AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 119:156 –174 (2002) Asymmetrical Spondylolysis Charles F. Merbs* Department of Anthropology, Arizona State University, Tempe, Arizona 85287-2402 KEY WORDS unilateral spondylolysis; incomplete spondylolysis; fatigue fracture; interarticularis; pedicle; spina bifida ABSTRACT The objective of this study was to examine examples of spondylolysis in which the pattern of separation was clearly asymmetrical, in order to learn more about the process of bone separation that produces this condition. Although the primary focus was on unilateral complete separation, examples of asymmetry represented by incomplete separation and by complete bilateral separation where the separation sites are in different locations on the two sides were included. Two collections were used, one consisting of Canadian Inuit skeletons curated at the Canadian Museum of Civilization, and the other of 48 individual examples of asymmetrical spondylolysis from sites in a variety of localities curated by several different institutions. The first collection was studied primarily to observe early manifestations of spondylolysis, particularly incomplete separation, while various patterns of asymmetrical complete separation were the focus of the second. The results indicate that asymme- try is part of the earliest osteological picture of spondylolysis, with right-sidedness predominating, a condition perhaps related in some way to handedness. The right-side predominance appears to decrease with age. The ratio of unilateral to bilateral separation ranges from 3–33% in different studies, and a significant number of the unilateral separations have spina bifida occurring in the same vertebra. Overall, the specimens examined here, considered along with clinical cases, nicely illustrate a progression of spondylolysis. A unilateral separation may heal, it may progress to bilateral separation, or it may remain as a permanent condition, producing a pattern of degenerative changes that can include spondylolisthesis. A unilateral healing of bilateral complete separation is likely a rare phenomenon, at least after the separations have reached a certain level of maturity. Am J Phys Anthropol 119:156 –174, 2002. © 2002 Wiley-Liss, Inc. Spondylolysis in its classic form is a fracture of the lower lumbar region related to erect posture (Neugebauer, 1885; Poirier, 1911; Thieme, 1950; Merbs, 1996a). As such, it is a uniquely hominid condition. It has also been observed that spondylolysis occurs with higher frequency in individuals engaged in certain activities that appear to put unusual stress on the lower back (Merbs, 1996a,b). The classic presentation of spondylolysis is complete separation through the isthmus between the superior and inferior articular processes (interarticularis) on both sides of the arch of a lower lumbar vertebra (Merbs, 1996a). The result of this symmetrical fracturing is a “separate neural arch” and a potential for anterior slippage (spondylolisthesis) above the point of defect. This is consistent with the stresses of erect posture being transmitted equally to the two sides of the neural arch. Spondylolysis has other presentations, however, particularly with respect to vertebra affected, location of defect, degree of separation, and unilateral involvement. Many of these include some expression of asymmetry, a situation more likely to reflect specific activity patterns than simple erect posture. Asymmetrical spondylolysis has been given little attention in anthropological studies, usually being relegated to passing mention or descriptions of individual specimens. With the exception of a study by Waldron (1992), unilateral spondylolysis is simply included among the totals obtained in specific studies. Miles (1975) noted, however, that unilateral spondylolysis could be associated with spondylolisthesis, despite the fact that the arch is not completely separated. In his study of Alaskan Eskimos and Aleuts, Stewart (1953) recorded examples of spondylolysis in which the separation took place through a pedicle rather than interarticularis, noting that such separations were nearly always paired with separation through the isthmus on the opposite side, thus making them asymmetrical. It was noted in a study of Canadian Inuit skeletons (Merbs, 1995) that complete separation on one side could be paired with incomplete separation on the opposite side, and that incomplete separation could begin from above downward on one side of an arch, and from below upward on the opposite side. It thus seems likely © 2002 WILEY-LISS, INC. *Correspondence to: Charles F. Merbs, Department of Anthropology, Arizona State University, Tempe, AZ 85287-2402. E-mail: email@example.com Received 8 October 2001; accepted 5 February 2002. DOI 10.1002/ajpa.10100 Published online in Wiley InterScience (www.interscience.wiley. com). ASYMMETRICAL SPONDYLOLYSIS that asymmetry can be a property of spondylolysis from its earliest expression, and that in some cases it can remain as a permanent condition in adults, even contributing to olisthesis. The objective of this study was to examine asymmetrical expressions of spondylolysis to better understand the morphology and dynamics of this condition. Of particular interest was unilateral spondylolysis, and its development and progression. Questions considered included: 1) to what extent does unilateral separation represent an early stage of bilateral separation, 2) how does it become established as a permanent condition, and 3) is it ever the result of bilateral separation where one side has healed? Excluded from the objectives were any attempts to identify the kinds of activities or habitual behaviors that could have produced the asymmetries observed, except for noting the predominance of right-handedness in humans generally. The stimulus for the study came from an initial finding in a study of Canadian Inuit skeletons (Merbs, 1983, 1995, 1996b) that classic spondylolysis (bilateral complete separation through interarticularis) accounted for barely more than half of all affected intact vertebrae, and that asymmetry of some type was obvious even in some of these. The focus was thus directed toward discovering the significance of this asymmetry in the Inuit and what it could contribute to a greater understanding of spondylolysis in general. Also included as a separate series are additional examples of asymmetrical spondylolysis that came to the author’s attention during other studies, or were brought to his attention by other researchers. PREVIOUS STUDIES Archaeological skeleton collections In a study of 786 Aleut and Eskimo skeletons from Alaska, Stewart (1953) found 259 (33%) vertebrae with spondylolysis, 25 (9.7%) of these showing unilateral separation. Nine (36%) of the 25 unilateral cases included spina bifida, compared with just 5 of 234 (2.1%) with bilateral separation (from Stewart, 1953, Table IV, p. 941). Stage of development of the individual does not appear to be a factor in the occurrence of spina bifida here, since only two individuals, both with bilateral separation, were less than 18 years of age. The distribution of unilateral cases without spina bifida is L2 ⫽ 1, L4 ⫽ 3, and L5 ⫽ 12, while all nine with spina bifida were L5. Among the bilateral cases were 15 which combined complete separation through a pedicle (anterior or posterior to the transverse process) with complete separation through interarticularis on the opposite side, thus producing an extremely asymmetrical separate neural arch. Even the one example in Stewart (1953) of separation through a pedicle on both sides (p. 944, his Fig. 6) is very unusual, resulting in a markedly asymmetrical arch. 157 Gunness-Hey (1982) found unilateral separation in 3 of 45 (6.7%) vertebrae with spondylolysis from Kodiak Island, Alaska, and Simper (1986) found a higher frequency of unilateral cases (4 of 34; 11.8%) in Greenland Eskimo skeletons. Two of the Greenland cases (L1 and L3) included spina bifida. The study by Snow (1948) of Archaic Indian skeletons from Indian Knoll, Kentucky, produced only two examples of unilateral involvement among 62 (3.2%) vertebrae with spondylolysis. In both cases, the defect was on the left side in L5. Stewart (1979) found 47 vertebrae with spondylolysis in a study of Arikara Indian skeletons from South Dakota, nine (19.1%) of them unilateral. All vertebrae involved were L5, with the defect on the left side in 3 and the right side in 6. Among European studies, Vyhnánek (1989) found 42 cases of spondylolysis in 227 Slavic-Avaric skeletons dating to the 7th– 8th centuries in Slovakia and 49 in 803 ancient Slavic skeletons from Slovakia and the Czech Republic. The percentage of cases that are unilateral is given only for males: 7.1% for Slavic-Avaric, and 15.3% for ancient Slavic. All of the unilateral cases involved interarticularis, most of the affected vertebrae were L5, and no significant side difference was reported. Waldron (1992) found 5 examples of unilateral involvement among 54 (9.3%) spondylolytic vertebrae from an Anglo-Saxon and several Medieval sites in England. Three were male and two female, all with ages in the 15–35-year range. The left side was involved three times (all L5), and the right twice (L5 and L2). Gaballah and Badawy (1980) found 5 examples of unilateral separation among 17 (29.4%) spondylolytic vertebrae, and one separated arch in 90 ancient Egyptian skeletons. The specific lumbar vertebrae were not always identifiable, but unilateral lysis was combined with spina bifida in both cases of L5 involvement. Most archaeological series contain too few examples of unilateral spondylolysis to provide useful frequency values, but sometimes descriptions of individual cases can be informative. An example is illustrated by Miles (1975, p. 33) from Mug House (Mesa Verde, CO), involving a female estimated to have been approximately 29 years of age when she died. Pictured is an L5 with complete separation through interarticularis on the right side and a much expanded intact interarticularis on the left side. A greatly enlarged inferior zygapophysial facet on the left side indicates that the L5-S1 joint on this side had undergone considerable degenerative change, and first-degree olisthesis (less than 1/3 the anteroposterior diameter of the body) relative to the sacrum was reported (Miles, 1975, p. 31). Although the author felt that spondylolysis had originally been bilateral with the left side eventually healing, this is not supported by the published radiograph (Miles, 1975, p. 33). However, the vertebra does have the appearance of a long-standing unilateral separation, resulting eventually in severe osteoarthritis and olisthesis. 