Genetic and genomic analyses of musculoskeletal differences between BEH and BEL strains.

Berlin high (BEH) and Berlin low (BEL) strains selected for divergent growth differ threefold in body weight. We aimed at examining muscle mass, which is a major contributor to body weight, by exploring morphological characteristics of the soleus muscle (fiber number and cross sectional area; CSA), by analyzing the transcriptome of the gastrocnemius and by initiating quantitative trait locus (QTL) mapping. BEH muscles were four to eight times larger than those of BEL. In substrain BEH+/+, mutant myostatin was replaced with a wild-type allele; however, BEH+/+muscles still were two to four times larger compared with BEL. BEH soleus muscle fibers were two times more numerous (P < 0.0001) and CSA was two times larger (P < 0.0001) compared with BEL. In addition, soleus femoral attachment anomaly (SFAA) was observed in all BEL mice. One significant (Chr 1) and four suggestive (Chr 3, 4, 6, and 9) muscle weight QTLs were mapped in a 21-day-old F2 intercross (n = 296) between BEH and BEL strains. The frequency of SFAA incidence in the F2 and in the backcross to BEL strain (BCL) suggested the presence of more than one causative gene. Two suggestive SFAA QTLs were mapped in BCL; however, their peak markers were not associated with the phenotype in F2. RNA-Seq analysis revealed 2,148 differentially expressed (P < 0.1) genes and 45,673 single nucleotide polymorphisms and >2,000 indels between BEH+/+ and BEL males. In conclusion, contrasting muscle traits and genomic and gene expression differences between BEH and BEL strains provide a promising model for the search for genes involved in muscle growth and musculoskeletal morphogenesis.

Stanniocalcin-2 inhibits skeletal muscle growth and is upregulated in functional overload-induced hypertrophy.

AIMS: Stanniocalcin-2 (STC2) has recently been implicated in human muscle mass variability by genetic analysis. Biochemically, STC2 inhibits the proteolytic activity of the metalloproteinase PAPP-A, which promotes muscle growth by upregulating the insulin-like growth factor (IGF) axis. The aim was to examine if STC2 affects skeletal muscle mass and to assess how the IGF axis mediates muscle hypertrophy induced by functional overload. METHODS: We compared muscle mass and muscle fiber morphology between Stc2-/- (n = 21) and wild-type (n = 15) mice. We then quantified IGF1, IGF2, IGF binding proteins -4 and -5 (IGFBP-4, IGFBP-5), PAPP-A and STC2 in plantaris muscles of wild-type mice subjected to 4-week unilateral overload (n = 14). RESULTS: Stc2-/- mice showed up to 10% larger muscle mass compared with wild-type mice. This increase was mediated by greater cross-sectional area of muscle fibers. Overload increased plantaris mass and components of the IGF axis, including quantities of IGF1 (by 2.41-fold, p = 0.0117), IGF2 (1.70-fold, p = 0.0461), IGFBP-4 (1.48-fold, p = 0.0268), PAPP-A (1.30-fold, p = 0.0154) and STC2 (1.28-fold, p = 0.019). CONCLUSION: Here we provide evidence that STC2 is an inhibitor of muscle growth upregulated, along with other components of the IGF axis, during overload-induced muscle hypertrophy.

Resolving candidate genes of mouse skeletal muscle QTL via RNA-Seq and expression network analyses.

BACKGROUND: We have recently identified a number of Quantitative Trait Loci (QTL) contributing to the 2-fold muscle weight difference between the LG/J and SM/J mouse strains and refined their confidence intervals. To facilitate nomination of the candidate genes responsible for these differences we examined the transcriptome of the tibialis anterior (TA) muscle of each strain by RNA-Seq. RESULTS: 13,726 genes were expressed in mouse skeletal muscle. Intersection of a set of 1061 differentially expressed transcripts with a mouse muscle Bayesian Network identified a coherent set of differentially expressed genes that we term the LG/J and SM/J Regulatory Network (LSRN). The integration of the QTL, transcriptome and the network analyses identified eight key drivers of the LSRN (Kdr, Plbd1, Mgp, Fah, Prss23, 2310014F06Rik, Grtp1, Stk10) residing within five QTL regions, which were either polymorphic or differentially expressed between the two strains and are strong candidates for quantitative trait genes (QTGs) underlying muscle mass. The insight gained from network analysis including the ability to make testable predictions is illustrated by annotating the LSRN with knowledge-based signatures and showing that the SM/J state of the network corresponds to a more oxidative state. We validated this prediction by NADH tetrazolium reductase staining in the TA muscle revealing higher oxidative potential of the SM/J compared to the LG/J strain (p<0.03). CONCLUSION: Thus, integration of fine resolution QTL mapping, RNA-Seq transcriptome information and mouse muscle Bayesian Network analysis provides a novel and unbiased strategy for nomination of muscle QTGs.

