A successful mitosis and cytokinesis requires the formation of a strictly bipolar mitotic spindle to ensures equal segregation of sister chromatids to the daughter cells and proper formation of the cleavage furrow between segregating chromatids. Spindle assembly requires a coordinated effort within the cell to completely rearrange the microtubule network from an organized interphase array to the bipolar mitotic spindle. Spindle assembly factors as well as microtubule dynamics are crucial to this process. Drugs, such as taxol and vinblastine, have been used to disrupt microtubule dynamics. Taxol specifically stabilizes plus end dynamics resulting in multipolar spindles, suggesting that the formation of stable microtubules results in spindle assembly failure. However when added to cells during anaphase, the bipolar spindle remains bipolar. This suggests that multipolar spindles formed in taxol cells are not solely due to microtubule stabilization but result from morphological changes to the microtubule cytoskeleton that occur during the G2/M transition. Here, we show that, upon treatment with taxol, microtubules redistribute to the cell cortex, where asters form. These asters eventually converge to form a multipolar spindle. Disruption of dynein does not interfere with the process, however, spindle assembly factors such as NuMA and HSET are involved in organizing the microtubule asters. Another microtubule stabilizing drug, vinblastine was also analyzed. We show that stabilization by vinblastine creates multipolar spindles by a different mechanism than taxol. Instead of a dramatic rearrangement of the microtubule network, we see multiple microtubule asters form around the chromatin. We find that the centrosomes continue to organize a subset of the microtubule population; however the remaining microtubules form asters independently. There appears to be both centrosome mediated and chromatin mediated spindle assembly occurring. While it is traditionally thought that the centrosomes organize the bipolar mitotic spindle, recent work has suggested that this is not always the case. Venous egg extracts and laser ablation of centrosomes have demonstrated that bipolar spindles can form in the absence of centrosomes. The model for this is a chromatin mediated pathway that organizes randomly oriented microtubules. However, somatic animal cells do not normally enter mitosis with their microtubule network unorganized. Here, we have addressed these inconsistencies in a mammalian cell system, African Green monkey kidney cells or BSC-1 cells. When centrosomes are surgically removed from the cell, we find that while bipolar spindles can form in the absence of centrosomes, mistakes happen a significant amount of times. This suggests that centrosomes are indeed required to provide a strong bipolar cue to ensure that bipolar spindle assembly occurs without errors.