The purpose of the present study is to demonstrate that the use of an electron applicator with energy-dependent source-to-collimator distances (SCDs) will significantly improve the dose homogeneity for abutted electron fields in segmented-field electron conformal therapy (ECT). Multiple Coulomb scattering theory was used to calculate and study the P(80-20) penumbra width of off-axis dose profiles as a function of air gap and depth. Collimating insert locations with air gaps (collimator-to-isocenter distance) of 5.0, 7.5, 11.5, 17.5 and 19.5 cm were selected to provide equal P(80-20) at a depth of 1.5 cm in water for energies of 6, 9, 12, 16 and 20 MeV, respectively, for a Varian 2100EX radiation therapy accelerator. A 15 x 15 cm(2) applicator was modified accordingly, and collimating inserts used in the variable-SCD applicator for segmented-field ECT were constructed with diverging edges using a computer-controlled hot-wire cutter, which resulted in 0.27 mm accuracy in the abutted edges. The resulting electron beams were commissioned for the pencil-beam algorithm (PBA) on the Pinnacle(3) treatment planning system. Four hypothetical planning target volumes (PTVs) and one patient were planned for segmented-field ECT using the new variable-SCD applicator, and the resulting dose distributions were compared with those calculated for the identical plans using the conventional 95 cm SCD applicator. Also, a method for quality assurance of segmented-field ECT dose plans using the variable-SCD applicator was evaluated by irradiating a polystyrene phantom using the treatment plans for the hypothetical PTVs. Treatment plans for all four of the hypothetical PTVs using the variable-SCD applicator showed significantly improved dose homogeneity in the abutment regions of the segmented-field ECT plans. This resulted in the dose spread (maximum dose-minimum dose), sigma, and D(90-10) in the PTV being reduced by an average of 32%, 29% and 32%, respectively. Reductions were most significant for abutted fields of nonadjacent energies. Planning segmented-field ECT using the variable-SCD applicator for a patient with recurrent squamous cell carcinoma of the left ear showed the dose spread, sigma, and D(90-10) of the dose distribution in the PTV being reduced by an average of 38%, 22% and 22%, respectively. The measured and calculated dose in a polystyrene phantom resulting from the variable-SCD, segmented-field ECT plans for the hypothetical PTVs showed good agreement; however, isolated differences between dose calculation and measurement indicated the need for a more accurate dose algorithm than the PBA for segmented-field ECT. These results confirmed our hypothesis that using the variable-SCD applicator for segmented-field ECT results in the PTV dose distribution becoming more homogenous and being within the range of 85-105% of the 'given dose'. Clinical implementation of this method requires variable-SCD applicators, and the design used in the present work should be acceptable, as should our methods for construction of the inserts. Dose verification measurements in a polystyrene phantom and the recommended improvements in dose calculation should be appropriate for quality assurance of segmented-field ECT.