H55N polymorphism as a likely cause of variation in citrate synthase activity of mouse skeletal muscle.

Citrate synthase (CS) is an enzyme of the Krebs cycle that plays a key role in mitochondrial metabolism. The aim of this study was to investigate the mechanisms underlying low activity of citrate synthase (CS) in A/J mice compared with other inbred strains of mice. Enzyme activity, protein content, and mRNA levels of CS were studied in the quadriceps muscles of A/J, BALB/cByJ, C57BL/6J, C3H/HeJ, DBA/2J, and PWD/PhJ strains of mice. Cytochrome c protein content was also measured. The results of the study indicate that A/J mice have a 50-65% reduction in CS activity compared with other strains despite similar levels of Cs mRNA and lack of differences in CS and cytochrome c protein content. CS from A/J mice also showed lower Michaelis constant (K(m)) for both acetyl CoA and oxaloacetate compared with the other strains of mice. In silico analysis of the genomic sequence identified a nonsynonymous single nucleotide polymorphism (SNP) (rs29358506, H55N) in Cs gene occurring near the site of the protein interacting with acetyl CoA. Allelic variants of the polymorphism segregated with the catalytic properties of CS enzyme among the strains. In summary, H55N polymorphism in Cs could be the underlying cause of low CS activity and its high affinity for substrates in A/J mice compared with other strains. This SNP might also play a role in resistance to obesity of A/J mice.

Energetic limitations on suction feeding performance in centrarchid fishes.

Energetic analysis of ecologically relevant behaviors can be useful because animals are energetically limited by available muscle mass. In this study we hypothesized that two major determinants of suction feeding performance, the magnitudes of buccal volumetric expansion and subambient buccal pressure, would be correlated with, and limited by, available muscle mass. At least four individuals of three centrarchid species were studied: largemouth bass (Micropterus salmoides), bluegill (Lepomis macrochirus) and green sunfish (Lepomis cyanellus). Buccal pressure was measured directly via cannulation of the buccal cavity with a catheter-tipped pressure transducer. Buccal expansion was estimated from lateral high-speed video (500 or 1000 Hz) sequences and published data on internal kinematics of largemouth bass. These estimates were calibrated from silicone casts made of the buccal cavity post-mortem. Estimated work and power were found to be significantly correlated with muscle mass over all individuals. The slopes of these relationships, estimates of mass-specific muscle work and power, were found to be 11+/-2 J kg(-1) and 300+/-75 W kg(-1), respectively. These estimates are consistent with observations made of in vivo and in vitro muscle use and with digital particle image velocimetry measurements of water flow in feeding centrarchids. A direct trade-off between mean pressure and change in volume was observed, when the latter was normalized to muscle mass. We conclude that available muscle mass may be a useful metric of suction feeding performance, and that the ratio of muscle mass to buccal volume may be a useful predictor of subambient buccal pressure magnitude.

Mono- versus biarticular muscle function in relation to speed and gait changes: in vivo analysis of the goat triceps brachii.

The roles of muscles that span a single joint (monoarticular) versus those that span two (biarticular) or more joints have been suggested to differ. Monoarticular muscles are argued to perform work at a joint, whereas biarticular muscles are argued to transfer energy while resisting moments across adjacent joints. To test these predictions, in vivo patterns of muscle activation, strain, and strain rate were compared using electromyography and sonomicrometry in two major elbow extensors, the long and lateral heads of the triceps brachii of goats (Capra hircus), across a range of speed (1-5 m s(-1)) and gait. Muscle recordings were synchronized to limb kinematics using high-speed digital video imaging (250 Hz). Measurements obtained from four goats (25-45 kg) showed that the monoarticular lateral head exhibited a stretch-shortening pattern (6.8+/-0.6% stretch and -10.6+/-2.7% shortening; mean+/-s.e.m. for all speeds and gaits) after being activated, which parallels the flexion-extension pattern of the elbow. By contrast, the biarticular long head shortened through most of stance (-16.4+/-3.4%), despite elbow flexion in the first half and shoulder extension in the last half of stance. The magnitude of elbow flexion and shoulder extension increased with increasing speed (ANCOVA, P<0.05 and P<0.001), as did the magnitude and rate of active stretch of fascicles in the lateral head (P<0.001 for both). In all individuals, shortening fascicle strain rates increased with speed in the long head (P<0.001), and, in three of the four individuals, strain magnitude increased. Few independent effects of gait were found. In contrast to its expected function, the biarticular long head appears to produce positive work throughout stance, whereas the monoarticular lateral head appears to absorb work at the elbow. The biarticular anatomy of the long head may mitigate increases in muscle strain with speed in this muscle, because strain magnitude in the second phase of stance (when the shoulder extends) decreased with speed (P<0.05).

