Tetracycline is known to be effective in eliminating periodontopathogens and have collagenolytic activity. This study was performed to observe the desorption kinetics and structural changes of tetracycline-treated barrier membranes for guided tissue regeneration. Four kinds of barrier membranes were tested : $Tefgen^{(R)}$(American Custom Medical, USA) and $Gore-Tex^{(R)}$(W.L. Gore & Associates Inc., USA) as nonresorbable membranes ; Resolut(polyglycolide & polylactide copolymer, W.L. Gore & Associates Inc., USA) and $Biomend^{(R)}$(collagen, Collatec Co., USA) as resorbable membranes. The membranes were cut into discs(diameter : 4mm) and were immersed in 5% tridodecylmethylammonium chloride(TIMAC) ethanol and air-dried. The membrane discs were absorbed with $100{mu}g/ml tetracycline solution(pH8) for one minute and dried. For desorption kinetics, TC treated discs were immersed in phosphate buffered saline solution (PBS, pH 7.4). PBS was exchanged daily and TC concentration was measured by absorbance at 276nm on UV spectrophotometer. To measure remaining antibacterial activity, discs of 1 day to 4 weeks after desorption were placed on Mueller Hinton agar containing Bacillus cereus and incubated aerobically in $37^{circ}C$ for twelve hours and the inhibition diameters were measured. To observe the structural change of membranes after TIMAC treatment or immersion in PBS, the membrane discs were examined under SEM. The results were as follows : 1. Total amounts of TC absorbed into membrane discs($0.7536mm^2$) were $2000{mu}g$, $1800{mu}g$, $2625{mu}g$ and $2499{mu}g$ for $Tefgen^{(R)}$, $Gore-Tex^{(R)}$, $Biomend^{(R)}$ and $Resolut^{(R)}$. 2. The concentration of TC released from barrier membrane discs was maintained over $4{mu}g/ml$ until the fifth day in nonresorbable membranes and $Resolut^{(R)}$, but until the fourth day in $Biomend^{(R)}$, Until the ninth day in nonresorbable membranes and until the seventh day in resorbable membranes, the TC concentration was maintained over $1{mu}g/ml$. 3. The four membrane discs in the first day showed similar size of inhibition zone. One to four weeks later, the inhibition zone was much smaller in resorbable membrane discs than nonresorbable membrane discs. 4. Any structural change due to treatment of TIMAC was not observed on the nonresorbable membranes. $Resolut^{(R)}$ did not show any structural change except fibrillar loosening during immersion period, but Biomend showed destruction of membrane structure from the first week of immersion. This study indicates that tetracycline treated barrier membranes lead to the sustained release of tetracycline for over 7 days. This slow release pattern of tetracycline may contribute to the favorable clinical outcome of guided tissue regeneration.