Acta Botanica Neerlandica 15 (1966) 1-33

Remarks on the position, the

delimitation and the subdivision

of the Rubiaceae

C.E.B. Bremekamp

(Botanical Museum and Herbarium, Utrecht)

(received October 22nd, 1965)

Introduction

It is often assumed that the delimitation and the subdivision of the various families which have been distinguished in the Angiosperms, do offer serious no longer difficulties. They would belong to those of for which objects study already long ago a fairly satisfactory so- found. If wish be with this lution was we to acquainted solution, the would have would be look such works only thing we to do, to up as Benthamand Hooker’s “Genera Plantarum” and Engler and Brand's “Nattirliche Pflanzenfamilien”. Some improvements might still be

but these would be of minor desirable, importance only. These as- be sumptions, however, are to regarded as dangerous illusions.

That the very serious nature of the shortcomings found in the delimitation and subdivision of these families, especially of the larger ones, is so often overlooked, is apparently due to an attitude of mind which is observed in of a comparatively large part the taxonomists,

viz. a lack of interest in the development of a truly natural classi- fication. This is not incomprehensible. Most of them spend the major part of their time in the elaboration of floras covering areas of more but of or less limited extent, and they are rarely aware the fact that the the which is in knowledge of families obtained this way, remains in necessarily incomplete. Moreover, the elaboration of a flora the most essential point is the construction of serviceable keys to the species as well as to the groups of higher rank, not the exact deli- mitation of these the latter end material is groups; to usually more than the of flora has required compiler a at his disposition. However, we must not overlook the fact that a key, in order to be serviceable, need not reflect the degree of affinity between the units with which it is dealing; in reality, such keys are often entirely or almost entirely this therefore artificial, and applies also to classifications which are

will of based on such keys. To illustrate this, I discuss here some the aid of which in the Rubiaceae suitable the characters by very keys to the genera have been constructed, but which when they subse- used the quently were for elaboration of a classification, led to entirely unsatisfactory results. time have To the characters which for a long played an important either part in the subdivision of the Rubiaceae, belong the presence of their kind of which to of uni- or pluriovular ovary cells, fruits,

1 2 C. E. B. BREMEKAMP

this end divided in and and the of their were dry fleshy ones, nature which viz. seeds, of also two kinds were distinguished, winged and

of the that wingless ones. However, even a superficial study groups have been distinguished by the aid of these characters, will show that As most of them are entirely unnatural. the differences are morpho- of this have been logically hardly any value, might expected. The number of ovules cell shows continuous of per ovary a range there variation, and so long as is no reason to assume a correlation between a definite number of ovules and one or more other characters, it is therefore attach the of not justified to to presence one or two ovules cell taxonomic value per ovary a greater than to that of three

or of another number of them; we know, in fact, that the number of cell within ovules per ovary may vary the limits of a single genus of the Pavetta L” in (see my “Monograph genus Fedde, Repertorium 37: remarks 10. 1943, and my on the genus Anotis DC and the in the of the Cruckshanksieae last chapter present publication) as well

as in genera that on account of their many common features are to be in Tarenna Gaertn. and Pavetta regarded as nearly related, so e.g. L

(see p. 7 of my “Monograph of the genus Pavetta L”). the between and be That difference dry fruits fleshy ones can not

the of used for characterization natural groups is also easily com- This distinction is prehensible. an ecological, not a morphological one, and in both kinds of fruits quite considerable morphological differ-

ences are found; in some instances the differences between a dry and than those between kinds of a fleshy fruit may even be smaller two fruits kinds of dry or two fleshy ones. In the genus Mussaenda L species with dry fruits are found side by side with species producing if fleshy ones; even these species are divided over two genera, as sometimes is done, this does not make much difference, for in that

case we will have to admit that the two kinds of fruits are found in

related very nearly genera. The same applies to the distinction between winged and wingless seeds: this is too an ecological, not a morphological distinction. From

the morphological point of view these wings are by no means all alike; in the the tribe in which the with seeds Cinchoneae, genera winged

were in the which be brought together, they are genera may rightly

regarded as near allies of Cinchona L, in position as well as in the

structure of the cells of which they are mainly composed, entirely different from those found in which that are the genera were trans- ferred by me to the Rubioideae. Another drawback of the classifications accepted in the works is mentioned above, that they are based mainly on earlier classifi- which cations were drawn up when a much smaller number of genera

were known, and that the authors often assumed that the subfamilies

and even the tribes based on this more limited number would suffice

for the of the discovered This incorporation subsequently genera. ap- is the of the that parently explanation fact so often a genus has been In the the squeezed into a tribe into which it does not fit. case of

Rubiaceae there are even quite a number of genera of which it is POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 3

extremely doubtful whether they really belong to this family. This point will be discussed in the second chapter.

Many taxonomists, moreover, seem to feel a certain reluctance to subfamilies which contain accept but a single tribe, and tribes which contain but a single genus. This reluctance, however, is entirely We unjustified. know several families and even some orders which

contain but a or a small number of and there is single very genera, that accordingly no reason whatever for assuming a subfamily or a tribe should contain than of the always more one group next lower rank. Whether is a taxon to be classified as a subfamily or a tribe

evidently depends upon the numberand the importance of the charac- in it differs ters which from its nearest allies and on nothing else, and

where our haveoverlooked the of predecessors presence some differences undervalued their is or importance, it doubtless our duty tocorrect them. With regard to what has been said in the preceding paragraph, it of of for is, course, importance to find ways and means estimating the taxonomic value of the of groups characters that are to be used for the distinction of the tribes and subfamilies and, naturally, of the In the of families too. case the latter the number of diagnostic charac- is often ters extremely small and their value not rarely dubious. With regard to the Rubiaceae this point will be dealt with in the chapters and one two, but at this point it is perhaps worth while to draw the attention to the fact that in the description of this family given in the works mentioned above we do not find a single character that

can be as This of regarded a truly general one. is, course, due to the fact that this family, in the delimitation accepted in these works, elements comprises which do not really belong to it. The inclusion of such elements, moreover, has often led to entirely false conclusions

with regard to the affinity of the families. In the case of the Rubiaceae, for the inclusion of instance, the genera Henriquezia Spruce ex Bth. and Platycarpum Humb. et Bonpl. has led to the assumption that there should be a comparatively close affinity between this family and the

However, when these two are it Bignoniaceae. genera excluded, ap- pears that not a single argument is left which might possibly be adduced in favour of this view. After the exclusion of the genera Carlemannia Bth. and Sylvianthus Hook. f. the affinity with the Capri- and their allies has become less and if the foliaceae pronounced, genus Dialypetalanthus Kuhlm. had been left in the Rubiaceae, this would have

been rather a good argument for assuming a close affinity with the ! It Myrtales is therefore of great importance that such misclassifications corrected. Several other of are examples genera which were incorpo-

~ ~ rated in the Rubiaceae on insufficients ground, will be given in the chapter dealing with the delimitation of the family. It of be denied can, course, not that the number of characters

which related taxa have in decreases common, with the increase in rank of the taxon to which they belong; it is certainly no wonder that families but have, as a rule, a very small number of general characters. However, groups based on a single general character

should always be regarded with some distrust. 4 C. E. B. BREMEKAMP

Verdcourt (Bull. Jard. hot. de I’fitat, Bruxelles 28; 228. 1958) discusses Ixoroideae and it because the my subfamily rejects pollination

mechanism which is a general feature of this group, is thought to be also in which other circles of present some genera belong to affinity, that and a point will be discussed in chapter three, further because

this pollination mechanism is regarded by him as a single character. This The mechanism is is certainly not so. type of pollination which that characteristic for the Ixoroideae requires the anthers are in the

bud in with the of the which contact part style serves as a temporary depository for the pollen grains; this means that they must be found

at the the latter and that The same height as they open introrsely. “ part ofthe style which serves as receptaculum pollinis” must be covered

with short hairs, otherwise it would be unable to fulfil its function.

Finally the flowers must be protandrous. This type of pollination mechanism therefore much is a more complex arrangement and ac- cordingly of much greater taxonomic importance than most other which pollination mechanisms, e.g. than the heterostylism to Verdcourt I himself attaches a relatively high value. As have discussed some

the time ago relatively small taxonomic value of this kind ofpollination mechanism in “On in my paper pollen dimorphism heterostylous

Psychotrieae, especially in the genus Mapouria Aubl.” (Grana palyno- 4: 54 it logica and 55, 1963) seems superfluous to enter into details. Suffice it to state that heterostylism is found in several but distantly related that it is restricted of the and families, always to part genera that it is in these but feature. With the even genera rarely a general

kind of pollination mechanism found in my subfamily Ixoroideae this is quite different. It is apparently restricted to this subfamily of the

Rubiaceae and to a few other families, viz. the Campanulaceae, the

Goodeniaceaeand the Compositae, and in the two last-mentioned families

the mechanism is not even quite the same, for the hairs on the style here different the have a function; moreover, receptaculum pollinis Goodeniaceae “Pollenbecher” of the (the of Schonland) occupies a dif-

ferent position and shows a different structure, whereas in the Com- positae the receptaculum pollinis is lacking. In contradistinction with heterostylism this kind of pollination mechanism and its variations in all the in which feature. are groups they occur, a fully general

The combination of characters on which it rests, is therefore to be

regarded as taxonomically highly important.

My subfamily Rubioideae was accepted by Verdcourt l.c. as a na- tural the tribes which that united in one, although at moment were this in group, had but one character common, viz. the presence of I add here that of the other subfamilies raphides. may one two recog- nized by Verdcourt, viz. the Cinchonoideae, has in the delimitation

accepted by him not a single general character, for the absence of

be this would be raphides can certainly not regarded as such; pos- sible if it could be that mechanism only proved they possess some the and that this which prevents development of this kind of crytals, mechanism is in all tribes been moreover, the same the that have included in this subfamily. 5 POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE

That the presence of raphides is to be regarded as a taxonomically follows from the fact that it is in the tribes very important feature, in which it occurs, a general character, whereas this kind of crystals is completely absent in the other tribes of the Rubiaceae and also in all the families which are to be regarded as their nearest allies. In itself, their is sufficient that the however, presence not to prove affinity of the tribes in which they occur, is so close that they are to be united in a single subfamily. In this connection it is noteworthy that most

of these tribes agree also in another character, viz. in the valvate aestivation of the corolla lobes. There tribes in are, however, two which another kind of aestivation of the corolla lobes is found, viz. the Hamelieae, where the aestivation is imbricate, and the Hillieae, where it is contorted. The inclusion of the Hamelieae in the Rubioideae is perhaps permitted, as at least in one of the tribes in which the aestivation is said to be valvate, viz. the Psychotrieae, at least two

genera are included, viz. Notopleura Brem. and Naletonia Brem., in which the aestivation is imbricate. These of which the two genera,

first was regarded by Hooker as a section of Psychotria L, are in all other normal members of the tribe “Notes respects Psychotrieae (see my on the Rubiaceae of Surinam” in Rec. d. trav. bot. Need. 31: 284

and 289. 1934). The case of the Hillieae, however, is an entirely The Hillia in the different one. inclusion of the genus Jacq. Cincho- neae was made possible by Hooker by means of a slight change in the definition of the latter; according to him the seeds of the Cinchoneae

would be “alate or appendiculate”, but the word “appendiculate” has been added with the sole of the obviously purpose legitimizing inclusion of this but this genus, can not be regarded as an acceptable of the solution difficulty caused by the neccessity of finding a place for this aberrant for the tuft of the the seed genus, hairs at top of can be with That the certainly not homologized a wing. wings with which the seeds in the other which and at genera by Hooker, an de earlier date already by Candolle, were included in this tribe,

of different could would prove to be kinds, not be foreseen at that of time, but this excuse can not be brought forward in the case the

genus Hillia, as the latter does not show the slightest resemblance to of the with seeds any genera winged that rightly or wrongly were in it for brought together this tribe. Schumann makes even worse,

notwithstanding the fact that according to his definition of the Chinchoneae the seeds he includes the are always winged, too genus Hillia in this tribe. Hillia, however, occupies a very isolated position in the family. One of its most remarkable features is found in the structure of the testa cells, which show a rather striking resemblance with those of the Gleasonia another genus Standi., genus occupying a isolated in rather position; their other characters, however, these two show resemblance whatever. Hillia well Gleasonia genera no as as are therefore referred by me to subfamilies of their own.

