NOTES RELATING TO THE FLORA OF BHUTAN: XXXIV.

The typification of Rivca ornuta (Roxb.) Choisy (Convolvulaceae), based on Lettsomia ornata Roxb., is discussed. R. ornata is shown to have been frequently (though not consistently) misapplied to a taxon from S India and Sri Lanka which has also been called Convolvulus candicans Roth ex Roem. & Schult.. nom. illegit.; it is here newly described as R. wightiana R.R. Mill. Lettsomia ornata was based on a specimen of a plant introduced to Calcutta Botanic Garden from Uttar Pradesh. This is synonymous with Rivca ornata var. griffithii C.B. Clarke, not Clarke's 'typical variety' which Prain formally named var. typica. Lectotypifications are proposed for two species names in Argyreia Lour. Calystegia hederacea Wall.. Convolvulus arvensis L. and Dichondra repens J.R. & G. Forst. are recorded from Bhutan for the first time.


Introduction
For many years it has been generally accepted that there are two species of Rivea Choisy in the Indian subcontinent: R. hypocrateriformis (Desr. ex Lam.) Choisy, and R. ornata (Roxb.) Choisy. R. hypocrateriformis, the type species of the genus (Manitz, 1976), poses no taxonomic problems. This paper deals with R. ornata and other names synonymized with it by various authors.
Two taxa have been confused within R. ornata since its basionym. Lettsomia ornata Roxb., was transferred to Rivea by Choisy (1833). One of these (which includes the type of the basionym) is an erect shrub, hardly ever developing twining habit. This occurs along the sub-Himalayan tract as far east as Myanmar (Burma) and possibly Laos and Thailand. The other taxon, from southern India and Sri Lanka, is a vigorous climber. Other differences exist and the taxa are better regarded as distinct species. The name R. ornata has often, though not consistently, been incorrectly applied and stabilization of its application is desirable. A summary of the current nomenclature and typification of taxa mentioned in this paper is given below, with examples of the differing usages of each name where applicable. Syn.: Convolvulus canclicans Roth ex Roem. & Schult., Syst. Veg. 4: 273 & 790 (1819) nom. illegit. non C. canclicans Willd. ex Roem. & Schult., op. cit. 302 (1819); Roth, Nov. PI. Sp. 106 (1821). For type see below.

