Both plants and animals rely on nucleotide-binding domain leucine-rich \nrepeat-containing (NLR) proteins to respond to invading pathogens and \nactivate immune responses. An emerging concept in NLR biology is that \n“sensor” NLR proteins are often paired with “helper” NLR proteins to mediate \nimmune signalling. However, the degree to which NLRs form signalling \nnetworks beyond sensor and helper pairs is poorly understood. In this thesis, \nI discovered that a large NLR immune signalling network with a complex \narchitecture mediates immunity to oomycetes, bacteria, viruses, nematodes, \nand insects. Helper NLRs in the NRC (NLR-required for cell death) family are \nfunctionally redundant but display distinct specificities towards diverse sensor \nNLRs. Several sensor NLRs, including Rx, Bs2 and Sw5b, signal via \ninterchangeable NRC2, NRC3 or NRC4, whereas some other sensor NLRs \nhave a more limited downstream spectrum. For example, Prf signals via \ninterchangeable NRC2 or NRC3 but not NRC4, and Rpi-blb2 signals via only \nNRC4. These helper/sensor NLRs form a unique phylogenetic superclade, \nwith the NRC clade sister to the sensor NLR clades. The network has \nemerged over 100 million years ago from an NLR pair that diversified into up \nto one half of the NLRs of asterids. I propose that this NLR network increases \nevolvability and robustness of immune signalling to counteract rapidly evolving \nplant pathogens.