSTOM); ibid., Page 22034 (E = no. 22032 at ORSTOM: some larger leaves lanceolate); ibid., Page 22033 (E); ibid., Page 22030 (E). Rivière des Lacs, 100 m before the fork in the river which divides into Creek Pernod, 22°10′45′′S 166°50′50′′E, 130 m, 29 xi 2002, Gardner et al., Third New Caledonia Araucaria Exped. TNCA 5000 (E); ibid., Gardner et al., Third New Caledonia Araucaria Exped. TNCA 5001 (E). Lake Suprin, 22°17.247′S 166°59.504′E, 238 m (Hope, 2015); Lake Boulet, 22°16.973′S 166°58.587′E, 222 m (Hope, 2015).

USA (Cultivated). Florida: Miami: Montgomery Botanical Center (Coral Gables), cultivated in nursery, 20 xii 2009, J. Silba B-654 (NY–image).

Bioregion: New Guinea & Melanesia

Ecoregion: AA0113 New Caledonia Rain Forest.

Ecology

Retrophyllum minus as delimited here, and by de Laubenfels (1972), is exclusively a low-altitude species. It grows along lake and shallow river margins in areas subject to seasonal flooding, from near sea level up to 350 m but typically below 170(–250) m, very often as a swamp tree. The observation in the protologue, “Habite dans la Nouvelle-Calédonie, au sommet de très-hautes montagnes. . .” presumably applies to material of Retrophyllum comptonii, which does inhabit mountain summit habitats as well as lowlands. Before the separation of Podocarpus comptonii from P. minor by Buchholz (1949), the name Podocarpus minor was applied indiscriminately to New Caledonian plants of Retrophyllum from both lowland swamps and high mountains. Retrophyllum minus grows in single-species stands or with other tall shrubs, such as Dacrydium araucarioides Brongn. & Gris. Retrophyllum comptonii also occurs at some localities, especially in the Plaine des Lacs. Plants from lowland swamp habitats (true Podocarpus minor, = Retrophyllum minus) were often given the informal designation “type lacustre”, especially on labels of specimens collected by Bernier, while plants later segregated as Podocarpus comptonii (= Retrophyllum comptonii) were labelled “type sylvestre”. This species has been regarded as a rheophyte by van Steenis (1981: 168), and its seeds float, aided by their porous epimatium (Carlquist, 1974: 223). Jaffré (1988) described an association in which the two dominant species were Retrophyllum minus and Dacrydium guillauminii J.Buchholz, which are otherwise two of the rarest plant species of New Caledonia. Retrophyllum minus was, according to Jaffré (1988), found on unstable soil in small coves in the 1- to 5-m-wide alluvial zone along the riverbanks. Other associates include Cloezia aquarum (Guillaumin) J.W.Dawson, Homalium kanaliense Briq., Melaleuca brongniartii F.Muell., Neocallitropsis pancheri (Carrière) de Laub., Pancheria communis Baker f. and Xanthostemon aurantiacus Schltr. Fruiting periods of Retrophyllum minus tend to coincide with periods of high water levels (Cornu et al., 2001).

Analysis of late Quaternary sediment cores from Lac Xere Wapo in the eastern Plaine des Lacs region has revealed the presence of Retrophyllum minus pollen and wood as far back as 120,000 years ago, but that there have been some marked fluctuations in abundance; the species was abundant in the earliest zone, XW-4 (120,000–90,000 years ago), but there was an abrupt rise and fall in abundance in zone XW-3 (90,000–80,000 years ago) and a dramatic decline in zone XW-2 (80,000–45,000 years ago) that is countered by a rise in Araucaria pollen. In the most recent zone (XW-1, 45,000 years ago to present), Retrophyllum again becomes one of the dominant species but Dacrydium is the most dominant (Stevenson & Hope, 2005).

Chromosome number: 2n = 20 (Hair & Beuzenberg, 1958, two separate counts as Podocarpus minor and P. palustris)

Mycological associations

See under Retrophyllum comptonii.

Conservation status: EN B1ab(iii,v)+2ab(iii,v); C2a(i) (Thomas, 2010b)

At the Madeleine site, which has been subjected to much trampling by the public, who have used the rare trees as diving platforms (Province Sud, 2012), there has been extensive planting since 1999 of Retrophyllum minus (1200 trees: Province Sud, 2012) and other rare gymnosperms, such as its associates Dacrydium guillauminii and Neocallitropsis pancheri. Germination success of Retrophyllum minus there has been very good (Cornu et al., 2001; Province Sud, 2012).

Uses

No uses of Retrophyllum minus are reported in recent literature. Sebert (1874), however, commented that, among the New Caledonian Podocarpaceae, Podocarpus minor was “l'espèce la plus interessante sous le rapport industriel”; this opinion of course also embraced Retrophyllum comptonii, which at that time was included in a broad concept of Podocarpus minor as previously stated.

Group C. Bark exfoliating as large plates. Vegetative shoots monomorphic. Terminal buds naked. Adult leaves distichous and heterofacially flattened in 2 ranks, relatively thin. Resin canals (1)3(5) median plus 2–6 lateral ones round leaf margin. Pollen cones in racemose inflorescences subtended by leaves, the raceme branches opposite, individual cones ± pedicellate, subtended by leaf-like bracts; microsporophylls spirally arranged, lanceolate. Female cones borne on specialised lateral reproductive shoots, sylleptic; peduncle broadened distally, in its vegetative portion bearing small spreading bracts that may or may not resemble reduced leaves; cone axis with 6–8 sterile bracts below fertile bract. Fertile bract with its basal half firmly adnate to epimatium, navicular. Receptacle formed from fertile bract and last sterile bract, cylindrical. Seed not beaked at micropylar end. Germination hypogeal. Seedling phyllotaxis spiral. – South America.

5. Retrophyllum rospigliosii (Pilg.) C.N. Page, Notes Roy. Bot. Gard. Edinburgh 45: 380 (22 Feb. 1989 [‘1988’]). – Podocarpus rospigliosii Pilg., Notizbl. Bot. Gart. Berlin 8: 273 (1 Feb. 1923). – Decussocarpus rospigliosii (Pilg.) de Laub., J. Arnold Arbor. 50: 347 (15 Jul. 1969). – Nageia rospigliosii (Pilg.) de Laub., Blumea 32(1): 211 (3 Feb. 1987). – Type: Peru, Oxapampa, N. Esposto 556 (holo B–extant, images seen; photos of holo at A, F [negative #11579, and a second duplicate photo of the B holotype mounted on a herbarium sheet], GH, NY); iso S (2 sheets: SBT W 430–image, S–Regnell 7799–image, both fragments ex Berlin) and S spirit collection–database). Farjon (2010: 941) indicated that the holotype was at USM (Universidad Nacional Mayor de San Marcos, Peru) but, if there is a duplicate there, it is highly improbable that Pilger saw it and, if it exists, it will be another isotype. Figs 1C, F, J, M, 7A–H, 9D, E.

