BULL. BOT. .SURV. INDIA Vol. 7, Nos. 1-4, pp. 73-88, 1965

THE TRANSFER OF CYRTANDROMOEA FROM GESNERIACEAE TO SCROPHULARIACEAE, WITH NOTES ON THE CLASSIFICATIONOF THAT FAMILY

B. L. BURTT Royal Botanic Garden, Edinburgh

1. THE STRUCTURE AND AFFINITIES OF the Matang plant. The seedline raised had cotyledons CYRTANDROMOEA which were equal in size on germination and remained so throughout. Thus the doubts already felt about the In a recent attempt to classify and give keys to position of Cyrtandromoea were immediately confirmed, the Old World genera of Gesneriaceae, Cyrtandromoea .for no member of Cyrtandroideae is known to have was ten~~tivelyassigned a position in Cyrtandroideae- isocotylous seedlings (cf. Burtt, 1963, 205). Loxonieae, "for want of a better place". At the same The plants of Cyrtandromoea raised from Sarawak time attention was drawn to its resemblances to Brookea seed produced flowers in the late autumn of 1963 and Benth. and Rehmannia Libosch., two genera whose from one of them serial sections of the ovary were exclusion from Gesneriaceae was thdught justified. A prepared. The illustration (fig. 1) shows a selection thorough re-investigation of Cyrtandromoea was of somewhat diagrammatic cross-sections at different advocated (Burtt, 1963, 210, 217). levels of the ovary ; it had been taken from a flower The opportunity, one might say incentive, for, such that had been hand-pollinated but had fallen without a study has arisen from a visit paid to Malaya and fruit development. Pollen tubes are not visible in any Sarawak in the summer of 1962-after the above-quoted of the sections. The ovary is clearly bilocular through- paper was written. Cyrtandromoea grandis was found out its length ; the placentae are lamellar in form and in the Ulu Gombak near Kuala Lumpur, and another ovuliferous over the whole surface, species (possibly C. subsessilis) on Mt. Matang, near Examination of young fruits of Cyrtandromoea Kuching, Sarawak. gmndis, preserved in spirit, showed that the fruit wall is The genus has been introduced $to cultivation at thin while the placentae are fleshy (Plate I, H). These Edinburgh from a few seeds found in a broken fruit on characters' recall such genera as Fieldia (Gesnerioideae-

Fig. 1 Cyrtmdromoea subsessilis (Miq.) B. L. Burtt ? (B. 1944) ; series of trans- verse sections of ovary from apex (1) to base (12) X 59. 74 BULLETIN OF THE BOTANICAL SURVEY OF INDIA [VO~.7

Mitrarieae), though there the' ovary is unilocular. The seed structure of Cyrtandromoea also needed Although the possession of a unilocular ovary and elucidation. Endosperm is definitely present and the parietal placentae is usually given as a diagnostic feature inner and side walls of the testa have laminate thicken- of Gesneriaceae, a bilocular ovary does not absolutely ings (Plate I, J) : the outer wall remains thin and exclude a genus from thii family-it is, in fact, quite easily becomes broken, so that the testa appears to have remarkable how, many illustrations of Gesneriaceae a reticulate pitting. show, without comment, a bilocular ovary. Before considering .possible relationships between

Plate I Cyrtandromoea grandis Ridl. : A, basal portion of stem X 3 ; B, apical portion of atem X 4 ; C, corolla, dissected X 9 ;.D, gynoecium X 4 ; E, anthem, posterior and anterior views X3 4 ; F, pollen grain X 650 ; G, ovary in longitudinal &on XI* ; H, young fruit in transverse section X 3 ; 1, seed X 63 ; J, seed in section X 63. Cyrtandromoea subsessilis ? K, seedling X 4. (A-J from BUM & Woods, B 1616; K from Burtt & Woods, B 1944). 19651 BURTT : TRANSFER OF CYRTANDROMOEA FROM QESNERIACEAE TO SCROPHULARUCEAE 75

Cyrtandromoea and groups outside Gesneriaceae it dorsal corolla lobes are external in aestivation and the seemed desirable to find out whether there was an leaves opposite), and the convex lower lips of the affinity with any genera of the subfamily Gesnerioideae, bilabiate spurless corolla and the cymose inflorescence which has constantly isocotylous seedlings and seeds lead directly to the tribe Cheloneae. with endosperm. Ridley has inadvertently provided a Within Cheloneae Wettstein's key asks first oi all pointer in this direction, for the plant he described from whether the fruit is a capsule or berry and then, if a Sarawak as Cyrtandromoea minor proves to be the capsule, whether it is septicidal or loculicidal. This is Americag Chysothemis pulchella (Don) Decne., somewhat less than satisfactory ; for it separates which must have been cultivated in Kwhing. The Paulownia from Wightia, Uroskinnera from Brookea, resemblances between these two genera lie in the general Hemichaena from Leucocarpus. Cyrtandromoea can leaf shape, the tubular winged shortly toothed calyx have a more or less indehiscent thin walled berry or a and the external form of the gynoecium. The differ- weakly loculicidal capsule. If the first it is to be placed ences are manifold : Chrysothemis has bent filaments next to Leucocarpus D. Don, if the second next to and anthers cohering by their tips, a large unilateral Hemichaena Benth. Both genera are Central American nectary, bifid parietal placentae and spirally striate and monotypic, and both have been figured in Curtis's seeds with a funicular food-body. Even the resemblance Botanical Magazine : Leucocarpus in 1831 (t. 3067 as in the calyx is superficial, for the wings or ridges on Mimulus) and Hemichaena in 1875 (t. 6164). The the calyx of Chrysothemis run to the sinuses whereas best account of Leucocarpus is, however, David Don's, in Cyrtandromoea they are median to the calyx-teeth. published by Sweet (D. Don, 1831). If Cyrtandromoea belongs to Gesnerioideae, it The fruit of Leucocarpus is a white berry, but the should be in the affinity of Besleria with which it has skin is rather thin and most of the substance of the the annular disc in common. Now Besleria is a large fruit is due to the fleshy placenta. It is interesting and varied genus of some 140 species, of which very that Don made comparison with the fruit of Fieldia*. few are in cultivation, and of which even the herbarium The calyx of Leucocarpus is tubular and shortly toothed, material available to me is rather scanty. Fortunately but is not so long as in Cyrtandromoea and is not there is a fine revision of the genus by C. V. Morton accrescent after flowering, so that the fruit is not in- (1939). It appears that constant generic features of cluded within it. The ovary, I and especially the Besleria are that the anther-thecae are confluent at placentation (cf. Hartl, 1956: 204, Abb. 12), the style- their tips, that the two parietal placentae are ovuli- and bilamellate stigma are essentially the same in both ferous only on their inner faces, that the stigma is genera. In the stamens of Leucocarpus (Plate 11, E) bilobed or stomatomorphic and that the fleshy berry the anther thecae are widely divergent, and confluent has a thick verrucose exocarp and spirally striate seeds. on dehiscence, whereas they are subparallel in Cyrtan- All these features are in sharp disagreement with the dromoea and only slightly divergent at the base. The characters of Cyrtandromoea and the two genera are exserted fruit and the anthers provided the best dis- clearly not closely related. tinction between Leucocarpus and Cyrtandromoea. The only possible affinity remaining within Ges- The seeds of Leucocarpus are similar to those of nerioideae is with the tribe Mitrarieae. Here we find Cyrtandromoea in general appearance and have about a resemblance in fruit characters, for this group also the same quantity of endosperm. In sections, however, has a thin fruit wall and fleshy placenta. However they do not show the' distinctive laminated thickening the ovary of Mitrarieae is unilocular and the placentae of the inner and side walls characteristic of Cyrtandro- are parietal and bifid. In flowers, leaves and habit moea. It has not been possible to make wide enough there is little resemblance and Mitrarieae are an austral comparative studies to assess the value of this charac- pup, in contrast to Cyrtandromoea which is tropical. ter : in this affinity Schultes ( 1941) refers to. the seeds It is concluded that Cyrtandromoea has no close of Uroskinnera being enclosed in a thin membrane- afhity with other members of Gesneriaceae. * In reading this article on Leucocarpus I learned for bilafular OvV and other of Cyrtan- the fint be Don hd wi&bwn his p~viou~~ypm- &ornoen to saggat that an affinity is most likely be fa~lyMdymoearpOCCoc (bn, sa9ng found for it in the large polymorphic family Scrophu- accumte examination has fully convinced me that they must lariaceae ; in fact there is a notable superficial resem- be united to the Gcsnsriaceae, which again are hardly dis- blance to same cultivated species of ~j~~l~~.using tinguishable by MY tangible character from Scrophularinae." Thin implies, that Don recognized the family Gesncriaceat Wettstein's (1895) cla~ificationCyrtandromoea should in 182& ia accopkd publiertion only fmm 1821 be placed in subfamily Antirrhinoideae (because the (c., ~~x~~,viii, 173 : 1959). BULLETIN OF THE BOTANICAL SURVEY OF INDIA

