Acta Botanica Neerlandica 6 (1957) 351—377

On the Position of Platycarpum Humb. et Bonpl., Henriquezia Spruce ex Bth. and

Gleasonia Standl.

C.E.B. Bremekamp

(.Botanical Museum, Utrecht)

(;received February 15th, 1957)

Introduction

Some time the “Institute do ago Agronomico Norte, Belem, Para,

Brazil” sent me a set of specimens for identification among which I detected Bth. This a new species of Henriquezia Spruce ex discovery

induced me to study once more and now in somewhat more detail the relation between this and the related genus nearly Platycarpum Humb. and the these with et Bonpl. position two genera occupy the rather regard to habitually similar Gleasonia Standi., a subject I to which had already paid some attention at an earlier occasion,

and I be the on which had reported in a note which is to found at base of 16 of work “the African of Oldenlandia L p. my on Species

sensu Hiern et K. Schumann” in Verh. Kon. Ned. Akad. v. Wetensch.,

Sect. 2, 48, no 4, 1952. By the good office of Dr Bassett Maguire I

received on loan from the New York Botanical Garden the rich

materials by which, as a result of the Botanical Garden’s recent these in its expeditions to Tropical America, genera are represented herbarium, and during a recent visit to England I could study also the valuable collections in the herbarium of the Royal Botanic

Gardens, Kew. In the course of these studies I discovered among the

Platycarpum specimens too a new species. Descriptions of the latter

and of will be in my new Henriquezia found further on this paper. but of the main They are, however, secondary importance; object

of this study is to ascertain as precisely as possible the taxonomical of these three rather position puzzling genera.

In the note on p. 16 of my work on “the African Species of Olden-

landia L sensu Hiern et K. Schumann” I stated that Gleasonia Standi,

“on account of the large and flat exalbuminous seeds and also because of the of very peculiar structure the testa is to be referred to the have Henriquezieae, a tribe that will to be removed as a separate family the In of this to Tubiflorae.” support opinion I referred to Fig. c, d, and f e of Tab. V, which represent the testa of Henriquezia verticillata

Spruce ex Bth., of Platycarpum orinocense Humb. et Bonpl. and of two species of Gleasonia, viz. Gl. uaupensis Ducke and Gl. macrocalyx Ducke.

The study of the more abundant and better material that I had has led now at my disposition to a somewhat different standpoint. The between and the other differences Gleasonia two genera appeared be taxonomical to of far greater importance than the points of resem-

351 352 C. E. B. BREMEKAMP

and shake the blance, although this more detailed study did not I had arrived with the opinion at regard to genera Platycarpum and viz. that be Henriquezia, they are to regarded as representatives of a new family Henriqueziaceae, for which on account of its affinity with the Bignoniaceae, Pedaliaceae, Martyniaceae, Thunbergiaceae, Mendonciaceae and Acanthaceae, a position will have to be found in the order Tubiflorae, I that the Gleasonia be now see genus can not regarded as a very near ally, and that it is either to be left in the Rubiaceae, where it occupies, the seeds and the on account of exalbuminous large size of the coty- ledons, an anomalous position, or else to be referred to a new family, which however, on account of the structure of flower and fruit and of the presence of colleters on the inside of the stipules, would have be in the This to placed the near vicinity of Rubiaceae. implies, of course, that the points of resemblance between Gleasonia and the and which I based former genera Platycarpum Henriquezia, on my I view, are no longer accepted by me as homologies, but that now regard them as mere analogies. The with taxonomical change in my opinion regard to the position in of Gleasonia is due to a better insight the characters in which this

differsfrom the other and in in genus two ones those which it resembles them. In order facilitate I of all to a discussion, will first give an of enumeration the differences, as the latter are in this respect of more importance than the points of resemblance.

Comparison between Gleasonia and the genera Platycarpum and Henriquezia.

The principal differences between Gleasonia and the two other the absence in Gleosonia of the that genera are: 1) peculiar “glands” in Platycarpum and Henriquezia are found at the base of the petiole, and which in fact are found nowhere else in the whole kingdom of plants; 2) the presence of colleters or resin glands on the inside of the of character which this shares with the stipules Gleasonia, a genus great majority of the Rubiaceae, and their absence in Platycarpum and of the that of Gleasonia Henriquezia; 3) a different structure corolla, being actinomorphous with induplicate aestivation of the lobes, that of other with streak the two genera zygomorphous a longitudinal of hairs on the ventral side of the throat and with ascending aesti- which vation (cf. Fig. 1); 4) a different structure of the androecium,

Fig. 1. Floral diagrams of A. Gleasonia duidana, B. Henriquezia verticillata and C. Platycarpum orinocense. ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 353 in Gleasonia is subregular, in Platycarpum and Henriquezia distinctly zygomorphous, the filaments being inserted at unequal height, bent

at the base and of unequal length; 5) a different arrangement of the ovules those of Gleasonia in each cell (cf. Fig. 2), forming ovary a

doublerow which ascends to about midway the dissepiment, whereas the two to four ovules found in each of the cells of the ovary two other inserted the genera are at nearly same height; and 6) a different of the that of Gleasonia development capsule, (Fig. 9, p. 000) remaining whereas that of the other becomes either entirely inferior, two genera

11, or even almost semi-superior ( Henriquezia , Fig. p. 000) completely 10, Less of superior ( Platycarpum , Fig. p. 000). important points those difference are found in the arrangement of the pollen grains, of Gleasonia (Fig. 3) remaining united in tetrads, whereas they are free in the set two other genera, and also in their structure, those of Gleasonia those of the always being 3-colpate, other genera (Fig. 4) either 3- or 4- or 5- or 6-colpate, in the number and shape ofthe seeds,

those of Gleasonia being more numerous and more or less swollen and whereas the four seeds of the other angular, two to two genera are flat, and in the structure of the testa, the cells being convex in Gleasonia, and produced into papillae in Henriquezia (Fig. 5) and Platycarpum (Fig. 6). Before proceeding to the points of resemblance, it will be well to discuss these differences, the more important ones as well as the others, in some detail. The which in and “glands” Platycarpum Henriquezia are seen at the base of the petiole on the side facing outwards, are large glabrous and shiny spots slightly protruding beyond the surrounding tissue; in black herbarium material they at once draw the attention by the colour they assume in drying. Their exact nature is unknown: it is possible that they secrete a sugary fluid, and in that case they would

to the class of extra-floral but it also be belong large nectaries, may

that they are bacterial nodules; without living material it will pro- be difficult settle this Howsoever this these bably to point. may be, petiolar “glands” certainly are a very characteristic feature of these like of which found nowhere else. two genera, the is The of colleters resin the inside of the presence or glands on stipules with the of Gleasonia is a character which this genus shares majority of the Rubiaceae. In shown Krause fact, they are, as was by (Ueber

harzsecernierende Driisen an den Nebenblattern von Rubiaceen, in Ber. d. d. bot. Gesellsch. 27: 446-452. 1909), found in all tribes of this family with the exception of the Rubieae (Galieae). Their absence

in of the the in this tribe might be an argument favour view that leaf-like of these modified supernumerary parts plants are no stipules, is as now commonly accepted, but that they are true leaves. Krause colleters would be in the According to present genus but mistake. I did Henriquezia too, this must be a Although not in- nitida vestigate the species quoted by him, the apparently very rare H. I H. Spruce ex Bth., studied three other species, viz. verticillata Spruce

ex Bth., H. macrophylla Ducke and H. longisepala Brem. v. infra, and in but I addition several species ofPlatycarpum, in none of them could 354 C. E. B. BREMEKAMP

therefore find the slightest trace of these structures. I suppose that the specimen studied by Krause will have been misidentified, and in other that it belonged reality to some genus.

