THE ANATOMICAL RECORD 207:339-348 (1983) Transverse Periosteal Sectioning and Femur Growth in the Rat JAMES B. McLAIN AND PETER S. VIG Bowman Gray School of Medicine, Wake Forest Uniuersity, Winston-Salem, NC 27103 (J.B.M.) and Department of Orthodontics, the Uniuersity of Michigan, A n n Arbor, MI 48109 R S . V) ABSTRACT Circumferential cuts through the periosteal covering of long bones have been demonstrated to transiently increase epiphyseal growth. This effect appears to be independent of vascular changes accompanying surgery and has been hypothesized to relate to releasing tension in the periosteal envelope. This study was designed to address problems of previous investigations by controlling for the effects of the surgical procedure and by using regression analyses to analyze intra- and interanimal variations in the length and proportionality of the femur in experimental, sham, and control SpragueDawley male rat littermates. Experimental animals received circumferential periosteal sectioning of the right femur and no operation to the left limb. A sham operation without periosteal sectioning was performed on the right femur in the sham group. Right to left differences were analyzed using two multiple regression models; one involved three absolute length measurements as the dependent variables, while the other used the three ratios of these length measurements as the dependent variables. The ratio measures were utilized to reflect changes in bone proportionality. Circumferential periosteal section was followed by a n alteration in the shape of rat femurs a t 2 weeks postsurgery with a slight retardation of the length dimension from medial epicondyle to head of the femur and a n overgrowth of the length dimension from the lateral epicondyle to the greater trochanter. The sham procedure produced a proportional decrease in all length measurements. The experimental procedure was also associated with surface bone apposition at the site of section. At 3 weeks postsurgery, normalization of bony contours between sham, experimental, and control groups had occurred; however, there were still some statistically significant decreases of length dimensions in the sham and experimental groups. In the experimental group the length measurement involving the weight-bearing head of the femur remained reduced a t 3 weeks postsurgery. It is hypothesized that the functional demands of the long bone play a n important role in the effect of periosteal regulation on growth. In situations where a normal tensive force is exerted on the bone, the periosteal envelope will act to restrain epiphyseal growth. When the bone is under a normal compressive force the release of periosteal tension is not a quantitatively significant stimulus to epiphyseal growth and the effects of surgical intervention and muscle trauma will play a more important role in the growth response of the epiphyses. Form and function interactions that play a n important role in normal bony development must act through a variety of mediating factors from muscle and soft tissue influences at the gross level to biochemical processes at the cellular level. The role of the 0 1983 ALAN R. LISS. INC. periosteal covering of bones in this interaction remains a n area of interest for investigation. It has been suggested that the growing epiphyses place tension on the fiReceived January 21,1983; accepted May 13, 1983. 340 J.B. McLAIN AND P.S. VIG brous layer of the periosteum and that essentially it is stretched over the underlying shaft (Lacroix, 1951; Moss, 1978). The hypothesis of a “neutral zone” suggests that a region exists near the midshaft of a long bone where the tension on the fibrous layer of the periosteum from the two growing epiphyses is equalized and is therefore essentially zero. A slower growing periosteum could then exert a compressive force on the underlying epiphyses and thus control or regulate the growth of the long bone. Hueter (1862) and Volkmann (1862) noted the relatively diminished growth of bones that are under compressive forces. Others have demonstrated the ability to retard longitudinal growth of long bones by the application of compressive loadings on the epiphyseal plates (Mueller, 1922; Strobino, et al., 1952; Bylander et al., 1981, 1983). One method of accomplishing this, by the stapling of one epiphyseal plate, results in an overgrowth of the contralateral epiphysis (HallCraggs and Lawrence, 1969). It may be hypothesized that the stapling procedure moves the neutral zone toward the stapled epiphysis, thus reducing the force on the opposite epiphysis and permitting more growth. Roskjaer (1977) noted that the highest turnover of periosteal cells occurs adjacent to the fastest growing epiphysis. Attempts