158 C.F. MERBS Vyhnánek (1989, p. 76) illustrates an unusual case of unilateral spondylolysis affecting L5 in an adult male from the 11th century Slavic site of Abraham. The separation is through an extremely narrow interarticularis which curves inward to constrict the neural canal (stenosis), while the right isthmus, although unusually broad, is more normal in appearance. The transverse process on the left side is short and tapering, while that on the right side is longer and blunt. Also, the left pedicle contains what the author calls a retrosomatic foramen. Part of the remarkable asymmetry of this vertebra may be due to its possessing a transitional morphology, essentially lumbar on the left side, but with some sacral features on the right. If the affected vertebra is correctly identified as L5, it means that it is the 24th vertebra below the skull and thus an example of partial cranial shift at the lumbosacral border. How much of the asymmetry can be attributed to the border-shifting and how much to the unilateral nature of the spondylolysis is then difficult to say. Snow (1948, p. 522) illustrates a bilateral separation from Indian Knoll which appears to have existed for a time as a unilateral separation, based on the markedly different appearance of interarticularis on the two sides. The isthmus on the left side appears relatively normal in width, with the separated edges well-corticated and only slightly irregular, while that on the right side is twice as wide, poorly corticated, and extremely irregular. This suggests that separation occurred first on the left side, with the separation edges undergoing considerable cortication, but no significant remodeling, before separation developed on the right side. The great thickening of the right isthmus, due apparently to callus development, and the incomplete cortication of the separation edges on this side, suggest that attempted repair was taking place when death intervened. The right transverse process appears to contain a foramen similar to that reported by Vyhnánek (1989, p. 76), but it could be just an artifact. In what appears to be the only study dealing specifically with unilateral spondylolysis in archaeological skeletons, albeit just five cases, Waldron (1992) made the following observations with respect to the side showing the spondylolysis: 1) the lamina may be underdeveloped, 2) the body may show asymmetrical posterior wedging, 3) the spinous process may be deviated toward the opposite side, and 4) radiological evidence of sclerosis may be observable on the opposite side. He concluded, however, by saying that “these findings do not invariably accompany the condition” (p. 180). Anatomical skeleton collections Willis (1923) found 8 of 31 (25.8%) spondylolytic vertebrae to have unilateral separation in a study of 748 skeletons in the Todd Collection at the Smithsonian Institution. All involved L5, and in each case the lysis was on the right side, in 2 of the 8 (12.5%) in combination with spina bifida. Increasing his sample to 1,520 individuals, Willis (as reported by Stewart, 1931) found 20 of 80 (25%) spondylolytic vertebrae to have unilateral involvement, but it is not clear how many of these included spina bifida. The side distribution for the lysis was 17 (85%) right and 3 (15%) left, and the vertebrae involved were L1 ⫽ 1, L3 ⫽ 2, L5 ⫽ 15, and L6 ⫽ 2. By adding data from the Terry Collection at the Cleveland Museum of Natural History to those previously published by Willis (1923), Roche and Rowe (1951) were able to report on the vertebrae of 4,200 skeletons representing European-Americans and African-Americans. In this expanded series, unilateral separations accounted for 32 of 183 (17.5%) spondylolytic vertebrae, with right-side separations occurring more frequently than left (22 to 10). The vertebrae affected were L2 ⫽ 3, L3 ⫽ 1, L4 ⫽ 1, and L5 ⫽ 27. The L5 cases probably included two previously reported by Stewart (1931) as L6. Six of the 32 (18.8%) unilateral cases included spina bifida, compared with just 3 of 151 (2%) with bilateral separation. All nine spina bifida cases involved L5. The authors noted that “in almost all unilateral separations encountered the arch was asymmetrical, apparently because of the difference in length between the two isthmi. Hence those with wide separations were more asymmetrical than those with narrow separations” (Roche and Rowe, 1952, p. 491). Roche and Rowe (1952, p. 492) further noted that “no significant difference could be demonstrated between the ages of those with bilateral separations and those with unilateral separations.” In a report on South African Bantu skeletons in what is now known as the Raymond Dart Collection at the University of the Witwatersrand, South Africa, Shore (1929) described a 45-year-old Xosa male with unilateral (right) spondylolysis of L4. He noted the presence of a rough line on the bone surface in the position of the defect on the opposite side, marked asymmetry with the left lamina being much thicker and longer than the right, and slight deflection of the spinous process to the left. The asymmetry, he concluded, demonstrates that “the intervertebral stresses were transmitted mainly by way of the left arch and pedicle” (Shore, 1929, p. 437). He also described a similar case involving the right side of L5 in a 36-year-old Shangaan male. Again the left lamina is wider and the left pedicle thicker than the right, and the spinous process is deflected to the left, but in this case the spondylolysis is bilateral. He suggested that the defect on the right side predated and probably existed for some time before the appearance of that on the left. Eisenstein (1978) found 6 unilateral examples among 18 (33.3%) spondylolytic vertebrae from 485 skeletons in the Dart Collection. The defects were equally divided between right and left sides, and all of the identified vertebrae were L4 (2/4, 50%) or L5 (4/13, 30.8%). Eisenstein (1978, p. 489) noted that “there was no reason ASYMMETRICAL SPONDYLOLYSIS to suspect that the intact opposite pars interarticularis represented a healed fracture.” Clinical studies Spondylolysis including unilateral involvement has been of considerable interest to the field of sports medicine because of its possible association with specific athletic activities. Letts et al. (1986), for example, described 10 cases of unilateral spondylolysis in “highly competitive adolescent athletes” engaged in repetitive training and competition exercises involving flexion/extension of the lumbar spine. All defects occurred at interarticularis in L4 (1 case) or L5 (9 cases). Rossi (1978) found 30 examples of unilateral separation among 225 (13.3%) vertebrae with spondylolysis in athletes, most of them male. Both studies found the lesion associated most commonly with gymnastics, but also with hockey (Letts et al., 1986) and diving (Rossi, 1978). In their study of living North Alaskan Eskimos, Kettelkamp and Wright (1971) found 8 vertebrae with unilateral separation among 45 (17.8%) with spondylolysis in 43 individuals. The units affected were L4 (2/4, 50%) and L5 (6/41, 14.6%), and all separations occurred at interarticularis. They also recorded a narrowing of interarticularis on the defective side in 5 of the 8 cases. Lowe et al. (1987) reported that seven of their patients with unilateral lysis showed changes in the opposite hemiarch, including hypertrophy and a healed defect at interarticularis, along with arch deformity and sclerosis. They also found increased scintigraphic uptake at interarticularis in some of the affected vertebrae, indicating that bone activity was taking place. Maldague and Malghem (1976) approached the problem from a different direction, looking at 50 individuals with unilateral arch hypertrophy of unknown origin. In 32 (64%) of these, the apparent cause turned out to be a separation at interarticularis on the opposite side. They also observed that a vertebra affected by spondylolysis on one side and arch hypertrophy on the other side also showed a distinctive asymmetry of the arch, which included a tilting of the spinous process. They were concerned that a defect like this could cause rotational instability. Porter and Park (1982) also observed asymmetry of the neural arch in the five cases of unilateral spondylolysis that they studied. The asymmetry does present one advantage in the clinical context, however: it can easily be detected on simple anteroposterior radiographs, thus allowing prediction of the presence of a unilateral arch defect (Maldague and Malghem, 1976; Porter and Park, 1982). Although it is clear that separation can develop in both hemiarches simultaneously (Merbs, 1995), clinical studies have also confirmed that unilateral separation can progress to bilateral separation. Hadley (1963) described a 14-year-old male with relatively mild back pain, but no sign of spondylolysis when he was first examined. Two years later the symptoms 159 had become more severe, and radiographs showed a well-established lysis on the left side of L5. The right interarticularis was “crossed by a band of decreased density” that could be related to stress fracturing. Aland et al. (1986) described a 25-year-old individual in which complete separation at interarticularis on the left side of L4 ultimately led to pedicle separation on the right. Radiographs identified spondylolysis along with sclerosis of the right pedicle, and computed tomography (CT) showed a lucency in this pedicle. A follow-up 4 months later showed massive overgrowth of bone on the pedicle, similar to the callus development seen at a fracture site, and an excisional biopsy of the pedicle did indeed reveal an established fracture and nonunion. The authors concluded that spondylolysis through the left interarticularis produced instability, leading to hypertrophy and a stress fracture of the right pedicle. Persistence of the instability resulted in nonunion of this fracture. The pain was apparently due primarily to a continued overgrowth of bone impinging on the root of the fourth lumbar nerve. Garber and Wright (1986) described a case involving a 26-year-old which appears identical to that of Aland et al. (1986), even to the vertebra and side affected. Four years later, the condition had progressed to complete bilateral separation through interarticularis. In a restudy of young athletes, Letts et al. (1986) discovered that 5 unilateral cases had healed, 3 were basically unchanged, and 2 had progressed to bilateral separation. Albers and Yochum (1980) also described a case of unilateral separation progressing to bilateral. A 21-year-old female with a 5-year history of low back pain aggravated by athletic activities was identified as having a radiolucent defect at left interarticularis identified as spondylolysis, with sclerosis and enlargement of the right pedicle. Radiographs taken 3 years later showed bilateral separation at interarticularis. The pedicle was still enlarged but not sclerotic. This case suggests that a change in stress focus had occurred on the right side, from the pedicle to interarticularis, or that interarticularis was simply the more vulnerable of the two sites. The progression from unilateral to bilateral separation is not limited to young individuals, as illustrated in the study by O’Beirne and Horgan (1988) of a 40-year-old man with a 10-year history of lower back pain who presented with complete spondylolysis of the right hemiarch. The left side was still essentially intact, but the lamina showed a thickening on either side of the defect, suggesting the existence of a stress fracture that was progressing toward nonunion. “In the presence of unilateral spondylolysis,” according to O’Beirne and Horgan (1988, p. 221), “the remainder of the neural arch is exposed to abnormal stresses, and it seems plausible to suggest that the defect in the contralateral lamina was, in fact, a stress fracture sustained secondary to spondylolysis. This view is supported by the 160 C.F. MERBS hypertrophic appearance of the bone on either side of the defect, consistent with repeated attempts at healing.” Although the time necessary for unilateral separation to produce observable changes in the opposite hemiarch is uncertain, Maldague and Malghem (1976) felt that in the cases they studied the separation was “not of recent origin.” Nevertheless, Sherman et al. (1977) found such changes even in very young individuals. Among patients with unilateral spondylolysis and “reactive sclerosis and hypertrophy” of the opposite pedicle and lamina whom they studied was one as young as 8 years of age, and 10 between 12–21 years. The vertebrae and side