I. The position of the Rubiaceae

In order to determine the taxonomic value of the various cha- 6 C. E. B. BREMEKAMP racters that are met with in a family, it is desirable to know some- the of these characters in the thing of distribution families which are to be regarded as its nearest allies. With this end in view we will consider briefly the position which is to be assigned to the Rubiaceae in the system of the Angiosperms. In more recent times this problem Bot. has been discussed at some length by Utzschneider (Mitt. Staatssamml. Mtinchen 3: 96-98. 1951) and by Wagenitz (Bot. Jahrb. 79: 17-35. 1959).

the that the the Wagenitz comes to conclusion order of Rubiales as

definedin i.e. the Engler’s “Syllabus”, as a group comprising Rubiaceae, and be Caprifoliaceae, Adoxaceae, Valerianaeeae Dipsaceae, can not regarded natural because of resemblance between the as a one, the points

Rubiaceae and the other families included in this to be group appear of less importance than those between the Rubiaceae and some of the

families that were included in the order which at that time was known as the Contortae. These points of resemblance become even more striking if this order is given a more natural delimitation by

the Oleaceae to order of its the and the removing an own, Oleales, Buddleiaceae to the order formerly known as the Tubiflorae, where it be in the of the in which may put vicinity Scrophulariaceae, a family

some of its representatives formerly had been included, e.g. by de On of the of resemblance which the Jussieu. account many points Rubiaceae show with the Loganiaceae and their nearest allies, they are this of introduces included by Wagenitz in group families, for which he

the name Gentianales Bindley emend. Wagenitz. This order, therefore, comprises in his delimitation the families Loganiaceae, from which the

Buddleioideae are excluded, Rubiaceae, Apocynaceae, Asclepiadaceae, Gentia-

and the rather The naceae aberrant Menyanthaceae. other families of the in the former Rubiales are kept together an order for which he uses

emend. name Dipsacales Bindley Nakai. The points in which the Rubiaceae resemble the other families of the

Gentianales sensu Wagenitz and those in which they differ from the follows. leaves of Dipsacales sensu Nakai may be summarized as The the Gentianales in the decussate more are, except Menyanthaceae, or, those ofthe either decussate rarely, verticillate, Dipsacales or alternate; in the Gentianales and here they are, moreover, simple entire, though in too an exception to this rule is found in the Menyanthaceae, whereas the either Dipsacales they are simple or, more rarely, pinnately com- pound and often dentate or more deeply incised. The corolla is in the Gentianales always actinomorphic, and in the Dipsacales either less actinomorphic or, more often, more or distinctly zygomorphic. The androecium is in the Gentianales always isomerous, in the Dipsa-

isomerous far cales either or oligomerous. The endosperm is, as as in the can be judged at present, in the Gentianales always nuclear,

Dipsacales always cellular. The stipules are in the Gentianales often provided with colleters, whereas these appendages of the stipules are in the Dipsacales always absent. True glandular hairs are, on the other hand, but rarely met with in the Gentianales, whereas they are of in the Another of common occurrence Dipsacales. point similarity POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 7 between the Rubiaceae and the other families of the Gentianales may be in the of perhaps seen presence alcaloids; in the Dipsacales these substances absent the are or at least very rare. Alcaloids belonging to of group the tryptophans seem to be confined to the Loganiaceae, the

Apocynaceae and the Rubiaceae. However, as they are not found in all

the genera belonging to these families, their presence would only be if be of importance to the taxonomist it could proved that they are

derived from precursors which are generally present in the Gentianales and absent in the Dipsacales. So far the most important points of resemblance between the Rubiaceae and the other families of the

Gentianales be the of the nuclear endo- seem to presence colleters and the difference the absence of sperm, most important intraxylary

phloem in the Rubiaceae and its presence in the other families, and the of the which in Rubiaceae inferior least position ovary, the is or at

more or less so, and in the Gentianales always superior. In these two

the the points Rubiaceae agree with Dipsacales. The taxonomic of the inferior should importance ovary, however,

not be overrated. In the great majority of the Rubiaceae the ovary is the Rubi- indeed completely inferior, but in Gaertnereae, a tribe of the still oideae nearly related to the Psychotrieae, the ovary, though at first

partly inferior, becomes in the later stages of its development almost and in other tribes the completely superior, some genera belonging to

proportion between the part of the ovary which rises above the place of insertion of and corolla and below the calyx that latter, may vary

considerably. In the genus Mitrasacmopsis Jovet, which by its author included in the the in the was Loganiaceae-Spigelieae, ovary is, as almost the of Gaertnereae, completely superior; presence raphides proves that this in which genus can not belong to the Loganiaceae, a family that raphides are entirely unknown, but it is to be transferred to the

Hedyotideae, a tribe of the Rubioideae; it is, in fact, a near ally of my in which the is also almost genus Diotocranus, ovary completely su- the perior; two genera even may prove to be identical. which wish Another point deserves our attention when we to esti-

mate the taxonomic value of the inferior is that ovaries of ovary,

this kind are found in widely different families, and that in some of these the otherwise rather families, e.g. in uniform Gesneriaceae, su-

perior as well as inferior ovaries are met with. The example of the

is Gesneriaceae the more instructive as this family belongs to the Tubi-

order which all other with florae, an in the families are provided superior ovaries.

The considerations expounded in the two preceding paragraphs lead

to the conclusion that in the Rubiaceae the of inferior presence an ovary be and can regarded only with some reserve as a general character, that this is of that character, moreover, not so paramount importance it would in itself be sufficient to exclude this family from the Gentianales and to compel us to regard it as more nearly related to the Dipsacales. The absence ofan intraxylary phloem in the Rubiaceae forms doubt-

less a more serious obstacle against the inclusion of this family in the

Gentianales, as this is a character which is less irregularly distributed. 8 C. E. B. BREMEKAMP

Considered in connection with the admittedly less important inferior it would the creation of ovary perhaps justify a monotypic order

Rubiales. The latter would, however, be more closely related to the families which would be left in the Gentianales than to the Caprifolia-

ceae and the other families of the Dipsacales. Apart from the affinity with the Gentianales and the Dipsacales that the be In this with Campamlales should also considered. con- nection the resemblance between the pollination mechanism of the Ixoroideae and that found in the Campanulaceae, the Goodeniaceae and

in the the Compositae, which was discussed introduction, would perhaps

deserve more attention than it has received so far.

II. The delimitation of the Rubiaceae

In the introduction to this paper we have already pointed out that

the genus Dialypetalanthus Kuhlm. had to be excluded from the Rubi-

aceae because it shows in the structure of the flowers with their poly-

merous androecium an unmistakable affinity with the families brought together in the Myrtales. We might have added that nowhere else in the Rubiaceae indications of with this order any an affinity is to be Rizini For found (cf. et Occhioni in Lilloa 17: 253. 1949). a similar reason Henriquezia Spruce ex Bth. and Platycarpum Humb. et Bonpl. had be excluded. These show in the to two genera slightly zygomor- phic corolla and in the quite distinctly zygomorphic androecium and the also in large exalbuminous seeds a high degree of similarity with

the Bignoniaceae and some other families of the Tubiflorae, whereas elsewhere in the Rubiaceae if we a few other anomalous , except genera I show have been included insufficient which, as will hereafter, on

grounds, no indications of such a relationship can be detected (see “On the of Humb. my paper position Platycarpum et Bonpl., Henrique- zia Spruce ex Bth. and Gleasonia Standi, in Acta Bot. Neerl. 6: 351 377. The Carlemannia Bth. and Hook. f. 1957). genera Sylvianthus were excluded already at a much earlier date by Solereder (Bull. Herb. the dentate Boiss. 1: 173-178. 1893). On account of or serrate leaves,

the rudimentary stipules and the oligomerous androecium they are

to be removed to a position in the vicinity of the Caprifoliaceae, i.e. to

the these in Rec. d. bot. neerl. Dipsacales (see my note on genera trav. 36: 372. 1939).

The genera mentioned above are not the only ones that will have

to be of other is excluded, whereas the position some genera extremely doubtful and should be reconsidered.

included in the Antho- The genus Opercularia Gaertn., now usually

account of its unilocular and the spermeae, was on ovary oligomerous androecium referred by de Jussieu (Ann. Mus. d’Hist. Nat. 4; 118.

to of its the but Richard 1824) a family own, Operculariaceae, (Mem. sur la famille des Rubiacees: 62-67. 1830), who distinguished in this of viz. Gaertn. and Pomax group species two genera, Opercularia Soland. in the ex A. Rich., included it as a separate tribe Rubiaceae, and this example was followed by de Candolle in his “Prodromus”. Richard rejected the decision taken by de Jussieu, because in his POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 9

number of would in the Rubiaceae and be- opinion the stamens vary in of with unilocular ovaries cause that family more examples genera would for in the occur. Richard, however, was mistaken, genera which on good grounds have been referred to this family the androe-

cium is isomerous and the is divided in two always ovary always or cells. this rule found in the Eleuther- more Exceptions to are genus anthus F. the v. Mull., which is doubtless a near ally of genera Oper- cularia and in the Tammsia which will be dis- Pomax, genus Karst., cussed and in which have been included in the hereafter, some genera Ellis and Gardenieae, viz. in the type genus Gardenia the nearly related f. and Casasia A. and in genera Macrosphyra Hook, Rich., also Villaria

Rolfe. However, as Richard (op. cit. 214-215) already pointed out, the which in these the ovules structures on genera are inserted, though

distinctly separated from each other in the basal part of the ovary, in the for this are upper part not rarely united; reason they can

very well be regarded as dissepiments, in which case the ovary would have be described more-locular. The to as imperfectly 2- or genera

Opercularia, Pomax and Eleutheranthus ought certainly to be studied in

I that such will confirm the more detail; expect a study standpoint of de and that further his Jussieu, as a result family Operculariaceae will prove to belong to the Dipsacales. Other of doubtful Tammsia genera position are Karst., Pentagonia

