Hip and Thigh Anatomy of the Clouded Leopard (Neofelis nebulosa) with Comparisons to the Domestic Cat (Felis catus)



The clouded leopard (N. nebulosa) is an endangered felid native to Southeast Asia. It is considered to be the largest and most acrobatic of the arboreal cats. To date, published studies have focused on cranial morphology, particularly osteology and dentition. This study describes the hip and thigh morphology of N. nebulosa and compares it to the domestic cat (F. catus). A number of statistically significant differences were observed between N. nebulosa and F. catus that were primarily associated with muscle attachment surface areas rather than differences in overall morphology. F. catus had proportionally larger attachment areas for Mm. gluteus profundus, tensor fasciae latae, and biceps femoris, while N. nebulosa had larger attachment areas for Mm. gluteus medius, rectus femoris, and semitendinosus. By comparison, surface area analysis of nonfelid species (C. latrans and V. vulpes) showed more differences when compared to felids demonstrating that this analytical technique is useful for differentiating closely-related species from more distant ones. These results suggest that muscle map area analysis may be useful for comparing phylogenetic relationships between and within groups of different species and may also suggest variations in locomotor habits. Anat Rec, 2012. © 2012 Wiley Periodicals, Inc.


Understanding morphometric variations in any vertebrate, in part, requires study of its muscle morphology. One method for doing this involves creating maps of muscle attachment sites, called muscle mapping. This technique may also aid in interpreting the action of a given muscle by mapping its attachment sites on bones. Muscle mapping has been performed extensively in humans (Clemente,1997; Gilroy et al.,2008), but published data for animals are uncommon considering the number of living species (Sisson and Grossman,1953; Schon,1968; Swindler and Wood,1973; Domning,1977,1978; Evan de Lahunta,1980; Spoor and Badoux,1988; Fisher et al.,2008,2010). One of the most complete animal muscle maps has been published for the domestic dog given its value in veterinary medicine (Evans and de Lahunta,2004).

Morphometric comparisons of species have primarily focused on skeletal analyses and less on muscle morphology. Differences in dentition, for example, are commonly used to identify taxonomic relationships among various species (Werdelin,1987; Caumul and Polly,2005). In addition, osteological indices, such as forelimb to hindlimb ratios may indicate locomotor and predatory behavior differences (Gonyea,1976; Anyonge,1996).

This study examines the hip and thigh muscle morphology of the clouded leopard (N. nebulosa). Clouded leopards belong to the Pantherinae lineage and are native to Southeast Asia. The presence of unique cloud-shaped dark spots on their lighter colored pelts makes them distinguishable from other species in the region. They are medium-sized, weighing between 16 and 23 kg and ranging from 60 to 106 cm in head-body length (Wozencraft,2005). They have short stocky limbs and a tail, that is, usually equal to the body length which is used extensively for maintaining balance while climbing (Nowell and Jackson,1996; Sunquist and Sunquist,2002; Nowak,2005).

Clouded leopards are arboreal animals that mostly prefer primary forests. In the trees, they are among the most acrobatic of felids, displaying arboreal talents similar to those of the smaller South American margay of South America and the marbled cat of Southeast Asia. However, the latter two species are three to four times smaller in body size (Nowell and Jackson,1996; Sunquist and Sunquist,2002; Nowak,2005). Hemmer (1968) has observed the arboreal behavior of clouded leopards in captivity and reported that they are able to balance on tree branches with no difficulty, climb head-first down tree trunks, move along horizontal branches with their backs to the ground, and hang down from tree branches using their hind feet. They are able to catch prey both on the ground and in the trees with ease. Their prey ranges from small animals such as birds, chickens, and monkeys to larger animals such as deer, cattle and wildboar (Davies,1990; Nowell and Jackson,1996; Sunquist and Sunquist,2002; Nowak,2005). They prefer feeding in the trees to protect their food from other predators (Austin and Tewes,1999; Ghose,2002; Austin et al.,2007).

Clouded leopards are listed as an endangered species (Appendix I of CITES) and are classified as vulnerable by the IUCN (Baillie and Groombridge,1996; Nowell and Jackson,1996; Nowak,2005). The biggest threat to survival of the species results from human activity. Deforestation has decreased their numbers and caused them to move into more sparsely forested areas. Pelts and other body parts, such as teeth and bones, are sold in the black market or used for traditional medicine, reducing the clouded leopard population in Southeast Asia. In addition, clouded leopard meat has been featured in high-end restaurants and markets (Cat News,1987).

