Upsala J Med Sci 79: 28-38, 1974 Haemostatic Plug Formation in the Rabbit Mesentery A Methodological Study D. BERGQVIST Department of Experimental Medicine, Pharmacia AB, and Department of Surgery, University Hospital, Uppsala, Sweden ABSTRACT A methodological study has been carried out using the rabbit mesentery for the investigation of the initial haemostatic mechanism in the microvasculature. Heaemo- static plug formation in proximal arteriolar ends differ from that in venules in many respects. The arteriolar haemostatic plug formation time is shorter than the venular and it is vessel size dependent in arterioles but not in venules. The frequency of rebleeding is higher and the rebleeding time is longer in venules than in arterioles. It is stressed that. the same vessel type and size should be used in comparative studies if comparable results are to be obtained. It is also shown that most of the rebleedings appear within 10 minutes of the cessa- tion of the primary bleeding and that the haemostatic plug formation time shows a skew distribution, which should be taken into account in statistical calcula- tions. The importance of using an adequate experi- mental technique including a sharp and swift transec- tion of the vessel is also stressed. INTRODUCTION The mesenteric preparation of different animal species has been used for various microcirculatory investigations, particularly since the report of the method by Chambers & Zweifach (6). Several authors have used this method when studying the initial haemostatic process (1, 7 , 12, 16, 20, 21, 22, 23, 27, 29, 30, 36). Among these authors Hugues (22) made the most careful methodological study, although all have used the model to study different aspects of the platelet reaction in the early haemo- static mechanism. With a few exceptions (8, 9, 28) most investigators agree on the fundamental role played by the platelets in sealing a bleeding microvessel. Apitz ( 1 ) called the aggregate of platelets >>die blutstillende Thrombose. and Zucker (36) called it >>the platelet plug,,. Their histo- logical findings were verified by the ultrastructural Upsala J M e d Sci 79 studies on the shaemostatic platelet plugx made by Kjaerheim & Hovig (27), who for the first time showed that the plug was composed of densely packed platelets. Thus many investigations, dealing with different aspects of the haemostatic mechanism have been carried out in transected and punctured mesenteric microvessels. However, there are many discrepancies in the results, especially in the eva- luation of arteriolar versus venular haemostasis and the occurrence and nature of rebleedings (1, 7 , 12, 16, 22, 30, 36). Because of these contra- dictory findings it is not possible to use the mesenteric model for haemostatic studies without further methodological investigations. The aim of this investigation therefore was to: 1. standardise the mesenteric model for studies of the initial haemostatic mechanism 2. study the influence of vessel type and size on the haemostatic plug formation time 3. study the rebleeding pattern 4. establish the importance of these factors to form a basis for further physiological, patho- physiological and pharmacological studies. MATERIALS AND METHODS 112 New Zealand white rabbits of both sexes ( 2 . 7 + 0.5 kg), fed on a standard diet (Ewos pellets, Astra- Ewos AB, Sodertalje or Teknosan pellets, Ferrosan AB, Malmo, Sweden), were used. The animals were starved at least 12 hours before the experiment. Prior to anaesthesia bleeding time, whole blood coagulation time and haematocrit were performed in all rabbits as a screening method to ex- clude abnormal animals. The. bleeding time was measured in the following way: the rabbit ear was warmed for 2 minutes with a heating lamp and veins of 0.2-0.5 mm diameter Haemostatic plug formation in the rabbit mesentery 29 were transected. The mean of four bleeding times was recorded in every rabbit. Whole blood coagulation time was determined by breaking off a short length of blood-filled capillary tube every 30 second until a fibrin strand had formed. Two measurements were performed in every rabbit. Haematocrit was measured in triplicate with a micro- haematocrit centrifuge (10 000 g for 5 min; International Equipment Co., Boston, USA). The animals were anaesthetised with 20% urethane in 0.9% saline (Kebo, Stockholm, Sweden) by intravenous injection (4). Through a small midline incision the distal part of ileum was exteriorised over a siliconised glass