Fig. 7. Retrophyllum rospigliosii. A, Branching arrangement (Pennington 1433). B, Juvenile foliage shoot (Roldán et al. 3168). C, Adult male shoot (Pennington 1433). D, Young male cones (Pennington 1433). E, Mature male cone (Pennington 1433). F, Microsporophyll and microsporangia, adaxial view (Carr 15666). G, Microsporophyll and microsporangia, abaxial view (Pennington 1433). H, Adult female branch (Bunting 4939). Magnifications: A, B, × 0.67; C, H, × 1.5; D, E, × 3; F, G, × 15. Scale bars: A, B, 6 cm; C, H, 2 cm; D, E, 1 cm; F, G, 1 mm. Drawn by Claire Banks.

?Torreya bogotensis [Linden, Cat. n. 25 (1870) et ex Thurb. in Bull. Torrey Bot. Club 1: 7 (1870) nom. inval.]; Linden ex K.Koch, Wochenschr. Vereines Beförd. Gartenbaues Königl. Preuss. Staaten 14(25): 199 (1871), nom. utique rej. (Mill, 2010; Brummitt, 2011: 1207). – Type not indicated.

Podocarpus montanus sensu Knuth in Fedde, Repert. Spec. Nov. Regni Veg. Beih. 43: 95 (1926) non (Humb. & Bonpl. ex Willd.) Lodd. (1836).

Etymology

Podocarpus rospigliosii is named after Dr Carlos Rospigliosi Vigil, who led the first expedition sent out in May 1918 by the newly founded Museo de Historia Natural in Lima, of which Rospigliosi was the inaugural director, and in which both Nicolás Esposto (collector of the type of P. rospigliosii) and Rospigliosi participated. In 1922, Rospigliosi sent specimens of Podocarpus rospigliosii to Pilger in Berlin, who described it as new some months later.

Vernacular names: pino, pino hayuelo, pino de montaña, pino colombiano (Colombia: Villamizar & Suevara 63, UDBC), pino hembrito (Colombia, Santander: Cardenas & Oliveros 198, COL–database), chaquiro crespo and chaquiro colorado (Colombia: Vásquez Correa, 2010), pino negro (Colombia: Villamizar & Suevara 63, UDBC), diablo fuerte (Cajamarca, Peru: Vicuña-Miñano & León, 2003), pino romerón, pino silvestre (throughout Colombia); pino de Pacho (Cundinamarca, Colombia); romerillo macho (Becerra-Montalvo & Zevallos-Pollito, 2014), romerillo blanco, diablo fuerte, alcumano, ulcumano, utcumanu (NW Peru: Vicuña-Miñano & León, 2003; Vicuña-Miñano, 2005), pino criollo (Mérida, Venezuela: Aranguren & Márquez, 2011)

Some of the above, such as diablo fuerte and ulcumanu, are shared with other species of Podocarpaceae.

Distinguishing features

Retrophyllum rospigliosii is distinguishable from the only other currently known South American species of its genus, R. piresii, by its more lanceolate leaves that are acute to shortly acuminate and frequently abruptly narrowed just below the tip, and by its more shortly ellipsoid to subglobose female cones that are narrowed at the base. The foliage shoots are at all stages arranged ± alternately, whereas in Retrophyllum piresii the adult shoots (the only state known) are opposite. The peduncle of the female cone bears reduced leaves, not scales as in Retrophyllum piresii. Several of these characters were not noted when the original description of Retrophyllum piresii was drawn up and serve as additional justification for the recognition of two separate species. Eckenwalder (2009), however, reduced Retrophyllum piresii to synonymy within R. rospigliosii on the grounds that, in his opinion, the specimen known to him fell within the range of variation of the latter species. Retrophyllum rospigliosii can be distinguished from the Fijian R. vitiense and the New Guinea R. filicifolium (the only other species in the genus that have all leaves not imbricate) by its bark exfoliating in large plates (not small pieces), the naked buds (in R. vitiense and R. filicifolium protected by bud scales), the foliage shoots lacking scale leaves, the blade of the microsporophyll longer than the microsporangia (not equalling or shorter than them), the more acute female cone peduncle bracts that lack even a narrow scarious margin, the much more pronounced zigzag arrangement of the decurrent leaf bases (which in fresh material are bright green and contrast strongly with the brownish colour of the rest of the branchlet) and the ellipsoid to subglobose (not obovoid) female cones. Prumnopitys montana (Humb. & Bonpl. ex Willd.) de Laub., which has in the past been confused at least once with Retrophyllum rospigliosii (Knuth, 1926: see synonymy), has spirally arranged leaves that appear alternate-distichous and the shoot flattening is homofacial, not heterofacial, so that the leaves along both sides of the shoots are turned with their abaxial surfaces uppermost; the laminas are also almost all obovate, rather than lanceolate as in R. rospigliosii.

Evergreen tree 20–45 m; dbh to 250 cm (Vicuña-Miñano & León, 2003). Trunk of mature trees buttressed at base, unbranched for up to 28 m although branching in young trees begins at c.3 m. Bark of trunk smooth, bluish-black, exfoliating in large plates; wood reddish-yellow. Crown ovoid or umbrella-shaped, much branched. Branchlets of last two orders erecto-patent, straight or slightly curved; ultimate branchlets usually alternate, adult ones 25–140(–210) mm with 5–32(–50) pairs of leaves, juvenile ones up to 270 mm (perhaps more) with up to at least 40 pairs of leaves. Leaf scars absent. Buds naked. Leaves on last two orders of branching all of one type (scale leaves absent). Leaves on penultimate axes soon caducous except when subtending branch bases or when branch does not develop further branching, arranged in 2 stepped opposite-distichous parallel ranks; decurrent bases crossing over to opposite side of branchlet and forming a green zigzag pattern that contrasts strongly with the dark brown bark when fresh; lamina larger than those of ultimate shoots, ± ovate-elliptic. Juvenile leaves of ultimate shoots unequally amphistomatic, 4–5 mm apart, opposite-distichous, pectinate, subsessile; lamina horizontally spreading, mid to deep green, narrowly lanceolate or lanceolate, 15–23 × 3–6 mm, straight, coriaceous, initially dorsally keeled when unfolding but both surfaces becoming flat, with vein-like striae between rows of stomata, midrib slightly raised on both surfaces, apex subacute or obtuse, base obtuse. Adult leaves equally or unequally amphistomatic, 3–5 mm apart throughout on most foliage shoots but up to 10 mm apart on leader shoots and sometimes in lower 1/3 of foliage shoots, opposite-distichous, pectinate, not imbricate, subsessile (free part of petiole 0.3–0.5 mm); lamina spreading and forming a shallow ‘V’ along branch, erecto-patent, diverging at 45–60(–70)°, mid to deep green, elliptic, lanceolate, ovate, ovate-elliptic, obovate-elliptic or obovate, 7–14 × 3–5 mm, straight, thinly coriaceous, the margin only slightly thickened, both surfaces flat with vein-like striae alternating with the stomatal rows, midrib only slightly raised but distinct on both surfaces, the blade frequently (but not always) abruptly narrowed distally below the acute or shortly blunt-acuminate apex that is never mucronate, the base obtuse or cuneate, the petiole decurrent along the whole length of the internode.