affected were L2 (1 left), L3 (1 left), L4 (2 left), and L5 (2 right, 5 left). All but the youngest had reported back pain. Lusins et al. (1994) used standard CT and singlephoton emission computed tomography (SPECT) to evaluate three cases of unilateral spondylolysis. In the youngest, a 16-year-old male with a 5-year history of back pain after exercise, CT showed L5 lysis on the left side and degenerative changes in the right L5-S1 zygapophysial joint, while SPECT showed increased activity in this joint. In the second, a 33-year-old male with a 10-year history of chronic back pain, CT showed L5 lysis on the left side and some degree of hypertrophy of interarticularis on the right. Although the L5-S1 zygapophysial facets appear normal, SPECT disclosed increased diffuse bone activity in this area. In the third, a 46-year-old female, CT showed L5 lysis on the left side and hypertrophy of interarticularis on the right. Although the zygapophysial joints at this level appeared normal on standard radiographs, SPECT showed increased activity anteriorly and to the right of the midline. The authors concluded from the three cases that following unilateral separation, stress will be transmitted unevenly to the pedicle, interarticularis, and articular facets on the opposite side, and that this can be confirmed by positive SPECT scans. If the process observed by Lusins et al. (1994) were to continue beyond the stage they observed, the result presumably would have been zygapophysial degeneration on the side opposite the spondylolysis, and perhaps olisthesis. Gunzburg and Wagner (1988) reported on a case of L4 right unilateral spondylolysis in a 41-year-old Turkish woman with an olisthesis of approximately 20%. The authors noted the difference between the typical olisthesis associated with complete bilateral spondylolysis and degenerative olisthesis which occurs in the absence of spondylolysis, and they decided to apply the “degenerative” label to their case. Present also were degenerative zygapophysial changes and osteophyte development on vertebral disk margins. Based on the case of a 32-year-old woman with a 12-year history of lower lumbar pain and spondylolysis on the left side of L5, Kornberg (1988) felt that unilateral spondylolysis was responsible for prema- ture degenerative changes. Besides L5-S1 zygapophysial degeneration on the right side and grade I (less than one-third the anteroposterior body diameter) olisthesis, the affected vertebra also showed spina bifida. Slippage allowed by the lysis on one side appeared to produce greater stress on the opposite intact side, leading to joint remodeling that could include the equivalent of degenerative olisthesis. A rotation of L5 to the right was also seen on CT. MATERIALS AND METHODS The data analyzed here are derived from two different sources. The first is a study of vertebral columns from Inuit archaeological sites in the Canadian Arctic which included detailed observations on spondylolysis, spondylolisthesis, and related phenomena. The second consists of individual cases of asymmetrical spondylolysis from archaeological contexts encountered by the author during other studies or brought to his attention by other researchers. The Canadian Inuit skeletons are primarily from the Nunavut Territory (formerly the eastern part of the Northwest Territories). Most are from sites in the Kivallo district west of Hudson Bay, with others from Labrador (part of the province of Newfoundland) and the Northwest Territories (District of Mackenzie). In all, 417 Inuit skeletons (children through old adults) were found to include vertebrae that could be used in the study. Sites producing the largest number of skeletons with observable vertebrae include Tunermiut (Native Point, Southampton Island), 118; Silumiut, 112; Kamarvik, 68; Saglek Bay (Labrador), 38; Kulaituijavik, 15; Inuksivik (including Ukusialik), 12; and Naujan, 9. Also included are 11 skeletons representing the Caribou Inuit who lived in the Barren Grounds west of Hudson Bay. The material ranges in time from the Thule culture period, beginning about AD 1000 in Arctic Canada, through the historic period (beginning 17th century), to the first half of the 20th century. The Sadlermiut of Southampton Island represent an isolated group that became extinct during the winter of 1902–1903. Sex and age determinations are based on standard criteria (Buikstra and Ubelaker, 1994), with the condition of vertebrae specifically eliminated as an age determinant. Sex in nearly all cases was based on shape of the pelvis. Where possible, skeletons were placed in four approximate age categories: adolescent, 13–18 years; young adult, 18 –30 years; middle adult, 30 – 45 years; and old adult, over 45 years. The second series is composed of individual examples of asymmetrical spondylolysis involving at least one complete separation. The specimens come from the states of Alaska, Arizona, California, and New Mexico, the Canadian provinces of British Columbia and Newfoundland, and the Sudan in Africa. Where direct observation of the entire skeleton was not possible, information on age and sex obtained from the primary observer was used. To this series were 161 ASYMMETRICAL SPONDYLOLYSIS TABLE 1. Canadian Inuit lumbar vertebrae with complete or incomplete spondylolysis listed by pattern of symmetry or asymmetry Vertebra affected L1 L2 L3 L4 L5 L6 S1 Totals Pattern Symmetrical Asymmetrical 1A 1B 2A 0 0 5 6 5 1 4 21 0 0 0 5 1 0 2 8 0 0 0 1 1 0 0 2 0 21 0.184 6 2 0.070 Spondylolysis type1 2B 0.018 0 2 2 4 0 1 5 2 2 16 0 16 0.140 3A 3B Totals 0 1 2 19 38 1 0 61 0 0 0 1 5 0 0 6 2 5 7 33 55 4 8 114 0.018 0.044 0.061 0.289 0.482 0.035 0.070 1.000 67 47 0.588 0.412 61 0 0.535 0 6 0.053 1 1A, incomplete separation, unilateral; 1B, incomplete separation, bilateral; 2A, complete separation on one side, incomplete separation on other side; 2B, complete separation, unilateral; 3A, complete bilateral separation at interarticularis on both sides (classic form); 3B, complete bilateral separation at interarticularis on one side, through a pedicle or lamina on the other side. added Canadian Inuit cases that met the criteria. This second series totaled 48 cases which were grouped as follows: unilateral complete separation (31 cases), complete separation combined with incomplete separation (5 cases), complete separation at interarticularis combined with complete separation through a pedicle or lamina (9 cases), and three complete separations in the same vertebra (3 cases). Data collected for vertebrae with spondylolysis included the following where possible: vertebra affected and evidence of border-shifting (based on morphology and number of units below the skull), side and location of defect, complete vs. incomplete separation, cortication and width of interarticularis at defect site, symmetry of the arch and shape of the spinous process, presence of spina bifida, degenerative changes involving disk or zygapophysial joints, and spondylolisthesis. Specimens were radiographed when necessary. RESULTS The Canadian study produced 119 lumbar vertebrae with some manifestation of spondylolysis, but five had to be eliminated from the asymmetry study because one hemiarch was missing or too damaged to make the required observations. The remaining 114 were then divided into three basic categories as follows: type 1, showing incomplete separation, but no complete separation; type 2, complete separation on just one side; and type 3, complete separation on both sides. The categories were also subdivided (Table 1). Based on whether separation was incomplete or complete, as well as the part and side affected, 69 (60.5%) of the affected vertebrae are considered symmetrical (types 1B, 3A) and 45 (39.5%) asymmetrical (1A, 2A, 2B, 3B). If the direction of incomplete separation (downward from the superior margin or upward from the inferior margin) is also considered, two of the type 1B cases (Table 2, cases 1 and 5) would also be classified as asymmetrical, increasing this category to 47 (41.2%). Although spondylolysis may occur in older individuals, it seems more a condition of youth, and for this reason incomplete spondylolysis (types 1A, 1B, and 2A) in adolescents and young adults was chosen as the starting point for studying early stages of asymmetry. A previous study (Merbs, 1995) found at least one vertebra with incomplete separation in 18 Canadian Inuit skeletons, 16 of them adolescent or young adult. These 16 became a focus of the present study. Expansion of the series to include Canadian Inuit skeletons from other sites produced no additional examples of incomplete separation in the relevant age group. Among the 16 skeletons are 31 vertebrae with spondylolysis (Table 2). For study purposes, however, 7 affected vertebrae were eliminated, 5 because the only separations present are complete (case 4, L4; case 5, L5 and S1; case 7, L5; case 11, L5), and 2 (case 8, L5; case 10, L5) because one side is missing, thus making it impossible to determine if the separation was asymmetrical. That brought the final total of vertebrae in this series with asymmetrical separation to 24. Thirty-one incomplete separation sites are represented in these 24 vertebrae, 16 of them proceeding downward from the superior margin, 12 proceeding upward from the inferior margin, and 3 proceeding from both margins simultaneously (Fig. 1). The ratio of downward to upward separations (counting the last three as both) varies by level, with upward occurring more frequently only at L3 (3:2). The two occur equally at L4 (8:8) and L5 (4:4), and all five separations at S1 are downward. Incomplete separations occur more frequently on the right side than the left (19:11), and this is true at every level: L3 ⫽ 3:1; L4 ⫽ 9:5; L5 ⫽ 4:3; and S1 ⫽ 3:2. In case 4 (Table 2), an incomplete separation on the left side of L5 is combined with complete separation on the right, thus adding L5 to the list of vertebrae showing greater expression on the right side and changing the ratio to 20:10. In both cases where the fracture lines proceed in opposite directions on the two sides (cases 1 and 5), it is downward on the left side and upward on the right side (Fig. 2) (Table 2). 162 C.F. MERBS TABLE 2. Incomplete spondylolysis in adolescent and young adult Canadian Inuit1 Spondylolysis Case no. Burial no. Sex 1. 2. 3. 4. SIL-111 KU-24A SIL-65 SIL-43 Male ? Female Female 13–17 13–17 13–17 18–20 5. SIL-17 Male 18–20 years 6. SIL-105 Male 18–20 years 7. KA-96 Male 18–20 years 8. SIL-33 Female 21–30 years 9. SIL-112 Male 21–30 years 10. SIL-131 Male 21–30 years 11. SIL-142 Male 21–30 years 12. 13. 