plate in a movable, electrically heated microscope table. During the experiment the preparation was continuously super- fused with thermostated Tyrode's solution of 37.5-39.o"C and p H 7.4. The bowels were covered with moistened swabs covered with plastic foil to minimise heat loss by evaporation. The excess fluid was removed by suction. With the electrically heated table and a heating lamp the rectal temperature of the animals was kept at 37.5-39.0"C during the experiment. The temperature was continuously monitored with a thermistor probe (Laboratorio Richerche Elettroniche, Milan, Italy). For observations on the mesenteric microcirculation a Leitz BIOMED intravital microscope with a Leitz Ultrapak 6 . 5 X objective at a total magnification of 81 x was used. The diameter of the vessels was measured with a calibrated ocular graticule. The diameter was measured before transection and the vessel type was identified, i.e. arteriole or venule. In the first part of the study (35 animals) a small opthalmic knife was used for transection of the micro- vessels and in the second part (77 animals) this was performed with a disposable Gillette scalpel blade, roo, 80 "I or 5 0 Y 4 0 shape E/11. Vessels used were in the transparent part of the mesentery giving the upper limit of their sizes as about 110 pm. The time from transection until the bleeding stopped was recorded as well as the frequency and time for rebleedings. Every transected vessel was observed for about 20-30 minutes. Statistical methods The mean value, the standard deviation and the regression coefficient was calculated according to standard methods (34). The haemostatic plug formation time had a skew distribution (vide infru). For statistical purposes the original values are rendered more normally distributed by logarithmic transformation. In testing differences in haemostatic plug formation times between the different vessel ends the sign test was used (lo), and significance was considered when p<0.05. In testing differences in frequencies of rebleeding the rank-sum test was used (10). Both the sign test and the rank-sum test are distribution free tests, i.e. avoid assumptions about the nature of the distribution function of the data. Dejkitions Haemostatic plug: the platelet mass formed at the ends of a transected bleeding vessel. Primary bleeding time or primary haemostatic plug formation time: the time from transection until the bleeding f r s t stops (PHT). Rebleeding: bleedings occurring at intervals through the initially efficient haemostatic plug. Total rebleeding time: the total of all the rebleeding times recorded in one vessel. Total bleeding time or total haemostatic plug formation time: the sum of the primary bleeding time and the total rebleeding time (THT). 0 A r r e r d e s Venules 0 Fig. 1 . The frequency (in per cent) of non-bleeding vessels in relation to the vessel diameter when an 10 20 M LO 9 60 ;O % gb l i b 1iO 120 ophthalmic knife is used for the Vessel dameter i,u) transection. I Upsalu J Med Sci 79 30 D . Bergqvist 0 Arreraks Venules 0 0 0 F i g . 2. The ratio of proximal to distal bleeding vessel ends after tion to the vessel diameter. 1 z E 9 t 10 N, 30 LO 50 60 70 80 90 TOO 110 120 transection with a scalpel in rela- Vessel dtamerer (pl RESULTS The bleeding time from transected ear veins is 86?31 seconds, the coagulation time 136?78 seconds and the haematocrit 3 8 . 7 c 3 . 6 . Non of the animals was rejected from the study because of an extreme value. After transecting a microvessel in the mesentery cleanly and swiftly with a sharp knife, bleeding will immediately start from the transected vessel. Within a few seconds a haemostatic plug starts to grow because of platelets adhering to the injured vessel and aggregating to each other. The plug continues to grow until the bleeding stops after a variable length of time. When a haemostatic plug, sufficient to stop bleeding, was formed at a proximal arteriolar end, blood will continue fo flow into the nearest vessel branch. In both venular ends and in the distal arteriolar end the blood flow will either be re- versed, or stopped if anastomoses are absent. This primary haemostatic plug is not always stable and one o r more rebleeding may occur. In such a case new channels are opened up through the plug and no fragmentation of the plug occurs. With a blunt blade a certain number of vessels fail to bleed at all. Fig. 1 shows the percentage of non-bleeding vessels in the different vessel size groups based on calculations from the experiments where the oph- thalmic knife was used (35 animals). At least for arterioles there is a greater number of non-bleeding vessels when the diameter is small ( r = - 0 . 