ACTA BOTANICA NEERLANDICA, Vol. 4 (5), 1955

Notes on some Acanthaceous Genera and

Species of Controversial Position

BY

C.E.B. Bremekamp

(.Botanical Museum, Utrecht)

{received October 12th, 1955)

I. The position of the Stenandriopsis S. Moore

The created S. Moore in of genus Stenandriopsis was by Journ. Bot. 44: 153. 1906 for a collected first by Vaughan Thompson and Baron in afterwards by an unspecified part of Madagascar. As the plate by which the description is accompanied depicts the specimen Baron the latter is be collected by (n. 6708), to regarded as the type. its author the but this Stenandriopsis was referred by to Justicieae, tribe is apparently accepted by him in the delimitation it received in

Bentham and Hooker’s “Genera Plantarum”, and as it is in this

sense a most heterogeneous mixture, this does not greatly enlighten Of is that Moore it with Crossandra us. more importance compares

Salisb. and Stenandrium i.e.with to Nees, genera belonging my subfamily and referred by me respectively to the Acantheae and the

in the Malesian Aphelandreae. However, my paper on “The Acantheae of

Area. I. General Considerations” in Proc. Kon. Ned. Akad. v. Weten-

58: I the sch., Ser. c. 166. 1955, pointed out that it can not belong to

Acantheae as the corolla throat lacks the incision in the adaxial side

which is characteristic for that tribe. It the can not belong to Aphe-

landreae either as the corolla limb is subactinomorphous instead of As I had that time distinctly bilabiate. to rely at entirely on Moore’s

and the which the latter is I description on plate by accompanied, was unable arrive but I that the to at a conclusion, suggested genus might tribe of Acanthoideae. Since then I have had the represent a new my opportunity to inspect in the herbarium of the British Museum of Natural the material which the for History on genus was based, which best thanks I tender my to the Keeper, and now I am able to definite express a more opinion. the it appeared that stem was not provided with articulations,

all the that parts were free from cystoliths, and that pollen grains were but not monocolpate, as described by Moore, tricolpate, there can be doubt that the the Acanthoideae no genus belongs to sensu meo. That the pollen grains were regarded by Moore as monocolpate, is ACANTHACEOUS GENERA AND SPECIES OF CONTROVERSIAL POSITION 645

doubtless due to the circumstance that he made his observations on immature when ones. Such grains shrink, as a rule, considerably the flowers are dried, and in shrinking the wall is at one side drawn in; this dent was apparently mistaken by Moore for a normal feature of the whereas the of the three pollen grain, presence true colpae was overlooked.

As the absence of an incision in the adaxial side of the corolla

throat and the subactinomorphous nature of the corolla limb were conclusion that the either confirmed, my genus can not belong to the Acantheae or to the Aphelandreae is to be accepted. The structure of the flower be similar that of the flower of proved to very to Haselhoffia from Lindau, a genus differing Stenandriopsis in the presence of four instead of ovules the cells and in the two in each of ovary ellipsoidal 12-colpate instead of globose and 3-colpate pollen grains. It apparently distinct tribe be characterized follows: represents a which may as

Stenandriopsideae, nova tribus Acanthoidearum corolla limbo subactinomorpho instructa ad tribum Haselhoffiearum accedens, ovario loculo ovulis duobus utroque munito, granulis pollinis globosis et 3-colpatis ab ea diversa, ovarii structura cum tribu Acanthearum et

cum tribu Aphelandrearum quadrans, a tribu Acanthearum corollae fauce

adaxiali tribu limbo facie Integra, a Aphelandrearum corollae sub- actinomorpho distincta. Habitat insulam Madagascar dictam. Genus typicum: Stenandriopsis S. Moore in Journ. of Bot. 44 :153, tab. 478 B. 1906.

S. Moore The type species of the genus Stenandriopsis is St. Thompsoni this the and is only species of which specimens were available to me.