Research on these animals has been limited because of the low rate of captive breeding success and excessive intersexual aggression. As a result, numerous studies have concentrated on their reproductive physiology (Wildt et al.,1986, Yamada and Durant,1989; Brown et al.,1995, Howard et al.,1998, Wielebnowski et al.,2002, Pukazhenti et al.,2006). The great length of the canine teeth in N. nebulosa led a number of researchers to study cranial morphology in an effort to find similarities with extinct sabretooth cats (Therrien,2005; Christiansen,2006;2008; Freeman and Lemen,2007; Slater and van Valkenburgh,2008). These studies focused on skull and jaw muscle attachment sites, but the cranial musculature has not been described. Other studies analyzed postcranial bone morphometrics to evaluate locomotor behavior, resulting in the classification of N. nebulosa as an ambush predator (Gonyea,1976; van Valkenburgh,1987; Anyonge,1996; Serrat et al.,2007; Meachen-Samuels and van Valkenburgh,2009). Gonyea (1978) included N. nebulosa in the study of elbow and wrist joint morphology of carnivores in an attempt to find a correlation with locomotion and the use of the forelimbs during hunting.

During an extensive review of the literature no published detailed anatomical description for N. nebulosa appears to exist nor has its morphology been compared to other felids. Studying muscle morphology and identifying muscle attachment sites on bones may also be useful for interpreting the functional characteristics of musculature. Biomechanical analysis of muscle attachment sites has been used previously to evaluate potential mechanical advantages (McGowan,1999; Kardong,2009).

In this study, the hip and thigh muscle morphology and, in particular, the muscle attachment site locations of the clouded leopard (N. nebulosa) were located, analyzed and compared to those of the domestic cat (F. catus). Given its relative phylogenetic distance and locomotor differences of N. nebulosa compared to F. catus, we initially hypothesized that the muscle morphology of the two species would be somewhat different from each other. While a few differences in muscle morphology do exist, analysis of muscle attachment site surface area proved to be a more useful way to define the two species. The addition of fox (Vulpes vulpes) and coyote (Canis latrans) pelvic muscle maps revealed the phylogenetic significance of the analysis. Although morphological differences were identified between the two species, biomechanical analysis is required before it will be possible to determine whether muscle attachment site variations are related to differences in locomotor behavior.


Two N. nebulosa specimens were obtained through the Smithsonian Institution Osteopreparation laboratory and are currently housed in the research collection of the Biological Sciences Department of Northern Illinois University (NIU 2105-2106). Both animals were captive bred animals that died of natural causes. Following necropsy, the animals were transported to Northern Illinois University in 2008 for dissection. Five preserved specimens of F. catus were obtained from NASCO Biological Supply Company and were studied as a comparison to N. nebulosa. Two coyotes (C. latrans) and one fox (V. vulpes) were wild specimens that died from various head trauma and obtained courtesy of Willowbrook Wildlife Center in Naperville, IL. These two species were included in this study as a comparison to the felid specimens. However, unlike the felid dissections canid dissections focused solely on identification of muscle attachment sites.

Initially, the skin was removed from the clouded leopards, coyotes, and foxes, after which the specimens were placed in a 50% ethanol and normal saline solution. The hip and thigh muscles of the domestic cats and the clouded leopards were dissected. Only the pelvic muscles of the coyote and fox were analyzed during this study. A detailed photographic record was made of all dissections using an Olympus E520 digital camera.

In selected specimens, short tufts of muscle fibers were left at the attachment sites on the bone to help more accurately identify attachment site locations and to aid in drawing the muscle maps after the dissections were completed. An example of this can be seen on Fig. 1. To further identify some muscle attachment sites associated with the vertebral column, the site boundaries were marked with fine lead wire before obtaining a radiogram of the site using a GE Mobile 100-15 portable X-ray machine and high speed films (Alpha Images Medical X-ray Films AXB-90).

Figure 1.

An example of leaving short tufts of muscle fibers on the pelvis during a dissection to facilitate accurate identification of muscle attachment sites. The picture is of F. catus pelvis.