to stimulate long bone growth in cases of limb inequality have been a stimulus to the study of periosteal stripping operations. Unlike a simple circumferential cut through the fibrous layer of the periosteum, most of these studies have involved partial or complete stripping of the periosteum from the shaft of the long bone. These methods and others that have attempted to increase endosteal blood flow have generally produced unpredictable results (Wu and Miltner, 1937; Lacroix, 1951; Brodin, 1955; Jansen and Langenskiold, 1956; Bei-trand and Trillat, 1948; Jenkins et al., 1975; Larson et al., 1961; Sola et al., 1963; Shui and Wong, 1964; Yabsley and Harris, 1965; Pease, 1952; Frejka and Fait, 1958; Chan and Hodgson, 19701. The unpredictability of the stripping operations may relate to the possibility that two control systems are being disturbed. In a study of rabbits Whiteside and Lesker (1978) showed that only in animals with traumatized muscle did periosteal dissection appear to severely compromise collateral circulation and that in animals with untraumatized muscle neither extraperiosteal nor subperiosteal dissection reduced blood flow. There- fore in those experiments involving a greater degree of muscle trauma, it may be hypothesized that endosteal blood flow increased, and greater stimulation of the growing epiphyses would be expected. If the periosteum is under tension and does restrict epiphyseal growth, surgical release of this tension should stimulate growth, at least until the periosteum heals. Crilly (1972) and Auer et al. (1982) observed measurable clinical retraction of the margins of circumferentially cut periosteum in long bones. Hellquist (19721, in experiments which involved stripping of the periosteum from rabbit maxillae, found consistent regeneration of the resected periosteum that was more adherent to the underlying bone, as well as transiently more cellular and thicker. These histological changes could easily affect the elasticity of the fibrous layer of the periosteum and its ability to stretch over the underlying bone. In fact, Hellquist observed a deviation of the snout toward the operated side when periosteum was stripped from the left nasofrontoand premaxillo-maxillary bones. Crilly (1972) attempted to differentiate the effects of periosteal tension, vascular damage, and muscle trauma on long bone growth in a set of experiments utilizing the radius of the chicken. At 17 days he found a n initial overgrowth of 27.3% for transverse fracture, 2.8% for muscle trauma, 24.8% for transverse periosteal section, and only 1.9% for longitudinal periosteal section. Results at 120 days demonstrated differences that were much reduced-8.5%, 1.0%, 5.7%, and O.O%, respectively. He concluded that the major factor involved in longitudinal overgrowth was the release of the periosteal tension and that little was gained with the addition of transverse fracture of the shaft. He recognized that there was some influence from changes in endosteal blood flow, but concluded that this contribution was negligible. The negligible results with longitudinal periosteal section coincide with Whiteside and Lesker’s (1978) observations that the collateral circulation is not severely compromised unless muscle is traumatized. One severe limitation of Crilly’s work was that experimental and sham radii were used for comparison. This has the potential for exaggerating the amount of overgrowth, since both experimental and sham radii received surgery to place radiopaque implants. This design precludes any assessment of the effect of the surgical approach and inadvertent muscle damage on the growth of the ra- TRANSVERSE PERIOSTEAL SECTIONING dii. It is therefore possible that an observed increase in growth may have been a feature of the experimental design and not a valid test of his hypothesis. The purpose of the present study was to differentiate the effects of surgery alone from those of surgery together with periosteal section on growth of the rat femur. The rat femur was selected since it is a weight-bearing bone. In contrast to Crilly's studies on chick radii, we also sought to detect the possible influence of function on the growth behavior of a long bone. MATERIALS AND METHODS Six sets of six male Sprague-Dawley outbred albino rat littermates were used for this study. All 36 animals were 21 days old at the time of surgery. The planned time of sacrifice for three litters (18 animals) was 14 days after surgery (i.e., 35-day-old animals), and for the remaining three litters (18 animals) the date of sacrifice was 21 days after surgery (i.e., 42-day-old animals). The litters were randomly assigned to 2- and 3-week groups using a table of random numbers. In each of the litters three