Grass pea (Lathyrus sativus L.) is one of the essential food legumes in some developing countries with great forage production potential. The main goal of this study was to evaluate the influence of different sowing dates on yield and yield components of various grass pea ecotypes i n the semi -arid region of Iran. Three years experiments (2009 to 2011) were performed through factorial based on randomized complete block design with three replications. Three levels of sowing dates included: mid April (S1), early May (S2) and end of May (S3) and four ecotypes of grass pea which originated from different regions (Syria (E1), Ethiopia (E2), Oshnaveye (local ecotype) (E3) and Mashhad (local ecotype) (E4)) were employed in this study. Pod number per ground m-2 , pod weight per ground m - 2 , seed number per pod and thousand seed weight were measured in this study. The results showed that all study parameters were influenced by different sowing dates and grass pea ecotypes. Second sowing date resulted in the highest biological (680 g m - 2 in 2010) a nd seed yield (254 g m -2 in 2011). There was significant negative correlation between average maximum temperatures and biological and seed yield of grass pea across all ecotypes, years and locations. Syria and Mashhad ecotypes achieved highest biological a nd seed yield across study ecotypes under first and second sowing dates. In conclusion, insignificant delay in grass pea sowing date for avoiding lower temperatures and higher temperatures during early and end of growth period could be suitable pattern, fu rther to selection of appropriate ecotype to achieve highest yield

Keywords