It is noteworthy that although colleters are by no means confined to the Rubiaceae (cf. Hanstein in Bot. Ztg 26: 697-713, 720-735, 744-761 and 768-787. in this of rather 1868), they are family a peculiar be structure which returns only in the Cunoniaceae, a family that may

related to the Rubiaceae and in the Viola. That the distantly , genus is colleters of Gleasonia too are of this type, certainly a weighty argu- in favour of the view that this either the Rubia- ment genus belongs to else for which will have ceae themselves or to a new family a place

to be found in the near vicinity of the latter. That the corolla of Gleasonia is actinomorphous is in itself perhaps

but in connection with the not very important, valvate, or more precisely induplicate, aestivation of its lobes it is certainly not easily reconcilable with the view this would be of that genus a near ally Platycarpum and Henriquezia, where the corolla is more or less distinctly the zygomorphous with a longitudinal strip of hairs on ventral side of the throat, and where the aestivation is ascending-imbricate. of corolla would exclude the The zygomorphism the not genera

Platycarpum and Henriquezia from the Rubiaceae, for slightly zygomor-

phous corollae are in this family certainly not unknown, although

the side the throat. they never show a band of hairs on ventral of

A slightly unequal length of the filaments is here, as an accompanying either. As of such feature, not uncommon examples a slight zygomor-

phism of the androecium corresponding to a curvature of the upper of the refer the Pallasia part corolla, we may to genera Klotzsch, Macrocnemum Wedd., Ferdinandusa Pohl and Posoqueria Aubl. This it does form of zygomorphism, however, is of little importance as the the not obscure fundamentalsymmetry of plan according to which the flower is built. In this respect it is of hardly more value than the of the that is often with in zygomorphism calyx so met genera belong- ing to this family, and which is due to the development of one of the

calyx lobes into a foliaceous appendage, a form of zygomorphism

that does the the and is not affect other parts of flower, moreover restricted few flowers inflorescence. The androecium of to a per Platycarpum and Henriquezia, however, shows a fundamentally different form of zygomorphism, which finds its expression in the insertion

the filaments and in of of at unequal height a curvature their basal Rubiaceae part. Among the genera for which a position in the has been the number of those that with claimed, are provided a zygomor-

phous androecium of this kind is very small indeed, and there is in all these instances (e.g. Tammsia Karst., Aitchisonia Hemsl.) good

reason to doubt the correctness of their classification; in none of them, the filaments bent their base. moreover, are at The the difference in arrangement of the ovules is another argument the against view that Gleasonia would be a near ally of Platycarpum and indica- Henriquezia (cf. Fig. 2), but the arrangement itself gives us no of tion with regard to the taxonomic position either of Gleasonia or ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 355

the two other for similar to that in Gleasonia genera, arrangements seen as well as arrangements of the kind observed in Platycarpum and found in families different families Henriquezia are belonging to very and orders.

The of Gleasonia is of for which capsule (Fig. 9, p. 364) a type we

might easily find a place within the range of those that are met with

Fig. 2. Ovary of A. Gleasonia duidana, B. Henriquezia longisepala, and C. Platycarpum orinocence. Note the the in the cells of ridge opposite placenta ovary Henriquezia and Platycarpum.

the but the almost among Rubiaceae, semi-superior or completely

superior capsules found in Henriquezia (Fig. 11, p. 373) and Platycarpum of aberrant kind. Nevertheless their (Fig. 10, p. 372) are a more deviation from the ordinary inferior type of capsule, although accen- the width of which these tuating once more the gap separates two

genera from Gleasonia, would in itself not be sufficient to exclude these from the deviation tribe genera Rubiaceae, as a similar occurs in the in the Gaertnerieae, whose position family must, on account of its very close resemblance to the Psychotrieae, be regarded as well-established.

The less important differences mentioned at the end of our list, in are found the pollen characters and in number, shape and structure of the seeds. That the pollen grains of Gleasonia (Fig. 3) remain united in tetrads, is somewhat doubtless a unexpected feature, which sharply differen- tiates this genus from Platycarpum and Henriquezia, but whose taxo- nomical be overrated. It importance should not is certainly not a character on account of which the would have to be excluded genus the in unknown from Rubiaceae, for pollen grains united tetrads are not in this the Randia family; they are e.g. a general character of genus of the indi- (Houst.) L sensu Fagerlind. The uncommonly large size vidual 70 in and 110 in grains (diam. /x Gl.uaupensis /jl Gl.duidana) is in this respect perhaps of somewhat greater importance, as the

of the Rubiaceae the rather even pollen grains are, on whole, small; the that far found largest ones are known so (they were in the genera L. and those Richardia (Houst.) Palicourea Aubl.) are not so bulky as of Gleasonia. That the pollen grains of Platycarpum and Henriquezia too are of of resemblance between these very large, is, course, a point 356 C. E. B. BREMEKAMP

Fig. 3. Pollen tetrad of Gleasonia uaupensis.

Fig. 4. Pollen grains ofA. Platycarpum orinocense, and B. Henriquezia macrophylla. The figure to the right of A and that between the

two pollen grains of B are optical sections ofthe wall. 357 ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA

genera and Gleasonia, but not one of any importance, as large pollen with in different circles of That the grains are met very affinity. pollen grains of Platycarpum and Henriquezia (Fig. 4) are usually with than three is the provided more colpae, on other hand worth

noting, not only because it constitutes a difference with Gleasonia, but also because pollen grains with more than three colpae (or pores) characterize of rank far that ofthe usually groups a exceeding genus. of similar It is also noteworthy that pollen grains a very type are met with in several of the families of the Tubiflorae with which the of resemblance genera Platycarpum and Henriquezia show points

Pedaliaceae, Mendonciaceae, see on the structures (.Bignoniaceae , pollen with “The Position of the met in these families my paper on genus

Thomandersia Baill.” in Rec. d. trav. bot. neerl. 39: 166—175. 1942).

in of the The differences number and shape seeds are not very either and somewhat obscured the important moreover by very striking points of resemblance, viz. the absence of endosperm and the of the that strong development cotyledons, points we will discuss

hereafter. The seeds of Gleasonia are more numerous than those of

the two other where each of the fruit cells contains but genera, one to four seeds, and they are angular and less strongly compressed. The in the the of difference structure of testa is doubtless greater importance. The testa cells of Platycarpum (Fig. 5) and Henriquezia

Fig. 5. Testa of Platycarpum orinocense. Fig. 6. Testa of Henriquezia verticillata.

those of (Fig. 6) are drawn out into papillae or short hairs; Gleasonia The of the are merely convex. presence papillae on testa was known

so far only for Platycarpum, but a renewed study of the testa of Henri-

quezia verticillata Spruce ex Bth. has led to the conclusion that they in this That overlooked are present genus too. they were originally

was due to the circumstance that the seed investigated at that time 358 C. E. B. BREMEKAMP was fully mature; at that stage the outer wall of the testa cells has the wall shown in is the reticulate disappeared; my original figure inner I here in one. For this reason insert a new figure which the outer wall with the papillae is shown. An exactly corresponding structure of the testa is met nowhere in the Rubiaceae, but this would in itself not be enough to exclude these genera, as the variability in of the is in that the structure testa so large this family it is as yet decide what would fall impossible to outside its range. In the Gleasonia the of the be of genus structure testa proves to an entirely differentkind. The two figures given in my earlier commu- nication are unsatisfactory, as the seed of Gl. uaupensis Ducke of which

I studied the testa, was not yet fully mature, and as that of Gl. macro- Ducke that the sclerization of the wall had calyx was so young outer not yet begun. Ripe seeds of Gl. macrocalyx were this time too not available, so that I am still unable to tell in how far the structure of their testa agrees with that seen in the two other species. The wartlets the of which earlier on testa, one was figured in my publication, are at any rate a feature by which these seeds differ from those of the two other species. Of the latter, Gl. duidana Standi., the species on which the and Gl. mature seeds genus was founded, uaupensis, fully could now be studied. The testa cells of these two species have a convex outer which be with wall, proves to provided an irregular network of thickenings (Fig. 7), which in fully mature seeds extends

Fig. 7. Testa of A. Gleasonia duidana, and B. Gl. uaupensis.

over the lateral walls too. The nearest approach to this structure

found in is Hillia Jacq., a genus which occupies a rather isolated in the but i.a. of the position family, which, on account presence of raphides, can certainly not be regarded as a near ally (see for this work “the African of Oldenlandia L genus my on Species sensu

Hiern et K. Schumann” p. 14 and 15 and Tab. V fig. a); the cells the of different and their of Hillia testa, however, are a shape, lateral ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 359

walls are much thicker. The structure of the testa cells shows that

the if it is be included in the will have genus Gleasonia, to Rubiaceae, to be referred to a tribe or even to a subfamily ofits own, and that it be the done certainly can not put in Rondeletieae, as was by Standley. In this tribe the basal wall of the testa cells is with provided very which that it shows the large round or oval pits, means structure which is characteristic for the subfamily to which this tribe belongs, viz. the Cinchonoideae the subdivision of sensu meo (see my paper on the 5 family in “Rapports et Communications aux Sections 2, 4, et 6, Huitieme Congres de Botanique: 113, 1954), and these large pits are in Gleasonia completely absent.

