Over the period 1982 to 1988, 31 consecutive patients at the Hand Surgery Unit of the Sheba Medical Centre were subjected to elbow joint arthrolysis to treat restriction of range of motion solely due to trauma. This retrospective study aims to evaluate the relative influence of the followings factors on functional outcome : sex, age, type of original injury and initial management, presence of para-articular ossification, delay between injury and arthrolysis, and the use of manipulation and a continuous passive motion device (CPM) following surgery. The range of motion was recorded prior to arthrolysis and after operation (excluding one patient who subsequently underwent arthrodesis for intractable pain). Follow-up averaged 15.3 months (± 5.4). In the 24 patients with extension deficit (> 20°), the mean improvement was of 26.9° (> 23.1°); in the 21 patients with flexion deficit the mean improvement was of 21.2° (> 18°). The mean improvement for total range of motion in the series overall was of 35.2° (± 23.8°). 90 % showed an improvement of at least 10° and 30 % attained normal ROM. All of these improvements in range were statistically highly significant (p < 0.0001). None of the variables had predictive value with regard to improvement of flexion. With regard to improvement in extension, the only variable of value was the use of a continuous passive motion device following surgery ; those patients subjected to CPM showed a mean improvement of 32.6° (± 19.0°), while those without averaged 12.8° (± 27.5°) (p < 0.01). Respective rates of improvement beyond 10° were 88.2 % vs. 28.6 %, while the respective incidences of patients attaining normal extension were 64.7 % vs. 14.3 % (p = 0.03).