Excised embryonic axes from seeds of three taxa, namely, Citrus suhuiensis cv. limau madu, Citrumelo (Citrus paradisi x Poncirus trifoliate) and Fortunella polyandra, were desiccated in a laminar airflow, over silica gel, and ultra-rapidly. Desiccation sensitivity (WC50) was estimated for each taxon using the quantal response model. High desiccation tolerance (WC50 = 0.11 g water per g dry mass. g/gdw) was observed for limau madu embryonic axes desiccated in a laminar airflow and ultra-rapidly (WC50 =0.10 g/gdw). Desiccation tolerance was substantially lower (WC50 = 0.19 g/gdw) for silica gel dehydration. Similarly, high desiccation tolerance (WC50 = 0.15 g/gdw) was associated with F. polyandra embryonic axes when desiccated in a laminar airflow, while a lower desiccation tolerance (WC50 = 0.17 g/gdw) was observed with silica gel dehydration. Ultra-rapid desiccation led to the highest desiccation tolerance (WC50 = 0.14 g/gdw). The dehydration rate, however, had no influence on desiccation tolerance (WC50 ~ 0.14 g/gdw) for Citrumelo embryonic axes. After each desiccation period, embryonic axes were directly immersed in liquid nitrogen (LN) followed by rapid rewarming. Normal seedling recovery of 80 to 83% for excised embryonic axes of limau madu was observed for laminar airflow and ultra-rapid dehydration, but for silica gel dehydration, 57% recovery was obtained. Similarly, for Citrumelo, high recoveries of 100% and 97% were obtained from axes desiccated in a laminar airflow and using ultra-rapid dehydration, respectively, whereas a lower value was associated with silica gel dehydration (80%). For F. polyandra, 50% recovery was obtained both for laminar airflow and ultra-rapid dehydration, while much lower recovery (43%) was associated with silica gel dehydration. Regardless of the drying method employed, axis survival percentages following exposure to LN were commensurate with the desiccation sensitivity pattern.