Inflorescence evolution in Santalales: integrating morphological characters and molecular phylogenetics

Abstract

PREMISE OF THE STUDY:The sandalwood order (Santalales) includes members that present a diverse array of inflorescence types, some of which are unique among angiosperms. This diversity presents not only interpretational challenges but also opportunities to test fundamental concepts in plant morphology. Here we used modern phylogenetic approaches to address the evolution of inflorescences in the sandalwood order. METHODS:Phylogenetic analyses of two nuclear and three chloroplast genes were conducted on representatives of 146 of the 163 genera in the order. A matrix was constructed that scored nine characters dealing with inflorescences. One character, "trios", that encompasses any grouping of three flowers (i.e., both dichasia and triads) was optimized on samples of the posterior distribution of trees from the Bayesian analysis using BayesTraits. Three nodes were examined: the most recent common ancestors of (A) all ingroup members, (B) Loranthaceae, and (C) Opiliaceae, Santalaceae s.l., and Viscaceae. KEY RESULTS:The phylogenetic analysis resulted in many fully resolved nodes across Santalales with strong support for 18 clades previously named as families. The trios character was not supported for nodes A and C, whereas it was supported for node B where this partial inflorescence type is best described as a triad. CONCLUSIONS: Essentially every major inflorescence type can be found in Santalales; however, the dichasium, a type of partial inflorescence, is rarely seen and is not plesiomorphic for the order. In the family Erythropalaceae, inflorescences are mostly in small, axillary fascicles or cymes. Successive families show both cymose and racemose types and compound systems (e.g., thyrses). Inflorescences in Amphorogynaceae and Viscaceae are not dichasial and in general are difficult to compare to "standard" inflorescences.

Parasitic angiosperms: How often and how many?

“…This publication is mentioned only because at that time most botanists were focused on rbcL and few realized that 18S rDNA had sufficient phylogenetic signal to address questions within angiosperm orders, families, etc. Phylogenetic studies on Santalales from the 1990s to present have increased both taxon and gene sampling such that today nearly all genera in the order are represented by at least some nuclear and chloroplast sequences (Nickrent & al., ).…”

“…() and Nickrent & al. (). Previously the early‐branching, non‐parasitic families included Erythropalaceae, Strombosiaceae, and Coulaceae (clades A, B and D), but new data strongly support a relationship between Strombosiaceae and Octoknemaceae (clades B and C).…”

“…As shown in Nickrent & al. (), many of the relationships among families along the “spine” of the Santalales tree have now been resolved with strong support. It was stated in APG IV () that alterations of APG III must await future resolution; it is here proposed that that time has come for Santalales.…”

“…Previous analyses with less sequence data did not resolve the interrelationships among the families of “Olacaceae s.l.” (clades A–G), even though the support for the individual families was high (e.g., Su & al., ). With new data, the E–G clade now has strong support (Nickrent & al., ), which introduces some interesting questions regarding the origin of parasitism in the order. For Olacaceae, parasitism has been documented for two of the three genera in the family, Olax L. (Pate & al., ) and Ptychopetalum Benth.…”

“…That clade received strong support in Nickrent & al. (), and these two families share a number of morphological synapomorphies. Indeed Linnaeus first classified the North American genus Comandra Nutt.…”