Pollen morphologicalevidence for subdivision and affinities of Lecythidaceae Jan Muller Rijksherbarium, Leiden Summary A in is of demarcation subdivision ofthe pollen types encountered Lecythidaceae proposed. The presence a and line between an original colpate a derived syncolpate pollen type is confirmed. The significance of pollen characters for taxonomic subdivision is evaluated and it is concluded that the subdivision proposed in by Niedenzu 1892 agrees best with the pollen evidence. Pollen morphology does not yet provide any clear indications of wider affinities of the family, except in a negative sense. Introduction of In his monograph the genus Barringtonia (Lecythidaceae), Payens (1967) stated, after reviewing the various proposals for subdivision by former authors, that 'a satisfactory be taxonomie subdivision of the family is still wanting and may not even possible'. Although some of these authors, notably Niedenzu (1892) and Pichon (1945), paid attention to pollen morphology, it was Erdtman (1952) who first pointed out that a line the characterized the clear demarcation runs through family, one part being by syntricolpate Planchonia pollen type, the other by the tricolpate Lecythis type. This Melchior who based his evidence was not taken into account by (1964), subdivision of the family on Pichon (194s). this character In the following it will be attempted to evaluate the significance of for taxonomicsubdivision in Lecythidaceae and in addition to comment briefly on the wider affinities of this family as far as these can be deduced from pollen morphology. This note is dedicated to Professor Van Steenis on the occasion of his retirement and in grateful recognition of his promotion of palynological studies at the Rijksherbarium. POLLEN TYPES of the of A preliminary account pollen morphology the genus Barringtonia was already published in Payen's monograph (I.e., p. 169—171). Since then I have been able to of the extend my observations by studying the pollen all species of related genera Careya, Chydenanthus, Combretodendron, and Planchonia. Their pollen proved to be basically that be included Planchonia similar to ofBarringtonia and can in a main type although the variability within this type is considerably larger than suggested by Erdtman (1952). the A detailed account of this variability will form subject of a later paper. of the of as far as available, was In addition, the pollen remaining genera Lecythidaceae, all the cursorily examined, proving to belong to the Lecythis main type. In following scheme for which the subdivision of Niedenzu (1892) is adopted (table 1), a slightly 1 revised subdivision of the pollen types in Lecythidaceae is presented ). The in 'A. *) opportunity is taken here to correct a most unfortunate error Payens, I.e., p. 170: Calyp- trocalyx main type' should read ‘A. Calyptrata main type'. 351 VOL. 352 BLUMEA XX, No. 2, 1972 of species 'folia neocaledonica calyptrata lougilongifolia papehpapch procera remaining Barringtonia Careya Chydenanthus CombretodendronCombretodendron Planchonia B. B. B. B. B. type groove) typetype groove)groove) calyptrata asiatica marginal marginal Barringtonia (without Barringtonia (with I Asteranthus Crateranthus NapoleoneaNapoleonea Bertholletia Cariniana Chytroma CorythophoraCorythophora Couratari Couroupita Eschweilera Gustavia Holopyxidium A. B. Foetidia Grias Lecythis TABLE PLANCHONIOIDEAE FOETIDIOIDEAE NAPOLEONOIDEAE LECYTHIDIOIDEAE TYPE TYPE MAIN MAIN PLANCHONIA (syncolpate) LECYTHIS (colpate) II.IL I. J. Muller: Pollen morphology Lecythidaceae 353 From this scheme it will be clear that Erdtman's assertion that a pollen morphological line be drawn within the confirmed. demarcation can family can be However, before this fact can be utilized for taxonomic subdivision it is necessary to evaluate the character involved and whether other characters are correlated with to investigate any it. First of all, the fact that the difference between the Planchonia and the Lecythis main is viz. be pollen types a qualitative one, syntricolpate versus tricolpate, must discussed. While in none of the species transitions have been observed, this of course does not imply that earlier in the phylogeny this has also not been the case. In fact, it is likely that the syntricolpate condition has been gradually derived from the tricolpate one. Evidence for this point of view can be found in recent observations on Passifloraceae pollen by Presting (1965) and Pacque (oral communication) which show that varying and degrees of syncolpatism may occur, even within one species that, in that family at the condition is derived A second least, syncolpate a one. argument to consider the condition be based the fact that this characterizes the tricolpate as original can on type