The maintenance of biodiversity in intensively farmed landscapes in Australia relies mainly on small patches and linear strips of native vegetation - what's been left behind to delineate paddock and property boundaries, alongside roads and stock routes, where soils were unsuitable for farming, or wherever landholders wanted to retain trees. The agricultural matrix itself provides some habitat opportunities but these are limited, especially in areas where irrigation is used.This project set out to answer the simple question: what is more important for the conservation of insect and spider biodiversity in these landscapes -is it the amount of native vegetation, where that vegetation is, or the condition it is in? As with all ecological studies, the question is simple but the answers are complex. Taking three &quote;typical&quote; cotton landscapes that included either or both irrigated and dryland cropping systems, we sampled arthropods and analysed the species richness and functional diversity of the ants, and the morphospecies (identifiable as distinct species but not named) diversity of all other arthropod groups. We used 2 sampling techniques: pitfall trapping and suction sampling of arthropods from ground vegetation. We also recorded vegetation condition and local habitat variables, and calculated landscape metrics based on the proportions of major land use types within circular zones of 100, 500 and 1000m radius of sampling sites. On average there were no differences between landscapes in vegetation condition. However, condition was significantly higher in linear strips than in patches in the medium native vegetation landscape, with this trend being reversed at higher levels of native vegetation. This is almost certainly due to the impact of livestock grazing, which is more diffuse at higher levels of native vegetation, and more or less absent at lower levels.Different components of the arthropod fauna appear to respond in different ways to what happens in the landscape, so there is no &quote;one size fits all&quote; answer to the question posed above. Ant assemblages were different in each landscape, but did not respond to the type of habitat they were found in, e.g. discrete patches or linear strips of vegetation. Similarly, other ground active arthropod groups showedno habitat specificity, and differences werefound only between the verylow (Jandowae) and low (Broadwater) native vegetation landscapes. Arthropods in the ground vegetation did respond to the amount of native vegetation in the landscape, although these relationships were not statistically significant. They also responded to habitat type, but only at the high native vegetation landscape, Callandoon, where irrigation infrastructure (dams, channels etc.) is thought to be responsible through removal of marginal habitats.There was a clear pattern of change in the ants within crops during the growingseason. Some species persist in the paddock during fallow seasons, and this might be assisted by stubble retention and no---till practices.Strategic management of retained native vegetation in intensively farmed landscapes such as these, including narrow roadside strips, can provide essential habitat resources for a range of biota, including arthropods that deliver ecosystem services of benefit to both production and conservation.