Objective: To observe the effects of surgical delay procedure on the survival of perforator flap with three angiosomes in rat, and to explore its possible mechanism., Methods: The flap model was a perforator flap with three angiosomes which located on the right dorsal side of a rat based on the right deep circumflex iliac vessel. The two connection areas between the three angiosomes were successively named choke zone (CZ) 1 and CZ 2 beginning from the pedicle to the remote area. A total of 110 SD rats were divided into routine flap group (RF, n = 40), delay only group (DO, n = 30), and delay flap group (DF, n =40) according to the random number table. (1) In group RF, 30 rats were selected according to the random number table, and flap surgery was performed directly. Six rats were sacrificed on post operation day (POD) 0, 1, 2, 3, 7 respectively to collect the full-thickness skin samples at both CZs for HE staining to measure the vascular density and diameter. The rest 10 rats underwent flap surgery immediately after a catheter was successfully implanted into their external jugular vein. A volume of 1.5 mL sodium fluorescein solution (100 g/L) was injected to the 10 rats on POD 0 (5 rats) or POD 1 (5 rats) each time with a 2-day interval to learn the change in flap circulation. Each rat was injected for 4 times. The flap survival rate of the 10 rats was calculated on POD 7, and the configuration and distribution of the vessels in the flap were observed through angiography with the improved perfusion method of lead oxide-gelatin. (2) In group DO, the right thoracodorsal perforators of all the rats were surgically ligated through a small skin incision, and 6 rats were sacrificed on POD 0, 1, 2, 3, 7 respectively. The skin samples of each rat at the same area as in group RF were harvested to measure the vascular density and diameter. (3) In group DF, rats were treated with ligation surgery as in group DO, and then they were assigned and treated as in group RF on POD 7 with corresponding indexes detected later. Data were processed with group t test, analysis of variance with factorial design, and SNK test., Results: (1) Significant differences of vascular density at both CZ 1 and CZ 2 were found on POD 7 among the three groups ( with F values respectively 2. 69 and 2. 76, P values below 0.05). The vascular density values of CZ 1 and CZ 2 of rats in group DF were (29 ± 7) and (31 ± 8) per mm on POD 7, which were significantly higher than those of group RF [(23 ± 5) and (23 ± 3) per mm2, with q values respectively 5.67 and 6.01, P values below 0.05] and those within group DF on POD 0 (with q values respectively 6.42 and 7. 14, P values below 0. 05). On POD 3 and 7, the vascular diameter values of CZ 1 of rats in groups RF and DF were significantly higher than those of group DO (with q values from 8. 15 to 11.13, P values below 0.05). The vascular diameter values of CZ 2 of rats in group DF onPOD 0, 1, 2, 3,7 [(65 ± 8), (63 ± 13), (69 ± 9), (67 ± 8), (64 ± 13) 230m] and in group DO on POD 3 and 7 were significantly higher than those in group RF [respectively (46 ± 10) , (40 ± 9), (43 ± 13), (46 ± 12), (47 ± 11) µm on POD 0, 1,2, 3, 7 ] at corresponding time point (withqval- ues from 7.29 to 10.79, P values below 0.05). The difference in vascular diameter between CZ 1 and CZ 2 was statistically significant in groups RF and DO on POD 3 and 7, and in group DF on POD 0, 1 , and 2 (with q values from 5.32 to 9.56, P values below 0.05). Compared with that on POD 0 within each group, the vascular diameter of CZ 1 in groups RF and DF and that of CZ 2 in group DO increased significantly on POD 3 or 7 (with q values from 6.12 to 8.13, P values below 0.05). (2) In groups DF and RF, blood from the pedicle ran through CZ 1 and covered the dynamic territory successfully within POD 7. On POD 0, the blood within all flaps was blocked for about 3 min after going through CZ 1 at 1 cm distal from CZ 2 in group DF and around CZ 2 in group RF. (3) Flap survival rate of rats in group DF was (95 ± 12) % , which was statistically higher than that of group RF [(80 241 9) % , t = 2.91, P <0.01]. All the partial flap necrosis occurred in potential territory. (4) Compared with the vessels in the left dorsal side without surgery, the vessels of CZ 1 in group RF were dilated obviously, and the boundary between vascular trees became indistinct, but the vessels in CZ 2 changed slightly; the vessels in both CZs in group DF were dilated dramatically., Conclusions: The delay method could enhance the survival of potential territory in perforator flap with three angiosomes, and it acted mainly by dilating the choke vessels in CZ 2 before flap surgery.