Bth., Hippotis Ruiz et Pav. and Sommera Schlecht. These four genera were regarded by Schumann (Natiirl. Pflanzenfam. IV, 4: 69. 1891) as nearly related because of the peculiar striation of their leaves. This so-called moire or watered striation, however, is not restricted to these genera, but is found also in the Malesian species of the genus Antirhea Gommerc. South-American of (see my paper on a species this genus in Acta Bot. Need. 8: 480. 1959), and as to the four genera mentioned in it is certain the preceding sentence, by no means that

they may be regarded as nearly related. The genus Tammsia with its unilocular ovary and parietal placentas differs considerably from the three other in combination with the anisomerism between ones; calyx and corolla and the zygomorphism of the latter and of the these characters make the of this in the androecium, position genus Rubiaceae extremely doubtful. Of one of the species of Pentagonia it is that it this is which is said has pinnatifid leaves; too a character never met in the genera which on good grounds have been included in this and if this has been referred this family, species correctly to genus, the latter will doubtless have to be referred to another family. Hippotis and Sommera less and are aberrant. The strongly zygomorphic calyx the less strongly, although still distinctly zygomorphic corolla found in Hippotis, a combination of characters which by Richard (op. cit.;

was with that met with in the P. 176) compared genus Spathodea

Beauv. are doubtless unusual features, but would (Bignoniaceae) , they in themselves not suffice exclude this from the and to genus Rubiaceae, fit in before Sommera seems to even better this family. However, a final decision be these four should be studied in can taken, genera more detail. 10 C. E. B. BREMEKAMP

The Aitchisonia for which has been found in genus Hemsl., a place the Paederieae, though on what grounds is not clear, is, on account of the insertion of the stamens at various heights in the corolla tube and of the presence of glandular hairs, certainly to be excluded from this of des family. As I have pointed out on p. 153 my “Monographic

Boiv. Baill. et Brem.” 16: genres Cremocarpon ex Pyragra (Candollea 147—177. 1958), the combination of raphides and glandular hairs which is found in this is and does not in genus, vary rare occur any other genus provided with interpetiolar stipules and a bilocular ovary with ovule in each the For a single ascending of cells. this reason we will have to assume that Aitchisonia represents an as yet undescribed

family, which probably will have to be put in the Dipsacales in the vicinity of the Caprifoliaceae. in- The genera Gaertnera Lam. and Pagamea Aubl. were originally in the but and cluded Loganiaceae, were transferred by Baillon Schu- mann to the Rubiaceae, where they were placed, on account of the solitary ascending ovules and the valvate aestivation of the corolla of Surinam” lobes, in the Psychotrieae. In my “Notes on the Rubiaceae

(Rec. d. trav. bot. neerl. 31: 248. 1934) I returned them to the Lo- th held ganiaceae, but in my communication to the 8 Botanical Congress in of which of Paris, an abstract was published on p. 113 the “Rap- 5 6: under ports et Communications aux Sections 2, 4, et 113. 1954” the title “Les sous-familles les tribus des Rubiacees” I et changed my attitude with to these and the decision taken regard genera, accepted in favour of the view that by Baillon; a strong argument they are to be included in the Rubiaceae is the of with presence stipules provided colleters, whereas a strong argument in favour of their exclusion from the Loganiaceae is to be found in the absence of intraxylary

the of and the valvate aestivation of the phloem; presence raphides corolla lobes that the Rubioideae. prove they belong to subfamily In of 1934 I also excluded the Perama wich my paper genus Aubl., indeed shows unusual viz. many features, rudimentary stipules, paral- the of but lel-nerved leaves, presence two calyx lobes, which on of the in the account position they occupy diagram were regarded by Baillon (Hist. d. PI. 7: 392. 1880) as bracteoles, the sometimes trimeric the trilocular and the seeds. The corolla, ovary triquetrous in the position of this genus was once more discussed by me introduc- tion work “TheAfrican of Oldenlandia L Hiern to my on Species sensu

et K. Schumann” In 13 I that on (1952). a note on p. pointed out of the of is be included in the account presence raphides the genus to subfamily Rubioideae, but that it can not be returned to the Spermaco- the tribe in which it had been Hooker well ceae, put by as as by Schumann, as it differs from the latter in the ascending ovules, the rudimentary stipules and the structure of the pollen grains, and that it is of its unusual features be referred tribe on account many to to a of its own, the Perameae.

III. The subdivision of the Rubiaceae

take into consideration that the If we family Rubiaceae, as was POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 11

pointed out in the preceding chapter, still contains several genera of which it be doubted their inclusion may rightly whether in this family is justified, it will be no matter of surprise that the delimitation of the in which divided older works in which groups the family was in the of the still a classification family was given, was far from satisfactory, for if it really had been satisfactory, the fact that these genera were included in it and that, as will be shown in this chapter, an even

much larger number was referred to the wrong tribe, would be these mistakes of inexplicable. My attempts to correct my predecessors has led the of but to development a new classification, before entering into it brief of the older details, seems appropriate to give a survey

ones.

The oldest subdivision of the Rubiaceae which deserves our at-

is that de 196 ofhis “Genera Plantarum” tention, given by Jussieu on p. He of the de- (1789). distinguished ten groups, mainly on account

hiscence or indehiscence of the fruit, the number of locules it contains and the number of seeds in the latter, though the number of stamens,

the position of the leaves and the habit are also taken into consi-

in the last is deration; one of groups, the one, special emphasis laid the form of the it the with on inflorescence; comprises genera capitate

flowers. few of the de can Only a groups distinguished by Jussieu be regarded as natural or more or less so. Group I (Fructus dicoccus

dispermus. Stamina saepius quattuor. Folia plerumque verticillata; caulis plerumque herbaceus) comprises the genera which are now included in the Rubieae in the ( Stellatae, Galieae) and, addition, genus Anthospermum. Group II (Fructus dicoccus dispermus. Stamina quattuor, rarius

quinque aut sex. Folia saepius opposita, mediante vagina ciliata; caulis plerumque herbaceus) agrees in the main with the Spermacoceae, though Knoxia Houstonia also been the genus and, by mistake, the genus have included. 111 bilocularis Group (Fructus monocarpus polyspermus. Stamina quattuor. Folia opposita; caulis herbaceus aut frutescens) comprises which for the referred to the genera greater part are by me Hedyotideae. Stamina Group IV (Fructus monocarpus bilocularis polyspermus. rather quinque. Folia opposita; caulis saepe frutescens) forms a unnatural for referred the mixture, it contains genera which at present are to Hedyotideae, the Gardenieae, the Mussaendeae, the Cinchoneae and the Rondeletieae. V bilocularis Group (Fructus monocarpus dispermus. Stamina sex aut plura. Folia opposita; caulis frutescens aut arboreus) com- but viz. Hillia and of which Hillia prises two genera, Duroia, occupies isolated in Duroia in all a very position the family, whereas agrees essential with Gardenieae. points the Group VI (Fructus monocarpus bilocularis dispermus. Stamina quattuor. Folia opposita; caulis plerumque contains which included in the frutescens) genera at present are Ixoreae,

the Coussareae and the Guettardeae. Group VII (Fructus monocarpus bilocularis dispermus. Stamina quinque. Folia opposita; caulis frutescens aut mixture of included in the arboreus) comprises a genera now Psychotrieae, the Paederieae, the Anthospermeae, the Ixoreae, the Chiocceae and the and which do the de- Vanguerieae two genera not answer viz. Chimarrhis Simira. scription and Group VIII (Fructus monocarpus 12 C. E. B. BREMEKAMP

multilocularis loculis monospermis. Stamina quattuor aut quinque aut Folia caulis which plura. opposita; saepe frutescens) comprises genera are in the now included Guettardeae, Chiococceae, Ixoreae, Vanguerieae and

Psychotrieae. Group IX (Fructus monocarpus multilocularis loculis polyspermis. Stamina quinque aut plura. Folia saepius opposita; fru-

tices aut contains but three viz. Hamelia Hamelieae herbae) genera, ( ) Patima and Sabicea (since then included in the Mussaendeae, from

where Sabicea removed to tribe of its the was by me a own; position Fatima of is still uncertain). Group X (Flores aggregati supra re- rarius ceptaculum commune aut coadunati. Folia opposita; arbores aut frutices, rarius herbae) comprises a number of unrelated genera which the are now divided over Anthospermeae, Gardenieae, Coussareae, Psycho-

trieae, Morindeae and Naucleeae and one genus of which the position still is uncertain, viz. Cephalanthus. In a remark made on p. 210 he unites these in four viz. with groups more comprehensive ones, one fruits which with the II and didymous corresponds groups I, 111, with two-seeded with the VI and one fruits, corresponding groups VII,

one with many-seeded fruits, corresponding with the groups IV and V,

and with fruits with the VIII one plurilocular corresponding groups his X there is in this clas- and IX; for group apparently no place sification. De with few additions Jussieu’s ten groups were accepted a and emendations by Jaume de Saint-Hilaire in his “Exposition des families naturelles” (I. 2: 428-453. 1805), whereas de Candolle’s original division of the Rubiaceae in four tribes, viz. the Stellatae, Coffeaceae, Chinchonaceae and Guettardaceae (Ann. du Mus. 9: 217. 1827) in the main the made de was apparently based on division by Jussieu in the above quoted remark. la In his “Memoire sur famille des Rubiacees” (Paris, 1830) A.

Richard distinguished 11 tribes, of which the first seven are charac- terized the of ovule in each of the by presence a single ovary cells,

whereas in the four other tribes each ofthe ovary cells contains several ovules. The tribes with uni-ovular cells the ovary are Asperuleae (= Rubieae), Anthospermeae, Operculariaceae, Spermacoceae (in this tribe

the genus Cephalanthus was included), Coffeaceae (contains genera which the are now distributed over Psychotrieae, Coussareae, Paederieae, Ixoreae

and Vanguerieae), Guettardaceae (contains the genera with a drupe

bi- several i.e. containing a or plurilocular pyrene or pyrenes, apart

from the Guettardeae such totally different genera as Saldinia, Nonatelia, Retiniphyllum. Lasianthus and Morinda) and the Cordiereae (said to be intermediate between the Guettardaceae and the Gardeniaceae it ; com-

prises three rather widely different genera, viz. Cordiera, Myrmecodia and The tribes with cells the Tricalysia). pluri-ovular ovary are in main with de Hamelieae (agreeing the Jussieu’s group IX), the

Isertieae (a small tribe differing from the preceding one in the fruit,

which is 4-5 it contains the a drupe containing pyrenes; genera Isertia, Gongalea Metabolus and the (= Gonzalagunia), Anthocephalus) ,

Gardeniaceae to with two locules, each with several (said possess a berry it is mixture of the Gardenieae and the seeds; a genera belonging to

Mussaendeae and contains in addition several genera which belong to POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 13 other of the is tribes or which position still uncertain), and the Cinchoneae (with bilocular capsules with several seeds in each locule; it comprises besides Cinchoneae the genera belonging to the Rondeletieae, the Nau- cleeae and the and isolated like Hedyotideae generaoccupying an position, Sickingia and Hillia). Richard based his classification therefore in the main the number of ovules found in of cells on each the ovary and on the nature of the fruit, but the characters by which the various of little kinds of fruits are distinguished, are morphologically or no

and importance, practically no attempt has been made to correlate these with differences characters observed in other parts of the plants.

It is therefore not to be wondered that but few of the tribes dis-

tinguished by him can be accepted as natural ones or more or less natural ones. This applies to the tribes Asperuleae, Anthospermeae.

