“…Lack of conservatism in chemical defense could result from multiple nonexclusive sources of variation, including abiotic factors or inducibility (Kim & Jander, ; Himanen et al ., ), or conflicting selection for different compounds imposed by specialist/generalist herbivores or conspecific/heterospecific competitors (Kliebenstein et al ., ; Lankau & Strauss, ; Müller, ) across habitats. That said, studies using common‐garden approaches also find low to moderate phylogenetic signal in secondary compounds (Agrawal et al ., ; Johnson et al ., ; Karinho‐Betancourt et al ., ), and for a subset of our data ( n = 15) we found a significant, moderate correlation between total glucosinolates measured in the field and in the glasshouse. Oenothera 's flavonoids (Johnson et al ., ), Quercus ’ tannins (Pearse & Hipp, ), Inga ’ s phenolics (Kursar et al ., ), Datura 's alkaloids (Karinho‐Betancourt et al ., ), and Bursera 's terpenes (Becerra et al ., ) also have weak phylogenetic signal.…”