Muscle Attachment Site Area Measurements

Using a graph pad with accompanying software (Wacom), the exact muscle attachment location was drawn on actual size digital photographs of bone specimens. Several photographs at different angles were taken of each bone to determine the optimal view for each attachment site. To obtain the best view of the pelvis, three views were photographed: (1) the face of ilium was aligned parallel with the camera lens, (2) the face of ischium was placed parallel with the camera lens, and (3) a face-on view of the ischial tuberosity. For the femur four views were analyzed with reference to anatomical position on the animal: anterior, lateral, medial, and posterior. The optimal view was first chosen after which each muscle attachment site was outlined then digitized. The areas of the selected muscle attachment sites were determined using ImageJ 1.30v software. The digitized surface areas were then expressed as proportional surface areas by dividing each muscle surface area by the sum of all the measured muscle surface areas on the respective bone (pelvis or femur) of each specimen. The n number for N. nebulosa pelvi was two, which is a rather small sample size. Therefore, after verifying muscle attachment sites by dissection on these two specimens it was possible to identify these same sites on three additional pelvic bone specimens from the Field Museum, Chicago. Our initial studies showed a large number of similarities between N. nebulosa and F .catus. Therefore, to demonstrate that this type of surface area analysis is indeed species-related, the pelvic muscle attachment areas were analyzed for C. latrans and V. vulpes (Fig. 2).

Figure 2.

The pelvic muscle attachment surface area means (±SEM) analyzed from the parallel-ilium view for N. nebulosa (n = 5), F .catus (n = 5), C. latrans (n = 2), and V. vulpes (n = 1). The number 1 denotes a significant difference between F. catus and N. nebulosa (P ≤ 0.05). The number 2 denotes a significant difference between both of the felids and C. latrans (P ≤ 0.05). The number 3 denotes a significant difference between one of the felids and C. latrans (P ≤ 0.05).

Skeletal Defleshing Procedure

After each dissection was completed, the bones of the specimens were defleshed for surface analysis. Each specimen was incubated at 75°C in a 1% solution of sodium hydroxide and sodium chloride. At various times during the incubation period, the bones were rinsed periodically with water, scrubbed with a test tube brush, and then placed back in the solution. When the muscle fibers were seen to loosen, the bones were removed from the alkali solution and rinsed with distilled water. The bones then were incubated at 37°C overnight in a phosphate buffer solution (PBS, 0.17M NaCl, 0.003M KCl, 0.013M Na2PO4, and 0.002M KH2PO4, pH 7.4) containing trypsin (1 g/L). Additional trypsin (1 g) was added every 6 hrs until all soft tissue was removed. The bones were then placed overnight in a 50% ethanol solution to remove lipids.

Ilium and Ischium Proportional Determinations

Ilium and ischium lengths have been used previously to study a number of biomechanical parameters (Smith and Savage,1956; Anemone,1993). Therefore these measurements were determined for F. catus and N. nebulosa. The ilium was measured from the most cranial point on the iliac crest to the midpoint of the acetabulum. The ischium was measured from the midpoint of the acetabulum to the most caudal point on the ischial tuberosity (Fig. 3). The two lengths were summed, and each ilium and ischium length was divided by this sum to obtain a percentage value. The ratio of the ilium length to the ischium length was also determined. The mean ratio value differences were analyzed by the analysis of variance test (ANOVA).

Figure 3.

Bony landmarks used to determine ilium and ischium lengths.

Statistical Analysis

The percent surface area values for the muscle attachment site surface areas were calculated by dividing each muscle attachment sites surface area by the total amount of surface area of muscles occupying on the surface of a given bone (pelvis or femur). This analysis will be referred to simply as the “surface area.” Because ratios are not normally distributed, all ratios were arcsine transformed before use in statistical analysis (Zar,1999).


To avoid repetition, the muscle descriptions below are based on dissections of N. nebulosa. Unless otherwise noted, the same or very similar morphological characteristics were also observed during the dissections of F. catus.