Spermacoceae and perhaps to the Opercularieae.

In the classification Candolle in the elaborated by de “Prodro-

mus” (4: 342-343. 1830) 13 tribes are distinguished, of which the first five with each other in the of several seeds in the agree presence the loculi of fruit, whereas the other eight tribes have but a single seed in each of the loculi. To the first the Cinchonaceae group belong (whith bilocular capsules and winged seeds) subdivided in Naucleae (with sessile fruits) and Cinchoneae (with pedicellate fruits), the Gar-

deniaceae (with bilocular or occasionally by abortion unilocular fleshy fruits and seeds without wings) subdivided in Sarcocephaleae (with

but sessile, capitate fruits) and Gardenieae (with pedunculate or sessile,

not capitate fruits), the Hedyotideae (with bilocular capsules and seeds withoutwings) subdivided in Rondeletieae(with triangular, free stipules) the base and the Hedyoteae (with stipules that are at united into a

sheath and at the top split in several narrow processes), the Isertieae (with drupes containing 2-6 pyrenes) and the Hamelieae (with berries

containing several locules). The second group comprises the Cordiereae the (with baccate, plurilocular fruits), Guettardaceae (with drupes con-

taining 2-10 pyrenes and cylindrical seeds) subdivided in Morindeae

(with capitate or connate fruits) and Guettardeae (with more or less distinctly pedicellate fruits), the Paederieae (with but slightly fleshy fruits of the which which fleshy parts are soon lost, after the two

flattened pyrenes remain attached to the top of the bipartite axis;

seeds with a fleshy endosperm), the Coffeaceae (fruits said to be bilocular

the berries with two semiglobose seeds, latter with a longitudinal the flat subdivided groove on side; endosperm horny) in Coffeae (with and distinctly pedicellate fruits) Cephaelideae (with capitate fruits sur-

rounded by bracts), the Spermacoceae (with more or less dry fruits

4 containing 2 to pyrenes, and with a bilamellar stigma) subdivided in Cephalantheae (with sessile flowers), Euspermacoceae (with pedicellate and into and flowers fruits splitting 2-4 mericarps Putorieae (with a

more or less fleshy fruit which does not split), the Anthospermeae (with less fruit which into a more or dry splits two mericarps or, more rarely, fleshy and bilocular, and with long and hairy stigmata), the Stellatae which into (with a more or less dry fruit splits two mericarps and and with or, more rarely, fleshy bilocular, capitate stigmata) 14 C. E. B. BREMEKAMP

and the Operculariaceae (with unilocular one-seeded connate fruits the with which finally open at top two valves). If the classification ofde that of we compare Candolle with Richard,

we see that two new tribes have been added, viz. the Paederieae, which

is be natural and the which to regarded as a one, Hedyotideae, com- prises two subtribes, the Rondeletieae and the Hedyoteae, and as these subtribes show but this tribe is ob- a very superficial resemblance, viously an entirely artificial one. Some of the other tribes too are

subdivided, but in the case of the Gardeniaceae, the Guettardaceae and the Coffeaceae these subdivisions merely emphasize the unnaturalness of the tribes which divided in this are way. In the classification of Lindley (The Vegetable Kingdom: 761 and the is 768. 1846) the name of family changed into Cinchonaceae. This

has apparently been done because one of the tribes, viz. that of the tribe raised Stellatae, has been excluded; this was by Lindley to family

has been and the rank. The tribe Cordiereae suppressed, name Cof- feaceae has been replaced by Psychotridae, but otherwise there are no differences of with the classification of de Candolle. any importance In Bentham and Hooker’s “Genera Plantarum” (II, 1. 1837) Hooker the ofthe number of ovules too splits family on account per of of ovary cell, but instead two main groups, he distinguished three them. Series A the tribes with comprises pluri-ovular ovary cells, series B those with collateral ovules, and series G those with uni-ovular

cells. Series A and series subdivided in ovary C, moreover, are each A in sub- two subseries; series according to the nature of the fruit a

series with dry fruits and one with fleshy fruits, and series C according

to the position of the embryo in the seed in a subseries in which the radicle points downwardsand one in which the radicle points upwards.

In the first subseries of series A two of the subtribes created by de

Candolle, viz. the Naucleae and the Rondeletieae are raised to the

rank of tribes, whereas an entirely new tribe, that of the Henriquezieae, is the in added; however, two genera brought together this tribe

show no distinct affinity with the Rubiaceae, and the tribe therefore excluded and in the rank of in the was by me, placed a family vicinity of the Bignoniaceae (see chapter II). In the second subseries de Can- the Mus- dolle’s Isertieae are suppressed, whereas two new tribes,

saendeae and the Catesbaeeae are added. Series B comprises two tribes,

the Cruckshanksieae and the Retiniphylleae, of which the first shows an

the tribe which it unmistakable resemblance to Hedyotideae, a in was

sunk by Schumann, whereas the second was created for the aberrant which Richard of genus Retiniphyllum, by was regarded as a near ally Nonatelia, whereas it was put by Schumann in the Gardenieae; both has authors, however, were mistaken, for as Retiniphyllum no raphides,

it be the can no near ally ofNonatelia, a genus belonging to Psychotrieae,

and as it does not show the pollination mechanism which is typical

for the Gardenieae, it can not be included in the latter either. In the of C Morindeae removed from the first subseries series the are neigh-

bourhood of the Guettardeae, with which in fact they show no near and transferred the second whereas four affinity, to subseries, new POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 15

tribes are added, viz. the Knoxieae, the Chiococceae, the Alberteae and the

Vanguerieae. In the second subseries de Candolle’s Coffeae are split into three tribes, viz. the Ixoreae, the Psychotrieae and the Coussareae. In comparing Hooker’s classification with that of de Candolle, it strikes us that great importance is attributed to the aestivation of the corolla lobes, a point which in the classification of de Candolle played no and which Hooker the Ixoreae the part, enabled to separate from Psy- chotrieae and the Coussareae, and that in the series with uni-ovular cells attention is the of the in the seed ovary paid to position embryo

(radicle either superior or inferior), i.e. to the insertion of the ovules; of attention drawn in the case the Guettardeae, moreover, the is to the absence of the The total number of or poor development endosperm. tribes distinguished by Hooker is 25 against 13 by de Candolle. The classification adopted by K. Schumann in Engler & Prantl’s “Natiirliche Pflanzenfamilien” (IV, 4: 16. 1891) is in the main founded that on of Hooker. Schumann too begins with a division based on the

of but in Hooker but number ovules, contrast with he accepts two main for which he introduces the Cinchonoideae and groups, names Coffeoideae. The Cinchonoideae are divided in Cinchoninae (with dry

fruits) and Gardeninae (with fleshy fruits), the Coffeoideae on account of the position of the radicle in Guettardinae (radicle superior) and Psychotriinae (radicle inferior). In the Cinchoninae he includes six tribes, viz. the Condamineae, Oldenlandieae (= Hedyotideae ), the Rondeletieae, the

Cinchoneae, the Henriquezieae and the Naucleeae, the same tribes, there- fore, as Hooker included in his subseries 1 of series A. The Gardeninae

are divided in two tribes, the Mussaendeae and the Gardenieae, cor- responding with the tribes VII and X of subseries 2 of Hooker’s series A; the two other tribes which Hooker recognized in this sub- the the series, Hamelieae and Catesbaeeae, are included by Schumann which is for in the Gardenieae, no improvement, these tribes have very

little in common with each other and with the true Gardenieae. Of

the two tribes for which Hooker created his series B, the first, the included in his with Cruckshanksieae, are by Schumann Oldenlandieae,

which whereas the they are doubtless nearly related, Retiniphylleae were him in included by his Gardenieae, which, as pointed out above, was a mistake. The Guettardinae correspond to Hooker’s series C subseries 1,

and the same five the Vaneueneae comprise tribes, Alberteae, Knoxieae, , Guettardeae and Chiococceae. The Psychotriinae similarly correspond to the Hooker’s series C subseries 2, and comprise same eight tribes, viz. the Ixoreae, Psychotrieae, Paederieae, Anthospermeae, Coussareae, Morindeae, Spermacoceae and Galieae. The differences with the classification elabo-

rated by Hooker are therefore but few, and the characters on which main the classification rests, are in the the same.

The classifications discussed above are all based in the first place number of ovules whereas for the subdivision on the per ovary cell,

of the main much was attached the of the groups weight to nature in fruits, which to this end were distinguished dry and fleshy ones, a distinction which is morphologically unsound and therefore taxono-

mically of little or no value, and subsequently also to the position of 16 C. E. B. BREMEKAMP

the radicle, which is certainly of importance, but ofwhich nevertheless

the importance has been overrated, for the Guettardeae show but little

affinity to the other tribes of the Guettardinae, and the Ixoreae show

more affinity with the latter than with the other tribes of the

in Psychotriinae, the group which they were included by the authors of these classifications. This means that the Guettardinae and the be artificial Psychotriinae are to regarded as groups. and More recently, so far only tentatively, two new classifications

from were proposed, which deviate considerably the older ones. The of is first these two new classifications that which I myself expounded th at one ofthe meetings of the section for taxonomy of the Bth8 Botanical

Congress held in Paris in 1954, but as of this classification only a very brief report was published (Rapports et Communications aux sections 5 et 6: and as I have since then 2,4, 113-114), changed my views I will first discuss the other of the on some points, one, as latter a more detailed report has been published. This is the one the of proposed by Verdgourt (Remarks on classification the Ru- biaceae in Bull, du Jard. bot. de I’fitat, Bruxelles, 28: 209-281. 1958). Verdcourt distinguishes three subfamilies l.c. 250—252), viz. Rubioi- deae (raphides present and seeds albuminous), Cinchonoideae (raphides and hairs and absent seeds albuminous; never septate) Guettardoideae

(raphides absent, seeds exalbuminous or with traces only ofalbumen).

The Rubioideae are subdivided into Psychotrieae, Coussareae, Morindeae, Knoxieae, Paederieae, Coccocypseleae Schradereae, Urophylleae, Craterispermeae, , Argostemmateae, Ophiorrhigeae, Hamelieae, Cruckshanksieae, Hedyotideae, and Rubieae. This therefore in- Anthospermeae, Spermacoceae subfamily cludes the tribes which were divided by me over two subfamilies, viz. the Rubioideae and the Urophylloideae. The last-mentioned sub- family is included by him on the ground that the styloid crystals of calcium from the oxalate differ too little true raphides, a view with which but few specialists in this field will agree. In my opinion there is little tribes included in this sub- very resemblance between the in the family and those which were united by me subfamily Rubio- tribe ideae. With the introduction of the new Craterispermeae and with the restoration of the Cruckshanksieae I The Cinchonoideae agree. comprise in Vercourt’s delimitation the tribes Naucleeae, Cinchoneae, Rondeletieae

which the Condamineae (in are included), Mussaendeae, Catesbaeeae, Gardenieae, Ixoreae (divided in Ixorinae, Cremasporinae and Heinseniinae,) Retiniphylleae, Alberteae, Vanguerieae and Chiococceae. These tribes were divided for reasons which I will expound hereafter, by me over two which in subfamilies, the Cinchonoideae and the Ixoroideae, my reduction the opinion have but little in common. The of tribe Cre-

of the Ixoreae is the masporeae to a subtribe hardly justifiable; genus Verdcourt’s remarks this Heinsenia was not seen by me, but on genus do with the Ixoreae. With to his not prove a near affinity regard reduction of the Condamineae to the Rondeletieae I feel rather sceptical, but I admit that both tribes known and deserve are very imperfectly of the a thorough revision. In his third subfamily, that Guettardoideae, POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 17 he recognizes but one tribe, the Guettardeae. In this point there is full

agreement between his classification and mine.