14. SIL-169 SIL-170 SIL-171 Male Male Male 21–30 years 21–30 years 21–30 years 15. KA-121 Male 21–30 years 16. SN-6 Male 21–30 years Age years years years years Vertebra Left Right L5 L3 L4 L3 L4 L5 S1 L4 L5 S1 L4 S1 L3 L4 L5 S1 L4 L5 L3 L4 L5 S1 L4 L5 L4 L5 L4 L5 L4 L5 L5 IS-IA NS IS-IA Œ NS NS IS-IA Œ NS IS-IA CS-IA CS-IA NS IS-IA NS NS CS-IA NS IS-IA CS-IA IS-IA Œ IS-IA X IS-IA NS CS-IA IS-IA Œ IS-IA NS NS NS NS NS IS-IA Œ IS-IA Œ IS-IA Œ IS-IA CS-IA CS-IA IS-IA IS-PD Œ CS-IA NS IS-IA Œ IS-IA IS-IA Œ IS-IA CS-IA IS-LM IS-IA X NS IS-IA IS-IA NS IS-IA Œ NS NS NS IS-IA Œ IS-IA IS-IA Œ IS-IA Œ IS-IA Œ 1 Case numbers correspond to those in Merbs (1995, p. 2329). Sites: KA, Kamarvik; KU, Kulaituijavik; SIL, Silumiut; SN, Silumiut North. Defect location: IA, pars interarticularis; PD, pedicle; LM, lamina. NS, no separation (neural arch intact); IS, incomplete separation; , defect begins on superior margin and proceeds downward; Œ, defect begins on inferior margin and proceeds upward; CS, complete separation; X, missing or damaged (no observation possible). Fig. 1. Distribution by vertebra and side of incomplete and complete spondylolysis in adolescent and young adult Canadian Inuit. Although the incomplete separations provide interesting information on fatigue fracture activity that was taking place at time of death, 57.7% (41/71 separation sites) of the spondylolysis observed in these adolescents and young adults was already complete (Fig. 1). The greatest contributor of complete separations is L5 (28), followed by L4 and S1 (5 each), and L3 (2 cases). This suggests that most spondylolysis begins at L5, then proceeds upward and downward from that level. Cases of complete bilateral separation among the Canadian Inuit (Table 1, type 3A) in general provide little information about the relative timing of the separations on the two sides. In nearly all, the parts involved had undergone considerable resorption and remodeling, which obscured any evidence that one side separated significantly ahead of the other. Among the 13 young adults in this category, more give the impression of having separated first on the right side, but the evidence is inconclusive. A clearer example is a case of L5 bilateral separation in a young adult male. Unfortunately, the morphology of the superior aspects of the separated edges had been erased through contact with acidic soil that formed in the grave (Merbs, 1997), but the inferior aspects (arch portion) are nicely preserved and show complete cortication. Nevertheless, the right side displays noticeably more remodeling than the left, suggesting that separation occurred first on the right side. Based on a matching of osteophytes and other degenerative changes, eight clear examples of olisthesis following bilateral complete spondylolysis can be identified in the Canadian Inuit series, with slippage ranging from 3–14 mm. The L4/5 level is affected three times, L5/S1 four times, and L5/L6 once. The last, involving L5 and a lumbarized sacral vertebra, shows anterior slippage of 37% relative to the diameter of the vertebral bodies (Merbs, 1983, p. ASYMMETRICAL SPONDYLOLYSIS 163 Fig. 3. Anterior view of fifth lumbar vertebra and sacrum from Point Hope, Alaska, showing asymmetrical spondylolisthesis. Fig. 2. Case 1, Table 2. Bilateral incomplete spondylolysis of fifth lumbar vertebra. Top: Dorsal view shows separation extending laterally from superior margin of left interarticularis. Bottom: Inferior view shows separation on inferior margin of right interarticularis. 125, his Fig. 63B). Since olisthesis associated with separate neural arch can be identified only by degenerative changes that take time to develop, and then only under ideal conditions, it is likely that other cases went unidentified, especially in younger individuals. The Canadian series contains no obvious examples of asymmetrical olisthesis, but one was observed involving L5/S1 in an old adult female from Point Hope, Alaska (Fig. 3). Although the amount of slippage appeared to be approximately equal on both sides, L5 was tilted sharply downward on the left side and large osteophytes had developed. The opposing body surfaces were actually eburnated on this side, indicating the absence of any disk material. This is indicative of unequal pressure on the two sides, that on the left being much greater than that on the right. In the next phase of the analysis, Canadian Inuit cases of spondylolysis types 2A, 2B, and 3B were added to similar cases obtained from outside this area. Excluded were the six 3B cases involving L1 or L2. Although the mechanical forces and their effects are similar in these two vertebrae to those seen in the lower lumbar vertebrae, the etiology of the condition may be quite different. The defect in each case presents as typical complete separation through interarticularis on one side or the other, but is very atypical in that the margins of the defect are faceted, giving the appearance of a joint in a situation where no movement is possible. The etiology of this situation is not clear, but an error in fetal development was suggested (Miki et al., 1991). The addition of the Canadian Inuit cases involving L3–L6 brings the total number of specimens in this series to 48 (Table 4). Included are 31 vertebrae with complete separation on one side and the other side intact (series I, numbers 1–31), 5 vertebrae with complete separation on one side and incomplete separation on the other side (series II, numbers 32–36), 9 vertebrae with complete separation through interarticularis on one side and complete separation through some other part of the arch on the other side (series III, numbers 37– 45), and 3 vertebrae with two complete separations on one side and one on the other side (series IV, numbers 46 – 48). In only one case are two vertebrae (numbers 4 and 33) from the same individual. The 48 affected vertebrae are distributed as follows: L3 ⫽ 2, L4 ⫽ 5, L5 ⫽ 37, and L6 ⫽ 4. Males outnumber females 28 to 19 (sex not determined for one). Involved are 68 separation sites located in the following parts: interarticularis ⫽ 51, pedicle ⫽ 11, lamina ⫽ 4, superior articular process ⫽ 1, and inferior articular process ⫽ 1. Separation is complete at 63 sites and incomplete at 5. In series I and II, complete separation occurs more often on the right side (20:16). There is a marked age difference, however, with right separations predominating (13:6) in younger individuals (numbers 1–16, 32–34), and left separations (7:10) in older individuals (17–31, 35–36). In 11 cases of unilateral separation (series I and II), the laminae of the affected vertebra had not fused (spina bifida), resulting in a separated hemiarch. The separated hemiarch was missing for eight of these, presumably not recovered, not identified as human, or lost during many decades of curation. In three cases of unilateral separation, the opposite side had broken postmortem, and only one of these arches was available 164 C.F. MERBS TABLE 3. Summary of vertebrae affected in adolescent and young adult Canadian Inuit skeletons with incomplete spondylolysis Vertebrae affected Incomplete ⫹ no separation Incomplete ⫹ complete separation Incomplete ⫺ direction difference Incomplete ⫺ symmetrical Complete separation - unilateral Complete separation - bilateral One side damaged or missing Totals L3 L4 L5 S1 Totals 4 6 4 1 1 3 17 1 3 3 3 2 2 31 2 2 1 4 for study. Although spina bifida cannot be ruled out for the two missing arches, its occurrence seems unlikely. In all cases of missing arches or hemiarches, the presence of zygapophysial facets on the vertebra below indicated that they once did exist. Based on information limited to recovered parts, spina bifida occurred in 33.3% (11/33) of vertebrae with unilateral spondylolysis (series I and II), but in 47.1% of the younger age group (8/17) compared with 18.8% (3/16) in the older group. Vertebrae with unilateral separation may also show evidence of spondylolisthesis. This should be relatively easy to identify in series I and II individuals, but observations are incomplete because of poor preservation or missing adjacent vertebrae. It was observed in just 1 of 18 (5.6%) younger individuals and 6 of 17 (35.3%) older individuals, with the amount of slippage ranging from 3– 8 mm (Fig. 4). The spondylolisthesis associated with unilateral separation closely resembles that seen in degenerative spondylolisthesis (Merbs, 2001). In both cases, the arch remains attached to the body, unlike in classic spondylolisthesis following bilateral spondylolysis, where the arch becomes separated from the body. As in degenerative olisthesis, the amount of slippage associated with unilateral separation appears to be closely related to the severity of the osteoarthritis present. Unlike in degenerative olisthesis, however, where both sides are affected, it is only the side opposite the separation that exhibits severe osteoarthritis in unilateral spondylolysis. As indicated by the grooving and eburnation seen in severe cases of degenerative olisthesis, considerable linear vertical movement is taking place, but little if any lateral movement. With time, this movement and the pathological alteration of the joints that it produces can lead to a forward migration of the joints (olisthesis). Throughout the process, however, the articular facets retain the curvilinear shape (when viewed from above) normally associated with joints at this level in the column. Because the condition is relatively stable with limited anterior movement other than that associated with the olisthesis, the amount of slippage occurring is relatively easy to measure (Merbs, 2001). Based on the cases of long-standing unilateral separation studied here, joint movement at the affected level appears to be much less stable. The normally curvilinear zygapophysial facets become 11 1 2 2 11 1 1 5 flattened, thus allowing more lateral movement. The shape assumed by the facets also indicates that vertical movement at the joints is not entirely linear, as in degenerative olisthesis, but is combined with anterior movement. The inferior facets actually become convex when viewed from the side, reflecting the superior anterior arc of movement taking place. Because of the anterior movement taking place in unilateral separation which is not directly associated with olisthesis, the amount of slippage occurring is more difficult to measure, depending to a greater extent than in degenerative olisthesis on how the vertebrae are positioned when the measurement is taken. It seems likely that the amount of slippage observed in living patients with respect to unilateral olisthesis would also vary somewhat depending on the positioning of the patient. The basic asymmetry of long-standing unilateral spondylolysis may also be reflected in a wedging of the body of the affected vertebra, with the intact side reduced in height. In case 24 (Table 4), for example, the body measures 26 mm in height on the side with the separation, compared with 21 mm on the intact side. Osteophyte formation indicative of disk degeneration is also more prevalent on the intact side. All four cases of L6 spondylolysis (cases 6, 18, 24, and 30) involved vertebrae that showed partial caudal shift, with essentially lumbar morphology on one side and sacral morphology on the other (Fig. 5). The transitional vertebra was separate from the sacrum and appeared relatively mobile. In each case the separation occurred on the “lumbar side,” even in the two cases where the hemiarches had not united dorsally (Merbs, 1983, p. 126, Fig. 64). The morphology of separation involving interarticularis in series I cases (Table 4) ranges broadly in length, regularity, and cortication of the edges. The appearance at one end of the range is that of a fresh fracture with rough edges, minute projections on one edge usually matching quite well with indentations on the other edge. The edges are in contact over their entire length, and the fit is often too tight to observe what might have been taking place along this line. The defective interarticularis appears broader than the intact one because of the addition of reactive bone at the ends of the fracture line. Although these early-stage cases are usually associated with younger individuals, they may also appear in old 165 ASYMMETRICAL SPONDYLOLYSIS 1 TABLE 4. Lumbar vertebrae with asymmetrical patterns of spondylolysis Spondylolysis Location Age Sex Vertebra Left Right Series I, unilateral complete separation 1. Pueblo Bonito, NM Child ? L5 CS-IA NS 2. Native Point, NT2 Adolescent Female L5 NS CS-IA 3. Amchitka Island, AK Young adult Male L5 CS-IA NS 4. Silumiut, NT2 Young adult Female L4 NS CS-IA 2 5. Native Point, NT Young adult Female L6 NS