The three Madagascar species which were described by Benoist in leaves. Not. Syst. 8: 151. 1939, are provided with pseudo-verticillate Leaves in this characteristic feature of the arranged way are a genera Salisb. Crossandra and Blepharis Juss., and as the corolla limb of Crossandra is like that of Stenandriopsis subactinomorphous, these three in species might perhaps belong either to Crossandra or to a new genus the latter’s vicinity. The adaxial side of the corolla throat, however, shows Crossandra if with in a deep incision, and my supposition regard least to the position of these species is right, such an incision or at

an indication of the latter should be present here too.

II. The Lindau position of the genus Rhombochlamys

The of the of which so far position genus Rhombochlamys Lindau, two species were found in South America, was also discussed by me in “The Acantheae of the Malesian Area. I. General my paper on Ser. Considerations” in Proc. Kon. Ned. Akad. v. Wetensch., C,

58; 164-165. 1955, in connection with the delimitation of my sub- I the the tribe Rhombo- family Acanthoideae. came to conclusion that which created Lindau for the of this chlamydeae, was by reception genus, Lindau can not belong in the vicinity of the Andrographideae, as thought,

but that there is good reason to insert it in the vicinity of the Aphelan- 646 C. E. B. BREMEKAMP

dreae. Lindau paid too much attention to the pollen characters, which,

moreover, he seems to have observed in this case inaccurately, and the resemblance with the neglected apparently very striking Aphelan- dreae, which nevertheless had not escaped his attention, for he duly remarks “Aeusserlich hat die Gattung mit Aphelandra Aehnlichkeit”.

However, here too a definite conclusion could not be reached because with three of the four main no information was available regard to characters in which the subfamily Acanthoideae differs from the sub- articulations in family Ruellioideae, viz. the absence or presence of the the absence of and the absence stem, or presence cystoliths, or presence of in pores the pollen grains. When Lindau had been right in placing this tribe in the vicinity of the Andrographideae, it would have fallen in my subfamily Ruellioideae, and the stem ought to show articulations, be and cystoliths ought to present, the pollen grains ought to be with provided pores. The of the the types two species pertaining to genus Rhombochlamys in the unfortunately are no longer available; they were lost fire by which the Berlin-Dahlem Herbarium was devastated. However, I

could in the Kew Herbarium in study an Acanthacea collected Colum- bia, i.e. in the country in which the types of the two species were found, them. rather and identified by Standley as belonging to one of The

condition of this did allow full verification of the poor specimen not a in its characters could be identification, but so far as compared with

the description they proved to agree with it, so that there is no reason to doubt the correctness of the identification, which probably will have been based on more At rate the con- a complete specimen. any choice dition of the specimen was sufficient to allow a in two of the three pairs of alternatives mentioned above. The stem proved to be without articulations, and cystoliths were absent; the pollen structure unfortunately could not be studied. The absence of articulations and of in with mentioned in the cystoliths conjunction the presence, of four monothecous description, anthers, however, indubitably prove that Lindau erred when he assigned to this genus a place in the vicinity of the Andrographideae, and that it certainly is to be referred to the Acanthoideae, where it will have to be placed near or perhaps even in the Ifit should that the do Aphelandreae. appear pollen grains not possess germ pores, they would hardly differ from the 3-colpate type that is with in the the the met great majority of Acanthoideae, Aphelan- dreae included, and then the only point in which the Rhombochlamydeae would differ instead of from the latter, would be the presence of four ovules in each of the and this would be sufficient two ovary cells, hardly to maintain them as a separate tribe.

“ III. The position of the african Hemigraphis ” species

In my “Materials for a Monograph of the Strobilanthinae (Acantha-

” ceae)” in Verb. Ned. Akad. v. Wetensch., Afd. Natuurk., 2nd Sect. artificial 41, no 1, 1944, I rejected the former tribe Strobilantheae as an artificial referred Strobilanthes its group, and 81. and nearest allies, among ACANTHACEOUS GENERA AND SPECIES OF CONTROVERSIAL POSITION 647

which Hemigraphis Nees occupies a prominent position, as a subtribe

Strobilanthinae to a widened tribe Ruellieae. The subtribe Strobilanthinae

1° the proved easily recognizable by presence of two rows or, more bundles of hairs which the of the is retained rarely, by upper part style against the adaxial side of the corolla, and 2° by the laterally flattened