animals were designated as experimental and three as sham animals. The experimental or sham procedure was limited to the right femur, and for each animal the left femur served as the unoperated comparison for that animal. Changes in the normal function of these contralateral limbs could not be accounted for in this design; however, the animals continued normal activity and diets throughout the study period. The experimental procedure consisted of a dorsal approach to the midshaft region of the femur. A sharp curette was used to cut and gently reflect the perios- Fig. 1. Length measurements used for the study. A) the longest distance from the most proximal point on the head of the femur to the most distal point of the medial condyle; B) the shortest distance from the most distal point on the neck of the femur to the most proximal 341 teum completely around the shaft. The cut margins of the periosteum retracted leaving an approximately 2-mm wide area of bony shaft exposed. The completeness of section was confirmed by the tactile feeling of the curette against the bone. The area of section approximated the region where the distal extent of the Adductor brevis and proximal end of the Vastus intermedium muscles insert into the shaft of the femur. This approach appeared to be the most atraumatic available to the midshaft region. The skin was sutured and the animals were returned to their cages for routine care. The sham procedure consisted of an identical surgical approach with reflection of all muscles, but did not include periosteal section. No surgical procedures were carried out on the left femurs. An additional series of unoperated femurs from 42-day-old animals, utilized in a similar surgical study of the mandible, were utilized to gauge normal intraanimal variation between right and left femurs. Animals were sacrificed according to their schedule at either 2 or 3 weeks with an overdose of intraperitoneal pentobarbital. Both femurs and the mandibles from each animal were recovered and defleshed by Sarcophagus beetle larvae. The bones with cartilage intact were placed in warm tap water in sealed containers. After approximately two rinsings with water the articular cartilage cleanly separated from the underlying bone. The bones were then dried for examination. Direct measurements were made on each femur by the same investigator using a Helios caliper. Figure 1 shows the location of the three length measurements (A, B, and C) on a drawing of a 42-day-old rat femur. All point on the intercondylar notch; and C) the longest distance from the most proximal point of the greater trochanter to the most distal point of the lateral condyle. Shaded area indicates site of section. 342 J.B. McLAIN AND P.S. VIG measurements were taken on three separate occasions by the same investigator, and measurement error did not exceed 0.5% for any dimension. Data were analyzed using the Statistical Analysis System’s multivariate regression analysis package (GLM). Each of the three independent recordings of the measurements were kept separate for analysis purposes and provided the pure error sum of squares for the regression models generated. Two separate statistical models were developed for this study. The first model included the differences between right and left femurs for the three length measurements A, B, and C a s the dependent variables; the individual animals served as the independent variables. In the second model, which was used to test for similar proportionality of the bones in the experimental, sham, and control groups, the dependent variables were the differences between right and left femurs for the ratios A:B, A:C, and B:C, and the independent variables were the individual animals. By utilizing difference measurements between right and left femurs for each animal, in effect each animal served as its own control. Separate comparisons of each group with the other two, i.e., experimental to sham, experimental to control, and sham to control, were made using the “contrast statement” capabilities of the GLM statistical program. The control group allowed assessment of the normally occurring intraanimal variation between right and left femurs, while the use of a sham group allowed separation of effects due to sectioning and surgical procedure and the effect of surgical intervention alone. These factors combine to make this design very sensitive to changes due to treatment. Data for the femurs were analyzed separately for the animals sacrificed at 2 and 3 weeks postsurgery to eliminate statistical problems associated with growth effects. For the animals sacrificed at 2 weeks no intraanimal variation group existed, and sham and experimental groups were compared. For animals sacrificed a t 