Convolvulus candicans Roth ex
~ N o n C. candicans Sol. ex Sims in Bot. Mag. t. 1603Mag. t. (1813 type, from material cult. K in 1776 probably originating from America; =Ipomoea tiliacea (Willd.) Choisy); nee C. candicans Willd. ex Roem. & Schult., op. cit. 302 (1819 nom. illegit. & Ipomoea candicans Rottler mss. nom. inval.: type, Tranquebar, Klein, B-Willd. 3692, microfiche seen, and LINN-Smith 307.16, microfiche seen; renamed Lettsomia uniflora Roxb., Fl. Ind. 2: 85 (1824), =R. hypocrateriformis.) Typification of Lettsomia ornata Roxb. Roxburgh (1814) indicated that he had received material oi Lettsomia ornata from Hardwicke in 1802, from 'Hindoostan '. Thomas Hardwicke (1756-1835 was Captain of the 1st Company of the 3rd Battalion of the military service of the East India Company at Cawnpore from 1797-1803(Dawson, 1946. These dates agree with the information in Roxburgh (1814Roxburgh ( , 1824 concerning the origin of the plant. Hardwicke presumably collected the plant in the vicinity of Cawnpore from where we know he sent it to Roxburgh for cultivation . Prain (1894: 88) stated that Hardwicke 'did not, so far as can be ascertained, collect in Southern India', a crucial fact in relation to the subsequent confusion that has arisen over the application of Roxburgh's name to plants from S India. No specimen of L. ornata has yet been traced that bears Roxburgh's own handwriting. Those in K-W and G-DC bear Wallich's labels; that in K-W is undated, that in G-DC is dated 1830 but not by Wallich. Thus, I refrain from designating either of them as lectotype of L. ornata since it is possible that a specimen from the Calcutta Botanic Garden collection of L. ornata may exist elsewhere that can be more directly attributed to Roxburgh. However, Wallich 1369 (in the broad sense) has historically been considered to be the type of L. ornata. Roxburgh (1814ornata. Roxburgh ( , 1824ornata. Roxburgh ( , 1832 clearly intended the name L. ornata to apply to a plant from N India. Despite slight inconsistencies in his description as published (Roxburgh, 1824(Roxburgh, , 1832, it is equally clear that he intended the name to apply to a shrub which, if it twined at all, only did so occasionally at the ultimate branches. Roxburgh's plant was not a vigorous climber although many later authors, beginning with Wight (1848), have misapplied the name to such a plant.
The Roxburgh drawing at K corresponding to his manuscript description is numbered 1535 and labelled Ipomoea ornata (Sealy, 1957: 354). This combination has never been validly published. Roxburgh (or possibly Carey when editing Roxburgh's manuscript for publication) clearly changed his mind as the species was eventually published as Lettsomia ornata Roxb. Sealy (1957) stated that Roxburgh's drawings were numbered consecutively according to his Flora Indica manuscript and sent to Kew at intervals from 1796 onwards; no. 1463 was sent between 1803 and 1804 (Sealy,op. cit. 300 '. Most of these appear to be field records; only a single gathering representing them, Wallich 1369 1 localized 'Hetaunra', has been traced (K-W; G-DC, microfiche seen). In his footnote. Wallich mentioned a climbing, twining habit, and indicated that the plant was synonymous with what he called 'Convolvulus candicans Roth'. The K-W example of Wallich 1369/1, labelled 'Lettsomia ornata Roxb. Legi ad Hetounra Decemb. 1820', has been examined. It (and the G-DC example) agrees exactly with the description given in Wallich's footnote. Wallich's specimens are mature fruiting plants; they do not exactly match the plant cultivated from Hardwicke's introduction (Wallich 1369/2) although they appear to be conspecific and were so treated by Prain (1894). However, Wallich 1369/1 was collected in 1820, many years after Roxburgh prepared his manuscript description and also several years after Roxburgh's death in 1815. Hence, Wallich 1369/1 cannot represent type material of L. ornata Roxb. Clarke (1883: 183) seems to have doubted whether Roxburgh's L. ornata was native to N India, for he wrote, 'Roxburgh says he obtained seeds of this from Cawnpore (possibly from a garden?); no examples have been seen from the South Deccan '. Roxburgh (1824) made no mention of garden origin for Hardwicke's collection, so Clarke could not have known if the plant had come from a garden (rather than being subsequently planted at Calcutta). Garden origin of Hardwicke's plants of L. ornata is here considered unlikely; Hardwicke was an enthusiastic naturalist and made many botanical drawings 'on the spot' in the field (Britten, 1906). Without definite information, it is impossible to discount the possibility of garden origin, although Prain (1896) records that plants identical to Roxburgh's original concept of L. ornata were subsequently collected over a wide area of the sub-Himalayan tract.

Convolvulus candicans Roth
The nomenclature and typification of C. candicans Roth ex Roem. & Schult. is summarized above. Benjamin Heyne, a German Moravian missionary, went to Tranquebar in 1792 and was superintendent of the Bangalore Gardens from 1802 1808 (Stafleu & Cowan, 1979). Tranquebar (Tarangambudi, Tamil Nadu) and Bangalore (southeastern Karnataka) are both in the southern Deccan; most, if not all, of Heyne's collections were made in this region, rather than in N India where Hardwicke's plants of L. omata had been collected. All the sheets labelled C. candicans in K-W belong to R. hypocrateriformis

Subsequent widening of the application ofR. ornata (Roxb.) Choisy
It has been established above that the origin of Hardwicke's plants that served as the basis for the type collection of L. ornata was Uttar Pradesh and that the plant had a shrubby, virtually non-twining habit. However, the name has been frequently misapplied to a different, strongly climbing plant from the Deccan and Sri Lanka. Some of this confusion can be attributed to Wallich's editorial footnote to Roxburgh's description, which mentioned C.  Roem. & Schult.), was described as a twiner ('caule volubili'). Later authors no doubt confused the two taxa and in so doing attributed a climbing or twining habit to C. candicans Roth ex Roem. & Schult. and thus to R. omata (Roxb.) Choisy with which it was generally synonymized.
[correctly (Roxb.) Wall.] and Argyreia omata [Roxb.] Sweet, he included two other elements: Convolvulus candicans Roth, and herb. Wallich 2253, named Convolvulus tarita. C. ornatus sensu Choisy embraced both Wallich 1369/2 (type of L. omata Roxb.) and Wallich 1369 1 (Nepal) as both are in G-DC. C. tarita belongs to R. hypocrateriformis, as discussed below: hence it must be excluded from any concept of R. ornata. Choisy appears to have been influenced when drawing up his description by Roth's C. candicans (and Wallich's footnote saying that the latter was supposedly the same as R. ornata). His transfer of L. omata Roxb. to Rivea, together with the other included elements, created a confused taxon. in effect " R. omata sensu Choisy non L. ornata Roxb. '. R. ornata sensu Choisy, including plants from both north and south India, has been similarly treated by the following authors among others: Don (1838: 254); Voigt (1845: 351); Drury (1866: 295); Woodrow (1898: 170); Vickery in Hara et al. (1982: 108);Naithani (1985: 448); Haridasan & Rao (1987: 645). Drury (lac. cit.) stated that the plant was a climber; thus, the Deccan plant formed the basis of his species concept. Woodrow's concept is not known as his 1898 checklist lacked descriptions, but he was dealing with the flora of western India mainly around Mahabaleshwar and thus the plant known to him would have been the Deccan taxon; however he followed Clarke's nomenclature (1883) in its inclusive sense. Naithani (1985: 448) described R. ornata as a "climbing shrub"; his description is otherwise too brief and inadequate to identify the taxon he was dealing with, but his work was concerned with an area at the western end of the sub-Himalayan tract and thus he would have been dealing with R. ornata in the sense Roxburgh used the basionym.