abaxial stigma lobe. I stated l.c. that this subtribe is confined to Asia

and the adjoining part of Oceania, where it is represented by a few Hemigraphis species. The position of Bentham’s African Strobilanthes species, for which Clarke in Thiselton-Dyer’s “Flora of Tropical Africa” 162. had created I had (5: 1899) a new genus Acanthopale, And. discussed a year before in my paper “Ueber Dischistocalyx T. ex Bth. und Acanthopale C. B. Clarke ()” in Bot. Jahrb. 73: is 126-150. 1943, where I came to the conclusion that Acanthopale no had near ally of Strobilanthes, as Clarke assumed, but that it should

these be placed near Dischistocalyx, as Lindau had done, and that two Ruellia L. Of the Strobilanthes genera belong to a group centered round

species that had been described from Madagascar I stated in my

” “Materials for a Monograph of the Strobilanthinae (p. 29) that they

had been referred this and that do not wrongly to genus, they even belong to the Strobilanthinae. This apparently applies also to the Mada- that since then have been described. The gascar species descriptions of these , however, are in most cases not sufficiently detailed to allow determination of the which a genus to they really belong. Of the three African species that Clarke in Flora of Tropical

Africa 5: 58. 1899 had referred to Hemigraphis Nees, viz. H. abyssinica (Nees) C. B. Clarke, H. tenera (Lindau) C. B. Clarke and H. Schwein- furthii C. B. Clarke, and of the two African species that since then had

been added to this group, viz. H. origanoides Lindau and H. prunelloides S. I said l.c. that this because Moore, they coud not belong to genus

their flowers are arranged in small axillary clusters, their calyx seg- and ments are nearly free, their stamens about equal in length, their pollen grains globose. I suggested that they might belong to Synnema

Bth., i.e. to the subtribe Hygrophilinae. At the time I worked “Materials for of the on my a Monograph

Strobilanthinae” I had to rely solely on the descriptions of these species

given in the literature, but recently I found the opportunity to study specimens in the Herbarium of the Royal Botanic Gardens, Kew,

for which opportunity I wish to thank the Director. This study

confirmed that these do to the my supposition plants not belong which Strobilanthinae, for the hairs on the adaxial side of the corolla by in this the is retained in its be and group style place, proved to absent, the abaxial stigma lobe is not laterally flattened but filiform. That they

really belong to the Hygrophilinae is proved by the shape and structure

of the pollen grains, which are in all these species not only globose, in what itself is of no great importance, but also ornamented with a

fairly large number ofbands, all of themofthe same widthandseparated from each other and with by narrow furrows, provided four equatorial of the germ pores situated in the furrows. They are therefore type that is characteristic for this subtribe. 648 C. E. B. BREMEKAMP

In referring these species to the Hygrophilinae I returned to the stand- point that the earlier authors had taken in. The longest-known

species of this group, Hemigraphis abyssinica, was originally described 11: by Nees in DC, Prodr. 83. 1847 as Polyechma abyssinicum, and the

“ Hochst. was him in his Hygrophileae where genus Polyechma placed by , it is the and preceded by genera Hemiadelphis Nees Physichilus Nees, and followed by Glossochilus Nees, Nomaphila Bl. and Hygrophila R.Br., which that it is included of with means in a group genera which, the of the known all of exception imperfectly Glossochilus, are part my subtribe Hygrophilinae. Anderson (in Journ. Linn. Soc. 7 : 22. 1864) convinced between was so fully of the near affinity the Polyechma species listed by Nees and the species which the latter had referred that he sank the whole in S. Moore to Hygrophila, group Hygrophila. had assumed when too originally a similar attitude, viz. he described that a plant Clarke afterwards referred to Hemigraphis abyssinica, as

related Cardanthera justicioides, and some nearly ones as C. africana

S. Clarke (T. And.) Moore var. Schweinfurthii; on the latter afterwards based his That the Cardanthera Buch- Hemigraphis Schweinfurthii. genus Ham., which is identical with Synnema Bth., belongs to the Hygrophilinae

can not be doubted.

