We report on the effect on performance of varying the length of the capillary during throughout in-capillary derivatization (TICD) capillary electrophoresis (CE). Performance was evaluated by on-line coupling with a sample and CE runbuffer loading device that was newly introduced for this study. The device was assembled with a low cost using two 5 mm inner diameter (ID) disposable polyethylene syringes. First, a sequence was manually formed consisting of a 200 muL run buffer solution plug, a 100 muL sample plug and another 200 muL run buffer solution plug. Each plug was separated from its neighbor by a 100 muL air plug. When each plug reached the injection point where both a platinum-wire anode and the end of the separation capillary tube were located, 340 V/cm separation voltage (electrophoresis voltage) and 34 V/cm injection voltage were applied to the capillary for 3 s. Then the analytes were derivatized during migration in 50 mum ID capillaries filled with 2 mm o-phthalaldehyde (OPA)/N-acetylcysteine (NAC) in a 20 mm phosphate-borate buffer (pH 10), followed by separating and detecting of OPA derivatives by absorbance of 340 nm. Derivatization, separation, and detection were performed systematically using capillaries which varied in length from 5 to 80 cm. In the case of TICD-CE of a mixture containing 1 mm aspartic acid (Asp) and 20 mm m-nitorophenol (MNP) as a test solution, it was determined that peak area and peak width ratios of Asp to MNP did not depend on capillary length. Enantiomeric separations Of DL-alanine (Ala) and Asp were examined using a run buffer consisting of a 45 mum beta-cyclodextrin (CD) - 2 mm OPAMAC - 20 mm phosphate-borate buffer (pH 10). Even though the resolution of these enantiomeric pairs decreased with decreasing capillary length, as expected, the peaks corresponding to both enantiomeric amino acids were identified even when a 5 cm capillary was used. An 8-component amino acid mixture was also tested with 5 cm and 10 cm capillaries.