Pollen cones subtended by a foliage leaf, lateral in leaf axils of current growth, solitary or in groups or primary clusters of up to 3, often with an opposite pair of cones below; common peduncle present when cones in clusters but otherwise absent, individual cones shortly pedicellate and subtended by a pair of bracteoles; peduncle (when present) with 4–6 decussate, ovate, acute scales; cones greenish brown when young turning pinkish brown, ovoid when young becoming cylindrical, straight or curved from base, 8–12 mm when shedding pollen; microsporophylls 28–36, decussate in 7–9 whorls of 4, lamina longer than microsporangia, triangular-lanceolate, narrowly lanceolate or lanceolate, with entire or finely crenulate whitish hyaline scarious margin, the apex acute and incurved; microsporangia reniform; pollen snow-white or cream.

Female cones borne on current year's growth, terminal on specialised lateral reproductive shoots bearing reduced, ovate, acute leaves. Peduncle deflexed-patent to pendulous, its lower portion bearing 1 or 2 pairs of ovate-elliptic, leaf-like peduncle bracts and passing into the cone axis which bears 3 or 4 pairs of decussate, elliptic to obovate leaf-like sterile bracts (sometimes with interspersed small scales) with acute apex and glaucous on adaxial surfaces; apical part of cone axis (above last pair of sterile bracts) forming an indistinct cylindrical receptacle that becomes violet when ripe and is often shed with the cone together with the fertile bract. Fertile bract initially erect, with its basal part adnate to epimatium, distal part navicular and free. Female cone shortly ellipsoid to subglobose, 16–30 × 12–18 mm (the seed proper c.15 × 11 mm); epimatium violet and ± glaucous especially when not fully ripe, somewhat asymmetrical with the adaxial side rather more convex than the abaxial, with a narrowed base and very short conical distal crest that sometimes breaks off and with a straight, conical beak at micropylar end; micropylar beak with two short prongs that point slightly obliquely (towards the abaxial side) downwards towards the last sterile bract; cone resiniferous within; seed proper rounded at the distal end, narrowed to a conical spine at the micropylar end.

Germination apparently hypogeal (first foliage shoot emerging at soil level, cotyledons not visible). Seedlings with alternate-spiral phyllotaxis, the main axis with relatively long internodes and rather widely spaced, alternate-spiral leaves similar to those of the foliage shoots; foliage shoots with 12–25 pairs of leaves (the lower ones with fewer pairs than the upper), the leaves in 4 ranks (2 sets of 2) that are flattened distichously to form a shallow ‘V’ along the shoot with the largest leaves near the middle of the shoot.

Notes

Adult foliage shoot length (and consequently number of leaves) is very variable, being dependent at least partly on position on the tree and reproductive condition; individual ultimate foliage shoots in the upper or fertile regions of the tree are much shorter than those lower down (which also have longer leaves than those higher up). Neill & Salinas 7202 and Morales et al. 867, both from Ecuador, are atypical in having adult foliage leaves diverging from the axis mostly at angles greater than 65°, with many almost at a right angle especially in the former specimen. The adult foliage shoots of Neill & Salinas 7202 are also unusually long (to > 20 cm, with up to 50 pairs of leaves), whereas those of Morales et al. 867 are much shorter. The habitat and altitude (tropical forest at 600–675 m) of these two gatherings are also atypical for Retrophyllum rospigliosii. The habitat and foliage angle of both specimens are more characteristic of Retrophyllum piresii, which has not been recorded from Ecuador. However, Neill & Salinas 7202 has the alternate branching and lanceolate, acute leaves that are typical of Retrophyllum rospigliosii and consequently it is assumed to be a slightly divergent form of that species. In Morales et al. 867 most branches are opposite as in Retrophyllum piresii, but a few are alternate. Another specimen [Peru. Puno: Río Candamo, fila at mouth of Río Guacamayo, ridge top forest with cloud forest aspects, 13°30′S 69°50′W, 870 m, 28 v 1992, Gentry et al. 77310 (MO)] is also problematic. It is again from a rather low altitude for Retrophyllum rospigliosii, and the branches are opposite as in R. piresii. Unfortunately, it is sterile so it is not known whether it would have the distinctive long ellipsoid female cones that are so characteristic of Retrophyllum piresii. It was determined as the latter species [on the field label, as Nageia piresii (Silba) de Laub.] by de Laubenfels in 1994, which remains the current determination, but its true status remains uncertain.

Distribution

Northern and Western South America. W Venezuela (Táchira, Mérida, Trujillo), E Colombia (Antioquia, Cundinamarca, Magdalena, Norte de Santander, Santander del Sur), Ecuador (Napo, one collection; also see Jørgensen & Léon-Jánez, 1999), NW & C Peru (Gray & Buchholz, 1948: 120; Torres-Romero, 1988; Brako & Zarucchi, 1993: 2; Vicuña-Miñano, 2005: 286), NE Bolivia (Dpto. La Paz: Prov. Bautista Saavedra and Prov. Franz Tamayo, 3 collections: Zenteno-Ruiz, 2007 and more recently collected specimens cited below). Andean Region (Northern Andean Province). TDWG: 82 VEN 83 BOL CLM ECU PER. Map: Fig. 8.

Fig. 8. Global distribution of Retrophyllum rospigliosii.