In order to understand why Clarke in opposition to the three authors quoted above referred these plants to Hemigraphis, we should realize that Clarke did this in the of not place genus neighbourhood

Strobilanthes, as Bentham and Hooker had done in their “Genera

Lindau did in Nat. Plantarum” (2: 1065. 1876) and as Pflanzenfam.

IV, 3 b: 300. 1895, but in that of Ruellia; his tribe Strobilantheae was restricted those of Ruellieae in which the cells contain to genera my ovary but one or two ovules, whereas his Euruellieae differed from his Hygro- the instead philinae in actinomorphous or subactinomorphous of

distinctly bilabiate corolla. As the Asiatic species for which the genus African Hemigraphis was created, as well as the species with which we

with are here concerned, are provided actinomorphous or subactino- ovules in morphous corollas, and as they possess more than two each of the it is clear that had be referred in this classi- ovary cells, they to

fication to the Euruellieae, and a certain similarity in habit, caused

by the comparatively small stature of these plants and by the small

size and white colour of the flowers, will have led him to consider due the them congeneric. His mistake therefore was to circumstance

that he was not acquainted with the really important features of the Strobilanthinae, and that he overrated the importance of the differences in the of the degree symmetry of corolla (actinomorphous or nearly far been referred so versus zygomorphous). In the genera that so have find all transitions to the Hygrophilinae we between the distinctly

bilabiate corolla of Brillantaisia Pal. Beauv. and corollas that are no

more distinctly zygomorphous than some of those that are found in the other subtribes of the Ruellieae. Moreover, we should not forget that the corolla when we speak in Acanthaceae of an actinomorphous we mean a corolla with an actinomorphous limb; the corolla consi- dered whole is the androecium as a always zygomorphous, being ACANTHACEOUS GENERA AND SPECIES OF CONTROVERSIAL POSITION 649

adnate to its adaxial side; it is not be expected that the limb of such a corolla will ever be fully actinomorphous; there is only a more or less close approximation to this state. Apart from the but slightly zygomorphous corolla (it is cautiously

described by him as “not 2-Hpped”) there is in the description of the Clarke in the “Flora of Africa” genus Hemigraphis given by Tropical but one point which, if it could be confirmed, would exclude the

“ ” possibility that the African Hemigraphis species might belong to the Hygrophilinae, viz. the subtrigonous shape of the pollen grains. in This however, proved to be a mistake; reality the pollen grains and with four are globose provided equatorial germ pores, which, as that the this stated above, irrefutably proves plants belong to subtribe. At this character not have influenced Clarke’s any rate, can decision, because at that time it was not yet known that the presence of four is character that is confined this subtribe. equatorial germ pores a to It was even generally believed that the pollen grains of the Hygro- with three indeed philinae were provided pores; so they had been described by Lindau in Naturl. Pflanzenfam. IV, 3 b: 295. 1895. The of four for the first time used presence pores was as a diagnostic of the in the Surinam Genera character Hygrophilinae my “Key to and based Pollen which is be found in Species on Characters”, to my Surinam” “Notes on the Acanthaceae of in Rec. d. trav. bot. Neerl.

” 35: 136. 1938 (see also the “Key to the subtribes of the Ruellieae in Tremekamp and “A ofthe Nannenga-Bremekamp, preliminary survey Ruelliinae (Acanthaceae ) of the Malay Archipelago and New Guinea” in

Ned. Natuurk. n

the other character that is never met with in this hand, one genus nor, in other the Strobilanthinae. This is the in fact, any genus belonging to arrangement of the “flowers in small axillary clusters”. Although

Clarke was not acquainted with the two main diagnostic characters of the Strobilanthinae to which I have already referred, and although therefore nevertheless he lacked a reliable guide, it is strange that he the accepted arrangement of the flowers in small axillary clusters as admissible in the for this kind of genus Hemigraphis, arrangement of the flowers had been of the his doubtless one principal reasons why

had referred of this either predecessors plants group to Hygrophila or to in the latter’s genera vicinity.