earliest Angiosperm pollen grains which occur in the early Cretaceous, while syncolpate later record the Planchonia types appear in the geological (Muller, 1970). Furthermore, other characters also. the and Lecythis main types differ in In the Lecythis type, grains are the axis in the generally smaller, polar rarely exceeding 45// length, endoapertures are often more pronounced as in Eschweilera, Foetidia, Grias, and Napoleona, and the exine is of than is in the Planchonia main generally a more simple structure found type, being thin less distinct typically composed of a endexine, a layer of more or columellae, and a generally rather thin tectum which may be smooth (Bertholletia, Couroupita, Eschweilera, Holopyxidium), fmely reticulate ( Chytroma, Couratari, Grias, Lecythis, Napoleona), foveolate scabrate-verrucate (Foetidia), or (Crateranthus, Gustavia). In contrast, the large (45 —60/u polar diameter) grains of the Planchonia main type lack defined while the generally a clearly endoaperture, syncolpate ektoapertures often show a number of specialized structures in the marginal zones. Especially striking in the in The this respect are polar thickenings occurring the Barringtonia asiatica type. exine with structure on the mesocolpia may show a heavy tectum funnel-like depressions, supported by columnate structures, comparable to the Tilia structure type described by and Godwin and Chambers (1961) Praglowski (1971). However, simple exine structures main is also occur in the Planchonia type and it especially the Barringtonia calyptrata type which be considered least evolved and which must as approaches certain representatives of the and the Lecythis main type especially pollen of Crateranthus. As can be seen on Plate I, where the pollen of two representatives of the B. calyptrata type is compared to that of Crateranthus in both the columellae rather and congolensis, are conspicuous regu- a medium almost smooth tectum larly distributed, supporting thick, (figs. 5, 6). An additional of is the of the striking point similarity presence scattered verrucae on aper- tural membranes This a feature which the Planchonia (figs. 4, 7, 8). is occurs frequently in main in main in type, but the Lecythis type so far only has been encountered Crateranthus pollen. As already stated, the step from tricolpate (fig. 7) to syntricolpate (fig. 1), which is the differencebetween the of Crateranthus main pollen grains and the two representatives of the Barringtonia calyptrata type shown, although at present qualitatively definable, need be construction not a fundamental one. Moreover, the general of the ektoapertures from endo- (apart the syncolpatism) and the absence of marginal grooves and of distinct further The the apertures are points of agreement. highly complex pollen of Barringtonia asiatica is much less similar and side type probably represents a separately evolving branch in the phylogeny of the family. BLUMEA VOL. 354 XX, No. 2, 1972 In this connection it is of interest to review briefly the geographical distributionof the pollen types. The Planchonia main type is restricted to the Old World tropics, ranging from Africa and the The Madagascar to west Pacific. Lecythis main pollen type is found the and The in American tropics, tropical Africa, Madagascar. genus Crateranthus is whilethe African, Barringtonia calyptrata type is restricted to New Guinea, North Australia, the Solomon and Caledonia. obvious that this distribution Islands, Fiji, New It is pattern does not the idea that the latter could be related to support group closely Crateranthus, unless assumes a relict distribution for one the B. calyptrata type. On the other the transatlantic distributionof the main hand, Lecythis pollen type agrees well with the view that this be close the may to the ancestral pollen type of family. The of the Planchonia could then have been origin main type in Africa, presumably at a time when transatlantic not more. this of migration was possible any In connection it is interest mentionthat the earliest which the to pollen grains can be assigned to B. asiatica type date from the Paleocene of Borneo (Muller, 1970) and from the Eocene of India (Venkata- & from chala Kar, 1968), while Payens (I.e., p. 172) cites fossil wood ofBarringtonia the Eocene ofIndia. Puzzling in this context is a recent record of fossil Barringtonia leaves from the Eocene of Alaska (Wolfe, 1972). In it would that the main conclusion, appear two pollen types could be utilized, in combinationwith for subdivision macromorphological characters, a of the family, since they probably reflect a Cretaceous split in the phylogeny, with one branch finding its main development in South America and to a lesser extent in Africa, while the other branch has probably migrated eastwards