I will of the remarks now give a summary on the classification of the Rubiaceae made me in various of by papers on genera or groups this genera belonging to family, and expound the classification at which I have arrived. is still finally This, however, a provisional one, for there still considerable are a number ofgeneraof which the position remains uncertain. As I unable continue work in this am to my I leave of field, must the study these genera to others. In the above-mentioned read the Bth8th Botanical paper at Congress I distinguished six subfamilies, viz. the Cinchonoideae, the Urophylloideae, the Ophiorrhizoideae, the Guettardoideae, the Ixoroideae and the Rubioideae.

that the Since then I have added a seventh subfamily, of Gleasonioideae “On the (see my paper position of Platycarpum Humb. et Bonpl., Henriquezia Spruce ex Bth. and Gleasonia Standi.” in Acta Bot. Need.

6: 351-377. and in the I will 1957), present publication add an eighth the Hillioideae. of the tribes included in the one, Here, moreover, one Ophiorrhizoideae, the Ophiorrhizeae, is transferred to the Urophylloideae, that the so for remaining tribe, the Pomazoteae a new subfamily name had to be coined, viz. Pomazotoideae.

The Cinchonoideae in the are delimitation proposed by me easily

recognizable by the structure of the testa, which consists of cells with the vary large pits, occasionally accompanied by ordinary, very minute ones; the large pits, however, are always confined to the basal in the thick lateral walls wall; very they are totally lacking. A second important diagnostic character is to be found in the elongated form of the placenta. Other general characters, though not confined to this

are the multiovular cells and the absence of subfamily, always ovary raphides. Genera which show another kind of placentation and whose

testa cells show another kind of of also sculpture, and, course, those in which This raphides are present, should be excluded. applies e.g. to the which for time known genus a long was as Sickingia Willd.,

but of which the correct name is Simira Aubl. “The (see my paper on identity of Simira tinctoria Aubl. in Acta Bot. Neerl. 3: 150-153. 1954),

to the Gleasonia to the and to several genus Standi., Hedyotideae genera with provided raphides which were included in the other tribes;

these genera do not show the typical structure of the testa cells found

in the true Cinchonoideae nor the type of placentation which is charac- teristic for the latter.

The Urophylloideae are still imperfectly characterizable. Their feature most striking are the large, thick-walled testa cells which differ from those found in the Cinchonoideae by the absence of the peculiar

that Testa of the large pits which are typical for subfamily. cells same kind the as those of Urophylloideae, however, are found also in the Coccocypseleae and the Schradereae, tribes which differ from those included in the the of Urophylloideae by presence raphides. That this kind of would be crystals present in Urophyllum Wall., as reported by Solereder, 18 C. E. B. BREMEKAMP is in this and its allies calcium oxalate is in a mistake; genus present the form ofstyloids.

The third subfamily, that of the Pomazotoideae, is characterized of which the shows unusual by a testa consisting of cells basal wall an kind of dots, which seem to consist of bundles of very small pits. In this the cells and subfamily too ovary are always pluri-ovular raphides absent. are always

The fourth is that of the Gleasonioideae. It with subfamily agrees the next subfamily, that of the Guettardoideae, in the exalbuminous seeds, but differs from them, and indeed from all other subfamilies of the Rubiaceae in the large size of the cotyledons. Other points of difference with the Guettardoideae are the absence of crystals in the wall of the hairs, the pluri-ovular ovary cells and the capsular

fruit. In the wall structure ofthe testa cells they show some resemblance the with which also in the to Hillioideae, they agree pluri-ovular ovary cells and in the capsular fruits, but from which they differ in the ab- of in the dehiscent and in the sence raphides, loculicidally capsule

absence of the tuft ofhairs at the top of the seed.

The fifth subfamily, that of the Guettardoideae agrees, as stated above, with the preceding subfamily in the exalbuminous seeds, but

it differs conspicuously from the latter in the smaller size of the co- the uni-ovular cells and the fruits. A tyledons, ovary drupaceous

character in which this subtribe differs from all the other ones, is the

of of calcium oxalate in the wall of the hairs with presence crystals the various These which parts are covered. crystals are absent only

in the genus Machaonia, which, however, on account of the entirely is be excluded. In different fruits, to this subfamily too raphides are always absent.

The sixth subfamily, that of the Ixoroideae, comprises all those

tribes the of the in which upper part style acts as a receptaculum that the pollinis, which means hairs by which it is covered, collect

already before the flower opens, the pollen from the anthers; when flower the be removed the opens, pollen grains can by visiting insects, flowers reached which may transport them to which have the stage in which the Other stigmata are ready to receive them. important

characters, though not confined to this subfamily, are the simple

interpetiolar stipules, the insertion ofthe stamens in the corolla throat

and the absence ofraphides. At one time I thought that the pollination this also mechanism which is characteristic for subfamily, was found

in the in their and in the Naucleeae former delimitation genus Cepha- be lanthus, but this view afterwards appeared to erroneous.

The seventh subfamily is that of the Rubioideae, which are cha-

racterized the of feature which it shares with by presence raphides, a the next subfamily, that of the Hillioideae, and by the either valvate imbricate aestivation of the corolla lobes. or, rarely, POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 19

the The eighth and, at least for the moment, last subfamily, Hil- with the lioideae, agrees, as stated above, preceding one in the of but differs from the latter the contorted presence raphides, by aestivation of the corolla lobes and from all other Rubiaceae in the

of tuft of hairs the of the seed. In of presence a at top the structure

the testa cells it shows some resemblance with the Gleasonioideae or it least to those species of Gleasonia of which fully mature seeds were available.

Of these eight subfamilies the Chinchonoideae, Gleasonioideae, Guettardoi- Rubioideae and Hillioideae doubtless natural deae, Ixoroideae, are groups, but this can not yet be said with the same degree of certitude of the and the the Urophylloideae Pomazotoideae, for although tribes that are

referred in of to them, can apparently not be accommodated any the other it is subfamilies, as yet not fully certain that the points in which of sufficient unite of them they agree, are importance to some in a more comprehensive group.

Now that of the subfamilies has been a preliminary survey given, we will turn our attention to the tribes for whose accommodation they

were established.

In the Cinchonoideae I include seven tribes, the Cinchoneae, the Naucleeae, the Condamineae, the Rondeletieae, the Sipaneae, the Mus-

saendeae and the Sabiceeae. In read at the Bth8th Botanical Con- my paper I included in this also gress subfamily a tribe which I called the

Manettieae but this was an as I had out , inaccuracy, for, pointed in work “The African Oldenlandia” already my on Species of (p. 14),

the genus Manettia Lis in the possession of raphides, and is therefore

to be included in the Rubioideae.

The Cinchoneae were in the delimitation accepted by the older authors rather artificial of A a conglomerate genera. comparatively number of them to be excluded account of the large are on presence “ of These TheAfrican of raphides. genera are (see Species Oldenlandia”,

p. 14) Bouvardia Salisb., Heterophyllaea Hook, f., Hindsia Bth., Hyme-

nopogon Wall., Manettia L, Danais Commerc. ex Vent., Coursiana Ho-

molle and Hillia Jacq.; they were included in this tribe either on ac- of their count winged seeds, a character which is of little importance

so long as it is not proved that the wings may be regarded everywhere

as homologous structures, or else (Hillia) because of their adnate Other that be placentas. genera are to removed, are Crossopteryx Fcnzl . , and these show Coptosapelta Korth.; as genera the pollination mecha-

nism which is typical for the Ixoroideae, they will have to be included in that subfamily. Coptosapelta would, according to Solereder, be but this is provided with raphides, a mistake. The position of some other is still uncertain. On the other of the genera hand, most genera

which so far have been included in the Naucleeae, will have to be

removed to the as differ from the other of Cinchoneae, they genera 20 C. E. B. BREMEKAMP

this tribe in the inflorescence These capituliform only. genera are Adina Salisb., Mitragyne Korth., Uncaria Schreb., Neonauclea Merr. and probably also Anthocephalus A. Rich, and Breonia A. Rich. On account of the be capituliform inflorescence they might perhaps united in a subtribe.

The Naucleeae restricted to the Nauclea L in the are by me genus sense in which it was defined by Merrill, i.e. as congeneric with Afzei. In this delimitation characterized the Sarcocephalus they are by

of fruits and but presence connate ovaries, fleshy compound slightly The has that it is winged seeds. testa not yet been studied, so not fully certain that the inclusion of this tribe in the Cinchonoideae is The of the L remains justified. position genus Cephalanthus as yet As each of the cells contains uncertain. ovary but a single, pendulous it is doubtful whether this be included in the ovule, very genus can Cinchonoideae. It has sometimes been referred to the Spermacoceae, but this is the attached certainly a mistake, as ovules are not to the middle ofthe dissepiment, but to its top. Before its position can be determined, this will have be studied in detail. genus to more

The Condamineae differ from the Cinchoneae in the horizontal, not

vertical, position of the ovules and in the not or but slightly winged seeds. characters with In these two they agree the Rondeletieae, from which they differ in the valvate aestivation of the corolla lobes.

The Rondeletieae differ from the Condamineae in the aestivation of

the corolla lobes, which in this tribe is either imbricate or contorted.

these of aestivation in imbricate However, as two types are, contrast to

and valvate, but rarely met with in nearly related genera, it is not that further will lead of tribe. impossible study to a splitting up this

Of the genera which so far have been included in the Rondeletieae, will have be removed other some to for reasons. This applies to the

Cham, which is genus Deppea et Schlecht., according to Solereder with which provided raphides, Sickingia Willd., as I have shown

the elsewhere, is identical with Simira Aubl., and which on account of

structure of the testa is removed by me to the Urophylloideae, Gleasonia

Standi., which is the type of my subfamily Gleasonioideae, and Sipanea Aubl. Hook, and Limnosipanea f, which on account oftheir herbaceous

the growth, a very unusual feature in Cinchonoideae, are transferred to

a tribe Sipaneae.

The Sipaneae are a small tribe containing but three genera; on of their herbaceous somewhat isolated account growth they occupy a in the These position subfamily. genera are Sipanea Aubl., Virecta Vahl and Limnosipanea Hook. f.