The question to what genus the five African “Hemigraphis” species be are to transferred is difficult to decide, at least when we do not

in the in accept Hygrophila R.Br. wide sense which it was understood for in by Anderson and by Clarke, that case they are doubtless to be included in that This is in genus. delimitation, however, my opinion too wide, and I advocate the restoration of the genera Hemiadelphis Nees and Bl. if the Nees, Physichilus Nomaphila and, necessary, creation of one or more new ones.

In “Materials for of the Strobilanthinae” I referred my a Monograph

“ ” these African Hemigraphis species in the “Index Specierum” (l.c. 650 C. E. B. BREMEKAMP

Bth. At that time this seemed the p. 136—142) to Synnema to me most because the of the plausible solution, type species Polyechma, genus to which of the had been referred Nees one species by reminded me of this genus, and also because S. Moore had referred two of these species stated to Cardanthera, which, as above, is synonymous with Synnema.

This decision, however, is to be rescinded. Synnema Bth. is well charac- terized by the very large number of ovules and by the small size of and these with in the retinacula, characters are not met the African

“ Hemigraphis” species. have had referred of the Nees, as we seen, one species to Polyechma and it therefore be that this would Hochst., might possible genus prove to be the most suitable place for them. This, however, has also proved an illusion. P. coeruleum the of the Hochst., type species genus, appears be with ovules and with to provided numerous weakly developed and in these it retinacula, undoubtedly very important points agrees therefore with and be identical with the latter. At Synnema, might any whether the in which rate, two are identical, case Polyechma would be the for the distinct correct name genus, or though nearly related, so that returned much is certain Hemigraphis abyssinica can not be to this

as its retinacula are well and its ovules not genus, developed, very numerous. A more suitable be found in the place can perhaps genus

Physichilus Nees, in which a similar arrangement of the flowers is met with, but where the corolla is more distinctly zygomorphous.

“ ” However, whether we refer the African Hemigraphis species to R.Br. Nees Hygrophila sensu T.And., to Physichilus or to a new genus in the vicinity of the latter, is for the moment ofsecondary importance only, the main point is that they do not belong to Hemigraphis nor to

other of the Strobilanthinae but to the any genus Hygrophilinae.

IV. The position of the genus Stenothyrsus G. B. Clarke

The based in monotypic genus Stenothyrsus was by Clarke Journ.

As. Soc. Beng. 74: 650. 1908 on a plant collected by Curtis in Perak, Malay Peninsula. It is said to be related to Hemigraphis Nees, from which it would differ mainly by the nature of its inflorescence, which is described of short as a spike consisting opposite cymes made up of 6 7 flowers. inflorescences of this kind or However, as are entirely in the the that this unknown Strobilanthinae, supposition genus would be related to Hemigraphis seemed to me unfounded, but as the descrip- tion clue its real and had gave me no to position, as I no material of this hand when I “Materials for species at was working on my a Monograph of the Strobilanthinae” I left it out of consideration. Since then I have found occasion to inspect in the Kew Herbarium the which based. As it specimen on the genus was appeared that there were no rows or bundles of hairs on the adaxial side of the corolla of the in its for retaining the upper part style place, and that the rather abaxial stigma lobe, though long, was not laterally flattened,

conclusion that the the my genus can not belong to Strobilanthinae was fully confirmed. A study of the globose pollen grains revealed the ACANTHACEOUS GENERA AND SPECIES OF CONTROVERSIAL POSITION 651

of four situated in and of 24 presence equatorial germ pores groves or size and from 28 meridional bands, all of the same and shape separated other which contained the each by narrow groves, four of germ pores. that the has be transferred There can therefore be no doubt genus to to the Hygrophilinae. The abaxial side of the corolla throat is provided corolla limb in the African with a very distinct palate, but the is, as

“ Hemigraphis” species dealt with above, subactinomorphous, and this have induced Clarke it related will doubtless to regard as to Hemigraphis ; the of distinct is another feature presence a palate, however, pointing in the direction of the Hygrophilinae. On account of the terminal long-

and pedunculate narrowly paniculiform inflorescence it occupies a rather isolated position in this subtribe, but the cymous partial in the inflorescences are structure fully comparable to axillary inflores- cences found in the other genera. Noteworthy diagnostic characters the the are two spinules at base of each theca and the strongly pilose dorsal side of the connective.

