Eighteen L. inconspicuous accessions collected from different countries were evaluated for variations of seed weight, seed protein content, and electrophoretic patterns of the total seed proteins analyzed under reducing conditions. They exhibited a reasonable genetic variability for the evaluated traits. This genetic variability revealed that improvement through simple selection for these traits is possible. The variation between the seed size of this accessions was attributed to the development process or the life cycle of the plant, and the environmental condition to which the mother plant is exposed. On the other hand, the variation in protein content among the different accessions may be due to genotype and/or seasonal influences. The relationship between protein content and 100 seeds weight in the evaluated accessions was reversible, the accession showed the lowest quantity of the total seed proteins was the accession that exhibited highest weight of 100 seeds and nearly vice versa. Each accession gave a different electrophoretic pattern except the two accessions collected from Iran, exhibited an identical one. The difference in 100 seed weight and total protein content of these accessions indicated that they are not genetically identical. The variation in the electrophoregram of the evaluated accessions located in the bands with molecular weight more than 98 kDa, the heavy subunits of alpha-lathyrin subunits and the region molecular weight around 70 kDa. The results of cluster analysis based of SDS/PAGE under reduction conditions indicated that genetic diversity between Turkish, Syrian, and Iranian and Australian accessions is pronounced, and Turkish accessions are closer to both Syrian and Iranian accessions than the relation between Syrian and Iranian. This suggested that crosses between the Iranian and Syrian accessions could create more genetic variability than crosses with Turkish accessions. The distribution of Turkish and Syrian accessions between more than one clusters revealed that genetic diversity and geographic distribution were independent of each other. PCA showed that all accessions were separated on the first principal component, indicating that the accessions showed a good association, due, probably, to parallel evolution.