CS-IA 6. Kagamil Island, AK Young adult Female L5 NS CS-IA 7. Unalaska Island, AK Young adult Male L3 NS CS-IA 8. Grasshopper, AZ Young adult Male L5 CS-IA NS 9. Paako, NM Young adult Female L5 CS-IA NS 10. Pottery Mound, NM Young adult Male L5 NS CS-IA 11. Puye, NM Young adult Male L5 NS CS-IA 12. Sapawe, NM Young adult Male L5 NS CS-IA 13. Silumiut, NT2 Young adult Male L5 CS-IA NS 14. Grasshopper, AZ Young adult Female L5 NS CS-IA 15. Old Walpi, AZ Young adult Male L6 NS CS-IA 16. Semna South, Sudan Young adult Male L5 CS-IA NS 17. St. Lawrence Island, AK3 Middle adult Male L5 NS CS-IA 18. Abiquiu, NM Middle adult Female L5 NS CS-IA 19. Kuskokwim, AK3 Middle adult Male L4 CS-IA NS 2 20. Mackenzie, NWT Middle adult Female L5 CS-IA NS 21. Gran Quivira, NM Middle adult Male L6 CS-IA NS 22. Kamarvik, NT2 Middle adult Male L5 CS-IA NS 23. Pueblo Bonito, NM Middle adult Female L5 CS-IA NS 24. Semna South, Sudan Middle adult Male L5 NS CS-IA 25. Native Point, NT2 Middle adult Female L6 CS-IA NS 2 26. Silumiut, NT Middle adult Male L5 CS-IA NS 27. Puye, NM Middle adult Male L5 NS CS-IA 28. Port au Choix, NFL Old adult Female L5 NS CS-IA 29. Navajo Reservoir, NM Old adult Male L5 NS CS-IA 30. Hawikuh, NM Old adult Female L5 CS-IA NS 31. Pastolik, AK3 Old adult Male L5 CS-LM NS Series II, complete separation ⫹ incomplete separation 32. Silumiut, NT2 Young adult Female L5 IS-IA CS-IA 33. Pecos Mission, NM Young adult Female L5 IS-IA CS-IA 34. Ridley Island, BC Young adult Female L5 IS-PD CS-PD 3 35. Seward Peninsula, AK Middle adult Male L3 CS-IA IS-PD 1 36. Native Point, NT Old adult Male L4 IS-IA CS-IA Series III, complete separation/interarticularis ⫹ complete separation pedicle or lamina 37. Amchitka Island, AK3 Adolescent Male (?) L5 CS-PD CS-IA 38. Umnak Island, AK Young adult Male L4 CS-IA CS-PD 39. Inuksivik, NT2 Young adult Male L5 CS-LM CS-IA 2 40. Silumiut NT Middle adult Male L4 CS-IA CS-PD 41. Kamarvik, NT2 Middle adult Male L5 CS-IA CS-PD 42. Saglek, LAB Middle adult Female L5 CS-IA CS-PD 43. Mocho, CA Middle adult Female L5 CS-IA CS-IP 44. Kamarvik, NT2 Old adult Male L5 CS-PD CS-IA 45. Umnak Island, AK Old Adult Female L5 CS-PD CS-IA Series IV, one complete separation ⫹ two complete separations 46. Pastolik, AK Young adult Male L5 CS-IA CS-IA⫹CS-LM 47. Native Point, NT2 Young adult Male L5 CS-IA CS-IA⫹CS-LM 3 48. St. Lawrence Is., AK Middle adult Female L5 CSCS-PD SP⫹CS-IA Spinabifida Olisthesis ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫹ ? ? ⫹ ⫺ ⫹ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ? ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ 3 mm ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ 7 mm 3 mm 8 mm 3 mm ⫺ ⫺ 8 mm 6 mm ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ? ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ? ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ? ? ⫺ 11 mm ? 7 mm ⫺ ⫺ ⫺ 1 AK, Alaska; AZ, Arizona; BC, British Columbia, Canada; CA, California; LAB, Labrador, Newfoundland, Canada; NFLD, Newfoundland, Canada; NM, New Mexico; NT, Nunavut Territory, Canada; NWT, Northwest Territories, Canada; CS, complete separation; IS, incomplete separation; IA, interarticularis; PD, pedicle; LM, lamina; SP, superior articular process; IP, inferior articular process. 2 Included in Canadian Inuit portion of study. 3 Illustrated in Stewart (1953, p. 943–944). adults (Fig. 6). Many of these early-stage cases probably reunite with subsequent remodeling, eventually making them difficult or impossible to detect. If the separation persists, the osseous picture can change dramatically. Through resorption and remodeling, the affected interarticularis may become very narrow and the edges of the separation smooth and well-corticated. A distinct gap may develop (Fig. 7). In sharp contrast, the intact side, particularly the pedicle, may increase markedly in size through new bone development, and the zygapophysial joint on this side develops arthritis. As the arthritis becomes more severe, the affected vertebra slips forward. Perhaps the most variable feature of the changes taking place is the appearance of the “intact” interarticularis which may undergo both fracture development or attempted repair, perhaps even in episodic fashion. 166 C.F. MERBS Fig. 4. Case 24, Table 4. Superior view of fifth lumbar vertebra and sacrum from Semna South, Sudan, showing complete separation through right interarticularis and spondylolisthesis of 7 mm (exposed marker). Note deviation of spinous process toward intact side. Fig. 7. Case 26, Table 4. Inferior view of fifth lumbar vertebra from Silumiut, Nunavut Territory, Canada, showing complete separation through left interarticularis. Note severe osteoarthritis of right zygapophysial joint. Fig. 5. Case 21, Table 4. Superior posterior view of sixth lumbar vertebra from Gran Quivira, New Mexico, with complete separation through left interarticularis. In terms of morphology, vertebra is lumbar on left side and sacral on right. Note resorption of left lamina and protrusion of defect area into neural canal. Fig. 6. Case 28, Table 4. Posterior view of fifth lumbar vertebra from Port au Choix, Newfoundland, showing complete separation through right interarticularis. The arch in series I cases (Table 4) usually shows some asymmetry attributable to the factors just described, although the range may vary considerably, from barely detectable to considerable. A distinctive feature of this asymmetry when the vertebra is viewed dorsally is a deviation of the spinous process away from the midline. Several vertebrae show no obvious deviation and others are too damaged to make the observation, but 13 of the 14 that do show deviation are deflected toward the intact side. Sometimes deviation of the spinous process may also be observed in vertebrae with unilateral incomplete separation. The vertebrae in this study provide little information regarding the spinous process tilt reported by clinicians (Maldague and Malghem, 1976). In some cases the process was too damaged to make the observation, but most often it simply lacked a distinct vertical aspect that could express tilt. In those cases where it was possible to make an observation (cases 9, 16, 19, and 24), the top of the process tilted toward the defect, as in the clinical cases. Another source of asymmetry in vertebrae with unilateral separation is a deviation of the arch into the neural canal at the defect site, thus reducing the size of the neural canal (stenosis). This appears be due, in part at least, to a buildup of bony callus at the fracture site (Fig. 5). Series II (Table 4), consisting of five vertebrae which combine complete separation on one side with incomplete separation on the other side, are particularly interesting because they potentially represent vertebrae in transition from unilateral to bilateral separation. In four of these (cases 32–35), the incomplete separation appears to have been in the process of developing at time of death. In the three young adults represented (cases 32–34), the complete separation looks to be of recent origin, while in the fourth, a middle adult, it likely existed for a much longer period before death. ASYMMETRICAL SPONDYLOLYSIS In case 32 (cross-listed in Table 2 as case 4), the complete separation has the irregular shape of a recent fracture and is poorly corticated along its entire length. The separation line appears to have three distinct parts when viewed from the side, suggesting that it developed as the coalescence of two separate partial separations, one extending downward from the superior margin of articularis and the other up from the inferior margin, eventually joining across the center of the isthmus by a horizontal fracture line (Merbs, 1995, p. 2,331, his Fig. 4). Case 33 is similar to 32, but with the addition of spina bifida. Also, the fracture line is less complex. Case 34 exhibits complete spondylolysis through the pedicle between the superior articular and transverse processes of L5 on the right side. The arch also fractured postmortem between the pedicle and the body on the left side, and the body part was not recovered. The surfaces of the two separations present an interesting contrast, the fracture edges of the cancellous bone at the postmortem break still being sharp, while those associated with the spondylolysis show the effects of resorption (Fig. 8). The left hemiarch is intact, but the surface of the bone shows evidence of an incomplete separation through the pedicle extending from the margin of the neural canal to near the tip of the transverse process. Viewed radiographically, the separation appears as a narrow band of radiolucency bordered by irregular patches of sclerotic bone. If allowed to continue, the result would likely have been complete bilateral separation through both pedicles. With the passage of additional time, the separation surfaces would have become corticated and the separation sites remodeled, perhaps, given the large surface areas involved, even developing into pseudarthroses. Case 35 is similar to that just described, but it involves L3 in an older individual, and the complete separation is at interarticularis, not through a pedicle. Although a gap did not develop at the site of the complete separation, the surfaces are well-corticated, and considerable resorption and remodeling took place, to produce an extremely narrow isthmus. A radiograph of the affected vertebra published by Stewart (1953, p. 944, his Fig. 7) shows a radiolucent line extending across at least part of the pedicle, from the margin of the neural canal to near the tip of the transverse process, similar to that in case 34. There appears to be a big difference in the timing of the separations in the two individuals, however, with the complete separation in case 34 barely in advance of the incomplete separation, while the complete separation in case 35 had to exist long enough for major changes to take place at the separation site. Case 36, which involves L4 in an old adult male with severe osteoporosis, is similar to case 35, but with the incomplete separation far more advanced (Fig. 9). Although the complete separation on the right side retains the irregular appearance of a fracture line, some cortication and resorption have 167 Fig. 8. Case 34, Table 4. Fifth lumbar vertebra from Ridley Island, British Columbia, showing complete separation through left pedicle and partial separation through right pedicle. Top: Posterior view, showing surface of recent separation (upper left) and postmortem break (right center), and partial separation (arrow) through left pedicle. Middle: Inferior view, showing incomplete separation (arrows) extending from margin of neural canal to near end of tip of transverse process. Bottom: Radiograph, showing opacity (sclerosis) in region of incomplete separation (arrow). taken place, and a narrow gap (1–3 mm) has formed. Interarticularis on the left side is much broader, due to a lateral extension of new bone that gives the 168 C.F. MERBS Fig. 9. Case 36, Table 4. Posterior view of fourth lumbar vertebra from Native Point, Nunavut Territory, Canada, showing complete separation through right interarticularis and nearly complete separation through left interarticularis. Note extensive development of new bone on left side, giving appearance of a second transverse process. appearance of a second transverse process. A separation extends laterally from the margin of the neural canal, through interarticularis, and through all but the lateral tip of the process formed by the new bone. The separated surfaces are broad, irregular, and poorly corticated. Considerable bone activity appears to have been taking