“…The partitioning behaviour of trace elements between melt and minerals has been studied for decades (e.g., Jensen 1973;Hart and Dunn 1993;Blundy and Wood 2003;Dubacq and Plunder 2018), either through the use of glass/groundmass-mineral pairs in rapidly quenched rocks (Onuma et al, 1968(Onuma et al, , 1981Fujimaki, 1986;Minissale et al, 2019), (Hart and Dunn, 1993;Dygert et al, 2014), or theoretical methods (Blundy and Wood, 1994;Van Westrenen et al, 2000;Dubacq and Plunder, 2018). Extensive literature is therefore available on partitioning behaviour for minerals commonly crystallizing from tholeiitic magmas such as clinopyroxene (e.g., Skulski et al, 1994;Johnson, 1998;Green et al, 2000;Hill et al, 2011;Bédard, 2014;Dygert et al, 2014;Shepherd et al, 2022), olivine (e.g., Irving, 1978;Beattie et al, 1991;Beattie, 1993;Taura et al, 1998;Bédard, 2005;Evans et al, 2008) and plagioclase (e.g., Bédard, 2006;Aigner-Torres et al, 2007;Tepley et al, 2010;Sun et al, 2017;Dygert et al, 2020), but data remain limited for minerals solely formed from silica-undersaturated magmas (e.g., leucite, melilite, nepheline). Published partition coefficients between feldspathoids and melt are based mainly on phenocryst-matrix observations (Onuma et al, 1981;Ewart and Griffin, 1994;Foley and Jenner, 2004;Arzamastsev et al, 2009;…”