Differential muscle function between muscle synergists: long and lateral heads of the triceps in jumping and landing goats (Capra hircus).

We investigate how the biarticular long head and monoarticular lateral head of the triceps brachii function in goats (Capra hircus) during jumping and landing. Elbow moment and work were measured from high-speed video and ground reaction force (GRF) recordings. Muscle activation and strain were measured via electromyography and sonomicrometry, and muscle stress was estimated from elbow moment and by partitioning stress based on its relative strain rate. Elbow joint and muscle function were compared among three types of limb usage: jump take-off (lead limb), the step prior to jump take-off (lag limb), and landing. We predicted that the strain and work patterns in the monoarticular lateral head would follow the kinematics and work of the elbow more closely than would those of the biarticular long head. In general this prediction was supported. For instance, the lateral head stretched (5 +/- 2%; mean +/- SE) in the lead and lag limbs to absorb work during elbow flexion and joint work absorption, while the long head shortened (-7 +/- 1%) to produce work. During elbow extension, both muscles shortened by similar amounts (-10 +/- 2% long; -13 +/- 4% lateral) in the lead limb to produce work. Both triceps heads functioned similarly in landing, stretching (13 +/- 3% in the long head and 19 +/- 5% in the lateral) to absorb energy. In general, the long head functioned to produce power at the shoulder and elbow, while the lateral head functioned to resist elbow flexion and absorb work, demonstrating that functional diversification can arise between mono- and biarticular muscle agonists operating at the same joint.

Functional morphology of prey capture in the sturgeon, Scaphirhynchus albus.

Acipenseriformes (sturgeon and paddlefish) are basal actinopterygians with a highly derived cranial morphology that is characterized by an anatomical independence of the jaws from the neurocranium. We examined the morphological and kinematic basis of prey capture in the Acipenseriform fish Scaphirhynchus albus, the pallid sturgeon. Feeding pallid sturgeon were filmed in lateral and ventral views and movement of cranial elements was measured from video sequences. Sturgeon feed by creating an anterior to posterior wave of cranial expansion resulting in prey movement through the mouth. The kinematics of S. albus resemble those of other aquatic vertebrates: maximum hyoid depression follows maximum gape by an average of 15 ms and maximum opercular abduction follows maximum hyoid depression by an average of 57 ms. Neurocranial rotation was not a part of prey capture kinematics in S. albus, but was observed in another sturgeon species, Acipenser medirostris. Acipenseriformes have a novel jaw protrusion mechanism, which converts rostral rotation of the hyomandibula into ventral protrusion of the jaw joint. The relationship between jaw protrusion and jaw opening in sturgeon typically resembles that of elasmobranchs, with peak upper jaw protrusion occurring after peak gape.

Constitutive expression of Yes-associated protein (Yap) in adult skeletal muscle fibres induces muscle atrophy and myopathy.

The aim of this study was to investigate the function of the Hippo pathway member Yes-associated protein (Yap, gene name Yap1) in skeletal muscle fibres in vivo. Specifically we bred an inducible, skeletal muscle fibre-specific knock-in mouse model (MCK-tTA-hYAP1 S127A) to test whether the over expression of constitutively active Yap (hYAP1 S127A) is sufficient to drive muscle hypertrophy or stimulate changes in fibre type composition. Unexpectedly, after 5-7 weeks of constitutive hYAP1 S127A over expression, mice suddenly and rapidly lost 20-25% body weight and suffered from gait impairments and kyphosis. Skeletal muscles atrophied by 34-40% and the muscle fibre cross sectional area decreased by ≈40% when compared to control mice. Histological analysis revealed evidence of skeletal muscle degeneration and regeneration, necrotic fibres and a NADH-TR staining resembling centronuclear myopathy. In agreement with the histology, mRNA expression of markers of regenerative myogenesis (embryonic myosin heavy chain, Myf5, myogenin, Pax7) and muscle protein degradation (atrogin-1, MuRF1) were significantly elevated in muscles from transgenic mice versus control. No significant changes in fibre type composition were detected using ATPase staining. The phenotype was largely reversible, as a cessation of hYAP1 S127A expression rescued body and muscle weight, restored muscle morphology and prevented further pathological progression. To conclude, high Yap activity in muscle fibres does not induce fibre hypertrophy nor fibre type changes but instead results in a reversible atrophy and deterioration.