Now that we have finished our discussion of the indications the

differences between Gleasonia and the and Henri- genera Platycarpum

quezia may give us with regard to the taxonomical position of these

it will be well to attention the of genera, pay some to points resem- blance in order to find out whether the latter do not invalidate the conclusions we have arrived at. main of resemblance between The points the three genera are the found in large size of the pollen grains, a feature to which, as we have already pointed out, not much value can be attached, in the

absence of endosperm in the ripe seed, in the structure of the embryo with its large cotyledons and its small axial part, in the shape and of of in the mode dehiscence the capsule, structure of the testa, in

the pattern formed by the stomata and the surrounding epidermis cells the lower side of the and in the of resin on leaf, presence large cells in the twigs just outside the cambium. Exalbuminous seeds far known in ofthe subfamilies were so not any of the Rubiaceae, although the amount of endosperm in the seeds of

the Guettardoideae is but small, the embryo being of a larger size than is usual in the family. That the seeds of Gleasonia, Platycarpum and with each other in that Henriquezia agree being exalbuminous, means we experience with each of these genera in this respect the same when find for them the difficulty we try to a place among Rubiaceae, but as exalbuminous seeds occur in widely different circles of affinity, does that it certainly not mean they should of necessity be regarded

as nearly allied. That the in with the axial cotyledons are very large comparison part of the embryo is another feature on account ofwhich it is difficult find for these the for where to genera a place among Rubiaceae, even in the embryo of plants belonging to this family fairly large cotyledons in the Simira Aubl. are noted, as genus ( Sickingia Willd.), they prove be than the axial to hardly longer part. Howsoever this may be, this taken feature too can not be as proof of a near affinity between

and Gleasonia the genera Platycarpum and Henriquezia, as embryos of this kind confined taxonomical are certainly not to a single group. The resemblance in the shape of the capsule and in its mode of dehiscence need detain The resemblance in not us very long. shape is rather superficial, for the capsule of Gleasonia is completely inferior and by no means so strongly flattened as the semi-superior or entirely 360 C. E. B. BREMEKAMP

of the two other and the resemblance in superior capsules genera, the mode of dehiscence, which in all three is loculicidal, is taxonomi-

of loculicidal return in almost cally no importance, as capsules every order of the Angiosperms.

The in the I resemblance the structure of testa, by which was

originally so strongly impressed that I overemphasized its taxonomical after detailed be importance, appears now, a more investigation, to of It is that in all three the a rather superficial nature. true genera ofthe the walls testa cells are strengthened by a network of thickenings,

but this network is in Gleasonia met with in the convex outer wall

and the lateral whereas it is found in the other ones, two genera on the basal wall; however, it must be admitted that, at least in Platycar- in the which of the pum, it is also present papillae, are part outer wall.

However, as testa cells with reticulate walls return in very different families, their taxonomical importance should not be overrated. It adduced for these can not even be as an argument excluding genera in Hillia the consists from the Rubiaceae, as the genus Jacq. too testa of reticulate cells.

The position of the stomata with regard to the surrounding epi-

the underside the is in all three the dermis cells on of leaf genera form with the same (Fig. 8). The pattern they these cells is one that

underside the leaf of A. Gleasonia Fig. 8. Epidermis on the of macrocalyx, B. Platy- C. carpum negrense, and Henriquezia verticillata.

is for the Rubiaceae. does not that typical This, however, prove they belong to this family, for this pattern is not confined to it. It returns i.e. in which of the e.g. in some of the Bignoniaceae, a family to two be allied. three genera are supposed to resin cells in the is similar The presence of very large phloem a character. Although such cells are apparently not of common occur- unknown in that rence in the Rubiaceae, they are not family; they

found in the Gardenieae. occur also in several other are e.g. They families. Precise knowledge with regard to these structures is, however,

herbarium suitable for difficult to obtain, as specimens are not very their study.

Besides the points in which Gleasonia resembles Platycarpum as well there is also in which it resembles of as Henriquezia a point only one ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 361

viz. It is them, Platycarpum. a very striking one, and it is found in the which in both these and stipules genera are large intrapetiolar.

Especially the intrapetiolar position is a most remarkable feature, time for intrapetiolar stipules, although at one thought to be not uncommon in the Rubiaceae, are in reality, as I have pointed out at an earlier occasion (Bull. Jard. Bot. Buitenz. Ser. 3: 13: 425. 1935) In the had very rare. fact, among genera in which they been recorded in the older literature Capirona Spruce ex Bth. proved to be the only in one which they are actually present. I must admit, however, that the of Gleasonia that time description was at overlooked by me, and add that with the of reliable regard to stipules Platycarpum no infor- I mation was as yet available. myself found them at that time in Didymoecium Brem. and recently also in Suteria DC. It is noteworthy

that these genera belong to different tribes and partly even to different and that neither show in subfamilies, Gleasonia nor Platycarpum other well-marked resemblance of them. The respects a to any one taxo- nomical value of the intrapetiolar stipules must therefore be regarded and their in Gleasonia well in as very small, presence as as Platycarpum can therefore hardly be interpreted as indicating affinity between them. that It seems we are confronted here with a mere analogy, and this supposition finds support in the fact that the intrapetiolar of their stipules Platycarpum appear to owe origin to the splitting of an originally closed sheath, whereas those of Gleasonia seem to be free from the The of start. stipules Gleasonia, moreover, differ, as we

have already seen, fundamentally from those of Platycarpum in the of colleters. In the absence of the latter there presence is complete agreement between the stipules of"'Platycarpum and those of Henriquezia, which at first sight look quite different, in the first place because

they are rather narrow, and in the second place because there are them each two of to leaf. There is, however, agreement in so far that in sheath Henriquezia too the stipules are at first united into a which completely envelops the terminal bud.

The position of the genus Gleasonia

If the we try to determine position of the genus Gleasonia by the

aid of the data given above, we come to the conclusion that the latter

of the agrees in most its characters with notion we have, in the course of time, arrived at with regard to the family Rubiaceae; especially im- in this portant respect are the decussate leaves, the stipules provided with the flowers colleters, actinomorphous with inferior ovary and

sympetalous corolla, and the simple whorl of stamens alternating with corolla The in the lobes. points which it deviates from our notion

of the this family are the exalbuminous seeds, structure of the embryo with its large cotyledons and comparatively small axis, and also,

though this seems to be of somewhat less importance, in the large

size of the pollen grains. The problem with which we are confronted, therefore be formulated in this should extend the limits may way: we far of the family Rubiaceae so that it becomes possible to include the 362 C. E. B. BREMEKAMP

is it to create for this aberrant genus Gleasonia, or preferable genus

a new family? is of The absence of endosperm doubtless a character great taxo- the nomical importance. It is true that amount may vary to some definite it does in the Rubiaceae extent in a group, as e.g. themselves, where the with but we have one subfamily, Guettardoideae, a compara- albuminous tively thin layer of it, but the difference between and

exalbuminous seeds is nevertheless, as a rule, well-marked, and it

that has in to what happens but rarely one any difficulty deciding

be Moreover, to a certain rank category a seed should referred. up in well-defined as a but one of we find taxonomically groups, rule,

