Analysis of deviance table testing the effects of rearing density (1, 2, 4 or 8 larvae/pot), larval phenotype (green, intermediate or black) and weight (used as a covariate) on (a) encapsulation, (b) capsule melanization and (c) hemocyte densities of <i>Anticarsia gemmatalis</i> larvae.

<p>To verify if the overall immunological responses were a function of density <i>per se</i>, or of contact among larvae, we analyzed the density either as a continuous or categorical explanatory variable (contrasting the factor levels). The factor levels were amalgamated if they were not significantly different. (<i>*P</i><0.05, <i>**P</i><0.01,<i>***P</i><0.001, <i>n.s.</i> = not significant).