8 5 for arterioles; p O . O S ) . The rest of the results are based on calculations from the animals in which sharp scalpel blades were used for transection. After transection of a venule bleeding usually occurs from both ends of the transected vessel, but it seldom occurs from the distal end of a transected arteriole as can be seen from Fig. 2. N o correlation could be demonstrated between the ratio (number of proximal bleeding vessel ends per number of distal bleeding vessel ends) and the vessel size (r for arterioles is - 0.37 and r for venules is 0 . 6 4 ; p > 0 . 0 5 ) . The total and primary haemostatic plug formation times for different vessel sizes are given in Table I and this is demonstrated graphically in Figs. 3 and 4. It can be seen, that the haemostatic plug formation time, both the total and the primary, increases with increasing arteriolar size. There is a good correlation between arteriolar diameter and the haemostatic plug formation time (Table 11). On the other hand there is no correlation between venular size and haemostatic plug formation time (Table 11). From Figs. 3 and 4 and Table I it can also be seen that the total venular rebleeding time is much greater than the total rebleeding time for the proximal arteriolar end, and that there is a great variation as far as the distal arteriolar end is concerned. The haemostatic plug formation time for distal Upsala J Med Sci 79 Haemostatic plug,formation in the rabbit mesentery 3 1 Table I. Total ( T H T ) and primary ( P H T ) haemostatic plug formation times f o r arterioles and venules The mean value and S.D. are given Arterioles Venules Vessel Proximal end Distal end Proximal end Distal end size Olm) THT PHT THT PHT THT PHT THT PHT 7 3 7 2 4 9 2 7 2 2 9 62$.73 5 6 1 6 7 13 3 5 2 3 6 3 1 2 3 5 102+ 127 861-88 200f256 166f238 277f281 2322266 20 41 2 4 6 32_+31 138k216 1172209 2702300 2012273 2782268 2241246 26 4 5 k 4 3 3 7 2 3 5 113f203 8 8 2 1 9 7 290+304 221F279 2322255 195i-245 33 6 6 k 5 4 4 9 2 4 2 1782259 1772260 2622218 1762187 2932287 245+278 40 7 5 f 7 3 6 2 2 6 5 2572318 252+317 1892381 1452128 353f322 283f306 46 8 5 2 7 5 6 7 f 6 1 3852396 2142324 2502239 126+134 391f302 2952308 53 99+86 7 7 2 6 2 2622332 1582276 2211-204 1492134 386+314 3102307 59 9 9 2 7 1 8 8 f 6 8 206f281 1881-280 2542273 1952269 3182287 2612277 66 1231103 101F98 363_+364 314_+354 2522287 2291296 3472312 260_+279 79 139k91 121+89 2342270 2342270 3832310 3142287 3532293 283f275 92 1 6 4 2 9 9 1471-92 366f365 3652377 2641-271 2292261 302k266 230k243 106 1992109 1842106 4 3 6 t 3 3 0 4092350 arteriolar ends is significantly longer than that for the proximal ends ( P < O . O O l ) , but the corresponding venular haemostatic plug formation times d o not differ significantly ( p > O . O S ) . The haemostatic plug formation time for proximal arteriolar ends is statistically shorter than that for venules ( P < O . O O l ) . The difference between venular and distal arteriolar haemostatic plug formation time is insignificant for the primary time (P>O.OS) and of borderline significance for the total time (O.Ol97%) rebleedings have occurered, and within 5 minutes more than 70% of the rebleedings have occurred. Every line represents one vessel size and the time for re- bleeding t o occur is independent of the vessel size. The distribution of the haemostatic plug formation times is skew, that is, even in normal rabbits some vessels bleed for a very long time. The longest haemostatic plug formation times are seen in distal arteriolar ends and in venules. In Figs. 8 and 9 the distribution curves are given for vessels of the sizes 26 and 53 p m . The skew distribution is rendered more normal after logarithmic trans- formation, DISCUSSION rorai reb/eed(ng r m e isec / Since urethane is known to induce haemolysis ( 5 ) with the release of adenosine diphosphate, ADP, the use Of anaesthesia in haemostatic studies may be theoretically unacceptable. ADP is, F i g . 