these kinds of seed. That the seeds of Gleasonia are undoubtedly

for this exalbuminous, is therefore a strong argument removing genus will have from the Rubiaceae. If we will not go that length, we to the widen the delimitation of family, for as there is no place for our be in of the a new one will have to created, genus any subfamilies, and to make place for the latter, the limits of the family will have delimitation of the is to be extended. If we decide that the family of first of all have find better left unchanged, we will, course, to out there for the in of the other families of whether is a place genus one the Rubiales, for in view of the important points of resemblance with it this order. the Rubiaceae s.s. it need not be doubted that belongs to

with the As there is here but one family exalbuminous seeds, viz. Valerianaceae, this question is easily settled: in the Valerianaceae there the latter have leaves is certainly no place for our genus as exstipulate and zygomorphous flowers with an incomplete androecium and value the gynoecium. The great taxonomical of points of resemblance between Gleasonia and the Rubiaceae creates a strong link between them, and brings them in opposition to the rest of the Rubiales, which

form that would have to be as means that they a group recognized formulate also a suborder. For this reason we might our problem

should for Gleasonia a new and for the in this way: we create family, latter and the Rubiaceae s.s. a new suborder, or should we leave the

in the of which in this the delimitation should genus Rubiaceae, case be widened?

for formulated in the In order to find a solution the problem pre- find whether the differences ceding paragraph, we will have to out between Gleasonia and the subfamilies that so far have been recognized

of taxonomical than within the Rubiaceae, are greater importance between of for if the differences between those existing any two them, of the be of there some latter would prove to greater importance, for Gleasonia. would certainly be no reason to create a new family

other the between these subfamilies would If, on the hand, differences prove to be of less importance than those between Gleasonia and the whole then the creation of a mere would group, subfamily hardly be sufficient. The main points in which Gleasonia differs from the whole set of subfamilies are the exalbuminous seed and the structure of the embryo

It with its large cotyledons and short axis. differs moreover from the ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 363

Cinchonoideae sensu meo in the absence of the large oval or orbicular

pits on the basal walls of the testa cells, from the Urophylloideae by the

much thinner, not finely perforated walls of the testa cells, from the

Ophiorrhizoideae in the absence of the peculiar wartlets on the walls of the testa cells, from the Guettardoideae in the entirely different fruit and also in the absence of crystals of calcium oxalate in the walls the from the of hairs by which the various parts are covered, Ixoroideae different of the of the which in the totally structure upper part style, and does not function as “receptaculum pollinis”, from the Rubioideae by the absence of raphides. That Gleasonia is fundamentally different from each of the subfami- lies that hitherto have beenrecognized in the Rubiaceae, can therefore and not be doubted, when we were to obey our first impulse, we would hardly hesitate to decide that it is more fundamentally different from whole than two of them from each other. We the set any are should realize, however, that all these subfamilies consist of more of them of and that than one genus, most even a very large number,

the differences between them would doubtless be more numerous and therefore more impressive if each of them was represented, like

the which Gleasonia a For group to belongs, by single genus only. it decidethe in this reason seems hardly possible to question an entirely and the introduction of objective way, as a new subfamily answers and will our present needs, probably meet with less resistance than the introduction of a new family would do, it seems prudent for the time further. being to go no The exposition given above is summarized in the following des- cription of the new subfamily.

Gleasonioideae nova subfamilia Rubiacearum a subfamiliis aliis seminibus exalbuminosis et embryone cotyledonibus parte axili ma- joribus instructo faciliter distinguenda.

Arbores. Folia opposita, petiolata. Stipulae magnae, intrapetiolares, simplices. Inflorescentiae terminales, corymbi- vel paniculiformes.

Flores pentameri, actinomorphi, magni. Ovarium biloculare, utroque loculo ovulis pluribus instructo; ovula in serie verticali duplici dimidio inferiore septi inserta. Calyx regularis, in lobos longos partitus; lobi anthesin decidui. Corolla satis tubo albi, post magna, extus sericea, infundibuliformi sensim in faucem ampliato, lobis aestivatione in- duplicativa. Stamina fauceinserta; antherae exsertae. Granula pollinis

in in dua tricolpata, magna, tetrades unita. Stylus stigmata exeuns, breviter exsertus. Fructus (Fig. 9) capsula complanata, loculicide

loculo dehiscens, utroque seminibus pluribus instructa. Semina angu- losa, exalbuminosa; embryo cotyledonibus crassis parte axili majoribus instructus; testa e cellulis convexis composita, pariete externa reti- culatione laxa invigorata. unico Americae T incola. Subfamiliaadhuc e genere constans, ropicalis As this but it also contains but subfamily comprises a single genus, describe if it should a single tribe. It seems superfluous to the latter; of would have be Gleasonieae. receive a name, this, course, to Gl. The genus Gleasonia Standi, comprises at present three species, 364 C. E. B. BREMEKAMP duidana Standi., Gl. uaupensis Ducke and Gl. macrocalyx Ducke, all three confined to Tropical South America, the first to Amazonian Venezuela, the two others to Amazonian Brazil. Gl. duidana and Gl. uaupensis are now comparatively well-known; they are doubtless nearly related species, differing but slightly in the length of the petioles, the size of the leafblade and of the stipules, the length and width of the calyx

Fig. 9. Capsule of Gleasonia uaupensis.

lobes, the size of the the size of the corolla, pollen grains (cf. p. 356) and the of the Gl. Ducke structure testa; uaupensis, moreover, is, as

already reported, a myrmecophilous species, the concave stipules sheltering small red ants. The third species, Gl. macrocalyx, differs conspicuously from the two others in the absence of the hirsute in the much size ofthe and in the indumentum, larger fruit, presence of wartlets the this is still on testa; species unfortunately very imper- in that far have been fectly known. As all specimens so collected, the

corolla and them had already been shed, as none of were provided with the characters of the and fully ripe fruits, very important pollen of the mature testa could not yet be studied. So long as these characters it is decide whether this are unknown, hardly possible to species

really belongs to the genus Gleasonia. However, as its seeds are exal- buminous and as its embryo is provided with large cotyledons, it is

not to be expected that itwill have to be referred to anothersubfamily.

The position of the genera Platycarpum and Henriquezia

The absence of endosperm and the large size the cotyledons reach in comparison with the axial part of the embryo are two important

characters in which, as we have seen, the genera Platycarpum and Henriquezia resemble Gleasonia and in which they all three differ from the general plan of the Rubiaceae, but Platycarpum and Henriquezia. from the latter also in in differ some important points which Gleasonia

is These the consistent with it. points are 1) absence of colleters on the inside of the the of band of hairs stipules, 2) presence a on the ventral side of the corolla throat and the distinctly zygomorphous and the almost androecium, 3) semi-superior or even completely ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 365

superior capsule. Each of these deviations from the general plan

would already in itself form an important argument against the in- clusion of these but their becomes genera, put together weight so that it will overwhelming crush every attempt in this direction. The of of hairs the ventral side of presence a strip on the corolla throat and the distinctly zygomorphous androecium, moreover, point towards circle an entirely different ofaffinity, viz. that of the Tubiflorae, where we find in the Bignoniaceae, Pedaliaceae, Thunbergiaceae, Mendon- ciaceae and Acanthaceae of families with which the a group genera and remarkable Platycarpum Henriquezia show a quite resemblance in in is habit; fact, this resemblance much more pronounced than that with the Rubiaceae.