Scaling of in vivo muscle velocity during feeding in the largemouth bass, Micropterus salmoides (Centrarchidae).

Many vertebrates undergo large increases in body size over the course of a lifetime, and these increases are often accompanied by changes in morphological and physiological parameters. For instance, in most animals, increases in size with growth are accompanied by decreases in the maximum speed of shortening (V(max)) in locomotor muscles. Curiously, in muscles involved in suction feeding, V(max) shows no decreases with size in vitro, despite the fact that timing of kinematic events involved in suction feeding (e.g., time to peak gape) slow with increased size. The goal of this study was to examine whether muscular speed in vivo varies with size during suction feeding in the largemouth bass (Micropterus salmoides). The dorsal epaxial musculature of 10 individual bass (varying from 123 to 685 g and from 18.1 to 32.0 cm standard length [SL]) was implanted with sonometric crystals to measure muscle length during feeding on elusive prey (large goldfish). No relationship was found between the mean individual or maximum speed of shortening with mean individual log-transformed SL. However, mean magnitude of shortening and maximum shortening magnitude showed nonsignificant increases with SL ([Formula: see text] and 0.06, respectively). Average duration of shortening was found to increase with log-transformed SL. The size invariance of observed shortening velocity in the epaxial muscles during feeding may stem from size invariance of imposed loads during suction feeding. This is in contrast to what is normally seen in locomotor systems where loads on muscles often increase with body size.

QTL Analysis of Type I and Type IIA Fibers in Soleus Muscle in a Cross between LG/J and SM/J Mouse Strains.

Properties of muscle fibers, i.e., their type, number and size, are important determinants of functional characteristics of skeletal muscle, and of the quality of meat in livestock. Genetic factors play an important role in determining variation in fiber properties, however, specific genes remain largely elusive. We examined histological properties of soleus muscle fibers in two strains of mice exhibiting a twofold difference in muscle mass, LG/J and SM/J, and their F2 intercross. The total number of muscle fibers (555 ± 106; mean ± SD) did not differ between the strains or between males and females. A higher percentage of type I fibers was observed in the LG/J compared to the SM/J strain (P < 0.001) in both males (45 ± 3 vs. 37 ± 4%) and females (58 ± 4 vs. 41 ± 3%). Across strains, females had a higher percentage of type I fibers than males (P < 0.001), and the sex effect was greater in the LG/J strain (strain-by-sex interaction, P < 0.001). The cross-sectional area (CSA) did not differ between type I and type IIA fibers, but was greater in the LG/J than the SM/J strain (1365 ± 268 vs. 825 ± 229 µm(2), P < 0.001). Three significant quantitative trait locus (QTL) affecting CSA for type I and type IIA fibers mapped to chromosomes (Chr) 1, 6, and 11 and three suggestive QTL for percentage of type I fibers mapped to Chr 2, 3, and 4. Within each significant QTL, regions of conserved synteny were also implicated in variation of similar traits in an analogous study in pigs. Our results provide the evidence that the intercross between the SM/J and LG/J strains is a promising model to search for genes affecting muscle fiber properties.

Feeding muscles scale differently from swimming muscles in sunfish (Centrarchidae).

The physiological properties of vertebrate skeletal muscle typically show a scaling pattern of slower contractile properties with size. In fishes, the myotomal or swimming muscle reportedly follows this pattern, showing slower muscle activation, relaxation and maximum shortening velocity (V(max)) with an increase in body size. We asked if the muscles involved in suction feeding by fishes would follow the same pattern. We hypothesized that feeding muscles in fishes that feed on evasive prey are under selection to maintain high power output and therefore would not show slower contractile properties with size. To test this, we compared contractile properties in feeding muscles (epaxial and sternohyoideus) and swimming muscle (myotomal) for two members of the family Centrarchidae (sunfish): the bluegill (Lepomis macrochirus) and the largemouth bass (Micropterus salmoides). Consistent with our predictions, the V(max) of myotomal muscle in both species slowed with size, while the epaxials showed no significant change in V(max) with size. In the sternohyoideus, V(max) slowed with size in the bluegill but increased with size in the bass. The results indicate that scaling patterns of contractile properties appear to be more closely tied to muscle function (i.e. locomotion versus feeding) than overall patterns of size.

