Band spreading originating in open sampling tubes is presented. This phenomenon is emphasized particularly with the injection of large sample volumes. Whether the sampling loop is coiled or straight, the variance values an be from 1.3 to 1.8 times, larger than the minimal value which is equal to Vi2/12 (theoretical value for an njected volume, Vi). Two methods are discussed that decrease this additional band spreading due to open loops. The first method keeps the sample loop on-line with the mobile phase only for the time, ti = Vi/Q where Q is the flow rate). In this way, it is necessary to choose a sample loop volume at least twice as large as the volume to be injected. The second method packs sample loops with glass beads. Diagrams are given that ilustrate the choice of glass bead sizes and tube inner diameters according to the injected volume, the additional and broadening, and technological constraints. Sample volume from 0.5 up to 100 cm3 can be easily infected with resulting variance to 1.01 times the minimum variance equal Vi2/12.