Muscles Originating from the Pelvis

  • 1M. gluteus medius is located cranial to M. gluteus superficialis, caudal to M. sartorius and superficial to M. gluteus profundus. It is a thick muscle covering a third of the wing of the ilium extending beyond the caudal dorsal iliac spine and is also attached to the transverse processes of the last sacral and first caudal vertebrae. It inserts on the lateral proximal tip of the greater trochanter of the femur. In N. nebulosa, the ridge along the dorsal ilium between the cranial and caudal dorsal iliac spines forms the cranial origin of M. gluteus medius and is more pronounced and elevated compared to F. catus. The surface area at the origin site was greater for N. nebulosa as compared to F. catus (Table 1, Figs. 4A, 5A, 6(1), and 7A). The insertion area on the femur, however, was slightly greater for N. nebulosa (Table 2).
  • 2M. gluteus profundus originates from the medial and ventral surface of the wing of the ilium extending dorsally to the border of the origin of M. tensor fasciae latae. It attaches cranially to the body of the ilium and extends to the border of the M. gemellus cranialis attachment site. It is deep to M. gluteus medius and M. piriformis. M. gluteus profundus inserts onto the anterior distal surface of the greater trochanter. In N. nebulosa, the iliac site of origin appeared more rugose when compared to F. catus, however, the surface area on the pelvis was greater for F. catus than for N. nebulosa (Table 1, Figs. 5, 6(2), 7, and 8A).
  • 3M. tensor fasciae latae originates as a thin, short strip on the ventral lateral surface of the ilium (tuber coxae) and lies dorsal to M. sartorius. The muscle blends with and inserts on the fascia lata of the thigh. In F. catus, the muscle origin consists primarily of fleshy fibers, while in N. nebulosa the attachment is more aponeurotic. The surface area of origin on the pelvis was greater for F. catus than for N. nebulosa [Table 1, Figs. 4A and 6(3)].
  • 4M. sartorius is located on the cranial, medial thigh and originates as a thin strip on the ventral border of the iliac crest beginning from the cranial ventral iliac spine and extending to the body of the ilium. The muscle inserts onto the medial surface of the patella and the proximal medial tibia and the tibial crest. It is superficial to M. vastus medialis and M. rectus femoris. (Figs. 4, 6(4), and 9).
  • 5M. rectus femoris is a member of the quadriceps femoris muscle group that lies between M. vastus lateralis and M. vastus medialis and covers M. vastus intermedius. It originates on the tuberosity for M. rectus femoris, which in N. nebulosa is more ventral to the acetabulum than in F. catus where the tuberosity is more cranial to the acetabulum. The surface area of the origin was greater in N. nebulosa than F. catus [Table 1, Figs. 4B, 5B,C, and 6(5)]. M. rectus femoris contributes to formation of the quadriceps tendon along with the other quadriceps femoris muscles that inserts on the patella. The patellar ligament, in turn, inserts onto the tibial tuberosity.
  • 6M. capsularis (articularis coxae) originates on the ilium, dorsal to the rectus femoris tuberosity and cranial to the acetabulum. It inserts onto the anterior femur just distal to the greater trochanter. In N. nebulosa, the insertion site was seen as an indentation on the bone. In F. catus, the site was smoother and less noticeable. On the femur, the insertion site was slightly larger for N. nebulosa (Table 2, Figs. 5B,C, 6(6), and 7B).
  • 7M. gemellus cranialis is located deep to M. gluteus superficialis and originates on the ischial spine which borders the origin of M. gluteus profundus caudally and M. gemellus caudalis cranially. The fibers cover the medial surface of the tendon of M. obturatorius internus. The muscle inserts into the trochanteric fossa along with M. gemellus caudalis and M. obturatorius internus. The origination site area on the pelvis was greater for F. catus than N. nebulosa; a more prominent ischial spine was also noted on the F. catus pelvis (Table 1, Figs. 5C, 6(7), and 8B).
  • 8M. gemellus caudalis originates on the dorsal border of the ischium between the ischial spine and the cranial border of the ischial tuberosity and is covered by M. gluteus superficialis and M. biceps femoris. Its fibers also surround the tendon of M. obturatorius internus. It extends to the proximal femur where it inserts into the trochanteric fossa. In F. catus, the origin of M. gemellus caudalis is thinner with more space between it and M. quadratus femoris than in N. nebulosa. (Figs. 5B,C, 6(8), and 8B).
  • 9M. biceps femoris covers a large area on the lateral aspect of the thigh and is located caudal to M. caudofemoralis and cranial to M. semitendinosus. It is considered to be part of the hamstring muscle group. It originates on the dorsal aspect of the ischial tuberosity and inserts onto the lateral margin of the patella and the proximal lateral surface of the proximal half of the length of the tibia by the crural fascia. A tendinous slip also joins M. biceps femoris to the calcaneal tendon. After reflecting M. biceps femoris and M. gluteus superficialis, a thin slip of muscle called M. abductor cruris caudalis (tenuissimus) can also be seen. This muscle inserts into M. biceps femoris at the halfway point on the thigh and into the crural fascia. The surface area for the origin on the pelvis was larger for F. catus [Table 1, Figs. 4A, 5A, and 6(9)].