3 weeks postsurgery, the intraanimal limb variation was available from animals taken from a parallel study involving the mandible. The intraanimal limb variation was quite small at 4 days of age and one would not expect the magnitude of this variation to have been larger a t 3 days with smaller femurs. RESULTS Gross observations of the femurs taken from animals sacrificed 14 days and 21 days postsurgery are shown in Figures 2 and 3, respectively. At 14 days postsurgery both the experimental and sham animals demonstrate a decrease in length A on the operated side (Fig. 4); however, the experimental animals demonstrate less of a n effect than do the sham animals with the difference significant at P < 0.0001 (Table 1).For Measurements B and C, the experimental group demonstrates a n increased length on the operated side, while the sham group demonstrates a decreased length on the operated side. The length differences between the right and left sides between the two groups were significant at P < 0.0001. The ratio measurements A:B, A:C, and B:C represent a n attempt to measure distortion or “bending” of the femur attributable to the interventions (Fig. 5). A reduction in ratio A:B in the experimental group 14 days postsurgery reflects the relative overgrowth of B and undergrowth of A. In contrast, the sham animals showed reduction in B on the operated side, which was relatively greater than the reduction of length A. Ratio A:C demonstrates a decrease in A relative to C in the experimental group. For the sham group, lengths A and C have been reduced in similar proportions so that the ratio value A:C is the same for operated and unoperated sides. Ratio B:C remains the same for operated and unoperated sides in the experimental animals, indicating that both lengths B and C were increased in similar proportions. The sham animals were characterized by a relatively greater reduction in length B than in C, resulting in a decrease in ratio B:C. The statistical analyses for the ratio measures appear in Table 1. To summarize, for animals aged 35 days, which were sacrificed 14 days postsurgery, the following observations were made. 1) In experimental animals the operated side was characterized by a deformation of the bone with a slight decrease in length A (involving Fig. 3. Dried femurs from 42-day-old rats, 21 days following surgery. Periosteal sectioning was performed in the midshaft region of the right femur. Left femur served as intraanimal control. A) ventral view; B) dorsal view; R) right (experimental side); L) left (control side); a) the lateral ridge forming the insertion of the Vastus lateralis and Vastus intermedius in the control limb; b) the lateral ridge in the operated limb appears to extend more proximally and laterally; c) some animals exhibited bony alteration at the site of section, while an equal number show no alteration. The corresponding dorsal surface appears essentially normal. Fig. 2. Dried femurs from 35-day-old rat, 14 days following surgery. Periosteal sectioning was performed in the midshaft region of the right femur only. Left femur served as intraanimal control. A) ventral view; B) dorsal view; R) right (experimental side); L) left (control side); a) the lateral ridge forming the insertion of the Vastus lateralis and Vastus intermedius in the control limb; b) the lateral ridge in the operated limb appears more proximal and exhibits some concavity; c) the site of previous periosteal sectioning appears as roughened new bone formation on the ventral surface and a smoother deposition on the dorsal surface. Figure 3. 344 J.B. McLAIN AND P.S. VIG 22.00 19.00- 21.80 i8.8C18.60- .... __.. .. 2- 18.40\ 20.20- 18.00- 20 00- 17.80- 19 80- 17.60- 19.60- 17.40- 19.40- ~.-. 1 ~ R L R EXP L 20.8020.6020.4020.20- -. ..~ . 20.00- tL 19.80 17.00- SHRM -. - 17.20' 19 2 0 - 21.20:1.00- 18.20- R L L R R DIP EXP A Fig. 4. Length measurements A, B, and C for animals sacrificed 14 days postsurgery (35 days old). Vertical scale is in millimeters. Dots represent means, bars represent standard deviations, and dotted lines represent SHAM B mv L C maximum and minimum values. EXP, experimental animals (N = 8); SHAM, sham animals (N = 7); R, right; L, left. Death following surgery, N = 3. significantly different from each other, but were significantly different from the control animals a t P < 0.0001. The ratio measurements for animals sacriVariable' Sum of squares F value P > F' ficed 21 days postsurgery are presented graphically in Figure 7. This figure indicates A 0.49968 249.84 0.0001 B 2.74523 368.76 0.0001 that right and left femurs exhibited similar C 2.18508 291.34 0.0001 ratios for the three length measurements for A:B 0.00417 118.17 0.0001 both the sham and control groups. The experA:C 0.00117 63.33 0.0001 imental group was characterized by dissimiB:C 0.00036 10.34 0.0031 lar ratios for the right and left femurs. :Degrees of freedom for each variable = 1 In the experimental group length A was Probability of obtaining the F value by chance significantly decreased relative to both lengths B and C, and length C was also sigthe head of the femur) and with a n increase nificantly decreased relative to length B (Tain lengths B and C. 2) In experimental ani- ble 2). mals measurements B and C were similarly To summarize, for animals aged 42 days, affected by the operation. 