Specimens seen and other records. Venezuela. Mérida: otherwise unloc., 7000 ft [2135 m], 1846, Funck & Schlim 1208 (P–image); 1928, Pittier 12756 (S–database); [distributed] 1959, Bernardi s.n. (K). Campo Elias: Los Quebraditos, above Jají, 2590 m, 21 iv 1944, Steyermark 55999 (K, MO, NY, S–database). Bosque de San Eusebio, 2100–2400 m, 30 x 1962, Ruiz Terán 1148 (K). La Azulita: Environs de La Carbonera, 2000–2500 m, ix 1952, Humbert 25588 (NY–database, previously determined as Podocarpus harmsianus and Prumnopitys harmsiana). La Carbonera, 2250 m, 30 i 1953, Bernardi 335 (NY); La Cabonera [sic], 15 km NW of Ejido, 2200 m, 19 vi 1953, Little Jr. 15245 (NY–database). La Carbonera, ± 25 km W. of Mérida, 2400 m, 22 iv 1965, Breteler 4495 (COL–photo, MO, NY, RB–image, S–database). La Carbonera, 17 ii 1972, Dobson 1277 (CONN–database). La Carbonera, 2400 m, 6 i 1980, de Laubenfels 754 (A, male and bark; K–database, L–image, male and bark); ibid., 2400 m, 6 i 1980, de Laubenfels 755 (A, K–database, L–image); ibid., 2400 m, 6 i 1980, de Laubenfels 756 (A, L–image, female; K–database); La Azulita road, 50 km from Mérida, 1700 m, 11 iii 1980, Sobel & Strudwick 2130 (NY). Rio Capaz, Via Azulita, 20 vii 1951, Curran 2147 (NY). Libertador: Environs de Mérida, 1600–2200 m, 14 ix 1952, Humbert 26148 (P–image, seedlings). Santos Marquina: La Mucuy, 18 x 1955, García-Barriga 15490 (COL–image and photos). Táchira: Jáuregui: Savana Grande between La Grita and Yeguines, 1911, Jahn 99 (S–database). Junín: Municipia Delicias, Caserío Villa Páez y sus alrededores, 1900–2300 m, 6 iii 1962, Ruiz Terán 918 (K, MO). Entre Villa Páez y Betania, cerca de la frontera de Colombia, 2000–2400 m, 15 xi 1975, Bunting 4898 (NY); ibid., 2050–2350 m, 11 x 1976, Bunting 4939 (NY, K–database); faldas debajo del Páramo de Tamá, cerca de la frontera Colombo-Venezolana, arriba de Betania y Tamá, cerca de la Quebrada Buena Vista, 2300–2450 m, 22–24 v 1967, Steyermark et al. 98706 (S–database); Trujillo: Boconó: P. N. Guaramacal, alrededores de la Laguna de los Cedros, 1830 m, 9 i 2004, Stergios et al. 20755 (E, F–n.v., MER–n.v., MERF–n.v., MO–n.v., PORT–n.v., NY–n.v., US–database (barcode 00772960)–n.v., VEN–n.v.).

Colombia. Unloc., 1828–1890, Hb. Triana 665 (BM). Cundinamarca [Prov. Gualivá]: Albán: Albán, 1962, Goitia s.n. (UDBC 10797–image and photos). Vereda Santa Ana, finca San Pablo, 1800 m, 1 ix 1985, Schmidt-Mumm 324 (COL–photo). San Francisco: Cordillera Oriental, Finca « El Carmero » El Tablazo entre Subachoque y San Francisco, 1900–2100 m, 26 i 1944, García-Barriga 11038 (COL–photo, MA–n.v.); ibid., 26 i 1944, García-Barriga 11041 (COL–image). Vereda “San Miguel”, 2100 m, 18 iv 1983, Barrera 50 (COL–photo). Sasaima: Estación San Bernardo, entre Sasaima y Albán, 1700–1800 m, 2–5 viii 1945, Dugand & Jaramillo 3962 (COL–photo). Sasaima, vereda San Bernardo, carretera a Villeta, entre los kilometros 66–70, 1860 m, 30 vi 1945, Garcia-Barriga 11586 (COL–photo). Sasaima, vereda de Apocentos (antiqua hacienda de Apocentos), 10–12 vii 1960, García-Barriga 17259 (NY, COL–photo). Sasaima, San Bernardo, quebrada La María y Río Dulce, 1750–1950 m, 20–30 xi 1962, García-Barriga 17584 (NY). Sasaima, San Bernardo, quebrada La María y Río Dulce, 1750–1950 m, 20–30 xi 1962, García-Barriga 17585 (BM, COL–photo). Sasaima, Vereda San Bernardo, orillas del Río Dulce, 1780 m, 31 xii 1962, García-Barriga 18014 (COL–image and photos). Sasaima, Hacienda San Pablo, 1800–1900 m, 1949, Schneider 905 (S–database). Sasaima, Vereda las Mercedes, 1200 m, García-Barriga 20159 (COL–n.v.). Sasaima, Vereda de Guane, entre los ríos Gualibá y Guane, Finca La Concepción, 8–11 i 1974, García-Barriga 20468 (COL–image). Vergara: Mpio. de San Francisco [sic], Vergara, 2 ii 1981, Barrera 45 (COL–photo), 46 (COL–photo). [Prov. Guavio] Guasca: Guasca, i 1936, Pérez-Arbeláez 4798 (COL–photo). [Prov. Rionegro] Pacho: Mpio. Pacho, otherwise unloc., 2000 m, ix 1962, Goitia s.n. (UDBC 2060–images and photos); ibid., 1965, Mahecha s.n. (UDBC 5436–image and photos). Pacho, 18 x 1934, Pérez-Arbeláez 3150 (COL–2 sheets, photos). Pacho, Vereda Patasía, Hacienda Patasía, 1650–1820 m, 4 viii 1947, García-Barriga 12492 (COL–photo, K, NY). Pacho, Hacienda de Patasía, 1800 m, 21 vi 1948, Uribe Uribe 1734 (COL–photo). Pacho, 1859 m, 1 vi 1944, Bohorquez R. 475 (COL–photo). [Prov. Tequendama] Tena: Laguna de Pedro Palo, finca La Laguna, por carreteable desde Patio Bonito, 7 iii 1988, Franco et al. 2430 (COL–photo). [Prov. Ubaté] Ubaté: Mpio. Ubaté, without precise locality, Mahecha s.n. (UDBC 11190–image and photos). Antioquia: [Subreg. Norte] Angostura: Vereda El Guásimo, Finca El Guásimo, 2312 m, 10 viii 2007, A.M. Vásquez 3 (MEDEL–database). [Subreg. Oriente] Guatapé: Vereda Santa Rita, Finca Vertedero (Empresas Publicas de Medellín), 1967 m, 22 viii 2007, A.M. Vásquez 5 (MEDEL–database). [Subreg. Suroeste] Fredonia: Fredonia, 1850 m, 2 viii 1930, Archer 541 (US–n.v., photo NY; PH–database, S–database; seed separate, with Soukup 1801 from Peru, see below). Jardín: Vereda El Salado, Río Docató, Finca Las Manguitas, 5°33.1′N 75°53.0′W, 1500 m, 11 iv 1990, Betancur et al.1817 (COL–photo, MO, NY). [Subreg. Valle de Aburra] Envigado: Vereda Santa Catalina, 2050 m, 7 viii 2007, Vélez-Puerta 1999 (MEDEL–database). Medellin: El Poblado, Carrera 49, primera entrada a EAFIT, 1530 m, 14 xii 2005, Vélez 6946 (MEDEL–3 sheets, database). Huila: [Subreg. Subsur] Acevedo: Parque Nacional Natural Cueva de los Guácharos, centro de potrero de Leonidas, 2090 m, 11 ix 1979, Henao 237 (COL–photo). Magdalena: Santa Marta: Sierra Nevada de Santa Marta, Cerro San Lorenzo, 1850 m, 11 vii 1968, Villamizar & Guevara 63 (UDBC 2223–image and photos); ibid., 2100 m, 11 ix 1965, Espinal-T. & Delgado-F. 1790 (COL–photo). Quindio: Salento: Barsinal, 1970 m, i 1853, Triana 1800 (BM). Pereira: Acosta: Parque Regional Natural Ucumari, Sectores La Pastora y Los Chorros, 2200 m, 9 ii 2000, Roldán et al. 3168 (MO). Risaralda: Santa Rosa de Cabal: Vereda La Suiza, Sector Riobarbo, Finca Lisbrán, 1841 m, 10 xii 2007, Rodríguez-C. et al. 505 (COL–image). Norte de Santander: [Subreg. Occidente] Abrego: Linea divisoria entre los Deptos. Santander del Norte y Cesar, entre Abrego y Las Jurisdicciones (Cerro de Croque), 3440–3750 m, 22–23 v 1969, García-Barriga & Jaramillo-Mejía 19866 (MA–n.v., NY, COL–photo; assigned to Norte de Santander by Torres-Romero, 1988, and Vargas Rincón, 2011). [Subreg. Sur-oriente] Toledo: Hoya de Samaria, 2000–2100 m, 29, 30 x 1941, Cuatrecasas et al. 12768 (COL–photo; cited by Vargas Rincón, 2011, as Decussocarpus rospigliosii). Santander del Sur: [Prov. Comunera] Gámbita: Vereda de Guausa, margen izquierdo carretera, 2070 m, 28 vi 1999, Cardenas & Oliveros 198 (COL–image). [Prov. Soto] Piedecuesta: Piedecuesta, Cristales en bosquete, 2100 m, 15 xi 1989, Roncancio 23 (UDBC 2915–image and photos); ibid., 12 xii 1989, Roncancio 024 (COL–photo). Tona: Tona, 2295 m, 20 xi 1989, Roncancio 055 (COL–photo); Tona, Vereda Guarumales, finca La Plazuela, 7°09′00′′N 72°59′29′′W, 2288 m, 13 xi 2003, Quirós Quirós 75 (COL–2 sheets, image and photos).