  • 10M. semitendinosus originates on the ischial tuberosity between M. biceps femoris and M. semimembranosus. The muscle gives rise to a tendon that inserts on the medial surface of the tibial crest. Its surface area of origin was larger for N. nebulosa than F. catus (Table 1, Figs. 4, 5A, 6(10), and 9).
  • 11M. quadratus femoris is located deep to M. gluteus superficialis and M. biceps femoris. In N. nebulosa, it originates on the lateral surface of the ischium, bordering the origin of M. gemellus caudalis and the margin of the ischial tuberosity. The muscle scar at the origin appeared to have a higher degree of rugosity than in F. catus. In N. nebulosa, the muscle inserts just lateral to the lesser trochanter, while in F. catus the insertion site is more cranial to the lesser trochanter. The insertion sites in both species are not within the trochanteric fossa as reported previously (Rozenweig,1990), but rather are distal to it and distal to the insertions of Mm. obturatorius internus, gemelli, and obturatorius externus (Figs. 5, 6(11), and 8B).
  • 12M. semimembranosus is part of the hamstring muscle group originating on the ischial tuberosity caudal to the origin of M. semitendinosus. The muscle insertion has a femoral and tibial head with the femoral head inserting onto the medial side of linea aspera of the femur and on the medial epicondyle. The tibial head inserts upon the medial side of proximal tibia. (Figs. 4B, 5A,B, 6(12), 8, and 9).
  • 13M. obturatorius externus originates on the external bony margin of the obturator foramen next to the pubic symphysis and ventral to the origin of M. quadratus femoris. It inserts by a tendon into the trochanteric fossa of the femur at a point between the insertions of Mm. gemelli and obturatorius internus and M. quadratus femoris. This insertion site was larger for N. nebulosa than F. catus (Table 2, Figs. 5B,C, 6(13), and 8B).
  • 14M. gracilis is located on the medial thigh and originates as an aponeurotic attachment on the caudal third of the pubic symphysis. It has a thin insertion on the tibia, proximal and medial to the tibial crest. [Figs. 4B and 6(14)].
  • 15M. adductor femoris is formed by M. adductor brevis and M. adductor magnus. Both muscles arise from the pubic symphysis on either side of M. gracilis origin. M. adductor brevis originates from the cranial half, and M. adductor magnus from the entire span of the pubic symphysis. The muscles insert on the lateral side of the femur. M. adductor brevis extends along half the length of the linea aspera of the femur, while M. adductor magnus inserts onto the entire length of the of the linea aspera and is located more laterally than M. adductor brevis (Figs. 4B, 5A,B, 6(15), and 8B).
  • 16M. adductor longus is located on the medial side of the thigh. It arises from a small area on the pubis lateral and cranial to the pubic symphysis and inserts onto the posterolateral side of the linea aspera next to M. adductor brevis insertion. The muscle is proximal to M. adductor femoris and can be seen following the removal of M. sartorius and M. gracilis. The insertion site was larger for F. catus than N. nebulosa (Table 2, Figs. 4B, 6(16), and 8B).
  • 17M. pectineus arises from the pubis between M. iliopsoas and M adductor longus. The muscle inserts onto the proximal medial lip of linea aspera of the femur. The origin site was larger for N. nebulosa than F. catus, and the insertion site was larger for F. catus (Tables 1 and 2, Figs. 4B, 6(17), and 8B).
  • 18Mm. psoas minor and major originate from the ventral bodies and transverse processes of thoracic vertebrae 12 and 13 as well as all of the seven lumbar vertebrae. M. psoas major inserts with iliacus on the lesser trochanter of the femur. M. psoas minor inserts on a small tuberosity on the ventral border of the pelvis near the origins of Mm. iliacus and pectineus. M. psoas minor insertion on the pelvis was slightly larger for F. catus than N. nebulosa (Table 1, Figs. 4B, 6(19), and 8A).
  • 19M. iliacus originates from the ventral border of the ilium, between the origins for Mm. sartorius and pectineus. It inserts on the lesser trochanter along with M. psoas major (Figs. 4B, 6(19), and 8A).
  • 20M. obturatorius internus originates on the medial margin of obturator foramen. It covers the foramen opening and occupies about half the border of the foramen circumference. Because a clear view of the origin could not be photographed in the whole pelvis specimen for N. nebulosa (sawing the specimen in half was not an option), the origin site surface area could not be measured. The muscle gives rise to a tendon that passes through the lesser ischiatic foramen. As it approaches its insertion, the tendon is surrounded by M. gemellus cranialis and M. gemellus caudalis. All three muscles then insert together within the trochanteric fossa (Figs. 5A,C and 8B).
Figure 4.