3) In sham animals which were sacrificed 21 days postsurgery, the operated side was characterized with a the following was observed. 1)In both the reduction in all three length measurements sham and experimental animals, the operwith no bony deformation. 4) In sham ani- ated side was characterized by a reduction in mals lengths A and C were similarly affected all three length measurements. 2) The sham by the operation, while length €3 appeared to group experienced a proportional reduction be decreased by the operation to a greater in all three length measurements on the opdegree. erated as compared to the unoperated side. At 21 days postsurgery both experimental 3 ) The experimental animals showed a relaand sham groups showed decreased length tively greater decrease in length A than in dimensions in the operated limbs compared lengths B or C on the operated side. 4)The to the contralateral unoperated limb (Fig. 6). experimental group also showed a relatively The control group demonstrated minimal greater decrease in length C than in length right to left variation in unoperated femurs. B on the operated side. Measurement A, involving the head of the DISCUSSION femur, was diminished to a greater extent in the experimental group than in the sham; The results are summarized and presented however, both groups showed decreased in Figure 8. As predicted from reviewing prelengths when compared to the control group vious investigations, the stimulatory effect (P < 0.0001; Table 2). For measures B and C of periosteal section on long bone growth was the experimental and sham groups were not greatest nearest the time of sectioning (2 TABLE 1. Multiple regression analysis results comparing right to left limb differences between experimental and sham animals 14 days postsurgery (35 davs old) 345 TRANSVERSE PERIOSTEAL SECTIONING 1.00 I l l L R / ,9s .R5 .90 an R R L EXP ,85 L R L R EXP 1 R SHRM EXP SHRM AIB BIC L SHAM A/C Fig. 5. Ratio measurements A:B, B:C and A:C for animals sacrificed 14 days postsurgery (35 days old). 25.00 22.00 24.50 21.50 24.00 21.00 23.50 20.50 23.00 20.00 22.50 19.50 22.00 19,oo 21.50 18,Sfl 21.00 R L c @ R L s l y R L w m 1R.flO R L LXP R L SHAM A R L W CONTROL maximum and minimum values. EXP, experimental animals (N = 8); SHAM, sham animals (N = 8); CONTROL, control animals (N = 20); R, right; L, left. Death following surgery, N = 2. TABLE 2. Multiple regression analysis results comparing right to left limb differences between experimental, sham, and control anirnals 21 days DostsurLrerv (42 davs old) Variable' A B C A:B A:C B:C Contras? Exp-Control Sham-Control Exp-Sham Exp-Control Sham-Control Exp-Sham Exp-Control Sham-Control Exp-Sham Exp-Control Sham-Control Exp -Sh a m Exp-Control Sham-Control Exp-Sham Exp-Control Sham-Control Exp-Sham 1ONTRol C B Fig. 6. Length measurements A, B, and C for animals sacrificed 21 days postsurgery (42 days old). Vertical scale is in millimeters. Dots represent means, bars represent standard deviations and dotted lines represent SilN Sum of squares F value P > F* 4.90671 2.65219 0.24083 1.33203 1.66965 0.01333 1.72811 1.44586 0.00880 0.00169 0.00004 0.00086 0.00122 0.00034 0.00019 0.00000 0.00013 0.00007 457.82 247.46 22.47 65.09 81.59 0.65 326.00 272.76 1.66 17.20 0.39 8.69 41.31 11.47 6.47 0.04 2.73 1.47 0.0001 0.0001 0.0001 0.0001 0.0001 0.4222 0.0001 0.0001 0.2017 0.0001 0.5340 0.0043 0.0001 0.0012 0.0131 0.8375 0.1027 0.2299 'Degree of freedom for each variable = 1. 'Contrast statements comparing two groups at a time: Exp-Control = experimental vs. Fontrol; Sham-Control = sham vs. control; Exp-Sham = experimental vs. sham. Probability of obtaining the F value by chance. 346 J.B. McLAIN AND P.S. VIG 1.15 1.10 1.05 R w m EXP AIB L EXP CONTROL SHAM B/C R L SHAM R L CUNTROl A/C Fig. 7. Ratio measurements A:B, B:C and A:C for animals sacrificed 21 days postsurgery (42 days old). ,SO ,401 I 30 ,20#lo- ._._. :.:. ...... ........ ..... ..... ..... ....... ... 