The Mussaendeae in delimitationbut small tribe. too are my a

Apart from the genus Mussaenda L and its nearest allies, the genera and Asemanthia Ridl. it the Isertia Aphaenandra Miq. comprises genera and and few other Schreb. Cassupa Humb. et Bonpl. perhaps a ones POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 21 which are less well known, but the great majority of the genera in this which by the earlier authors were included tribe, are to be trans-

ferred to other ones, often even to tribes belonging to other subfamilies;

some of them do not even belong to the Rubiaceae. The characteristic

the features of genera which are left by me in the Mussaendeae, are the bifid or bipartite stipules, the valvate aestivation of the corolla lobes and the insertion of the ovules discoid Of the on a placenta. genera which were included in this tribe by Schumann, Schradera Vahl, Lucinaea DC, Mycetia Reinw., Myrioneuron R. Br. and Coccocypselum

ofthe [P. Br.] Sw. are on account presence ofraphides to be removedto the whereas Arn. Korth. 1 Rubioideae, Acranthera ex Meisn., Coptophyllum ),

Pleiocarpidia K. Sch. (Aulacodiscus Hook. f. non Ehrenb.), Urophyllum

l ) The name Coptophyllum Korth. used by Miquel for two specuies belonging to the Pomazota Ridl. should be In of the genus rejected. my “Monograph genus Pomazota Ridley” (Journ. Arnold Arbor. 28: 186-203.1947) I did this because Coptophyllum Korth. is a later homonym of Coptophyllum Gardner. I overlooked at that time that Coptophyllum Korth. is inserted in the list of nomina generica in it that the of of conservanda; fact, never occurred to me name a genus which but three species had been described, all of them, moreover, plants which had but rarely been mentioned in the literature, might have been entered in this list! For the I did into the of the same reason not go very deeply question generic identity

of the species described by Korthals; I even included it provisorily in Pomazota,

viz. with the remark that it was a “species non satis nota”. As since then, however, in the Korth. have been it some new combinations genus Coptophyllum proposed, seems indicated to discuss this question somewhat more in detail. The of the which Korth. type specimen species on Coptophyllum was founded, is no longer available, and was probably already lost at the time at which Miquel second The of the is described a species. description type species very incomplete, difficult and so incomplete that it will be probably even impossible to identify this species. In the description of the first of the two species which were included in this the with by Miquel genus, generic name was provided an interrogation mark. That Miquel was not certain that this species really belonged to the same described means that he had to genus as the plant by Korthals, evidently rely the of the stated is only on description latter, which, as above, very incomplete. If he the he have felt for had seen type specimen, would not any doubt, Pomazota is That the mark an easily recognizable genus. Miquel omitted interrogation when he described his second species, is probably to be explained by assuming that he compared it with his first species and found it to be congeneric with that one.

IfI am right in assuming that Miquel had to rely only on the description given

and it to that this then he must by Korthals, seems me can hardly be doubted, have been struck by the fact that the corolla was described by Korthals as naked inside, the stipules as obtuse and the fruit as indehiscent, characters that are not found in the two species described by himself, and, in fact, in none of the other in the it is that species described genus Pomazota; true, however, Korthals put an interrogation mark behind “indehiscens”.

Now, is it permitted to assume that there can be no doubt with regard to the taxonomic of the described Korthals? This identity genus by question can certainly not be answered in the affirmative, and we will therefore have to admit that Miquel made mistake in his new this of doubtful a referring two species to genus identity.

It is possible, of course, that Korthals made a mistake in describing the inside

of the corolla tube as naked and another one in describing the stipules as obtuse,

but in order this would have which in all other to prove we to produce specimens would with the and such respects agree description given by Korthals, specimens been found. We will have concede therefore that have as yet not to Coptophyllum

Korth. is anunidentifiable name, and that its inclusion in the list ofnomina generica conservanda was not justified. For the two species described by Miquel the generic

name Pomazota has to be accepted. 22 C. E. B. BREMEKAMP

Wall., Pauridiantha Hook. f. and Praravinia Korth. have to be excluded because their testa cells do not show the structure which is typical for the subfamily to which the Mussaendeae belong; the position of the Tammsia genera Karst., Pentagonia Bth., Hippotis Ruiz et Pav. and

Sommera Schlecht. is, as was pointed out in chapter II, as yet uncertain; of Tammsia it is dubious whether it even very can be left in the family.

The from Sabiceeae are excluded by me the Mussaendeae because of the the inflorescences and the simple stipules, axillary very narrow cells. This tribe contains but viz. Sabicea Aubl. testa a single genus,

In the Urophylloideae I include the Urophylleae, the Pauridi- antheae and, provisionally, the Simireae and the Ophiorrhigeae. Whether

Verdcourt’s Heinsiinae, which were included by him in the Mus-

this this saendeae, may perhaps be shifted to subfamily, in which case have raised subtribe would to be to the rank of a tribe, remains to be seen. Following Schumann’s example. I would include in this taxon besides the genus Heinsia DC the genus Bertiera Aubl. The testa cells of Bertiera African of Tab. (The Species Oldenlandia, Vllb) agree in structure with those of the Urophylloideae (op. cit. Tab. IV b-h and Tab. Vb).

The their Urophylleae are characterized by dioecism, the pluri-

locular ovary with in each ovary cell two fully distinct, subpeltate which axis placentas are attached near the and shift as the fruit is

the ripening to the central part of dissepiment, and the more or less This tribe distinctly spreading stigmata. comprises the genera Urophyl- lum Wall., Antherostele Brem., Didymopogon Brem., Maschalocorymbus Brem., Pravinaria Brem., Praravinia Korth., Raphidura Brem., Leucolo- phus Brem., Lepidostoma Brem., Crobylanthus Brem., Pleiocarpidia K. Sch. and Stichianthus Val.

The Pauridiantheaehave bisexual, heterostylous flowers, a usual- with in ly bilocular, more rarely a quadri- or a quinquelocular ovary each of the cells a false dissepiment extending from the top for some

distance downwards, axillary placentas which at the top are incised show by the false dissepiment and therefore an obcordate shape, and

stigmata which are either cohering or free, but never spreading. To

this tribe belong the genera Pauridiantha Hook, f., Pampletantha Brem., Stelechantha Brem., Commitheca Brem., Poecilocalyx Brem. and Rhipi- dantha Brem. and possibly also Temnopteryx Hook. f. and Pentaloncha Hook. f.

Simireae but Simira The comprise a single genus, viz. Aubl., bet- known under the Willd. “The ter name Sickingia (see my paper on identity of Simira tinctoria Aubl. in Acta Bot. Neerl. 3: 130-133. 1954).

This shows in the structure of its cells genus testa a striking resem- the but blance with the genera belonging to two preceding tribes, in it differs from them. Its other respects considerably most striking POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 23 characters found in the in which the ovules inserted are way are on the and in the seeds which in the half extend placenta large upper into a wing. On account of the horizontal position of the ovules referred the Sickingia was originally to Rondeletieae, but in reality the resemblance is rather superficial, for the number of the ovules is much than in the but The smaller Rondeletieae and they form two rows. form and the mode of insertion of the wing is quite different from that found in the seeds of other member of the The any family. genus

doubtless occupies a very isolated position in the family, but in order to decide whether it has correctly been referred to the Urophylloideae or not, it will have to be studied in more detail.

The the three which included same applies to genera by me were in the Ophiorrhizeae, viz. Ophiorrhiza L, Spiradiclis 81. and Virectaria Brem. On the account of thick-walled testa cells they are provisionally referred to this subfamily. In the structure of the testa cells they show African of a very striking similarity (The Species Oldenlandia :

Tab. IV a, b and c), but in other respects they differ rather con-

spicuously. According to Verdcourt Ophiorrhiza would contain raphi- des, but it is not impossible that he mistook styloid crystals for raphides. Of he “leaves Spiradiclis, however, says contain optically active ob- into that longs which separate numerous needle-shaped crystals”, so

we will have to assume that here indeed raphides are present. In that this least be case genus at would have to returned to the Hedyotideae.

The Pomazotoideae far the contain so but a single tribe, Po-

mazoteae, which is characterized by a testa consisting of thin-walled cells with basal which either a wall is minutely dotted or more or

less coarsely tuberculate. It comprises the genera Pomazota Ridl. (The African Species of Oldenlandia: Tab. IV i), Klossia Ridl., Siderobombyx

Brem., Xanthophytum Reinw. ex 81. (op. cit. Tab. VI f), Lerchea L (op. cit. Tab. IV and also j) perhaps Paedicalyx Pierre ex Pitard, Keenania Hook, f., Campanocalyx Val., Polysolenia Hook. f. and Leptomischus Drake. Most of these still genera are very imperfectly known; some of them were originally included in the Hedyotideae, from which

they had to be removed on account of the absence of raphides, whereas others the from had been referred to Mussaendeae, which, as

stated above, they differ in the structure of the testa.

The Gleasonioideae contain far but so a single genus, though the it is not excluded that species which so far have been referred to the will than latter, prove to belong to more one genus. The testa

of Gleasonia macrocalyx Ducke is provided with most peculiar wartlets

each consisting of several cells, but the structure of the mature cellwall could be studied the available seeds as yet not as were too young. In

the two species of which ripe seeds were available, viz. G. duidana the Standi., type species, and G. uaupensis Ducke, the testa does not show but the walls show distinct network of cross-bars any wartlets, a “On the of Humb. (see my paper position Platycarpum et Bonpl., 24 C. E. B. BREMEKAMP

Bth. and Henriquezia Spruce ex Gleasonia Standi, in Acta Bot. Need.

6: 351-377. 1957, fig. 7). The pollen grains are in this genus larger than in other of the Rubiaceae in which could be any genus they cit. measured, and they are, moreover, united in tetrads (op. fig. 3).

Pollen grains that are united in tetrads, however, are not confined

to this genus; they are e.g. a general feature in the genus Randia

(Houst.) L sensu Fagerlind.

The fifth subfamily, that of the Guettardoideae, also contains but one tribe, that of the Guettardeae, but this tribe comprises a number of subfamilies. Of the larger genera than the two preceding

genera which were included in this tribe by Schumann, Machaonia

can, as stated above, not be maintainedin this subfamily, and whether F. with its different Timonius Rumph. (= Abbottia v. Mull.) totally ovaries and fruits has rightly been referred to the same tribe as Guettarda looks dubious. L, very

the In the sixth subfamily, Ixoroideae, a quite considerable

number of tribes are to be included, viz. the Coptosapelteae, Acranthereae, Gardenieae, Ixoreae, Chiococceae and Cremasporeae, Vanguerieae.

The tribe of the Coptosapelteae comprises in my delimitation

the Fenzl. to genera Coptosapelta Korth. and Crossopteryx According Verdcourt (1.c.; 239), the latter would have testa cells of which wall for the is “pitted as in Cinchona”, but this is an error, the pits

smaller than in the and not confined are latter, they are, moreover, to the basal ball, but are found also in the lateral ones (see his plate XVI, fig. QJ. In Hymenodictyon Wall, and Corynanthe Welw., which,

according to Verdcourt, would have the same pollination mechanism

and the of the is not as Coptosapelta Crossopteryx, upper part style distinctly swollen and lacks the short hairs which in the Ixoroideae

and enable this part to function as receptaculum pollinis. Hymenodictyon allies of Corynanthe, moreover, can not be regarded as near Coptosapelta and Crossopteryx, for the aestivation of their corolla lobes is valvate, whereas it is in Coptosapelta and Crossopteryx contorted. Other char- the acteristic features of Coptosapelta and Crossopteryx are subglobose

capsules, the pluriovular ovary cells and the short stipules.

