مطالعه ضرايب همبستگي صفت‌هاي زراعي و تجزيه عليت عملكرد دانه در چاودار

Study of Correlation Coefficients of Agronomic Traits and Path Analysis of Seed Yield in Rye

چاودار (Secale cereale L.) يكي از گرامينه‌هاي مهم و با ارزش است كه براي توليد نان يا مصرف علوفه مورد استفاده قرار مي‌گيرد. به ‌منظور بررسي روابط بين عملكرد دانه و اجزاء عملكرد، 64 ژنوتيپ مختلف چاودار در آزمايش لاتيس ساده با دو تكرار در سال 1395 در منطقه مراغه مورد ارزيابي قرار گرفتند. تجزيه همبستگي نشان داد كه عملكرد دانه با صفت‌هاي وزن سنبله، وزن دانه سنبله، شاخص برداشت سنبله، تعداد دانه در سنبله، تعداد سنبله در واحد سطح و شاخص برداشت همبستگي مثبت و معني‌دار داشت. با استفاده از تجزيه رگرسيوني گام‌به‌گام سهم هر يك از صفت‌هايي كه بيشترين تأثير را در عملكرد دانه داشتند مشخص گرديد صفت‌هاي وزن دانه سنبله و تعداد سنبله در واحد سطح به مدل رگرسيوني عملكرد دانه واردشده و 75 درصد از تنوع موجود بين ژنوتيپ‌ها را تبيين كردند. بر اساس تجزيه عليت نيز دو صفت مذكور به‌ترتيب با ضرايب 0/71 و 0/60 بيشترين تأثير مستقيم را روي عملكرد دانه داشتند. از طرف ديگر وزن سنبله بيشترين تأثير مستقيم و غيرمستقيم را (از طريق افزايش وزن دانه سنبله) بر افزايش عملكرد دانه داشت. به‌طوركلي از اين مطالعه مشخص شد كه دو صفت وزن دانه سنبله و تعداد سنبله در واحد سطح در اولويت اول و صفت‌هاي مرتبط با ويژگي‌هاي سنبله، وزن هزار دانه، قطر ساقه، شاخص برداشت، طول ريشك و ارتفاع گياه در الويت بعدي بر افزايش عملكرد دانه مؤثر مي‌باشند و مي‌توان از آن‌ها به‌عنوان شاخص‌هايي براي انتخاب در جهت بهبود عملكرد دانه استفاده نمود.

Background and Objectives Rye (Secale cereale L.) is one of the important and valuable Gramineae that is used for bread making and fed in some areas. Identifying correlation coefficients among the measured traits is very important and aids to evaluating the magnitude of interrelationships between how traits can facilitate the application of indirect selection in plant breeding programs. The target of this research was estimating the correlation coefficients, and path analysis among some important traits for selection criteria for improving grain yield in rye. Materials and Methods In order to evaluate the relationships between seed yield and its components, 64 different rye genotypes were evaluated in simple lattice design with two replications during 2016 in Maragheh region. Phenotypic correlation coefficients were calculated for ten characters during the growing season and the correlation coefficients between grain yield and different characters were subjected to path coefficient analysis separately for partitioning these values into direct and indirect effects. Stepwise regression technique was used to obtain the best fitted model as well as to estimate direct and indirect effects via path coefficient analysis. Applying bootstrap method verified the obtained results from conventional method. Results Correlation analysis showed that seed yield was associated significantly and positively with spike weight, weight of seeds per spike, harvest index of spike, seed number per spike, number of spike per area and harvest index traits. Using stepwise regression analysis, the share of the most important traits was determined and weight of seeds per spike and number of spike per area entered into regression model and explained 75 percent of the existing variation among genotypes. The results derived from path analysis indicated that, two above mentioned traits had significant effects on seed yield with 0.71 and 0.60 coefficients, respectively. On the other hand, spike weight had the most direct and indirect effects through increasing weight of seeds per spike on seed yield. Discussion In general, from this study we found that two traits including weight of seeds per spike and number of spike per area as the first order and related traits to spike, thousand seed weight, stem diameter, harvest index, awn length and plant height were in the second order for increasing seed yield and could be used as the selection indices for improving seed yield. Also, the used method for obtaining path coefficients was proved useful in analyzing correlation coefficients in breeding programs for evaluation of interrelated variables.