Synthetic control over pore size and pore connectivity is the crowning achievement for porous metal-organic frameworks. The same level of control has not been achieved for molecular crystals, which are not defined by strong, directional intermolecular coordination bonds. Hence, molecular crystallization is inherently less controllable than framework crystallization, and there are fewer examples of 'reticular synthesis'-where multiple building blocks can be assembled according to a common assembly motif.Here, we apply a chiral recognition strategy to a new family of tubular covalent cages, to create both 1-D porous nanotubes and 3-D diamondoid pillared porous networks.The diamondoid networks are analogous to metal-organic frameworks prepared from tetrahedral metal nodes and linear, ditopic organic linkers. The crystal structures can be rationalized by computational lattice energy searches, which provide an in silico screening method to evaluate candidate molecular building blocks. These results are a blueprint for applying the 'node and strut' principles of reticular synthesis to molecular crystals.Despite many advances in supramolecular chemistry, it is still challenging to control molecular crystallization to create a specific, useful property. 1,2 This is important in the emerging area of porous molecular solids, 3 which have practical advantages such as solution processability. The crystal packing in porous molecular crystals defines the pore dimensions, which in turn define properties such as guest selectivity. 4,5 The same challenge-control over solid state structure-applies to all 2 functional molecular crystals because crystal packing defines physical properties such as electronic band gap and thermal or electrical conductivity.A central paradigm in crystal engineering is to synthesize building blocks, or 'tectons', with strong, directional interactions, such as hydrogen bonding 6 or metal-ligand binding, 7 which direct assembly into a targeted three-dimensional superstructure (Fig. 1). 1,2,8,9 For metal-organic frameworks (MOFs) and porous coordination polymers (PCPs), directional metal-ligand bonds are used to do this (Fig. 1a). [10][11][12][13][14] Likewise, hydrogen bonding can be used to create organic molecular crystals with defined network structures (Fig. 1b). 9,15,16 We have used chiral recognition to assemble porous organic cages 3 (POCs) into structures with 3-D pore channels (Fig. 1c). 3 POCs are rigid molecules with a permanent internal void that is accessible to guests via 'windows' in the cage. [17][18][19] Control of structure and function for POCs can be difficult, however, because slight changes in the molecular structure 19 or the crystallization solvent 20 can cause a profound change in the crystal packing. Chiral window-towindow interactions (Fig. 1e,f) can direct these POCs to assemble into 3-D pore networks in several cases, 19,21,22 but this is not ubiquitous. For example, some cages require specific solvents to template the window-to-window packing. 20 The chiral cage CC3-S (...