Superficial views of N. nebulosa hip and thigh: (A) lateral view, (B) medial view.

Figure 5.

Deep muscles of N. nebulosa lateral thigh. (A) Mm. biceps femoris and caudofemoralis are reflected and M. gluteus medius has been removed. (B) Mm. vastus lateralis and semimembranosus are cut at their origins and reflected to show Mm. vastus intermedius, rectus femoris, and adductor femoris. (C) M. piriformis has been removed to show the hip rotators.

Figure 6.

Lateral pelvic muscle maps for Neofelis nebulosa, Felis catus, Canis latrans, and Vulpes vulpes. (1) gluteus medius, (2) gluteus profundus, (3) tensor fasciae latae, (4) sartorius, (5) rectus femoris, (6) capsularis, (7) gemellus cranialis, (8) gemellus caudalis, (9) biceps femoris, (10) semitendinosus, (11) quadratus femoris, (12) semimembranosus, (13) obturatorius externus, (14) gracilis, (15) adductor femoris, (16) adductor longus, (17) pectineus, (18) psoas minor, (19) iliacus.

Figure 7.

(A) Lateral and (B) anterior femoral muscle maps for Neofelis nebulosa and Felis catus.

Figure 8.

(A) Medial and (B) posterior femoral muscle maps for Neofelis nebulosa and Felis catus.

Figure 9.

Medial tibial muscle maps for (A) Neofelis nebulosa and (B) Felis catus.

Table 1. Comparisons of percentage surface area means of F. catus and N. nebulosa pelvic muscle attachment sites
MusclesF. catus(5)N. nebulosa(5)
  • N numbers for each species are listed in parentheses beside the name.

  • a

    Significant difference for F. catus and N. nebulosa (P ≤ 0.05).

  • b

    Significant difference for F. catus and N. nebulosa (P ≤ 0.001).

Pelvic muscle attachment sites (% mean ± SEM of surface area)
 Gluteus mediusa24.6 ± 0.426.3 ± 0.4
 Gluteus profundusb28.8 ± 0.226.3 ± 0.4
 Tensor fascia lataea8.6 ± 0.37.3 ± 0.2
 Articularis coxae3.4 ± 0.13.2 ± 0.1
 Gemellus cranialisb5.5 ± 0.24.7 ± 0.1
 Gemellus caudalis11.1 ± 0.310.2 ± 0.3
 Biceps femorisb11.1 ± 0.37.1 ± 0.2
 Sartorius10.7 ± 0.59.9 ± 0.4
 Iliacus14.0 ± 0.214.5 ± 0.5
 Rectus femorisa7.1 ± 0.27.8 ± 0.2
 Psoas minora2.8 ± 0.12.5 ± 0.1
 Pectineusa5.0 ± 0.15.5 ± 0.2
 Adductor longus6.5 ± 0.16.5 ± 0.1
 Adductor femoris19.8 ± 0.320.4 ± 0.6
 Obturatorius externus23.3 ± 0.324.3 ± 0.4
 Quadratus femoris14.3 ± 0.314.5 ± 0.1
 Semitendinosusb4.6 ± 0.37.1 ± 0.3
 Semimembranosus9.3 ± 0.29.5 ± 0.4
Table 2. Comparisons of percent surface area means of F. catus and N. nebulosa femoral muscle attachment sites
MusclesF. catus (4)N. nebulosa (4)
  • N numbers for each species are listed in parentheses beside the name.

  • a

    Significant difference for F. catus and N. nebulosa (P ≤ 0.05).

  • b

    Significant difference for F. catus and N. nebulosa (P ≤ 0.001).

Femoral muscle attachment sites (% Mean ± SEM of surface area)
 Gluteus profundus5.0 ± 0.14.4 ± 0.3
 Vastus lateralisb11.5 ± 0.216.3 ± 0.4
 Articularis coxaea7.0 ± 0.38.3 ± 0.2
 Vastus intermedius38.5 ± 1.038.2 ± 0.4
 Gluteus mediusa3.4 ± 0.14.9 ± 0.3
 Piriformis2.9 ± 0.23.3 ± 0.1
 Gluteus superficialisb3.9 ± 0.25.0 ± 0.1
 Vastus medialis24.7 ± 1.423.7 ± 0.8
 Obturatorius internus  and gemelli3.9 ± 0.24.3 ± 0.3
 Obturatorius externusb3.6 ± 0.054.3 ± 0.1
 Quadratus femoris5.7 ± 0.25.4 ± 0.2
 Pectineusa7.5 ± 0.16.9 ± 0.2
 Adductor longusa10.7 ± 0.29.7 ± 0.2
 Adductor femoris29.6 ± 0.728.7 ± 0.7
 Semimembranosus12.6 ± 0.610.9 ± 0.3

