The high intensity of free-electron lasers now allows for the possibility of obtaining measurable diffraction from biological samples with a single X-ray pulse. An important consequence of diffract-before-destroy imaging is that the sample is destroyed and therefore must be replaced preferably at the repetition rate of the FEL. This presents an interesting challenge; the sample must be rapidly replaced within the X-ray focus at the proper particle density and degree of hydration without damaging or denaturing the sample. If particle number density is too high, for example due to clustering or evaporation, the diffraction pattern resulting from coherent illumination of multiple particles may be discarded when sorting for 3D reconstruction. If number density is too low the hit rate, percentage of pulses with measurable scattered intensity, may also be too low to collect a complete data set. Evaporation will also leave behind less volatile material and this change of concentration may be damaging to the sample. On the other hand the similarity in electron density for water and biological material provides poor contrast for fully hydrated material. It is often also necessary to consider sample consumption. While high, near unity, hit rate can be obtained using liquid jets, a liquid flow rate greater then 1 microliter per minute must be maintained. Several sample injection possibilities, drop on demand, aerosols, liquid jets, aerodymamic lenses, have been explored and a review of these results is presented.