Plate I1 Leucocarpus alatus D. Don : A, habit X 3 ; B, calyx X ; C, corolla, dissected X 3 ; D, gynoecium X 8; E, anthers, posterior and anterior views X 2d ; F, pollen grain X 650 ; G, fruit in longitudinal sectiqn X % ; H, fruit in transverse section X 4 ; I, seed X 40. (A, B, C, G & H from British Flower Garden. ser. 2(2) : t. 124, 1831). The map indicates the general arsas of Cyrtandromoea (cross-hatched) and Leucocarpus (qtraight lines). In the inset X marks districts where Cyrtandromoea is known to occur, not individual records. this may well be the thin outer walls of the cells of lobes and full yellow flower, red-spotted in the throat the testa. A similar structure is depicted by Harz (1885) like Mimulus guttatus, most easily distinguishes it. in Capsicum annuum. Mature fruits have not been available for examination Hemichaena differs from Leucocarpus in the and we may hope they, or the seeds, will yield some anther-thecae being parallel, in its capsular fruit and additional point of difference. more deeply cleft calyx. In anthers, therefore, it is Having reached the decision that Cyrtandromoea closer to Cyrtandromoea from which the deeper calyx- is to be transferred from Gesneriaceae to Scrophularia- 19651 BURTT : TRANSFER OF CYRTANDROMOEA FROM OESNERIACEAE TO SCROPHULARIACEAE 77

ceae-Scrophularioideae-Mimuleae the phytogeogra- then, with no claim to represent original research phical aspects of the change are worth comment. (though some has' been necessary). A compilation Instead of having to deal with a gesneriaceous Malay- prepared for working purposes is simply being made sian endemic of no obvious affinity, we have a scrophu- available for general use. A critical personal judgment lariaceous genus, still a Malaysian endemic but clearly on the tribes deserving acceptance has not been allied to Leucocarpus in Central America. In fact, attempted ; in each case one possible course which to use the terminology of van Steenis (1962), Cyrtan- seems convenient for purposes of discussion has. been dromoea and Leucocarpus form a tropical amphi- adopted. transpacific pair. Any addition to the list of such Without a considerable amount of bibliographic genera (Steenis, 1962, 247-253) is interesting': two research one cannot arrive at a tribal nomenclature that points make the present instance particularly so. First, is likely to be correct. A reference to an earlier (not S~ro~hulariaceaeis a family not previously represented necessarily the original) use of each name is given : on the list ; secondly this family is often thought of as the interested student must make his own way from a largely temperate one, yet here is good evidence of there. The names Scrophularioideae for the subfamily its early tropical existence. and Scrophularieae for the tribe containing the genus Recently (Burtt, 1962) I dealt with Rhyncho- Scrophularia are now dictated by the International glossurn; which is also a tropical amphi-transpacific Code of Botanical Nomenclature (Art. 19). genus. These twd examples are the only ones amongst The most generally accepted classification of the Tubiflorae. Scrophulariaceae, both in circumscription and sub- division, is still basically that represented by Bentham's 2. NOTES ON THE CLASSIFICATION OF THE later views (1876), views which .were considerably SCROPHULARIACEAE modified from those he had expressed earlier (1835 a Having established the affinity of Cyrtandromoea & b, 1846). If we look first at the circumscription of with Leucocarpus and its immediate allies we may well the family, we find that Wettstein (1895) made one ask where this little group stands in relation to the major change-the inclusion of the mainly S. African rest of the family Scrophulariaceae? Wettstein's key family Selaginaceae (excl. Globulariaceae) as a tribe led us to the correct affinity, but his tribe Cheloneae Selagineae, placed next to Manuleae. British botanists is a heterogeneous assemblage of genera and Leucocarpw have generally rejected this view (e.g. Hutchinson, has been transferred by later authors either to Gratioleae 1959 ; Faulks, 1963), but June11 (1961) has shown or to Mimuleae, if that tribe be kept distinct. that there ,are no really significant differences between At this point, therefore, a brief consideration of Selagineae and Manuleae. If we expect family differ- Scrophulariaceae as a whole is not out of place. The ences of a really fundamental nature in the Tubiflorae family is so "well-known" that its enormous variability it may be we shall end up with only one family. There- is sometimes forgotten ; furthermore its essentially tem- fore I consider that decisions can only be. taken against perate nature is often exaggerated. More widely the background of the classification of the whole group. recognized, however, is the difficulty of discriminating Extensive re-modelling may be necessary and in this between Scrophulariaceae and allied families such as Junell's work will provide valuable data.