When Humbold and Bonpland described the genus Platycarpum 104. (Plantae aequinoctiales 2: 81, tab. 1809) they were apparently so impressed by the resemblance meant in the preceding paragraph that their without hesitation in they placed new genus any the Big- noniaceae. On account of the inferior the of the ovules ovary, position and the of this of presence stipules, is, course, an unacceptable decision, but these characters were either misunderstood or overlooked by them. It be doubted that the can not they regarded ovary as superior, for they described the calyx as inferior, probably because they found the scar of the latter near the base of the capsule, and did not realize that far the of the latter its by greater part owes origin to the very strong development of the portion inside the calyx; and that they the overlooked presence of stipules is also easily comprehensible, as the latter are early deciduous and, as a rule, present only round the terminal bud; it is quite possible that all the branches they had collected ended in in inflorescences, and that case probably not a have been That the single stipule would present. seeds are described with membranaceous as provided a margin, is not correct either; it seeds where the inside the is applies only to young space testa not filled the A yet completely up by embryo. similar mistake was made Bentham he described the seeds of by when Henriquezia as marginate, though to be quite fair to him I must admit that he cautiously added

the words “ut videtur”. Here too in the fully mature seeds the whole

inside the space the testa is occupied by embryo, and there is no trace Bentham of a wing. J. D. Hooker (in et Hooker, Gen. PL 2: 12.

1873), to be sure, was entirely mistaken when he described the seeds of the Henriquezieae, the tribe created by him for the reception of the and genera Henriquezia Platycarpum, as “latissime alata”. The authors who like Fenzl (Denkschr. d. k. bayer. bot. Gesellsch. 3: 239 et 265. 1841), de Candolle (Prodr. 9: 233. 1845) and Bureau (Monographic des Bignoniacees: 80-81 et 103. Paris 1864) in the half next century occupied themselves with Platycarpum, accepted Humbold and Bonpland’s classification, but they too overlooked the of and the inferior of the presence stipules position ovary. When Bentham described his new genus Henriquezia (Kew Journ. of Bot. 6: 337. he mentioned its 1854), duly near affinity with Platy- carpum, and referred it, just as his predecessors had done with the 366 C. E. B. BREMEKAMP latter, to the Bignoniaceae, and like them he failed to specify his reasons. It is well when look the he to remember this we at arguments adduced allied. in support of his view that the two genera are nearly They read follows; and as “The five equal and perfect stamens, the short broad fruit with very few seeds, are the same, as well as the general

and leaves.” If the to the habit simple two genera really belonged Bignoniaceae, these characters would suffice to contrast them with the rest of the family, and this in its turn would bring them in closer contact with each other. However, as soon as we drop the notion characters deal that they belong to this family, these loose a great of then be features of rather their value, for they appear to a general in different circles of This nature, which return very affinity. applies especially to the first-named characters, the five perfect stamens and of short broad the simple leaves, but even the presence and fruits with small number of seeds not be a can regarded as a very special character, and it looses, moreover, much of its value when we realize that the of the in other capsules two genera are respects quite different, of be almost and for that Platycarpum proves to completely superior in dehiscence the right and left half of each valve fold backwards

until they meet, whereas the capsule of Henriquezia is only semi- superior and its valves in separating do not appreciably change their form. This difference in the mode of dehiscence was unknown to the earlier authors who had no ripe fruits at their disposition. When

Bentham said that the stamens in the two genera are all equal, he in It is was mistaken, for they are reality quite distinctly unequal. rather unfortunate that this inequality escaped his attention, and also that the overlooked the bent at the base of the filaments, for the

of androecium of this kind is one of the presence a zygomorphous of resemblance between these and the most important points genera Bignoniaceae and their nearest allies. between The more important points of resemblance Henriquezia and because Platycarpum were omitted by Bentham, partly, no doubt, he regarded them as general characters of the Bignoniaceae, the family which the referred and because to genera were by him, partly, they the first the were overlooked. To group apparently belong flat, exalbuminous seeds, which erroneously were regarded as winged, and the structure of the embryo with its large and flat cotyledons; to the second the remarkable “glands” at the base of the petiole, the

of the but an- presence stipules, complete distinctly zygomorphous droecium, and the collateral ovules. That he overlooked the impor- tance of the petiolar “gland” is rather surprising, because the presence of verticillata of this organ is in his description Henriquezia duly recorded; that he would in have overlooked its presence Platycarpum, seems I that the omission therefore hardly possible. suppose was merely accidental.

the of difference between The way in which he summarizes points the is either. to him two genera not very satisfactory According Henriquezia differs from Platycarpum in “the semi-adherent calyx with and four and the of the a persistent base only lobes, shape corolla, ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 367

much than that of The so more Bignoniaceous Platycarpum" . calyx,

however, is not semi-adherent but totally adherent; only when the

fruit the shifts from the to begins to grow out, calyx gradually top and the and its base a position midway between the latter base; can I hardly be called persistent: in this respect see no difference at all of between the two genera. That some the Platycarpum species too

but four he could as these had possess calyx lobes, not know, species not yet been collected. The corolla of Henriquezia is much larger than that of Platycarpum, but in shape it is hardly different from the latter.

The really important points of difference were as yet overlooked; in of the in the number of they are found the nature stipules, ovules

and in the characters of the fruit. The stipules ofHenriquezia are narrow there of them each those of and are two to petiole, Platycarpum are wide and intrapetiolar; in Henriquezia they are, moreover, inserted on the petioles and shed with the latter, whereas in Platycarpum they

are inserted on the stem, and leave a scar that surrounds the stem cells of and passes above the base ofthe petioles. The ovary Henriquezia. Henri- contain four ovules, those of Platycarpum two. The capsule of quezia is semi-superior, oblate-lenticular, and provided with valves that after dehiscence retain their shape, whereas that of Platycarpum is almost completely superior, discoid, and provided with valves of which the right and left half after dehiscence fold back until they

meet.

J. D. Hooker (in Bentham et Hooker, Gen. PI. 2: 12. 1873)

accepted Bentham’s view with regard to the near affinity between the and this them into two genera, brought out by uniting a

tribe, for which he introduced the name Henriquezieae. He differed, however, from Bentham and the latter’s predecessors with regard to the place that would have to be assigned to them, and instead the he of to Bignoniaceae referred his new tribe to the Rubiaceae. It is not quite clear why he did this, for the points in which they differ from the general plan of this family and in which they

the and the exal- agree with Bignoniaceae, the zygomorphous corolla buminous mentioned in his The seeds, are duly description. zygo-

morphism of the flower is even somewhat overemphasized, for the

corolla is described as bilabiate, whereas the lobes are in reality but

and united into and lower slightly unequal certainly not an upper a The distinct of the other lip. very zygomorphism the androecium, on is that hand, not specially mentioned. We must assume therefore he did not attach much weight to these characters, and thought the of the which he described and the position ovary, as semi-superior,

presence of stipules, which he was the first to mention, of greater That this would be importance. so, however, seems very dubious, for semi-superior and even entirely inferior ovaries, although not

present in the Bignoniaceae themselves, are not unknown in the order to which the latter belong. As Bentham already remarked, they of occur in one the subfamilies of the Gesneriaceae. Stipules, it is true,

do not occur in this order, but as their distribution in the various orders is rather their suffi- erratic, presence can not be regarded as a 368 C. E. B. BREMEKAMP

cient ground to exclude these genera from an order in which stipules far absent. were so regarded as The diagnostic characters which Hooker mentioned in his key to the have been taken from the is characterized genera, calyx. Henriquezia by “Calycis limbus circumscisse deciduus” and Platycarpum by “Calycis limbus 5-lobus”. The expression “circumscisse deciduus” is not quite

I the clear. can not believe that it refers to the way in which upper part of the calyx is separated from the hardly noticeable base, for in this respect there is not much difference with Platycarpum; I prefer that it refers the left the the surface to suppose to scar by calyx on the is of fruit; this scar found midway between the top and the base, i.e. in where circumscissa” would the place a “capsula open: in the of this there is indeed marked difference between position scar a very the two genera, for in Platycarpum it is found quite near to the top of the pedicel. That in Platycarpum too species with four calyx lobes are found, was, as mentioned above, in Hooker’s time unknown. Schumann (in FI. Bras. VI, 6: 133. 1889 and in Engler and Prantl, Nat. Pflanzenfam. IV, 14: 39. 1891) adopted Hooker’s classifi- cation, but he too failed to make this decision clear. Nevertheless he seems to have felt the difficulty caused by the zygomorphism of the for in this I wish his remark that Hooker’s flower, light to see of Henriquezieae form by means some other genera with zygomorphous flowers, like Capirona, Coutarea and Ferdinandusa, a link between the

