In this pape r e int r duce the categor214 (a.k.a. chr omatic) r) ge quer2L (CRQs) in the context oflar1P disk-r esident data sets, motivated by the fact that CRQsar conceptually simple and emer4 often in DBMSs. On the basis of spatial data str uctur3N and R-tr4L in par ticular e p r pose a multi-tr e index that follo s the br ad concept of augmenting nodes ith additional infor ation to accelerw e quer31P Augmentation is examined ithr espect to maximal/minimal points in subtr15L the pr oper23P of hichar exploited by the pr3 osed sear hing algor6 hm to e#ectively pr une the sear h space. Detailed exper21z - talr esults, ith bothr eal and synthetic data, illust rw e the significant per1z2 ance gains (up to an orer of magnitude) due to the pr4 osed method, compar2 to ther2P lar r ange quer (follo ed by the filter4 g .rN . categor654 and to a naive R-tr ee augmentation method. 1