Suction feeding mechanics, performance, and diversity in fishes.

Despite almost 50 years of research on the functional morphology and biomechanics of suction feeding, no consensus has emerged on how to characterize suction-feeding performance, or its morphological basis. We argue that this lack of unity in the literature is due to an unusually indirect and complex linkage between the muscle contractions that power suction feeding, the skeletal movements that underlie buccal expansion, the sharp drop in buccal suction pressure that occurs during expansion, the flow of water that enters the mouth to eliminate the pressure gradient, and the forces that are ultimately exerted on the prey by this flow. This complexity has led various researchers to focus individually on suction pressure, flow velocity, or the distance the prey moves as metrics of suction-feeding performance. We attempt to integrate a mechanistic view of the ability of fish to perform these components of suction feeding. We first discuss a model that successfully relates aspects of cranial morphology to the capacity to generate suction pressure in the buccal cavity. This model is a particularly valuable tool for studying the evolution of the feeding mechanism. Second, we illustrate the multidimensional nature of suction-feeding performance in a comparison of bluegill, Lepomis macrochirus, and largemouth bass, Micropterus salmoides, two species that represent opposite ends of the spectrum of performance in suction feeding. As anticipated, bluegills had greater accuracy, lower peak flux into the mouth, and higher flow velocity and acceleration of flow than did bass. While the differences between species in accuracy of strike and peak water flux were substantial, peak suction velocity and acceleration were only about 50% higher in bluegill, a relatively modest difference. However, a hydrodynamic model of the forces that suction feeders exert on their prey shows that this difference in velocity is amplified by a positive effect of the smaller mouth aperture of bluegill on force exerted on the prey. Our model indicates that the pressure gradient in front of a fish that is feeding by suction, associated with the gradient in water velocity, results in a force on the prey that is larger than drag or acceleration reaction. A smaller mouth aperture results in a steeper pressure gradient that exerts a greater force on the prey, even when other features of the suction flow are held constant. Our work shows that some aspects of suction-feeding performance can be determined from morphology, but that the complexity of the behavior requires a diversity of perspectives to be used in order to adequately characterize performance.

Muscle function and power output during suction feeding in largemouth bass, Micropterus salmoides.

Muscle power output is thought to limit suction feeding performance, yet muscle power output during suction feeding has never been directly measured. In this study, epaxial activation and strain, hyoid depression, and intra-oral pressure were simultaneously measured during suction feeding in the largemouth bass (Micropterus salmoides). A mechanical model of muscle force transmission between the neurocranium and oral cavity was used to estimate muscle stress, work, and power. The epaxials shortened from rest an average of 9% of their length, with the highest efforts producing greater than 20% strain. Onset of shortening was simultaneous with or shortly after (< 10 ms) onset of activation. Maximal net power for individual fish ranged from 17 to 137 W kg(-1). Muscle power was significantly correlated with rectified EMG area (r = 0.80; p < 0.0001). The power required for cranial expansion was significantly correlated with epaxial power (r = 0.81; p < 0.0001), and the power exponent of this relationship ( approximately 1 for 3 of the 4 fish) implies that epaxial power accounts for most of the power of cranial expansion. The limitations imposed by the kinematic requirements and loading environment of suction feeding (short delay between activation and strain, maximal stress occurring after shortening, operation at lengths shorter than resting length) may prevent maximal muscular power production.

In vitro estimates of power output by epaxial muscle during feeding in largemouth bass.