Muscles Originating on the Vertebral Column

  • 1M. caudofemoralis originates from the transverse processes of the second, third, and fourth caudal vertebrae. It inserts onto the lateral border of the patella. (Figs. 4A and 5A).
  • 2M. abductor cruris caudalis originates by two heads from the first and second caudal vertebrae. It is a thin muscle that inserts into the crural fascia along with M. biceps femoris (Fig. 5A).
  • 3M. gluteus superficialis is positioned cranial to M. caudofemoralis and originates from the spinous processes of the third sacral and first caudal vertebrae and from the thoracolumbar fascia. It inserts onto the lateral proximal side of the femur just distal to the greater trochanter. This insertion site was larger for N. nebulosa than F. catus (Table 2, Fig. 4A).
  • 4M. piriformis which is considered to be part of the gluteal muscle group is a triangular muscle that originates from the dorsal the transverse ridges of the sacrum. It passes through the greater sciatic foramen superficial to M. gluteus profundus. The sciatic nerve emerges deep to M. piriformis. It inserts on the greater trochanter proximal to the insertion of M. gluteus medius (Figs. 5A,C and 7A).

Muscles Originating on the Femur

  • 1M. vastus lateralis, located on the lateral side of the thigh, originates from a transverse line on the anterior femur just distal to the greater trochanter and extends to the lateral lip of linea aspera. The muscle joins the quadriceps femoris tendon and attaches to the lateral patella, then joining with the patellar ligament which inserts on the tibial tuberosity. The origin site was larger for N. nebulosa (Table 2; Figs. 4A, 5A,B, and 7).
  • 2M. vastus medialis originates on the anterior-medial femur over the entire length to the medial epicondyle. The muscle inserts onto the medial side of the patella (Figs. 4B, 7B, 8A, and 9).
  • 3M. vastus intermedius originates on the anterior femur and is deep to M. rectus femoris and lateral to M. vastus medialis. Its origin extends almost to the knee joint. The muscle inserts onto the proximal patella (Figs. 5B and 7).

Comparing Pelvic Muscle Attachment Areas Among C. latrans,V. vulpes, F. catus, and N. nebulosa

The pelvis is a relatively flat bone, that is, easily photographed to determine muscle attachment site surface area measurements as compared to the femur. Initial attachment site area measurements for F. catus and N. nebulosa indicated that they were more similar than expected, leading us to question whether homologous attachment site area measurements in more phylogenetically distant species might also be similar. Therefore, pelvic muscle attachment site area measurements were made in two nonfelids, C. latrans (n = 2) and V. vulpes (n = 1). The data obtained for the fox (V. vulpes) was from a single animal; therefore, no statistical analysis was performed on the data for this species. However, the results for V. vulpes appeared quite similar to that of C. latrans (Fig. 2).

Except for Mm. gemellus cranialis and biceps femoris, C. latrans differed significantly from F. catus and N. nebulosa in all cases (Fig. 2). The area for the M. gluteus medius site in C. latrans was much larger from that of N. nebulosa and F. catus. By contrast, Mm. gluteus profundus, tensor fasciae latae, articularis coxae, and gemellus caudalis originated from smaller attachment sites in C. latrans as compared to F. catus and N. nebulosa. M. biceps femoris origin site in F. catus was closer to the canids than to N. nebulosa.

Comparisons of Ilium and Ischium Lengths for F. catus and N. nebulosa

Because ilium and ischium lengths have been demonstrated to be biomechanically significant observations (Smith and Savage,1956; Anemone,1993), these measurements were taken and compared between N. nebulosa and F. catus. The ratio of ilium to ischium lengths was significantly higher for F. catus (1.61 ± 0.02) compared to N. nebulosa (1.40 ± 0.01). The ilium length based on the total ilium and ischium lengths was significantly greater for F. catus (0.62 ± 0.002) than for N. nebulosa (0.59 ± 0.0001), and the ischium was significantly longer for N. nebulosa (0.42 ± 0.0001) than for F. catus (0.39 ± 0.002).