place along the fracture line when death occurred, but whether the defect was healing or proceeding to complete separation is difficult to say based on its morphology. An interesting feature of this vertebra is that the inferior articular processes form a more oblique angle with the body, at 125° (measured between the inferior surface of the body and the facet surface of the process) compared with the 100 –105° usually seen. A raised facet had formed on the inferior surface of the L4 spinous process and on the superior surface of the L5 process, suggesting that the unusual angulation of the L4 arch may have been caused by forces generated by the spinous process contact. This also implies that movement was possible between the arch and the body portion of L4, allowing the increased angulation to take place, and that the small amount of healing seen on the left side occurred after complete separation. The L4 described here is actually 1 of 3 vertebrae affected by spondylolysis in this individual. Bilateral complete separation through interarticularis is present in L5 and incomplete separation at interarticularis on the left side in L6, a lumbarized sacral unit with essentially lumbar morphology on the left side and sacral morphology on the right. The spondylolysis at L5 resulted in an estimated anterior olisthesis of 14 mm (37%) relative to L6 (Merbs, 1983, p. 125, Fig. 63B), which in turn led to extreme osteophyte development (disk degeneration) on the superior anterior margin of the body. The problem at L5 likely contributed to that at L4. The nine vertebrae that make up series III (Table 4) combine complete separation through interarticu- Fig. 10. Case 39, Table 4. Posterior view of fifth lumbar vertebra from Inuksivik, Nunavut Territory, Canada, showing complete separation through right interarticularis and complete separation through left lamina adjacent to inferior articular process, with extensive new bone formation. laris on one side with separation through a pedicle (cases 37, 38, 40, 41, 42, 44, and 45), a lamina (case 39), or an inferior articular process (case 43) on the other side. In five of the pedicle cases the separation is located where the pedicle attaches to the vertebral body (cases 37, 38, 41, 42, and 45); in the other two it lies between the superior articular and transverse processes. In some cases the line of separation is irregular, having the appearance of a recent fracture, and the separation site has been enlarged by reactive bone. In case 39, for example, the separation through interarticularis on the right side appears to have been in existence for some time before death, with considerable remodeling having occurred (Fig. 10), while that through the lamina immediately adjacent to the inferior articular process on the left side presents the appearance of a healing fracture. The separation line is extremely irregular, with its edges in tight contact, and the entire area has been considerably expanded by new bone. Case 44 presents a very different appearance, with both the interarticularis and pedicle separations showing considerable maturity (Fig. 11). An indication of this is that a large faceted area has formed between the portion of the pedicle still attached to the body and that attached to the arch, indicating that considerable movement occurred at this unusual joint (pseudarthrosis). Case 43 is the most unusual in series III, because neither pedicle nor lamina is involved. The separation on the right side extends through the inferior articular process, while that on the left side is through interarticularis, but at the extreme lower margin of the superior articular facet (Fig. 12). Although the left separation site still has the irregular appearance of a fracture line, cortication and minimal remodeling have taken place. In contrast, the right side looks like an active fracture site, with considerable new bone contributing to a very thick 169 ASYMMETRICAL SPONDYLOLYSIS Fig. 11. Case 44, Table 4. Fifth lumbar vertebra from Kamarvik, Nunavut Territory, Canada, showing complete separation through right interarticularis and left pedicle between superior articular and transverse processes. Right: Posterior view of arch. Left: Anterior view of arch, showing facet that formed at pedicle separation site. arch in this area. What is clearly an unstable situation has led to 11 mm of olisthesis. All three vertebrae in series IV (Table 4) combine a single complete separation on one side with two on the other. Cases 46 and 47 are virtually identical, each involving a young adult male with bilateral complete separation of L5, with an additional separation through the right lamina midway between the spinous process and the inferior articular process. The separations through interarticularis are completely corticated, but neither displays any obvious signs of olisthesis. The two vary with respect to laminar separation, however. In case 46, the younger of the two, this separation appears to be of recent origin, presenting as an irregular fracture line with poor cortication and some reactive bone still present, while in case 47 the fracture site has smooth edges and the separation surface is completely corticated (Merbs, 1983, p. 123, Fig. 60). In addition, L4 in this individual has bilateral complete separation at interarticularis and is estimated to have slipped forward 4 mm relative to L5. Case 48 was described by Stewart (1953, p. 944, his Fig. 6) as having defects “through the pedicle on both sides,” but the total effect is still markedly asymmetrical. The separation line on the right side begins at the neural margin of the arch just anterior to the superior articular process and proceeds laterally and dorsally in approximately a straight line, leaving most of the transverse process still attached to the body. The pedicular separation on the left side also begins at the neural margin just anterior to the superior articular process, but it is shorter and curves sharply following the anterior margin of the process. A second, horizontal separation crosses interarticularis, but well above the usual location seen in classic spondylolysis. These two separations effectively isolate the superior articular process from the rest of the vertebrae. Unfortunately, this small part was not available for study, but most of its morphology could be reconstructed from its adjacent parts. Based on the presence of reactive bone, level of cortication, and degree of remodeling, the separation dividing the arch from the body on the left side appears to have occurred first with the development of a pseudarthrosis similar to that seen in cases 40 – 42 and 44 (Table 4). The separation on the right side and that isolating the left superior articular process occurred later but in undetermined order. Both of the separations on the left side were still experiencing bone activity when death occurred. DISCUSSION Although the traditional view of spondylolysis is that of bilateral complete separation through interarticularis, a process that effectively divides the arch from the body in relatively symmetrical fashion, this study indicates that most spondylolysis probably presents an asymmetrical picture as it develops, with the final symmetrical effect achieved only after considerable remodeling has taken place. Some cases, however, such as those involving unilateral complete separation and bilateral complete separation where different parts of the arch are affected on the two sides, may exhibit asymmetry as a permanent condition. 170 C.F. MERBS Fig. 12. Case 43, Table 4. Fifth lumbar vertebra and sacrum from Mocho, California, showing complete separation through left superior articular process and right inferior articular process. Top: Inferior view of L5, with parts separated to show edges. Bottom: Superior view of L5 and sacrum, with parts articulated. Most cases of spondylolysis are thought to result from fatigue or stress fracturing of some part of the vertebral arch, and the bare-bone appearance of the condition in its incomplete form, and even frequently in its complete form, is certainly consistent with this etiology. Stress fractures (spondylolysis included) characteristically have a “subradiological” period during which radiographs will appear normal (Elliott et al., 1988). Presumably this is true of their bare-bone appearance as well. However, the presence of a defect may be detected using forms of imaging that visualize bone activity at a particular site. For example, Bellah et al. (1991) found an abnormal focus of radiotracer uptake using single-photon emission computed tomography in the lumbar spines of 71 young patients with symptoms of low back pain possibly related to spondylolysis. In 66 of these, the abnormalities appeared in the lower lumbar region, localized to interarticularis. Perhaps even before the fractures become identifiable on radiographs, they may be visible on the bare bone surface of archaeological specimens (Merbs, 1995). They are seldom noted, however, probably for the following reasons: 1) spondylolysis usually occurs with low frequency, and only a fraction of these are likely to represent incomplete separation; 2) incomplete separation occurs primarily in adolescents and young adults, age groups that may be poorly represented in archaeological series; or 3) they are simply overlooked. The high frequency of this condition observed in the Hudson Bay Inuit collections thus represents an unusual opportunity to study this phenomenon in the context of asymmetrical expressions and spondylolysis formation. It should be noted, however, that even in the young individuals of these collections, incomplete separations occur primarily above and below L5, with complete separations occurring most frequently at L5. This suggests that the primary activity resulting in spondylolysis has already taken place, even in these young individuals, and the near absence of incomplete separations in older individuals seems to confirm this (Merbs, 1995). Nevertheless, the incomplete separations that are present in the young individuals, despite their transient nature, can provide interesting insight into the entire process. The pattern that emerges suggests that the stresses responsible for spondylolysis in its early stages tend to be greater on the right side than the left. They also appear to be greater on the inferior margin at the L4 level and the superior margin at S1, and greater on the inferior margin on the right side above S1. The right-sidedness at this level of involvement suggests that the asymmetry of the stresses may in some way be related to handedness, with activities emphasizing the dominant hand, usually the right, being somehow transmitted to the lower back. In the two vertebrae where the separation lines go in opposite directions, the pattern is consistent with greater stress on the left superior margin and the right inferior margin at interarticularis. Limiting the asymmetry analysis to complete separation (i.e., the general approach of archaeological and anatomical studies) has produced rates of unilateral separation to bilateral separation that range from a high of 33.3% (6/18) in the anatomy skeletons of the South African Dart Collection (Eisenstein, 1978), to a low of 3.2% (2/62) in Archaic Indian skeletons from the Indian Knoll site in Kentucky. The Canadian Inuit frequency of 21.2% (18/85) lies near the midpoint of this range. Although the literature contains little information on sidedness in unilateral separation, the only relatively large series, the combined Todd and Terry Collections, produced a right to left ratio of 22 to 10 (Roche and Rowe, 1951). In the unilateral series