Background: Physical therapists assess joint movement by observation and palpation. New imaging technologies that enable vision of bones and joints during functional activities can be used to analyze joint kinematics and review traditional assumptions. Purpose: The purpose was to measure relationships between flexion, rotation, and translation at the knee and to validate this visually. Design: This was a prospective, observational study. Methods: Twenty-five healthy participants over 45 years of age (13 males) knelt from upright kneeling to full flexion with the foot free. Fluoroscopy recorded movement at 30 frames per second of x-ray. A CT scan provided 3-dimensional data which was registered to the fluoroscopy frames to provide a moving model. Motion in 6-degrees-of-freedom was analyzed for coupling of movements. Results: Participants reached 142 (6)° flexion in kneeling (mean [SD]). Posterior femoral translation was coupled to flexion (r = 0.96). From 90° to 150° flexion, the femur translated posteriorly by 36 (3) mm to finish 23 (3) mm posterior to the center of the tibia at 150° flexion. From 90° to 150° flexion, the femur externally rotated from 8 (6)° to 16 (5)°. Flexion was coupled to rotation (r = 0.47). Abduction was less than 3° and lateral translation was less than 3mm. Visually, the femur appeared to translate posteriorly until the femoral condyles rested on the posterior rim of the tibial plateau, with concurrent external rotation so that the popliteal fossa aligned with the posterior margin of the medial tibial plateau. Limitations: A limitation of the study is that knee flexion can include squat and lunge as well as kneeling. Conclusion: Deep flexion requires femoral posterior translation and external rotation. These findings invite review of the concave-convex rule as it may apply to manual therapy of the knee.