This study provides, for the first time, a detailed description of hip and thigh muscle morphology in the endangered species N. nebulosa and compares it with that of F. catus. While the results show the overall morphology to be similar for the two felid species, muscle attachment areas showed some variation which may prove to be related, in part, to differences in locomotor habits exhibited by the two animals. By contrast, when felid and nonfelid species were compared, the differences were far greater in every respect, suggesting that muscle map surface area analysis may also be a useful phylogenetic tool for making interspecies comparisons.

Comparisons of the hindlimb muscle morphology between N. nebulosa and F. catus were surprisingly similar given that the phylogenetic distance between the two Felidae subgroups is fairly large. Phylogenetic analyses suggest that the Panthera lineage (N. nebulosa) diverged around 10.8 Ma as an early branch of the modern felids while the domestic cat lineage (Felis catus) branched subsequently between 6.7 and 6.2 Ma (Mattern and McLennan,2000; Johnson et al.,2006). Subtle differences were seen between the two subgroups. For example, N. nebulosa M. tensor fasciae latae has a more aponeurotic origin compared to the direct muscle attachment in F. catus and the M. rectus femoris origin was closer to the acetabulum in N. nebulosa and closer to the iliac crest in F. catus.

Muscle attachment surface areas were more varied in comparisons between F. catus and N. nebulosa. Several previous studies have examined variations in muscle attachment site areas in various types of carnivores and correlated them with locomotor characteristics (Taylor,1974; Anemone,1993; Wang,1993; Heinrich and Houde,2006). These studies, however, only observed and compared relative surface areas without calculating actual surface area measurements. In this study, muscle attachment site surface area differences may not only be related to differences in locomotion behavior and muscle biomechanics but also may be useful for making phylogenetic comparisons between different species.

It is possible that analyzing some of these muscle attachment sites in three dimensions may provide additional useful information. A flat bone viewed in a two-dimensional plane may not benefit from three-dimensional analysis. By contrast, a curved surface with a high degree of rugosity might benefit from this type of analysis. A three-dimensional analysis may also enable the assessment of the attachment site rugosity, which is believed to be a measure of the muscle's force (Bryant and Seymour,1990; Zumwault,2005). Work is currently in progress to further evaluate this type of analysis.

The N. nebulosa pelvis showed a larger muscle attachment surface area for M. rectus femoris compared with F. catus. In previous studies, comparisons of various carnivores with different locomotor behaviors have shown that arboreal species, such as the ringtail and the African palm civet, have a larger muscle attachment surface area for M. rectus femoris than those seen on the pelvises of more cursorial species (Heinrich and Rose,1997; Heinrich and Houde,2006). Thus, the rectus femoris muscle in N. nebulosa and other arboreal species may play an important role in climbing.

Mm. gluteus profundus and tensor fasciae latae muscle attachment sites were larger on the F. catus pelvis compared to N. nebulosa. This appears to be due to differences between the two species in the length of the ilium and ischium. F. catus has a longer ilium while N. nebulosa has a longer ischium. Previous studies have shown that muscles originating from a longer section of bone have longer muscle fibers (Anemone,1993). Longer muscles move the joint over a greater distance more rapidly than muscles with shorter fibers (Gowitzke and Milner,1980; Anapol,2004). Therefore, it is possible that in F. catus, M. gluteus profundus moves the hip joint more rapidly to abduct and rotate the thigh when compared to N. nebulosa.

Our initial analysis of muscle map surface areas showed a high degree of similarity between N. nebulosa and F. catus. This prompted us to wonder whether map areas for the same muscles would also be similar for more distantly related species. To test this we compared pelvic muscle maps for C. latrans and V. vulpes with those of the felids. The results show that while there were minor differences in the muscle attachment site surface areas observed between felids, the differences were greater when felids were compared to canids. In particular, the muscle surface areas for the origins of Mm. gluteus profundus, tensor fasciae latae, articularis coxae, and gemellus caudalis were considerably smaller, and the area for the gluteus medius muscle was much larger in the canids compared to felids. These results further demonstrate that the surface area analysis may be a useful tool for evaluating phylogenetic relationships.


The authors thank Charley Potter and John Ososky and the Smithsonian Institution for facilitating acquisition of the N. nebulosa specimens. They also thank Virginia Naples for the helpful comments during the preparation of this manuscript. Ian Johnson provided figure drawings. Jennifer Nevis at the Willowbrook Wildlife Center provided the coyote and fox specimens.