Orobanchaceae, Gesneriaceae, Bignoniaceae. " Hans Hallier ( 1903-summarized in Diels, 1908) Discussion on interrelationships in this group of considered the classificatioi~ of Scrophulariaceae, but families can only be pursued with profit in terms of many of his ideas seemed to have been based on super- their constituent tribes. I have myself erred in talking ficial resemblance or on single characters. If Planta- about the Scrophulariaceous or Gesneriaceous nature ginaceae cannot be maintained as a separate family of such genera as Brookea, Charadrophila and (Hallier included it in Scrophulariaceae) there is little Rehmannia. Unless one suggests whereabouts in these hope for classification at this level in Gamopetalae. families a genus is to be placed such discussions are The latest significant suggestion concerning the really meaningless. In investigating Cyrtandromoea it circumscription of Scrophulariaceae has been Breme- has, therefore, been necessary to obtain an overall kamp's (1953) transfer of the subfamily Nelsonioideae- picture of the tribal classification of Scrophulariaceae. from Acanthaceae to tribal rank in Scrophulariaceae. When a number of working notes had been put to- Two points may be made before discussing Breme- gether, it seemed that other botanists might also find kamp's reasoning : first, that Bremekam~was writing them useful and that their retention in a private file as a specialist on Acanthaceae and the specialist is often was somewhat wasteful. They are published here, inclined to tie more concerned with the purity of his 78 BULLETIN OF THE BOTANICAL SURVEY OF INDIA [Vol. 7 own family than with that of the one which he is woody capsule valves that recurve on dehiscence are making the transfer : secondly that he found it im- characteristically acanthaceous. possible to fit Nelsonieae into the current classification As in the case of Selagineae, it is of less importance of Scrophulariacehe. Although he placed Nelsonieae to decide here- and now on the formal position of as a tribe next to Rhinantheae, this means abandoning Nelsonioideae than to learn as much as we can about the presently accepted Rhinanthoideae. Further dis- them, and to be able to recognize them as an entity. cussion of this point comes later. Meanwhile I wish Useful discussion of affinities is Personales must be in to examine part of Bremekamp's argument because it terms of tribes rather than of the families themselves. has considerable bearing on one's attitude to classifica- We shall shortly see, in considering the tribes of tion at tribal level. Scrophulariaceae, that a much greater bar to clear Bremekamp qu6tes Robert Brown's (1810, p. 473) discussion than the final positioning of these groups is the view that Acanthaceae are certainly a natural family problem set by the rather numerous anomalous genera. but not easily circumscribed. His comment is, "when We can now turn to the major current subdivision a group is presented to us as a natural one its members of Scrophulariaceae into subfamilies. Both Bentham must possess at least one common character for other- (1876) and Wettstein (1895) had a subfamily Pseudo- wise it is impossible to recognize its naturalness and solaneae ( = Pseudo-solanoideae) for the tribes Verbas- when it is provided with one or, preferably, more com- ceae and Leucophylleae. Pennell, h*owever, abandoned mon characters it must of course be possible to define this when he accepted the idea put forward by it". Surely these remarks apply to an artificial group, Robertson (1888) that the short-tubed, actinomorphic, held together merely by the possession of one or two or nearly actinomorphic, corolla of this group is a characters. A natural group might well be formed secondary development from a long tubular corolla. of a series of genera showing changes from one to the This is an outstanding instance of evolutionary views, next so that no obvious diagnostic character held as distinct from the discovery of any new characters, throughout. If there were good evidence of the close being allowed to change the classification. Robyns a&ty of one genus to the next, however, the group (1931) also rejected the idea that this group had any would certainly be j;dged "natural" by most taxono- close relationship to Solanaceae, largely because mists (even without any evolutionary interpretation). Verbascum and its allies lack any trace of the oblique I do not -suggest that such genera would have no orientation of the ovary so characteristic of Solanaceae. characters in common, merely 'that they might be very While both these points of view may contain a subtle and only apparent on intensive investigation, good deal of truth, they do nothing to minimize the .as it may be are those that link Nelsonioideae to the distinctness of Verbasceae and Leuc~ph~lleaefrom the rest of Acanthaceae. There is no contradiction in say- rest of Scrophulariaceae. It is well to remember that ing that a group is natural but difficult to define, and Nees (1834, p. 78) went so far as to regard ''Verbas- as an example I give the tribe Klugieae in Gesneriaceae cinae" as a distinct family. The verbascum gmup as (Burtt, 1963). The genera Rhynchoglossum (incl. a natural unit should probably exclude the thud tribe. Klugia-Burtt, 1962), Monophyllaea, Epithema, Moul- ~ptosimeae,which is composed of S. African plants tonia seem to me to form a natural group. It needs, with elongate corolla tube. however, the addition of an anatomical character- It remains to point out that the name, Pseudo- the presence of medullary vascular bundles-for its easy solanoideae, used for the Verbascum group as a sub- definition. Without that R. Brown's comment 'natural family is untenable under our current Code of Nomen- but hard to define' would certainly apply. clature. The subfamily name must be dwived from At present I cannot accept Bremekamp's transfer the name of. its. type genus. of this subfamily to Scrophulariaceae, interesting as are The distinction between s~ro~hularioideae(as the resemblances he points out. Johri &' Hardw Singh we must call the Antirrhinoideae of various authors] (1959) have come to the conclusion, in so far as and ~hinanthoideaelies in the aestivation of the corolla Elytraria may be taken as representative of Nelsonioi- lobes. To the extent that the major and obvious groups deae, that the group stands apart from the rest of of allied general fall wholiy into one subfamily or the .Acanthaceae but .does not show a true approach to other, it is clearly justified. But there are a good many ScrophuIariye. The samk conclusion has been genera in Scrophulariaceae whose affinity is uncertain, reached by Mohan Ram & Masand (1963) .from a and the po~sibilit~that these are currently placed in study of Nelsonizl campestris. I would add that the the wrong subfamily on this single character must not 19651 BURTT : TRANSFER OF CYRTANDROMOEA FROM OESNERIACEAE TO SCROPHULARUCEAE 79 be ignored. In any case there is scope for a thorough between them, is by no means constant, The Pseudo- re-study of the family to see if this division into 2 sub- solanoideae are a much more compact group but seem families can be set on a wider and surer foundation. to lack a valid name at the subfamily level. For our The aestivation of the corolla-lobes was not always present purposes there is much to be-said for omitting accurately abserved. Pennell has effected the transfer the rank of subfamily and going straight from the of Capra7ia and Sopuhia from Digitaleae to Gratioleae family down to tribes. on more accurate observation. Hart1 (1955) has Bentham's (1876) twelve tribes were : (i) Leuco- pointed out that Lindenbergia, currently placed in phylleae, (ii) Aptosimeae, (iii) Verbasceae, (iv) Cal- Scrophularioideae-Gratioleae, has rhinanthoid aestiva- ceolarieae, (v) Hemimerideae, (vi) Antirrhineae, tion. The affinity of this genus deserves closer study. (vii) . Cheloneae, (viii) Manuleae, (ix) Gratioleae, My own observations on living material of (x) Digitaleae, (xi) Gerardieae, (xii) Euphrasieae. Isoplexis canariensis indicate that the lateral lobe%are Wettstein (1895) united Leucophylleae with Verbas- overlapped by the upper qnes, AQ~vice versa as has ceae, but brought the number up to twelve again by always been said. Whether the stock at present in the inclusion of Selagineae. cultivation at Edinburgh is exceptional in this respect It is fair to say that there are three of these tribes I do not yet know. which are highly heterogeneous : they are Cheloneae Lagotis has been placed in Veroniceae by Pennell (which must on its present circumscription be called (1943) with the knowledge that the flower aestivation Scrophularieae), Gratideae and Digitaleae. The is that of the other subfamily. It is worth noting that remainder seem to be reasonably homogeneous. corolla form (in respect of proportion of upper and Bellini (1907) set out a reclassification of Scrophu- lateral lobes-the latter are notably narrow and acute) lariaceae. He called the family Personatae and recog- is comparable in Isoplexis and some species of Lagotis. nized two subfamilies, Scrophulariaceae and Rhinan- Scrofella Maxim. placed by Wettstein in Antirrhinoideae thaceae, with 12 tribes in the first and 3 in the second -Cheloneae has very shallow upper and lateral lobes (including Orobanchaceae as a tribe). Bellini's system united into an upper lip, and (at least on herbarium was based almost entirely on one set of characters material) the aestivation character is unworkable. derived from the nectaries. Thus the genera of the Scrofella should probably be placed in Veroniceae near two subfamilies normally established on aestivation are Wulfenia and Calorhabdos. Finally it is to be remem- mixed up. Although the names of these tribes are bered that Bentham himself noted that aestivation was often the same as those that were already in use, their inconstant in Rhinanthoideae-Gerardieae-Escobodinae, content mdy be very different. Mimuleae, for instance, a group he placed on its general affinities. is composed of Mirnulus, Paulownia, Maurandia, In pointing out these difficulties in the primary Lophospermum and Mmus. Most of the other tribes classification of Scrophulariaceae I am not suggesting are much reduced from their normal content. that the character is of no value ; it obviously is. Digitaleae, however, includes Scrophularia and other Fwermore all Gesneriaceae examined have rhinan- genera from Cheloneae, others from Gratioleae (reduced thoid aestivation except Monophyllaea (which is to a single genus) and even Boea and Besleria from scrophularioid) and Rhynchoglossum in which the rim Gweriaceae. Bellini may be followed in just one of the palate meets a corresponding rim on the upper particular : in the recognition of a separate tribe, lip and the lobes project outwards from 'these rims Collinseae, for Collinsia and Tonella ; this step obtains without any overlapping. Both these genera belong to confirmation from the study of the embryology of the tribe Klugieae (Burtt, 1963). In all known Acan- Collinsia by Crete (1958). thaceae-Nelsonioideae the aestivation is scrophularioid, The conspectJs of Scrophulariaceaeby Rouy (1909) though Bremekamp (1953) thinks that the affinity of has received little attention ; yet, though adding the group is with Rhinantheae. The aestivation of little to botanical knowledge, he made some important Cyrtandromoea is scrophularioid. classificatory changes and raised the number of tribes I think there is little doubt that the two conven- to 16. This he did by reviving Veroniceae (Bentham, tional subfamilies, Scrophularioideae (as we should 1835), though he attributed it to himself as a new hve to call the Antirrhinoideae) and Rhinanthoideae tribe, and by inaugurating Limoselleae, Hemiphragmeae are still too large and varied to be useful units for dis. (later also proposed by Pennell (1943), who seems to cussion and, as we have seen, the character of corolla- have been unaware of Rouy's paperj and Torenieae. aestivation which alone has been offered as a distinction , Rouy's Dodartieae includes the earlier Gratioleae Benth. 80 BULLETIN OF THE BOTANICAL SURVEY OF INDIA [VO~.7 and his Rehmannieae the earlier Digitaleae Benth., Halleria Linn. at least as to the type genus in each instance. Rouy Teedia Rudolphi. also established numerous new subtribes, whose con- Phygelius E. Mey. sideration lies beyond our present scope. Russelia Jacq.-tribe Russelieae Pennell ( 1919 ) . The most significant studies on Scrophulariaceae Freylinia Pangelli. in recent years have been by the late F. W. Pennell, Ixianthes ,&nth. though his overall contribution to the classification of Anastrabe E. Mey. the family was always restricted by the regional limita- Bowkeria Harv. tions of his individual papers. Brookea Benth. In his account of the Scrophulariaceae of the Wightia Wall.-transferred to Paulownieae. south-eastern U.S.A., Pennell (1919) established the separate tribes Paulownieae, Russelieae and Angelonieae Collinsia Nutt.-tribe Collinsieae Bellini ( 1907). The following year, in dealing with the familjr in Tonella Nutt.-qibe Collinsieae Bellini ( 1907). Colombia (Pennell, 1920, p. 142) he accepted Mimuleae Scrophularia L. as a tribe distinct from Gratioleae. Later again, in Scrofella Maxim.-to be transferred to Veroniceae? Scrophulariaceae of the Western Himalaya (Pennell, Chelone Linn. 1943), we find Mimulus once more included in Penstemon Mitch. Gratioleae : the only other tribal change in this paper Chionophila Benth. was the proposal to establish Hemiphragmeae as an Tetranema Benth. independent tribe ; but as we have already seen this Brandisia Wight-tradsferred to Paulownieae. tribe had been established by Rouy (1909). Paulownia Sieb. & Zucc.-tribe Paulownieae We are primarily concerned in this paper with Pennell (1919). two or three tribes : Cheloneae, Gratioleae and Uroskinnera Lindl. Mimuleae (if accepted as distinct). Cheloneae includes Berendtiella Wettst. & Harms (= Berendtia A. Scrobhularia and it will therefore be referred' to in Gray)-transferred to Mimuleae or Gratioleae. future by its correct name : Scrophularieae. It was Hemichaena Benth.-transferred to Mimuleae or always a rather heterogeneous assemblage of genera Gratioleae. which has been to some extent broken up since Even with the exclusions noted here, Scrophularieae Wettstein's account. It may therefore be helpful to remains a rather mixed assemblage, as has been poirited list his genera and their later treatment. There is no out recently by Thieret (1954). Perhaps the genera doubt that this tribe includes some of the most interesting with .well-developed staminodes, ~crophdaria itself, genera of the family ; many of them are woody plants, Penstemon, Chelone and (?) Uroskinnera, should an unusual occurrence in Scrophulariaceae. Three eventually be kept separate from the remainder, but stand perilously near the borderline between Scrophula- that course leaves a lot of problems which cannot be riaceae and Bignoniaceae : Paulownia studied by Nakai settled immediately. The degree of interrelationship' (19&--who makes it an independent family). Westfall amongst the woody South African genera is a problem (1949-favouring Bignoniaceae) and S. H. Hu (1959- in itself. favouring Scrophulariaceae) ; Wightia studied by van Rouy (1909) grouped the first five genera in the Steenis (1949), who keeps it in Scrophulariaceae, and above list as a sub-tribe Leucocarpinae, but only because J. K. Maheshwari (1961) who prefers to place it in .of the baccate indehiscent fruit. The sub-tribal name Bignoniaceae ; and Brandisia (see H. L. Li, 1917). may be, noted for future use, but this concept, which The conflicting views of these authgrs emphasise how places Halleria and Teedia with Leucocarpus but difficult it is to make comparisons at the family level excludes Hemichaena, has no botanical validity. in the present state of our knowledge. The transfer of ~kucorarpus, Hemichaena and Wettstein's genera in this tribe were: Berendtiella to Mimuleae is much more justified, the Synapsis Griseb.-transferred to &gnoniaceae most important characters which these groups have in (Urban, 1926). common being the tubular toothed calyx and bilamellate Leucocarpus D. Don-transferred to Mimuleae or stigma. These features are so far from those of Gratiola Gratioleae. itself that I am inclined to accept the tribe Mimuleae Dermatocalp 0erst.-a little known genus, perhaps as a useful unit. If this is done, then the subtribe Bignoniaceae (cf. - Monachino, 1949). Leucocarpinae Rouy may be re-defined on its cymose inflwescence and will then serve to distinguish ment are the open corolla, porose anther-dehiscence Leucocarpw, Hemichaena and Berendtiella from and unilaterally beaked capsules. Mimulinae. The chief criticism of such a course is Such annotations and commentaries on various that it takes group distinctions at this point beyond any aspects of the classification of Scrophulariaceae could available classification for the rest of Scrophulariaceae. no doubt be greatly extended ; but that would be out This criticism is not altogether valid, there are a good of place here. I have simply tried to put down some Inany subtribal names available (though, the sub-tribes useful notes, chiefly with reference to that part of the themselves may need re-definition), and other tribes are famil? where Cyrtandrornoea belongs, and also to give much more uniform than are the Cheloneae and some hint of the amount of work that needs doing Gratioleae, A start must be made somewhere. before we can achieve a satisfactory tribal system. It has been already suggested that much of the My mental picture of Scrophulariaceae is of a difficulty in putting forward a classification of S,cro- large family which is especially prolific in temperate and phulariaceae stems from the rather numerous anomalous warm temperate regions and in a group of genera genera. Some have already been mentioned. Brookea associated with damp habitats in warmer climates (e.g: Benth., a Bornean endemic genus, is another. Hallier the tribe Torenieae). But there are also many tropical (1908) wished to place it in Gesneriaceae, but I have and subtropical genera, often woody. It is amongst previously (Buflt, 1963, p. 217) rejected this view and these genera that the most interesting problems of the this rejection is now confirmed by the observaticn at family arise. This is consistent with the view that the Edinburgh that the seedlings have equal cotyledons. Scrophulariaceae is in origin a tropical family, despite Brookea, as originally suggested by Bentham, may he its great development in temperate regions. allied to the Central American Uroskinnera Lindl. There follows a list of 21 tribes which it seems This latter genus differs in having a well developed most reasonable to take as a prelbninary pattern for staminode, as is found in Penstemon and Chelone ; but further studies of the family. I have not considered on Wettstein's key neither Brookea nor Uroskinnera the genus Buddleia and its, close allies in these notes. should be placed in Scrophularieae (his Cheloneae) at The latest views on these prohleins are suinmarized by aM, as both have terminal racemose, not cymose. Leenhouts (1962, p. 265) and there is no additional inflorescences. It is this feature, also, which keeps them information to be added. Leenhouts retains Buddleia apart from Cyrtandromoea. Uroskinnsra has been in Loganiaceae, though Hart1 (1956) and Wagenitz revised by R. E. Schultes (1941), but he does not dis- (1959) have recently treated it under Scrophulariaceae, cuss its affinities. and Hutchinson (1959) regards it as a separate fawily. Bremekamp (1953) discusses the position of the !. Verbasceae ; Bentham (1835). S.W. African genus Hiernia S. Moore (1880) in relation 2. Leucophylleae : Bentham ( 1876). to the Scrophulariaceae-Acanthaceae boundary. If 3. Aptosimeae : Bentham (1876). Rhinanthoideae were raised to faillily rank, he says, 4. Calceolarieae : Bentham (1846). Nelsonioideae should be a subfamily of it and "for the 5. Antirrhineae : Bentham (1835). genus Hiemia perhaps a third subfamily might be 6. Hemimerideae : Bentham (1835). created. For the moment however I do not want to go 7. Scrophularieae SO far". 8. Russelieae : Pennell (1919). Hiernia has got into this discussion on families and Paulownieae : Pennell ( 1919). Collinsieae : Bellini ( 1907). :subfamilies under false pretences. The best, and cer- -tainly the busiest, of taxonomists make mistakes, and Minluleae : Pennell ' (1920). the simple fact is that Spencer Moore misplaced Hiernia Gratioleae : Bentham (1835). in Acanthaceae. The genus, which does seem to merit Torenieae : Rouy ( 1909). independence, should be placed in Scrophulariaceae-- Limoselleae : Rouy ( 1909). Gerardieae and not far from Graderia Benth., as Manuleae : Bentham (1876). already pointed out by Engler & Gilg (1903), and Selagineae : Wettstein ( 1895). Xylocalyx Balf. fil. (cf. Carter, 19621. The ~eculiar Digi taleae : Bentham ( 1835). , -- I woody calyx in the latter genus is, of course, quite dis- 18. Hemiphragmeae : Rouy ( 1909). tinctive and it also has anthers with both thecae fertile 19. Veroniceae : Bentham (1835). (though unequal in size). Apart from these points the 20. Gerardieae : Bentham (1835). differences are of a minor character. Points of agree- 21. Rhinanthae : Bentham (1835). 82 BULLETIN OF THE BOTANICAL SURVEY OF INDIA [Vol. 7