V. The position of the genus Oreothyrsus Lindau

of received Among a batch Acanthaceae for identification from the

Director of the Philippine National Herbarium, Manila, I noticed a with twice plant opposite long-pedunculate, once or more often pseudo-dichotomously ramified inflorescences and with globose,

The these characters sparsi-porous pollen grains. combination of was in that observed so far only two New-Guinean species Lindau in

and Schumann Lauterbagh, Nachtrage zur Flora der deutschen

had referred Schutzgebiete in der Siidsee, 308-309, 1905, to a new and it seemed therefore indicated the genus Oreothyrsus, to put Phi- in this is lippine plant at least provisionally genus. It true that the of the latter does in with that pollen structure not agree every respect of the two New-Guinean species, but for the moment it seems allowed Further to regard the differences as of secundary importance only. on this then it we will return to question, and we will discuss in some detail. A final decision, however, will have to be postponed until the

New-Guinean species can be studied on actual specimens. At present Lindau’s the this is impossible, as specimens, as Director of the Botanical Museum and Herbarium, Berlin-Dahlem kindly informed

and as no new collections are available. For me, are no longer extant, critical examination of the which the has been a position to genus the referred, we will have to rely on description given by Lindau and data on the which the study of the Philippine species has brought to light. Before entering into details it seems therefore desirable to give of the latter. Here it is. a description samarensis Brem Oreothyrsus n. spec, a congeneribus hactenus notis foliis basi rotundatis, floribus majoribus, calyce 5-partito,

labio corollae superiore integro, granulis pollinis nec alveolatis nec scrobiculatis, capsulis majoribus distinguenda.

Herba e basi decumbente erecta, 20-30 cm alta. Caulis simplex, primum tomentellus, deinde puberulo-pubescens, 0.9-1.6 mm diam.. 652 C. E. B. BREMEKAMP

internodiis Folia 2-6 cm longis. petiolo primum tomentello, deinde dense puberulo-pubescente 1.0-2.5 cm longo instructa; lamina oblonga, 3.5-7.5 lanceolato-oblonga vel ovato-lanceolata, plerumque cm longa 1.5-3.5 interdum 10 3 et cm lata, tamen usque ad cm longa et cm lata,

apicem obtusum versus plerumque paulum contracta, basi rotundata et ibi interdum paulum asymmetra, margine integra, paulum discolor, sicc. subtus dilute supra saturate et olivacea, supra glabra et cystolithis parvis dense lineolata, subtus costa nervisque primumtomentella, deinde inter puberulo-pubescens, nervos glabra, nervis utroquelatere costae 4-7, venulis inconspicuis. Inflorescentiae oppositae, pedunculo puberulo-

pubescente 2-4 cm longo elatae, semel vel saepius bis dichasiales, abortu florum terminalium tamen semel vel bis pseudo-dichotomae, ramulis

ultimis monochasialibus. Bracteae subulatae, infimae usque ad 3 mm 1.5 Bracteolae longae, superiores usquead mmdecrescentes. nullae.Pedi- celli 1.5-3.0 mm longi. Calyx aequaliter 5-partitus; lobi 7 mm longi et basi 0.6 anthesin ad 9 mm lati, post usque mm accrescentes, acuti, Corolla 6 tubo 3 subglabri. mm longa, extus subglabra, mm longo fauce 1.5 labiis etiam apice parce barbato, mm, 1.5 mm longis; labium superius integrum, sine rugula; labium inferius 3-partitum, lobo 1.0 lobis mediano mm lato, lateralibus 0.3 mm latis. Stamina 2 fila-

antherae mentis glabris 2.2 mm longis instructa; dorsifixae 1.2 mm

altae, apice basique obtusae, paulum asymmetrae. Granula pollinis

(cf. fig.) globosa, sparsipora, poris 6-8 virgis circumdatis, virgis a

Pollen grain of Oreothyrsus samarensis Brem.