Convolvulus tarita Buck.-Ham.
Convolvulus tarita Buch.-Ham. ex Wall., Numer. List 67, no. 2253(1830, is a nomen nudum. Prain (1894: 89) remarked that there was no material of its "type". Buchanan-Hamilton in herb. Wallich 2253, at Calcutta. The specimen which should be in K-W under this Wallich number is also missing (IDC undated; confirmed by present author. 1992). There is at E a specimen, originally presented to the University of Edinburgh in 1830 with Buchanan-Hamilton's herbarium, which bears two labels. Of these, the lower is one of Francis (Buchanan) Hamilton's partly printed labels (the handwriting is probably that of an amanuensis), numbered 515. with the name 'Convolvulus Tarita', and "Tarita Sanscritar. Habitat in Magadhar dumetis"; the other, immediately above and entirely handwritten, reads 'Convolvulus Tarita. Mongger 10th Sept. 1811'. Thus, this specimen is virtually certain to be original material of C. tarita. The 'Magadhar' label agrees with entry no. 515 in Hamilton's catalogue (Hamilton, 1830) which gives additional information ("Brihat Tarachandi oflicinarum Hindice") but does not mention a collection from Mongger. Another specimen (G-DC, microfiche seen: labelled 'Mongger 10th Sept. 1811") is obviously a duplicate but apart from the identical handwriting, locality and date there is nothing to link it with Buchanan-Hamilton. It is conspecific with the E specimen.  (E) has been compared with the two syntypes of R. ornata var. griffithii C.B. Clarke and other material definitely referable to that taxon. and with material of R. hypocrateriformis. All plants of var. griffithii have flowering calyces c. 14mm long; those of R. hypocvateriformis are 8-10mm long. There are also subtle, less constant differences in the shape of the leaf base: R. ornata var. griffithii most commonly has a very broadly and shallowly cordate leaf base, that of R. livpocratcriformis is more narrowly and more deeply cordate.  has leaves of the latter type, and flowering calyces 9 10mm long. It is concluded, therefore, that Convolvulus tarita is synonymous, not with R. ornata. but with R. hvpocratcriformis. This confirms Prain's suspicion, based on his field knowledge, that only one Rivea, R. hypocrateriformis, grew in Bihar (Prain. 1894: 88-89). As a nomen nudum which has never been validated, the transfer of this synonym from R. ornata to R. hypocrateriformis has no nomenclatural significance, but it is important as C. tarita clearly formed the basis of the 'Monghir' records of R. ornata in Choisy (1833R. ornata in Choisy ( , 1845. Don (1838) and Voigt (1845).