3. SYSTEMATIC REVISION OE! CYRTANDROMOEA Cyrtandromoea acuminata C. B. Clarke. However, we The first species of Cyrtandromoea to become now rule that he should never have used this name at known to botanists was the Javan C. decurrens, all, for in synonymy he included one of Miquel's described by Blume as a species of Loxonia in 1826. Sumatran species, Busea subsessilis, which provides the Three years later the name Cyrthandra [sic] acuminata earliest valid epithet. appeared in Wallich's Numerical List (no. 808 : 1829) The next important point to note in Clarke's treat- with reference to a specimen collected in Pena~gby ment is that he reduced Cyrtandra acuminata Kun to George Porter, the Superintendent of the Botanic a synonym of Cyrtandromoea decurrens, the Javan Garden there from 1822-1834 (he had previously been species. Furthermore his description (largely) and his overseer at Calcutta Botanic Garden). The subse- figure (entirely) of C. decurrens are based on Kurz's quent fate of this name will be dealt with later : it is specimen, This (no. 26098 in herb. C. B. Clarke now mentioned here because the plant is a Cyrtandromoea at Kew) is labelled Nicobar Islands, and so cited by and this was one of the earliest discoveries of. the Clarke ; but Kurz published the locality as Mt. Harriet, genus. Nothing happened then for nearly thirty years : Port Blair, Andaman Islands, and the spemimen matches until 1855 when Zollinger gave a new gerrerico-specific excellently with his description. Two sterile sheets at description for Blume's plant under the name CyrtandG- Calcutta have 'Andamans' crossed out and Katchall, moea decurrens*. Three years later, in April 1858, Nicobars substituted. Kurz's name is *also crossed out part five of the second volume of Miquel's Flora Indine as collector, but none other is '@ven. Batavae was published and in it the gemus Busea, Clarke's identification of Kurz's plant with including Blume's Loxonia decurrens. Miquel was C. decurrens seems improbable and the long narrow evidently unaware of Zollinger's publication and his inflorescences and white calyces are all against it. The name becomes a Fynonyrn of Cyrtandromoea. He question which arises is whether these distinctions war- added three new species of the genus, all from Sumatra ; rant the immediate proposal of a new name or whether they included a B. acuminata which, however, had no the matter should be allowed to stand over until the connection with Wallich's Cyrtandra acuminata. plant is rediscovered. In view of the uncertainty as to Bentham, when treating Gesneriaceae for Genera its origin I have adopted the latter course. Plantarum correctly reduced Busea to Cyrtandromoea Clarke transferred Miquel's species of Busea to and noted that Wallich's Cyrtardra acuminata (still a Cyrtandromoea and, as he already had C. acuminata nomen nudumj belonged to the same genus. In 1875 Benth. & Hook. f., renamed Busea acuminata as Cyrtrandra acuminata received a valid description from Cyrtandromoea miqueliaana. Little need be said of the hand of Sulpice Kure. But, although he took up the other species. C. cymulosa C. B. C1. seems to me Wallich's name and included Wallich's record, it is quite' inadequately distinguished from the plant we must clear that the basis for his description, and thereforc now call C. subsessilis, and I make the reduction accord- the type of the name, was his own specimen. ingly. It must be admitted, however, that none of the Now we come to Clarke's revision in 1883, and species of Cyrtandromoea is well known and most are his' treatment of plants with the epithet acuminata is very poorly represented in herbaria. It may well be one of the most important points. To start with he that other reductions will need to be made. followed some common practices of the time : he Cyrtandromoea entered the Bornean list when accepted Wallich's nomen nudum as a valid name Ridley described C. minor from Kuching. But this and Bentham's mention of it under Cyrtandromoea must have been an escape or a cultivated plant for it as effective transfer to that genus, and as it was pub- proves to be the American gesneriad Chrysothemis lished in Genera Plantarum he cites it as Cyrtandromoea pulchella (Don.) Decne. However the genus does acuminata Benth. & Hook. . f. From this Cyrtandra occur in Borneo : there are several specimens collected acuminata Kurz was excluded. Under the modern by J. & M. S. Clemens in the Kinabulu area and as Code there is clearly no basis for this citation : in fact already mentioned Woods and myself collected a species Clarke was now providing the first description of on Mt. Matang last year. We now have this in cul- Wallich's plant and his species should be cited as tivation at Edinburgh. It is a slender plant with pure