5 . The stability of the haemostatic plug is shown by plotting the frequency of rebleeding against the total rebleeding time. Origo represents an absolutely stable plug. Upsala J Med Sci 79 Huemostutic p l u g formation in the ruhbit mesentery 3 3 Table 11. Correlation coefficients between the vessel diameter in pm and different parameters f o r the four vessel segments studied NS: not significant. pco.001 Correlation ( r value) between vessel size and ~ Proximal Distal Proximal Distal arteriolar end arteriolar end venular end venular end primary haemostatic plug formation time 0.99*** 0.92*** 0.51 NS 0.23 NS Total haemostatic plug formation time 0.99 * * * 0.84*** 0.42 NS 0.31 N S - 0.48 NS Frequency of rebleeding 0.29 NS 0.06 NS -0.34 NS Total rebleeding time 0.60 NS 0.01 NS -0.19 NS 0.57 NS 100 3 - c u k w m e w 5 9 y 50 u P F d - 9 % 0 $ - 2 2 60 IN) I80 240 300 360 120 A80 540 600 I m e fsec I Ilme ( S K I given above the curve). 1 120 F i g . 6 . The number of re- bleeding arteriolar haemo- static plugs expressed as per cent of total number of re- bleeding plugs plotted against the time between the first arrest of bleeding and the start of the first rebleed- ing. Every line represents L&J 5$ 6& one vessel size (diameter 3-742854 Upsala J M e d Sci 7 9 34 D. Bergqvist h 2 0 a e c 0 I Promma/ venules I I r m fsec) rime fsec) however, rapidly metabolized and removed from plasma (17, 18, 24). We have evaluated the effect of different anaesthetic agents on the formation and stability of t h e haemostatic plug using the mesenteric model and found no significant differences between urethane, chloralose and mebumal sodium anae- sthesia or neurolept analgesia (4). As rabbits anaesthetised with urethane were stable in acid- base balance and blood pressure, and as urethane was the agent most easy to handle in our hands, we decided to use it in our further studies. It must be borne in mind, however, that our haemostatic studies were made while the mesenteric Fig. 7. The number of re- bleeding venular haemostatic plugs expressed as per cent of total number of rebleed- ing plugs plotted against the time between the first arrest of bleeding and the start of the first rebleeding. Every line represents one vessel size (diameter given above the curve). preparation was continuously superfused with Tyrode’s solution. The superfusion milieu is standardised as far as possible, for ions, pH, temperature and flow. Because of the superfusate it must be stressed, that the haemostatic plug formation time obtained with the present method may not be comparable to bleeding times obtained with other techniques, for instance described by Duke ( 1 1 ) or I v y (25). Comparisons between these two types of techniques should thus be very care- ful. Using a continuous superfusate, blood and various tissue factors are washed away. There are also differences between the tissues themselves, ffpsulu J Med Sci 79 Haemostatic plug formation in the rabbit mesentery 35 Fig. 8. The distribution of the arteriolar haemostatic plug formation time shown L 200 3w 100 500 i.e. anatomical localisation, thickness, forces of pressure and pulling and numbers of transected vessels. Another factor of importance is the transection technique. The phenomenon of nonbleeding vessels has been described before but not the difference between vessel types ( 1 , 22, 3 3 , 36). Endothelial adhesion was previously considered to be the main haemostatic mechanism in capillaries (15, 33), but Jsrgensen & Borchgrevink ( 2 6 ) showed histo- logically that platelet plugs were formed in capillaries too. From our study it is evident (Fig. I), that the frequency of non-bleeding vessels, when using a blunt knife, is somewhat higher in arterioles than in venules, and that the frequency is correlated to the diameter of the arteriole. The theory of endo- thelial adhesion and the frequent occurrence of non-bleeding vessels are probably due to the fact that the importance of sharp knives and swift 6W 760 /set/ for two vessel sizes. transections in previous haemostatic experiments wasnot considered. Inconclusion, theseresultsmade us use disposable knife blades in the main study. The frequency of non-bleeding vessels could then be neglected and could well be compared with Hugues' (22) results. Some authors studying the initial haemostatic mechanism have used puncture of microvessels instead of transection (7, 14, 19, 36). In those studies the size of the puncture hole, the thickness and elasticity of the vessel wall and the adjacent tissue will probably influence the bleeding time. When a vessel is punctured there is a combination of a haemostatic plug outside the vessel and a microthrombus inside it. After transection of a venule bleeding occurred from both ends of the transected vessel, while bleeding was less common from the distal