Sample size for morphological traits of pigeonpea

<p>The objectives of this study were to determine the sample size (i.e., number of plants) required to accurately estimate the average of morphological traits of pigeonpea (<em>Cajanus cajan </em>L.) and to check for variability in sample size between evaluation periods and seasons. Two uniformity trials (i.e., experiments without treatment) were conducted for two growing seasons. In the first season (2011/2012), the seeds were sown by broadcast seeding, and in the second season (2012/2013), the seeds were sown in rows spaced 0.50 m apart. The ground area in each experiment was 1,848 m2, and 360 plants were marked in the central area, in a 2 m × 2 m grid. Three morphological traits (e.g., number of nodes, plant height and stem diameter) were evaluated 13 times during the first season and 22 times in the second season. Measurements for all three morphological traits were normally distributed and confirmed through the <em>Kolmogorov-Smirnov </em>test. Randomness was confirmed using the <em>Run Test</em>, and the descriptive statistics were calculated. For each trait, the sample size (n) was calculated for the semiamplitudes of the confidence interval (i.e., estimation error) equal to 2, 4, 6, ..., 20% of the estimated mean with a confidence coefficient (1-?) of 95%. Subsequently, n was fixed at 360 plants, and the estimation error of the estimated percentage of the average for each trait was calculated. Variability of the sample size for the pigeonpea culture was observed between the morphological traits evaluated, among the evaluation periods and between seasons. Therefore, to assess with an accuracy of 6% of the estimated average, at least 136 plants must be evaluated throughout the pigeonpea crop cycle to determine the sample size for the traits (e.g., number of nodes, plant height and stem diameter) in the different evaluation periods and between seasons. </p>