Systematic Significance of Pollen Nuclear Number in Euphorbiaceae, Tribe Euphorbieae
Total Page:16
File Type:pdf, Size:1020Kb
Amer. J. Bot. 69(3): 407-415. 1982. SYSTEMATIC SIGNIFICANCE OF POLLEN NUCLEAR NUMBER IN EUPHORBIACEAE, TRIBE EUPHORBIEAE1 GRADY L. WEBSTER, EARLENE RUPERT,2 AND DARYL KOUTNIK Department of Botany, University of California, Davis, California 95616; and Agronomy Department, Clemson University, Clemson, South Carolina 296312 ABSTRACT Pollen nuclear number is determined in 139 species of 5 genera in the Euphorbieae, subtribe Euphorbiinae. The 111 new determinations are tabulated along with previous reports, and the results indicate that the distribution of binucleate (II) and trinucleate (III) pollen is strongly associated with the taxonomic groupings within the Euphorbieae. Although binucleate pollen is probably primitive within the tribe Euphorbieae, as suggested by the nuclear condition in Neoguillauminia, the situation in Euphorbia still requires further elucidation. Within Euphorbia, the morphologically most primitive species studied have III pollen despite the fact that II pollen is presumably the original condition for the subtribe Euphorbiinae. In Euphorbia, II pollen only is reported from nine sections and III pollen only from ten sections, while in four sections (Esula, Goniostema, Aphyllis, and Deuterocalli) both II and III pollen have been found. The New World species of Euphorbia nearly all have III pollen, whereas the vast majority of the African succulents have II pollen. The genera of New World origin, Chamaesyce and Pedi- lanthus, have III pollen, while the African genera Monadenium and Synadenium have II pollen. Independent derivations of III pollen from II pollen appear to have occurred in sections Gon- iostema, Aphyllis, and Deuterocalli (all of subg. Euphorbia). There is no evidence that reversals from III to II pollen have occurred. IN HIS SURVEY of pollen nuclear number in the time indicated II pollen in Monadenium and angiosperms, Brewbaker (1967) indicated that Neoguillauminia, III pollen in Chamaesyce, the Euphorbiaceae were one of 32 families in Pedilanthus, and Synadenium; and both II and which both binucleate (II) and trinucleate (III) III pollen in Euphorbia. Although a large num- pollen grains occur; at the same time, he noted ber of nuclear number determinations of that Euphorbia is one of only five angiosperm species of Euphorbia had been made by 1973, genera in which both II and III pollen have we have delayed publication of further details been recorded. Webster and Rupert (1973) re- so that a more representative sampling could ported pollen nuclear numbers in 66 species be made of species belonging to the more than of Euphorbiaceae belonging to 42 genera, and 30 sections of Euphorbia recognized by Bois- confirmed the results of Brewbaker. In the tab- sier (1862) and Pax and Hoffmann (1931). ulation of Webster and Rupert, 18 tribes were shown to have II pollen, three to have III pol- METHODs-Determinations of nuclear num- len, while in three tribes both II and III pollen ber were made from fresh pollen collected in were recorded. With the somewhat dubious the Department of Botany greenhouses at Da- exception of the Australian tribe Ricinocar- vis, and also from cyathia preserved in FAA peae (which requires further study), the mixed and stored in alcohol. Pollen was stained in records of II and III pollen are restricted to the iron-acetocarmine saturated with chloral hy- subfamily Euphorbioideae (s. str. of Webster, drate after pretreatments with concentrated 1975), specifically to the two tribes Hippo- HCl and glacial acetic acid in ethyl alcohol maneae and Euphorbieae. (Rupert and Webster, 1972). Pollen which Pollen nuclear numbers were reported by stained with difficulty was treated in alcoholic Webster and Rupert (1973) only for seven HCl-carmine solution as recommended by species in six genera of tribe Euphorbieae. Gardner and Rattenbury (1974). For observa- Reports on nuclear number made up to that tions of starch grains and lipids, pollen was stained with IKI and Sudan IV (Baker and Baker, 1979). Visual observations, especially 1 Received for publication 1 December 1980; revision of critical taxa, and photographs were made accepted 18 January 1981. with a Zeiss microscope using Nomarski in- We thank M. Barbercheck, G. Breckon, M. Craig, B. Dehgan, J. Lavranos, S. Lynch, and R. Pearcy for their terference illumination. For some species, ob- assistance. This study has been supported in part by NSF servations were repeated on two or more grants GB-7109, GB-25347, and DEB-7724263. accessions. 