than from the proximal arteriolar end, which almost ,urn Proxmal venuk 26,um Fig. 9. The distribution of formation time shown for Roxrmal 5 3 p m Dsra! 2 6 p m the venular haemostatic plug I I I I I I I na 200 3aJ 403 Mo 600 700 .do & fscCl two vessel sizes. Vpsala J Med Sci 79 36 D. Bergqvist always bled (Fig. 2). Although this phenomenon has been briefly mentioned by various authors (1, 13, 22, 33, 36), no systematic study of this problem has been carried out. Fig. 2 shows that the number of observations from distal arteriolar ends is less than from the other vessel ends, a fact that must be remembered in the further discussion. The pressure differences, that exist between proximal and distal arteriolar ends, could in part explain this observation. A pressure difference should also exist between the both venular ends, but this will not be so pronounced as in arterioles and probably is of no importance since the venular ratio is near 1 . There are contradictory results in the literature as far as the relationship between the haemostatic plug formation time and vessel type and diameter is concerened ( 1 , 22, 33, 36). Hugues (22) observed the importance of the vessel type, but apart from his study and as far as the vessel size is con- cerned, no systematic studies on these problems have been carried out. Hugues (22) found the longest bleeding times in proximal arteriolar ends, whereas bleeding times from the other vessel seg- ments were of the same magnitude and about half that seen in the proximal arteriolar ends. In his calculations Hugues did not use the 11% of the transected vessels, where the bleeding time exceeded 10 minutes, which was seen in venules and distal arteriolar ends-a finding that fits well with our experience (vide infru). This is probably the explanation why Hugues found the primary haemostatic plug formation time referred to above, where the proximal arteriolar end showed the longest bleeding time. He pointed out that the most reproducible results were obtained from the proximal arteriolar ends, but we suggest that valuable in- formation also can be obtained when the venules are taken into consideration. Thus we found aspirin to have no effect on arteriolar haemostatic plug formation time, whereas venular haemostatic plug formation time was significantly shortened (2). In his study Hugues (22) also studied the vessel contraction in connection with haemostatic plug formation. He found no correlation between the degree of contraction and the bleeding time. We found that the haemostatic plug formation time is longer and the rebleeding tendency greater in venules than in arterioles, and that the time for haemostatic plug formation is significantly longer in the distal than in the proximal arteriolar end. Upsala J Med Sci 79 These investigations clearly show, that in studies of haemostatic plug formation one must be careful to differentiate between arterioles and ven.ules as well as taking into account the vessel diameter at least in arterioles. It can be concluded, that the same vessel size and type should be used in com- parative studies, if comparable results are to be obtained. This fact has not always been taken into consideration (12, 16, 30). The literature concerning the nature and occurrence of rebleedings is contradictory ( 1 , 7, 12, 16, 20, 22, 30, 36), and no systematic study on this phenomenon has been made. The rebleeding always started with the opening up of new channels through the formed plug as pointed out by Zucker (36). However, some authors have remarked on the frequency of fragmentation both in rebleedings (35) and during the growth of primary plugs (3 1, 32). We found the lowest frequency of rebleeding in the distal arteriolar ends, but the length of the total rebleeding time varied. This time and the frequency of rebleeding was fairly constant for the proximal arteriolar end, while both parameters were longer for the two venular segments. Both proximal vessel ends showed some correlation between the frequency of rebleeding and the total rebleeding time. On the other hand the rebleeding pattern was independent of the vessel size in the range investigated. Our results, showing that re- bleedings are more frequent in venules, speak against the hypothesis claimed by Aptiz ( l ) , that pressure is of prime importance for the rebleedings to occur. Thus it is also of considerable importance t o define the vessel type when studying the re- bleeding pattern, whereas the vessel size seems to be