Ecuador. Napo: El Chaco: 4 km al E de Lumbaqui, 1 km al E del puente sobre Río Aguarico, arriba del Río Aguarico, 0°05′N 77°20′W, 600 m, 5 viii 1986, Neill & Salinas 7202 (NY, K–database, MO–database; atypical, see comments). Sucumbios: Gonzalo Pizarro: Lumbaqui, derecho de Vi OCP, 0°04′16′′N 77°18′00′′W, 675 m, 29 v 2002, Morales et al. 867 (MO, atypical, see above). Zamora-Chinchipe: Chinchipe: Parque Nacional Podocarpus, La Esmeralda (Coperativa San Francisco de Numbala Alto), 4°22′S 79°03′W, 2250 m, i 1995, Palacios & Tirado 13026 (MO). Región de la Cordillera del Cóndor, sector sur, Parroquia San Francisco de Vergel, Cuenca alta del Río Vergel, Pica Sangola, 4°44′09′′S 78°56′59′′W, 2200 m, 16 iii 2005, Quizhpe et al. 1092 (MO); ibid., cerca a La Canela, Sendero a Las Palmales, 4°35′44′′S 78°59′08′′W, 2190 m, 5 iii 2007, Quizhpe & Wisum 2461 (MO–database). Road south of Podocarpus National Park and Tapichalaca Reserve, c.4 km N of Valladolid, Río Valladolid, upper Río Chinchipe watershed, 4°30′43′′S 79°07′24′′W, 2150 m, 3 xii 2006, Neill & Jost 15337 (MO–database). Palanda: Camino paralelo a quebrada Flor Amarilla, cruce sobre la quebrada San Francisco, limite con el Parque Nacional Podocarpus, 4°00′00′′S 79°13′00′′W, 1900–2200 m, i 1995, Palacios & Tirado 13159 (MO–database).

Peru. Cajamarca: San Ignacio: Chirinos: La Palma, 10 km NW de Chirinos, 5°25′S 78°53′W, 1780 m, 5 ii 1988, Gentry et al. 61207 (MO). Huarango: El Triumfo-Convento, 5°13′S 78°40′W, 1000–1400 m, 30 vi 1996, Campos & Rodríguez 2825 (MO). San José de Lurdes: Buenos Aires, 4°55′S 78°51′W, 1870 m, 28 x 1995, R. Vásquez & Jaramillo 20459 (MO). Pasco: Oxapampa:Huancabamba: Parque Nacional Yanachaga-Chemillén, Parcela Permanente 1 ha Oso Playa, 10°17′58′′S 75°36′35′′W, 2200 m, 31 xi [sic] 2006, Monteagudo et al. 12942 (MO–image); ibid., 2200 m, same date [with same error!], Monteagudo et al. 12986 (MO–image); ibid., 3 xi 2006, Monteagudo et al. 13078 (MO–image); ibid., 6 xi 2006, Monteagudo et al. 13230 (MO–image); ibid., arbol no. 403, 11 xi 2006, Monteagudo et al. 13353 (MO–image, NY–database). Parque Nacional Yanachaga-Chemillén, remedición de parcela Oso-Playa 1.0 ha, 10°17′58′′S 75°36′56′′W, 2200 m, 20 vi 2008, Monteagudo et al. 16407 (MO–database). Oxapampa: Oxapampa, Esposto 556 (holo B–image, iso GH–image, NY–image, fragment of holo S). Oxapampa, otherwise unloc., viii 1942, Soukup 1801 (US–n.v.; image NY; sterile shoot; loose seed is annotated, “this seed belongs in packet with W.A. Archer 541, Fredonia, Colombia”; S–database). Oxapampa, R.T. Pennington 1422 (E); ibid., R.T. Pennington 1426 (E); ibid., R.T. Pennington 1427 (E); ibid., R.T. Pennington 1431 (E); ibid., R.T. Pennington 1433 (E). Chorobamba, 10°33′S 75°33′W, 1700 m, 5 vii 1985, D.N. Smith & Brack E. 7634 (NY, F–image, MO–n.v.; cultivated). Palcazú: Comunidad Nativa Santa Rosa de Palcazú, sector Santo Domingo, 10°26′50′′S 75°03′25′′W, 470 m, 13 x 2008, Huamán et al. 313 (MO–database, cultivated). Chontabamba: 2 de Mayo, Hacienda La Victoria, 10°35′S 75°28′W, 1890 m, 28 v 1982, D. Smith et al. 1752 (NY). Villa Rica, viii 1955, Soukup 4401 (GH). Villa Rica, 10°43′S 75°16′W, Chacras above town, 1500 m, 21 xi 1998, Daza & T.D. Pennington 16464 (E). Centro Bocaz, camino y trocha a Purus, 10°38′S 75°11′W, 1590 m, 19 ix 2003, Perea 381 et al. (NY–database). Base of road towards Pajonal/Villa Rica, near Oxapampa, 10°38′05′′S 75°20′16′′W, 2093 m, 17 ix 2001, Weigend 5795 et al. (NY–database). Junín: Junín:Junín: Caserio de Pampas Raimondi, 11°05′S 76°0′W, 2040 m, 23 xi 1973, Espejo 29ELE (K–database). Ucayali: Coronel Portillo:Iparia: falda dentro las cuencas del Rio Ariapo y Rio Iparia, afluentes del Rio Ucuyali, Reserva Comunal el Sira, 1550–1600 m, 9°27.85′S 74°33.95′W, 1 xi 2009, Graham 5179 (E, MOL–image).