407 This content downloaded from 169.237.8.35 on Mon, 16 Sep 2013 18:15:23 PM All use subject to JSTOR Terms and Conditions 408 AMERICAN JOURNAL OF BOTANY [Vol. 69 Species newly reported in Table 1 are rep- cluding Rauh (1967) and Uhlarz (1974), have resented by voucher specimens in the Univer- indicated that E. cylindrifolia and E. decaryi sity of CaliforniaHerbarium at Davis (DAV), are atypical for Goniostema; hence the evi- or for cultivated living plants, by Davis or dence from pollen nuclear number reinforces Berkeley plant accession numbers. The ar- the argument for referring these species to a rangementof subgenerictaxa within Euphor- different section or subsection, as is suggested bia follows the systems of Boissier (1862) and by Anton (1974). Pax and Hoffmann(1931), as modifiedby Web- It seems probable that a future taxonomic ster (1967, 1975, and ined.). revision will resolve the apparently anomalous occurrence of both II and III pollen in sect. REsULTs-In Table 1 we summarizeour de- Goniostema. However, this may not be so with terminations, as well as those published by regard to sect. Aphyllis (sect. Tirucalli of Bois- previous investigators (omitting some which sier). If the section is treated in the broad sense are definitelyknown to be incorrect).It is clear, of Rauh (1967), who combines Boissier's sec- on the basis of reports on 22 species (17 count- tions Tirucalli and Arthrothamnus under the ed by us), that pollen is III in Chamaesyce. On latter name, three species have II pollen while the other hand, counts of seven species of two species-E. aphylla and E. platyclada Monadenium all indicate II pollen. Our one Rauh-have III pollen (Fig. 6). However, even additional count in Pedilanthus confirms the in the narrow circumscription used in Table 1, report of Brewbaker (1967), so that the genus both II and III pollen are recorded from sect. would appear to have III pollen, although ad- Aphyllis; we have confirmed Schiirhoff' s ditional counts would be desirable. Restudy of (1924) record of III pollen in E. aphylla but Synadenium indicates II pollen; although fur- find II pollen in E. dregeana Mey. ex Boiss. ther determinations of nuclear number are and E. schimperi Presl. Not only Boissier, but needed, that genus might be expected to agree subsequent workers such as Berger (1907), in pollen nuclear number with Monadenium, Rauh (1967), and Jacobsen (1970), have which it resembles both morphologically and placed these species in the same taxonomic cytologically (Jones and Smith, 1969). group. Sampling within Euphorbia has now provid- In the American species of subg. Agaloma, ed pollen nuclear data for 106 species (86 pollen appears to be III in all of the 17 species counted by us) belonging to 24 sections in four examined by us (Fig. 5). Although Lyon (1898) subgenera. Although certain taxa of critical reported II pollen in E. corollata L., this record phylogenetic position have still not been in- appears to be incorrect; we have found only vestigated, we believe that the coverage given III pollen in South Carolina material of that in Table 1 is reasonably comprehensive. As species. Our observations on two species of indicated by the summaryin Table 2, the dis- subg. Poinsettia confirm Brewbaker's report tributionof pollen nuclear numberis strongly of III pollen, and suggest that the reports of correlatedwith subgenericposition. Subgenus II pollen for this subgenus by Sharma (1955) Esula, which includes taxa that presumably and Parkash (1974) are in error. In the 13 are primitivewithin the genus (Webster, 1967), species of Chamaesyce studied by us, the pol- has III pollen in at least 36 of the 38 species len is uniformly III; this is not surprising in (in five sections) which have been studied. The view of the presumed derivation of Chamae- reports of II pollen in E. rothiana Spreng. and syce from North American taxa of Euphorbia E. terracina L., by Srivastava (1952) and subg. Agaloma (Webster, Brown and Smith, D'Amato (1947), respectively, appear dubious 1975). Although we have made only one new and should be rechecked. It is now clear that count on Pedilanthus, this agrees with Brew- subg. Esula pollen is trinucleate in the great baker's record of III pollen. majority, if not all, of the taxa. All seven species of the African genus Mon- Subgenus Euphorbia, which includes the adenium agree in having II pollen, which sup- various succulent taxa of Africa, Madagascar, ports the derivation of Monadenium and Syn- and India, contrasts strongly with subg. Esula adenium from the African succulent taxa of in having II pollen in 42 out of the 46 species Euphorbia subg. Euphorbia as suggested by studied (Fig. 3, 4). However, here again there Pax (1921). Apparently our earlier record of are exceptions. Within the Madagascarian III pollen for Synadenium cupulare (Boiss.) sect. Goniostema, two species (E. cylindrifolia Wheeler was in error, as reexamination of Marn.-Lap. & Rauh and E. decaryi Guill.) have material of this species indicates II pollen. III pollen, while the other nine taxa examined Study of additional taxa of Synadenium would have II pollen. However, various authors, in- be desirable. This content downloaded from 169.237.8.35 on Mon, 16 Sep 2013 18:15:23 PM All use subject to JSTOR Terms and Conditions March, 1982] WEBSTER ET AL.-POLLEN OF EUPHORBIACEAE 409 TABLE 1. Distribution of binucleate and trinucleate pollen in the Euphorbieaea Nuclear Taxon number Source of specimen or literature reference Subtribe NEOGUILLAUMINIINAE Neoguillauminia cleopatra (Baill.) Croiz.