Arn. U\The Acranthereae contain but a single genus, viz. Acranthera Meisn. of the Acranthera Arn. ex (see my “Monograph genus ex Meisn. in Journ. of the Arnold Arbor. 28; 261-308. 1947). In this “remains included between the genus the receptaculum pollinis empty between the anthers and can be reached only through the windows

the On the I projecting tips of connectives” (I.c.: 273). same page of the of the remarked “the connection stamens by means projecting

tips of the connectives with the top of the style is a feature so entirely in the unparalleled Rubiaceae that one might feel inclined to regard

the position of the genus with regard to the rest of the family as

similar to that of the Asclepiadaceae with regard to the Apocynaceae, of connection between the although in this way the importance the POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 25 anthers and the style is doubtless over-emphasized. It is perhaps more

the in which in the L readily comparable to way genera Ceropegia and Dichaelia Harv. ( Asclepiadaceae ) the corolla-tips cohere.” Other characteristic features of the this species belonging to genus are the herbaceous growth, the bilocular ovary with in each cell two elongated the and attached their placentas inserted near axis over whole length to the dissepiment, the reduplicate-valvate aestivation of the corolla lobes, the insertion of the stamens near the base of the corolla tube, the usually baccate, rarely more or less capsular fruits and the numer- ous small seeds provided with a fleshy endosperm and with a testa of cells with wall. consisting a very densely punctate

My next tribe is that of the Cremasporeae, which in the main

with the Alberteae of but does not include the agrees previous authors, Alberta E. Hook. f. genera Mey, Nematostylis and Belonophora Hook. f. Verdcourt (l.c. : 248) also excludes Heinsenia K. Sch. and Lampro- thamnus Hiern. Heinsenia be allied may to Belonophora, but the argu- of which excluded ments on account Lamprothamnus was by him, are

Before the of these not fully convincing. exact position genera can be will be in The three determined, they have to studied more detail. first-mentioned excluded because do not genera were by me they show the pollination mechanism which is typical for the Ixoroideae.

Verdcourt this would be of of According to an example degeneration the typical mechanism, but for this assumption no grounds are ad- duced. The characterized the of typical Cremasporeae are by presence a single pendulous ovule in each of the cells of the bilocular ovary and by the contorted aestivation of the corolla lobes.

The Gardenieae in delimitation much smaller than are my a group in that The delimitation this accepted by my predecessors. given to tribe by Hooker was widened by Schumann, who included in it the of Hamelieae, a tribe rather unnatural composition which nevertheless had been accepted by all his predecessors since the time of de Jus- sieu. The of this the Hamelia removed type taxon, genus Jacq., was by me, on account of the presence of raphides, to the subfamily Rubioideae. in In the delimitation accepted by Hooker and that ac- Schumann the tribe of the Gardenieae the cepted by comprises genera Tarenna Gaertn. (= Webera Schrel. and Chomelia L.), Enterospermum Hiern., Tricalysia A. Rich., Empogona Hook. f. and Morelia A. Rich., which, on account of the structure of the drupe, were transferred by the Ixoreae 6-11 of of the Pavetta me to (see p. my “Monograph genus in L” Fedde’s Repert. 37: 1-208. 1934). Other genera that are to be Aubl. and related Cladoceras excluded, are Posoqueria the genus

Brem. (in Hooker’s leones Plantarum t. 3411. 1940), which possess an entirely different pollination mechanism; the exact position of

remains The these two genera as yet uncertain. true Gardenieae are easily recognizable by their many-seeded, comparatively large fruits which are provided with a thick, leathery or more or less woody pericarp and a gelatinous endocarp in which the numerous seeds are 26 C. E. B. BREMEKAMP

in which the male embedded. They are not rarely dioecious, case flowers with of which the are provided a style upper part serves as remarks this of in “receptaculum pollinis” (see my on group genera of Fruits of the my “Monograph the genus Pavetta L” p. 11-12). kind described above found in the Gardenia Randia are genera Ellis, Houst., Rosenbergiodendron Fagerl., Tocoyena AubL, Genipa L, Alibertia A. L. f. other Rich., Ibetralia Brem., Duroia and perhaps some ones.

The Ixoreae are in my delimitation a larger group than in that of restricted this tribe with uni-ovu- my predecessors, who to genera lar the ovary cells. As number of ovules per ovary cell appeared to in natural from vary perfectly genera from one to two ( Pavetta L) or

to several Tarenna this restriction one ( Gaertn.), was dropped by me, and a comparatively large number of genera were, as stated above, transferred from the Gardenieae to this tribe. The two tribes resemble each other in the contorted aestivation of the corolla lobes. This is doubtless an important character, and for this reason the genera which show aestiva- Phyllomelia Griseb. and StrumpfiaJacq., animbricate tion of the corolla lobes, are to be excluded. Hooker said of these

that “dubiae and too the genera they were affinitatis”, today question of their taxonomic position has not yet been decided. Some other of doubtful discussed in of the genera position were my “Monograph Pavetta 8 and 9. genus L”, p.

In the Chiococceae the cells uni-ovular and the ovary are always ovules always pendulous; another noteworthy feature is the insertion the the of The of stamens near base the corolla tube. only exception to this rule is found in the genus Placocarpa Hook. f. which for this reason will have to be excluded. The aestivation of the corolla lobes is either valvate or imbricate.

tribe be this the The last to included in subfamily are Vangue- rieae. The genera belonging to this tribe resemble those belonging to the Chiococceae in the uni-ovular cells and the ovary pendulous ovules, which however are, as a rule, inserted below the top of the with of these furthermore dissepiment; the majority genera they agree in the valvate aestivation of the corolla lobes. One of the two most important points of difference between these two tribes is found in the insertion of the stamens; in the Chiococceae they are always inserted near the base of the corolla tube, in the Vanguerieae always in the corolla throat. The other important difference is found in the form of the “receptaculum pollinis”, which in the Vanguerieae is short and swollen and at the base usually more or less free from the rest of the

whereas it shows in the the but style, Chiococceae same elongated, slightly swollen form as in the other tribes of this subfamily. The

Bth. lacks the at the of the genus Craterispermum swelling top style, and as it is provided with raphides and with heterostylous flowers it

removed Verdcourt the was by to Rubioideae. POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 27

The Rubioideae are the largest subfamily of the Rubiaceae, and in the in this less comprise survey given paper no than nineteen and this number will increase when the of tribes, probably genera doubtful become better known. position They are, as we have seen, characterized the of and the valvate by presence raphides by or, rare- ly, imbricate aestivation of the corolla lobes. The only other sub- in which the of has been is that family presence raphides reported, created for of aberrant Hillia the accommodation the genus Jacq., where the corolla lobes show a contorted aestivation. The nineteen subfamilies of the Rubioideae are the Hedyotideae, Cruckshanksieae, Argo- stemmatideae, Coccocypseleae, Schradereae, Hamelieae, Spermacoceae, Anthosper-

Lathraeocarbeae meae, Rubieae, Perameae, Psychotrieae, Triainolepideae, ,

Gaertnereae Knoxieae and , Coussareae, Paederieae, Morindeae, Craterispermeae.

the The Hedyotideae are by far largest tribe of this subfamily. They in delimitation the which show valvate comprise my genera a aes- tivation of the corolla lobes, a character in which they agree with most of the other feature which tribes, pluri-ovular ovary cells, a in several returns other ones, a peltate placenta attached to the middle of the mode dissepiment, a of attachment which is found also in the

Hamelieae, non-connivent anthers opening by slits, a point in which the and they differ from Argostemmatideae, a testa consisting of cells with a thin or at least not thick wall. In work “The very my on African of Oldenlandia” Species I transferred the genera Bouvardia

Salisb., Heterophyllaea Hook, f., Hindsia Bth., Manettia Mutis ex L,

Danais Commerc. ex Vent., Coursiana Homolle, Hymenopogon Wall., Myrioneuron R. Br. and Mycetia Reinw. from the Cinchonoideae to the

Hedyotideae. The first seven genera had, on account of the winged seeds, been included in the Cinchoneae, whereas Myrioneuron and Mycetia, on account of their fleshy fruits, had been included in the Mussaendeae; all these and of genera, however, possess raphides a testa consisting the for Cin- cells which lack large pits which are characteristic the chonoideae. Myrioneuron and Mycetia occupy a somewhat isolated po- sition in the of the Hedyotideae on account fleshy fruits and on account of the simple, oblong stipules. The that excluded from the genera were by me Hedyotideae com- prise Lerchea L, Pomazota Ridl. and Xanthophytum Reinw. ex BL, which

transferred the 81. were to Pomazotoideae, Virectaria Brem., Spiradiclis

and Ophiorrhiza L, which were removed to the Ophiorrhizeae, a tribe which in the Carle- was provisionally included by me Urophylloideae,

mannia Bth. and Sylvianthus Hook, f, which, as pointed out in Chapter

are to be excluded from the the 11, family, genera Carphalea Juss., Cruckshanksia Hook, et Arm, Dirichletia Klotzsch, Anotis DC and Jackia the the Wall., which on account of placenta arising from basal part

of the dissepiment and the small number of ovules are included in the and Clarkella Hook. f. and Cruckshanksieae, Argostemma Wall., per- haps Neurocalyx Hook, f., which on account of their connivent anthers the the the basal the and insertion of stamens on part of corolla tube

are referred to a tribe Argostemmatideae. The position of the genus 28 C. E. B. BREMEKAMP

Neurocalyx is not yet clear. If Airy Shaw’s section Thyrsoideae (Kew Bull. 1937: this it will be 281) really belongs to genus, have to trans-

ferred to the Rondeletieae, for in this section the walls of the testa cells show the structure which is typical for the Cinchonoideae (see “The

” African Species of Oldenlandia Tab. II f).

The Cruckshanksieae, a tribe first recognized by Hooker, differ

as stated above, from the Heydyotideae by the small number of ovules in the ovary cells (1 to 3or 4) and the rod-like placentas rising from the base of the dissepiment. To this tribe belong the genera Cruck- shanksia Hook, et Arm, Carphalea Juss., Anotis DC and probably the known Wall. The Dirichletia Klotzsch imperfectly Jackia genus is, according to Homolle (Bull. Soc. Bot. 93: 613. 1936), identical with Carphalea.

The Argostemmatideae differ from the Hedyotideae in the con- nivent with anthers opening a short slit or a pore and in the insertion of the stamens at the base of the corolla tube. To this tribe belong the nearly related genera Argostemma Wall, and Clarkella Hook. f. and perhaps Neurocalyx Hook. f. (see the remarks made above on the position of the section Thyrsoideae).

The Coccocypseleae differ from the Hedyotideae in the simple sti- pules, the subcapitate flowers and the structure of the testa cells with

” their thick walls (“The African Species of Oldenlandia Tab. Ill g), which are of the same type as those found in the Schradereae (op. cit. Tab. HI h and To this tribe the i). belong genera Coccocypselum [P. Br.] Sw. and Lipostoma D. Don.

The Schradereae the most tribe the Rubio- are deviating among ideae with pluri-ovular ovary cells and valvate aestivation of the corolla lobes. The two included in this viz. Schradera genera tribe, Vahl and Lucinaea DC, were originally referred to the Mussaendeae.

They are glabrous epiphytic shrublets with rather thick shoots, fleshy leaves and of large deciduous stipules, usually climbing by means adhesive rootlets; the inflorescences are capituliform; the flowers are

with the fruits baccate and the seeds provided a cupular calyx; are are provided with a testa consisting of thick-walled cells which show

~ a close resemblance to the testa cells of Coccocypselum.