of less importance within the range under in- vestigation. We have found, that the rebleeding pattern is a useful parameter, for instance in pharmacological studies. Thus, the rebleeding fre- quency was significantly decreased in rabbits treated with aspirin (2), ether ( 4 ) and sulfinpyrazone or phenylbutazone (unpublished results), but the haemostatic plug formation time was normal in all groups except the aspirin group in which the venular haemostatic plug formation time was shortened. As can be seen from Figs. 6 and 7 the majority of rebleedings (>97%) occurred within 10 minutes of the formation of the primary haemostatic plug. Using this experimental model system it would thus Haemostatic plug formation in the rabbit mesentery 37 be sufficient in normal, untreated rabbits to observe for a period of 10 minutes after primary haemo- statis. The distribution of haemostatic plug formation time has not been studied in detail before, although Hugues (22) found that in I I % of transected venules and distal arteriolar ends the bleeding continues for more than 10 minutes. In these cases he found fragmentation from already formed plugs or no plugs at all. In our studies it was extremely rare to find a total absence of plugs. On the other hand the growth rate of the plugs in venules and distal arteriolar ends was very slow, when bleeding was prolonged (3). As can be seen from Figs. 8 and 9 the haemostatic plug formation time shows a skew distribution in our material, the skewness being more pronounced for venules and distal arteriolar ends. Since prolonged bleeding times are a part of the normal bleeding time distribu- tion it is our opinion, that they must be included in calculations on haemostatic plug formation time. In our material logarithmic transformation of the haemostatic plug formation time render it more normally distributed and in our opinion this method should be used, when statistical comparisons are made. CONCLUSIONS I . A swift and clean transection with a sharp knife should be used in experiments on haemostatic plug formation. 2. The ratio of bleeding proximal ends to bleeding distal ends is between 2 and 3 for arterioles and about 1 for venules. This ratio is not correlated to the vessel size. 3. Total and primary haemostatic plug formation time in proximal arteriolar ends are statistically shorter than those in venules and distal arteriolar ends. 4. Total and primary haemostatic plug formation time in arterioles but not in venules are correlated to the vessel diameter. 5. The difference in haemostatic plug formation time between arterioles and venules decreases with increasing vessel size. 6. The total rebleeding time is longer in venules than in proximal arteriolar ends. Distal arteriolar ends show a variable pattern. 7. The frequency of rebleeding in proximal arteriolar and venular ends is correlated with total , rebleeding time. 8. There is no correlation between vessel diameter and total rebleeding time or frequency of re- bleeding. 9. The frequency of rebleeding is lowest for distal arteriolar ends, followed by the proximal arteriolar ends and both venular ends. 10. Most rebleedings (>97%) have occurred within 10 minutes after the formation of the primary haemostatic plug. 1 I . The haemostatic plug formation time shows a skew distribution which can be rendered more normal by logarithmic transformation. REFERENCES 1 . Apitz, K.: Die Bedeutung der Gerinnung und Thrombose fur die Blutstillung. Virchows Arch Path Anat 308: 540, 1942. 2. Arfors, K.-E.,Bergqvist,D.,Bygdeman,S., McKenzie, F. N. & Svensjo, E.: The effect of inhibition of the platelet release reaction on platelet behaviour in vitro and in vivo. Scand J Haematol 9: 322, 1972. 3. Bergqvist, D. & Arfors, K.-E.: The growth rate of haemostatic plugs in the rabbit mesentery. Proc. VII Conf. Microcirculation, Aberdeen, 1972. Karger, Basel, 1973. p. 186. 4. Bergqvist, D., McKenzie, F. N. & Arfors, K.-E.: Influence of various anaesthetic agents on the formation and stability of haemostatic plug in the rabbit mesentery. Upsala J Med Sci, in press. 5 . Bree, M. & Cohen, B. Effects of urethane anaesthesia on blood and blood vessels in rabbits. Lab Animal Care 15: 254, 1965. 6 . Chambers, R. & Zweifach, B. W.: Topography and function of the mesenteric capillary circulation. Am J Anat 75: 173, 1944. 7. Chen, T . I . & Tsai, C . : The mechanism of haemo- stasis in peripheral vessels. J Physiol 107: 280, 1948. 