We recorded for the first time the presence of the Hymenoepimecis genus in Colombia. Previously Pádua et al. (2015; 2020)d dua (2022) studied individuals of this genus were found in several countries of Central and South America, except in Colombia. A female of H. castilloi was previously found in a malaise trap in the Peruvian Andes (Pádua et al. 2020), but its hosts were unknown and now was confirmed as Leucauge mariana. Since Colombia is the biological corridor between Central and South America, many species must be present in this territory.
Eberhard (2013) found that cocoon webs of L. mariana parasitized by Eruga ca. gutfreundi (Hymenoptera: Ichneumonidae) differ from those parasitized by H. tedfordi. The main differences were that Eruga ca. gutfreundi cocoon webs showed a 3D structure with multiple radial lines radiating in many directions that converged in the cocoon. While, H. tedfordi cocoon webs were planar with strengthened radial and frame lines, and occasionally a sparse tangle of lines below. In this case, H. castilloi cocoon webs are similar to those of H. tedfordi, finding horizontal planar webs with reinforced radii that converged to a hub, and the cocoon hanging on a vertical line, with occasionally a sparse tangle below. However, the two cocoon webs reported away from aggregates for H. castilloi were more similar to the typical cocoon web design of L. argyra Walckenaer parasitized by H. argyraphaga Gauld (Eberhard 2001) and to those highly reduced cocoon webs of H. tedfordi (Eberhard 2013) with only a few reinforced radii converging to the hub. Also, the complete lack of adhesive spirals and the increase of lines connected to vegetation in order to reinforce the structure observed in the cocoon webs of L. mariana was observed in all interactions between Leucauge and Hymenoepimecis wasps (Eberhard 2001; Sobczak et al. 2019, Pádua et al. 2016; Santos-Murgas et al. 2022; Gonzaga et al. 2015; Kloss et al. 2022). We suggest that these features of cocoon webs may be a modification of the web that promotes stability over prey interception and climate conditions, as previously observed in cocoon webs built by Agelena limbate Thorell (Araneae: Agelenidae) (Matsumoto 2009) and by spiders of genus Cyclosa (Araneae: Araneidae) (Kloss et al. 2016).
We found variation among H. castilloi cocoon webs. When parasitism occurs in web aggregates, the modified web was a vertical line that connected the different webs or an irregular and dense web with a sparse tangle below the hub. However, isolated cocoon web was formed by 3 to 4 reinforced radii converging to a hub, similar to that observed in L. argyra (Eberhard 2001), L. roseosignata Mello-Leitão (Sobzack et al. 2009), L. henryi Mello-Leitão (Pádua et al. 2016)d venusta Walckenaer (Santos-Murgas et al. 2022), but different from the design of the cocoon webs in L. volupis which presented a complex tangle below the hub in all cocoon webs (Gonzaga et al. 2015, Kloss et al. 2022). In isolated cocoon webs and in aggregates of L. mariana webs, the parasitoid cocoon was always hanging on a vertical line and the web had lines connected to the vegetation at multiple lines. Leucauge individuals sometimes cooperate and can group under specific conditions forming colonies (Salomon et al. 2010). These colonies consist of individual orb-webs connected by a shared framework silk that are joined to the vegetation with thick silk threads, in these structures we found the cocoons hanging on a vertical line. Probably this structure is stable enough that do not need reinforcement making less evident the cocoon web modification, while in isolation there is a clear pattern of a modified web with reinforced silk. However, further investigations are needed to identify the punctual web modifications of cocoon webs in aggregates.
Korenko et al. (2022) studied the two lines of defense of the parasitoids to avoid predation in their most vulnerable stage (pupae). The cocoon spun by the parasitoid larva itself (1) and a cocoon web provided by the spider host (2) due to larvae manipulation. In this way, hanging in aggregates could be a good option for the parasitoid to keep away from the danger of being eaten by some guilds like scavengers that showed capability of eating the pupae when it is exposed to the soil. Nevertheless, there are quite a few hyperparasitoids species that can parasitize on pupae that are in aggregates of another Hymenoepimecis species: Trichonephila clavipes Linnaeus (Araneae: Araneidae) is an orb weaver that can form web aggregates and this spider has been parasitized by H. bicolor. However, Pádua et al. (2022) registered hyperparasitism on H. bicolor by wasps Lymeon sp. (Ichneumonidae: Cryptinae), Neotheronia lineata Fabricius, and N. donovani Gauld (Ichneumonidae: Pimplinae), suggesting that aggregation is not a warranty to avoid parasitism. Also, there is a predation risk by pupae provided by kleptoparasitic spiders, as previously observed in cocoon of parasitoid Polysphincta nr. purcelli Gauld (Kloss et al. 2016), leaving this strategy of parasitizing in aggregates to be negative for parasitoids and positive for spiders by reducing the risk of parasitism.
According to Kloss et al. (2016), one of the most important reasons for decreasing pupae survivorship is that the rain and any falling object that can break the web structure and make the pupae fall to the ground where they are at higher risk of predation. On the other hand, in aggregates, the pupae would stay surrounded by other webs, giving it more structural support than a single cocoon web. Also, there is an intraspecific variation in the cocoon web of Leucauge mariana that can be found with modification when it is isolated and without a pattern of modified web when is in a forming web aggregate. Following these findings, there is not only a difference in cocoon webs between different Hymenoepimecis-Leucauge species interactions but also in different population conditions. However, it is important to analyze parasitoidism in dense spider aggregation after the fact that there is no construction of a modified web.
Eberhard and Gonzaga (2019) proposed that, since the mechanism of manipulation was increasing the ecdysone levels on the spider, the modifications in the same spider-host will be similar despite of the parasitoid species. In this case, we noticed that H. castilloi and H. tedfordi manipulated L. mariana to build similar cocoon webs when they were isolated. On the other hand, L. mariana built a different type of cocoon web under the influence of E. ca. gutfreundi, this fact can be attributed to the behavior of the individuals due to altitudinal differences. Since this is the first report of this kind of parasitism at high altitudes there are no studies to compare. However, we also found that different conditions in the spider population (aggregates and isolated) can change the build pattern of the cocoon web. Further studies using phylogenetic, morphological and ecological analysis are needed to understand the evolution of behavioral manipulation from the wasp.