- - white flowers and seems to represent C. subsessilis (or * C. B. Clarke sugrsted (in - DC. Mon. Phan. v, p. a species very closely allied). The frequently basal 183 : 1883) that Zollinger's description was based on a mixture flowers and woody stem of the Kinabulu plants suggest of C. decurrens and the plant Clarke called C. acuminat~. I see no reason to suppose this was so. that they may be C. pndi~Ridl. 19651 BURTT : TRANSFER OF CYRTANDROMOEA FROM GESNERIACEAE TO SCROPHULARIACEAE 83

I thought I saw Cyrtandromoea in two other places enumeration of the species, are given with some doubt. in Sarawak : along the path by the Pelagus Rapids on In fact it cannot be over-emphasised that our knowledge the Rejang River, in an area of secondary growth with of the individual species of Cyrtandromoea is very Brookea, and in the Semengoh Forest Reserve outside incomplete and that classification within the genus is Kuching, also by a path. The plants were sterile, still in a preliminary stage. however, and my knowledge of the Sarawak flora as It is, nevertheless, clear that Sumatra has a greater a whole is too scanty to assert positively that they could diversity of forms than any other area, though the not have been any other genus. most distinctive species is the mainland C. megaphylla. The distribution of the various species is shown The two species which appear to have the widest dis- in the accompanying table. I have included as positive tributions are C. grandis and C. subsessilis, from records even those determinations which, in the Thailand & Burma to Borneo.

GEOGRAPHICAL DISTRIBUTION OF SPECIES OF CYRTANDROMOEA Andamans Burma Thailand Malay #enins. Sumatra Borneo Java or Nicobars 1. dispar A - A - + - - 2. rnegaphylla 3. miqueliana 4. ungustifolia 5. subintegra 6. decurrens - - - - + - +. 7. grandis -. - + + + + - 8. grandiflora - + + - + - -

9. (Cyrtandra ucuminata) + - - - - - A 10. sumatrana 1 1. subsessilis

Cyrtandromoea Zoll. Verz. Ind. Arch. 3 : 58 (1885) ; anthers wJth more or less parallel thecae hastately Benth. in Benth. & Hook. f. Gen. Pl., 2 : 1020 divergent at the base. Disc inconspicuous, annular. (1876) ; C.B.Cl. in DC. Mon. Phan. 5(1) : 184 Ovary conic or cylindric, short, 2-celled, placentae (1883) et in Hook. f. F1. Brit. Ind. 4 : 370 (1884) ; T-shaped ovuliferous over the whole surface. Style Ridey, F1. Malay Penins. -2 : 542 (1923) ; Lemee, long, slender ; stigma large, dorsoventrally bilamellate. Dict. Gen. 2 : 473 (1930). Syn. : Busea Miq. Fruit a thin-walled berry with fleshy placentae or a F1. Ind. Bat. 2 : 732 (1856). thin-walled loculicidal capsule ; always included within Habit : herbaceous or shrubby, slightly branched the enlarged calyx. Seeds numerous, with endosperm ; or more usually unbranched, with stems up to 10 ft testa of ripe seed appearing deeply reticulate, the inner high ; the stems sometimes quadrangular, with or with- and side walls having laminated thickenings and the out narrow wings, when woody with thin grey bark. outer wall being thin and eventually breaking dow~i. Leaves opposite, those of a pair equal or rarely (C. Seedlings with cotyledons equal in size throughout their disbar) very unequal with one reduced to 1 cm long 3 life (C. subsessilis, B. 1944). often drying blackish. Inflorescences cymose, few- flowered in the upper leaf axils (rarely solitary), or Pollen has 'been examined for Cyrtandrolnoea many flowered from the lower woody part of the stern grandis (B. 1616) and Pauline Woods has provided and then sometimes densely clustered sometimes elon- the following description :-Pollen finely reticulate, gate and pseudoracemose. Bracts small. Calyx tubular, simplibaculate homobrochate with lamina about 1p in shortly toothed, enlarged and usually conspicuously diameter. Colpi usually long, frequently pointed at' veined in fruit, red, white or green. Corolia infundi- the apices and covered by an almost psilate membrane. buliform, the limb bilabiate, the floor of the throat Measurements : PM (shortest distance between two (always?) with 2 raised ell ow ridges. Stamens 4 , colpi) = 5p (3.8-6) ; E (equatorial axis) = 21p 2 upper filaments longer than the 2 lower, all straight ; (20.8-22) ; P (polar axis) = 19~(17.6-22). 84 BULLETIN OF THE BOTANICAL SURVEY OF INDIA [VO~..7

KEY TO THE SPECIES OF CYRTANDROMOEA Mat. F1. Mal. Pen. Gamopet. 998 (1909), F1. Ma1 la. Leaves very unequal, one of each pair only about Pen. 2 : 543 (1923). cm long. 1. C. dispar MALAYA. PERAK : Water-fall Hill, 670 m, Ib. Leaves of a pair more or less equal and well developed herbaceous, 3 ft high, fl. white, comm. Jan. 1884, ...... 2. Wray 43 (holo. K) ; Gumong Haram, May 1884, 2a. Leaf-blade cordate at the base, ovate. : 2. C. rnegaphylla Scortechini 584 (CAL). PAHANG Gunong 2b. Leaf-blade narrowed or decurrent at base, lanceolate Benimbun, Nov. 1908, Ridley 13596 (K. BM) ; Fraser to elliptic ...... 3. Hill, upon the Selangor Border, 1200-1310 m, 16-30 3a. Plant covered with short thick (probably viscid) Sept. 1922, Burkill & Holltum SFN 8560. (K); hairs; fruit round, filling the calyx which is scarcely Renglet, Cameron Highlands, 1080 m, 22 Apr. 1930, inflated. 3. C. miquelianu ; 3b. Plant without a dense uniform indurnentum, pube- Henderson SFN 23663 (K) Sungei Yet, Fraser Hill, scent, scabro-pubescent (especially on leaf-stalk and 1110 m, fls. white, 28 Aug. 1923, M. Nur 11112. peduncle) or more or less glabrous...... 4. SELANGOR : Ginting Peras, May 1896, Ridley 4a. Leaves narrow, breadth not exceeding one quarter 7602 (CAL). of the length. 4. C. ang~cstifolia The petiolate leaves, ovate, cordate at the base, 4b. Leaves at least a third as long as wid&> ...... 5. are quite distinctive of this species, which has not yet 5a. Leaf-margin subentire ; fruit barely half the length of the calyx. 5. C. subintegra been collected outside the central areas of the Malay 5b. Leaf margins clearly serrate or dentate...... 6. Peninsula. 6a. Leaf-blade decurrent on ptiole (i.e. petiole winged) Cyrtandromoea miqueliana C.B .Cl. in DC. Mon. and continuing downwards to form narrow wings on Phan. 5 (1) : 187 (1883). Syn. : Busea acuminata the quadrangular stem (especially clearly seen on : young shoots). 6. C. decurrenr Miq. F1. Ind. Bat. 2 733 (1858)-non C. acurni- 6b. Leaf-blade attenuate into a distinct petiole ; shoots nata C.B.Cl. not quadrangular when young (sometimes becoming SUMATRA. Palembang, Teysmann (fide Miquel) so with narrow wings later)...... 7. 1196 H.B. (hol6. U, iso. L). Afd. Sikoto, Bukit Tinggi, 7a. Main inflorescences at base of woody stem in dense T 1000 m, 15 Jun. 1918, Bunnemeijer 3009 (L). bunched cymose clusters, a few flowers in upper They leaf axils. 7. C. grutrdis Clarke did not see flowers of this species. 7b. Flowers axillary, solitary or in small cymes, or in are present on Bunnemeijer 3009 and the following elongate pendulous- inflorescences on lower part of details may be added to the description : pedicel the stem...... 8. 1-1.2 cm, with short spreading hairs ; calyx 1-1.3 cm 8a. Corolla 5 cm long or more. 8. C. grandiflora long including the 1 mm .long teeth, shortly hairy like 8b. Corolla not exceeding 4 cm ...... 9. the rest of the plant ; corolla tube 4 cm long ; 4 mm 9% Flowers in long pendulous inflorescences wide at base expanding to 1 cm just below the throat ; 9. C. sp. (Cyrtandru acumiraata). ; 9b. Flowers axillary or in short cymes...... lo. upper lobes 5 mm long and 6 mm wide lower lobes 10a. Calyx bristly pubescent on and between the ridges. 7-8 rnm long and 7 mm .wide ; filaments, 2 long 2.6 cm, 10. C. srrmafrunu 2 shorter 2.2 cm ; style 2.5 cm; glabrous, expanded lob. Calyx glabrous or pubescent on the ridges. into a broadly bilamellate stigma ; ovary more or less 11 . C. su bscssilis cylindric, 5 mm long, glabrous. Cyrtandromoea dispar C.B.Cl. in DC. Mon. Phan C~rtandromoea angustifolia (Miq.) C.B.Cl. in DC. 5(1) : 187 (1883). Mon. Phan. 5(1) : 186 (1883). Syn. : Busea SUMATRA. Sine loc. Korthals 122 (synt. L), angustifolia Miq., Fl. Ind. Bat. 2 : 735 (1858). 123 p.p. (2 sheets, L). SUMATRA. Priamon, Teysmann 1193 HB (synt. C; B. Clarke quotes Korthals 650, which did not U, L). Sine loc., Ko~thals123 p.p. (Synt. L). come with material borrowed from Leiden. The two Since these specimens have already been given sheets of No. 123 were not seen by C. B. Clarke (a specific rank, no useful purpose is served by trying to third sheet of this number is C. angustifolia). Mate- re-evaluate them in the absence of further material or rial is still inadequate to confirm that this plant is in information. its correct genus. Cyrtandromoea subiitegra C.B.C1. in DC. Mon. Phan. Cyrtandmmoea megaphylla Hemsl. in Hook. Ic. P1. 5(1) r 187 (1883). 16 : tab. 1555 (1886) ; Ridley in J.. Str. Br. Roy. SUMATRA. Sine lot. Korthals 126 (holo. L). As. Soc. 43 : 86 (1905) et in J. As. Soc. Beng. This seems not to have been recollected. The 74(2) : 788 (1908), reimp. in King & Gamble, type sheet is in fruit and floral characters are unknpwn. 19651 BURTT : TRANSFER OF CYRTANDROMOEA FROM GESNERIACEAE TO SCROPHULARIACEAE 85