Recent work has employed video and sonometric analysis combined with hydrodynamic modeling to estimate power output by the feeding musculature of largemouth bass in feeding trials. The result was an estimate of approximately 69 W kg(-1) of power by the epaxial muscle during maximal feeding strikes. The present study employed in vitro measurements of force, work and power output by fast-twitch epaxial muscle bundles stimulated under activation conditions measured in vivo to evaluate the power output results of the feeding experiments. Isolated muscle bundles from the epaxial muscle, the sternohyoideus and the lateral red or slow-twitch muscle were tied into a muscle mechanics apparatus, and contractile properties during tetanic contractions and maximum shortening velocity (Vmax) were determined. For the epaxial muscles, work and power output during feeding events was determined by employing mean stimulation conditions derived from a select set of maximal feeding trials: 17% muscle shortening at 3.6 muscle lengths/s, with activation occurring 5 ms before the onset of shortening. Epaxial and sternohyoideus muscle displayed similar contractile properties, and both were considerably faster (Vmax approximately 11-13 ML s(-1)) than red muscle (Vmax approximately 5 ML s(-1)). Epaxial muscle stimulated under in vivo activation conditions generated approximately 60 W kg(-1) with a 17% strain and approximately 86 W kg(-1) with a 12% strain. These values are close to those estimated by hydrodynamic modeling. The short lag time (5 ms) between muscle activation and muscle shortening is apparently a limiting parameter during feeding strikes, with maximum power found at an offset of 15-20 ms. Further, feeding strikes employing a faster shortening velocity generated significantly higher power output. Power production during feeding strikes appears to be limited by the need for fast onset of movement and the hydrodynamic resistance to buccal expansion.

Ontogeny of suction feeding capacity in snook, Centropomus undecimalis.

The ontogeny of suction feeding performance, as measured by peak suction generating capacity, was studied in the common snook, Centropomus undecimalis. Suction pressure inside the buccal cavity is a function of the total expansive force exerted on the buccal cavity distributed across the projected area of the buccal cavity. Thus, the scaling exponent of peak suction pressure with fish standard length was predicted to be equal to the scaling exponent of sternohyoideus muscle cross-sectional area, found to be 1.991, minus the scaling exponent for the projected buccal cavity area, found to be 2.009, equal to -0.018. No scaling was found in peak suction pressure generated by 12 snook ranging from 94 to 314 mm SL, supporting the prediction from morphology. C. undecimalis are able to generate similar suction pressures throughout ontogeny.

Muscle activation and strain during suction feeding in the largemouth bass Micropterus salmoides.

Activation and strain in the sternohyoideus (SH) were measured in vivo in five largemouth bass Micropterus salmoides. The SH is thought to actuate lower jaw depression, hyoid depression and suspensorial abduction during suction feeding in teleost fish. Sonomicrometry was used to measure fascicle shortening and lower jaw kinematics, while activity was measured by electromyography (EMG). SH fascicles shortened by an average of 11% during suction feeding. In three fish SH fascicles consistently shortened during fast lower jaw depression, but in two individuals they contracted isometrically or lengthened slightly during fast lower jaw depression. The SH continued shortening after peak gape, presumably actuating hyoid depression and lateral expansion of the buccal cavity. Onset of SH relengthening and onset of lower jaw elevation were simultaneous, as were the return of the SH to resting length and gape closure. Activation followed the onset of shortening by an average of 23 ms, although the muscle was active an average of 15 ms before the onset of rapid shortening. SH fascicles reached sustained shortening velocities averaging -2.5 fascicle lengths per second, and generally increased shortening velocity after peak gape. The shortening velocities measured in this study suggest that the SH actively shortens to generate power during suction feeding. This study is the first direct measurement of in vivo muscle function during suction feeding, the most common mechanism of prey capture among aquatic vertebrates.

Morphology predicts suction feeding performance in centrarchid fishes.

Suction feeding fish differ in their capacity to generate subambient pressure while feeding, and these differences appear to relate to morphological variation. We developed a morphological model of force transmission in the fish head and parameterized it with measurements from individual fish. The model was applied to 45 individuals from five species of centrarchid fishes: Lepomis macrochirus, Lepomis punctatus, Lepomis microlophus, Micropterus salmoides and Pomoxis nigromaculatus. Measurements of epaxial cross-sectional area, epaxial moment arm, buccal area and buccal area moment arm were combined to estimate pressure generation capacity for individual fish. This estimation was correlated with pressure measured in fish feeding on elusive prey to test the model's ability to predict pressure generation from morphology. The model explained differences in pressure generation found among individuals (P<0.001, r2=0.71) and produced a realistic estimate of normalized muscle stress during suction feeding (68.5+/-6.7 kPa). Fish with smaller mouths, larger epaxial cross-sectional area and longer epaxial moments, such as L. macrochirus (bluegill sunfish), generated lower pressures than fish with larger mouths, smaller cross-sectional area and shorter moments, such as M. salmoides (largemouth bass). These results reveal a direct trade-off between morphological requirements of feeding on larger prey (larger mouth size relative to body depth) and the ability to generate subambient pressure while suction feeding on elusive prey.