Bolivia. La Paz: Franz Tamayo: Parque Nacional Madidi (ANMI), Sector Macheineo, Camino viejo de la carretera de Mohima a Pata, 14°39′18′′S 68°34′30′′W, 1873 m, 27 v 2011, Gardner et al. (2011): 30 (E). Parque Nacional Madidi (ANMI), Camino hacia el Rio Tuichi, 14°36′52′′S 68°33′53′′W, 1811 m, 29 v 2011, Gardner et al. (2011): 37 (E). Parque Nacional Madidi (ANMI), Santa Cruz de Valle Ameno, 14°38′23′′S 68°31′30′′W, 1605 m, cultivated, 29 v 2011, Gardner et al. (2001): 38 (E). Santa Cruz: Andrés Ibáñez: “Western slopes of the Andes”, in swampy places about Santa Cruz, 0°–40°S, vii 1860, Pearce s.n. (BM).

Colombia (cultivated). Cundinimarca: Bogotá: Bogotá, Parque de La Independencia; jardines y parques, 2620 m, 20 iii – 20 iv 1946, Duque-Jaramillo 2964 (COL–photo); ibid., Duque-Jaramillo 3108 (COL–photo, NY). Santafé de Bogotá, Parque Nacional, 2650 m, i 1963, Goitia UDBC 2129 (UDBC–image and photos). Avenida de los Cerros Orientales, 2700 m, 1 xi 1973, Mahecha 1391 (UDBC 5985) (UDBC–image and photos). Bogotá, D.E. Calle 74 con carrera 11, 8 vii 1979, E. Forero & J. Forero G. 6207 (NY, COL–photo). [Prov. Sabana Occidente] Subachoque: Vereda Casablanca, finca “California”, 2700 m, plantada junto a una cerca, como ornamental, 12 vi 2003, Hernández Schmidt 1190 (COL–photo). Vereda Tobal, 2800 m, cultivada en un jardín, 9 xi 2003, Hernández Schmidt 1384 (COL–photo). Quindio: Calarcá (introduced): Jardin Botánico del Quindio, Sede principal, Sector del Sendero de los Caobos, 1490 m, 1 iv 1999, Matallana Tobón 523 (COL–photo). Armenia (introduced): City of Armenia, Parque Aborigen, 1550 m, planted as an ornamental, 7 x 1993, Wright 047 (E–3 sheets).

Peru (cultivated).Lima: Lima, botanic garden, Asplund s.n. (S–database); ibid., 1940, Asplund 12027 (S–database).

Scotland.Midlothian: (Cultivated, origin Venezuela, cuttings from Arnold Arboretum via J. Silba). Royal Botanic Garden Edinburgh, accession 19951955*A, Frachon 1337 (E). (Cultivated, origin Colombia, R. Nicholson 348/92). Royal Botanic Garden Edinburgh, accession 20001721*A, Frachon 1340 (E).

Bioregions: Northern Andes, Central Andes

Ecoregions: NT0105 Bolivian Yungas, NT0109 Cauca Valley Montane Forests, NT0118 Cordillera Oriental Montane Forests, NT0121 Eastern Cordillera Real Montane Forests, NT0136 Magdalena Valley Montane Forests (in which it and Podocarpus oleifolius D.Don are two of the most dominant trees), NT0153 Peruvian Yungas, NT0159 Santa Marta Montane Forests, NT0175 Venezuelan Andes Montane Forests (where again it is one of the most dominant trees along with Prumnopitys montana and Podocarpus oleifolius together with assorted dicot trees).

Ecology

Montane forests, cloud forest; (600–)1500–3750 m (generally at higher altitudes than Retrophyllum piresii; the three records from below 1000 m relate to the three atypical specimens from Ecuador and Peru discussed above, all of which have some features of Retrophyllum piresii). Retrophyllum rospigliosii grows best on slightly sloping ground, fertile river beds and small depressions, on wet, deep, fertile acid clay or clay/sand soils with good to slow drainage; it is not tolerant of either swampy or drought conditions. It requires constant humidity and cloudy conditions with an annual rainfall of 1500–2500 mm and an annual average temperature of 10–18°C (Nieto & Rodriguez, 2003). It plays an important role in intercepting rainfall: a Retrophyllum rospigliosii plantation at La Mucuy (Mérida, Venezuela) intercepted 42% of precipitation, almost the same percentage as the 45% intercepted by nearby montane cloud forest (Ataroff, 2002). The seeds of Retrophyllum rospigliosii are recalcitrant, with a viability in storage at 12°C of 3–6 months but with a drastic drop in viability after 4 months (Ceballos-Freire & López-Ríos, 2007).

In Cajamarca, Peru, Retrophyllum rospigliosii occurs in undisturbed and fragmented primary forest together with Cecropia polystachya Trécul, Cinnamomum triplinerve (Ruiz & Pav.) Kosterm., Heliocarpus americanus L., Ladenbergia magnifolia (Ruiz & Pav.) Klotzsch, Otoba parvifolia (Markr.) A.H.Gentry, Piptocoma discolor (Kunth) Pruski, Ruagea insignis (C.DC.) T.D.Penn., Virola calophylla Warb., etc. (Vicuña-Miñano & León, 2003).