Whether the characters given are those of a species Sung, Trang, c. 900 my shrub 1.3 cm in evergreen or just of an individual specimen is quite uncertain. forest, flowers from base of stem, white with yellow Cyrtandromoea decurrens (Bl.) Zoll., Verz. Ind. Arch. streaks in throat, 17 Apr. 1928, Kerr 15273 (K, BM, L). 3 : 58 (1885) ; C.B.Cl. in DC. Mon. Phan. 5(1 ) : MALAYA. KELANTAN : Kwala Aring, Yapp 184 (1883 p.p. icon. excl.) et in Hook. f. F1. Brit. 163 (K). PERAK : Ulu Bubong, 120-180 my open Ind. 4 : 370 (1884), p.p. Syn. : Loxonia decurrens jungle, herb 4-6 ft, flowers only near ground, creamy Blume, Bijdr. 776 (1826). Busea decurrens (Bl.) white with bright yellow stamen, fruit reddish-yellow, Miq., F1. Ind. Bat. 2 : 733 (1858). 4 inch, diameter, enclosed in a dark red calyx, Jan. JAVA, Sine loc. Blume (holo. L) ; Kuhl &? van 1886, King's collector 10150 (CAL) ; Gunong Batu Hasselt (L) ; Reinwardt .(L) ; Zollinger (L) ; Hali- Puteh, 900-1200 m, dense jungle, rich soil in rocks, moen, Hasskarl (L.).. Bantam, Gunong Kantjana, 9 shrub 6-8 ft, leaves rich green with velvety gloss, flowers June 1912, Koorders 41175 (L), Bantam, Bodjong- white with large dark blue [sic] calyx, fruit claret manik, 14 June 1912, Koorders 40936 (L). Buitenzorg, colour, Aug. 1885, King's Collector .8155 (CAL) ; Gunong Wiroe, boven Nangila, z. W. v. Poerseda- Kroh Forest Reserve, Tapeh, flat semi-swamp, 5 ft Leuwiliang, 500 m, 4 Feb. 1929, Bakhuizen van den tall, calyx tube red with 5 points, corolla tube white, Brink 7121 (L). Buitenzorg, Tjianten z. W. van 11 March 1947, Wyatt-Smith (Kepong FN -63151). Leuwiliang, 1918, Beumee 849 (L). SE1,ANGOR : base of Bukit Hitam, shrub. flowers SUMATRA. Pulau Simaloe, 13 May 1918, at base of stern, calyx red, corolla white, May 1896, Achmad 460 (L). Ridley 7576 (K) ; Ulu Gombak road, March 1915, C. decurrens may have the largest leaves of any Ridley (K) ; Sungei Buloh, 12 Aug. 1908, Ridley 13360 in the genus. Those on the sp6cbnen marked as the (K, BM), ibid., 6 March 1915, Ridley (K) ; Welds type at Leiden have the lamina 30 c~rilong and 15 cm Hill, Kuala Lumpur, shrub 4 ft tall, calyx red, corolla broad, the winged petiole 6 cm long and 1.2 cnl broad. white. 12 Dec. 1920, Ridley (K) ; Ulu Gombak, sandy Those on allinger's specimen are not much over a soil in stream valley, tall weak slender unbranched third that size. Inflorescences may be few flowered and woody stem to 8 ft about 2 cm diam. at base, leafy axillary to leaves in the upper part or many-flowered towards top, largest cymes at ground level, a few on the lower part of the stem. Both types probably smaller ones above, calyx red, corolla creamy white, occur on the same plant (as they do in C. grandis) floor of throat translucent except for two yellow ridges, but the material available does not demonstrate this stamens and pistil white, 15 Apr. 1962, Burtt & Woods, adequately. B. 1616 (E). KEMAMAN : Bukit Kajang, 150 my The relationship between this species and C. grandis 28 Nov. 1935, Corner SFN 30710 (K). NEGRI may be very close and I have relied on little but the SEMILAN : Johol, calyx deep red, corolla white, broadly wingec! petiole to determine the specimen cited stems 7 ft, on banks in forest, 18 Jan. 1917, Ridley (K). above from the island of Simaloe, and the absence of JOHORE : Sungei Sedili, Mersing road, low altitude, such a winged petiole in the North Sumatran specimens 30 Sept. 1936, Corner SFN 31937 (K) ; Gunong Panti quoted under C. grandis- Examination of living plants west, low altitude, 14 Apr. 1936, Corner SFN 30960 (K). should show other differelices if the species are really N. SUMATRA. Sibolangit, T 350 m, 26 Nov. distinct. 1917, Loerzing 5430 (L) ; ibidem, T 400 m, 19 Dec. Cyrtandromoea grandis Ridl. in J. Str. Br. Roy. As. 1917, Loerzing 5474 (L). Ober-Deli, T 450 m, Soc. 43 : 87 (1905) et in J. As. Soc. Beng. 74(2) : mittleres Petanital, Gestrueppwildnesse bis Wald, meist 789 (1908), reimp. in King & Gamble, Mat. F1. im Schatten und auf feuchtem Boden ; halbstrauch 1-2 Mal. Pen., Garnopet. 999 (1909), F1. Mal. Pen. In hoch, 19 Feb. 1929, Loerzing 15276 (L). 2 : 543 (1923) ; E. C. Barnett in Craib, F1. Siam. The determination of the Sumatran specimens Enum. 3(3) : 210 (1962). is tentative ; see also the comments under C. decurrens. THAILAND. Kwae Noi Basin, Neeckey (near The occurrence of C. grandis in Sabah (North Wengka), 150 m ; shrub 1 m with few leaves at top, Borneo) is very. probable, but none of the material flowers on stem usually at base, pure white with yellow examined so far enables an absolutely safe detennina- Patch in throat, in shady places ; Karin 397 (I,). tion to be made. Although the characteristic basal Nahvn Sritamarat, Ronpibun Hill, c. 600 myin jungle, inflorescences are present there is nowhere in the * m, kwers mainly at base, calyx red, corolla cream, coYectors' notes any reference to the red calyx which Feb* 1922, Eryl Smith 397 (BM), 471 (BM); Kao seems to be a well-marked feature of the Peninsula 86 BULLETIN OF THE BOTANICAL SURVEY OF INDIA [Vol. 7 plant. Nevertheless the following specimens may be: material alone the Sumatran specimens (of whi~h, tentatively assigned to tbis species : however, Bartlett 10529 and Loerzing 13536 diverge in SABAH (NORTH BORNEO). Sandakan and having relatively broader more deeply toothed leaves) vicinity, Sept.-Dec. 1920, Ramos 1394 (K). Mt. must be referred to this species. Kinabalu : Dallas, 900 m, plant 3 ft, white with Cyrtandromoea sp. Syn. : Cyrtandra acuminata Kurz yellow stripe on side of labellurn, 24 Nov. 1931, in Journ. of Bot. 13 : 329 (1875)---excl.syn. Wall. ; Clemens 26871 (K, BM) ; Dahobong River, 1050- non Cyrtandromoea acuminata C.B.Cl. Cyrtandro- 1200 m, mar'gin among boulders, 4-5 ft, flowers white, moea decurrens ; C.B.Cl. in DC. Mon. Phan. 5(1) : fruit green, 11 Sept. 1933, Clemens 40323 (L, K, BM) ; 184 (1883) p.p. quoad syn. : kurzianum, Descr. p.p. Penubakan, ridge east of Dahobong River, 1200 m, et tab. 21. steep forest, 4-5 ft, flowers white, 2 Nov. 1933, Clemens ANDAMAN ISLANDS. Mt. Harriet above Port 50069 (K, BM) ; West canyon jungle, flower totally Arthur (ex Kurz-see remarks on p. 81). white, rocky stream margin, 4 Jan. 1933, Clemens The long pendulous inflorescences from the lower 30623 (BM, NY). part of the stem are the most distinctive feature of Cyrtandromoea grandiflora C.B.CI. in DC. Mon. this plant. In Fitch's illustration the inflorescence is Phan. 5(1) : 186 (1883) et in Hook. f., F1. Brit. shown as a raceme, and Clarke so describes it. This Ind. 4 : 371 (1884) ; E. C. Barnett in Craib, F1. is not so. Here, as throughout the genus, the inflores- Siam. Enum. 3 (3) : 210 (1962). cence is basically cymose. The bracts are arranged BURMA. Tenasserim ; Moulmein, Lobb (holi in subopposite pairs, one of them has no axillary bud, K) ; Moolyet, 1500 m, 31 Jan. 1877, Gallatly 265 and therefore appears to subtend the terminal flower ; (GAL) ; Thoungyaen, Beddome (BM). THAILAND. the other subtends the'branch which continues the deve- Payap : Doi Angka, east slope the Wang drainage, lopment of the inflorescence. The inflorescence is in S.E. of the Pa-Ngem, c. 1800 m, fl. white with yellow structure a true scorpioid cyme (cf. definition in lines, 30 Aug. 1927, Garrett 430 (K, L) ; Doi Angka, Lawrence, 1951, p. 755), though there is no hint of Me Ka Pak.drainage, c. 1650 myshrub 7 ft, A. white, scorpioid coiling, 21 July 1934, Garrett 891 (K, L) ; Doi Angka (Inta- Cyrtandromoea sumatrana Ridl. in J. Fed. Mal. St. nan), c. 1500 m, shrub about 3 m high, fls.' white Mus. 8(4) : 68 (1917). with 2 yellow ridges on lower lip, in evergreen forest, SUMATRA. Barong Bharu, W. side of Barisan 16 July 1922, Kerr 6295 (K, BM). Rachaburi : Kao Range, T 1200 m, Robinson & Boden Kloss (holo. BM, Ri Gai, Kanburi, straggling shrub 3 m, high evergreen iso. K). foreit, 1400 m, 1 Feb. 1926, Kerr 10373 (BM). Cyrtandromoea subsessilis (Miq.) B. L. Burtt, comb. SUMATRA. Pajakumbah, northern slope of Mt. nov. Syn. : Busea subsessilis Miq., F1. Ind. Bat. Sago, 1500-1600 m, in primary forest ; flowers white, 2 : 733 (1858). Cyrtandromoea acuminata C.B.Cl. tube with violet tinge, mouth with yellow dot, leaves vio- in DC. Mon. Phan. 5(1) : 185 (1883), in Hook. let at underside ; 30 June 1955, Meijer 3652 (L). Paja- f., F1. Brit. Ind. 4 : 370 (1884) ; Ridley in Journ. kumbah, Mt. Sago, c. 1500 mymountain forest ; flowers As. Soc. 74(2) : 788 (1909), F1. Malay Penins. with light red-brown stalk, white tube with red tinge, 2 : 542 (1923) ; Barnett in Craib, F1. Siam. Enum. white mouth with yellow bands ; I6 March 1956, 3 (3) : 210 (1962)-nomen illegitimum. ? Cyrtan- Meijer 4815 (L). East Coast, Deleng Si Naboen dromoea cymulosa C.B.Cl. in DC. Mon. Phan. (ascent from Kampong Goeroe Kinajan), Karoland ; 5(1) : 186 (1883), in Hook. f., F1. Brit. Ind. 3-5 ft tall, calyces greenish white, inflated ; 25-26 June 4 : 371 (1884). 1927, Bartlett 8613 (L). Tapianoeli, vicinity of BURMA. Tenasserim, Mergui, Grifith 225 (K) ; Loemban Loboe, Toba, 14 0ct.-14 Nov. 1936, Rahrnnt ibidem, Heifer 1203 (K); Wagou, flower pink and Si Boeea (Bartlett distr. 10529). N. Sumatra, Sibajak, white, April 1911, Meebold 15457 (CAL). 1700 my im etwas lichteren Urwald ; Kraut 60 cnl THAILAND. Pattani, Betong, c. 300 my shrub hoch ; Blueten auffallend, trompeten formig, weiss mit 1.5 m, fl. pinkish-white, edge of evergreen forest, 31 hellrot, 14 AU~.1908, Loerring 13536 (L). July 1923, Kerr 7427 (K, BM). Ronpibun Hill, 480 This species has not previously been recorded from m, ~etalswhite, sepals reddish or green, 1.5 m, high, Sumatra. A knowledge of the living plants or careful E;~ZSmith 424 (BM). Ban Kabuli, Toh Moh, 1500 comparison of material in alcohol, is needed before &is my herb in evergreen forest, 23 Apr. 1931, Lak~hnakara extension of range is absolutely certain. On dried 783 (BM). 1965J BURTT : TRANSFER OF CYRTANDROMOEA FROM OESNERIACEAE TO SCROPHULARIACEAE 87

