This paper presents a method for passively breaking droplets into multiple sizes with precise control over the number and size of each droplet. Two techniques are described: (i) breakup at a T junction and (ii) flow past isolated obstacles. The T junction method allows for the generation of daughter droplets of a predetermined size ratio, with the ability to repeat the process without increasing droplet polydispersity until drop radii are on the order of the channel width. Smaller droplets can be produced at higher flow rates. An analytical model is presented to understand the transition from stretching to non-breaking of drops in extensional flow. For situations where minimal device real estate is important, an obstacle-based method is used to break a predetermined fraction of droplets. The study demonstrates the use of microfluidic technology to generate highly uniform emulsion droplets. Water droplets are dispersed in a continuous phase of hexadecane, with a surfactant added to stabilize the droplets. The droplets are generated at a T junction, where the continuous phase shears off droplets of distilled water, creating an inverse emulsion with a narrow distribution in droplet size. The droplet size and distribution can be tailored by exploiting extensional flow near the stagnation point of the T junction. The ratio of the flow rates in the side arms is inversely proportional to the ratio of the lengths of the side arms, allowing for asymmetric breakup of droplets. The critical conditions for droplet breakup at a T junction are determined by examining both non-breaking and breaking conditions. The critical capillary number for breaking is found to depend on the initial extension of the droplets. The study also shows that asymmetric configurations may require adjustments to the Rayleigh-Plateau stability criterion. The T junction method is more reliable than the single obstruction design, as it allows for active control of drop size distribution. The obstacle-based method is more efficient in terms of device real estate but is less controllable. The study demonstrates the ability to design dispersions with controlled volume fractions and drop sizes using geometrically mediated drop breakup.This paper presents a method for passively breaking droplets into multiple sizes with precise control over the number and size of each droplet. Two techniques are described: (i) breakup at a T junction and (ii) flow past isolated obstacles. The T junction method allows for the generation of daughter droplets of a predetermined size ratio, with the ability to repeat the process without increasing droplet polydispersity until drop radii are on the order of the channel width. Smaller droplets can be produced at higher flow rates. An analytical model is presented to understand the transition from stretching to non-breaking of drops in extensional flow. For situations where minimal device real estate is important, an obstacle-based method is used to break a predetermined fraction of droplets. The study demonstrates the use of microfluidic technology to generate highly uniform emulsion droplets. Water droplets are dispersed in a continuous phase of hexadecane, with a surfactant added to stabilize the droplets. The droplets are generated at a T junction, where the continuous phase shears off droplets of distilled water, creating an inverse emulsion with a narrow distribution in droplet size. The droplet size and distribution can be tailored by exploiting extensional flow near the stagnation point of the T junction. The ratio of the flow rates in the side arms is inversely proportional to the ratio of the lengths of the side arms, allowing for asymmetric breakup of droplets. The critical conditions for droplet breakup at a T junction are determined by examining both non-breaking and breaking conditions. The critical capillary number for breaking is found to depend on the initial extension of the droplets. The study also shows that asymmetric configurations may require adjustments to the Rayleigh-Plateau stability criterion. The T junction method is more reliable than the single obstruction design, as it allows for active control of drop size distribution. The obstacle-based method is more efficient in terms of device real estate but is less controllable. The study demonstrates the ability to design dispersions with controlled volume fractions and drop sizes using geometrically mediated drop breakup.