Phytochemistry

Ninety-one compounds have been isolated and identified from the essential oil of the ‘fruits’ of Retrophyllum rospigliosii (Quijano-Celis et al., 2010). The essential oil from the leaves yielded 68 compounds (Quijano-Celis et al., 2013). In both cases, limonene (38% in ‘fruits’, 43% in leaves) and α-pinene (16% in ‘fruits’, 18% in leaves) comprised the largest fractions. Retrophyllum rospigliosii contains several norditerpene dilactones that were previously known only in Podocarpus and/or Nageia, including nagilactones E, F and G, β-sitosterol and a derivative of the latter (Amaro-Luis & Carroz, 1988), as well as a unique compound, rospiglioside (Amaro Luis et al., 2006). As well as rospiglioside, Amaro Luis et al. (2006) isolated ferruginol, sugiol, sugiol acetate, totarol, totarol acetate, 4β-carboxy-19-nortotarol and 16-hydroxy-4β-carboxy-19-nortotarol from the leaves of Retrophyllum rospigliosii, but the phenolics of the wood do not appear to have been studied. Its biflavones are similar to those found elsewhere in Podocarpaceae and include sciadopitysin, podocarpusflavone A, podocarpusflavone B, amentoflavone, 7,4′,7′′,4′′′-tetra-O-methyl-amentoflavone, 7,4′,7′′-tri-O-methyl-amentoflavone, 7,7′′-di-O-methyl-amentoflavone, sequoiaflavone and heveaflavone (Chaabi et al., 2007; Amaro-Luis et al., 2008). The biflavones act as phosphodiesterase inhibitors and may therefore have potential as therapeutic drugs for a wide variety of medical conditions (Rahimi et al., 2010 and work cited therein). However, the biflavones and other compounds of Retrophyllum rospigliosii assayed were found to lack any antimycotic activity against three pathogenic fungi (Niño et al., 2007).

Plant–animal interactions

In Colombia, the seeds of Retrophyllum rospigliosii are eaten by the endangered yellow-eared parrot or loro orejiamarillo, Ognorhynchus icterotis (Massena & Souanc) (Botero-Delgadillo & Páez, 2011). Whether this activity effects dispersal is not known.

Retrophyllum rospigliosii is the one of the hosts of at least three members of the wood-boring beetle family Scolytidae, namely Araptus impensus (Wood), Platypus secus Wood (Wood & Bright, 1992, as Podocarpus ‘raspigliosii’) and Thamnophthorus impensus Wood (Schedl, 1965). Some of the paratypes of Platypus secus were collected from Retrophyllum rospigliosii logs at the beetle's type locality in La Carbonera Experimental Forest, Mérida, Venezuela (Wood, 1971, as Podocarpus ‘raspigliosii’).

Mycological associations

Retrophyllum rospigliosii is the only known host of Tripospora venezuelensis E.Müll. (Coryneliales), which is known only from the type collection from Mérida, Venezuela (Müller & Dennis, 1965; Benny et al., 1985). The mycorrhiza of Retrophyllum rospigliosii have been investigated by Furman (1970) and Guerrero Forero & Hodson de Jaramillo (1988). The latter team found that Retrophyllum rospigliosii had a mixture of vesicular arbuscular mycorrhiza that included several species of Acaulospora Gerdemann & Trappe (Acaulosporaceae) and Glomus Tul. & C.Tul. (Glomeraceae).

Conservation status (global): VU (A2acd) (Gardner & Thomas, 2013)

Logging of this valuable timber tree has reduced or fragmented most of the formerly rather extensive stands of this species. Whole mountainsides of the species in Peru were clear-felled in the 1980s for timber, and the species is there now reduced to scattered individuals (M. Gardner, Royal Botanic Garden Edinburgh, pers. comm., 7 Feb. 2012). Yaguana et al. (2012) have commented that the forest at Numbala (Ecuador), at the edge of Podocarpus National Park, is one of the last remnants of its type. Retrophyllum rospigliosii is regarded as NT in Colombia (Barragan Bedoya & Valdés Torres, 2011).

Uses

The fruits are reported as edible (Gray & Buchholz, 1948; Buchholz & Gray, 1957; and several specimen labels). The tree is often cultivated in gardens and parks in Colombia (Nieto & Rodriguez, 2003) and Peru (Brako & Zarucchi, 1993), and is used in forestry in Colombia (Nieto & Rodriguez, 2003, who also give details of recommended practices for seed harvesting and silviculture) as well as in Venezuela (Ruiz Terán 918 in sched., K, MO). The wood is of very good quality and is widely used in construction in Venezuela and Colombia. It is easily worked and is used to make furniture, boxes, cabinets, veneer, pencils, paper pulp and other items (Nieto & Rodriguez, 2003). It is also used to make handcrafted decorative objects (Feuillet Hurtado et al., 2011). Its characteristics have been described recently by Vásquez Correa & Alcántara Vara (2009) and Vásquez Correa et al. (2010). The logging cycle for the species was estimated at 109 years by Becerra-Montalvo & Zevallos-Pollito (2014).

6. Retrophyllum piresii (Silba) C.N. Page, Notes Roy. Bot. Gard. Edinburgh 45: 380 (22 Feb. 1989 [‘1988’]). – Decussocarpus piresii Silba, Phytologia 54: 461, t. 1 (10 Dec. 1983). – Nageia piresii (Silba) de Laub., Blumea 32(1): 211 (3 Feb. 1987). – Type: Brazil: Território de Rondônia, Serra Pacas Novos, Seringal S. Luiz, 50 km east of Rio Pacas Novos, 14 viii 1976, N.A. Rosa, J.M. Pires, W. Rodriques, S. Barros 856 [misprinted in protologue as ‘Rosa & Pires 586’] (holo US, iso K–database, L–image, MG–image, MO, NY, R–2 sheets, images, RB–image). Fig. 9A–C.

Fig. 9. Retrophyllum piresii. A, Branching arrangement (Rosa et al. 856). B, Adult female branch (Rosa et al. 856). C, Mature female cone (Rosa et al. 856). Retrophyllum rospigliosii. D, Young female cone (Vasquez et al. 20459). E, Mature female cone (Bunting 4939). Magnifications: A, B, × 0.67; C, × 1.5; D, E, × 3. Scale bars: A, 6 cm; B, 2 cm; C, D, E, 2 cm. Drawn by Claire Banks.

Etymology

Retrophyllum piresii is named after João Murça Pires (1917–1994), Brazilian botanist, who published (with C. Mainieri) a revision of Podocarpus in Brazil (Silvic. Sao Paulo 8: 1–24, 1973) and who was one of those who collected both the type and the only two other currently known specimens of wild origin, made a few days earlier and on the same day.

