Given that many of these compounds are soluble in ethanol, which was the solvent used for their extraction in the present study, our measure of Artemisinic-acid antifungal activity probably includes the effect of several of these or related compounds. As a perspective of this work, as well as addressing to what extent they are synthesized by the host itself or by symbiotic microorganisms would give further insight into the study of antifungals in spiders. The synthesis of the abovementioned cuticular antifungals certainly requires resources that are obtained from the host diet, such as amino-acids and fa y acids, and thus we predicted that only animals in good physiological condition would be able to produce effective antifungals. However, we found no relationship between energetic body-condition and intensity of antifungal activity. There are two possible explanations for this finding: that these antifungals are not costly to produce and individuals in a range of conditions can maintain high levels, which seems unlikely given their chemical composition, or that cuticular antifungals are so important in infection avoidance that individuals cannot allow to reduce their production. Given that living solitary can be costly in terms of foraging efficiency and predation risk, the la er interpretation can also explain why experimentally isolated individuals lost energetic condition compared to individuals kept in a group in our laboratory experiment. Isolated individuals paid an energetic cost and not a cost in antifungal activity, probably because Anemarsaponin-BIII dietary restriction can be overcome whereas loss of antifungals is too risky. If the same resources are shared between cuticular antifungals and metabolic function, a trade-off between disease prevention and physiological condition may result. As we ignore the plasticity of up- and down regulation of cuticular antifungals, we cannot discard the possibility that differences in antifungal activity could be detected in a long-term experiment. We predicted that investment in antifungal protective activity would be higher in crowded nests, especially given the potentially higher risk of contagion that exists when there is high genetic relatedness among nest members, as in D. ergandros. However, we found no relationship between nest density and cuticular antifungal activity, neither in nature nor under laboratory conditions, suggesting that spiders constitutively express cuticular antifungal activity against the tested fungus. It is possible that cuticular antifungal activity, unlike immune response, is not a dynamic trait that can be regulated under varying risk of infection. Other unmeasured components of spider defense, such as haemolymph immune response or melanization, might be adjusted under different densities, as occurs in other arthropods, but this remains to be tested. Despite the benefits of group living in terms of foraging, reproduction or predator avoidance, our results show that living at high densities is costly in terms of reducing energetic reserves. However, we only found this pa ern under natural conditions, where food was not artificially supplemented and under natural predation risk.