Cinchoneae and the Bignoniaceae. This is, if taken literally, a rather for it that the kind of strange assertion, presupposes zygomorphism found in the flowers of the Bignoniaceae and in those of Platycarpum and Henriquezia is homologous with the zygomorphism shown by the flowers ofCapirona, Coutarea and Ferdinandusa. However, that a charac- found in order in of the of be ter one part genera a single family can in regarded as homologous with a character observed a family be- longing to a differentorder, is certainly a rare exception, for it demands that the character is of the of the part general plan next higher unit, viz. the one that comprises both these orders; in our case this is hardly to be expected. Moreover, the resemblance between the zygomorphism shown by the flowers of the Bignoniaceae and of Platycarpum and Henri- and that in the flowers of Coutarea and Ferdinandusa quezia seen Capirona, is very superficial indeed, and can therefore not be regarded as The characteristic features of the indicating homology. most zygo- morphism shown by the flowers of the Bignoniaceae and of Platycarpum and in the of the Henriquezia are found insertion stamens at different height and in the bent at the base of the filaments, and these features in flowers of are completely absent the Capirona, Coutarea and Ferdi- nandusa, where the zygomorphism is confined to a slight curvature and difference the of the corolla a corresponding in length of the filaments. The first kind of zygomorphism is characteristic for a large part of the Tubiflorae, whereas the second is but a slight modification of the actinomorphism that is typical for the Rubiales. The zygo- link between morphism of Platycarpum and Henriquezia is therefore no that of the and that of the Coutarea and Bignoniaceae genera Capirona, ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 369

Ferdinandusa and via the latter with the actinomorphism of the other Rubiaceae, but it is fully identical with that of the Bignoniaceae and quite different from that of Capirona, Coutarea and Ferdinandusa. The characters used Schumann in his the viz. by key to genera, for Platycarpum “Flowers nearly hypocrateriform, vertical. Calyx lobes singly deciduous. Leaves decussate”, and for Henriquezia “Flowers

bilabiate, horizontal. Calyx circumscissile. Leaves verticillate”, are

inadequate. Neither in the shape of the corolla nor in the position

of the flower any well-marked difference is observable; the calyx is shed in both in the verticillate genera same way; and species with

leaves found in too. are, as we now know, Platycarpum In his of the Steyermark. monograph genus Platycarpum (Am. Journ. of Bot. 39: 418-429. 1952) also followed Hooker in

the to the Rubiaceae but he too no referring Henriquezieae , gave In his to the he introduced arguments. key genera a new diagnostic viz. the number of ovules in character, the ovary cells, 2 in 4 in is reliable Platycarpum, Henriquezia. This a difference, though less discernible than the differences in the easily stipules {PI.: as many

as the leaves and wide; H. : twice as many as the leaves and narrow) and in the capsule {PL: discoid and almost entirely superior, the and left half each valve back until and H.: right of folding they meet; oblate-lenticular and semi-superior, the right and left half of the

valves not folding back), which are not mentioned. From this of the literature and from remarks survey the I have inserted here and there conclude that the we may genera Platycarpum

and Henriquezia are near enough to each other and at the same time

other united into sufficiently distinct from genera to be a taxonomical least of tribe will group to which at the rank a have to be assigned. Further it clear that this be inserted in the seems group can not Bignoniaceae, as they differ from the latter in such important points

as the inferior the collateral ovules and the of ovary, presence stipules. The peculiar kind of zygomorphism shown by the flower, however, with this than suggests a nearer affinity family with the family Rubia-

ceae, to which they were referred by Hooker. From the general plan

of the latter differ not in the of the of hairs they only presence strip on the ventral side of the corolla throat and in the zygomorphism of

the androecium but also in the absence of colleters on the inside of

the stipules, in the exalbuminous seeds and in the structure of the

embryo with its large cotyledons and small axial part. The two latter characters found also in the are genus Gleasonia, which, as we have is be referred either pointed out, to to the Rubiaceae themselves or else related and their in the to a very closely family, presence Henriquezieae

can therefore not be regarded as a factor of decisive importance, but the combination of the first-named characters, the peculiar kind of which finds its in the of zygomorphism expression presence a strip of hairs ventral side of the on the corolla throat and in the zygomor- and of is phous androecium, the absence colleters, doubtless of so outstanding significance that it fully justifies the exclusion of these from the Rubiaceae. The of genera peculiar type zygomorphism ex- 370 C. E. B. BREMEKAMP

hibited by their flowers, moreover, points in a different direction, viz. in that of the Tubiflorae.

The of the presence an inferior ovary makes it impossible to return the but it does from group to Bignoniaceae, not prevent us referring it to the order of which the latter form a part, as in one of the other families, the Gesneriaceae, such ovaries are by no means uncommon. this kind of The collateral ovules form no obstacle either, as arrange-

ment of the ovules is certainly not unknown in the Tubiflorae; it occurs in here two families that are to be regarded as near allies of the Bigno- niaceae, viz. the Thunbergiaceae and the Mendonciaceae, and also in the

Labiatae. The of at first to offer presence stipules might seem, sight, a greater difficulty, as stipules are indeed entirely unknown in the but should realize that the in which Tubiflorae, we groups they occur, are of very different taxonomical rank, and that they are rather the various orders. That the irregularly spread over Tubiflorae were

so far regarded as an order in which stipules are absent, is therefore sufficient for in which no ground excluding a group they are present. That the and the genera Platycarpum Henriquezia can not belong to

Rubiaceae, would not necessarily mean that there is no place for them in the Rubiales, i.e. in the order to which the Rubiaceae belong. This is doubtless that deserves a possibility careful consideration. It is cer- that of the families that tainly easy to see they can not belong to one far have been in this for there is so recognized order, only one among that in of them agrees with our genera the presence exalbuminous seeds, and this family, the Valerianaceae, differs from them in the exstipulate leaves and in the incomplete androecium and gynoecium. The that form distinct the other possibility they might a family is, on

hand, not so easily discarded, as the order is but poorly defined, so that it is difficult what it. At to say might or might not belong to any rate, as the Rubiaceae are the only family with some of whose represen- tatives show certain and this our two genera a resemblance, as resem- blance may, on good grounds, be regarded as a mere analogy, the insertion of such a new family in this order can not be regarded as a plausible solution of our problem. The peculiar type of zygomorphism shown the flowers of feature that by our two genera is, at any rate, a looks entirely out of place in this order. The considerations lead the conclusion that the foregoing to genera which Platycarpum and Henriquezia form a well-defined group, for there be in the but which fits well seems to no place Rubiales, very be into the order Tubiflorae, although it can apparently not referred On to one of the families that so far have been recognized. account of the large exalbuminous seeds it would have to be referred to a position in the neighbourhood of the Bignoniaceae, Pedaliaceae, Marty- niaceae, Thunbergiaceae, Mendonciaceae and Acanthaceae; in the structure of the there with the pollen grains is moreover some resemblance Pedaliaceae and the Bignoniaceae; in the collateral ovules with the Thunbergiaceae and the Mendonciaceae; in habit and especially in the form of the fruit, which is flattened perpendicular to the dissepiment, remind Fenzl of the Tecomeae they one, as already remarked, (Bigno- ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 371

and in the seeds of of the Peda- niaceae); large and flat, wingless some liaceae.