MALAYA. KELANTAN : Chaning, white, 2 corolla with some long stalked golden glandular hairs. Feb. 1917, Ridley (K). KEDAH : Koh Mai Forest It has not so far proved easy to cultivate : some buds Reserve, 4 Apr. 1938, Kiah SFN 25172 (K, BM). fell without opening and the only flower to open failed PAHANG : Ulu Sungei Kuantan, 180 rn, calyx to set fruit after hand-pollination. It is noticeable that red, corolla white, 11 June 1934, Symington & Kiah though the young shoot behind the growing point is SFN 28779 (K). PENANG : Pwter (Wallich No. circular, the stem at the base of the young pl&nts 808, K, BM) ; ibidem, Maingay (Kew. distr. 123212, became 4-angled and narrowly winged. K, L) ; Penaru Bukit, 240 m, 4 Oct. 1886, Curtis 1016 (K) ; Penang Hill, 22 Aug. 1879, King (CALI. SPECIES EXCLUDENDAE PERAK : Waterfall Hill, Taiping, 1892, Ridley 2917 1. Cyrtandromoea minor Ridley in J. As. Soc. (EM) ; Taiping reservoir, 150 m, 27 Mar. 1924, Straits Br. 49 : 20 (1908) = Chrysothtmis pulchella M. Hanifl SFN 13132 ; Larut, 300-600 m, in dense (Donn) Decne., a native of Central America evidently jungle in rich soil, rocky, local, herbaceous plant more cultivated at one time in Kuching, Sarawak. The type like a shrub 6-8 ft leaves soft middle green, fls. white, of C. minor has been examined at Singapore. fr. red glossy, Dec. 1883, King's Collector [Kunstler] 2. Cyrtandromoea repens Ridley in J. Roy, As. 5342 (CAL) ; Larut, 240-300 m, open jungle, moist Soc. Straits Br. 57 : 74 (1911) = Gomphostermna localities, herb 3-5 ft, leaves soft and fleshy light green, curtisii Prain (Labiatae). The type of C. repens has fl. stem dark blue, fls. white, Aug. 1881, Kun~tler2139 been examined at Singapore. (CAL) ;Larut hills, 1892, Ridley 2917 (CAL) ;Tapah, 3. Busea aperifolia (Bl.) Miquel, F1. Ind. Bat. 2 : Nov. 1908, Ridley (BM) ; Temango, July 1909, Ridlev 733 ( 1858) = Chirita aspcrifolia (Bl.) B. L. Burtt in ; 14267 (BM) near Ulu Selama, stems 2-4 ft long, In Notes Roy. Bot. Gard. Edinburgh; 24 : 41 (1962). a clearing, 10 Jan. 1900, Yapp 624 (K) SELANGOR : Ulu Langat, Me Nuang Casing, Feb. 1912, Boden Kloss (K, BM) ; Ginting Sempah, March 1917, Ridlej~, Robinson d Kloss (K).; Bidai, March 1917, Ridley, I am particularly indebted to Rosemary Smith, Robinson B Kloss (K) ; Ulu Gombak, 28 Dec. 1920, both for preparing the illustrations and for other help ; Ridley (K). KEMAMAN : Bukit Kajang, 150 m, 4 to Heather Prentice for anatomical preparations ; to Nov. 1935, Corner 30202 (K). NEGRI SEMBILAN : Pauline Woods for the examination and illustration of Bukit Tangga, 20 Dec. 1920, Ridley (K). the pollen grains ; and to Patrick Woods for help in SUMATRA : Paiembang, Teysmann 1196 H.R: the field and for the colour slide on which the habit (holo. U). figure of C. grandis is based. I am grateful to the authorities of the heharia cited for allowing me to The nomenclature of this species has already been examine the material in their care. discussed. Despite the number of specimens quoted I also wish to thank Dr. H. Santapau, Director there is still room for critical comparisons of plants of the Botanical Survey of India, for honouring me with from various parts of the apparent range. Much better an invitation to contribute to this volume, which wery- herbarium material, and better notes are required, and one confidently expects will mark the beginning of a if possible seeds for cultivation. The Sarawak ~ plant new period of fruitfulness in botanical endeavour by grown at Edinburgh appears to belong to this species, the staff of the Indian National Botanic Garden and the but I have no knowledge of the living plant in other National Herbariuin. areas and I consider that this material should retain. a mark of interrogation against the determination. LITERATURE CITED It is: SARAWAK. Mt. Matang, 550 m, in secondary BANERJU, I. Endosperm in Scrophulariacaae. I. Zndian Bot. SOC. 40 : 1-11, 1961. growth on steep slope, herbaceous plant, leaves and BELL,N, CTiferi pr Una cla81ifiFYione stems bright green (quickly darkening in the press), Personatae (Scrophulariaceae et Rhinantaceae) . Ann. flower white, fruiting calyx green, 29 May 1962, Burtt Bor,, Roma 6 : 131-145, 1907.