Vernacular name: pinheiro-da-amazónia (Secco et al., 2009)

Distinguishing features

This species shares many characters with its South American congener Retrophyllum rospigliosii but, as was noted in the protologue, it can be distinguished by its elliptic (not lanceolate) leaves with ± parallel sides, and the more ellipsoid female cones that are not narrowed at the base. Additional characters not described by Silba that distinguish it from Retrophyllum rospigliosii are the bracts on the female peduncle all very reduced and scale-like (not foliar), the obtuse leaf apices (misleadingly described in the protologue as ‘bluntly acute’) and the leaves never abruptly narrowed just below the apex (normally so in Retrophyllum rospigliosii). The opposite branching with leaves on penultimate branches opposite and not forming stepped ranks also separates Retrophyllum piresii from the vast majority of specimens of R. rospigliosii, although a few specimens of the latter approach R. piresii in these features, as noted above. Further differences may become apparent when more material is collected (male and juvenile collections of Retrophyllum piresii are particularly desirable to complete the description). The two species also appear to be ecologically separated, with Retrophyllum piresii inhabiting lowland swamp forest at altitudes of less than 1000 m and R. rospigliosii normally occurring in montane and cloud forests, nearly always above 1500 m, although there are one or two low-altitude records as discussed under that species.

Evergreen tree c.30 m tall and up to 2.5 m dbh, adult tree branching at c.15 m above base. Trunk (fide Farjon, 2010) straight and erect, in large trees lacking branches for 15 m or more. Bark (imperfectly known) medium to dark brown outside, purplish to dark brown inside, smooth on branchlets; wood reddish brown. Crown rounded (fide Farjon, 2010: 940). Branchlets of last two orders erecto-patent, slightly curved; ultimate branchlets usually opposite, 50–130 mm with 24–38 pairs of leaves, penultimate ones 180–210 mm. Leaf scars absent. Buds obconical, naked. Leaves on last two orders of branching all of one type (scale-leaves absent). Leaves on penultimate axes normally soon caducous although persistent when subtending branch bases and when axis has not developed further branching, opposite and spreading or orientated so that both leaves of each pair are suberect, not forming stepped ranks. Juvenile foliage not known. Adult foliage leaves unequally amphistomatic, 4–5 mm apart, opposite-distichous, pectinate, not imbricate, subsessile (petiole 0.2–0.5 mm); lamina spreading and forming a shallow ‘V’ along branch, erecto-patent, diverging at 65–80°, mid to deep green, elliptic, oblong-elliptic or obovate-elliptic (never lanceolate), 9.5–11 × 2–2.5 mm in middle of shoot and 5–6 × 1.7–2.5 mm near base, straight, thinly coriaceous, the margin only slightly thickened, both surfaces flat with vein-like striae alternating with the stomatal rows, midrib only slightly raised on both surfaces or not at all, sometimes obscure distally on abaxial surface, the apex obtuse and the blade never abruptly narrowed below it, usually with an indistinct mucro, the base obtuse, the petiole decurrent along the whole length of the internode. Pollen cones unknown. Female cones borne on previous year's growth, terminal on specialised lateral shoots bearing very reduced scale-leaves or bracts. Peduncle c.15–16 mm, about 2/3 length of cone, deflexed-patent to pendulous, its lower portion bearing at least 3 pairs of decussate scale-like, shortly ovate or elliptic bracts, its apical part together with the fertile bract forming an indistinct, cylindrical receptacle that is shed with the cone. Fertile bract with basal part adnate to epimatium, distal part free and navicular. Female cone with epimatium pecan-brown turning reddish when mature, 19–22 × 11–14 mm, distinctly ellipsoid with a shortly conical distal crest, not narrowed at the base, the seed proper slightly smaller.

Taxonomic notes

Silba's protologue, and later description (Silba, 1986: 78), are both very brief and give little detail concerning many aspects of its morphology, but no other taxonomic account of this species appears to have been published until now except that of Farjon (2010), which added some details of trunk and habit. Further collections of the species from the type locality are badly needed, particularly of the unknown juvenile stages and male cones.

Distribution

W Brazil (SW Rondônia: Serra dos Pacaás Novos), near the border with lowland NE Bolivia. More recently recorded from Bolivia and Peru in similar habitats (D.J. de Laubenfels, pers. comm., 24 October 2005). These records require confirmation; one problematic sterile specimen, determined as Retrophyllum piresii by de Laubenfels, is discussed above under R. rospigliosii. For the time being, the species is regarded as endemic to Brazil and known from the two wild gatherings (both female) and one sterile cultivated collection cited here. Map: Fig. 10.

Fig. 10. Global distribution of Retrophyllum piresii.

Specimens seen. Brazil. Rondônia: Guajará-Mirim: Serra Pacaás Novos, Seringal S. Luiz, 50 km E of Rio Pacas Novos, 14 viii 1976, Rosa et al. 856 (holo US, iso K, L–image, MG–image, MO, NY, R–image, RB–image); Serra Pacas Novos, 11°13′S 63°35′W, 250 m, 14 viii 1976, Rosa et al. 856 (iso INPA–image with different label information as indicated). Serra dos Pacos [sic] Novos, 30 min. de vôo de helicopter de S. Luz–Rio Pacos Novos, 11°15′S 23°50′W [sic: presumably an error for 63°50′W], 7 viii 1976, Pires 850 & Rosa (RB–2 sheets both numbered RB253293, images). Região do rio Pacaas Novos, Projeto RADAM/BRASIL, ponto 15 à 250–300 m de altitude, 11°13′S e 63°35′W, 14 viii 1976, Rodrigues et al. 9646 (INPA–2 sheets, images; female; third collector listed as ‘S.N. Rosas’ but probably a misprint for N. Rosa). Brazil (cultivated). Pará: Belém: Belém, Museu Emilio Goeldi, Parque Zoobotânico Emílio Goeldi, procedente de Rondônia, 30 x 2008, Kinupp 3426 (RB–image; also photos of living plant).

Bioregion

Amazonia. Ecoregion: NT0135 Madeira-Tapajós moist forests; perhaps also in NT0166 Southwest Amazon moist forests.

Ecology

Lowland tropical swamp forest; 250–300 m.

Phenology

Reputedly fertile in January (de Oliveira, 2012).

Conservation status (global, IUCN 3.1): DD (two independent assessments: de Oliveira, 2012; Gardner, 2013)

The only confirmed locality from which wild material has been gathered is within a National Park. Insufficient is known about the extent and area of occurrence, or of potential threats, to award an assessment other than DD and this will not change until further research can be done on the species in the field. The species is in cultivation (see cited specimens).

Uses

None have been reported although it is likely to have wood of high quality similar to that of R. rospigliosii.