Definite misapplication qf'R. ornata to S Indian plants
Wight's account of R. ornata (Wight. 1848) very clearly refers to the Deccan plant because he stated "stems climbing". Indeed, the only specimen he cited was from "Balaghaut mountains, Madras'. This almost certainly refers to the hills in the Cuddapah district NW of Madras (formerly called Balaghaut), rather than the Palghat district of Kerala or the Balaghaut Range in Maharashtra; the only Rivea known from either of these latter areas is R. hypocrateriformis, not R. ornata or the southern taxon confused with it (cf. Vajravelu, 1990) whereas the southern ' R. ornatci has frequently been collected from the Cuddapah area (see citations below, under R. wightiana R.R. Mill). The accompanying illustration, t. 1356, is not one of the c.400 in Wight's work that were copied from Roxburgh's drawings; it may well have been made from the Balaghaut specimen. An unlocalized specimen (K) is a very close match to the illustration but is not identical to the plant drawn; it may be a duplicate of the same gathering as was used to prepare the drawing. Wight's work apparently represents the first definite (mis)application of the name R. ornata to plants from southern India and probably represents the principal source of all the later confusion, caused by other workers copying him. A list of works using the name in this incorrect sense is given above under R. ornata auct. non (Roxb.) Choisy. R. ornata Clarke (1883: 183) for the first time discriminated two taxa within Rivea ornata. Possibly because Wight had previously done so, Clarke unfortunately attached the name Rivea ornata ( Roxb.) Choisy (including its basionym. Lettsomia ornata Roxb.) to the southern taxon. From this, which he (like Wight) appeared to regard as 'typical R. ornata\ he separated the northern taxon as var. griffithii C.B. Clarke. Clarke's action means that, under the current ICBN (Greuter et al., 1994). the southern plants should, in theory, be given the autonym R. ornata var. ornatci whenever varietal rank is awarded to them. However, the type of the basionym, and thus of R. ornata var. ornata R. ornata var. typica appeared to be 'strictly confined to South India and Ceylon". The northern plant, var. griffithii C.B. Clarke, in whose synonymy Prain included Lettsomia ornata Roxb. and Argvreia ornata (Roxb.) Sweet, was cited from several localities in the sub-Himalayan region 'from the Sivaliks to the Sikkim Terai'.

Discrimination of taxa within
Prain, as he made abundantly clear in his footnote (op. cit. 88-89), had finally realized that the northern and southern plants were not only probably specifically distinct, but that the application of the specific epithet ornata had been confused. He writes, '... the plant figured by Roxburgh in his Ic. Ined.. and described by him. is not the South Indian but the sub-Himalayan plant. It is to the latter that Roxburgh's trivial name of 'ornata' ought therefore rightly to belong; but as its application to Roth's Convolvulus candicans has now become stereotyped, it will be necessary to allow the name Rivea ornata Choisy, to continue to designate the plant from Southern India, and be preferable to name the North-Indian one Rivea Roxburghii ...'. Shortly afterwards (Prain, 1896: 537), he noted that H. Haines had brought 'further large suites of this very distinct variety [R. ornata var. griffithii] from the Duars ... These make it more probable than ever that Roxburgh's plant is specifically distinct from Convolvulus candicans Roth". He stated that Haines had found it growing 'at least as far as the Assam frontier', and that King's collectors had brought back large quantities of the same taxon from the Southern Shan States (Myanmar). He noted that, 'From no part of its wide area, (from the Sivaliks to the Shan Hills) come any specimens showing the slightest tendency to connect Roxburgh's plant with Roth's one from Southern India".
Prain's publication (1894: 89) of the name Rivea roxburghii does not constitute valid publication under Art. 34.1 of the current ICBN (Greuter et al., 1994). since it was merely a provisional name given in the future tense; it has never appeared in Index Kewensis. His taxonomic conclusions and circumscription of the taxa are. nevertheless, correct. The name R. roxburghii Prain was validated in Brandis (1906: 484) by a short English description, together with a reference back to Prain's first mention of the name, which should be cited (and added to Index Kewensis) as R. roxburghii Prain ex Brandis. Although not formally typified, it was an avowed nomen novum for Lettsomia ornata Roxb. and thus should also be typified by Wallich 1369 2. Hence R. roxburghii is also a taxonomic synonym of that part of R. ornata (Roxb.) Choisy which includes the basionym, but it is not synonymous with R. ornata in its misapplied sense.
Nevertheless, for the past century, Prain's conclusions appear to have been overlooked. Nearly all standard Indian regional Floras have continued to uncritically follow Flora of British India and either misapply Rivea ornata to the southern taxon (e.g. Watt, 1892: 558;Trimen. 1895: 205-6: Cooke, 1908Gamble, 1922 Raizada (1976) decided that the Rivea occurring in the Dehra Dun area was not the same as R. omata ( Roxb.) Choisy and referred it to R. laotica Ooststr.. described from Laos by van Ooststroom (1957). Raizada was correct in discriminating two taxa. but it appears that by R. omata (Roxb.) Choisy he actually meant the southern plant {R. omata auct. non (Roxb.) Choisy). His treatment was followed by Babu (1977). However, Deva & Jain (1979). in a critical investigation based partly on fresh, field-collected material and partly on available specimens, regarded the Dehra Dun Rivea as R. omata var. griffithii. They considered that R. laotica was synonymous with R. omata var. griffithii. and simply the latter taxon at the easternmost limit of its range. If their conclusions are correct, which is quite likely given that Prain cited collections of R. omata from the Shan States (Myanmar), the correct name at species rank for R. laotica is R. omata (Roxb.) Choisy, or, if R. omata were to be conserved with a new type to preserve historic usage. R. roxburghii Prain ex Brandis.