& Woods B. 1944 (El.. . BENTHAM, G. In Lindley, Botmicai Regisre? 21 : Sub tab. 1770. 183Sa : (June). The following descriptive note from the plant - Scrophularineae indicae : a synopsis of the East Indian grown at Edinburgh may be added : calyx 1.5 cm Scrophulurineae, London 1835b : (after July). long, strongly 5 winged ; corolla, including lobes 2.3 - Scrophulariaceae, in De CandoUe, Prodromw systematis cm, white with 2 yellow ridges on floor ; outside of universalIiJ regni vegetabiib, I0 : 186-586, 1846. Paris. 88 BULLETIN OF THE BOT'ANICAL SURVEY OF INDIA [VO~.7

-In Bentham & Hooker, J. D., Gertera Plantartcnt, 2 : MAHESHWARI.J. K. The genus Wightia Wall. in India. 1876. Scrophulariaceae, pp. 913-980 : Gesneriaceae, with a discussion of its systematic position. Bull. bot. pp. 990-1025. London. Slcrv. India 3 : 31-35, 1961. BREMEKAMP.C. E. B. The delimitation of the Acanthaceae. MOHANRAM H. Y., & MASAND,PUSHBA. The embryology Proc. Acad. Sci. Netherlands, ser. C 56 : 533-546, 1953. of Nelsonia campestris R. Br. Phytomorphology 13 : 82- BR~N.R. Prodromus florae Novae Hollandiae. London 91, 1963. 1918. MONACHINO,J. V. ,A note on Schlegia and Dermatocalyx. BURIT, B. L. Studies in the Gesneriaceae of the Old World : Phytologia 3 : 102-105, 1949. xxiii : Rhynchoglossum & Klugia. Notes R. bot. Gdn. MOORE,S. Hiernia. J. Bat. 18 : 196-197, 1880. Edinb. 24 : 167-171, 1962. MORTON,C. V. A revision of Besleria. Contr. U. S. nut. - ditto, xxiv : Tentative keys to the tribes & genera. Herb. 26(9) : 395-474, 1939. Ibid. 24 : 205-220, 1963. NAKAI,T. Families of Japanese flowering plants. J. lap. Bot. CARTW.S. Revision of Xylocalyx Balf. f. (Scrophulariaceae). 24 : 13, 1949. Kew Bull. 16 : 147-152, 1962. NEES VON ESENBECK,C. G, Monograph of the East Endian CRETE, P. Developpement de l'albumen et de l'embryon Solaneae [& Verbascineae] Trans. Linn. Soc. Lond. chez le Collinsia bicolar Benth. (Scofulariacees) . 17 : 37-82, 1834. Phytomorphology 8 : 302-305, 1958. PE'NNELL,F. W. Scrophulariaceae of the South-eastern DIELS, L. Scrophulariaceae in Engler & Prantl, Die United States. Proc. Acad. nut. Sci. Philad. 71 : 224-291, natuerliche Pflanzenfam., Ergaengzungshefte 2 : 31 1, 1908. 1919. Leipzig. - Scrophulariaceae of Colombia-I. lbid. 72 : 136188, DON, D. Descriptions of two new genera of Nepaul plants. 1920. Edinb. Phil. J. 7 : 83-86,, 1822. - The Scrophulariaceae of eastern temperate North - Leucocarpus alatus in R. Sweet, British Flower Gurdett, America. Acad. nat. Sci. Philad. Mongr. No. 1. 1935. (ser. 2) 2 : t. 124, 1831. London. - The Scrophulariaceae of the western Himalayas. Acad. ENGLER,A. & GILO, E. Scrophulariaceae in CL Warburg, nat. Sci. Phtladelphia, Mongr. No. 5, 1943. Kunene-Sambesi Expedition H. Baum. 363-364, 1903. ROBERTSON,C. Zygomorphy and its causes. Bot. Gaz. Berlin. 13 : 146-151, 203-208, 224-230, 1888. PAULKS,P. I. : The systematisation of the angiosperms. ROBYNS,W. L'organization florale des Solanawes zygmor- 1963. Aberdeen. phes. Mem. Acad. R. Belg. 10 : 1-82, 1931. HALLIER,H. Ueber die Abgrenzung und Verwandschaft der ROUY, G. "Conspectus" des tribus et des genres de la eimlnen Sippen bei den Scrophulariaceen. Bull. Herb. famille des Scrofulariacees. Rev. gen. Bot. 21 : 194-207, Boiss. (ser. 2) 3 : 181-207, 1903. 1909. HARTL,D. Das Vorkommen rhinanthoider Knospendeckung SCHULTES,R. E. A synopsis of the genus Uroskinnera. bei Lindenbergia Lehm. einer Gattung der Scrophularia- Harv. Univ. Bot. Mus. Led. 9 : 65-83, 1941. ceae-Antirrhinoideae. Ost. bot. Z. 102 : 80-83, 1955. WEENIS, C. G. G. J. VAN. Notes on the genus Wightia - Morphobgische Studien am Pistil der Scrophulariaceen (Scrophulariaceae) . Bull. Jard. bot. Buitenz. sep 3, Ibid. 103 : 185-242, 1956. 18 : 213-227, 1949. HARZ, C. ~dwirtschajtlicheSamenkurtde. 1885. Berlin. - The land-bridge theory in botany. Blumea 11 : 236- Hu, SHIU-YING. A monograph of .the genus Paulonia. 372, 1962. Quart. J. Taiwan Mus. 12 : 1-54, 1959. THIERET,J. W. Tribes and genera of Central American HUTCHINSON.J. Families of Flowering plants. Ed. 2. 1959. Scrophulariaceae. Ceiba 4 (3) : 164, 1954. Oxford. TIEGHEM,P. VAN. St~cturede I'etamine chez les Scrofularia- JOHRI, B. & HARDEVSINGH. The morphology, embryology cees. Ann. Sci. Nat. Ser. 8, 17 : 363-371, 1903. and systematic position of Elytraria acaulis (L.f.) Lindau. URBAN,I. Synapsis ilicifolia. Fedde, Rep. Sp. Nov. 22 : 369- Bot. Notiskr, 112 : 225, 1959. 371, 1926. IUNELL,S. Ovarian morphology and taxonomical position WAGENITZ,R. Die systematische Stellung der Rubiaceae : of Selagineae. Svensk bot. Tidskr. 55 : 168-192, 1961. ein Beitrag zur System der Sympetalen. Bot. Jahrb. LAWRENCE,G. H. M.. Taxonomy of Vascular Plants. New 79 : 17-35, 1959. York 1951. WESTFALL.J. jJ. Cytological and embryological evidences LEENHOUTS,P. W. Loganiaceae, in Flora Malesia~la (ser. for the reclassification of Paulownia. Amer. I. Bot. 1) 6 (2) : 294-296, 1962. 36 : 805, 1949. Lr, HUI-LIN. The relationship and taxonomy of the genus WEITSTEIN,R. VON. Scrophulariaceae in Engler & Prantl, Brandisia. I. Arn. Arb. 28 : 127-136, 1947. Die ~raturlichePflanzenfam, 4 : 3B, 39-107, 1895.