0- . . ..... ......... ..... -.20- -,30** -.SO J Y 4.1 A B 14 DAYS *+ ** * C A B 21 DAYS C Fig. 8. Mean differences, right minus left femurs for measurements A, B, and C in experimental (EXP),sham and control groups, sacrificed at 14 and 21 days postsurery. **, P = 0.05; t i ,experimental and sham groups are significantly different from one another when double asterisks are outside the bracket. t , Double asterisks indicate a significant difference between the starred group and the control group. weeks postsurgery), while the effect changed from stimulatory to inhibitory a t 3 weeks postsurgery (presumably following the complete healing and reattachment of the periosteum). At 2 weeks, lengths B and C showed relative increases in the experimental limbs. This is especially so when compared to the sham group and is also supported by viewing the distribution of intraanimal limb variation in unoperated controls a t 3 weeks. One may hypothesize that a relationship exists between the functional demands on particular segments of the epiphyseal plates and the effects observed when the periosteum is cut. Bylaner et al. (1981) demonstrated a differential effect of Blount stapling of human femur epiphyses with a greater growth-arresting effect medially than laterally. It is interesting to note in this study the decrease in length A in the experimental group a t 2 weeks. This measure involves the head of the femur-an area which one would suspect is always under compressive loading force during function in the rat. Measurement C involves the greater trochanter of the TRANSVERSE PERIOSTEAL SECTIONING femur, an area associated with the insertions of the Gluteus medius, Gluteus profundus, and Piriformis muscles. These muscles as well as those inserting a t the lesser trochanter and intratrochanteric notch are aligned in such a fashion a s to produce a natural tension on the covering periosteum. Measurement B could be considered as middle ground in terms of tension and compression between measures A and C. Consistent with these anatomical observations, length C appears more susceptible to stimulation with periosteal section a t 2 weeks than does length B, and length A shows a decrease in length greater than normal intralimb variation. This decrease in length A would be attributed to the surgical procedure itself and a decrease in the normal function of the bone. While the differences exhibited in this study were significant, the magnitudes were small. This is in contrast to Crilly’s (1972) work using chick radii, where differences as great as +27.0% for the experimental side were observed. At 2 weeks the largest differences in this study were about + 1.0% for the experimental side. Two explanations can be offered to account for these differences. First, Crilly compared experimental to sham limbs, both having had surgery to insert bone markers; thus he failed to account for the inhibitory effect of surgery, muscle reflection, and scarring in his “control” limbs. Second, it can be hypothesized that the functional aspects of the bones involved varies greatly between the two studies. The normal compressive loading of the bones during function may be quite different between the chick radii and rat femur, thus providing another covariable that is known to be of importance in the regulation of bone growth. While only two points in time were examined in this study, what was observed here was consistent with other investigations indicating a peak of effect a t the time of periosteal reattachment and a decrease in effect as growth continues. Local bony alterations also diminished from 2 to 3 weeks postsurgery. Better defined and larger new bone deposits were evident a t 2 weeks, while at 3 weeks remodeling had produced a recovery to essentially normal contour for most of the specimens. It is tempting to propose that at 3 weeks postsurgery not only has the stimulatory effect of periosteal sectioning ceased, but that an inhibitory effect is also emerging related perhaps to the healing and reattachment of the periosteum to the femur shaft. Moss (1978) suggests that in man collagen cross- 347 linking with age reduces periosteal elasticity to the point that chondrogenesis ceases in the epiphyseal plates. Further histological investigations are being carried out to determine which changes in the periosteum a t the site of section posthealing might account for this restrictive effect. Hellquist’s (1972) observations of thicker, more adherent periosteum in rabbit maxillae following sectioning may hold the answer. Another observation, the greater susceptibility of length A to decrease upon surgical intervention in both the sham and experimental groups, may indicate a greater dependence on normal function for the maintenance of length A (involving the head of the femur). While supporting both Lacroix (1951), Crilly (19721, and Moss (19781, who postulated that the periosteal