The and united genera Platycarpum Henriquezia. therefore are to be into that will have receive the a group to rank of a family, and for which a place will have to be found in the order Tubiflorae. As the two united Hooker into tribe genera were already by a Henriquezieae, the name Henriqueziaceae seems to be indicated for this family. It is true that Hooker might perhaps better have derived the name for his

tribe from Platycarpum, as that name has much older rights than

In later Baillon 7: 345 Henriquezia. fact, some years too (Hist. d. PI. et 487. united the into but he the diffe- 1880) genera a group, as judged between them of minor he rences importance, did not regard this group as a tribe but as a and for this he of the genus, genus had, course, to use This if name Platycarpum. means that Baillon’s view were to be accepted, that of the name of the tribe as well as the family would have to be derived from Platycarpum. However, as there is no reason to assume

that later will botanists return to Baillon’s standpoint, it seems safe to the the base name of family on that of Hooker’s tribe. The family

may be described as follows:

Henriqueziaceae familia nova Tubiflorarum praesentia stipularum et “glandulae petiolaris” a familiis aliis diversa, seminibus exalbu- minosis ad Bignoniaceas, Pedaliaceas, Martyniaceas, Thunbergiaceas,

Mendonciaceas et Acanthaceas accedens sed ab eis absque stipulis et ovario infero glandula petiolari distincta, structura pollinis cum ovulis collateralibus Bignoniaceis et Pedaliaceis, cum Thunbergiaceis et Mendonciaceis habitu forma congruens, et capsularum inter Bignoniaceas

Tecomeis nonnullis, seminibus complanatis et exalatis Pedaliaceis ali- quibus comparanda. Arbores vel mediocres. Folia decussata parvae vel verticillata, petiolata et stipulata, simplicia et integerrima; petiolus ad basin dorso “glandula” laevi, sicc. nigrescente instructus. Stipulae sine colletris. Flores in inflorescentias terminates, paniculiformes dispositi.

Ovarium inferum, biloculare, utroque loculo ovulis 2-4 collateralibus fere 4- vel anthesin instructum. Calyx ad basin 5-partitus, post deciduus. Corolla semper 5-mera, paulum zygomorpha, extus pubes-

cens, tubo in faucem campanulatam ampliato, intus ad insertionem staminum barbate et virga pilorum e basi lobi mediani descendente

percurso, lobis subaequalibus obtusis, aestivatione adscendente im- bricatis. Stamina 5 ad altitudines inaequales inserta, inclusa; filamenta basi curvata, inaequilonga; antherae dorsifixae, basi sagittatae, loculis Granula parallelis. pollinis satis magna, colpis 3-6 instructa. Discus annularis. Stylus glaber, ad altitudinem antherarum in

stigmata 2 exeuns. Capsula semi-supera et oblato-lenticularis vel fere

tota supera et discoidea, dissepimento angusto instructa, loculicide loculo dehiscens, utroque semina 2 vel 4 continens. Semina valde

complanata sed exalata, exalbuminosa, cellulis testae in papillas

exeuntibus; embryo cotyledonibus planis et magnis et axi brevi instructus.

in Generibus 2 parte Americae Tropicalis Amazonica dicta distributa. 372 C. E. B. BREMEKAMP

The of the diffe- two genera are easily distinguishable by means rences given in the following key.

Key to the Genera of the Henriqueziaceae

inserted the 1. Stipules as many as leaves, on stem, intrapetiolar and ovate. Flowers comparatively small, i.e. corolla less than 2

cm long. Ovary cells with 2 ovules. Pollen grains 3- or 4-colpate

(Fig. 4). Capsule discoid, almost entirely superior, i.e. the scars and the the of calyx corolla near top of pedicel; valves consisting until of a right and left half which after dehiscence fold back

meet Testa cells with walls they (Fig. 10). straight (Fig. 5) . . Platycarpum

of orinocense before after Fig. 10. Capsule Platycarpum and dehiscence; a. scar of

calyx; b. scar of corolla.

1 : twice inserted the linear. Stipules as many as leaves, on petioles,

Flowers large, i.e. corolla more than 3 cm long. Ovary cells with

3 4 ovules. 5- 6- or Pollen grains or colpate (Fig. 4). Capsule

oblate-lenticular, semi-superior, i.e. the scar of the calyx midway

between the top and the base, and that of the corolla midway between the top and the scar left by the calyx; valves not consis- ting of mobile halves (Fig. 11). Testa cells with undulating walls (Fig- 6) Henriquezia

There are still a number of species of which flower and fruit are

either unknown or but imperfectly known. However, in view of the

great uniformity shown by those that could be studied in sufficient

detail, there seems to be good reason to expect that the other ones will not deviate too much. Still, with regard to the characters of the

pollen grains and especially of the testa cells some reserve should be

the be made, as pollen grains could studied only in a small number

of and in species, as the genus Henriquezia seeds were available only from the The type species. most reliable characters nevertheless are those of the and the latter also useful stipules, are the most ones, as when neither flowers fruits they may serve even nor are available; even when they themselves have been shed, their scar can enlighten the and this is us with regard to position they occupied, enough.

In the the the be structure of pollen grains difference seems to ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 373

than greater the key indicates, for, as the figures show, they differ not only in the number of colpae but also in size, outline and relief. It does latin of not seem necessary to give new descriptions the two for genera, although the existing ones are not entirely satisfactory, be extended they can easily and corrected by means of the data contained in the key and in the description of the family. The revised Steyermark genus Platycarpum was not so long ago by l.c., who brought the number of species to five. A sixth species be and will described below, of a seventh species material was collected

Fig. 11. Capsule of Henriquezia verticillata; a. scar of calyx; b. scar of corolla. 374 C. E. B. BREMEKAMP

inBritish Guiana (Paharaina Mts, Mt Ayanganna, Maguire e.a. 40666

~ NY) ; it resembles PI. Duckei Steyermark, but has much larger capsules;

as no flowers were available, it seems better to leave it unnamed.

based his to the the Steyermark key species mainly on presence or

absence of an indumentum, and on the latter’s distribution and qua- is be that will natural lity. It not to expected this lead to a fully classi-

fication, but in view of the near affinity between the species, a better

arrangement will probably be difficult to obtain, and it can not be that far denied that the species so have been described, are in this identifiable. it the that way easily Moreover, was only way was open

to him, as he had to include his Pl. decipiens, of which neither flowers

known. Pl. could nor fruits are My own Froesii easily find a place

in his key; to this end we would have to insert before his division Pl. orinocense and Pl. another leading to Duckei one contrasting these of whose two species on account their opposite leaves with Pl. Froesii,

leaves are ternate.

One might perhaps have preferred to base the key in the first

instance on the number of calyx lobes, were it not for Pl. decipiens,

which in this case could not have been placed, as its calyx is unknown. The with 5-merous Pl. Pl. and species calyx are orinocense, negrense Pl. Froessii, those with 4-merous calyx Pl. Duckei and Pl. rhododactylum; in both these there is with in the groups one species ternate leaves, first Pl. Froesii, and in the second Pl. rhododactylum. It is possible, however, that the taxonomical value of the difference in the number

of lobes is not so as one be inclined to in calyx high might suppose; Pl. orinocense the number of lobes and calyx seems to vary, such of with of variability is, course, not easily reconcilable our idea a taxonomically important character. It is perhaps worth noting that

the the in the a comparison of position calyx lobes occupy 4-merous and in the 5-merous calyx, leads to the conclusion that the anterior

lobe of the 4-merous is with anterior calyx homologous the two ones of the 5-merous calyx. be described follows: The new species may as Froesii Brem. foliis ternatis Pl. rhododac- Platycarpum n. spec, cum Woodson sed foliis tylo et Steyermark congruens, calyce 5-mero et

scabridis et subtus molliter minoribus, supra pubescentibus, numero ab faciliter minore nervorum percursis eo distinguenda, calyce 5-mero foliis ad Pl. Ducke sed foliis ternatis et parvis negrense accedens, supra subtus inflorescentia scabridis et pubescentibus, tomentella ab eo Pl. orinocensi Humb. ad 5-mero diversa, a et Bonpl., quern calyce et indumento foliorum accedit, foliis et ternatis et minoribus, numero minore nervorum percursis recedens.

