Many geophytes have a high ornamental value. They are preferably propagated by micropropagation because in this way large quantities of uniform, disease-free starting material are produced in a short period of time. In comparison with shoots, bulblets have several clear advantages as starting material. Therefore, in vitro bulblet formation is an important target for improvement of tissue culture of geophytes. The research described in this thesis was carried out with the lily cultivars ‘Santander’ and ‘Stargazer’. Commercially, lily is the second geophyte in the global flower industry. The size of lily bulblets regenerated in vitro has a direct effect on the performance in the field. After planting, large bulblets sprout with a stem and gain twice as much weight compared with small bulblets that sprout with a rosette. In the present study we studied basic and applied aspects of the following topics: (1) new methods for sterilization during initiation, (2) the effect of scale-related factors on bulblet growth, (3) growth enhancement by moderate abiotic stresses, and (4) the effect of CO2 removal from headspace of tissue culture containers on lily bulblet growth and as a control Arabidopsis thaliana seedling growth. Contamination is an everlasting problem in tissue culture laboratories. Lily bulbs are underground organs and contain therefore more contaminants as compared with aerial organs. During initiation, operators cause additional contamination in two ways that have as yet not been recognized adequately. (1) Rinsing explants with sterile water after surface-sterilization is the generally advised method to remove the residues of decontaminants. However, when scales are heavily contaminated, the surface-sterilization does not kill microorganisms in all scales. Surface-sterilization is usually done with batches of 10 to 30 scales and the contaminated scales may cross-contaminate uninfected scales during rinsing in water. We have tested the rinsing water and found heavy bacterial contamination in the 2th and especially the 3rd rinse. The contaminated rinsing water resulted in a high incidence of cross-contamination. Cross contamination was reduced almost fully by rinsing in diluted NaClO solution (0.03%) instead of sterile water. There was no negative effect of diluted NaClO on growth. (2) A second way of introducing contamination by the operator is the entering of microorganisms during detachment from mother bulbs via the vascular bundles caused by negative hydrostatic pressure within the bulb tissue. By detaching scales from bulbs submerged in 0.03% NaClO, hydrostatic-pressure related contamination was strongly reduced. The growth of bulblets increased by 22% and 17% when cross contamination and negative-hydrostatic pressure related contamination were prevented. Storage organ formation is controlled by interacting environmental, biochemical and genetic factors. We studied various aspects of the effect of the scale explant. We found that large explants produced larger bulblets than small explants. When bulblets were excised and cultured in vitro, growth was improved by 33% when a small piece of the original scale explant was left attached to the bulblet. The position in the scale from where the explant was excised affected the growth of the regenerating bulblets. Basal-scale explants improved bulblet growth by 40-50 % compared with apical-scale explants. This might be related to the physiological state of the tissues: there was more starch and there were more vascular bundles present in basal scale explants. Furthermore, excision of an explant from the middle of the scale improved bulblet growth by 40-50 % compared with explants excised from the edge of the scale. In general, the middle scale explants were heavier and contained wider vascular bundles. Plants in stressful conditions tend to allocate a higher proportion of biomass to below-ground biomass (roots and storage organs) as compared to above ground biomass. We investigated the effect of moderate abiotic stresses on lily bulblets grown in vitro. In general, lily bulblets showed an increased growth after moderate stresses. Hot air increased growth by 30%, hot water by 40%. We also examined the effect of drought and anaerobiosis. Drought stress increased growth of bulblets by 40% in the cultivar ‘Stargazer’, but significantly decreased bulblet growth in cv ‘Santander’. Anaerobiosis increased growth in ‘Stargazer’ and ‘Santander’ by 32% and 65%, respectively. We also showed that a moderate stresses treatment protects lily bulblets against future severe abiotic stresses. In general, the in vitro situation is not favorable for plants. Composition of the headspace (high humidity, strongly fluctuating CO2- and O2-levels and accumulation of gases like ethylene), low light, and wounding are unfavorable conditions that plants have to deal with. We examined the effect of CO2 starvation on growth in vitro with and without addition of 3% sucrose to the medium. A CO2-poor headspace reduced the growth of bulblets, leaves, roots and scale explants strongly, also in the presence of 3% sucrose. CO2 removal from the headspace decreased growth of Arabidopsis seedlings by 50 % on medium with 3% sucrose. It seems unlikely that the growth reduction on medium with 3% sucrose is caused solely by the lack of sucrose production in photosynthesis when CO2 is removed. Indeed, we found evidence that the low CO2 resulted in heavy stress that in turn may reduce growth. Fv/Fm in lily and Arabidopsis dropped when CO2 was removed. Occurrence of reactive oxygen radicals (ROS) was examined in Arabidopsis seedlings by staining with nitroblue tetrazolium (NBT). ROS was virtually absent in ex vitro growing seedlings and very abundant in seedlings grown under CO2 starvation. Seedlings grown under normal tissue culture conditions showed an intermediate presence of ROS. We hypothesize that low levels of CO2 may results in ROS in in vitro seedlings which reduces growth.