analyzed here (Table 4), defects are also more numerous on the right side (20 to 16). Age may be a factor, however, as right defects are much more numerous (13:6) among younger individuals but not among older individuals (7:10). ASYMMETRICAL SPONDYLOLYSIS Spina bifida and spondylolysis occur together with higher-than-expected frequency according to the clinical literature (Amuso and Mankin, 1967; Fredrickson et al., 1984; Oakley and Carty, 1984), but this association has been difficult to replicate in studies of archaeological skeletons (Waldron, 1993). When limited to unilateral separation, however, particularly involving L5, an association with spina bifida does seem likely. Stewart (1953), working with archaeological skeletons from Alaska, and Roche and Rowe (1951), working with the Todd and Terry Collections, came up with spina bifida-unilateral spondylolysis distributions easily significant at the 0.001 level, using the 2 test. This association is supported by the present study, where the two conditions occur in the same vertebrae in one-third of cases (Table 4, series I and II). Age may be a significant factor, however, as the frequency is much higher in younger individuals than in the older group (47% to 19%). All of the unilateral spondylolysis-spina bifida cases found by Stewart (1953), Simper (1986), and Roche and Rowe (1951) involved L5, a common site for spina bifida, and all 11 in the present study involve the last lumbar vertebra, L5 (9 cases) or L6 (2 cases). Given the different etiology of the two conditions, the relationship between spina bifida and unilateral spondylolysis seems purely mechanical, the discontinuity in the arch tending to weaken it, but also acting as a block to the transmission of stresses from one side to the other. Little information is available on asymmetrical spondylolysis involving transitional lumbosacral vertebrae, particularly when the morphology of the affected vertebra is already asymmetrical, i.e., lumbar on one side, sacral on the other. To some extent, the case described by Vyhnánek (1989, p. 76, his Fig. 8) meets this description, with the single separation through interarticularis occurring on the lumbar side. Four vertebrae in the present study also involve transitional lumbosacral vertebrae, in each case a partially lumbarized sacral unit, and in all four (two with spina bifida), the defect is on the “lumbar” side (Table 4). Because of its basic morphology, the sacral side allows less movement than the lumbar side, and thus there is less likelihood of stresses that could lead to spondylolysis concentrating at interarticularis on the sacral side. An intact arch may progress to unilateral spondylolysis and then to bilateral separation, but extracting information regarding the exact nature of this process from skeletons is difficult. Here the clinical data can be especially valuable. The identification of unilateral separation in a patient allows for a close monitoring of the intact side of the arch, to record any changes that take place. The earliest changes that have been observed are scintigraphic uptake at interarticularis, indicating that bone activity was taking place (Lowe et al., 1987), and the appearance on radiographs of a band of decreased density probably related to fatigue fracturing (Hadley, 1963). Archaeological equivalents of this would 171 appear to be cases 32 and 35 (Table 4), both of which display narrow bands of lucency on radiographs as well as external changes consistent with fatigue fracturing. More general clinical observations included hypertrophy, sclerosis, and arch deformity (Lowe et al., 1987). In some cases, what appears to have been a developing defect healed with no apparent lasting effect (Letts et al., 1986); in others, the condition progressed to bilateral separation (Albers and Yochum, 1980; Aland et al., 1986; Garber and Wright, 1986; Letts et al., 1986). In the presence of unilateral separation, it would appear that the opposite side of the neural arch is exposed to abnormal stresses that could lead to fatigue fracturing. The hypertrophic appearance in the vicinity of the fracture site reported by several clinicians is consistent with attempts at healing (Lowe et al., 1987; O’Beirne and Horgan, 1988). An excellent example of this is the 25-year-old individual described by Aland et al. (1986), in whom sclerosis and lucency were observed radiographically on the side opposite a unilateral spondylolysis, followed by massive overgrowth of bone on the pedicle. Four months later, the bone mass was excised and revealed an established fracture with nonunion. Although the phenomenon of unilateral separation leading to abnormal stresses and fracturing in the opposite hemiarch appears to be associated primarily with younger individuals, it was also observed in a 40year-old man with a 10-year history of lower back pain (O’Beirne and Horgan, 1988). Bilateral complete separation with subsequent healing on just one side to produce unilateral spondylolysis has been reported in both the clinical and archaeological literature, and it appears to be a real phenomenon as long as the fracture site involved remains active. After cortication and remodeling of the separation edges, and given the potential for movement between the separated arch and the body, healing would seem an unlikely occurrence. Asymmetry of the neural arch is commonly associated with unilateral spondylolysis, but the nature and degree of asymmetry seem to vary considerably. The kinds of asymmetry usually noted are as follows: 1) tilting of the vertical axis of the spinous process with the anterior aspect closer to the side with the defect; 2) deviation of the spinous process from the midline toward the intact side; 3) a shorter, smaller lamina, usually on the side of the defect; and 4) a wider (or narrower) interarticularis at the defect site compared with the opposite side. Arch asymmetry has the advantage of being easily detected on simple anteroposterior radiographs of living individuals, thus allowing the clinician to predict the presence of a unilateral arch defect. Arch asymmetry was actually used as a sorting device by Maldague and Malghem (1976), who discovered that unilateral arch hypertrophy of unknown cause, as identified in 32 of 50 individuals, turned out to be associated with a separation at interarticularis on the opposite side. Roche and Rowe (1952) attributed 172 C.F. MERBS asymmetry to the gap in the isthmus on the defective side, noting that those with a wide gap were more asymmetrical than those with a narrow gap. At best, however, this explanation can account for only part of the asymmetry observed, as even cases where the separated edges are in contact usually show some arch asymmetry. Spinous process tilt turned out to be the most elusive of the factors of asymmetry examined in the present study. Part of this was due to missing arches and poor preservation, but some was due to the morphology of the spinous processes involved, which lacked a readily recognizable vertical linear dimension. In the four cases where it was possible to make an observation, the top of the process did tilt toward the defect. Deviation of the spinous process away from the midline provided more information, most of it consistent with clinical expectations. All but 1 of 14 showing deviation did so toward the intact side. This appears consistent with the intact side bearing the brunt of stresses generated in the arch following unilateral separation. The size and shape of the lamina on the defective side relative to the intact side proved to be highly variable and difficult to quantify. Narrower, shorter laminae on the defective side appeared to correlate roughly with the longevity of the defect: the longer the separation existed, the smaller the lamina on the side of the defect. The size discrepancy thus appears due to resorption and remodeling with reduced stress on the defective side, contrasted perhaps with hypertrophy on the intact side. The width of the isthmus at the site of separation was found to vary greatly, ranging from much broader than that on the intact side to much narrower. In those cases where a distinct gap is visible at the separation site, the isthmus tends to be narrow. Most of the unilateral separations observed in this study were consistent with lying along a continuum representing duration of the defect, and (perhaps even more important) phase of healing or nonhealing. Even in cases where the fatigue fracture appears to still be developing, bone reaction at the site tended to increase the diameter of the isthmus. With progression of the fracture to some separation, and for some time thereafter, the bone reaction may be intense, resulting in an even greater widening of the isthmus. The separation line tends to be very irregular, giving the appearance of a fresh fracture. The minute projections on one edge tend to match quite well with indentations on the other edge, and cortication, which is difficult to observe because of the tight fit, is slight to nonexistent. If union does not occur, reactive bone is no longer produced, and the separation edges begin to lose their complex topography and develop a cortex. This is followed by resorption and remodeling, not only of the new bone that formed, but of the original bone as well. The result is a narrowing of the isthmus which may or may not be accompanied by a gap forming at the separation. Although the early stages of this process tend to be found in younger individuals, they may even be found in an old adult, as illustrated by case 28. With the stresses normally shared by both zygapophysial joints shifted to the side opposite the defect, the joint on that side becomes severely stressed. Using CT imaging, Lusins et al. (1994) observed degenerative changes in the zygapophysial joint on the intact side in an individual only 16 years of age. With SPECT imaging they also demonstrated the presence of bone activity on the intact side in this and two other patients, even in the absence of observable changes on CT images. The authors concluded from their cases that following unilateral separation, stress to the back will be transmitted unevenly to the pedicle, interarticularis, and articular facets on the opposite side, ultimately affecting the zygapophysial joint on this side. A similar situation was described by Gunzburg and Wagner (1988) and Kornberg (1988), with the similarity to degenerative spondylolisthesis noted in both reports. The study of archaeological cases disclosed some differences, however; the degenerative olisthesis associated with unilateral separation occurred upward and forward along a curvilinear plane, while that associated with bilateral separation was essentially a simple anterior migration of the joint. Other indicators of increased stress on the intact side in unilateral separation are a reduction in body height and greater disk degeneration. Some of this asymmetry may account for what Kornberg (1988) referred to as a rotation of L5 seen on the CT scans of one of his patients. The clinical literature frequently associates pain with unilateral separation; in fact, it is often that pain which brings a case to the attention of a clinician (Lusins et al., 1994; Kornberg, 1988; Albers and Yochum, 1980; O’Beirne and Horgan, 1988; Sherman et al., 1977; Hadley, 1963). Although archaeological specimens are hardly ideal for assessing the pain factor, it can be noted that several elements of osseous morphology are likely sources. The first of these is the stress fracturing itself, a painful condition wherever it occurs. The second is the inward curvature of the arch at the defect site that could