scutellis 45 diam. Staminodia nulla. Discus 0.5 separatis, fi mm altus. 1 Ovarium glabrum mm altum, utroque loculo ovulis 2. Stylus 4.0

mm longus, glaber; stigma vix distinctum. Capsula stipitata 15 mm 3 longa et mm diam., glabra, 4-sperma; retinacula acuta vel vix bidenticulata. conspicue Semina corrugata; testa e cellulis tracheidali- bus composita. Habitat insulam Filippinam Samar dictam.

Samar: Mt alt. 100 Edano 3180 Sarawag, m, (PNH 15323), typus, fl. 8.XII. Mt alt. 150 id. 1951; Purog, m, 3444 (PNH 15433), 3477 (PNH 15424), 3489 (PNH 15402), all fr. 24.XH.1951.

As it is as yet not absolutely certain that this species really belongs to Oreothyrsus I have mentioned in certain features that , my description

be as charactersofthis may regarded general genus. I have also included number of characters that omitted a are usually from species descrip- tions because for of they are typical groups higher rank; they have

been inserted here because the to which the is group genus belongs, still unknown. before inclusionof However, attempting to justify my ACANTHACEOUS GENERA AND SPECIES OF CONTROVERSIAL POSITION 653 the in this and Philippine plant genus, before entering into a discussion of the of the it desirable position latter, seems to review our knowledge of the and the two New-Guinean species of grounds on account of which the author of the referred genus it to a definite group. The his genus, which was referred by its author to Porphyrocominae , in was described by him the following terms: 5-lobus. Corolla labio “Calyx parva, bilabiata, supero bidentato, infero 3-lobo. Stamina 2 antheris aequialte affixis, muticis. Pollinis granula globosa, scrobiculata, poris 8 instructa. Capsula stipitata, 4- —Herbae. Inflorescentiae pedunculatae, dichotome-ramo- sae.” (“Stamina 2 antheris aequialte affixis” should doubtless be read “Stamina as 2 thecis aequialte affixis”, but this was a mere slip of the pen). The described above differ in Philippine species appears to three points from this generic description: the calyx is 5-partite instead of 5- the of the corollais entire instead of upper lip bidentate, and the pollen grains are ornamented with bands alternating with shields The instead of scrobiculate. two first-named points are doubtless of minor and be importance may left out of consideration, but the the difference in the relief of pollen grains is not so easily dismissed. In the remarks added by Lindau to his generic description the pollen grains are described as sparsiporous, and their surface is said to be with “Gruben die durch Leisten sind, provided . . ~ getrennt oder welche oft zu kurzen Spitzchen Hockerchen sich erheben.” This might suggest a structure as is found in the pollen of the Ruelliinae, which of a group in some the genera are provided with sparsiporous but the bilabiate corolla makes it pollen, distinctly very improbable the that genus should belong to this group. However, our impression that the of should be of this pollen Oreothyrsus type, may be entirely That will wrong. this is possible, have to be admitted when we recollect that Landau’s description of the pollen structure of Lepidagathis Willd. , „ suggested similarity with the pollen structure ot Barleria L, and was indeed intended to suggest this, for Lindau referred these rather

the same Here the is widely differing genera to group. too similarity for the almost very slight indeed, imperceptible sculpture of the the Lepidagathis pollen is doubtless totally different from coarse reticu- lation seen in that'of Barleria. It is therefore by no means impossible that the reticulation of the pollen ofthe two New-Guinean Oreothyrsus the species too is very faint, and that spinules by which the grains are covered deserve more attention. This certainly was not Lindau’s opinion for it was doubtless on account of the reticulation of the pollen that he the in grams put genus the vicinity of the South-American which he had described earlier genus Poikilacanthus, at an occasion, which referred his and he to Porphyrocominae. The latter, however, are mixture of which do have a genera not seem to a single character in and common, as Poikilacanthus, moreover, does not show a definite of the included in affinity to any genera this subtribe, it is difficult to I see why it was put there. Its seeds, which could not study myself, are figured by Lindau as smooth and distinctly marginate, and remind 654 C. E. B. BREMEKAMP

