Agricultural landscape homogenization has detrimental effects on biodiversity and key ecosystem services. Increasing agricultural landscape heterogeneity by increasing seminatural cover can help to mitigate biodiversity loss. However, the amount of seminatural cover is generally low and difficult to increase in many intensively managed agricultural landscapes. We hypothesized that increasing the heterogeneity of the crop mosaic itself (hereafter “crop heterogeneity”) can also have positive effects on biodiversity. In 8 contrasting regions of Europe and North America, we selected 435 landscapes along independent gradients of crop diversity and mean field size. Within each landscape, we selected 3 sampling sites in 1, 2, or 3 crop types. We sampled 7 taxa (plants, bees, butterflies, hoverflies, carabids, spiders, and birds) and calculated a synthetic index of multitrophic diversity at the landscape level. Increasing crop heterogeneity was more beneficial for multitrophic diversity than increasing seminatural cover. For instance, the effect of decreasing mean field size from 5 to 2.8 ha was as strong as the effect of increasing seminatural cover from 0.5 to 11%. Decreasing mean field size benefited multitrophic diversity even in the absence of seminatural vegetation between fields. Increasing the number of crop types sampled had a positive effect on landscape-level multitrophic diversity. However, the effect of increasing crop diversity in the landscape surrounding fields sampled depended on the amount of seminatural cover. Our study provides large-scale, multitrophic, cross-regional evidence that increasing crop heterogeneity can be an effective way to increase biodiversity in agricultural landscapes without taking land out of agricultural production.

Field experiments to test the hypothesis that a cannibalistic species is territorial are rare. We conducted two field experiments to test the hypothesis that adult females of the Mediterranean tarantula, Lycosa tarentula (L.) are territorial. In a relatively longterm experiment we placed a female intruder in an artificial burrow within the hypothesized territory of a resident. The intruders disappeared from their burrows at 7ϫ the rate of spiders in a reference and two control treatments. Residency status, not relative size, determined whether the intruder or the resident remained, and evidence suggests that the winner frequently cannibalized the loser. We also conducted a short-term field experiment in which we induced encounters between females. The results were consistent with territorial defense because escalation was more likely if spiders were similar in size, and cannibalism, the outcome of one-third of the encounters, occurred only after escalation. Thus, adult females of the Mediterranean tarantula are territorial, and cannibalism may be a consequence of territorial defense.

We conducted a field experiment to test for food limitation in immature stages, and its consequences for mature females, in the territorial, cannibalistic spider Lycosa tarentula (L.). Randomly selected antepenultimate juveniles were provided supplemental prey until they matured, at which time supplemental feeding ceased. Immature stages of L. tarentula are food‐limited. Supplemented juvenile spiders decreased foraging activity, disappeared at a lower rate and grew faster than the control spiders, which had been exposed only to ambient prey levels. Fed juvenile females were less hungry at maturity, as judged by an index of body condition, and showed higher mating success as adults, as judged by cohabitation rates with mature males. Foraging theory predicts that in order to compensate for residual effects of food limitation, adult female spiders that had experienced a shortage of prey as juveniles – the controls – would have to exhibit a greater increase in foraging activity upon maturing than the prey‐supplemented group. Contrary to expectation, the control females did not increase their foraging activity, but the previously fed females did forage more actively as adults. Furthermore, the difference in mass gain during the mating period between the two groups was opposite from what the difference in change in foraging activity would predict. Control females, the spiders that had not changed their foraging activity, gained mass more rapidly than the previously fed females, with the result that the two groups were similar in mass by the end of the mating period. We hypothesize that an increased rate of sexual cannibalism may have been one mechanism by which control females compensated for the food limitation that they had experienced as immatures.

1. Predation-exclusion experiments have highlighted that top-down control is pervasive in terrestrial communities, but most of these experiments are simplistic in that they only excluded a single group of predators and the effect of removal was evaluated on a few species from the community. The main goal of our study was to experimentally establish the relative effects of ants and birds on the same arthropod assemblage of canopy trees.2. We conducted 1-year long manipulative experiments in an organic citrus grove intended to quantify the independent effects of bird and ant predators on the abundance of arthropods. Birds were excluded with plastic nets whereas ants were excluded with sticky barriers on the trunks. The sticky barrier also excluded other ground dwelling insects, like the European earwig Forficula auricularia L.3. Both the exclusion of ants and birds affected the arthropod community of the citrus canopies, but the exclusion of ants was far more important than the exclusion of birds. Indeed, almost all groups of arthropods had higher abundance in ant-excluded than in control trees, whereas only dermapterans were more abundant in bird-excluded than in control trees. A more detailed analysis conducted on spiders also showed that the effect of ant exclusion was limited to a few families rather than being widespread over the entire diverse spectrum of spiders.4. Our results suggest that the relative importance of vertebrate and invertebrate predators in regulating arthropod populations largely depends on the nature of the predator-prey system.

Ants and spiders are ubiquitous generalist predators that exert top-down control on herbivore populations. Research shows that intraguild interactions between ants and spiders can negatively affect spider populations, but there is a lack of long-term research documenting the strength of such interactions and the potentially different effects of ants on the diverse array of species in a spider assemblage. Similarly, the suitability of family-level surrogates for finding patterns revealed by species-level data (taxonomic sufficiency) has almost never been tested in spider assemblages. We present a long-term study in which we tested the impact of ants on the spider assemblage of a Mediterranean citrus grove by performing sequential 1-year experimental exclusions on tree canopies for 8 years. We found that ants had a widespread influence on the spider assemblage, although the effect was only evident in the last 5 years of the study. During those years, ants negatively affected many spiders, and effects were especially strong for sedentary spiders. Analyses at the family level also detected assemblage differences between treatments, but they concealed the different responses to ant exclusion shown by some related spider species. Our findings show that the effects of experimental manipulations in ecology can vary greatly over time and highlight the need for long-term studies to document species interactions.