impinge on the root of the lumbar nerve at that level, a possibility noted by Aland et al. (1986). Pain at later stages could be associated with degenerative changes. CONCLUSIONS Spondylolysis is more a process than a condition, and its classic presentation in archaeological skeletons, i.e., bilateral separation through interarticularis, is essentially the end product of that process. Trying to understand the process by studying only its final effect has severe limitations. The study of incomplete spondylolysis can provide greater insight into the process by which a separation develops in the vertebral arch, but few examples of this stage of ASYMMETRICAL SPONDYLOLYSIS separation in archaeological specimens have been described. Asymmetrical spondylolysis, most likely produced by unequal stresses being applied to the two sides, presents another opportunity to examine those stresses, particularly as they affect the two sides of a vertebra in different ways. The earliest examples of spondylolysis recognizable in bare bone, based on the study of this phenomenon in Canadian Inuit skeletons, are frequently asymmetric, usually from side to side, but sometimes in vertical direction as well (downward from above or upward from below). Based on the presence of complete separation in Inuit adolescent and young adult skeletons, the process usually begins at the L5 level, with additional separations then developing above (L3 and L4) and below (S1). These additional separations in the adolescent and young adult Inuit appear as incomplete spondylolysis, or stress fractures in progress. The presence of bony callus at the separation sites reflects an effort at healing, and clinical studies suggest that fracturing and healing may be episodic, resulting ultimately in complete healing or complete separation. The incomplete separations occur more frequently on the right side, suggesting that activities reflecting handedness may play a role. The study of examples of asymmetrical spondylolysis in archaeological skeletons involving complete separation on at least one side produced a similar picture with respect to side involvement, the separation occurring more frequently on the right side, particularly in younger individuals. It is clear from the clinical literature that cases of unilateral spondylolysis may heal, progress to bilateral separation, or develop into a permanent unilateral separation. Although it is difficult to demonstrate these dynamic processes in dead bone, the presence of external discontinuities and opacities on radiographs suggestive of stress fracturing, along with callus development in the vicinity of the defects, indicates that the defect sites were indeed active when death occurred. Although attempts to find correlations between spondylolysis and spina bifida in the archaeological context similar to those noted in clinical studies have been largely unsuccessful, unilateral spondylolysis would appear to be an exception. In this study, the two conditions were found to occur together with a much higher frequency than expected. This may be due to greater stress being concentrated on one side of the arch or the other because of the discontinuity in the midline, thus increasing the likelihood of separation on the stressed side. Another important factor is that if separation does occur on one side, the discontinuity of the arch will minimize its effect on the opposite side, making it less likely that separation will occur on that side as well. The examples of unilateral separation that remain as such, whether they are of archaeological or clinical origin, eventually develop a distinctive pat- 173 tern of morphology that reflects the unequal stresses on the two sides. The extent to which unequal stresses contributed to the unilateral separation to begin with is unclear, but the separation is clearly related to morphological changes that subsequently took place. The severe osteoarthritis that develops in the zygapophyseal joint opposite the defect attests to the fact that this side bears a disproportionate amount of stress. Ultimately this can lead to an anterior slippage, similar in many respects to that seen in degenerative olisthesis. Other changes taking place that contribute to the overall asymmetry of the affected vertebra include size differences in the thickness of the laminae, width of the isthmi at interarticularis and height of the body, deviation and tilt of the spinous process, and impingement of the arch on the neural canal. Taken along with clinical examples, the archaeological specimens of unilateral separation, whether incomplete or complete, present a dynamic picture of the stresses occurring in the lower lumbar region in humans and the bony response that takes place. ACKNOWLEDGMENTS Special thanks go to Ethne Barnes, Jerome Cybulski, Michael Fong, and Sonja Jerkic for bringing fascinating specimens to my attention and allowing me to include them in this study. Thanks also go to the various institutions that allowed me to study the vertebrae used here, and especially the Canadian Museum of Civilization in Hull (Quebec, Canada). LITERATURE CITED Albers VL, Yochum TR. 1980. Reactive sclerosis of a pedicle due to unilateral spondylolysis. ACA J Chiroprac 14:56 –59. Aland C, Rineberg BA, Malberg M, Fried SH. 1986. Fracture of the pedicle of the fourth lumbar vertebra associated with contralateral spondylolysis. J Bone Joint Surg [Am] 68:1454 –1455. Amuso SJ, Mankin HJ. 1967. Hereditary spondylolisthesis and spina bifida. J Bone Joint Surg [Am] 49:507–513. Bellah RD, Summerville DA, Treves ST, Micheli LJ. 1991. Lowback pain in adolescent athletes: detection of stress injury to the pars interarticularis with SPECT. Radiology 180:509 –512. Buikstra JE, Ubelaker DH. 1994. Standards for data collection from human skeletal remains. Fayetteville, AR: Arkansas Archeological Survey Research Series, no. 44. Elliott S, Hutson MA, Wastie ML. 1988. Bone scintigraphy in the assessment of spondylolysis in patients attending a sports injury clinic. Clin Radiol 39:269 –272. Eisenstein S. 1978. Spondylolysis: a skeletal investigation of two population groups. J Bone Joint Surg [Br] 60:488 – 494. Fredrickson BE, Baker D, McHolick WJ, Yuan HA, Lubicky JP. 1984. The natural history of spondylolysis and spondylolisthesis. J Bone Joint Surg [Am] 66:699 –707. Gaballah MF, Badawy ZH. 1980. A further anatomical study of spondylolysis in human lumbar vertebrae. Egypt J Anat 3:9 – 18. Garber JE, Wright AM. 1986. Unilateral spondylolysis and contralateral pedicle fracture. Spine 11:63– 66. Gunness-Hey M. 1982. The Koniag Eskimo presacral vertebral column: variations, anomalies and pathologies. Ossa 7:99 –118. Gunzburg R, Wagner J. 1988. Degenerative spondylolisthesis with unilateral spondylolysis. A case report. Int Orthop 12: 139 –141. Hadley LA. 1963. Stress fracture with spondylolysis. AJR 90: 1258 –1262. 174 C.F. MERBS Kettelkamp DB, Wright DG. 1971. Spondylolysis in the Alaskan Eskimo. J Bone Joint Surg [Am] 53:563–566. Kornberg M. 1988. Spondylolisthesis with unilateral pars interarticularis defect and contralateral facet joint degeneration. Spine 13:712–713. Letts M, Smallman T, Afanasiev R, Gouw G. 1986. Fracture of the pars interarticularis in adolescent athletes: a clinical biometrical analysis. J Pediatr Orthop 6:40 – 46. Lowe J, Libson E, Ziv I, Nyska M, Floman Y, Bloom RA, Robin GC. 1987. Spondylolysis in the upper lumbar spine. J Bone Joint Surg [Br] 69:582–586. Lusins JO, Elting JJ, Cicoria AD, Goldsmith SJ. 1994. SPECT evaluation of unilateral spondylolysis. Clin Nucl Med 19:1–5. Maldague BE, Malghem JJ. 1976. Unilateral arch hypertrophy with spinous process tilt: a sign of arch deficiency. Radiology 121:567–574. Merbs CF. 1983. Patterns of activity-induced pathology in a Canadian Inuit population. Ottawa: National Museum of Man Mercury Series, Archaeological Survey of Canada paper no. 119. Merbs CF. 1995. Incomplete spondylolysis and healing: a study of ancient Canadian Eskimo skeletons. Spine 20:2328 –2334. Merbs CF. 1996a. Spondylolysis and spondylolisthesis: a cost of being an erect biped or a clever adaptation? Yrbk Phys Anthropol 39:201–228. Merbs CF. 1996b. Spondylolysis of the sacrum in Alaskan and Canadian Inuit skeletons. Am J Phys Anthropol 101:357–367. Merbs CF. 1997. Eskimo skeleton taphonomy with identification of possible polar bear victims. In: Haglund WD, Sorg MH, editors. Forensic taphonomy: the postmortem fate of human remains. Boca Raton: CRC Press. p 249 –262. Merbs CF. 2001. Degenerative spondylolisthesis in ancient and historic skeletons from New Mexico Pueblo sites. Am J Phys Anthropol 116:285–295. Miki T, Tamura T, Senzoku F, Kotani H, Hara T, Masuda T. 1991. Congenital laminar defect of the upper lumbar spine associated with pars defect. A report of eleven cases. Spine 16:353–355. Miles JS. 1975. Orthopedic problems of the Wetherill Mesa populations. Washington, DC: National Park Service, Wetherill Mesa Studies, Publications in Archaeology 7G. Neugebauer FL. 1885. Neuer Beitrag zur Aetologie und Casuistik der Spondyl-olisthesis. Arch Gynaekol 25:182–252. Oakley RH, Carty H. 1984. Review of spondylolisthesis and spondylolysis in paediatric practice. Br J Radiol 57:877– 885. O’Beirne JG, Horgan JG. 1988. Stress fracture of the lamina associated with unilateral spondylolysis. Spine 13:220 –222. Poirier P. 1911. Traité d’anatomie humaine. 2nd ed. Volume 1. Paris: Masson. Porter RW, Park W. 1982. Unilateral spondylolysis. J Bone Joint Surg [Br] 64:344 –348. Roche MB, Rowe GG. 1951. The incidence of separate neural arch and coincident bone variations. A survey of 4,200 skeletons. Anat Rec 109:233–252. Roche MB, Rowe GG. 1952. The incidence of separate neural arch and coincident bone variations: a summary. J Bone Joint Surg [Am] 34:491– 494. Rossi F. 1978. Spondylolysis, spondylolisthesis and sports. J Sports Med 18:317–340. Sherman FC, Wilkinson RH, Hall JE. 1977. Reactive sclerosis of a pedicle and spondylolysis in the lumbar spine. J Bone Joint Surg [Am] 59:49 –54. Shore LR. 1929. A report on a specimen of spondylolisthesis found in the skeleton of a Bantu native of South Africa: with further specimens illustrating an anomalous mode of development of the lower lumbar vertebrae. Br J Surg 16:431– 439. Simper LB. 1986. Spondylolysis in Eskimo skeletons. Acta Orthop Scand 57:78 – 80. Snow CE. 1948. Indian Knoll skeletons of site Oh2, Ohio County, Kentucky. Lexington, KY: University of Kentucky Reports in Anthropology, volume 4, no. 3, part 2, p 370 –532. Stewart TD. 1931. Incidence of separate neural arch in the lumbar vertebrae of Eskimos. Am J Phys Anthropol 16:51– 62. Stewart TD. 1953. The age incidence of neural-arch defects in Alaskan natives, considered from the standpoint of etiology. J Bone Joint Surg [Am] 35:937–950. Stewart TD. 1979. Patterning of skeletal pathologies and epidemiology. In: Laughlin WS, Harper AB, editors. The first Americans: origins, affinities, and adaptations. New York: WennerGren Foundation for Anthropological Research. p 257–274. Thieme FP. 1950. Lumbar breakdown caused by erect posture in man. Ann Arbor: Anthropological Papers, Museum of Anthropology, University of Michigan, no. 4. Vyhnánek L. 1989. Einige Beurteilungsprobleme der intrinsischen (konstitutionellen) Spondylopathien im alten Knochenmaterial. Archaol Mus 2:63– 82. Waldron T. 1992. Unilateral spondylolysis. Int J Osteoarch 2:177–181. Waldron T. 1993. A case-referent study of spondylolysis and spina bifida and transitional vertebrae in human skeletal remains. Int J Osteoarch 3:55–57. Willis TA. 1923. The lumbo-sacral vertebral column in man, its stability of form and function. Am J Anat 32:95–123.