covering of long bones exerts a restraining force on the epiphyses, this investigation also demonstrates that the functional demands of individual long bones may modify the periosteal influence and may play a more dominant role in the growth regulation of the epiphyses. These are matters for further investigation. ACKNOWLEDGMENTS This project was supported through a National Research Service Award Individual Fellowship, National Institute of Dental Research grant F32 DE05166. LITERATURE CITED Auer, J.A., R.J. Martens, and E.H. Williams (1982) Periosteal transection for correction of angular limb deformities in foals. J. Am. Vet. Med. Assoc., 181:459466. Bertrand, P., and A. Trillat (1948) Le Traitement des inegalities de longueur des membres inferieurs pendant la croissance. Rev. Chir. Orthop., 34t264-311. Brodin, H . (1955) Longitudinal bone growth: The nutrition of the epiphyseal cartilages and t h e local blood supply. An experimental study in rabbit. Acta Orthop. Scand. 20 Suppl.:1-92. Bylander, B., L. Hansson, and G. Selvik (1983) Pattern of growth retardation after blount stapling: A roentgen stereophotogrammetric analysis. J. Pediatr. Orthop., 3:63-72. Bylander, B., G. Selvik, L. Hansson, and S. Aronson (1981) A roentgen stereophotogrammetric analysis of growth arrest by stapling. J. Pediatr. Orthop., 1:81-90. Chan, K.P., and A.R. Hodgson (1970) Physiologic leg lengthening. A preliminary report. Clin. Orthop., 68:55-62. Crilly, R.G. (1972) Longitudinal overgrowth of chicken radius. J. Anat., 112:11-18. Frejka, B., and M. Fait (1958) Clinical evaluation of linear growth stimulation. In: Septieme Congress International de Chirurgie Orthopedique. Imprimerie des Sciences, Bruxelles, pp. 616-661. Hall-Craggs, E.C.B., and C.A. Lawrence (1969) The effect of epiphysial stapling on growth in length of the rabbit’s tibia and femur. J. Bone Joint Surg. (Br., 51B:359-365. 348 J.B. McLAIN AND P.S. VIG Hellquist, R. (1972) Facial skeleton growth after periosteal resection. Scand. J. Plast. Reconstruct. Surg., 10 (Suppl.):1-98. Hueter, C . (1862) Anatomische Studien a n den extremitatengelenken neugeborener und erwachsener. Virchows Arch. Pathol. Anat. Physiol. Klin. Med., 25: 572-599. Jansen, K., and A. Langenskiold (1956) Inhibition and stimulation of growth. Acta Orthop. Scand., 26:296319. Jenkins, D.H.R., D.H.F. Cheng, and A.R. Hodgson (1975) Stimulation of bone growth by periosteal stripping. J. Bone Joint Surg. (BrJ, 57B:482-484. Lacroix, P. (1951) Organization of Bones. Churchill, London. Larson, R.L., P.J. Kelly, J.M. Janes, and L.F.A. Peterson (1961) Suppression of the periosteal and nutrient blood supply of the femora of dogs. Clin. Orthop., 21:217-225. MOSS,M.L. (1978) The design of bones. In: Scientific Foundations of Orthopaedics and the Surgery of Trauma. R. Owen, J.W. Goodfellow, and P.G. Bullough, eds. W. Heinemann, London, pp. 59-66. Mueller, W. (1922) Experimentelle Untersuchungen ueber mechanisch bedingte Umbildungsprozesse am wachsenden und fertigen knochen und Ihre Bedeutung fuer di Pathologie des knochens, insbesondere die Epiphysenstoerungen bei rachitisaehnlichen Erkrankungen. Bruns’ Beitr. Klin. Chir., 127r251-290. Munro, I.R. (1978) The effect of total maxillary advancement on facial growth. Plast. Reconstruct. Surg., 62751-762. Pease, C.N. (1952) Local stimulation of growth of long , bones. A preliminary report. J. Bone Joint Surg. (Am., 34A:1-24. Roskjaer, M. (1977) Sutures and periosteum of growing intramembranous bone. Ph.D. Thesis. University of Nijmegen, Nijmegen. Shui, M.H., and W.T. Wong (1964) Perinsteal stripping in the stimulation of long bone growth in rabbits. J. West. Pacific Ortbop. Assoc., 4t421. Sola, C.K., F.S. Silberman, and R.L. Cabrini (1963) Stimulation of the longitudinal growth of long bones by periosteal stripping. J. Bone Joint Surg. (Am.), 4.5A: 1679-1684. Strobino, L.J., G.O. French, and P.C. Cnlonna (1952) The effect of increasing tensions on the growth of epiphyseal bone. Surg. Gynecol. Obstet., 95694-700. Taylor, J.F., and E. Warrell (1977) The effect of local trauma on tibia1 growth. J. Bone Joint Surg. (Br.), 59B:503. Volkmann, R. von (1862) Chirurgische Erfahnungen uber Knockenverbiegungen und Knockenwachstum. Virchow Arch. Pathol. Anat., 24t512-540. Whiteside, L.A., and P.A. Lesker (1978) The effects of extraperiosteal and subperiosteal dissection. I. On blood flow in muscle. J. Bone Joint Surg. (Am.), 60A.23-26. Wu, Y.K., and L.J. Miltner (1937) A procedure for stimulation of longitudinal growth of bone. J. Bone Joint Surg. (Am.), 19t909-921. Yabsely, H.H., and W.R. Harris (1965) The effect of shaft fractures and periosteal stripping on the vascular supply to the epiphyseal plates. J. Bone Joint Surg. (Am.), 47At551-566.