Within-plant variation in glucosinolate concentrations of Raphanus sativus across multiple scales

Variation in chemical defenses remains underexplored. In particular, little is known about patterns of variation at small scales within leaves and spatial variation of induction. I examined variation in glucosinolate concentrations in the leaves of Raphanus sativus at several different spatial scales in two related experiments. I used samples equivalent in area to the amount an intermediate-sized caterpillar might eat in 1 d, a smaller scale than used in most previous studies. I examined variation due to induction and leaf age and small-scale spatial variation within leaves. The mean and variance of glucosinolate concentrations were higher in induced plants, young leaves, and the proximal half of leaves. Higher glucosinolate concentrations in the proximal half of leaves are previously unreported. Small-scale variation was extreme, accounting for 57% of the total random variation, and spatially random. There was no spatial autocorrelation found at scales as small as 1-2 cm. The high degree of small-scale, spatially random variation in glucosinolate concentrations in leaves is previously unreported. This small-scale variation and the variation caused by induction may have significant effects on herbivores and could be an important component of plant defense.