Arbor parva, circ. 4 m alta. Ramuli novelli ferrugineo-tomentosi,

veteriores cortice nigrescente, sicc. plicatulo vestiti. Folia ternata;

petiolus glandula basali excepta ferrugineo-tomentosus, 2-4 mm

longus; lamina oblonga, 6.5-8.5 cm longa et 2.2-3.8 cm lata, apice

obtusa, basi acuta, margine recurvata, coriacea, paulum discolor, sicc. dense brunnescens, supra costa strigosa excepta primum pilis

satis longis sparsa, mox pilis rejectis scabrida, subtus costa tomentella, 375 ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA nervis venulisque pubescens, costa subtus prominente, nervis utroque latere costae 11 vel 12 subtus prominulis, venulis dense reticulatis utrimque distinguendis. Stipularum calyptra longe ferrugineo-villosa. Inflorescentia pedunculo ferrugineo-tomentello 2-3 cm longo elata, 7-8 ramulis rachide cm longa, ternatis, verticello primo et interdum secundo foliis ordinariis vel magnitudine aliquantulo redactis suffultis, ramulis aliis pedicellisque ebracteatis; rachis, ramuli pedicellique ut pedunculus ferrugineo-tomentelli. Calyx 5-merus. Corella rubra et

fragrans dicta, matura nondum visa. Capsula etiam ignota. Habitat Amazoniam Brasiliensem.

Brazil: Amazonas; Rio Negro, Preto, Matupiry, Froes 22812, Nov. “in 1947, type (U); open country”. Froes the 22812 was distributed by “Institute Agronomico do Norte” under the name “Platycarpum negrense Ducke”, but it is easily distinguishable from that spedies by the ternate instead of opposite, the side the lower on upper scabrid and on softly pubescent leaves and by the somewhat greater length ofthe hairs by which the various parts of the inflorescence are covered. By its ternate leaves it resembles Pl. rhododactylum Woodson et Steyermark, but the latter has a 4-merous

and its leaves are the side and calyx on upper entirely glabrous on the lower for the which nearly so, very small papillae by they are covered the aid of on this side are seen only by a lens. The genus Henriquezia has received but little attention in recent and the number of its has increased but years, species slowly. It is now the latest H. Ducke. six, addition being macrophylla Below a seventh „ _ species will be described, viz. H. longisepala.

In this genus too the species are all closely related, and it is there- fore difficult to them in natural arrange groups. They are, however, easily distinguishable, and this is mainly due to the following sets of differences, 1) stipules deciduous, i.e. shed shortly after the whorl of leaves to which they belong, begins to expand, or stipules persistent, the i.e. shed at same time as the leaves to which they belong; 2) leaves 4- the 5-, or 3-nate; and 3) nerves on underside of the leaf prominent Less in the the or not. important differences are found length of calyx in the of the leaf and the number of lobes, shape lateral nerves, in the of nature the indumentum, and in the colour the leaves assume in

drying. Deciduous stipules are found in H. verticillata Spruce ex Bth. with 4- or 5-nate leaves and in H. obovata prominent nerves, Spruce ex Bth. with 4-nate leaves of a somewhat different shape but also with and in H. prominent nerves, oblonga Spruce ex Bth. with 3-nate leaves and non-prominent nerves; persistent stipules are met with in H. Ducke with 4-nate leaves macrophylla and prominent nerves, in H. K. Sch. with 3-nate leaves and the Jenmanii prominent nerves, with the only species acute, on underside greyish-tomentellous leaves, in H. nitida Spruce ex Bth. with 3-nate leaves and non-prominent and in Brem. nerves, H. longisepala also with 3-nate leaves and non- but from prominent nerves, differing H. nitida, and in fact from all the other the species of which flowers are known, in the greater length of the which corolla calyx lobes, are nearly as long as the tube, and 376 C. E. B. BREMEKAMP from H. nitida in particular in the ferrugineous-tomentellous inflo- rescence.

It is not impossible that part of the specimens that have been iden- tified with H. verticillata (Froes 21518 et 22784, both from the Rio

Ducke 21683 et all three from will Negro, 81, 22813, Manaos) prove undescribed to belong to a nearly related, but as yet species. The leaves of these plants are 4-nate, wider and with a rounded instead of acute base, and provided with a smaller number of nerve pairs

instead and the is 3 (7—9 of 11), corolla slightly larger (tube cm long from instead of 2.3 cm). However, more material, preferably the type locality, will have to be compared, before it will be possible to decide whether these differences fall outside the normal of range variability. in this the in My new species offers respect no difficulties; points which it differs from its allies are numerous and well-marked. Nor can it be doubted that it belongs to the Henriqueziaceae; this follows of the ”at the of its from the presence “gland base the petiole, slightly zygomorphous corolla with the streak of hairs running down the ventral side of the campanulate throat, the five stamens inserted at unequal height and provided with filaments showing a bent at the base; that it belongs to Henriquezia follows from the number and shape of the stipules and their insertion on the basal part of the petiole, the comparatively large size of the corolla, the 5-colpate pollen grains, and the of 3 4 ovules in each ofthe cells. presence or ovary Brem. Henriquezia longisepala n. spec, stipulis persistentibus H. H. K. H. nitida cum macrophylla Ducke, Jenmanii Sch. et Spruce ex Bth., foliis ternatis cum H. oblonga Spruce ex Bth., H. Jenmanii H. H. nitida et H. nitida, nervis non prominentibus cum oblonga et congruens, a H. macrophylla foliis ternatis, multo minoribus et nervis H. nervis non prominentibus instructis, a Jenmanii non prominentibus foliis et insuper apice rotundatis, a H. nitida inflorescentia ferrugineo- tomentella, a H. oblonga stipulis persistentibus et foliis multo minoribus hactenus distinguenda, a speciebus omnibus quarum calyx notus est, insuper longitudine sepalorum diversa. Arbor circ. 4 alta. Ramuli novelli obtuse parva, m triangulares, ferrugineo-tomentelli, 4-5 mm diam., internodiis 1.0-4.5 cm longis; ramuli veteriores cortice vestiti. Folia griseo-brunneo opaco ternata; petiolus glandula basali excepta ferrugineo-tomentellus 1.5—1.7 cm longus; lamina oblonga vel oblongo-oblanceolata, 12-18 cm longa et 4.5-7.0 cm lata, apice rotundata et mucronulata, basi acuta et in in foliis maturis petiolum decurrens, coriacea, utrimque glabra, supra nitida, sicc. chryseo-lutea, subtus pallidior, costa supra latitudine conspicua, prominula, longitudinaliter striata, subtus prominente, nervis utroque latere costae plerumque 12, utrimque distinguendis inter sub lente subtus sed non prominentibus, nervos supra reticulata, laevis. Stipulae circ. 2 cm longae, acutissime exeuntes, marginibus incurvatae, persistentes. Inflorescentia pedunculo ferrugineo-tomen- tello 4 3 etiam cm longo instructa; rachis cm longa, ferrugineo- 15 tomentella. Bracteae ramulos suffulcientes circ. mm longae et

4 mm latae, infimae 3-fidae, aliae integrae, ad anthesin deciduae. ON PLATYCARPUM, HENRIQUEZIA AND GLEASONIA 377

Flores pedicello ferrugineo-tomentello circ. 8 mm longo elati. Ovarium 4 ferrugineo-tomentellum mm altum, utroque loculo ovulis 3 vel 4 instructo. Calycis lobi lineari-lanceolati, 2.5-3.0 cm longi, tertia 3 ad 6.5 parte inferiore circ. mm lati, costati, medium mm lati, tertia parte superiore sensim attenuati, extus ferrugineo-tomentelli, intus griseo-tomentosi. Corolla miniata, extus griseo-tomentosa, tubo campanulato 3.0 cm longo et parte superiore 1.5 cm diam., lobis circ. 11 1.3 cm longis et 0.9 cm latis. Stamina filamentis glabris et 13 mm longis et antheris 7 mm longis instructa. Granula pollinis depresse 75 diam. Discus tomentellus. globosa, 5-angularia et 5-colpata, [i

Stylus glaber 22 mm longus; stigmata 2 linearia 1 mm longa. Capsula nondum visa.

Habitat Amazoniam Brasiliensem.

Brazil: Rio Sao Froes Amazonas; Negro, Felipe, Igara pe Touri, 28781, 27 Sept. 1952 (ex Inst. Agron. do Norte), type (U).

which this is The studies on paper based, were made possible by a grant from the “Netherlands Organization for Pure Research (Z.W.O.)”.