Specific defense protects against some parasite genotypes but not others, whereas non-specific defense is effective against all genotypes of a parasite. Some empirical studies observe hosts with variability only in non-specific defense, other studies find only specific defense. I analyse a model with combined specific and non-specific defense to determine the conditions that favor detectable variation in each form of defense. High variation in non-specific defense is often maintained when resistance increases in an accelerating way with investment, whereas low variation tends to occur when resistance increases at a decelerating rate with investment. Variation in specific defense rises as the parasite pays a higher cost to attack a broad host range (high cost of virulence), as the number of alternative specificities declines, and as the average level of non-specific defense increases. The last condition occurs because greater non-specific protection tends to stabilize the gene frequency dynamics of specific defense. Selection favors a negative association between costly components of specific and non-specific defense-hosts defended by one component are favored if they have reduced allocation to other costly components. A negative association confounds the measurement of costs of resistance. Individuals with specific defense may have reduced investment in costly non-specific defense. This leads to an apparent advantage of specifically defended hosts in the absence of parasites and a measured cost of resistance that is negative.