MOORE AND STEINS separated the amino acids obtained from protein hydrolysates by column chromatography using a succession of buffers with different ~1-1 and salt concentrations. In physiological solutions many more ninhydrin positive substances than the usual zo protein amino acids may be' present and HAWLTON~ using a similar system to SPACIWRN et al. established the time of elution of 140 ninhydrin-positive substances. A number of these substances were unresolved from each other.
.4n alternative method of elution is to use a continuous gradient of pH and salt concentration. Multichambered devices (Varigrads) provide a convenient method for the production of such gradients. The theory and practice of these devices has been considered. by PETERSON AND SOBER 3. A system using such a gradient has been developed by the Technicon Co. but this leaves many substances unseparated from each other,
In some circumstances it may be necessary to improve the separation in regions containing substances of interest without losing resolution in other regions. It seemed reasonable to approach this problem by decreasing the slope of the gradient immediately preceding the elution of such regions without seriously interfering with the rest of the gradient.
Hence we have the problem of altering the initial concentration in the chambers in order to achieve the "desired" gradient. The direct algebraic solution, using data from PETERSON AND SOBERS was attempted but was found to be laborious and often to lead to negative values for some of the initial concentrations.
Another method was to make a judicious guess at the initial concentrations, calculate the gradient which would result and compare it with the one desired. A series of such guesses and comparisons should enable us to reach the desired curve and only positive concentrations will be tried. IIowever, since several chambers contribute at any one time an accurate fit by this process will be laborious. If we could remove the labour from the calculation stage of the guessing method it would be much more useful.