is one of those of some of the Ruellieae, where the marginal thickening due the of hairs. Howsoever this much to presence mucous may be, so is certain that seeds of this kind are not found in any of the groups with The which the Porphyrocominae are supposed to be related. finely reticulated pollen grains of Poikilacanthus are ellipsoidal and provided with 4—6 equatorial pores, and in the two last-named characters they doubtless different the of are widely from globose sparsiporous ones

Oreothyrsus. In fact, there seems to be no reason whatever to regard related. these two genera as nearly

The armature of spinules by which the pollen grains of the two New-Guinean is in far of it species are covered, so importance as may have been responsible for the disappearance of the relief shown by the pollen grains of the Philippine species. A similar phenomenon is observed in the genus Rhacodiscus Lindau, where the pollen grains of covered with but the some species are spinules lack typical relief of

the the to which this is to be whereas Justiciinae, group genus referred, relief the pollen of some other species shows this quite clearly; this pollen, however, lacks the spinules. If a similar development had then the taken place in the genus Oreothyrsus, structure seen in the will pollen grains of the Philippine species have to be regarded as the

original one. This structure, however, is essentially the same as that with shown by the globose or ellipsoidal pollen grains provided three

that are met with in several subtribes of the and it pores Justicieae, be therefore that in of the latter for the might one a place genus is be To Oreothyrsus to found. determine this place we will have to turn

our attention to other taxomonically important parts. Among the latter the seedcoat occupies a very prominent place. The seeds of the two New-Guinean species are unfortunately not

known in sufficient detail so that we will have to base our conclusion of Here the on those the Philippine one. seedcoat proves to consist

of cells whose walls show tracheidal thickenings. This structure has

far been observed in the and in two of which so Rhytiglossinae genera in tribe the exact position the could not yet be determined, viz. in Polytrema C. B. Clarke and in Leptostachya Nees emend. C. B. Clarke et On of the different Stapf. account totally pollen type a very near the and affinity between Rhytiglossinae Oreothyrsus seems excluded.

Still it is worth while to remember that Lindau had drawn the atten-

tion the in habit between his New- to very striking similarity

Guinean species and the plant for which Nees had used the name

Raphidospora dichotoma (Bl.) Nees, but which, as Lindau rightly

then it remarked, can not belong to that genus; since was found out it is that a Strophacanthus, viz. Str. membranifolius (Miq.) Brem. Stropha-

it should be is to the canthus, remembered, a genus belonging Rhyti- habitual glossinae. The similarity rests i.a. on the structure of the

in both inflorescences, which genera are once or twice dichasially branched with suppressed terminal flowers. However, the difference between the 2-porous discoid pollen grains of the Rhytiglossinae with

their clearly differentiated marginal zone and the globose, sparsiporous

pollen grains of Oreothyrsus seems to be so fundamentalthat it would ACANTHACEOUS GENERA AND SPECIES OF CONTROVERSIAL POSITION 655

between this hardly be justifiable to assume a very near affinity genus and the Rhytiglossinae.

In the genera Polytrema and Leptostachya the pollen structure comes that it in the nearer to of Oreothyrsus than does Rhytiglossinae. The pollen grains of Polytrema are globose and probably sparsiporous, and those and of Leptostachya shortly ellipsoidal or subglobose provided, with of three pores situated in furrows. The inflorescence Leptostachya, which is either a spike or a panicle consisting of spikes, is however so unlike the once or twice pseudo-dichotomously branched inflores- of that between these the cences Oreothyrsus, genera too affinity can be The of not very near. inflorescences Polytrema, on theother hand, are is still cymous, and this genus, which unfortunately very imperfectly be known, might therefore more nearly related.

made The investigations on which this paper was based, were

possible by a grant from the Netherlands Organization for Pure Research (Z.W.O.).