Since no Canadian evidence exists on Canadian IPO lockups, this thesis examines the design, impact and market behavior of two types of lockups for Canadian IPOs listed on the TSX during the 1997-2005 period. We find that the existence of lockup information in an issuer's prospectus does not significantly reduce IPO underpricing by underwriters. We also find that IPO firms with dual-class share structures tend to have higher proportions of shares locked up and longer escrow lockup periods, and that larger firms have higher proportions of shares locked up as "escrow shares". These findings partially support the evidence reported by Brau et al. (2005) that firms with less transparency will provide a signal that the interests of their insiders are better aligned with outside investors by including lockups in their prospectuses. Significant negative (cumulative) abnormal returns are found only for high-tech firms immediately around unlock days. Lower abnormal trading volumes and relative spreads are found after unlock days for only the sample of IPOs with escrow lockups with stipulated non-zero lockup length.  Keywords: IPO, lockup, escrow, abnormal market behavior.  JEL Classification: G10, G15