Lcctoiypification of R. ormila var. griffithii
Clarke evidently saw three specimens (syntypes) that he referred to his new variety: two from Griffith's Herbarium (one "from the Calcutta Botanic Garden', the other supposedly collected by Griffith in Sikkim, where he never actually went), and the third collected by Falconer. In the type cover at K there are only two sheets: Falconer's, and Griffith H.E.FC. 5854 cSikkim'). The origin ot H.E.I.C. 5854 may have been Bhutan where Griffith did collect, but no specimen with a journal ticket has been found to confirm this. The third syntype, located by the present author at K in 1992, is simply labelled • Rivea omata H.B.C. ex Hb. Griffith'. Prain (1894: 88. fn.) wrote. "The specimens in Griffith's Herbarium are from the Roxburghian plants of the Calcutta Botanic Garden'. The third syntype. with that origin stated, is very likely also an isotype of l.ettu/mia omata Roxb.. or at least a 'clonotype' of it (i.e. a specimen taken at a later date than the original type specimen, but from the same living accession at Calcutta). All three sheets bear Clarke's handwritten determination. 'Rivea omata var. griffithii'. The Griffith specimens are in fruit, while the Falconer one bears three short flowering stems (with buds but no fully open flowers) and two separate leaves. Clarke separated his variety principally on the basis of peduncle length (M-4 'Ain.\ as opposed to ll/ -.in or less' in his concept of 'typical R. omata'). leaves less silky beneath and corolla limb 'glabrous without even in the bud". The last character could only have come from the Falconer specimen, as the Griffith ones have no flowers. The peduncles uu the Falconer sheet are 1.5-4.5 inches (45-110mm) long; on both Griffith sheets the fruiting peduncles are about 1-2 inches (20-50mm) long. Although Clarke named the variety after Griffith, this should not necessarily be used as a reason to favour one specimen over the other in choosing a lectotype. Since the Falconer sheet is a closer match in peduncle length to Clarke's measurements, and also the only one showing the diagnostic glabrous corolla exterior, it, and not either of the Griffith specimens, is here selected as the lectotype of Rivea ornata var. griffithii. Prain (1894: 88) stated that Falconer's specimen was collected in the Sivaliks.
Prain transferred Lettsomia ornata Roxb. and Argyreia ornata (Roxb.) Sweet from the synonymy of his R. ornata var. typica to the synonymy of var. griffithii. Thus, from 1894 var. griffithii became a taxonomic synonym of Roxburgh's Lettsomia ornata and hence of R. ornata (and of R. roxburghii Prain ex Brandis).

Nomenclatural consequences
Taxonomically, R. ornata var. griffithii C.B. Clarke (lectotype, Falconer s.n.. K) is identical to Lettsomia ornata Roxb. (type, Roxb. in WaUich 1369/2, K-W). Thus, if the northern and southern taxa are regarded as varieties of one species, it is the northern one that should be called var. ornata (syn. var. griffithii). General usage on the other hand has been to call the southern one 'var. tvpica" or simply R. ornata, and to incorrectly include the type of the basionym in the concept of the southern taxon.
There are three possible options open to resolve the nomenclature. These have different consequences but all upset historic usage to a greater or lesser degree. They are: 1. Strictly follow Art. 10.4 and retain the existing nomenclatural type of Lettsomia ornata. WaUich 1369/2, without rejection of the name as allowed for by Art. 56.1. Nomenclatural consequences: Northern taxon remains R. ornata (Roxb.) Choisy at species rank (syn. R. roxburghii Prain ex Brandis); at varietal rank, becomes R. ornata var. ornata (syn. R. ornata var. griffithii C.B. Clarke). Southern taxon, as a species, requires a nomen novum (typified by the type of Convolvulus candicans Roth ex Roem. & Schult.) or else description as a new species, with new type; as a variety, it needs description as a new variety of R. ornata (Roxb.) Choisy. Roxb. under Art. 56.1. Nomenclatural consequences: Northern taxon at specific rank becomes R. roxburghii Prain ex Brandis (next available valid name); as a variety of the southern taxon, both a new specific epithet for the latter and a new varietal combination would be required. Southern taxon at species rank requires either a nomen novum (typified as under option 1) or a description as a species nova with a new type; at varietal rank it would have to be described as a new variety of R. roxburghii. 3. Conserve the name Rivea ornata under Art. 14.1 with a new, conserved type (different to WaUich 1369/2), chosen from S Indian material, to preserve historic usage as much as possible.