The Hamelieae are in my delimitation restricted to the genera and those in Hamelia Jacq. Hoffmannia Sw., i.e. to which raphides are resemble the tribes in the present. They preceding pluri-ovular ovary cells, but differ from them in the imbricate aestivation of the corolla lobes. The differencebetween the valvate and the imbricate aestivation of the corolla lobes, however, is apparently of less importance than that these of between two types aestivation and the contorted one; in the Psychotrieae in which the aestivation is normally valvate, we

least in which it is imbricate. For this the meet at two genera reason POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 29 rather aberrant structure of the inflorescence found in the Hamelieae, a cyme with monochasial branches, is perhaps of more importance. The should be studied in detail. two genera, however, more

In the tribes the cells contain but now following ovary a single or, exceptionally ( Triainolepideae ), two or three collateral ovules and the aestivation with the of the Brem. is, exception genera Naletonia and Notopleura Brem. belonging to the Psychotrieae, where it is imbricate, always valvate.

The Spermacoceae are characterized by ovules which are attached to the middle of the dissepiment; in the seed the radicle points

downwards, and in this respect there is agreement with the following tribes with the exception of the Knoxieae and the Craterispermeae where

the radicle points upwards. The flowers are usually though not always

4-merous, and the fruits are either schizocarpous or capsular. The united in sheath fili- stipules are a truncate cylindrical fringed with oblate and form appendages. The pollen grains are usually pluricol- with short The Perama Aubl. excluded because pate colpae. genus was ofits parallel-nerved leaves, rudimentary stipules, the two-lobed calyx, the pendulous ovules and the globose tricolporate pollen grains.

The Anthospermeae differ from the Spermacoceae in the position

of the ovules, which are inserted on the basal part of the dissepiment; tribes with the in this respect they resemble the following exception of the Knoxieae and Craterispermeae. The style is usually split in two and the inserted the basal very long stigmata, stamens are usually on the but like the far part of corolla tube, stigmata they project beyond Eleutheranthus the mouth of the tube. The genera F. v. Mull., Opercularia

Gaertn. and Pomax Soland., all three with unilocular ovaries, are

apparently to be excluded, but before their position can be determined, should be studied in detail. The Boiv. they more genus Cremocarpon Boill. of the ex was transferred by me, on account horny endosperm, des Boiv. to the Psychotrieae (see my “Monographic genres Cremocarpon Candollea 16; 147-177. In the ex Baill. et Pyragra Brem.” in 1958). true Anthospermeae the endosperm is always fleshy.

known the The Rubieae, originally as Stellatae and afterwards as the the Galieae, are characterized by the usually verticillate leaves, rudimentary stipules, the usually rudimentary calyx, the didymous the seeds the or by abortion occasionally globose fruits, adhering to pericarp, the often strongly curved embryo and the globose pluri- colpate pollen grains with their long colpae. That the stipules would but the not be rudimentary on contrary strongly developed and transformed into parts which are indistinguishable from true leaves, look in the few Galium does not very probable to me, as species

which decussate distinct small are possess leaves, though stipules pre- leaves sent, and as in the species with verticillate never intermediate

structures are met with. That at the nodes never more than two 30 C. E. B. BREMEKAMP

shoots be valuable axillary are formed, can not regarded as a argu- ment for the stipular nature of the leaves which do not bear shoots in their axil, for in plants with verticillate leaves but rarely more

than one or two of the axillary buds found at the same node develop into shoots. That the leaves in whorls of 6 well of may occur as as

8, and occasionally even of 7, is not favourable either for the view Of that part of them are stipules. the genera which so far have been the included in this tribe Didymaea Hook. f. seems to be only one of dubious position; it has well-developed stipules and seeds which do

not adhere to the pericarp; it might perhaps be transferred to the Anthospermeae but this deserves further , study.

The Perameae, a tribe created for the accommodation of the Perama Aubl. which included in genus originally was the Spermacoceae,

but which differs from the genera which are left by me in that tribe, by the parallel-nerved leaves, the rudimentary stipules, the insertion of the ovules at the base of the dissepiment and the two-lobed calyx. A two-lobed calyx is found also in Declieuxia H.B.K. and Congdonia

two which were referred to the but Miill.-Arg., genera Psychotrieae,

which occupy a rather isolated position in that tribe.

The number Psychotrieae comprise a large of genera, even though

some of those which by previous authors were included in this tribe,

had to be removed to other This to the ones. applies genus Triainolepis Hook, f., which on account of the collateral ovules and the pluri-

locular was made the of a new the Gaertnera pyrene type tribe, genera Lam. and Pagamea Aubl., which on account of the cylindrical stipular sheath and the almost transferred completely superior ovary, were

to a tribe of their own, the Gaertnereae, and the genus Thiersia Baill.

= Evea I show Rubiaceae of ( Aubl.), which, as could (“Notes on the Surinam” in Rec. d. trav. bot. need. 31: 278. 1934) is to be included in Faramea i.e. in the tribe Coussareae. That the Declieuxia Aubl., genera

H.B.K. and Congdonia occupy a rather isolated position in the Psycho-

trieae, has already been pointed out. The most characteristic feature

of the which left in this tribe is the with two genera are by me drupe which differ from those that or more pyrenes, are usually met with, in the thinness of the sclerenchymatous envelope. That in two of the viz. in Naletonia Brem. and in the valvate genera, Notopleura Brem.,

aestivation of the corolla lobes is substituted by an imbricate one, has already been mentioned.

The Triainolepideae differ from the Psychotrieae by their collateral ovules and the of is by presence a single pyrene which, moreover,

provided with a thick sclerenchymatous envelope; another note-

feature the 5 7 filiform worthy are stipules with their 3, or lobes colleter des Triai- which are crowned by a (see my “Monographic nolepidees, tribu nouvelle des Rubioidees” in Proc. Kon. Ned. Akad. Ser. 59: The tribe three v. Wetensch. C, 1-21. 1956). comprises POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 31

viz. Hook. Paratriaina Brem. and genera, Triainolopis f, Thyridocalyx Brem.

The Lathraeocarpeae agree with the Triainolepideae in the struc-

of the fruit with its but differ ture bony plurilocular pyrene, they from them the of but ovule in each of the by presence a single ovary cells and by the pluricolporate instead of tricolporate pollen grains; the united with the base of the sessile leaves stipules, moreover, are and are produced into 1 or 2 stiff points. From the Psychotrieae and most of the other tribes of the Rubioideae they differ in the fleshy instead of called “Les Lathrae- horny endosperm (see my paper 1’Etat ocarpees, tribu nouvelle des Rubioidees” in Bull. Jard. Bot. de Bruxelles 27: 159-166. This tribe contains but 1957). a single genus, viz. Lathraeocarpa Brem.

The Gaertnereae are an often discussed group which on account of the almost less entirely superior ovary seems to occupy a more or anomalous position in the Rubiaceae, and which on account of this

feature was referred by various authors, and at one time by me too,

the with a to Loganiaceae. They are, however, provided raphides, form of crystals which is never met with in the true Loganiaceae, and the is that lack intraxylary phloem which a characteristic feature of family. Another point in which they differ from the Loganiaceae in the of colletors the The almost presence on stipules. entirely superior

is such feature as was ovary, moreover, not an important originally this condition is found also in of assumed, because some the genera

of the In most of their characters the Gaertnereae Hedyotideae. agree from with the Psychotrieae ; the most important difference, apart the is the unionof the almost entirely superior ovary, stipules into a long tribe contains but viz. Gaertnera cylindrical sheath. The two genera, Lam. and Pagamea Aubl.

The Coussareae the but come very near to Psychotrieae, are easily distinguishable from them by the nature of the dissepiment, which is thin which functions for short time and the very and a only, by of with seed. This tribe contains presence a single pyrene a single Aubl. but two genera, viz. Coussarca Aubl. and Faramea

The Paederieae differ from the Psychotrieae by the greater length the In its this of the stigmata and by dry fruits. present delimitation tribe The Aitchisonia makes a rather unnatural impression. genus

Hemsl. is on account of the insertion of the stamens at various heights in the corolla tube and of the of hairs presence glandular certainly second to be excluded; it may, as I have pointed out in the chapter,

belong to the Dipsacales, though the combination of glandular hairs far been found in of the families included and raphides has so not any

in that order. The other doubtless to the but genera belong Rubiaceae, it they differ so strongly from Paederia L that seems hardly allowed 32 C. E. B. BREMEKAMP

include in to them the same tribe. to the which They belong genera

should be studied in more detail.

The Morindeae with the in agree Spermacoceae the way in which ovules from the are attached to the dissepiment, but they differ them in habit and in the nature of the stipules, which though usually united in a sheath, are never provided with filiform appendages; in the of genus Didymoecium Brem., probably a near ally Rennellia Korth., Tribrachya Korth. and Zeuxanthe Ridl., the stipular sheath is produced into lobes intrapetiolar (see my paper on “Didymoecium genus novum Rubiacearum Morindearum” in Bull. Jard. Bot. de Buitenzorg, Ser. 13: 425-428. where critical is of the 3, 1935, a survey given occur- of in rence intrapetiolar stipules the Rubiaceae). The genera which

were united in this tribe, are on the whole still imperfectly known.

The Knoxieae and the Craterispermeae differ from the other tribes of the Rubioideae with uni- biovular cells or, rarely, ovary in of the position the ovules; the latter are in these two tribes attached

to the top of the dissepiment; in the seeds the radicle accordingly Knoxieae L points upwards. The (Knoxia and Pentanisia Harv.) are herbaceous whereas the of the plants, only genus Craterispermeae is the the (Craterispermum Bth.) shrubby. In Knoxieae stipules are usually and in the the fringed Craterispermeae entire; inflorescences, moreover,

are in the Knoxieae terminal and in the Craterispermeae axillary.

that The eighth subfamily, of the Hillioideae, comprises but a single

the and this contains in tribe, Hillieae, tribe my delimitation but a viz. Hillia features of this single genus, Jacq. Noteworthy genus are

the of a character in which it with the presence raphides, agrees genera which are brought together in the Rubioideae, the contorted aestivat-

ion of the corolla lobes, a character in which it differs from all of

and the of of hairs the of the them, presence a tuft on top seed, a

character that is found nowhere else in the Rubiaceae. The wall of

the testa cells shows structure Tab. V in work a peculiar (see a my African of on “The Species Oldenlandia”), which reminds one more

that in in its other or less of found the genus Gleasonia Standi., but Hillia show characters does not any similarity whatever with this genus. the contorted aestivation of the corolla Among genera showing a

lobes there is one other genus which according to Solereder would contain This is the Cham, Schlecht. Ho- raphides. genus Deppea et

wever, as Solereder in some other instances too mistook styloid crystals for it is that this is The there- raphides, not impossible an error. genus, fore, should be reinvestigated.

The in the of the results that since survey given preceding pages

the appearance of the last general treatment of the Rubiaceae, viz. that of K. Schumann in Engler and Brand’s “Natiirliche Pflanzenfami-

lien” (1891), have been obtained by the monographic treatment of a POSITION, DELIMITATION AND SUBDIVISION OF THE RUBIACEAE 33 number of and will have left doubt with genera tribes, no regard to of the insufficiency the older classifications, but it will have shown at the same time that before a fully satisfactory classification can be obtained, a considerable number of genera will have to be reinves- As I unable continue work in this tigated. am myself to my field, I leave this must to a younger generation.