Peltate leaf architecture has evolved from conventional bifacial leaves many times in flowering plant evolution. Characteristics of peltate leaves, such as the differentiation of a cross zone and of a radially symmetric, margin-less petiole, have also been observed in mutants of genes responsible for adaxial-abaxial polarity establishment. This suggests that altered regulation of such genes provided a mechanism for the evolution of peltate leaf structure. Here, we show that evolution of leaf peltation in Tropaeolum majus, a species distantly related to Arabidopsis thaliana, was associated with altered expression of Tropaeolum majus FILAMENTOUS FLOWER (TmFIL), a gene conferring abaxial identity. In situ hybridization indicates that adaxial and abaxial domains are established in early leaf primordia as in species with bifacial leaves. Upon initiation of the cross zone by fusion of the blade margins, localized expansion of TmFIL to the upper leaf side could be seen, indicating a local loss of adaxial leaf identity. The observed changes in expression are consistent with a role of TmFIL in radialization of the petiole and circularization of the leaf blade margin by the cross zone. In addition, expression was observed in segment primordia and during expansion of the bifacial blade, suggesting additional roles for TmFIL in leaf development.

This paper is meant to be a contribution to a better understanding of the focus system of Tundra Yukaghir. It recapitulates the major findings that were made with respect to the topic discussed and presents a number of new empirical facts some of which contradict or form an essential addition to the preexisting views. 4 SA stands for 'Subact' in terms of Hengeveld and Mackenzie (2008) 5 Modified focal nouns receive pronominal focus marking. (Krejnovič 1958:46) 6 The same is true for the pronominal focus markers in 1 st and 2 nd person.

The switch reference (SR) system in Yukaghir exhibits certain characteristics that supplement and partly correct some of the typological generalizations formulated in Haiman and Munro (1983: ix–xv). The central observation made here is that SR in Yukaghir is not primarily a syntactic but a semantic phenomenon which interacts with syntactic and pragmatic factors. It is assumed that this finding is cross-linguistically valid. This analysis is based on a number of published primary and secondary sources as well as on unpublished fieldwork data.