It is well-‐known that we cannot process and consciously perceive all information present in a complex visual scene. Attention selects information for further in-‐depth processing and gates access to awareness. Hence, it is important to understand what factors control and guide visual attention, and especially, how we select visual information based on our top-‐down goals. Classically, top-‐down attentional guidance has been thought to operate on specific feature values of a sought-‐after object. However, a more recent relational account has proposed that top-‐down attention is guided by the relative differences between the features of relevant and irrelevant items in the visual field. Research has found that relational attentional guidance seems to be the preferred search mode compared to feature-‐specific search modes. However, no study to date has attempted to explore whether these different search modes can be adopted simultaneously to select a conjunction stimulus that differs in two features from irrelevant items. This study used a variant of the spatial cueing paradigm to test whether attention was biased to the specific colour and/or size of a conjunction target, or to its relative colour and/or size. The results of a first experiment showed that attention is preferentially biased to the relative target colour and size in search for a conjunction stimulus. Experiments 2 and 3 demonstrated that observers have the ability to simultaneously adopt a feature-‐specific search mode along the target feature dimension of size, and a relational search mode along the target feature dimension of colour (Experiment 2), and vice versa (Experiment 3). Moreover, a control experiment (Experiment 4) showed that observers are able to adopt an exclusively feature-‐specific search mode along both target feature dimensions of colour and size. The implications of these findings for current theories of attention are discussed.