Reject Lettsomia ornata
(This option has the same effect as the action recommended by Prain, 1894: 88.) Nomenclatural consequences: Northern taxon at species rank cannot be called R. ornata (Roxb.) Choisy as Roxburgh's type is now excluded. For the same reason, R. roxburghii Prain ex Brandis also cannot be used, unless it too is conserved with a new type. As variety, becomes R. ornata Choisy var. griffiihii C.B. Clarke. Southern taxon: because Roxburgh's type is excluded, the citation R. ornata (Roxb.) Choisy becomes incorrect. The combination is treated as dating from Choisy (1833). to be cited as R. ornata Choisy. Austin's presently incorrect typification of R. ornata remains incorrect, being superseded by the new conserved type. As a variety, the southern taxon becomes R. ornata Choisy var. ornata. It is debatable whether R. ornata has been sufficiently 'persistently misapplied' for a formal conservation or rejection proposal to succeed, although one or other might help to stabilize the nomenclature. Usage has varied according to author, some restricting it to the type taxon, others widening it to include the non-type S Indian taxon as well as the type taxon, others again misapplying it to the non-type taxon (see lists of usages under each name above). R. K. Brummitt (pers. comm., 1992), after reading an early draft of this paper, considered that there seemed to be a case for making a proposal under what was then Art. 69, but that it would probably be rejected because the epithet ornata had on occasion been used in a broad sense including both northern and southern taxa. Another reviewer of a later version of the paper suggested that an Art. 14 proposal to conserve the name with a new type should be considered in order to maintain historic usage.
Article 57 requires that 'a name that has been widely and persistently used for a taxon or taxa not including its type is not to be used in a sense that conflicts with current usage unless and until a proposal to deal with it under either Art. 14.1 or 56.1 has been submitted and rejected'. As I consider that (a) specific rank is appropriate for both taxa and (b) the name R. ornata has been used sufficiently often to denote either the northern taxon only or both it and the southern taxon, as an interim measure in this paper I shall use R. ornata for the taxon containing the historically accepted nomenclatural type of Lettsomia ornata Roxb. The southern taxon requires a new name. Although it has often been referred to by the illegitimate name Convolvulus candicans Roth ex Roem. & Schult., the epithet candicans is not chosen here for use in Rivea in spite of its descriptive aptness (the stems of the southern plant are often much more densely white-tomentose than in R. ornata from N India). I consider that to continue usage of candicans would create further confusion as the epithet has been used in closely allied genera (e.g. Ipomoea) to denote other taxa (and also in the totally unrelated but nomenclaturally very confusable Theaceous plant, Lettsomia candicans (Tul.) Choisy, now classified in the genus Freziera Willd.). Instead, the epithet wightiana has been chosen, since it was Robert Wight who first definitely misapplied the name R. ornata to the Deccan plant. A proposal to conserve R. ornata with a new type under Art. 14.1 will be submitted for consideration by the Spermatophyta Committee. If this is eventually accepted, R. wightiana R.R. Mill (the southern taxon) will revert to R. ornata Choisy while R. ornata (Roxb.) Choisy in the present account will become R. roxburghii Prain ex Brandis.
(Umbelliferae). D. repens is a very distinctive member of Convolvulaceae, small and prostrate with reniform or cordate-orbicular leaves and very small axillary greenish or whitish flowers. It is widespread in the warmer regions of both hemispheres. There are numerous records from Yunnan (China) but this seems to be its first record from the eastern Himalayan region. It has not so far been recorded from Nepal or further west in the Himalaya. Nevertheless, it can be easily overlooked on account of its small size.

A C K N O W L E D G E M E N T S
I wish to thank the Directors of the Herbaria at BM and K for the loan of specimens and for facilitating the study of other material. I am also very grateful to Dr R. K.
Brummitt (Royal Botanic Gardens, Kew) for reading a draft of the portion of this paper dealing with the typification of Rivea ornatu and making valuable comments. The comments of an anonymous referee have also been of areat benefit.