8 . Cruz, W. 0.: Studies on hemostasis in vivo. Henry Ford Hospital Symp. Blood Platelets. Little, Brown & Co., Boston, 1961. 9. Cruz, W. 0.: The significance ofa smooth muscle com- ponent in hemostatis. Proc SOC Exptl Biol Med 119: 876, 1965. 10. Dixon, W. J . & Massey, F . J., Jr: Introduction to Statistical Analysis, McGraw-Hill, New York, 1957. 1. Duke, W . W.: The relation of blood platelets to hemorrhagic disease; Description of a method for determining the bleeding time and coagulation time and report of three cases of hemorrhagic disease relieved by transfusion. J Am Med Assoc 55: 1185, 1910. 2 . Evans, G . , Packham, M. A., Nishisawa, E. T., Mustard, J . F. & Murphy, E. A . : The effect of acetylsalicylic acid on platelet function. J Exptl Med 128: 877, 1968. 13. French, J. E., Macfarlane, R. G. & Sanders, A. G . : The structure of haernostatic plugs and experimental thrombi in small arteries. Brit J Exptl Path45: 467, 1964. Upsala J M e d Sci 79 38 D . Bergqvist 14. Herrmann, R . , Frank, J . D. & Marlett, D. L.: An in vivo technique for assessing the formation of a hemostatic platelet plug. Proc SOC Exp!l Biol Med 128: %0, 1968. 15. Herzog, F.: Die Rolle der Capillaren bei der Blut- stillung. Pfliigers Arch Gesamt Physiol207: 476, 1925. 16. Hirsh, J . , Buchanan, M., Glynn, M. & Mustard, J . F.: Effect of streptokinase on hemostatis. Blood 32: 726, 1968. 17. Holmsen, H. & Stormorken, H.: Kinetic studies on the breakdown of adenosine diphosphate in human plasma. Scand J Clin Lab Invest Suppl. 84, p. 138. 1965. 18. Holmsen, I . & Holmsen, H.: Breakdown of adenosine nucleotides in whole blood. 111 Congr Int SOC Thromb Haemost. Washington 1972. Abstract volume p. 206. 19. Honour, A. J . & Mitchell, J . R. A,: Platelet clumping in injured vessels. Brit J Exptl Path 45: 75, 1964. 20. Hovig, T., Rowsel!, H. C., Dodds, W. J., Jsrgensen, L. & Mustard, J. F.: Experimental hemostatis in normal dogs and dogs with congenital disorders of blood coagulation. Blood30: 636, 1967. 21. Hovig, T., Dodds, W. J., Rowsell, H. C. & Mustard, J. F.: The transformation of hemostatic platelet plugs in normal and factor IX deficient dogs. Amer J Path5.3: 355, 1968. 22. Hugues, J . : Contribution a I’etude des facteurs vasculaires et sanguins dans I’hemostase spontanee. Arch Internat Physiolll: 565, 1953. 23. Hugues, J . : Metamorphose visqueuse des plaquettes e t formation du clou hemostatique. Thromb Diath Haemorrh 3: 34, 1959. 24. Ireland, D. M. & Mills, D. C. B.: Degradation of adenosine diphosphate in plasma. Biochem J 92: 30P, 1%4. 25. Ivy, A. C., Shapiro, P. F. & Melnick, P.: The bleeding tendency injaundice. Surg Gynec Obstet60: 781,1935. 26. Jsrgensen, L. & Borchgrevink, C.: The platelet plug in normal persons. I. The histological appearance of the plug 15 to 20 minutes and 24 hours after the bleeding and its role in the capillary hemo- stasis. Acta Path Microbiol Scand57: 40, 1963. 27. Kjaerheim, A. & Hovig, T.: The ultrastructure of haemostatic blood platelet plugs in rabbit mesenterium. Thromb Diath Haemorrh 7: I , 1962. 28. Macfarlane, R. G.: Critical review: the mechanism of haemostasis. Quart J Med N S 10: 1 , 1941. 29. Marr, J . , Barboriak, J . J . & Johnson, S. A.: Relation- ship of appearance of adenosine diphosphate, fibrin formation and platelet aggregation in the haemostatic plug in vivo. Nature 205: 259, 1965. 30. Packharn, M. A,, Warrior, E. S., Glynn, M. F., Senyi, A. S. & Mustard, J . F.: Alteration of the response of platelets t o surface stimuli by pyra- zole compounds. J Exptl Med 126: 171, 1967. 31. Roskam, J . : Role of platelets in the formation of a hemostatic plug. Blood platelets. Little, Brown & Co., Boston, Mass., 1961. 32. Roskam, J . , Hugues, J., Bounameaux, Y. & Salmon, J.: The part played by platelets in the formation of an efficient hemostatic plug. Thromb Diath Haemorrh 3: 510, 1959. 33. Sanders, A . G.: In vivo observations on haemostasis in the hamster. Symp. Zoo1 SOC Lond 27: 109, 1970. 34. Snedecor, G. W. & Cochran, W . G.: Statistical methods. The Iowa State Univ. Press. Ames, Iowa, USA. 6th ed. 1967. 35. Stormorken, H. & Owren, P. A.: Physiopathology ofhemostatis. Sem Hemat8: 3, 1971. 36. Zucker, M. B.: Platelet agglutination and vaso- constriction as factors in spontaneous hernostasis in normal, thrombocytopenic, heparinized and hypo- prothrombinemic rats. Amer J Physiol148: 275, 1947. Received M a y 16, 1973 Address for reprints: David Bergqvist Department of Experimental Medicine Pharmacia AB, Box 604 S-751 25 Uppsala 1 Sweden Upsala J Med Sci 79